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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2c2dcf09-3f4c-4e7f-b1e1-934bd3e56e9d | depth-pruning-with-auxiliary-networks-for | 2204.10546 | null | https://arxiv.org/abs/2204.10546v1 | https://arxiv.org/pdf/2204.10546v1.pdf | Depth Pruning with Auxiliary Networks for TinyML | Pruning is a neural network optimization technique that sacrifices accuracy in exchange for lower computational requirements. Pruning has been useful when working with extremely constrained environments in tinyML. Unfortunately, special hardware requirements and limited study on its effectiveness on already compact models prevent its wider adoption. Depth pruning is a form of pruning that requires no specialized hardware but suffers from a large accuracy falloff. To improve this, we propose a modification that utilizes a highly efficient auxiliary network as an effective interpreter of intermediate feature maps. Our results show a parameter reduction of 93% on the MLPerfTiny Visual Wakewords (VWW) task and 28% on the Keyword Spotting (KWS) task with accuracy cost of 0.65% and 1.06% respectively. When evaluated on a Cortex-M0 microcontroller, our proposed method reduces the VWW model size by 4.7x and latency by 1.6x while counter intuitively gaining 1% accuracy. KWS model size on Cortex-M0 was also reduced by 1.2x and latency by 1.2x at the cost of 2.21% accuracy. | ['Rowel Atienza', 'Josen Daniel De Leon'] | 2022-04-22 | null | null | null | null | ['keyword-spotting'] | ['speech'] | [ 2.26375729e-01 4.12743241e-01 -8.44896957e-02 -4.51218218e-01
-2.33786345e-01 -8.08810145e-02 2.04494208e-01 4.74736273e-01
-9.71459508e-01 6.10463202e-01 -3.78291935e-01 -7.89724290e-01
-1.71032120e-02 -6.71691716e-01 -6.98403478e-01 -2.27832943e-01
-1.59947462e-02 -1.63844615e-01 3.91204685e-01 9.29388031e-02
4.39385802e-01 4.83282745e-01 -1.68414009e+00 3.46410483e-01
4.84288633e-01 1.40006101e+00 4.47120667e-01 9.09298956e-01
-1.01959877e-01 7.72059500e-01 -1.06569767e+00 2.20787693e-02
4.74307269e-01 3.04501265e-01 -5.72276890e-01 -6.86240196e-01
6.27585113e-01 -5.19809902e-01 -1.57744616e-01 8.45925450e-01
6.37543142e-01 -1.01821370e-01 5.59864528e-02 -1.21015799e+00
1.21263094e-01 9.63819981e-01 -7.13767350e-01 4.81170177e-01
-1.63630962e-01 -4.30052020e-02 3.07963848e-01 -6.27401352e-01
-9.90014337e-03 1.00115132e+00 8.70677650e-01 5.05841374e-01
-7.83426166e-01 -1.18364608e+00 -2.24764735e-01 2.08031282e-01
-1.70613122e+00 -7.25028872e-01 2.85547465e-01 2.36687973e-01
2.07695127e+00 4.73765790e-01 6.51202738e-01 3.92066479e-01
5.33593655e-01 5.08605540e-01 1.00800681e+00 -6.73687816e-01
4.42398518e-01 1.62965983e-01 5.14610052e-01 9.44200218e-01
8.44777107e-01 -1.25044435e-01 -7.89456546e-01 -6.12988397e-02
3.98639351e-01 -8.03839341e-02 -9.38166231e-02 3.49370211e-01
-5.39742410e-01 5.33093870e-01 2.59909272e-01 5.10094985e-02
-1.96193188e-01 8.37497592e-01 6.17394328e-01 3.31631869e-01
-2.37697195e-02 4.57204729e-01 -8.26582074e-01 -6.49045825e-01
-1.21481323e+00 -1.33725554e-01 8.27043235e-01 1.08224618e+00
5.22466004e-01 6.59355998e-01 4.80456889e-01 5.90047538e-01
2.63511986e-01 4.42158669e-01 8.23766053e-01 -5.99008441e-01
5.97200215e-01 9.25142407e-01 -2.03918591e-01 -7.49126315e-01
-8.46274078e-01 -6.37364805e-01 -7.56128848e-01 5.70851684e-01
6.35973439e-02 -2.08796903e-01 -1.22416818e+00 1.59458923e+00
9.10768509e-02 -1.15340285e-01 6.38359264e-02 2.82682419e-01
6.69851601e-01 5.32260895e-01 -8.96191515e-04 9.44588110e-02
1.34439731e+00 -9.18564379e-01 -4.96256769e-01 -3.41196477e-01
7.84727395e-01 -8.22648406e-01 9.60327327e-01 7.09492624e-01
-1.20328605e+00 -4.32530314e-01 -1.95036674e+00 -8.51702020e-02
-4.67277944e-01 4.19104159e-01 7.96348572e-01 1.13827097e+00
-1.22959673e+00 6.53703511e-01 -1.10285890e+00 -2.71904409e-01
6.26042187e-01 1.24764061e+00 -4.05242108e-02 3.75545442e-01
-4.97437894e-01 1.09841669e+00 5.06805003e-01 -2.05501392e-01
-7.36850619e-01 -9.74455893e-01 -6.03406370e-01 3.56695443e-01
3.83840829e-01 -9.86813977e-02 1.22284818e+00 -4.82098728e-01
-1.55653083e+00 2.38205105e-01 1.31710684e-02 -1.12374282e+00
2.49378174e-03 -5.26318669e-01 -5.77172399e-01 1.20817490e-01
-4.60348606e-01 8.99281085e-01 6.83003902e-01 -3.16919327e-01
-8.28876376e-01 -2.11687520e-01 8.95619541e-02 -8.39996561e-02
-6.57499611e-01 -3.41360837e-01 3.24880145e-02 -4.07112151e-01
-5.51994964e-02 -8.21228564e-01 -1.27775803e-01 -7.14138150e-02
-1.31516695e-01 1.48939148e-01 1.10546720e+00 -5.41230798e-01
1.37344337e+00 -1.74252105e+00 -6.67539179e-01 1.34476736e-01
4.15616095e-01 7.76797950e-01 1.54709667e-01 -2.03241017e-02
1.95848092e-01 1.72037333e-01 -2.83187237e-02 -4.35922705e-02
-3.83089840e-01 2.21631881e-02 4.17778678e-02 4.11428362e-01
1.12568222e-01 7.34587967e-01 -2.73046970e-01 -2.55831629e-01
2.17076853e-01 5.37626266e-01 -6.65965140e-01 -3.50567877e-01
2.05744117e-01 -7.78484702e-01 5.93643729e-03 7.20617712e-01
5.33452809e-01 6.21224977e-02 8.56708437e-02 -3.23442638e-01
-5.71807683e-01 4.81511295e-01 -1.14329302e+00 1.61877322e+00
-8.54220986e-01 1.07843769e+00 1.11042164e-01 -8.13284874e-01
9.98434603e-01 1.14465542e-01 -8.05930719e-02 -9.33565259e-01
6.42350674e-01 2.50551492e-01 4.46341425e-01 -4.93648760e-02
6.41289115e-01 3.81781340e-01 4.28900048e-02 4.13470179e-01
1.20768420e-01 1.85292676e-01 -7.62221813e-02 -7.07899686e-03
1.62215805e+00 -2.08050251e-01 5.76107860e-01 -6.87779546e-01
2.64262915e-01 6.30654627e-03 4.55738097e-01 7.77485013e-01
-7.58236572e-02 -2.91713942e-02 1.47415325e-01 -6.60772145e-01
-8.73040259e-01 -4.59347039e-01 -1.48592547e-01 7.67397940e-01
-4.20462415e-02 -9.29143250e-01 -7.43951321e-01 -2.85923272e-01
-2.45884418e-01 9.18776155e-01 -3.06642115e-01 -3.67160410e-01
-6.93720162e-01 -7.70459533e-01 1.09284496e+00 6.68027222e-01
8.74467015e-01 -6.51806176e-01 -2.07380962e+00 2.95707136e-01
8.03762019e-01 -9.26463127e-01 6.62035868e-02 9.63496864e-01
-1.43550503e+00 -8.49949956e-01 -4.21371795e-02 -5.54406464e-01
6.27581716e-01 1.70803562e-01 9.86797452e-01 3.73755664e-01
-7.77124226e-01 -1.84313387e-01 -1.71469182e-01 -9.84027863e-01
-1.43679515e-01 3.04892540e-01 1.66078493e-01 -8.97449553e-01
5.71984112e-01 -7.25095749e-01 -5.97616136e-01 -9.58013013e-02
-7.27024078e-01 2.62712896e-01 1.00724757e+00 7.50469506e-01
2.91918725e-01 1.20493054e-01 4.19571936e-01 -6.46807134e-01
4.93288279e-01 -1.73488662e-01 -8.74860942e-01 6.80065900e-02
-1.12963176e+00 4.12983686e-01 6.55885279e-01 -7.10532725e-01
-6.84757888e-01 2.71741509e-01 -3.14010791e-02 -1.03056923e-01
2.93505937e-01 1.79163709e-01 1.48334503e-01 -4.90191221e-01
7.46754766e-01 -8.91604647e-02 -5.73347174e-02 -5.16369283e-01
-9.54233557e-02 7.83920944e-01 3.07213843e-01 -1.15421183e-01
2.46929139e-01 2.33589202e-01 1.91798761e-01 -1.05705607e+00
-8.52355957e-02 -5.40340133e-02 -1.73187122e-01 2.09056586e-01
3.60392869e-01 -8.78642619e-01 -1.00142205e+00 1.04956031e-01
-9.88831401e-01 -4.57832128e-01 -9.42416564e-02 3.03767890e-01
8.51761848e-02 3.44912186e-02 -3.78732175e-01 -8.03714693e-01
-1.19822490e+00 -1.10670006e+00 5.51305652e-01 3.70305300e-01
-6.70659363e-01 -2.42731467e-01 -6.62357688e-01 -1.90845892e-01
8.99303615e-01 -1.67109892e-01 1.08388269e+00 -5.49712360e-01
-3.98469925e-01 -3.46142143e-01 -2.52303898e-01 2.68171221e-01
-1.22975923e-01 -2.15140417e-01 -9.25021768e-01 -1.38229206e-01
1.91836163e-01 -1.69345751e-01 6.83446586e-01 2.71930546e-01
1.07434142e+00 -2.08611012e-01 -4.16869819e-01 6.15420461e-01
1.80145156e+00 6.92417264e-01 2.66621500e-01 4.01883930e-01
4.11623657e-01 -6.78371936e-02 4.18240011e-01 5.76466858e-01
-1.32954121e-01 4.72924143e-01 7.11191773e-01 2.40766734e-01
-4.18467581e-01 7.89267495e-02 3.18087131e-01 9.54957485e-01
3.08873385e-01 1.27865802e-02 -1.12786412e+00 5.27819097e-01
-1.34373975e+00 -3.34801704e-01 2.53082030e-02 2.12947536e+00
9.01151955e-01 6.59744561e-01 -3.63095969e-01 6.99724138e-01
2.84106463e-01 -1.46998540e-01 -6.43357456e-01 -1.16620445e+00
3.10102254e-01 1.04009223e+00 1.05125511e+00 5.49301267e-01
-5.76203763e-01 8.24453354e-01 5.26452112e+00 8.84327352e-01
-1.40340555e+00 2.10545540e-01 4.34917599e-01 -9.38815117e-01
5.28018832e-01 -9.69800800e-02 -1.31036675e+00 4.01967257e-01
1.56316626e+00 -8.82034749e-02 2.65636533e-01 1.21021676e+00
-1.41493395e-01 -6.75330698e-01 -1.10382116e+00 1.29677558e+00
4.53001782e-02 -1.50047898e+00 -1.39616132e-01 -6.69201761e-02
2.62780339e-01 1.90322563e-01 -1.13081530e-01 3.77489299e-01
-1.16022229e-01 -1.22144949e+00 8.56711745e-01 -2.09453739e-02
8.02516997e-01 -1.22307813e+00 8.83998811e-01 3.66062433e-01
-1.13957810e+00 -2.07544222e-01 -5.20492256e-01 -5.93912959e-01
-1.77337542e-01 4.49654102e-01 -1.22249854e+00 -2.65418530e-01
1.07530379e+00 -1.94091216e-01 -7.41750658e-01 9.24958646e-01
1.23766847e-01 5.03780663e-01 -1.00569177e+00 -5.75278759e-01
2.39125922e-01 4.62728947e-01 1.64852485e-01 1.34022486e+00
3.83242548e-01 8.51909667e-02 -5.33267498e-01 2.71007478e-01
-1.53856009e-01 -1.24462180e-01 -2.91318744e-01 1.07638665e-01
8.40136349e-01 1.29424369e+00 -9.31758463e-01 -4.18209553e-01
-1.18377171e-01 7.47479558e-01 3.01693916e-01 -2.51417309e-01
-9.21159625e-01 -1.07365584e+00 5.81576228e-01 1.62140235e-01
3.68615478e-01 -1.59982085e-01 -6.32834256e-01 -4.22164112e-01
6.87361136e-02 -7.77149498e-01 -3.07118684e-01 -4.59514618e-01
1.24362461e-01 8.60986829e-01 3.17515507e-02 -7.00733066e-01
-1.31464466e-01 -5.92313111e-01 -3.00654233e-01 7.88968384e-01
-1.06984377e+00 -6.23690248e-01 -5.13600171e-01 9.05540809e-02
7.09918559e-01 -4.06684339e-01 9.31721687e-01 3.96307766e-01
-6.30900323e-01 9.69897389e-01 -2.21908495e-01 -4.12294388e-01
8.33720118e-02 -8.66096616e-01 5.64973295e-01 7.73351312e-01
-4.72984128e-02 1.08151340e+00 6.04193032e-01 -5.25462031e-01
-1.65279174e+00 -1.04831505e+00 7.34443545e-01 1.07610896e-01
2.02247575e-01 -5.56827128e-01 -5.49167395e-01 3.24090004e-01
3.88383478e-01 -1.20311752e-01 5.43622196e-01 -1.97285965e-01
-9.48504880e-02 -3.69196892e-01 -1.46707344e+00 8.25086057e-01
8.27387810e-01 -1.49129078e-01 -2.81994820e-01 -7.90429041e-02
7.33519018e-01 -4.64974493e-01 -6.07177615e-01 2.69851416e-01
7.48081744e-01 -9.13572133e-01 6.61403537e-01 -7.04958662e-02
6.04172386e-02 -1.57681674e-01 -4.00477082e-01 -6.50893092e-01
1.32014081e-01 -8.11926544e-01 -5.94152868e-01 7.91096985e-01
5.33661962e-01 -4.64958042e-01 1.09170246e+00 4.72927362e-01
-2.33920261e-01 -1.18132401e+00 -1.01844418e+00 -7.50267863e-01
-4.24403131e-01 -9.58499908e-01 6.18965685e-01 3.62263680e-01
4.62401509e-02 6.35061264e-01 -9.11704674e-02 2.18734462e-02
3.60483974e-01 -5.61712742e-01 4.72296804e-01 -9.62585449e-01
-1.59500673e-01 -4.26914960e-01 -6.05732620e-01 -9.31477606e-01
-3.77504170e-01 -5.65043986e-01 -1.65086880e-01 -9.48651314e-01
-1.37235507e-01 -4.09490854e-01 -3.20830017e-01 8.10957074e-01
4.45103437e-01 2.91586339e-01 1.94635242e-01 -3.49071294e-01
-3.16233844e-01 1.09820805e-01 1.69437736e-01 -9.14423466e-02
-3.15063804e-01 -2.44017974e-01 -5.86550653e-01 1.06466269e+00
1.25729001e+00 -7.36772239e-01 -9.26339567e-01 -7.68941700e-01
4.48776811e-01 -3.06661338e-01 -1.32167265e-01 -1.59043717e+00
5.24822295e-01 4.86323565e-01 5.87287486e-01 -9.43536341e-01
6.82434082e-01 -8.80744874e-01 2.86955871e-02 1.13538849e+00
-2.04649032e-03 5.72613060e-01 1.04516554e+00 2.94710487e-01
2.50495344e-01 -7.17064023e-01 7.16834366e-01 1.82792079e-02
-8.08097005e-01 -3.22097719e-01 -4.30767953e-01 -3.98942232e-01
1.02974010e+00 -7.05283165e-01 -2.62509733e-01 -1.80415343e-02
-1.85217455e-01 -7.00627938e-02 9.54258293e-02 1.88214868e-01
8.41997683e-01 -8.17850888e-01 -3.88164334e-02 3.83013338e-01
-4.38158184e-01 -1.26124591e-01 -1.19687833e-01 5.84488451e-01
-1.07151675e+00 1.00424933e+00 -4.10838813e-01 -3.08654726e-01
-1.85024548e+00 1.06204964e-01 1.08774625e-01 -9.23134163e-02
-6.87550843e-01 1.21831131e+00 -4.60043103e-01 2.47415379e-01
5.39028466e-01 -8.80330265e-01 -1.05708139e-02 -3.04989815e-01
8.63153338e-01 9.09161389e-01 5.86546481e-01 9.81650800e-02
-6.60017431e-01 2.58432090e-01 -3.75813961e-01 -9.79351625e-02
1.41251338e+00 3.11227798e-01 -2.00175308e-02 1.66681647e-01
1.24616110e+00 -8.80329497e-03 -8.82890046e-01 1.40172735e-01
3.33283812e-01 -3.52338143e-02 5.90192914e-01 -9.72318351e-01
-9.91769612e-01 7.92332649e-01 1.03436506e+00 -1.58772632e-01
1.43274426e+00 -5.54190934e-01 8.10630679e-01 8.57104182e-01
4.88565117e-01 -1.21763778e+00 -3.14736694e-01 4.21224743e-01
6.09767497e-01 -8.43214810e-01 5.92703938e-01 -2.69347012e-01
-2.78551001e-02 1.19426715e+00 8.02909434e-01 -2.07443327e-01
6.91344917e-01 1.16190398e+00 -3.09330791e-01 -1.91380903e-01
-1.06563723e+00 4.09864157e-01 -2.21956983e-01 5.22227883e-01
3.15916806e-01 5.51550090e-02 -4.06707823e-01 5.40692210e-01
-5.93671679e-01 5.89314327e-02 5.78349411e-01 1.40352368e+00
-6.55782402e-01 -8.95224154e-01 -4.09304984e-02 8.64335120e-01
-7.48014331e-01 -5.86085677e-01 -1.01742692e-01 1.15915382e+00
2.67510831e-01 9.47256684e-01 2.59452879e-01 -8.16639781e-01
7.87813887e-02 3.24059762e-02 5.48431993e-01 -6.22783363e-01
-1.08076346e+00 -2.32831880e-01 2.15178967e-01 -7.72885442e-01
6.49311319e-02 1.15411401e-01 -1.37859809e+00 -4.47625428e-01
-4.87982333e-01 -1.07140400e-01 1.28317857e+00 5.77392876e-01
6.58845365e-01 8.43180954e-01 -3.58728975e-01 -5.05984187e-01
-5.74170887e-01 -9.44030225e-01 -1.75282106e-01 -6.95042193e-01
2.01990843e-01 -4.84706670e-01 -1.69106461e-02 -3.02142531e-01] | [8.377518653869629, 2.7998499870300293] |
7cc3fe9b-ee7f-48f3-bb31-2066f8a4ec26 | a-new-method-using-deep-learning-to-predict | 2305.02475 | null | https://arxiv.org/abs/2305.02475v1 | https://arxiv.org/pdf/2305.02475v1.pdf | A new method using deep learning to predict the response to cardiac resynchronization therapy | Background. Clinical parameters measured from gated single-photon emission computed tomography myocardial perfusion imaging (SPECT MPI) have value in predicting cardiac resynchronization therapy (CRT) patient outcomes, but still show limitations. The purpose of this study is to combine clinical variables, features from electrocardiogram (ECG), and parameters from assessment of cardiac function with polarmaps from gated SPECT MPI through deep learning (DL) to predict CRT response. Methods. 218 patients who underwent rest gated SPECT MPI were enrolled in this study. CRT response was defined as an increase in left ventricular ejection fraction (LVEF) > 5% at a 6-month follow up. A DL model was constructed by combining a pre-trained VGG16 module and a multilayer perceptron. Two modalities of data were input to the model: polarmap images from SPECT MPI and tabular data from clinical features and ECG parameters. Gradient-weighted Class Activation Mapping (Grad-CAM) was applied to the VGG16 module to provide explainability for the polarmaps. For comparison, four machine learning (ML) models were trained using only the tabular features. Results. Modeling was performed on 218 patients who underwent CRT implantation with a response rate of 55.5% (n = 121). The DL model demonstrated average AUC (0.83), accuracy (0.73), sensitivity (0.76), and specificity (0.69) surpassing the ML models and guideline criteria. Guideline recommendations presented accuracy (0.53), sensitivity (0.75), and specificity (0.26). Conclusions. The DL model outperformed the ML models, showcasing the additional predictive benefit of utilizing SPECT MPI polarmaps. Incorporating additional patient data directly in the form of medical imagery can improve CRT response prediction. | ['Weihua Zhou', 'Amalia Peix', 'Jiangang Zou', 'Ernest V. Garciaf', 'Diana Paeze', 'Claudio T Mesquitad', 'Quiying Sha', 'Xinwei Zhang', 'Chen Zhao', 'Zhuo He', 'Kristoffer Larsena'] | 2023-05-04 | null | null | null | null | ['specificity'] | ['natural-language-processing'] | [ 9.65372100e-02 -1.37193024e-01 -5.30728638e-01 -4.09297824e-01
-1.02174091e+00 -7.11611390e-01 -4.57717180e-02 2.64281929e-01
-3.22853655e-01 9.85898018e-01 4.16460812e-01 -7.12224126e-01
-2.90201008e-01 -5.63061476e-01 -1.40941799e-01 -6.27698183e-01
-4.12756026e-01 1.00052595e+00 -2.86019474e-01 4.59774643e-01
-2.66687162e-02 5.95923007e-01 -4.35603201e-01 6.84182703e-01
5.48026681e-01 9.95971620e-01 3.52981776e-01 1.02648401e+00
4.67224360e-01 9.91539061e-01 -1.02490835e-01 3.95431608e-01
1.35583177e-01 -9.24052894e-01 -7.31191874e-01 -1.32175103e-01
1.68993980e-01 -5.36107302e-01 -2.41703480e-01 -1.47861661e-02
1.22283554e+00 -5.05262792e-01 5.64949512e-01 -5.51798284e-01
-1.76113650e-01 6.34807706e-01 -1.86278924e-01 7.86471009e-01
-1.24453314e-01 4.42509085e-01 5.51319242e-01 -1.04552817e+00
5.42339385e-01 3.82986635e-01 1.15869212e+00 3.45799208e-01
-1.46315360e+00 -3.19857448e-01 -5.01850426e-01 -8.15217048e-02
-1.14024520e+00 3.20845060e-02 3.70998204e-01 -8.88900101e-01
1.05091929e+00 5.39414883e-01 1.07908881e+00 1.91179127e-01
5.47631502e-01 -5.87769076e-02 1.12484419e+00 -2.84138113e-01
-8.19236860e-02 -7.91819319e-02 5.62430359e-02 5.60806096e-01
4.93641905e-02 5.18999577e-01 -5.67283258e-02 -5.46921253e-01
1.08371806e+00 -1.87105145e-02 -5.61511934e-01 -2.93386906e-01
-1.66220558e+00 8.06767821e-01 6.14076495e-01 2.55145639e-01
-6.29559278e-01 2.17360824e-01 7.13284612e-01 2.23475188e-01
2.87875712e-01 4.67005342e-01 -7.88501263e-01 -1.15654141e-01
-9.19762671e-01 -9.75786969e-02 3.41094106e-01 8.84487629e-02
1.33563787e-01 1.65729940e-01 -5.30817330e-01 8.55062783e-01
4.06940132e-01 7.51898825e-01 5.83629727e-01 -9.55902576e-01
4.27496433e-01 5.75993478e-01 -8.00068397e-03 -4.90277380e-01
-1.07331443e+00 -1.11549914e+00 -1.15856862e+00 1.90966874e-01
1.93854675e-01 -5.37227333e-01 -9.12846148e-01 1.18738580e+00
-3.61584932e-01 -1.77445248e-01 -1.50955141e-01 1.37661994e+00
8.30518305e-01 2.60620713e-01 6.36802197e-01 -4.76436287e-01
1.26446462e+00 -4.71749365e-01 -1.25419959e-01 -1.04270704e-01
1.43099689e+00 -3.70165110e-01 8.91921520e-01 6.37576953e-02
-1.08669579e+00 -5.66567838e-01 -8.27673078e-01 4.68653798e-01
4.59065378e-01 6.91306829e-01 3.97620052e-01 9.94096041e-01
-1.03062737e+00 8.47178876e-01 -1.04008126e+00 -3.22637677e-01
7.84103394e-01 7.29779780e-01 -4.25331354e-01 9.71084833e-02
-1.04670894e+00 9.74334657e-01 1.49313673e-01 1.42337695e-01
-6.26797378e-01 -1.07231390e+00 -4.66165066e-01 2.07346249e-02
-3.24139267e-01 -1.60168946e+00 7.37559736e-01 -1.05228686e+00
-1.27753425e+00 1.01259089e+00 -4.30774949e-02 -3.06419343e-01
5.37298024e-01 9.94274244e-02 3.27689108e-03 4.58457023e-01
1.53256759e-01 5.00384569e-01 1.64117858e-01 -8.01801085e-01
-2.90006071e-01 -6.96804762e-01 -4.39527899e-01 4.31218147e-01
1.50946558e-01 2.09524259e-01 3.05457115e-01 -4.44344312e-01
5.81573427e-01 -1.13843668e+00 -4.52629179e-01 -2.92286098e-01
-1.23512961e-01 3.96489233e-01 3.59649986e-01 -1.03250170e+00
1.14648914e+00 -1.72126281e+00 -6.89622536e-02 2.43622288e-01
6.53326333e-01 2.45542273e-01 2.31880963e-01 6.07002489e-02
-5.76770127e-01 7.76458085e-01 -2.90267974e-01 2.56140739e-01
-7.29069948e-01 -1.12369999e-01 7.41215721e-02 4.68564242e-01
3.60475061e-03 1.61098206e+00 -4.84151572e-01 -2.02093884e-01
5.40550351e-01 3.53630304e-01 -5.18210232e-01 -1.35339364e-01
3.86725366e-01 1.24569833e+00 -2.08571419e-01 4.45771635e-01
3.20163757e-01 -7.13866174e-01 7.55598962e-01 -1.15746990e-01
1.19170032e-01 2.08800379e-02 -3.97570670e-01 1.34555256e+00
-1.81774139e-01 3.38492125e-01 -4.85730022e-01 -6.93342030e-01
9.84403193e-01 7.41593897e-01 9.83339727e-01 -5.70157230e-01
1.17735520e-01 4.03561592e-01 6.14694059e-01 -3.77415657e-01
-5.56963682e-01 -8.16431224e-01 2.79950857e-01 4.93760407e-01
-2.36021608e-01 -3.87807563e-02 -4.80973512e-01 -1.92127332e-01
1.06551909e+00 6.30001426e-02 2.59048879e-01 -3.95066559e-01
5.41972578e-01 3.08383495e-01 6.71350420e-01 1.00610113e+00
-2.29478478e-01 1.15699446e+00 7.65099347e-01 -1.04116213e+00
-1.20178235e+00 -1.05110049e+00 -5.45421243e-01 2.43092924e-01
-4.56127346e-01 -2.85932690e-01 -8.72821361e-02 -6.31616414e-01
8.29283614e-03 -9.67484899e-03 -3.96464080e-01 8.52381811e-02
-9.14881766e-01 -1.37177253e+00 5.81203818e-01 1.20748925e+00
2.43390843e-01 -9.28095579e-01 -8.69107902e-01 5.81808567e-01
-3.51119310e-01 -6.60697460e-01 4.55930308e-02 1.28870264e-01
-1.88486123e+00 -1.01901650e+00 -1.23570371e+00 -6.23059392e-01
3.44937056e-01 -3.46182972e-01 1.33957362e+00 2.28849515e-01
-2.37473145e-01 9.11116749e-02 -3.85014080e-02 -1.01308718e-01
-3.25263470e-01 1.51550338e-01 -1.43275231e-01 -4.22927409e-01
-2.54678577e-01 -6.24340653e-01 -1.08506787e+00 3.90033424e-01
-1.39976248e-01 2.45969042e-01 7.72776484e-01 1.20442557e+00
9.79555249e-01 -8.07371378e-01 9.40876544e-01 -8.92374039e-01
1.52583510e-01 -3.01138937e-01 -1.01035737e-01 -1.12318769e-01
-1.06856191e+00 -8.42831016e-01 2.43297324e-01 -3.86570320e-02
-4.95234519e-01 1.44020095e-01 9.08479244e-02 -2.96665221e-01
-9.00545195e-02 9.71187234e-01 3.03137839e-01 6.62797913e-02
8.84839594e-01 -7.04862326e-02 1.22131094e-01 -1.96686372e-01
-2.31468081e-01 3.47431391e-01 1.73662409e-01 -2.84017026e-01
1.99917212e-01 2.84586489e-01 5.15701771e-01 -3.58295321e-01
-3.20775956e-01 -2.56925970e-01 -7.86399007e-01 -1.54771507e-01
8.07774484e-01 -1.00213575e+00 -5.72469413e-01 1.75229132e-01
-7.34654665e-01 -6.35104239e-01 -5.29069424e-01 1.15593731e+00
-6.48095965e-01 2.62309134e-01 -7.15557039e-01 -3.73310626e-01
-8.92105699e-01 -9.27596569e-01 6.02710724e-01 -1.93887770e-01
-5.18937230e-01 -1.14256322e+00 2.45851725e-01 3.91378641e-01
6.75969303e-01 6.88108325e-01 1.39255834e+00 -5.15540659e-01
-3.85388553e-01 -3.33246201e-01 -3.91565859e-01 3.81932467e-01
6.79556429e-02 -5.53446829e-01 -7.41284549e-01 -1.92348137e-01
-1.27929682e-02 -3.63454781e-02 6.72178030e-01 1.23024559e+00
1.05904973e+00 3.11094970e-01 -1.58874407e-01 7.55041420e-01
1.39546359e+00 5.14646769e-01 7.43787169e-01 2.44574621e-01
7.64682293e-01 9.67831444e-03 1.07039530e-02 3.13505858e-01
3.25119972e-01 6.40682578e-01 1.32136643e-01 -4.55659896e-01
-3.55695188e-01 5.34786172e-02 -9.05906782e-02 4.41681594e-01
-5.77631772e-01 1.22755997e-01 -1.62701654e+00 2.69942135e-01
-1.73163223e+00 -6.21200323e-01 -8.56456876e-01 2.05202246e+00
1.56449512e-01 1.41937971e-01 3.82336453e-02 -9.26608220e-02
6.62460327e-01 -2.61224389e-01 -5.71507931e-01 -3.82962406e-01
-3.43026429e-01 3.82408828e-01 3.88524890e-01 4.13849384e-01
-9.57658231e-01 2.32509151e-01 6.57757902e+00 -2.06209242e-01
-1.39011395e+00 3.22855532e-01 1.38102734e+00 6.57897145e-02
-2.25540195e-02 3.53582948e-01 5.65049499e-02 1.32176682e-01
1.02900445e+00 -2.49091927e-02 -1.07094795e-02 2.55931884e-01
6.54825866e-01 -1.80071592e-01 -8.63857567e-01 9.36481118e-01
-5.82838282e-02 -1.60816205e+00 -3.49027961e-01 5.60417324e-02
5.66341281e-01 4.70579296e-01 5.31820469e-02 2.54570961e-01
-6.13498330e-01 -1.28047013e+00 1.87772438e-01 7.31075048e-01
1.50942385e+00 -1.53279543e-01 1.40340662e+00 2.29819909e-01
-8.61535192e-01 -2.68114239e-01 1.21195488e-01 -4.73812908e-01
2.66033620e-01 6.33319795e-01 -1.29954493e+00 7.10583508e-01
4.23437387e-01 8.08772504e-01 -3.38823050e-01 9.71832454e-01
-1.01979105e-02 1.04767871e+00 -1.07994601e-02 5.69970012e-01
-2.15776220e-01 1.77165538e-01 5.31378686e-01 8.58655214e-01
2.04003602e-01 4.15929914e-01 2.60203630e-01 8.92349303e-01
1.33641005e-01 3.91536534e-01 -2.94582754e-01 2.05623478e-01
-5.89514822e-02 1.14899993e+00 -8.60063314e-01 -5.44197679e-01
-4.38364983e-01 4.49235320e-01 -3.67086142e-01 3.49230975e-01
-6.17123306e-01 2.00933442e-01 -2.69528657e-01 7.88389921e-01
-3.57020944e-01 1.15804926e-01 -1.30719030e+00 -8.68197262e-01
-2.08987355e-01 -6.77979290e-01 5.76591253e-01 -1.05881798e+00
-1.15269792e+00 7.60731697e-01 -3.30582052e-01 -1.27043617e+00
-2.25007400e-01 -3.92247885e-01 -5.27251959e-01 1.85266650e+00
-1.16347075e+00 -1.14362550e+00 -3.38348776e-01 4.35728990e-02
1.68363795e-01 -2.42138892e-01 1.34276760e+00 1.71321362e-01
-2.45533347e-01 7.85571411e-02 -2.88792491e-01 3.05666119e-01
6.80081189e-01 -1.29690945e+00 -1.16402633e-01 1.59429714e-01
-5.61740041e-01 4.43035007e-01 -1.61572754e-01 -9.63757336e-01
-8.45023453e-01 -1.10397840e+00 1.26397073e+00 -6.80336475e-01
3.66243608e-02 6.91803753e-01 -5.11388659e-01 6.85079038e-01
-2.64756829e-01 4.73315746e-01 9.76604223e-01 4.03011814e-02
4.12290126e-01 6.62469342e-02 -9.77668822e-01 -2.81686783e-02
6.03341460e-01 -5.12178600e-01 -2.94541538e-01 1.51403159e-01
-2.68459588e-01 -6.44071341e-01 -1.44352639e+00 1.14355028e+00
9.11637008e-01 -6.76181853e-01 9.01479185e-01 -7.97586024e-01
4.44037288e-01 -2.02666566e-01 8.03783443e-03 -9.39516246e-01
-8.11471105e-01 -1.92942888e-01 2.35420883e-01 3.22866052e-01
7.02409685e-01 -7.41850793e-01 8.89489889e-01 8.37049365e-01
-4.24601465e-01 -1.05554640e+00 -6.81433260e-01 -1.94771394e-01
4.47617412e-01 -1.43143743e-01 5.95528148e-02 9.53984022e-01
9.19748545e-02 3.72298956e-01 -6.90613538e-02 2.58366633e-02
1.17841586e-01 1.41298875e-01 1.73652843e-01 -1.44592535e+00
-4.13684428e-01 -1.90395400e-01 -4.93827313e-01 -2.79705465e-01
-2.36550719e-01 -1.62394106e+00 -7.21122503e-01 -2.00944257e+00
8.71965349e-01 -9.61129844e-01 -7.59067237e-01 5.79958677e-01
-2.68960297e-01 6.12807155e-01 3.27101141e-01 6.36030734e-01
2.49016941e-01 1.23103820e-02 1.28451538e+00 1.83144554e-01
-5.57579935e-01 1.67950973e-01 -6.63588822e-01 6.87520862e-01
1.26521957e+00 -7.11180151e-01 -3.76833141e-01 -3.16974640e-01
2.93999255e-01 9.94829178e-01 9.53450620e-01 -1.10785341e+00
-3.26324016e-01 3.12303066e-01 1.18793952e+00 -4.07934725e-01
1.02209345e-01 -4.60726231e-01 4.91195917e-01 1.07065320e+00
-7.27657676e-02 4.04167354e-01 2.87491888e-01 4.41649817e-02
-4.10874896e-02 2.62452483e-01 6.63773596e-01 -3.83324623e-01
-1.01138525e-01 4.12820995e-01 -5.47614038e-01 5.85949495e-02
6.17286623e-01 -3.04099143e-01 -2.01579988e-01 -2.70086050e-01
-1.64872217e+00 -1.98532656e-01 4.65875538e-03 -9.04933736e-02
6.52681231e-01 -1.21539235e+00 -1.24250829e+00 7.25112557e-02
-1.64742887e-01 -3.87464166e-01 5.53730786e-01 1.84750378e+00
-9.51138198e-01 6.45145893e-01 -4.24160540e-01 -1.22303939e+00
-9.88911033e-01 8.59680250e-02 1.08168352e+00 -5.22498429e-01
-1.10526145e+00 3.40508938e-01 2.63927460e-01 -2.64917314e-01
-4.79862005e-01 -2.28363335e-01 -9.64542851e-02 -2.29791433e-01
1.80829287e-01 3.55686724e-01 3.52036536e-01 -5.09701252e-01
-5.33858776e-01 7.28433907e-01 4.67039794e-01 -1.57785460e-01
1.10960758e+00 -1.12204947e-01 -6.62974268e-02 4.36907083e-01
8.29837859e-01 -2.97914952e-01 -7.36816764e-01 1.04045093e-01
-2.44359136e-01 -2.50457048e-01 3.84082586e-01 -1.67301309e+00
-9.94154394e-01 9.16557729e-01 1.46767747e+00 -2.88221598e-01
1.13576889e+00 -1.19640373e-01 4.13835794e-01 -1.91147715e-01
3.52037959e-02 -3.51409167e-01 -2.96371639e-01 1.12903908e-01
7.78665066e-01 -1.04218698e+00 -8.94980412e-03 -3.13629717e-01
-9.81047571e-01 1.28877580e+00 1.65561900e-01 -4.36922628e-03
7.74410486e-01 2.33174264e-02 4.47917640e-01 -1.35361701e-01
-5.10970652e-01 3.79223406e-01 1.53455943e-01 3.90102834e-01
8.98445725e-01 4.50147599e-01 -6.79100692e-01 9.59086835e-01
3.70912820e-01 4.82442409e-01 2.95454621e-01 7.13022768e-01
-3.30444694e-01 -1.21180904e+00 -1.25878811e-01 8.59292686e-01
-6.67163074e-01 -2.24460512e-01 -1.44532219e-01 5.25790036e-01
2.59287089e-01 6.30938709e-01 -2.66259283e-01 -1.00217693e-01
2.90393978e-01 5.46739042e-01 3.87560070e-01 -7.23942876e-01
-1.02939224e+00 3.18153769e-01 1.51587322e-01 -1.76628530e-01
-3.48829985e-01 -6.74441636e-01 -1.20239329e+00 1.58181250e-01
-5.29095531e-02 -1.29539430e-01 4.17327583e-01 6.70430958e-01
5.02306759e-01 9.81202602e-01 5.80242515e-01 -5.84708393e-01
-1.79449305e-01 -1.07639039e+00 -6.15750551e-01 -1.67696401e-01
1.87545314e-01 -2.11002812e-01 -9.58964378e-02 2.44868100e-01] | [14.397130966186523, 3.3160696029663086] |
51d7b11e-32c0-4280-8d60-736abded083b | u-net-convolutional-networks-for-biomedical | 1505.04597 | null | http://arxiv.org/abs/1505.04597v1 | http://arxiv.org/pdf/1505.04597v1.pdf | U-Net: Convolutional Networks for Biomedical Image Segmentation | There is large consent that successful training of deep networks requires
many thousand annotated training samples. In this paper, we present a network
and training strategy that relies on the strong use of data augmentation to use
the available annotated samples more efficiently. The architecture consists of
a contracting path to capture context and a symmetric expanding path that
enables precise localization. We show that such a network can be trained
end-to-end from very few images and outperforms the prior best method (a
sliding-window convolutional network) on the ISBI challenge for segmentation of
neuronal structures in electron microscopic stacks. Using the same network
trained on transmitted light microscopy images (phase contrast and DIC) we won
the ISBI cell tracking challenge 2015 in these categories by a large margin.
Moreover, the network is fast. Segmentation of a 512x512 image takes less than
a second on a recent GPU. The full implementation (based on Caffe) and the
trained networks are available at
http://lmb.informatik.uni-freiburg.de/people/ronneber/u-net . | ['Thomas Brox', 'Philipp Fischer', 'Olaf Ronneberger'] | 2015-05-18 | null | null | null | null | ['thermal-image-segmentation', 'dichotomous-image-segmentation', 'video-polyp-segmentation', 'skin-cancer-segmentation', 'multi-tissue-nucleus-segmentation', 'pancreas-segmentation'] | ['computer-vision', 'computer-vision', 'computer-vision', 'medical', 'medical', 'medical'] | [ 3.04212719e-01 3.01219225e-02 4.02780205e-01 -3.77031565e-01
-6.36735022e-01 -5.29128492e-01 2.17638880e-01 -4.51594740e-02
-1.27730668e+00 1.03058481e+00 -6.06918454e-01 -2.92492837e-01
2.07149982e-01 -2.87554651e-01 -9.01741028e-01 -8.30298185e-01
-1.37677118e-01 7.58579910e-01 6.19913220e-01 1.57631487e-01
3.16270053e-01 6.16021216e-01 -1.07982194e+00 2.53268123e-01
2.12334216e-01 1.01118708e+00 7.78050900e-01 1.03377473e+00
1.37377962e-01 4.68693525e-01 -4.77422476e-01 -6.33846745e-02
2.90805340e-01 -1.97033808e-01 -1.13166893e+00 -2.15230659e-01
6.24151468e-01 -3.62456918e-01 -9.62845832e-02 8.26402426e-01
7.12132871e-01 1.72860119e-02 1.93847477e-01 -9.30241644e-01
-1.55356616e-01 3.29138428e-01 -2.60371834e-01 7.30669379e-01
-4.84923750e-01 1.84456572e-01 5.12122631e-01 -7.73327947e-01
1.08409560e+00 8.22463930e-01 9.02707756e-01 8.25974166e-01
-1.29823363e+00 -6.21792972e-01 -1.24681063e-01 8.27130228e-02
-1.05509830e+00 -4.16165173e-01 2.97787726e-01 -4.47503626e-01
1.15325463e+00 -7.80381709e-02 9.78346109e-01 7.82713056e-01
1.38496146e-01 5.87597013e-01 1.17698383e+00 -3.68677348e-01
2.25310072e-01 -3.05690914e-01 2.97392488e-01 8.37840796e-01
1.68476403e-01 -6.53317804e-03 -8.03833157e-02 -2.80770827e-02
1.14185560e+00 1.01812989e-01 -3.84432226e-01 -1.40712276e-01
-1.48775697e+00 5.15305996e-01 5.65802217e-01 6.05438650e-01
-2.87041925e-02 5.63047409e-01 6.08821750e-01 4.63835858e-02
3.26938897e-01 4.10288066e-01 -8.46827269e-01 -1.08052529e-01
-1.04816639e+00 2.71937609e-01 6.64296865e-01 6.12100661e-01
8.77685070e-01 -2.44842634e-01 2.92124540e-01 5.16707242e-01
7.62457177e-02 1.08834267e-01 4.52822447e-01 -1.46877968e+00
-1.17188282e-01 2.78343648e-01 -4.10807654e-02 -3.66094321e-01
-8.29914391e-01 -2.91524887e-01 -8.84768903e-01 7.08523989e-01
9.10289407e-01 -4.67517167e-01 -1.04574513e+00 1.53961670e+00
4.34542537e-01 3.07273269e-01 -4.32495385e-01 9.69528198e-01
5.58557034e-01 2.99203098e-01 -1.66395798e-01 1.52422115e-01
1.29306090e+00 -1.01748097e+00 -4.81074929e-01 -1.54221788e-01
8.21306109e-01 -6.30048990e-01 7.21930623e-01 4.78643835e-01
-1.32885265e+00 -3.64396840e-01 -1.05713272e+00 -4.48671132e-01
-6.50218964e-01 2.31595010e-01 6.87067568e-01 3.14176559e-01
-1.59937036e+00 1.04931426e+00 -1.39405310e+00 -5.38611352e-01
1.01675618e+00 9.56405580e-01 -6.43787146e-01 8.64730701e-02
-5.09560108e-01 6.88903511e-01 5.83958387e-01 2.64627896e-02
-7.93500066e-01 -8.73296738e-01 -3.48112673e-01 -2.50620097e-01
-1.15814582e-01 -6.49657488e-01 1.08063555e+00 -9.09974039e-01
-1.48744226e+00 1.33950150e+00 4.80683185e-02 -9.01988924e-01
5.84037066e-01 -1.93060674e-02 3.52326006e-01 5.24726808e-01
-1.03648365e-01 1.29654014e+00 3.20035189e-01 -9.46301639e-01
-6.37041569e-01 -5.91458917e-01 6.84980722e-03 -3.44061792e-01
8.10719002e-03 2.59409457e-01 -4.36307669e-01 -3.71092737e-01
-2.37484649e-01 -1.07259917e+00 -3.28010947e-01 4.08776611e-01
-2.66767591e-01 6.15474917e-02 9.91975546e-01 -6.57374799e-01
3.92250866e-01 -1.81362402e+00 -9.46484655e-02 -1.08168371e-01
4.09413278e-01 5.05957782e-01 -4.63185795e-02 6.68442696e-02
-1.45439282e-01 -5.51624335e-02 -5.28882444e-01 -6.72183335e-01
-2.80867517e-01 2.21644655e-01 3.34303565e-02 9.46276844e-01
1.04195010e-02 9.57125485e-01 -7.20422447e-01 -4.68764275e-01
1.23527288e-01 7.97781050e-01 -6.15195513e-01 -3.28472070e-02
-3.45380977e-02 7.58176863e-01 7.47670829e-02 4.40783381e-01
7.20159173e-01 -6.31334722e-01 1.71089396e-01 -1.51504830e-01
-2.83744007e-01 3.23391706e-02 -6.62123621e-01 1.89168859e+00
-2.14542255e-01 9.89868581e-01 5.59741557e-01 -1.22751677e+00
5.14538527e-01 2.83506185e-01 4.64126617e-01 -3.15073907e-01
3.53777915e-01 4.08101469e-01 6.95498288e-02 -1.29684955e-01
-3.17418799e-02 1.28628790e-01 4.88829285e-01 5.41720927e-01
4.13978636e-01 -3.38133648e-02 7.71529436e-01 1.06226951e-01
1.32396245e+00 3.95411670e-01 -3.73834334e-02 -4.61036325e-01
3.10651273e-01 1.51620775e-01 5.97151995e-01 4.94196475e-01
-4.37145501e-01 7.40509033e-01 4.45531130e-01 -8.68880510e-01
-1.51583648e+00 -5.27156889e-01 -3.32028627e-01 1.04865158e+00
-2.16029942e-01 -1.53882399e-01 -1.05309641e+00 -4.12517220e-01
-4.63893533e-01 -5.86792501e-03 -7.05846906e-01 5.27885139e-01
-8.25336337e-01 -7.61225760e-01 7.43421078e-01 5.91570139e-01
5.60216606e-01 -1.33305061e+00 -1.04711068e+00 2.56120473e-01
2.74446547e-01 -1.19881666e+00 -6.03637984e-03 6.96846008e-01
-1.00795424e+00 -1.18052220e+00 -9.81371820e-01 -1.12397659e+00
9.20546234e-01 -6.81929514e-02 1.05386484e+00 5.84028542e-01
-6.50543332e-01 -1.81383733e-02 2.93671656e-02 -2.08989829e-01
1.42412893e-02 4.03077930e-01 -2.81959772e-01 -7.53006876e-01
1.18429646e-01 -8.29599619e-01 -8.96841526e-01 2.80149102e-01
-8.69547784e-01 3.34693730e-01 5.00436604e-01 1.05440891e+00
8.72185647e-01 -3.76501054e-01 3.06235909e-01 -1.06261301e+00
5.85872233e-02 -1.88413784e-01 -9.59553480e-01 -1.62778363e-01
-2.34193727e-01 -3.66631955e-01 7.30658710e-01 -2.98709869e-01
-5.21471918e-01 4.41200256e-01 -6.40972972e-01 -2.36817777e-01
-6.02699816e-01 2.22540051e-02 4.85648781e-01 -5.70841074e-01
3.72002453e-01 1.01777827e-02 6.98670000e-02 -4.76486355e-01
2.26636734e-02 1.22821651e-01 6.77769244e-01 -4.44771051e-01
2.83760369e-01 1.06947887e+00 1.74249366e-01 -5.87761939e-01
-6.43851519e-01 -4.97100264e-01 -9.53029573e-01 -6.58202842e-02
1.13728952e+00 -5.48991740e-01 -8.00861657e-01 5.39660752e-01
-1.31572652e+00 -1.13808250e+00 -1.64955243e-01 4.16436255e-01
-7.80406892e-01 1.75719798e-01 -1.03746521e+00 -3.65588158e-01
-6.37999058e-01 -1.04985380e+00 1.05404580e+00 3.48633438e-01
-5.72998337e-02 -1.04671991e+00 2.88377225e-01 2.15624258e-01
4.23224896e-01 2.10389033e-01 6.48997128e-01 -7.84398973e-01
-6.04103088e-01 -6.26941845e-02 -5.28041720e-01 2.81383336e-01
-2.63430953e-01 2.21984506e-01 -1.00212955e+00 -4.78499740e-01
-1.86294109e-01 -5.24643302e-01 1.05248010e+00 7.08292067e-01
1.37091625e+00 -2.24923026e-02 -4.39114511e-01 1.10855770e+00
1.59610748e+00 1.16835944e-02 5.53882539e-01 5.71941078e-01
6.47882104e-01 5.37494004e-01 5.09662293e-02 1.04670718e-01
1.34878665e-01 4.69549775e-01 5.75334370e-01 -3.99350882e-01
-5.00855967e-02 7.51149774e-01 -5.12178577e-02 6.30174577e-01
-3.27564210e-01 1.11162558e-01 -1.03977132e+00 7.01532662e-01
-1.69971299e+00 -8.28359842e-01 -1.95250019e-01 1.89894569e+00
9.85562623e-01 8.63611698e-02 1.80361375e-01 -8.85662735e-02
6.71632409e-01 -2.84374744e-01 -4.47429150e-01 -9.69124362e-02
-1.22025199e-01 5.50504804e-01 6.78174436e-01 5.73194802e-01
-1.17289364e+00 9.90730286e-01 5.83789158e+00 8.14934075e-01
-1.22053766e+00 3.91632855e-01 1.07077694e+00 -2.79541343e-01
3.57444406e-01 -1.33860901e-01 -9.40977454e-01 5.73227227e-01
1.02654517e+00 2.86258459e-01 4.16785628e-01 6.49990499e-01
1.98051939e-03 -2.15274990e-01 -9.12460625e-01 8.18850756e-01
-4.23169255e-01 -1.89933145e+00 -5.63074172e-01 3.01232874e-01
7.98060656e-01 8.14025462e-01 -1.06396146e-01 -3.96717824e-02
2.58039623e-01 -9.72525299e-01 4.43677783e-01 3.51574302e-01
7.29522765e-01 -7.44416952e-01 8.91496122e-01 2.57734895e-01
-8.42112303e-01 2.49947160e-01 -7.99216509e-01 9.94317010e-02
1.30192429e-01 6.44477308e-01 -6.96533918e-01 7.66300736e-03
9.64697242e-01 5.57142556e-01 -5.50194263e-01 1.26104450e+00
1.84906751e-01 6.04892373e-01 -6.87062204e-01 2.44753972e-01
4.76480901e-01 -2.52808392e-01 2.31246457e-01 1.54268956e+00
3.02140355e-01 -8.64216462e-02 -2.34844722e-03 8.29755843e-01
-2.90475100e-01 -2.80826569e-01 -3.90211672e-01 6.91028237e-02
1.53524265e-01 1.88211119e+00 -1.37176847e+00 -3.68668228e-01
-1.28623024e-01 9.15060103e-01 6.18808687e-01 2.42403492e-01
-5.91576397e-01 -4.07938629e-01 4.06125456e-01 1.26499720e-02
7.14960456e-01 -2.66310751e-01 -2.17158735e-01 -7.00464964e-01
-4.39692855e-01 -4.03211743e-01 -7.21318321e-03 -8.59636307e-01
-9.86592174e-01 7.35399127e-01 -4.55048740e-01 -6.30535185e-01
-2.60626227e-02 -1.17882466e+00 -7.11911142e-01 6.20568097e-01
-1.53025508e+00 -1.08702183e+00 -3.30457687e-01 4.94171768e-01
1.96097031e-01 1.81273952e-01 7.60531902e-01 5.00128329e-01
-5.93751252e-01 1.68006226e-01 1.59902081e-01 4.14025456e-01
5.89272499e-01 -1.49071908e+00 5.36628664e-01 6.79674983e-01
-7.61877820e-02 7.24631488e-01 4.58500832e-01 -3.29736352e-01
-8.24995875e-01 -1.13907862e+00 8.24554980e-01 -2.36917466e-01
7.87911594e-01 -3.19563627e-01 -9.73896742e-01 1.03886282e+00
6.28745496e-01 5.96832156e-01 6.72222257e-01 -4.57022548e-01
1.89204425e-01 2.00062886e-01 -1.16434336e+00 3.87123466e-01
8.14787745e-01 -1.96400151e-01 -1.63919419e-01 6.70739233e-01
4.90740061e-01 -5.14597297e-01 -8.39303911e-01 2.65476536e-02
4.35146302e-01 -1.10216153e+00 8.98174405e-01 -3.19712192e-01
4.35468525e-01 -3.42907280e-01 1.59874886e-01 -1.02449167e+00
-5.05283363e-02 -8.80501449e-01 9.41082314e-02 7.53619015e-01
1.86888054e-01 -6.88294470e-01 9.49781835e-01 2.09089413e-01
-3.72431636e-01 -1.17984140e+00 -1.18290615e+00 -5.92965424e-01
2.74440706e-01 -6.76923096e-02 6.68920726e-02 5.97353756e-01
-2.77583748e-01 1.26750931e-01 1.41253546e-01 -2.60928512e-01
7.47297525e-01 -1.15589723e-01 5.79035878e-01 -1.21702063e+00
-1.25931293e-01 -4.15257841e-01 -5.18019557e-01 -1.01834190e+00
2.32444227e-01 -7.89244473e-01 2.57675070e-02 -1.34547341e+00
1.65389806e-01 -4.85498428e-01 -2.02078506e-01 5.73502958e-01
1.72357976e-01 8.90861869e-01 6.77519068e-02 1.87885642e-01
-8.34824622e-01 1.33946175e-02 1.19452190e+00 2.01513380e-01
3.07111382e-01 -1.52521446e-01 -1.39997557e-01 1.07446086e+00
1.08414590e+00 -6.01619005e-01 1.45292774e-01 -5.78497291e-01
2.33364329e-01 -2.94115692e-01 7.29663014e-01 -1.33389258e+00
5.03193319e-01 5.33414185e-01 6.43360615e-01 -6.57710552e-01
4.66229260e-01 -7.72004724e-01 -1.14749677e-01 5.86226702e-01
-4.18502748e-01 1.95662081e-01 3.32119137e-01 1.50065452e-01
8.77136663e-02 -4.48126823e-01 1.15419030e+00 -4.42191750e-01
-4.77401227e-01 5.59952080e-01 -4.79691058e-01 7.99891502e-02
8.34933698e-01 -1.82940125e-01 -6.62469149e-01 1.08854614e-01
-8.06396246e-01 -1.57633685e-02 8.57579768e-01 -6.23022854e-01
3.69119138e-01 -1.01585674e+00 -3.57431859e-01 6.59447489e-03
-3.79845798e-01 3.07316661e-01 1.59497187e-01 1.31713963e+00
-1.28588247e+00 4.29162621e-01 -7.03149319e-01 -6.76361799e-01
-1.27319694e+00 2.08619252e-01 5.94757617e-01 -5.50089538e-01
-7.70168662e-01 1.30855203e+00 1.74205229e-01 -5.10858655e-01
8.00080672e-02 -4.88555074e-01 -5.40325716e-02 -2.95140326e-01
7.42682695e-01 3.57928723e-01 1.72856688e-01 -4.42827791e-01
-4.16810572e-01 4.53489363e-01 -1.46643087e-01 -1.60319537e-01
1.77422535e+00 -1.55445207e-02 -4.71907794e-01 2.39645839e-01
1.43184042e+00 -4.58955646e-01 -1.82617855e+00 4.61519100e-02
4.04497571e-02 -1.37469679e-01 2.90288627e-01 -7.05594897e-01
-1.41906881e+00 1.13153839e+00 7.38612294e-01 1.45644680e-01
8.26995850e-01 -1.38036132e-01 1.03117502e+00 5.94345033e-01
2.76297212e-01 -1.01514447e+00 -1.17588572e-01 6.97003484e-01
5.10389805e-01 -1.18786693e+00 -1.27711101e-02 9.86469537e-03
-2.36746490e-01 1.42874467e+00 5.85753083e-01 -6.69932842e-01
5.88350058e-01 7.84888923e-01 7.85227343e-02 -2.84552604e-01
-8.97952318e-01 -2.94099003e-01 -1.70109853e-01 8.18078637e-01
4.69451755e-01 -3.85541975e-01 -4.12569940e-02 1.38169006e-01
4.16651182e-02 2.60539502e-01 4.98427570e-01 1.01715863e+00
-5.27073681e-01 -9.55569327e-01 2.63341703e-02 4.28038090e-01
-1.01381481e+00 -2.41258487e-01 -1.14573285e-01 8.86247873e-01
3.32772881e-01 2.91774005e-01 4.77852017e-01 1.84840962e-01
-2.87074029e-01 -1.06797606e-01 8.34374130e-01 -5.92202365e-01
-8.09330404e-01 2.49412432e-01 -1.99698851e-01 -7.53246784e-01
-7.23966658e-01 -4.04034495e-01 -1.69477308e+00 -4.58813548e-01
-6.12904057e-02 -1.65338293e-01 8.03932548e-01 9.45173740e-01
2.88045853e-01 5.41658700e-01 -1.37204364e-01 -1.76027000e+00
1.62217617e-01 -9.47520792e-01 -6.09591305e-01 4.82674455e-03
3.47270846e-01 -2.84777313e-01 -3.88845235e-01 5.27254343e-01] | [14.312260627746582, -3.115811824798584] |
61cda6ca-a051-44ac-ba9b-c33247cbdcb6 | sound-event-localization-and-detection-for | 2209.01802 | null | https://arxiv.org/abs/2209.01802v2 | https://arxiv.org/pdf/2209.01802v2.pdf | Sound Event Localization and Detection for Real Spatial Sound Scenes: Event-Independent Network and Data Augmentation Chains | Sound event localization and detection (SELD) is a joint task of sound event detection and direction-of-arrival estimation. In DCASE 2022 Task 3, types of data transform from computationally generated spatial recordings to recordings of real-sound scenes. Our system submitted to the DCASE 2022 Task 3 is based on our previous proposed Event-Independent Network V2 (EINV2) with a novel data augmentation method. Our method employs EINV2 with a track-wise output format, permutation-invariant training, and a soft parameter-sharing strategy, to detect different sound events of the same class but in different locations. The Conformer structure is used for extending EINV2 to learn local and global features. A data augmentation method, which contains several data augmentation chains composed of stochastic combinations of several different data augmentation operations, is utilized to generalize the model. To mitigate the lack of real-scene recordings in the development dataset and the presence of sound events being unbalanced, we exploit FSD50K, AudioSet, and TAU Spatial Room Impulse Response Database (TAU-SRIR DB) to generate simulated datasets for training. We present results on the validation set of Sony-TAu Realistic Spatial Soundscapes 2022 (STARSS22) in detail. Experimental results indicate that the ability to generalize to different environments and unbalanced performance among different classes are two main challenges. We evaluate our proposed method in Task 3 of the DCASE 2022 challenge and obtain the second rank in the teams ranking. Source code is released. | ['Jun Yang', 'Mark D. Plumbley', 'Feiran Yang', 'Qiuqiang Kong', 'Ming Wu', 'Yin Cao', 'Jinbo Hu'] | 2022-09-05 | null | null | null | null | ['sound-event-detection', 'direction-of-arrival-estimation', 'room-impulse-response', 'sound-event-localization-and-detection'] | ['audio', 'audio', 'audio', 'audio'] | [ 2.29087025e-01 -6.22316957e-01 9.33398485e-01 -4.03035611e-01
-1.30452955e+00 -6.96476281e-01 3.37114125e-01 -3.06532457e-02
-5.58257639e-01 6.02832854e-01 3.95511359e-01 -8.45668763e-02
-3.43411565e-01 -3.60337853e-01 -7.94325292e-01 -6.27699018e-01
-5.11546910e-01 8.60195160e-02 4.81698126e-01 -1.64571539e-01
-2.96726227e-02 4.79421884e-01 -1.90725183e+00 7.14361131e-01
1.89214513e-01 1.11033142e+00 4.76268739e-01 1.15244055e+00
2.55876392e-01 7.86426902e-01 -9.82770443e-01 4.37451929e-01
3.93168837e-01 -4.50627804e-01 -3.55376899e-01 -6.43832564e-01
6.03370905e-01 5.85222691e-02 -2.38379255e-01 2.89236307e-01
1.31869113e+00 8.32801580e-01 4.08179939e-01 -1.49988663e+00
-5.23600867e-03 6.09064102e-01 1.36088073e-01 6.70110464e-01
3.51624757e-01 6.48780540e-02 7.23357558e-01 -1.31409514e+00
6.30333573e-02 9.13366199e-01 9.94840026e-01 3.86503309e-01
-1.10585058e+00 -9.74218607e-01 -1.01154886e-01 4.29472536e-01
-1.46421361e+00 -6.43397629e-01 7.78508186e-01 -4.17806834e-01
1.14631081e+00 7.38461137e-01 2.43547976e-01 1.66959560e+00
-3.55196029e-01 3.56934786e-01 1.12061679e+00 -4.33271229e-01
6.37722850e-01 1.24187969e-01 1.68130934e-01 -5.56788109e-02
-4.61705089e-01 4.33419466e-01 -1.18252921e+00 -1.71341464e-01
3.75312626e-01 -4.51613277e-01 -4.33914572e-01 2.57322848e-01
-1.12494850e+00 3.35499793e-01 1.94099024e-01 3.02601337e-01
-4.86003488e-01 2.14757875e-01 5.63305795e-01 2.79245824e-01
3.58115673e-01 6.23864233e-01 -6.60109937e-01 -4.52730566e-01
-8.16228509e-01 4.06192660e-01 8.23671937e-01 8.43674779e-01
1.94329992e-01 4.86945897e-01 -6.29504263e-01 9.70500648e-01
1.42768240e-02 6.21523082e-01 5.83247125e-01 -7.78378546e-01
6.86747611e-01 -1.98369741e-01 1.19216822e-01 -8.30193996e-01
-5.96162856e-01 -6.83433771e-01 -5.47530890e-01 9.17826369e-02
1.63828313e-01 -3.36578518e-01 -8.52620423e-01 2.06244946e+00
1.29880711e-01 7.68065214e-01 2.49987990e-02 8.83920312e-01
1.02559686e+00 9.09575403e-01 -2.72203311e-02 5.28693870e-02
1.02597606e+00 -8.41561913e-01 -5.98938525e-01 -1.44355327e-01
3.24204713e-01 -7.67661870e-01 1.15541077e+00 6.52098596e-01
-9.76098359e-01 -1.12274671e+00 -9.32518840e-01 4.07340467e-01
-4.20164615e-01 1.74877971e-01 3.49360496e-01 6.41702175e-01
-9.12665486e-01 1.19764768e-01 -6.81325972e-01 -1.70499563e-01
3.18428501e-02 3.58245745e-02 -3.76335055e-01 2.35564336e-01
-1.38355792e+00 2.13296011e-01 2.88803637e-01 1.74126327e-01
-1.34854662e+00 -1.07803881e+00 -7.53951550e-01 1.59457654e-01
2.08979353e-01 -2.13569194e-01 1.31940830e+00 -5.17520905e-01
-1.29934931e+00 1.98605403e-01 2.96160042e-01 -5.19132555e-01
4.15290594e-01 -2.69419700e-01 -9.76167440e-01 2.52475329e-02
1.36009634e-01 3.46755058e-01 7.98198402e-01 -1.21155369e+00
-7.60584891e-01 1.05265558e-01 -4.52107489e-01 1.66423991e-01
-3.12774509e-01 2.43542939e-01 -1.89958975e-01 -8.36139798e-01
-1.50731117e-01 -7.63523877e-01 -8.73411745e-02 -6.52793109e-01
-4.03847426e-01 -2.73293667e-02 4.92129445e-01 -5.90611577e-01
1.27917027e+00 -2.65092373e+00 -3.55173200e-01 2.53593564e-01
-1.80882424e-01 6.66381866e-02 -6.12188876e-01 5.56779385e-01
-6.26075447e-01 -3.31433058e-01 -8.55959877e-02 -7.92482674e-01
7.85431713e-02 -1.67473942e-01 -7.16356575e-01 1.65477723e-01
1.17510170e-01 1.90935135e-01 -8.04972112e-01 -5.18480502e-02
1.19620591e-01 3.44442904e-01 -6.37711346e-01 3.09362859e-01
5.18066287e-02 5.78433633e-01 -1.11390136e-01 3.05003703e-01
6.83154106e-01 3.54188621e-01 -4.14779216e-01 -2.44855538e-01
-4.39317495e-01 5.73259413e-01 -1.73737764e+00 1.77773690e+00
-8.27016473e-01 6.53526843e-01 -5.15975468e-02 -7.26196885e-01
9.09448385e-01 6.32098734e-01 2.84787029e-01 -6.96410000e-01
-7.23788813e-02 2.78183371e-01 -6.21042699e-02 -6.40663803e-01
5.35739183e-01 5.41597456e-02 -3.32210273e-01 2.78437734e-01
2.92381376e-01 -7.84139335e-02 5.96229285e-02 3.33085805e-02
1.50328505e+00 -4.72965129e-02 -2.12870106e-01 -8.86876360e-02
2.32669801e-01 -3.02049458e-01 5.79471588e-01 1.29280281e+00
-5.58311753e-02 1.16721153e+00 1.03233516e-01 -2.12587744e-01
-6.75492764e-01 -1.34772003e+00 -1.37082294e-01 1.39901674e+00
-3.66641074e-01 -2.61230022e-01 -4.60907429e-01 -5.37636757e-01
-2.53666639e-01 1.21393335e+00 -7.00270534e-01 -2.21551463e-01
-6.69294477e-01 -6.08485520e-01 1.15540552e+00 7.58068085e-01
3.42787176e-01 -1.44611013e+00 -7.85855234e-01 3.63962948e-01
-2.45311365e-01 -1.28549337e+00 -4.06088650e-01 7.78242171e-01
4.20015194e-02 -7.18919516e-01 -3.55716586e-01 -7.97756433e-01
4.62960824e-02 1.25745818e-01 1.02563596e+00 -5.86413145e-01
-4.48176324e-01 7.58965969e-01 -6.32012367e-01 -7.16382802e-01
-4.07954097e-01 -1.92375571e-01 3.66876125e-01 2.57513583e-01
-1.73036382e-01 -9.62841332e-01 -5.34312606e-01 4.40641701e-01
-9.45214391e-01 -2.60654539e-01 2.05791995e-01 7.01713324e-01
4.18867081e-01 1.48590818e-01 9.81730759e-01 -3.41518730e-01
6.24154270e-01 -5.84495902e-01 -3.60419124e-01 -4.60864231e-02
1.06970616e-01 -5.08175969e-01 7.98883379e-01 -6.79243684e-01
-1.12139034e+00 6.01227582e-02 -3.01709056e-01 -5.81487417e-01
-5.52636564e-01 2.53772199e-01 -1.60291791e-01 2.86307186e-01
1.08506346e+00 4.10663992e-01 -5.82769632e-01 -6.12172067e-01
8.97985995e-02 6.98560655e-01 8.25799465e-01 -4.70633715e-01
6.00744307e-01 6.14484474e-02 -1.18091285e-01 -8.04773092e-01
-5.99606454e-01 -5.23285329e-01 -1.40596464e-01 -2.36095101e-01
7.42334247e-01 -1.07185078e+00 -3.42490941e-01 6.10202968e-01
-1.22675025e+00 -4.82930392e-01 -5.27475715e-01 9.83218908e-01
-4.02444720e-01 -2.66521603e-01 -1.41475663e-01 -9.21953499e-01
-1.52882665e-01 -7.92793691e-01 9.20748472e-01 -1.55401751e-01
-4.73585427e-02 -4.83916551e-01 6.71716511e-01 1.61715344e-01
5.55311799e-01 1.69588104e-01 4.80033219e-01 -1.29007852e+00
-2.08144307e-01 -2.22964615e-01 3.10781479e-01 8.52245569e-01
7.90803209e-02 -3.24254215e-01 -1.47694433e+00 -1.17889851e-01
1.46347418e-01 -4.02327359e-01 8.05523992e-01 2.64272183e-01
1.25933993e+00 -1.00425087e-01 7.08919317e-02 5.68575263e-01
1.03595364e+00 4.35205549e-01 4.48271304e-01 1.27361208e-01
4.54305410e-01 4.83241051e-01 5.72124004e-01 7.06141174e-01
1.40489966e-01 8.37148547e-01 4.04848546e-01 -8.87390971e-02
-5.39180875e-01 -1.85302690e-01 4.59311336e-01 1.01121128e+00
1.71490416e-01 -7.16121078e-01 -9.82701302e-01 8.25957894e-01
-1.38177073e+00 -1.11470139e+00 -2.44386390e-01 2.34749866e+00
6.70658171e-01 -4.02521454e-02 2.47163279e-03 4.51304704e-01
6.74339235e-01 -6.47872537e-02 -7.69799203e-02 -2.63560712e-01
-2.66266316e-01 6.23247802e-01 2.88660266e-02 2.72156745e-01
-1.09271598e+00 5.18168807e-01 5.80098724e+00 9.77074325e-01
-1.13508272e+00 3.07235807e-01 1.78350866e-01 -4.39569473e-01
5.83889969e-02 -4.35445905e-01 -6.25136316e-01 6.38812125e-01
1.09492397e+00 4.01235431e-01 3.37682992e-01 4.49815452e-01
3.19694191e-01 1.01877436e-01 -1.17032623e+00 1.13079345e+00
1.35003254e-01 -1.14040899e+00 -3.27598304e-01 -4.29247499e-01
5.66372335e-01 2.58610576e-01 3.12474817e-01 7.32515216e-01
-9.30941626e-02 -9.23984110e-01 1.05425262e+00 5.24802625e-01
7.09430695e-01 -7.49253392e-01 5.82730472e-01 1.63530096e-01
-1.32284880e+00 -2.82590687e-01 5.63524142e-02 -6.29249662e-02
2.78801560e-01 3.98034781e-01 -1.19722056e+00 6.06972277e-01
1.15132070e+00 1.13253675e-01 -5.59636235e-01 1.32134306e+00
-4.71939296e-02 1.19194829e+00 -5.87025046e-01 7.81033039e-02
5.31104542e-02 5.60781121e-01 1.04327178e+00 1.42666602e+00
6.05116308e-01 -9.43826512e-02 8.93152431e-02 6.27990782e-01
7.23412037e-02 1.49586108e-02 -5.48883498e-01 4.08823192e-01
8.78288925e-01 1.23716903e+00 -4.12729979e-01 1.77794416e-02
-1.59922630e-01 7.56808341e-01 -9.18838978e-02 4.54751849e-01
-1.15444005e+00 -6.67144597e-01 3.89882058e-01 -8.01845118e-02
4.41251963e-01 -2.10460871e-02 1.21090531e-01 -3.98478538e-01
-2.33806781e-02 -8.50371897e-01 1.46926165e-01 -9.74305570e-01
-1.23334074e+00 1.09601510e+00 2.27357447e-01 -1.57890069e+00
-1.12180836e-01 -1.89672157e-01 -8.26079488e-01 8.34008574e-01
-1.15261185e+00 -9.78201687e-01 -5.11335433e-01 7.67013967e-01
7.15061903e-01 -4.52809215e-01 9.64073420e-01 7.67045557e-01
-5.60978591e-01 9.12352502e-01 5.18755801e-02 -8.50343108e-02
9.56283391e-01 -1.16492927e+00 4.64362651e-01 8.67619812e-01
3.84919822e-01 1.30350351e-01 7.11321235e-01 -3.50851417e-01
-8.09268713e-01 -1.52348554e+00 6.58141315e-01 -5.07755637e-01
5.70236206e-01 -1.00422883e+00 -7.86524534e-01 4.95481104e-01
-6.04321547e-02 4.61330324e-01 7.58497596e-01 -1.47413164e-01
-2.64838576e-01 -3.17738682e-01 -8.52889419e-01 2.54735440e-01
1.04587853e+00 -4.57412332e-01 -5.85976720e-01 2.92039782e-01
9.15601552e-01 -4.17473495e-01 -4.32372212e-01 4.52926546e-01
2.53755778e-01 -9.38052118e-01 1.05082548e+00 -4.89190489e-01
9.55543891e-02 -3.63865763e-01 -6.80938244e-01 -1.37418437e+00
-2.00890198e-01 -6.19263768e-01 1.49241373e-01 1.64272273e+00
4.03186113e-01 -5.84251761e-01 1.75056413e-01 -8.59619230e-02
-6.84711933e-01 -2.47787774e-01 -1.26365137e+00 -1.02718163e+00
-4.99348164e-01 -1.18421221e+00 5.18823385e-01 7.42942333e-01
-6.99498117e-01 3.00729930e-01 -3.99151057e-01 6.83818221e-01
3.15198362e-01 -2.10587740e-01 7.74553895e-01 -1.04478705e+00
-6.02817595e-01 1.67625118e-02 -3.73919904e-01 -6.12687647e-01
-7.28663132e-02 -8.00540209e-01 6.82365656e-01 -9.58623230e-01
-2.69754916e-01 -7.00467527e-01 -8.56848240e-01 5.56361437e-01
4.04428169e-02 2.35740215e-01 1.72630087e-01 -2.04074457e-01
-6.66599572e-01 7.57760346e-01 6.37681603e-01 2.57868946e-01
-5.11075497e-01 5.49587309e-01 -2.38032341e-01 5.51136315e-01
7.51905441e-01 -7.71303058e-01 -6.18058622e-01 -4.00486231e-01
2.48619631e-01 1.14019617e-01 9.20091808e-01 -1.69893968e+00
3.46577078e-01 1.99900031e-01 2.57082820e-01 -8.01058888e-01
6.50148153e-01 -8.11001062e-01 2.94310212e-01 2.47078319e-03
-5.86567223e-01 2.27051880e-02 7.77880073e-01 5.96725643e-01
-2.76572704e-01 -1.65952798e-02 4.30021405e-01 1.62792176e-01
-6.58027172e-01 -3.02491151e-03 -4.72095072e-01 2.89568067e-01
8.15161407e-01 1.50672108e-01 -9.88838077e-02 -2.46194974e-01
-7.35638857e-01 -1.61389738e-01 -4.69864190e-01 6.83547080e-01
6.19701684e-01 -1.39206386e+00 -1.04661071e+00 4.28685993e-01
1.25268206e-01 -6.86354861e-02 8.03625166e-01 6.38068199e-01
-1.46789506e-01 1.57283008e-01 -1.02977999e-01 -5.55297673e-01
-1.13993680e+00 1.39292315e-01 4.17217046e-01 -1.06055290e-02
-4.68462080e-01 1.30787027e+00 3.86008739e-01 -7.13685274e-01
6.15983605e-01 -4.16454226e-01 -2.69342303e-01 9.97988358e-02
5.64089537e-01 6.01833344e-01 4.82863486e-01 -3.27723384e-01
-6.34232998e-01 1.87414046e-02 3.55221987e-01 -5.32006562e-01
1.52344787e+00 -2.02764645e-02 3.27275574e-01 6.90619290e-01
1.18308854e+00 5.13833284e-01 -1.03569150e+00 -7.00303838e-02
-3.59199196e-01 -9.98538807e-02 2.07089514e-01 -1.25323188e+00
-7.91399479e-01 7.76895642e-01 1.12743044e+00 2.44616389e-01
1.41635895e+00 -1.07536167e-01 5.51394522e-01 2.49885961e-01
1.64864734e-01 -1.01745403e+00 2.03858197e-01 6.26767159e-01
1.19534945e+00 -8.30322564e-01 -6.33706808e-01 5.03338464e-02
-5.79927325e-01 7.68890977e-01 8.17392468e-01 2.73948498e-02
5.95810354e-01 5.65818846e-01 9.85284224e-02 -3.25023732e-03
-9.14898038e-01 2.29014438e-02 2.71748543e-01 6.95402741e-01
-2.10579168e-02 -1.57141313e-01 3.82221937e-01 1.27784193e+00
-3.50924551e-01 -3.45165104e-01 1.61339730e-01 9.28736746e-01
-8.65643099e-02 -6.52972877e-01 -5.87993324e-01 1.62482336e-02
-2.52587169e-01 -4.33393449e-01 1.04592800e-01 6.74716592e-01
3.72656107e-01 1.08004797e+00 2.33407110e-01 -6.16004765e-01
8.77577424e-01 1.20041616e-01 2.07615271e-01 -5.36981642e-01
-9.89535749e-01 -2.29948089e-02 2.01643050e-01 -5.41090846e-01
-3.04083675e-02 -7.00499892e-01 -1.10034359e+00 4.21594471e-01
-3.23027335e-02 2.28518665e-01 8.58155489e-01 5.73886573e-01
5.45703590e-01 1.42879891e+00 1.02417266e+00 -9.61305320e-01
-5.52568138e-01 -1.13972974e+00 -6.67531252e-01 2.89158493e-01
5.18851459e-01 -5.31620860e-01 -6.70513988e-01 -2.45601423e-02] | [15.149701118469238, 5.244039535522461] |
883e0833-6b58-40c9-9161-ccd8d8cd4afc | challenges-and-applications-of-automated | 2108.07865 | null | https://arxiv.org/abs/2108.07865v1 | https://arxiv.org/pdf/2108.07865v1.pdf | Challenges and Applications of Automated Extraction of Socio-political Events from Text (CASE 2021): Workshop and Shared Task Report | This workshop is the fourth issue of a series of workshops on automatic extraction of socio-political events from news, organized by the Emerging Market Welfare Project, with the support of the Joint Research Centre of the European Commission and with contributions from many other prominent scholars in this field. The purpose of this series of workshops is to foster research and development of reliable, valid, robust, and practical solutions for automatically detecting descriptions of socio-political events, such as protests, riots, wars and armed conflicts, in text streams. This year workshop contributors make use of the state-of-the-art NLP technologies, such as Deep Learning, Word Embeddings and Transformers and cover a wide range of topics from text classification to news bias detection. Around 40 teams have registered and 15 teams contributed to three tasks that are i) multilingual protest news detection, ii) fine-grained classification of socio-political events, and iii) discovering Black Lives Matter protest events. The workshop also highlights two keynote and four invited talks about various aspects of creating event data sets and multi- and cross-lingual machine learning in few- and zero-shot settings. | ['Erdem Yörük', 'Reyyan Yeniterzi', 'Jakub Piskorski', 'Vanni Zavarella', 'Hristo Tanev', 'Ali Hürriyetoğlu'] | 2021-08-17 | null | https://aclanthology.org/2021.case-1.1 | https://aclanthology.org/2021.case-1.1.pdf | acl-case-2021-8 | ['learning-word-embeddings'] | ['methodology'] | [-1.95059463e-01 1.34194707e-02 -3.72362971e-01 -4.94543850e-01
-1.20769846e+00 -6.83571935e-01 1.31614351e+00 7.86660194e-01
-6.79355860e-01 6.99885964e-01 1.33574140e+00 -4.77954373e-03
5.07162809e-02 -8.26015532e-01 -3.66192251e-01 -2.26738781e-01
-6.54463917e-02 8.66606534e-01 2.91455865e-01 -7.76019573e-01
3.97965640e-01 3.63319755e-01 -1.33388257e+00 9.10081506e-01
2.70064861e-01 3.71747404e-01 -3.18803072e-01 6.37197316e-01
-4.93562907e-01 1.20841002e+00 -5.55712521e-01 -7.48043478e-01
-1.12543859e-01 -4.36837733e-01 -1.01262248e+00 -4.31552202e-01
3.36427718e-01 2.06549823e-01 -4.84022856e-01 8.90052021e-01
8.66002262e-01 5.66055179e-02 4.29051578e-01 -8.15031111e-01
-3.14221442e-01 1.39511740e+00 -5.10151744e-01 9.85139072e-01
5.86496115e-01 -4.44707662e-01 1.31236255e+00 -8.32479835e-01
1.41479266e+00 1.68659949e+00 9.34483409e-01 -1.25667024e-02
-1.03601325e+00 -7.51210809e-01 -3.50969322e-02 3.07023317e-01
-7.15760052e-01 -8.09840560e-01 6.58819973e-01 -8.22521508e-01
1.26346016e+00 2.13909194e-01 6.88323200e-01 1.66079128e+00
3.35315973e-01 7.65961230e-01 1.09590364e+00 -5.87514520e-01
1.94031175e-03 2.94748187e-01 4.85569477e-01 4.95429516e-01
5.81906587e-02 -6.67732954e-02 -9.20225501e-01 -8.66717994e-01
3.37330326e-02 -2.94991374e-01 2.48611033e-01 1.09776959e-01
-1.40857100e+00 1.51847577e+00 -3.78734916e-02 8.25178146e-01
-4.21325743e-01 -2.57506073e-01 1.23636210e+00 5.94938517e-01
1.22736561e+00 4.07326609e-01 -4.36617732e-01 -4.62810367e-01
-8.94239962e-01 6.14385486e-01 1.03884363e+00 1.81571022e-01
2.98979491e-01 -2.26373017e-01 -2.13003770e-01 1.14855206e+00
4.05975021e-02 3.47748131e-01 5.08548856e-01 -4.73971307e-01
8.46209168e-01 5.32192588e-01 1.44819245e-01 -1.36661375e+00
-9.45638835e-01 -6.37780456e-03 -2.91545689e-01 -2.07716733e-01
3.68816614e-01 -7.31954634e-01 -4.84177232e-01 1.46735680e+00
6.48962975e-01 -1.69091642e-01 9.95772630e-02 6.62138283e-01
1.31241667e+00 9.11632776e-01 1.00004092e-01 -4.59354371e-01
1.88522863e+00 -5.12409866e-01 -9.59833384e-01 -3.45667660e-01
5.61268508e-01 -1.25736058e+00 5.47302365e-01 -1.86526582e-01
-1.00583315e+00 4.12211977e-02 -6.68638527e-01 -2.37133265e-01
-7.57938206e-01 -3.82737905e-01 5.23570478e-01 3.28772336e-01
-2.97174931e-01 3.09825689e-01 -6.44244075e-01 -1.06628144e+00
1.84550539e-01 -4.30285782e-01 -2.81001359e-01 2.79293031e-01
-1.70872128e+00 1.16038406e+00 3.67614686e-01 -7.31347382e-01
-5.80457211e-01 -8.50555837e-01 -9.17428374e-01 -4.18074518e-01
1.48425326e-01 -1.76889342e-04 1.01030421e+00 -6.54729724e-01
-9.27258730e-01 1.88195908e+00 1.23369813e-01 -7.21491277e-01
5.35907388e-01 -3.10051203e-01 -1.00574875e+00 -2.58639485e-01
6.85294449e-01 -4.50136103e-02 2.06004754e-01 -4.22095180e-01
-9.68553901e-01 -3.35884720e-01 -1.62156388e-01 1.30818248e-01
-2.46408999e-01 1.35082531e+00 4.52917069e-01 -6.50692940e-01
-1.87285274e-01 -7.04504073e-01 2.29520295e-02 -8.47954392e-01
-4.21144664e-01 -5.84288776e-01 6.40732765e-01 -1.04055166e+00
1.04386675e+00 -2.04818940e+00 -4.55588102e-02 -3.09652448e-01
-9.21013877e-02 -7.44349286e-02 1.73110798e-01 1.22929013e+00
-9.67271850e-02 -1.02109246e-01 4.99158472e-01 1.89074948e-02
2.92086869e-01 1.26492172e-01 -8.59989583e-01 8.42192411e-01
-9.45163816e-02 4.81764257e-01 -1.19719136e+00 -4.72976476e-01
7.51420250e-03 2.27507934e-01 -6.51584193e-02 -4.17378128e-01
1.57639682e-01 5.95236979e-02 -3.97065759e-01 1.35303974e-01
2.23171577e-01 1.69864729e-01 4.04507637e-01 -6.71481043e-02
-8.53497684e-01 1.06929052e+00 -1.07991540e+00 1.33029854e+00
-1.98844552e-01 1.28303111e+00 1.28283650e-01 -1.12538838e+00
8.10855269e-01 5.99315345e-01 4.10714120e-01 -6.34041369e-01
4.12176698e-01 2.75537550e-01 -2.73708016e-01 -5.79162776e-01
7.20113695e-01 -4.76639658e-01 -8.46617877e-01 6.86026454e-01
1.78841770e-01 3.34432721e-01 4.87414807e-01 2.47146681e-01
1.13235414e+00 -3.33552122e-01 7.83221483e-01 -6.47982180e-01
1.74121618e-01 2.74045259e-01 8.78708422e-01 6.71385229e-01
-3.54207844e-01 2.14074299e-01 6.39486074e-01 -1.32872748e+00
-1.32510424e+00 -6.67778134e-01 -4.80778486e-01 1.83557391e+00
-4.09720212e-01 -4.89246488e-01 -3.46028805e-01 -7.27927864e-01
-4.73809838e-02 7.35010684e-01 -7.83798456e-01 5.88009953e-01
-1.08719146e+00 -1.33007288e+00 8.17288518e-01 -8.21686909e-02
7.39409700e-02 -1.29884398e+00 -7.59193599e-01 5.72426856e-01
-8.19524646e-01 -1.11703479e+00 -2.12734759e-01 3.26161236e-01
-1.62811279e-01 -1.37990618e+00 -4.90992576e-01 -1.16512918e+00
-1.96793288e-01 -2.07161158e-01 1.24007571e+00 -6.69120371e-01
-6.46797538e-01 -4.32468317e-02 -4.51044589e-01 -7.06240535e-01
-7.27217495e-01 1.10297993e-01 3.40883911e-01 -2.13390335e-01
8.86683643e-01 -3.54380190e-01 -5.86922630e-04 -1.51872188e-01
-4.28318828e-01 -1.51999131e-01 -3.73589471e-02 8.15501869e-01
1.81627706e-01 -3.33520055e-01 8.25623274e-01 -1.14650440e+00
8.95614922e-01 -7.17079997e-01 -3.14061016e-01 -1.36107448e-02
3.30998376e-02 -2.41906643e-01 1.98942706e-01 -1.97097391e-01
-1.15010619e+00 -4.99353051e-01 -3.49456787e-01 6.20040059e-01
-2.76803106e-01 5.41124046e-01 4.30653840e-01 5.45372665e-01
1.30167782e+00 -2.04183117e-01 -4.88398284e-01 -4.93876547e-01
4.40782756e-01 9.83297646e-01 1.92113504e-01 -3.11704725e-01
5.89942575e-01 7.75091112e-01 -7.03756750e-01 -1.22480726e+00
-1.37389767e+00 -8.42401564e-01 -1.95171982e-01 -2.24567100e-01
9.68885303e-01 -1.16719985e+00 -3.79915647e-02 2.93157101e-01
-1.35588253e+00 2.67791990e-02 -5.35966992e-01 5.40411949e-01
-1.63669795e-01 5.61222248e-02 -9.68131781e-01 -5.99120498e-01
-4.95554566e-01 -3.83261174e-01 7.72869587e-01 -2.22434867e-02
-7.63681352e-01 -1.04580164e+00 9.01543796e-01 4.63386804e-01
4.82356027e-02 1.04263103e+00 6.72543526e-01 -1.20250976e+00
5.21777689e-01 -2.47905567e-01 -8.71835127e-02 -3.08758855e-01
-1.62375689e-01 -5.61449565e-02 -7.68363357e-01 -1.77550882e-01
-2.00188428e-01 -6.85347736e-01 9.57100511e-01 4.56727445e-01
-2.43750393e-01 -4.77541864e-01 -6.02958381e-01 1.28907396e-03
1.07585037e+00 5.12122139e-02 3.36509615e-01 1.00098515e+00
3.39493267e-02 7.38700032e-01 3.61267805e-01 7.66478360e-01
3.81494910e-01 5.99106967e-01 -2.38121748e-01 2.67376155e-01
-1.65632859e-01 -1.02019655e-02 5.77393711e-01 9.30907011e-01
1.91319287e-02 1.38154954e-01 -1.16136837e+00 1.21789753e+00
-1.92700636e+00 -1.79414821e+00 -6.41851962e-01 1.64163709e+00
9.32184100e-01 2.95964062e-01 2.23261327e-01 8.85012299e-02
1.04050434e+00 7.87603259e-01 1.80081993e-01 -8.71291280e-01
-3.74655873e-01 3.91646475e-02 4.35740411e-01 4.08998847e-01
-1.57640612e+00 9.27243590e-01 6.69075918e+00 6.96304917e-01
-8.14267397e-01 7.62441695e-01 2.88297236e-01 -2.89133936e-01
6.44188724e-04 -1.17376089e-01 -1.14336252e+00 3.79890800e-01
1.13266444e+00 -5.06989896e-01 8.22506323e-02 6.83373749e-01
3.03882480e-01 3.35294098e-01 -5.64279199e-01 4.35956836e-01
1.83085427e-01 -2.01085687e+00 -4.91983622e-01 -5.59627861e-02
9.51193929e-01 1.06758511e+00 -5.49218595e-01 5.80864012e-01
8.27872753e-01 -4.01533335e-01 9.79889572e-01 6.97639287e-02
4.43989068e-01 -8.70426893e-01 8.27626705e-01 2.83085793e-01
-9.03908968e-01 -1.91770971e-01 -3.16044182e-01 -3.11832458e-01
6.17622733e-01 1.01381671e+00 -4.16962266e-01 2.92473793e-01
9.66687262e-01 6.27082884e-01 2.12115478e-02 7.85123050e-01
-1.66821942e-01 6.83859825e-01 -3.26711059e-01 -3.39317948e-01
6.13324404e-01 4.34353530e-01 1.25006652e+00 1.89792633e+00
4.25363285e-03 6.09157644e-02 4.22955334e-01 1.62558541e-01
-5.25679961e-02 3.91292602e-01 -6.23410463e-01 -1.29367173e-01
5.08219182e-01 1.27552724e+00 -9.12430584e-01 -3.03730607e-01
-2.50466466e-01 4.17562276e-01 4.64869797e-01 -1.72147885e-01
-7.70098746e-01 -5.65774143e-01 7.14188039e-01 3.39944661e-01
1.50494903e-01 9.69350785e-02 3.43860760e-02 -1.09703910e+00
-4.55406427e-01 -8.86021852e-01 9.58320677e-01 -5.66159151e-02
-1.59384096e+00 5.83313406e-01 1.45616651e-01 -6.03164375e-01
-2.61806399e-01 -8.89289081e-02 -7.72794247e-01 5.61335266e-01
-1.05105901e+00 -1.10473537e+00 2.44399607e-01 2.65054911e-01
6.65303528e-01 -5.76071739e-01 9.43977177e-01 5.07340729e-01
-2.87342399e-01 1.44936487e-01 3.56273383e-01 5.62730432e-01
1.15799296e+00 -9.04942930e-01 5.93334079e-01 4.60086256e-01
3.35202903e-01 6.69726729e-02 1.13983238e+00 -7.07750022e-01
-6.61355674e-01 -1.08238447e+00 2.07292342e+00 -4.44339901e-01
1.49204206e+00 -5.60816228e-01 -2.20103264e-01 6.56810701e-01
6.43045485e-01 -4.57803965e-01 9.73402023e-01 6.97389126e-01
-4.68019873e-01 1.94636539e-01 -1.12420917e+00 3.58752966e-01
6.97868943e-01 -5.25907278e-01 -1.38113856e+00 1.21011186e+00
4.99418557e-01 -3.84272605e-01 -5.73128581e-01 -1.29144058e-01
5.04360318e-01 -5.76891720e-01 9.05243218e-01 -1.09569263e+00
6.43449306e-01 2.59117037e-01 -1.21557340e-01 -1.25079215e+00
-5.81281543e-01 -6.96526468e-01 2.82525688e-01 1.48209667e+00
3.90221983e-01 -4.44654793e-01 3.17482382e-01 -2.45768830e-01
-2.45203435e-01 5.03580868e-02 -1.40663266e+00 -5.37409544e-01
1.57918513e-01 -3.68335754e-01 -4.84844930e-02 1.67941523e+00
3.92772555e-01 8.19009900e-01 -5.66429913e-01 -1.69022456e-01
5.42312443e-01 1.12390801e-01 5.78556895e-01 -1.49280119e+00
6.30185381e-02 -4.87015873e-01 -5.04643321e-01 -1.13537677e-01
1.07855126e-01 -9.52532947e-01 -1.11611150e-01 -1.83386981e+00
2.74750262e-01 8.59871060e-02 -6.00972734e-02 3.48976135e-01
3.40414494e-01 4.42238271e-01 -1.51831105e-01 1.95119381e-01
-6.40808523e-01 4.63384353e-02 5.10662317e-01 -2.46308506e-01
-2.27992505e-01 -3.19539815e-01 -6.49612725e-01 1.11449516e+00
5.64679205e-01 -9.34248745e-01 4.86168057e-01 -2.63295144e-01
7.93741584e-01 -1.99915543e-01 2.63936847e-01 -6.54113948e-01
-2.63951570e-02 -2.86729783e-01 6.11579977e-02 -7.20657110e-01
1.06982581e-01 2.16443036e-02 1.25392806e-03 4.67044055e-01
-5.14168918e-01 -8.62705186e-02 1.59481302e-01 4.96812284e-01
-3.10860723e-01 -1.79691941e-01 9.19699669e-01 -4.36598688e-01
-6.37031078e-01 -2.23687798e-01 -8.87279749e-01 8.47758770e-01
1.10882854e+00 5.26719511e-01 -8.15646052e-01 -2.23719597e-01
-7.78210044e-01 2.46661287e-02 -1.15784273e-01 7.43339837e-01
-1.05072118e-01 -1.47011900e+00 -1.60621667e+00 -3.40470195e-01
1.69344187e-01 -9.70880270e-01 2.41484940e-01 9.84068930e-01
-6.18518233e-01 3.69574130e-01 -4.93230857e-02 1.93576496e-02
-1.46896899e+00 2.25464985e-01 -5.48507683e-02 -7.28035510e-01
-1.13636434e+00 7.91879475e-01 -3.21117461e-01 -6.51261866e-01
7.00510815e-02 1.23879589e-01 -5.80554545e-01 1.15864575e+00
1.08368385e+00 4.37334418e-01 -8.68726820e-02 -1.05844736e+00
-7.63290584e-01 5.47544882e-02 -2.11612478e-01 -1.80125430e-01
1.85323620e+00 -3.75316627e-02 -1.07748687e-01 8.84604335e-01
1.06169200e+00 2.13824347e-01 -3.92625332e-01 -4.50233519e-01
3.67311537e-01 -4.75008041e-03 1.41494617e-01 -7.83595443e-01
-3.94964337e-01 4.99246240e-01 3.38860810e-01 9.30260301e-01
1.69739157e-01 4.31944430e-01 9.41832483e-01 1.03648178e-01
1.15393147e-01 -1.80986512e+00 -2.18306482e-01 1.04438221e+00
1.14846087e+00 -1.11866605e+00 1.52304843e-01 7.34403357e-03
-5.32373428e-01 1.15934229e+00 -2.26972997e-01 -4.17471319e-01
7.99631536e-01 4.43051875e-01 1.23693407e-01 -6.51980162e-01
-7.88440585e-01 -2.01225579e-01 1.62518308e-01 3.53679121e-01
7.78577030e-01 6.31427243e-02 -1.00432980e+00 6.36509180e-01
-3.51667821e-01 -3.89840722e-01 4.68724787e-01 8.41073573e-01
-7.70369768e-01 -8.66911352e-01 -4.17658895e-01 3.65478367e-01
-1.15450168e+00 -1.96817741e-01 -6.44131422e-01 9.76467907e-01
3.10704321e-01 8.91119361e-01 1.80539399e-01 -1.17877424e-01
5.07886946e-01 3.27870011e-01 4.52599712e-02 -6.63731396e-01
-1.20460570e+00 3.78207602e-02 1.24362814e+00 -1.98899716e-01
-8.13476443e-01 -1.15801895e+00 -8.81436884e-01 -7.80364335e-01
2.44094077e-02 5.36228657e-01 7.49739468e-01 1.04136395e+00
9.02590007e-02 4.39329624e-01 3.99518996e-01 -7.81093955e-01
-2.99488544e-01 -1.17665648e+00 -5.00445247e-01 5.65899789e-01
-1.91745460e-02 -4.41771299e-01 -3.83917630e-01 -4.60316204e-02] | [9.061468124389648, 9.736250877380371] |
4526414b-678b-4b8e-b6ed-f5707003c8d9 | contextualized-word-embeddings-enhanced-event | 1904.11942 | null | http://arxiv.org/abs/1904.11942v1 | http://arxiv.org/pdf/1904.11942v1.pdf | Contextualized Word Embeddings Enhanced Event Temporal Relation Extraction for Story Understanding | Learning causal and temporal relationships between events is an important
step towards deeper story and commonsense understanding. Though there are
abundant datasets annotated with event relations for story comprehension, many
have no empirical results associated with them. In this work, we establish
strong baselines for event temporal relation extraction on two under-explored
story narrative datasets: Richer Event Description (RED) and Causal and
Temporal Relation Scheme (CaTeRS). To the best of our knowledge, these are the
first results reported on these two datasets. We demonstrate that neural
network-based models can outperform some strong traditional linguistic
feature-based models. We also conduct comparative studies to show the
contribution of adopting contextualized word embeddings (BERT) for event
temporal relation extraction from stories. Detailed analyses are offered to
better understand the results. | ['Rujun Han', 'Bashar Alhafni', 'Nanyun Peng', 'Mengyue Liang'] | 2019-04-26 | null | null | null | null | ['temporal-relation-extraction'] | ['natural-language-processing'] | [ 3.33288193e-01 1.54322103e-01 -6.94492400e-01 -4.88571256e-01
-5.85191309e-01 -5.15577734e-01 1.22093022e+00 6.49381876e-01
-5.02383769e-01 9.42953944e-01 1.23884213e+00 -2.14048192e-01
-2.18128592e-01 -1.01006806e+00 -6.04876459e-01 -4.84255637e-04
-4.86743987e-01 5.19928448e-02 2.79883593e-01 -4.54934150e-01
3.90457839e-01 2.33261496e-01 -1.36013305e+00 8.40491831e-01
5.86724877e-01 7.49704778e-01 -2.15905800e-01 5.07173955e-01
1.56366415e-02 1.74646449e+00 -4.17971492e-01 -5.75074792e-01
-3.32913727e-01 -4.36813504e-01 -1.27973807e+00 -4.10783321e-01
-1.94571033e-01 -5.26056826e-01 -8.87588680e-01 4.46188509e-01
4.24201906e-01 5.98126948e-01 9.05377626e-01 -1.16667151e+00
-1.13744259e+00 1.48490596e+00 -4.31189418e-01 9.98269916e-01
9.24210966e-01 3.57471108e-02 1.45493770e+00 -6.73731804e-01
1.00899673e+00 1.16635048e+00 7.34390616e-01 2.43227616e-01
-7.96802938e-01 -6.73691869e-01 3.27513635e-01 1.11659431e+00
-1.01313233e+00 -3.76721889e-01 1.01734090e+00 -4.33304161e-01
1.57581806e+00 9.52253044e-02 8.30040276e-01 1.92902184e+00
5.84517792e-02 9.83265400e-01 1.03176641e+00 -4.32654202e-01
-8.04744661e-02 -2.92819560e-01 5.63244343e-01 2.74432510e-01
-3.69707285e-03 1.87664330e-01 -1.18092918e+00 1.44134939e-01
5.69508910e-01 -1.21043153e-01 -3.43650043e-01 4.19517815e-01
-1.52228248e+00 1.02491736e+00 4.09606755e-01 6.21084809e-01
-5.25885284e-01 4.08427835e-01 8.41379046e-01 -2.09788699e-02
6.43583357e-01 4.51103747e-01 -3.26018751e-01 -6.61322773e-01
-7.10677266e-01 4.30854589e-01 6.54443860e-01 7.58750677e-01
-4.53233309e-02 -1.11285359e-01 -5.35536766e-01 6.58541679e-01
-6.84524328e-02 -2.49962091e-01 4.34199959e-01 -6.20034993e-01
6.94968164e-01 6.26585126e-01 -1.50097057e-01 -1.26091802e+00
-6.70269191e-01 4.39246967e-02 -4.46896583e-01 -4.43838179e-01
4.15428996e-01 -2.09464714e-01 -2.77031571e-01 1.75396097e+00
1.01086296e-01 6.32950425e-01 8.37867409e-02 7.88065612e-01
1.43755496e+00 6.90816998e-01 2.96991408e-01 -2.87646890e-01
1.66387999e+00 -6.91195786e-01 -1.28322983e+00 -2.40747184e-01
6.00263476e-01 -5.81676066e-01 1.12171662e+00 4.84199077e-02
-1.01708090e+00 -6.82999194e-02 -1.18550849e+00 -4.71122980e-01
-6.53993070e-01 -1.59774467e-01 1.31095922e+00 4.95440587e-02
-3.25169891e-01 5.79226315e-01 -8.41418743e-01 -7.06342995e-01
8.33013117e-01 -3.49766403e-01 -2.48441115e-01 3.87718575e-03
-1.95648253e+00 1.23905003e+00 9.43148911e-01 -1.76105008e-01
-8.57873976e-01 -1.00817490e+00 -1.12493587e+00 -1.53083116e-01
6.46109581e-01 -3.60323191e-01 1.31563628e+00 -2.80213416e-01
-1.05158317e+00 9.00070071e-01 -3.05743396e-01 -8.12652230e-01
1.79977834e-01 -5.14212906e-01 -7.64086425e-01 1.64410517e-01
1.63533852e-01 4.11823660e-01 1.43654481e-01 -8.01555574e-01
-4.20892775e-01 3.90579887e-02 5.01192331e-01 1.27311349e-01
-4.77459878e-01 7.14678526e-01 9.68694910e-02 -9.56166804e-01
-2.03440860e-01 -3.12735230e-01 2.85255332e-02 -3.38210464e-01
-4.99860078e-01 -7.61474729e-01 5.36472023e-01 -7.45992184e-01
1.64513803e+00 -1.89537168e+00 -7.16703385e-02 -6.54210687e-01
2.65156236e-02 -3.99361581e-01 1.61420405e-01 9.03742373e-01
-4.17424083e-01 2.38046914e-01 -2.05201685e-01 -1.22999571e-01
1.71949238e-01 2.11511001e-01 -8.07652295e-01 2.09438667e-01
6.05484426e-01 1.39058816e+00 -1.15945792e+00 -6.76072121e-01
2.39825379e-02 4.37606722e-01 -3.71939778e-01 -1.18031008e-02
-2.45209813e-01 1.26208469e-01 -4.07138616e-01 3.43150645e-01
-7.21035227e-02 -1.40285015e-01 -5.03326356e-02 -3.59627008e-01
-1.36984795e-01 1.07592535e+00 -8.20478380e-01 1.61067903e+00
-2.92548686e-01 1.22478557e+00 -1.02311468e+00 -8.34005773e-01
5.04553974e-01 6.74512148e-01 2.83917159e-01 -4.96360302e-01
4.15409565e-01 -4.11467731e-01 -2.50560585e-02 -8.89489174e-01
7.55363464e-01 -4.62716848e-01 -3.61530960e-01 6.76383197e-01
1.15461275e-01 -1.60138234e-01 5.58031738e-01 4.92552251e-01
1.19784951e+00 2.44136170e-01 7.83696711e-01 8.74229074e-02
1.23969615e-01 3.14231247e-01 4.57871705e-01 5.44424355e-01
-2.96848297e-01 5.73552132e-01 7.57070482e-01 -4.76039052e-01
-7.69205213e-01 -1.08763397e+00 -2.92298287e-01 9.52974558e-01
-6.12701029e-02 -7.29669869e-01 -1.76256210e-01 -5.08929074e-01
-3.31160218e-01 1.46103024e+00 -1.06200838e+00 1.11580066e-01
-8.63863647e-01 -9.28076148e-01 8.62251282e-01 1.01262999e+00
4.68274981e-01 -1.32884216e+00 -7.75838077e-01 2.65486270e-01
-5.58601499e-01 -1.44640720e+00 -1.65809095e-01 1.67461574e-01
-4.06501979e-01 -1.35678005e+00 3.58464904e-02 -5.82169056e-01
-1.80176243e-01 -6.42911196e-02 1.06191766e+00 -5.25236607e-01
-6.67246729e-02 1.25378683e-01 -7.84659684e-01 -6.86481178e-01
-1.31026357e-02 -2.86944099e-02 -1.22165889e-01 -5.02666295e-01
6.53153241e-01 -8.18656981e-01 -2.46320084e-01 -2.71575063e-01
-9.01038706e-01 3.11326087e-01 -4.41684313e-02 6.04885757e-01
2.59285212e-01 3.01371574e-01 7.27888525e-01 -8.50744963e-01
9.16755080e-01 -9.98620629e-01 2.56098092e-01 -1.61310032e-01
-2.26909637e-01 -2.57635057e-01 4.81261432e-01 -6.90392196e-01
-1.45184541e+00 -6.84953034e-01 -1.09434567e-01 3.14613193e-01
-3.05310130e-01 9.21178401e-01 1.02065198e-01 9.32662308e-01
8.50153387e-01 -1.08864322e-01 -9.00739014e-01 -9.03890207e-02
7.87415385e-01 3.07856321e-01 8.01522493e-01 -5.98943353e-01
6.08320057e-01 6.83856547e-01 -2.15472832e-01 -4.14299011e-01
-1.47192574e+00 -3.41077805e-01 -6.74265444e-01 -4.22117442e-01
1.19154692e+00 -7.83708334e-01 -5.24281383e-01 1.15228079e-01
-1.44747019e+00 -3.37616384e-01 -5.10804415e-01 7.64216959e-01
-5.67584634e-01 -2.74029691e-02 -9.26199317e-01 -7.10839987e-01
1.41839743e-01 -4.03701842e-01 6.22272611e-01 1.62753686e-01
-9.80463862e-01 -1.21704066e+00 1.76967010e-01 3.45701993e-01
3.71816568e-02 9.73826706e-01 8.47327769e-01 -8.68146241e-01
-1.89961821e-01 -1.04593851e-01 -3.08108270e-01 -5.59413254e-01
1.26283422e-01 -6.45031929e-02 -1.03766549e+00 5.25554240e-01
-6.85222670e-02 -6.08614028e-01 1.16562033e+00 4.63734448e-01
8.51491213e-01 -3.00808012e-01 -3.90664577e-01 1.62619218e-01
1.13410759e+00 3.54774475e-01 7.89864480e-01 6.92482710e-01
6.75009310e-01 7.32156873e-01 6.36503339e-01 7.28626847e-01
8.06749880e-01 3.81369770e-01 9.77861360e-02 3.36023867e-01
-4.09940630e-01 -4.90399778e-01 5.72326779e-01 5.40761232e-01
-2.93667257e-01 -4.93350446e-01 -1.07236028e+00 1.12644255e+00
-1.89833426e+00 -1.64453363e+00 -6.65790737e-01 1.13269997e+00
1.25338662e+00 3.82648081e-01 1.80141218e-02 5.14189899e-01
4.35537577e-01 8.40958297e-01 -2.16665462e-01 -3.99361491e-01
-4.75489408e-01 3.33871692e-01 2.29049232e-02 3.13870698e-01
-1.29251862e+00 1.09075594e+00 6.39127350e+00 8.87860715e-01
-4.33112353e-01 3.72479767e-01 3.33763599e-01 -4.65559393e-01
-3.93483877e-01 2.14386776e-01 -4.19831306e-01 -6.46547675e-02
9.36762869e-01 -3.70535642e-01 1.23652726e-01 5.02167463e-01
3.22078109e-01 -2.48264998e-01 -1.43864489e+00 7.98497617e-01
1.91934615e-01 -1.73138070e+00 -1.49523392e-01 -4.11103755e-01
6.72427177e-01 -1.94260255e-01 -2.96189338e-01 4.34504390e-01
5.76527297e-01 -1.22546363e+00 9.93402719e-01 5.27980328e-01
4.97446001e-01 -7.82066405e-01 7.56119609e-01 -2.54082214e-03
-1.27616966e+00 1.04066998e-01 1.55014724e-01 -7.53187418e-01
8.06722224e-01 5.51770329e-01 -6.75194740e-01 5.73234737e-01
7.16722369e-01 1.36286092e+00 -4.51446742e-01 5.02775073e-01
-8.17879558e-01 1.11663365e+00 -2.71977093e-02 -2.43720129e-01
1.68293342e-01 4.12280262e-01 7.80884445e-01 1.45529068e+00
-1.09760776e-01 6.80543542e-01 -1.70139194e-01 1.12358451e+00
-1.04325622e-01 -5.33641540e-02 -8.29725385e-01 -5.52833974e-01
4.60178673e-01 9.12784100e-01 -7.88885474e-01 -3.24296981e-01
-4.86787051e-01 7.04908252e-01 5.51283121e-01 1.82947665e-01
-1.18026924e+00 -1.84723452e-01 4.59784180e-01 -1.10901892e-02
1.29709631e-01 -3.49992305e-01 -6.33888602e-01 -1.14943445e+00
-1.25801772e-01 -3.14642280e-01 8.81283581e-01 -8.96838725e-01
-1.66012704e+00 3.40369582e-01 6.22499049e-01 -6.73795938e-01
-2.17407078e-01 -2.39735290e-01 -9.54802811e-01 3.81219178e-01
-1.47506773e+00 -1.22548270e+00 -6.26078201e-03 4.56672281e-01
7.39138722e-01 6.69921935e-02 7.61839032e-01 4.10515293e-02
-6.24323010e-01 2.04586685e-01 -5.91506660e-01 3.83302838e-01
5.64460099e-01 -1.14782083e+00 3.15065205e-01 9.61143672e-01
2.96403736e-01 4.78726953e-01 8.05440485e-01 -8.73229027e-01
-9.26391780e-01 -9.50824440e-01 1.28934455e+00 -6.76961541e-01
1.35808790e+00 -2.01125979e-01 -9.26817477e-01 1.09996498e+00
8.57783079e-01 -5.36766946e-01 1.02675438e+00 6.49278164e-01
-5.32433450e-01 5.07510126e-01 -8.64619613e-01 1.01500607e+00
1.34695840e+00 -5.89609265e-01 -1.36227107e+00 3.60536039e-01
1.10810709e+00 -2.82601833e-01 -1.04498470e+00 2.75698811e-01
3.60167563e-01 -7.16287673e-01 1.04859293e+00 -8.77711952e-01
1.36546528e+00 -7.73074031e-02 -2.67456174e-01 -1.06536031e+00
-2.50641197e-01 -3.85088295e-01 -4.19354647e-01 1.84180784e+00
4.24805373e-01 -3.04030627e-01 2.17030555e-01 3.27008724e-01
-2.27106780e-01 -8.47366452e-01 -8.45421016e-01 -6.54646754e-01
2.21555635e-01 -1.21948171e+00 6.31067514e-01 1.56803119e+00
6.62930131e-01 6.27950668e-01 -2.21689910e-01 8.94595310e-02
7.51159266e-02 8.89749005e-02 1.28732264e-01 -7.79851198e-01
-1.40925214e-01 -6.55534983e-01 -9.40714180e-02 -3.76967221e-01
5.91104329e-01 -9.65784788e-01 -2.72790462e-01 -1.93354952e+00
6.17442489e-01 1.01776980e-01 -2.42315739e-01 5.67551136e-01
-4.29740131e-01 3.28628980e-02 -9.56939608e-02 -2.59151042e-01
-5.30387759e-01 7.09878862e-01 1.10599601e+00 -1.22070312e-01
-1.65739894e-01 -6.21050298e-01 -6.28997266e-01 9.89030302e-01
1.09203637e+00 -6.30972028e-01 -5.63007832e-01 -3.05371284e-01
5.51193893e-01 9.49685723e-02 6.17027283e-01 -5.81095874e-01
1.18164621e-01 -6.61360443e-01 2.11402893e-01 -8.60881925e-01
3.43225986e-01 -2.35863611e-01 -6.91008940e-02 -6.65686429e-02
-8.67945075e-01 7.56051275e-04 4.99959737e-01 5.32390773e-01
-2.95160204e-01 -1.95232317e-01 1.45853609e-01 8.44183192e-02
-9.76683795e-01 -8.47609937e-02 -5.98500848e-01 4.66080606e-01
1.17180169e+00 -1.48148879e-01 -5.55662513e-01 -6.03419840e-01
-5.30319214e-01 2.11194307e-01 -1.20352440e-01 6.35198653e-01
7.83354580e-01 -1.65143371e+00 -1.07056701e+00 -7.22283602e-01
2.92698920e-01 -7.23795742e-02 2.29205593e-01 7.67189741e-01
-1.41338050e-01 3.84956151e-01 1.17275223e-01 7.23709315e-02
-1.00104654e+00 7.53385067e-01 -1.85975984e-01 -5.51576555e-01
-7.60767698e-01 1.06841671e+00 -7.05193952e-02 7.35429972e-02
-3.64413857e-02 -4.89842743e-01 -7.49446988e-01 5.76489151e-01
7.51439989e-01 4.27800745e-01 -4.65114504e-01 -6.11935914e-01
-5.05267084e-01 -1.65632330e-02 1.35831237e-02 -5.08087158e-01
1.78207397e+00 9.43647418e-03 -1.23412378e-01 1.03855920e+00
8.90269816e-01 -4.60714363e-02 -9.01604295e-01 -2.33915031e-01
3.28526914e-01 -2.77551442e-01 1.09085158e-01 -7.63358116e-01
-4.18635637e-01 7.21286178e-01 -2.08117366e-01 3.32046032e-01
9.89202917e-01 3.48351985e-01 9.88793969e-01 5.13634495e-02
1.15541495e-01 -9.26687956e-01 1.56408593e-01 7.54846454e-01
1.28064108e+00 -1.06967533e+00 2.54670262e-01 -4.46501464e-01
-9.04802978e-01 1.17232275e+00 4.95431572e-01 -2.30105028e-01
4.38578188e-01 3.89207065e-01 -5.05043507e-01 -5.77150226e-01
-1.02153838e+00 -5.94632030e-01 4.42410797e-01 4.68767166e-01
7.72274196e-01 5.12095802e-02 -5.92832446e-01 1.12468350e+00
-6.83051467e-01 -3.27436514e-02 5.89250684e-01 8.94220233e-01
-2.60820761e-02 -7.87762344e-01 -3.16204160e-01 3.52142394e-01
-6.31602943e-01 -5.22731304e-01 -5.58983743e-01 1.10418248e+00
2.14165524e-01 1.30884242e+00 5.79390787e-02 -1.48219153e-01
3.49913090e-01 2.87228912e-01 7.50978589e-01 -5.54168940e-01
-6.54398978e-01 -3.97327572e-01 8.81405830e-01 -3.80770087e-01
-8.11501205e-01 -1.03727639e+00 -1.59369147e+00 -5.17682970e-01
-1.53077096e-01 -3.68299425e-01 -6.28795177e-02 1.19105291e+00
-1.62810192e-01 1.14199197e+00 -7.44862296e-03 -5.50059438e-01
3.63928318e-01 -1.09804046e+00 -2.43202016e-01 5.37171304e-01
2.41290289e-03 -8.76203299e-01 -1.32631555e-01 3.76932383e-01] | [10.885753631591797, 8.911446571350098] |
238656a3-6a60-4d3a-8768-9fd5e5671e6d | task-specific-experimental-design-for | 2306.05484 | null | https://arxiv.org/abs/2306.05484v1 | https://arxiv.org/pdf/2306.05484v1.pdf | Task-specific experimental design for treatment effect estimation | Understanding causality should be a core requirement of any attempt to build real impact through AI. Due to the inherent unobservability of counterfactuals, large randomised trials (RCTs) are the standard for causal inference. But large experiments are generically expensive, and randomisation carries its own costs, e.g. when suboptimal decisions are trialed. Recent work has proposed more sample-efficient alternatives to RCTs, but these are not adaptable to the downstream application for which the causal effect is sought. In this work, we develop a task-specific approach to experimental design and derive sampling strategies customised to particular downstream applications. Across a range of important tasks, real-world datasets, and sample sizes, our method outperforms other benchmarks, e.g. requiring an order-of-magnitude less data to match RCT performance on targeted marketing tasks. | ['Christopher Frye', 'Ilya Feige', 'Gary Willis', 'Alexander Adam', 'Tobias Schwedes', 'Kim Moore', 'Bethany Connolly'] | 2023-06-08 | null | null | null | null | ['causal-inference', 'experimental-design', 'marketing', 'causal-inference'] | ['knowledge-base', 'methodology', 'miscellaneous', 'miscellaneous'] | [ 5.26823878e-01 2.15946823e-01 -1.01674843e+00 -3.84902954e-01
-7.71950603e-01 -6.78620219e-01 8.86848569e-01 2.92954922e-01
-5.44424713e-01 9.94258225e-01 6.69416785e-01 -1.03182387e+00
-5.57753980e-01 -6.49044514e-01 -1.07181323e+00 -3.30860287e-01
-4.12191689e-01 4.74064589e-01 -2.88725406e-01 1.13978237e-01
5.17900288e-01 2.89080143e-01 -1.34189594e+00 3.21427554e-01
7.72858143e-01 6.25898719e-01 -1.90380022e-01 4.78205383e-01
2.85346478e-01 4.78369325e-01 -4.78054494e-01 -4.12375689e-01
1.11408725e-01 -5.13140023e-01 -7.87341654e-01 -3.37811857e-01
4.09457833e-01 -4.23869997e-01 5.04739722e-03 6.91130757e-01
8.44061077e-01 -9.01565999e-02 7.81524003e-01 -1.40257502e+00
-6.71460032e-01 1.24221969e+00 -6.49731517e-01 1.97579011e-01
3.97128999e-01 4.46450621e-01 1.20593131e+00 -2.14925840e-01
4.99786139e-01 1.65945792e+00 4.99150723e-01 3.14828902e-01
-1.80299282e+00 -9.92141366e-01 3.89835119e-01 -2.99854837e-02
-5.99082112e-01 -6.24206424e-01 3.21019530e-01 -4.45696205e-01
7.44564176e-01 3.56190443e-01 4.83466566e-01 1.83684230e+00
4.68447864e-01 6.01595581e-01 1.56251764e+00 -2.73445398e-01
6.65414929e-01 -1.85520664e-01 -2.21489131e-01 -1.38519760e-02
6.87236249e-01 7.99183726e-01 -4.14102703e-01 -4.49255466e-01
6.42470658e-01 -2.39305347e-02 -2.45793864e-01 -5.21282852e-01
-1.49768019e+00 1.36340821e+00 4.18343484e-01 -8.06875378e-02
-5.94474792e-01 4.98655111e-01 5.33883870e-01 3.97428185e-01
3.84456992e-01 8.04850340e-01 -8.20382535e-01 -1.05513021e-01
-7.95714438e-01 8.17774594e-01 8.29954088e-01 6.61324441e-01
1.03248283e-01 -3.87207896e-01 -5.43956637e-01 4.56253320e-01
2.49833494e-01 4.79383767e-01 5.21039069e-01 -1.05264914e+00
5.61795413e-01 2.99431622e-01 3.48336756e-01 -6.14695311e-01
-5.70139647e-01 -2.92671978e-01 -7.77958333e-01 8.23556632e-02
6.58333540e-01 -3.68211269e-01 -9.27257955e-01 1.84681463e+00
4.57282305e-01 -1.41655421e-02 -3.28971863e-01 8.69656742e-01
4.91278887e-01 1.39249742e-01 7.52407908e-01 -3.60788107e-01
1.41839039e+00 -2.30804443e-01 -6.22917831e-01 -4.85843152e-01
7.89412379e-01 -7.64594018e-01 1.19201970e+00 3.92643690e-01
-8.65824997e-01 1.57962441e-01 -8.04459512e-01 2.51730025e-01
-2.63331234e-01 -5.46731591e-01 1.12837207e+00 9.83456492e-01
-4.71208006e-01 7.85839498e-01 -2.13237181e-01 -1.12274297e-01
7.67426968e-01 3.29197794e-01 -6.39698580e-02 -1.06873594e-01
-1.35889053e+00 8.72917652e-01 1.60215095e-01 -1.20860059e-03
-1.13851333e+00 -1.42859888e+00 -6.18506074e-01 1.27490580e-01
8.34140718e-01 -9.29149687e-01 1.58946133e+00 -5.43594658e-01
-1.18102729e+00 4.67684031e-01 2.95696050e-01 -5.74398279e-01
7.30599821e-01 -2.55719237e-02 -1.35998219e-01 -5.31692088e-01
3.51742029e-01 5.67671895e-01 6.84885442e-01 -9.40856576e-01
-7.65907168e-01 -3.90124917e-01 2.97334909e-01 1.91068292e-01
1.56527817e-01 3.55256438e-01 3.63711506e-01 -7.34243155e-01
-4.49253231e-01 -9.45023596e-01 -7.62406051e-01 -5.24901092e-01
-4.30813521e-01 -3.71424377e-01 2.38737538e-01 -1.24099858e-01
1.23326814e+00 -1.68386948e+00 -2.00481504e-01 -2.22272314e-02
2.19811112e-01 -1.88338026e-01 -1.17544934e-01 4.69516337e-01
-4.31184381e-01 5.42622209e-01 -1.03751272e-01 3.59966904e-01
2.83369839e-01 -2.59188712e-01 -3.57330978e-01 4.97195870e-01
1.73538640e-01 1.10746598e+00 -9.94579792e-01 -3.72118831e-01
1.09284185e-01 -6.22634590e-02 -7.72017062e-01 -1.44994617e-01
-4.03131336e-01 2.52061635e-01 -6.11293793e-01 2.47921303e-01
3.69006544e-01 -2.11169854e-01 4.42791760e-01 1.13801979e-01
-9.73896906e-02 9.55342114e-01 -1.26746845e+00 1.30022407e+00
-5.84409177e-01 1.81247264e-01 -2.08069414e-01 -1.20359313e+00
3.16146135e-01 3.55377913e-01 3.84456575e-01 -8.59926999e-01
1.49179935e-01 2.16611370e-01 5.18230379e-01 -4.27985072e-01
1.36430077e-02 -5.74578047e-01 -3.57527345e-01 7.28414595e-01
-3.92283440e-01 -3.90351117e-02 1.40812308e-01 -1.01199389e-01
1.34046686e+00 1.26843169e-01 7.63618648e-01 -5.22434771e-01
-3.08365226e-01 3.70546222e-01 5.36475003e-01 1.03476024e+00
-1.14816902e-02 2.27603778e-01 7.94699550e-01 -4.34953213e-01
-9.84060407e-01 -7.18171537e-01 -4.61849272e-01 1.01016319e+00
-4.92753327e-01 8.96262899e-02 -4.77891684e-01 -7.71175981e-01
4.66455847e-01 1.12150300e+00 -8.61565113e-01 -1.78663850e-01
-3.92094195e-01 -1.25200772e+00 4.10287023e-01 2.77118921e-01
6.00254722e-02 -1.22058141e+00 -8.19687247e-01 4.48006123e-01
-4.46564518e-02 -7.02273548e-01 -3.76143634e-01 1.20998397e-01
-1.17377687e+00 -1.28394258e+00 -6.81287169e-01 6.39978796e-02
1.38571545e-01 1.64288908e-01 1.29512596e+00 -4.10711884e-01
-2.28884906e-01 4.41285297e-02 2.80320607e-02 -9.06276762e-01
-4.09718156e-01 -1.58449724e-01 -6.49140552e-02 -3.19009572e-01
5.66736341e-01 -4.55180943e-01 -1.06699789e+00 2.42853999e-01
-7.69811988e-01 -2.80267715e-01 9.67162549e-01 8.50999534e-01
1.83313712e-01 -1.01865925e-01 1.00613606e+00 -1.34401333e+00
1.11449432e+00 -7.61921406e-01 -6.38483286e-01 5.70101812e-02
-1.00266278e+00 1.12888575e-01 3.55046004e-01 -8.27529550e-01
-1.01076031e+00 -3.73707861e-01 5.20052493e-01 1.98888376e-01
-2.97173530e-01 8.56172800e-01 -1.38120607e-01 4.69187647e-01
1.17118025e+00 -7.77589083e-01 -2.76757572e-02 -3.86793971e-01
5.45204401e-01 5.99424720e-01 -2.34386399e-01 -6.89132035e-01
2.28285044e-01 2.56002963e-01 1.62976414e-01 -2.17864096e-01
-7.27372050e-01 4.89192419e-02 -4.83595021e-02 1.49105832e-01
5.28681338e-01 -9.68906820e-01 -1.21233714e+00 -1.32098243e-01
-7.61413038e-01 -8.24491262e-01 -4.07996148e-01 9.07269716e-01
-6.30073726e-01 -1.45518824e-01 -3.07400882e-01 -6.64152622e-01
-1.51956290e-01 -1.18071878e+00 8.94850492e-01 -8.58828202e-02
-6.86171651e-01 -7.90645719e-01 1.62945405e-01 2.77133554e-01
2.49292344e-01 3.60906154e-01 1.02225697e+00 -4.53404456e-01
-4.35492128e-01 -1.57943845e-01 -2.62228966e-01 -1.84304878e-01
7.94457570e-02 -9.93152265e-04 -8.51233423e-01 -1.02720775e-01
-3.51937972e-02 -3.85758311e-01 7.28417218e-01 1.25952387e+00
1.20218217e+00 -6.99687004e-01 -5.73388100e-01 -3.82589437e-02
1.14507747e+00 2.90703058e-01 5.81530631e-01 5.11197150e-01
3.73105049e-01 1.06226039e+00 9.62695599e-01 2.92613685e-01
4.70170677e-01 6.90005124e-01 1.42184287e-01 -1.06672868e-01
6.78020567e-02 -3.58745217e-01 2.52220988e-01 -2.12840363e-01
-5.41366227e-02 -5.74490316e-02 -8.04072678e-01 7.43824005e-01
-1.88322866e+00 -9.72999930e-01 -5.73198140e-01 2.64062929e+00
1.14394510e+00 3.78577501e-01 4.62089628e-01 -8.11441168e-02
4.79772955e-01 -8.01275820e-02 -6.14284396e-01 -7.82557845e-01
1.83950409e-01 2.81284600e-01 1.02783525e+00 8.13633427e-02
-8.95944417e-01 4.95272249e-01 7.20947933e+00 7.69627810e-01
-8.47330630e-01 7.92623535e-02 8.82624209e-01 -3.65751952e-01
-6.10675156e-01 4.11073387e-01 -4.81717825e-01 5.23876965e-01
1.38126934e+00 -4.16050464e-01 2.47440606e-01 6.33758903e-01
7.44059682e-01 -2.24696144e-01 -1.57880485e+00 3.54785770e-01
-6.90285623e-01 -1.28619468e+00 -2.62019753e-01 4.41834807e-01
7.60189474e-01 -1.22842833e-01 1.04231939e-01 1.70270503e-01
9.91836727e-01 -1.37493622e+00 7.61897564e-01 -6.09997585e-02
8.69226456e-01 -5.81891000e-01 6.01553261e-01 2.54067689e-01
-2.37692758e-01 -3.63759756e-01 -3.08576167e-01 -4.79228079e-01
5.43625876e-02 9.13249671e-01 -8.90456140e-01 2.49111220e-01
9.05999660e-01 4.89035755e-01 -2.67489254e-01 9.64403152e-01
-3.49379987e-01 9.85415280e-01 -2.32094511e-01 -2.70056635e-01
2.10542649e-01 6.22548424e-02 2.10658669e-01 1.07254922e+00
1.25253469e-01 1.39113769e-01 -3.08707774e-01 8.31286550e-01
-1.03900708e-01 1.98450625e-01 -9.71005499e-01 -3.45257580e-01
5.64801931e-01 9.08462703e-01 -5.39638400e-01 -1.66200012e-01
-4.23090488e-01 1.20285928e-01 5.49398176e-03 3.50623250e-01
-7.24098742e-01 8.38579610e-02 6.98864460e-01 1.00129321e-01
1.98592976e-01 3.26785982e-01 -6.43645644e-01 -7.56575108e-01
-2.57349253e-01 -1.53843236e+00 6.27516389e-01 -5.08811295e-01
-1.31784713e+00 -3.44034702e-01 4.79691505e-01 -9.66737807e-01
-2.98828453e-01 -3.89796704e-01 -4.22992170e-01 8.90822291e-01
-1.20190430e+00 -7.28149891e-01 4.00878847e-01 1.59355015e-01
3.81375581e-01 3.18611652e-01 7.91422069e-01 1.72859550e-01
-5.42916238e-01 3.64389360e-01 -1.94057330e-01 -3.97524029e-01
9.57230687e-01 -1.29620993e+00 4.62741613e-01 5.01276195e-01
-2.16654077e-01 7.47609496e-01 1.07365668e+00 -8.78227055e-01
-1.36696017e+00 -6.29691422e-01 9.59658623e-01 -5.55309594e-01
1.14249337e+00 -3.90828788e-01 -4.44286972e-01 8.09727907e-01
1.10754482e-01 -6.77127242e-01 6.34458125e-01 9.00346994e-01
-3.80716532e-01 -1.16431132e-01 -9.72654283e-01 1.06039786e+00
1.28889382e+00 -3.19222845e-02 -5.61642766e-01 4.76644844e-01
8.55840683e-01 -2.43648753e-01 -1.16182804e+00 5.66216767e-01
6.85725033e-01 -6.72785401e-01 1.07281244e+00 -1.06480908e+00
9.25082564e-01 1.67181805e-01 2.94797748e-01 -1.48099804e+00
-2.68188626e-01 -7.08520353e-01 3.13876480e-01 1.09505415e+00
7.84635305e-01 -7.85328686e-01 5.84532440e-01 8.56846094e-01
3.46734166e-01 -4.82134342e-01 -6.56988084e-01 -5.55003047e-01
3.37059349e-01 -5.44861615e-01 9.17232096e-01 1.27766812e+00
-1.25811309e-01 7.05033541e-01 -2.50326216e-01 -5.10520712e-02
9.95399356e-01 4.30921048e-01 7.52694905e-01 -1.22310948e+00
-2.94299185e-01 -7.56359875e-01 1.00550845e-01 -5.29802322e-01
-1.87048346e-01 -5.12236297e-01 -7.94975534e-02 -1.33377528e+00
3.27752411e-01 -8.26311529e-01 -1.51785731e-01 3.82184923e-01
-3.65080595e-01 -2.72729397e-01 -9.33199301e-02 -8.13070759e-02
2.76847035e-02 2.96301275e-01 1.26380801e+00 -8.28443617e-02
-4.00306344e-01 4.20438707e-01 -1.36425328e+00 4.82451975e-01
9.22172070e-01 -1.01719952e+00 -6.38112664e-01 -3.76373082e-02
6.16106331e-01 1.85312390e-01 6.15050435e-01 -1.39404938e-01
-1.68724984e-01 -8.37117970e-01 2.46550500e-01 -1.34263918e-01
-3.68445307e-01 -6.15390897e-01 3.98288399e-01 5.51575720e-01
-8.08323383e-01 2.81191692e-02 1.96794301e-01 7.73432851e-01
2.60493785e-01 -1.81296632e-01 2.13858664e-01 -1.94784239e-01
-1.84336111e-01 7.04443976e-02 -2.48210862e-01 2.89167553e-01
8.75152886e-01 1.00067362e-01 -5.67938328e-01 -2.11840019e-01
-1.54170185e-01 4.14953887e-01 1.67309105e-01 4.78137702e-01
2.37269953e-01 -1.10742009e+00 -9.83226717e-01 -5.20697713e-01
1.32745951e-01 -1.17392801e-01 2.42570907e-01 1.05535638e+00
-9.88658424e-03 6.65434301e-01 -7.88113102e-03 -3.13718647e-01
-1.00669968e+00 1.01449931e+00 -2.38190293e-02 -4.33373928e-01
-6.20515347e-01 4.99836922e-01 5.20270944e-01 -4.08354580e-01
7.22334683e-02 -5.40113807e-01 -1.49626955e-01 1.81434169e-01
5.29759467e-01 4.05309707e-01 -9.98981763e-03 3.24515253e-01
-2.47339860e-01 3.79774463e-03 -6.81914091e-02 -3.11092347e-01
1.48195922e+00 3.68981659e-02 -3.17808948e-02 4.13172275e-01
8.11799884e-01 -2.63291657e-01 -1.24804163e+00 4.55161147e-02
3.43067110e-01 -4.76360381e-01 3.46616507e-01 -1.01900566e+00
-6.00331247e-01 5.53469062e-01 2.97680527e-01 4.34307516e-01
9.82990026e-01 -5.71322516e-02 5.51701263e-02 -9.26482603e-02
3.74911934e-01 -1.06259060e+00 -4.51886892e-01 -4.35709566e-01
9.96573567e-01 -1.39184284e+00 5.17008245e-01 -1.22356147e-01
-7.03530371e-01 3.89745384e-01 1.79894399e-02 -2.81990021e-02
5.81972599e-01 1.25828600e-02 -3.36316228e-01 -4.80606318e-01
-1.02816832e+00 -1.06465518e-01 2.64807969e-01 4.87258136e-01
6.21610463e-01 6.52236879e-01 -1.07404447e+00 3.77090961e-01
-2.01870054e-01 2.90309548e-01 5.39249063e-01 7.71484733e-01
1.85106233e-01 -1.10482693e+00 -4.42647278e-01 8.77222240e-01
-7.72961974e-01 -2.09403142e-01 -4.14127588e-01 1.17879164e+00
-2.49936894e-01 1.19829762e+00 -1.59953192e-01 1.22651860e-01
6.33566439e-01 -2.13389680e-01 4.57653999e-01 -5.66850543e-01
-5.72535157e-01 1.67143926e-01 6.03111148e-01 -9.05048132e-01
-7.14201510e-01 -9.48515058e-01 -7.75664389e-01 -7.76197195e-01
-2.56423414e-01 -1.18476085e-01 7.26086199e-01 8.02542090e-01
3.73093635e-01 4.26445454e-01 5.23080647e-01 -5.13178825e-01
-1.05524170e+00 -1.02883983e+00 -3.97714913e-01 3.62439007e-01
1.95807874e-01 -8.62493336e-01 -3.60287130e-01 -2.61798263e-01] | [8.076125144958496, 5.3788652420043945] |
2291b0dd-a845-42cc-b4b8-17ad69fc80ec | youngsheldon-at-semeval-2021-task-5-fine | null | null | https://aclanthology.org/2021.semeval-1.130 | https://aclanthology.org/2021.semeval-1.130.pdf | YoungSheldon at SemEval-2021 Task 5: Fine-tuning Pre-trained Language Models for Toxic Spans Detection using Token classification Objective | In this paper, we describe our system used for SemEval 2021 Task 5: Toxic Spans Detection. Our proposed system approaches the problem as a token classification task. We trained our model to find toxic words and concatenate their spans to predict the toxic spans within a sentence. We fine-tuned Pre-trained Language Models (PLMs) for identifying the toxic words. For fine-tuning, we stacked the classification layer on top of the PLM features of each word to classify if it is toxic or not. PLMs are pre-trained using different objectives and their performance may differ on downstream tasks. We, therefore, compare the performance of BERT, ELECTRA, RoBERTa, XLM-RoBERTa, T5, XLNet, and MPNet for identifying toxic spans within a sentence. Our best performing system used RoBERTa. It performed well, achieving an F1 score of 0.6841 and secured a rank of 16 on the official leaderboard. | ['W.b. Vasantha', 'Ilanthenral Kandasamy', 'Mayukh Sharma'] | 2021-08-01 | null | null | null | semeval-2021 | ['toxic-spans-detection'] | ['natural-language-processing'] | [-3.35617691e-01 -2.05504358e-01 -1.08149379e-01 -5.49943782e-02
-1.04502809e+00 -6.45727634e-01 6.85388684e-01 6.70964956e-01
-7.65910983e-01 9.14003968e-01 3.59375894e-01 -3.65106732e-01
6.49564564e-02 -6.67948306e-01 -4.37785953e-01 -3.89337093e-01
-2.71709591e-01 2.28658184e-01 -1.40050352e-01 -2.79072791e-01
8.32417846e-01 5.16732693e-01 -9.55171466e-01 6.35383666e-01
9.55093026e-01 8.75292063e-01 1.59870192e-01 8.62264812e-01
-1.05461448e-01 1.02601910e+00 -1.18086541e+00 -3.16701084e-01
-5.01422696e-02 -2.60446936e-01 -1.08764958e+00 -5.18746912e-01
1.67240784e-01 1.72016054e-01 -2.24815413e-01 7.07812607e-01
5.40652215e-01 1.48703447e-02 8.83015275e-01 -8.93906415e-01
-6.06418967e-01 1.20783961e+00 -5.55879533e-01 5.05351007e-01
4.03219312e-01 2.32632980e-01 1.32368898e+00 -8.76802027e-01
6.35401011e-01 1.23559964e+00 6.84287667e-01 5.36189497e-01
-9.47153032e-01 -7.29344189e-01 1.41616479e-01 4.47561353e-01
-1.39678013e+00 -3.17986220e-01 2.60529518e-01 -7.57621467e-01
1.38336241e+00 2.11911425e-01 4.81734037e-01 1.30284917e+00
7.27173448e-01 6.90431297e-01 1.12490392e+00 -2.77396351e-01
1.26145422e-01 -6.23766556e-02 4.89196271e-01 6.48139536e-01
5.55001618e-03 -2.96464980e-01 -1.11859322e+00 -1.58212841e-01
-1.90578416e-01 -4.00601685e-01 -9.20471177e-02 8.47733855e-01
-1.17726505e+00 8.31918716e-01 2.21130967e-01 5.47358513e-01
-5.16889215e-01 2.31613427e-01 8.63425791e-01 3.90595883e-01
7.72536218e-01 8.73926222e-01 -5.85135698e-01 -2.50351369e-01
-1.16057980e+00 1.97758049e-01 6.68526530e-01 6.13911629e-01
3.22850674e-01 -1.23579614e-01 -1.05151772e+00 7.64360309e-01
4.19149697e-02 1.47114590e-01 4.35826808e-01 -5.48463762e-01
6.72796428e-01 2.89792627e-01 5.73966764e-02 -5.42782843e-01
-7.77836382e-01 -4.72850055e-01 -5.58377206e-01 -1.24794416e-01
3.73855501e-01 -2.96183705e-01 -9.39872921e-01 1.58872128e+00
-3.86014611e-01 -1.25487790e-01 -5.41956089e-02 4.70072776e-01
5.24053574e-01 8.29968154e-01 6.61476552e-01 -1.69418260e-01
1.29106510e+00 -8.28202546e-01 -5.05419314e-01 -8.63661468e-02
1.00804532e+00 -8.75587165e-01 1.03203869e+00 7.16419995e-01
-1.02012646e+00 -3.86728197e-01 -1.02028906e+00 -5.56027032e-02
-6.07013643e-01 2.12139592e-01 1.22850925e-01 3.68197978e-01
-9.53389108e-01 1.01860988e+00 -2.83078939e-01 -3.79798919e-01
2.34511733e-01 2.08068460e-01 -1.78553045e-01 2.47278139e-01
-1.68302989e+00 1.36861348e+00 4.76259619e-01 -1.98421434e-01
-1.30681098e+00 -7.45987654e-01 -6.76885068e-01 1.14322856e-01
-8.44419524e-02 -5.35658121e-01 1.09925985e+00 -1.87771499e-01
-1.12025297e+00 1.01760101e+00 -6.33532256e-02 -7.78386295e-01
4.08823609e-01 -5.09319246e-01 -4.67577934e-01 5.09710461e-02
4.12338227e-01 4.38467413e-01 4.27325964e-01 -7.03431249e-01
-6.28052592e-01 -1.79154098e-01 -8.36211294e-02 -1.21142738e-01
-3.64787072e-01 5.74241042e-01 9.14393216e-02 -5.21919370e-01
-5.60039639e-01 -4.28874731e-01 -1.75247848e-01 -7.20760107e-01
-9.85412121e-01 -7.90886760e-01 2.86569923e-01 -8.96709979e-01
1.63655519e+00 -1.96747613e+00 -1.07342727e-01 1.76361471e-01
1.30254671e-01 2.00625464e-01 -4.62271988e-01 8.41569126e-01
-6.77648112e-02 5.27321815e-01 5.16723022e-02 -4.58204091e-01
6.84025362e-02 -2.47917101e-01 -3.21908832e-01 5.44607222e-01
2.66358942e-01 5.63056767e-01 -9.51439559e-01 -6.29870176e-01
-9.01297927e-02 1.03365004e-01 -1.54638961e-01 1.77638829e-01
-3.02302510e-01 8.48972499e-02 -3.12326819e-01 6.95392609e-01
4.02867317e-01 2.65253484e-01 -1.35842249e-01 1.64342776e-01
-5.84181368e-01 7.65207767e-01 -3.77946019e-01 1.58119071e+00
-6.80778861e-01 6.38258636e-01 -2.30109960e-01 -7.50301123e-01
1.07663167e+00 2.95331925e-01 2.93149889e-01 -6.40361130e-01
1.39471129e-01 1.20464191e-01 -1.27798775e-02 -6.48732126e-01
7.51560092e-01 -2.06621587e-01 -5.84238827e-01 3.19105148e-01
3.26136611e-02 2.38747761e-01 7.58687496e-01 3.96407932e-01
1.55706167e+00 -2.33348832e-01 1.61243588e-01 -4.60403800e-01
5.71020365e-01 -8.54478851e-02 5.92407286e-01 7.04607964e-01
-2.42029727e-01 5.06740272e-01 9.42021608e-01 -9.71155763e-02
-8.72480154e-01 -1.00003862e+00 -2.13303734e-02 1.31259894e+00
-3.99950594e-01 -9.41724598e-01 -7.51216114e-01 -8.29604268e-01
-1.09628798e-03 1.28549206e+00 -6.84343696e-01 -1.52285933e-01
-4.31130499e-01 -5.58067083e-01 1.13506818e+00 4.23936963e-01
1.98231414e-01 -1.47189438e+00 -4.22646642e-01 2.91311771e-01
-2.64066786e-01 -7.18734682e-01 -5.91607988e-01 6.45861804e-01
-2.82421529e-01 -9.74605978e-01 -4.24669415e-01 -5.45556605e-01
9.13736746e-02 -3.96429539e-01 1.17926335e+00 -3.23476419e-02
-2.72791505e-01 -3.44018847e-01 -4.20441538e-01 -5.36231518e-01
-3.51624459e-01 3.60464096e-01 2.47870296e-01 -3.38988572e-01
2.82557696e-01 -1.57359898e-01 -3.28405559e-01 -1.38100848e-01
-4.22942549e-01 -3.86826336e-01 4.73601639e-01 5.43021858e-01
2.25489780e-01 -2.33452305e-01 1.03173900e+00 -8.14947426e-01
1.03482854e+00 -6.33367181e-01 -1.45266622e-01 2.68154889e-01
-5.03660798e-01 -6.35518581e-02 9.63869870e-01 -2.26943761e-01
-6.79648876e-01 -1.33534372e-01 -5.67601144e-01 -5.71371950e-02
-2.73432489e-02 6.34464025e-01 -1.92626640e-01 4.86323208e-01
6.19519174e-01 2.96565462e-02 -5.82649171e-01 -7.29217052e-01
4.45571125e-01 9.84497845e-01 4.22583848e-01 -4.88650024e-01
4.75402921e-01 -1.43626437e-01 -2.55002320e-01 -7.47413754e-01
-1.14292562e+00 -6.14416301e-01 -6.11738384e-01 -1.93961367e-01
9.81675446e-01 -8.73861194e-01 -7.60369182e-01 5.07769227e-01
-1.59883606e+00 -2.84003854e-01 -9.25786570e-02 2.96169907e-01
-3.05775434e-01 2.24201679e-01 -9.53344285e-01 -7.99676418e-01
-1.01243603e+00 -6.36200905e-01 8.11320901e-01 -1.98624213e-03
-8.04783702e-01 -8.84340703e-01 2.14619741e-01 2.98995823e-01
2.29108974e-01 1.02344222e-01 1.15956020e+00 -9.57538188e-01
2.62644768e-01 -6.82457462e-02 -1.06529228e-01 3.88884634e-01
-2.54586458e-01 2.14722112e-01 -8.80411983e-01 -2.68230285e-03
-3.47618610e-01 -5.51513314e-01 1.38472104e+00 1.93126768e-01
1.35268068e+00 -3.90582055e-01 -2.81943113e-01 1.93320572e-01
1.12625289e+00 -1.75060201e-02 5.52826226e-01 6.48685098e-01
3.70562881e-01 5.88684380e-01 7.92651236e-01 7.40340114e-01
5.68490252e-02 2.58092165e-01 3.06487769e-01 3.30607623e-01
-5.05442806e-02 -3.05945843e-01 1.00682688e+00 6.87477469e-01
3.10152560e-01 -6.21553004e-01 -1.00225294e+00 8.38769734e-01
-1.52584386e+00 -1.01994622e+00 -4.08015013e-01 1.67731333e+00
1.05567825e+00 4.57292587e-01 2.02613100e-01 2.07183182e-01
5.64086437e-01 3.08377653e-01 -2.33487874e-01 -1.25013888e+00
-1.19832888e-01 3.00532758e-01 4.86609906e-01 3.40546668e-01
-9.60994840e-01 1.14529574e+00 7.08390522e+00 1.22497857e+00
-1.03768551e+00 1.12053588e-01 6.72480226e-01 -3.53954643e-01
-1.23710901e-01 -3.39572996e-01 -9.23421025e-01 6.51648223e-01
1.52311623e+00 -3.42416048e-01 9.17230919e-02 5.22141874e-01
7.39695311e-01 -2.69066542e-01 -1.15260875e+00 4.58321571e-01
1.56436905e-01 -1.24055457e+00 -9.84476041e-03 -8.79981369e-02
4.89065021e-01 3.28694254e-01 -4.71373610e-02 5.31447649e-01
2.60701478e-01 -1.37076104e+00 9.92186427e-01 6.69519663e-01
4.92110133e-01 -1.03291142e+00 6.92974269e-01 4.79158551e-01
-7.91056693e-01 -4.00632024e-02 -4.90205497e-01 -2.59803087e-01
7.00622201e-02 9.79421377e-01 -7.68751740e-01 4.55791026e-01
5.73049664e-01 7.86633849e-01 -5.34722745e-01 1.07519984e+00
-6.10272765e-01 1.06300735e+00 -2.24228725e-02 -1.71912417e-01
4.38814312e-01 1.79180682e-01 4.12568927e-01 1.80176854e+00
2.61820734e-01 -2.65125781e-01 6.24080375e-02 7.60204971e-01
-4.82004702e-01 3.86437595e-01 -4.01004612e-01 -2.08183050e-01
5.02212584e-01 1.42706800e+00 -4.54455316e-01 -4.63074654e-01
-7.10826553e-03 7.88250625e-01 4.63454157e-01 -6.71237260e-02
-8.19032133e-01 -7.61138916e-01 6.54737771e-01 1.65370740e-02
4.58787940e-02 -1.12689778e-01 -7.42581844e-01 -6.94904447e-01
-3.44731569e-01 -5.52950263e-01 5.79405129e-01 -6.64018035e-01
-1.53828204e+00 7.30281293e-01 -3.81777763e-01 -8.14400434e-01
6.67793006e-02 -6.30957365e-01 -1.17764437e+00 1.15296733e+00
-1.28001332e+00 -8.56589079e-01 2.48385921e-01 2.31449768e-01
6.81147993e-01 -1.02182575e-01 8.06958973e-01 1.59258336e-01
-7.70757318e-01 6.59254193e-01 8.32190067e-02 2.64378607e-01
9.78858531e-01 -1.29675984e+00 4.47816581e-01 7.11801291e-01
-1.33659080e-01 5.13267159e-01 6.98872626e-01 -7.84645736e-01
-7.59496987e-01 -1.47427881e+00 1.72582197e+00 -3.81618351e-01
1.06497407e+00 -4.15272593e-01 -6.87357247e-01 3.32090825e-01
5.74173212e-01 -7.03445256e-01 6.68635309e-01 3.73747528e-01
-5.15772998e-01 1.49346337e-01 -1.06005847e+00 2.30331182e-01
8.26660514e-01 -5.70915580e-01 -6.58343613e-01 6.29696071e-01
6.93365872e-01 2.09610328e-01 -1.04477382e+00 -5.84909171e-02
1.61857203e-01 -6.91860318e-01 6.48182869e-01 -7.90662587e-01
8.43403637e-01 5.49899274e-03 9.01712701e-02 -1.65189588e+00
-4.08826143e-01 -4.75287497e-01 1.32290363e-01 1.50685203e+00
7.80314505e-01 -2.36472115e-01 3.72274011e-01 1.50589526e-01
-4.43054765e-01 -9.61469710e-01 -8.09889615e-01 -1.01773024e+00
5.89289606e-01 -4.68198895e-01 2.26981804e-01 6.79945111e-01
6.25925124e-01 6.93805337e-01 -5.20675659e-01 -2.17839465e-01
3.82383734e-01 3.50216143e-02 -4.68960479e-02 -9.75237846e-01
9.35291722e-02 -6.45633221e-01 2.54283994e-02 -4.37983066e-01
5.80582261e-01 -1.26938128e+00 2.69777566e-01 -1.63909447e+00
4.60747123e-01 -9.05994847e-02 -6.30749404e-01 7.81851590e-01
-1.95936322e-01 2.54348725e-01 3.19819480e-01 -9.94706228e-02
-7.65128970e-01 3.57209265e-01 7.92197287e-01 -3.94304603e-01
-4.64554206e-02 -1.28542095e-01 -8.27793658e-01 6.25338376e-01
1.32931054e+00 -6.16272569e-01 2.72102743e-01 -3.23713720e-01
2.55979866e-01 -2.65017837e-01 -5.02662435e-02 -9.81828272e-01
7.53695071e-02 -1.60786718e-01 4.39673096e-01 -9.96151924e-01
-2.10029469e-03 1.32915020e-01 -4.04642880e-01 7.86561668e-01
-6.89319491e-01 1.27860561e-01 1.37041301e-01 -2.19789594e-02
-6.15541488e-02 -4.94998366e-01 7.02451110e-01 -9.07519609e-02
-4.94910717e-01 9.03280154e-02 -9.48056281e-01 1.94564730e-01
8.66188228e-01 3.40928674e-01 -4.58406061e-01 -4.17024456e-02
-6.29382312e-01 5.52651286e-01 1.54079452e-01 3.40398133e-01
6.55093670e-01 -8.49709928e-01 -1.11274123e+00 -4.30695355e-01
1.70200199e-01 -5.84408462e-01 1.39249906e-01 7.69160211e-01
-4.46989626e-01 5.81974208e-01 -1.83039218e-01 7.03005940e-02
-1.36845803e+00 5.79628170e-01 1.16908468e-01 -7.54102528e-01
-3.16544324e-01 9.10792172e-01 -2.14338124e-01 -1.49569049e-01
2.11506486e-01 -1.45269468e-01 -5.28395474e-01 5.90175986e-01
6.19808137e-01 7.37475693e-01 1.63151160e-01 -5.77477455e-01
-6.50129437e-01 3.52001250e-01 -1.11388542e-01 -5.53505868e-03
1.26290190e+00 1.84861317e-01 -4.10935849e-01 5.21365047e-01
1.26460230e+00 1.58832565e-01 -4.84603256e-01 1.43545493e-01
5.01474082e-01 7.67908469e-02 -2.57176366e-02 -1.23286498e+00
-6.31870806e-01 9.61032748e-01 2.80619971e-02 2.14389712e-02
8.18567336e-01 4.90001356e-03 1.02990317e+00 3.19814950e-01
4.96743359e-02 -1.46182573e+00 1.40793473e-01 1.09294760e+00
9.11909640e-01 -6.06604278e-01 -9.53742415e-02 3.91405486e-02
-6.81120574e-01 1.07746816e+00 5.14259875e-01 -1.77362397e-01
2.39676759e-01 2.10134342e-01 6.36498109e-02 -1.54418975e-01
-1.19702244e+00 -2.34330505e-01 1.02100506e-01 2.63195276e-01
7.46761560e-01 2.14907423e-01 -9.47263896e-01 8.04179788e-01
-4.97150481e-01 -2.70333141e-01 6.50298238e-01 5.18378139e-01
-7.96343565e-01 -9.93189752e-01 -3.36222798e-01 6.36569500e-01
-9.32581425e-01 -4.25216496e-01 -7.93263555e-01 4.34945196e-01
3.62068206e-01 1.30196357e+00 -1.30568564e-01 -7.36493289e-01
3.45290631e-01 3.78423780e-01 1.65172771e-01 -8.01371098e-01
-1.25498462e+00 -1.88771442e-01 6.13875806e-01 -3.83848786e-01
2.16220945e-01 -7.47337043e-01 -1.49946058e+00 -5.20668745e-01
-6.97888583e-02 5.58800995e-01 5.81620336e-01 9.68370914e-01
8.62638727e-02 6.20301247e-01 7.22157538e-01 -4.62587357e-01
-7.95980930e-01 -1.38667929e+00 -8.50715995e-01 1.85519695e-01
7.75602981e-02 -1.44458428e-01 -2.33443648e-01 -1.68761447e-01] | [8.93812084197998, 10.647504806518555] |
a3c1c1dd-3ba8-4121-9040-75fb8ecd8a30 | engage-the-public-poll-question-generation | null | null | https://aclanthology.org/2021.acl-long.3 | https://aclanthology.org/2021.acl-long.3.pdf | Engage the Public: Poll Question Generation for Social Media Posts | This paper presents a novel task to generate poll questions for social media posts. It offers an easy way to hear the voice from the public and learn from their feelings to important social topics. While most related work tackles formal languages (e.g., exam papers), we generate poll questions for short and colloquial social media messages exhibiting severe data sparsity. To deal with that, we propose to encode user comments and discover latent topics therein as contexts. They are then incorporated into a sequence-to-sequence (S2S) architecture for question generation and its extension with dual decoders to additionally yield poll choices (answers). For experiments, we collect a large-scale Chinese dataset from Sina Weibo containing over 20K polls. The results show that our model outperforms the popular S2S models without exploiting topics from comments and the dual decoder design can further benefit the prediction of both questions and answers. Human evaluations further exhibit our superiority in yielding high-quality polls helpful to draw user engagements. | ['Lemao Liu', 'Baolin Peng', 'Jing Li', 'Yuji Zhang', 'Keyang Ding', 'Zexin Lu'] | 2021-08-01 | null | null | null | acl-2021-5 | ['poll-generation'] | ['natural-language-processing'] | [ 1.34899855e-01 6.05967164e-01 -3.15338284e-01 -5.49755812e-01
-1.50955451e+00 -6.45918190e-01 9.38576281e-01 -4.45570722e-02
-1.88504174e-01 1.07593191e+00 9.40074563e-01 -5.82685947e-01
3.65667015e-01 -8.55354369e-01 -4.99112040e-01 -2.17135191e-01
7.13362753e-01 6.51520491e-01 1.51674375e-01 -5.18553436e-01
3.22222054e-01 -5.18197000e-01 -1.33299792e+00 7.08907545e-01
1.07752526e+00 7.26280093e-01 4.18443769e-01 1.03853714e+00
-6.49077415e-01 9.82175052e-01 -7.96222866e-01 -9.36505318e-01
-1.45154953e-01 -7.20761776e-01 -9.58223939e-01 2.43440196e-02
2.81353980e-01 -2.00380653e-01 -5.39254285e-02 7.27212787e-01
7.42757618e-01 4.36007865e-02 4.88413006e-01 -8.94703448e-01
-7.53949344e-01 1.24652040e+00 -1.81510627e-01 -2.14236185e-01
9.44866240e-01 7.48631209e-02 1.62550223e+00 -8.15944970e-01
8.24879229e-01 1.45079648e+00 5.10775387e-01 8.38512778e-01
-9.48650539e-01 -5.45009017e-01 1.79778531e-01 -6.56897798e-02
-7.65075088e-01 -4.09253031e-01 8.14098418e-01 -2.52713293e-01
3.10326278e-01 7.22631931e-01 6.07892454e-01 1.77268600e+00
-1.23787448e-02 1.48405409e+00 1.07691693e+00 -1.89445674e-01
3.19618881e-01 4.58423495e-01 2.00580448e-01 3.71855348e-01
-1.92526668e-01 -7.88661957e-01 -6.74842179e-01 -5.28493166e-01
6.65267408e-02 -2.14176714e-01 -5.38992323e-02 4.89759386e-01
-1.09036064e+00 1.16465569e+00 -4.05368358e-02 -4.98933624e-03
-3.05289775e-01 -1.66760951e-01 2.40048319e-01 3.56097609e-01
6.93228006e-01 7.06784010e-01 -5.44667602e-01 -5.16525507e-01
-9.52497721e-01 7.73671329e-01 1.50791764e+00 1.18243897e+00
7.01173663e-01 -2.27657795e-01 -7.64887512e-01 1.08016908e+00
3.51837903e-01 5.96167326e-01 7.13260651e-01 -9.47949529e-01
6.69208467e-01 6.30404294e-01 3.23164046e-01 -1.02302516e+00
-1.23195387e-01 -3.38478506e-01 -6.32966936e-01 -9.03080881e-01
4.49626267e-01 -7.82172382e-01 -4.16065753e-01 1.46889925e+00
3.03804010e-01 -1.42505407e-01 -1.05521932e-01 6.95136487e-01
1.24763012e+00 8.92584443e-01 -9.43023413e-02 1.48900390e-01
1.71043217e+00 -1.04877698e+00 -9.73216176e-01 -4.31655079e-01
4.51230735e-01 -1.06731009e+00 1.42541623e+00 2.53819317e-01
-1.23064339e+00 -6.27470315e-01 -4.10683960e-01 -3.55599433e-01
-8.85646343e-02 3.82767856e-01 5.99863052e-01 7.23849297e-01
-8.22783232e-01 1.06602818e-01 -3.18479300e-01 -2.89888561e-01
1.83859780e-01 -1.28987860e-02 4.12704676e-01 3.01534146e-01
-1.54549432e+00 1.21575870e-01 -1.93375707e-01 -5.08037359e-02
-5.54213941e-01 -5.79451144e-01 -7.67214298e-01 -1.18127078e-01
5.30037403e-01 -8.18654239e-01 1.80298460e+00 -5.79739332e-01
-1.92140496e+00 6.74894333e-01 -6.67099655e-01 -4.33435857e-01
6.30306959e-01 -3.04012895e-01 -2.22096324e-01 -6.27563000e-02
4.09500420e-01 7.52660632e-01 6.38580918e-01 -8.71882617e-01
-7.36002147e-01 2.72699326e-01 2.51665205e-01 1.36826873e-01
-5.28906584e-01 -2.90898178e-02 -5.39485633e-01 -5.73551655e-01
-1.13324888e-01 -1.12151706e+00 -4.62878913e-01 -5.99248230e-01
-8.56723905e-01 -7.93937504e-01 5.15991747e-01 -8.35028768e-01
1.34672952e+00 -1.61670065e+00 -2.37800747e-01 -1.78217500e-01
6.04465790e-02 -1.58118848e-02 -7.95489699e-02 7.24452972e-01
5.48710227e-01 2.45625377e-01 1.59824237e-01 -5.77190697e-01
5.53746045e-01 1.18995070e-01 -7.02869236e-01 -1.16472945e-01
6.68155402e-03 1.31315219e+00 -1.02322817e+00 -4.99413133e-01
-3.51696372e-01 1.01725217e-02 -1.09423149e+00 5.03723204e-01
-7.96053410e-01 3.27708751e-01 -9.06603098e-01 3.99094552e-01
3.30718040e-01 -7.04064608e-01 2.24256232e-01 3.31081361e-01
1.09719850e-01 1.10450363e+00 -8.09783578e-01 1.53781557e+00
-6.95947468e-01 7.10200667e-01 1.29228517e-01 -3.10196668e-01
1.24942470e+00 4.22360510e-01 2.37171933e-01 -6.70065522e-01
-2.62113381e-02 3.07776272e-01 -3.61443847e-01 -6.46551669e-01
1.14493835e+00 6.70316741e-02 -7.54534245e-01 7.37781644e-01
-2.38977090e-01 -4.74152386e-01 4.61751372e-01 5.88483334e-01
9.72718000e-01 -5.12453727e-02 -1.30326254e-02 -2.02021137e-01
5.83132684e-01 -9.54411924e-02 2.47262031e-01 1.14752495e+00
2.03302220e-01 8.03795099e-01 7.58290410e-01 -3.61130275e-02
-8.31880808e-01 -7.06539154e-01 2.44171321e-01 1.26629817e+00
-1.95652872e-01 -7.19827354e-01 -7.16491580e-01 -9.00741816e-01
-1.47004992e-01 8.80335331e-01 -3.47165316e-01 4.15466607e-01
-6.04750216e-01 -5.68214953e-01 2.88880646e-01 1.48676455e-01
3.23655725e-01 -1.22946489e+00 1.54531449e-01 3.97294790e-01
-1.01021445e+00 -1.31085229e+00 -6.83826566e-01 -3.79097462e-01
-3.48578453e-01 -7.78105795e-01 -1.10632336e+00 -7.67244637e-01
3.98072332e-01 6.54760227e-02 1.43088698e+00 -2.47227892e-01
1.84073478e-01 3.20789456e-01 -4.56157804e-01 -5.35991013e-01
-5.71143448e-01 8.36732209e-01 -4.96888727e-01 1.92004785e-01
5.46225846e-01 -5.69278836e-01 -7.60138929e-01 3.36439937e-01
-6.73834443e-01 2.60336906e-01 4.12320226e-01 6.96542203e-01
-9.31976084e-03 -6.87025607e-01 1.01909339e+00 -1.59664035e+00
1.42703271e+00 -9.93088067e-01 -3.37539881e-01 -4.10950370e-02
-3.52654487e-01 1.57589954e-03 6.50214016e-01 -3.50421518e-01
-1.29488993e+00 -3.13003033e-01 -7.46646762e-01 6.29343688e-01
-8.26733336e-02 5.39625287e-01 6.16413802e-02 6.80079341e-01
5.70412695e-01 2.74979562e-01 -3.77015114e-01 -6.56735063e-01
4.54712719e-01 1.12283659e+00 1.19663410e-01 -6.16084695e-01
5.66200435e-01 8.40678215e-02 -7.34690845e-01 -9.86414373e-01
-1.37009716e+00 -7.98071444e-01 8.54456127e-02 -3.59751910e-01
6.79984570e-01 -1.05925381e+00 -1.08545411e+00 2.06221104e-01
-1.26879621e+00 -2.66126722e-01 -2.65931875e-01 3.35643589e-02
-3.86232048e-01 2.95422494e-01 -9.18868959e-01 -1.18602610e+00
-4.74271208e-01 -9.36970830e-01 1.48734868e+00 3.61294359e-01
-9.21073973e-01 -1.13162506e+00 1.57511622e-01 1.05306375e+00
4.12356585e-01 -1.39718264e-01 6.43335462e-01 -8.34034145e-01
-6.33280277e-01 -1.79709345e-01 -4.62267511e-02 7.02790320e-02
-7.23421797e-02 -3.67593288e-01 -9.93886888e-01 8.74706954e-02
-3.91268283e-02 -6.62802994e-01 8.89168859e-01 3.83670838e-03
1.29497504e+00 -9.34174836e-01 -3.61675955e-02 1.22682070e-02
8.66220355e-01 -3.23538214e-01 5.34402847e-01 -1.92991551e-02
4.46395874e-01 7.76075721e-01 3.83129835e-01 7.11391270e-01
1.17671907e+00 3.75792265e-01 9.24311986e-04 4.12747115e-02
-7.01045543e-02 -8.39017689e-01 7.04349816e-01 1.60012710e+00
5.19806445e-01 -6.52281582e-01 -5.65004468e-01 7.99457371e-01
-1.74905396e+00 -8.23298752e-01 -5.65142632e-01 1.50250506e+00
1.33032727e+00 1.67734787e-01 9.63262320e-02 -1.05972901e-01
4.51629013e-01 5.79220235e-01 -1.43358216e-01 -4.82779562e-01
-1.10246077e-01 2.93677360e-01 1.77980319e-01 7.72822738e-01
-7.61923969e-01 8.17635238e-01 5.75545263e+00 1.03997290e+00
-6.94598794e-01 1.08604483e-01 8.61461282e-01 3.15793231e-02
-1.10613060e+00 3.61417420e-02 -1.48984754e+00 7.83001661e-01
1.08814049e+00 -2.02800646e-01 -1.74089391e-02 8.34896088e-01
3.81463379e-01 -5.84437326e-02 -7.14601994e-01 5.89199305e-01
7.19928518e-02 -1.52061081e+00 1.69116735e-01 -1.25070289e-01
1.32797515e+00 -1.47886202e-01 8.94770473e-02 7.86252677e-01
9.11089003e-01 -6.74139917e-01 7.50645876e-01 5.73945105e-01
3.38288695e-01 -3.49171549e-01 7.43648410e-01 7.81074882e-01
-8.82181525e-01 1.24182329e-02 -3.61456692e-01 -2.37957105e-01
4.52809453e-01 8.61864686e-01 -1.41115665e+00 3.32836658e-01
1.29800484e-01 5.61637878e-01 -6.37741745e-01 7.27170348e-01
-4.82555419e-01 1.22770762e+00 -1.77655205e-01 -1.02902007e+00
6.08838618e-01 -1.03297189e-01 3.54821682e-01 1.46070337e+00
4.88613993e-01 -1.89579323e-01 3.67018610e-01 8.67403507e-01
-5.30501544e-01 3.73593271e-01 -2.92109102e-01 -1.10888161e-01
4.26236957e-01 1.37787545e+00 -4.58226383e-01 -4.24989372e-01
-3.48006546e-01 8.96850824e-01 2.89177522e-02 4.44438398e-01
-6.14065349e-01 -3.11342806e-01 3.84563029e-01 3.07320565e-01
1.11759335e-01 -1.70120448e-02 -3.38093132e-01 -1.35343361e+00
1.52467191e-01 -1.27569902e+00 2.97351301e-01 -6.52280211e-01
-1.42577696e+00 3.76805335e-01 -4.13256258e-01 -8.47014666e-01
-7.17845917e-01 -2.01325193e-01 -6.32404149e-01 9.62677002e-01
-1.75290549e+00 -9.37889457e-01 -1.86146617e-01 2.69584864e-01
1.17437255e+00 1.01455912e-01 6.40569150e-01 3.62775058e-01
-1.81780428e-01 5.22869289e-01 -1.65712964e-02 2.37289220e-01
8.07513237e-01 -1.47173750e+00 9.17678893e-01 3.86482239e-01
9.68610421e-02 6.56754553e-01 7.91418850e-01 -7.03065872e-01
-1.43368804e+00 -1.12399185e+00 1.94458497e+00 -6.26537621e-01
7.11165726e-01 -7.88010895e-01 -3.99468899e-01 4.03607607e-01
5.98698437e-01 -6.91600978e-01 9.22264636e-01 2.92600036e-01
2.49029696e-01 2.34044120e-01 -6.14376545e-01 1.07966542e+00
6.23452842e-01 -7.20073819e-01 -5.52913964e-01 5.48013389e-01
8.52359891e-01 -3.68676901e-01 -3.67795408e-01 -1.56571835e-01
4.77507502e-01 -7.90586293e-01 6.19465232e-01 -5.41327775e-01
8.58582139e-01 1.69966578e-01 1.09439470e-01 -1.19113731e+00
2.01324314e-01 -1.48887622e+00 -1.72181688e-02 1.48106670e+00
9.77433801e-01 -2.88022548e-01 1.26409578e+00 7.06287622e-01
-6.56211078e-02 -7.43164957e-01 -5.47110021e-01 -5.36790639e-02
-9.24470127e-02 -6.12112939e-01 6.74190998e-01 6.67371690e-01
9.56250504e-02 8.70839655e-01 -7.13304996e-01 -3.52682441e-01
1.72858089e-01 3.54788810e-01 1.14249229e+00 -1.08608305e+00
-4.17672306e-01 -2.75229037e-01 5.81586063e-01 -2.12690282e+00
6.53582215e-02 -9.22051907e-01 2.83447057e-01 -1.64279914e+00
1.22348249e-01 -4.47626889e-01 4.86834258e-01 -1.64823309e-01
-6.61743402e-01 2.50941753e-01 8.26152340e-02 -2.43308738e-01
-1.08473313e+00 6.54949546e-01 1.62549365e+00 -3.30399275e-02
-7.60720521e-02 5.92380226e-01 -1.20463026e+00 5.99947453e-01
5.91130972e-01 -4.31120872e-01 -3.62710088e-01 -2.71583706e-01
1.02390981e+00 5.01425743e-01 -4.98506054e-02 -6.21691525e-01
3.15622717e-01 -6.60154000e-02 -3.81054506e-02 -8.03299308e-01
4.01728988e-01 -2.20917001e-01 -3.92992318e-01 8.17429349e-02
-1.12373710e+00 -2.21479714e-01 -3.39629918e-01 6.95569634e-01
-2.57296652e-01 -4.02098864e-01 1.98387221e-01 -4.26658064e-01
-4.28563915e-02 1.13704227e-01 -8.58789444e-01 6.34794891e-01
2.63603151e-01 3.26417774e-01 -1.40460745e-01 -1.30068839e+00
-6.01365089e-01 5.83654583e-01 -1.58550873e-01 6.74073458e-01
3.80240023e-01 -1.37800920e+00 -1.06348050e+00 1.26454383e-01
-6.51098043e-02 -8.24232250e-02 2.27521092e-01 4.76873338e-01
-1.13715239e-01 8.40387702e-01 5.68167329e-01 -4.53353226e-01
-1.02761602e+00 -3.31878096e-01 -3.23546767e-01 -7.29112267e-01
-1.80432513e-01 1.17457449e+00 -5.45072816e-02 -7.84504771e-01
1.64608836e-01 -4.62739289e-01 -5.72104633e-01 6.46388233e-01
5.30857325e-01 1.91464886e-01 -7.93341398e-02 -1.29532367e-01
2.98462719e-01 -9.79920328e-02 -1.02213100e-02 -2.97456384e-01
1.20100701e+00 -4.47507471e-01 6.40258342e-02 5.41766226e-01
1.22289240e+00 6.56322837e-01 -1.02406442e+00 -4.41132933e-01
2.68789321e-01 -2.58820415e-01 -4.46711659e-01 -5.75431049e-01
-5.97199738e-01 7.43117929e-01 -3.60142589e-01 7.09881604e-01
5.90305865e-01 7.20563233e-02 1.63256252e+00 5.74313641e-01
8.57581422e-02 -1.22013831e+00 4.08690929e-01 9.15202856e-01
7.04882443e-01 -1.27207613e+00 -4.27804261e-01 -2.63212264e-01
-9.40504849e-01 1.07287335e+00 3.45246822e-01 -1.60885602e-02
4.82052445e-01 1.15329504e-01 2.02513218e-01 -5.00608496e-02
-1.19573700e+00 -1.74556136e-01 3.84683043e-01 1.66844606e-01
7.75299728e-01 4.82040681e-02 -6.23532593e-01 1.07306814e+00
-9.68591392e-01 -7.44414851e-02 7.77014494e-01 5.21857202e-01
-4.19725180e-01 -1.35883188e+00 1.08913640e-02 6.31575406e-01
-8.30718279e-01 -2.85727888e-01 -6.26490414e-01 1.60940513e-01
-2.78905928e-01 1.30119681e+00 -1.81195319e-01 -2.34316394e-01
1.10466458e-01 3.28022242e-01 -3.93737793e-01 -9.47988272e-01
-1.03600430e+00 4.19062041e-02 7.27559388e-01 -2.36080453e-01
-2.19380692e-01 -7.64573991e-01 -5.97532988e-01 -2.75605232e-01
-1.83177084e-01 7.72501051e-01 5.87530136e-01 8.22687030e-01
3.81758898e-01 5.84565222e-01 9.04821873e-01 -2.31763497e-01
-7.51490772e-01 -1.16023767e+00 -2.19177499e-01 4.52735931e-01
3.77368212e-01 2.78733790e-01 -3.73066634e-01 8.06305930e-02] | [12.141853332519531, 8.37349796295166] |
8899b258-5af3-4726-b491-ccfcbf3636bb | pattern-mining-for-anomaly-detection-in | 2306.10857 | null | https://arxiv.org/abs/2306.10857v1 | https://arxiv.org/pdf/2306.10857v1.pdf | Pattern Mining for Anomaly Detection in Graphs: Application to Fraud in Public Procurement | In the context of public procurement, several indicators called red flags are used to estimate fraud risk. They are computed according to certain contract attributes and are therefore dependent on the proper filling of the contract and award notices. However, these attributes are very often missing in practice, which prohibits red flags computation. Traditional fraud detection approaches focus on tabular data only, considering each contract separately, and are therefore very sensitive to this issue. In this work, we adopt a graph-based method allowing leveraging relations between contracts, to compensate for the missing attributes. We propose PANG (Pattern-Based Anomaly Detection in Graphs), a general supervised framework relying on pattern extraction to detect anomalous graphs in a collection of attributed graphs. Notably, it is able to identify induced subgraphs, a type of pattern widely overlooked in the literature. When benchmarked on standard datasets, its predictive performance is on par with state-of-the-art methods, with the additional advantage of being explainable. These experiments also reveal that induced patterns are more discriminative on certain datasets. When applying PANG to public procurement data, the prediction is superior to other methods, and it identifies subgraph patterns that are characteristic of fraud-prone situations, thereby making it possible to better understand fraudulent behavior. | ['Christine Largeron', 'Vincent Labatut', 'Rosa Figueiredo', 'Lucas Potin'] | 2023-06-19 | null | null | null | null | ['anomaly-detection', 'fraud-detection'] | ['methodology', 'miscellaneous'] | [ 6.49970025e-02 1.55060425e-01 -3.49651396e-01 -2.57098883e-01
-2.51007855e-01 -5.21986723e-01 3.85976523e-01 7.56479740e-01
4.60790545e-02 5.65720439e-01 -2.97900438e-02 -4.17014480e-01
-5.54224610e-01 -1.31276274e+00 -5.42857647e-01 -6.96730137e-01
-4.08294857e-01 5.78834951e-01 1.09695673e-01 -1.60184965e-01
1.45083457e-01 6.67870045e-01 -1.28411269e+00 4.30496156e-01
1.04777026e+00 8.68801355e-01 -4.43408877e-01 -1.01963907e-01
-1.71304122e-02 8.99869680e-01 -4.13867712e-01 -8.96423757e-01
1.90865651e-01 -3.35338682e-01 -6.21457338e-01 2.51143366e-01
1.61266640e-01 -9.13582742e-02 -4.09432888e-01 1.18127489e+00
-1.85660690e-01 -5.75764120e-01 7.13794589e-01 -1.45544600e+00
-1.16271302e-01 6.89744830e-01 -5.85987926e-01 1.37844130e-01
6.43555760e-01 -3.84025611e-02 1.54460287e+00 -6.04294538e-01
8.65571022e-01 8.33398700e-01 7.92122900e-01 -1.71224819e-03
-1.60186160e+00 -4.50670540e-01 6.95770308e-02 3.50694895e-01
-1.16003489e+00 -6.22651912e-03 1.00873363e+00 -5.15257835e-01
7.22033620e-01 4.73843575e-01 7.53421307e-01 9.68080997e-01
2.02821061e-01 8.82451653e-01 1.14492857e+00 -3.63042921e-01
2.33166158e-01 -8.11891723e-03 4.04905200e-01 6.44791782e-01
1.00513601e+00 -1.24999471e-02 -4.80200410e-01 -7.15280890e-01
3.68714899e-01 3.94143939e-01 -5.15875816e-01 -8.66060853e-01
-1.04635310e+00 9.89145994e-01 2.45607927e-01 5.76736689e-01
-8.21745157e-01 -7.84252584e-02 5.15023470e-01 5.79981208e-01
4.22604263e-01 3.79070103e-01 -3.93581688e-01 1.42591253e-01
-8.00783396e-01 4.31784391e-01 1.05393255e+00 6.84002578e-01
9.08901572e-01 -2.13589758e-01 -1.31769747e-01 4.83341575e-01
1.84971556e-01 -5.31396596e-03 5.45823909e-02 -4.82057929e-01
5.92767835e-01 1.47317743e+00 -4.10148613e-02 -1.19538999e+00
-5.51106632e-01 -5.62056124e-01 -8.81850123e-01 5.96707724e-02
7.75776744e-01 4.84031409e-01 -6.68087661e-01 1.34754038e+00
2.20447868e-01 -3.00152540e-01 -1.64947420e-01 7.14925289e-01
2.16893464e-01 -9.55512524e-02 -1.04639046e-01 -4.12096292e-01
1.26571202e+00 -2.30534524e-01 -6.42829955e-01 1.84492946e-01
1.02924204e+00 -4.87661481e-01 9.34601188e-01 7.10845768e-01
-4.21321481e-01 2.97013968e-01 -8.08300734e-01 6.80627823e-01
-4.90295857e-01 -1.72272757e-01 1.03422201e+00 8.93472552e-01
-6.19849801e-01 1.08163917e+00 -6.51532650e-01 -3.40184689e-01
4.40868348e-01 1.66071132e-01 -5.12514234e-01 -4.47160006e-01
-1.06579983e+00 7.11087227e-01 4.28920507e-01 5.74940480e-02
-2.43014634e-01 -3.29981565e-01 -7.98647583e-01 2.19518155e-01
7.61341572e-01 -3.41327608e-01 6.66312873e-01 -8.20529342e-01
-4.67348248e-01 8.32724452e-01 2.45130867e-01 -7.26849794e-01
8.82086515e-01 3.12068105e-01 -5.98450780e-01 1.59129649e-01
3.60692628e-02 -4.94038194e-01 7.30504751e-01 -1.12188923e+00
-2.99871743e-01 -6.91739917e-01 2.00968713e-01 -4.80301619e-01
-3.57713819e-01 -8.92678648e-02 -4.22302708e-02 -9.26811278e-01
4.98137176e-01 -8.32673609e-01 -1.49198353e-01 -5.57121992e-01
-6.78231955e-01 -3.39583457e-01 7.78797507e-01 -7.38849878e-01
1.51421165e+00 -2.01454067e+00 -1.46296650e-01 9.13319945e-01
3.16235036e-01 -9.34254229e-02 3.34735334e-01 8.01810026e-01
-1.67360067e-01 3.78276706e-02 -5.49962997e-01 1.58268631e-01
4.33742367e-02 4.34798867e-01 -1.01939507e-01 7.35401809e-01
1.75290480e-01 6.77751243e-01 -9.19514298e-01 -2.90621877e-01
-3.99259552e-02 -1.39730245e-01 -4.13420737e-01 -9.80920121e-02
-1.39049798e-01 2.68375903e-01 -5.61155915e-01 1.19089508e+00
7.24385500e-01 -2.37187952e-01 8.66657674e-01 -1.15945891e-01
2.96617270e-01 2.74519324e-01 -1.21044421e+00 1.20742464e+00
-1.05490088e-02 1.55109406e-01 -1.71462655e-01 -1.39238226e+00
1.10108781e+00 1.91346288e-01 8.38614881e-01 -5.91267824e-01
-1.08385488e-01 5.25835514e-01 1.57845527e-01 -3.53345275e-01
1.93077490e-01 1.51082397e-01 -2.17699498e-01 6.97530389e-01
-2.09617510e-01 3.36942673e-01 5.80866039e-01 4.14014697e-01
1.61161506e+00 -1.66713238e-01 5.25459886e-01 -3.14609855e-01
5.07461190e-01 9.47957113e-02 7.85793304e-01 6.11305952e-01
1.11625642e-01 3.86296213e-01 1.25600696e+00 -7.82534063e-01
-8.41824591e-01 -6.39966965e-01 -1.93158999e-01 3.19238067e-01
-6.21391311e-02 -6.79303408e-01 -6.13102436e-01 -1.31860852e+00
8.82679880e-01 6.11274123e-01 -5.94259799e-01 -7.56109655e-02
-4.70870852e-01 -1.08336747e+00 1.60914093e-01 2.32616603e-01
1.09492518e-01 -1.05793977e+00 -3.59430552e-01 5.03736258e-01
-1.63152888e-01 -1.15894449e+00 9.70189571e-02 1.12538815e-01
-1.14408350e+00 -2.02530384e+00 -2.53710508e-01 -9.81104597e-02
9.04657304e-01 1.34474412e-01 1.28236914e+00 6.87917590e-01
-3.23993981e-01 2.84143865e-01 -5.08912444e-01 -1.56583473e-01
-7.19501019e-01 5.55646494e-02 -2.33032554e-03 5.10825276e-01
6.46398067e-01 -7.34242916e-01 -3.62477601e-01 3.62153500e-01
-9.43199515e-01 -3.28676075e-01 8.41072083e-01 8.47139001e-01
2.71415502e-01 5.53687178e-02 5.79199791e-01 -1.54934645e+00
4.95863825e-01 -6.13611937e-01 -6.89866364e-01 3.42810869e-01
-1.36539459e+00 8.17774683e-02 6.04245067e-01 1.06798366e-01
-5.87601423e-01 -6.19749799e-02 3.27054441e-01 -2.17526689e-01
-1.29149795e-01 6.90378070e-01 -2.33262539e-01 -2.32335806e-01
3.07734847e-01 4.48679328e-02 9.76915192e-03 -8.13049316e-01
-2.65498966e-01 4.69782799e-01 1.23776563e-01 -3.42296988e-01
7.98521161e-01 4.25934702e-01 2.23956227e-01 -3.66740763e-01
-4.32221711e-01 -7.26465285e-01 -6.20058656e-01 -2.89256096e-01
1.86491698e-01 -5.12895942e-01 -7.62083530e-01 3.94346505e-01
-9.25904393e-01 1.54950038e-01 -1.17856964e-01 4.13716167e-01
-4.20790672e-01 8.98067474e-01 -7.20141470e-01 -9.32012558e-01
-1.00732058e-01 -6.63001776e-01 6.06269836e-01 -4.63088751e-01
-2.69429326e-01 -6.92213655e-01 -3.67418975e-02 3.84477049e-01
2.19188780e-02 8.56707692e-01 1.30002213e+00 -1.06048465e+00
-6.52881861e-01 -5.71010947e-01 -1.90850139e-01 3.03137720e-01
1.39956728e-01 -9.90787894e-02 -6.56376719e-01 -4.30115223e-01
-2.39996880e-01 2.19059199e-01 1.01069903e+00 -3.76886278e-02
1.10784173e+00 -5.87023437e-01 -4.07160103e-01 2.12928534e-01
1.48987794e+00 -1.08288243e-01 6.48695469e-01 6.23431921e-01
3.96422625e-01 7.69707620e-01 7.33437300e-01 5.34239113e-01
3.12275648e-01 8.53035152e-01 8.04056644e-01 -3.38774920e-02
4.21699971e-01 -1.50146455e-01 2.25775510e-01 4.06556398e-01
-5.60027719e-01 9.02210176e-02 -9.14264917e-01 5.48003078e-01
-2.20930576e+00 -8.79162014e-01 -8.98034930e-01 2.42988873e+00
4.12110806e-01 3.75545919e-01 4.43370491e-01 7.12215483e-01
7.43692577e-01 7.26397261e-02 -1.91464767e-01 -2.82123178e-01
-2.75345564e-01 4.94431928e-02 7.41395473e-01 -9.31352377e-02
-9.22375202e-01 3.21828246e-01 5.57164478e+00 4.18691874e-01
-5.35217524e-01 -3.59876826e-02 4.25899059e-01 2.37629056e-01
-6.29326880e-01 3.53804320e-01 -2.12644815e-01 6.43775344e-01
5.41233182e-01 -4.03235704e-02 2.86697924e-01 1.00252414e+00
3.87089998e-02 -1.34350508e-01 -1.08632112e+00 6.70036435e-01
-5.34155518e-02 -1.07055914e+00 -5.85295521e-02 5.78034401e-01
4.26882207e-01 -2.44196653e-01 -6.72126472e-01 8.56059492e-02
1.03004694e-01 -7.24907160e-01 5.19329011e-01 4.87238765e-01
2.97905535e-01 -8.19013596e-01 1.16960347e+00 1.99133605e-01
-7.77845621e-01 -2.99623042e-01 -1.87182844e-01 5.82569838e-03
-5.79950996e-02 1.16953182e+00 -8.09866548e-01 1.13662779e+00
7.88488865e-01 6.69706047e-01 -5.05947530e-01 1.12414026e+00
-3.79979879e-01 6.12718344e-01 -2.04640865e-01 2.46099502e-01
8.22367221e-02 -4.76835907e-01 6.33855462e-01 1.18762231e+00
4.05667782e-01 -2.06697702e-01 6.03723899e-02 8.65887463e-01
-7.24183023e-02 2.48651668e-01 -9.38482523e-01 -7.71673769e-02
1.14711419e-01 1.22680986e+00 -7.60411382e-01 -2.06274167e-01
-7.42537260e-01 6.98457181e-01 2.26116300e-01 2.38337606e-01
-4.50993359e-01 -2.36466706e-01 5.90632677e-01 4.97550547e-01
3.76674831e-01 -3.80401798e-02 -1.53279766e-01 -1.39859045e+00
5.25626838e-01 -9.46958542e-01 9.50360000e-01 1.47831067e-02
-1.39317203e+00 2.97447115e-01 -1.53703600e-01 -1.48223948e+00
-3.22186172e-01 -6.47752106e-01 -5.79349220e-01 3.43843520e-01
-1.73362470e+00 -9.81597185e-01 -3.41430098e-01 7.83848643e-01
-1.61723256e-01 -1.17111482e-01 8.86753976e-01 3.20243329e-01
-4.39892441e-01 4.10418928e-01 9.27097723e-02 2.34049916e-01
3.42379510e-01 -1.31207395e+00 2.06186712e-01 9.47315037e-01
5.06832600e-01 4.26798940e-01 8.24522495e-01 -8.00372124e-01
-1.49767876e+00 -7.81112194e-01 1.11583996e+00 -2.17993662e-01
8.37764919e-01 -3.99542630e-01 -1.33236837e+00 6.08009338e-01
-2.98863173e-01 1.38642982e-01 5.73872924e-01 3.66190791e-01
-5.78267932e-01 -2.46224165e-01 -1.25426722e+00 -1.01090129e-02
1.14594412e+00 -2.69445509e-01 -4.42740947e-01 4.73970413e-01
1.30559191e-01 -2.16643643e-02 -1.12822950e+00 4.70785469e-01
4.88583624e-01 -1.48479223e+00 7.04061508e-01 -5.77475190e-01
4.33837213e-02 -1.73969731e-01 1.22776106e-01 -1.03009248e+00
-2.65554219e-01 -5.69790006e-01 -5.02686739e-01 1.24249363e+00
4.08282220e-01 -1.02092493e+00 1.01941288e+00 5.16686440e-01
1.95250362e-01 -5.93727112e-01 -1.01836324e+00 -1.13360250e+00
-6.51832938e-01 -3.50996137e-01 8.92948031e-01 1.54946268e+00
3.60861540e-01 -2.32722104e-01 -3.09753805e-01 -3.13295797e-02
9.19392467e-01 6.26362026e-01 7.99118578e-01 -1.72467184e+00
-2.66541302e-01 -4.87656832e-01 -8.61936808e-01 8.86381865e-02
3.10337134e-02 -1.02321482e+00 -5.96878231e-01 -1.39026070e+00
1.75260395e-01 -6.58761859e-01 -5.37457466e-01 7.39308655e-01
2.54333336e-02 2.50307322e-01 1.15531586e-01 5.25348723e-01
-1.48160428e-01 1.82464510e-01 6.10747516e-01 -2.62496352e-01
-4.55438904e-03 4.65985805e-01 -5.32328129e-01 7.50327468e-01
6.83200896e-01 -7.48948812e-01 2.44135588e-01 2.60068864e-01
4.18258280e-01 -2.01117285e-02 5.80414116e-01 -6.86372399e-01
-2.64293514e-02 -1.53590545e-01 5.79346642e-02 -5.41846752e-01
-2.99158305e-01 -1.08812046e+00 3.96787137e-01 9.18009102e-01
2.64387518e-01 7.15412498e-02 -3.51206452e-01 8.91433954e-01
-5.47229707e-01 -3.50408435e-01 1.86273277e-01 -1.87008202e-01
-6.44280374e-01 2.88125813e-01 -1.04965970e-01 -3.87735426e-01
9.82279539e-01 -2.02423826e-01 -1.61753818e-01 -3.09767127e-01
-5.28738081e-01 6.08174428e-02 7.56322503e-01 1.54649943e-01
4.40815389e-01 -1.42832422e+00 -7.99347043e-01 1.15524568e-01
5.79728246e-01 -3.01197350e-01 -3.72386947e-02 1.25406814e+00
-4.33206320e-01 2.72114515e-01 -4.72585931e-02 -5.86800814e-01
-1.22661269e+00 6.28558874e-01 3.85484751e-03 -7.93911815e-01
-9.32088256e-01 -6.64226636e-02 -1.87728196e-01 -2.52867579e-01
-3.87378410e-03 -4.04067218e-01 -1.97119117e-01 2.72593558e-01
3.93393748e-02 4.40977782e-01 5.06010413e-01 -3.54869753e-01
-6.98095500e-01 2.12776393e-01 -5.75167648e-02 6.52643740e-01
1.54367030e+00 2.57631302e-01 -5.49922228e-01 5.18125221e-02
7.20459819e-01 3.13275635e-01 -7.76023388e-01 -3.55144858e-01
9.25078332e-01 -8.63813877e-01 -3.74310166e-01 -5.99343061e-01
-1.32586503e+00 3.72753084e-01 -6.73518032e-02 9.95261550e-01
1.13040161e+00 -1.21658426e-02 5.39391994e-01 2.15659052e-01
8.50169778e-01 -9.73332167e-01 -3.40167701e-01 8.07287544e-02
7.97879040e-01 -1.16698682e+00 2.66773462e-01 -8.46981645e-01
-4.55543041e-01 1.38982904e+00 1.50045291e-01 -1.51268512e-01
1.75749749e-01 -6.62453324e-02 -3.46563160e-01 -5.28362274e-01
-4.42413300e-01 -2.24340290e-01 1.32092476e-01 4.04638439e-01
1.67731106e-01 2.92059064e-01 -8.77450347e-01 5.99009633e-01
1.08283289e-01 -2.42789723e-02 6.79782391e-01 7.49790788e-01
-1.15234941e-01 -1.41174936e+00 -3.15507263e-01 8.34754348e-01
-6.33970141e-01 8.61287192e-02 -6.23087943e-01 1.18254399e+00
-1.36047155e-01 7.13294804e-01 -1.98167831e-01 -2.36172855e-01
7.20745146e-01 2.09342375e-01 3.88465554e-01 -3.63609016e-01
-7.93188214e-01 -2.10976124e-01 5.52678466e-01 -8.91225815e-01
-4.65896249e-01 -9.30294037e-01 -7.69900084e-01 -4.42508399e-01
-5.14519572e-01 2.90458053e-01 4.27826881e-01 7.82000899e-01
3.54863375e-01 1.93519384e-01 8.26651812e-01 -1.06839180e-01
-6.14586115e-01 -8.54938209e-01 -1.14827371e+00 1.10217547e+00
-2.45204680e-02 -8.55923891e-01 -5.40096819e-01 -4.03876096e-01] | [6.86381721496582, 5.820218086242676] |
4dd04bbd-78f9-442d-be29-b3a3183a320b | osteosarcoma-tumor-detection-using-transfer | 2305.09660 | null | https://arxiv.org/abs/2305.09660v1 | https://arxiv.org/pdf/2305.09660v1.pdf | Osteosarcoma Tumor Detection using Transfer Learning Models | The field of clinical image analysis has been applying transfer learning models increasingly due to their less computational complexity, better accuracy etc. These are pre-trained models that don't require to be trained from scratch which eliminates the necessity of large datasets. Transfer learning models are mostly used for the analysis of brain, breast, or lung images but other sectors such as bone marrow cell detection or bone cancer detection can also benefit from using transfer learning models, especially considering the lack of available large datasets for these tasks. This paper studies the performance of several transfer learning models for osteosarcoma tumour detection. Osteosarcoma is a type of bone cancer mostly found in the cells of the long bones of the body. The dataset consists of H&E stained images divided into 4 categories- Viable Tumor, Non-viable Tumor, Non-Tumor and Viable Non-viable. Both datasets were randomly divided into train and test sets following an 80-20 ratio. 80% was used for training and 20\% for test. 4 models are considered for comparison- EfficientNetB7, InceptionResNetV2, NasNetLarge and ResNet50. All these models are pre-trained on ImageNet. According to the result, InceptionResNetV2 achieved the highest accuracy (93.29%), followed by NasNetLarge (90.91%), ResNet50 (89.83%) and EfficientNetB7 (62.77%). It also had the highest precision (0.8658) and recall (0.8658) values among the 4 models. | ['Khandaker Tabin Hasan', 'Raisa Fairooz Meem'] | 2023-05-16 | null | null | null | null | ['cell-detection'] | ['computer-vision'] | [-1.93624243e-01 2.70723552e-01 -1.84054375e-01 1.36999384e-01
-6.24427795e-01 1.45726010e-01 3.32206786e-01 2.89576501e-01
-7.76813030e-01 1.07588875e+00 3.57749201e-02 -3.21086317e-01
-9.16091129e-02 -8.46772194e-01 -3.63287836e-01 -8.27257276e-01
1.28351361e-01 6.18017495e-01 7.33175218e-01 -1.40788509e-02
1.56084076e-01 3.79002899e-01 -1.03061438e+00 3.77956092e-01
6.07163846e-01 7.06641316e-01 4.33292508e-01 6.66320801e-01
5.63063584e-02 9.38528121e-01 -3.37601751e-01 2.18540907e-01
-1.20736589e-03 -3.65678757e-01 -8.20115566e-01 -8.30085650e-02
-2.58173972e-01 -1.99941710e-01 -1.93074167e-01 7.50178039e-01
7.49606252e-01 -1.72719866e-01 1.11920464e+00 -1.10417032e+00
-4.03485782e-02 1.29002884e-01 -8.36432099e-01 6.66185021e-01
-5.03295548e-02 2.94835344e-02 3.71688277e-01 -7.77940691e-01
6.74156606e-01 6.68888092e-01 8.46238792e-01 5.31784415e-01
-8.14438820e-01 -9.40800428e-01 -8.82061064e-01 4.47250992e-01
-1.44592464e+00 -1.45385219e-02 5.67960143e-02 -6.93613350e-01
9.69362557e-01 1.76331133e-01 5.75468600e-01 6.97636127e-01
8.09754670e-01 5.70842683e-01 1.46031547e+00 -6.19416893e-01
2.67213941e-01 3.94021481e-01 6.41079992e-02 7.47327387e-01
1.96631044e-01 1.02992989e-01 -2.60332853e-01 1.17245659e-01
9.27030861e-01 -3.56654846e-03 -1.69087455e-01 5.65690622e-02
-9.94986355e-01 8.82577598e-01 7.04460561e-01 9.11327362e-01
-4.45167065e-01 3.48096117e-02 6.53263748e-01 3.99125606e-01
1.68595999e-01 -2.73276567e-01 -4.66719538e-01 4.80446555e-02
-7.10625291e-01 -9.03116390e-02 3.12319726e-01 5.20754337e-01
5.81627727e-01 3.57991531e-02 1.76839441e-01 1.02280724e+00
3.53471249e-01 1.68446362e-01 1.22448707e+00 -1.00766001e-02
1.39441743e-01 7.30597258e-01 -3.66146833e-01 -4.97810125e-01
-4.81081247e-01 -4.82638866e-01 -1.03755164e+00 3.87839943e-01
2.83312649e-01 8.36292133e-02 -1.28537619e+00 1.21984482e+00
1.80775493e-01 1.36935949e-01 2.06951618e-01 4.57648039e-01
1.08606553e+00 7.22676814e-01 4.43576068e-01 5.15356250e-02
1.53749990e+00 -8.93815339e-01 -3.25546831e-01 1.61147207e-01
9.91303861e-01 -8.42820823e-01 7.42471814e-01 1.51892468e-01
-6.79090381e-01 -4.58227396e-01 -1.06155157e+00 3.92425239e-01
-5.53440452e-01 2.17353389e-01 4.10738379e-01 5.72643280e-01
-1.25401044e+00 3.58984500e-01 -8.80802155e-01 -1.03577983e+00
6.07975900e-01 7.32029974e-01 -7.49090552e-01 -1.05108500e-01
-8.34010839e-01 1.06094348e+00 6.01094961e-01 -4.26560491e-01
-1.06541431e+00 -5.08054852e-01 -3.32662493e-01 -1.57678425e-01
-5.48594780e-02 -7.71176994e-01 1.07765019e+00 -1.00638855e+00
-1.28076291e+00 1.10613394e+00 3.72607559e-01 -5.05096138e-01
5.38364887e-01 2.71829188e-01 -3.22599202e-01 2.92920142e-01
2.31531247e-01 8.45648527e-01 1.10604316e-01 -9.08677399e-01
-8.34076583e-01 -3.19133550e-01 -4.27789569e-01 9.48992819e-02
-2.28865474e-01 6.78552836e-02 -1.48335144e-01 -6.09327137e-01
3.98387872e-02 -1.02653182e+00 -7.09760189e-02 -9.14904028e-02
-1.55670807e-01 -4.45701927e-01 1.06515741e+00 -6.92726195e-01
8.80019069e-01 -1.91190076e+00 -3.83808494e-01 2.79246479e-01
4.31008600e-02 5.11833549e-01 1.20885715e-01 4.38864648e-01
-4.43598002e-01 1.56539768e-01 3.82464305e-02 4.32708651e-01
-5.14887393e-01 1.39671415e-01 6.52592778e-01 4.58404839e-01
-1.68365926e-01 8.03470552e-01 -6.30949855e-01 -1.02564096e+00
3.39998692e-01 3.37632805e-01 -1.53772071e-01 -2.68720705e-02
3.62272382e-01 3.35734516e-01 -4.86800194e-01 7.83582747e-01
2.71802127e-01 -3.48342180e-01 -2.22306684e-01 -8.64587426e-02
2.11531609e-01 -4.63517100e-01 -7.90869832e-01 1.14146495e+00
-5.13988078e-01 6.78695321e-01 -1.07538089e-01 -1.25559211e+00
7.26753592e-01 7.63483286e-01 7.96253085e-01 -7.50103831e-01
5.10736406e-01 5.13338983e-01 3.64525646e-01 -7.06385314e-01
-2.93667138e-01 -6.22275293e-01 4.10498142e-01 2.76979029e-01
3.46879363e-02 1.44119367e-01 1.90481856e-01 5.97020388e-02
1.32689965e+00 -2.04789743e-01 7.73099184e-01 -4.98234004e-01
8.31936419e-01 -4.69533950e-02 3.20624501e-01 1.41840056e-01
-4.29197401e-01 4.58313942e-01 1.40371725e-01 -3.98956895e-01
-8.81191254e-01 -1.12147713e+00 -3.40596169e-01 7.31641114e-01
-2.52342701e-01 1.26526400e-01 -6.40340030e-01 -6.58185482e-01
-2.19541773e-01 3.92046601e-01 -8.26375782e-01 -1.00925446e-01
-3.72137308e-01 -8.97814989e-01 4.23537880e-01 5.24356604e-01
8.46257925e-01 -1.30522799e+00 -6.52265728e-01 2.44288854e-02
2.06601974e-02 -7.08467126e-01 1.38554424e-01 2.49339148e-01
-1.20900285e+00 -1.26461232e+00 -1.06566715e+00 -1.19901347e+00
8.61806214e-01 -1.10795960e-01 8.06477964e-01 4.15031046e-01
-5.93316495e-01 2.22762272e-01 -4.81123239e-01 -8.19926202e-01
-6.40198946e-01 9.93335322e-02 -3.40657443e-01 -3.15873414e-01
5.59314966e-01 -5.05410850e-01 -5.67569494e-01 3.41133416e-01
-7.87398577e-01 1.06204145e-01 1.11718035e+00 1.03944135e+00
5.45381069e-01 7.80780837e-02 9.47976828e-01 -8.34929287e-01
4.34247106e-01 -6.80832982e-01 8.35974440e-02 7.36225322e-02
-6.38203681e-01 -4.54194367e-01 4.45847034e-01 -1.66303113e-01
-6.87387705e-01 -1.29363552e-01 -1.15984410e-01 -2.96257753e-02
-2.37871751e-01 5.03587663e-01 3.01803917e-01 -1.38270512e-01
8.09126198e-01 1.32531837e-01 3.00901830e-01 -8.64404589e-02
-6.10972941e-01 1.05118072e+00 1.72733024e-01 -1.96913943e-01
3.98979098e-01 3.75179589e-01 1.13520026e-01 -1.04895794e+00
-1.88405707e-01 -8.10059369e-01 -2.36613363e-01 -3.46659213e-01
1.04379106e+00 -7.68774629e-01 -7.01585114e-02 5.65806925e-01
-5.05815089e-01 -3.78943294e-01 -2.59901434e-01 8.28462601e-01
-5.09505153e-01 1.92803934e-01 -8.71400833e-01 -4.23383594e-01
-8.65909815e-01 -1.10637856e+00 6.27212346e-01 3.12791884e-01
-2.05956370e-01 -1.00113761e+00 5.63092604e-02 3.18231344e-01
5.65920293e-01 4.49364185e-01 1.11346853e+00 -8.70847225e-01
-2.23635018e-01 -5.53746819e-01 -3.09204578e-01 2.78833598e-01
3.06089491e-01 -8.31616148e-02 -7.55387783e-01 -2.96551496e-01
-1.95504948e-01 -4.83461201e-01 7.39316106e-01 5.62003732e-01
8.10478926e-01 2.41082102e-01 -8.41134608e-01 8.39761943e-02
1.74118948e+00 7.18514204e-01 1.03716254e+00 7.79842257e-01
1.71214491e-01 2.84937292e-01 3.99680078e-01 1.60530827e-03
-2.71558184e-02 2.52165765e-01 3.75886917e-01 -2.72355855e-01
-5.07541418e-01 1.25452891e-01 4.16435041e-02 9.32688296e-01
-3.36022824e-01 -3.40528339e-01 -1.28799391e+00 7.44985878e-01
-1.37401843e+00 -8.02904069e-01 -1.85535252e-01 2.05505824e+00
7.30667770e-01 2.91043133e-01 1.59966126e-01 6.04632616e-01
7.92082906e-01 -6.13129616e-01 -2.64306366e-01 -2.54017800e-01
2.67224997e-01 7.45803475e-01 4.08155471e-01 3.86842191e-02
-1.00454938e+00 5.73877871e-01 5.58968163e+00 1.14302897e+00
-1.22312343e+00 2.72014052e-01 9.22727764e-01 2.61056423e-01
4.33231771e-01 -1.78106442e-01 -4.04935300e-01 5.19194722e-01
1.07126069e+00 -1.16678681e-02 -4.00025308e-01 6.91354036e-01
9.68360007e-02 -7.46306062e-01 -8.58889401e-01 9.64309990e-01
-2.61122376e-01 -1.04484856e+00 -1.24287894e-02 2.59163320e-01
5.85138977e-01 2.13720843e-01 -3.49436224e-01 5.83158374e-01
1.66289419e-01 -1.21997595e+00 -1.27974693e-02 3.34287435e-01
7.55534410e-01 -6.91852987e-01 1.47252762e+00 5.63214242e-01
-9.16132689e-01 5.37918136e-02 -4.35993403e-01 1.47547886e-01
-2.10708275e-01 3.83646727e-01 -1.35786271e+00 4.70311046e-01
9.19099808e-01 4.40002352e-01 -5.06739318e-01 1.16908348e+00
3.18829179e-01 8.05210412e-01 -4.00200039e-01 -3.48633945e-01
2.70394355e-01 5.27682640e-02 -1.02906106e-02 1.09411526e+00
6.03925884e-01 3.03693265e-02 4.65063937e-02 1.02318577e-01
1.39212772e-01 6.68535233e-01 -4.25735235e-01 2.74476022e-01
1.54586732e-01 1.21977067e+00 -1.28473318e+00 -4.57872689e-01
-4.18278784e-01 4.70248729e-01 -7.57131428e-02 -3.08667887e-02
-8.40767264e-01 -4.20832515e-01 -3.52918744e-01 5.25181830e-01
1.29317999e-01 4.84524697e-01 -1.48382142e-01 -5.23829401e-01
-6.42555118e-01 -7.66138792e-01 7.15972364e-01 -9.48116481e-01
-1.11173713e+00 6.05364442e-01 2.47284323e-01 -1.21001530e+00
2.74029803e-02 -7.15034127e-01 -7.14338541e-01 6.02219760e-01
-1.18726444e+00 -1.23827612e+00 -5.94357371e-01 5.05409777e-01
6.63358986e-01 -5.47373950e-01 8.39635909e-01 3.79771501e-01
-4.21466798e-01 3.90919089e-01 6.35294318e-02 2.27812737e-01
5.83830476e-01 -1.17034495e+00 -5.97460568e-01 2.38810703e-01
-4.59757209e-01 1.38511389e-01 3.86380136e-01 -5.25117218e-01
-5.04611969e-01 -9.61718261e-01 6.11735523e-01 1.55581042e-01
4.65746433e-01 3.50555092e-01 -7.39265442e-01 5.47064126e-01
4.45155680e-01 1.75770879e-01 1.03069055e+00 -3.99508625e-01
2.40657836e-01 -1.33497462e-01 -1.44890249e+00 3.36232901e-01
3.06996882e-01 -1.64649952e-02 -4.68544245e-01 4.53018665e-01
-2.50103295e-01 -1.85121864e-01 -1.11225522e+00 5.41631341e-01
4.66395050e-01 -1.09428966e+00 7.52658248e-01 -2.97162026e-01
5.11855721e-01 -2.33944491e-01 -4.24465798e-02 -1.05725610e+00
-3.91105920e-01 3.00825775e-01 3.61888051e-01 8.96819293e-01
4.66097623e-01 -8.25324953e-01 1.25219405e+00 1.47106182e-02
-1.95867762e-01 -1.19857693e+00 -1.04739654e+00 -7.81328559e-01
4.37605560e-01 1.86020806e-01 1.07063770e-01 9.74193454e-01
-1.41774535e-01 4.32062536e-01 3.56183499e-01 -3.60980064e-01
2.80846328e-01 -3.50324422e-01 6.58457935e-01 -1.23585486e+00
-5.77978566e-02 -3.24612111e-01 -9.73503828e-01 -1.83903739e-01
-3.56688917e-01 -9.74476337e-01 -3.73927861e-01 -1.98641813e+00
6.86781466e-01 -5.55876911e-01 -5.81360281e-01 6.75861180e-01
1.29765138e-01 5.16112447e-01 -2.03340039e-01 3.80566210e-01
-7.98894987e-02 7.17712343e-02 1.37654364e+00 -7.39869401e-02
2.00611241e-02 4.05455306e-02 -1.92269862e-01 7.60798752e-01
1.25288057e+00 -4.57352251e-01 -6.17395997e-01 1.66209131e-01
-2.60780424e-01 1.46771530e-02 1.94704428e-01 -1.50707197e+00
2.36149188e-02 -3.69580574e-02 6.77387416e-01 -6.66033208e-01
7.76205510e-02 -8.55350912e-01 4.32248086e-01 1.13840067e+00
1.11979283e-01 3.26287337e-02 1.22508354e-01 3.96061689e-01
-3.31546515e-01 -5.67175031e-01 1.13841128e+00 -5.53455174e-01
-8.44785750e-01 1.76441908e-01 -6.56839728e-01 -3.57177824e-01
1.79235613e+00 -8.52521956e-01 -1.53956234e-01 -1.58955008e-01
-9.49128568e-01 -1.41571626e-01 1.85802355e-01 -6.99200034e-02
4.55127537e-01 -1.23177636e+00 -8.52132559e-01 -9.39513966e-02
1.33820057e-01 -6.46941178e-03 2.39138231e-01 1.22872508e+00
-1.19855714e+00 4.19035435e-01 -5.64749658e-01 -6.97051644e-01
-1.46289086e+00 2.12221354e-01 5.40736735e-01 -6.73320472e-01
-5.15198350e-01 8.31004024e-01 2.76532501e-01 -1.42402872e-01
-1.00985318e-01 -2.55278558e-01 -5.46203554e-01 -2.26659372e-01
1.35139108e-01 6.51616991e-01 3.20462316e-01 -7.21891582e-01
-3.42676282e-01 5.98071635e-01 -3.54792506e-01 2.38475680e-01
1.33759689e+00 4.82463241e-01 -1.53804049e-01 3.38754326e-01
1.24798512e+00 -2.96703160e-01 -4.47322637e-01 1.11555733e-01
1.01403072e-01 -1.27783092e-02 1.46887198e-01 -9.04535830e-01
-1.07271957e+00 7.67112136e-01 1.14983904e+00 -5.49722426e-02
1.16817391e+00 -4.86754216e-02 8.14808965e-01 1.54013997e-02
4.41506386e-01 -1.10319149e+00 3.53445917e-01 1.24875769e-01
6.66225672e-01 -1.15005815e+00 2.48996019e-01 -3.88077497e-01
-3.03211838e-01 1.16776454e+00 7.26143599e-01 -3.38052243e-01
9.70217705e-01 2.63897240e-01 1.44875869e-01 -2.96391547e-01
-6.63044214e-01 -8.24044272e-03 -9.38689038e-02 7.79375076e-01
6.66591644e-01 3.19803730e-02 -6.90062106e-01 3.23246151e-01
-1.58357516e-01 6.18796229e-01 5.05178213e-01 1.31193531e+00
-7.09157526e-01 -9.56813395e-01 -3.83521795e-01 8.67624164e-01
-7.25622952e-01 2.32325003e-01 -1.43870190e-01 1.43274915e+00
2.78514266e-01 6.68543220e-01 -6.67991191e-02 -2.66248345e-01
4.93032224e-02 -6.08826764e-02 4.09646749e-01 -5.56082606e-01
-5.99791050e-01 1.27041563e-01 -1.60380453e-02 1.12504840e-01
-5.59811175e-01 -3.62953603e-01 -1.51594138e+00 -1.21894524e-01
-5.53108513e-01 2.93380409e-01 6.74877346e-01 8.26529860e-01
-7.77618587e-02 6.67653680e-01 1.66996598e-01 -2.22009629e-01
-3.69478822e-01 -1.45710838e+00 -6.94022238e-01 1.68400913e-01
-2.07977265e-01 -7.40275562e-01 -2.05833584e-01 2.83484846e-01] | [15.174031257629395, -2.7532410621643066] |
ff2b17b5-2cc8-42d4-8d16-c5d49bc04d6e | fusion-of-complex-networks-based-global-and | 2106.10701 | null | https://arxiv.org/abs/2106.10701v1 | https://arxiv.org/pdf/2106.10701v1.pdf | Fusion of Complex Networks-based Global and Local Features for Texture Classification | To realize accurate texture classification, this article proposes a complex networks (CN)-based multi-feature fusion method to recognize texture images. Specifically, we propose two feature extractors to detect the global and local features of texture images respectively. To capture the global features, we first map a texture image as an undirected graph based on pixel location and intensity, and three feature measurements are designed to further decipher the image features, which retains the image information as much as possible. Then, given the original band images (BI) and the generated feature images, we encode them based on the local binary patterns (LBP). Therefore, the global feature vector is obtained by concatenating four spatial histograms. To decipher the local features, we jointly transfer and fine-tune the pre-trained VGGNet-16 model. Next, we fuse and connect the middle outputs of max-pooling layers (MP), and generate the local feature vector by a global average pooling layer (GAP). Finally, the global and local feature vectors are concatenated to form the final feature representation of texture images. Experiment results show that the proposed method outperforms the state-of-the-art statistical descriptors and the deep convolutional neural networks (CNN) models. | ['Zhengrui Huang'] | 2021-06-20 | null | null | null | null | ['texture-classification'] | ['computer-vision'] | [ 3.35955292e-01 -7.13382006e-01 -2.01957956e-01 -4.48183745e-01
-6.79942071e-01 -2.69054603e-02 2.98604965e-01 3.49523174e-03
-2.33211979e-01 4.07813907e-01 -1.84926882e-01 1.53009862e-01
-5.27159214e-01 -1.25214553e+00 -4.55101013e-01 -1.30340528e+00
-1.80363730e-01 -3.00403446e-01 3.86647284e-01 -3.93132260e-03
5.75659335e-01 6.66133165e-01 -1.75083983e+00 6.62517965e-01
5.53584039e-01 1.97835469e+00 4.07622568e-02 4.30319577e-01
-1.47870645e-01 7.27427304e-01 -4.69171494e-01 1.53909028e-01
-1.69845611e-01 -2.97456563e-01 -5.76014817e-01 2.07579494e-01
3.06501478e-01 -1.12838484e-01 -5.35806119e-01 1.20155644e+00
3.93409759e-01 3.32137674e-01 5.48694372e-01 -1.01670289e+00
-9.70844150e-01 8.01114589e-02 -6.53755665e-01 1.14889652e-01
1.05801150e-01 -5.50537631e-02 6.83373868e-01 -8.44491422e-01
4.56150740e-01 1.38828647e+00 4.29720551e-01 -1.41734198e-01
-9.90377665e-01 -5.65309346e-01 -1.93356216e-01 5.37450910e-01
-1.62142015e+00 -2.86207348e-01 9.40512896e-01 7.08419308e-02
6.84715927e-01 4.73219573e-01 5.43675363e-01 8.28236341e-01
7.62567401e-01 6.26545429e-01 1.34191608e+00 -2.78298497e-01
-1.51721016e-01 -4.94296670e-01 1.69735506e-01 1.22448218e+00
9.96197611e-02 -1.58840388e-01 -5.50484240e-01 3.80865997e-03
9.97335494e-01 4.32988673e-01 -1.12896964e-01 -5.97734600e-02
-1.16225338e+00 6.25992000e-01 7.10130751e-01 4.25252050e-01
-5.93940258e-01 1.00925073e-01 3.68285894e-01 4.05397564e-01
2.95094788e-01 -1.98533803e-01 -1.08527787e-01 8.37278590e-02
-5.22290409e-01 2.78885253e-02 4.81350094e-01 4.36977595e-01
1.46351695e+00 -2.87475854e-01 -4.99848247e-01 1.11863613e+00
2.70661503e-01 7.72532046e-01 5.02647758e-01 -5.07663131e-01
4.09606695e-01 7.10077584e-01 -4.27769750e-01 -2.07318139e+00
-1.42011359e-01 -1.17356084e-01 -1.29886019e+00 -6.81010336e-02
1.79957405e-01 2.57407039e-01 -1.02260172e+00 1.10648668e+00
-1.05582895e-02 6.84075281e-02 -9.68569070e-02 9.78422582e-01
8.13607514e-01 7.77930558e-01 -6.93576559e-02 5.84772937e-02
1.49822950e+00 -8.40671837e-01 -5.95643818e-01 1.35417739e-02
1.05051771e-01 -8.45594406e-01 6.88289046e-01 4.41784829e-01
-5.39513528e-01 -1.05510807e+00 -1.25430274e+00 2.70269010e-02
-6.05686307e-01 4.66898054e-01 5.90755820e-01 1.70016900e-01
-6.43129349e-01 7.71668077e-01 -9.66461599e-01 -4.08658274e-02
5.89151919e-01 4.33770061e-01 -6.70715332e-01 -2.86818653e-01
-1.08446705e+00 4.08854365e-01 5.18196702e-01 6.59521163e-01
-4.21712697e-01 1.03255659e-01 -1.15906096e+00 3.21464032e-01
5.95580228e-02 -1.28487647e-01 3.94140333e-01 -9.82201219e-01
-1.60039842e+00 5.07782340e-01 -2.46055096e-01 3.72983575e-01
-1.57504827e-01 3.94234478e-01 -6.75990820e-01 3.86548132e-01
-6.10950217e-02 3.35297197e-01 9.05356824e-01 -1.02210593e+00
-8.80441070e-01 -3.86716783e-01 -4.57720667e-01 7.11364374e-02
-4.69174832e-01 -6.48395196e-02 -6.06073856e-01 -5.76102495e-01
7.03393638e-01 -3.89611900e-01 -2.41504777e-02 -1.18091784e-01
-6.06709003e-01 -1.27558261e-01 1.13547885e+00 -7.81628013e-01
1.29163432e+00 -2.48835468e+00 1.05029449e-01 9.38154399e-01
2.23457873e-01 -7.68956244e-02 -3.91019017e-01 6.84628934e-02
1.42308742e-01 -2.85338308e-03 -5.63262329e-02 -2.53230836e-02
-9.73659381e-02 2.99796939e-01 1.32791027e-02 4.35975283e-01
4.38566744e-01 8.13580155e-01 -4.76196110e-01 -7.00984120e-01
4.79976892e-01 5.12077570e-01 -1.14335954e-01 4.10719141e-02
1.68127835e-01 1.84244737e-01 -9.47349906e-01 7.22175777e-01
1.10530221e+00 -1.06799692e-01 2.26864725e-01 -6.70869470e-01
-8.19406181e-04 -2.45228946e-01 -1.28999698e+00 1.39025974e+00
-4.31420058e-01 3.37209880e-01 -3.13048273e-01 -1.25147843e+00
1.54908979e+00 -2.65299201e-01 3.42199951e-01 -1.04256737e+00
3.35339516e-01 2.75964499e-01 -1.82604805e-01 -7.78042197e-01
1.77336946e-01 2.57861733e-01 -3.43228340e-01 2.17057630e-01
3.47970277e-01 3.56837481e-01 -4.61503081e-02 -6.05200887e-01
7.29741275e-01 2.53596026e-02 5.30556180e-02 -1.93064481e-01
9.35969174e-01 -3.45758349e-01 5.29017210e-01 6.95761383e-01
6.36024624e-02 6.48687184e-01 6.76730275e-01 -9.21043575e-01
-5.70059717e-01 -9.76857603e-01 -1.59031644e-01 1.01749325e+00
5.04844785e-01 -7.27079064e-02 -5.52977502e-01 -6.92326725e-01
2.17591614e-01 -3.24119896e-01 -9.14632022e-01 -3.08286428e-01
-5.54899275e-01 -7.62969613e-01 4.17320281e-01 3.39313149e-01
1.20991611e+00 -1.28818953e+00 -2.20736787e-01 2.97383457e-01
-1.49709001e-01 -6.61613822e-01 -3.11434507e-01 1.24546431e-01
-4.53893572e-01 -9.55663860e-01 -5.11296570e-01 -1.19286478e+00
6.89647913e-01 2.21097112e-01 3.03128451e-01 2.65209973e-01
-4.31166261e-01 -2.14809284e-01 -3.92045975e-01 3.09495807e-01
2.65773475e-01 -4.85789888e-02 -2.58419216e-01 7.81599581e-01
5.75330019e-01 -5.35927832e-01 -6.52480841e-01 3.34424198e-01
-8.46962273e-01 -1.86284669e-02 9.46372688e-01 1.22323465e+00
1.03093004e+00 4.20837402e-01 2.19314739e-01 -3.38532776e-01
6.96993947e-01 -1.66171744e-01 -2.71329194e-01 5.34895718e-01
4.41984721e-02 1.04450949e-01 7.27069974e-01 -3.90325189e-01
-9.85845268e-01 -1.25408396e-01 1.78328541e-03 -2.00056985e-01
-2.66694307e-01 9.58942354e-01 -1.97881490e-01 -6.47820354e-01
2.58238792e-01 6.62994266e-01 2.07221061e-01 -5.26029050e-01
1.24988496e-01 9.98633683e-01 5.27437389e-01 -7.32573271e-01
3.06228757e-01 3.38571966e-01 1.34070501e-01 -8.03336620e-01
-6.66365743e-01 -6.56576753e-02 -6.29377663e-01 -1.08351037e-02
9.46469724e-01 -4.45506185e-01 -1.07813466e+00 1.15102470e+00
-9.09246743e-01 -3.63078006e-02 2.96150178e-01 4.22027469e-01
-5.30491590e-01 4.23114151e-01 -7.58034110e-01 -2.96085060e-01
-2.44143903e-01 -1.26344836e+00 1.14364040e+00 6.33997202e-01
4.38819170e-01 -8.57448161e-01 -3.54654223e-01 -3.18138361e-01
6.67148590e-01 5.55557132e-01 1.04247725e+00 -2.96009302e-01
-5.31285048e-01 -3.69079173e-01 -8.01153898e-01 5.29187500e-01
5.51802754e-01 1.71173990e-01 -7.09875166e-01 -9.54746976e-02
-1.38292357e-01 -3.06001216e-01 1.25509584e+00 3.27005357e-01
1.94731677e+00 -3.94713461e-01 -4.14690197e-01 9.11289990e-01
1.61338925e+00 1.94646865e-01 7.35562623e-01 3.96531701e-01
8.23069930e-01 3.21796447e-01 4.53706115e-01 3.96440327e-01
3.95400077e-01 3.95574152e-01 4.71692905e-02 -2.37151340e-01
-7.82721639e-02 -1.66984215e-01 -1.12857275e-01 1.10602546e+00
-4.32718098e-01 -5.64293116e-02 -5.81510961e-01 1.01942025e-01
-1.84407020e+00 -8.85763586e-01 2.48651072e-01 1.94236588e+00
6.88629508e-01 6.39748424e-02 -5.16089380e-01 1.15319071e-02
9.32190776e-01 5.22715628e-01 -1.92956269e-01 -3.53290588e-01
-5.20160973e-01 7.89509654e-01 4.84858423e-01 1.75655216e-01
-1.38757813e+00 8.63434076e-01 5.45757484e+00 1.46543837e+00
-1.62496769e+00 -2.28901446e-01 8.48051488e-01 5.56497335e-01
8.51995796e-02 -1.87843069e-01 -3.34662586e-01 5.19287109e-01
4.12618339e-01 1.30761817e-01 6.49566948e-01 5.98767757e-01
-1.19635478e-01 -1.73935100e-01 -4.14844662e-01 1.19810581e+00
1.99055523e-01 -1.19231856e+00 3.71039480e-01 2.13429350e-02
6.24105334e-01 -2.85296947e-01 3.14793736e-01 1.96840644e-01
-1.42064437e-01 -1.09789777e+00 3.93718153e-01 1.26721489e+00
9.80819821e-01 -9.30436671e-01 1.06984317e+00 -1.61407650e-01
-1.59189010e+00 -3.12473983e-01 -8.53363633e-01 -2.06496716e-02
-4.37249035e-01 6.22705519e-01 2.27383614e-01 9.39097285e-01
8.29102695e-01 8.67796183e-01 -8.03215027e-01 9.30700481e-01
-1.55237600e-01 2.40532443e-01 -3.02039117e-01 -2.31273979e-01
3.73836488e-01 -1.94168463e-01 -1.90383732e-01 1.20786679e+00
5.13995647e-01 9.28766206e-02 3.18981737e-01 6.57808661e-01
-1.89053658e-02 2.13259518e-01 -2.02396989e-01 -8.80970340e-03
3.48198295e-01 1.73040831e+00 -7.92444766e-01 -5.01302242e-01
-3.06889296e-01 1.07344306e+00 3.65153253e-01 5.51530659e-01
-4.42916751e-01 -1.35965705e+00 4.70481873e-01 -6.68597162e-01
5.71159840e-01 -2.98210144e-01 -1.73948959e-01 -1.03023255e+00
1.43513575e-01 -7.62747228e-01 1.98504344e-01 -5.96011996e-01
-1.33145392e+00 6.36685252e-01 -4.91793543e-01 -1.17370844e+00
1.85982659e-01 -9.62160051e-01 -7.38017797e-01 1.19434083e+00
-1.35183871e+00 -1.54594219e+00 -7.70948827e-01 6.68101788e-01
-1.32052094e-01 -7.11316690e-02 8.42808247e-01 1.75745606e-01
-6.58331454e-01 6.19854331e-01 2.92308629e-01 5.04411936e-01
6.00660980e-01 -7.25515425e-01 5.78506365e-02 4.82931256e-01
-3.63097221e-01 6.03833973e-01 -1.49214417e-01 -5.43109894e-01
-1.59126532e+00 -1.08797741e+00 7.01826215e-01 2.59135932e-01
4.64798152e-01 -3.17342818e-01 -1.05853832e+00 3.44575614e-01
-1.70528442e-01 5.73271334e-01 3.07400942e-01 -2.58684993e-01
-3.09623271e-01 -5.23686707e-01 -1.05674076e+00 1.63958207e-01
7.83092320e-01 -7.45861053e-01 -3.25125039e-01 -6.27193972e-02
3.23084742e-01 -3.72811556e-01 -1.18688023e+00 4.99513686e-01
9.97730374e-01 -1.00129437e+00 7.20166266e-01 -3.65466028e-01
5.74830353e-01 -4.00129706e-01 -3.89426559e-01 -1.13047326e+00
-7.99394429e-01 1.64610282e-01 4.22733337e-01 9.93550360e-01
1.14564955e-01 -9.17722166e-01 4.75027263e-01 -2.73520559e-01
-5.10530770e-02 -1.01084733e+00 -9.88463223e-01 -4.97991353e-01
-3.96468133e-01 -5.81689030e-02 9.36638296e-01 9.13037956e-01
-6.83986163e-03 -3.11531156e-01 -2.64862806e-01 1.02316573e-01
4.78794605e-01 4.46533382e-01 3.78807396e-01 -1.05881715e+00
1.14421383e-01 -6.98908627e-01 -9.51467812e-01 -8.23014438e-01
7.04135299e-02 -9.04474676e-01 1.06145874e-01 -1.32416582e+00
3.51110280e-01 -5.81913292e-01 -9.87477243e-01 8.80706608e-01
-1.81308299e-01 6.31871402e-01 -1.53849050e-01 4.55815345e-02
-6.03606462e-01 6.64311051e-01 1.56123948e+00 -2.94912785e-01
2.53043249e-02 -4.05968010e-01 -6.13324821e-01 3.32673699e-01
6.55844033e-01 -2.05820918e-01 3.43118995e-01 -2.99277514e-01
-3.50318044e-01 -4.92844284e-02 6.03773117e-01 -1.21359646e+00
3.59385252e-01 -4.51119155e-01 1.06324029e+00 -2.81022966e-01
1.10933393e-01 -4.49176759e-01 -2.07264945e-01 4.10395652e-01
-6.79842383e-02 -1.66125298e-01 1.63735285e-01 2.18044788e-01
-8.78603280e-01 2.98461974e-01 5.29038608e-01 1.57793999e-01
-9.64453399e-01 6.98435485e-01 -1.91186830e-01 -7.53101110e-01
6.25580728e-01 -3.35825771e-01 -7.51979709e-01 1.12934500e-01
-6.91589117e-01 -1.46971981e-03 5.16722053e-02 5.44832647e-01
1.06570494e+00 -1.92964053e+00 -4.71671551e-01 9.83818412e-01
2.08955824e-01 -2.60901064e-01 7.94758439e-01 8.70881617e-01
-8.05744708e-01 1.89890385e-01 -9.48845029e-01 -6.46461189e-01
-8.94821763e-01 2.53592104e-01 4.28833753e-01 -8.21068510e-02
-2.19447598e-01 6.17267013e-01 -6.86746538e-02 -2.51980811e-01
-1.75633490e-01 -2.50810385e-01 -4.42453146e-01 -6.67882897e-03
5.71700215e-01 1.99681386e-01 1.46759346e-01 -7.91698933e-01
-3.91297370e-01 1.21550524e+00 9.56373289e-03 2.08615825e-01
1.32896531e+00 2.50754412e-02 -7.77678907e-01 -4.33981009e-02
1.88077688e+00 -2.33660623e-01 -9.27642107e-01 -5.36794186e-01
-3.50433439e-01 -5.47772706e-01 1.66895941e-01 -5.05009234e-01
-1.31906939e+00 9.90877867e-01 7.42838204e-01 1.30115137e-01
1.47769809e+00 -2.74139583e-01 7.28591383e-01 4.13309157e-01
2.57874191e-01 -1.01056659e+00 1.31570265e-01 6.37685478e-01
7.38933921e-01 -8.42813313e-01 -1.82180434e-01 -2.51769930e-01
-2.68060625e-01 1.72842991e+00 6.60107017e-01 -4.83127743e-01
8.80621254e-01 1.64277328e-03 3.45041938e-02 -3.39776576e-01
-2.65443414e-01 -1.71952114e-01 6.24648333e-01 4.78618294e-01
3.44914943e-02 3.25194418e-01 -1.81582496e-01 8.47050548e-01
1.85055342e-02 -1.60949696e-02 -1.53637156e-01 7.98818469e-01
-7.13881433e-01 -9.34701800e-01 -3.68724048e-01 7.69714713e-01
-3.22920084e-01 7.86986798e-02 4.38035093e-02 4.60567117e-01
4.89509255e-01 7.79578328e-01 4.40817863e-01 -1.11546302e+00
2.20015615e-01 -2.55213857e-01 4.20693308e-01 -1.44856453e-01
-4.01150100e-02 5.95434494e-02 -3.54342431e-01 -7.53844976e-01
-2.71825433e-01 -1.72523797e-01 -9.71196413e-01 -3.70763779e-01
-2.68009990e-01 3.68256047e-02 6.22034371e-01 9.59007800e-01
2.89610147e-01 7.08812594e-01 9.96682346e-01 -1.16600704e+00
-3.17225426e-01 -1.02770007e+00 -8.64102781e-01 4.85070139e-01
2.91708559e-01 -7.84514785e-01 -3.28334302e-01 -1.35884419e-01] | [10.35673999786377, -0.3542247712612152] |
3383c7e2-036a-4d78-988e-0690fdc1e78d | 3d-color-homography-model-for-photo-realistic | null | null | https://link.springer.com/article/10.1007/s00371-017-1462-x | https://link.springer.com/article/10.1007/s00371-017-1462-x | 3D color homography model for photo-realistic color transfer re-coding | Color transfer is an image editing process that naturally transfers the color theme of a source image to a target image. In this paper, we propose a 3D color homography model which approximates photo-realistic color transfer algorithm as a combination of a 3D perspective transform and a mean intensity mapping. A key advantage of our approach is that the re-coded color transfer algorithm is simple and accurate. Our evaluation demonstrates that our 3D color homography model delivers leading color transfer re-coding performance. In addition, we also show that our 3D color homography model can be applied to color transfer artifact fixing, complex color transfer acceleration, and color-robust image stitching. | ['Han Gong; Graham Finlayson; Robert Fisher; Fufu Fang'] | 2019-03-01 | null | null | null | the-visual-computer-2019-3 | ['image-stitching'] | ['computer-vision'] | [ 4.61270809e-01 -4.15234476e-01 3.13000351e-01 6.41984865e-02
-4.62967545e-01 -8.23426008e-01 5.55064499e-01 -6.12535715e-01
-2.11510770e-02 3.23689938e-01 -1.34589344e-01 -3.67806137e-01
3.95414591e-01 -6.20102346e-01 -9.80476677e-01 -3.95766318e-01
4.56420660e-01 2.24281877e-01 3.70902508e-01 -3.58424515e-01
5.66883504e-01 8.11823368e-01 -1.26554298e+00 1.15100080e-02
8.56222093e-01 7.81487107e-01 -1.36731446e-01 9.77089226e-01
-1.40767902e-01 2.88729012e-01 -4.55264181e-01 -5.79579592e-01
8.23133349e-01 -9.11069989e-01 -6.58822298e-01 3.77794176e-01
9.55806196e-01 -5.57671189e-01 -2.81534582e-01 1.17847395e+00
1.85168684e-01 -2.42897362e-01 4.49611306e-01 -1.60203314e+00
-1.00308549e+00 -2.74005264e-01 -1.12285757e+00 -4.65502352e-01
6.92346811e-01 -1.53554991e-01 3.28951657e-01 -8.90822232e-01
7.93580830e-01 1.25334001e+00 6.46376252e-01 4.31950927e-01
-1.51784313e+00 -6.98170364e-01 -5.46104968e-01 -6.04854804e-03
-1.15985382e+00 1.90833013e-03 1.11299443e+00 -2.43985221e-01
3.95344049e-01 6.62643135e-01 1.23760915e+00 4.65523690e-01
6.19603634e-01 3.79769027e-01 1.69729531e+00 -9.28240716e-01
1.91717576e-02 -5.24893515e-02 -3.42878699e-01 8.00036311e-01
1.90833628e-01 7.05209076e-01 -5.13440430e-01 -2.18701482e-01
1.43991339e+00 1.50335208e-01 -4.52068567e-01 -8.59411716e-01
-1.10037076e+00 3.08706135e-01 3.52785319e-01 -2.25696322e-02
3.75823490e-02 5.70054770e-01 -7.95100257e-02 4.74927306e-01
5.52424610e-01 3.86222035e-01 6.31699711e-02 -1.73059583e-01
-7.89491415e-01 -4.43355031e-02 6.48581207e-01 1.36745453e+00
1.12282360e+00 1.91058412e-01 1.73530400e-01 4.88340378e-01
4.88449000e-02 1.05992544e+00 3.61703783e-02 -1.58794868e+00
5.10842204e-02 4.60571200e-01 2.34420225e-01 -1.11310911e+00
2.73965776e-01 1.29221410e-01 -5.19143224e-01 1.05107594e+00
2.06484661e-01 1.81935579e-01 -8.39939296e-01 1.24056888e+00
2.17724696e-01 2.98430771e-01 -1.39905497e-01 8.53707790e-01
5.40964305e-02 6.51097775e-01 -5.52020550e-01 4.66951989e-02
1.16243303e+00 -8.13655019e-01 -7.61681080e-01 4.00703773e-02
9.06025693e-02 -1.42768991e+00 1.13894475e+00 3.22514206e-01
-1.53702950e+00 -4.58251745e-01 -9.85013664e-01 -4.72813457e-01
-8.13763812e-02 -1.74567699e-01 5.90430915e-01 8.84221554e-01
-1.45500195e+00 4.42671150e-01 -4.46849138e-01 -3.80396158e-01
-2.07396269e-01 8.03117678e-02 -6.21949434e-01 -3.06567758e-01
-8.37623775e-01 9.47369218e-01 1.01120874e-01 -2.65652329e-01
-3.14928323e-01 -1.03479385e+00 -5.12185633e-01 -4.67028804e-02
4.17188965e-02 -8.56982708e-01 9.55252111e-01 -1.46365142e+00
-2.07138228e+00 1.27456641e+00 -1.97869286e-01 5.57122529e-02
7.69788742e-01 -2.49902159e-01 -2.11863846e-01 3.01871419e-01
-1.91178322e-01 5.96037805e-01 1.18973756e+00 -1.90622151e+00
-5.62970638e-01 -2.09765196e-01 -3.30987483e-01 3.68228197e-01
-6.77832663e-02 -2.98690833e-02 -9.50799227e-01 -6.35217071e-01
3.35024387e-01 -1.09861338e+00 2.15250269e-01 5.84721565e-01
-4.30353820e-01 9.14732575e-01 1.18950367e+00 -7.58108079e-01
8.66898358e-01 -2.25327826e+00 1.34606868e-01 4.50462610e-01
8.28745142e-02 8.29790086e-02 -3.17124814e-01 8.36491168e-01
-4.10341442e-01 -9.16808322e-02 -3.17401558e-01 -2.27573603e-01
-2.43116558e-01 2.15638876e-02 -5.22323489e-01 4.84304726e-01
-7.02815056e-02 9.12301838e-01 -8.45670223e-01 -3.37405413e-01
4.85523641e-01 8.65258873e-01 -5.77155352e-01 3.04515213e-01
2.30832100e-01 4.05022293e-01 8.64755809e-02 6.17274702e-01
1.28765333e+00 1.79350913e-01 4.91659902e-02 -5.92658937e-01
-3.19668502e-01 -4.62317199e-01 -9.17336762e-01 1.83371222e+00
-3.92395526e-01 8.55732083e-01 9.51105952e-02 -5.95719405e-02
1.00134683e+00 -3.05240173e-02 6.15113318e-01 -6.88720703e-01
2.06827015e-01 1.56805411e-01 -5.52094936e-01 7.51383826e-02
8.24739099e-01 -9.72487628e-02 2.65864193e-01 7.98380494e-01
-2.88597018e-01 -9.82747793e-01 -1.87888175e-01 3.02991360e-01
5.23602366e-01 4.33775872e-01 1.26269776e-02 -3.85492921e-01
4.70838666e-01 -1.19099580e-02 2.40746826e-01 3.58963341e-01
8.80092978e-02 1.09872293e+00 4.97514933e-01 -3.59120905e-01
-1.70100999e+00 -1.15331101e+00 1.30966291e-01 5.37705243e-01
7.50566781e-01 -1.67594045e-01 -9.31682289e-01 -1.90092295e-01
6.81254417e-02 5.25413096e-01 -5.78878403e-01 -2.47936994e-01
-6.38102829e-01 -6.94315508e-02 3.57170492e-01 2.04141513e-01
7.66876459e-01 -2.75603682e-01 -6.36918008e-01 -1.05108768e-01
-3.75948637e-03 -8.72395933e-01 -1.15290976e+00 -3.04749131e-01
-1.02222192e+00 -1.27861261e+00 -9.83080566e-01 -9.13677871e-01
1.02577281e+00 1.14070117e+00 1.10585201e+00 2.36387774e-01
-3.21373254e-01 1.02253151e+00 -4.62283105e-01 -4.56268042e-02
-7.79021025e-01 -7.95982420e-01 -3.28277946e-01 3.18025887e-01
-7.42594376e-02 -4.71660525e-01 -8.66205037e-01 6.21783316e-01
-1.21106577e+00 5.74828327e-01 2.17316434e-01 7.24157989e-01
6.27570152e-01 -2.78662622e-01 -6.66898012e-01 -8.84247065e-01
4.60072964e-01 5.68872273e-01 -9.87722576e-01 5.26873827e-01
-6.76179409e-01 -5.30228019e-02 4.20661747e-01 -3.95041347e-01
-1.39636123e+00 3.91084045e-01 2.67020851e-01 -7.71784723e-01
3.21563184e-01 -2.76650578e-01 1.91455483e-02 -9.02991891e-01
5.26321530e-01 3.83262455e-01 5.94909787e-01 -4.48793352e-01
8.86379242e-01 1.42941549e-01 7.66364932e-01 -5.99161446e-01
1.56198287e+00 9.71936107e-01 3.46360058e-01 -6.62593722e-01
-1.49207078e-02 -1.44763932e-01 -8.09638500e-01 -4.52433497e-01
6.64741218e-01 -7.15223193e-01 -8.09541821e-01 7.78408170e-01
-1.21500623e+00 -5.14175475e-01 -2.79470950e-01 2.26332694e-01
-8.32761109e-01 7.47539997e-01 -6.91659868e-01 -2.77067155e-01
-2.14860886e-01 -9.36028898e-01 1.29399323e+00 3.69858481e-02
1.77657828e-01 -1.03762186e+00 2.66832739e-01 -6.46729991e-02
5.38078725e-01 3.99223059e-01 8.74017537e-01 8.18374753e-01
-8.95661473e-01 -2.55619317e-01 -4.81439024e-01 2.86895603e-01
3.19659919e-01 5.89320123e-01 -8.26360345e-01 -4.84615296e-01
-9.11201090e-02 2.41135098e-02 5.45769930e-01 2.60598063e-01
7.00329840e-01 8.15256983e-02 -1.50968447e-01 1.12579215e+00
1.78320467e+00 4.96800780e-01 1.08025205e+00 3.17857563e-01
8.51425290e-01 3.30187023e-01 5.07848144e-01 2.81938892e-02
2.63423249e-02 8.50184441e-01 2.68554419e-01 -8.71816754e-01
-5.05832553e-01 -5.41074097e-01 3.34027261e-01 6.30011737e-01
-3.21877360e-01 2.47025758e-01 -6.51806712e-01 -1.10784739e-01
-1.38074040e+00 -8.85471165e-01 -3.13073814e-01 2.47632766e+00
6.40565634e-01 -4.24015075e-01 -1.28500670e-01 8.00758451e-02
8.75397146e-01 -2.66787797e-01 -3.08127970e-01 -4.91585940e-01
-2.18093634e-01 7.94423372e-02 9.66731489e-01 9.60225224e-01
-5.37037253e-01 1.09150529e+00 7.30441999e+00 5.31227946e-01
-1.18469620e+00 2.58803442e-02 3.94911110e-01 4.26437318e-01
-7.04828620e-01 3.30997944e-01 -4.33555171e-02 1.71433523e-01
5.67141511e-02 -5.03240526e-01 8.35916877e-01 4.62084234e-01
5.61942682e-02 -2.02078819e-01 -9.93414462e-01 1.36534989e+00
5.25535166e-01 -1.18035460e+00 3.84695053e-01 2.53444910e-01
1.00420582e+00 -6.73304677e-01 5.21563351e-01 -5.05620003e-01
2.70908982e-01 -3.81464720e-01 8.80652964e-01 2.70527631e-01
1.41753078e+00 -6.49044156e-01 1.23064853e-02 -3.25819254e-01
-1.11545026e+00 3.24266613e-01 -3.44286889e-01 2.94119030e-01
2.28716820e-01 2.43563831e-01 -4.39936221e-01 6.59127772e-01
6.03651702e-01 6.58448100e-01 -5.70260346e-01 1.24940443e+00
-2.57112771e-01 -7.88147897e-02 -4.07423861e-02 4.50631350e-01
-1.10150963e-01 -8.97573888e-01 4.06302482e-01 8.18231761e-01
7.47240543e-01 2.48052165e-01 -3.33855659e-01 1.04935765e+00
1.19234659e-01 -3.35183740e-02 -7.91713297e-01 3.28813374e-01
2.62588590e-01 9.92119551e-01 -8.06965411e-01 -2.44770363e-01
-3.54401648e-01 1.89633071e+00 -2.67931521e-01 8.25609207e-01
-8.94429386e-01 -4.04850423e-01 7.81783521e-01 3.58137600e-02
-8.73290598e-02 -3.25604349e-01 -3.97888809e-01 -1.28695273e+00
-3.64184976e-01 -6.01412714e-01 -2.97074914e-01 -1.44269156e+00
-9.57207143e-01 5.33752978e-01 -1.99253652e-02 -1.84029651e+00
1.02846950e-01 -8.00171793e-01 -5.57397068e-01 8.70561302e-01
-1.63525414e+00 -1.27332175e+00 -6.73680961e-01 8.81604493e-01
9.67947394e-02 9.62397978e-02 6.58331513e-01 1.86932653e-01
1.62562728e-01 4.18894678e-01 3.86871547e-01 -1.37929976e-01
1.05882609e+00 -1.29881835e+00 6.52493954e-01 1.01089084e+00
-2.41403654e-01 4.70217586e-01 7.32044876e-01 -4.98105019e-01
-1.95374644e+00 -7.59295523e-01 2.66520023e-01 -5.39145589e-01
2.44576141e-01 -2.44362488e-01 -8.39765966e-01 7.06879973e-01
6.06596529e-01 -1.17739879e-01 3.56369913e-01 -6.16442204e-01
-5.31035721e-01 -3.87263596e-01 -1.13416743e+00 6.30365431e-01
9.64321077e-01 -8.03685129e-01 -2.50140458e-01 -4.73557748e-02
5.39784789e-01 -7.60040641e-01 -6.32711947e-01 -2.40391523e-01
9.19580638e-01 -1.31413698e+00 1.20405138e+00 5.00893295e-02
3.66731972e-01 -5.33009529e-01 6.73529878e-02 -1.48384333e+00
-4.05371517e-01 -1.00336134e+00 5.40280998e-01 9.53587770e-01
-2.01367009e-02 -7.63877273e-01 6.33760393e-01 8.17437410e-01
-5.17078489e-02 2.57502854e-01 -7.50254989e-01 -8.37377727e-01
4.92891334e-02 6.84396029e-02 5.13608634e-01 1.10513377e+00
1.14091728e-02 -3.24473143e-01 -8.02560508e-01 1.54702021e-02
9.46269453e-01 4.38604414e-01 1.02238357e+00 -8.98508430e-01
-2.50992656e-01 -3.60061854e-01 -3.28467667e-01 -1.01804996e+00
-1.80909202e-01 -5.53722858e-01 -1.82675645e-01 -1.15947402e+00
2.02269718e-01 -3.25782120e-01 5.74258417e-02 1.26569822e-01
-2.32455116e-02 6.80105925e-01 5.30121267e-01 2.71364301e-01
2.00919017e-01 4.16755736e-01 1.82666194e+00 -7.84748793e-02
-3.27248394e-01 -4.30510670e-01 -3.33018094e-01 3.14894527e-01
6.47903979e-01 -2.37718314e-01 -5.10177732e-01 -7.22691536e-01
2.59088665e-01 2.91476518e-01 4.18001324e-01 -7.20985711e-01
1.08073212e-01 -2.14832112e-01 4.36692536e-01 -3.71165633e-01
5.34259975e-01 -1.37309265e+00 6.50698364e-01 6.64826691e-01
-1.78037435e-02 4.49199796e-01 1.47381991e-01 3.33446354e-01
-2.58439302e-01 1.47259682e-01 9.12207365e-01 -5.37087023e-02
-6.93711102e-01 1.66701987e-01 -8.01760256e-02 -2.81622797e-01
1.15531063e+00 -7.78403521e-01 -3.73635381e-01 -6.21781945e-01
-1.73936531e-01 -5.55955768e-01 1.40119851e+00 2.30419233e-01
1.02618635e+00 -1.75822127e+00 -4.54594284e-01 7.71933138e-01
-9.80820972e-04 -6.60964012e-01 6.97127134e-02 5.36348879e-01
-1.48850489e+00 -7.39956228e-03 -7.72317648e-01 -6.88978851e-01
-1.38628662e+00 6.52614295e-01 5.01378775e-01 4.98264790e-01
-7.90816128e-01 5.02452850e-01 3.35740894e-01 -1.58929080e-01
-1.08723983e-01 -1.37806714e-01 6.27255619e-01 -5.75955153e-01
5.50297976e-01 5.17968059e-01 -1.93767682e-01 -5.83336234e-01
-7.13434741e-02 1.56345284e+00 4.10160512e-01 -6.03527129e-01
8.91619802e-01 -3.21279824e-01 -5.15533268e-01 6.92546070e-02
1.44026387e+00 2.60944515e-01 -1.49881232e+00 4.48758006e-02
-7.04128385e-01 -1.40260041e+00 6.49553984e-02 -6.52821064e-01
-1.35171986e+00 8.80815029e-01 8.16371083e-01 9.73624066e-02
1.52294922e+00 -4.54335332e-01 6.76070571e-01 -2.25741304e-02
5.04014313e-01 -8.34219038e-01 2.46548474e-01 2.26473197e-01
1.02874911e+00 -8.10895860e-01 1.16493339e-02 -8.53820443e-01
-5.10209262e-01 1.36780977e+00 4.79105592e-01 -2.60537267e-01
4.37880665e-01 3.90025169e-01 4.08368230e-01 1.58331588e-01
-2.14063287e-01 7.82511756e-02 2.44853109e-01 8.03553224e-01
2.38900095e-01 -1.37332022e-01 -2.25996256e-01 -7.92005777e-01
7.00505897e-02 3.83887887e-02 9.17857945e-01 8.54479432e-01
-1.68683186e-01 -1.48076832e+00 -9.27095294e-01 -4.89703000e-01
1.81586653e-01 -6.67942911e-02 -5.87847292e-01 8.95533442e-01
-3.04110408e-01 5.16888142e-01 9.66082439e-02 -4.86608297e-01
5.74218392e-01 -1.06942423e-01 9.54374969e-01 -2.98546944e-02
-2.01185286e-01 4.92947549e-01 -4.57761943e-01 -9.65598643e-01
-5.16239583e-01 -6.73978552e-02 -8.58350635e-01 -7.88695514e-01
-1.35492146e-01 -1.34399757e-02 8.64220858e-01 2.61304192e-02
4.10780907e-01 4.38484885e-02 9.40352142e-01 -1.21177399e+00
9.98840760e-03 -1.60012305e-01 -9.27153707e-01 8.57822061e-01
3.41815233e-01 -4.02625918e-01 -2.91506976e-01 6.29234612e-01] | [9.72194766998291, -2.412860155105591] |
731b74c0-5568-4de9-965f-c3a492fe8863 | lung-nodule-classification-by-the-combination | 1712.02198 | null | http://arxiv.org/abs/1712.02198v2 | http://arxiv.org/pdf/1712.02198v2.pdf | Lung Nodule Classification by the Combination of Fusion Classifier and Cascaded Convolutional Neural Networks | Lung nodule classification is a class imbalanced problem, as nodules are
found with much lower frequency than non-nodules. In the class imbalanced
problem, conventional classifiers tend to be overwhelmed by the majority class
and ignore the minority class. We showed that cascaded convolutional neural
networks can classify the nodule candidates precisely for a class imbalanced
nodule candidate data set in our previous study. In this paper, we propose
Fusion classifier in conjunction with the cascaded convolutional neural network
models. To fuse the models, nodule probabilities are calculated by using the
convolutional neural network models at first. Then, Fusion classifier is
trained and tested by the nodule probabilities. The proposed method achieved
the sensitivity of 94.4% and 95.9% at 4 and 8 false positives per scan in Free
Receiver Operating Characteristics (FROC) curve analysis, respectively. | ['Taro Sekiyama', 'Masaharu Sakamoto', 'Kun Zhao', 'Hiroki Nakano'] | 2017-11-19 | null | null | null | null | ['lung-nodule-classification'] | ['medical'] | [ 2.08108768e-01 3.32995892e-01 -5.92226863e-01 -3.78118753e-01
-6.62096143e-01 -1.92575365e-01 2.01100260e-02 1.55384973e-01
-1.87660471e-01 6.51853263e-01 -1.42652467e-01 -5.74897051e-01
-2.30102032e-01 -9.80127633e-01 -3.99975508e-01 -5.93193114e-01
7.68482238e-02 5.38314819e-01 4.63356614e-01 2.05745101e-01
-2.79040933e-01 3.86621118e-01 -1.44423866e+00 1.01737988e+00
9.00373638e-01 1.23926687e+00 -2.62191832e-01 9.23449993e-01
-6.58624023e-02 1.22635472e+00 -4.55640137e-01 -2.36364037e-01
2.73293734e-01 -4.69323397e-01 -6.53287232e-01 1.66474313e-01
3.55300456e-01 -4.87548560e-01 -5.38853884e-01 9.04350400e-01
4.53156710e-01 -6.10802650e-01 8.31631303e-01 -1.28508055e+00
-1.46609187e-01 6.32059991e-01 -6.16390705e-01 3.97419542e-01
-6.03735089e-01 -9.67098624e-02 6.28962100e-01 -8.45538974e-01
7.10856542e-02 6.80304766e-01 9.15741146e-01 4.11467135e-01
-4.85431433e-01 -7.58338571e-01 -3.68348151e-01 -6.24087779e-03
-1.24732125e+00 7.46047199e-02 1.65044412e-01 -5.49016953e-01
6.03672683e-01 6.55431569e-01 1.02411187e+00 4.55547988e-01
6.97055519e-01 6.95776403e-01 6.48876667e-01 -1.99479342e-01
-2.66783059e-01 2.05914512e-01 1.19587407e-01 9.21414137e-01
1.02077496e+00 1.79361224e-01 9.04676393e-02 -2.12782428e-01
8.68103921e-01 5.43081701e-01 1.32333040e-01 -4.20370735e-02
-1.34852803e+00 7.86984921e-01 8.52280140e-01 4.19341773e-01
-3.01586270e-01 2.33256564e-01 4.27508444e-01 3.49007584e-02
3.87164831e-01 1.23854823e-01 -3.26821834e-01 6.65267587e-01
-8.08153629e-01 1.11561306e-01 7.52517581e-01 5.13542891e-01
2.12765217e-01 -2.08099082e-01 -7.20518351e-01 5.65154552e-01
3.70436281e-01 2.64552414e-01 8.96531522e-01 -4.81124133e-01
-2.52970420e-02 9.07784700e-01 -2.44578160e-02 -6.52496696e-01
-6.97481334e-01 -1.02862585e+00 -1.29295206e+00 1.74733534e-01
4.00808394e-01 -2.59704351e-01 -1.31171143e+00 1.12201095e+00
1.80493578e-01 2.05102190e-02 1.19558014e-01 5.89567840e-01
1.12279904e+00 -3.24096903e-02 1.01573013e-01 -1.77029639e-01
1.66888082e+00 -1.01849532e+00 -7.00186968e-01 2.92842984e-02
1.02175462e+00 -7.22541332e-01 3.46843660e-01 7.42019713e-02
-8.49193454e-01 -4.84532595e-01 -1.12171292e+00 4.93680805e-01
-4.62026782e-02 6.61553264e-01 5.90607762e-01 7.79322147e-01
-9.53232467e-01 4.04632747e-01 -9.67613697e-01 -1.91203207e-01
9.28807437e-01 6.85999095e-01 -1.97726831e-01 2.71130297e-02
-1.09741879e+00 8.68734896e-01 3.69680554e-01 1.85582221e-01
-8.95398140e-01 -8.29799533e-01 -4.38165516e-01 2.55551308e-01
6.96064765e-03 -9.55080032e-01 1.74200177e+00 -1.19074917e+00
-7.84561694e-01 8.01045775e-01 1.86983332e-01 -6.19930565e-01
6.10835373e-01 2.42729813e-01 -3.74766111e-01 -4.98780869e-02
-7.67008513e-02 4.65246260e-01 5.05407572e-01 -6.64419711e-01
-9.25062597e-01 -2.78727055e-01 -2.51785427e-01 2.19441399e-01
-7.89801553e-02 -3.89361560e-01 1.39948785e-01 -4.10406828e-01
4.67333913e-01 -8.58658016e-01 -4.46849227e-01 2.86254317e-01
-4.42555755e-01 -1.36585236e-01 1.05240333e+00 -3.56510818e-01
9.62659538e-01 -1.91254067e+00 -7.25035846e-01 1.93231508e-01
9.92612541e-01 4.57310826e-01 3.22729737e-01 -3.76858145e-01
-5.35816967e-01 5.16662240e-01 9.29923579e-02 5.58427334e-01
-5.67992270e-01 -1.18426748e-01 6.31668091e-01 4.65248674e-01
5.12690127e-01 1.03495455e+00 -7.81184256e-01 -7.16013014e-01
2.54017204e-01 2.66769588e-01 -5.38288832e-01 1.71912134e-01
3.16504776e-01 -8.11158046e-02 -4.12553757e-01 7.95769453e-01
7.10997224e-01 -7.52826452e-01 2.43502945e-01 -5.04032195e-01
2.29067877e-01 -1.15897030e-01 -8.94730628e-01 8.15277398e-01
-2.10492626e-01 5.60578525e-01 -4.50478435e-01 -8.07212651e-01
8.40802670e-01 6.31678581e-01 6.55506849e-01 -3.39953601e-01
3.90104741e-01 5.95389605e-01 9.18261409e-01 -6.90501750e-01
-8.89340267e-02 -5.52229881e-01 6.91860989e-02 1.78137541e-01
8.06543306e-02 2.28876434e-02 -1.85609847e-01 -7.79653937e-02
1.33132148e+00 -6.19778633e-01 8.76664877e-01 -1.75039396e-01
3.79769176e-01 1.98182002e-01 4.64049101e-01 8.98368716e-01
-5.22816062e-01 6.63208008e-01 7.91217744e-01 -8.07238579e-01
-8.48369896e-01 -1.00498772e+00 -4.51960951e-01 3.48316044e-01
-7.25587681e-02 1.93991348e-01 -4.69579041e-01 -1.08260751e+00
9.98844206e-02 -2.69827489e-02 -9.19779420e-01 -4.87129658e-01
-3.85535687e-01 -1.15218329e+00 6.97558105e-01 7.23957419e-01
6.80899382e-01 -6.80336833e-01 -7.06639230e-01 1.22728199e-01
-8.09729844e-02 -4.55666453e-01 8.37381110e-02 5.40908635e-01
-9.68153596e-01 -1.55718899e+00 -7.53325701e-01 -9.55595791e-01
8.16764355e-01 3.02106202e-01 1.11103547e+00 5.43016851e-01
-6.06235206e-01 -3.17303151e-01 -9.93049592e-02 -8.27375293e-01
-5.20025134e-01 2.61533797e-01 -2.77667135e-01 -1.64355949e-01
5.27403772e-01 -1.30309924e-01 -8.40241492e-01 3.29176575e-01
-7.38973796e-01 1.42908588e-01 1.19297993e+00 1.05354452e+00
5.88350296e-01 -8.20501766e-04 8.23082805e-01 -1.11840117e+00
1.56232595e-01 -6.71089470e-01 -2.58092910e-01 2.05275893e-01
-3.29085886e-01 -3.08537126e-01 1.76845551e-01 -3.48310113e-01
-7.54257321e-01 4.20990258e-01 3.11575104e-02 -3.49666148e-01
-1.12485521e-01 4.04284984e-01 3.88530910e-01 -1.55015469e-01
1.00037968e+00 -3.20932031e-01 3.23804080e-01 1.47535801e-01
-4.62174147e-01 9.41868126e-01 1.16556190e-01 3.37144285e-01
5.28825879e-01 2.87123740e-01 1.33244261e-01 -2.22842202e-01
-1.14851069e+00 -5.41710675e-01 -4.61844057e-01 -5.34306765e-01
7.40460575e-01 -1.00183928e+00 -4.85519230e-01 4.34751362e-01
-8.10646772e-01 1.08700879e-01 -4.18976665e-01 9.69764888e-01
-2.51869261e-01 -2.77298123e-01 -4.98480916e-01 -5.48481047e-01
-4.75318581e-01 -1.18795133e+00 7.84549654e-01 5.13993621e-01
-1.41189292e-01 -5.66609561e-01 -1.34565473e-01 -1.50697893e-02
7.82614172e-01 4.88633037e-01 9.53820825e-01 -1.35485125e+00
-4.26765293e-01 -1.00279236e+00 -6.13602161e-01 3.57242763e-01
5.96635461e-01 2.08810627e-01 -9.68348324e-01 -8.88424218e-02
7.53783509e-02 -1.38162732e-01 9.14714098e-01 7.05132961e-01
1.60707068e+00 -8.06638896e-02 -7.75861800e-01 3.30765337e-01
1.56882429e+00 3.16031843e-01 6.54551566e-01 -1.48705766e-01
4.86784130e-01 9.60159153e-02 1.99058354e-01 2.26773188e-01
-1.89887598e-01 -1.28276825e-01 8.15571308e-01 -5.40965974e-01
-4.36878383e-01 1.62136853e-01 -4.49160069e-01 5.98480344e-01
-3.40999179e-02 -4.24443334e-01 -1.23816931e+00 5.81022441e-01
-1.57955492e+00 -6.21051908e-01 -4.38775867e-01 1.99067020e+00
5.60036004e-01 3.73111784e-01 -2.11519763e-01 2.25347817e-01
1.05372155e+00 -3.51265639e-01 -5.08969307e-01 8.46493766e-02
1.49797142e-01 1.55843854e-01 6.47455454e-01 1.20388471e-01
-1.41536331e+00 1.61725566e-01 6.86477900e+00 6.93699539e-01
-1.17875874e+00 1.02298722e-01 1.28990006e+00 -1.43118158e-01
-1.56994034e-02 -3.24027628e-01 -5.21386027e-01 1.37783632e-01
8.96407187e-01 -2.59105831e-01 -5.99233210e-01 8.50879192e-01
-2.83086360e-01 -2.72437423e-01 -8.75591397e-01 5.71107447e-01
-6.30960092e-02 -1.47026861e+00 -6.27349392e-02 2.59045814e-03
7.92657077e-01 1.29376471e-01 1.09775908e-01 3.99456263e-01
1.53309196e-01 -1.32750332e+00 -9.42787230e-02 5.25210023e-01
9.45121467e-01 -5.67381382e-01 1.69272804e+00 2.71647304e-01
-8.68382812e-01 -2.66986582e-02 -2.25060999e-01 3.32619995e-02
-5.68639457e-01 9.50232923e-01 -1.50039160e+00 4.79658365e-01
7.42585719e-01 4.57981437e-01 -6.88995898e-01 1.53163731e+00
3.54713291e-01 6.81310356e-01 -3.85532230e-01 -2.86958933e-01
1.24966130e-01 6.54084563e-01 1.76378384e-01 1.00226271e+00
4.59461719e-01 7.87585899e-02 -6.29618689e-02 5.06077111e-01
-2.33856872e-01 1.21374026e-01 -4.83557463e-01 1.05903536e-01
1.58774897e-01 1.61748767e+00 -1.12111747e+00 -4.89115208e-01
-3.15596372e-01 3.19242626e-01 -1.67687505e-01 -7.30014145e-02
-1.13122284e+00 -5.34268081e-01 -4.60479185e-02 3.18130821e-01
2.03423053e-01 6.87983751e-01 -5.25000989e-01 -8.84906292e-01
-2.34040692e-01 -5.09352803e-01 4.64738041e-01 -4.21094328e-01
-1.29320300e+00 6.00359917e-01 -3.46821696e-01 -1.66824377e+00
-1.29715085e-01 -7.68250823e-01 -7.64363825e-01 6.59992933e-01
-1.42038143e+00 -1.08205962e+00 -8.52959573e-01 1.54483885e-01
3.28598171e-01 -1.10066153e-01 7.30858326e-01 2.86053300e-01
-3.11204284e-01 6.76896691e-01 -1.62066147e-02 5.00212729e-01
7.03142822e-01 -1.10693598e+00 -2.37316519e-01 3.35628867e-01
-5.81759095e-01 -6.52960092e-02 9.14564505e-02 -7.58529067e-01
-5.77506185e-01 -1.50022781e+00 8.92867208e-01 -2.23554403e-01
2.71823794e-01 1.97414860e-01 -8.47683191e-01 5.95149755e-01
-1.56094208e-01 7.46408641e-01 1.03274441e+00 -2.76142448e-01
-1.78922683e-01 -4.55328338e-02 -1.51288772e+00 1.49499446e-01
4.92903560e-01 1.29330903e-01 -1.74072310e-01 5.86747825e-01
7.27279186e-01 -7.69875765e-01 -9.28664684e-01 1.13453472e+00
8.01717341e-01 -9.76920903e-01 7.94178784e-01 -7.38853395e-01
6.19100153e-01 -3.55826885e-01 -9.20754224e-02 -8.23612630e-01
-5.92405558e-01 4.45556819e-01 3.21242698e-02 3.99402469e-01
8.53250325e-01 -3.52067381e-01 1.04650378e+00 -5.80479577e-03
-1.84270173e-01 -1.32250440e+00 -8.76429200e-01 -2.78549224e-01
1.72306538e-01 1.57807991e-02 4.64089483e-01 6.85397208e-01
-4.90866750e-01 -1.63837463e-01 8.85674208e-02 1.27916843e-01
2.75070935e-01 -1.84626356e-01 -3.60421725e-02 -1.51325989e+00
1.71491206e-02 -4.43470925e-01 -7.21769571e-01 6.46709464e-03
-5.08011341e-01 -1.03884423e+00 -2.32072044e-02 -1.64227664e+00
6.49323404e-01 -4.26927805e-01 -7.02018917e-01 5.64970613e-01
-3.64625692e-01 5.53759217e-01 -2.58675069e-01 3.22878748e-01
-4.44721162e-01 6.16938667e-03 1.49334168e+00 -3.80981356e-01
1.12435661e-01 5.93209207e-01 -8.54993701e-01 8.80749702e-01
1.13760412e+00 -7.40396500e-01 -2.18263835e-01 -3.54367346e-02
1.67089269e-01 2.40692422e-01 4.09286141e-01 -1.38216186e+00
2.11767033e-01 -6.23779278e-03 1.30315351e+00 -1.06178093e+00
-1.33686200e-01 -7.61058748e-01 1.22777797e-01 1.34140790e+00
-2.07534790e-01 -3.71013582e-01 1.97005376e-01 5.25432169e-01
-3.61283213e-01 -2.39642367e-01 9.70046878e-01 -3.54284734e-01
-3.16598229e-02 4.21830773e-01 -6.95577025e-01 -4.96627629e-01
1.34690344e+00 -2.79401720e-01 -2.81945735e-01 -2.22513620e-02
-8.65123868e-01 1.18929371e-01 -2.12690651e-01 8.92639756e-02
6.32858992e-01 -1.53696299e+00 -9.15545523e-01 2.24358067e-01
1.61661521e-01 3.14800203e-01 5.43263555e-01 1.22170198e+00
-7.04369307e-01 3.22623312e-01 -3.61657917e-01 -1.04192781e+00
-1.15441763e+00 3.09276849e-01 1.21375251e+00 -6.28100216e-01
-2.43906397e-02 8.58231425e-01 2.02023834e-01 -5.21162927e-01
3.72364223e-02 -5.41947305e-01 -2.97633529e-01 3.63611765e-02
5.40576458e-01 2.39445999e-01 6.24031544e-01 -5.25956079e-02
-3.81419033e-01 1.75769344e-01 -3.54394972e-01 4.37879771e-01
8.89450788e-01 5.15600920e-01 -1.92663938e-01 3.94747913e-01
1.12875402e+00 -4.75358903e-01 -5.09856641e-01 -8.60560015e-02
-1.76965967e-01 -1.49011746e-01 1.04076535e-01 -1.01856911e+00
-1.31508529e+00 7.03382194e-01 1.20440495e+00 2.69333452e-01
1.09862733e+00 1.28780782e-01 3.94320488e-01 2.99023569e-01
-3.59571457e-01 -3.56326401e-01 1.70371413e-01 2.87428558e-01
4.49926913e-01 -1.49102640e+00 3.00927330e-02 -6.41162932e-01
-4.00644332e-01 1.38922763e+00 1.18641484e+00 -4.89642590e-01
1.08358669e+00 5.37654519e-01 1.02828026e-01 -2.86509812e-01
-1.00165963e+00 -8.56395960e-02 2.58157730e-01 5.06274879e-01
6.90292835e-01 3.96566600e-01 -2.61099666e-01 1.00791073e+00
1.38027743e-01 4.48244393e-01 6.49677753e-01 9.51677144e-01
-8.44014347e-01 -6.32812023e-01 -3.85156751e-01 1.55107093e+00
-7.16991246e-01 1.49609167e-02 -2.86846548e-01 1.00761890e+00
4.72617596e-01 4.53293592e-01 5.13331950e-01 -6.01498127e-01
3.58717322e-01 1.03487931e-01 2.52838850e-01 -6.74271524e-01
-9.85234261e-01 1.92167431e-01 -1.18008599e-01 -2.20988125e-01
-3.94431621e-01 -2.44264737e-01 -1.07662225e+00 5.68655320e-02
-8.78031611e-01 -1.96759254e-02 3.35746676e-01 7.52328455e-01
7.40952343e-02 1.25772679e+00 6.74865901e-01 -2.27420464e-01
-6.04568779e-01 -1.16545570e+00 -5.23802459e-01 -1.69682696e-01
6.60333753e-01 -5.87622702e-01 -4.47539955e-01 -1.84119955e-01] | [15.387218475341797, -2.210686445236206] |
aa03b978-9591-4c71-bea4-859060c4f107 | segnetr-rethinking-the-local-global | 2307.02953 | null | https://arxiv.org/abs/2307.02953v1 | https://arxiv.org/pdf/2307.02953v1.pdf | SegNetr: Rethinking the local-global interactions and skip connections in U-shaped networks | Recently, U-shaped networks have dominated the field of medical image segmentation due to their simple and easily tuned structure. However, existing U-shaped segmentation networks: 1) mostly focus on designing complex self-attention modules to compensate for the lack of long-term dependence based on convolution operation, which increases the overall number of parameters and computational complexity of the network; 2) simply fuse the features of encoder and decoder, ignoring the connection between their spatial locations. In this paper, we rethink the above problem and build a lightweight medical image segmentation network, called SegNetr. Specifically, we introduce a novel SegNetr block that can perform local-global interactions dynamically at any stage and with only linear complexity. At the same time, we design a general information retention skip connection (IRSC) to preserve the spatial location information of encoder features and achieve accurate fusion with the decoder features. We validate the effectiveness of SegNetr on four mainstream medical image segmentation datasets, with 59\% and 76\% fewer parameters and GFLOPs than vanilla U-Net, while achieving segmentation performance comparable to state-of-the-art methods. Notably, the components proposed in this paper can be applied to other U-shaped networks to improve their segmentation performance. | ['Min Zhu', 'Fengjie Wang', 'Chengrui Gao', 'Junlong Cheng'] | 2023-07-06 | null | null | null | null | ['medical-image-segmentation'] | ['medical'] | [ 1.96443528e-01 2.97805250e-01 -1.90699637e-01 -3.92258048e-01
-4.10613447e-01 -2.46896237e-01 -1.18623078e-02 2.26381451e-01
-8.02572668e-01 3.62727255e-01 -4.00876962e-02 -5.11733472e-01
8.36075693e-02 -7.84557104e-01 -6.16043150e-01 -5.67898929e-01
-4.42380086e-02 5.16236993e-04 7.87966192e-01 -1.50262371e-01
2.46565230e-02 3.95774931e-01 -1.07535017e+00 1.56469628e-01
1.02568305e+00 1.07800305e+00 3.85551363e-01 5.62538326e-01
-2.21392646e-01 7.28909373e-01 -5.15626073e-01 -2.40016550e-01
1.37554362e-01 -4.61785108e-01 -8.36228907e-01 3.42445485e-02
7.29779676e-02 -5.15341222e-01 -7.07690418e-01 1.06757212e+00
6.96340740e-01 -3.37723014e-03 2.43509501e-01 -6.53971136e-01
-5.95273614e-01 8.57491672e-01 -6.60494387e-01 4.41590875e-01
-2.38882482e-01 1.57817200e-01 6.81659520e-01 -3.39230478e-01
4.65563983e-01 9.32410479e-01 7.06220329e-01 6.62346363e-01
-9.65605974e-01 -5.56171000e-01 3.89117151e-01 1.83413789e-01
-1.47887039e+00 -2.74136990e-01 5.62536716e-01 -9.53997970e-02
8.39935482e-01 1.95085227e-01 5.67965388e-01 5.25686502e-01
2.17006177e-01 1.01657915e+00 5.02209485e-01 -5.72653525e-02
5.80896772e-02 -1.67348459e-01 2.93622762e-01 8.33793581e-01
5.49246557e-02 -2.84482419e-01 8.28946009e-03 3.68393064e-01
1.06309426e+00 3.16587687e-01 -4.77879137e-01 -6.97219446e-02
-1.23544228e+00 7.01514125e-01 1.13137949e+00 5.73972702e-01
-1.43949434e-01 2.77551353e-01 4.74767119e-01 8.45399648e-02
3.76221657e-01 3.04401666e-01 -4.28196818e-01 1.15625463e-01
-1.09752154e+00 -1.98026687e-01 4.86931533e-01 9.75014746e-01
7.69397557e-01 -1.34700969e-01 -4.61986393e-01 8.00275266e-01
6.40566647e-02 1.17521040e-01 7.73301601e-01 -5.85273385e-01
3.30540925e-01 8.69858444e-01 -3.90092731e-01 -8.93782437e-01
-8.28220606e-01 -7.86486208e-01 -1.17526376e+00 -3.05516273e-01
1.79741204e-01 -3.74329329e-01 -1.35989738e+00 1.53607845e+00
2.10542262e-01 3.33727539e-01 -2.26699281e-02 8.67705226e-01
1.18952179e+00 5.46637535e-01 -4.54357229e-02 1.70699626e-01
1.34134078e+00 -1.31671047e+00 -6.61832988e-01 -2.95570880e-01
9.25877333e-01 -6.00719929e-01 7.39171147e-01 -7.72703961e-02
-1.27228153e+00 -6.71412945e-01 -1.15816009e+00 -3.75535935e-01
-2.64799237e-01 1.49750099e-01 6.31374359e-01 5.98226488e-01
-1.27703869e+00 7.32616246e-01 -1.01529765e+00 -5.93091324e-02
6.72256529e-01 7.17889607e-01 2.20853481e-02 -6.52251467e-02
-1.10476005e+00 5.39710641e-01 5.32314062e-01 4.17163312e-01
-4.99658376e-01 -7.78228045e-01 -9.22694921e-01 3.20356637e-01
3.74510467e-01 -6.30605400e-01 1.15464425e+00 -8.32956135e-01
-1.29542077e+00 5.03088117e-01 -9.06936377e-02 -4.83463883e-01
4.26945865e-01 3.38375680e-02 -1.94675401e-01 2.75138855e-01
5.42712025e-02 1.09422958e+00 2.62864470e-01 -9.22345638e-01
-6.57930911e-01 -3.12921077e-01 2.19947368e-01 1.95984229e-01
-4.25654411e-01 -2.47477844e-01 -1.12896812e+00 -8.77612591e-01
3.55953395e-01 -7.49016881e-01 -6.25035703e-01 1.06643900e-01
-4.89564061e-01 6.65667504e-02 7.64967561e-01 -5.91873944e-01
1.47956657e+00 -2.40734243e+00 -6.60646334e-02 9.77693573e-02
4.67937738e-01 6.41490281e-01 -1.78213567e-01 -4.63860817e-02
1.85740843e-01 1.30227923e-01 -5.15681565e-01 -2.71119505e-01
-4.02628541e-01 3.69476676e-01 2.09086716e-01 3.97507370e-01
1.44342855e-01 1.22832549e+00 -8.38423431e-01 -7.51904309e-01
3.44506770e-01 4.84736174e-01 -8.79401803e-01 -2.43825391e-02
2.53068209e-02 3.92617077e-01 -5.13877034e-01 5.42422414e-01
7.37375617e-01 -4.72680181e-01 1.20979816e-01 -3.73062223e-01
-1.15790561e-01 1.53384566e-01 -8.75542521e-01 1.79914665e+00
-4.83881682e-01 4.58151847e-01 2.36821428e-01 -1.18197536e+00
8.25146616e-01 2.78990507e-01 6.74116790e-01 -8.27269793e-01
4.57267702e-01 1.97954059e-01 2.42512986e-01 -3.54343563e-01
2.95236021e-01 1.27550229e-01 -1.08615411e-02 6.03207238e-02
2.01905239e-02 2.43520111e-01 2.60465890e-01 1.01476029e-01
1.01479661e+00 -2.27932349e-01 1.86740443e-01 -3.49396884e-01
6.63203478e-01 -2.09292799e-01 7.32796848e-01 7.17529237e-01
-2.88197100e-01 8.40245366e-01 5.60025871e-01 -4.14176077e-01
-6.55299485e-01 -8.12524199e-01 -2.53166497e-01 8.90076399e-01
5.94585240e-01 -3.19121748e-01 -9.78493273e-01 -6.81327105e-01
-3.00305039e-01 3.07718217e-01 -6.51095867e-01 -6.18846081e-02
-8.40228140e-01 -9.07956481e-01 7.79443443e-01 8.59347522e-01
9.60493982e-01 -1.00139344e+00 -7.85293639e-01 5.48249424e-01
-1.12635627e-01 -1.09501636e+00 -8.70334148e-01 2.63077170e-01
-9.84105110e-01 -1.01610053e+00 -1.03772688e+00 -1.03599346e+00
8.95406663e-01 2.88865268e-01 8.88220191e-01 3.85518283e-01
-3.53525758e-01 -3.06656003e-01 -4.52855378e-01 -1.33220568e-01
-2.99897511e-02 4.84207153e-01 -7.04826832e-01 -1.45516366e-01
7.60845393e-02 -4.66104358e-01 -1.10798204e+00 4.91299719e-01
-1.10516214e+00 3.29140157e-01 8.27538610e-01 9.20475602e-01
6.17275119e-01 -7.50430301e-02 5.43947518e-01 -1.06558335e+00
3.27735037e-01 -3.67899060e-01 -3.46175104e-01 2.32627064e-01
-4.15136486e-01 -6.86047301e-02 6.98935628e-01 -2.74147898e-01
-7.74346471e-01 3.94401513e-02 -5.35706043e-01 -2.75673896e-01
1.42494678e-01 4.48754698e-01 -1.27202734e-01 -1.87754273e-01
2.41815910e-01 3.45990062e-01 1.04448207e-01 -4.62337106e-01
2.97141761e-01 6.55403554e-01 5.91794133e-01 -1.20099179e-01
2.34436661e-01 4.94488150e-01 -2.33553365e-01 -6.20291293e-01
-8.15756500e-01 -4.91149932e-01 -5.89156985e-01 8.81457925e-02
9.91530240e-01 -8.57957304e-01 -7.81749070e-01 6.37506187e-01
-1.05890608e+00 -4.42370236e-01 -2.46687368e-01 3.38377059e-01
-1.20659970e-01 4.65532959e-01 -9.50746059e-01 -1.39030173e-01
-5.95017076e-01 -1.63593149e+00 9.59861875e-01 5.13472378e-01
8.23429301e-02 -9.82176423e-01 -4.76716459e-01 2.68271178e-01
6.40888214e-01 8.41691941e-02 8.64940822e-01 -4.82348889e-01
-5.46886623e-01 -1.01128928e-01 -8.06056023e-01 3.49926919e-01
2.54030317e-01 -3.11800361e-01 -6.90819621e-01 -3.44762146e-01
-9.65056568e-02 1.70964211e-01 1.22720134e+00 6.99389160e-01
1.54409897e+00 -2.38270268e-01 -4.63793039e-01 1.05074441e+00
1.30917287e+00 4.11632210e-01 6.21754169e-01 1.52033478e-01
9.05154109e-01 2.22762793e-01 3.70270163e-02 2.03308403e-01
5.42118728e-01 4.25350010e-01 4.52747881e-01 -7.95642734e-01
-3.21821213e-01 5.50199188e-02 -1.25879660e-01 1.05902278e+00
1.88786194e-01 -3.98293227e-01 -8.40836048e-01 6.94112062e-01
-1.99231577e+00 -4.00680810e-01 9.61330906e-02 1.85117853e+00
8.92345846e-01 1.48101598e-01 -1.64081797e-01 -3.06236092e-02
7.53136516e-01 2.69354165e-01 -6.75154805e-01 -2.90909797e-01
-1.04221543e-02 3.59722048e-01 8.77871513e-01 3.59527707e-01
-1.25783479e+00 8.45464230e-01 6.06409216e+00 1.01960123e+00
-1.31669986e+00 1.81165501e-01 9.94525552e-01 -1.86759859e-01
-1.36187971e-01 -3.42646688e-01 -6.38501942e-01 4.83421177e-01
5.84828019e-01 1.26654655e-01 1.42068580e-01 6.12143636e-01
-9.05064587e-03 4.17137295e-02 -8.27271342e-01 8.79960716e-01
-7.66644925e-02 -1.48935020e+00 7.08195008e-03 -1.13017932e-01
6.52696848e-01 1.53666794e-01 -1.26759723e-01 2.57399827e-01
8.92582163e-02 -9.75062251e-01 4.20071810e-01 1.24468371e-01
7.52178669e-01 -8.62417638e-01 1.04662251e+00 2.10754931e-01
-1.28191459e+00 2.22702902e-02 -3.10232759e-01 1.32310778e-01
3.47928047e-01 6.11977637e-01 -4.71337348e-01 6.26631081e-01
6.27600372e-01 7.53502965e-01 -6.44681573e-01 1.21979511e+00
5.26074953e-02 4.45433259e-01 -3.23905021e-01 2.30669510e-04
7.85979986e-01 -3.15058157e-02 1.22593068e-01 1.25449991e+00
2.44647995e-01 1.36665106e-01 2.67512679e-01 6.29161060e-01
-3.02848637e-01 6.08495064e-02 -2.48812679e-02 3.76851827e-01
2.78324813e-01 1.04285502e+00 -1.17384923e+00 -4.40867305e-01
-4.84688312e-01 9.27050114e-01 2.08687320e-01 2.07566589e-01
-9.48110998e-01 -5.96301377e-01 5.66402912e-01 1.27271980e-01
6.05894566e-01 -3.48465256e-02 -5.51149368e-01 -1.02642739e+00
-3.74818780e-02 -5.65111220e-01 3.16654295e-01 -2.68576562e-01
-8.01630437e-01 9.23485279e-01 -2.83993959e-01 -9.10621405e-01
3.19368720e-01 -3.67021054e-01 -5.27846575e-01 6.03474379e-01
-1.66476238e+00 -9.77557242e-01 -3.59154880e-01 5.18477678e-01
4.99376863e-01 2.36717761e-01 4.33588654e-01 6.55347645e-01
-9.50299263e-01 8.39472711e-01 1.44350976e-01 4.68877763e-01
5.05555809e-01 -1.12827241e+00 6.48769021e-01 8.63903463e-01
-2.53974646e-01 6.54838324e-01 1.08103670e-01 -4.71116513e-01
-1.01630402e+00 -1.35724545e+00 5.78933060e-01 2.87378579e-01
3.27188432e-01 -2.11781427e-01 -8.72393012e-01 4.84001130e-01
1.21479169e-01 3.06653887e-01 4.43601310e-01 -2.39781424e-01
-2.56149494e-03 -2.36763477e-01 -1.12744832e+00 7.66871929e-01
1.00427461e+00 -1.15063630e-01 -1.71539247e-01 9.63755921e-02
1.03876555e+00 -7.19146013e-01 -8.19927514e-01 6.39003158e-01
3.14320832e-01 -9.75328267e-01 9.28969085e-01 -4.51942682e-02
4.35150027e-01 -3.70155305e-01 2.25382045e-01 -8.90490770e-01
-4.03957456e-01 -5.32867372e-01 2.59666175e-01 8.72470737e-01
4.16758031e-01 -7.64038980e-01 8.70977044e-01 3.73105824e-01
-5.23332357e-01 -1.27321839e+00 -1.00976884e+00 -3.68871510e-01
1.52745089e-02 -1.94982693e-01 6.28809750e-01 6.77319586e-01
-2.21598044e-01 1.56718895e-01 -2.20158309e-01 2.02735178e-02
3.04608226e-01 5.70607744e-02 2.21321225e-01 -7.57590771e-01
-2.96772450e-01 -6.68698430e-01 -4.14262235e-01 -1.63299513e+00
-2.71704078e-01 -9.40703154e-01 1.63668230e-01 -1.58074737e+00
1.26441792e-01 -6.08107269e-01 -4.76742059e-01 6.87317431e-01
-3.81508738e-01 5.32660723e-01 2.68238157e-01 1.95281301e-02
-6.48352265e-01 4.43691969e-01 1.72447515e+00 -1.41500682e-01
-2.38168031e-01 -7.15958253e-02 -8.47853482e-01 7.11010456e-01
7.15500653e-01 -2.35470027e-01 -4.89306420e-01 -8.19570303e-01
-2.78985769e-01 1.50264546e-01 2.14543983e-01 -1.07944715e+00
4.68127728e-01 2.47849733e-01 2.71529675e-01 -7.10014164e-01
-1.07703507e-02 -7.11931884e-01 -2.47916549e-01 6.50370777e-01
-1.84639707e-01 -3.22636403e-02 3.08491141e-01 3.17002445e-01
-3.95145625e-01 -1.84882790e-01 9.47731376e-01 -1.84681594e-01
-6.61924779e-01 5.88877201e-01 -3.42753202e-01 -1.49345733e-02
1.04891479e+00 -3.75957042e-01 -2.73499310e-01 1.43343378e-02
-6.56405210e-01 5.04890501e-01 3.34208131e-01 1.98116168e-01
5.80780745e-01 -8.79301310e-01 -3.92283767e-01 3.03365588e-01
-1.63599238e-01 6.05954528e-01 7.51261950e-01 1.04134452e+00
-1.03602779e+00 5.17621458e-01 -1.37996972e-01 -5.83908617e-01
-9.35028076e-01 5.55368125e-01 4.78622705e-01 -4.66088772e-01
-1.00103354e+00 9.55943882e-01 6.54801846e-01 -3.04025620e-01
2.17073232e-01 -6.02116525e-01 -2.12897897e-01 -5.09004183e-02
5.01872003e-01 1.83345988e-01 1.01622418e-01 -3.79103214e-01
-4.62562948e-01 5.17092884e-01 -3.38883609e-01 3.59859139e-01
1.17366493e+00 -1.40363559e-01 -1.77835375e-01 -1.20751113e-01
1.35294688e+00 -5.19855142e-01 -1.39798272e+00 -3.32522392e-01
-2.59238452e-01 -6.72388002e-02 4.09053445e-01 -6.11349225e-01
-1.85605264e+00 8.67663860e-01 6.32488549e-01 -4.76991795e-02
1.35602570e+00 -1.25908971e-01 1.25936806e+00 9.06464085e-02
4.31400836e-02 -9.01941717e-01 -2.82847643e-01 3.79717648e-01
3.46329838e-01 -1.00429606e+00 -7.98215121e-02 -6.15953565e-01
-4.51769710e-01 1.00503409e+00 6.17413819e-01 -2.09171444e-01
8.23964059e-01 4.91741657e-01 1.52025253e-01 -1.36202231e-01
-3.69510353e-01 -2.61170775e-01 1.54938444e-01 2.98783362e-01
6.19406343e-01 -7.28729889e-02 -5.04203737e-01 6.52517557e-01
7.31477141e-02 9.46136639e-02 4.14747119e-01 8.11252117e-01
-2.57500380e-01 -9.21503663e-01 1.36650085e-01 6.08283758e-01
-6.87674224e-01 -2.98935115e-01 1.78108126e-01 7.37011075e-01
2.96784729e-01 7.82791555e-01 2.94582665e-01 -2.58115500e-01
3.23772520e-01 -5.76599121e-01 1.92599103e-01 -4.74128217e-01
-9.38132226e-01 3.73356253e-01 -1.94015414e-01 -6.04118049e-01
-3.35429579e-01 -2.89824933e-01 -1.41177320e+00 -2.13420913e-01
-3.53441149e-01 -3.71884294e-02 4.15370554e-01 8.57802093e-01
5.33658206e-01 1.09792531e+00 3.79141420e-01 -7.78923213e-01
-2.89221019e-01 -8.39477539e-01 -4.01283592e-01 1.67076185e-01
3.89115542e-01 -2.74287254e-01 9.87839550e-02 -1.30116686e-01] | [14.620824813842773, -2.601945161819458] |
76816490-7876-4a82-92bd-7c3ff6065fab | tablegpt-few-shot-table-to-text-generation | null | null | https://aclanthology.org/2020.coling-main.179 | https://aclanthology.org/2020.coling-main.179.pdf | TableGPT: Few-shot Table-to-Text Generation with Table Structure Reconstruction and Content Matching | Although neural table-to-text models have achieved remarkable progress with the help of large-scale datasets, they suffer insufficient learning problem with limited training data. Recently, pre-trained language models show potential in few-shot learning with linguistic knowledge learnt from pretraining on large-scale corpus. However, benefiting table-to-text generation in few-shot setting with the powerful pretrained language model faces three challenges, including (1) the gap between the task{'}s structured input and the natural language input for pretraining language model. (2) The lack of modeling for table structure and (3) improving text fidelity with less incorrect expressions that are contradicting to the table. To address aforementioned problems, we propose TableGPT for table-to-text generation. At first, we utilize table transformation module with template to rewrite structured table in natural language as input for GPT-2. In addition, we exploit multi-task learning with two auxiliary tasks that preserve table{'}s structural information by reconstructing the structure from GPT-2{'}s representation and improving the text{'}s fidelity with content matching task aligning the table and information in the generated text. By experimenting on Humans, Songs and Books, three few-shot table-to-text datasets in different domains, our model outperforms existing systems on most few-shot settings. | ['Ting Liu', 'Xiaojiang Liu', 'Wei Bi', 'Bing Qin', 'Xiaocheng Feng', 'Yawei Sun', 'Heng Gong'] | 2020-12-01 | null | null | null | coling-2020-8 | ['table-to-text-generation'] | ['natural-language-processing'] | [ 6.66606247e-01 4.51415628e-01 -2.78187156e-01 -3.53926629e-01
-1.45919740e+00 -5.07778108e-01 5.38441181e-01 8.97057503e-02
-2.15044782e-01 1.09285939e+00 6.97235823e-01 -2.21483365e-01
2.58047760e-01 -1.19042349e+00 -1.11712325e+00 -9.07656252e-02
6.46478236e-01 1.05080605e+00 1.19387276e-01 -9.22157288e-01
2.04336837e-01 -2.86343634e-01 -1.36917663e+00 9.25572455e-01
1.02236390e+00 8.05749655e-01 3.92065704e-01 5.64019978e-01
-8.47523391e-01 1.07294095e+00 -6.80861175e-01 -7.87677109e-01
4.54316467e-01 -8.48421872e-01 -8.10348988e-01 -1.98519781e-01
4.90553260e-01 -4.40305382e-01 -2.89324403e-01 7.12534308e-01
6.84839785e-01 3.54491860e-01 6.58115447e-01 -1.04311347e+00
-1.19916308e+00 1.36874008e+00 -3.73544335e-01 2.69533880e-02
5.54258943e-01 2.06082270e-01 1.21586967e+00 -1.14561784e+00
9.65291142e-01 1.45541167e+00 4.99296963e-01 9.60949123e-01
-1.23038983e+00 -6.11290276e-01 -1.39463738e-01 8.63269269e-02
-1.33536005e+00 -8.71882379e-01 4.93978769e-01 -1.01529002e-01
1.53876698e+00 9.80369970e-02 1.08877175e-01 1.33923054e+00
2.49819264e-01 8.11662912e-01 5.54777086e-01 -4.78836536e-01
3.51258823e-05 3.65168273e-01 -3.50369066e-01 8.22056234e-01
2.33996868e-01 -2.95716941e-01 -8.62038195e-01 1.79741800e-01
4.67086583e-01 -2.66586870e-01 7.60591254e-02 -8.20065513e-02
-1.30466378e+00 7.68615842e-01 -1.12911001e-01 1.68936178e-01
5.96616231e-03 -4.66374643e-02 6.21992052e-01 4.82448399e-01
2.32049838e-01 4.69659597e-01 -5.58440864e-01 -1.62894294e-01
-8.75437081e-01 4.56563443e-01 8.90504658e-01 1.87704563e+00
8.45979869e-01 3.69042486e-01 -9.53725755e-01 8.14853191e-01
-1.51455775e-01 6.94010913e-01 7.74322093e-01 -3.84140015e-01
1.49712205e+00 6.50511682e-01 -1.32922277e-01 -6.91669643e-01
5.74161150e-02 2.95536267e-03 -1.17070663e+00 -3.38690072e-01
3.01933795e-01 -2.48245999e-01 -7.80320287e-01 1.60523927e+00
-5.64249093e-03 -3.08082283e-01 5.55857956e-01 1.87371626e-01
1.05362618e+00 1.05128813e+00 -2.79035598e-01 -1.93873182e-01
1.33292544e+00 -1.10493433e+00 -9.73656118e-01 -5.95695972e-01
9.56130624e-01 -6.75083220e-01 1.62952101e+00 -2.20972411e-02
-1.25267005e+00 -7.97613978e-01 -9.72038329e-01 -7.24020362e-01
-6.99591458e-01 1.79350242e-01 3.48734021e-01 5.22209704e-01
-8.93987894e-01 5.67376792e-01 -2.82306999e-01 -4.91342187e-01
4.72931832e-01 4.76833843e-02 -3.03054661e-01 -2.49005854e-01
-1.43500721e+00 8.42665792e-01 6.83838665e-01 -2.63839453e-01
-7.72628844e-01 -1.02869856e+00 -1.31394660e+00 3.76292199e-01
8.67735505e-01 -9.03832436e-01 1.39371169e+00 -2.83609360e-01
-1.53397501e+00 6.77326620e-01 -3.46985638e-01 -5.72747767e-01
4.72542644e-01 -1.19574249e-01 -2.15887189e-01 -4.01959866e-01
4.65424120e-01 6.64588034e-01 6.29841745e-01 -8.83972466e-01
-5.64101756e-01 -2.42596313e-01 -2.97981620e-01 3.08140308e-01
-1.51488125e-01 -7.31768012e-02 -4.09719497e-01 -9.26354706e-01
-2.09325299e-01 -3.73829246e-01 -3.91588248e-02 -8.71468335e-02
-5.73904514e-01 -3.16369921e-01 5.38464963e-01 -7.64857829e-01
1.41862595e+00 -1.69840908e+00 -9.54107940e-02 -3.72519463e-01
-2.27613524e-01 2.40135491e-01 -4.11905557e-01 9.18304741e-01
1.65685400e-01 2.98427641e-01 -3.92262489e-01 -5.10128200e-01
2.32685804e-01 2.40124494e-01 -8.81912291e-01 -4.02671456e-01
5.15442193e-01 1.30714643e+00 -7.94437647e-01 -7.50839770e-01
5.62112518e-02 -2.13867985e-02 -8.08848679e-01 4.31138635e-01
-5.37410855e-01 3.00558712e-02 -3.07626843e-01 5.90270877e-01
3.02858204e-01 -1.49313554e-01 -7.04914033e-02 -1.01990357e-01
2.42197037e-01 5.38223982e-01 -1.25263560e+00 2.23930144e+00
-6.52149200e-01 2.50809491e-01 -5.45070767e-01 -7.03070760e-01
1.00205946e+00 4.21887398e-01 8.62218812e-02 -8.67859900e-01
2.44133081e-02 1.22433074e-01 -1.44574255e-01 -5.77039659e-01
8.04575562e-01 -3.52581114e-01 -5.09334683e-01 8.45941782e-01
5.79277515e-01 -4.86333638e-01 5.56942761e-01 5.56532025e-01
1.03617072e+00 4.77389812e-01 4.76047903e-01 6.00803569e-02
4.36916888e-01 1.52117729e-01 3.47138524e-01 1.01164985e+00
2.44958729e-01 7.35231519e-01 3.85934889e-01 -3.43826205e-01
-1.31792068e+00 -7.97467232e-01 2.61941493e-01 1.49398899e+00
-2.69027561e-01 -7.98997223e-01 -8.01676750e-01 -5.70676744e-01
-2.13182166e-01 1.37161934e+00 -7.12700069e-01 -4.12017286e-01
-7.90711701e-01 -4.29752946e-01 8.45368207e-01 6.64018393e-01
4.77329910e-01 -1.22920489e+00 -2.86674500e-01 4.54326570e-01
-4.79627728e-01 -1.26878536e+00 -1.04099381e+00 2.93941230e-01
-7.78845370e-01 -6.45494998e-01 -4.12049621e-01 -8.30258548e-01
6.12318337e-01 1.62179917e-01 1.33238983e+00 -2.40243465e-01
-9.10562798e-02 -1.56109005e-01 -2.97438055e-01 -6.20147347e-01
-8.22642207e-01 1.69265777e-01 -1.37082845e-01 -3.46145779e-01
2.17538416e-01 -4.94632155e-01 1.46856457e-01 8.77214298e-02
-1.08941853e+00 5.83245993e-01 6.87197745e-01 8.97100091e-01
6.08267725e-01 -1.39807165e-01 6.02742076e-01 -1.40324461e+00
7.45678782e-01 -2.53963530e-01 -3.37168992e-01 7.82096684e-01
-7.32795417e-01 6.59537256e-01 1.18935180e+00 -2.78287500e-01
-1.38802576e+00 -1.84867278e-01 1.24688894e-02 -2.44655192e-01
1.94976702e-01 3.23974550e-01 -5.40173829e-01 6.93494081e-01
9.30097520e-01 6.36961281e-01 -3.24937165e-01 -3.48936707e-01
6.71405911e-01 5.73694885e-01 6.64751112e-01 -8.57680440e-01
1.13976312e+00 1.60895392e-01 -2.40492150e-01 -2.67598599e-01
-1.24090946e+00 -3.04325428e-02 -7.93071926e-01 4.23165560e-01
7.38479853e-01 -1.00158668e+00 -1.50406376e-01 1.43750146e-01
-1.40705967e+00 -3.15440297e-01 -6.59780025e-01 -7.15057924e-02
-7.61920333e-01 1.08044975e-01 -4.79390323e-01 -5.87744534e-01
-8.94850016e-01 -7.35349834e-01 1.28724825e+00 2.82522012e-03
-3.17690700e-01 -8.94434631e-01 1.07140191e-01 3.37464184e-01
3.76558304e-01 -5.93961924e-02 1.42079163e+00 -9.35294330e-01
-5.88274062e-01 -7.91738927e-02 -8.86047482e-02 3.10153179e-02
3.25526208e-01 -2.60831207e-01 -8.19587290e-01 -6.63326085e-02
-9.90068316e-02 -8.31606150e-01 6.93019748e-01 -1.74183637e-01
8.80120575e-01 -7.56102562e-01 -5.53283356e-02 7.34656334e-01
1.31061697e+00 2.18239248e-01 5.77269852e-01 4.03809287e-02
8.60373437e-01 6.02340102e-01 5.95142186e-01 4.78537142e-01
4.79640245e-01 5.61716437e-01 -2.55001456e-01 3.52799803e-01
-4.47695553e-01 -1.30865765e+00 4.47947502e-01 1.01549053e+00
4.14423645e-01 -4.64336276e-01 -7.56133854e-01 4.42995578e-01
-1.84226191e+00 -1.08822834e+00 3.89189571e-01 2.01339102e+00
1.31184673e+00 4.66949880e-01 -3.03493947e-01 -1.83262885e-01
6.42562866e-01 2.24270195e-01 -6.29949510e-01 -4.22908276e-01
-3.50531638e-01 2.59250969e-01 2.38473251e-01 4.01905864e-01
-6.20058477e-01 1.52022636e+00 5.58479786e+00 1.07878125e+00
-5.98355591e-01 -6.30787760e-02 5.80049992e-01 -1.67187348e-01
-6.35961413e-01 1.62584066e-01 -1.43334365e+00 2.28176266e-01
9.84816253e-01 -9.59048212e-01 6.22763455e-01 7.52235532e-01
-2.86576673e-02 3.33909869e-01 -1.49349725e+00 1.01834357e+00
6.73116505e-01 -1.47320700e+00 9.50943172e-01 -4.26113546e-01
7.62636960e-01 -4.92476612e-01 -2.51176447e-01 1.11060393e+00
3.83227319e-01 -1.16557765e+00 9.07287598e-01 3.51322085e-01
1.38383937e+00 -5.84350526e-01 3.90160620e-01 6.42971814e-01
-1.27659106e+00 2.58833647e-01 -7.47774124e-01 -7.56281763e-02
1.63392752e-01 1.35797903e-01 -1.14085317e+00 6.13300562e-01
3.09629887e-01 6.63735211e-01 -6.32946730e-01 1.94096103e-01
-3.19901913e-01 1.00682475e-01 3.14478367e-03 -3.03253293e-01
9.67727229e-02 -1.02170289e-01 3.73414218e-01 1.13739097e+00
5.55105805e-01 1.97033152e-01 2.48278528e-02 1.31920063e+00
-5.84200680e-01 2.52392769e-01 -1.11557102e+00 -3.51393610e-01
5.81131279e-01 1.08812261e+00 -4.86247092e-01 -7.47335315e-01
-4.76660967e-01 9.84408200e-01 4.14874077e-01 1.11617424e-01
-5.85382342e-01 -7.34987259e-01 2.55100816e-01 4.93788943e-02
2.30222777e-01 9.87114087e-02 -6.40666842e-01 -1.43995404e+00
2.39568219e-01 -1.20581400e+00 3.53049189e-01 -9.75772500e-01
-1.05836856e+00 8.54733944e-01 1.29262224e-01 -1.14096081e+00
-7.48519599e-01 -3.24012756e-01 -6.47397757e-01 1.03096962e+00
-1.11809337e+00 -1.29349482e+00 5.50647303e-02 7.00025618e-01
1.29854310e+00 -6.30485415e-01 1.02404642e+00 -1.79520864e-02
-5.14421463e-01 1.02574062e+00 -1.11911476e-01 3.98203969e-01
1.00555158e+00 -1.30409753e+00 1.21205103e+00 1.07759249e+00
2.77627438e-01 6.19386256e-01 6.24286234e-01 -9.53646719e-01
-1.67859006e+00 -1.32630134e+00 1.34248698e+00 -8.34595561e-01
7.07403302e-01 -9.83290493e-01 -8.72745574e-01 9.31149960e-01
4.22215194e-01 -2.59779900e-01 6.21622384e-01 -1.73311740e-01
-4.92000192e-01 -1.91918358e-01 -1.03589296e+00 7.88084090e-01
1.24146712e+00 -6.63337111e-01 -8.95847619e-01 3.56274873e-01
1.19599354e+00 -7.21133590e-01 -6.47933602e-01 1.35345623e-01
3.77560228e-01 -3.38826865e-01 5.87628841e-01 -9.49751794e-01
9.78346944e-01 4.48985882e-02 -3.02069783e-01 -1.32229996e+00
-1.18719712e-01 -1.06485868e+00 -2.82259196e-01 1.53412199e+00
9.40996706e-01 -3.55257615e-02 8.27004969e-01 6.90500021e-01
-3.76515299e-01 -6.48201525e-01 -7.80101895e-01 -6.96036279e-01
2.31557921e-01 -2.40004122e-01 8.39133203e-01 7.70434618e-01
7.46177062e-02 1.27251828e+00 -7.89433718e-01 -5.63020408e-01
2.95487195e-01 7.92285148e-03 1.08319759e+00 -8.27919543e-01
-3.92191440e-01 -4.07175310e-02 5.28038442e-01 -9.94720578e-01
1.52994767e-01 -1.24770308e+00 3.06437612e-01 -1.73290193e+00
3.83532017e-01 -2.88910344e-02 2.26693407e-01 7.58768559e-01
-3.56754422e-01 -3.89509499e-01 3.92898411e-01 2.37591378e-02
-5.23494065e-01 7.89273381e-01 1.52701676e+00 -3.83423775e-01
-2.05890372e-01 -1.94189489e-01 -1.14553738e+00 1.16946429e-01
5.25357604e-01 -7.31646776e-01 -7.88881660e-01 -7.85363913e-01
4.78501618e-01 4.99648780e-01 -3.19472104e-01 -1.03195500e+00
4.42999363e-01 -2.91023016e-01 3.11432451e-01 -6.15206540e-01
1.85611174e-01 -4.69006121e-01 -1.94398269e-01 2.04496786e-01
-7.96533942e-01 2.91874915e-01 4.04905498e-01 4.26915675e-01
-4.56455275e-02 -3.40850919e-01 6.88916981e-01 -5.18652976e-01
-4.05600429e-01 4.33629006e-01 -1.96686491e-01 6.99043274e-01
5.75347126e-01 -1.70157328e-01 -5.08880019e-01 -4.10823643e-01
-1.08128525e-01 3.31121922e-01 1.98393047e-01 7.37829328e-01
6.28281236e-01 -1.40146720e+00 -9.27967966e-01 4.89651948e-01
4.98049796e-01 3.82881373e-01 1.15134947e-01 1.40126497e-01
-1.95328817e-01 7.74651170e-01 -3.94414783e-01 -1.98424663e-02
-8.22507620e-01 7.28080809e-01 6.49615005e-02 -7.92240679e-01
-5.84188104e-01 8.11772346e-01 2.47901231e-02 -7.33267963e-01
2.91799635e-01 -7.12765336e-01 6.47940412e-02 -1.07295280e-02
7.71990299e-01 1.15036309e-01 2.63723701e-01 -1.68893561e-01
8.92367512e-02 2.59582669e-01 -5.86671770e-01 -2.81882137e-01
1.28976512e+00 -7.25003257e-02 1.74816817e-01 6.84506476e-01
8.99111986e-01 -3.91765013e-02 -1.00899136e+00 -5.80435634e-01
1.75845712e-01 -2.57805169e-01 -4.87122595e-01 -9.77917552e-01
-4.75490838e-01 1.05808949e+00 -9.02433228e-03 -4.72282171e-01
7.58942425e-01 -2.70665675e-01 1.49989128e+00 1.17201817e+00
3.54498565e-01 -1.27940488e+00 3.90012205e-01 1.03434968e+00
9.53636348e-01 -1.31175351e+00 -2.18294099e-01 -2.13011771e-01
-1.09813440e+00 1.00423741e+00 9.11668777e-01 2.38853991e-01
1.28806174e-01 3.59141469e-01 -1.10489689e-01 2.15092123e-01
-1.39559019e+00 -6.87726587e-02 1.93375975e-01 8.19451451e-01
5.14972687e-01 -4.41931516e-01 8.21060464e-02 1.02166247e+00
-6.69715941e-01 1.29721016e-01 6.50170445e-01 9.82747734e-01
-5.31040251e-01 -1.18497455e+00 -9.57418792e-03 6.27972722e-01
-2.66617447e-01 -7.10923493e-01 -4.88680899e-01 6.62529528e-01
-5.51466979e-02 7.70524621e-01 -9.20204595e-02 -3.34787011e-01
5.79908371e-01 3.92593056e-01 4.39177752e-01 -1.33274126e+00
-6.30699456e-01 -1.55047864e-01 2.33066350e-01 -3.67350936e-01
3.98658246e-01 -3.00448090e-01 -1.17714477e+00 -3.34325314e-01
1.58369139e-01 1.85456529e-01 2.10525379e-01 9.68246281e-01
3.00485283e-01 5.38237214e-01 3.21367890e-01 -2.47798309e-01
-9.39242601e-01 -1.29404199e+00 -5.10510623e-01 5.14107645e-01
2.03118268e-02 -1.24606937e-01 -1.74689386e-02 2.92143971e-01] | [11.619741439819336, 8.829852104187012] |
2713b05d-64e2-41ea-9506-603691d1ec52 | listen-carefully-and-tell-an-audio-captioning | 2006.15406 | null | https://arxiv.org/abs/2006.15406v4 | https://arxiv.org/pdf/2006.15406v4.pdf | Listen carefully and tell: an audio captioning system based on residual learning and gammatone audio representation | Automated audio captioning is machine listening task whose goal is to describe an audio using free text. An automated audio captioning system has to be implemented as it accepts an audio as input and outputs as textual description, that is, the caption of the signal. This task can be useful in many applications such as automatic content description or machine-to-machine interaction. In this work, an automatic audio captioning based on residual learning on the encoder phase is proposed. The encoder phase is implemented via different Residual Networks configurations. The decoder phase (create the caption) is run using recurrent layers plus attention mechanism. The audio representation chosen has been Gammatone. Results show that the framework proposed in this work surpass the baseline system in challenge results. | ['Maximo Cobos', 'Pedro Zuccarello', 'Javier Naranjo-Alcazar', 'Sergi Perez-Castanos'] | 2020-06-27 | null | null | null | null | ['audio-captioning'] | ['audio'] | [ 7.02613175e-01 4.43374842e-01 3.49412322e-01 -3.07677180e-01
-1.30444479e+00 -4.56991464e-01 5.20737410e-01 -6.20425791e-02
-2.32279524e-01 7.22616196e-01 7.53107727e-01 8.56367052e-02
2.31986254e-01 -2.31497064e-01 -7.69538641e-01 -4.47302639e-01
1.05930455e-02 7.15560913e-01 3.09047569e-02 -2.10412502e-01
6.16223440e-02 -6.84093758e-02 -1.64128041e+00 9.41481829e-01
-4.45382018e-03 1.07209551e+00 5.10500789e-01 1.26849937e+00
-1.84925005e-01 1.23670483e+00 -7.08207309e-01 -1.37868300e-01
-1.82745486e-01 -7.26846457e-01 -1.06822276e+00 -1.94083497e-01
2.60622911e-02 -2.13075578e-02 -1.16715711e-02 6.71850383e-01
7.85843611e-01 2.47700419e-02 4.52453107e-01 -1.13989353e+00
-2.99095660e-01 1.29835379e+00 8.34850296e-02 1.30651444e-02
8.01645398e-01 -1.23104647e-01 1.17073405e+00 -8.13699663e-01
4.38484877e-01 1.16972578e+00 4.16328520e-01 7.73152232e-01
-1.31113875e+00 -6.46709025e-01 -1.38879806e-01 3.43887240e-01
-1.52517211e+00 -6.64564669e-01 9.38265920e-01 -5.75349867e-01
8.21890295e-01 4.06682909e-01 2.99784899e-01 1.35397279e+00
1.83273368e-02 6.16278589e-01 6.40326560e-01 -6.38896048e-01
3.30339104e-01 3.20032924e-01 6.39144033e-02 -1.93776086e-01
-3.92987013e-01 -3.03551406e-01 -5.99035084e-01 -1.48748025e-01
2.91729301e-01 -6.78372741e-01 -3.89077872e-01 1.64814532e-01
-1.11248183e+00 8.31332982e-01 5.71204275e-02 4.09709692e-01
-7.01058269e-01 5.46404123e-01 8.42956245e-01 3.22572231e-01
2.79925942e-01 6.12635672e-01 -1.07818261e-01 -3.71784717e-01
-1.03469825e+00 2.03881294e-01 8.51086199e-01 7.29378283e-01
4.30737406e-01 1.63474083e-01 -4.54468697e-01 8.80692124e-01
4.41621780e-01 1.65223345e-01 7.87218332e-01 -7.04758167e-01
5.21833837e-01 4.41251248e-02 1.68860376e-01 -5.58087289e-01
-2.72446364e-01 -4.27936763e-01 -7.11791217e-01 -9.32227448e-02
-9.75321382e-02 -3.54785502e-01 -9.57840681e-01 1.57012069e+00
-2.56381035e-01 5.09403646e-01 3.46385628e-01 9.90348160e-01
1.18415117e+00 1.27765775e+00 2.85652310e-01 -1.02670081e-01
1.45740795e+00 -9.46808457e-01 -1.12293172e+00 -2.25770548e-01
3.83477658e-01 -1.05264068e+00 9.18171942e-01 3.65206897e-01
-1.15803277e+00 -6.58262670e-01 -9.87524927e-01 2.41156425e-02
-1.53406546e-01 2.17697829e-01 7.53076631e-04 3.28558058e-01
-1.28030872e+00 1.52501211e-01 -3.90110910e-01 -1.53443202e-01
-2.62019694e-01 4.78339255e-01 -3.75348985e-01 5.38104951e-01
-1.55245936e+00 5.82623661e-01 8.54869306e-01 6.62293658e-02
-1.14079475e+00 -3.60873401e-01 -1.01699865e+00 4.40363526e-01
-1.76119842e-02 -6.76921010e-01 1.90951598e+00 -1.48316467e+00
-1.95150125e+00 6.08157814e-01 -4.59738225e-02 -1.10021853e+00
3.13614130e-01 -2.78612554e-01 -4.62469041e-01 3.63161981e-01
-7.35378638e-02 9.39661264e-01 1.01900876e+00 -1.10832798e+00
-3.44073445e-01 2.83215970e-01 -1.90737396e-01 3.83131146e-01
1.91360623e-01 3.43090236e-01 -3.84785086e-01 -5.48810184e-01
-2.84829140e-01 -9.25504446e-01 1.02258697e-01 -7.72096992e-01
-4.69620526e-01 -7.07784593e-02 7.16240883e-01 -8.34728420e-01
1.33690178e+00 -2.28185439e+00 4.04236205e-02 5.42353652e-02
-3.42583448e-01 2.32127756e-01 -1.99576199e-01 6.50836527e-01
-5.90723813e-01 8.14274028e-02 -2.08757907e-01 -5.95792770e-01
1.42274395e-01 -2.23334730e-01 -7.78427899e-01 5.66441193e-02
3.18494171e-01 6.93310261e-01 -5.62752962e-01 -5.13557076e-01
7.47091472e-02 8.77607107e-01 -5.02151549e-01 7.64456153e-01
-5.37401915e-01 3.03876966e-01 -1.40416339e-01 2.56728027e-02
1.88170299e-01 -1.07322715e-01 -1.99473396e-01 -5.44912815e-02
-1.26239359e-01 7.85468340e-01 -1.05932796e+00 1.77615106e+00
-6.79045737e-01 1.04164302e+00 2.13855222e-01 -7.40106225e-01
1.00587606e+00 1.14384961e+00 1.89347982e-01 -2.92149186e-01
2.16308251e-01 2.13454336e-01 -2.75313742e-02 -5.70096493e-01
5.17254055e-01 -2.87583351e-01 -1.98639616e-01 5.00772953e-01
1.55406982e-01 -1.91917375e-01 -1.42013505e-01 1.71080306e-01
9.75303888e-01 2.00366288e-01 1.36537552e-01 2.39176750e-01
8.11261117e-01 -2.06498969e-02 -1.05608642e-01 5.16641140e-01
3.06749314e-01 1.26510930e+00 4.60182875e-01 -5.39406687e-02
-1.13019669e+00 -5.48615396e-01 2.13439345e-01 1.12864053e+00
-4.31306660e-01 -5.94599009e-01 -9.14206803e-01 -9.60386395e-02
-7.24177599e-01 7.65250683e-01 -6.48831666e-01 -8.20570290e-02
-5.30021012e-01 3.74540053e-02 6.88116908e-01 2.51625359e-01
2.46959120e-01 -1.66326427e+00 -4.27041024e-01 7.21383572e-01
-5.51342726e-01 -1.20120382e+00 -4.90206897e-01 4.12994474e-01
-3.36793184e-01 -4.18593287e-01 -9.10294592e-01 -9.81034994e-01
1.04611814e-02 -3.17768037e-01 1.01192880e+00 -4.13029343e-01
6.48231581e-02 3.69378209e-01 -6.58178091e-01 -6.31155550e-01
-7.84892917e-01 4.91189480e-01 -2.27554888e-01 4.36541349e-01
1.31319180e-01 -4.75075901e-01 -3.06134611e-01 -6.87491000e-02
-9.70534563e-01 3.09080839e-01 6.87269270e-01 6.72384918e-01
5.57287991e-01 -4.02372181e-01 8.66928399e-01 -6.67448342e-01
9.72247303e-01 -5.30943811e-01 -2.66554624e-01 3.21005806e-02
9.61899608e-02 2.63621509e-01 7.37910926e-01 -5.14005780e-01
-8.07826519e-01 6.93805218e-01 -7.37767756e-01 -4.51580614e-01
-3.74119341e-01 6.67353034e-01 -2.46533617e-01 5.76142371e-01
4.54608709e-01 2.04618469e-01 -1.17784560e-01 -6.69749677e-01
6.01574853e-02 1.28178322e+00 7.52259493e-01 -2.14487240e-01
5.43387711e-01 -1.42897949e-01 -3.43243599e-01 -8.83358061e-01
-8.05975616e-01 -6.31324351e-01 -3.45311463e-01 -3.24690789e-01
1.07117307e+00 -1.18164241e+00 -5.51589251e-01 -1.21209435e-01
-1.66821229e+00 -1.42894074e-01 -2.87081510e-01 6.63055539e-01
-9.58781064e-01 -1.58197731e-01 -3.82638663e-01 -1.14152026e+00
-7.69896924e-01 -1.09716892e+00 1.36880398e+00 2.15515383e-02
-6.02459013e-01 -6.14415288e-01 4.34102893e-01 3.28779995e-01
4.26311076e-01 1.43450975e-01 5.47711611e-01 -1.22916150e+00
-2.41105497e-01 -2.56471246e-01 8.90150480e-03 3.79840493e-01
-1.38389364e-01 -2.12703630e-01 -1.55599141e+00 5.67080043e-02
-2.59623170e-01 -4.63067472e-01 7.09815204e-01 2.84479409e-01
1.16633928e+00 -5.16037703e-01 1.39829814e-01 1.78728387e-01
1.17740977e+00 3.23001921e-01 8.52823555e-01 3.06979060e-01
2.55622417e-01 7.25444436e-01 3.86815041e-01 3.47655654e-01
-8.83797407e-02 9.22605574e-01 5.90209842e-01 7.80507475e-02
-3.31300497e-01 -5.47825992e-01 7.73506939e-01 8.60005260e-01
4.32046324e-01 -4.41343635e-01 -9.09706354e-01 5.79605281e-01
-1.82520163e+00 -1.24256277e+00 -7.61406571e-02 2.07308125e+00
8.29121053e-01 2.04736724e-01 1.49510145e-01 5.26637137e-01
1.00941849e+00 -1.28929630e-01 -1.64322704e-02 -9.68201458e-01
1.76232547e-01 2.99253225e-01 -1.40050761e-02 7.41194904e-01
-9.60109413e-01 7.32135296e-01 6.17886257e+00 4.15038913e-01
-1.38326561e+00 2.27826327e-01 3.35773408e-01 7.36331195e-02
-1.31369531e-01 -2.04149172e-01 -6.32748783e-01 5.46754539e-01
1.74972117e+00 -1.65637881e-01 3.04531842e-01 6.56014204e-01
5.24712861e-01 1.93327159e-01 -1.28046536e+00 1.21856272e+00
1.97053581e-01 -1.20491052e+00 3.04027796e-01 -3.35584402e-01
1.86288312e-01 -9.36283022e-02 -2.49428060e-02 4.60945636e-01
-2.16028959e-01 -1.21541762e+00 9.51832056e-01 5.36567748e-01
9.12588120e-01 -7.27611244e-01 1.15491021e+00 2.14411974e-01
-1.28191960e+00 1.09471560e-01 5.29966727e-02 -2.14946568e-01
4.70696837e-01 2.04058781e-01 -1.73497272e+00 2.95999676e-01
4.10279512e-01 2.25454569e-01 -2.96687573e-01 1.44707978e+00
-1.47051126e-01 1.18691874e+00 -2.08368897e-01 -7.07624927e-02
3.43596816e-01 3.10498238e-01 7.78206587e-01 1.71069181e+00
4.06883061e-01 -1.98908836e-01 -6.19225651e-02 6.65827215e-01
-1.76034421e-01 4.90530401e-01 -5.85593998e-01 -2.74603516e-01
2.79584587e-01 8.81740630e-01 -4.61375266e-01 -2.13452592e-01
1.09919742e-01 1.00047731e+00 -1.91249818e-01 1.86489239e-01
-9.02003944e-01 -8.43139470e-01 2.71173388e-01 3.59334379e-01
2.96920061e-01 1.94227576e-01 2.12632284e-01 -7.48287022e-01
-1.36633739e-01 -6.67732775e-01 2.22728163e-01 -1.44182551e+00
-6.68027639e-01 1.16380847e+00 9.50477943e-02 -1.28705370e+00
-9.79020655e-01 -1.92252025e-01 -8.04131567e-01 9.55264091e-01
-1.20434999e+00 -1.00935137e+00 -1.33655667e-01 3.71654958e-01
8.85849893e-01 -1.70867100e-01 1.00839090e+00 5.27019262e-01
-2.91482508e-01 3.04826677e-01 -2.77099073e-01 2.50198752e-01
8.11859012e-01 -1.09207642e+00 1.84382230e-01 4.82481420e-01
3.89061868e-01 1.42678738e-01 1.38671541e+00 -3.83441001e-01
-9.83559906e-01 -1.26694500e+00 1.36181724e+00 -2.36125272e-02
6.95264161e-01 -6.49691045e-01 -9.81339216e-01 7.55908251e-01
7.74205863e-01 -3.73674005e-01 7.21324086e-01 -4.26532477e-01
-1.45951118e-02 -6.32199645e-03 -6.48028672e-01 1.81901023e-01
1.09468840e-01 -7.81028390e-01 -8.41965497e-01 2.85064667e-01
1.11102891e+00 -4.35601413e-01 -3.68636072e-01 1.06983691e-01
3.93740267e-01 -3.51938665e-01 6.28125370e-01 -4.74289775e-01
4.30449069e-01 -5.05869627e-01 -1.77815378e-01 -1.05561554e+00
-6.08674437e-02 -1.19663656e+00 7.47156590e-02 1.59730673e+00
7.98422098e-01 -4.38429676e-02 4.12321389e-01 1.65003031e-01
-3.21489096e-01 -1.17197998e-01 -1.10970163e+00 -5.10576785e-01
-4.67864990e-01 -8.56096447e-01 4.77346003e-01 3.11643600e-01
-3.30210179e-02 1.03750014e+00 -7.48318791e-01 1.98297322e-01
2.39239670e-02 -2.77047545e-01 5.94343722e-01 -1.32653284e+00
-2.47382179e-01 -2.47861996e-01 -5.05172431e-01 -7.22304404e-01
4.01687413e-01 -8.75487268e-01 4.95709807e-01 -1.67965603e+00
-8.22466463e-02 7.66645893e-02 -1.65402055e-01 5.34497678e-01
4.30915236e-01 5.68077326e-01 2.51005024e-01 5.53395227e-02
-5.69435179e-01 4.97206718e-01 6.31816566e-01 -4.00756717e-01
-4.84949231e-01 1.84247226e-01 -4.57234979e-01 2.78681219e-01
1.06694043e+00 -6.75060272e-01 -2.32531026e-01 -1.10238403e-01
5.20201504e-01 4.10985827e-01 3.31959546e-01 -1.28172708e+00
3.44289035e-01 3.83585781e-01 -7.45553598e-02 -5.21904647e-01
8.06687295e-01 -9.85571682e-01 6.15141332e-01 2.19070822e-01
-1.00997114e+00 6.55339360e-02 3.11819494e-01 3.11891407e-01
-6.71907902e-01 -6.56610250e-01 5.05441606e-01 7.92894810e-02
-2.75373846e-01 -2.06308678e-01 -7.04739869e-01 -1.88576490e-01
6.98959112e-01 -2.24311743e-02 2.11946204e-01 -9.84049737e-01
-1.20495403e+00 -1.08033568e-01 -2.13151455e-01 5.56189358e-01
6.47461593e-01 -1.48837614e+00 -1.14207232e+00 -1.16414959e-02
1.37466177e-01 -2.07401812e-01 3.34144644e-02 4.87471402e-01
-4.64416414e-01 7.95394480e-01 -1.04088053e-01 -5.03015757e-01
-1.42113376e+00 2.73911029e-01 3.09043020e-01 -2.16733575e-01
-5.32354712e-01 5.73649228e-01 -5.62784448e-02 2.45286196e-01
7.20731497e-01 -7.18557686e-02 -7.54613400e-01 3.41612607e-01
8.80450070e-01 -7.57226273e-02 1.97911680e-01 -9.85628366e-01
-1.99893564e-01 2.33233541e-01 -1.37997009e-02 -8.03533137e-01
1.37758613e+00 -1.13472179e-01 8.90702307e-02 9.11045611e-01
1.40193343e+00 -6.81946874e-02 -7.36051619e-01 2.44814843e-01
7.18809962e-02 2.54037052e-01 1.83499560e-01 -9.37152565e-01
-4.30135816e-01 1.08791268e+00 6.80697024e-01 3.92035812e-01
1.03445959e+00 -7.28903431e-03 6.17948771e-01 3.19621325e-01
-1.06188819e-01 -9.55722749e-01 -3.00967302e-02 7.07546711e-01
1.35197651e+00 -9.93333757e-01 -4.77164775e-01 5.72938137e-02
-9.46445763e-01 1.22300470e+00 3.11039746e-01 5.24393395e-02
4.86272067e-01 2.75328964e-01 1.40094697e-01 7.50003532e-02
-1.21057522e+00 -3.48963380e-01 5.47493100e-01 4.45677966e-01
7.44935572e-01 -1.38177916e-01 -6.95789307e-02 1.04519701e+00
-5.40212691e-01 7.37452367e-03 7.41241634e-01 5.01009524e-01
-4.01422113e-01 -9.82679605e-01 -7.03869998e-01 1.53005598e-02
-7.60258913e-01 -1.51635066e-01 -8.61331582e-01 2.66361445e-01
-9.95578244e-02 1.15929699e+00 1.54584497e-01 -5.26852489e-01
1.84373885e-01 5.62599421e-01 -1.85305595e-01 -9.32623267e-01
-1.13995314e+00 3.34419101e-01 2.27323711e-01 -2.17416599e-01
-3.64492506e-01 -3.25827569e-01 -1.33012176e+00 6.07009172e-01
-2.35932142e-01 9.26317573e-01 1.12833226e+00 6.50622845e-01
3.56376261e-01 7.63022661e-01 5.78693807e-01 -9.60316539e-01
-1.44966573e-01 -1.34184968e+00 -1.91719934e-01 1.23714127e-01
7.15884328e-01 -1.56404659e-01 -5.07285655e-01 5.09143114e-01] | [15.27840805053711, 4.8929572105407715] |
9ae42bcf-258c-4ba2-9c95-9a44a067da2c | grounded-word-sense-translation | null | null | https://aclanthology.org/W19-1808 | https://aclanthology.org/W19-1808.pdf | Grounded Word Sense Translation | Recent work on visually grounded language learning has focused on broader applications of grounded representations, such as visual question answering and multimodal machine translation. In this paper we consider grounded word sense translation, i.e. the task of correctly translating an ambiguous source word given the corresponding textual and visual context. Our main objective is to investigate the extent to which images help improve word-level (lexical) translation quality. We do so by first studying the dataset for this task to understand the scope and challenges of the task. We then explore different data settings, image features, and ways of grounding to investigate the gain from using images in each of the combinations. We find that grounding on the image is specially beneficial in weaker unidirectional recurrent translation models. We observe that adding structured image information leads to stronger gains in lexical translation accuracy. | ['Lucia Specia', 'Pranava Madhyastha', 'Chiraag Lala'] | 2019-06-01 | null | null | null | ws-2019-6 | ['multimodal-machine-translation', 'grounded-language-learning'] | ['natural-language-processing', 'natural-language-processing'] | [ 5.79782605e-01 2.20628873e-01 -3.13847601e-01 -5.93255796e-02
-1.09993780e+00 -7.72525132e-01 9.49372053e-01 5.66423349e-02
-3.08139741e-01 6.22354627e-01 7.37592161e-01 -6.75722659e-01
2.52127498e-01 -5.30846775e-01 -8.86266470e-01 -5.34158409e-01
4.43144172e-01 1.96518317e-01 -1.67122662e-01 -4.12823290e-01
2.38921568e-01 8.38174745e-02 -1.11393356e+00 6.34476960e-01
6.64088249e-01 5.00539660e-01 2.58887291e-01 4.27843243e-01
-2.77731091e-01 6.27494752e-01 -4.76488173e-01 -5.25599301e-01
2.12640762e-01 -1.01179433e+00 -1.00094926e+00 3.40776682e-01
7.50414431e-01 -4.50376198e-02 -2.63752282e-01 9.44662333e-01
4.98660952e-01 6.05701376e-03 5.98656654e-01 -1.23308623e+00
-1.12793934e+00 6.75162673e-01 -6.25817001e-01 3.71578902e-01
7.21574247e-01 2.70993918e-01 1.26263785e+00 -1.16713285e+00
1.06275916e+00 1.43304813e+00 2.18249738e-01 4.70865935e-01
-1.48054314e+00 -3.00760478e-01 2.64143944e-01 1.03693902e-01
-1.01789188e+00 -5.97943306e-01 7.55188644e-01 -4.29777652e-01
8.79801095e-01 1.31895900e-01 4.67347592e-01 1.29362082e+00
1.45762861e-01 6.05811775e-01 1.62767851e+00 -8.28805864e-01
2.41308822e-03 1.90761849e-01 2.03666035e-02 8.04396927e-01
3.27725798e-01 1.14097469e-01 -8.94680142e-01 1.64770827e-01
7.09847331e-01 -3.04750830e-01 -2.48826101e-01 -4.83225197e-01
-1.70316887e+00 9.47040677e-01 6.08248055e-01 4.01493251e-01
-2.21461535e-01 2.92247832e-01 2.66074315e-02 5.36739290e-01
4.59370136e-01 6.42743349e-01 -1.03339674e-02 1.33501425e-01
-7.52314687e-01 1.45645306e-01 4.74101752e-01 8.38765681e-01
8.44925642e-01 6.76436815e-03 -5.08702636e-01 6.14970863e-01
1.59681842e-01 6.60305977e-01 7.76787475e-02 -8.03459346e-01
8.22904825e-01 5.11041522e-01 1.72590882e-01 -8.39774728e-01
-3.27483296e-01 -2.61477202e-01 -2.85684317e-01 7.66778551e-03
4.65554595e-01 -1.22913770e-01 -9.83042538e-01 1.89104331e+00
-6.40490744e-03 -1.55805305e-01 9.57505628e-02 1.08287156e+00
8.72908831e-01 7.28633583e-01 2.87321359e-01 -1.18731208e-01
1.31353724e+00 -8.70970368e-01 -5.91172755e-01 -5.26447177e-01
6.96650624e-01 -1.02539957e+00 1.62836587e+00 -3.39328021e-01
-9.20827806e-01 -2.55515993e-01 -9.72326159e-01 -4.03828889e-01
-3.31369340e-01 4.13314439e-02 2.88708627e-01 3.60989273e-01
-1.16381299e+00 1.37940004e-01 -6.03332937e-01 -7.22063780e-01
2.92350352e-01 -2.64752433e-02 -3.01646620e-01 -3.76628935e-01
-1.00147021e+00 1.13035047e+00 -2.72530109e-01 -1.86360106e-01
-6.64624095e-01 -3.07578921e-01 -9.37688112e-01 -2.25747973e-01
3.34954053e-01 -1.04734969e+00 1.09737396e+00 -1.41728616e+00
-9.10785675e-01 1.40822577e+00 -5.68287194e-01 -2.52680957e-01
3.34855467e-01 -3.75856571e-02 -3.16964053e-02 1.64404944e-01
3.52630556e-01 9.89493310e-01 8.36982548e-01 -1.54502296e+00
-2.48403639e-01 -4.97725040e-01 2.99957871e-01 4.57240999e-01
-1.58347577e-01 -5.60535304e-03 -3.04244816e-01 -7.05206335e-01
-4.19601053e-02 -1.33407199e+00 -3.92108085e-03 -3.72286849e-02
-3.77264202e-01 1.43015962e-02 5.23065925e-01 -6.97747886e-01
7.29735017e-01 -2.03887057e+00 5.00107586e-01 -6.42682565e-03
1.81122452e-01 -2.73747861e-01 -5.29613793e-01 6.85446024e-01
-1.99983623e-02 2.68003643e-01 -3.04245012e-04 -2.85984457e-01
-1.16765000e-01 2.50168353e-01 -5.75538337e-01 2.96985775e-01
5.43428957e-01 1.58782339e+00 -7.60107160e-01 -5.13233066e-01
-1.22258119e-01 4.99057204e-01 -3.04361999e-01 1.45933339e-02
-2.68641204e-01 5.19940674e-01 -2.40913063e-01 5.48736751e-01
1.27957165e-01 -4.92013514e-01 1.64965048e-01 -1.42194375e-01
-4.47366945e-02 3.84099543e-01 -6.61788166e-01 1.74667585e+00
-6.87779486e-01 1.16153550e+00 -1.88716009e-01 -7.32851326e-01
7.13514328e-01 1.64831653e-01 1.17615886e-01 -1.29332626e+00
-1.16021670e-01 8.39216486e-02 1.28650963e-01 -4.37886089e-01
6.00298166e-01 -5.54248273e-01 3.06360163e-02 6.09427691e-01
-2.23086059e-01 -2.60611507e-03 -5.07472605e-02 3.61239046e-01
7.47656405e-01 2.87126064e-01 1.51191428e-01 -2.22928703e-01
-1.20837837e-01 4.72023755e-01 -9.62548777e-02 6.61451042e-01
6.71920478e-02 7.11911798e-01 5.56905329e-01 -2.24375069e-01
-1.15197194e+00 -1.06807482e+00 3.31423610e-01 1.36700737e+00
4.34056729e-01 -1.20488971e-01 -5.98721504e-01 -4.54061359e-01
-2.61150628e-01 7.59828091e-01 -7.06182182e-01 -1.67496458e-01
-6.43727720e-01 -5.84703505e-01 2.27279261e-01 6.36542857e-01
2.22786188e-01 -1.27732277e+00 -7.78198540e-01 -2.74578989e-01
-6.29895151e-01 -1.28135026e+00 -5.90181231e-01 -4.07923153e-03
-9.45630312e-01 -7.97065258e-01 -7.96290398e-01 -9.04950202e-01
6.80316806e-01 6.61468029e-01 1.45106316e+00 1.77176565e-01
6.16515579e-04 7.15126038e-01 -5.39649427e-01 -3.41003507e-01
-3.95277768e-01 2.86683291e-01 -3.31077337e-01 -1.14667267e-01
1.05938345e-01 -9.68490317e-02 -4.55605984e-01 1.98386222e-01
-1.03834200e+00 3.77519697e-01 8.41002285e-01 9.35449004e-01
4.58365083e-01 -8.78145635e-01 3.62103015e-01 -7.01814532e-01
9.69457388e-01 -3.31695825e-01 -1.44982547e-01 4.98851031e-01
-4.93816048e-01 2.88673490e-01 -1.58871617e-02 -4.90679204e-01
-6.92548871e-01 -6.09223247e-02 1.91501319e-01 -1.85148403e-01
1.30612195e-01 5.00061870e-01 5.32175265e-02 -1.41140064e-02
8.37258518e-01 7.00410008e-02 4.98394202e-03 -6.00568652e-02
7.48576283e-01 3.32488209e-01 2.05553219e-01 -7.46374786e-01
7.58591831e-01 4.33331370e-01 -1.06843961e-02 -8.30630541e-01
-7.45488763e-01 -2.47986242e-01 -5.20564556e-01 -1.51381522e-01
1.08373499e+00 -9.29903626e-01 -2.19769403e-01 -2.59329438e-01
-1.20064819e+00 -5.30129194e-01 -2.92241782e-01 1.34490296e-01
-7.44669855e-01 7.59512857e-02 -3.30429345e-01 -4.70995545e-01
-1.43845290e-01 -1.27769363e+00 1.44693184e+00 -9.90557075e-02
-3.36334556e-01 -8.70330155e-01 9.45045725e-02 4.74176139e-01
2.92671353e-01 3.28112304e-01 1.18724942e+00 -2.37868473e-01
-8.49327266e-01 3.58251810e-01 -4.31168348e-01 -2.09564492e-01
1.28645718e-01 -3.60484928e-01 -8.25927734e-01 -2.47303754e-01
-2.52696425e-01 -4.36852336e-01 1.06328940e+00 1.67207390e-01
3.71289819e-01 -4.20355946e-01 -2.05410987e-01 3.35709959e-01
1.42501009e+00 -8.62367228e-02 7.41259158e-01 3.75762880e-01
8.69903207e-01 8.13570917e-01 5.15634656e-01 -2.31535912e-01
4.56405967e-01 1.01688790e+00 2.56107509e-01 -6.81119323e-01
-4.43239272e-01 -5.32553256e-01 4.56741899e-01 2.50880271e-01
6.56806007e-02 -4.03432608e-01 -1.19032538e+00 7.29714394e-01
-1.80256617e+00 -9.32885170e-01 1.45768583e-01 1.83721697e+00
7.65654385e-01 -1.30090013e-01 2.92998821e-01 -6.88574463e-02
7.09051073e-01 1.97040603e-01 -1.49291381e-01 -4.11771178e-01
-4.13190901e-01 1.55971989e-01 3.29477668e-01 7.14209855e-01
-5.94195664e-01 1.08894718e+00 6.15975285e+00 3.50786209e-01
-1.38722849e+00 2.00781479e-01 7.70373940e-01 -1.41685665e-01
-8.18062186e-01 9.63985473e-02 -4.35680479e-01 4.67927568e-02
6.17408276e-01 1.59328580e-01 5.72576404e-01 3.04742754e-01
2.96696991e-01 -1.00958779e-01 -1.38212979e+00 8.92516732e-01
3.00337642e-01 -1.55454719e+00 4.47804570e-01 1.59525007e-01
8.88173342e-01 -9.61732045e-02 3.12504917e-01 2.75832713e-02
8.84863809e-02 -1.31214213e+00 1.02484274e+00 3.54237795e-01
9.65526700e-01 -3.69754434e-01 4.96485084e-01 -6.56900257e-02
-7.36330569e-01 5.93292080e-02 -1.53142298e-02 -2.13308260e-01
2.33295023e-01 2.31404719e-03 -8.67004931e-01 2.44846717e-01
4.18013513e-01 5.30499518e-01 -8.49204302e-01 5.31980753e-01
-1.06809311e-01 6.38247728e-01 2.03687493e-02 -1.38164937e-01
4.15352970e-01 -6.32284582e-02 4.96509314e-01 1.11386740e+00
3.33901644e-01 -1.04682736e-01 1.38039455e-01 9.09431994e-01
-1.23647600e-01 2.50958741e-01 -1.02090347e+00 -3.34004521e-01
2.34569207e-01 9.12697136e-01 -8.68836820e-01 -5.89579903e-02
-6.64289653e-01 1.07758474e+00 3.80550385e-01 7.11173713e-01
-5.78641295e-01 2.16464326e-01 4.96164739e-01 5.33039808e-01
2.92060107e-01 -3.85859877e-01 -6.53270721e-01 -1.25759578e+00
7.49812946e-02 -1.08268416e+00 1.97620347e-01 -1.20099831e+00
-1.02531660e+00 5.23537457e-01 4.96887742e-03 -9.86396492e-01
-2.27934122e-01 -6.13818645e-01 -4.13853496e-01 8.61797214e-01
-1.41085076e+00 -1.53342140e+00 -2.49418721e-01 5.67856014e-01
5.57936251e-01 1.36087388e-01 5.60081661e-01 -3.81663926e-02
-2.88656980e-01 5.30889452e-01 -3.10365647e-01 8.02469626e-02
8.09457064e-01 -8.72304022e-01 4.67519164e-01 9.04600203e-01
8.60331833e-01 6.96240008e-01 9.46801066e-01 -6.59364820e-01
-1.58930266e+00 -8.26090932e-01 1.13837600e+00 -7.78058708e-01
5.58231652e-01 -4.51414526e-01 -5.39699674e-01 7.85049796e-01
6.83043420e-01 -3.11006337e-01 5.09248972e-01 1.91952363e-01
-6.33698285e-01 2.45363027e-01 -8.23060274e-01 1.00505781e+00
1.16109526e+00 -8.02754939e-01 -5.44701636e-01 1.89439490e-01
1.03910756e+00 -3.51334661e-01 -3.75789076e-01 1.04004979e-01
2.43396461e-01 -6.50184095e-01 9.96693492e-01 -7.68381357e-01
8.15510333e-01 -2.46185720e-01 -4.80404913e-01 -1.36225784e+00
-2.79330879e-01 -4.46662158e-01 3.80848140e-01 9.74499762e-01
7.91717708e-01 -3.72232169e-01 5.42783856e-01 2.27962285e-01
1.35780156e-01 -8.09236825e-01 -8.46419573e-01 -3.91762435e-01
2.69382119e-01 -1.55307814e-01 3.74326497e-01 9.60816264e-01
3.25042717e-02 9.96232212e-01 -3.74998450e-01 -7.09689036e-02
1.99085325e-01 5.55773914e-01 6.44102573e-01 -5.77261090e-01
-1.96151644e-01 -3.53106618e-01 -2.35404134e-01 -9.30591106e-01
7.37192407e-02 -9.15229142e-01 1.01815909e-02 -1.86748266e+00
4.88612473e-01 -9.02611911e-02 -1.09049240e-02 6.23976052e-01
-4.69437629e-01 7.81243324e-01 6.90339029e-01 2.85269469e-01
-6.15292609e-01 1.96100563e-01 1.41114867e+00 -3.23570102e-01
-1.32553756e-01 -5.04341125e-01 -1.08006632e+00 8.81242827e-02
7.73642957e-01 -2.83063531e-01 -4.33344781e-01 -9.41814721e-01
6.76691473e-01 1.45871088e-01 7.89194047e-01 -3.14605862e-01
-2.26154417e-01 -2.70357877e-01 4.03337210e-01 -8.84827599e-02
2.57870972e-01 -5.12236655e-01 -9.03612077e-02 1.88709676e-01
-7.50023723e-01 6.88758969e-01 2.64457554e-01 4.84787077e-01
-1.31629393e-01 1.25230372e-01 6.17496669e-01 -2.17218265e-01
-5.51906943e-01 -1.62030846e-01 -2.05139488e-01 4.61786330e-01
7.48584569e-01 -1.79822996e-01 -4.65993732e-01 -7.66848028e-01
-6.80628300e-01 -1.40281841e-01 8.71386111e-01 7.28491366e-01
6.47420108e-01 -1.53156984e+00 -8.42044950e-01 -1.63286537e-01
4.90113050e-01 -6.37705445e-01 -3.23304236e-01 8.96119535e-01
-2.07656488e-01 4.15761441e-01 -2.85303444e-01 -7.01648772e-01
-1.22303748e+00 5.23438632e-01 3.35765034e-01 3.78038138e-02
-4.18569297e-01 6.25028193e-01 4.05197233e-01 -2.45896205e-02
-3.13450173e-02 -3.30009103e-01 9.67079177e-02 1.12844929e-01
1.09975837e-01 7.16091394e-02 7.03439415e-02 -9.98000443e-01
-3.71625751e-01 7.71745980e-01 1.83258533e-01 -6.75751865e-01
9.95067120e-01 -3.33465308e-01 9.23087522e-02 5.32989323e-01
1.15868092e+00 5.49203493e-02 -1.00643981e+00 -2.18690038e-01
1.26464572e-02 -3.44883889e-01 -1.57510728e-01 -9.42842722e-01
-9.00890648e-01 1.00246036e+00 6.37948692e-01 -3.88549604e-02
9.13817823e-01 5.62344790e-01 5.95337570e-01 2.80402511e-01
3.07077140e-01 -6.20444298e-01 3.72855157e-01 3.13286066e-01
1.10876095e+00 -1.46394098e+00 -1.04270894e-02 -1.70184791e-01
-9.66533184e-01 7.30776608e-01 3.38744283e-01 1.77324280e-01
-7.74986073e-02 -1.61699697e-01 4.94002968e-01 -3.90800387e-01
-6.52325213e-01 -5.46036661e-01 6.00589871e-01 6.54756427e-01
6.14036202e-01 6.29374906e-02 -2.67421514e-01 -2.52623204e-02
-2.85842508e-01 -3.78003299e-01 3.80454600e-01 8.39557767e-01
-3.03390980e-01 -9.74311709e-01 -4.48874414e-01 1.71328217e-01
-2.13426545e-01 -3.54299992e-01 -1.07505393e+00 7.10159838e-01
5.03043197e-02 1.08217096e+00 7.79219493e-02 -1.98424518e-01
2.26967603e-01 1.90033630e-01 8.13873470e-01 -6.86509848e-01
-3.78806502e-01 7.85726085e-02 2.29886353e-01 -5.92673600e-01
-4.61749017e-01 -7.14728475e-01 -1.20840490e+00 -6.42098710e-02
-1.07357778e-01 -1.87415272e-01 5.68680823e-01 1.00712490e+00
3.55525970e-01 1.67567551e-01 2.14298964e-01 -6.07226074e-01
-2.20890924e-01 -7.00382769e-01 1.63274139e-01 7.26976097e-01
5.97057641e-01 -5.65277696e-01 -1.92193985e-01 3.89654309e-01] | [11.361414909362793, 1.427054762840271] |
4fdaeb2d-2e4c-47fe-adf4-828776488dd3 | cross-validation-is-all-you-need-a | 2306.13990 | null | https://arxiv.org/abs/2306.13990v1 | https://arxiv.org/pdf/2306.13990v1.pdf | Cross-Validation Is All You Need: A Statistical Approach To Label Noise Estimation | Label noise is prevalent in machine learning datasets. It is crucial to identify and remove label noise because models trained on noisy data can have substantially reduced accuracy and generalizability. Most existing label noise detection approaches are designed for classification tasks, and data cleaning for outcome prediction analysis is relatively unexplored. Inspired by the fluctuations in performance across different folds in cross-validation, we propose Repeated Cross-Validations for label noise estimation (ReCoV) to address this gap. ReCoV constructs a noise histogram that ranks the noise level of samples based on a large number of cross-validations by recording sample IDs in each worst-performing fold. We further propose three approaches for identifying noisy samples based on noise histograms to address increasingly complex noise distributions. We show that ReCoV outperforms state-of-the-art algorithms for label cleaning in a classification task benchmark. More importantly, we show that removing ReCoV-identified noisy samples in two medical imaging outcome prediction datasets significantly improves model performance on test sets. As a statistical approach that does not rely on hyperparameters, noise distributions, or model structures, ReCoV is compatible with any machine learning analysis. | ['Anne Martel', 'Jianan Chen'] | 2023-06-24 | null | null | null | null | ['noise-estimation'] | ['medical'] | [ 5.85437775e-01 -1.59204692e-01 -1.43856823e-01 -5.58152676e-01
-1.46799898e+00 -6.83203697e-01 1.71242148e-01 8.51797998e-01
-5.67313969e-01 7.58258283e-01 1.83675349e-01 -2.13173330e-01
-4.12043035e-01 -5.49769819e-01 -4.74905878e-01 -8.07026625e-01
1.55928686e-01 5.09051323e-01 9.92754623e-02 3.36694926e-01
1.95595529e-02 3.39452624e-01 -1.50752056e+00 5.90025783e-01
6.07892096e-01 9.93432403e-01 -1.62836701e-01 6.30975246e-01
2.39060968e-02 5.63160002e-01 -6.29207134e-01 -8.73180032e-02
2.23417748e-02 -6.63929582e-01 -8.43594134e-01 -2.63797671e-01
3.45549136e-01 4.02994663e-01 1.70175523e-01 1.17049730e+00
8.68134201e-01 -1.06376767e-01 7.06928313e-01 -1.00678349e+00
-6.70115277e-02 6.01304054e-01 -2.28376493e-01 1.70138791e-01
-1.40795559e-02 4.32374030e-02 7.94136167e-01 -3.99561018e-01
8.42451870e-01 9.07088757e-01 1.22384036e+00 6.21635318e-01
-1.92376006e+00 -7.27901816e-01 -2.28512704e-01 -5.60275018e-02
-1.24204183e+00 -3.11160922e-01 4.20048207e-01 -7.16751456e-01
5.10559738e-01 7.15655327e-01 3.10646266e-01 1.20784080e+00
2.26572588e-01 3.27075899e-01 1.40810740e+00 -4.61523980e-01
6.36038065e-01 -4.56622802e-02 5.41267216e-01 3.94445270e-01
5.82330227e-01 5.37617095e-02 -3.41869175e-01 -6.57537758e-01
2.22635284e-01 -1.78003788e-01 -2.94513524e-01 -4.45551217e-01
-1.30151343e+00 7.99461305e-01 -1.13994805e-02 3.27185601e-01
-2.32350677e-01 -6.07918426e-02 9.52163339e-01 3.05814624e-01
5.99575520e-01 8.28538656e-01 -7.38630176e-01 -1.53677821e-01
-8.82830203e-01 8.55161846e-02 7.32910573e-01 3.93145263e-01
4.59201396e-01 -4.83904064e-01 -4.80277181e-01 1.17406607e+00
-1.67442501e-01 3.50059986e-01 5.77628374e-01 -9.56901491e-01
-1.05637632e-01 4.80125964e-01 -3.06631159e-02 -6.80683017e-01
-1.06930339e+00 -7.99411714e-01 -9.67022538e-01 1.86614186e-01
5.43180645e-01 -2.26923786e-02 -1.05645323e+00 1.72931457e+00
3.29801530e-01 -7.51436129e-02 -7.87117928e-02 5.61625779e-01
9.15460646e-01 -1.56562567e-01 4.25061971e-01 -5.75857282e-01
1.44669914e+00 -3.73436481e-01 -8.33722234e-01 -4.12137546e-02
1.18799579e+00 -6.60473824e-01 1.08352089e+00 7.05965042e-01
-6.03797197e-01 -3.53394449e-01 -6.53413415e-01 1.06487520e-01
-1.84884131e-01 1.30118147e-01 3.09632212e-01 8.25893879e-01
-6.60110533e-01 7.69917667e-01 -9.62059379e-01 -4.36792850e-01
7.36540556e-01 2.87115365e-01 -5.87492406e-01 -1.24358386e-01
-7.00305998e-01 7.06641793e-01 2.04327837e-01 2.03941539e-02
-7.19154537e-01 -9.48126614e-01 -8.43040049e-01 -1.17751390e-01
2.28693500e-01 -6.19313002e-01 1.15527725e+00 -6.35734081e-01
-8.20466995e-01 1.07857144e+00 -1.43996611e-01 -4.03622270e-01
5.54800391e-01 -1.04605570e-03 -4.35449183e-01 -1.16898239e-01
2.17714325e-01 1.23062573e-01 4.71797049e-01 -1.41917467e+00
-4.83447939e-01 -5.01637280e-01 -7.51868188e-01 -4.11484301e-01
-5.59385633e-04 2.70626470e-02 -7.25803971e-02 -6.98796213e-01
4.84918833e-01 -9.62841690e-01 -4.84414190e-01 -3.95600587e-01
-4.86035317e-01 -2.17787996e-02 4.25835639e-01 -3.53078991e-01
1.51863360e+00 -2.08619142e+00 -2.27249146e-01 4.09278363e-01
4.03994948e-01 1.44612521e-01 -6.00867234e-02 3.30448686e-03
-3.24405760e-01 4.31538641e-01 -3.54006916e-01 -3.09509546e-01
-2.86035269e-01 2.08470047e-01 2.11092159e-01 5.79094946e-01
1.36775803e-02 6.22935414e-01 -9.96140182e-01 -4.03857023e-01
9.53894630e-02 2.58423060e-01 -4.69950378e-01 5.33293039e-02
-2.92761084e-02 6.80286050e-01 -1.28759190e-01 5.26346684e-01
6.50000751e-01 -4.51615930e-01 5.55179000e-01 -5.44600427e-01
2.40968391e-01 2.06419349e-01 -1.09966326e+00 1.55895495e+00
-2.57733226e-01 3.45037729e-01 -5.05582765e-02 -1.03014362e+00
7.73466527e-01 1.45881787e-01 6.16996169e-01 -6.14715934e-01
2.74851799e-01 4.18725312e-01 -3.41757722e-02 -7.32863367e-01
-1.16939917e-01 -5.58987796e-01 -2.34231353e-01 2.19043463e-01
8.28524828e-02 8.75493661e-02 1.93560615e-01 -1.92421213e-01
1.67770398e+00 -3.09496582e-01 4.49467361e-01 -4.74306941e-01
2.17716038e-01 1.67791590e-01 9.43065464e-01 1.31219995e+00
-4.20866609e-01 1.09039760e+00 8.00451994e-01 -4.28714663e-01
-7.90509641e-01 -8.99427593e-01 -8.02637577e-01 9.61399972e-01
-2.10539266e-01 -5.14101565e-01 -8.21825147e-01 -1.10820901e+00
5.35803959e-02 4.76283729e-01 -1.07496417e+00 -1.47219852e-01
-3.29792529e-01 -1.44657540e+00 5.95292687e-01 2.95875341e-01
-2.33134791e-01 -1.08747005e+00 -4.13672894e-01 1.96454316e-01
-2.41014779e-01 -9.88405347e-01 -7.16740415e-02 8.63358438e-01
-7.19421804e-01 -1.68899035e+00 -5.01586795e-02 -6.80678487e-01
7.30369210e-01 -6.59426153e-02 1.33158326e+00 1.97749287e-01
-5.86465716e-01 2.94667035e-01 -5.26503623e-01 -5.27054846e-01
-8.12865376e-01 2.22423434e-01 1.58488285e-02 -1.20052718e-01
6.43561959e-01 -3.35190952e-01 -5.80527842e-01 3.18822026e-01
-9.17341053e-01 -4.12438065e-01 3.54927212e-01 1.07259011e+00
1.13907838e+00 1.39833212e-01 7.77988136e-01 -1.50828147e+00
4.68114078e-01 -5.60233355e-01 -4.15457159e-01 2.40160897e-01
-8.23408246e-01 1.73527062e-01 3.97062868e-01 -2.31468603e-01
-6.09838188e-01 2.19024613e-01 -2.76575297e-01 5.83347976e-02
-4.38620895e-01 3.20977300e-01 8.15701932e-02 1.13776259e-01
1.19186330e+00 -3.16449225e-01 1.90053254e-01 -6.49561286e-01
-2.07497805e-01 6.46249533e-01 4.23897088e-01 -3.53965223e-01
2.23450005e-01 4.97399807e-01 2.60134757e-01 -4.90519971e-01
-1.21660936e+00 -7.27032244e-01 -5.26611626e-01 6.43639565e-02
9.30105209e-01 -7.13729739e-01 -7.31311023e-01 3.37183833e-01
-7.07513690e-01 -3.61960649e-01 -6.18938267e-01 6.04726017e-01
-3.53186727e-01 2.13850200e-01 -4.14379627e-01 -6.41994715e-01
-3.39008361e-01 -1.31601727e+00 1.14963102e+00 -1.59896538e-01
-5.37793159e-01 -8.50940466e-01 2.77770579e-01 4.01520461e-01
2.65806973e-01 5.34631729e-01 1.04762089e+00 -1.05268729e+00
9.40285549e-02 -3.04812729e-01 -1.03855841e-01 5.21581829e-01
2.22283915e-01 -3.41411412e-01 -1.18706024e+00 -2.80561596e-01
5.47109880e-02 -3.68411362e-01 1.10759151e+00 7.20522821e-01
1.64647770e+00 1.84543192e-01 -5.22665501e-01 4.86352116e-01
1.40807223e+00 1.66260656e-02 5.66026330e-01 5.14968753e-01
4.55304861e-01 4.56800580e-01 4.63569134e-01 3.41053158e-01
-1.43863708e-02 4.11640853e-01 2.88032293e-01 -2.81367391e-01
-1.35298833e-01 9.25904736e-02 -3.80396158e-01 3.69227290e-01
3.65779936e-01 6.73038827e-04 -1.01190960e+00 5.60076714e-01
-1.76669109e+00 -5.92040539e-01 -6.12013757e-01 2.30423093e+00
9.56190228e-01 2.31885538e-01 1.58625131e-03 3.61677647e-01
6.51090801e-01 -2.65827864e-01 -4.88293022e-01 -7.09166285e-03
-2.25612685e-01 4.02358055e-01 5.44751048e-01 3.30070317e-01
-1.33407080e+00 2.92429626e-01 7.32302475e+00 8.43938410e-01
-7.18979061e-01 4.00766462e-01 1.01035678e+00 -1.88593894e-01
-1.08291253e-01 -2.51980990e-01 -4.96358991e-01 4.86756116e-01
9.00436878e-01 3.66373479e-01 5.75434826e-02 7.42017567e-01
3.99556220e-01 -4.44709957e-01 -1.14575732e+00 1.01361501e+00
-5.34076020e-02 -1.23690343e+00 -3.49132419e-01 -1.64348066e-01
8.10864985e-01 2.16456186e-02 -1.45348087e-01 2.11950362e-01
3.06950301e-01 -1.24185669e+00 2.82023042e-01 5.76299965e-01
7.44929790e-01 -6.00368023e-01 1.30517304e+00 1.37350768e-01
-5.12964070e-01 -1.96091130e-01 -2.45290741e-01 2.14410722e-01
-1.28129825e-01 1.40037906e+00 -6.72873497e-01 3.82745385e-01
1.01223457e+00 2.98778445e-01 -7.72908270e-01 1.22807717e+00
2.74331152e-01 9.86832201e-01 -2.12557688e-01 3.74626935e-01
-3.56420636e-01 1.15298226e-01 2.09915802e-01 1.32592797e+00
1.18257470e-01 -1.53441112e-02 1.68447539e-01 3.60238969e-01
-1.70972407e-01 2.62446672e-01 -3.03964853e-01 3.69980454e-01
3.83585393e-01 1.29198444e+00 -1.06650007e+00 -1.93578944e-01
-1.75528288e-01 5.43242574e-01 2.35145524e-01 5.36402911e-02
-5.29079378e-01 -8.68926719e-02 6.68723762e-01 3.02593738e-01
2.42909249e-02 2.56946772e-01 -7.86963582e-01 -8.26051354e-01
-5.32976463e-02 -1.04990506e+00 6.84718132e-01 -3.06266487e-01
-1.69431555e+00 6.60308480e-01 -2.60800213e-01 -1.20533061e+00
2.42356658e-01 -3.85442048e-01 -5.30280210e-02 4.48758394e-01
-1.17363632e+00 -7.67166078e-01 -5.17948985e-01 -3.41459587e-02
1.53678194e-01 1.11990839e-01 1.11393964e+00 3.72986794e-01
-5.53150415e-01 6.45069122e-01 4.34436738e-01 1.82622567e-01
1.02783036e+00 -1.27327526e+00 2.09468603e-02 5.09115875e-01
-7.30797872e-02 3.87449294e-01 7.62668908e-01 -8.10946286e-01
-7.03169644e-01 -1.27427101e+00 7.20884323e-01 -7.66738892e-01
4.39829618e-01 -4.34493512e-01 -1.09126830e+00 4.67698485e-01
-5.33932984e-01 5.40045142e-01 1.13236809e+00 5.55950522e-01
-3.92892182e-01 -1.61766499e-01 -1.36934817e+00 7.46036321e-02
1.16050971e+00 -2.91761488e-01 -1.23483948e-01 6.10809684e-01
3.76760274e-01 -4.23217297e-01 -1.10628986e+00 8.53736579e-01
4.05619144e-01 -1.06946874e+00 6.38216913e-01 -8.57980072e-01
7.84556344e-02 -3.47482502e-01 -1.90910041e-01 -1.19775736e+00
-4.54016894e-01 -3.28805029e-01 4.15146977e-01 1.11492920e+00
6.32777452e-01 -3.42070132e-01 8.35913956e-01 4.27352160e-01
2.65155849e-03 -6.79535806e-01 -1.05322945e+00 -6.49548650e-01
6.35017306e-02 -7.83632576e-01 3.43965292e-01 1.04946852e+00
-1.89857483e-01 1.74476340e-01 -1.67897433e-01 9.54505503e-02
7.38510132e-01 -1.66159302e-01 5.28847396e-01 -1.64708292e+00
-1.26571178e-01 -1.71430141e-01 -5.00768304e-01 1.02088757e-01
8.13115686e-02 -1.04683352e+00 2.49007747e-01 -1.47256911e+00
5.42331755e-01 -8.43792021e-01 -6.35754287e-01 6.45153046e-01
-5.21504402e-01 7.44309962e-01 -2.52261311e-01 8.27446505e-02
-8.61575782e-01 1.36830911e-01 7.96480894e-01 -1.12228962e-02
-1.28728956e-01 1.72641035e-02 -7.27377355e-01 6.77283764e-01
7.95993984e-01 -1.05313396e+00 -2.05925107e-01 8.69606156e-03
2.56860316e-01 -2.83630371e-01 3.11520547e-01 -1.28000164e+00
-1.21545747e-01 6.01198301e-02 5.43102562e-01 -3.41145992e-01
-3.13641250e-01 -8.05683315e-01 4.61621314e-01 5.33622980e-01
-6.03834331e-01 -1.21350609e-01 1.53014511e-01 6.57363415e-01
-3.37347612e-02 -4.03205454e-01 1.06262183e+00 -1.33222923e-01
-1.82344377e-01 -7.50645846e-02 -3.46851766e-01 3.11931968e-01
8.67632210e-01 3.73283438e-02 -4.02711570e-01 2.17152968e-01
-1.25728548e+00 7.09755719e-02 3.37873340e-01 2.52878249e-01
1.60633862e-01 -1.15891051e+00 -7.54127800e-01 2.59125501e-01
5.05270958e-01 -1.47992168e-02 4.65889484e-01 9.42482591e-01
-3.40450644e-01 5.61534017e-02 1.87093154e-01 -9.24504280e-01
-1.60824656e+00 4.87737507e-01 5.34031689e-01 -7.09472001e-01
-5.39792538e-01 7.15679288e-01 1.42571211e-01 -6.86823964e-01
1.69462934e-01 -4.20155734e-01 -3.85293178e-02 2.20714524e-01
5.39364338e-01 4.28876638e-01 7.90039539e-01 -2.52485603e-01
-4.25597936e-01 4.25951988e-01 -1.09373137e-01 3.36236626e-01
1.38812244e+00 5.79088852e-02 -2.56343693e-01 5.34916520e-01
1.30828619e+00 -5.85058481e-02 -8.57051075e-01 -1.68383345e-01
2.88016140e-01 -2.32557312e-01 8.75252038e-02 -1.14853024e+00
-8.60557497e-01 2.04801887e-01 1.04056871e+00 3.55972975e-01
1.24246824e+00 2.37333458e-02 3.87733459e-01 9.21686292e-02
1.55640274e-01 -1.24822652e+00 -1.73468336e-01 2.66160101e-01
5.33752024e-01 -1.56648195e+00 -1.38546629e-02 -5.48333466e-01
-3.57365400e-01 7.70186722e-01 2.61400759e-01 3.00600767e-01
8.83438766e-01 4.83557224e-01 3.61188710e-01 -3.88459861e-01
-5.74658394e-01 -1.91446990e-01 1.58976823e-01 7.78763652e-01
4.40430105e-01 1.11581326e-01 -6.29469335e-01 9.21529889e-01
-1.59639031e-01 6.73485398e-02 3.84358793e-01 8.08610022e-01
-3.88349444e-01 -1.21377659e+00 -5.58229625e-01 1.17965341e+00
-8.75409782e-01 3.68892848e-02 -1.89568363e-02 5.96376240e-01
5.04377365e-01 1.15883684e+00 -8.77165943e-02 -3.16582829e-01
5.85916102e-01 3.99211258e-01 2.71513522e-01 -6.95703030e-01
-8.56625676e-01 1.79970283e-02 2.72095054e-01 -5.92050433e-01
-3.48082155e-01 -8.54523897e-01 -1.19072056e+00 1.23738758e-01
-4.17738050e-01 2.80060291e-01 6.77483380e-01 8.13998342e-01
4.97385651e-01 8.52783501e-01 4.14329350e-01 -2.67383426e-01
-4.87021238e-01 -1.03888667e+00 -7.14787424e-01 9.40195024e-01
4.90126163e-01 -6.98229790e-01 -5.75888753e-01 -7.40599334e-02] | [14.922709465026855, -2.4548425674438477] |
892a8774-3a63-4ac7-acb6-156e87506a63 | calibration-aware-margin-loss-pushing-the | 2307.04047 | null | https://arxiv.org/abs/2307.04047v1 | https://arxiv.org/pdf/2307.04047v1.pdf | Calibration-Aware Margin Loss: Pushing the Accuracy-Calibration Consistency Pareto Frontier for Deep Metric Learning | The ability to use the same distance threshold across different test classes / distributions is highly desired for a frictionless deployment of commercial image retrieval systems. However, state-of-the-art deep metric learning losses often result in highly varied intra-class and inter-class embedding structures, making threshold calibration a non-trivial process in practice. In this paper, we propose a novel metric named Operating-Point-Incosistency-Score (OPIS) that measures the variance in the operating characteristics across different classes in a target calibration range, and demonstrate that high accuracy of a metric learning embedding model does not guarantee calibration consistency for both seen and unseen classes. We find that, in the high-accuracy regime, there exists a Pareto frontier where accuracy improvement comes at the cost of calibration consistency. To address this, we develop a novel regularization, named Calibration-Aware Margin (CAM) loss, to encourage uniformity in the representation structures across classes during training. Extensive experiments demonstrate CAM's effectiveness in improving calibration-consistency while retaining or even enhancing accuracy, outperforming state-of-the-art deep metric learning methods. | ['Yifan Xing', 'Joe Tighe', 'Ying Nian Wu', 'Jun Fang', 'Qingming Tang', 'Linghan Xu', 'Qin Zhang'] | 2023-07-08 | null | null | null | null | ['metric-learning', 'image-retrieval', 'metric-learning', 'retrieval'] | ['computer-vision', 'computer-vision', 'methodology', 'methodology'] | [-2.27221809e-02 -4.80887353e-01 -2.70105392e-01 -8.63516867e-01
-1.16971970e+00 -6.75060153e-01 3.48332554e-01 2.85252839e-01
-5.83013117e-01 5.37708938e-01 -2.01761067e-01 -1.95559219e-01
-5.77089846e-01 -5.24148226e-01 -4.52947825e-01 -7.74313986e-01
7.74227008e-02 1.69321954e-01 6.38676584e-02 1.13982283e-01
2.51863122e-01 6.56193256e-01 -1.49191928e+00 1.85936823e-01
9.40531611e-01 1.41810870e+00 -2.85965413e-01 3.91840219e-01
3.41558270e-02 3.72758716e-01 -6.93450272e-01 -7.57757127e-01
6.32061422e-01 -3.62282872e-01 -2.80779719e-01 -1.62272856e-01
8.45632017e-01 -4.17960048e-01 -3.81671727e-01 1.18660235e+00
5.72530687e-01 -9.67252105e-02 9.73419309e-01 -1.43888128e+00
-8.04609239e-01 3.03952962e-01 -6.48513675e-01 1.66915268e-01
-9.23322439e-02 3.07092573e-02 1.44607258e+00 -8.34937453e-01
1.78937450e-01 8.37156534e-01 8.43750417e-01 2.95763463e-01
-1.28344405e+00 -7.79956102e-01 1.24412537e-01 1.50360078e-01
-1.61588931e+00 -1.68172866e-01 7.62630522e-01 -5.21713436e-01
3.23238403e-01 1.69813231e-01 3.12859923e-01 8.37119043e-01
4.95614260e-01 5.98773777e-01 9.13700342e-01 -3.04395139e-01
3.27777773e-01 3.39505762e-01 1.40972033e-01 5.84706247e-01
5.73724687e-01 2.86609679e-01 -4.77123290e-01 -2.18414277e-01
5.76650500e-01 -3.26711982e-02 -3.95444483e-01 -1.09636641e+00
-9.12251115e-01 1.03300452e+00 7.11453438e-01 2.49356732e-01
-2.97141429e-02 1.48071945e-01 3.27083677e-01 6.59624219e-01
4.43887830e-01 5.85054219e-01 -3.25843751e-01 -3.90228480e-02
-9.67290878e-01 3.21964711e-01 5.80114722e-01 6.74677193e-01
6.02170408e-01 -2.85217732e-01 -3.68554235e-01 9.66901541e-01
-4.31347303e-02 2.47912362e-01 5.07363737e-01 -8.49334300e-01
4.36583787e-01 6.69981301e-01 1.03491269e-01 -1.04897738e+00
-1.56142414e-01 -7.64467359e-01 -6.89207375e-01 4.25967932e-01
6.23692989e-01 1.30216390e-01 -8.81962657e-01 2.05016613e+00
3.40420216e-01 -2.22785696e-01 -2.16482952e-01 9.68520045e-01
2.09599912e-01 2.09626496e-01 6.01199232e-02 5.07662110e-02
9.71954644e-01 -4.81973350e-01 -3.58190387e-01 -9.27437618e-02
6.79413855e-01 -7.49664724e-01 1.43692422e+00 1.80782959e-01
-7.32355952e-01 -3.82362127e-01 -1.54291677e+00 -5.74085340e-02
-4.90130842e-01 7.22359121e-02 4.49824840e-01 7.61494398e-01
-6.31702363e-01 7.84277737e-01 -5.92284322e-01 1.12961130e-02
4.49520916e-01 4.14513499e-01 -4.03384000e-01 -3.49970728e-01
-1.06814182e+00 9.46563363e-01 1.38301969e-01 -1.34132937e-01
-7.27022111e-01 -1.08135402e+00 -5.54318368e-01 2.26601496e-01
1.57152697e-01 -3.28745514e-01 1.01353872e+00 -9.08867240e-01
-1.04861355e+00 9.53770041e-01 4.16220456e-01 -2.78208286e-01
8.94334912e-01 -3.02349329e-01 -3.34795564e-01 -1.27175629e-01
-8.29894617e-02 5.92030346e-01 8.70446980e-01 -1.26454079e+00
-5.68376482e-01 -5.63821554e-01 3.11795808e-02 2.06027180e-01
-8.26762378e-01 -2.59531856e-01 -3.19297135e-01 -6.62133276e-01
2.55701423e-01 -8.19868028e-01 2.04514310e-01 7.17015147e-01
-1.63562223e-01 -2.50211865e-01 5.84838808e-01 -4.19363648e-01
1.36635375e+00 -2.39251065e+00 -2.45828498e-02 3.92196804e-01
1.97240412e-01 2.01008990e-01 -2.49835059e-01 6.71593025e-02
1.74894035e-02 -6.17921762e-02 -4.04893577e-01 -1.87968537e-01
2.09904045e-01 1.67322248e-01 -2.33869597e-01 7.89534152e-01
3.31702620e-01 5.60262859e-01 -7.51194775e-01 -4.34531212e-01
1.72940880e-01 6.05560899e-01 -4.72749949e-01 3.13629836e-01
1.85887605e-01 -5.50052375e-02 -2.43710726e-01 6.76122487e-01
5.96816480e-01 -3.04189444e-01 -1.63922101e-01 -2.88639784e-01
3.41033787e-01 -4.57112975e-02 -1.19542825e+00 1.55792344e+00
-4.60501194e-01 7.26720214e-01 -2.68950224e-01 -1.05306494e+00
9.94018257e-01 -1.69253960e-01 5.75873971e-01 -8.37006092e-01
1.27990648e-01 4.24211562e-01 2.09596351e-01 -2.40123682e-02
3.35205883e-01 5.03857546e-02 -3.05654984e-02 3.21884155e-01
7.52436463e-03 -3.26621421e-02 1.08383608e-03 -1.32567078e-01
9.57252860e-01 -2.67272532e-01 6.75496534e-02 -4.11242187e-01
2.09293976e-01 -3.65078479e-01 5.85735738e-01 8.33470166e-01
-5.64365983e-01 8.06684554e-01 5.21766484e-01 -3.79366070e-01
-1.35432696e+00 -1.39794528e+00 -7.24654675e-01 9.93765473e-01
3.15967500e-01 -3.11421719e-03 -7.29045630e-01 -8.12635660e-01
5.82717240e-01 4.33171034e-01 -9.11655307e-01 -6.83082461e-01
-2.87628263e-01 -9.04528618e-01 5.27024329e-01 6.65547609e-01
4.91628468e-01 -4.12236631e-01 -6.91294312e-01 -3.65062319e-02
1.90620139e-01 -8.78954887e-01 -7.14326560e-01 3.43851060e-01
-6.30029678e-01 -1.31113338e+00 -8.74492705e-01 -6.74382985e-01
6.83195412e-01 1.42270893e-01 1.16046166e+00 -8.14255625e-02
-4.14403051e-01 2.73761094e-01 -2.18631253e-01 -2.72974104e-01
-1.64652124e-01 1.98980168e-01 2.12118123e-02 5.79702780e-02
6.27760649e-01 -2.46678129e-01 -8.75843704e-01 5.91064155e-01
-8.92342210e-01 -6.49668574e-01 6.26198292e-01 1.01273835e+00
5.42473853e-01 7.38223456e-03 4.71156299e-01 -5.52313685e-01
7.78737724e-01 -2.52079278e-01 -7.59339035e-01 6.81187570e-01
-1.22849083e+00 2.12585166e-01 4.58118290e-01 -8.09162378e-01
-4.25017357e-01 -3.93982261e-01 3.14487994e-01 -5.31111062e-01
4.15625483e-01 2.97081798e-01 -1.19683526e-01 -3.52367491e-01
9.35999930e-01 -1.93260357e-01 2.14616302e-02 -1.88095868e-01
2.52989680e-01 7.91230321e-01 6.01808965e-01 -6.97596192e-01
7.97909737e-01 2.67983556e-01 -6.50708750e-02 -3.13326329e-01
-1.12710309e+00 -5.68891525e-01 -4.23837543e-01 -7.93451816e-02
5.13466060e-01 -7.09612131e-01 -3.12777698e-01 3.31762105e-01
-6.88890398e-01 -2.61615843e-01 -4.25223142e-01 4.82281387e-01
-4.12567466e-01 1.39464363e-01 -3.22880238e-01 -5.42554617e-01
-2.47627795e-01 -1.23819160e+00 9.77210820e-01 1.23401634e-01
-1.57427013e-01 -8.66191804e-01 3.89239639e-02 8.26955065e-02
5.44493079e-01 3.29430908e-01 9.90904689e-01 -6.86201096e-01
-3.00557196e-01 -6.24545872e-01 -5.65619469e-01 7.99016297e-01
2.08390757e-01 4.02563065e-02 -8.51204991e-01 -5.32496989e-01
-1.81470349e-01 -4.28346604e-01 8.75785947e-01 3.23092967e-01
1.53061938e+00 -1.84575114e-02 -8.54552258e-03 8.67257297e-01
1.46083391e+00 1.94607358e-02 6.35083377e-01 4.62596446e-01
4.86810803e-01 5.69432557e-01 6.58092797e-01 3.56346279e-01
-6.80412576e-02 7.71577656e-01 2.93734133e-01 8.72735772e-03
2.27611344e-02 -1.40712008e-01 6.07297048e-02 4.12089199e-01
6.82126343e-01 -2.06906259e-01 -8.22734475e-01 6.31117046e-01
-1.65054727e+00 -5.80016792e-01 5.25522470e-01 2.67222714e+00
9.86479342e-01 3.56308460e-01 -3.89984734e-02 3.68727297e-01
6.89762712e-01 -7.16128051e-02 -9.66132104e-01 -2.43916526e-01
-1.21184401e-01 2.05043945e-02 7.81484365e-01 3.72895122e-01
-1.12973773e+00 5.11596501e-01 6.19824314e+00 9.37754393e-01
-1.18790805e+00 -1.12998515e-01 8.53454053e-01 -1.41818956e-01
-2.44684070e-01 -2.79213309e-01 -7.40053833e-01 5.91394186e-01
8.00287306e-01 -3.12179625e-01 2.17180148e-01 9.75830495e-01
-2.72719651e-01 1.35779813e-01 -1.32738078e+00 1.19825137e+00
1.86459169e-01 -1.05106103e+00 -1.02633545e-02 1.42979786e-01
6.94228292e-01 -5.25100939e-02 4.88400161e-01 2.55015820e-01
1.34166762e-01 -8.99628699e-01 6.79977715e-01 2.14911640e-01
1.03778768e+00 -9.44088697e-01 8.18643212e-01 -1.66540951e-01
-7.79279292e-01 -4.25850064e-01 -5.87542236e-01 6.42715931e-01
-3.54931235e-01 8.52854371e-01 -4.08748627e-01 1.75204396e-01
7.84351587e-01 3.42496425e-01 -7.02264488e-01 1.37209618e+00
1.32444754e-01 2.28492856e-01 -3.21772993e-01 -4.59679461e-04
2.32676581e-01 -1.81321338e-01 1.58707589e-01 9.62177098e-01
4.00877148e-01 -2.46671155e-01 -9.65077057e-02 6.22387648e-01
-4.05117482e-01 2.70880852e-02 -4.44114357e-01 8.56159180e-02
7.99365699e-01 8.50279689e-01 -3.08357030e-01 -2.93347053e-02
-1.90138236e-01 8.39891136e-01 3.43272239e-01 2.40722924e-01
-1.03394902e+00 -6.66316092e-01 9.78802323e-01 9.22513306e-02
2.69747019e-01 -1.29652634e-01 -5.83042622e-01 -9.95823026e-01
4.66403931e-01 -7.76977956e-01 5.95940232e-01 -3.44093680e-01
-1.67244899e+00 3.39749157e-01 7.11583495e-02 -1.38758337e+00
7.47390836e-02 -7.48782158e-01 -2.78314561e-01 5.96677005e-01
-1.54860938e+00 -9.18549418e-01 -4.54096377e-01 4.70274031e-01
1.04101583e-01 -2.72661328e-01 5.94362020e-01 5.55333018e-01
-3.89040112e-01 1.41580582e+00 5.31737745e-01 1.84929058e-01
1.10639703e+00 -1.27890098e+00 -9.18097273e-02 6.67049587e-01
2.63855428e-01 5.10595202e-01 6.60083294e-01 -1.04885865e-02
-1.06569552e+00 -1.06229568e+00 5.66838205e-01 -4.57968801e-01
6.08327806e-01 -3.55103225e-01 -1.08189762e+00 2.66073942e-01
-3.92785728e-01 3.03683192e-01 8.50243926e-01 1.57759145e-01
-9.75418448e-01 -7.14841604e-01 -1.44241250e+00 5.17223179e-01
8.05037498e-01 -6.44534647e-01 -2.80342668e-01 3.08892578e-01
6.61901295e-01 -2.82593280e-01 -1.03482890e+00 6.94236040e-01
8.36498082e-01 -9.52454031e-01 1.07366502e+00 -6.28518581e-01
3.06679517e-01 -1.41387969e-01 -6.37395740e-01 -1.18682885e+00
-2.42033422e-01 -2.02774033e-01 -2.27629036e-01 1.14270306e+00
3.33891332e-01 -4.81422096e-01 8.36066842e-01 9.06564116e-01
3.00736167e-02 -9.83388841e-01 -1.23862338e+00 -1.13528562e+00
5.92660546e-01 -2.29341805e-01 6.23162866e-01 9.50635314e-01
-3.73171836e-01 -5.96093796e-02 -5.42597026e-02 1.19105704e-01
9.03585255e-01 -2.05095261e-02 6.06683433e-01 -1.28826153e+00
-2.10870385e-01 -6.84683442e-01 -8.14621389e-01 -5.71832418e-01
-1.84489880e-03 -6.38751328e-01 -9.98208858e-03 -1.07578146e+00
1.17910467e-01 -8.00958395e-01 -9.03536916e-01 3.52238804e-01
-1.69115961e-01 3.00295413e-01 -3.42087038e-02 2.95505881e-01
-5.07925153e-01 6.84002101e-01 8.78057182e-01 -2.95129061e-01
-1.45743892e-01 -2.27798030e-01 -7.48975813e-01 2.05540478e-01
6.13846421e-01 -5.64436913e-01 -5.38672328e-01 -6.01079464e-01
-2.82354504e-02 -5.07946432e-01 2.59148180e-01 -1.27991247e+00
1.53859019e-01 -1.05307415e-01 5.39838970e-01 -1.97526470e-01
1.47047460e-01 -8.75253499e-01 -2.45203242e-01 3.19494128e-01
-6.80579841e-01 9.14701298e-02 3.78801906e-03 5.85815370e-01
-2.93885946e-01 -6.96541592e-02 1.23530209e+00 4.48170543e-01
-3.97674561e-01 5.08301795e-01 4.62372363e-01 3.43154103e-01
1.09663820e+00 -2.49396592e-01 -2.88496405e-01 -1.54745087e-01
-1.25543013e-01 2.31493458e-01 6.23908997e-01 5.72318137e-01
4.56760168e-01 -1.55945444e+00 -5.80521047e-01 1.13201804e-01
6.88161552e-01 -1.66481361e-01 -7.84582943e-02 5.04491508e-01
-3.96104366e-01 2.13864837e-02 -9.74232927e-02 -8.61653984e-01
-1.02724552e+00 3.92043501e-01 5.47653675e-01 -2.36218721e-01
-2.45551124e-01 9.57490325e-01 9.83810276e-02 -3.66210401e-01
6.67823136e-01 -2.48611942e-01 4.11568642e-01 -1.16609111e-02
5.49788594e-01 2.38939956e-01 4.40852523e-01 -1.42543659e-01
-2.72730529e-01 6.18512332e-01 -3.17173094e-01 -1.74503587e-02
1.04204571e+00 1.34502128e-01 3.51528168e-01 3.79129738e-01
1.63188350e+00 -4.41935748e-01 -1.61770439e+00 -2.50422359e-01
9.20290351e-02 -8.84407938e-01 2.26701781e-01 -8.05217385e-01
-1.07417285e+00 8.89632881e-01 1.20860434e+00 3.47897932e-02
1.02063537e+00 -3.27899992e-01 7.11729228e-01 3.73999238e-01
3.30405861e-01 -1.32631147e+00 2.26354644e-01 8.43478590e-02
8.33210349e-01 -1.57714736e+00 1.95582911e-01 -3.61632043e-03
-5.35993099e-01 9.58890676e-01 6.37016177e-01 -1.04850411e-01
7.24174619e-01 1.59232020e-02 9.82492268e-02 7.09927604e-02
-4.76172805e-01 1.87842667e-01 5.35887122e-01 5.25439322e-01
3.62327397e-01 1.94928367e-02 -2.05760539e-01 1.13630921e-01
-1.33640677e-01 -2.10671961e-01 1.06626727e-01 8.37932289e-01
-3.66121203e-01 -1.05379331e+00 -1.78459495e-01 6.89777553e-01
-3.73752117e-01 1.68931320e-01 -3.52224290e-01 1.00430727e+00
-6.15214333e-02 6.95707142e-01 3.65605325e-01 -4.25836951e-01
3.48818213e-01 -6.30403906e-02 5.51744998e-01 -2.86480457e-01
-2.50724226e-01 -3.17059129e-01 -3.74145001e-01 -5.37700117e-01
-7.41176754e-02 -4.72892076e-01 -7.61730850e-01 -4.32381213e-01
-5.25968671e-01 3.45640667e-02 7.54946411e-01 6.14127517e-01
2.82548755e-01 5.69200628e-02 9.52597022e-01 -2.13121459e-01
-1.35323238e+00 -7.20074654e-01 -6.01644397e-01 7.94639945e-01
4.86972034e-01 -9.88586307e-01 -7.79127955e-01 -4.33465600e-01] | [9.463810920715332, 3.0676050186157227] |
2ed81471-6d73-45c5-8c38-59f478fd431f | stc-spatio-temporal-contrastive-learning-for | 2202.03747 | null | https://arxiv.org/abs/2202.03747v2 | https://arxiv.org/pdf/2202.03747v2.pdf | STC: Spatio-Temporal Contrastive Learning for Video Instance Segmentation | Video Instance Segmentation (VIS) is a task that simultaneously requires classification, segmentation, and instance association in a video. Recent VIS approaches rely on sophisticated pipelines to achieve this goal, including RoI-related operations or 3D convolutions. In contrast, we present a simple and efficient single-stage VIS framework based on the instance segmentation method CondInst by adding an extra tracking head. To improve instance association accuracy, a novel bi-directional spatio-temporal contrastive learning strategy for tracking embedding across frames is proposed. Moreover, an instance-wise temporal consistency scheme is utilized to produce temporally coherent results. Experiments conducted on the YouTube-VIS-2019, YouTube-VIS-2021, and OVIS-2021 datasets validate the effectiveness and efficiency of the proposed method. We hope the proposed framework can serve as a simple and strong alternative for many other instance-level video association tasks. | ['Liqing Zhang', 'Chengjie Wang', 'Ying Tai', 'Yabiao Wang', 'Liang Liu', 'Hang Zhou', 'Jinlong Peng', 'Zhangxuan Gu', 'Zhengkai Jiang'] | 2022-02-08 | null | null | null | null | ['video-instance-segmentation'] | ['computer-vision'] | [-3.44184898e-02 -2.41386518e-01 -4.34779555e-01 -4.58528429e-01
-6.66294396e-01 -3.43766361e-01 7.01485455e-01 6.78988695e-02
-6.51299357e-01 6.40707672e-01 -1.05567552e-01 7.08585829e-02
1.28776893e-01 -4.22768265e-01 -6.79385304e-01 -5.48839629e-01
-1.75716519e-01 -1.77970544e-01 7.84769058e-01 2.44771332e-01
1.63524061e-01 2.64959753e-01 -1.49969900e+00 3.45214158e-01
5.88881791e-01 1.33429480e+00 2.39163473e-01 4.75415945e-01
-9.70790088e-02 8.21440697e-01 -2.13855177e-01 -3.67475301e-01
4.77039367e-01 -4.01043206e-01 -6.94179118e-01 3.23774487e-01
7.16558397e-01 -4.29366320e-01 -3.96324277e-01 8.35017264e-01
3.56843144e-01 4.20018643e-01 3.51525277e-01 -1.33294368e+00
-1.30667478e-01 2.65159637e-01 -8.69922757e-01 5.50718844e-01
3.01496804e-01 3.76883060e-01 9.09328401e-01 -9.03093517e-01
8.03843617e-01 8.26157868e-01 6.41109347e-01 3.84920478e-01
-8.71719778e-01 -6.67177677e-01 5.73707998e-01 4.71911043e-01
-1.30695093e+00 -3.03882569e-01 8.50493014e-01 -5.38442791e-01
6.73889935e-01 2.28816211e-01 1.01095915e+00 1.03442860e+00
-1.41569329e-02 1.15030074e+00 9.65182304e-01 -5.65294102e-02
2.22098500e-01 1.59497969e-02 1.09648690e-01 8.75679493e-01
5.69185168e-02 -6.42895997e-02 -5.31638622e-01 2.52947479e-01
8.64960194e-01 1.86948359e-01 -3.11607152e-01 -4.44991469e-01
-1.43048227e+00 3.57599765e-01 4.05260921e-01 3.25106680e-01
-5.25732994e-01 3.56399626e-01 7.23900378e-01 -2.19638590e-02
6.58720613e-01 2.41092239e-02 -3.63411278e-01 -1.71829745e-01
-1.62597895e+00 7.93015584e-02 4.72061753e-01 1.14577675e+00
5.57177663e-01 1.27062827e-01 -5.47522247e-01 6.05393410e-01
2.57684678e-01 -8.51622075e-02 3.06815714e-01 -1.04646480e+00
3.74480009e-01 4.70198452e-01 1.48641393e-01 -1.07898879e+00
-2.43072450e-01 -4.80652839e-01 -3.93629253e-01 8.28947723e-02
3.42355669e-01 7.96628296e-02 -8.96370232e-01 1.49977672e+00
7.62848914e-01 1.16709983e+00 -3.83441776e-01 1.04085457e+00
1.07325268e+00 5.98166943e-01 4.00036603e-01 -3.55690897e-01
1.32097971e+00 -1.31753147e+00 -8.62955868e-01 1.67832658e-01
5.42041361e-01 -5.09951711e-01 8.98013711e-01 1.29547387e-01
-1.36154354e+00 -8.60295057e-01 -9.44077790e-01 -7.30107948e-02
-3.11283499e-01 2.58402169e-01 6.10637844e-01 3.84307832e-01
-8.19809973e-01 3.55697632e-01 -9.47749257e-01 -2.25907326e-01
7.16755927e-01 2.68370658e-01 -4.03626293e-01 -2.59460378e-02
-9.61863637e-01 3.38397890e-01 6.12593949e-01 3.15998137e-01
-8.27225924e-01 -6.95544660e-01 -1.04316723e+00 -6.54105917e-02
6.32168591e-01 -4.57954705e-01 9.21268940e-01 -1.03408980e+00
-1.28349948e+00 7.47346461e-01 -2.25376070e-01 -5.10838866e-01
6.98110223e-01 -4.36488152e-01 -2.65715510e-01 4.75716114e-01
5.19013889e-02 7.60639668e-01 8.74456942e-01 -1.15508556e+00
-9.07158792e-01 -6.41148314e-02 3.04828525e-01 2.55410552e-01
-5.79853714e-01 7.76704103e-02 -1.27951634e+00 -8.16967309e-01
-2.57978551e-02 -7.66259789e-01 -1.85608715e-01 2.45428249e-01
-3.12749505e-01 -1.98216349e-01 1.11077523e+00 -8.48431468e-01
1.69414151e+00 -2.16050673e+00 9.03988630e-02 6.96044043e-02
2.32280985e-01 4.53406274e-01 -1.40699685e-01 -1.08145989e-01
-2.62438525e-02 -1.03765048e-01 -2.51532972e-01 -4.49044615e-01
-2.83461511e-01 -4.53651436e-02 1.86860338e-01 6.65132225e-01
3.79712641e-01 1.01789057e+00 -1.09045660e+00 -1.00025070e+00
6.39685631e-01 5.55985570e-01 -7.11958408e-01 7.89577737e-02
-2.61606157e-01 5.08531570e-01 -4.20254529e-01 8.65259588e-01
5.12132704e-01 -2.38702789e-01 1.90287083e-02 -6.60122931e-01
-2.25888669e-01 -3.01776017e-04 -1.25167108e+00 1.88008511e+00
-1.88379914e-01 6.27862155e-01 2.43233703e-02 -1.23847783e+00
3.91804814e-01 4.85496134e-01 1.00206172e+00 -7.51089871e-01
1.90733477e-01 3.07216328e-02 -3.68901968e-01 -7.14970112e-01
6.07780516e-01 3.92369062e-01 1.47164196e-01 1.17071643e-01
9.92952436e-02 3.53132129e-01 5.67088366e-01 2.58055419e-01
9.50245321e-01 8.42713654e-01 2.37533480e-01 -1.16681933e-01
7.96982586e-01 -2.21573580e-02 8.61141145e-01 5.08072734e-01
-6.08439565e-01 6.44777000e-01 2.84141451e-01 -4.52094555e-01
-7.61225045e-01 -7.58877397e-01 -9.42171216e-02 6.97556734e-01
6.09818280e-01 -5.19729197e-01 -6.94257855e-01 -1.05292332e+00
-2.57334888e-01 3.91195685e-01 -5.63188136e-01 3.07171375e-01
-9.14728463e-01 -4.46778446e-01 2.95819044e-01 7.04144299e-01
7.68704772e-01 -9.40193653e-01 -5.16248107e-01 4.53457654e-01
-3.40859175e-01 -1.63899505e+00 -7.65358865e-01 -2.99920976e-01
-8.18261743e-01 -1.20908105e+00 -9.14594293e-01 -8.50057006e-01
6.65784597e-01 4.24323231e-01 8.55478346e-01 2.09897757e-01
-3.88724983e-01 5.70811987e-01 -3.93049300e-01 1.00266114e-01
2.26429597e-01 -9.80909392e-02 -1.10252157e-01 1.89313695e-01
1.66636556e-01 -2.03931734e-01 -8.95092368e-01 3.78173798e-01
-8.89344692e-01 2.30074123e-01 2.83409745e-01 7.41517484e-01
8.84103537e-01 -7.70887062e-02 4.69569802e-01 -7.31975317e-01
1.23213366e-01 -4.48591888e-01 -6.03851497e-01 3.35843980e-01
-2.73851335e-01 -4.66595680e-01 3.83309901e-01 -5.35967767e-01
-1.10896575e+00 3.67760599e-01 4.71059941e-02 -9.16112483e-01
1.27411948e-03 4.36248362e-01 4.00984064e-02 -8.80161896e-02
1.76837608e-01 3.49351525e-01 -1.75495833e-01 -2.42808416e-01
3.88179332e-01 3.88384074e-01 6.26354337e-01 -4.03581202e-01
7.11840868e-01 6.45955443e-01 -1.63876936e-01 -7.10665464e-01
-7.95818329e-01 -7.97460556e-01 -8.08959126e-01 -6.52709901e-01
1.29429066e+00 -1.14745307e+00 -5.62610447e-01 3.04062068e-01
-1.01849651e+00 -5.02805173e-01 -1.03182778e-01 6.61486983e-01
-6.45904362e-01 6.59786642e-01 -7.05301583e-01 -7.07712233e-01
-1.88609630e-01 -1.21214354e+00 1.11321330e+00 2.80033529e-01
-1.25491723e-01 -1.00663400e+00 -1.92871094e-01 5.17702460e-01
1.00515801e-02 3.62976581e-01 2.84562528e-01 -3.76056284e-01
-8.68918002e-01 -1.83695033e-01 -5.04478812e-01 3.70652884e-01
1.72856480e-01 3.71992916e-01 -6.53014719e-01 -2.59447515e-01
-3.70941341e-01 -7.31584206e-02 9.36934412e-01 5.55350482e-01
1.60293865e+00 -2.14047715e-01 -4.35761839e-01 6.35784149e-01
1.39144707e+00 3.24983567e-01 5.64624786e-01 4.51924622e-01
9.32271898e-01 3.65665466e-01 1.19264555e+00 4.54008073e-01
4.33435142e-01 9.37820613e-01 4.30686057e-01 -3.78388137e-01
-3.95389289e-01 2.31866315e-01 4.69330668e-01 7.17768013e-01
-3.75117600e-01 -1.67054489e-01 -5.34404278e-01 6.24463558e-01
-2.18228650e+00 -1.37410533e+00 -1.64914533e-01 2.05997872e+00
6.24340475e-01 9.40448940e-02 2.83903331e-01 1.15846202e-01
6.85184777e-01 4.02361244e-01 -3.22106212e-01 1.01757832e-01
1.21460602e-01 7.16241030e-03 2.69912660e-01 2.26895407e-01
-1.55438435e+00 1.08720982e+00 5.59548473e+00 8.36234152e-01
-1.22430372e+00 1.86609089e-01 7.20084846e-01 -3.10026467e-01
1.06059268e-01 -2.30775058e-01 -6.28433704e-01 7.43430972e-01
3.91113520e-01 2.34461322e-01 7.42887035e-02 7.30775416e-01
5.29560328e-01 -9.74848643e-02 -1.05550122e+00 1.10031199e+00
3.48126106e-02 -1.68972862e+00 -1.05326541e-01 -2.16970310e-01
7.20990777e-01 -3.12738299e-01 -4.16116975e-02 3.37652296e-01
-3.53796601e-01 -5.10123968e-01 8.63367498e-01 4.85288680e-01
7.86225021e-01 -6.31780982e-01 5.60293674e-01 -6.09590523e-02
-1.75507700e+00 1.18999854e-02 1.18031465e-01 2.64359087e-01
4.53656286e-01 3.20696026e-01 -4.71267521e-01 7.25626469e-01
9.91422772e-01 1.29395783e+00 -4.45900589e-01 1.36497426e+00
-1.20027803e-01 5.81818044e-01 -2.13090524e-01 3.33404213e-01
5.36680460e-01 -1.30684763e-01 4.10667807e-01 1.62480843e+00
2.21933648e-01 2.88970113e-01 3.31023335e-01 5.72143555e-01
-5.75261889e-03 5.73318526e-02 -2.99722105e-01 -3.46539654e-02
2.90881574e-01 1.53530467e+00 -9.80507374e-01 -6.70207262e-01
-7.86634386e-01 1.10069990e+00 3.03951968e-02 5.71937501e-01
-1.52601290e+00 -5.04156202e-02 5.79554617e-01 1.82815924e-01
6.26974940e-01 -3.38866293e-01 -1.03692777e-01 -1.30820966e+00
1.11606911e-01 -6.51362538e-01 4.53209311e-01 -6.63270175e-01
-9.23800290e-01 4.75438833e-01 1.04496077e-01 -1.64222002e+00
-4.10662442e-02 -2.84368366e-01 -6.69055343e-01 3.23203564e-01
-1.77743411e+00 -1.19860804e+00 -5.79118669e-01 9.52197194e-01
9.62185860e-01 3.14766094e-02 1.54234335e-01 6.80782914e-01
-9.90774572e-01 5.83171248e-01 -3.84089679e-01 3.14170957e-01
5.99168360e-01 -8.90455782e-01 -3.84898148e-02 1.07185340e+00
2.07832828e-01 3.78332317e-01 4.81474578e-01 -7.23977923e-01
-1.13658190e+00 -1.53984666e+00 5.52094877e-01 -2.29771152e-01
5.34509242e-01 -1.27229154e-01 -8.57222438e-01 6.60334229e-01
2.96966374e-01 5.60340941e-01 4.87806320e-01 -4.04585809e-01
-1.59531161e-01 -1.62465394e-01 -1.09937286e+00 6.86028481e-01
1.22537994e+00 -2.54720509e-01 -3.44059229e-01 2.94811636e-01
6.02398276e-01 -5.26696026e-01 -9.27763820e-01 5.97238839e-01
6.19410872e-01 -8.97386909e-01 1.14953768e+00 -5.06598353e-01
5.20359874e-01 -6.07597530e-01 -2.04665046e-02 -6.33868456e-01
-2.77288795e-01 -6.15192413e-01 -5.29135942e-01 1.27362216e+00
2.62441374e-02 -1.96852580e-01 8.87926996e-01 6.20691895e-01
-2.59226412e-01 -1.05905056e+00 -8.88168931e-01 -8.28470647e-01
-6.36575758e-01 -5.79887152e-01 2.01236024e-01 1.01769364e+00
-2.47355506e-01 -2.49011084e-01 -4.68997777e-01 2.39361331e-01
7.73045182e-01 6.36808202e-02 8.00848663e-01 -8.33327174e-01
-2.58028388e-01 -4.16702241e-01 -5.25557578e-01 -1.20728076e+00
1.62419319e-01 -6.41586483e-01 -1.09337002e-01 -1.53884184e+00
1.71911150e-01 -5.56425393e-01 -7.14774966e-01 5.33658028e-01
-4.04791951e-01 6.81956410e-01 2.10703239e-01 2.31776640e-01
-1.17911386e+00 4.27789092e-01 1.31334066e+00 -9.04125422e-02
-2.38964930e-01 -1.84044376e-01 -1.36978954e-01 6.09640121e-01
5.25501728e-01 -2.27114514e-01 -5.17653704e-01 -2.15877622e-01
-3.60129088e-01 4.41490067e-03 4.71876949e-01 -1.02096629e+00
3.57615948e-01 -2.71607727e-01 3.97631615e-01 -7.82159626e-01
6.11728668e-01 -8.91807497e-01 1.65826112e-01 3.57776284e-01
-1.85972035e-01 -3.03561855e-02 2.52222717e-01 7.55349159e-01
-4.55817878e-01 4.93268371e-02 7.48520315e-01 -3.02522276e-02
-1.35089993e+00 7.66763985e-01 -1.26865476e-01 -9.94088799e-02
1.70065379e+00 -6.60572767e-01 -1.66647099e-02 -1.00393675e-01
-6.75003171e-01 4.44218934e-01 3.05481642e-01 5.54531455e-01
7.42361009e-01 -1.27117991e+00 -4.77256477e-01 2.56344397e-02
9.46181789e-02 -1.32402360e-01 4.15602803e-01 1.38066971e+00
-4.77569371e-01 2.22872704e-01 -1.38315141e-01 -8.25047612e-01
-1.48651743e+00 5.51946342e-01 1.29586190e-01 -1.52737647e-01
-8.06134999e-01 9.43791747e-01 2.79867530e-01 1.30452618e-01
3.81742448e-01 -1.65619239e-01 -2.79431015e-01 2.92314053e-01
5.78016222e-01 2.80689716e-01 -1.71046600e-01 -8.47598255e-01
-5.31993270e-01 4.82423633e-01 -9.38922390e-02 4.67243418e-02
1.28045654e+00 -3.23322415e-01 2.58053213e-01 2.48919114e-01
1.08209991e+00 -2.41135195e-01 -1.67046380e+00 -2.46991873e-01
-5.25668487e-02 -5.69088161e-01 8.02747682e-02 -3.26180875e-01
-1.55836904e+00 6.72997653e-01 4.54596788e-01 2.28555165e-02
1.26883888e+00 -1.94274709e-01 9.62194443e-01 -2.31200159e-01
3.57089132e-01 -1.20245147e+00 3.03107679e-01 8.30514207e-02
5.96678138e-01 -1.34467995e+00 9.97786522e-02 -5.68814039e-01
-5.96588671e-01 1.10260129e+00 8.95995080e-01 -7.53167132e-03
6.78853333e-01 9.13739577e-02 -1.89091757e-01 -6.25381246e-02
-5.87884247e-01 -3.27051461e-01 5.24433017e-01 3.18848193e-01
5.39530814e-01 -3.84983540e-01 -6.37220800e-01 2.92062700e-01
6.59302831e-01 2.01561436e-01 2.39486814e-01 9.87363935e-01
-1.44868255e-01 -8.58278871e-01 4.58296901e-03 4.30010468e-01
-5.47989130e-01 -2.92637069e-02 1.83616176e-01 1.04753244e+00
1.57392979e-01 6.72751904e-01 -1.00217061e-02 -3.91886592e-01
1.80938933e-02 -1.34387031e-01 3.74299735e-01 -4.81096476e-01
-7.08439291e-01 2.64270008e-01 2.68375594e-02 -1.15234470e+00
-1.03226626e+00 -7.70775378e-01 -1.41154957e+00 6.85370043e-02
-3.13354313e-01 -2.39759058e-01 4.41894650e-01 8.64807487e-01
3.73950422e-01 6.79403126e-01 4.50299412e-01 -1.09509671e+00
1.39783531e-01 -4.29574788e-01 -2.98575222e-01 5.18937111e-01
2.80216396e-01 -8.15807760e-01 -1.72727853e-01 3.57279688e-01] | [9.121580123901367, -0.04811060428619385] |
9bae6724-f9dd-46bd-84c1-0c676bf9d14f | sam3d-zero-shot-3d-object-detection-via | 2306.02245 | null | https://arxiv.org/abs/2306.02245v1 | https://arxiv.org/pdf/2306.02245v1.pdf | SAM3D: Zero-Shot 3D Object Detection via Segment Anything Model | With the development of large language models, many remarkable linguistic systems like ChatGPT have thrived and achieved astonishing success on many tasks, showing the incredible power of foundation models. In the spirit of unleashing the capability of foundation models on vision tasks, the Segment Anything Model (SAM), a vision foundation model for image segmentation, has been proposed recently and presents strong zero-shot ability on many downstream 2D tasks. However, whether SAM can be adapted to 3D vision tasks has yet to be explored, especially 3D object detection. With this inspiration, we explore adapting the zero-shot ability of SAM to 3D object detection in this paper. We propose a SAM-powered BEV processing pipeline to detect objects and get promising results on the large-scale Waymo open dataset. As an early attempt, our method takes a step toward 3D object detection with vision foundation models and presents the opportunity to unleash their power on 3D vision tasks. The code is released at https://github.com/DYZhang09/SAM3D. | ['Xiang Bai', 'Zhe Liu', 'Xiaoqing Ye', 'Zhikang Zou', 'Hongcheng Yang', 'Dingkang Liang', 'Dingyuan Zhang'] | 2023-06-04 | null | null | null | null | ['3d-object-detection'] | ['computer-vision'] | [-2.37179875e-01 -7.34920707e-03 -1.08192861e-01 -3.60491425e-01
-4.89912778e-01 -5.77962160e-01 8.72052133e-01 -3.63530755e-01
-4.18966651e-01 -5.82391955e-02 -4.76796879e-03 -5.30568838e-01
3.56511354e-01 -5.03314793e-01 -4.60813284e-01 -3.19926739e-01
1.14026956e-01 4.53799784e-01 6.23745799e-01 -3.24040294e-01
1.57922640e-01 4.68052983e-01 -1.64664328e+00 2.12129816e-01
5.59905648e-01 7.08128095e-01 4.98167902e-01 8.70621324e-01
-3.45762253e-01 5.49390852e-01 -1.81193039e-01 -3.90249431e-01
4.61499959e-01 -3.45481694e-01 -8.44068587e-01 1.27540827e-01
8.34212303e-01 -4.26459074e-01 -3.54680061e-01 1.07728398e+00
4.90676284e-01 -1.11442693e-01 5.24287462e-01 -1.21874630e+00
-9.67639983e-01 5.44919848e-01 -7.40523279e-01 4.81848180e-01
1.43936738e-01 4.41251427e-01 1.23924804e+00 -1.03913403e+00
6.93827152e-01 1.43017602e+00 6.36145353e-01 7.65383303e-01
-1.11088681e+00 -2.46103927e-01 1.05404459e-01 2.46101409e-01
-1.02152538e+00 -5.23763657e-01 6.45195723e-01 -5.74760973e-01
1.24206352e+00 7.88204074e-02 7.48601437e-01 1.05717123e+00
-2.41587535e-01 1.42795491e+00 1.19420266e+00 -3.54888529e-01
-1.08648613e-01 4.06363048e-03 2.92302072e-01 9.22027588e-01
6.67717159e-02 -3.72545719e-02 -6.44471765e-01 3.75312001e-01
6.18730307e-01 -2.84371316e-01 -1.75155938e-01 -6.38830885e-02
-1.19327581e+00 7.74664283e-01 6.48229718e-01 2.24072129e-01
6.66042492e-02 2.71437317e-01 2.52577603e-01 3.79413180e-02
6.12144768e-01 2.10593238e-01 -3.29195231e-01 -1.08246535e-01
-9.70500946e-01 1.95818126e-01 6.88359261e-01 1.02469647e+00
4.65791702e-01 -6.98825791e-02 -1.17249504e-01 8.24482262e-01
5.45895040e-01 6.12289131e-01 1.78390458e-01 -1.19362390e+00
1.31595373e-01 6.57373071e-01 -1.04314141e-01 -5.33045948e-01
-4.40420121e-01 -2.98851043e-01 -4.03422743e-01 5.20541310e-01
7.03751504e-01 1.56855792e-01 -1.12858176e+00 1.50035703e+00
4.62990880e-01 1.71734706e-01 -2.36275449e-01 1.16045952e+00
1.20876610e+00 6.39213741e-01 -5.34971505e-02 4.15238887e-01
1.45921552e+00 -1.07261598e+00 -1.55220658e-01 -4.85342234e-01
4.90846008e-01 -8.40554178e-01 1.24695063e+00 3.10160249e-01
-1.18077171e+00 -6.70177937e-01 -8.53687048e-01 -4.75735396e-01
-3.37792575e-01 -9.14014876e-02 9.99631345e-01 6.37533307e-01
-1.30043685e+00 2.28938773e-01 -9.52611923e-01 -7.12718904e-01
8.24975967e-01 3.03931683e-02 -2.59079456e-01 -1.95184320e-01
-8.68389785e-01 1.05329978e+00 2.17764035e-01 2.15076581e-01
-1.00652158e+00 -4.33317035e-01 -7.17362583e-01 -3.65311086e-01
5.35997331e-01 -9.26564693e-01 1.29445386e+00 -6.78116441e-01
-1.22594261e+00 1.52888298e+00 -1.39522225e-01 -4.80014503e-01
6.02201223e-01 -1.95524946e-01 1.26910165e-01 1.58484593e-01
9.87296402e-02 1.12868416e+00 8.33493829e-01 -9.86570239e-01
-7.51642346e-01 -4.59516943e-01 2.89505303e-01 4.77091968e-02
-9.55850929e-02 2.32025996e-01 -6.98937297e-01 -2.78331161e-01
5.29110692e-02 -8.81438911e-01 -1.48803934e-01 4.95529115e-01
-3.54976922e-01 -5.63068628e-01 6.89532399e-01 -3.36145282e-01
6.12400830e-01 -2.05920982e+00 9.02696550e-02 -4.72373575e-01
5.34382522e-01 6.10155702e-01 -3.26398522e-01 1.32837117e-01
5.20288229e-01 2.92557776e-01 -3.51492375e-01 -5.12973487e-01
2.35339299e-01 9.76727828e-02 -2.02904761e-01 4.19432044e-01
4.92944062e-01 1.40358615e+00 -8.79517138e-01 -3.72120172e-01
3.06632966e-01 5.70696712e-01 -5.15882254e-01 -3.96455824e-03
-4.04931426e-01 1.46894291e-01 -4.53811735e-01 8.94232571e-01
5.61261535e-01 -4.54160899e-01 -4.46884811e-01 -5.64815514e-02
-5.09345889e-01 2.04257950e-01 -5.38863838e-01 1.88912010e+00
-3.89681123e-02 9.30346787e-01 3.91583323e-01 -7.74328768e-01
8.17185462e-01 -3.42150107e-02 2.46825427e-01 -6.35487258e-01
4.06375557e-01 1.53767601e-01 1.95618242e-01 -6.78393304e-01
4.46837068e-01 -2.54668314e-02 1.17823333e-01 2.48389825e-01
9.50044394e-02 -5.87886691e-01 2.49099687e-01 4.55494314e-01
9.77646351e-01 4.57128495e-01 -3.21330093e-02 -8.27097073e-02
2.80592710e-01 4.73817706e-01 1.16360597e-01 7.98246145e-01
-6.81517661e-01 6.97523117e-01 2.78713971e-01 -2.28999928e-01
-1.08985877e+00 -1.15486908e+00 -2.71161675e-01 9.39934075e-01
1.69701710e-01 -2.27302104e-01 -6.69975579e-01 -6.17529213e-01
1.95732698e-01 4.88429427e-01 -5.36406338e-01 9.15723443e-02
-2.63276577e-01 -7.59330094e-01 8.07912707e-01 5.17248213e-01
6.97044611e-01 -1.00766468e+00 -6.66621685e-01 -6.71875849e-02
3.85963321e-02 -1.29108059e+00 -2.77362287e-01 1.15045952e-02
-5.88528931e-01 -9.07370925e-01 -9.89748359e-01 -9.17568684e-01
2.94170797e-01 7.73780584e-01 1.07267511e+00 1.21270098e-01
-5.62262416e-01 5.83034694e-01 -4.01764303e-01 -6.72656834e-01
-3.63854468e-01 -8.14199299e-02 9.34099928e-02 -3.22668105e-01
6.62145734e-01 -3.21597636e-01 -5.78760445e-01 1.67069465e-01
-5.56102991e-01 2.13863492e-01 3.47136319e-01 3.00889552e-01
4.77699727e-01 -6.25884116e-01 6.39174521e-01 -4.95308250e-01
3.06349188e-01 -4.04574513e-01 -6.62387669e-01 -4.79664430e-02
-3.36318284e-01 -2.52915382e-01 1.00551844e-01 -1.86012104e-01
-9.17117178e-01 -2.94345301e-02 -5.54362237e-01 -3.82270783e-01
-3.23132664e-01 3.09809715e-01 9.67798159e-02 8.83942246e-02
5.23688972e-01 6.60196021e-02 5.84259965e-02 -7.69379973e-01
9.17750716e-01 8.37256014e-01 6.47079110e-01 -4.07254964e-01
7.23700285e-01 8.74943733e-01 -1.03802711e-01 -1.09325612e+00
-1.13178992e+00 -5.36812365e-01 -7.63786495e-01 -1.57110840e-01
1.20027924e+00 -9.88342881e-01 -5.24804711e-01 7.18509257e-01
-1.19949806e+00 -5.88227510e-01 -1.16984881e-01 2.55659252e-01
-4.47870553e-01 5.46504259e-01 -6.88986123e-01 -8.68334055e-01
-2.01149896e-01 -1.04094589e+00 9.64565337e-01 4.87274915e-01
-6.77370131e-02 -8.83204162e-01 -1.06435634e-01 7.50559151e-01
2.30900839e-01 -2.81255040e-02 6.39810562e-01 -4.80252236e-01
-7.76791871e-01 -6.55505434e-02 -5.64785540e-01 5.00273824e-01
-2.39745960e-01 1.12916276e-01 -1.21940196e+00 3.59704159e-02
-1.35882407e-01 -4.03312504e-01 1.37097394e+00 4.93804723e-01
8.15855265e-01 4.14134532e-01 -2.21172094e-01 7.67210722e-01
9.55202162e-01 -9.24391448e-02 3.03538412e-01 2.85594761e-01
8.44813585e-01 5.39460599e-01 3.85749102e-01 5.22768684e-02
7.35762537e-01 5.23517370e-01 6.65361762e-01 -1.68180868e-01
-7.93048322e-01 -2.51789421e-01 3.52475643e-01 7.24505186e-01
-1.18539684e-01 -2.79017597e-01 -1.17893422e+00 6.81037188e-01
-1.69534230e+00 -9.41268146e-01 -5.88410079e-01 1.71316743e+00
5.22055507e-01 4.84895587e-01 2.52334058e-01 -2.89544135e-01
4.73124444e-01 3.44063759e-01 -7.11476624e-01 -5.34937561e-01
-4.06596273e-01 -1.95011228e-01 7.13051558e-02 4.42065418e-01
-1.05215418e+00 1.36359000e+00 5.77917957e+00 5.96392274e-01
-1.02397430e+00 1.69663325e-01 6.01572454e-01 -1.95016656e-02
-1.91815764e-01 -4.31578234e-02 -1.17829120e+00 1.85861349e-01
5.82054973e-01 1.13979794e-01 3.36774707e-01 8.34840357e-01
4.41889256e-01 -2.29513720e-01 -9.51305568e-01 9.55554128e-01
2.04182565e-01 -1.38010764e+00 -1.67316601e-01 -5.45368865e-02
5.26920915e-01 9.45901275e-01 1.24557666e-01 2.26745188e-01
3.04333597e-01 -8.86612594e-01 8.72489214e-01 3.54607612e-01
6.41464770e-01 -2.65174627e-01 4.09864038e-01 6.29462361e-01
-1.02233088e+00 1.98729143e-01 -6.56739175e-01 -2.01316580e-01
4.65637058e-01 5.37073255e-01 -9.51084852e-01 4.16554250e-02
8.99543703e-01 1.03023040e+00 -6.40553832e-01 1.23407948e+00
-5.26174486e-01 8.50519061e-01 -4.66492385e-01 -1.07553430e-01
4.66050982e-01 -1.09927624e-01 9.82556939e-01 1.34500837e+00
1.65389508e-01 4.60881889e-02 1.88129440e-01 1.15503740e+00
-1.32964164e-01 -1.67761639e-01 -7.12529957e-01 -2.53972501e-01
2.34340191e-01 1.44559956e+00 -1.04898119e+00 -3.04738909e-01
-7.13879943e-01 1.02881467e+00 2.78563231e-01 1.43953606e-01
-8.48238647e-01 -1.35297135e-01 7.41914630e-01 3.72737879e-03
4.88481581e-01 -5.77284873e-01 -4.46621418e-01 -1.27987385e+00
-2.61032522e-01 -5.50984025e-01 9.87505764e-02 -1.07785285e+00
-1.44725120e+00 4.85353559e-01 -2.68017918e-01 -7.63404667e-01
2.57462740e-01 -9.30830359e-01 -6.98751748e-01 7.25037754e-01
-1.62559128e+00 -1.49331951e+00 -1.60658807e-01 4.54323411e-01
8.56494486e-01 -3.28653418e-02 4.08323020e-01 9.39721018e-02
-5.37362754e-01 1.51361793e-01 -2.76263475e-01 2.95335829e-01
5.38784504e-01 -1.14755237e+00 8.86661530e-01 9.39190924e-01
7.19393373e-01 4.22623843e-01 5.21885991e-01 -5.04988849e-01
-1.67248368e+00 -8.25976849e-01 8.53395998e-01 -1.01937962e+00
8.35721612e-01 -6.01968765e-01 -8.39577854e-01 7.10163176e-01
1.12832390e-01 5.77916484e-03 2.52257913e-01 1.30553931e-01
-6.48933649e-01 2.43336290e-01 -9.04791832e-01 6.96813703e-01
1.40562272e+00 -5.18615901e-01 -9.43627596e-01 2.18340993e-01
7.73321629e-01 -2.74788082e-01 -4.96227473e-01 1.80282399e-01
4.35077608e-01 -1.01196837e+00 1.15456295e+00 -4.72883373e-01
4.09868628e-01 -4.63565648e-01 -2.00679317e-01 -9.07744169e-01
-3.27791691e-01 -5.77381074e-01 -1.88616246e-01 1.33981407e+00
4.42194492e-01 -4.42286640e-01 6.54237151e-01 4.11066264e-01
-4.95470047e-01 -7.28292465e-01 -9.34991598e-01 -8.23507547e-01
1.41932815e-01 -9.03764129e-01 2.03972861e-01 6.70046031e-01
-1.72064558e-01 6.34413540e-01 -6.86116144e-02 6.37102723e-02
7.68582880e-01 2.13448569e-01 8.84288192e-01 -1.21721315e+00
-3.19378257e-01 -7.99005151e-01 -3.62687081e-01 -1.45347917e+00
4.27842177e-02 -1.52242351e+00 4.16356884e-02 -1.92222321e+00
2.64680773e-01 -2.97593951e-01 -1.77282900e-01 7.34707654e-01
-1.01171531e-01 8.28324497e-01 6.16876423e-01 3.28259826e-01
-8.05282235e-01 3.12031418e-01 1.35972178e+00 -2.78613508e-01
-1.02943160e-01 7.67790247e-03 -8.53841424e-01 1.07329655e+00
8.32067370e-01 -1.78091496e-01 -2.23903269e-01 -8.84661973e-01
2.04520419e-01 -3.61907959e-01 7.53428340e-01 -6.78990066e-01
1.47984341e-01 1.10848397e-01 1.53651118e-01 -7.29931533e-01
4.26971018e-01 -3.75155509e-01 -5.19853950e-01 4.55379814e-01
-1.08048908e-01 -2.89764613e-01 3.17110270e-01 3.64239693e-01
9.02495161e-02 -8.30961615e-02 8.06468546e-01 -3.87985915e-01
-1.28261912e+00 4.80678737e-01 -2.57969588e-01 2.86368072e-01
7.79378176e-01 -3.18363160e-01 -3.94247681e-01 -3.70893143e-02
-7.28364229e-01 4.79536951e-01 5.04681349e-01 7.60951996e-01
6.53912544e-01 -7.97394753e-01 -9.18462396e-01 9.68027562e-02
3.79758030e-02 -1.40203480e-02 2.39032265e-02 1.04572833e+00
-4.29571062e-01 3.39722276e-01 -6.71690926e-02 -1.04214668e+00
-1.17181158e+00 5.19575477e-01 2.72620827e-01 3.58026534e-01
-1.00826323e+00 1.42313230e+00 1.57885596e-01 -2.66464651e-01
8.19348469e-02 -4.42353785e-01 -5.67849912e-02 1.85763985e-01
6.42117202e-01 3.04163963e-01 -2.73415655e-01 -6.28808141e-01
-4.85510111e-01 7.14903414e-01 -1.69480056e-01 -4.18142863e-02
1.62406564e+00 -3.80301297e-01 -1.91063881e-01 5.96175790e-01
9.73740935e-01 -1.24937832e-01 -1.63424599e+00 -1.88581675e-01
1.45148151e-02 -2.49949291e-01 2.36424327e-01 -9.41473007e-01
-8.94446135e-01 1.33189738e+00 4.15117115e-01 3.32024306e-01
7.15900362e-01 6.58516228e-01 9.24689114e-01 1.30501792e-01
4.28974062e-01 -8.83253038e-01 1.90301295e-02 8.99668992e-01
6.28668368e-01 -1.65635800e+00 -1.89826638e-01 -1.56492889e-01
-5.97113371e-01 9.66921866e-01 5.18693149e-01 -2.24445716e-01
5.84306419e-01 2.96041727e-01 2.38112524e-01 -3.29273671e-01
-6.30577385e-01 -8.51695001e-01 4.17489648e-01 8.64750683e-01
4.21794742e-01 -5.76162757e-03 -8.91670063e-02 4.14322436e-01
-9.24511701e-02 -4.70210053e-02 4.23616201e-01 4.11705106e-01
-9.09379125e-01 -9.70070004e-01 -1.43956989e-01 2.30521947e-01
-3.19834739e-01 -2.57179588e-01 -6.37506902e-01 6.99968934e-01
2.62910753e-01 9.62297976e-01 -2.63265576e-02 -1.48095116e-01
2.44627580e-01 6.68747574e-02 5.71597636e-01 -7.39983737e-01
-2.74513543e-01 6.91782534e-02 -6.25860021e-02 -4.73065734e-01
-4.41486806e-01 -6.63691521e-01 -1.51813245e+00 -2.62552768e-01
-2.17025265e-01 -2.87334919e-01 8.93624902e-01 8.96886647e-01
2.59594113e-01 1.96476623e-01 7.76443258e-02 -1.13710093e+00
-4.30456102e-01 -6.63947284e-01 -2.96790183e-01 7.95713440e-02
3.45823020e-01 -4.65311170e-01 -4.04595673e-01 -1.04546592e-01] | [9.831748008728027, 1.2322089672088623] |
6330d7db-acd8-4757-97ce-dc469f6663c5 | demand-response-by-aggregates-of-domestic | 2307.02218 | null | https://arxiv.org/abs/2307.02218v1 | https://arxiv.org/pdf/2307.02218v1.pdf | Demand Response by Aggregates of Domestic Water Heaters with Adaptive Model Predictive Control | This paper describes an intelligent management algorithm for an aggregate of domestic electric water heaters called to provide a demand response service. This algorithm is developed using Model Predictive Control. The model of the entire aggregate is dynamically identified using a recursive polynomial model estimation technique. This allows the control to be adaptive, i.e., able to adjust its decisions to the system characteristics, which vary over time due to the daily distribution of users hot water consumption. To answer the demand response requirements, aggregated power variations are realized by modifying the temperature set-points of the water heaters without compromising the users comfort. The developed approach allows tracking a regulation signal and mitigating the so-called rebound, i.e., the recovery of energy needed by the aggregate at the end of the service to return to the baseline thermal state. Analyses in a simulation environment allow the validation of the potentialities of the proposed method. | ['M. Rapizza', 'D. Cirio', 'F. Silvestro', 'S. Massucco', 'F. Conte'] | 2023-07-05 | null | null | null | null | ['management'] | ['miscellaneous'] | [ 4.67156954e-02 2.61499882e-01 -2.27648243e-02 -1.98020209e-02
6.52667657e-02 -7.72228956e-01 3.55213761e-01 3.35503221e-01
1.32073238e-01 5.86723685e-01 -4.11802679e-01 -1.86084643e-01
-5.83068252e-01 -8.57367396e-01 -1.68753594e-01 -1.18681967e+00
1.61947936e-01 5.48060238e-01 -5.51351719e-02 -1.12109847e-01
1.46529928e-01 7.92567313e-01 -1.41217756e+00 -4.20582354e-01
1.01645648e+00 8.46342087e-01 3.41624141e-01 4.27128941e-01
2.40544498e-01 3.20212282e-02 -4.05876696e-01 7.66405702e-01
-7.16300588e-03 -3.27532351e-01 -1.89547360e-01 2.58821011e-01
-6.98205233e-01 -3.54424238e-01 2.18396738e-01 6.70349121e-01
3.49205464e-01 4.49986637e-01 8.22192669e-01 -1.40800691e+00
2.59856433e-01 1.66014954e-01 1.01157285e-01 -2.02102646e-01
-7.66183138e-02 1.62313983e-01 5.26199043e-01 -1.76557794e-01
1.05933733e-02 6.28659368e-01 1.82699729e-02 2.59509951e-01
-1.50305247e+00 -2.99031943e-01 -1.05512538e-03 2.21222475e-01
-1.45240831e+00 -1.32416770e-01 7.57704854e-01 -4.10404652e-01
8.69463921e-01 5.75539112e-01 9.27286148e-01 1.19463541e-01
3.99992138e-01 2.88433164e-01 1.00299609e+00 -3.41860592e-01
8.78846705e-01 3.40033233e-01 -7.77329579e-02 -9.74243060e-02
3.36766183e-01 -2.87628800e-01 2.56480932e-01 -3.18624943e-01
2.31661409e-01 -2.49646738e-01 -5.04885077e-01 -4.08269763e-01
-1.08618349e-01 2.02237993e-01 1.92964464e-01 8.28120828e-01
-7.25369811e-01 -9.45571363e-02 2.11468711e-01 -6.79700822e-02
3.84606987e-01 3.75067979e-01 -6.13550484e-01 -1.44622251e-02
-7.96669066e-01 4.63435054e-02 9.04426754e-01 5.96286952e-01
6.03379190e-01 3.12639058e-01 8.74395445e-02 2.36529335e-01
3.85210901e-01 6.38641238e-01 4.59692329e-01 -6.55273199e-01
-1.26508102e-01 6.85674310e-01 7.85547554e-01 -6.12967134e-01
-6.34591401e-01 -3.63328964e-01 -5.79685748e-01 3.16006005e-01
1.03384636e-01 -7.58968234e-01 -5.03559589e-01 1.32470155e+00
6.76388860e-01 -3.44448656e-01 -3.48213427e-02 5.30417562e-01
-3.29229534e-01 1.17149842e+00 2.12411612e-01 -8.01028073e-01
1.09396994e+00 -2.57520545e-02 -1.13252521e+00 3.01191568e-01
5.16087711e-01 -2.66380429e-01 2.99984485e-01 1.06837459e-01
-7.52866149e-01 -2.05340102e-01 -8.53256881e-01 7.81875134e-01
-4.22121644e-01 2.93990135e-01 -3.79331738e-01 5.44459462e-01
-9.63965118e-01 5.31862020e-01 -8.77844930e-01 -3.81544441e-01
-5.39637983e-01 3.76632094e-01 2.42299467e-01 6.00455523e-01
-1.18833709e+00 1.12108231e+00 6.26828015e-01 6.97224677e-01
-4.29064900e-01 -9.37309444e-01 -4.01985079e-01 5.22364140e-01
3.46108600e-02 -4.25850242e-01 1.12748504e+00 -9.42467093e-01
-1.83784568e+00 -1.83867421e-02 -3.43883038e-02 -6.96708858e-02
5.49087286e-01 1.13218121e-01 -5.30564845e-01 6.96804672e-02
-5.37858009e-01 -4.39500570e-01 4.25772578e-01 -1.03556156e+00
-5.07952988e-01 -2.63514608e-01 -5.25533259e-01 4.09840941e-01
-4.05453980e-01 -4.74420518e-01 2.68624038e-01 1.44349158e-01
-1.45068094e-01 -1.00620401e+00 -3.68025392e-01 -4.57686633e-01
-2.55315334e-01 -4.60990101e-01 1.15923142e+00 -6.01339042e-01
1.19490361e+00 -1.77324367e+00 4.07287508e-01 9.59440768e-01
-5.99565685e-01 1.92203462e-01 3.87810767e-01 1.02807355e+00
-2.26280063e-01 -1.57253221e-01 -3.30851048e-01 9.71827134e-02
2.31619075e-01 2.14773238e-01 1.16854340e-01 4.80236173e-01
-5.07173315e-03 5.46058267e-03 -5.51518202e-01 2.63352603e-01
6.80893362e-01 3.50562543e-01 -2.30483282e-02 5.21898866e-01
-2.93840230e-01 3.31944734e-01 -8.47469270e-01 3.70521867e-03
4.82319832e-01 4.66225207e-01 6.86967134e-01 -4.58268464e-01
-7.07793951e-01 -3.36757809e-01 -1.43010128e+00 7.07792342e-01
-8.51363301e-01 1.59334898e-01 5.57820916e-01 -1.20640993e+00
1.11817122e+00 7.01212823e-01 9.68945920e-01 -6.52518153e-01
2.62096643e-01 4.82485414e-01 -2.31966287e-01 -3.89792234e-01
4.09323514e-01 2.06759065e-01 1.24185272e-01 2.36708626e-01
-4.76052910e-01 -4.39689398e-01 6.62009344e-02 -2.92690992e-01
5.89983582e-01 6.06778637e-03 1.27246752e-01 -8.68476093e-01
9.41253662e-01 -1.77803159e-01 6.02478862e-01 -3.17124754e-01
1.71119586e-01 -6.24796629e-01 4.26059693e-01 -9.98743251e-02
-1.10375619e+00 -2.46507034e-01 -2.81941801e-01 5.03216863e-01
2.05675602e-01 4.74748880e-01 -8.05490851e-01 5.50083891e-02
-7.91839883e-02 1.37867486e+00 -1.68047637e-01 -3.57297421e-01
-5.71121156e-01 -9.12093461e-01 -4.63948667e-01 -1.97966155e-02
3.77002180e-01 -5.56881189e-01 -1.20561397e+00 6.89512074e-01
7.50165358e-02 -7.52542913e-01 5.10538276e-03 3.43345463e-01
-8.97698641e-01 -9.80476856e-01 -4.92752165e-01 -2.47292846e-01
8.80217135e-01 -5.63034952e-01 6.78384066e-01 1.32968992e-01
-4.73732017e-02 5.47864377e-01 -1.45046443e-01 -2.01888412e-01
-6.50008500e-01 3.08169782e-01 -2.79310197e-02 2.62196034e-01
-8.97515789e-02 -4.23222721e-01 -5.52698851e-01 5.07715344e-01
-1.03976512e+00 -1.09948337e-01 1.59093700e-02 1.16519354e-01
4.00019228e-01 4.94463235e-01 8.17138851e-01 -5.66953361e-01
8.32166255e-01 -5.92587471e-01 -1.19344020e+00 4.97451037e-01
-9.85633790e-01 2.43757233e-01 9.34015453e-01 -9.65215266e-02
-1.19770837e+00 4.82384026e-01 2.23862737e-01 1.73705116e-01
-2.13777378e-01 1.43234348e-02 -6.82938099e-01 4.79170941e-02
-1.67483717e-01 3.20791483e-01 -3.13905962e-02 -4.50845033e-01
8.76058340e-02 5.42392612e-01 2.01905087e-01 -4.45072621e-01
9.81367648e-01 -2.64647752e-01 4.79380429e-01 -6.33848965e-01
3.41104746e-01 -2.80198008e-01 -4.14500743e-01 -6.94863856e-01
5.39575338e-01 -4.69584316e-01 -1.01116788e+00 8.12840044e-01
-7.54577279e-01 -4.95178640e-01 -2.91294996e-02 5.85366674e-02
-5.02093613e-01 1.64853670e-02 -1.52560382e-03 -1.32011318e+00
-6.78606749e-01 -9.04995859e-01 3.26607168e-01 7.74772882e-01
-1.35261118e-01 -1.17584252e+00 8.54237378e-02 -3.84022653e-01
6.67932689e-01 6.43098593e-01 1.00124037e+00 -2.44904280e-01
-3.80755275e-01 -2.71134466e-01 6.45899951e-01 2.68333465e-01
1.97405264e-01 4.01236385e-01 -7.04535246e-01 -5.03594220e-01
4.43987697e-02 3.69820178e-01 -1.82324916e-01 2.59786427e-01
7.08015859e-01 -3.53232414e-01 -4.58114952e-01 6.84935227e-02
1.93539953e+00 7.68106401e-01 4.27759618e-01 2.57073104e-01
-4.07151319e-02 7.10351646e-01 6.46246076e-01 8.83161783e-01
4.22168434e-01 7.99610376e-01 8.00264537e-01 -1.69091240e-01
9.43082929e-01 3.40183139e-01 1.97091743e-01 2.25899160e-01
1.27526686e-01 -4.52083379e-01 -7.17053771e-01 5.11780798e-01
-1.68214715e+00 -5.87889850e-01 -3.19149762e-01 2.17242575e+00
4.34937030e-01 -2.83855855e-01 -9.29542556e-02 4.02679265e-01
7.99738467e-01 -2.00370491e-01 -5.71770728e-01 -1.08682013e+00
3.98145199e-01 -1.90391019e-01 9.15646553e-01 5.73459268e-01
-3.36635560e-01 -1.80908144e-01 5.08006573e+00 4.19645190e-01
-9.95899081e-01 -4.35089439e-01 6.17220700e-01 1.24887526e-01
-4.30758625e-01 -3.76593508e-02 -3.47719997e-01 6.82032764e-01
1.31950235e+00 -7.26287246e-01 4.46296185e-01 6.74692571e-01
1.09031820e+00 -6.33156896e-01 -8.69345844e-01 1.67822406e-01
-4.99437094e-01 -5.53346276e-01 -6.33367181e-01 -4.26885076e-02
6.40943229e-01 -4.75337952e-01 -3.34507167e-01 -3.67989168e-02
-1.70585170e-01 -2.56688837e-02 4.22943830e-01 9.47878003e-01
2.19319426e-02 -1.20273352e+00 7.69146442e-01 9.04956162e-01
-1.20877862e+00 -3.90396953e-01 2.60464311e-01 2.25891635e-01
5.16034305e-01 6.58294559e-01 -7.30253458e-01 6.77942693e-01
2.20035806e-01 -2.97387928e-01 -7.74646327e-02 9.22447443e-01
8.96578562e-03 4.49237287e-01 -7.31163800e-01 -5.29813349e-01
-1.34175718e-02 -6.03859425e-01 4.18387830e-01 5.56208909e-01
4.10545647e-01 3.20923746e-01 -1.13854147e-01 7.94048190e-01
6.85032427e-01 2.29105189e-01 -1.24011576e-01 2.76484221e-01
5.99040687e-01 1.32604361e+00 -4.93847311e-01 -2.61878043e-01
3.46455604e-01 7.30668485e-01 -5.30739367e-01 5.83954811e-01
-7.42140889e-01 -5.44691205e-01 5.07403910e-01 2.29271948e-01
-6.64012507e-03 -1.56847879e-01 -1.36675507e-01 -3.52325261e-01
1.13379084e-01 -4.89781192e-03 -2.43018270e-02 -5.29159844e-01
-6.98296070e-01 5.19140251e-02 3.34434777e-01 -9.42766130e-01
-8.44316065e-01 -1.49455398e-01 -9.11569357e-01 1.17061830e+00
-1.33372819e+00 -4.59194690e-01 -1.39087081e-01 3.37108523e-01
1.90456510e-01 3.10326993e-01 8.73666942e-01 2.44668975e-01
-9.35572267e-01 -8.38104170e-04 9.33829010e-01 -6.04009211e-01
7.33371228e-02 -1.15112936e+00 -4.71069515e-01 8.32544863e-01
-1.27471113e+00 -1.66740865e-01 1.15062106e+00 -3.58337492e-01
-1.59599793e+00 -6.87716663e-01 8.09663951e-01 5.87633431e-01
5.90845942e-01 -1.92242250e-01 -9.39395845e-01 5.36681637e-02
4.13096339e-01 -4.31706488e-01 2.53147721e-01 -5.96171975e-01
6.64536595e-01 -6.39271140e-01 -1.61192000e+00 2.33137786e-01
-3.03890944e-01 -2.43810877e-01 1.22043438e-01 1.32369131e-01
-5.26153892e-02 -5.67152537e-02 -1.07731962e+00 3.31055790e-01
4.66239840e-01 -4.56558645e-01 1.11680828e-01 2.16383293e-01
-4.27283227e-01 -3.78639936e-01 2.48852745e-01 -1.69155395e+00
-1.44456804e-01 -6.90649092e-01 -1.06393434e-01 1.40866053e+00
2.93792129e-01 -7.70358980e-01 2.82252073e-01 1.53847039e+00
9.05797072e-03 -4.50992137e-01 -1.18411791e+00 -2.53348619e-01
-3.14906120e-01 1.50921956e-01 6.20553493e-01 6.02100432e-01
2.31918961e-01 -2.12239757e-01 -3.42861228e-02 7.71794677e-01
5.30256152e-01 -8.02097190e-03 3.04916799e-01 -8.04543436e-01
-4.25158069e-02 -2.45055541e-01 9.82843265e-02 -5.14139980e-02
5.41148447e-02 -2.43588045e-01 3.18345696e-01 -1.68947458e+00
-4.68840033e-01 -2.34580025e-01 -2.22354591e-01 2.93131828e-01
4.86400425e-02 -4.38711286e-01 1.17038809e-01 -1.23749219e-01
2.74491161e-01 7.49491930e-01 7.42534697e-01 6.80783950e-03
-6.07543528e-01 4.56427485e-01 2.93012500e-01 2.91056812e-01
1.06022096e+00 -2.27196500e-01 -6.51374638e-01 -6.59020692e-02
5.22023719e-03 4.06185269e-01 7.12359091e-03 -1.25660205e+00
1.37739688e-01 -4.25827742e-01 -3.15424055e-02 -5.90344489e-01
-1.94361463e-01 -1.87128055e+00 1.05648327e+00 1.01193511e+00
-2.00927760e-02 1.33952156e-01 1.50612563e-01 4.12984520e-01
1.37800634e-01 -3.24716181e-01 7.42869437e-01 5.65311730e-01
-5.75453162e-01 -2.33055741e-01 -1.04134250e+00 -7.89669096e-01
1.59580660e+00 1.50258597e-02 -9.98255536e-02 -3.95514220e-01
-6.56019211e-01 8.70502830e-01 1.18321240e-01 1.91073090e-01
-2.59960532e-01 -8.83985519e-01 -3.78206372e-01 1.04648927e-02
-1.66605711e-01 -2.72170663e-01 6.57667279e-01 6.01873279e-01
-4.72867846e-01 2.09815085e-01 -1.82635382e-01 -2.91969389e-01
-1.05821598e+00 2.73306310e-01 1.08267570e+00 -2.09446818e-01
-1.05517440e-01 -2.54461855e-01 -4.33587283e-01 6.29223734e-02
-2.82422245e-01 -4.03141916e-01 -3.14058542e-01 4.99314189e-01
1.73726514e-01 5.92923939e-01 3.05229574e-01 -2.01776862e-01
-2.93294758e-01 4.65946168e-01 8.91328990e-01 2.84802198e-01
1.28390038e+00 -3.48684758e-01 -2.09249765e-01 5.02063811e-01
9.55962241e-01 -3.99970174e-01 -1.21449184e+00 2.01578334e-01
-4.39792722e-02 1.07544124e-01 1.49734482e-01 -7.74073184e-01
-1.06851149e+00 6.43535629e-02 8.37101519e-01 8.93072546e-01
1.64064610e+00 -6.85993254e-01 2.03322738e-01 2.26460934e-01
1.46720096e-01 -1.64027655e+00 -7.72459269e-01 1.97043896e-01
6.15738451e-01 -2.10600063e-01 1.77093763e-02 -4.86798659e-02
-2.62543112e-01 1.50206542e+00 4.15194422e-01 1.25932526e-02
7.76922941e-01 3.77335429e-01 -4.05912459e-01 3.52845162e-01
-5.78594387e-01 1.97248310e-01 -9.11819786e-02 1.33400127e-01
7.41749778e-02 4.84649748e-01 -8.76569033e-01 2.39252076e-01
5.51434696e-01 1.02607116e-01 5.53013504e-01 6.54357135e-01
-6.56071007e-01 -8.97333145e-01 -5.90670288e-01 -2.86080595e-02
4.91728038e-02 6.10105217e-01 1.52644813e-01 8.33554626e-01
3.42334926e-01 9.60262358e-01 3.07544947e-01 2.00665921e-01
9.00647700e-01 3.76123995e-01 -2.49292895e-01 5.77213727e-02
-5.82777917e-01 1.08329713e-01 2.69094855e-01 -3.38851243e-01
-2.36496657e-01 -4.83994961e-01 -1.45290697e+00 -8.74941871e-02
-2.62981057e-01 5.54491878e-01 1.21036685e+00 9.97752488e-01
2.24938020e-01 5.58292449e-01 1.47082520e+00 -6.99227750e-01
-5.82990229e-01 -7.48607278e-01 -9.07581866e-01 -3.97389680e-02
1.03975408e-01 -2.98359543e-01 -5.74149549e-01 -3.87252986e-01] | [5.716127872467041, 2.5003879070281982] |
2e5badb6-d359-4b32-add2-74bb8bcac066 | cascader-cross-modal-cascading-for-knowledge | 2205.08012 | null | https://arxiv.org/abs/2205.08012v2 | https://arxiv.org/pdf/2205.08012v2.pdf | CascadER: Cross-Modal Cascading for Knowledge Graph Link Prediction | Knowledge graph (KG) link prediction is a fundamental task in artificial intelligence, with applications in natural language processing, information retrieval, and biomedicine. Recently, promising results have been achieved by leveraging cross-modal information in KGs, using ensembles that combine knowledge graph embeddings (KGEs) and contextual language models (LMs). However, existing ensembles are either (1) not consistently effective in terms of ranking accuracy gains or (2) impractically inefficient on larger datasets due to the combinatorial explosion problem of pairwise ranking with deep language models. In this paper, we propose a novel tiered ranking architecture CascadER to maintain the ranking accuracy of full ensembling while improving efficiency considerably. CascadER uses LMs to rerank the outputs of more efficient base KGEs, relying on an adaptive subset selection scheme aimed at invoking the LMs minimally while maximizing accuracy gain over the KGE. Extensive experiments demonstrate that CascadER improves MRR by up to 9 points over KGE baselines, setting new state-of-the-art performance on four benchmarks while improving efficiency by one or more orders of magnitude over competitive cross-modal baselines. Our empirical analyses reveal that diversity of models across modalities and preservation of individual models' confidence signals help explain the effectiveness of CascadER, and suggest promising directions for cross-modal cascaded architectures. Code and pretrained models are available at https://github.com/tsafavi/cascader. | ['Tom Hope', 'Doug Downey', 'Tara Safavi'] | 2022-05-16 | null | null | null | null | ['knowledge-graph-embeddings', 'knowledge-graph-embeddings'] | ['graphs', 'methodology'] | [ 2.77166795e-02 -1.01385675e-02 -4.96582389e-01 -2.38481715e-01
-8.05771232e-01 -5.41444600e-01 5.92512727e-01 5.83622754e-01
-5.11531115e-01 4.98391896e-01 5.40110528e-01 -3.18804711e-01
-5.38609743e-01 -7.91358292e-01 -7.56275475e-01 -3.88579339e-01
-1.86865613e-01 5.90751052e-01 2.03191549e-01 -2.51070291e-01
-4.37938049e-02 1.63350299e-01 -1.16862822e+00 4.60432589e-01
7.95340419e-01 8.17462504e-01 -1.93858027e-01 7.18111515e-01
-1.35394990e-01 8.45777869e-01 -2.28099778e-01 -8.29644382e-01
2.28466913e-02 -1.55983999e-01 -8.03805351e-01 -6.45118535e-01
6.16853058e-01 -1.99334696e-01 -6.03057206e-01 8.74295056e-01
7.01832771e-01 -2.93798093e-02 5.50171614e-01 -1.10979676e+00
-9.61973190e-01 8.18304002e-01 -5.25096774e-01 1.31873950e-01
1.10366561e-01 -1.53515339e-01 1.60435975e+00 -9.28757310e-01
4.80037838e-01 1.03712547e+00 7.52664685e-01 4.80619282e-01
-1.44223237e+00 -5.99235833e-01 1.50819778e-01 4.19126987e-01
-1.54497111e+00 -4.20204967e-01 6.72177672e-01 -2.86531270e-01
1.06744301e+00 2.35808492e-01 6.64162159e-01 9.53019798e-01
1.37868121e-01 8.20710778e-01 8.91287804e-01 -1.75064608e-01
-5.74601181e-02 -3.92994359e-02 2.68103331e-01 1.08030415e+00
5.02159595e-01 -1.42792806e-01 -9.64974344e-01 -4.61984366e-01
5.20741940e-01 -7.34829828e-02 -2.62496799e-01 -5.84599257e-01
-1.35935712e+00 7.64447927e-01 7.03817070e-01 1.94988310e-01
-2.66053319e-01 4.12737966e-01 4.52326506e-01 2.66402870e-01
5.72123349e-01 6.00391209e-01 -6.94133282e-01 1.17431983e-01
-7.93108642e-01 -1.24785993e-02 5.75391471e-01 6.95681870e-01
4.92625445e-01 -3.83544087e-01 -3.42915297e-01 1.04798317e+00
2.74223715e-01 1.87636659e-01 3.14879447e-01 -6.25655413e-01
4.43094254e-01 9.29485500e-01 -3.11793625e-01 -1.05807149e+00
-5.74852288e-01 -9.24165606e-01 -9.59249258e-01 -2.81031817e-01
2.02161267e-01 2.12186232e-01 -7.86342323e-01 1.88571239e+00
2.55790412e-01 9.56367329e-02 8.80010147e-03 7.36060679e-01
9.93318021e-01 3.53759557e-01 2.95539260e-01 1.43654853e-01
1.37889576e+00 -9.07722175e-01 -2.99652219e-01 -5.30732796e-02
8.34414721e-01 -4.51750249e-01 1.02892995e+00 3.38015795e-01
-9.10321414e-01 -2.82055974e-01 -9.63990092e-01 -3.39876771e-01
-4.84052896e-01 2.14989409e-01 8.62481356e-01 3.39184970e-01
-1.41115272e+00 5.72783411e-01 -8.58214796e-01 -2.86620617e-01
5.96928239e-01 5.73057532e-01 -5.11424422e-01 -2.04135254e-01
-1.42210829e+00 7.29697585e-01 3.79190832e-01 1.03864625e-01
-5.73019564e-01 -1.02351320e+00 -7.09225893e-01 1.07233785e-01
2.66378880e-01 -1.14487016e+00 8.22986066e-01 -5.15817463e-01
-1.23953569e+00 6.75096154e-01 -7.47174351e-03 -5.04399657e-01
5.15378058e-01 -4.11850244e-01 -5.20714104e-01 1.25749871e-01
-1.28384039e-01 7.08320320e-01 3.44633043e-01 -8.49869490e-01
-3.07626367e-01 -4.81503189e-01 1.34364575e-01 4.33241606e-01
-8.49523485e-01 -3.68930101e-01 -7.59643674e-01 -5.56676984e-01
-1.95971001e-02 -8.34898889e-01 -1.59303695e-01 -2.84343332e-01
-4.31883574e-01 -3.87935311e-01 2.96428919e-01 -7.08605230e-01
1.53170729e+00 -1.89271295e+00 3.38033736e-01 2.54999548e-01
6.57207251e-01 3.16007107e-01 -3.45967889e-01 5.29938281e-01
1.91465959e-01 2.43962631e-01 2.86180712e-02 -1.75233036e-01
-1.62257135e-01 2.04608357e-03 -6.39781579e-02 2.72432506e-01
1.85386255e-01 1.44241416e+00 -1.10403156e+00 -4.66482878e-01
3.55885401e-02 5.95134318e-01 -6.56465471e-01 -1.63284406e-01
-1.28030241e-01 1.85548246e-01 -3.78050685e-01 6.66807532e-01
2.09666654e-01 -7.91781306e-01 6.38006032e-01 -6.80005729e-01
4.45146650e-01 3.91124487e-01 -7.51220584e-01 1.67076385e+00
-3.72951537e-01 3.67017716e-01 -4.09293473e-01 -8.75940621e-01
5.77495575e-01 2.24132463e-03 5.18874466e-01 -5.59569180e-01
-1.06843889e-01 2.45331958e-01 1.08804293e-01 -3.38095240e-02
5.17793834e-01 6.71571344e-02 1.38848554e-02 1.46796033e-01
1.99340120e-01 4.06530410e-01 1.65296003e-01 7.56262422e-01
1.39338422e+00 -1.67637587e-01 1.85643196e-01 -1.31608441e-01
3.17539006e-01 -9.47695524e-02 3.97342056e-01 7.45664954e-01
3.33663262e-02 3.86612773e-01 5.41729450e-01 -3.05579156e-01
-7.43030608e-01 -1.08563972e+00 6.83023632e-02 1.39598739e+00
1.62385911e-01 -7.13922381e-01 -2.22650304e-01 -8.68196547e-01
2.99761236e-01 4.70708579e-01 -7.19585836e-01 -5.05468190e-01
-3.25706959e-01 -9.56067264e-01 9.67794538e-01 7.93152809e-01
3.45235020e-01 -5.38600445e-01 7.44233187e-03 2.48737216e-01
-1.14047468e-01 -1.09851098e+00 -4.32724118e-01 2.22793370e-02
-9.77847576e-01 -1.16729593e+00 -6.14299774e-01 -5.62269449e-01
7.65019178e-01 1.47538424e-01 1.36649573e+00 1.78824097e-01
-1.96194395e-01 5.21242917e-01 -3.59546244e-01 -6.68860078e-02
-1.90189853e-01 3.75055730e-01 1.79514676e-01 -2.08905086e-01
2.00640783e-01 -4.02062595e-01 -9.17871416e-01 1.10777877e-01
-8.02338302e-01 1.63528979e-01 9.57498968e-01 1.03389072e+00
6.49951577e-01 -6.18469603e-02 8.20826888e-01 -1.09811878e+00
6.46851540e-01 -4.52685058e-01 -3.67383510e-01 6.50294363e-01
-1.09269273e+00 4.42613780e-01 5.08459866e-01 -2.19193429e-01
-7.39381433e-01 -2.91218638e-01 -2.86945459e-02 -4.57881123e-01
4.91317630e-01 8.31445456e-01 8.18495080e-02 -1.34999052e-01
5.19487023e-01 1.50137290e-01 -9.17250738e-02 -3.65743309e-01
6.78253174e-01 4.26446706e-01 3.65754694e-01 -6.05264485e-01
5.28666258e-01 2.77173996e-01 -5.74168041e-02 -3.57064128e-01
-8.93052340e-01 -5.84965348e-01 -3.58260095e-01 -1.17291346e-01
5.27127504e-01 -1.20648885e+00 -6.19688392e-01 3.42079282e-01
-7.60273337e-01 -2.60263443e-01 7.26758093e-02 3.95430982e-01
-7.42076561e-02 2.51184374e-01 -8.97156835e-01 -3.56083244e-01
-6.48812711e-01 -7.80865312e-01 9.35640812e-01 -6.55792654e-03
-1.26988053e-01 -1.14570367e+00 1.02438428e-01 7.20074058e-01
4.00948495e-01 -1.69642195e-02 1.29145002e+00 -8.14355791e-01
-5.21343946e-01 -2.04886839e-01 -2.54413724e-01 2.69447148e-01
4.44450788e-02 -1.04617283e-01 -7.28309274e-01 -3.94634336e-01
-1.08171892e+00 -4.30502445e-01 1.31765532e+00 2.11319566e-01
1.15102077e+00 -2.00623676e-01 -5.54924011e-01 4.87883270e-01
1.37169695e+00 -3.51998061e-01 3.17844123e-01 8.94902050e-02
9.41184998e-01 3.07554990e-01 2.00830564e-01 5.48285507e-02
7.52042115e-01 6.34859681e-01 2.31642291e-01 -1.33295944e-02
-2.95143783e-01 -5.97705185e-01 4.11119998e-01 1.08284760e+00
-4.69737612e-02 -2.82214373e-01 -1.01038861e+00 7.02582896e-01
-1.89079893e+00 -6.83884501e-01 6.65545091e-03 2.26420832e+00
9.37438965e-01 1.86386600e-01 3.65962982e-02 -1.60565630e-01
5.86338103e-01 1.04534559e-01 -7.23172963e-01 -1.94883794e-01
-1.55697614e-01 6.36798069e-02 6.22424126e-01 4.76115912e-01
-9.91335332e-01 9.85943675e-01 5.81595135e+00 8.49392712e-01
-8.69478106e-01 2.27400109e-01 7.35373795e-01 -3.26853752e-01
-5.06047428e-01 -1.83659896e-01 -8.35497081e-01 1.92883864e-01
9.14460182e-01 -2.04493105e-01 3.96572143e-01 5.11686265e-01
-2.93946385e-01 2.52993822e-01 -1.09466374e+00 8.45648706e-01
8.02124441e-02 -1.58905149e+00 3.43600065e-01 1.94027886e-01
8.28337967e-01 5.39881527e-01 1.34146005e-01 5.25261521e-01
5.64512312e-01 -8.96701157e-01 3.10502082e-01 4.38230157e-01
8.44391763e-01 -6.43600464e-01 7.58835852e-01 1.20198265e-01
-1.14870453e+00 9.26495809e-03 -2.35183001e-01 2.13627264e-01
-2.15732217e-01 9.29487109e-01 -9.71270680e-01 8.28563452e-01
6.39314294e-01 6.83645189e-01 -7.71535635e-01 8.46754968e-01
-2.95731574e-01 6.78512633e-01 -3.27232897e-01 -1.41115457e-01
4.74754535e-02 9.94522944e-02 3.75068694e-01 1.31230450e+00
4.30960506e-02 5.71086258e-02 1.54232206e-02 4.36931580e-01
-6.72414660e-01 2.71963298e-01 -5.04935384e-01 -3.74511868e-01
5.49384594e-01 1.38324642e+00 -5.63958347e-01 -3.34924698e-01
-3.99730235e-01 8.65305483e-01 7.63043582e-01 2.40795806e-01
-7.47142255e-01 -2.35968158e-01 5.65160036e-01 6.67858347e-02
1.79094911e-01 -1.17478989e-01 -1.48968473e-01 -1.12978649e+00
-2.19143648e-02 -7.20020175e-01 8.33440900e-01 -3.81513655e-01
-1.60286748e+00 4.40824687e-01 -1.79721743e-01 -9.48879361e-01
1.13187104e-01 -6.69602215e-01 -1.13246448e-01 4.72988337e-01
-1.53557193e+00 -1.41143405e+00 -1.24515861e-01 4.13549095e-01
4.94809411e-02 -7.59753436e-02 8.27987373e-01 5.14969528e-01
-5.39813042e-01 9.91732597e-01 3.50737363e-01 2.56410778e-01
8.36428583e-01 -1.32198095e+00 2.28482038e-01 5.27654707e-01
5.27172089e-01 8.94783795e-01 1.89118028e-01 -6.73816144e-01
-1.50416660e+00 -1.27643871e+00 8.98428500e-01 -6.97124243e-01
8.08152676e-01 -3.34527493e-01 -8.52683604e-01 5.66948950e-01
-1.01431556e-01 2.55852669e-01 1.06027389e+00 7.78132796e-01
-8.57419789e-01 -3.86342406e-01 -7.72168100e-01 6.03540242e-01
1.30072105e+00 -6.20822370e-01 -2.24592224e-01 2.32498690e-01
6.41418219e-01 -2.79752433e-01 -1.18993175e+00 8.33040237e-01
8.29481840e-01 -7.09264040e-01 1.05841696e+00 -9.86943066e-01
5.06905437e-01 -3.46802711e-01 -9.90982056e-02 -1.27819824e+00
-6.00510001e-01 -1.80018753e-01 -5.94628334e-01 1.07433569e+00
6.22545719e-01 -7.18546927e-01 7.85789251e-01 4.04084414e-01
1.23194093e-02 -1.27612376e+00 -7.20951974e-01 -6.58891201e-01
-1.56985652e-02 -3.24352056e-01 2.75373548e-01 1.10240424e+00
3.70154940e-02 6.25782430e-01 -2.72320539e-01 2.84886241e-01
5.54143429e-01 1.68032184e-01 5.07785678e-01 -1.18034554e+00
-5.71079731e-01 -6.25162542e-01 -5.50963998e-01 -8.70488107e-01
1.19395174e-01 -1.38705206e+00 -2.87553936e-01 -1.94245040e+00
5.40659368e-01 -4.64178920e-01 -1.02219856e+00 8.81723821e-01
-5.25740504e-01 4.74593431e-01 7.96422437e-02 2.52105206e-01
-1.15119064e+00 5.66629946e-01 8.83917630e-01 -3.11492354e-01
-1.07960664e-01 -4.20160621e-01 -1.07970691e+00 4.56584662e-01
7.70816147e-01 -2.14172080e-01 -6.26075089e-01 -7.61299849e-01
7.22590506e-01 -2.13553786e-01 5.01186669e-01 -7.34042585e-01
3.61931503e-01 3.04917574e-01 5.82945943e-01 -2.56420761e-01
2.40824491e-01 -3.76882732e-01 2.46286079e-01 4.64622259e-01
-5.88915944e-01 5.38025284e-03 3.06060523e-01 8.79103184e-01
-1.04226634e-01 4.15101171e-01 4.49788898e-01 2.80844212e-01
-7.21432686e-01 4.73368168e-01 1.95169985e-01 6.15070648e-02
6.38297617e-01 2.60890126e-01 -6.04309380e-01 -2.04514369e-01
-5.61275005e-01 4.64985430e-01 2.94416338e-01 5.72690785e-01
5.42535603e-01 -1.32233667e+00 -7.47914016e-01 -6.94242641e-02
5.31046331e-01 -2.14647457e-01 3.11344713e-01 1.02124131e+00
-3.73151928e-01 4.53131914e-01 2.39903346e-01 -3.61200035e-01
-1.28402138e+00 2.58349985e-01 1.84207857e-01 -8.00685883e-01
-3.21711481e-01 1.14789736e+00 1.02686107e-01 -5.84274530e-01
5.68909124e-02 -2.39986442e-02 -6.05288185e-02 -7.72306323e-02
3.09183359e-01 3.89409930e-01 3.92555028e-01 -3.18672091e-01
-7.28606343e-01 4.08813864e-01 -4.65559125e-01 2.15185270e-01
1.23346865e+00 1.96900547e-01 -1.93428934e-01 1.92432791e-01
1.20161021e+00 1.27820089e-01 -7.37265527e-01 -5.19174039e-01
9.06485617e-02 -1.48648739e-01 2.68719912e-01 -1.23384118e+00
-1.03982937e+00 6.00783229e-01 3.60380143e-01 -1.70301139e-01
1.11686432e+00 2.80027837e-01 7.10426390e-01 4.94294047e-01
4.68499035e-01 -7.85755575e-01 6.17758110e-02 2.75520205e-01
6.91210389e-01 -1.27213788e+00 1.61869720e-01 -2.21382797e-01
-6.09255314e-01 7.66846359e-01 5.59067070e-01 1.21238843e-01
7.09564447e-01 -6.20943457e-02 -1.39634043e-01 -4.11695242e-01
-1.04860163e+00 -1.76795721e-01 6.72630191e-01 9.05280337e-02
6.69654667e-01 3.35502326e-01 -4.65238482e-01 6.84892714e-01
2.48909798e-02 -1.33031392e-02 -1.49987653e-01 5.62848985e-01
-1.23299666e-01 -1.26637900e+00 2.04649746e-01 9.66352820e-01
-4.67756987e-01 -4.44301486e-01 -5.40790737e-01 5.71140587e-01
-3.79479714e-02 9.04183745e-01 -2.41480559e-01 -7.60825038e-01
2.52175361e-01 1.08172037e-01 5.13247430e-01 -5.06165206e-01
-4.97636765e-01 -1.85826167e-01 2.60803550e-01 -4.99084800e-01
-2.26645291e-01 -4.40592915e-01 -1.12900257e+00 -3.17906946e-01
-4.07033890e-01 1.50328308e-01 3.72290581e-01 6.14319503e-01
9.22008634e-01 5.86711943e-01 2.70771712e-01 -1.40053734e-01
-5.79200566e-01 -8.11617434e-01 -3.77834350e-01 3.72124225e-01
-4.14851774e-03 -5.55886269e-01 -3.36161628e-02 -2.26919279e-01] | [8.683392524719238, 7.785478115081787] |
1f274158-869b-4a3a-bfb6-cd2a10c4a6cd | w-transformers-a-wavelet-based-transformer | 2209.03945 | null | https://arxiv.org/abs/2209.03945v1 | https://arxiv.org/pdf/2209.03945v1.pdf | W-Transformers : A Wavelet-based Transformer Framework for Univariate Time Series Forecasting | Deep learning utilizing transformers has recently achieved a lot of success in many vital areas such as natural language processing, computer vision, anomaly detection, and recommendation systems, among many others. Among several merits of transformers, the ability to capture long-range temporal dependencies and interactions is desirable for time series forecasting, leading to its progress in various time series applications. In this paper, we build a transformer model for non-stationary time series. The problem is challenging yet crucially important. We present a novel framework for univariate time series representation learning based on the wavelet-based transformer encoder architecture and call it W-Transformer. The proposed W-Transformers utilize a maximal overlap discrete wavelet transformation (MODWT) to the time series data and build local transformers on the decomposed datasets to vividly capture the nonstationarity and long-range nonlinear dependencies in the time series. Evaluating our framework on several publicly available benchmark time series datasets from various domains and with diverse characteristics, we demonstrate that it performs, on average, significantly better than the baseline forecasters for short-term and long-term forecasting, even for datasets that consist of only a few hundred training samples. | ['Abdenour Hadid', 'Tanujit Chakraborty', 'Lena Sasal'] | 2022-09-08 | null | null | null | null | ['univariate-time-series-forecasting'] | ['time-series'] | [-1.03431419e-01 -7.50828326e-01 -9.12187770e-02 -3.57968420e-01
-5.79080522e-01 -5.08311212e-01 6.03580356e-01 -3.49805295e-03
1.51045382e-01 3.50883186e-01 4.41744566e-01 -5.27219355e-01
-3.08523029e-01 -7.53111362e-01 -5.34968376e-01 -7.00397015e-01
-6.32861853e-01 1.58145502e-01 -3.98646183e-02 -3.75086457e-01
-9.96206552e-02 3.31884712e-01 -1.48267841e+00 2.37443745e-01
7.15924799e-01 1.50446403e+00 -1.78318620e-01 2.25534275e-01
1.03676347e-02 1.00135088e+00 -3.17959785e-01 -1.60089612e-01
3.32758911e-02 -1.79218695e-01 -3.40346575e-01 -1.25409305e-01
-8.27888325e-02 -1.86587155e-01 -5.97762287e-01 5.71974158e-01
2.86394000e-01 1.80717170e-01 3.66187543e-01 -1.28417683e+00
-6.17524564e-01 6.54073298e-01 -5.20636737e-01 7.11393833e-01
-4.30750735e-02 -3.79500836e-02 1.16048348e+00 -8.74943495e-01
5.72692938e-02 1.01321781e+00 1.02768695e+00 -7.45526850e-02
-1.18455482e+00 -9.26131248e-01 1.45302445e-01 7.48404086e-01
-9.07000840e-01 -3.66591245e-01 1.32818317e+00 -3.90825629e-01
1.32049882e+00 4.23434116e-02 7.01207519e-01 1.39434767e+00
7.25644708e-01 7.82728672e-01 9.49446797e-01 -5.06037325e-02
4.84556258e-02 -6.38724685e-01 1.34495556e-01 2.80458331e-01
-5.61899781e-01 2.38367468e-01 -5.89257002e-01 -2.24263594e-01
5.89041293e-01 5.71820617e-01 -1.97928220e-01 -3.83709408e-02
-1.28721595e+00 7.89847195e-01 2.23725140e-01 6.54272437e-01
-7.80690312e-01 5.74493408e-02 9.46517408e-01 8.39276254e-01
9.81139541e-01 9.45725366e-02 -6.81586802e-01 -4.80879009e-01
-8.91091406e-01 -5.33804633e-02 4.20994431e-01 4.44493383e-01
9.95335728e-02 6.90075934e-01 -6.69620335e-02 8.44966114e-01
-2.04196408e-01 5.98943532e-01 9.71129417e-01 -3.96844655e-01
5.01158834e-01 3.08469683e-01 -7.51814991e-02 -1.21512532e+00
-3.28278750e-01 -9.32759583e-01 -1.45247722e+00 -3.48571122e-01
1.16832852e-01 1.76744312e-01 -6.88410878e-01 1.77371800e+00
2.89083682e-02 8.65425825e-01 -8.75652954e-02 4.78093714e-01
4.97628063e-01 1.09913611e+00 -6.67817369e-02 -4.94838923e-01
1.15191615e+00 -4.74021584e-01 -9.75581169e-01 1.62529722e-01
1.57068789e-01 -7.65553296e-01 9.48254168e-01 6.15547001e-01
-8.80541205e-01 -7.48019814e-01 -7.83740878e-01 1.01870961e-01
-2.18484893e-01 4.42300364e-02 5.00596464e-01 -8.41963887e-02
-8.27852130e-01 8.29010189e-01 -1.30787647e+00 -6.27477765e-02
2.03367203e-01 1.01104096e-01 -1.73198506e-01 5.76936528e-02
-1.55436456e+00 7.71787226e-01 -5.25381081e-02 1.66491166e-01
-8.74407172e-01 -1.00854492e+00 -8.51172209e-01 3.27280551e-01
-1.34220058e-02 -1.91779658e-01 1.17332208e+00 -6.56478941e-01
-1.31782329e+00 2.70561039e-01 -3.19347680e-01 -8.93966556e-01
2.58982837e-01 -3.28798443e-01 -1.00584090e+00 -1.94189832e-01
-6.46413937e-02 -3.59399676e-01 1.16373861e+00 -3.40310931e-01
-4.65363473e-01 -4.17216539e-01 -3.81277859e-01 -2.88870364e-01
-7.81038642e-01 1.08710632e-01 2.37724975e-01 -1.21070385e+00
5.59552945e-02 -6.41426086e-01 -1.58689748e-02 -3.25724155e-01
-1.09134108e-01 -6.14917338e-01 1.42192984e+00 -9.11733925e-01
1.33890200e+00 -2.35219097e+00 1.44702300e-01 4.36850488e-02
2.60001570e-01 1.58478945e-01 -2.74927557e-01 9.07168388e-01
-5.81679523e-01 -4.32151526e-01 -9.33496468e-03 -3.47201616e-01
2.59137843e-02 2.61926174e-01 -1.46462750e+00 5.72194636e-01
1.94679096e-01 8.94808412e-01 -8.00145030e-01 1.69956833e-01
5.31236291e-01 8.51692796e-01 -4.99397703e-02 1.43218398e-01
-3.88989821e-02 5.85957527e-01 -3.46356124e-01 4.33960438e-01
3.87989938e-01 -3.65075499e-01 -1.42569706e-01 -4.60201979e-01
-2.36177444e-01 4.97130543e-01 -4.29719567e-01 1.47049797e+00
-6.21859491e-01 8.32555175e-01 -5.18859506e-01 -1.60143530e+00
1.12412560e+00 7.88459063e-01 1.25852990e+00 -1.08868515e+00
-7.70831034e-02 2.55284905e-01 -6.98957592e-02 -4.88958687e-01
5.10420501e-02 -3.41882139e-01 1.18060216e-01 4.81038570e-01
1.42992884e-01 1.94476157e-01 -1.35934711e-01 -2.76223749e-01
1.32581770e+00 -8.20067227e-02 2.41185516e-01 -7.13230968e-02
4.28333700e-01 -2.55969465e-01 6.50586724e-01 3.95063497e-02
-1.10488003e-02 1.96723238e-01 1.99373394e-01 -9.94014978e-01
-9.73076344e-01 -9.86626506e-01 -1.29291281e-01 1.12293088e+00
-4.54978734e-01 -4.96473759e-01 6.22306354e-02 -1.36085317e-01
-1.12722302e-02 6.87295854e-01 -6.56941831e-01 -4.75706816e-01
-7.74694502e-01 -5.89364946e-01 4.97380674e-01 7.95053482e-01
3.29272538e-01 -1.05527139e+00 -6.07014239e-01 6.22725666e-01
-4.47407275e-01 -1.14920843e+00 -4.54676181e-01 3.81670833e-01
-1.08982193e+00 -7.80328810e-01 -7.89152086e-01 -5.98649442e-01
-5.57975983e-03 3.11467320e-01 1.13252330e+00 -3.68046463e-01
-1.12186551e-01 3.69179636e-01 -5.50792515e-01 -5.08821189e-01
2.29729321e-02 -2.17290908e-01 2.71245956e-01 2.55635262e-01
4.23502982e-01 -1.30807137e+00 -4.87109661e-01 2.62404531e-01
-9.09694254e-01 -2.53446609e-01 1.91635862e-01 1.02177572e+00
6.05551302e-01 4.96499300e-01 8.62219334e-01 -2.11962357e-01
7.06695020e-01 -8.01155865e-01 -4.96505678e-01 2.12157905e-01
-3.59281212e-01 -1.21155649e-01 1.34278858e+00 -7.27934837e-01
-5.96088588e-01 -5.21369159e-01 -9.24089625e-02 -9.83649611e-01
3.72628659e-01 1.07652569e+00 4.14207578e-01 3.97546440e-01
2.77966112e-01 8.29466701e-01 -1.10999823e-01 -6.73432589e-01
-6.73727170e-02 2.62733877e-01 5.95823407e-01 -5.29004037e-01
8.66881192e-01 5.27490258e-01 -4.62787366e-03 -7.30906129e-01
-8.87715638e-01 -4.62842405e-01 -3.71009022e-01 -7.40655884e-02
2.96015471e-01 -9.87117708e-01 -4.90640193e-01 5.77141881e-01
-1.06609428e+00 -2.26414874e-01 -3.26795936e-01 5.36259115e-01
-4.17868972e-01 1.38116077e-01 -7.56270170e-01 -6.06558919e-01
-6.27655387e-01 -7.47763634e-01 9.16384995e-01 -4.65353131e-01
-1.41644716e-01 -1.33196068e+00 2.74013996e-01 -2.19544813e-01
9.86031950e-01 5.44499099e-01 1.17638958e+00 -5.92702508e-01
-1.83759797e-02 -3.48480642e-01 1.77293986e-01 4.59704220e-01
5.23331046e-01 -7.76775256e-02 -8.28700125e-01 -5.19063473e-01
5.20398855e-01 -1.35864139e-01 7.76477337e-01 5.60781837e-01
1.69507802e+00 -2.70570606e-01 -1.05349451e-01 6.81044996e-01
1.02110660e+00 4.09489512e-01 4.53637868e-01 -4.72249575e-02
5.59952438e-01 4.11225289e-01 3.28153402e-01 7.77002990e-01
4.64062512e-01 6.39770567e-01 4.11458880e-01 6.51762187e-02
2.74811953e-01 -2.93164968e-01 5.92808843e-01 1.55262959e+00
-2.74788707e-01 -2.01165415e-02 -1.00342631e+00 6.52588844e-01
-1.77183056e+00 -1.38795435e+00 9.03702602e-02 2.00791621e+00
5.70915282e-01 8.32409412e-02 1.62667021e-01 6.69469178e-01
2.90091813e-01 4.60876226e-01 -8.92164886e-01 2.29235482e-03
-9.01168138e-02 5.63907146e-01 -7.77019635e-02 -1.06846675e-01
-1.17452836e+00 4.12039310e-01 5.84528255e+00 6.50777161e-01
-1.86442208e+00 1.01645857e-01 5.31865478e-01 -8.72553047e-03
-3.64245147e-01 -4.02668834e-01 -6.22584857e-02 6.38093948e-01
1.21301770e+00 -6.69726968e-01 5.27490616e-01 5.72321355e-01
5.28876960e-01 8.59539390e-01 -1.10367167e+00 1.20816779e+00
-2.25026995e-01 -1.17027748e+00 -5.71766496e-02 -1.91301048e-01
4.31285650e-01 3.11565876e-01 6.16258323e-01 5.38333654e-01
7.29306117e-02 -1.15911913e+00 6.44239604e-01 3.53561521e-01
7.99633265e-01 -7.96373665e-01 6.40583575e-01 4.74527776e-01
-1.76450908e+00 -3.79368395e-01 -2.31005147e-01 -2.47767493e-01
1.65491983e-01 1.08456707e+00 -1.95998624e-01 7.34058917e-01
9.49063003e-01 1.62444615e+00 -2.01099459e-02 7.14942336e-01
2.59505272e-01 1.19501686e+00 -3.20082903e-01 4.76332694e-01
1.74285546e-01 -1.74345598e-01 4.75789040e-01 7.38664031e-01
8.35744679e-01 2.15269402e-01 1.41082123e-01 4.07494634e-01
8.56835544e-02 -1.80296879e-02 -5.54368675e-01 -3.29065949e-01
3.78269076e-01 8.28947902e-01 -3.48986387e-01 -2.96519995e-01
-6.39961302e-01 6.71040416e-01 1.27584696e-01 4.81194645e-01
-1.08730972e+00 -1.25943899e-01 7.22033978e-01 -2.95842644e-02
5.15448689e-01 -4.66077119e-01 1.21447042e-01 -1.49885643e+00
3.54978681e-01 -1.12045765e+00 5.93633294e-01 -5.79794824e-01
-1.66036665e+00 8.26151252e-01 -8.78972188e-02 -1.82922900e+00
-4.08555537e-01 -3.93722177e-01 -8.83004963e-01 7.46439278e-01
-1.66187906e+00 -1.24393547e+00 -1.36968300e-01 1.07891023e+00
7.23833621e-01 -3.09773266e-01 9.34505522e-01 6.44174039e-01
-4.01815563e-01 3.42039466e-01 3.82211357e-01 1.46974817e-01
4.69721705e-01 -9.67585504e-01 5.66446245e-01 7.69690156e-01
3.57892483e-01 5.89320838e-01 6.96858525e-01 -2.49218479e-01
-1.67970395e+00 -1.26199663e+00 8.68704498e-01 -6.86404482e-02
1.07911181e+00 -2.50629216e-01 -1.12976575e+00 8.65864456e-01
2.24780142e-01 3.65814328e-01 7.68925607e-01 1.84647605e-01
-6.83882117e-01 -6.08753741e-01 -6.91923380e-01 1.62147328e-01
8.02178979e-01 -8.27416778e-01 -6.83363557e-01 3.38565767e-01
5.55497527e-01 -1.73264846e-01 -1.03742230e+00 5.63873827e-01
5.98071218e-01 -8.94448698e-01 1.08205795e+00 -6.71301126e-01
6.44952416e-01 4.59098890e-02 -2.88059503e-01 -1.65552127e+00
-5.38288295e-01 -8.00228298e-01 -5.51243007e-01 1.06769776e+00
-2.00817838e-01 -9.68433797e-01 2.26236388e-01 1.07558839e-01
-2.86033362e-01 -8.68932784e-01 -1.19693661e+00 -9.28197205e-01
1.16940670e-01 -7.28587329e-01 8.99946690e-01 1.27861118e+00
3.66358310e-02 1.85612813e-01 -8.19258630e-01 1.12482511e-01
5.32372117e-01 5.84443390e-01 4.12129313e-01 -1.37727225e+00
-2.93446749e-01 -4.35140133e-01 -3.99105906e-01 -9.85033095e-01
3.18874389e-01 -7.36745298e-01 -2.98846692e-01 -1.03491831e+00
-1.71401158e-01 -2.34274924e-01 -9.65064049e-01 5.98648012e-01
1.54375851e-01 -2.01916844e-02 -1.48811460e-01 3.41279030e-01
-1.95724040e-01 1.10125351e+00 9.79393423e-01 -3.16107363e-01
1.12540205e-03 2.02975869e-01 -2.12926134e-01 4.90449458e-01
6.37336791e-01 -2.60020018e-01 -6.82724774e-01 -5.97898781e-01
4.95343059e-02 5.53423345e-01 4.32241470e-01 -9.93979454e-01
1.40334845e-01 -1.34340405e-01 2.29982659e-01 -8.37784171e-01
5.71240604e-01 -1.13153243e+00 3.70949924e-01 5.14156520e-01
-2.97392607e-01 5.72203040e-01 3.45341951e-01 5.76814592e-01
-8.34638536e-01 8.32729578e-01 3.60016555e-01 2.49231547e-01
-7.44149387e-01 6.92419291e-01 -2.27282882e-01 -1.51540712e-01
8.61553550e-01 1.37628525e-01 -1.78005353e-01 -7.86788464e-01
-5.91667533e-01 9.07567888e-02 -2.24991754e-01 6.99751496e-01
7.20154583e-01 -1.65408123e+00 -8.24631453e-01 5.37513018e-01
1.10375695e-01 -4.60579246e-01 4.09217805e-01 1.13148355e+00
2.29002520e-01 4.90171254e-01 -2.75898665e-01 -6.72267079e-01
-1.08470929e+00 6.36330366e-01 2.38103881e-01 -5.11059761e-01
-1.09194458e+00 4.97333705e-01 4.57899608e-02 -5.74808717e-02
2.28662148e-01 -8.79849374e-01 -2.77749956e-01 7.36987367e-02
6.57956541e-01 2.61031866e-01 1.13404326e-01 -6.30558848e-01
-3.77124161e-01 7.38967478e-01 3.98849361e-02 2.53952652e-01
2.02260900e+00 1.04361475e-01 -2.34243974e-01 9.73470807e-01
1.32732654e+00 -1.90149143e-01 -1.02284336e+00 -7.38509893e-01
2.02815514e-02 -2.72579730e-01 9.13424790e-02 -4.00338143e-01
-1.32268977e+00 1.19978654e+00 3.91607463e-01 8.25120330e-01
1.59546590e+00 -3.48669380e-01 1.32317758e+00 3.61357003e-01
5.65593183e-01 -5.10967016e-01 9.24454927e-02 8.53994489e-01
1.09266376e+00 -1.03838015e+00 -2.07656130e-01 1.33522078e-01
-3.07406873e-01 1.27411842e+00 -3.24101299e-02 -3.89546335e-01
1.11596322e+00 2.82673717e-01 5.59416227e-02 -1.06242500e-01
-1.27316713e+00 2.03638181e-01 5.43351293e-01 3.20407689e-01
5.71364164e-01 4.81953621e-02 1.74354360e-01 6.83759987e-01
-2.74467945e-01 9.90091041e-02 -1.04203172e-01 5.57961285e-01
-5.16226813e-02 -1.00311649e+00 -3.00236255e-01 6.63951457e-01
-6.43147588e-01 3.21028568e-02 2.16728151e-01 3.45171541e-01
-3.61075640e-01 9.37237561e-01 2.70099670e-01 -5.48438251e-01
2.36471862e-01 1.40094846e-01 1.33187935e-01 -9.14072841e-02
-4.72717166e-01 1.80299014e-01 -2.52610058e-01 -7.08187282e-01
-5.83242238e-01 -7.43656456e-01 -8.50917935e-01 -5.07641792e-01
1.39090583e-01 1.01479523e-01 3.17699015e-01 1.10784912e+00
4.12730873e-01 7.86475182e-01 1.10261405e+00 -8.11731935e-01
-8.64913344e-01 -1.03399873e+00 -5.18778920e-01 6.42575860e-01
8.37229908e-01 -7.29109943e-01 -3.30167323e-01 1.01085044e-01] | [7.031061172485352, 2.914599657058716] |
a363cbda-41cc-453a-aa0d-31ea91b7d4c8 | multilingual-content-moderation-a-case-study | 2302.09618 | null | https://arxiv.org/abs/2302.09618v1 | https://arxiv.org/pdf/2302.09618v1.pdf | Multilingual Content Moderation: A Case Study on Reddit | Content moderation is the process of flagging content based on pre-defined platform rules. There has been a growing need for AI moderators to safeguard users as well as protect the mental health of human moderators from traumatic content. While prior works have focused on identifying hateful/offensive language, they are not adequate for meeting the challenges of content moderation since 1) moderation decisions are based on violation of rules, which subsumes detection of offensive speech, and 2) such rules often differ across communities which entails an adaptive solution. We propose to study the challenges of content moderation by introducing a multilingual dataset of 1.8 Million Reddit comments spanning 56 subreddits in English, German, Spanish and French. We perform extensive experimental analysis to highlight the underlying challenges and suggest related research problems such as cross-lingual transfer, learning under label noise (human biases), transfer of moderation models, and predicting the violated rule. Our dataset and analysis can help better prepare for the challenges and opportunities of auto moderation. | ['Malihe Alikhani', 'Ajay Divakaran', 'Sabit Hassan', 'Katherine Atwell', 'Karan Sikka', 'Meng Ye'] | 2023-02-19 | null | null | null | null | ['cross-lingual-transfer'] | ['natural-language-processing'] | [ 3.21962714e-01 1.43369228e-01 -2.02849656e-01 -2.84998089e-01
-6.08960152e-01 -7.92528093e-01 7.41551578e-01 2.83597469e-01
-1.27660424e-01 5.65564275e-01 8.19581330e-01 -4.34034467e-01
1.71267893e-02 -1.77566279e-02 -3.30713630e-01 -3.73546362e-01
3.14992443e-02 8.38495642e-02 -1.97411984e-01 -4.95677710e-01
5.48901975e-01 1.33037806e-01 -1.32555580e+00 7.34533131e-01
1.19959342e+00 5.53976357e-01 -4.34314609e-01 4.45599586e-01
1.12875171e-01 1.04736054e+00 -8.40038061e-01 -7.78347731e-01
9.77053940e-02 -7.03302979e-01 -8.66786957e-01 -6.09878898e-02
9.99429166e-01 -2.40447968e-01 -1.60370827e-01 1.35230732e+00
7.16793060e-01 -2.58952558e-01 6.47525370e-01 -1.50345385e+00
-1.16647255e+00 1.15330696e+00 -7.18329430e-01 3.16010267e-01
4.31892753e-01 1.37027934e-01 9.71675932e-01 -6.23831272e-01
8.80345821e-01 1.46327317e+00 7.92461932e-01 8.38257611e-01
-9.73407805e-01 -1.06362605e+00 2.40398943e-01 2.19595075e-01
-1.03835547e+00 -7.20262587e-01 6.50331855e-01 -1.02973354e+00
2.61812329e-01 3.34027171e-01 2.25276574e-01 1.87035668e+00
1.22602917e-01 5.63269377e-01 1.37143981e+00 -1.30193964e-01
5.35805896e-03 5.21310091e-01 3.25327933e-01 4.43730474e-01
2.47502849e-01 -2.35307395e-01 -6.75719917e-01 -6.82987630e-01
-3.65479141e-01 -3.57902467e-01 -3.67159963e-01 2.25918487e-01
-8.76400411e-01 1.17236233e+00 2.14783207e-01 4.81744558e-01
-1.30501166e-01 -3.70218545e-01 1.07476664e+00 4.08136994e-01
6.14268303e-01 7.02287912e-01 -2.62042552e-01 -5.78132086e-02
-9.68786240e-01 3.16578716e-01 7.24689841e-01 7.00269699e-01
2.56079644e-01 -9.73741408e-04 -3.12706560e-01 8.30960512e-01
2.07433775e-01 6.47373497e-01 5.18269002e-01 -8.01205397e-01
5.69143713e-01 2.28092492e-01 4.33495902e-02 -1.42034090e+00
-4.88796741e-01 -2.94106781e-01 -6.75337493e-01 5.18500768e-02
4.12242174e-01 -6.92716002e-01 -4.65716481e-01 1.90692532e+00
3.77303720e-01 -1.21346369e-01 -4.26216871e-01 6.71748221e-01
6.52840257e-01 4.42583293e-01 3.11324835e-01 -3.99498910e-01
1.35638499e+00 -4.64403629e-01 -9.19749975e-01 -1.53679848e-01
7.62477458e-01 -8.95403624e-01 1.12790632e+00 5.99196911e-01
-8.62303972e-01 -1.15279399e-01 -9.43416595e-01 -1.06400602e-01
-5.17780542e-01 -4.56578523e-01 2.73913354e-01 1.05639446e+00
-7.23849714e-01 3.27390581e-01 -1.02776915e-01 -3.64599109e-01
4.19928044e-01 -1.10866062e-01 -3.40316266e-01 1.26770467e-01
-1.64502919e+00 1.11988449e+00 4.89516044e-03 1.12147637e-01
-9.26569879e-01 -9.18189108e-01 -7.18706608e-01 -4.29046512e-01
2.37071618e-01 1.98653294e-03 1.08091652e+00 -1.35351717e+00
-1.11540568e+00 1.02559757e+00 1.59101233e-01 -2.18914196e-01
5.35323441e-01 -3.73315424e-01 -7.28759170e-01 6.88050985e-02
4.05201435e-01 2.35634238e-01 1.01759100e+00 -1.16339409e+00
-5.27316093e-01 -4.18064445e-01 8.63940567e-02 -5.21177873e-02
-8.51491392e-01 8.09429765e-01 6.14992201e-01 -5.59293449e-01
-5.64219832e-01 -1.01215100e+00 3.65854770e-01 -4.65440840e-01
-7.39042699e-01 -1.60976246e-01 1.11888301e+00 -1.27868819e+00
1.55150461e+00 -2.14572501e+00 1.33317620e-01 -6.67379890e-03
5.67296803e-01 4.63469177e-01 -4.85435352e-02 6.37009859e-01
-1.06956564e-01 5.73164046e-01 -6.79851323e-02 -2.08756208e-01
4.74435575e-02 -4.56575900e-01 -7.02058673e-01 7.77213573e-01
-1.05058566e-01 4.48220044e-01 -9.42475140e-01 -3.92502040e-01
-5.52632570e-01 3.29686165e-01 -7.46072471e-01 1.25345141e-01
-3.82823274e-02 4.35114026e-01 8.51260684e-03 7.05470026e-01
6.09954238e-01 5.87993786e-02 3.52155358e-01 8.52402151e-02
-3.74645621e-01 6.08981252e-01 -7.64895618e-01 9.83220458e-01
-5.54309376e-02 8.15441489e-01 5.64735711e-01 -6.29765868e-01
6.82313263e-01 3.06840181e-01 3.70299160e-01 -5.49334824e-01
2.56674767e-01 1.17458485e-01 3.69994819e-01 -8.75549614e-01
6.61898613e-01 -2.65820563e-01 -4.98626113e-01 6.92097723e-01
-2.60519773e-01 3.31448078e-01 -5.79583198e-02 5.76252520e-01
9.70705032e-01 -3.28520298e-01 6.51460066e-02 -3.81675601e-01
2.74006754e-01 -1.34620741e-01 7.42431879e-01 6.79313362e-01
-8.05328846e-01 2.76604295e-01 9.58978713e-01 2.59774625e-02
-9.17890966e-01 -4.67987359e-01 -2.60373414e-01 1.49457777e+00
-1.55966818e-01 -5.88462591e-01 -8.94540846e-01 -9.98239756e-01
-6.29032478e-02 8.20608079e-01 -1.00364506e+00 -3.60586971e-01
-3.27048689e-01 -6.48147523e-01 1.03138638e+00 -5.24459630e-02
2.28625134e-01 -1.06344688e+00 -3.05838734e-01 -9.51557457e-02
-5.44078290e-01 -1.02491295e+00 -6.63235605e-01 -6.24226741e-02
7.91686848e-02 -1.17171323e+00 -1.42158389e-01 -4.32283789e-01
2.49111935e-01 3.22822064e-01 6.78914487e-01 2.37267286e-01
-4.92102057e-02 2.31990784e-01 -6.14165604e-01 -5.30239820e-01
-7.95445263e-01 9.20991376e-02 3.33504587e-01 2.04703972e-01
5.64268172e-01 -4.05727297e-01 -2.70205021e-01 7.36775696e-02
-9.22135830e-01 -1.66004926e-01 5.06622531e-02 5.59256554e-01
-3.52779657e-01 -2.31913686e-01 7.82857656e-01 -1.55928385e+00
9.64272201e-01 -1.15304470e+00 2.09776927e-02 1.74108028e-01
-6.47315145e-01 -5.20535588e-01 4.98829424e-01 -6.71847701e-01
-1.01908612e+00 -6.47769988e-01 1.51102826e-01 1.17096402e-01
-1.47931173e-01 4.82254326e-01 -1.46180630e-01 2.40120322e-01
1.10863411e+00 -4.58107948e-01 -3.35077047e-02 -3.42798948e-01
3.77909452e-01 1.21954668e+00 1.16670720e-01 -5.11577129e-01
9.16580796e-01 3.11868697e-01 -6.68319583e-01 -8.52722943e-01
-1.26664352e+00 -3.90183628e-01 -4.83903110e-01 -4.45897609e-01
9.34256554e-01 -8.27121198e-01 -5.81704974e-01 4.50812846e-01
-1.28726244e+00 -3.53230178e-01 4.57393140e-01 1.10824578e-01
-2.27587596e-02 5.78902960e-01 -9.31403577e-01 -7.19276488e-01
-5.81026733e-01 -7.69005120e-01 6.04322255e-01 -1.77088782e-01
-8.53187621e-01 -8.13737154e-01 3.01727146e-01 8.09750140e-01
4.55047816e-01 6.13431990e-01 1.13836324e+00 -8.88820469e-01
3.45624268e-01 8.22029635e-03 -1.83096267e-02 3.68701696e-01
8.65473300e-02 2.42134169e-01 -9.93235886e-01 -2.14591593e-01
-7.40369689e-03 -9.53569710e-01 3.88681173e-01 -1.24972999e-01
7.27773309e-01 -6.79372966e-01 3.65052447e-02 8.92379060e-02
8.76674235e-01 -2.69709695e-02 5.30946374e-01 2.68380284e-01
9.12194133e-01 1.10642612e+00 3.77932966e-01 5.76175630e-01
3.99894357e-01 3.50017160e-01 3.08385879e-01 2.80782819e-01
2.98174173e-02 -3.38864833e-01 9.29849684e-01 9.50096250e-01
3.34597081e-01 -6.44289479e-02 -9.76914108e-01 4.55781549e-01
-1.76951444e+00 -1.32976282e+00 -4.64105070e-01 1.78022265e+00
1.22458041e+00 -1.02259461e-02 4.58712995e-01 4.47971858e-02
9.74658072e-01 2.50056714e-01 -2.59713948e-01 -9.19323146e-01
-1.44392982e-01 -3.47844183e-01 1.12937801e-01 6.03502274e-01
-1.05616677e+00 9.34239566e-01 6.67838287e+00 3.16160083e-01
-1.36393845e+00 5.62412262e-01 5.87108791e-01 -5.23044579e-02
-4.67550814e-01 -1.08763479e-01 -9.72401261e-01 7.93523312e-01
1.11750674e+00 2.18042471e-02 4.34917063e-01 7.82060921e-01
3.28975469e-01 3.18947047e-01 -9.26485837e-01 6.97200716e-01
7.05167711e-01 -8.28684986e-01 -3.78124893e-01 1.32035121e-01
9.78777945e-01 5.26544005e-02 3.76123428e-01 5.28391242e-01
3.30355912e-01 -9.14860904e-01 1.00338840e+00 4.09111101e-03
5.85861564e-01 -4.44499522e-01 3.55467260e-01 3.87381822e-01
-2.28231460e-01 -2.62028813e-01 -2.35088617e-01 -1.28104195e-01
-1.83916353e-02 4.68525261e-01 -6.62258327e-01 -8.50216448e-02
6.69479012e-01 6.61774933e-01 -7.50815451e-01 6.58139110e-01
-3.99629593e-01 8.84718955e-01 2.58744985e-01 -3.21910344e-02
1.79362193e-01 1.22276589e-01 5.32016814e-01 1.67328703e+00
-7.35166073e-02 -2.11918727e-01 7.74027333e-02 5.24383068e-01
-1.94596156e-01 1.92912936e-01 -7.10286260e-01 -4.21365082e-01
6.02405906e-01 1.34795570e+00 -3.44477326e-01 -2.16677785e-01
-4.09337372e-01 7.54748583e-01 6.05450869e-01 2.32323438e-01
-9.56172347e-01 -2.78225839e-01 7.88173079e-01 3.57859910e-01
-1.04582302e-01 -1.38763323e-01 -5.81505835e-01 -1.19511306e+00
-2.77554274e-01 -1.74976802e+00 4.55166161e-01 -4.99737471e-01
-1.57481599e+00 5.57069957e-01 -7.07273036e-02 -7.28645265e-01
-2.90780198e-02 -4.48824793e-01 -4.13365364e-01 4.24047709e-01
-1.22173989e+00 -1.18284953e+00 -1.99008390e-01 5.15742123e-01
1.53689951e-01 2.20848378e-02 5.85332215e-01 5.80648422e-01
-6.96423888e-01 7.05799818e-01 -2.13865519e-01 1.82468474e-01
1.34423220e+00 -9.62894976e-01 -1.91198558e-01 9.71633494e-01
-3.39702904e-01 6.62442207e-01 9.92807865e-01 -1.02978528e+00
-1.02693021e+00 -9.72954214e-01 1.32998419e+00 -9.54552114e-01
1.30654085e+00 -8.71535361e-01 -8.85043263e-01 7.34737933e-01
6.24328554e-01 -6.46970153e-01 1.18675840e+00 6.10046029e-01
-1.14715815e+00 2.62967438e-01 -1.05837572e+00 8.35993409e-01
1.19670677e+00 -7.14425564e-01 -5.84032297e-01 5.76472521e-01
7.22974598e-01 -2.57289708e-02 -5.94840467e-01 -4.34727129e-03
4.56029356e-01 -8.14524949e-01 2.76733190e-01 -9.46150839e-01
8.19505334e-01 3.48450392e-02 -7.64939561e-02 -1.28021216e+00
-6.82894707e-01 -1.00515771e+00 -8.81411508e-02 1.76593065e+00
4.51917768e-01 -2.95950830e-01 1.43870056e-01 8.59422565e-01
-2.78176516e-02 -2.24962294e-01 -6.00852430e-01 -4.26389039e-01
4.60326940e-01 -3.61055881e-01 2.85754770e-01 1.79637468e+00
4.06253099e-01 6.66225135e-01 -1.07413626e+00 1.00460067e-01
4.18345600e-01 -1.62446305e-01 7.16515839e-01 -1.15782034e+00
-1.57925308e-01 -4.98699039e-01 -1.76180806e-02 -1.83178857e-01
5.96383035e-01 -9.06978011e-01 -3.95958759e-02 -1.01014996e+00
4.82423216e-01 -3.68782729e-01 1.89892750e-03 5.49186110e-01
-3.12066585e-01 4.40973371e-01 2.62946159e-01 3.31532508e-02
-6.83909714e-01 2.81745285e-01 8.53891790e-01 -1.42928988e-01
-5.69827072e-02 -4.82778281e-01 -1.59218836e+00 6.98765516e-01
8.88277829e-01 -6.68775380e-01 -3.23700011e-01 -3.28012407e-01
7.19659150e-01 -4.10999924e-01 4.97759171e-02 -4.32810873e-01
-6.26736432e-02 -4.46798921e-01 -7.70177245e-02 -3.40813063e-02
-2.76896935e-02 -5.80923557e-01 -2.04461798e-01 4.07253683e-01
-6.91239655e-01 1.63167909e-01 -8.37631077e-02 3.93446922e-01
1.82214603e-01 -3.56826484e-01 9.24426138e-01 6.15049638e-02
-1.67918578e-01 1.90133348e-01 -7.54932463e-01 5.91257691e-01
9.06530976e-01 2.84648359e-01 -8.94094944e-01 -6.05659068e-01
-4.83705521e-01 7.46718422e-02 3.03299665e-01 6.91940963e-01
2.58199692e-01 -1.15764511e+00 -1.14258671e+00 -2.23919258e-01
2.86594778e-01 -1.01705694e+00 1.35136977e-01 1.00513446e+00
-1.33538246e-03 4.96049505e-03 -1.94274142e-01 5.96574470e-02
-1.48486030e+00 7.78241634e-01 9.70207602e-02 1.25787795e-01
-3.17958981e-01 6.84090137e-01 1.38120398e-01 -3.34115624e-01
3.34292471e-01 3.94934863e-01 -3.47890705e-01 5.15619218e-01
9.64127362e-01 4.16341394e-01 -1.47220835e-01 -9.65226710e-01
-3.20048392e-01 -1.32774919e-01 -3.90756041e-01 -5.34935184e-02
1.13819349e+00 -2.16594607e-01 -5.18209517e-01 5.87010920e-01
1.18663228e+00 7.26704538e-01 -6.60316586e-01 -2.30852634e-01
2.49291062e-01 -4.31983769e-01 -2.23215640e-01 -1.07173729e+00
-4.50545549e-01 7.52514303e-01 2.36346856e-01 6.09591365e-01
6.30535483e-01 7.07753608e-03 8.84333670e-01 -2.31000558e-01
-4.44756523e-02 -1.49151790e+00 4.37409014e-01 6.80762708e-01
1.22074711e+00 -1.16924143e+00 -1.78889543e-01 -2.39894524e-01
-9.90289211e-01 7.51611590e-01 8.66307616e-01 1.60968304e-01
7.39372194e-01 7.01170042e-02 2.56219774e-01 -4.20725763e-01
-7.23474920e-01 2.03823179e-01 1.97642505e-01 6.79337025e-01
7.84170806e-01 2.67331451e-01 -6.69591427e-01 5.98097205e-01
-4.28299248e-01 -6.83575034e-01 7.16028810e-01 5.83288133e-01
-3.72013003e-01 -9.30367947e-01 -4.68721181e-01 2.69542664e-01
-9.27405536e-01 -2.70460188e-01 -1.21269929e+00 4.55319732e-01
5.12098670e-01 1.30686176e+00 -3.76116157e-01 -6.57696843e-01
3.44105139e-02 2.85314530e-01 9.65426788e-02 -8.44621181e-01
-1.13038337e+00 -2.44240433e-01 6.18587494e-01 -1.47583544e-01
-2.47520670e-01 -1.04473150e+00 -7.29528069e-01 -8.21662128e-01
-2.51848370e-01 1.79866344e-01 6.36298060e-01 1.06414998e+00
4.64385927e-01 5.18786050e-02 7.19038844e-01 -2.54372925e-01
-6.18310213e-01 -1.30109358e+00 -3.85781854e-01 9.28747296e-01
3.78511071e-01 -3.89293909e-01 -5.53085446e-01 4.51626107e-02] | [8.79279899597168, 10.488719940185547] |
f04a6e90-5013-47a9-b737-1da453dfddda | robust-mitosis-detection-using-a-cascade-mask | 2109.01878 | null | https://arxiv.org/abs/2109.01878v2 | https://arxiv.org/pdf/2109.01878v2.pdf | Robust Mitosis Detection Using a Cascade Mask-RCNN Approach With Domain-Specific Residual Cycle-GAN Data Augmentation | For the MIDOG mitosis detection challenge, we created a cascade algorithm consisting of a Mask-RCNN detector, followed by a classification ensemble consisting of ResNet50 and DenseNet201 to refine detected mitotic candidates. The MIDOG training data consists of 200 frames originating from four scanners, three of which are annotated for mitotic instances with centroid annotations. Our main algorithmic choices are as follows: first, to enhance the generalizability of our detector and classification networks, we use a state-of-the-art residual Cycle-GAN to transform each scanner domain to every other scanner domain. During training, we then randomly load, for each image, one of the four domains. In this way, our networks can learn from the fourth non-annotated scanner domain even if we don't have annotations for it. Second, for training the detector network, rather than using centroid-based fixed-size bounding boxes, we create mitosis-specific bounding boxes. We do this by manually annotating a small selection of mitoses, training a Mask-RCNN on this small dataset, and applying it to the rest of the data to obtain full annotations. We trained the follow-up classification ensemble using only the challenge-provided positive and hard-negative examples. On the preliminary test set, the algorithm scores an F1 score of 0.7578, putting us as the second-place team on the leaderboard. | ['Rutger H. J. Fick', 'Saima Ben Hadj', 'Stéphanie Petit', 'Ali Mammadov', 'Alireza Moshayedi', 'Capucine Bertrand', 'Jules Dedieu', 'Gauthier Roy'] | 2021-09-04 | null | null | null | null | ['mitosis-detection'] | ['medical'] | [ 4.02204514e-01 5.51231384e-01 -1.07575633e-01 -3.08019109e-04
-1.14472294e+00 -7.20286012e-01 5.05722344e-01 1.49490908e-01
-7.32153356e-01 9.16094542e-01 2.71911584e-02 -2.04452083e-01
4.68878210e-01 -7.27695048e-01 -8.23784471e-01 -9.96385336e-01
1.69894502e-01 9.65461433e-01 7.59719551e-01 1.30937442e-01
1.11683421e-01 3.79888743e-01 -1.18787956e+00 3.95706356e-01
5.10376334e-01 7.18675792e-01 -7.92131480e-03 9.88637745e-01
3.58257294e-01 5.53097248e-01 -6.94410980e-01 -3.52048308e-01
1.43982977e-01 -6.24041677e-01 -8.05843711e-01 3.67039107e-02
8.07233304e-02 -2.11269364e-01 -2.62488037e-01 6.68486774e-01
4.99363780e-01 -1.96258724e-01 7.16212690e-01 -9.17428792e-01
-1.08034343e-01 6.26481414e-01 -7.86959708e-01 4.83562231e-01
7.08292425e-02 5.57586372e-01 8.54380786e-01 -7.75394440e-01
1.19238389e+00 6.33896589e-01 7.10256755e-01 8.20495605e-01
-1.43068576e+00 -6.61097467e-01 -2.21378401e-01 -1.59516171e-01
-1.28171933e+00 -4.18913156e-01 2.70429850e-01 -3.49947393e-01
8.04248989e-01 7.35714007e-03 9.04741585e-01 1.09408653e+00
1.61163107e-01 5.56541443e-01 9.85478818e-01 -3.46055359e-01
4.05115664e-01 -6.62160218e-02 6.47659376e-02 5.60275197e-01
-4.35305871e-02 7.52599314e-02 -3.00040215e-01 6.13843948e-02
7.23123908e-01 -3.00651789e-01 -2.75192678e-01 -1.96020812e-01
-1.39965498e+00 6.73208654e-01 4.50138181e-01 4.40230310e-01
-8.50795805e-02 1.50793254e-01 3.21341366e-01 -2.61792056e-02
4.90038991e-01 3.79587889e-01 -4.46554363e-01 6.85726777e-02
-1.14381468e+00 6.66447682e-03 6.09241724e-01 4.54847366e-01
8.13823164e-01 -5.09864748e-01 -3.19231689e-01 4.15590584e-01
5.92210777e-02 -6.34714961e-02 6.62404299e-01 -8.86800170e-01
-1.66791063e-02 7.08897650e-01 -1.91780016e-01 -2.23944247e-01
-6.84324741e-01 -6.24850214e-01 -9.04840827e-01 2.35301495e-01
9.88436878e-01 -2.54989892e-01 -1.50010633e+00 1.66144931e+00
3.89394253e-01 4.04547930e-01 -4.42747734e-02 9.63763118e-01
4.82956350e-01 3.06361586e-01 1.81559071e-01 -7.65584558e-02
1.40832281e+00 -9.54669237e-01 -1.53543234e-01 -2.03978732e-01
1.06861401e+00 -5.83243191e-01 8.08931708e-01 2.93584049e-01
-9.77983534e-01 -3.88130337e-01 -1.11370468e+00 -1.18525580e-01
-5.26123166e-01 3.93449724e-01 3.49677444e-01 2.30986446e-01
-1.37421238e+00 4.06752974e-01 -9.94980037e-01 -5.07027447e-01
5.85767150e-01 5.61565042e-01 -3.47016364e-01 2.14233436e-02
-7.92966902e-01 8.36386263e-01 5.70105791e-01 -7.68547133e-02
-1.29028857e+00 -6.09480381e-01 -5.94998121e-01 2.46255193e-02
1.54581308e-01 -9.13520753e-01 1.14474714e+00 -1.06750357e+00
-1.13680613e+00 1.37234259e+00 -2.53885269e-01 -5.46764314e-01
6.36294007e-01 5.93243897e-01 1.33235827e-01 2.19822630e-01
3.24940056e-01 1.30926979e+00 5.61606228e-01 -1.13911641e+00
-8.85282993e-01 -3.22268575e-01 -1.06789649e-01 7.00530410e-02
3.36831287e-02 -3.15154344e-01 -8.33529711e-01 -5.14725149e-01
3.09667420e-02 -1.15949678e+00 -2.13488325e-01 -1.16216920e-01
-5.65272868e-01 -2.51905303e-02 7.55852759e-01 -6.02874517e-01
8.03865016e-01 -2.40098763e+00 1.55816257e-01 2.76056319e-01
3.42091918e-01 1.04153063e-02 -1.59972280e-01 -1.38785601e-01
-2.29599074e-01 1.65691584e-01 -2.82159269e-01 -7.13065982e-01
-3.78847063e-01 7.90596604e-02 7.59968162e-02 7.24995315e-01
4.08655226e-01 9.34324741e-01 -9.43213820e-01 -5.06669700e-01
-1.89928077e-02 3.33295494e-01 -5.38724005e-01 3.44266649e-03
-3.97027403e-01 6.23264194e-01 -3.91673706e-02 6.81517720e-01
4.00330514e-01 -5.77477634e-01 1.43776074e-01 4.96964343e-02
2.72378772e-01 -8.39287266e-02 -9.06097949e-01 1.55088997e+00
-1.40184760e-01 7.82485843e-01 -7.26660192e-02 -7.21186042e-01
5.19708991e-01 2.01464087e-01 4.59335476e-01 -3.21037740e-01
3.11026484e-01 2.35482827e-01 2.21832216e-01 -3.82568948e-02
1.55539393e-01 -8.16033967e-03 -4.82586920e-02 5.66778779e-01
3.24923575e-01 4.90790792e-02 5.51359653e-01 2.64391959e-01
1.68202198e+00 5.45531698e-02 8.18834603e-02 2.46830657e-02
5.06655335e-01 2.05949813e-01 5.92107773e-01 4.23893929e-01
-2.35973835e-01 1.06692636e+00 9.46412444e-01 -4.50113714e-01
-1.34471774e+00 -8.45594168e-01 3.22016850e-02 1.09775078e+00
-1.58688366e-01 -2.50078976e-01 -9.41349506e-01 -1.03886628e+00
-2.19957560e-01 3.18568110e-01 -1.24630725e+00 -1.71707541e-01
-5.47938168e-01 -1.02285206e+00 7.13206053e-01 5.47888756e-01
3.07738006e-01 -9.97720003e-01 -5.75285435e-01 1.39495447e-01
-8.81742910e-02 -9.40224171e-01 -4.82114583e-01 6.74628615e-01
-5.44390082e-01 -1.42006254e+00 -8.25763047e-01 -8.92426074e-01
1.14676666e+00 -1.28439263e-01 1.09963989e+00 5.21233439e-01
-2.66861677e-01 -1.23727076e-01 -2.88346201e-01 -2.33657196e-01
-6.23853564e-01 4.97501940e-01 -4.20749396e-01 -5.23756854e-02
3.60384136e-01 -2.98110127e-01 -5.95659792e-01 2.57968068e-01
-7.86979735e-01 1.66170821e-01 7.10491955e-01 1.06983721e+00
9.47138369e-01 -6.64807931e-02 4.80770677e-01 -9.40186203e-01
5.62471338e-02 -3.69366705e-01 -6.84804738e-01 -4.84492294e-02
-1.71535805e-01 -2.71794856e-01 5.03345191e-01 -4.71849084e-01
-5.38752794e-01 6.43157840e-01 -1.02690324e-01 -3.97733867e-01
-1.24033332e-01 2.68116176e-01 7.52852391e-03 -6.50652796e-02
8.40794027e-01 1.43599272e-01 3.54271047e-02 3.96756679e-02
1.29592031e-01 3.29664588e-01 9.77483392e-01 -1.30833060e-01
7.68216848e-01 4.98420358e-01 4.63827811e-02 -2.79376090e-01
-8.29264402e-01 -3.57050598e-01 -6.70982480e-01 -2.37441316e-01
1.33415616e+00 -8.83631647e-01 -5.20345330e-01 5.66589773e-01
-1.02546418e+00 -8.14893246e-01 -5.26001155e-01 1.62993744e-01
-5.32261729e-01 -1.22237325e-01 -8.64807129e-01 -2.57596701e-01
-1.44527167e-01 -1.17716587e+00 1.22954655e+00 3.97128195e-01
-2.79783487e-01 -7.55811512e-01 1.84052914e-01 4.61428910e-01
1.13871761e-01 4.13991988e-01 7.69404531e-01 -1.01165366e+00
-6.03987694e-01 -4.75733608e-01 -1.23085387e-01 -5.69821149e-02
-2.00894043e-01 2.09195197e-01 -1.28392696e+00 -3.41322333e-01
-4.29819286e-01 -4.31885302e-01 1.19741642e+00 4.13642973e-01
1.28827727e+00 3.43591273e-01 -8.88689220e-01 7.17741311e-01
1.10088074e+00 -2.54839081e-02 6.15242660e-01 5.15214324e-01
2.72301346e-01 2.72431344e-01 3.55582476e-01 5.68778180e-02
3.32058638e-01 2.48877153e-01 4.74558860e-01 -4.84199017e-01
-3.21917802e-01 -1.79321498e-01 1.45096362e-01 -1.62963849e-02
8.52607042e-02 -4.02061105e-01 -1.05436730e+00 6.40471101e-01
-1.56244755e+00 -6.34234130e-01 1.45885319e-01 1.99360991e+00
8.49792480e-01 5.30823290e-01 2.95624495e-01 2.30751336e-01
7.86348939e-01 -2.40126029e-01 -6.84580743e-01 1.19976811e-01
-1.80507019e-01 2.56485492e-01 5.10590553e-01 3.88532192e-01
-1.30643880e+00 9.71492052e-01 6.18067551e+00 7.77734756e-01
-1.14261258e+00 1.80254057e-01 1.49179566e+00 -1.50699601e-01
-4.65927273e-02 1.81488112e-01 -9.73678291e-01 6.33219242e-01
1.04307449e+00 1.25448689e-01 2.69861162e-01 5.10431528e-01
2.83611827e-02 -5.67770720e-01 -1.21421623e+00 4.35124993e-01
-4.38207984e-02 -1.64199018e+00 -4.01707083e-01 3.90683174e-01
8.05091918e-01 2.47856870e-01 -3.41752082e-01 4.96277601e-01
5.79352558e-01 -1.16755641e+00 3.83575231e-01 3.78552616e-01
8.98003399e-01 -6.43290818e-01 1.19718909e+00 6.14367664e-01
-8.50652933e-01 -1.74550936e-02 -2.30902359e-01 2.29971141e-01
2.67989021e-02 6.81063175e-01 -1.08784962e+00 1.17159948e-01
4.10916984e-01 3.49359363e-01 -8.06667089e-01 1.12912560e+00
-1.71417683e-01 5.98216116e-01 -6.82889283e-01 2.62126535e-01
-4.59002219e-02 1.81711331e-01 2.05229416e-01 1.26659811e+00
2.10754156e-01 5.39603818e-04 1.04463369e-01 7.92419016e-01
-5.48146605e-01 -4.89000410e-01 -2.51157761e-01 1.83320284e-01
4.41188246e-01 1.65910113e+00 -1.39875901e+00 -5.36179483e-01
-3.79891284e-02 8.05848420e-01 4.29984152e-01 1.76784903e-01
-9.10779953e-01 -1.30552605e-01 1.98533796e-02 2.88466781e-01
3.85043561e-01 2.68391490e-01 -4.78341430e-01 -8.27250421e-01
-3.49925488e-01 -7.30637968e-01 5.27118444e-01 -7.78722644e-01
-9.86326933e-01 3.99727494e-01 -2.85530180e-01 -9.72602963e-01
-2.53655046e-01 -2.92835355e-01 -9.16647196e-01 7.28046894e-01
-1.18855512e+00 -1.41856635e+00 -5.12293756e-01 2.22707182e-01
4.39184636e-01 5.66182695e-02 7.23978937e-01 2.31498048e-01
-6.82627618e-01 5.42118609e-01 -3.44098806e-01 3.60205382e-01
7.75221646e-01 -1.42322922e+00 3.71776730e-01 7.51860023e-01
-8.78557861e-02 2.06808269e-01 4.42970753e-01 -7.47233033e-01
-9.03542697e-01 -1.22592473e+00 7.41686821e-01 -6.25468493e-01
6.73844635e-01 -5.16148806e-01 -8.36336374e-01 9.36972558e-01
3.45584527e-02 4.88986284e-01 5.97863674e-01 -3.33059371e-01
2.17381880e-01 2.05774695e-01 -1.16756845e+00 4.15200770e-01
7.96550691e-01 -2.72271156e-01 -2.14423329e-01 5.07243037e-01
5.48133850e-01 -8.77556264e-01 -7.67319560e-01 2.64751703e-01
2.58713692e-01 -8.12972426e-01 5.69640279e-01 -1.76270083e-01
4.75461930e-01 -6.12562180e-01 2.08513424e-01 -1.37654769e+00
-2.82594621e-01 -2.67846733e-01 3.58088501e-02 1.29188371e+00
7.29923666e-01 -2.75845528e-01 1.49230504e+00 7.19455779e-02
-3.27146173e-01 -1.05448341e+00 -1.11559772e+00 -2.41478428e-01
2.73598760e-01 -1.05788156e-01 3.72852594e-01 5.69261312e-01
-1.19582348e-01 5.56190193e-01 1.82532817e-01 4.14826758e-02
2.99339354e-01 -4.07936782e-01 9.31557059e-01 -9.74654257e-01
-5.33969641e-01 -5.01953781e-01 -4.48206753e-01 -6.56099141e-01
1.92197233e-01 -8.87061417e-01 1.01146616e-01 -1.22940254e+00
4.43622828e-01 -3.54895115e-01 -2.25255102e-01 7.86233187e-01
-2.23959938e-01 9.35857654e-01 -1.87320989e-02 2.16413811e-01
-5.79674065e-01 1.40753416e-02 1.27100599e+00 -1.24467105e-01
-3.68527174e-02 -1.58085585e-01 -4.98557001e-01 7.11814940e-01
6.66931570e-01 -4.19550478e-01 -2.06595808e-01 -1.67301409e-02
1.29620507e-01 5.11155836e-02 5.41698754e-01 -1.36312759e+00
4.05463815e-01 1.99954703e-01 1.28133094e+00 -6.54931307e-01
3.52487177e-01 -4.77797866e-01 2.62234211e-01 4.87787038e-01
-2.93685585e-01 -2.77267277e-01 2.02666536e-01 4.21875507e-01
1.06340677e-01 -8.25909898e-02 1.00404704e+00 -1.40966028e-01
-3.02255958e-01 1.65402517e-01 -5.40958941e-01 5.08374050e-02
1.41206431e+00 -3.98989856e-01 -6.45009398e-01 -1.11465789e-01
-1.07705140e+00 3.55556875e-01 9.89041090e-01 -1.65311575e-01
1.85421482e-01 -9.14144158e-01 -5.81670225e-01 4.00495023e-01
2.41371170e-02 5.64434171e-01 3.91024649e-02 9.28590715e-01
-5.24613142e-01 1.17206827e-01 -9.89346355e-02 -8.71450603e-01
-9.92458403e-01 5.42689800e-01 6.12971783e-01 -8.79621744e-01
-4.38614398e-01 1.06095243e+00 3.54129434e-01 -3.32851946e-01
7.34259561e-02 -2.05091044e-01 -1.64372548e-01 5.74076623e-02
3.14939886e-01 -3.56212929e-02 2.25549787e-01 -4.03456748e-01
-3.72027785e-01 4.58211720e-01 -1.49524271e-01 -1.24176674e-01
1.13353324e+00 2.74346709e-01 -1.63585972e-02 1.01876639e-01
9.54560757e-01 -2.02459216e-01 -1.50805533e+00 2.26876467e-01
-1.45877544e-02 3.02114990e-02 -1.41861783e-02 -9.79803205e-01
-1.20398688e+00 5.01770198e-01 7.06909180e-01 3.67535464e-03
1.28425193e+00 2.34463051e-01 5.98909795e-01 -3.86113930e-03
1.74805224e-01 -9.21858370e-01 3.60964984e-02 5.36462426e-01
3.21237236e-01 -1.10223329e+00 -1.74931720e-01 -1.95582464e-01
-2.95491993e-01 1.02411282e+00 8.86461794e-01 -3.36330146e-01
2.83869833e-01 6.08114362e-01 9.60585475e-02 -1.86540186e-01
-1.12634122e+00 -1.79737270e-01 -1.68318331e-01 6.73292279e-01
3.86367887e-01 -1.44031450e-01 -1.44063160e-01 5.01565337e-01
-3.48299593e-01 3.58880639e-01 7.96334982e-01 7.36922801e-01
-3.80585402e-01 -1.01432669e+00 -2.40821600e-01 7.58428514e-01
-4.86517042e-01 9.28217247e-02 -6.30034804e-01 9.34691727e-01
4.13474143e-01 4.38595653e-01 5.20326734e-01 -4.10827279e-01
9.96054783e-02 -2.47582607e-02 2.21815944e-01 -6.91716135e-01
-7.07300723e-01 3.52520466e-01 1.41255092e-02 -2.86465049e-01
-1.78462844e-02 -7.26339698e-01 -1.58972180e+00 -3.35665166e-01
-2.70605713e-01 -1.09909326e-02 3.72831404e-01 9.65947151e-01
2.41349608e-01 7.65701056e-01 3.65387976e-01 -1.15551507e+00
-2.30784789e-01 -1.08076322e+00 -3.68524760e-01 1.74635798e-01
5.19063652e-01 -4.48160946e-01 -4.64244574e-01 3.13096881e-01] | [14.999314308166504, -3.099364757537842] |
b44dda11-e233-4866-857c-1bea51ee0761 | necessary-and-sufficient-conditions-for-exact | 2208.07983 | null | https://arxiv.org/abs/2208.07983v1 | https://arxiv.org/pdf/2208.07983v1.pdf | Necessary and sufficient conditions for exact closures of epidemic equations on configuration model networks | We prove that the exact closure of SIR pairwise epidemic equations on a configuration model network is possible if and only if the degree distribution is Poisson, Binomial, or Negative Binomial. The proof relies on establishing, for these specific degree distributions, the equivalence of the closed pairwise model and the so-called dynamical survival analysis (DSA) edge-based model which was previously shown to be exact. Indeed, as we show here, the DSA model is equivalent to the well-known edge-based Volz model. We use this result to provide reductions of the closed pairwise and Volz models to the same single equation involving only susceptibles, which has a useful statistical interpretation in terms of the times to infection. We illustrate our findings with some numerical examples. | ['Grzegorz A. Rempala', 'Eben Kenah', 'Istvan Z. Kiss'] | 2022-08-16 | null | null | null | null | ['survival-analysis'] | ['miscellaneous'] | [ 4.41294014e-02 2.12157905e-01 1.92243457e-01 3.47651929e-01
3.24547976e-01 -7.17197299e-01 4.55460101e-01 4.94157284e-01
-2.42476970e-01 1.00734174e+00 -3.51463407e-01 -6.71978891e-01
-7.95679867e-01 -1.03500903e+00 -5.20020247e-01 -1.06631505e+00
-7.91775882e-01 8.07499111e-01 4.53066677e-01 -5.80649972e-01
-9.95487943e-02 4.77654815e-01 -9.76623535e-01 -7.93287098e-01
8.81500065e-01 5.04971981e-01 -3.60790640e-01 1.06368792e+00
1.13700166e-01 2.29245439e-01 -6.67064786e-01 -6.03553593e-01
2.63684779e-01 -3.45181078e-01 -5.85755706e-01 -3.45607758e-01
-3.81147683e-01 -4.28179130e-02 -2.95334458e-01 8.17158520e-01
1.51754305e-01 -3.86474431e-01 8.72862577e-01 -1.75781429e+00
-5.75743735e-01 3.89459997e-01 -7.13416874e-01 2.91209936e-01
2.47604012e-01 -4.13496882e-01 5.75026453e-01 -2.39838377e-01
7.79823482e-01 1.21649253e+00 1.05346715e+00 3.00789446e-01
-1.54899156e+00 -3.88891131e-01 -2.27993265e-01 -3.62862974e-01
-1.64239383e+00 -1.50361592e-02 2.99261719e-01 -4.50830311e-01
6.67338192e-01 4.30468798e-01 9.92043853e-01 7.21774042e-01
6.27297997e-01 5.08860424e-02 1.20300114e+00 -4.08144742e-01
2.70908415e-01 -3.80875692e-02 5.45633495e-01 5.22992611e-01
1.14555335e+00 1.40900612e-02 2.16110334e-01 -6.85773432e-01
1.04501057e+00 1.55192256e-01 -2.63968736e-01 -5.51101744e-01
-7.79803574e-01 9.69171703e-01 8.43816698e-02 4.46114391e-01
-3.23076308e-01 6.23135455e-02 1.27476370e-02 4.97265071e-01
5.60880959e-01 1.21556036e-02 -2.64468759e-01 3.16198409e-01
-4.87880260e-01 2.88338423e-01 1.54134309e+00 8.68806362e-01
4.41111058e-01 -1.17518850e-01 3.26148242e-01 3.35137308e-01
3.97972941e-01 9.09826756e-01 -6.83745801e-01 -7.83997655e-01
-5.63725419e-02 3.78412426e-01 2.22386330e-01 -9.01068509e-01
-4.45486873e-01 -4.55394447e-01 -1.36610687e+00 -1.64093390e-01
5.92750788e-01 -9.15091857e-02 -4.62399185e-01 2.15843916e+00
1.00596786e-01 1.82562411e-01 1.47542357e-01 -3.55251282e-02
2.20215425e-01 7.36776769e-01 9.19335112e-02 -9.34215963e-01
7.59369791e-01 8.98187310e-02 -6.98276401e-01 6.50100529e-01
4.72224087e-01 -3.12340528e-01 -1.71195678e-02 9.19006094e-02
-1.11548638e+00 3.34631205e-01 -7.13472962e-01 6.14869893e-01
-2.09257886e-01 -6.33219600e-01 4.49995279e-01 6.94841146e-01
-1.73228490e+00 4.80767608e-01 -6.05611503e-01 -1.18341422e+00
-1.35359526e-01 4.91227895e-01 -1.67691514e-01 1.51387248e-02
-1.29255843e+00 5.39557993e-01 -2.05732495e-01 1.62908629e-01
-7.11195767e-01 -6.57346368e-01 -5.37080526e-01 1.00686230e-01
1.96610704e-01 -9.50301647e-01 9.15634274e-01 -5.44503510e-01
-8.62060428e-01 7.69293547e-01 -2.51931936e-01 -5.01592934e-01
4.71991777e-01 4.95919168e-01 -1.75283849e-01 3.10769439e-01
-1.93753034e-01 6.10694066e-02 2.87417531e-01 -1.49636281e+00
-5.85617758e-02 -4.41187412e-01 4.15768147e-01 -4.25901800e-01
5.88294156e-02 2.41735637e-01 -1.84239075e-02 -3.91636342e-01
-1.28530383e-01 -9.14591134e-01 -4.09530729e-01 -2.15660945e-01
-5.46498299e-01 -3.65554333e-01 4.74739283e-01 -4.36292201e-01
1.25121832e+00 -1.66949034e+00 3.45028818e-01 8.08874786e-01
5.52999914e-01 -1.10890768e-01 4.12132815e-02 1.04869056e+00
-1.13168143e-01 5.88745832e-01 -5.12490094e-01 -1.50881901e-01
-2.86521763e-01 2.44266406e-01 -6.96227849e-02 8.05293679e-01
9.35157314e-02 5.19214153e-01 -7.47091472e-01 -6.71965837e-01
-2.77619720e-01 5.83531916e-01 -5.44778645e-01 -2.00482219e-01
3.40270132e-01 3.36538017e-01 -4.22225803e-01 3.84548753e-01
1.04680431e+00 -2.05176502e-01 5.28269231e-01 3.50579351e-01
-1.66060090e-01 -4.97455090e-01 -9.86201048e-01 5.67958176e-01
-2.86788046e-01 3.48488718e-01 5.43873429e-01 -1.02992725e+00
5.38118243e-01 3.95946473e-01 6.88871801e-01 1.82594821e-01
4.00081724e-01 1.87031999e-01 1.96723625e-01 -9.19060484e-02
8.20992514e-02 -4.06858087e-01 -2.86492795e-01 3.13258886e-01
-4.08847034e-02 2.86212385e-01 4.28551435e-01 8.36760283e-01
1.50812149e+00 -6.26922309e-01 3.38623613e-01 -9.89173114e-01
2.70447165e-01 -6.63822591e-02 4.98546958e-01 8.71609330e-01
-2.63563901e-01 5.28511927e-02 1.07775915e+00 1.93200603e-01
-1.03622186e+00 -1.79553807e+00 -4.97890025e-01 -1.49315922e-02
5.96516669e-01 -4.48545069e-01 -9.16583598e-01 -4.19840217e-03
5.09488359e-02 1.65174767e-01 -9.00140464e-01 -1.28242284e-01
-2.50815749e-01 -9.90267515e-01 4.22678798e-01 -7.48721091e-03
2.91588396e-01 -4.05848593e-01 -1.37008801e-02 -3.74265052e-02
7.80934915e-02 -7.33188808e-01 -9.58122611e-02 -1.42471888e-03
-9.00030553e-01 -1.51770627e+00 -9.94659007e-01 -6.51441991e-01
7.87646055e-01 1.70144066e-01 1.05453956e+00 4.76379544e-01
-1.73459172e-01 8.86000097e-01 -7.11700842e-02 -1.25736222e-01
-8.57440770e-01 -1.10162042e-01 3.81164312e-01 -1.50080577e-01
1.11241639e-01 -8.24594378e-01 -5.96383154e-01 4.51047510e-01
-1.06905961e+00 -1.69990718e-01 1.13159597e-01 2.86183923e-01
1.70666188e-01 1.92667767e-01 8.55070412e-01 -6.78898156e-01
6.50036931e-01 -9.17648435e-01 -6.78294957e-01 5.02263665e-01
-4.75358069e-01 -1.05308786e-01 4.68565851e-01 -1.54252931e-01
-6.86107934e-01 -1.58837914e-01 2.06272259e-01 -7.52314553e-02
1.63714752e-01 5.47596693e-01 5.66283539e-02 -2.75182277e-01
2.66538989e-02 -1.24184391e-03 2.24256560e-01 -2.34962657e-01
-3.45743001e-02 6.92806244e-01 2.58533031e-01 -2.92406678e-01
9.81452346e-01 5.54771543e-01 6.91713631e-01 -1.18418622e+00
4.22795024e-03 -2.27683261e-01 -4.45432067e-01 -4.08776104e-01
7.43655801e-01 -5.45970142e-01 -1.28040779e+00 7.90282786e-01
-1.23690701e+00 -2.81431735e-01 -1.83235273e-01 9.44158286e-02
-5.42414188e-01 3.93579066e-01 -9.79935884e-01 -1.69422829e+00
1.51937529e-01 -5.80760598e-01 5.94010413e-01 1.36479855e-01
7.88836107e-02 -1.43953001e+00 6.35904253e-01 -3.57355058e-01
6.44040048e-01 6.12081409e-01 1.01878834e+00 -5.11217117e-01
-4.45484668e-01 -3.56449634e-01 -2.24060208e-01 -1.73448563e-01
-1.24027602e-01 4.65434819e-01 -1.32626235e-01 -3.36858094e-01
-1.70930434e-04 4.64651495e-01 8.27552557e-01 7.51624882e-01
4.40749973e-01 -2.95753330e-01 -8.18702400e-01 2.76685417e-01
1.88875663e+00 5.43285906e-02 5.33080637e-01 -3.50989372e-01
3.32147896e-01 8.06801260e-01 1.92054920e-02 3.23892891e-01
4.33450937e-01 6.54170573e-01 2.60616511e-01 9.53757539e-02
5.07807970e-01 -6.73312917e-02 2.52056003e-01 9.98313010e-01
-4.95388478e-01 -5.80173492e-01 -6.80311859e-01 6.12982392e-01
-1.74636745e+00 -1.10189915e+00 -7.12843418e-01 2.63031721e+00
6.86108530e-01 -9.27209258e-02 5.72238207e-01 1.11938762e-02
1.14908886e+00 -4.63585109e-01 3.12789902e-02 -5.74609876e-01
-3.90082657e-01 3.25426608e-01 1.13489473e+00 1.03992999e+00
-7.74539649e-01 2.60027021e-01 7.90860939e+00 4.46098834e-01
-4.37807322e-01 2.78060824e-01 4.85809118e-01 2.14176908e-01
-3.77770692e-01 2.81468660e-01 -6.10528469e-01 3.81324828e-01
1.11367834e+00 -7.65771210e-01 3.46515089e-01 2.86896944e-01
3.36664826e-01 -4.91370112e-01 -6.17621601e-01 4.67688560e-01
-2.09851637e-01 -9.11615491e-01 -1.65142417e-01 6.85659885e-01
6.69881344e-01 -5.32193661e-01 -3.04162294e-01 -6.00554310e-02
7.25652099e-01 -6.55599058e-01 1.43612817e-01 5.49714386e-01
7.44642675e-01 -7.26527750e-01 7.68719435e-01 1.45564929e-01
-1.34133875e+00 8.74769911e-02 -4.37319204e-02 -2.19255015e-01
6.27713323e-01 6.63171411e-01 -6.63104281e-02 7.95601666e-01
4.98880982e-01 5.24405003e-01 -2.84803808e-01 1.47085190e+00
3.75899374e-01 5.36633492e-01 -8.70728254e-01 8.46469253e-02
-2.76103973e-01 -6.11808479e-01 9.65672314e-01 7.06077099e-01
4.09827530e-01 3.00050020e-01 -2.74236470e-01 7.59657085e-01
2.50204742e-01 -1.87873423e-01 -9.07848537e-01 1.85076207e-01
3.05099875e-01 1.10790634e+00 -1.04552996e+00 4.37443033e-02
6.52630301e-03 7.14106560e-01 -1.74724147e-01 5.67599654e-01
-6.38334692e-01 -4.50169414e-01 6.85164034e-01 5.20529568e-01
2.01460510e-01 -3.20240736e-01 1.15596056e-01 -1.09658027e+00
-4.16570604e-01 3.64438668e-02 1.67068049e-01 -2.82229513e-01
-1.46402776e+00 4.46675867e-01 7.73763657e-01 -7.12599576e-01
-2.71101743e-01 -4.92488474e-01 -9.20175254e-01 8.15050125e-01
-9.13992405e-01 -7.03019440e-01 1.42501950e-01 4.56827909e-01
-3.87145936e-01 3.83464903e-01 6.94243670e-01 2.57954746e-01
-8.15629244e-01 2.71270037e-01 4.68000203e-01 -3.47869426e-01
1.02826573e-01 -1.27939308e+00 2.48771757e-02 8.23792756e-01
-6.75053179e-01 7.87520230e-01 1.22225356e+00 -9.79399443e-01
-1.44558799e+00 -7.89701819e-01 1.20266759e+00 -3.52729708e-01
7.54762352e-01 -6.10597432e-01 -5.80179214e-01 5.65859914e-01
2.53408074e-01 -2.27670714e-01 4.62941766e-01 -1.57680456e-02
1.86271593e-01 -3.17332111e-02 -1.36726785e+00 5.90672672e-01
1.24814045e+00 8.49040970e-02 -1.79376721e-01 3.62992018e-01
6.24648273e-01 3.29970896e-01 -8.60995233e-01 2.49621138e-01
3.96626323e-01 -7.75148392e-01 9.09368217e-01 -6.26369894e-01
1.43493369e-01 -1.73819900e-01 -1.39491737e-01 -1.10686779e+00
-1.71848044e-01 -9.17886257e-01 -2.40199198e-03 1.27837336e+00
3.66262734e-01 -1.08893239e+00 3.60842496e-02 2.23607659e-01
5.47326088e-01 -9.17900085e-01 -1.17883432e+00 -1.15378714e+00
4.71829057e-01 5.88916726e-02 3.16237092e-01 7.55288303e-01
2.66078621e-01 1.74204215e-01 -2.98833102e-01 1.38514429e-01
9.17719603e-01 -4.35531706e-01 3.83463264e-01 -1.87875605e+00
-2.88245350e-01 -4.36779112e-01 -6.76193118e-01 -5.83328128e-01
9.66132507e-02 -5.56176662e-01 -9.19925421e-02 -1.58663833e+00
4.92581218e-01 -8.53524685e-01 -9.74632278e-02 -2.24980935e-01
1.00130618e-01 2.91102648e-01 3.47263627e-02 2.47843638e-01
-4.59226012e-01 2.44713932e-01 6.69393361e-01 2.39378780e-01
-2.73077428e-01 4.99951512e-01 -6.22952580e-01 4.15357113e-01
7.98378706e-01 -4.91792977e-01 -4.08513010e-01 4.14861441e-01
8.13758194e-01 4.72044855e-01 7.08775759e-01 -6.57149792e-01
1.49750158e-01 -1.28038406e-01 -3.71193498e-01 -4.84423786e-01
2.27892369e-01 -7.74785995e-01 6.62580252e-01 8.84992480e-01
-5.31537132e-03 4.02772039e-01 4.68991995e-02 9.53351259e-01
3.63169670e-01 -1.01320967e-01 5.42933404e-01 2.74592906e-01
3.14000309e-01 3.30041170e-01 -5.68493426e-01 3.24552953e-02
1.40584326e+00 -2.22097993e-01 -4.80747432e-01 -9.69878852e-01
-8.81754100e-01 2.80046850e-01 7.55879819e-01 -3.81040543e-01
2.69871920e-01 -1.05742800e+00 -6.92839026e-01 -2.10849866e-01
-2.66442746e-01 -5.49528122e-01 4.26786810e-01 1.56361687e+00
-6.54245317e-01 3.70148569e-01 1.33003065e-04 -5.61103106e-01
-1.20650768e+00 6.85220480e-01 4.84440565e-01 -4.54156518e-01
-1.28014892e-01 1.92055702e-01 3.56424421e-01 1.40942086e-03
-6.52993694e-02 5.39067909e-02 5.41527160e-02 -3.33926469e-01
2.17206225e-01 4.83054012e-01 -4.68864381e-01 -7.40426242e-01
-5.98004401e-01 7.36281216e-01 2.73397446e-01 -1.86065912e-01
1.42359388e+00 -4.44708079e-01 -7.79208124e-01 5.57662725e-01
8.98796678e-01 1.57488689e-01 -6.96177006e-01 1.85712248e-01
-4.52732146e-01 5.29881269e-02 -4.98722970e-01 -2.96903402e-01
-9.94751990e-01 5.01027942e-01 2.32717693e-01 1.25297606e+00
1.15482676e+00 4.32447344e-01 3.43117565e-01 -8.35797191e-02
5.02964735e-01 -2.25646317e-01 -7.46525288e-01 3.64215404e-01
7.19993889e-01 -6.04231000e-01 -2.29045615e-01 -1.04491091e+00
1.64900824e-01 8.19463968e-01 2.39670306e-01 -5.66302359e-01
1.43670034e+00 5.50634742e-01 -7.17566609e-01 -1.86899036e-01
-7.20997453e-01 -1.36028364e-01 -3.97001356e-01 9.24633503e-01
2.41325602e-01 1.66620061e-01 -8.23829412e-01 5.46127498e-01
1.21225834e-01 -2.83481441e-02 8.53992939e-01 7.17801869e-01
-1.88416705e-01 -1.19204688e+00 -1.64179727e-01 5.90097725e-01
-4.00240839e-01 -8.31108317e-02 -7.48261988e-01 1.07530856e+00
-2.48317912e-01 1.00711381e+00 3.06081325e-01 -1.82619855e-01
6.08503185e-02 -1.43800542e-01 7.68300116e-01 -1.98425844e-01
-4.10198331e-01 -2.89407074e-01 -2.74018079e-01 1.36625953e-02
-5.44292808e-01 -5.39094567e-01 -7.04254806e-01 -1.28211415e+00
-4.79488969e-01 3.04737538e-01 1.86109185e-01 1.02365029e+00
2.30754152e-01 2.02085271e-01 6.87381506e-01 -4.45802450e-01
-2.90240169e-01 -8.96018088e-01 -1.15645623e+00 -1.93197787e-01
2.68146873e-01 -5.76572895e-01 -8.91816199e-01 -3.53408605e-01] | [5.991213321685791, 4.451955795288086] |
125e9a62-18d7-4c9e-a91d-129d07c52d24 | posetrack21-a-dataset-for-person-search-multi | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Doring_PoseTrack21_A_Dataset_for_Person_Search_Multi-Object_Tracking_and_Multi-Person_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Doring_PoseTrack21_A_Dataset_for_Person_Search_Multi-Object_Tracking_and_Multi-Person_CVPR_2022_paper.pdf | PoseTrack21: A Dataset for Person Search, Multi-Object Tracking and Multi-Person Pose Tracking | Current research evaluates person search, multi-object tracking and multi-person pose estimation as separate tasks and on different datasets although these tasks are very akin to each other and comprise similar sub-tasks, e.g. person detection or appearance-based association of detected persons. Consequently, approaches on these respective tasks are eligible to complement each other. Therefore, we introduce PoseTrack21, a large-scale dataset for person search, multi-object tracking and multi-person pose tracking in real-world scenarios with a high diversity of poses. The dataset provides rich annotations like human pose annotations including annotations of joint occlusions, bounding box annotations even for small persons, and person-ids within and across video sequences. The dataset allows to evaluate multi-object tracking and multi-person pose tracking jointly with person re-identification or exploit structural knowledge of human poses to improve person search and tracking, particularly in the context of severe occlusions. With PoseTrack21, we want to encourage researchers to work on joint approaches that perform reasonably well on all three tasks. | ['Jürgen Gall', 'Bernt Schiele', 'Shanshan Zhang', 'Di Chen', 'Andreas Döring'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['person-search', 'multi-person-pose-estimation'] | ['computer-vision', 'computer-vision'] | [-2.94225663e-01 -4.73344982e-01 1.49303135e-02 -2.78550833e-01
-7.56071031e-01 -7.57963955e-01 5.53818464e-01 1.22817587e-02
-7.83626616e-01 6.57108426e-01 3.30350816e-01 5.22041261e-01
5.75064607e-02 -1.67549536e-01 -5.17308831e-01 -2.09179550e-01
1.28623620e-01 1.09431207e+00 5.21445990e-01 2.16420546e-01
-2.89047062e-01 5.14227748e-01 -1.77200580e+00 -3.12638171e-02
2.54730523e-01 4.82365698e-01 -2.51946718e-01 9.24841583e-01
5.45236766e-01 9.80748236e-03 -6.60342693e-01 -1.02127218e+00
3.41865897e-01 1.66460112e-01 -5.50587535e-01 2.01407701e-01
1.45453203e+00 -5.63772976e-01 -3.23823661e-01 8.15552831e-01
8.49940538e-01 2.27724403e-01 4.09904808e-01 -1.51137853e+00
-6.33339658e-02 -1.62909552e-03 -7.52571523e-01 2.63869315e-01
1.00471532e+00 2.89863527e-01 6.02963328e-01 -7.07608461e-01
5.52504361e-01 1.81164253e+00 1.33388543e+00 6.35976017e-01
-1.21809542e+00 -6.20790184e-01 3.83200616e-01 -1.67552219e-03
-1.57751787e+00 -5.19809544e-01 1.36162236e-01 -6.22381210e-01
6.88622653e-01 5.53601503e-01 7.76771784e-01 1.43549335e+00
-4.41512495e-01 8.63534808e-01 8.29375625e-01 -2.48911768e-01
-4.36838895e-01 3.60081196e-01 4.55581784e-01 7.63839781e-01
8.68803680e-01 4.10839677e-01 -5.73206902e-01 -4.53805238e-01
5.69801867e-01 2.09734246e-01 1.10854805e-01 -5.30800045e-01
-1.34770834e+00 4.42160189e-01 2.71920741e-01 -3.73164634e-03
-1.97286665e-01 2.87406683e-01 5.53184032e-01 -1.02866530e-01
2.17285886e-01 -3.58727537e-02 -3.91763806e-01 -3.65527868e-02
-9.14091766e-01 1.05165076e+00 6.22383177e-01 1.33911049e+00
3.18655401e-01 -2.53263354e-01 -7.63681412e-01 7.58895278e-01
5.03853559e-01 7.85906076e-01 -6.50429539e-03 -7.32012391e-01
5.33443213e-01 6.46185279e-01 6.81564331e-01 -7.93155491e-01
-8.25336158e-01 -4.60153133e-01 -3.79338980e-01 -7.28184357e-02
8.64829540e-01 -3.14073235e-01 -7.76019752e-01 1.84697735e+00
7.91535020e-01 2.13668212e-01 -5.36682248e-01 1.12853622e+00
1.01844668e+00 -3.08623407e-02 4.33293998e-01 7.78870881e-02
2.12857652e+00 -1.09847021e+00 -5.87355137e-01 -5.72000146e-01
2.57299274e-01 -7.02234805e-01 4.55563694e-01 3.71232815e-02
-1.21017683e+00 -9.04499352e-01 -5.14716566e-01 1.27356559e-01
-4.35016423e-01 3.41767013e-01 5.18145144e-01 1.16520476e+00
-8.65017414e-01 2.39421666e-01 -6.25901520e-01 -6.96072102e-01
3.52061510e-01 5.10765433e-01 -3.37595731e-01 -2.04642098e-02
-8.76192689e-01 1.03304160e+00 2.74598122e-01 1.42962769e-01
-6.08792543e-01 -6.50857091e-01 -9.04481411e-01 -3.68794203e-01
5.31846523e-01 -1.11758018e+00 1.19400370e+00 -3.46000344e-01
-7.20678627e-01 1.25318885e+00 -3.14795494e-01 -2.96913475e-01
9.81753886e-01 -7.32113361e-01 -4.80200529e-01 -2.30594859e-01
4.04334486e-01 7.81239510e-01 7.60343432e-01 -1.17645919e+00
-8.75779629e-01 -9.25841987e-01 -2.32071713e-01 2.78913796e-01
-2.66773194e-01 7.56391644e-01 -1.11363173e+00 -6.30846620e-01
-3.59414130e-01 -1.36787677e+00 -1.62447155e-01 1.86329558e-01
-5.74671984e-01 -3.77936244e-01 7.81519473e-01 -8.64940405e-01
9.92852390e-01 -1.55950940e+00 2.76018381e-01 -2.67483313e-02
1.48810595e-01 2.57697701e-01 -2.91134492e-02 1.88093837e-02
2.09682658e-01 -2.43476257e-01 4.67491180e-01 -9.03287292e-01
3.68098825e-01 -1.36265159e-01 3.14982086e-01 7.38179564e-01
-2.67228693e-01 1.19825089e+00 -6.72068357e-01 -8.61264110e-01
3.01983178e-01 4.91041511e-01 -2.67854244e-01 -4.09721658e-02
1.55753139e-02 7.20798731e-01 -4.38732862e-01 1.05240703e+00
5.68985105e-01 -1.01875015e-01 -2.93454528e-02 -4.16896939e-01
-7.58763701e-02 -5.55309832e-01 -1.73879755e+00 1.48932552e+00
2.67567992e-01 3.26187640e-01 2.54773587e-01 -1.57978117e-01
3.64390671e-01 4.80051905e-01 7.47935057e-01 -2.42054895e-01
1.27621770e-01 -7.04784244e-02 -3.38986427e-01 -3.25749159e-01
7.99512088e-01 3.50360125e-01 -2.47646526e-01 2.84963697e-01
-5.06445542e-02 6.65361822e-01 6.36456192e-01 1.92408939e-03
7.97993183e-01 4.37021703e-01 2.78539598e-01 3.29719903e-03
5.34257114e-01 -9.41351429e-02 4.59822208e-01 1.01703990e+00
-7.89567828e-01 3.96369547e-01 -5.34661591e-01 -8.23455811e-01
-1.06346548e+00 -1.01209283e+00 -1.95019752e-01 1.63023674e+00
4.56890345e-01 -4.79659140e-01 -5.31713486e-01 -6.79887354e-01
4.49033558e-01 -1.59987956e-01 -4.39703822e-01 3.22358489e-01
-7.95521379e-01 -1.02337289e+00 7.76036680e-01 7.28547275e-01
3.08667451e-01 -8.46679568e-01 -5.79114676e-01 1.59747332e-01
-2.54052609e-01 -1.57420576e+00 -9.13143814e-01 -3.76322061e-01
-3.88245195e-01 -1.44165504e+00 -1.26281714e+00 -7.40684986e-01
4.39814419e-01 1.62378162e-01 1.25654650e+00 8.71727020e-02
-7.45181918e-01 1.06404173e+00 9.91744827e-03 -4.45942730e-01
5.40939160e-02 -2.39001021e-01 6.15002692e-01 -5.16917780e-02
5.29284775e-01 -3.67928743e-02 -6.78938687e-01 9.41806614e-01
-5.52998073e-02 -2.67914772e-01 2.05895349e-01 1.72841072e-01
3.79646599e-01 -2.36553282e-01 3.11833285e-02 -4.74905342e-01
3.10194552e-01 9.39544737e-02 -6.43704653e-01 6.57123744e-01
-8.24122652e-02 -4.81042087e-01 -2.24668264e-01 -7.13185072e-01
-8.47848117e-01 4.31904435e-01 5.50369248e-02 -3.71126175e-01
-4.66375768e-01 -5.08328319e-01 -1.94897026e-01 -3.73904794e-01
5.92085898e-01 -4.01613815e-03 -5.10471880e-01 -6.97894394e-01
3.18783522e-01 3.50266814e-01 8.71673524e-01 -6.35330200e-01
1.00596464e+00 5.91468811e-01 -9.93335769e-02 -5.47260404e-01
-8.65660727e-01 -1.00004506e+00 -1.13898265e+00 -5.72015345e-01
1.13121736e+00 -1.54081786e+00 -1.22928071e+00 5.49541056e-01
-1.30475783e+00 1.10630974e-01 1.38592243e-01 4.94054288e-01
-2.37938285e-01 5.54289281e-01 -5.36036432e-01 -1.30782986e+00
-4.15349364e-01 -9.35796380e-01 1.79719174e+00 3.64724249e-01
-5.31961739e-01 -8.88821006e-01 6.03214614e-02 7.34431684e-01
9.42873862e-03 5.88720620e-01 -2.79519260e-01 -5.57888627e-01
-6.55182183e-01 -7.07207203e-01 -2.15095609e-01 -5.38397610e-01
-2.21484721e-01 -3.61349165e-01 -9.11220312e-01 -8.84939671e-01
-7.22704947e-01 -3.75887334e-01 7.98138380e-01 4.94888365e-01
7.53450155e-01 -8.14847574e-02 -1.05913711e+00 4.99633342e-01
9.62543070e-01 -2.64109701e-01 1.00224048e-01 3.61077666e-01
9.44039702e-01 5.94185174e-01 5.10175288e-01 5.74048102e-01
7.11374700e-01 1.57035589e+00 1.10882670e-01 -2.19993051e-02
-4.34232384e-01 -8.05481002e-02 3.07600856e-01 -1.07198134e-01
-5.30072808e-01 -1.08513765e-01 -8.02454770e-01 3.70543063e-01
-1.92509341e+00 -1.14221466e+00 -5.09083331e-01 2.26829529e+00
5.03271818e-01 -8.55156500e-03 1.13039219e+00 -2.16224104e-01
1.22826517e+00 -7.34743699e-02 -4.93773848e-01 6.80719435e-01
-2.69533396e-02 -4.63338852e-01 6.26643777e-01 3.05228621e-01
-1.73306632e+00 6.81179225e-01 6.57344246e+00 4.42585588e-01
-1.47738233e-01 2.64404029e-01 4.50148620e-02 -3.39885086e-01
5.06269932e-01 -4.07316417e-01 -1.87242389e+00 5.20908356e-01
5.57900071e-01 3.32085639e-01 7.26728514e-02 5.86250484e-01
-1.81947246e-01 -1.61778703e-01 -1.23533106e+00 1.47342920e+00
2.94370562e-01 -7.26674259e-01 -3.53060901e-01 5.69138601e-02
6.31013215e-01 -3.61385405e-01 -3.26503105e-02 4.04475212e-01
3.17808211e-01 -6.93590701e-01 8.80100787e-01 6.28222227e-01
6.35024428e-01 -4.26814139e-01 5.38363934e-01 2.05571160e-01
-1.87787867e+00 -1.99951485e-01 -5.91954552e-02 2.30863243e-01
5.88107526e-01 -5.19208377e-03 -3.14485699e-01 4.97275651e-01
9.92364228e-01 4.80618954e-01 -1.12339008e+00 1.55288768e+00
4.10502285e-01 -1.79794759e-01 -5.01381040e-01 2.50921845e-01
-2.36757398e-01 1.90530077e-01 7.47841716e-01 1.40527463e+00
7.70674720e-02 -2.97898263e-01 9.82397258e-01 5.92413902e-01
2.53155977e-01 -2.58480996e-01 -1.70614913e-01 4.70949352e-01
4.10241038e-01 1.44306087e+00 -6.77865684e-01 -5.14118552e-01
-5.07856369e-01 1.00175118e+00 1.37774989e-01 2.61669457e-01
-1.11300647e+00 3.98486823e-01 7.70418942e-01 3.56959879e-01
1.84131891e-01 -1.55150682e-01 3.60537581e-02 -1.05131412e+00
1.74894020e-01 -8.01750779e-01 9.00109470e-01 -4.60432887e-01
-1.33049905e+00 3.17643732e-01 3.00674409e-01 -1.05091500e+00
-3.00186008e-01 -5.93503833e-01 -2.70536900e-01 8.57942939e-01
-1.13797784e+00 -1.87416983e+00 -5.75094640e-01 7.66250491e-01
5.02893984e-01 -2.90219486e-01 6.81925416e-01 8.07496250e-01
-8.78314614e-01 8.38411152e-01 -2.93140292e-01 4.50326860e-01
1.07754576e+00 -1.12170970e+00 4.89884436e-01 8.30094874e-01
1.15518577e-01 6.86917543e-01 8.24292660e-01 -1.21955454e+00
-1.28651142e+00 -1.29164219e+00 7.38387108e-01 -1.37062120e+00
2.02034280e-01 -6.15580797e-01 -4.00925547e-01 9.70908821e-01
-2.76391685e-01 7.95293748e-02 6.75187171e-01 5.77957749e-01
-2.49681830e-01 1.74843758e-01 -1.03007591e+00 4.89936292e-01
1.50504076e+00 -2.78843969e-01 -3.02411199e-01 6.96422279e-01
2.78341711e-01 -7.76622653e-01 -8.61312985e-01 5.17399132e-01
8.36208224e-01 -7.91440904e-01 1.87409532e+00 -5.03362298e-01
-6.91078424e-01 -4.63237315e-01 6.69381320e-02 -5.94607413e-01
-4.93701518e-01 -4.80279386e-01 -5.74851990e-01 1.15379655e+00
-1.16080083e-01 -1.29272610e-01 9.26597536e-01 9.76505995e-01
1.90289125e-01 -3.00362464e-02 -8.11818361e-01 -1.08772659e+00
-5.45390785e-01 -1.28147528e-01 5.06615937e-01 4.00164515e-01
-5.99923491e-01 9.03896764e-02 -9.00512993e-01 4.90888178e-01
1.27615166e+00 1.87246457e-01 1.28269982e+00 -1.63893354e+00
-2.70613164e-01 -3.23663503e-01 -4.50963140e-01 -8.48465979e-01
8.05304199e-02 -6.04487836e-01 -1.07387625e-01 -1.26276970e+00
8.22131276e-01 -2.42213681e-01 1.67966485e-01 4.34366733e-01
-6.27799451e-01 6.08034849e-01 4.46081668e-01 4.07596707e-01
-1.30356741e+00 1.73380777e-01 9.36894655e-01 -1.99723914e-01
3.48150246e-02 5.19096911e-01 -4.43603963e-01 7.68087506e-01
1.06591120e-01 -2.62342811e-01 2.47511193e-01 -3.44540000e-01
-2.31195062e-01 -1.00085199e-01 1.27370632e+00 -1.44656420e+00
5.14558196e-01 2.02721834e-01 1.06706631e+00 -1.03497505e+00
9.50311899e-01 -8.16580653e-01 4.97981638e-01 6.02230608e-01
-3.65577564e-02 2.62044400e-01 3.08859259e-01 7.14754105e-01
2.63818234e-01 4.52756658e-02 6.37408495e-01 -5.28407574e-01
-8.91392589e-01 7.59074271e-01 2.22883970e-01 1.11778781e-01
1.20751870e+00 -6.53230369e-01 -7.22527206e-02 -1.34487987e-01
-1.18161094e+00 4.75521207e-01 3.69000584e-01 7.95410752e-01
2.24024400e-01 -1.44729805e+00 -8.74155939e-01 -3.11947782e-02
2.09738418e-01 -3.44384283e-01 2.45662943e-01 7.55156457e-01
1.37411788e-01 7.14780390e-01 -1.61151379e-01 -7.74056017e-01
-2.09483290e+00 5.85949123e-01 4.32350308e-01 -3.85270983e-01
-4.94691014e-01 8.79558086e-01 8.26978981e-02 -5.24317384e-01
6.61332309e-01 2.99928606e-01 -2.63553500e-01 3.07682008e-01
8.09605956e-01 6.51534438e-01 -3.22938681e-01 -1.12238669e+00
-7.20120490e-01 9.27668989e-01 2.33923290e-02 4.46233265e-02
9.36522305e-01 -4.21192586e-01 3.08323473e-01 3.64514478e-02
6.09806240e-01 -1.39774054e-01 -1.43130279e+00 -4.14936513e-01
3.01185757e-01 -6.62937343e-01 -5.29322565e-01 -8.56181383e-01
-6.80569053e-01 2.91251093e-01 1.06331301e+00 -2.39438429e-01
3.85757297e-01 2.05777660e-01 6.05446458e-01 2.87573516e-01
5.61039209e-01 -1.17491686e+00 2.90516615e-01 3.22016239e-01
6.95665061e-01 -1.51833677e+00 4.52765435e-01 -4.91839290e-01
-4.80888903e-01 6.19451284e-01 1.09061706e+00 3.80112886e-01
1.13196678e-01 2.34484881e-01 -1.70539424e-01 -2.59110987e-01
-3.72462533e-02 -8.04268777e-01 8.26694191e-01 8.71382296e-01
2.79202819e-01 8.45118910e-02 2.71077275e-01 4.26471293e-01
-4.41304557e-02 -1.26014873e-01 -3.11653256e-01 7.31196463e-01
-3.13720018e-01 -1.08712888e+00 -1.25436652e+00 2.62564451e-01
-3.00627500e-01 3.85640293e-01 -5.08239508e-01 8.28445196e-01
3.58075619e-01 8.96462202e-01 -4.68587838e-02 7.21151009e-02
6.15006268e-01 1.77939132e-01 8.72476876e-01 -4.57639366e-01
-8.89525056e-01 3.46991234e-02 2.97461927e-01 -5.28206170e-01
-7.61898756e-01 -1.06203640e+00 -5.80640376e-01 -3.86762619e-01
-3.10991794e-01 -2.59859741e-01 3.38628650e-01 8.92478645e-01
-4.35029753e-02 3.72506440e-01 -1.92328662e-01 -1.21700966e+00
-3.90493095e-01 -9.59275246e-01 -4.31514084e-01 8.28206599e-01
1.44751623e-01 -9.76501524e-01 3.84394348e-01 3.99890207e-02] | [7.0065155029296875, -0.9945905804634094] |
962e0184-8c9a-4838-ab9f-37ae642f725f | improved-segmentation-of-deep-sulci-in | 2303.00795 | null | https://arxiv.org/abs/2303.00795v2 | https://arxiv.org/pdf/2303.00795v2.pdf | Improved Segmentation of Deep Sulci in Cortical Gray Matter Using a Deep Learning Framework Incorporating Laplace's Equation | When developing tools for automated cortical segmentation, the ability to produce topologically correct segmentations is important in order to compute geometrically valid morphometry measures. In practice, accurate cortical segmentation is challenged by image artifacts and the highly convoluted anatomy of the cortex itself. To address this, we propose a novel deep learning-based cortical segmentation method in which prior knowledge about the geometry of the cortex is incorporated into the network during the training process. We design a loss function which uses the theory of Laplace's equation applied to the cortex to locally penalize unresolved boundaries between tightly folded sulci. Using an ex vivo MRI dataset of human medial temporal lobe specimens, we demonstrate that our approach outperforms baseline segmentation networks, both quantitatively and qualitatively. | ['Ranjit Ittyerah', 'Paul A. Yushkevich', 'Ricardo Insausti', 'David A. Wolk', 'Marta Córcoles Parada', 'Carlos de la Rosa-Prieto', 'José Carlos Delgado González', 'Sandra Cebada-Sánchez', 'Maria del Pilar Marcos Rabal', 'Francisco Javier Molina Romero', 'Monica Muñoz', 'Maria del Mar Arroyo Jiménez', 'Maria Mercedes Iñiguez de Onzono Martin', 'Emilio Artacho-Pérula', 'Gabor Mizsei', 'Winifred Trotman', 'David J. Irwin', 'John A. Detre', 'Karthik Prabhakaran', 'M. Dylan Tisdall', 'Edward B. Lee', 'John Q. Trojanowski', 'Murray Grossman', 'Terry Schuck', 'John L. Robinson', 'Madigan L. Bedard', 'Laura E. M. Wisse', 'Pulkit Khandelwal', 'Sandhitsu Das', 'Long Xie', 'Sydney Lim', 'Sadhana Ravikumar'] | 2023-03-01 | null | null | null | null | ['anatomy'] | ['miscellaneous'] | [ 1.10655539e-01 4.50282097e-01 4.06892478e-01 -4.72527176e-01
-2.94318676e-01 -6.30736649e-01 2.41711140e-01 2.28344768e-01
-6.58609450e-01 6.34752750e-01 -8.42523426e-02 -3.18559617e-01
1.53696030e-01 -6.67759120e-01 -7.99098730e-01 -2.74713188e-01
-2.76349306e-01 5.75346649e-01 2.71415114e-01 1.31985158e-01
4.31867987e-01 7.40561366e-01 -6.90093756e-01 -2.69455820e-01
1.14142370e+00 8.18894804e-01 2.53771424e-01 3.69916737e-01
6.25698566e-02 2.24511474e-01 -2.09424630e-01 -2.83610076e-01
4.44184303e-01 -5.05537450e-01 -1.11295748e+00 1.24424905e-01
5.36626220e-01 -2.28643239e-01 8.47438723e-02 1.23797679e+00
-2.64496845e-03 -8.10764357e-02 1.01642060e+00 -6.35435462e-01
-1.60664305e-01 4.01262611e-01 -4.89168763e-01 3.89880389e-01
-2.35625520e-01 -3.17552909e-02 7.91676879e-01 -8.35646749e-01
8.75830889e-01 7.63709605e-01 6.43780828e-01 3.22580963e-01
-1.68083966e+00 -4.91793066e-01 1.35716036e-01 -1.88594356e-01
-1.29351950e+00 -3.95235777e-01 7.62637198e-01 -9.52190220e-01
5.72912693e-01 -2.16011286e-01 1.00031769e+00 2.93540895e-01
5.62887490e-01 4.84298497e-01 1.16828394e+00 -2.00737104e-01
5.58151364e-01 -4.10378963e-01 -2.39319149e-02 7.16692626e-01
2.23696485e-01 -2.39808023e-01 8.02640095e-02 -2.70910971e-02
1.41823208e+00 -3.62100452e-01 -7.08778352e-02 -6.82744980e-01
-9.58797276e-01 6.76296711e-01 7.73422241e-01 4.31603372e-01
-3.11177135e-01 1.86939687e-01 1.10742435e-01 -7.53361061e-02
4.46687788e-01 7.14891016e-01 -2.91057199e-01 2.29468301e-01
-1.22838104e+00 2.47278914e-01 5.23209810e-01 3.74265701e-01
7.16063201e-01 -8.52681771e-02 3.99762392e-01 7.85030782e-01
5.18102229e-01 -5.45521416e-02 2.08887890e-01 -1.31788576e+00
2.01024845e-01 7.64270008e-01 -2.54197836e-01 -7.55955815e-01
-5.25609791e-01 -3.01321536e-01 -8.61122429e-01 5.16832530e-01
8.04384530e-01 -1.83515519e-01 -1.13522363e+00 1.64573586e+00
3.60894173e-01 -3.40737283e-01 -4.65382993e-01 9.80198681e-01
1.24295846e-01 -3.09134778e-02 1.20853893e-01 -6.12175837e-02
9.80400741e-01 -4.12827283e-01 -2.15614632e-01 -4.63043749e-01
4.26328838e-01 -5.86700797e-01 9.60462809e-01 1.00824617e-01
-1.28291309e+00 -1.59167144e-02 -1.11107147e+00 -1.73033878e-01
-1.42065316e-01 -1.42364785e-01 4.82781082e-01 3.66107821e-01
-1.18665433e+00 7.41271079e-01 -1.12184143e+00 -1.17105745e-01
1.01266038e+00 6.97225511e-01 -5.12163758e-01 3.45930815e-01
-5.77320457e-01 9.09291327e-01 1.08234592e-01 -4.24092747e-02
-5.79145432e-01 -9.04316723e-01 -7.66114354e-01 1.06416464e-01
-3.96583602e-02 -6.51661336e-01 1.24774337e+00 -8.54810238e-01
-1.26430213e+00 1.09185612e+00 -4.01405506e-02 -3.75318319e-01
8.63090277e-01 8.06566775e-02 3.62899572e-01 4.83696967e-01
2.09251661e-02 8.88561010e-01 4.99439359e-01 -1.10358274e+00
3.21724638e-02 -7.01547623e-01 -2.52061933e-01 -3.11045721e-02
2.95775861e-01 -2.22267285e-01 -5.50989136e-02 -4.56721574e-01
6.45138204e-01 -7.55269647e-01 -6.02095842e-01 4.04178500e-01
-4.34388399e-01 2.04139218e-01 5.96304178e-01 -1.02183092e+00
5.95813394e-01 -2.09653473e+00 1.90015092e-01 6.47095144e-01
6.29989147e-01 -1.55192524e-01 2.19688997e-01 -1.09961994e-01
3.36645804e-02 1.78660795e-01 -7.18299031e-01 -1.84813932e-01
-3.33124906e-01 -1.15151353e-01 9.71901938e-02 7.55784810e-01
6.19716495e-02 8.83583367e-01 -6.65551782e-01 -6.92183793e-01
5.96530586e-02 3.88963997e-01 -6.99390352e-01 -2.13726703e-02
-3.21638286e-01 7.57572949e-01 -4.14724201e-01 3.29271764e-01
6.36091352e-01 -3.30271661e-01 3.53390276e-01 1.33380620e-02
-1.36398226e-01 1.10982031e-01 -7.05446005e-01 1.79657364e+00
-3.33137363e-01 6.00660563e-01 5.05737841e-01 -1.08926797e+00
6.29581094e-01 1.75483391e-01 7.23677516e-01 -5.64198732e-01
5.27621329e-01 4.72227842e-01 3.38710457e-01 -5.97658083e-02
-2.00016990e-01 -4.82767284e-01 3.14871937e-01 4.41746235e-01
5.10537066e-02 -6.36436462e-01 1.65204108e-01 2.30232760e-01
1.03359675e+00 6.37733936e-02 7.54047707e-02 -8.52027595e-01
2.64684439e-01 8.62216428e-02 5.19690335e-01 1.02329560e-01
-2.09528416e-01 9.02318478e-01 7.41470635e-01 -5.82829654e-01
-1.47057378e+00 -1.46824694e+00 -5.22105277e-01 3.40486079e-01
-1.10697299e-01 1.47941396e-01 -1.26391208e+00 -5.44434845e-01
-6.85841441e-02 3.15444261e-01 -6.62185431e-01 -4.62703174e-03
-7.15446174e-01 -5.82818806e-01 3.50809187e-01 4.21658516e-01
4.74702716e-01 -9.89412904e-01 -9.15928900e-01 3.29323202e-01
-8.64387900e-02 -1.12494946e+00 -5.75085521e-01 5.74645996e-02
-1.20753801e+00 -1.36686349e+00 -8.74695718e-01 -1.03976727e+00
1.19557059e+00 -3.36315811e-01 9.24216092e-01 2.57280946e-01
-6.62331820e-01 -5.26982173e-02 1.84765965e-01 -4.13280092e-02
-3.47556055e-01 1.32486716e-01 -2.32211411e-01 -3.25240880e-01
-4.88456562e-02 -1.02819455e+00 -8.04093122e-01 2.24841520e-01
-8.63878846e-01 2.57134408e-01 5.37978530e-01 4.72184598e-01
6.97039962e-01 -2.90193617e-01 3.84143829e-01 -7.86045671e-01
5.62847793e-01 -2.50144362e-01 -9.17840838e-01 8.21831450e-02
-5.19868076e-01 1.72060817e-01 6.15672469e-01 -3.02827656e-01
-7.10917294e-01 2.83379853e-01 -2.64456332e-01 3.21847424e-02
-5.49660400e-02 5.51914394e-01 8.57248604e-02 -6.02584422e-01
4.79516327e-01 3.03810854e-02 2.78600812e-01 -4.14164066e-01
7.59636983e-02 1.08564153e-01 6.97575688e-01 -5.55641949e-01
4.98757660e-01 7.17216194e-01 2.94891030e-01 -8.65383089e-01
-4.05047745e-01 -3.04579511e-02 -1.14974356e+00 -3.66243452e-01
9.07302320e-01 -2.28412554e-01 -6.12194240e-01 3.56966257e-01
-1.10009265e+00 -7.86704779e-01 -1.27813071e-01 3.57096314e-01
-8.22651148e-01 2.43634075e-01 -7.43914783e-01 -2.81930000e-01
-4.00486112e-01 -1.33572328e+00 8.14972460e-01 9.34797451e-02
-3.58708858e-01 -1.13287675e+00 1.82602242e-01 1.74254432e-01
3.65724653e-01 4.05626923e-01 1.26111174e+00 -3.13912213e-01
-6.20293200e-01 3.15341838e-02 -3.12257707e-01 3.20708275e-01
5.30118421e-02 1.03956081e-01 -5.61758101e-01 3.79258133e-02
-4.21664789e-02 -1.54063091e-01 7.78373837e-01 7.26020634e-01
1.10222054e+00 -1.21017590e-01 -3.35313588e-01 7.54665017e-01
1.43880582e+00 1.08072832e-01 5.73041499e-01 2.29067117e-01
5.26580393e-01 7.96227455e-01 -2.14584142e-01 -8.42384808e-03
3.28784972e-01 2.29340568e-01 1.02767244e-01 -1.46150246e-01
-2.22085506e-01 -6.48052394e-02 -1.79496661e-01 6.43669665e-01
-9.14473087e-02 4.02458847e-01 -1.23143411e+00 7.42591739e-01
-1.35389924e+00 -4.05045539e-01 -7.96037167e-02 2.27142072e+00
1.02121890e+00 2.88288772e-01 9.14734527e-02 -2.13219393e-02
4.27558005e-01 -3.49126369e-01 -7.23276138e-01 -2.83129960e-01
1.47663817e-01 4.74596947e-01 4.11567092e-01 8.94444525e-01
-9.49615061e-01 9.79607999e-01 7.30589056e+00 1.68232813e-01
-1.29627109e+00 5.65426052e-02 8.54772985e-01 1.36862025e-01
-2.26105914e-01 -9.81408656e-02 -1.70877889e-01 3.62329811e-01
4.37261611e-01 -1.20743662e-01 4.79903758e-01 4.36639398e-01
2.91239500e-01 -4.81050670e-01 -9.01909292e-01 5.03407598e-01
-3.80049258e-01 -1.44198966e+00 -2.75700122e-01 4.19128358e-01
8.18920016e-01 1.81199655e-01 1.28550874e-02 -3.61845642e-01
4.28219914e-01 -1.24977064e+00 6.85554266e-01 4.60818261e-01
7.32761741e-01 -6.49618030e-01 3.52987826e-01 1.49888381e-01
-9.53310609e-01 2.74578214e-01 -3.15129161e-01 3.21052223e-02
1.81439906e-01 8.24339449e-01 -1.08880782e+00 -1.32964715e-01
3.86749595e-01 1.52260587e-01 -5.36052644e-01 1.57253957e+00
-8.09345543e-02 4.82431084e-01 -5.92772007e-01 4.00711149e-01
2.11863160e-01 -5.35597086e-01 3.15065712e-01 9.19585764e-01
5.05190983e-04 9.32853445e-02 -3.50648090e-02 1.44279921e+00
-1.88661471e-01 3.83420527e-01 -4.53808397e-01 1.27398213e-02
1.93399161e-01 1.28219998e+00 -1.49412823e+00 -1.59471538e-02
-1.83903530e-01 7.95271873e-01 6.82102859e-01 4.17532235e-01
-1.80022880e-01 -4.25821133e-02 5.14752209e-01 5.80475628e-01
7.35342577e-02 -7.52676070e-01 -9.92519736e-01 -9.38257098e-01
1.89343512e-01 -2.97949910e-01 -6.93292916e-02 -4.51265872e-01
-9.61008549e-01 4.69450712e-01 -1.19154565e-01 -4.18064535e-01
1.58558483e-03 -5.20687044e-01 -8.60245585e-01 8.34802151e-01
-1.17943037e+00 -8.37697983e-01 -5.20790555e-02 2.69501269e-01
1.99911401e-01 2.86850572e-01 4.24956292e-01 3.19710970e-01
-2.77106524e-01 1.85797215e-01 -1.39555767e-01 3.98795128e-01
3.05942953e-01 -1.34850276e+00 5.90896606e-01 8.21737111e-01
-6.37290925e-02 8.51880014e-01 8.20024252e-01 -8.88013661e-01
-8.78177404e-01 -8.86621535e-01 7.00564861e-01 8.17682445e-02
7.34319985e-01 -4.97686982e-01 -1.04766619e+00 7.80644476e-01
-9.04745460e-02 1.68343276e-01 5.00309110e-01 -7.88928941e-02
-1.95848942e-01 1.14146583e-01 -1.32034266e+00 4.50613260e-01
6.37508869e-01 -2.32666045e-01 -5.25595546e-01 3.04307282e-01
1.23709939e-01 -1.86363533e-01 -1.06579971e+00 3.47319245e-01
7.06916571e-01 -8.00530851e-01 9.29991603e-01 -2.89992034e-01
4.68633354e-01 2.17475626e-03 2.71853566e-01 -1.40228605e+00
-1.88817799e-01 -2.59639859e-01 2.67549545e-01 6.18707955e-01
4.82928455e-01 -4.34264779e-01 1.11220598e+00 9.30154860e-01
-2.26428092e-01 -8.99800956e-01 -1.13300872e+00 -7.56580174e-01
8.77293229e-01 -9.15158615e-02 9.22405943e-02 7.75134265e-01
1.46274284e-01 1.82364266e-02 5.49912333e-01 -1.42863825e-01
9.39613879e-01 -1.47508025e-01 2.23866135e-01 -1.51773882e+00
2.03596607e-01 -1.01806068e+00 -4.10607159e-01 -8.46430659e-01
1.68434575e-01 -8.73752415e-01 2.03036696e-01 -1.67252731e+00
1.82727441e-01 -5.98306656e-01 1.81412920e-01 3.52341264e-01
1.52213484e-01 3.37463319e-01 -4.26746793e-02 2.13182390e-01
-1.51017383e-01 4.15937811e-01 1.54234743e+00 1.17501937e-01
-2.38689214e-01 -2.19017655e-01 -4.13716137e-01 1.06614411e+00
9.63964164e-01 -3.85060251e-01 -2.50781119e-01 -4.64882106e-01
1.70615211e-01 -1.36004552e-01 5.54836333e-01 -1.08840334e+00
3.40075076e-01 9.93035212e-02 5.90560675e-01 -2.26369888e-01
1.71499729e-01 -7.64820933e-01 -1.63993970e-01 5.70601165e-01
-2.64120787e-01 -1.33106500e-01 9.88871679e-02 1.46107869e-02
-7.41914883e-02 -6.27648979e-02 1.37933874e+00 -3.55286688e-01
-3.59565541e-02 4.31069106e-01 -6.34993434e-01 3.29541296e-01
9.33552384e-01 -1.63558245e-01 7.77315646e-02 -4.04049270e-02
-8.59866261e-01 3.97648104e-02 8.56596708e-01 -2.08459482e-01
6.89457953e-01 -1.03098249e+00 -5.22064626e-01 2.96899557e-01
-2.88270444e-01 2.76175022e-01 -2.50861645e-02 1.05256677e+00
-1.15929067e+00 1.29442707e-01 -4.67820615e-01 -4.06531215e-01
-8.77699018e-01 2.09369659e-01 8.36733282e-01 -7.64082149e-02
-7.96651006e-01 8.09342682e-01 3.00841928e-01 -4.71744061e-01
-1.86249986e-03 -5.73382676e-01 -2.68969536e-02 -3.15329343e-01
2.85633415e-01 1.00498065e-01 6.48920015e-02 -6.31942213e-01
-3.58830452e-01 5.19650280e-01 -8.14239681e-02 -3.63553673e-01
1.35972428e+00 -6.92768916e-02 -4.11156863e-01 2.62877077e-01
1.22379196e+00 -1.67707935e-01 -1.55605197e+00 -5.02064563e-02
2.60007560e-01 -2.27261797e-01 2.48471111e-01 -8.86509538e-01
-1.35454965e+00 9.05050635e-01 3.77850115e-01 8.35114419e-02
7.22393870e-01 1.06348589e-01 8.93052280e-01 -3.43331210e-02
3.94380271e-01 -1.09795392e+00 -7.65152648e-02 4.55140144e-01
7.63685405e-01 -9.88957942e-01 3.43329124e-02 -4.50168669e-01
-1.82070255e-01 1.03610063e+00 6.88404739e-01 -5.29865384e-01
1.00829792e+00 4.30387169e-01 1.20028012e-01 -4.90661055e-01
-1.53501213e-01 1.60738513e-01 3.97708416e-01 2.59730309e-01
5.56077957e-01 -1.08528659e-01 -3.38565975e-01 2.55228817e-01
-4.30985898e-01 -1.45339787e-01 4.05736595e-01 9.18911874e-01
-6.45309210e-01 -1.02401960e+00 -2.49238372e-01 6.76875412e-01
-6.07929468e-01 -5.00671193e-03 -6.25273287e-01 6.29483581e-01
7.74262697e-02 4.37068671e-01 2.65365869e-01 1.39155030e-01
3.78981829e-02 -6.59073964e-02 8.58004034e-01 -6.50023341e-01
-3.91432017e-01 2.55052477e-01 -4.35175478e-01 -4.05430466e-01
-1.99418366e-02 -8.57332408e-01 -1.73651254e+00 -1.17060035e-01
1.80463955e-01 -3.05223428e-02 8.81070197e-01 1.19253051e+00
1.67641088e-01 3.18277001e-01 2.04970717e-01 -9.73397017e-01
-9.55441222e-02 -8.26283336e-01 -7.86789775e-01 4.18275118e-01
2.92639315e-01 -6.85487092e-01 -2.29895502e-01 1.53116494e-01] | [14.246713638305664, -2.9852421283721924] |
f2382584-0854-4296-bdef-b8318c05cfe2 | source-free-domain-adaptation-for-semantic | 2103.16372 | null | https://arxiv.org/abs/2103.16372v1 | https://arxiv.org/pdf/2103.16372v1.pdf | Source-Free Domain Adaptation for Semantic Segmentation | Unsupervised Domain Adaptation (UDA) can tackle the challenge that convolutional neural network(CNN)-based approaches for semantic segmentation heavily rely on the pixel-level annotated data, which is labor-intensive. However, existing UDA approaches in this regard inevitably require the full access to source datasets to reduce the gap between the source and target domains during model adaptation, which are impractical in the real scenarios where the source datasets are private, and thus cannot be released along with the well-trained source models. To cope with this issue, we propose a source-free domain adaptation framework for semantic segmentation, namely SFDA, in which only a well-trained source model and an unlabeled target domain dataset are available for adaptation. SFDA not only enables to recover and preserve the source domain knowledge from the source model via knowledge transfer during model adaptation, but also distills valuable information from the target domain for self-supervised learning. The pixel- and patch-level optimization objectives tailored for semantic segmentation are seamlessly integrated in the framework. The extensive experimental results on numerous benchmark datasets highlight the effectiveness of our framework against the existing UDA approaches relying on source data. | ['Jun Wang', 'Wei zhang', 'Yuang Liu'] | 2021-03-30 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Liu_Source-Free_Domain_Adaptation_for_Semantic_Segmentation_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Liu_Source-Free_Domain_Adaptation_for_Semantic_Segmentation_CVPR_2021_paper.pdf | cvpr-2021-1 | ['source-free-domain-adaptation'] | ['computer-vision'] | [ 4.48983580e-01 6.47323951e-02 -3.35597575e-01 -4.94536430e-01
-8.01333368e-01 -5.32511830e-01 1.91849917e-01 -1.80276502e-02
-5.52801132e-01 7.43125439e-01 -2.46525601e-01 1.58018440e-01
1.51766360e-01 -9.85318899e-01 -9.14515078e-01 -8.91172767e-01
6.00688756e-01 4.91695702e-01 5.47689021e-01 -1.07406244e-01
-1.34318709e-01 1.95104912e-01 -1.20761037e+00 -3.79922278e-02
1.34069717e+00 1.25325739e+00 6.25729561e-01 9.26122218e-02
-5.32616556e-01 3.88120890e-01 -4.44769442e-01 -1.83199614e-01
2.75589615e-01 -4.13393736e-01 -7.93158233e-01 3.15690279e-01
1.45282313e-01 -1.71960011e-01 -6.19533136e-02 1.25571167e+00
3.79332721e-01 8.86137411e-02 3.28136474e-01 -1.12264192e+00
-7.10380971e-01 2.76607037e-01 -5.54010212e-01 -7.47099221e-02
-2.68522412e-01 1.58795327e-01 6.35623991e-01 -6.76688373e-01
8.76364887e-01 9.17906761e-01 5.75681984e-01 6.66068554e-01
-1.01487076e+00 -6.46184027e-01 5.46306670e-01 5.25387712e-02
-1.33046126e+00 -3.15719515e-01 1.25919735e+00 -3.90674025e-01
2.33728454e-01 -1.84892371e-01 6.60262287e-01 1.06037056e+00
-4.46118921e-01 9.96878028e-01 9.54267025e-01 -3.10244322e-01
3.84311199e-01 3.82401675e-01 1.67015940e-02 5.04930377e-01
1.04587205e-01 -1.41473085e-01 -4.88212049e-01 1.62804618e-01
9.51815069e-01 7.09399581e-02 -1.69798538e-01 -9.17142689e-01
-1.24613202e+00 6.62649751e-01 5.81759155e-01 2.05580652e-01
-3.78041208e-01 -4.66610610e-01 4.28213626e-01 1.10305786e-01
6.19886160e-01 1.03180081e-01 -8.62697542e-01 2.06388280e-01
-8.97191703e-01 -1.55958300e-03 4.92524415e-01 1.37262678e+00
1.18501723e+00 2.00446453e-02 2.57261898e-02 1.06311047e+00
2.86959112e-01 5.61371505e-01 6.44612193e-01 -7.07859337e-01
5.65227389e-01 9.94286180e-01 1.10795185e-01 -6.89270020e-01
-8.16099346e-02 -6.20172024e-01 -8.09352577e-01 3.69153847e-03
6.07638001e-01 -9.08366218e-02 -1.08438718e+00 1.81413043e+00
8.13129604e-01 2.77442843e-01 3.43565762e-01 8.71040523e-01
9.08341408e-01 5.57995498e-01 1.78854421e-01 -1.82789579e-01
9.93209839e-01 -1.32784951e+00 -5.18766642e-01 -5.06936729e-01
4.74651128e-01 -6.44429624e-01 1.23918426e+00 -3.24459672e-02
-8.51498067e-01 -7.56918490e-01 -1.07109141e+00 -1.26332343e-01
-4.11497563e-01 1.03828363e-01 3.54639262e-01 4.20582443e-01
-5.99396944e-01 3.27594459e-01 -8.07883143e-01 -3.30472738e-01
1.10731339e+00 2.55511314e-01 -4.27465260e-01 -3.38405252e-01
-1.13432693e+00 3.47909421e-01 7.93318689e-01 2.12485299e-01
-1.08934951e+00 -7.50549376e-01 -8.47871482e-01 -1.94442615e-01
6.48761094e-01 -5.50960958e-01 1.12404013e+00 -1.50949597e+00
-1.50113785e+00 9.45124090e-01 -3.81552912e-02 -1.77964583e-01
6.66694760e-01 -4.58815806e-02 -2.61916250e-01 9.70653296e-02
4.26758468e-01 7.00195968e-01 8.60723138e-01 -1.42293704e+00
-8.27484071e-01 -4.30248916e-01 -1.09431781e-02 2.81477898e-01
-5.04088521e-01 -4.68220383e-01 -7.94312716e-01 -5.95955729e-01
2.14688808e-01 -7.22917795e-01 -3.15432340e-01 2.61473060e-01
-3.35958391e-01 3.27194668e-02 1.11249804e+00 -6.63469791e-01
8.60675275e-01 -2.22620702e+00 1.53189152e-01 8.46142620e-02
-2.82939952e-02 5.55243373e-01 -2.39867970e-01 -8.11902434e-02
1.07616343e-01 -2.59329647e-01 -7.74050891e-01 -2.23703146e-01
-1.40296027e-01 4.47768092e-01 -9.73779410e-02 2.53957480e-01
2.61209995e-01 8.46992314e-01 -1.02509880e+00 -9.05795395e-01
2.03730464e-01 3.72049093e-01 -3.75040978e-01 5.03045738e-01
-6.76149666e-01 1.09159505e+00 -9.06644344e-01 7.90737271e-01
9.48751569e-01 -3.16436261e-01 -9.19463933e-02 -1.86363518e-01
1.58173338e-01 -1.03865169e-01 -1.16733921e+00 2.35741067e+00
-3.99295509e-01 8.12680647e-02 2.54085153e-01 -1.25245976e+00
1.08048415e+00 1.52567223e-01 5.61076343e-01 -8.16522419e-01
2.13252157e-02 4.74669516e-01 -3.00481439e-01 -4.36978549e-01
1.36123464e-01 -1.52067780e-01 -1.15100957e-01 2.07918376e-01
3.12502235e-01 5.44068310e-03 7.45539293e-02 3.61451507e-02
5.63625574e-01 6.51368618e-01 -2.69294027e-02 -9.01112109e-02
7.05352366e-01 4.35251772e-01 1.22020006e+00 3.90439361e-01
-3.38446319e-01 7.19418347e-01 2.71911979e-01 -3.47984433e-01
-1.14196455e+00 -9.70586240e-01 -2.09442303e-01 9.36286569e-01
5.16400814e-01 1.36652246e-01 -1.12349451e+00 -1.05909133e+00
-2.49072269e-01 3.13154072e-01 -5.80331802e-01 -2.37351522e-01
-3.91510487e-01 -6.53560519e-01 2.35731080e-01 7.00765967e-01
1.00788331e+00 -1.09058034e+00 -2.51531273e-01 3.55797112e-01
-3.19930971e-01 -1.22505891e+00 -5.68972647e-01 8.16580281e-02
-1.06614935e+00 -9.79296982e-01 -1.04898334e+00 -1.03649461e+00
9.54825759e-01 1.84404254e-01 8.74983430e-01 -9.06338766e-02
8.87218956e-03 1.37206912e-01 -4.03190017e-01 -3.74180377e-01
-3.65346253e-01 3.51325452e-01 -3.26612830e-01 3.85013819e-01
3.21983397e-01 -6.59473717e-01 -5.48032761e-01 4.43948984e-01
-1.10568249e+00 2.56232738e-01 5.36021531e-01 8.98692012e-01
1.18258905e+00 1.75856158e-01 8.78257036e-01 -1.30104995e+00
-3.78397740e-02 -5.76538503e-01 -7.14475393e-01 3.10021609e-01
-5.77763736e-01 -2.46644065e-01 6.31962538e-01 -4.87895429e-01
-1.44672215e+00 5.80105126e-01 -6.47964329e-02 -6.07069194e-01
-3.88259232e-01 4.99616265e-01 -9.36704993e-01 -1.07254364e-01
5.81148088e-01 4.24801528e-01 -1.65330917e-02 -6.96909904e-01
3.95617336e-01 5.98316550e-01 7.49044657e-01 -6.61640584e-01
9.06921864e-01 5.52842438e-01 -3.17749560e-01 -4.06862229e-01
-1.07250452e+00 -4.84877348e-01 -1.15134025e+00 -3.90612595e-02
9.49597299e-01 -1.20304823e+00 2.30983332e-01 9.85375822e-01
-1.03074169e+00 -4.87072945e-01 -5.80187261e-01 2.42789790e-01
-5.58179319e-01 3.28857154e-01 -1.18650019e-01 -3.69180828e-01
-1.14999786e-01 -1.21256971e+00 9.98782277e-01 6.64965749e-01
3.48109215e-01 -1.23203611e+00 -1.59744889e-01 6.19674683e-01
2.20328122e-01 4.27102864e-01 9.06081617e-01 -7.69955993e-01
-8.06788564e-01 -1.75697222e-01 -4.02837247e-01 6.01785421e-01
4.37810868e-01 -4.32611287e-01 -1.04086959e+00 -2.48744503e-01
-8.40087608e-02 -4.25061971e-01 6.59550548e-01 2.98168689e-01
1.14598680e+00 -1.19212136e-01 -3.34329933e-01 8.04263949e-01
1.45507431e+00 1.76799521e-02 3.60410124e-01 5.38952887e-01
1.10123181e+00 6.90845788e-01 1.04742348e+00 2.87598193e-01
5.74514389e-01 4.88943160e-01 4.21272665e-01 -5.02235413e-01
-2.73282081e-01 -6.17758274e-01 -6.23925440e-02 7.31459439e-01
2.43556693e-01 -7.88652971e-02 -8.94679606e-01 9.54418302e-01
-1.98638916e+00 -3.34353447e-01 -3.32117006e-02 2.13389444e+00
1.13463449e+00 2.61845272e-02 1.28586844e-01 -1.44562498e-01
7.89092898e-01 5.19606434e-02 -1.23597109e+00 2.91561812e-01
-2.42371619e-01 -6.00665845e-02 5.63914716e-01 1.71937361e-01
-1.27991211e+00 1.18973160e+00 5.21256542e+00 1.07124615e+00
-1.12088382e+00 4.68435079e-01 5.79915047e-01 1.56668797e-01
-2.91157603e-01 -1.19368387e-02 -6.18633807e-01 6.61833525e-01
6.62574828e-01 -1.32762268e-01 1.47276312e-01 1.23934615e+00
9.12571047e-03 -3.67076471e-02 -9.03487980e-01 8.07773113e-01
-2.16426790e-01 -1.21727061e+00 -1.33330882e-01 -6.49379119e-02
1.08337057e+00 2.29088757e-02 -1.14574492e-01 1.57044724e-01
3.11487347e-01 -5.52441776e-01 6.20327532e-01 4.12319213e-01
8.25803161e-01 -7.67104447e-01 7.12897003e-01 5.29178202e-01
-1.20427513e+00 8.33449513e-02 -6.01913691e-01 4.22625154e-01
8.19870606e-02 7.74160326e-01 -5.29178143e-01 7.29514718e-01
9.00363147e-01 1.06530535e+00 -4.99758035e-01 9.56548095e-01
-2.47495234e-01 5.95360339e-01 -3.13217163e-01 5.49426973e-01
1.90935180e-01 -3.55256200e-01 3.32633376e-01 7.16693699e-01
1.36780724e-01 -1.38496280e-01 4.19021338e-01 9.92819369e-01
-2.42089808e-01 2.89390475e-01 -3.97348464e-01 -7.44471699e-02
4.12552536e-01 1.14021099e+00 -7.83701777e-01 -2.91674107e-01
-4.55586463e-01 1.11016560e+00 2.61117667e-01 5.20362496e-01
-6.69815063e-01 -2.99424440e-01 5.19727230e-01 5.44642396e-02
4.09687281e-01 -8.39060172e-02 -4.75812435e-01 -1.27036369e+00
3.65060478e-01 -7.64875054e-01 4.89999205e-01 -6.70164049e-01
-1.35497117e+00 5.26367962e-01 -9.13420692e-02 -1.47345138e+00
8.81577358e-02 -2.98318803e-01 -4.34937984e-01 9.75258529e-01
-2.07996440e+00 -1.73888338e+00 -5.41787624e-01 9.05110717e-01
6.17651165e-01 -2.71525234e-01 6.76537514e-01 5.21758020e-01
-8.02112758e-01 5.87636530e-01 4.02502924e-01 2.39547238e-01
7.79422820e-01 -1.02162981e+00 1.70299575e-01 8.77055347e-01
-3.35512720e-02 2.12466031e-01 2.38404155e-01 -6.29180431e-01
-1.03836215e+00 -1.44309127e+00 5.11524796e-01 -2.39420488e-01
4.37393844e-01 -2.77464211e-01 -1.45186079e+00 6.29183948e-01
-2.27663592e-01 4.56864983e-01 6.06426001e-01 -1.57015830e-01
-3.00229281e-01 -3.58971119e-01 -1.22435915e+00 1.78410813e-01
1.07925344e+00 -4.58072692e-01 -3.90354216e-01 4.28447902e-01
8.68131042e-01 -6.14236057e-01 -9.27087963e-01 3.65169495e-01
1.25901595e-01 -7.69248903e-01 9.57115948e-01 -4.38604146e-01
3.72313261e-01 -5.51041603e-01 6.85073212e-02 -1.19768858e+00
1.14268228e-01 -1.02778755e-01 2.62117922e-01 1.69381809e+00
3.60286236e-01 -6.07317388e-01 9.02373314e-01 6.74367607e-01
-2.34300151e-01 -4.75617588e-01 -1.02393961e+00 -7.94599056e-01
3.43725353e-01 -2.06819743e-01 9.00033593e-01 1.11485636e+00
-6.32816076e-01 3.97396721e-02 -2.06549838e-01 3.61108333e-01
6.83118761e-01 4.29150403e-01 7.54492164e-01 -1.35424912e+00
-4.44785841e-02 -2.04028815e-01 -2.00128868e-01 -1.06093609e+00
3.81834954e-01 -9.63201106e-01 2.90232241e-01 -1.55871046e+00
1.33914128e-01 -9.50391293e-01 -4.99634057e-01 6.71010792e-01
-2.10762873e-01 7.30269924e-02 -8.15192312e-02 5.41604936e-01
-6.90767586e-01 8.25258493e-01 1.50492537e+00 -2.37322330e-01
-2.72948980e-01 1.15742497e-02 -8.31492662e-01 8.04212034e-01
6.35647178e-01 -6.96357906e-01 -7.65433431e-01 -6.44882739e-01
-2.78873146e-01 -1.39053181e-01 4.12967443e-01 -9.75773931e-01
1.76605508e-01 -4.43439245e-01 3.18231970e-01 -3.84060413e-01
3.67200002e-03 -1.09202218e+00 2.04531774e-02 -2.89264973e-03
-1.99212536e-01 -6.97783113e-01 3.04619193e-01 8.47668171e-01
-4.82701123e-01 -2.56338924e-01 1.12396777e+00 -1.95935711e-01
-1.16358590e+00 6.90289855e-01 3.84535104e-01 1.58967704e-01
1.11791337e+00 -3.38934720e-01 -3.36371437e-02 1.06024183e-01
-7.37170935e-01 4.31222707e-01 8.14628839e-01 3.88143718e-01
2.78321654e-01 -1.39891171e+00 -4.63878214e-01 3.54080260e-01
4.63129580e-01 1.03959751e+00 6.86136127e-01 5.83512902e-01
-2.40246415e-01 4.26883623e-02 -4.41787601e-01 -6.83514655e-01
-7.56383240e-01 6.88937128e-01 3.30149174e-01 -1.94158107e-01
-5.14905632e-01 9.28061843e-01 5.09558618e-01 -8.32958162e-01
1.25707939e-01 2.62254209e-04 -5.10245673e-02 -4.08222042e-02
2.67511517e-01 -4.16282900e-02 -3.22045237e-02 -7.61180699e-01
-2.74516016e-01 7.04876959e-01 -1.23098686e-01 1.90785602e-01
1.36220729e+00 -5.17821908e-01 -4.93697030e-03 3.93364817e-01
1.12181711e+00 -3.25626284e-01 -1.93828094e+00 -8.67255032e-01
-1.46998048e-01 -4.71076429e-01 1.00561425e-01 -9.24072981e-01
-1.53623629e+00 1.13700092e+00 5.50531805e-01 -3.61650825e-01
1.41270125e+00 2.57758442e-02 1.22361183e+00 1.67846009e-01
4.16614354e-01 -1.45751810e+00 4.86343727e-03 1.23692513e-01
5.49516737e-01 -1.48497760e+00 -2.42949709e-01 -6.00128591e-01
-8.03942263e-01 8.69929433e-01 9.57827747e-01 1.07314125e-01
6.17940247e-01 -1.69762343e-01 3.22731942e-01 1.42857760e-01
-1.51068404e-01 -2.39585683e-01 1.01268239e-01 1.01448333e+00
-8.19882378e-02 -1.55502155e-01 1.34220451e-01 1.08731580e+00
2.01865658e-01 2.66312003e-01 1.40885577e-01 9.85705256e-01
-2.76675254e-01 -1.50870931e+00 -2.28134751e-01 9.13501978e-02
-7.78757930e-02 1.01235412e-01 -1.86994463e-01 6.55470073e-01
4.73079413e-01 6.59036875e-01 -1.87621504e-01 -9.90684330e-02
4.73345160e-01 9.77418572e-02 4.79689389e-02 -7.02051818e-01
-1.22943066e-01 2.31497824e-01 -3.43655050e-01 -3.56717288e-01
-6.75751448e-01 -6.19390488e-01 -1.43107450e+00 1.17430493e-01
-2.68028706e-01 2.55210791e-02 5.70194125e-01 1.03803980e+00
4.50567454e-01 5.34130692e-01 6.45475507e-01 -5.54916263e-01
-2.13005617e-01 -6.49140060e-01 -7.34448493e-01 4.79084790e-01
3.40459049e-01 -6.70239627e-01 3.80839892e-02 5.20777047e-01] | [9.69644832611084, 1.3753103017807007] |
0fec68e3-d4e7-488d-ad2f-fb914ca30a4b | all-you-need-is-a-second-look-towards-tighter | 2004.12436 | null | https://arxiv.org/abs/2004.12436v1 | https://arxiv.org/pdf/2004.12436v1.pdf | All you need is a second look: Towards Tighter Arbitrary shape text detection | Deep learning-based scene text detection methods have progressed substantially over the past years. However, there remain several problems to be solved. Generally, long curve text instances tend to be fragmented because of the limited receptive field size of CNN. Besides, simple representations using rectangle or quadrangle bounding boxes fall short when dealing with more challenging arbitrary-shaped texts. In addition, the scale of text instances varies greatly which leads to the difficulty of accurate prediction through a single segmentation network. To address these problems, we innovatively propose a two-stage segmentation based arbitrary text detector named \textit{NASK} (\textbf{N}eed \textbf{A} \textbf{S}econd loo\textbf{K}). Specifically, \textit{NASK} consists of a Text Instance Segmentation network namely \textit{TIS} (\(1^{st}\) stage), a Text RoI Pooling module and a Fiducial pOint eXpression module termed as \textit{FOX} (\(2^{nd}\) stage). Firstly, \textit{TIS} conducts instance segmentation to obtain rectangle text proposals with a proposed Group Spatial and Channel Attention module (\textit{GSCA}) to augment the feature expression. Then, Text RoI Pooling transforms these rectangles to the fixed size. Finally, \textit{FOX} is introduced to reconstruct text instances with a more tighter representation using the predicted geometrical attributes including text center line, text line orientation, character scale and character orientation. Experimental results on two public benchmarks including \textit{Total-Text} and \textit{SCUT-CTW1500} have demonstrated that the proposed \textit{NASK} achieves state-of-the-art results. | ['Yuexian Zou', 'Meng Cao'] | 2020-04-26 | null | null | null | null | ['scene-text-detection'] | ['computer-vision'] | [ 3.30310673e-01 1.12392195e-01 6.95211440e-02 -2.52096802e-01
-7.13445485e-01 -5.46807170e-01 5.35140634e-01 -1.16912812e-01
-3.21616381e-01 4.28795367e-01 -1.81299627e-01 -3.76341939e-01
1.38462186e-01 -5.94172537e-01 -7.07175851e-01 -7.14295208e-01
6.23071611e-01 4.41615433e-01 5.37344873e-01 -3.69192399e-02
4.55368668e-01 4.84737724e-01 -8.58562827e-01 3.59366953e-01
8.97528589e-01 1.34244800e+00 2.88012862e-01 5.67900121e-01
-5.05543232e-01 4.11926776e-01 -7.17685640e-01 -5.03833532e-01
3.89649779e-01 -2.89745867e-01 -4.59121197e-01 1.33723781e-01
4.67505991e-01 -6.09588563e-01 -6.40870094e-01 9.18761551e-01
4.59523916e-01 2.32915178e-01 9.61027920e-01 -9.74344909e-01
-4.46694702e-01 6.07912004e-01 -1.17404211e+00 2.41485119e-01
-1.44428432e-01 8.85031745e-02 9.21904206e-01 -1.05564845e+00
3.86058778e-01 1.04758799e+00 7.66277552e-01 2.89335698e-01
-9.12537932e-01 -8.11693788e-01 4.94711995e-01 -4.11495328e-01
-1.53109515e+00 -8.06341767e-02 7.66386926e-01 -3.32285255e-01
9.23945725e-01 4.93273526e-01 4.29208875e-01 9.35850739e-01
3.46997917e-01 1.40305579e+00 7.95559287e-01 -2.66807526e-01
-6.77844062e-02 9.37366933e-02 3.12390655e-01 7.44109094e-01
2.29802489e-01 -4.45863813e-01 -8.96369889e-02 3.03004086e-01
1.20697391e+00 1.58590704e-01 -1.87688991e-01 8.09712112e-02
-1.14743328e+00 8.50970387e-01 4.86740977e-01 3.99238020e-01
7.75782317e-02 1.51093125e-01 5.09737432e-01 -2.35768571e-01
4.03947443e-01 1.70591280e-01 -2.64433622e-01 2.14860588e-01
-1.23568892e+00 1.86597183e-01 3.67269009e-01 1.37713397e+00
5.42260408e-01 1.97499976e-01 -3.62188101e-01 1.11161458e+00
1.74592063e-01 6.33665919e-01 5.93156159e-01 -1.52579114e-01
1.14031291e+00 8.48601162e-01 -1.74788103e-01 -9.16540504e-01
-7.29290009e-01 -2.90361315e-01 -1.12645257e+00 -1.86764877e-02
5.48591793e-01 -3.65397066e-01 -1.24066699e+00 8.15658927e-01
9.05069150e-03 -1.86514258e-01 -3.74236196e-01 8.55073154e-01
1.00172472e+00 8.92844737e-01 1.55009627e-02 1.37653008e-01
1.41720414e+00 -1.06028008e+00 -3.36705059e-01 -4.30539221e-01
8.51269782e-01 -9.57858682e-01 1.00633848e+00 4.70257759e-01
-1.03863919e+00 -5.64123392e-01 -9.48636591e-01 -3.76609296e-01
-5.81972420e-01 7.10634291e-01 3.85678083e-01 7.12552786e-01
-8.74831140e-01 4.48531471e-02 -6.52615070e-01 -3.59821081e-01
6.32456005e-01 7.51280189e-01 1.66466132e-01 1.77261919e-01
-7.02431083e-01 2.29484409e-01 6.49724483e-01 5.26287973e-01
-5.33125997e-01 -2.47707054e-01 -7.50519454e-01 1.09626815e-01
5.48500240e-01 -2.93758094e-01 8.19113374e-01 -8.36415827e-01
-1.47544265e+00 7.92436004e-01 8.56131539e-02 -1.58767596e-01
7.67212391e-01 3.02800816e-02 -2.16778010e-01 9.84887481e-02
2.23186880e-01 8.20399880e-01 1.17638481e+00 -1.05233598e+00
-8.46948564e-01 -4.85398471e-01 -2.33445928e-01 3.19669068e-01
-3.23888004e-01 1.09950542e-01 -9.84064996e-01 -1.20690954e+00
4.19880837e-01 -7.68524647e-01 -7.14895278e-02 -6.51220903e-02
-1.14243650e+00 -3.17777932e-01 1.15150297e+00 -6.87452257e-01
1.42328215e+00 -2.03649974e+00 1.79697275e-02 2.00659066e-01
4.09143835e-01 3.38510603e-01 3.05356026e-01 3.39180864e-02
-6.24572076e-02 3.72069925e-01 -1.18225977e-01 -4.18383598e-01
-3.87990698e-02 -2.00332269e-01 -3.03329468e-01 4.08506185e-01
5.06515503e-02 1.09974110e+00 -3.36652584e-02 -8.52477789e-01
7.05755472e-01 1.36493564e-01 -3.76972020e-01 -2.63510197e-01
-4.48703676e-01 1.48188815e-01 -1.25285816e+00 6.73966467e-01
8.86678517e-01 -3.23272318e-01 -2.47775868e-01 -1.57549739e-01
-3.51805776e-01 -4.37596768e-01 -1.27662885e+00 1.34450793e+00
1.27789974e-02 7.71281064e-01 3.12708877e-02 -1.19712400e+00
1.26064897e+00 -4.30479050e-02 5.61044633e-01 -6.30100191e-01
7.46887803e-01 1.19476974e-01 -3.35413098e-01 -3.96153063e-01
8.51541758e-01 2.89048642e-01 -3.19938898e-01 3.58010620e-01
-2.14646548e-01 -3.40236574e-01 -1.36535972e-01 3.90694626e-02
6.58310175e-01 9.77550000e-02 -1.07944096e-02 -2.32133374e-01
6.10180914e-01 1.37420326e-01 2.49231443e-01 8.63275945e-01
-1.57571092e-01 8.87863159e-01 7.48073399e-01 -5.72614074e-01
-1.21140873e+00 -7.49676406e-01 -6.00112021e-01 1.03165090e+00
4.26131815e-01 -2.19755262e-01 -9.99547899e-01 -7.49723911e-01
-2.69026220e-01 5.00479162e-01 -6.98239803e-01 3.34943801e-01
-9.20983493e-01 -1.04191887e+00 1.03500879e+00 8.71582031e-01
1.20313227e+00 -1.19905508e+00 -4.64501947e-01 2.26633608e-01
-1.16332218e-01 -1.24140489e+00 -7.35334635e-01 3.70560527e-01
-7.07766831e-01 -5.90512276e-01 -1.05948544e+00 -9.45931137e-01
7.68788278e-01 8.01959559e-02 3.69688213e-01 -7.49172047e-02
-6.92427814e-01 -6.84049353e-03 -3.91336113e-01 -4.72698867e-01
5.17859086e-02 1.93401501e-01 -5.63143373e-01 2.91006807e-02
2.84312487e-01 2.12029546e-01 -6.39639914e-01 5.77692926e-01
-1.09427130e+00 3.26506555e-01 6.24995410e-01 7.70009398e-01
8.58569086e-01 2.56837815e-01 2.26288587e-01 -5.46617627e-01
4.51371491e-01 4.56431322e-02 -5.97221613e-01 2.10325941e-01
-8.87976810e-02 -4.23630327e-01 9.92077231e-01 -3.83827031e-01
-1.07764101e+00 1.71443567e-01 3.74823175e-02 -4.73511398e-01
-3.85038406e-01 9.26042423e-02 -1.00878313e-01 1.95717439e-01
3.70679408e-01 6.86140299e-01 -4.74245608e-01 -1.90622419e-01
7.16417506e-02 9.73416686e-01 3.22811037e-01 -4.94716823e-01
5.24439216e-01 5.71857452e-01 -2.19435781e-01 -1.28814363e+00
-5.98333418e-01 -3.36674482e-01 -8.42593789e-01 -8.94664750e-02
1.40164196e+00 -6.78913891e-01 -7.31870413e-01 8.59024882e-01
-9.20977414e-01 -6.38839185e-01 3.74653973e-02 3.41470122e-01
-4.08202559e-01 6.10574722e-01 -6.61975682e-01 -7.44157016e-01
-5.58284938e-01 -1.37923717e+00 1.66460669e+00 4.51602191e-01
1.62525639e-01 -8.94979239e-01 -7.71499276e-01 6.50398731e-01
8.14675633e-03 1.41277641e-01 8.97130787e-01 -7.37805724e-01
-7.00088322e-01 -4.64653254e-01 -8.33649099e-01 1.53445020e-01
-3.27926069e-01 -1.11374334e-02 -8.23965311e-01 -1.87090293e-01
-2.63424277e-01 -1.83159158e-01 7.70255148e-01 8.47187936e-01
1.60234189e+00 1.15356289e-01 -5.43349743e-01 8.64400208e-01
1.24055934e+00 5.15591800e-01 6.65087402e-01 3.45152944e-01
9.62018132e-01 2.16467157e-01 4.73894209e-01 5.64856768e-01
1.92257881e-01 6.65377378e-01 1.20428316e-01 -3.34483206e-01
1.67922694e-02 1.80920959e-02 1.88632504e-04 3.31875443e-01
3.59814577e-02 -8.42847764e-01 -8.90321553e-01 1.46086678e-01
-1.66894317e+00 -5.74258983e-01 -5.34962177e-01 1.86182070e+00
2.20494464e-01 3.80167335e-01 -1.15153074e-01 1.48618758e-01
9.87193882e-01 1.39899254e-01 -8.15819383e-01 -4.95140731e-01
-1.53808624e-01 1.44832417e-01 6.41223133e-01 -3.68012302e-02
-1.43825662e+00 1.47245491e+00 3.92802429e+00 1.41071248e+00
-1.29723310e+00 -4.27507937e-01 1.12123692e+00 2.25561559e-01
2.31613830e-01 -2.80665189e-01 -1.32195401e+00 5.23097754e-01
7.07806572e-02 3.27564687e-01 2.13218704e-01 7.32975483e-01
1.65582895e-01 -3.44230413e-01 -7.56750524e-01 9.49843884e-01
2.10634023e-01 -1.17137611e+00 1.09337285e-01 1.54838741e-01
6.27052426e-01 -1.38092905e-01 1.96933582e-01 5.04097879e-01
-1.01132333e-01 -1.16740108e+00 7.49232113e-01 3.04322064e-01
1.36522985e+00 -4.10732299e-01 5.00036418e-01 4.62445050e-01
-1.69493532e+00 -1.43066077e-02 -5.27059376e-01 6.34452224e-01
-1.59052745e-01 1.87903687e-01 -7.22524047e-01 6.86800957e-01
8.83336246e-01 5.76978266e-01 -6.20655298e-01 8.58423471e-01
1.01362169e-01 5.71940064e-01 -7.12111115e-01 -4.84097362e-01
6.47245884e-01 -2.96610862e-01 4.92214918e-01 1.28559339e+00
2.03959361e-01 3.69533449e-01 3.95691574e-01 1.00993359e+00
-2.25981668e-01 4.97010767e-01 -1.69832841e-01 4.05037515e-02
6.46141022e-02 1.20025349e+00 -1.73029959e+00 -3.84046882e-01
-3.28098655e-01 1.18222535e+00 -3.13453302e-02 5.16989529e-01
-1.11590171e+00 -7.69542694e-01 -2.09303483e-01 1.40562579e-01
7.12564886e-01 -5.38713820e-02 -9.14996862e-01 -1.16747510e+00
1.92267731e-01 -5.08700550e-01 3.86549860e-01 -8.85619640e-01
-8.46130192e-01 5.80494940e-01 -4.18290719e-02 -1.12926733e+00
5.40675759e-01 -8.91495764e-01 -6.77920461e-01 9.36258912e-01
-1.04685128e+00 -1.37567735e+00 -5.33139825e-01 8.59790623e-01
1.18194580e+00 -1.81679934e-01 3.17185909e-01 8.79371166e-02
-1.09726536e+00 1.03712428e+00 4.01411772e-01 7.22829044e-01
3.60652685e-01 -1.09696782e+00 3.57223272e-01 5.12955248e-01
-2.25489154e-01 3.59434962e-01 1.89545348e-01 -8.99822474e-01
-1.10008740e+00 -1.35591221e+00 1.06806464e-01 -1.85331970e-01
5.04082739e-01 -6.63348734e-01 -7.69103348e-01 9.13184762e-01
-2.56125361e-01 1.13362730e-01 -5.30673787e-02 -4.86218661e-01
9.26563665e-02 1.33848310e-01 -1.14786875e+00 8.48872125e-01
6.37976468e-01 -8.09173509e-02 -3.98370415e-01 3.48207891e-01
4.72433954e-01 -7.66920924e-01 -6.64016128e-01 1.67697340e-01
4.37833607e-01 -9.11606967e-01 7.44131148e-01 4.47460040e-02
4.47930336e-01 -9.25172940e-02 2.53131445e-02 -3.97668123e-01
2.18535010e-02 -4.11733121e-01 4.97975022e-01 1.13451374e+00
3.39604735e-01 -7.03608394e-01 1.23143160e+00 5.22454143e-01
-5.13657928e-01 -8.05649877e-01 -1.07793891e+00 -3.97637248e-01
5.68183899e-01 -4.69792575e-01 4.11389232e-01 7.82343090e-01
-2.35970885e-01 1.57280773e-01 -1.16658270e-01 1.80194266e-02
2.74180621e-01 6.99822679e-02 6.40165925e-01 -1.00999165e+00
-9.94710699e-02 -8.14148784e-01 -3.17110986e-01 -1.76801372e+00
-2.29900613e-01 -9.03498113e-01 4.19368520e-02 -1.51362216e+00
1.75706968e-01 -7.44018376e-01 1.60161525e-01 4.80343640e-01
-1.27506033e-01 2.72710681e-01 6.09992370e-02 2.14439824e-01
-6.37048960e-01 3.83810133e-01 1.56633687e+00 -5.63920774e-02
-4.14064914e-01 5.54555394e-02 -3.46447110e-01 8.65936637e-01
6.23266399e-01 -5.85824139e-02 -1.21605560e-01 -4.49807823e-01
-2.95484737e-02 4.05852944e-01 2.96247393e-01 -9.18987274e-01
3.17297399e-01 1.06356867e-01 1.02573681e+00 -1.27299678e+00
3.75567943e-01 -6.03210568e-01 -5.66105783e-01 -1.75390057e-02
-1.67918637e-01 -2.63779879e-01 3.65996987e-01 4.52956319e-01
1.49073750e-01 -3.95369291e-01 8.69504988e-01 -4.14668806e-02
-4.91483510e-01 4.44557518e-01 -4.98537838e-01 1.51346669e-01
1.30714500e+00 -8.53399634e-01 -3.64316434e-01 9.29097738e-03
-6.62135720e-01 4.22350913e-01 1.51394755e-01 3.48810852e-01
6.51466191e-01 -5.63830793e-01 -4.63096440e-01 3.12174827e-01
-6.36339262e-02 7.24280179e-01 4.95327502e-01 8.67430985e-01
-7.04560339e-01 8.27666163e-01 2.20014527e-01 -8.41168106e-01
-1.03813350e+00 2.84275144e-01 5.71572185e-01 -4.67267990e-01
-1.11995018e+00 9.92858529e-01 7.34155595e-01 -3.36834937e-01
3.45344841e-01 -6.62458420e-01 -5.88935874e-02 -1.02673404e-01
7.69374073e-02 2.75891006e-01 -2.10746713e-02 -7.65197575e-01
7.11069405e-02 8.82518530e-01 -5.31969249e-01 1.88278243e-01
9.34758127e-01 -1.89056560e-01 1.91081896e-01 -5.52793592e-02
9.27989602e-01 -2.09672540e-01 -1.40014982e+00 -7.84717873e-02
-2.71921188e-01 -1.03300147e-01 4.41296324e-02 -8.80920887e-01
-1.02490795e+00 9.98778164e-01 5.30922949e-01 1.78509921e-01
9.77218032e-01 -2.22924836e-02 9.46585059e-01 4.13201213e-01
3.60354707e-02 -1.39378953e+00 2.11962000e-01 5.45635164e-01
8.12349737e-01 -1.15422797e+00 1.19928740e-01 -4.92316455e-01
-7.58817017e-01 1.61927342e+00 8.59484136e-01 -1.96017936e-01
5.45487761e-01 3.25366616e-01 -2.11302340e-01 -1.63442150e-01
9.89631191e-02 5.12613617e-02 2.41740063e-01 5.19442074e-02
3.01449209e-01 -3.37252505e-02 -1.99069291e-01 7.94182181e-01
-2.58525550e-01 -3.39421839e-01 3.96494985e-01 8.22630703e-01
-5.61884403e-01 -5.93829513e-01 -6.11499012e-01 1.03186035e+00
-6.31944120e-01 -2.15091959e-01 -3.75342458e-01 1.16708362e+00
2.63829947e-01 7.01095402e-01 1.74952596e-01 -1.57856300e-01
1.86811537e-01 -5.36387116e-02 8.82781968e-02 -5.28642237e-01
-7.76410639e-01 7.61170268e-01 -3.08882207e-01 -7.05085173e-02
3.74906778e-01 -7.16710806e-01 -1.85702550e+00 7.66497627e-02
-6.88149452e-01 -2.88266689e-01 6.16354287e-01 9.78833318e-01
-2.53815562e-01 6.26561105e-01 4.96079415e-01 -8.67710769e-01
-1.93831727e-01 -9.70970631e-01 -7.37505555e-01 5.06247245e-02
-2.06556007e-01 -3.38628888e-01 -1.75309673e-01 6.34299144e-02] | [12.072267532348633, 2.2621943950653076] |
e092f074-40f9-407a-aa24-2199cf223ea6 | spatially-covariant-lesion-segmentation | 2301.07895 | null | https://arxiv.org/abs/2301.07895v1 | https://arxiv.org/pdf/2301.07895v1.pdf | Spatially Covariant Lesion Segmentation | Compared to natural images, medical images usually show stronger visual patterns and therefore this adds flexibility and elasticity to resource-limited clinical applications by injecting proper priors into neural networks. In this paper, we propose spatially covariant pixel-aligned classifier (SCP) to improve the computational efficiency and meantime maintain or increase accuracy for lesion segmentation. SCP relaxes the spatial invariance constraint imposed by convolutional operations and optimizes an underlying implicit function that maps image coordinates to network weights, the parameters of which are obtained along with the backbone network training and later used for generating network weights to capture spatially covariant contextual information. We demonstrate the effectiveness and efficiency of the proposed SCP using two lesion segmentation tasks from different imaging modalities: white matter hyperintensity segmentation in magnetic resonance imaging and liver tumor segmentation in contrast-enhanced abdominal computerized tomography. The network using SCP has achieved 23.8%, 64.9% and 74.7% reduction in GPU memory usage, FLOPs, and network size with similar or better accuracy for lesion segmentation. | ['Jiahao Li', 'Chao Li', 'Dongdong Liu', 'Jinwei Zhang', 'Rongguang Wang', 'Hang Zhang'] | 2023-01-19 | null | null | null | null | ['tumor-segmentation'] | ['computer-vision'] | [ 6.35658264e-01 3.04999352e-02 -9.48527381e-02 -3.76845896e-01
-2.09287271e-01 -3.98448199e-01 3.84154826e-01 1.35693356e-01
-7.26901710e-01 6.21394634e-01 1.68400824e-01 -3.86706144e-01
-1.79179057e-01 -7.43150353e-01 -4.21977699e-01 -9.46326017e-01
-2.49848306e-01 1.42299876e-01 4.14832532e-01 1.90777898e-01
3.23248774e-01 7.18853652e-01 -8.34426403e-01 2.74888903e-01
9.16992664e-01 8.53863955e-01 4.46391463e-01 6.12629116e-01
4.87599932e-02 8.01669776e-01 -7.69569501e-02 8.26143399e-02
3.19205105e-01 -2.91145951e-01 -7.93365777e-01 1.62706599e-01
1.62745088e-01 -7.77937472e-02 -2.89590299e-01 1.17028332e+00
5.30740559e-01 2.40348250e-01 7.21507072e-01 -8.69273305e-01
-5.81058562e-01 4.84854907e-01 -8.26111674e-01 6.23504639e-01
-3.38240504e-01 1.30259201e-01 5.11843741e-01 -6.52850270e-01
6.58159435e-01 6.90569878e-01 7.05883861e-01 3.85699987e-01
-1.20332968e+00 -3.98537427e-01 -8.06187242e-02 -2.05367822e-02
-1.27987611e+00 -4.27720957e-02 6.56401098e-01 -5.17355442e-01
8.82698536e-01 3.84772062e-01 6.76191986e-01 6.42785490e-01
4.42745954e-01 5.21157682e-01 1.08385468e+00 -4.74176526e-01
7.84594417e-02 -1.98861778e-01 2.13862628e-01 9.44323719e-01
2.66759068e-01 1.05029540e-02 -3.22288014e-02 -1.36296883e-01
1.33287084e+00 5.86883686e-02 -4.34775829e-01 -4.99196917e-01
-1.43807757e+00 8.00266623e-01 8.94318759e-01 5.19895971e-01
-5.64080358e-01 1.42104432e-01 3.60841423e-01 -3.25723916e-01
2.98800707e-01 4.47773427e-01 -1.82136998e-01 2.93342173e-01
-9.27884877e-01 -2.05011711e-01 2.69151300e-01 6.00166619e-01
4.08406883e-01 1.97186902e-01 -3.29471499e-01 9.30594802e-01
2.12031007e-01 2.15242580e-01 7.40571260e-01 -8.63989592e-01
1.67227939e-01 7.12793410e-01 -2.09868744e-01 -1.28546715e+00
-9.42822933e-01 -6.49637461e-01 -1.12970674e+00 1.96125254e-01
4.18109149e-01 -5.77654876e-02 -1.26505709e+00 1.59429693e+00
4.28174108e-01 1.87235042e-01 -1.84028998e-01 1.20172584e+00
6.11885071e-01 4.43696946e-01 4.58023071e-01 6.01571128e-02
1.42652822e+00 -1.11195481e+00 -4.96860802e-01 -1.39327168e-01
7.42439032e-01 -7.23318160e-01 1.03865480e+00 6.77963644e-02
-1.15316820e+00 -1.88772500e-01 -1.12142777e+00 4.66740839e-02
-1.59244150e-01 3.56378585e-01 5.86277187e-01 5.71924746e-01
-9.97523904e-01 6.23006523e-01 -1.28727818e+00 -1.89805970e-01
6.25679970e-01 5.64501524e-01 -2.35627443e-01 5.29290177e-02
-8.08416188e-01 8.34309459e-01 5.43450177e-01 1.20001316e-01
-4.82772917e-01 -1.08035338e+00 -6.94257855e-01 8.67381878e-03
9.79505032e-02 -7.37194300e-01 8.69419754e-01 -1.12490857e+00
-1.36387360e+00 8.42403471e-01 1.59280047e-01 -4.96089637e-01
5.92063546e-01 2.68064171e-01 -6.40114397e-02 4.85931158e-01
4.32536192e-02 1.00536191e+00 5.47449589e-01 -9.18570876e-01
-4.97016460e-01 -3.27821553e-01 -1.87357664e-01 3.72663826e-01
5.07991575e-03 -3.14570889e-02 -5.41515231e-01 -9.80632782e-01
5.14222383e-01 -1.02947438e+00 -4.80030686e-01 3.00377816e-01
-4.16424960e-01 1.60443768e-01 8.84447813e-01 -9.37157333e-01
8.06618214e-01 -2.00410962e+00 8.65204483e-02 5.32260120e-01
2.43481278e-01 2.86413878e-01 7.13146850e-02 -3.31094861e-01
-3.42667550e-01 -3.48853916e-02 -6.31943643e-01 2.06360132e-01
-3.38036269e-01 2.50320852e-01 3.16293895e-01 6.27827048e-01
6.42553344e-02 1.07028556e+00 -7.72314548e-01 -6.38456762e-01
5.40561616e-01 6.55787587e-01 -6.40255213e-01 -1.77650645e-01
2.02643827e-01 8.18988800e-01 -4.15366292e-01 4.78131533e-01
6.41681969e-01 -5.02031565e-01 2.62606502e-01 -4.54857439e-01
-3.26630510e-02 -1.67844042e-01 -8.70142460e-01 1.81710815e+00
-3.19908321e-01 5.26674986e-01 2.58656263e-01 -1.11744225e+00
5.55393934e-01 2.48470083e-01 8.41408670e-01 -9.30837750e-01
4.20830220e-01 -5.26169576e-02 2.36950442e-01 -6.03957415e-01
2.02690810e-01 -6.49993494e-02 2.16542736e-01 4.00136083e-01
-2.29290292e-01 1.34447619e-01 1.00644358e-01 3.22620124e-02
8.11375916e-01 7.67400442e-03 3.02923620e-01 -8.54664087e-01
7.59412706e-01 2.13488787e-01 4.64151084e-01 6.17908537e-01
-3.39149565e-01 7.12984741e-01 4.00605857e-01 -7.18698263e-01
-1.16542876e+00 -9.72699761e-01 -3.46681416e-01 8.69321167e-01
1.09593878e-02 2.49936163e-01 -8.34896445e-01 -6.71905756e-01
-2.52083153e-01 4.73091692e-01 -7.20511317e-01 6.75623193e-02
-1.01903021e+00 -1.21250653e+00 3.56848538e-01 7.14938164e-01
7.54019320e-01 -1.08705199e+00 -9.15419340e-01 1.95683613e-01
2.08127499e-02 -1.09113812e+00 -6.91287637e-01 1.64099008e-01
-1.08740103e+00 -1.11534786e+00 -9.45539415e-01 -1.05997467e+00
1.28562534e+00 1.07429862e-01 8.61828089e-01 2.69493848e-01
-8.46594989e-01 1.80039816e-02 -5.14369048e-02 -3.21487524e-02
-1.95766568e-01 1.89219818e-01 -3.81506681e-01 -2.06033066e-01
-1.16095543e-01 -5.42600214e-01 -1.03652442e+00 1.49751455e-01
-1.02489281e+00 5.34190953e-01 6.33978426e-01 9.80424464e-01
6.50834024e-01 -2.76123166e-01 2.32148021e-01 -9.60855722e-01
4.89880055e-01 -2.73988843e-01 -5.56008458e-01 1.83633849e-01
-4.56497490e-01 -2.76310816e-02 4.61739957e-01 -3.19690138e-01
-1.08155298e+00 3.45285952e-01 1.43398583e-01 -1.21784285e-01
-8.17915946e-02 4.67988551e-01 6.00614607e-01 -4.02327359e-01
7.66477108e-01 2.89220631e-01 1.25793323e-01 -9.86813158e-02
2.29618534e-01 2.61615276e-01 6.70921981e-01 -4.15704697e-01
3.44977736e-01 6.93785250e-01 1.81470186e-01 -5.68144321e-01
-4.85615760e-01 -2.90395170e-01 -8.87966394e-01 -7.67322332e-02
1.13796949e+00 -4.42390859e-01 -4.86781090e-01 2.90654719e-01
-9.26200628e-01 -4.03135717e-01 -1.06467552e-01 6.83343649e-01
-2.89023668e-01 3.72091383e-01 -9.00557101e-01 -1.86589897e-01
-7.47440279e-01 -1.48409283e+00 6.66123688e-01 3.12914133e-01
-7.16664493e-02 -1.18753469e+00 -3.45791310e-01 7.80448392e-02
7.74358094e-01 4.47467655e-01 1.31057084e+00 -2.09411278e-01
-5.19472957e-01 5.44204898e-02 -7.42290854e-01 2.04655290e-01
2.30541438e-01 -2.32702017e-01 -5.09254456e-01 -3.44469547e-01
-2.21338794e-01 2.68117428e-01 6.00734353e-01 1.05378413e+00
1.45657182e+00 -1.43890753e-01 -3.50333571e-01 1.01792097e+00
1.53305173e+00 4.67989445e-01 3.67729366e-01 3.53566140e-01
7.70009279e-01 5.35112739e-01 2.41610353e-04 2.38102376e-01
3.47248055e-02 5.59801579e-01 3.91660392e-01 -5.71786523e-01
-4.33167636e-01 2.39938781e-01 -3.94831121e-01 8.40395451e-01
-1.07552119e-01 2.29812399e-01 -1.07858491e+00 6.12909675e-01
-1.69744933e+00 -4.97980118e-01 -2.32214496e-01 1.86164987e+00
8.73294592e-01 -1.94587186e-01 -2.13151112e-01 -2.03817576e-01
8.55619490e-01 -1.18774161e-01 -5.23401916e-01 -1.45900503e-01
1.88549861e-01 3.07661533e-01 8.72958481e-01 5.71286023e-01
-1.26440215e+00 7.09753871e-01 6.26305199e+00 6.90852463e-01
-1.51214695e+00 3.77023220e-01 8.57565284e-01 -6.32341206e-02
2.09973291e-01 -2.95308888e-01 -2.59194300e-02 3.52436304e-01
3.77723247e-01 2.41217799e-02 3.80711019e-01 7.99243987e-01
2.49825791e-01 -3.11127990e-01 -6.51336968e-01 8.91704321e-01
-1.45398349e-01 -1.65785396e+00 2.28021592e-02 -1.48128092e-01
9.67000723e-01 1.86216474e-01 2.41521388e-01 -2.42777869e-01
9.01544765e-02 -1.10857797e+00 4.67253149e-01 3.20824355e-01
6.08692944e-01 -7.91945279e-01 7.59465516e-01 9.23617855e-02
-9.67264295e-01 1.50707096e-01 -2.06028253e-01 3.30760568e-01
-9.28882882e-02 2.82928884e-01 -1.13940978e+00 2.61812180e-01
6.56409323e-01 4.66753453e-01 -4.93330419e-01 1.14286077e+00
-8.21868330e-02 3.05855840e-01 -3.16267192e-01 1.49942920e-01
7.96160698e-01 -2.81620115e-01 3.30560535e-01 1.38073218e+00
1.28692999e-01 1.27293423e-01 2.35220835e-01 7.78725743e-01
1.13318704e-01 2.90972054e-01 -2.40434691e-01 4.52593327e-01
1.62166357e-01 1.57508278e+00 -1.43497527e+00 -3.71185601e-01
-1.57422930e-01 7.72643507e-01 1.58865497e-01 4.36387390e-01
-1.01625979e+00 -3.05627704e-01 1.09669916e-01 7.02045262e-02
2.06683323e-01 -2.67597228e-01 -6.19851470e-01 -8.66219342e-01
-2.21137807e-01 -5.56605637e-01 3.36086422e-01 -6.60613120e-01
-9.79651868e-01 7.22621620e-01 3.99533249e-02 -8.96159589e-01
2.41023973e-02 -8.34801316e-01 -6.36075795e-01 8.21183920e-01
-1.57195330e+00 -1.05742359e+00 -5.92818439e-01 6.04237556e-01
2.70271182e-01 5.60423844e-02 6.13133132e-01 3.35456371e-01
-5.74479043e-01 2.56742328e-01 3.25623415e-02 2.83576131e-01
3.80897671e-01 -1.14871514e+00 -2.87336390e-02 8.48967731e-01
-2.35894993e-01 5.68959832e-01 3.33042920e-01 -5.17877162e-01
-8.72275770e-01 -1.18890870e+00 4.05563444e-01 1.01880684e-01
5.49341440e-01 5.77236451e-02 -8.60550225e-01 5.59516609e-01
3.30491573e-01 4.14657980e-01 4.75202233e-01 -5.09342194e-01
5.69976084e-02 2.13851705e-01 -1.47279239e+00 8.63736093e-01
6.54833257e-01 -1.64290637e-01 -2.33511016e-01 5.98596811e-01
2.94073284e-01 -8.10134888e-01 -8.91833961e-01 4.25565600e-01
4.23862785e-01 -7.43354201e-01 1.17627764e+00 -3.35158318e-01
3.28305274e-01 -2.06737280e-01 9.63000581e-02 -1.06085360e+00
-4.90523994e-01 -2.04552516e-01 4.34311330e-01 3.77243876e-01
3.94464403e-01 -7.07520902e-01 9.76380765e-01 7.76628613e-01
-3.78115714e-01 -1.00744534e+00 -9.23463106e-01 -2.64744043e-01
1.89037636e-01 -2.79530793e-01 3.65153521e-01 1.15335596e+00
-2.13138357e-01 -2.49870598e-01 1.38898920e-02 1.03114687e-01
6.76514447e-01 -1.69575110e-01 -6.41672965e-03 -7.32379854e-01
-7.10715260e-03 -7.37769186e-01 -3.18027407e-01 -5.66508889e-01
-1.58300966e-01 -1.27799809e+00 -2.12842435e-01 -1.41447115e+00
3.20676714e-01 -6.97707355e-01 -4.72169548e-01 6.72032833e-01
-7.04488009e-02 6.63356960e-01 1.29377797e-01 2.80302972e-01
-2.58476615e-01 1.00953922e-01 1.64096653e+00 -1.18770920e-01
-1.35378703e-01 -2.95002669e-01 -3.45785111e-01 8.93906832e-01
9.71105635e-01 -4.95254666e-01 -4.89446670e-01 -5.99465966e-01
-3.60122681e-01 3.17686726e-03 4.92174774e-01 -8.85300159e-01
3.78301084e-01 3.46297808e-02 8.64106655e-01 -3.29845041e-01
-1.01097070e-01 -7.98370600e-01 1.11942083e-01 8.61251175e-01
-4.05076265e-01 1.49370581e-01 2.52288729e-01 2.16262951e-01
-1.01883568e-01 -4.34645563e-01 1.13953149e+00 -4.08209026e-01
-5.40739417e-01 2.49168620e-01 -4.44670141e-01 -2.48464391e-01
1.31755984e+00 -2.99192756e-01 -1.89893529e-01 7.30833709e-02
-1.01514685e+00 -5.14250323e-02 1.78610414e-01 1.18048385e-01
6.13195300e-01 -1.23034835e+00 -4.87967402e-01 3.62858295e-01
-2.98307151e-01 -9.26811844e-02 4.86868680e-01 1.38698864e+00
-1.32646847e+00 6.33191526e-01 -5.55908203e-01 -9.12061632e-01
-1.27105892e+00 2.56004512e-01 6.49304390e-01 -4.62086558e-01
-9.82964694e-01 7.49028802e-01 2.12838382e-01 -3.33297819e-01
5.55062518e-02 -6.45183384e-01 -2.55337596e-01 -4.21307176e-01
2.14599758e-01 1.91423714e-01 2.66177207e-01 -7.40619421e-01
-3.22965950e-01 5.59410930e-01 -2.59799421e-01 3.59083563e-02
1.22431314e+00 1.09349400e-01 -2.10410029e-01 -3.19939643e-01
1.15184033e+00 -3.37807536e-01 -1.40716493e+00 -3.70892853e-01
7.38290399e-02 -3.13570917e-01 6.41363621e-01 -1.01263618e+00
-1.59940112e+00 7.32213557e-01 9.90599215e-01 -1.63074248e-02
1.11282766e+00 -2.86656290e-01 6.38645530e-01 5.39423106e-03
1.61423817e-01 -8.20034802e-01 -8.98270682e-02 2.02538252e-01
6.94266438e-01 -1.10944223e+00 1.78567842e-01 -4.66137797e-01
-6.80381715e-01 1.22264731e+00 6.26547515e-01 -4.74838018e-01
5.79841554e-01 4.27113950e-01 3.72406505e-02 -2.86063999e-01
-1.70369908e-01 1.86502680e-01 5.49418211e-01 5.09812534e-01
5.27793288e-01 1.75330922e-01 -4.46600705e-01 1.78847998e-01
1.14731938e-01 -1.70883119e-01 1.48116171e-01 1.05948687e+00
-2.67190903e-01 -8.41406286e-01 -2.48469517e-01 5.09173512e-01
-7.14772701e-01 -1.11925013e-01 1.98655859e-01 9.48129475e-01
1.66282356e-01 4.28544074e-01 1.83140337e-01 1.64814115e-01
8.03756788e-02 -3.94128487e-02 5.91517746e-01 -3.81848961e-01
-8.17769945e-01 1.82039142e-01 -2.21101537e-01 -4.31620419e-01
-5.34234643e-01 -4.11503851e-01 -1.58037388e+00 -1.86790470e-02
-4.71492670e-02 -6.41177967e-02 8.79387200e-01 7.47331262e-01
3.99232775e-01 9.40939784e-01 1.61941752e-01 -9.17781234e-01
-7.29879662e-02 -9.21102166e-01 -3.87582332e-01 5.15611708e-01
9.00372416e-02 -5.30433476e-01 6.48592487e-02 1.56704500e-01] | [14.518016815185547, -2.471959352493286] |
fec78fbc-d669-4b04-bd93-1413ea008254 | msv-challenge-2022-npu-hc-speaker | 2211.16694 | null | https://arxiv.org/abs/2211.16694v2 | https://arxiv.org/pdf/2211.16694v2.pdf | MSV Challenge 2022: NPU-HC Speaker Verification System for Low-resource Indian Languages | This report describes the NPU-HC speaker verification system submitted to the O-COCOSDA Multi-lingual Speaker Verification (MSV) Challenge 2022, which focuses on developing speaker verification systems for low-resource Asian languages. We participate in the I-MSV track, which aims to develop speaker verification systems for various Indian languages. In this challenge, we first explore different neural network frameworks for low-resource speaker verification. Then we leverage vanilla fine-tuning and weight transfer fine-tuning to transfer the out-domain pre-trained models to the in-domain Indian dataset. Specifically, the weight transfer fine-tuning aims to constrain the distance of the weights between the pre-trained model and the fine-tuned model, which takes advantage of the previously acquired discriminative ability from the large-scale out-domain datasets and avoids catastrophic forgetting and overfitting at the same time. Finally, score fusion is adopted to further improve performance. Together with the above contributions, we obtain 0.223% EER on the public evaluation set, ranking 2nd place on the leaderboard. On the private evaluation set, the EER of our submitted system is 2.123% and 0.630% for the constrained and unconstrained sub-tasks of the I-MSV track, leading to the 1st and 3rd place in the ranking, respectively. | ['Lei Xie', 'Jie Liu', 'Namin Wang', 'Li Zhang', 'Yue Li'] | 2022-11-30 | null | null | null | null | ['speaker-verification'] | ['speech'] | [-4.64184172e-02 -1.30321672e-02 1.47922747e-02 -6.31037712e-01
-1.44411850e+00 -5.43286741e-01 4.22325134e-01 -1.62517384e-01
-5.67433238e-01 4.25557941e-01 2.46377751e-01 -3.65432292e-01
2.56870300e-01 -7.45001361e-02 -6.81169271e-01 -7.26961970e-01
5.49048409e-02 3.63579392e-01 -2.64225334e-01 -1.80526301e-01
-2.60473341e-01 3.36779594e-01 -1.12679827e+00 1.61469262e-02
8.71300519e-01 1.06795311e+00 -1.33619398e-01 5.98165631e-01
2.61784017e-01 1.61098108e-01 -5.82339764e-01 -7.11489737e-01
4.98513551e-03 -8.26818869e-02 -6.52732849e-01 -1.53134093e-01
5.54279745e-01 -1.92371592e-01 -3.07556480e-01 1.01718259e+00
8.64230633e-01 1.03593469e-01 3.40221703e-01 -1.09713674e+00
-7.07656205e-01 1.09974897e+00 -4.41367149e-01 6.87573180e-02
5.82316108e-02 1.10092908e-01 9.96623635e-01 -1.23689806e+00
4.09941643e-01 1.18990254e+00 6.19134665e-01 8.29290390e-01
-1.03812873e+00 -1.06978679e+00 4.18360829e-01 4.23467487e-01
-1.80624008e+00 -1.21539044e+00 7.00413764e-01 -2.91299731e-01
8.08133543e-01 1.00215390e-01 -4.19747829e-02 1.11296082e+00
-3.81690264e-01 9.02280688e-01 7.15097547e-01 -3.94588649e-01
1.31067023e-01 3.61145139e-01 3.11353385e-01 4.33132380e-01
-8.53306055e-02 2.05432862e-01 -8.35923672e-01 -1.32891446e-01
-8.85919333e-02 -5.27051926e-01 -4.51118857e-01 5.27439527e-02
-1.05834472e+00 7.09020317e-01 2.51579672e-01 4.39427078e-01
-2.64433324e-01 -3.77506793e-01 4.98576730e-01 2.69465327e-01
4.25460160e-01 -8.43013898e-02 -8.50167096e-01 -5.63292764e-02
-1.14940405e+00 -7.64252543e-02 6.13073349e-01 8.69029939e-01
4.98617858e-01 5.21843255e-01 -4.65702176e-01 1.22216105e+00
6.02962971e-01 7.79628038e-01 6.20420933e-01 -4.14051741e-01
7.05338955e-01 1.63517758e-01 -1.47598654e-01 -1.05821177e-01
-8.32123756e-02 -7.36769140e-01 -8.57056856e-01 -3.66857275e-02
3.43078375e-01 -3.09026003e-01 -9.87787664e-01 2.13362265e+00
2.95738697e-01 1.47567540e-01 3.98835838e-01 7.70832360e-01
1.03557515e+00 6.06411815e-01 1.09047480e-01 -1.09991007e-01
1.53013015e+00 -9.76660252e-01 -5.23594379e-01 -1.22563295e-01
4.05953258e-01 -7.95629859e-01 9.78109360e-01 1.46975398e-01
-9.03028965e-01 -6.72619462e-01 -1.25913239e+00 3.76161300e-02
-2.12501168e-01 5.31351864e-01 -2.40139067e-01 8.78861904e-01
-1.31523275e+00 1.19798303e-01 -6.47479057e-01 -9.20903683e-03
3.26233059e-01 3.59024346e-01 -4.23932016e-01 -2.50751823e-01
-1.42137682e+00 9.06848252e-01 1.67176068e-01 3.78524721e-01
-1.19812679e+00 -8.88401985e-01 -8.25492501e-01 1.17019720e-01
1.76635608e-01 7.85114169e-02 1.31020296e+00 -5.90870380e-01
-1.78683424e+00 9.44673419e-01 -5.41293085e-01 -4.59236026e-01
5.37098289e-01 -1.21447414e-01 -9.14165914e-01 -4.52516437e-01
-1.20676234e-01 5.89771748e-01 7.42270768e-01 -1.01644564e+00
-5.24690449e-01 -3.84648263e-01 -5.20887733e-01 3.82020287e-02
-4.05656993e-01 3.34879816e-01 -7.18975127e-01 -5.53072691e-01
-8.81452560e-02 -9.36861455e-01 1.89495161e-01 -5.46711683e-01
-6.71704829e-01 -4.73888218e-01 8.20126474e-01 -1.21928334e+00
1.10409904e+00 -2.55655766e+00 1.54352963e-01 4.15824294e-01
-1.43762499e-01 5.01716733e-01 -4.39877242e-01 -3.44876503e-03
-1.37930214e-01 -7.01184720e-02 -2.25819886e-01 -8.97328377e-01
3.31715763e-01 -2.26607800e-01 -2.64126450e-01 4.07221109e-01
2.19702020e-01 6.61790848e-01 -3.25552166e-01 -1.80701733e-01
-1.72421604e-01 8.08774054e-01 -3.62822533e-01 9.94457155e-02
1.88567117e-01 4.71370757e-01 -5.69945276e-02 6.42557025e-01
9.99093771e-01 2.22148314e-01 2.19536513e-01 -1.16958730e-01
-1.15675785e-01 7.07591474e-01 -1.20032430e+00 1.55666101e+00
-3.98400664e-01 4.45702046e-01 4.73184347e-01 -5.73404014e-01
1.03097296e+00 6.59053266e-01 3.23976055e-02 -6.57613754e-01
1.12578720e-01 4.11629051e-01 1.62146777e-01 6.65289834e-02
4.55564588e-01 -8.48909914e-02 -1.43327385e-01 4.10558313e-01
2.88154542e-01 1.52510718e-01 -4.21386138e-02 1.47045165e-01
5.18309355e-01 -1.71917722e-01 3.14959362e-02 -2.87282377e-01
1.10155714e+00 -6.67888105e-01 7.83128500e-01 5.94071984e-01
-5.91963828e-01 4.75285918e-01 3.38488743e-02 6.21393099e-02
-6.29744768e-01 -1.04361212e+00 -3.38628471e-01 1.43625712e+00
-4.77757007e-01 -1.86197013e-01 -8.80823195e-01 -6.16808295e-01
-2.98716407e-02 7.64654279e-01 -5.67816019e-01 -1.36994764e-01
-6.06584609e-01 -4.59306866e-01 1.05529201e+00 3.05593640e-01
3.88663799e-01 -8.57447088e-01 3.90664726e-01 1.51783675e-01
-2.85579652e-01 -1.19135058e+00 -9.78584647e-01 1.92135930e-01
-3.23509157e-01 -5.97508132e-01 -9.31968331e-01 -9.96482074e-01
2.82749385e-01 -1.29220501e-01 8.11324775e-01 -7.14906827e-02
2.29282245e-01 -9.72284563e-03 -1.85312569e-01 -3.98542970e-01
-5.78637302e-01 4.98205423e-01 5.59651554e-01 3.79315645e-01
6.28810585e-01 -1.82978898e-01 5.94095290e-02 3.81425649e-01
-3.02565873e-01 -3.23000818e-01 3.03772569e-01 9.23714042e-01
4.50977504e-01 -3.90499771e-01 8.69108200e-01 -3.73075247e-01
3.85555089e-01 -1.35302499e-01 -1.00756383e+00 6.01885319e-01
-5.13042927e-01 1.39102653e-01 2.54846543e-01 -5.10748923e-01
-1.08430469e+00 3.63598093e-02 -4.95590210e-01 -3.34316850e-01
1.09600224e-01 4.48498845e-01 -7.57690489e-01 5.13069145e-02
3.87081593e-01 4.29134250e-01 -1.73931777e-01 -6.34374619e-01
3.09241116e-01 1.07806385e+00 6.66450739e-01 -4.71527308e-01
9.96740937e-01 -1.07405245e-01 -7.64992356e-01 -7.28761375e-01
-7.49966800e-01 -2.88088322e-01 -4.95267063e-01 7.50580505e-02
7.03660548e-01 -1.40196204e+00 -8.56349587e-01 7.37125278e-01
-1.04894161e+00 -3.22355598e-01 -6.77756965e-02 5.93316734e-01
5.63283078e-02 1.91758975e-01 -7.07266092e-01 -9.52626348e-01
-6.58940136e-01 -1.48593843e+00 9.89404440e-01 6.21347427e-02
3.58209461e-02 -5.90900064e-01 1.14351906e-01 6.07083619e-01
8.05336118e-01 -6.41501486e-01 5.20164132e-01 -9.75233674e-01
-2.89124906e-01 -2.17600033e-01 -1.16249762e-01 6.66415989e-01
-3.38507444e-02 -1.48283631e-01 -1.66433120e+00 -6.07298791e-01
-1.85868397e-01 -3.21257979e-01 9.54701304e-01 3.27456295e-01
8.56254876e-01 -9.81493294e-02 -5.23476191e-02 5.24579883e-01
7.50684202e-01 6.92273006e-02 3.60797465e-01 -5.36348261e-02
5.41557193e-01 5.22462130e-01 2.64016420e-01 1.45578578e-01
6.65788651e-01 1.07795513e+00 -1.71148106e-01 1.42800272e-01
-3.12297702e-01 -2.91094542e-01 7.64027894e-01 1.23415029e+00
2.40789309e-01 -3.95720266e-02 -9.21744168e-01 6.14196658e-01
-1.34835637e+00 -7.47471809e-01 1.98429674e-01 2.43294191e+00
1.01388216e+00 2.27702013e-03 1.83671400e-01 4.56349067e-02
9.47759211e-01 1.28362402e-01 -6.28704906e-01 -3.07185799e-01
-3.84635746e-01 1.26312539e-01 2.75296003e-01 8.46252799e-01
-1.16889930e+00 1.03957534e+00 5.71770954e+00 9.14339483e-01
-1.39461994e+00 5.75289905e-01 5.45395374e-01 -2.08114535e-01
-1.30081326e-01 -2.49849707e-01 -1.41196501e+00 4.24272716e-01
1.43046260e+00 -1.65547848e-01 4.87792522e-01 7.20580935e-01
1.61327757e-02 5.83444595e-01 -1.09897709e+00 8.52191567e-01
1.89681724e-01 -9.57998097e-01 -2.29756907e-01 2.05628440e-01
7.42076814e-01 6.22715712e-01 2.60458857e-01 8.24629664e-01
3.30066055e-01 -9.60664511e-01 1.02149451e+00 7.38541130e-03
1.11649704e+00 -7.76730537e-01 9.80128109e-01 1.16664611e-01
-1.18983579e+00 3.39611014e-03 -1.53036475e-01 5.12945414e-01
3.27861607e-01 5.95777512e-01 -9.78545785e-01 4.54875618e-01
6.05279982e-01 2.85804629e-01 -3.94343793e-01 8.05603981e-01
-3.17340255e-01 9.25785065e-01 -2.90090978e-01 1.09235279e-01
-1.38221323e-01 1.62500605e-01 5.41563392e-01 1.23697126e+00
3.08528781e-01 -4.53479946e-01 -5.74109331e-02 5.40743530e-01
-7.69290507e-01 5.35484068e-02 2.98487172e-02 1.22020967e-01
7.44524002e-01 1.10765254e+00 3.86081070e-01 -4.07583863e-01
-1.35697380e-01 7.17945039e-01 5.53644180e-01 5.28699160e-01
-7.82124639e-01 -2.19181910e-01 8.44997168e-01 -3.97393256e-01
6.01569653e-01 1.84816867e-02 -2.09573910e-01 -1.32825601e+00
1.15108170e-01 -1.05301571e+00 3.55065256e-01 -2.63094634e-01
-1.24938452e+00 9.85680282e-01 -4.58213061e-01 -9.45116043e-01
-4.32159811e-01 -3.22815061e-01 -5.73057711e-01 1.41792524e+00
-1.79175735e+00 -1.42121780e+00 1.01668216e-01 6.25114679e-01
4.22822028e-01 -6.16631567e-01 9.88066316e-01 7.51124978e-01
-9.51021731e-01 1.65708113e+00 2.44134381e-01 3.82680714e-01
1.02674222e+00 -7.97247708e-01 6.75471067e-01 1.05529141e+00
7.98041150e-02 5.27889848e-01 3.24710608e-01 -2.82992154e-01
-1.18023264e+00 -1.16718578e+00 1.35944438e+00 -4.27978039e-01
5.79362631e-01 -7.57177651e-01 -1.14432991e+00 5.62507987e-01
1.96268912e-02 9.74404067e-02 7.73737669e-01 5.84647357e-01
-9.28935945e-01 -3.74201685e-01 -1.20576167e+00 2.86449432e-01
5.58652937e-01 -1.03555715e+00 -3.85735393e-01 1.00270189e-01
8.05456042e-01 -3.18307728e-01 -8.06735396e-01 2.00667098e-01
6.51662171e-01 -3.14403087e-01 8.22014749e-01 -5.83916366e-01
-1.34929851e-01 -4.73798662e-01 -5.74302673e-01 -1.35694897e+00
-1.96163878e-01 -5.11271358e-01 3.60190943e-02 1.87600136e+00
9.54791963e-01 -6.16465986e-01 4.60525572e-01 4.32592839e-01
-1.43631235e-01 -1.93417832e-01 -1.48193336e+00 -8.43291581e-01
2.66521960e-01 -5.89510739e-01 8.27589393e-01 9.77091372e-01
-2.35272124e-01 3.33419025e-01 -4.81997341e-01 4.92392838e-01
8.77260625e-01 -1.52364314e-01 6.96498930e-01 -1.06314802e+00
-3.90792519e-01 -4.60435629e-01 -4.82267290e-02 -7.35267341e-01
4.31232870e-01 -1.11408818e+00 3.38179439e-01 -7.88242042e-01
2.02577457e-01 -4.15099889e-01 -7.43375242e-01 6.85544968e-01
-3.57917339e-01 2.95622557e-01 3.95842940e-01 9.38546434e-02
-3.92556190e-01 7.00584590e-01 6.18938267e-01 -4.11582619e-01
-2.73842126e-01 1.59675494e-01 -8.64264190e-01 1.73707828e-01
7.76727557e-01 -4.51265514e-01 1.90414339e-01 -5.07368028e-01
-5.94078720e-01 -2.09344327e-01 1.77209340e-02 -8.05999041e-01
2.13389322e-01 2.41706401e-01 2.88107485e-01 -5.18505871e-01
4.75624263e-01 -4.18657780e-01 -6.43449798e-02 4.24483538e-01
-4.25766885e-01 -3.29410762e-01 4.28429872e-01 1.33855686e-01
-3.31177324e-01 3.23095173e-03 1.05882585e+00 5.10424614e-01
-3.72545183e-01 3.07531178e-01 -7.10749328e-02 -4.69114557e-02
4.81414288e-01 2.32773304e-01 -3.46066505e-01 -3.86267632e-01
-7.90986419e-01 3.78229678e-01 7.82659426e-02 7.61232555e-01
2.88469344e-01 -1.36368001e+00 -1.42779613e+00 5.83982050e-01
3.49694848e-01 -3.79341692e-01 6.63695037e-01 7.71878779e-01
2.69841641e-01 5.23628712e-01 1.65517420e-01 -5.93073964e-01
-1.45858312e+00 1.79042146e-01 3.25100303e-01 -3.93749923e-01
-8.72725248e-02 1.13216949e+00 9.47884750e-03 -1.01582921e+00
5.24489105e-01 1.23526100e-02 3.53677236e-02 4.32420112e-02
8.59733164e-01 1.48737743e-01 4.78303105e-01 -1.09625816e+00
-9.70505655e-01 3.40934873e-01 -3.99641901e-01 -4.44609314e-01
1.26141775e+00 -1.35571077e-01 2.35322475e-01 2.29031220e-01
1.31050193e+00 4.33442116e-01 -1.06898451e+00 -6.87127411e-01
-2.20102128e-02 1.41509905e-01 2.24978536e-01 -1.08958209e+00
-1.11905193e+00 1.00907171e+00 8.09257865e-01 -3.18302542e-01
8.79766583e-01 2.26890277e-02 7.45961726e-01 3.40748370e-01
1.27288997e-01 -1.08018875e+00 -5.54843783e-01 8.15221608e-01
1.05590487e+00 -1.44885039e+00 -3.18303108e-01 -6.67543933e-02
-8.58310938e-01 5.86252272e-01 4.27001178e-01 4.55800921e-01
6.49723291e-01 2.47632563e-01 4.24072832e-01 4.11817670e-01
-6.70847595e-01 8.67995098e-02 5.93294799e-01 5.73553085e-01
4.56184477e-01 4.79903728e-01 2.27072135e-01 9.02641594e-01
-4.31202322e-01 -1.18718468e-01 6.92738518e-02 3.93751800e-01
-2.18303069e-01 -1.18198180e+00 -4.97303337e-01 3.43817808e-02
-4.16290849e-01 -3.59538972e-01 -3.19429725e-01 3.38711947e-01
-7.97598884e-02 1.16054761e+00 -2.17881516e-01 -5.60091376e-01
5.00225604e-01 4.69631642e-01 6.94901794e-02 -1.13922440e-01
-8.79987538e-01 6.91957250e-02 1.79522097e-01 -2.31969550e-01
-1.29472166e-02 -7.76224911e-01 -9.10744250e-01 -2.68894732e-01
-4.30882484e-01 3.87850940e-01 1.06553960e+00 8.75356495e-01
5.56706250e-01 4.05045271e-01 6.47934616e-01 -7.67132759e-01
-9.56704617e-01 -1.47134924e+00 -5.21016657e-01 8.33927244e-02
7.10273802e-01 -4.08783585e-01 -4.16207910e-01 -1.52268484e-01] | [14.344088554382324, 6.193714141845703] |
6dc7d948-a751-4160-a33b-1b15c79da89d | one-transform-to-compute-them-all-efficient | 2304.03412 | null | https://arxiv.org/abs/2304.03412v1 | https://arxiv.org/pdf/2304.03412v1.pdf | One Transform To Compute Them All: Efficient Fusion-Based Full-Reference Video Quality Assessment | The Visual Multimethod Assessment Fusion (VMAF) algorithm has recently emerged as a state-of-the-art approach to video quality prediction, that now pervades the streaming and social media industry. However, since VMAF requires the evaluation of a heterogeneous set of quality models, it is computationally expensive. Given other advances in hardware-accelerated encoding, quality assessment is emerging as a significant bottleneck in video compression pipelines. Towards alleviating this burden, we propose a novel Fusion of Unified Quality Evaluators (FUNQUE) framework, by enabling computation sharing and by using a transform that is sensitive to visual perception to boost accuracy. Further, we expand the FUNQUE framework to define a collection of improved low-complexity fused-feature models that advance the state-of-the-art of video quality performance with respect to both accuracy and computational efficiency. | ['Alan C. Bovik', 'Ioannis Katsavounidis', 'Cosmin Stejerean', 'Abhinau K. Venkataramanan'] | 2023-04-06 | null | null | null | null | ['video-quality-assessment', 'video-quality-assessment'] | ['computer-vision', 'time-series'] | [ 3.30748290e-01 -5.55938184e-01 -1.11415334e-01 -3.68506730e-01
-1.14304435e+00 -4.49507058e-01 4.40049648e-01 3.80573928e-01
-1.69803292e-01 2.11469516e-01 3.89252961e-01 -1.52952418e-01
-2.29949817e-01 -7.90090501e-01 -4.80516523e-01 -2.99903244e-01
-2.42639765e-01 -3.44062112e-02 3.80663663e-01 -2.08640277e-01
2.35765964e-01 3.16651225e-01 -2.03730297e+00 7.09694505e-01
9.15014565e-01 1.70239305e+00 -6.03337511e-02 1.03688526e+00
1.75929204e-01 1.07514477e+00 -4.47055846e-01 -6.77025437e-01
4.07287151e-01 -4.13318098e-01 -5.75780869e-01 2.66697183e-02
6.30579412e-01 -7.32234895e-01 -2.83707142e-01 9.39899206e-01
4.67193723e-01 -1.88384429e-01 2.52629876e-01 -1.27663195e+00
-1.58367068e-01 2.40194812e-01 -3.88637096e-01 4.17418987e-01
6.32487476e-01 2.51552701e-01 1.14927995e+00 -8.72394383e-01
5.24582624e-01 9.63173807e-01 5.85897386e-01 5.53406868e-03
-1.14899850e+00 -5.53770959e-01 -2.07573429e-01 8.19981575e-01
-1.27845120e+00 -7.42173374e-01 5.37192166e-01 -3.86134744e-01
1.05730772e+00 3.69153857e-01 9.83203292e-01 5.05787492e-01
3.15143377e-01 7.27795720e-01 8.55167568e-01 -2.34884307e-01
4.67440367e-01 -2.57721692e-01 -2.88331181e-01 5.09868622e-01
9.01522562e-02 1.02407321e-01 -1.02020872e+00 -2.02005833e-01
6.74590468e-01 -2.24031031e-01 -3.01197052e-01 -5.77820063e-01
-9.80766833e-01 6.17050767e-01 6.44937009e-02 1.30148306e-01
-6.10824406e-01 2.94596642e-01 5.18070161e-01 5.67547321e-01
5.79889715e-01 4.29425724e-02 -2.84042835e-01 -7.99074769e-01
-1.67473745e+00 4.57456499e-01 6.20647252e-01 6.80181623e-01
8.72766793e-01 1.09014630e-01 -2.28422210e-01 6.07538998e-01
4.89397973e-01 3.96128088e-01 9.51542407e-02 -1.41529751e+00
3.71620923e-01 6.78640127e-01 -1.61056016e-02 -1.13563180e+00
-1.69938892e-01 -5.36819637e-01 -5.96698523e-01 6.77011371e-01
7.90641978e-02 4.07273591e-01 -5.20536900e-01 1.42485058e+00
2.51385033e-01 2.65777349e-01 -1.45673618e-01 1.09010983e+00
4.75203007e-01 6.25126064e-01 -1.30503774e-02 -4.47548270e-01
1.27419162e+00 -8.86541307e-01 -6.80053651e-01 2.71131307e-01
4.13243860e-01 -8.92439365e-01 7.06886292e-01 9.59353089e-01
-1.56874800e+00 -7.95030117e-01 -1.40162289e+00 -1.47477373e-01
5.72782457e-02 -4.26404655e-01 4.72916722e-01 8.80732894e-01
-1.41594529e+00 8.74461591e-01 -9.25197542e-01 2.79894546e-02
6.79232240e-01 4.78405893e-01 -2.35108733e-01 -3.17071170e-01
-8.26370239e-01 7.26140380e-01 9.09446105e-02 -4.79098022e-01
-9.65932906e-01 -1.04066730e+00 -6.34489954e-01 3.99521679e-01
4.80264217e-01 -9.20888662e-01 1.21785855e+00 -1.21657526e+00
-1.60141766e+00 4.19269174e-01 -1.13255620e-01 -4.45852995e-01
5.95677316e-01 -4.90716010e-01 -6.64976180e-01 6.91229165e-01
-3.15605164e-01 4.81500447e-01 1.08021462e+00 -1.03080964e+00
-9.24262404e-01 -2.43383318e-01 1.18824534e-01 1.73426002e-01
-4.37964588e-01 3.46564442e-01 -7.17802107e-01 -5.84580779e-01
-5.10090105e-02 -3.19773316e-01 -2.30232719e-02 6.07140839e-01
4.55996662e-01 4.12669778e-02 8.26187551e-01 -7.75945246e-01
1.81614900e+00 -2.09156823e+00 3.88223439e-01 2.48120800e-01
5.80024123e-01 5.63720763e-01 -6.41378947e-03 2.62837023e-01
3.94560248e-01 1.01527438e-01 -1.21901043e-01 -5.41505873e-01
-5.85996062e-02 6.46758080e-02 1.40997946e-01 3.35247487e-01
3.94289762e-01 5.79828858e-01 -9.03304458e-01 -6.37746751e-01
4.99386013e-01 7.25245297e-01 -1.00460339e+00 4.09465432e-01
-5.80897219e-02 8.15935880e-02 -1.31718308e-01 8.06174695e-01
7.11837232e-01 -3.86357248e-01 1.08258948e-01 -4.84989196e-01
-3.54716808e-01 2.08095089e-01 -1.37988818e+00 2.10402155e+00
-2.83562064e-01 4.35775489e-01 1.43778309e-01 -6.16908967e-01
6.07267737e-01 6.23177230e-01 8.98181915e-01 -7.20205605e-01
1.97779641e-01 2.88319647e-01 -1.24520645e-01 -4.26707417e-01
7.35636413e-01 2.64184862e-01 3.09436232e-01 3.12826395e-01
3.64286810e-01 -9.69662741e-02 4.61542428e-01 2.02725232e-01
1.39821553e+00 4.80562687e-01 4.48601335e-01 -1.75884128e-01
5.47127843e-01 -2.17872798e-01 6.03778601e-01 3.40086102e-01
-5.94570816e-01 7.59717464e-01 2.59501934e-01 -3.75987202e-01
-1.38899851e+00 -1.27871239e+00 -1.73093565e-02 1.03736031e+00
-1.01832813e-02 -9.67887163e-01 -7.98231244e-01 -2.92717636e-01
-1.07467875e-01 -1.55343749e-02 -1.52742431e-01 1.47256739e-02
-5.16217709e-01 -3.66713822e-01 2.69547522e-01 3.16977888e-01
4.30957168e-01 -4.45748359e-01 -1.10325325e+00 5.35890400e-01
-1.32391199e-01 -1.25465322e+00 -2.49774024e-01 -2.87066072e-01
-6.73619807e-01 -9.89390433e-01 -6.21354461e-01 6.49480149e-02
-1.97478980e-01 3.64046305e-01 1.61794436e+00 3.09282959e-01
-1.80938035e-01 5.51195443e-01 -7.13153064e-01 -1.90108940e-02
-4.49486226e-01 -2.99118966e-01 -1.39415413e-01 1.73350349e-01
3.96350361e-02 -9.09286320e-01 -1.01968265e+00 5.09351343e-02
-1.06665945e+00 3.54413956e-01 4.04322773e-01 3.90127152e-01
7.29247332e-01 1.46786049e-01 3.18652660e-01 -2.98718035e-01
2.76054293e-01 -5.12857199e-01 -7.13623106e-01 1.09871887e-02
-8.45680594e-01 -5.55874854e-02 6.31787837e-01 -9.49484259e-02
-8.49532366e-01 -3.00202668e-01 -2.26127610e-01 -6.11098826e-01
1.41766816e-01 4.95939881e-01 -1.51830509e-01 -3.40975642e-01
3.67628276e-01 -4.90483344e-02 -5.00545129e-02 -4.37159330e-01
3.74370575e-01 5.83808661e-01 6.40152931e-01 -2.34783575e-01
6.24226630e-01 4.36499745e-01 3.39491576e-01 -5.26594996e-01
-5.02614796e-01 -6.71710730e-01 -4.23686534e-01 -6.75282955e-01
8.56799841e-01 -1.17406452e+00 -7.08866656e-01 3.19110483e-01
-9.85524654e-01 9.40133482e-02 -1.16509818e-01 4.58082646e-01
-7.90582359e-01 5.99551797e-01 -5.93772054e-01 -7.03187346e-01
-5.04507005e-01 -1.31632948e+00 1.08672130e+00 5.44547029e-02
-7.29868412e-02 -6.22586429e-01 7.97025561e-02 4.53839511e-01
7.16026902e-01 2.15231255e-01 7.53226280e-01 6.68929368e-02
-9.27731276e-01 1.56239614e-01 -4.54396456e-01 6.14902854e-01
-2.28917018e-01 2.01261207e-01 -9.90783274e-01 -3.59924644e-01
-8.46023709e-02 -1.85793757e-01 6.12053216e-01 3.11029375e-01
8.98655355e-01 1.19843803e-01 1.71714932e-01 8.39773595e-01
1.75717974e+00 -9.32564680e-03 6.88937008e-01 2.68379182e-01
3.86849374e-01 2.10651293e-01 6.21005416e-01 1.07692611e+00
6.29104376e-01 1.05180180e+00 6.04083836e-01 -5.37492000e-02
-5.14297426e-01 8.11799467e-02 4.70929921e-01 1.15793085e+00
-4.42346692e-01 -2.07109496e-01 -7.97792494e-01 2.77399689e-01
-1.85224044e+00 -1.00634396e+00 6.79946095e-02 2.32312822e+00
4.89435077e-01 1.49132490e-01 2.88021058e-01 8.38872969e-01
2.33856097e-01 1.21843427e-01 -2.96784818e-01 -4.17783350e-01
-1.91614535e-02 5.59086621e-01 2.97910661e-01 2.77160615e-01
-9.66892183e-01 4.89178419e-01 6.51184702e+00 8.56058776e-01
-9.63900208e-01 2.25772634e-01 4.37590539e-01 -1.89960241e-01
-2.76712030e-01 -1.13404654e-01 -3.76770914e-01 4.52810645e-01
1.38778973e+00 -2.52240777e-01 6.40927851e-01 7.30371058e-01
4.75093812e-01 -3.31285179e-01 -1.19374466e+00 1.34452498e+00
1.98417798e-01 -1.37908578e+00 9.46952254e-02 6.00693189e-02
5.21905065e-01 4.01121452e-02 1.76615536e-01 3.58909997e-03
-1.93598285e-01 -7.70547628e-01 9.26753759e-01 5.93943775e-01
7.65203893e-01 -8.04970324e-01 7.25262821e-01 -1.00984022e-01
-1.65514421e+00 -1.61078230e-01 -1.76044598e-01 -2.18797401e-01
5.97436786e-01 7.98831165e-01 -1.04365483e-01 8.63433659e-01
1.00991225e+00 7.30311632e-01 -6.87462568e-01 1.20704842e+00
1.96990639e-01 5.12084126e-01 -2.14451239e-01 5.06879866e-01
-3.69245782e-02 4.98717725e-02 3.90659779e-01 1.02761328e+00
6.46515489e-01 1.19005755e-01 2.49715179e-01 5.59592426e-01
6.02622814e-02 2.22862020e-01 -3.69821750e-02 4.33687009e-02
3.76986474e-01 1.21984315e+00 -3.66225094e-01 -5.79161942e-01
-7.34709382e-01 9.06799674e-01 2.08049014e-01 8.66941269e-03
-9.25156593e-01 9.10629630e-02 1.05743217e+00 1.06558122e-01
4.89464968e-01 -3.19858640e-01 -2.81501591e-01 -1.27613938e+00
1.24333739e-01 -1.06620085e+00 1.49546668e-01 -6.68658614e-01
-7.84898162e-01 5.34059286e-01 -2.22782716e-01 -1.71050990e+00
-3.67934376e-01 -4.20282900e-01 3.27937752e-02 6.03045821e-01
-1.99839532e+00 -1.06083930e+00 -5.45710981e-01 7.11495161e-01
5.73884726e-01 -1.18264761e-02 8.45210075e-01 8.55018079e-01
-1.05298296e-01 5.72971761e-01 -5.80330566e-02 -5.08765996e-01
5.51728964e-01 -7.91065454e-01 1.90735757e-01 1.00884998e+00
1.21430531e-01 -3.78239304e-02 7.89770067e-01 -2.44406015e-01
-1.90435934e+00 -8.24185193e-01 5.98011017e-01 -6.82511330e-02
5.18131137e-01 -8.25252980e-02 -8.24163914e-01 -1.89503655e-02
1.30199671e-01 3.58963937e-01 8.80025029e-01 -6.17075898e-02
-5.89358866e-01 -2.76174754e-01 -1.15434253e+00 3.85879517e-01
1.04282534e+00 -6.36042655e-01 -9.35973451e-02 -2.11315513e-01
6.29464447e-01 -2.16318235e-01 -1.46463597e+00 5.55091619e-01
8.16854239e-01 -1.78143466e+00 9.54702973e-01 8.82366300e-02
6.98090017e-01 -5.65374792e-01 -6.95733011e-01 -9.24559951e-01
-5.18188536e-01 -8.05807173e-01 -7.50105560e-01 1.04285514e+00
-1.04997404e-01 4.52070450e-03 7.78150618e-01 3.98327321e-01
-3.43549430e-01 -7.98410833e-01 -1.12835324e+00 -6.45174265e-01
-4.70125943e-01 -9.11335230e-01 6.88563228e-01 3.33522528e-01
1.16811030e-01 -1.86751690e-02 -5.49413919e-01 9.84064396e-03
5.84776103e-01 -1.22966453e-01 7.69374609e-01 -1.08479452e+00
-7.36067891e-01 -6.31864786e-01 -1.12537980e+00 -1.09144235e+00
-4.80837882e-01 -5.39557815e-01 -2.72596657e-01 -1.42681479e+00
2.20650986e-01 -1.96489245e-01 -6.05242014e-01 1.85556728e-02
-1.50981352e-01 7.47656643e-01 6.45857692e-01 1.69065341e-01
-1.14063871e+00 4.60321635e-01 9.82747078e-01 2.22532496e-01
2.28945747e-01 -3.89108181e-01 -3.23218912e-01 5.19469917e-01
3.52547318e-01 -2.00791538e-01 -3.32321465e-01 -6.37495041e-01
5.17322779e-01 4.00813162e-01 1.37275741e-01 -1.68498921e+00
1.45524740e-01 1.02846742e-01 2.70544440e-01 -5.20242691e-01
6.41032279e-01 -9.54590917e-01 3.26082885e-01 4.04558390e-01
2.49319356e-02 4.05147582e-01 -8.50046277e-02 2.72598863e-01
-4.63134468e-01 2.40513533e-01 8.29666674e-01 2.84228891e-01
-8.97577167e-01 4.64990824e-01 -2.54756957e-01 -1.32744700e-01
8.98345947e-01 -4.15774375e-01 -1.86719626e-01 -5.20650387e-01
-3.23246181e-01 -1.12784982e-01 7.51958609e-01 1.47958577e-01
7.89606571e-01 -1.30951798e+00 -8.49522591e-01 2.53839493e-01
1.22403443e-01 -5.12328327e-01 4.69612122e-01 7.46427417e-01
-6.72746241e-01 1.07981607e-01 -3.12477171e-01 -7.67745912e-01
-1.43413401e+00 6.51788056e-01 -1.53343812e-01 -5.25043249e-01
-3.96825910e-01 6.46600246e-01 -3.20475966e-01 3.95485759e-01
1.92383071e-03 -1.33913115e-01 -1.13613658e-01 -1.66600943e-01
9.26272094e-01 9.11398351e-01 3.92437518e-01 -8.88985634e-01
-1.58159584e-01 5.27453959e-01 1.70713872e-01 -1.62426367e-01
1.12362552e+00 -4.03312713e-01 5.98202311e-02 3.04729432e-01
1.17861390e+00 -2.37637475e-01 -1.39360547e+00 -1.44027814e-01
-9.77037940e-03 -8.92270982e-01 4.85662103e-01 -7.63970017e-01
-9.99518812e-01 7.10077703e-01 1.07950497e+00 1.21997811e-01
1.67091286e+00 -4.19898450e-01 8.95336211e-01 -3.61363709e-01
6.89531684e-01 -1.05614805e+00 1.40538216e-01 2.42977858e-01
5.31786799e-01 -1.26515651e+00 4.05102313e-01 -5.21755934e-01
-9.55551043e-02 1.10902357e+00 1.81502432e-01 -8.79724231e-03
7.19449520e-01 4.34858769e-01 6.44370541e-02 1.31185949e-01
-1.14858341e+00 -1.64513990e-01 4.45536315e-01 5.25623083e-01
6.56312346e-01 3.52211739e-03 -4.40044671e-01 3.36818516e-01
-1.03054633e-02 3.65174681e-01 3.20304602e-01 1.04030454e+00
-5.59202969e-01 -1.35120726e+00 -2.02681348e-01 4.80957597e-01
-6.88288391e-01 -3.75573844e-01 5.90932131e-01 2.71444023e-01
4.57693011e-01 1.31482875e+00 4.32679728e-02 -6.00467145e-01
2.01635897e-01 -3.01765531e-01 5.99806130e-01 -3.14175099e-01
-8.53559434e-01 2.35881537e-01 -3.10477316e-02 -1.18704903e+00
-6.72021151e-01 -5.45647800e-01 -8.74599397e-01 -4.99010205e-01
-2.20204350e-02 -1.44225746e-01 8.00661862e-01 9.18694079e-01
6.16723597e-01 6.32946253e-01 6.47858977e-01 -1.07861269e+00
-4.11735713e-01 -6.92868233e-01 -3.34838897e-01 5.38230836e-01
3.91509831e-01 -6.45393372e-01 -8.09240248e-03 2.26623625e-01] | [11.657958030700684, -1.8006197214126587] |
cd3a4fc5-1b87-42cc-a423-7fe4b09a5d6c | case-based-reasoning-with-language-models-for | 2301.11879 | null | https://arxiv.org/abs/2301.11879v2 | https://arxiv.org/pdf/2301.11879v2.pdf | Case-Based Reasoning with Language Models for Classification of Logical Fallacies | The ease and speed of spreading misinformation and propaganda on the Web motivate the need to develop trustworthy technology for detecting fallacies in natural language arguments. However, state-of-the-art language modeling methods exhibit a lack of robustness on tasks like logical fallacy classification that require complex reasoning. In this paper, we propose a Case-Based Reasoning method that classifies new cases of logical fallacy by language-modeling-driven retrieval and adaptation of historical cases. We design four complementary strategies to enrich input representation for our model, based on external information about goals, explanations, counterarguments, and argument structure. Our experiments in in-domain and out-of-domain settings indicate that Case-Based Reasoning improves the accuracy and generalizability of language models. Our ablation studies suggest that representations of similar cases have a strong impact on the model performance, that models perform well with fewer retrieved cases, and that the size of the case database has a negligible effect on the performance. Finally, we dive deeper into the relationship between the properties of the retrieved cases and the model performance. | ['Alain Mermoud', 'Hông-Ân Sandlin', 'Filip Ilievski', 'Zhivar Sourati'] | 2023-01-27 | null | null | null | null | ['logical-fallacies'] | ['miscellaneous'] | [-2.03699544e-01 4.14491653e-01 -6.77809596e-01 -3.80285054e-01
-8.47992778e-01 -7.20396161e-01 1.23894906e+00 7.00568497e-01
-3.29677135e-01 6.08228564e-01 6.51524067e-01 -1.10108435e+00
-3.21919262e-01 -8.71378481e-01 -6.92614973e-01 2.54969537e-01
1.91287547e-01 3.37076813e-01 5.18984139e-01 -5.14819622e-01
5.58852017e-01 2.33188853e-01 -1.33700466e+00 7.72356153e-01
1.02156436e+00 4.53874409e-01 -4.13083076e-01 1.47856042e-01
-5.77013850e-01 1.49211812e+00 -7.74659157e-01 -9.67392206e-01
-2.52668038e-02 -4.92891073e-01 -1.12277853e+00 -5.27373075e-01
3.04137379e-01 -3.78193617e-01 -2.83097357e-01 9.67321515e-01
5.65834641e-02 -3.23147982e-01 6.82628870e-01 -1.23202109e+00
-6.80386066e-01 1.26528776e+00 -3.62481117e-01 4.46128577e-01
6.81005776e-01 -2.45263264e-01 9.68015492e-01 -5.25055230e-01
1.13617742e+00 1.57496154e+00 9.53613579e-01 5.10177374e-01
-1.02183485e+00 -5.95538855e-01 6.12219036e-01 3.93672705e-01
-8.68355572e-01 -4.62909251e-01 8.28521967e-01 -5.26710570e-01
1.01060140e+00 4.50141311e-01 4.50925440e-01 1.33235693e+00
2.48225346e-01 5.82582653e-01 1.11292481e+00 -7.19586074e-01
3.90437573e-01 3.97742361e-01 6.20556712e-01 6.39859676e-01
1.08033407e+00 1.21059753e-02 -6.02751791e-01 -9.57078815e-01
3.23548347e-01 -2.64374435e-01 8.01905915e-02 -1.81933224e-01
-6.36375606e-01 1.22322702e+00 3.23071539e-01 5.80858529e-01
-3.30007553e-01 6.48061335e-02 3.87621433e-01 3.08284938e-01
7.26317108e-01 6.29991353e-01 -4.16025043e-01 2.07637111e-03
-8.93079340e-01 7.22189546e-01 1.11309350e+00 5.59347689e-01
1.77455470e-01 -3.38243604e-01 4.11308818e-02 6.63937211e-01
3.62167776e-01 4.08963621e-01 4.93794411e-01 -8.50775957e-01
7.55163908e-01 8.63453209e-01 3.71930957e-01 -1.32520950e+00
-3.37671161e-01 -4.32529390e-01 1.28182516e-01 9.70165059e-02
6.57063425e-01 1.05839275e-01 -5.36347210e-01 1.79151082e+00
2.05571592e-01 -5.55767536e-01 2.65759349e-01 6.99059486e-01
6.12801373e-01 3.23894531e-01 4.02330190e-01 -2.35942081e-02
1.36069298e+00 -4.48422492e-01 -6.95295572e-01 -7.56711662e-01
9.27444398e-01 -8.29030275e-01 1.10406756e+00 7.83719122e-02
-1.04008925e+00 2.22765937e-01 -1.05956841e+00 -1.08825313e-02
-3.85145992e-01 -2.42866546e-01 8.46725881e-01 5.05444586e-01
-2.52857357e-01 5.15341461e-01 -7.50858128e-01 -4.55261797e-01
4.53041345e-01 -4.42054987e-01 -7.69594610e-02 -8.00182968e-02
-1.48646951e+00 1.42201257e+00 2.88863808e-01 -3.41812074e-01
-2.70890594e-01 -5.30186176e-01 -7.98547745e-01 1.01570807e-01
5.76421201e-01 -6.04529142e-01 1.29840410e+00 -9.10609543e-01
-6.80150688e-01 7.63353765e-01 -1.59018159e-01 -6.28323495e-01
8.81406605e-01 -2.24361762e-01 -5.95283628e-01 -2.22337786e-02
4.79300827e-01 -9.34057310e-02 5.34832537e-01 -1.30759966e+00
-4.80289191e-01 -4.18681443e-01 3.50063920e-01 -1.55524060e-01
-2.39327207e-01 3.71721238e-01 1.84088692e-01 -6.27359033e-01
3.59074324e-01 -5.29257596e-01 -2.25996841e-02 -4.96890955e-02
-1.66372895e-01 -4.01426286e-01 4.77183104e-01 -8.80301476e-01
1.64744294e+00 -1.83218372e+00 -5.42952001e-01 4.03011501e-01
1.23547874e-01 1.10307865e-01 5.86896911e-02 7.50240326e-01
1.61263764e-01 8.92573714e-01 2.72193462e-01 3.12953651e-01
5.73951937e-02 6.29006550e-02 -7.47856021e-01 1.52731806e-01
3.61017175e-02 8.19272518e-01 -9.51792836e-01 -7.16839671e-01
-2.38609657e-01 9.52091068e-02 -6.67156816e-01 -1.94175452e-01
-3.59448284e-01 -3.58930618e-01 -8.48337173e-01 4.01381820e-01
4.08101201e-01 -5.48886478e-01 6.59635305e-01 4.87737507e-02
-2.04786919e-02 1.28231156e+00 -9.03366327e-01 1.21248007e+00
-3.57144803e-01 4.54556823e-01 -1.43788140e-02 -3.96414638e-01
6.11914277e-01 9.05743018e-02 -2.60125577e-01 -9.32717800e-01
-4.95310277e-02 4.30932999e-01 2.20637277e-01 -6.45116270e-01
2.14624807e-01 -2.98911303e-01 -9.69985425e-02 1.10121703e+00
-4.46429670e-01 -2.79704817e-02 3.63701195e-01 6.12895608e-01
1.07703149e+00 7.11350366e-02 5.75832129e-01 -3.17578316e-01
9.74606574e-02 6.15164757e-01 4.63431239e-01 1.14195812e+00
2.40014166e-01 1.11521846e-02 5.82069993e-01 -6.99026048e-01
-8.65012527e-01 -7.24982142e-01 -1.87450543e-01 8.71958673e-01
1.15574799e-01 -8.39316845e-01 -5.27668536e-01 -9.04036105e-01
2.06019461e-01 1.40433872e+00 -5.19535899e-01 -2.81339020e-01
-5.96996009e-01 -7.34885395e-01 7.18332350e-01 4.88957375e-01
3.86448145e-01 -6.58008516e-01 -1.02698517e+00 3.86141568e-01
-4.83951807e-01 -9.07341957e-01 3.87409508e-01 -8.07381123e-02
-9.25346732e-01 -1.35810494e+00 1.40127569e-01 -2.70770848e-01
5.44743001e-01 1.20344661e-01 1.24352705e+00 6.15777552e-01
6.39021769e-02 9.56835300e-02 -3.86765271e-01 -7.48979568e-01
-9.54903126e-01 -9.01884958e-02 -1.29945278e-01 -6.61992967e-01
6.39494061e-01 -3.43235880e-01 1.12413928e-01 4.44385000e-02
-8.82224083e-01 6.24258667e-02 1.78484440e-01 8.04342330e-01
-1.80093035e-01 -1.15315551e-02 4.22121435e-01 -1.30011094e+00
1.34307504e+00 -4.82915968e-01 -4.76932764e-01 6.26911104e-01
-9.74705219e-01 5.30575752e-01 3.33210081e-01 -5.19930959e-01
-1.34611559e+00 -8.61655712e-01 1.35146067e-01 2.90576369e-01
2.58609891e-01 9.92751539e-01 3.49680096e-01 2.89266706e-01
1.32194090e+00 -3.93593848e-01 1.02834731e-01 -4.58392084e-01
3.56440157e-01 5.31051874e-01 1.20359041e-01 -8.71028483e-01
8.14739466e-01 5.20799756e-01 -4.79196131e-01 -3.40967059e-01
-1.06059074e+00 6.94070309e-02 -3.14626694e-02 8.52842927e-02
1.30536065e-01 -6.82951510e-01 -2.42082834e-01 -3.90462838e-02
-1.38597810e+00 -2.62166649e-01 -9.18519273e-02 4.30088550e-01
-1.53511688e-01 3.63492429e-01 -6.83640599e-01 -1.04491210e+00
-2.09050894e-01 -6.35008216e-01 4.45981622e-01 -1.89015239e-01
-1.00999045e+00 -1.10740447e+00 1.35622881e-02 5.36014378e-01
5.64169168e-01 8.63837004e-02 1.68738461e+00 -1.16930377e+00
-2.72623956e-01 -4.50042307e-01 -2.95461714e-02 -2.39771217e-01
-5.27538499e-03 -6.44317269e-02 -7.90253520e-01 2.30555072e-01
1.73039705e-01 -3.58662874e-01 6.59815669e-01 5.19208722e-02
6.35252535e-01 -9.66831386e-01 -5.67945600e-01 -6.84264228e-02
1.13427842e+00 7.14773163e-02 4.46174502e-01 8.58603716e-01
1.24519793e-02 8.34402204e-01 6.06208444e-01 3.74268025e-01
2.29934484e-01 4.79648203e-01 2.08647829e-02 1.10083975e-01
-3.63288745e-02 -7.83241868e-01 7.64454901e-02 7.99805447e-02
2.62873024e-01 -2.70440966e-01 -1.33031368e+00 5.38435757e-01
-2.04090619e+00 -1.31332612e+00 -7.47899059e-03 1.96843004e+00
9.05906558e-01 6.39775574e-01 -5.70088625e-02 2.20959842e-01
5.36519229e-01 1.71620876e-01 -5.43744005e-02 -5.29475093e-01
-1.47978976e-01 -1.71487153e-01 2.03028157e-01 8.63111019e-01
-5.58233440e-01 8.53566766e-01 6.84937954e+00 6.27738655e-01
-1.03157270e+00 1.00200415e-01 5.68773746e-01 -9.07850862e-02
-1.11822331e+00 4.09838229e-01 -6.23908937e-01 3.85757178e-01
7.46809423e-01 -4.08872008e-01 2.21402258e-01 9.08034086e-01
1.20955773e-01 -6.12516403e-02 -1.11887228e+00 3.54160458e-01
2.34583810e-01 -1.86356831e+00 3.95542443e-01 -9.22570452e-02
3.58735591e-01 -1.70590341e-01 -2.67954975e-01 2.19869539e-01
5.95056772e-01 -7.13412344e-01 1.23400164e+00 2.51775146e-01
2.43286103e-01 -3.08783323e-01 8.64714622e-01 5.75164497e-01
-2.68482953e-01 -3.78789812e-01 -3.78415324e-02 -4.46710736e-01
1.82518587e-01 3.72520477e-01 -9.84982550e-01 2.87126631e-01
5.40106535e-01 1.82086393e-01 -8.23644757e-01 5.26928365e-01
-5.53095162e-01 9.16393399e-01 -4.23380464e-01 -3.37740332e-01
1.99320897e-01 2.59113669e-01 4.79843915e-01 1.08498657e+00
-1.06356457e-01 2.05667123e-01 -1.70097779e-02 1.10980284e+00
7.87502974e-02 -2.18402911e-02 -7.57156849e-01 -1.99628472e-01
7.69275904e-01 6.17520332e-01 -4.98389661e-01 -4.14767623e-01
-4.05894279e-01 1.00569338e-01 3.92309487e-01 3.67467940e-01
-5.89076340e-01 7.48195627e-04 3.44601840e-01 6.82757080e-01
-1.90020129e-02 6.73424304e-02 -6.94199681e-01 -1.19085670e+00
4.98101652e-01 -1.34773409e+00 7.78150856e-01 -8.05150747e-01
-1.48707354e+00 5.64634442e-01 4.82804805e-01 -7.48506308e-01
-8.80298018e-01 -4.49375510e-01 -6.40085757e-01 7.21616983e-01
-1.32681799e+00 -1.07207811e+00 1.88231580e-02 8.88793096e-02
2.39897773e-01 -7.42309168e-03 7.10141480e-01 -1.84419498e-01
-1.34000838e-01 4.01899517e-01 -3.46841127e-01 2.90370584e-01
6.41429901e-01 -6.55615211e-01 5.91182768e-01 1.03829348e+00
2.87664473e-01 1.30248725e+00 9.86298561e-01 -1.08712316e+00
-9.62285578e-01 -5.18720090e-01 1.51626968e+00 -8.85123610e-01
1.02624416e+00 -2.63257802e-01 -1.06108630e+00 7.35462129e-01
-6.96779862e-02 -3.82855147e-01 6.30743325e-01 6.95472598e-01
-1.15564573e+00 2.35119969e-01 -1.27188158e+00 8.97347152e-01
1.11987054e+00 -7.94506907e-01 -1.19828749e+00 3.78524214e-01
3.76583159e-01 -1.26447290e-01 -1.81437194e-01 1.96295634e-01
6.83244765e-01 -6.10129833e-01 8.88944089e-01 -1.41008961e+00
6.21960342e-01 6.09152727e-02 -1.33580253e-01 -8.24093044e-01
-3.86892796e-01 -4.13609147e-01 -1.40177712e-01 1.09395647e+00
9.37259495e-01 -7.83105731e-01 4.51689899e-01 1.17497861e+00
4.25016463e-01 -4.44726199e-01 -8.54350448e-01 -7.67664790e-01
3.81799221e-01 -4.50412631e-01 7.47065127e-01 1.06426716e+00
5.73224306e-01 5.11109769e-01 1.27513111e-01 -7.94666708e-02
4.01825011e-01 4.77770299e-01 5.41766524e-01 -1.51449859e+00
-1.43681258e-01 -3.60593826e-01 -8.48556384e-02 -5.51543653e-01
3.92212749e-01 -8.31785560e-01 -5.51238894e-01 -1.81799066e+00
3.89716685e-01 -6.12414718e-01 4.63851020e-02 5.93123436e-01
-1.70249760e-01 -4.42769587e-01 1.98138058e-01 4.29614425e-01
-4.83811229e-01 -6.01968952e-02 4.64504361e-01 -1.35296971e-01
-1.71980128e-01 -2.85857558e-01 -1.07015383e+00 1.04447126e+00
7.07886279e-01 -8.67255628e-01 -4.21672434e-01 -6.33084774e-01
7.00540423e-01 -2.80689672e-02 5.07464349e-01 -5.42151451e-01
1.79501072e-01 -5.23127854e-01 2.10926950e-01 -1.60106421e-01
5.63441850e-02 -6.07235610e-01 7.27784485e-02 7.45642424e-01
-8.94591987e-01 3.39716196e-01 2.52147645e-01 6.06059313e-01
1.75996181e-02 -6.04590416e-01 3.63962114e-01 -2.21456274e-01
-3.95613313e-01 -5.59880793e-01 -5.86268723e-01 4.65849936e-01
4.22320127e-01 -1.63603108e-02 -8.41854990e-01 -6.25720024e-01
-2.34119609e-01 -1.50610268e-01 5.90109408e-01 6.25126362e-01
2.96828449e-01 -1.09838831e+00 -9.32377696e-01 -1.13630727e-01
3.20153236e-01 -7.83893585e-01 -3.38440537e-01 5.07912815e-01
-5.13163388e-01 4.75112677e-01 -1.99037921e-02 -1.39310494e-01
-1.26632512e+00 5.14551759e-01 2.93342143e-01 -5.14499485e-01
-5.23193836e-01 4.40488279e-01 -2.16961652e-01 -2.59760886e-01
1.01252936e-01 -2.86070049e-01 -3.01405452e-02 -2.07791403e-02
6.89262569e-01 2.11896718e-01 2.07775205e-01 -1.20937854e-01
-6.58790946e-01 -1.42633960e-01 -5.55473864e-01 -4.09481943e-01
1.30718148e+00 1.66201860e-01 -9.66317356e-02 3.54403466e-01
4.86595333e-01 3.70488405e-01 -6.01795673e-01 -2.62975007e-01
5.06785214e-01 -5.43450177e-01 -9.07703303e-03 -1.32435417e+00
-3.04969579e-01 4.65544343e-01 -4.66893837e-02 4.40253109e-01
3.81637305e-01 2.85842091e-01 2.81860024e-01 6.12016380e-01
5.65771461e-01 -1.05794227e+00 -1.54545799e-01 4.67507720e-01
1.03720021e+00 -9.87161219e-01 3.59482169e-01 -6.62812114e-01
-5.62535584e-01 9.58221376e-01 4.63239551e-01 2.35587046e-01
3.58640969e-01 4.17860091e-01 3.38136286e-01 -5.29421687e-01
-9.77305651e-01 3.52024406e-01 -2.73276470e-03 3.02974254e-01
6.92884386e-01 -9.06690061e-02 -9.15907979e-01 9.60982621e-01
-3.27060938e-01 -7.68625438e-02 5.15472651e-01 1.29927862e+00
-2.69922286e-01 -9.56324279e-01 -5.03349125e-01 4.50114429e-01
-7.41133332e-01 -3.57317299e-01 -1.15399337e+00 1.15505791e+00
-3.68476331e-01 1.25546098e+00 -1.53337762e-01 -1.25958741e-01
2.43915275e-01 4.10598487e-01 3.84167731e-01 -4.82170969e-01
-9.42411840e-01 -4.17778671e-01 9.39813077e-01 -5.80051184e-01
-8.23729858e-02 -5.61926067e-01 -1.18096566e+00 -5.67789614e-01
-5.98856032e-01 5.35294354e-01 6.28737986e-01 1.13462651e+00
6.14723504e-01 -1.44923121e-01 -1.25225365e-01 2.02523351e-01
-1.23077917e+00 -8.49666178e-01 -1.94491167e-02 8.12694252e-01
8.24208558e-02 -7.30899155e-01 -5.78189790e-01 -2.91492969e-01] | [9.802037239074707, 8.270654678344727] |
522cb22c-5a46-4662-8b54-274cefe53729 | mosaic-masked-optimisation-with-selective | 2306.00906 | null | https://arxiv.org/abs/2306.00906v1 | https://arxiv.org/pdf/2306.00906v1.pdf | MOSAIC: Masked Optimisation with Selective Attention for Image Reconstruction | Compressive sensing (CS) reconstructs images from sub-Nyquist measurements by solving a sparsity-regularized inverse problem. Traditional CS solvers use iterative optimizers with hand crafted sparsifiers, while early data-driven methods directly learn an inverse mapping from the low-dimensional measurement space to the original image space. The latter outperforms the former, but is restrictive to a pre-defined measurement domain. More recent, deep unrolling methods combine traditional proximal gradient methods and data-driven approaches to iteratively refine an image approximation. To achieve higher accuracy, it has also been suggested to learn both the sampling matrix, and the choice of measurement vectors adaptively. Contrary to the current trend, in this work we hypothesize that a general inverse mapping from a random set of compressed measurements to the image domain exists for a given measurement basis, and can be learned. Such a model is single-shot, non-restrictive and does not parametrize the sampling process. To this end, we propose MOSAIC, a novel compressive sensing framework to reconstruct images given any random selection of measurements, sampled using a fixed basis. Motivated by the uneven distribution of information across measurements, MOSAIC incorporates an embedding technique to efficiently apply attention mechanisms on an encoded sequence of measurements, while dispensing the need to use unrolled deep networks. A range of experiments validate our proposed architecture as a promising alternative for existing CS reconstruction methods, by achieving the state-of-the-art for metrics of reconstruction accuracy on standard datasets. | ['Dushan N. Wadduwage', 'Chamira U. S. Edussooriya', 'A. Thieshanthan', 'Amashi Niwarthana', 'Tharindu Wickremasinghe', 'Pamuditha Somarathne'] | 2023-06-01 | null | null | null | null | ['image-reconstruction', 'compressive-sensing'] | ['computer-vision', 'computer-vision'] | [ 6.64321303e-01 -9.13577452e-02 -2.04384774e-01 -2.83055872e-01
-7.61957705e-01 -3.99900734e-01 7.40958154e-01 -3.84444535e-01
-3.96141708e-01 5.13017118e-01 3.63105208e-01 -1.80746973e-01
-4.26961809e-01 -5.64901412e-01 -9.05177236e-01 -8.20295632e-01
9.56435427e-02 5.22992313e-01 -2.49455109e-01 -1.03178062e-01
4.07235712e-01 4.43786055e-01 -1.31604409e+00 1.09657496e-01
7.09602416e-01 1.03049970e+00 6.02459013e-01 5.12527287e-01
1.15175113e-01 8.45855474e-01 -1.68275282e-01 -4.91386130e-02
5.97441196e-01 -8.15712571e-01 -3.70903105e-01 3.76119494e-01
5.41413784e-01 -5.30822337e-01 -6.56332016e-01 1.21705723e+00
4.70278770e-01 -8.33132211e-03 3.41238648e-01 -7.00504363e-01
-8.91030669e-01 7.37417698e-01 -2.86517352e-01 1.10738583e-01
4.20130074e-01 1.99008033e-01 9.70761597e-01 -1.10388887e+00
7.79879868e-01 7.66831994e-01 8.72111320e-01 3.11559707e-01
-1.46566916e+00 -3.14542830e-01 -6.64011389e-03 4.37243640e-01
-1.52099502e+00 -7.29065478e-01 1.13773620e+00 -2.24642158e-01
4.63632077e-01 4.13662493e-01 9.42122877e-01 1.28616178e+00
-1.60635099e-01 6.77948773e-01 1.27771592e+00 -6.43259048e-01
6.34921193e-01 -7.58692855e-04 -2.87036359e-01 5.66022098e-01
1.33307233e-01 2.09368274e-01 -7.11859882e-01 -1.18768379e-01
9.31451678e-01 1.30455896e-01 -8.07792425e-01 -8.92320037e-01
-1.27982700e+00 9.50160384e-01 4.01237369e-01 5.26877642e-01
-6.90159917e-01 2.33301997e-01 7.86748677e-02 4.56009418e-01
3.11754555e-01 6.29061580e-01 -1.62612930e-01 6.49529137e-03
-1.49646115e+00 5.11670336e-02 9.17707562e-01 6.67696953e-01
9.40172911e-01 3.42892528e-01 -1.98035270e-01 5.79046369e-01
1.80090070e-01 5.61812818e-01 5.57698905e-01 -1.19737983e+00
2.88073301e-01 -5.76456226e-02 1.25665531e-01 -1.19616246e+00
-1.55405663e-02 -8.86636257e-01 -1.11038744e+00 1.08750882e-02
2.30217516e-01 5.98143116e-02 -8.61405075e-01 1.77098989e+00
2.93454409e-01 6.70253873e-01 -1.58512056e-01 1.37972796e+00
1.46391466e-01 4.31736857e-01 -7.53300905e-01 -3.39452922e-01
7.94648647e-01 -7.78397977e-01 -7.17859030e-01 -4.49355096e-01
2.39954188e-01 -6.06307745e-01 1.03697515e+00 6.71313465e-01
-1.01560259e+00 -5.01933932e-01 -1.35491180e+00 -6.16569892e-02
2.12331310e-01 2.39278182e-01 4.12105471e-01 6.05588138e-01
-1.23368633e+00 8.90911281e-01 -8.71444225e-01 -9.97349247e-02
4.49868351e-01 -5.57374991e-02 -2.55424678e-01 -5.82779467e-01
-8.25456917e-01 8.08627665e-01 1.37088031e-01 2.25932524e-01
-1.23806858e+00 -8.01162779e-01 -7.64754355e-01 8.01808760e-02
5.21954775e-01 -7.93533623e-01 7.81951606e-01 -1.00651455e+00
-1.79606247e+00 5.75151861e-01 -2.24737339e-02 -6.74201846e-01
5.93349457e-01 -2.30389833e-01 -1.52135417e-01 4.52462763e-01
-1.66271888e-02 2.16992781e-01 1.50423479e+00 -1.37904286e+00
-1.20557658e-02 -2.85499483e-01 4.99930158e-02 1.00072749e-01
-3.49603981e-01 -7.48030186e-01 -3.61679822e-01 -5.61326027e-01
6.14158809e-01 -1.00061965e+00 -5.82618177e-01 1.56528100e-01
-4.06513035e-01 5.55182040e-01 6.43380761e-01 -6.65634751e-01
1.08513784e+00 -2.12406087e+00 5.58901310e-01 3.76422048e-01
2.34086826e-01 9.76309255e-02 -5.31060100e-01 6.07342660e-01
-9.97930840e-02 -3.37791890e-01 -5.82929730e-01 -6.61171615e-01
-3.10771801e-02 2.33099073e-01 -4.01387423e-01 1.12054324e+00
-1.66652218e-01 8.87310207e-01 -1.01093960e+00 7.98925459e-02
4.82339710e-01 6.94342971e-01 -7.68397212e-01 3.29104334e-01
-1.12838618e-01 8.08535516e-01 -2.86667526e-01 3.27891916e-01
7.97421157e-01 -5.69637477e-01 3.30889463e-01 -5.27095199e-01
-6.95045665e-02 3.39590311e-01 -1.44623256e+00 2.40049314e+00
-6.12143338e-01 5.71188092e-01 4.40798223e-01 -1.76473045e+00
7.72131562e-01 2.45552912e-01 8.46492887e-01 -6.24071538e-01
6.62395656e-02 4.58497733e-01 -2.24675849e-01 -4.74812835e-01
2.66351014e-01 -1.76191047e-01 4.44313079e-01 4.13602978e-01
2.65680403e-01 -4.09771949e-01 4.44783159e-02 1.39345706e-01
1.22616196e+00 2.22817108e-01 3.60777825e-01 -2.69927770e-01
4.51669395e-01 -2.73425370e-01 3.90713274e-01 1.15984666e+00
3.51309001e-01 1.03712559e+00 -8.48147720e-02 -5.32970548e-01
-1.26248181e+00 -8.61944437e-01 -3.86902988e-01 6.33202314e-01
1.62519023e-01 -3.82779360e-01 -6.75181806e-01 -4.07982051e-01
-2.05390751e-01 7.34759092e-01 -5.61281443e-01 -8.44520032e-02
-5.53747237e-01 -4.97307837e-01 1.73345387e-01 -2.07346886e-01
5.59630871e-01 -8.05440247e-01 -8.54598582e-01 4.23812926e-01
-3.67857963e-01 -1.33691120e+00 -4.40383762e-01 2.32653484e-01
-7.81653762e-01 -9.85445142e-01 -8.83921087e-01 -2.83788443e-01
6.97251141e-01 5.24887025e-01 1.12903368e+00 2.73451172e-02
-1.29523128e-01 7.57136583e-01 -5.17998576e-01 1.36584535e-01
-3.45773518e-01 -1.06973045e-01 -8.55221152e-02 5.48988461e-01
-5.58593497e-02 -1.12084472e+00 -8.69531810e-01 -1.45544978e-02
-1.12428176e+00 1.04793668e-01 6.76644623e-01 1.06407440e+00
6.05897784e-01 -3.58610094e-01 1.20485649e-01 -7.20688641e-01
3.83717149e-01 -5.79425752e-01 -4.12869990e-01 3.90171297e-02
-8.11815679e-01 3.12217891e-01 9.36416984e-01 -6.32219076e-01
-5.52908838e-01 2.29729906e-01 -2.15794016e-02 -1.00247920e+00
5.72527759e-02 8.22377443e-01 1.45866856e-01 -3.63121510e-01
8.29593122e-01 8.37323248e-01 3.12471718e-01 -5.11008978e-01
5.23967206e-01 2.91086286e-01 4.98004436e-01 -5.00257075e-01
8.31733465e-01 8.63862514e-01 2.94480175e-02 -8.68220270e-01
-1.04261112e+00 -4.12226558e-01 -5.03673136e-01 -5.22628091e-02
3.90994817e-01 -8.51733565e-01 -2.42271572e-01 1.94382489e-01
-8.92009914e-01 -4.60647166e-01 -7.62637198e-01 6.10714436e-01
-7.34049618e-01 7.48073459e-01 -4.46780384e-01 -5.69043994e-01
-2.08814844e-01 -1.20590246e+00 1.07010758e+00 -3.58540833e-01
1.03882812e-01 -7.95931220e-01 1.24822132e-01 1.93053350e-01
7.55935669e-01 -4.95965853e-02 2.23243952e-01 -5.70885129e-02
-7.94737220e-01 -9.66810584e-02 2.92630717e-02 5.97378612e-01
1.69424736e-03 -6.85508490e-01 -7.68478870e-01 -8.14993620e-01
7.50391185e-01 -3.84045124e-01 9.09779072e-01 6.41620934e-01
1.32381070e+00 -5.34103215e-01 1.65355727e-02 1.32551098e+00
1.61797333e+00 -3.01884919e-01 7.85746455e-01 2.89120466e-01
5.53315639e-01 2.37613097e-01 1.94799051e-01 7.74540842e-01
2.02916980e-01 6.70042396e-01 6.83435202e-01 1.03265062e-01
-2.71158695e-01 -3.11857849e-01 2.25659207e-01 8.46826315e-01
1.61345184e-01 -1.18735164e-01 -4.56850976e-01 4.98009205e-01
-1.67096734e+00 -9.95482802e-01 2.77532756e-01 2.27472281e+00
7.24665642e-01 -2.64358640e-01 -3.35354954e-01 2.01748967e-01
3.42496514e-01 6.44093752e-01 -8.41120362e-01 2.79144526e-01
-2.30947122e-01 3.58826756e-01 6.28907084e-01 5.47126114e-01
-7.30091929e-01 5.21047890e-01 5.68764830e+00 8.09065163e-01
-1.57845569e+00 2.65276611e-01 3.50140870e-01 -2.30764985e-01
-6.30019486e-01 3.10432553e-01 -3.07291687e-01 5.99008679e-01
8.65462363e-01 1.59571096e-01 1.18261433e+00 5.03780007e-01
9.95375216e-02 2.52595007e-01 -1.13774312e+00 1.37484276e+00
2.40495473e-01 -1.86535358e+00 1.00081535e-02 1.49010628e-01
7.42127478e-01 3.21359664e-01 9.88702029e-02 5.97062483e-02
6.50836304e-02 -9.23956454e-01 9.93569732e-01 5.63442647e-01
9.84706044e-01 -1.84428513e-01 2.95210868e-01 6.10432148e-01
-9.09995019e-01 -2.42632776e-01 -4.45588291e-01 -1.20495707e-01
3.90898317e-01 9.79020536e-01 -6.25960946e-01 4.92920637e-01
5.30895174e-01 9.92400110e-01 -1.61120117e-01 7.42828190e-01
-1.83796987e-01 6.82883859e-01 -5.76494277e-01 3.93495053e-01
3.93273652e-01 -6.22398973e-01 8.31141591e-01 9.21027303e-01
9.35566187e-01 1.06631465e-01 2.87699282e-01 1.06033993e+00
1.41171351e-01 -1.62181154e-01 -7.30435073e-01 7.24443197e-02
5.20560265e-01 1.08899152e+00 -3.86641532e-01 -1.90108269e-01
-2.94886231e-01 1.10085881e+00 1.91068575e-01 6.90866709e-01
-4.76079345e-01 2.85390615e-01 2.37730399e-01 3.95379871e-01
6.85119092e-01 -2.61151850e-01 -1.70679346e-01 -1.55589056e+00
2.77947307e-01 -1.30839658e+00 -1.14590125e-02 -6.19789720e-01
-1.11431170e+00 3.67541909e-01 -2.31999233e-01 -1.45475209e+00
-3.79590571e-01 -1.57226652e-01 -3.04885715e-01 6.90101504e-01
-1.72599280e+00 -9.03950274e-01 -3.80255401e-01 6.78658843e-01
4.82408732e-01 -2.56246597e-01 6.76089168e-01 2.34934896e-01
-1.16217665e-01 2.77105510e-01 2.73819745e-01 -2.46712789e-01
2.99510330e-01 -9.80563700e-01 1.08691975e-01 1.12461412e+00
6.37728810e-01 6.92408741e-01 9.28757071e-01 -3.79468858e-01
-1.96955729e+00 -6.65061176e-01 2.57556736e-01 1.04741916e-01
5.52815259e-01 -2.42946044e-01 -9.49487984e-01 6.54712319e-01
2.21432030e-01 5.69939375e-01 4.22856122e-01 -2.13341445e-01
-3.47166568e-01 -2.16184154e-01 -1.08241224e+00 4.18580115e-01
1.11620390e+00 -6.67791784e-01 -2.17742994e-01 3.73188943e-01
3.37516129e-01 -4.30419564e-01 -6.63241684e-01 3.62330824e-01
3.46923292e-01 -1.38837373e+00 1.31802142e+00 -3.20158340e-02
5.28146684e-01 -2.32476637e-01 -5.01914501e-01 -1.47950768e+00
-4.61952657e-01 -8.77665639e-01 -5.48037827e-01 5.25854766e-01
-2.18349434e-02 -5.36101580e-01 1.00208282e+00 8.84090364e-02
-3.85952294e-01 -8.77109110e-01 -1.13340580e+00 -6.31285787e-01
-2.73109168e-01 -5.34683585e-01 6.41414762e-01 1.11660671e+00
-4.51992095e-01 2.20109701e-01 -7.85102308e-01 4.29092348e-01
1.06386542e+00 2.53926635e-01 7.86863983e-01 -8.40350449e-01
-9.22983289e-01 -3.16549331e-01 -3.07307273e-01 -1.70243251e+00
8.43122881e-03 -9.40268576e-01 1.19371071e-01 -1.28954625e+00
-1.00617006e-01 -6.28533125e-01 -3.09119165e-01 -2.02740766e-02
1.39915437e-01 2.52205700e-01 2.89406627e-01 4.20134187e-01
-3.47283661e-01 7.99873233e-01 1.36015451e+00 -1.88365281e-01
5.20895161e-02 -1.64000377e-01 -6.64848328e-01 3.58244658e-01
4.08185333e-01 -4.16605413e-01 -6.93443120e-01 -8.39803755e-01
3.49183530e-01 4.58452463e-01 4.19521332e-01 -1.17210639e+00
5.11780858e-01 -8.42968524e-02 1.92401513e-01 -1.95844516e-01
5.79792142e-01 -9.21716154e-01 4.90651906e-01 3.42558414e-01
-4.16520298e-01 -4.31749135e-01 -5.06417453e-01 7.71079242e-01
-2.25665361e-01 -4.79852647e-01 8.03567052e-01 -3.29614788e-01
-4.58422840e-01 4.74308193e-01 1.27508432e-01 -1.45310774e-01
5.06664097e-01 -4.09227490e-01 2.40021825e-01 -6.16905689e-01
-5.63854277e-01 -2.92579561e-01 5.31177402e-01 -3.68472189e-02
8.01541626e-01 -1.25204003e+00 -8.93409491e-01 6.18986070e-01
-1.60837978e-01 5.25873788e-02 3.09247583e-01 9.86068189e-01
-4.47887391e-01 8.38938728e-02 2.31895670e-02 -7.87519336e-01
-4.39621031e-01 4.53210503e-01 3.44652176e-01 -2.62207806e-01
-1.01471293e+00 7.52272904e-01 -1.54394016e-01 -4.97098982e-01
1.50616452e-01 -8.73539671e-02 1.44392744e-01 -1.83263943e-01
5.61335802e-01 1.93459377e-01 2.45719720e-02 -3.65812689e-01
7.42514580e-02 5.55275738e-01 2.38373369e-01 -2.79478222e-01
1.64607012e+00 -3.87199640e-01 -2.16899037e-01 4.03181642e-01
1.53642178e+00 4.34238613e-02 -1.67095065e+00 -5.88846624e-01
-1.74224079e-01 -7.15994656e-01 4.45974380e-01 -4.83024776e-01
-1.39744043e+00 6.05319262e-01 4.75432634e-01 3.38454247e-01
1.13469517e+00 -2.44068652e-01 8.90298426e-01 3.74727488e-01
5.18629313e-01 -8.06385338e-01 1.95588335e-01 2.50811189e-01
1.11032557e+00 -1.19587910e+00 2.30755433e-01 -8.43151510e-02
-6.22950345e-02 1.11789775e+00 -6.52182475e-02 -5.87005198e-01
6.21192575e-01 2.93777525e-01 -1.95364565e-01 -2.39222541e-01
-4.01780248e-01 9.09573808e-02 2.31251642e-01 5.17663717e-01
-5.00073209e-02 -1.21260464e-01 -1.26466289e-01 -2.25712359e-02
-2.17270106e-01 3.11814696e-01 5.41572452e-01 7.07487285e-01
-2.41033897e-01 -1.05948055e+00 -4.61360842e-01 4.95104760e-01
-1.29098937e-01 -2.09014967e-01 1.75946161e-01 4.36833143e-01
-8.95108059e-02 7.89066970e-01 -2.05007166e-01 -1.73945159e-01
2.60887090e-02 -4.05918568e-01 7.24226177e-01 -4.57200170e-01
-6.87973052e-02 1.78012908e-01 -3.04109484e-01 -8.85075808e-01
-5.27621329e-01 -7.77578413e-01 -5.31754553e-01 -5.85741624e-02
-1.81406379e-01 1.13273464e-01 7.99722195e-01 1.03064668e+00
4.69571859e-01 2.00874850e-01 8.29405487e-01 -1.36204755e+00
-1.15699601e+00 -8.47737074e-01 -5.75689614e-01 5.06038964e-01
7.83828735e-01 -3.59961510e-01 -6.54662013e-01 -1.46977305e-01] | [11.34796142578125, -2.1808598041534424] |
ea959e0f-ef01-473b-b16d-0d96e7710cd7 | singing-voice-synthesis-system-based-on-time | null | null | https://aclanthology.org/O13-1009 | https://aclanthology.org/O13-1009.pdf | 基於時域上基週同步疊加法之歌聲合成系統 (Singing Voice Synthesis System Based on Time Domain-Pitch Synchronized Overlap and Add) [In Chinese] | null | ['Chia-Ping Chen', 'Ming-Kuan Wu'] | 2013-10-01 | singing-voice-synthesis-system-based-on-time-1 | https://aclanthology.org/O13-1009 | https://aclanthology.org/O13-1009.pdf | roclingijclclp-2013-10 | ['singing-voice-synthesis'] | ['speech'] | [-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] | [-7.428300857543945, 3.6735284328460693] |
5e1cd7b3-8ae6-4bc1-9ba5-07bd25718a02 | stacked-hybrid-attention-and-group | 2203.09811 | null | https://arxiv.org/abs/2203.09811v2 | https://arxiv.org/pdf/2203.09811v2.pdf | Stacked Hybrid-Attention and Group Collaborative Learning for Unbiased Scene Graph Generation | Scene Graph Generation, which generally follows a regular encoder-decoder pipeline, aims to first encode the visual contents within the given image and then parse them into a compact summary graph. Existing SGG approaches generally not only neglect the insufficient modality fusion between vision and language, but also fail to provide informative predicates due to the biased relationship predictions, leading SGG far from practical. Towards this end, in this paper, we first present a novel Stacked Hybrid-Attention network, which facilitates the intra-modal refinement as well as the inter-modal interaction, to serve as the encoder. We then devise an innovative Group Collaborative Learning strategy to optimize the decoder. Particularly, based upon the observation that the recognition capability of one classifier is limited towards an extremely unbalanced dataset, we first deploy a group of classifiers that are expert in distinguishing different subsets of classes, and then cooperatively optimize them from two aspects to promote the unbiased SGG. Experiments conducted on VG and GQA datasets demonstrate that, we not only establish a new state-of-the-art in the unbiased metric, but also nearly double the performance compared with two baselines. | ['Liqiang Nie', 'Yuan Cheng', 'Jianlong Wu', 'Xuemeng Song', 'Tian Gan', 'Xingning Dong'] | 2022-03-18 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Dong_Stacked_Hybrid-Attention_and_Group_Collaborative_Learning_for_Unbiased_Scene_Graph_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Dong_Stacked_Hybrid-Attention_and_Group_Collaborative_Learning_for_Unbiased_Scene_Graph_CVPR_2022_paper.pdf | cvpr-2022-1 | ['scene-graph-generation', 'unbiased-scene-graph-generation'] | ['computer-vision', 'computer-vision'] | [ 3.72447312e-01 3.54224652e-01 1.63216703e-02 -4.58374172e-01
-9.20370698e-01 -2.22699597e-01 6.47486567e-01 -8.94348845e-02
-1.74125910e-01 5.30884862e-01 3.97431910e-01 -1.29331961e-01
4.01444882e-02 -7.20006466e-01 -7.92307615e-01 -8.11586916e-01
3.65105450e-01 3.32358956e-01 1.56648204e-01 9.70393717e-02
-6.22513965e-02 -4.29850183e-02 -1.48025703e+00 3.51821482e-01
1.14963341e+00 1.17899156e+00 4.25974697e-01 2.09793165e-01
-9.14027542e-02 1.32757449e+00 -2.60895282e-01 -7.41172194e-01
-1.02150224e-01 -5.39930344e-01 -6.59444213e-01 4.86668646e-01
4.48931485e-01 -3.43395501e-01 -3.75050664e-01 1.26357651e+00
5.45785069e-01 3.86667363e-02 6.47481918e-01 -1.13844514e+00
-8.97147417e-01 8.79857361e-01 -6.62038505e-01 -2.52961189e-01
2.35162228e-01 2.81356961e-01 1.41663301e+00 -7.88653374e-01
5.84812641e-01 1.14934039e+00 3.06053847e-01 5.01622081e-01
-9.84855950e-01 -5.59054494e-01 7.20493972e-01 3.93750757e-01
-1.43218052e+00 -4.54895347e-01 1.03820550e+00 -3.46939236e-01
6.50729239e-01 1.75180227e-01 6.22646630e-01 1.13188386e+00
-1.64757431e-01 1.09431446e+00 8.93553078e-01 -2.12741017e-01
-8.49176347e-02 -8.29047058e-03 3.91979627e-02 9.86670732e-01
1.71180353e-01 -1.57151803e-01 -6.26050472e-01 3.07452232e-01
3.93467128e-01 -2.34908953e-01 -6.02310836e-01 -6.24966085e-01
-1.37343609e+00 6.28001094e-01 6.20608568e-01 -5.19254617e-02
-4.69450623e-01 -3.23475935e-02 1.95537776e-01 -1.75921932e-01
2.43982106e-01 1.66082591e-01 -8.20347667e-02 3.28599691e-01
-7.27658570e-01 6.01250641e-02 4.83136803e-01 1.12579942e+00
8.22122097e-01 -1.28809318e-01 -6.54858232e-01 7.02830493e-01
7.37967134e-01 2.23360851e-01 2.30686933e-01 -5.86786747e-01
8.33068192e-01 8.69173706e-01 -2.20519111e-01 -1.09388459e+00
-2.02037334e-01 -7.18528271e-01 -1.10881042e+00 -1.86052546e-01
6.18720092e-02 -6.52168365e-03 -9.73466277e-01 1.78463662e+00
3.32519084e-01 1.12682857e-01 2.03757867e-01 9.50313270e-01
1.18558538e+00 5.67868054e-01 3.04982454e-01 -1.82348788e-01
1.30450237e+00 -1.40514505e+00 -6.38924241e-01 -4.53110695e-01
4.51822430e-01 -4.78697568e-01 1.04629099e+00 3.08185279e-01
-1.09334993e+00 -7.39935398e-01 -1.13307381e+00 -2.95742452e-01
-1.24346167e-01 4.32619065e-01 6.39369607e-01 2.08707556e-01
-8.90378535e-01 1.44426078e-01 -5.70993483e-01 -1.46949425e-01
8.14094722e-01 1.39850289e-01 -3.98948967e-01 -2.79532373e-01
-1.03270912e+00 6.70551658e-01 7.52682567e-01 4.20878649e-01
-9.02325511e-01 -3.40204656e-01 -9.94831204e-01 1.04890049e-01
6.72632217e-01 -1.07056916e+00 8.68413448e-01 -8.08679700e-01
-1.29478419e+00 8.43495250e-01 -4.39341143e-02 -3.46891165e-01
6.18291914e-01 -1.16948433e-01 -3.13299388e-01 -4.85510901e-02
1.62985399e-01 8.74324858e-01 7.32511461e-01 -1.48211396e+00
-9.23742771e-01 -3.43381405e-01 3.34683180e-01 7.86587715e-01
-2.73237199e-01 -2.77162254e-01 -1.05086541e+00 -4.32177573e-01
2.35666290e-01 -5.78609109e-01 -2.58148193e-01 -3.49474639e-01
-8.64530444e-01 -3.25313389e-01 4.99828249e-01 -7.49470592e-01
1.29957211e+00 -2.17350435e+00 5.17042398e-01 7.70331472e-02
4.54373211e-01 1.20246433e-01 -1.42531097e-02 1.74474806e-01
8.23172852e-02 -6.73851818e-02 -3.35000545e-01 -7.53105283e-01
5.90840690e-02 2.48079717e-01 -3.01413417e-01 3.14222783e-01
3.28402281e-01 1.03599322e+00 -1.00410557e+00 -8.05856407e-01
2.23148718e-01 2.48284429e-01 -6.21720731e-01 4.99768525e-01
-3.89026999e-01 6.53573453e-01 -5.65272331e-01 6.74306393e-01
5.77750444e-01 -6.53530419e-01 2.24336773e-01 -7.17087984e-01
6.93251714e-02 1.54181764e-01 -1.02739596e+00 1.92150736e+00
-1.52493000e-01 1.78467840e-01 -1.90760374e-01 -1.27626991e+00
8.40688109e-01 3.44032161e-02 3.23744982e-01 -7.50093699e-01
1.48791358e-01 6.34218901e-02 1.61632262e-02 -5.99804401e-01
2.60669351e-01 1.28689989e-01 -8.65022466e-02 -1.82684455e-02
3.46085668e-01 6.56935424e-02 2.83647954e-01 2.92664737e-01
7.04505563e-01 4.09535646e-01 2.64633060e-01 9.89667103e-02
9.20127034e-01 -1.93958253e-01 6.14350080e-01 6.57457471e-01
-8.61849338e-02 7.90764987e-01 6.63450122e-01 -5.07936142e-02
-6.64867818e-01 -8.95462513e-01 3.59853595e-01 9.28184867e-01
7.30484545e-01 -5.10483742e-01 -6.48398757e-01 -1.03117621e+00
-3.16374779e-01 8.60505879e-01 -5.50431132e-01 -3.90249610e-01
-3.10628772e-01 -8.99159849e-01 3.79879564e-01 5.62093794e-01
9.22471046e-01 -9.05911982e-01 -2.49564186e-01 2.49022357e-02
-3.74124497e-01 -1.36055326e+00 -3.57161373e-01 -1.12323046e-01
-2.62155175e-01 -1.15469360e+00 -4.67104018e-01 -8.79711926e-01
6.69302464e-01 2.36069128e-01 1.11274600e+00 1.39995337e-01
1.73835516e-01 1.63587555e-01 -4.75171566e-01 -2.73517042e-01
-1.45912841e-01 1.92680568e-01 -4.64871109e-01 6.08790874e-01
2.22175330e-01 -3.62513423e-01 -5.79099476e-01 -8.43931362e-03
-5.61830461e-01 7.99736083e-01 7.99801350e-01 8.94839585e-01
8.83154392e-01 3.08384709e-02 4.49061632e-01 -1.04276204e+00
3.19401890e-01 -4.04484361e-01 -4.24553603e-01 6.68730497e-01
-5.75799584e-01 -1.04281068e-01 6.21875584e-01 -1.06630310e-01
-1.22661519e+00 2.07769111e-01 -1.84646934e-01 -4.74705577e-01
-6.47982806e-02 6.34519935e-01 -8.83871019e-01 1.22547515e-01
2.11657539e-01 5.30433297e-01 -1.27955511e-01 -2.67086118e-01
6.19612575e-01 6.08526647e-01 6.42968118e-01 -3.92449707e-01
7.17899740e-01 3.85330409e-01 -1.74533814e-01 -4.83506322e-01
-1.16803491e+00 -2.84280419e-01 -4.73437428e-01 -3.11623335e-01
1.07595551e+00 -1.19192839e+00 -5.24617493e-01 5.48286796e-01
-1.16465592e+00 -1.73450977e-01 -1.26334831e-01 4.62948620e-01
-5.40747166e-01 3.87689948e-01 -3.33126843e-01 -6.55140579e-01
-2.47072712e-01 -1.30960763e+00 1.33749199e+00 3.99281383e-01
2.57353574e-01 -7.67550766e-01 -4.39442873e-01 6.97140455e-01
-2.86682565e-02 1.40842751e-01 8.69099259e-01 -6.69589341e-01
-8.40773463e-01 5.10083437e-02 -7.20916271e-01 3.39552671e-01
1.08639106e-01 -1.99873567e-01 -1.02659619e+00 -1.91625819e-01
-1.92180529e-01 -3.90611053e-01 9.58431184e-01 8.38435665e-02
1.39427018e+00 -1.36405349e-01 -2.81200975e-01 9.21568096e-01
1.38234389e+00 1.39165401e-01 6.36895776e-01 1.62373334e-01
1.43144453e+00 7.09761918e-01 6.66717947e-01 1.56508595e-01
9.75606263e-01 5.91615498e-01 7.54763424e-01 -1.87347442e-01
-4.13996428e-01 -6.01205468e-01 1.71750292e-01 9.76572990e-01
-9.49842259e-02 -5.32341421e-01 -6.90357864e-01 4.24107522e-01
-2.09616184e+00 -6.49238884e-01 -9.77441221e-02 1.90267420e+00
7.41820991e-01 2.09769253e-02 -1.56501025e-01 -1.72570050e-01
6.89591527e-01 4.97109294e-01 -4.88346756e-01 3.13144773e-01
-3.81703138e-01 -1.85011670e-01 5.22795498e-01 3.37491751e-01
-1.27974653e+00 1.16211069e+00 5.29448318e+00 9.16759372e-01
-1.11894035e+00 -1.73187107e-02 7.39282012e-01 7.19579160e-02
-4.78088707e-01 1.01351254e-01 -7.00919271e-01 5.79802930e-01
3.47416610e-01 -7.76574537e-02 4.05705690e-01 6.07341468e-01
-1.04718640e-01 2.14528993e-01 -9.57819402e-01 1.23588669e+00
4.05249327e-01 -1.14315200e+00 5.26265681e-01 -8.94484762e-03
6.59826875e-01 -1.33517087e-01 -1.12425596e-01 3.15175563e-01
2.68466383e-01 -9.78960574e-01 1.09740031e+00 6.63382590e-01
6.82816446e-01 -6.56232536e-01 7.05635607e-01 5.20989418e-01
-1.24577773e+00 -7.43110776e-02 -3.61162812e-01 2.38488555e-01
3.05701256e-01 4.29792553e-01 -5.18577218e-01 1.18331373e+00
5.83466709e-01 1.07228231e+00 -8.61836255e-01 9.95537758e-01
-5.70751131e-01 4.00437266e-01 9.71154794e-02 1.36112496e-01
2.80233413e-01 -3.61845911e-01 3.53465468e-01 9.21347499e-01
2.14578778e-01 5.36498018e-02 3.44334215e-01 8.18576217e-01
-9.89079550e-02 2.41026476e-01 -5.01203001e-01 -1.04927525e-01
3.18783104e-01 1.27416694e+00 -4.67961788e-01 -4.52831477e-01
-7.12298870e-01 1.04877627e+00 6.20404899e-01 5.55183053e-01
-1.06535327e+00 -1.74701676e-01 2.68739551e-01 -2.74700075e-01
5.31289101e-01 8.20197687e-02 -2.97385305e-01 -1.45940244e+00
1.52258918e-01 -9.67796206e-01 4.96914417e-01 -9.46426928e-01
-1.16821456e+00 6.91174388e-01 -7.57515505e-02 -1.27396464e+00
-1.84814215e-01 -4.23597366e-01 -3.85804474e-01 8.10390592e-01
-1.58057117e+00 -1.82083261e+00 -5.84339917e-01 6.29730701e-01
4.87287730e-01 -1.68225393e-01 3.82105052e-01 4.65771437e-01
-9.01407063e-01 6.40619993e-01 -4.53669965e-01 1.19919315e-01
5.46006620e-01 -1.15461314e+00 5.21487743e-02 1.19557631e+00
3.42117965e-01 3.77251089e-01 4.33038563e-01 -6.53310061e-01
-1.33224154e+00 -1.37945795e+00 8.26277077e-01 -1.76219404e-01
5.51201463e-01 -3.53107125e-01 -9.39463675e-01 7.23487377e-01
2.56215006e-01 -1.75996199e-02 4.60714787e-01 2.27855146e-02
-2.90337116e-01 -2.63350189e-01 -6.53357744e-01 7.66032398e-01
1.45992804e+00 -4.86447483e-01 -5.04955471e-01 2.54382104e-01
7.94628501e-01 -6.32831097e-01 -5.63873887e-01 6.72718465e-01
2.56003380e-01 -1.13407373e+00 8.26555908e-01 -5.94203830e-01
5.86230874e-01 -5.72434425e-01 -2.36678809e-01 -1.14259326e+00
-5.33039451e-01 -3.17264795e-01 -2.54139036e-01 1.64575136e+00
3.25240135e-01 -5.67507565e-01 6.57646894e-01 2.50678807e-01
-4.33920383e-01 -8.82314980e-01 -5.74479759e-01 -2.86596984e-01
-4.85940069e-01 -4.25431877e-01 7.48291969e-01 7.19615757e-01
-3.79239857e-01 7.64412642e-01 -5.98504007e-01 4.56161350e-01
6.94663584e-01 3.63700897e-01 8.10242355e-01 -9.73915160e-01
-3.58145624e-01 -4.44686890e-01 -4.35923219e-01 -1.31170249e+00
1.03625976e-01 -8.27819526e-01 2.92033494e-01 -1.81788301e+00
3.78873020e-01 -2.41952658e-01 -4.07042295e-01 4.35919017e-01
-5.61449826e-01 2.47645810e-01 1.62660852e-01 2.40812287e-01
-1.03608596e+00 7.71622658e-01 1.42558277e+00 -3.12047780e-01
1.18264265e-01 -2.27544427e-01 -1.28989327e+00 6.75283134e-01
3.67585987e-01 6.01915419e-02 -8.73490036e-01 -6.35549903e-01
1.82944626e-01 -1.43970191e-01 5.10922968e-01 -8.89928758e-01
3.15262854e-01 -1.32426247e-01 2.35665977e-01 -6.57061338e-01
2.05093414e-01 -7.35952616e-01 8.21565688e-02 3.97057757e-02
-1.76805496e-01 -3.18042189e-01 -2.37458661e-01 7.61355758e-01
-5.21879256e-01 1.60751306e-02 5.91560781e-01 -7.17120618e-02
-1.12333810e+00 6.93598270e-01 2.52811760e-01 7.11717010e-02
1.04432583e+00 -1.70509815e-01 -4.40074980e-01 -2.92011797e-01
-5.77430189e-01 5.41347802e-01 4.75161433e-01 3.50439817e-01
4.21765536e-01 -1.47356093e+00 -7.23454714e-01 2.56576687e-01
4.29232568e-01 4.96044189e-01 5.65552235e-01 9.80400681e-01
-3.97778332e-01 2.51759529e-01 -4.94861342e-02 -6.26772344e-01
-1.10786211e+00 6.63937569e-01 3.41730714e-01 -4.25267786e-01
-5.85240602e-01 1.11698174e+00 7.58850932e-01 -9.46664345e-03
3.05650890e-01 -1.38947964e-01 -6.60821199e-01 3.73228759e-01
3.87212187e-01 -1.37298286e-01 -1.81427579e-02 -9.83799756e-01
-4.28281814e-01 5.92951834e-01 -6.11138111e-03 1.91376030e-01
1.03358746e+00 -4.53168631e-01 -6.59459606e-02 2.52333999e-01
1.09617150e+00 -1.55070677e-01 -1.41550517e+00 -2.54117757e-01
-3.22030604e-01 -3.89481783e-01 3.37677286e-03 -6.00706816e-01
-1.27518511e+00 8.31474245e-01 2.57132947e-01 3.58872622e-01
1.51421785e+00 1.44027382e-01 6.13409758e-01 -3.25186141e-02
2.20042914e-01 -8.85643423e-01 -1.02579549e-01 3.31554890e-01
8.16552877e-01 -1.33796203e+00 -5.37896045e-02 -7.98136592e-01
-1.01080966e+00 6.96133018e-01 8.28177750e-01 4.55098525e-02
1.92488819e-01 -2.51761705e-01 -7.47362822e-02 -2.34124511e-01
-8.15873742e-01 -6.41692519e-01 6.03831589e-01 6.31746948e-01
3.32607567e-01 -3.03037856e-02 -1.86997324e-01 8.14421177e-01
-1.00600868e-01 -5.92482239e-02 6.22542389e-02 4.08318043e-01
-1.61804840e-01 -8.60805273e-01 8.61749426e-02 5.25252581e-01
-1.77219734e-01 -1.67609870e-01 -3.19384038e-01 6.80504441e-01
4.74306196e-01 8.59678805e-01 -1.71765443e-02 -6.42161131e-01
4.05851185e-01 -1.89011678e-01 3.86213720e-01 -6.29156351e-01
-2.38343060e-01 1.66561291e-01 2.83193946e-01 -6.38337851e-01
-5.54841697e-01 -5.93853593e-01 -9.62980151e-01 2.16848448e-01
-2.72531450e-01 -5.65350279e-02 1.84657305e-01 1.14235497e+00
3.36161315e-01 9.27847385e-01 4.89487678e-01 -6.33318007e-01
-3.90000582e-01 -8.58470082e-01 -4.56697106e-01 5.69993258e-01
1.57818317e-01 -7.29225159e-01 -2.49493226e-01 1.39674708e-01] | [10.414712905883789, 1.5863338708877563] |
b2fbfbc5-e7ea-4a79-876d-2649a70c81e0 | applying-transformer-based-text-summarization | 2209.03791 | null | https://arxiv.org/abs/2209.03791v2 | https://arxiv.org/pdf/2209.03791v2.pdf | Applying Transformer-based Text Summarization for Keyphrase Generation | Keyphrases are crucial for searching and systematizing scholarly documents. Most current methods for keyphrase extraction are aimed at the extraction of the most significant words in the text. But in practice, the list of keyphrases often includes words that do not appear in the text explicitly. In this case, the list of keyphrases represents an abstractive summary of the source text. In this paper, we experiment with popular transformer-based models for abstractive text summarization using four benchmark datasets for keyphrase extraction. We compare the results obtained with the results of common unsupervised and supervised methods for keyphrase extraction. Our evaluation shows that summarization models are quite effective in generating keyphrases in the terms of the full-match F1-score and BERTScore. However, they produce a lot of words that are absent in the author's list of keyphrases, which makes summarization models ineffective in terms of ROUGE-1. We also investigate several ordering strategies to concatenate target keyphrases. The results showed that the choice of strategy affects the performance of keyphrase generation. | ['Dmitry Morozov', 'Anna Glazkova'] | 2022-09-08 | null | null | null | null | ['keyphrase-generation', 'keyphrase-extraction'] | ['natural-language-processing', 'natural-language-processing'] | [ 1.97020724e-01 8.93925950e-02 -3.40284824e-01 4.95964646e-01
-9.66693223e-01 -9.06110644e-01 1.03610826e+00 9.69930470e-01
-4.79439765e-01 1.07588804e+00 8.10728610e-01 -4.25418973e-01
-2.67153054e-01 -7.37014890e-01 -5.11031628e-01 -4.39423263e-01
3.00587565e-01 1.47064105e-01 1.54168591e-01 -3.50632608e-01
9.98985827e-01 6.11714900e-01 -1.52276504e+00 3.20757240e-01
1.05508256e+00 4.37211573e-01 2.95240611e-01 9.05293107e-01
-8.60502958e-01 8.02595258e-01 -1.25304437e+00 -3.17128658e-01
-9.01375636e-02 -6.91043973e-01 -8.92580271e-01 6.46348670e-02
5.83880424e-01 -1.44667447e-01 -5.06964266e-01 1.07486558e+00
2.85021365e-01 -1.41200885e-01 8.76418948e-01 -1.15840614e+00
-7.45491832e-02 1.09485149e+00 -4.42926168e-01 5.73960006e-01
7.50011325e-01 -5.43782711e-01 1.39345527e+00 -8.19544017e-01
8.75521004e-01 1.04163527e+00 2.16541782e-01 -6.88767210e-02
-8.02093744e-01 -3.88399243e-01 -2.06151739e-01 -1.54077038e-02
-1.19144773e+00 -4.48020607e-01 9.29084778e-01 -1.61476567e-01
8.21350873e-01 7.14152336e-01 7.32877314e-01 7.07329810e-01
6.68140113e-01 1.02626252e+00 8.11064780e-01 -8.28718245e-01
1.62219722e-02 -4.46354486e-02 5.85572839e-01 4.28271711e-01
6.73162282e-01 -5.58706522e-01 -6.95061445e-01 -4.56585824e-01
3.20327967e-01 -2.55628616e-01 -1.69557899e-01 3.38872254e-01
-1.56910527e+00 7.54130244e-01 -4.06448036e-01 5.54281771e-01
-6.55317783e-01 -3.28176111e-01 6.91078126e-01 3.32053363e-01
2.32975304e-01 1.07015228e+00 -2.46176541e-01 -2.18390003e-01
-1.42150724e+00 7.85470724e-01 1.08537531e+00 1.07943535e+00
4.74989325e-01 -2.55798221e-01 -8.10836613e-01 6.90970361e-01
-2.10590169e-01 3.61093789e-01 5.38287878e-01 -8.13781619e-01
6.56108916e-01 8.94535482e-01 2.93777943e-01 -1.24314666e+00
-8.05634931e-02 -3.61679703e-01 -6.63921118e-01 -6.87332332e-01
-5.19269183e-02 -2.83057868e-01 -8.84354293e-01 8.23354542e-01
-6.54254183e-02 -4.62950468e-01 1.37303159e-01 1.08527668e-01
1.46010029e+00 1.14483798e+00 -2.25127295e-01 -8.12962532e-01
1.56242204e+00 -8.65019202e-01 -1.18812740e+00 2.48098373e-03
3.12262833e-01 -1.43124592e+00 6.36560082e-01 4.54541475e-01
-1.10145700e+00 -3.81797284e-01 -1.08550918e+00 -8.36248398e-02
-4.10591245e-01 5.38105607e-01 2.82456845e-01 2.91250318e-01
-6.85266912e-01 6.04621112e-01 -2.99657941e-01 -3.55646044e-01
3.19022946e-02 -1.89212725e-01 -1.74880624e-01 2.03339338e-01
-1.00888884e+00 7.06848264e-01 7.54015267e-01 -5.94425857e-01
-3.03656012e-01 -8.36843312e-01 -3.70147616e-01 1.95459560e-01
6.83366477e-01 -4.51462060e-01 1.34477031e+00 -1.32202029e-01
-1.01181209e+00 5.36615014e-01 -2.40395606e-01 -3.50928545e-01
2.59783596e-01 -3.62567753e-01 -4.23183173e-01 6.43352568e-01
3.08574826e-01 3.09743166e-01 7.74426997e-01 -9.83826220e-01
-1.01462042e+00 2.19813913e-01 -6.40635490e-02 1.22266836e-01
-5.93068540e-01 2.29419857e-01 -5.31462610e-01 -1.19249940e+00
4.93156165e-02 -6.56886458e-01 4.35789023e-03 -7.74282992e-01
-8.97433281e-01 -5.10642469e-01 7.58697271e-01 -7.54753232e-01
1.94302797e+00 -1.73843932e+00 4.80995923e-02 3.36585104e-01
2.16523618e-01 1.10481910e-01 1.09749004e-01 1.17948425e+00
2.47213859e-02 6.18816435e-01 1.84030011e-01 2.59860098e-01
-5.59584685e-02 6.47092564e-03 -9.79694366e-01 -5.15821353e-02
-3.06754917e-01 7.75692940e-01 -9.38578367e-01 -1.09087646e+00
-9.61156264e-02 -2.31856361e-01 -9.63847563e-02 3.65557522e-01
-2.37036943e-01 -2.43755341e-01 -7.48456120e-01 4.61939991e-01
1.48212552e-01 5.34499958e-02 -2.25124091e-01 -3.60976726e-01
-4.55808580e-01 6.65062606e-01 -1.12014139e+00 1.23520958e+00
-1.54756114e-01 8.08047175e-01 -4.73001122e-01 -5.87382376e-01
7.36083865e-01 4.88391757e-01 6.44278288e-01 -2.65760839e-01
1.04650237e-01 2.72550434e-01 -2.68497407e-01 -1.54421508e-01
1.39163613e+00 2.84235686e-01 -3.46229941e-01 6.16114676e-01
-1.21668568e-02 -7.28840709e-01 1.16797280e+00 9.24810529e-01
1.02662182e+00 -3.71128052e-01 6.90518796e-01 -4.57986295e-01
6.75904870e-01 4.16201770e-01 4.36558798e-02 1.05236185e+00
5.86396933e-01 5.95973015e-01 8.86178434e-01 3.06385625e-02
-1.22167909e+00 -5.21372795e-01 5.48200235e-02 6.68167293e-01
-2.02927828e-01 -1.30659640e+00 -8.42721045e-01 -6.58930480e-01
-8.42213556e-02 8.82234752e-01 -3.48824263e-01 -9.35247242e-02
-5.05950689e-01 -6.52088284e-01 6.23200417e-01 -2.94839069e-02
1.33843541e-01 -1.02933502e+00 -4.71956670e-01 3.09033960e-01
-5.93638361e-01 -1.17677140e+00 -6.69296026e-01 4.78966273e-02
-6.74604535e-01 -9.80389357e-01 -1.02117300e+00 -6.18170500e-01
7.64620304e-01 5.05938947e-01 1.10392249e+00 7.39851072e-02
3.28746997e-02 2.63629973e-01 -8.62712920e-01 -7.14347422e-01
-7.60990322e-01 5.80823362e-01 -1.67472810e-01 -4.84759122e-01
1.62518844e-01 -2.44428605e-01 -1.01246245e-01 -3.27949524e-01
-1.20332515e+00 6.25921637e-02 7.10532367e-01 5.49277723e-01
4.43451852e-01 5.66060901e-01 3.71665835e-01 -7.91016400e-01
1.56269705e+00 -2.59627104e-01 -3.29799891e-01 4.58729804e-01
-7.38608539e-01 3.78004760e-01 7.85674274e-01 -5.36264300e-01
-6.74565315e-01 -5.55789232e-01 4.99349311e-02 7.93587789e-02
1.07047945e-01 8.57159734e-01 1.73483819e-01 3.41375589e-01
4.75804895e-01 7.64878631e-01 -3.13517749e-01 -6.39437616e-01
2.11776748e-01 8.69156778e-01 3.65781635e-01 -5.83222985e-01
8.45226228e-01 -8.14354047e-02 3.07990052e-02 -1.38571954e+00
-9.45053160e-01 -7.49264657e-01 -3.88761997e-01 -3.15366596e-01
4.40686375e-01 -7.68291950e-01 -2.41551504e-01 3.03569913e-01
-1.18637335e+00 6.04557753e-01 -5.14901817e-01 3.06370705e-01
-1.26230299e-01 9.61656213e-01 -5.23649395e-01 -3.18545878e-01
-8.21714640e-01 -9.61376131e-01 1.11447251e+00 4.78574008e-01
-9.06052530e-01 -4.35530752e-01 8.31311122e-02 1.34552732e-01
-2.80768462e-02 3.09269223e-02 1.21990526e+00 -1.03112745e+00
2.84421840e-03 -5.44403017e-01 6.01977371e-02 1.38038307e-01
3.84531111e-01 3.86157453e-01 -2.98149973e-01 -8.05830359e-02
-2.54060507e-01 6.70332387e-02 1.03842580e+00 2.87763298e-01
1.01554477e+00 -8.16272616e-01 -3.00219655e-01 5.34901097e-02
9.46423829e-01 1.62625462e-01 6.72055125e-01 5.03359079e-01
6.39245629e-01 5.47170341e-01 6.39994502e-01 5.02905965e-01
1.92082629e-01 3.24686974e-01 -2.14052096e-01 1.96420655e-01
-1.41163338e-02 -4.68576014e-01 4.62636918e-01 1.36542809e+00
6.51545376e-02 -5.02527952e-01 -7.28388965e-01 7.45208442e-01
-1.50034618e+00 -1.12568414e+00 -4.88433331e-01 1.76994157e+00
1.16342354e+00 3.87129366e-01 2.20412701e-01 5.48825264e-01
4.17543352e-01 5.28238475e-01 2.82817394e-01 -5.13807833e-01
-1.75129965e-01 2.20465004e-01 6.00638151e-01 4.09669399e-01
-9.21139479e-01 1.02961779e+00 6.08581400e+00 1.11501026e+00
-7.28488624e-01 -5.07830560e-01 1.62590131e-01 1.69392079e-01
-4.93896872e-01 8.45914185e-02 -1.20737135e+00 3.75980109e-01
8.73156965e-01 -9.57036257e-01 -2.67088294e-01 6.79231763e-01
2.86043525e-01 -5.77658653e-01 -8.02642167e-01 7.80576050e-01
2.30664551e-01 -1.46613908e+00 7.60962248e-01 9.81454998e-02
1.02164066e+00 -6.17955863e-01 -4.65343237e-01 -4.00930233e-02
1.10307939e-01 -6.36702478e-01 8.12980771e-01 5.51803708e-01
2.31744587e-01 -1.02499878e+00 7.90991366e-01 3.79891425e-01
-9.37274098e-01 2.75091052e-01 -3.10988277e-01 2.04171225e-01
-2.69432068e-02 9.77645397e-01 -8.85252595e-01 7.61372149e-01
2.94969797e-01 4.72700030e-01 -7.57687211e-01 1.12434411e+00
-4.05648679e-01 6.95636511e-01 -2.01203033e-01 -6.50509715e-01
4.27951604e-01 -2.61509538e-01 9.28689361e-01 1.61883259e+00
4.77705091e-01 1.41478941e-01 2.50616848e-01 3.76428604e-01
-3.29225481e-01 5.97461581e-01 -3.81549776e-01 -7.55546451e-01
6.01577997e-01 1.25520051e+00 -1.18930650e+00 -8.26888919e-01
-2.83938255e-02 6.56921864e-01 -5.26174068e-01 2.57295519e-01
-2.69400299e-01 -1.04067135e+00 1.59616917e-01 1.24559231e-01
3.08996439e-01 -3.17298681e-01 -1.84595436e-01 -1.11599374e+00
2.25808308e-01 -1.30291510e+00 2.46183738e-01 -5.92704237e-01
-8.77841473e-01 5.21129131e-01 3.76657695e-01 -1.18466914e+00
-3.51346821e-01 -1.76466197e-01 -6.74238443e-01 5.47300577e-01
-1.10677803e+00 -7.43161201e-01 6.27612555e-03 5.13478145e-02
9.12904501e-01 -2.90418148e-01 7.06017077e-01 -1.58173963e-01
-4.76524562e-01 2.00853869e-01 1.96365222e-01 2.22913727e-01
7.05244422e-01 -1.33282411e+00 4.43497986e-01 1.17817295e+00
3.63169968e-01 8.98412049e-01 1.24568379e+00 -1.05060661e+00
-1.39973748e+00 -6.71957910e-01 1.49615347e+00 -1.85140193e-01
7.32449234e-01 1.76569521e-02 -7.67813623e-01 1.87677443e-01
5.81912756e-01 -9.20130312e-01 4.96466696e-01 -9.91616473e-02
8.40122029e-02 9.22470819e-03 -5.59965074e-01 9.91306841e-01
3.37996006e-01 -3.93682957e-01 -1.26504004e+00 4.38657999e-01
6.91650689e-01 -3.18883032e-01 -8.40468645e-01 1.03959374e-01
4.23257321e-01 -2.57633626e-01 9.01317537e-01 -5.18230796e-01
8.67585123e-01 -2.37147480e-01 1.35201886e-01 -1.27936566e+00
-1.72371447e-01 -9.48363423e-01 -5.24176896e-01 1.55173993e+00
4.46744233e-01 -3.38913053e-02 4.74906713e-01 1.05293371e-01
1.82857916e-01 -4.07426178e-01 -4.17149812e-01 -6.09482884e-01
2.17757095e-02 -3.45427431e-02 4.85272497e-01 7.63622880e-01
3.08280945e-01 8.66702974e-01 -2.37043440e-01 -5.13996840e-01
5.50314188e-01 3.38689327e-01 7.61102438e-01 -1.28110659e+00
3.70311737e-01 -8.17530513e-01 1.42825663e-01 -8.91451836e-01
-1.01808131e-01 -6.65054858e-01 -9.91804004e-02 -1.99376059e+00
4.21714902e-01 2.33159229e-01 1.24982461e-01 3.34816664e-01
-4.53786999e-01 -4.05379355e-01 1.41138658e-01 4.25449014e-01
-3.26557934e-01 3.46452355e-01 1.28385448e+00 -3.58132988e-01
-5.55050611e-01 4.76972871e-02 -9.87497389e-01 6.24009907e-01
7.81081080e-01 -7.41858184e-01 -2.90351540e-01 2.99181342e-01
4.45222497e-01 6.43967241e-02 -1.14042722e-01 -7.29140759e-01
4.59740937e-01 -4.53842670e-01 6.87248781e-02 -1.05811238e+00
-2.48069838e-01 -3.14891398e-01 -7.57629126e-02 3.30995023e-01
-4.69815046e-01 4.61477429e-01 2.92060584e-01 6.60920963e-02
-3.08596969e-01 -7.62067914e-01 2.12521091e-01 -2.53097087e-01
-3.24693829e-01 -2.06046671e-01 -9.73465025e-01 3.01977396e-01
4.36988235e-01 -1.09504850e-03 -2.23856688e-01 -3.48557800e-01
5.36924340e-02 -3.65986973e-02 3.49135548e-01 4.41229731e-01
7.72635639e-01 -1.01191986e+00 -1.17388344e+00 -3.37029546e-01
1.16286620e-01 -2.67053038e-01 -4.22216952e-01 5.53105414e-01
-7.55812883e-01 9.14067328e-01 -6.74256235e-02 8.60203430e-02
-1.45134246e+00 2.68107295e-01 -4.32957858e-01 -6.61728859e-01
-6.31552458e-01 4.09314513e-01 -3.60610127e-01 7.09057152e-02
2.14664698e-01 -6.18818045e-01 -8.17124486e-01 8.92232060e-01
8.09934974e-01 4.12821859e-01 2.86254048e-01 -5.51116824e-01
-3.68722416e-02 2.83974290e-01 -5.96770346e-01 -3.20686787e-01
1.30804086e+00 4.45298627e-02 -6.43143773e-01 4.64155674e-01
1.09230173e+00 7.16512024e-01 -1.44117162e-01 -1.69388816e-01
4.90983367e-01 2.94907894e-02 4.23901200e-01 -4.73907113e-01
-4.79581803e-01 2.80847132e-01 -3.01496953e-01 4.88355756e-01
1.00362718e+00 -3.87844481e-02 8.84233892e-01 6.88567340e-01
-5.52432798e-02 -1.41853476e+00 1.77274887e-02 7.96565175e-01
8.70633543e-01 -8.02919567e-01 8.67318869e-01 -3.61780018e-01
-3.96795660e-01 1.32579827e+00 7.04674497e-02 1.59435317e-01
3.22723269e-01 1.06286243e-01 -1.67293176e-01 -3.14527750e-01
-5.63251972e-01 -4.47590202e-02 8.18719506e-01 -3.45347673e-01
4.68775988e-01 -1.52188137e-01 -1.04360735e+00 5.00406802e-01
-8.06591153e-01 -3.46878976e-01 1.06823516e+00 1.16773105e+00
-7.99552262e-01 -1.14189398e+00 -6.33755207e-01 9.07704115e-01
-1.15469658e+00 -2.83643961e-01 -1.08082116e+00 6.56258821e-01
-4.13504004e-01 1.08689642e+00 -2.73362815e-01 -2.20229343e-01
2.90499806e-01 9.78044197e-02 3.44995081e-01 -7.19161332e-01
-7.47217298e-01 4.16255176e-01 4.03266549e-01 1.52814880e-01
-3.19515079e-01 -6.62462950e-01 -1.04340982e+00 -4.35777754e-01
-4.87338901e-01 9.67440844e-01 4.76422340e-01 1.00137568e+00
1.02213979e-01 7.27649152e-01 5.87979019e-01 -4.09488350e-01
-3.18245441e-01 -1.23862767e+00 -2.38590106e-01 7.98714980e-02
3.50719154e-01 -6.64615259e-02 -1.58572316e-01 4.41892147e-01] | [12.345341682434082, 9.050810813903809] |
3de6da0c-0756-4b2a-8f07-6beced5e3dbf | a-lightweight-reconstruction-network-for | 2212.12878 | null | https://arxiv.org/abs/2212.12878v1 | https://arxiv.org/pdf/2212.12878v1.pdf | A Lightweight Reconstruction Network for Surface Defect Inspection | Currently, most deep learning methods cannot solve the problem of scarcity of industrial product defect samples and significant differences in characteristics. This paper proposes an unsupervised defect detection algorithm based on a reconstruction network, which is realized using only a large number of easily obtained defect-free sample data. The network includes two parts: image reconstruction and surface defect area detection. The reconstruction network is designed through a fully convolutional autoencoder with a lightweight structure. Only a small number of normal samples are used for training so that the reconstruction network can be A defect-free reconstructed image is generated. A function combining structural loss and $\mathit{L}1$ loss is proposed as the loss function of the reconstruction network to solve the problem of poor detection of irregular texture surface defects. Further, the residual of the reconstructed image and the image to be tested is used as the possible region of the defect, and conventional image operations can realize the location of the fault. The unsupervised defect detection algorithm of the proposed reconstruction network is used on multiple defect image sample sets. Compared with other similar algorithms, the results show that the unsupervised defect detection algorithm of the reconstructed network has strong robustness and accuracy. | ['Liqiang Zhu', 'Weibin Qiu', 'Weijie Wu', 'Jian Yao', 'Chao Hu'] | 2022-12-25 | null | null | null | null | ['defect-detection'] | ['computer-vision'] | [ 1.53244674e-01 1.12192728e-01 2.42384970e-01 -2.58135319e-01
-2.60234237e-01 5.31109929e-01 -3.19705367e-01 -1.41489208e-01
-1.61968768e-01 2.72798419e-01 -2.87390918e-01 -2.93873940e-02
-1.23359568e-01 -1.33130920e+00 -4.03251320e-01 -9.73678768e-01
1.59046873e-01 3.57637137e-01 2.84919381e-01 -1.75948918e-01
4.89245117e-01 4.44046736e-01 -1.65433240e+00 3.86382461e-01
8.46269846e-01 1.31200159e+00 7.05874979e-01 1.19871691e-01
-2.59600431e-01 9.09205019e-01 -8.07312250e-01 3.45089227e-01
2.11217985e-01 -6.40899658e-01 -6.01972640e-01 6.98360384e-01
-1.33248344e-01 -7.53840208e-01 -4.55668360e-01 1.23577929e+00
6.44362032e-01 6.36769906e-02 6.45462751e-01 -7.43901670e-01
-5.12549162e-01 1.70167014e-01 -5.07796288e-01 8.98307636e-02
-6.05950914e-02 -9.28308815e-03 5.00283659e-01 -9.94930506e-01
3.79567474e-01 1.16171753e+00 6.55863345e-01 2.71434397e-01
-7.28985190e-01 -3.85300219e-01 -4.33100164e-01 1.41942218e-01
-1.32948446e+00 -4.50593978e-02 1.28663290e+00 -3.98100942e-01
7.37192035e-01 -8.58527571e-02 4.66573030e-01 2.90098518e-01
4.32831228e-01 6.78637445e-01 5.72111309e-01 -6.19732261e-01
1.51207536e-01 -1.21526666e-01 -3.55342962e-02 1.07869709e+00
4.45012927e-01 2.23277017e-01 1.90974221e-01 2.68619239e-01
1.30224800e+00 5.71427763e-01 -3.71941507e-01 -6.01227805e-02
-6.75281584e-01 9.30129528e-01 6.69305503e-01 3.68539035e-01
-3.64386886e-01 -3.39625120e-01 4.35167968e-01 6.26721144e-01
5.10962546e-01 2.30692357e-01 -3.39209080e-01 4.80730861e-01
-7.05880165e-01 -2.77487755e-01 4.22142506e-01 4.59883451e-01
9.78680730e-01 5.73246539e-01 4.11917955e-01 1.20669329e+00
4.83570337e-01 3.36210787e-01 7.58049488e-01 -6.87097132e-01
2.51426518e-01 1.07987106e+00 -1.95443347e-01 -1.37463391e+00
-1.42962515e-01 -2.88989425e-01 -1.07297862e+00 7.29912996e-01
-1.04941707e-02 -5.32945059e-02 -1.02518487e+00 7.38334179e-01
2.28324503e-01 -2.62521446e-01 2.04231516e-02 9.26400721e-01
9.11410987e-01 7.66901255e-01 -6.40850961e-01 -1.35578394e-01
8.86077464e-01 -9.00475979e-01 -7.42614567e-01 -1.29067063e-01
5.88285208e-01 -6.57978475e-01 9.90788996e-01 7.90402889e-01
-9.85747159e-01 -7.73102701e-01 -1.36402273e+00 9.44768041e-02
1.59584321e-02 7.39873230e-01 5.61479688e-01 1.75273463e-01
-7.19569743e-01 7.89850593e-01 -5.63259065e-01 -3.41746360e-02
5.05669534e-01 3.60367090e-01 -4.53796089e-01 -5.35627484e-01
-7.49021053e-01 5.47080040e-01 5.15796721e-01 5.45864165e-01
-1.02882183e+00 -1.52205735e-01 -8.82237673e-01 -3.46889980e-02
8.43895376e-02 -1.32697627e-01 7.08225131e-01 -1.29603958e+00
-1.43632960e+00 7.79045165e-01 4.24977243e-01 -3.86275351e-02
2.14130968e-01 -1.50466273e-02 -5.00894189e-01 4.35610116e-01
2.14522615e-01 4.53663059e-02 8.68410707e-01 -1.38597655e+00
-7.06859589e-01 -2.75239885e-01 -3.77667636e-01 -3.54531556e-02
-1.71054497e-01 -2.09023803e-01 -2.30189905e-01 -6.55422270e-01
8.78088892e-01 -2.11752594e-01 -2.18504205e-01 2.13277236e-01
-3.90342802e-01 -9.12464559e-02 1.24781144e+00 -9.45988894e-01
8.59334946e-01 -2.59471345e+00 -2.36019477e-01 2.99072325e-01
9.52872932e-02 1.00218698e-01 -9.56329405e-02 2.98391193e-01
-2.19921753e-01 -1.14785396e-01 -6.34954095e-01 -1.99103281e-02
-5.96075833e-01 2.84689933e-01 2.07386419e-01 7.72571445e-01
2.18257844e-01 2.15058938e-01 -4.31260526e-01 -4.44405198e-01
3.37041765e-01 1.81664959e-01 -3.60297233e-01 5.09417593e-01
1.46767616e-01 2.53225267e-01 -4.63626832e-01 9.56517160e-01
9.00328159e-01 2.48829965e-02 -1.90434337e-01 -3.81170124e-01
8.01964402e-02 -5.28251473e-03 -1.26954675e+00 1.40330148e+00
-3.34521383e-01 4.26177233e-01 3.70175838e-01 -1.54932213e+00
1.55739748e+00 2.53604263e-01 5.73237836e-01 -8.49110365e-01
3.40742081e-01 4.62568074e-01 -1.45536736e-02 -1.09186506e+00
-2.16369459e-04 -3.76662284e-01 3.50840092e-01 4.62702513e-01
1.14123724e-01 -1.54176027e-01 -1.52416378e-01 -3.00744176e-01
1.12848794e+00 -2.31659129e-01 -2.61197239e-01 -3.34194869e-01
5.26891828e-01 -5.60271330e-02 9.00208652e-01 2.69354880e-01
1.61418505e-02 9.01885808e-01 2.97092229e-01 -7.67495096e-01
-1.11074269e+00 -8.84929776e-01 -1.91001669e-01 2.61168987e-01
5.38400412e-01 3.57429743e-01 -6.02177441e-01 -7.54649162e-01
-4.26230533e-03 1.95788309e-01 -5.27706623e-01 -5.17379880e-01
-6.62150502e-01 -6.73670292e-01 1.76543832e-01 3.70325893e-01
1.05384314e+00 -1.56025469e+00 -2.48320475e-01 1.69985190e-01
-1.01520799e-01 -3.56668442e-01 5.15579842e-02 1.02617688e-01
-1.37963557e+00 -1.45593965e+00 -4.92961466e-01 -1.65185797e+00
1.37521541e+00 3.19173783e-01 8.97554040e-01 8.52927268e-01
-5.99455237e-01 7.27987383e-03 -3.68889272e-01 -1.57801643e-01
-5.21135628e-01 -7.06888199e-01 -1.64358795e-01 7.90629461e-02
1.54861555e-01 -4.72169608e-01 -8.12758148e-01 4.18902218e-01
-1.00286949e+00 -2.92793751e-01 8.21139514e-01 1.25527656e+00
6.57309711e-01 1.22978926e+00 5.96399784e-01 -7.80522346e-01
4.59653586e-01 -3.10210258e-01 -3.15695107e-01 -2.32210279e-01
-7.41985261e-01 -1.59983918e-01 5.32253385e-01 -2.02821046e-01
-1.25234759e+00 -2.40095079e-01 -4.70197737e-01 -3.58886868e-01
-1.91435263e-01 5.77151716e-01 -4.63037610e-01 -3.74689698e-02
5.04046679e-01 4.66239840e-01 7.48385191e-01 -7.85753846e-01
-4.18141156e-01 8.50336432e-01 3.95242721e-01 -1.19703330e-01
4.03792858e-01 4.67524886e-01 -2.15057313e-01 -9.22168314e-01
-1.54810041e-01 -4.36399400e-01 -3.02881598e-01 -3.21623355e-01
4.84140545e-01 -7.54970908e-01 -4.03692096e-01 1.02379704e+00
-1.01406455e+00 -2.64443487e-01 -6.49700284e-01 6.88608825e-01
-4.36801761e-01 5.85909724e-01 -1.15390122e+00 -6.14620328e-01
-4.21469927e-01 -1.22510278e+00 8.06170464e-01 -6.10804260e-02
6.79112792e-01 -8.66290808e-01 -4.24797803e-01 3.20452720e-01
1.74221933e-01 2.07311332e-01 1.23176384e+00 -1.76042244e-01
-4.37075198e-01 -6.67007148e-01 -1.54792592e-01 1.24469078e+00
5.24209261e-01 -1.20542184e-01 -5.21721840e-01 -3.91891748e-01
9.85871732e-01 -3.18310231e-01 1.06771183e+00 4.85867321e-01
1.30942428e+00 -2.54037261e-01 -1.92990646e-01 3.53001148e-01
1.79594612e+00 5.64903080e-01 1.31251633e+00 4.07698452e-01
7.09696054e-01 6.75275743e-01 7.19433725e-01 2.01557934e-01
-2.36665875e-01 -6.55464008e-02 7.43023157e-01 -5.44040442e-01
-2.90255308e-01 -2.42191441e-02 2.78318137e-01 1.23096669e+00
-1.61773656e-02 -1.03815086e-01 -5.80536544e-01 6.84261620e-01
-1.39026403e+00 -7.79902875e-01 -3.03946495e-01 2.03701854e+00
4.64325994e-01 3.40911657e-01 -4.80754316e-01 7.85518408e-01
9.67064440e-01 -1.79404974e-01 -5.94725132e-01 -7.96031356e-02
-1.44161120e-01 2.32372135e-01 1.95570499e-01 1.91433057e-01
-8.41864586e-01 4.29603338e-01 6.33471060e+00 9.57582653e-01
-1.22217381e+00 -1.08854517e-01 6.17513359e-01 4.28471655e-01
-2.45827600e-01 -9.72644240e-02 -1.27864853e-01 5.30474961e-01
9.18475315e-02 6.26176894e-01 7.16609433e-02 9.69378352e-01
2.26660177e-01 -4.36844260e-01 -6.56893075e-01 1.14707959e+00
2.44897366e-01 -1.03744853e+00 6.44103065e-02 -6.26717955e-02
5.06084740e-01 -2.16401398e-01 -6.28618300e-02 8.99006948e-02
-6.19466528e-02 -9.05146062e-01 3.89661133e-01 4.96146798e-01
7.33398914e-01 -8.05708051e-01 1.16524529e+00 3.08785737e-01
-8.49353969e-01 -4.63630289e-01 -8.00281405e-01 -2.61018068e-01
-9.71527398e-02 1.28812718e+00 -4.49865013e-01 5.22373796e-01
7.93252170e-01 8.20493221e-01 -1.87543407e-01 9.96485114e-01
-2.07458928e-01 7.07486272e-01 -2.17659548e-01 5.02772033e-01
-2.33516563e-02 -3.85331482e-01 2.57279038e-01 5.67903817e-01
7.16744006e-01 -1.44150794e-01 2.13287666e-01 8.64559114e-01
-4.94245114e-03 1.85350835e-01 -7.30568707e-01 2.57013172e-01
2.02984527e-01 8.84364545e-01 -8.80717516e-01 -1.38525277e-01
-3.87142658e-01 1.04331875e+00 6.99620768e-02 1.70607761e-01
-1.66694090e-01 -6.97086394e-01 6.86265081e-02 4.93032634e-01
3.12918156e-01 2.85785291e-02 -3.10492843e-01 -8.64020228e-01
1.91454500e-01 -7.21826017e-01 2.57328123e-01 -8.53925705e-01
-1.44287503e+00 4.11833763e-01 -6.14723980e-01 -1.45253479e+00
3.57961833e-01 -7.84653187e-01 -1.03192115e+00 6.99523807e-01
-1.25380516e+00 -7.17778444e-01 -2.95762420e-01 4.47381526e-01
7.41209626e-01 -6.68186724e-01 5.45701265e-01 5.21282554e-01
-8.52513194e-01 2.22539023e-01 2.57791758e-01 3.62825245e-01
2.27594241e-01 -8.86957884e-01 -2.43711218e-01 9.78372574e-01
-5.30225456e-01 1.12544261e-01 3.30690324e-01 -9.39314961e-01
-1.12600303e+00 -1.01269078e+00 4.25487787e-01 4.15665299e-01
-6.86815456e-02 6.34079650e-02 -9.55706835e-01 2.67820060e-01
-4.72739376e-02 2.15002686e-01 1.43844113e-01 -5.17926216e-01
1.87841028e-01 -3.01298678e-01 -1.74581552e+00 1.20199490e-02
5.69328129e-01 -2.66621739e-01 -4.43767369e-01 2.73392975e-01
5.20289958e-01 -1.20334134e-01 -9.56599951e-01 7.50498354e-01
1.41668081e-01 -1.07641339e+00 8.13513517e-01 1.26564816e-01
7.62044728e-01 -4.11437929e-01 -2.14992650e-02 -1.06433213e+00
-4.74504411e-01 2.80209392e-01 2.71180302e-01 1.02393425e+00
1.87417284e-01 -7.05021441e-01 8.57320189e-01 -1.71030924e-01
-6.79457366e-01 -9.49778259e-01 -6.38360918e-01 -5.35759032e-01
-1.31920263e-01 -1.72072146e-02 3.69621515e-01 9.05038297e-01
-3.94585669e-01 -5.05934842e-02 -1.04960397e-01 1.21279873e-01
5.26772320e-01 5.24725206e-02 2.89564759e-01 -1.33114040e+00
-3.09052644e-03 -2.00108111e-01 -6.78373516e-01 -9.47790325e-01
-1.87463492e-01 -6.52415633e-01 3.85874182e-01 -1.88562584e+00
-1.59648523e-01 -7.77482212e-01 -3.49970996e-01 3.20251077e-01
1.10272765e-01 1.52288616e-01 -6.11563504e-01 5.75020313e-01
-2.45506335e-02 7.55842090e-01 1.71510518e+00 -5.62001944e-01
-8.36438760e-02 -4.66834642e-02 -2.96296597e-01 7.27741480e-01
9.11060929e-01 -3.99472684e-01 -4.53308821e-01 -5.48984706e-01
-9.36967432e-02 1.36646792e-01 1.93744704e-01 -1.14183271e+00
7.80692697e-02 6.32979050e-02 5.94210386e-01 -6.49906933e-01
-6.71019182e-02 -9.55009758e-01 -8.81838724e-02 6.97053730e-01
1.40520245e-01 -2.95650840e-01 -1.30349815e-01 4.93280172e-01
-6.93585634e-01 -6.65401459e-01 9.28929567e-01 -6.35132968e-01
-7.72132635e-01 3.06087464e-01 -3.65295112e-01 -4.79323357e-01
8.81259620e-01 -6.43430889e-01 1.89975705e-02 4.14262563e-02
-5.54663777e-01 -1.74134746e-01 4.21231240e-01 1.32473975e-01
1.35447145e+00 -1.43084371e+00 -6.39501333e-01 8.04931641e-01
-7.48614594e-02 6.00944877e-01 6.33377194e-01 5.48903167e-01
-1.28268623e+00 -2.94732600e-01 -3.97060364e-01 -5.13886869e-01
-6.32409632e-01 5.41940570e-01 6.60492063e-01 -9.83582959e-02
-1.10237586e+00 8.54410589e-01 9.42474008e-02 -4.89990801e-01
6.40260428e-02 -5.59689477e-02 -2.37137496e-01 -3.57441366e-01
4.15351421e-01 5.38603842e-01 2.88444161e-01 -5.35831213e-01
1.34358317e-01 4.59974885e-01 5.23093864e-02 4.46173608e-01
1.49134302e+00 -9.38533917e-02 -6.66679502e-01 2.08209023e-01
1.33555412e+00 -2.67033637e-01 -1.26527679e+00 -4.75619175e-02
-4.94409204e-01 -4.22495097e-01 5.23869753e-01 -5.26239932e-01
-1.94841504e+00 8.05745244e-01 1.08034325e+00 2.34433576e-01
1.41410303e+00 -3.20651412e-01 1.13934588e+00 2.55660534e-01
3.83418292e-01 -1.31457686e+00 5.68767369e-01 1.38193518e-01
8.79357755e-01 -1.44686747e+00 4.76599000e-02 -3.54700476e-01
-1.62220180e-01 1.38210499e+00 8.46110642e-01 -7.23017693e-01
9.77983832e-01 2.57781416e-01 1.10201843e-01 -7.85434842e-01
5.79684526e-02 1.12933554e-01 -2.56928474e-01 5.92527092e-01
5.79950325e-02 -2.27784261e-01 -4.29073811e-01 4.97316629e-01
2.65122741e-01 -7.51316100e-02 4.61273730e-01 1.20398843e+00
-8.69881392e-01 -8.24840188e-01 -4.86298323e-01 8.30406427e-01
-4.26305801e-01 2.52685964e-01 2.97768682e-01 6.06404603e-01
5.16273737e-01 9.75559413e-01 3.21449488e-01 -6.10873759e-01
3.08468640e-01 -4.23302144e-01 3.21290493e-01 -6.30036175e-01
-1.16716325e-01 -5.26540130e-02 -3.13360572e-01 -3.47312301e-01
-1.70841292e-01 -1.38219267e-01 -1.50175071e+00 -8.74762312e-02
-7.75005400e-01 1.28034949e-01 5.89659572e-01 8.35590065e-01
2.14111675e-02 6.82529390e-01 1.16668439e+00 -6.88806355e-01
-2.34901398e-01 -1.14976835e+00 -1.32679868e+00 5.49248815e-01
3.61113966e-01 -7.80702770e-01 -5.51699638e-01 1.32834107e-01] | [7.3997979164123535, 1.7890764474868774] |
c9361db7-d660-4b59-98ad-e71bfc173e9b | end-to-end-prostate-cancer-detection-in-bpmri | 2101.03244 | null | https://arxiv.org/abs/2101.03244v9 | https://arxiv.org/pdf/2101.03244v9.pdf | End-to-end Prostate Cancer Detection in bpMRI via 3D CNNs: Effects of Attention Mechanisms, Clinical Priori and Decoupled False Positive Reduction | We present a multi-stage 3D computer-aided detection and diagnosis (CAD) model for automated localization of clinically significant prostate cancer (csPCa) in bi-parametric MR imaging (bpMRI). Deep attention mechanisms drive its detection network, targeting salient structures and highly discriminative feature dimensions across multiple resolutions. Its goal is to accurately identify csPCa lesions from indolent cancer and the wide range of benign pathology that can afflict the prostate gland. Simultaneously, a decoupled residual classifier is used to achieve consistent false positive reduction, without sacrificing high sensitivity or computational efficiency. In order to guide model generalization with domain-specific clinical knowledge, a probabilistic anatomical prior is used to encode the spatial prevalence and zonal distinction of csPCa. Using a large dataset of 1950 prostate bpMRI paired with radiologically-estimated annotations, we hypothesize that such CNN-based models can be trained to detect biopsy-confirmed malignancies in an independent cohort. For 486 institutional testing scans, the 3D CAD system achieves 83.69$\pm$5.22% and 93.19$\pm$2.96% detection sensitivity at 0.50 and 1.46 false positive(s) per patient, respectively, with 0.882$\pm$0.030 AUROC in patient-based diagnosis $-$significantly outperforming four state-of-the-art baseline architectures (U-SEResNet, UNet++, nnU-Net, Attention U-Net) from recent literature. For 296 external biopsy-confirmed testing scans, the ensembled CAD system shares moderate agreement with a consensus of expert radiologists (76.69%; $kappa$ $=$ 0.51$\pm$0.04) and independent pathologists (81.08%; $kappa$ $=$ 0.56$\pm$0.06); demonstrating strong generalization to histologically-confirmed csPCa diagnosis. | ['Henkjan Huisman', 'Matin Hosseinzadeh', 'Anindo Saha'] | 2021-01-08 | null | null | null | null | ['deep-attention', 'clinical-knowledge', 'deep-attention'] | ['computer-vision', 'miscellaneous', 'natural-language-processing'] | [ 3.46823335e-01 6.62358105e-01 -5.06435394e-01 -2.82366574e-01
-1.62541652e+00 -5.81933856e-01 4.28366750e-01 1.75062880e-01
-4.54810560e-01 5.66374660e-01 1.06846362e-01 -5.26362717e-01
-3.55746299e-01 -5.97321510e-01 -4.52081382e-01 -7.29655206e-01
-6.93686485e-01 9.29513931e-01 -1.32829458e-01 3.61310303e-01
-5.74242175e-02 5.74171364e-01 -3.27264369e-01 2.90398430e-02
7.15477765e-01 1.10127223e+00 5.79788864e-01 7.22466648e-01
3.31033915e-01 5.72500348e-01 -1.94662511e-01 -4.27264512e-01
1.14071548e-01 -1.26415074e-01 -7.50938475e-01 -2.91706443e-01
7.78530896e-01 -1.98714644e-01 -4.48936284e-01 1.14246511e+00
4.96089190e-01 -5.06230175e-01 1.22929156e+00 -3.95670652e-01
-6.93736672e-01 5.69184721e-01 -9.37107265e-01 5.64379215e-01
-3.70588690e-01 3.93191099e-01 7.91867435e-01 -5.95575929e-01
7.59676635e-01 4.82144564e-01 1.11723638e+00 7.25729227e-01
-1.36019981e+00 -7.78903484e-01 -3.77005041e-01 -3.00191730e-01
-1.32608306e+00 4.27958257e-02 4.72280681e-01 -4.94179755e-01
8.81250858e-01 5.47031164e-01 6.01819515e-01 1.00534451e+00
1.23841274e+00 4.02063191e-01 9.93314683e-01 1.20070949e-01
1.96729600e-02 -1.11065976e-01 1.13573648e-01 9.77384984e-01
3.42617780e-01 1.50855049e-01 1.76303431e-01 -2.62619376e-01
1.24245501e+00 7.27279261e-02 -1.05247311e-01 -3.20192128e-01
-1.09173238e+00 8.06833565e-01 1.29203475e+00 2.61009693e-01
-3.11541706e-01 3.38067800e-01 4.48717743e-01 -4.43876833e-01
1.47693425e-01 6.68459058e-01 6.27557710e-02 3.49833578e-01
-7.44431734e-01 2.59627402e-01 2.74956077e-01 6.00440979e-01
-1.71982143e-02 -2.08931342e-01 -2.44806856e-01 9.40159142e-01
2.71453500e-01 8.73746514e-01 8.59900057e-01 -8.38563859e-01
-5.27343750e-02 5.53410590e-01 -2.88159311e-01 -8.38950157e-01
-1.11282003e+00 -1.31495512e+00 -1.38771963e+00 -1.37829214e-01
3.42064828e-01 2.87189424e-01 -1.46233678e+00 1.50848782e+00
-1.33455977e-01 -2.07924068e-01 -3.08863014e-01 8.13630998e-01
8.54613960e-01 2.89533306e-02 4.47695553e-01 -8.95726122e-03
1.82479191e+00 -5.66015005e-01 -7.37904012e-02 -5.95042169e-01
8.35725367e-01 -2.68813819e-01 7.45986223e-01 3.95607464e-02
-1.05846798e+00 -2.10633650e-01 -8.68132591e-01 6.38303608e-02
1.36253878e-01 5.70640683e-01 8.50588381e-01 6.62413478e-01
-1.14248991e+00 1.92777917e-01 -1.26100349e+00 -3.16175461e-01
8.99877310e-01 6.57437623e-01 -3.40081453e-01 -2.34237745e-01
-6.76884711e-01 1.17923963e+00 2.43536178e-02 2.20907345e-01
-1.57803690e+00 -1.18360078e+00 -7.82427251e-01 4.07981537e-02
1.14148647e-01 -8.81436169e-01 1.32639098e+00 -3.59678835e-01
-7.72636533e-01 1.18190038e+00 -2.34084483e-02 -8.39589536e-01
5.33848047e-01 4.96280611e-01 -2.52730101e-01 2.37990126e-01
6.42027855e-01 7.78496385e-01 2.10621268e-01 -1.07524610e+00
-5.25937676e-01 -1.00351167e+00 -6.18704617e-01 4.08839464e-01
1.67100757e-01 -4.64286536e-01 -5.26770651e-01 -4.15062159e-01
4.26390201e-01 -1.18263066e+00 -8.41824055e-01 -2.36281548e-02
-5.21491945e-01 1.36241332e-01 3.65209073e-01 -8.59296143e-01
6.37923479e-01 -2.06912613e+00 -5.95683139e-03 4.09202248e-01
7.71968603e-01 -1.44801319e-01 2.62228191e-01 -6.53952777e-01
4.17427626e-04 3.53475034e-01 -4.80555594e-01 -2.42140397e-01
-1.99851617e-01 -1.74028445e-02 3.65383983e-01 7.27544546e-01
2.26899952e-01 1.27146232e+00 -9.41057861e-01 -5.11189640e-01
2.61933386e-01 3.35963577e-01 -4.96083617e-01 -1.74128652e-01
3.39445800e-01 3.95223945e-01 -4.07326609e-01 1.03827417e+00
4.52767253e-01 -7.20866382e-01 5.85441709e-01 -2.68805295e-01
4.11756665e-01 -2.31557444e-01 -3.71274322e-01 1.77223194e+00
-4.76370901e-01 4.09957796e-01 6.05227292e-01 -7.33963192e-01
7.56623030e-01 2.00763159e-02 6.46253645e-01 -8.12124550e-01
-1.32982228e-02 5.09988248e-01 5.34349084e-01 6.41062809e-03
1.33638948e-01 -4.56935525e-01 -3.61352712e-01 -5.52196652e-02
2.11738899e-01 -3.21054041e-01 -2.32808650e-01 8.11598375e-02
1.78984892e+00 -5.09741187e-01 3.49606365e-01 -8.01742852e-01
3.44670475e-01 3.24461550e-01 7.04187751e-01 1.14487028e+00
-7.12470233e-01 5.79956293e-01 4.85263914e-01 -3.62548530e-01
-7.47578561e-01 -1.65754318e+00 -8.88325870e-01 5.86984992e-01
4.59134467e-02 1.48825586e-01 -2.21156865e-01 -8.74382555e-01
2.01910749e-01 3.47424448e-01 -9.85666394e-01 -1.06079690e-01
-6.33212030e-01 -1.11649227e+00 7.58644223e-01 8.66745591e-01
3.31604809e-01 -7.73376703e-01 -4.54555571e-01 2.22820550e-01
5.41296974e-02 -6.28464520e-01 -1.98164999e-01 6.50409222e-01
-1.06354904e+00 -1.11558664e+00 -1.21240628e+00 -7.94566572e-01
8.67460430e-01 -2.03073949e-01 1.21039486e+00 -4.46650952e-01
-8.35456848e-01 1.84226871e-01 3.75414431e-01 -1.84506312e-01
-2.26961106e-01 2.04840302e-01 -1.72929645e-01 -7.21331716e-01
4.85433817e-01 -2.30476126e-01 -8.33706200e-01 2.42382228e-01
-4.51723039e-01 -1.75946634e-02 1.43486345e+00 1.15065289e+00
9.82643485e-01 -2.65849978e-01 4.29015994e-01 -7.30554283e-01
3.40860575e-01 -5.54157436e-01 -1.53805509e-01 -7.73337185e-02
-5.65415204e-01 -4.49033707e-01 3.90851527e-01 -2.28188679e-01
-7.89219558e-01 2.00855553e-01 2.92821862e-02 -4.25960422e-01
-1.48674816e-01 5.90328395e-01 4.09990847e-01 -1.23062640e-01
9.54631150e-01 3.01177680e-01 6.15025386e-02 -4.25270805e-03
-2.74625514e-02 1.78913504e-01 1.10436690e+00 -3.08347166e-01
4.40140486e-01 4.83204901e-01 2.64760494e-01 -4.61295247e-01
-4.62464362e-01 -3.20272356e-01 -3.30041528e-01 -1.28354982e-01
1.02980053e+00 -1.03378582e+00 -5.81723511e-01 3.00077181e-02
-5.49375772e-01 -4.97166157e-01 -2.18368426e-01 5.81976354e-01
-5.22004128e-01 -6.85227141e-02 -9.50663924e-01 -4.05624747e-01
-7.33103395e-01 -1.53340197e+00 1.26190841e+00 3.12976629e-01
-4.25162792e-01 -8.53915811e-01 -1.36411041e-01 3.49273980e-01
7.01574147e-01 4.55279976e-01 9.94930804e-01 -7.02238798e-01
-4.14570928e-01 -6.15867198e-01 -6.21072233e-01 -2.60541439e-01
1.33781075e-01 -4.09049302e-01 -7.85475135e-01 -2.99896032e-01
-2.78709471e-01 -1.12952404e-01 8.96202147e-01 1.15415347e+00
1.38544679e+00 2.74880618e-01 -9.79829669e-01 6.75099194e-01
1.52934873e+00 5.02694070e-01 2.99224883e-01 2.13895097e-01
5.48815548e-01 -1.30785123e-01 1.18433200e-01 1.64005924e-02
2.43164524e-01 3.73714119e-01 7.37676382e-01 -1.06165096e-01
-1.47918388e-01 -9.07449890e-03 -2.72441924e-01 8.77828300e-02
-1.29329070e-01 3.54243964e-01 -1.62434518e+00 9.02511477e-01
-1.14299476e+00 -4.74224269e-01 -3.30682052e-03 1.95734739e+00
7.55422115e-01 5.39899886e-01 -3.47404301e-01 -6.53339922e-01
7.06919730e-01 1.94023345e-02 -7.76670754e-01 -7.94336870e-02
4.40392077e-01 3.12318683e-01 1.04289198e+00 4.04482633e-01
-1.30874288e+00 2.88936019e-01 5.69692850e+00 8.12539101e-01
-1.21336710e+00 3.31371963e-01 1.33102238e+00 -3.14530522e-01
1.78260788e-01 -5.92909515e-01 -7.01607883e-01 3.30430299e-01
9.13482606e-01 -3.37153375e-02 -1.29980832e-01 1.03399646e+00
-1.72961593e-01 -1.70355380e-01 -1.08660734e+00 9.39338088e-01
-8.36841203e-03 -1.71430457e+00 -5.41935742e-01 5.11564791e-01
6.36372864e-01 4.93692696e-01 3.88258666e-01 6.72612965e-01
5.14351010e-01 -1.54426622e+00 -7.54582137e-02 5.25495410e-01
1.38410282e+00 -5.46346188e-01 1.31182647e+00 5.05471788e-02
-1.04246449e+00 1.07635474e-02 -1.51221603e-02 5.08102596e-01
-2.39624038e-01 4.33126003e-01 -1.32789898e+00 4.64447498e-01
8.59308898e-01 5.33235371e-01 -6.87082410e-01 6.90267026e-01
1.10606179e-01 4.07524049e-01 -3.86758089e-01 6.62995279e-02
3.99033487e-01 5.96954048e-01 3.50959688e-01 1.33211207e+00
2.08836034e-01 1.86060145e-01 -1.16470426e-01 9.93728518e-01
-1.50393724e-01 -2.20675245e-01 -3.97844315e-01 1.79911226e-01
3.06591779e-01 1.58100188e+00 -8.90432239e-01 -1.31812990e-01
4.21626456e-02 6.19884670e-01 2.10266933e-01 1.39803172e-03
-7.66180396e-01 -3.60300466e-02 1.32894203e-01 2.49164909e-01
5.04244529e-02 1.56189710e-01 -5.91517866e-01 -7.24633574e-01
-3.01951408e-01 -4.33273226e-01 6.06625915e-01 -5.51127076e-01
-1.54838741e+00 5.34075201e-01 -3.61300796e-01 -8.85122001e-01
-2.51676619e-01 -8.71204793e-01 -6.07816637e-01 1.05744982e+00
-1.22731006e+00 -1.27819979e+00 -3.48592699e-01 1.02980472e-01
2.71243721e-01 1.26586268e-02 1.12684250e+00 -5.60403429e-03
-4.24761772e-01 6.42621279e-01 7.85814822e-02 5.20094633e-01
7.39589751e-01 -1.71525872e+00 -1.56753540e-01 2.59751409e-01
-8.59788835e-01 6.67993665e-01 1.79970235e-01 -8.73039544e-01
-1.51373911e+00 -1.37597716e+00 4.20962244e-01 -5.25415897e-01
7.81295419e-01 4.65156287e-02 -4.99110848e-01 9.96408284e-01
-2.64848739e-01 4.32637095e-01 7.68525660e-01 4.78211716e-02
4.07681428e-02 -1.63022518e-01 -1.62444210e+00 4.98426408e-01
7.46307135e-01 -2.84089386e-01 -3.27594995e-01 1.98262975e-01
1.61955982e-01 -8.66941035e-01 -1.54285920e+00 9.11422074e-01
5.31003356e-01 -5.42749166e-01 1.20980060e+00 -1.23363055e-01
5.84761858e-01 1.49940491e-01 -1.00311048e-01 -9.75199759e-01
-9.23802435e-01 3.18006933e-01 6.28234968e-02 4.41143215e-01
5.87556899e-01 -3.67508143e-01 1.21978319e+00 7.69063115e-01
-5.91359198e-01 -1.10290062e+00 -1.29226601e+00 -2.99324393e-01
5.98069549e-01 -3.64217550e-01 -1.50828451e-01 1.00974703e+00
-1.02299370e-01 -8.26693550e-02 3.80962312e-01 4.40143675e-01
8.31455886e-01 -9.37639698e-02 9.78060737e-02 -1.15990114e+00
-3.82295430e-01 -9.81463969e-01 -7.03468442e-01 -5.80635846e-01
-3.47673655e-01 -1.25291646e+00 -2.30949387e-01 -1.66909158e+00
8.58870804e-01 -5.66455126e-01 -5.66917658e-01 4.38291728e-01
-4.92463186e-02 5.58191419e-01 -1.50407061e-01 3.91243011e-01
-2.92282641e-01 1.82527155e-01 1.38722408e+00 -4.02206063e-01
1.11414567e-01 -1.31755367e-01 -9.21202540e-01 7.68129408e-01
6.63011611e-01 -3.11116755e-01 -1.62918139e-02 -8.04818869e-02
-3.55385810e-01 5.10310113e-01 7.72405505e-01 -1.17602575e+00
4.04070970e-03 -1.13794841e-01 1.27923119e+00 -7.52958894e-01
5.35913646e-01 -5.78080833e-01 9.73243788e-02 1.05382228e+00
-2.56740391e-01 -3.05582941e-01 4.37865853e-01 6.85984433e-01
1.63038671e-01 3.67887244e-02 8.12890589e-01 -6.15238607e-01
-5.17659307e-01 4.47998673e-01 -3.61819446e-01 -2.69345582e-01
1.07386720e+00 1.05348267e-02 -5.61701477e-01 4.89721745e-02
-1.27935469e+00 3.08068424e-01 1.90374374e-01 -1.66014358e-02
4.41127062e-01 -1.30013728e+00 -7.80910552e-01 -1.60496846e-01
1.24741852e-01 2.91503876e-01 6.32277310e-01 1.23833370e+00
-5.33991218e-01 8.37008357e-01 -1.49735920e-02 -1.16747093e+00
-8.61639082e-01 -2.61644162e-02 1.01966310e+00 -8.93937111e-01
-5.66135943e-01 1.19017816e+00 4.99108851e-01 -7.52119660e-01
5.25502004e-02 -5.29578924e-01 2.01068804e-01 -5.07507443e-01
3.97788197e-01 -9.41315293e-02 1.84839308e-01 -2.40834892e-01
-6.71397626e-01 2.71615565e-01 -6.28086507e-01 7.43427649e-02
1.16721463e+00 3.70152920e-01 7.44900182e-02 -6.97465390e-02
1.05513549e+00 -2.23655894e-01 -1.13739526e+00 -3.09599429e-01
-2.14664280e-01 -6.31529465e-02 3.98771018e-01 -1.46267164e+00
-9.89566267e-01 3.70172173e-01 1.25394022e+00 -9.76161733e-02
9.43997979e-01 5.40454566e-01 2.24475205e-01 -1.20262302e-01
3.33286971e-01 -4.13240165e-01 -2.41428033e-01 1.02928095e-01
6.50223434e-01 -1.52878797e+00 8.34341720e-02 -1.99160621e-01
-5.80912590e-01 8.38924348e-01 8.85082841e-01 -4.87256259e-01
5.21151125e-01 3.68588805e-01 -1.27211824e-01 -6.90216124e-01
-5.51802456e-01 3.54912996e-01 3.10720831e-01 6.15460873e-01
6.00675106e-01 7.58579910e-01 -3.36680233e-01 1.13869166e+00
-2.04199061e-01 -2.20261440e-01 3.12416852e-01 8.25033307e-01
-3.55118006e-01 -3.82218540e-01 -2.36684367e-01 1.13401484e+00
-5.39134800e-01 2.01033568e-03 -2.79629648e-01 1.07837665e+00
-1.22407362e-01 1.14047363e-01 2.94197500e-01 -5.40105700e-02
1.15919262e-01 -1.79192603e-01 3.64817500e-01 -6.42447829e-01
-6.88019216e-01 3.89110059e-01 -3.44643891e-02 -4.25252378e-01
1.29103407e-01 -6.75575674e-01 -1.40338838e+00 4.83147912e-02
-6.46819174e-02 -5.29779680e-02 5.70537388e-01 4.72437859e-01
2.68787831e-01 9.54448164e-01 1.67739034e-01 -7.69974470e-01
-5.28349042e-01 -1.15870488e+00 -7.39544570e-01 -1.52345002e-01
7.71487430e-02 -5.17518938e-01 -8.85337666e-02 -4.58664745e-01] | [14.910821914672852, -2.47361159324646] |
d07c0205-95b0-451f-8eba-f7d703622b85 | sixo-smoothing-inference-with-twisted | 2206.05952 | null | https://arxiv.org/abs/2206.05952v2 | https://arxiv.org/pdf/2206.05952v2.pdf | SIXO: Smoothing Inference with Twisted Objectives | Sequential Monte Carlo (SMC) is an inference algorithm for state space models that approximates the posterior by sampling from a sequence of target distributions. The target distributions are often chosen to be the filtering distributions, but these ignore information from future observations, leading to practical and theoretical limitations in inference and model learning. We introduce SIXO, a method that instead learns targets that approximate the smoothing distributions, incorporating information from all observations. The key idea is to use density ratio estimation to fit functions that warp the filtering distributions into the smoothing distributions. We then use SMC with these learned targets to define a variational objective for model and proposal learning. SIXO yields provably tighter log marginal lower bounds and offers significantly more accurate posterior inferences and parameter estimates in a variety of domains. | ['Scott Linderman', 'Andrew Warrington', 'Allan Raventós', 'Dieterich Lawson'] | 2022-06-13 | null | null | null | null | ['density-ratio-estimation'] | ['methodology'] | [ 1.59148201e-02 4.66472730e-02 -6.72698081e-01 -2.86599517e-01
-1.13719702e+00 -5.77786505e-01 1.19111967e+00 -2.92845130e-01
-2.63708770e-01 1.20093310e+00 2.20364138e-01 -2.99471051e-01
-2.39974484e-01 -6.23714030e-01 -6.61680222e-01 -7.48897493e-01
-1.51190341e-01 1.03271270e+00 5.03327906e-01 2.65119672e-01
2.66941786e-01 3.58607769e-01 -1.21118355e+00 -2.70016193e-01
7.57417142e-01 5.42413473e-01 4.10558023e-02 9.50114310e-01
2.66761690e-01 4.47624862e-01 -5.39858997e-01 -2.28963017e-01
1.33015499e-01 -5.53022921e-01 -2.19643742e-01 -3.15026604e-02
1.37877941e-01 -5.94356418e-01 -3.61776352e-01 1.31959784e+00
1.30815029e-01 4.46261436e-01 1.24804592e+00 -1.05817056e+00
-1.81372628e-01 5.79303265e-01 -4.08359498e-01 1.38307363e-01
1.20917112e-01 3.31625104e-01 8.86381865e-01 -4.74317849e-01
4.01330292e-01 1.63435030e+00 6.05065167e-01 5.13785243e-01
-1.99496484e+00 -5.16537607e-01 1.13990687e-01 -8.68636519e-02
-1.44666445e+00 -5.63242793e-01 3.52487892e-01 -4.38376695e-01
5.90791106e-01 1.61780819e-01 7.13019907e-01 1.31729341e+00
6.19483531e-01 6.52605355e-01 1.16971135e+00 -3.67538065e-01
8.00548792e-01 6.90218583e-02 -5.83807267e-02 4.25888449e-01
4.10288841e-01 8.16754580e-01 -4.14298862e-01 -8.07250500e-01
1.10548985e+00 -2.52440795e-02 -7.55879357e-02 -4.18999851e-01
-8.05659115e-01 1.15256524e+00 -2.37857118e-01 -3.60612959e-01
-3.48327518e-01 8.33165109e-01 7.69441798e-02 1.22726999e-01
3.18355978e-01 1.17217921e-01 -4.39755023e-01 -1.80660293e-01
-9.99740720e-01 9.55405653e-01 1.08783865e+00 7.20702171e-01
7.99726009e-01 2.67379075e-01 -4.21732336e-01 3.55372220e-01
8.91613901e-01 1.00479376e+00 -1.11328371e-01 -1.54290068e+00
1.75448701e-01 -4.60428089e-01 1.03544211e+00 -1.33502021e-01
1.51392102e-01 -3.14198643e-01 -3.00696969e-01 4.41756099e-01
8.17474186e-01 -3.83678317e-01 -1.25468051e+00 2.01797414e+00
3.59665006e-01 4.51079577e-01 -1.72778815e-01 5.92965841e-01
-4.63740140e-01 8.16481590e-01 8.27784687e-02 -5.07552207e-01
8.37910950e-01 -4.24707055e-01 -8.42367947e-01 -3.94935012e-01
3.08298841e-02 -6.28318548e-01 6.51849985e-01 3.12504649e-01
-1.12288094e+00 -3.36391330e-01 -8.47825289e-01 5.26317179e-01
1.23239867e-01 -1.85243174e-01 4.82947260e-01 7.98813879e-01
-8.50147724e-01 9.50943291e-01 -1.53837419e+00 -6.97134361e-02
2.88136840e-01 1.51211068e-01 5.27704716e-01 3.31128567e-01
-1.00760376e+00 1.15714872e+00 5.12662530e-01 -2.79064476e-01
-1.61268628e+00 -7.96691656e-01 -5.62644482e-01 -2.54571158e-02
3.89815122e-01 -7.20546365e-01 1.73002839e+00 -3.26374888e-01
-1.94334519e+00 1.06250860e-01 -4.47495967e-01 -9.04842675e-01
4.88708287e-01 -1.70907781e-01 -2.84357309e-01 -1.26411960e-01
-2.18477398e-02 4.56345290e-01 1.33993530e+00 -1.10581565e+00
-6.06757283e-01 -9.55776498e-02 -2.76728541e-01 3.16580646e-02
4.41319495e-01 -2.84670860e-01 -1.97856173e-01 -2.88488477e-01
-9.52671394e-02 -9.41034079e-01 -5.26804090e-01 1.98541693e-02
-3.71055365e-01 -2.42172569e-01 6.34004772e-01 -5.67565143e-01
1.27078903e+00 -1.64283121e+00 1.18316248e-01 4.81330872e-01
2.74371449e-02 -3.82250361e-02 9.26664770e-02 5.33626020e-01
4.18760598e-01 -2.41894394e-01 -1.44494832e-01 -2.97058046e-01
5.27928293e-01 3.85224760e-01 -9.17567432e-01 7.23830223e-01
-4.63051209e-03 6.73441529e-01 -8.53702426e-01 -3.69124472e-01
3.74227107e-01 2.30457425e-01 -7.27270663e-01 2.32835308e-01
-9.19891775e-01 6.36830330e-01 -6.17515385e-01 3.38871293e-02
5.39529264e-01 -9.62297395e-02 4.70523208e-01 2.25792155e-01
7.85542354e-02 2.93321788e-01 -1.21403897e+00 1.21311998e+00
-9.25317779e-02 2.11746842e-01 1.45975873e-01 -9.49855804e-01
7.55087078e-01 2.47274712e-01 9.37099680e-02 2.59784311e-01
3.16055715e-01 8.52150470e-02 -1.86334550e-01 -1.45449072e-01
2.38579750e-01 -7.17503190e-01 -3.81756395e-01 4.84338701e-01
1.75585419e-01 -7.03594267e-01 5.51950410e-02 7.29406998e-02
6.92882538e-01 4.53642011e-01 5.13882875e-01 -3.31014216e-01
-9.97621417e-02 -1.10294588e-01 7.80504823e-01 1.27259314e+00
-3.41226280e-01 1.89057812e-01 5.54703236e-01 -4.49704342e-02
-1.37422609e+00 -2.00260806e+00 -3.14043760e-01 8.22411716e-01
-1.82493523e-01 -2.54194736e-01 -6.11892879e-01 -2.97780365e-01
2.66822040e-01 1.14568019e+00 -5.92121303e-01 -3.68819386e-01
-3.89324546e-01 -7.20782399e-01 3.07291031e-01 4.90664303e-01
-7.48227164e-02 -3.06323200e-01 -2.08009228e-01 4.75534171e-01
5.03231920e-02 -6.97664976e-01 -6.31306052e-01 4.52187657e-02
-1.08631682e+00 -8.47213447e-01 -5.31068206e-01 -1.05306566e-01
3.68858635e-01 -5.22344351e-01 8.68059337e-01 -5.99578440e-01
1.61845207e-01 1.98832318e-01 3.48321229e-01 -4.72310871e-01
-9.06383514e-01 -3.77817631e-01 4.48468685e-01 -1.27499729e-01
3.04132015e-01 -6.44604146e-01 -1.45243421e-01 2.00989246e-01
-6.10059559e-01 -1.81969181e-01 2.40738481e-01 9.32630658e-01
4.97693092e-01 4.33352329e-02 4.41771477e-01 -6.57841384e-01
6.39514983e-01 -3.77493650e-01 -1.39839792e+00 2.00011224e-01
-4.52633023e-01 4.94105577e-01 5.51706672e-01 -7.38388419e-01
-1.38556731e+00 -7.36590773e-02 1.03629179e-01 -7.01029778e-01
-1.27910316e-01 2.79678792e-01 1.40465274e-01 4.34171826e-01
6.89291835e-01 1.62799433e-01 3.01415354e-01 -5.71990073e-01
5.13680458e-01 2.90493548e-01 7.11367846e-01 -1.00357676e+00
9.14593458e-01 5.72283566e-01 1.56025261e-01 -5.09026229e-01
-1.24119830e+00 1.61045879e-01 -2.24459410e-01 -9.63946953e-02
6.06616378e-01 -7.74255335e-01 -8.50322247e-01 1.08045317e-01
-8.75884831e-01 -6.10453248e-01 -5.06088078e-01 9.19279039e-01
-1.14395487e+00 2.17271164e-01 -5.27435660e-01 -1.53065920e+00
1.75249949e-01 -7.50680268e-01 8.63093734e-01 4.92361039e-01
-4.15210813e-01 -1.24402344e+00 4.92831558e-01 -2.28075907e-01
2.34171405e-01 1.32195756e-01 8.36691678e-01 -4.41923559e-01
-6.59697771e-01 -2.25926444e-01 2.84928948e-01 8.33395571e-02
-1.94778100e-01 2.24616662e-01 -6.75952971e-01 -6.77267730e-01
8.15576017e-02 -9.04076248e-02 9.37324286e-01 1.11360788e+00
8.36726487e-01 -3.24746221e-01 -7.13378549e-01 2.58367687e-01
1.33505428e+00 3.00610811e-01 4.43801373e-01 -2.66454309e-01
2.92340457e-03 1.85873881e-01 5.36178231e-01 7.54808664e-01
-6.22885786e-02 5.69449782e-01 -3.86091764e-03 7.50427127e-01
1.52006403e-01 -8.13784361e-01 6.31872594e-01 3.80115539e-01
1.60012707e-01 1.01428546e-01 -6.46951318e-01 6.42989039e-01
-1.99988234e+00 -1.23877275e+00 2.93123126e-01 2.50654674e+00
1.16942644e+00 5.63857138e-01 3.59589010e-01 -5.65463126e-01
8.54278743e-01 -1.11620286e-02 -8.78400624e-01 -2.47271135e-02
4.34022099e-01 4.73305106e-01 7.56487310e-01 1.06802154e+00
-9.73723173e-01 7.43737698e-01 8.53094769e+00 1.12791014e+00
-3.82994026e-01 2.79657394e-01 2.36992180e-01 -3.11382860e-01
-3.06196660e-01 5.93770623e-01 -1.37325788e+00 6.60909355e-01
1.36529374e+00 -5.18667102e-01 6.87639117e-01 6.48909450e-01
5.53267360e-01 -3.56511921e-01 -1.16620827e+00 4.29471195e-01
-6.25275135e-01 -1.39320350e+00 -1.45495489e-01 1.85246512e-01
1.17285979e+00 -6.22387789e-03 2.10802019e-01 4.52817470e-01
1.43918443e+00 -8.61461937e-01 9.01373744e-01 1.02591169e+00
5.98548412e-01 -8.35527480e-01 3.67285222e-01 6.70808017e-01
-9.18523729e-01 -1.81330502e-01 -6.56023741e-01 -1.51211828e-01
6.29589915e-01 6.36377990e-01 -8.82536829e-01 -1.48458019e-01
2.58077770e-01 3.77739936e-01 2.94796407e-01 1.17346239e+00
-3.26404661e-01 1.09702015e+00 -7.33464479e-01 -2.28130817e-01
-1.46426330e-03 -4.86257017e-01 8.60628188e-01 7.92137861e-01
2.70588189e-01 -3.27515841e-01 4.43903863e-01 1.43980694e+00
3.79528672e-01 -8.58599186e-01 -4.24608827e-01 -1.25366762e-01
1.07448733e+00 5.76185465e-01 -3.78320396e-01 -6.16359055e-01
-7.12972134e-02 3.12163323e-01 2.18164429e-01 7.19568491e-01
-1.19179583e+00 -3.15880805e-01 1.02457774e+00 -1.13630503e-01
6.74354017e-01 -4.22537506e-01 -1.01551868e-01 -1.08339429e+00
-6.14719093e-01 -6.51767731e-01 3.73334706e-01 -3.87355149e-01
-1.41395533e+00 -2.96548545e-01 8.47992539e-01 -9.64809716e-01
-8.82067502e-01 -3.82404029e-01 -7.48233914e-01 9.92788851e-01
-9.01779473e-01 -6.82209313e-01 7.47316718e-01 1.77942306e-01
4.09064054e-01 -2.86350474e-02 5.64919293e-01 -4.24370199e-01
-4.38777536e-01 2.35950172e-01 7.52758145e-01 -1.97764829e-01
3.58165562e-01 -1.43757856e+00 5.32555461e-01 7.47037351e-01
-1.53602645e-01 7.80527651e-01 1.49779868e+00 -9.50338900e-01
-1.09506810e+00 -9.44722116e-01 3.85094941e-01 -7.01088428e-01
1.10499561e+00 -2.43375838e-01 -6.82619154e-01 1.10323453e+00
-1.68113440e-01 -2.21485317e-01 1.69038564e-01 1.05813771e-01
-2.00221092e-01 1.61222160e-01 -1.14493918e+00 8.11095595e-01
5.36749542e-01 -5.98876774e-01 -8.10588896e-01 1.62616432e-01
4.37937319e-01 -4.14047062e-01 -9.31703925e-01 -2.52708867e-02
6.77250504e-01 -5.72525859e-01 9.67114329e-01 -8.34603429e-01
-1.89587250e-01 -5.76918602e-01 -2.94578850e-01 -1.63194716e+00
-2.14799762e-01 -1.13582540e+00 -7.22433269e-01 9.31340635e-01
3.37417901e-01 -8.92509818e-01 7.55501688e-01 4.98889208e-01
1.67198673e-01 -3.64109963e-01 -1.22828305e+00 -1.04269421e+00
4.04146314e-01 -4.20433640e-01 5.84988534e-01 2.72413403e-01
-9.26460102e-02 2.05115065e-01 -5.48822820e-01 3.00603390e-01
1.42198634e+00 2.09331736e-01 7.25374341e-01 -1.19627810e+00
-8.32079947e-01 -2.85915732e-01 3.63082550e-02 -1.51477146e+00
2.20977440e-01 -4.83836263e-01 3.57914180e-01 -1.07629430e+00
3.55744064e-01 -1.88301504e-01 1.06215037e-01 -8.24215077e-03
-1.41439095e-01 -1.35502502e-01 -4.57728170e-02 6.88559115e-02
-4.45741892e-01 8.48252237e-01 1.23088658e+00 1.69838250e-01
-2.13632360e-02 4.61035550e-01 -3.17787051e-01 9.39807296e-01
7.50111818e-01 -5.44654131e-01 -4.73158807e-01 1.33604735e-01
-7.96258226e-02 4.40196723e-01 5.51997185e-01 -8.07656288e-01
4.57667485e-02 -8.19357455e-01 4.97985035e-01 -9.34434354e-01
5.51293731e-01 -3.64277214e-01 5.23786485e-01 6.82699978e-01
-4.66752410e-01 -4.81593907e-01 2.45908578e-03 1.15215361e+00
3.70728046e-01 -4.11747962e-01 1.25204194e+00 -3.84288318e-02
-1.72858074e-01 2.34559447e-01 -8.68042588e-01 1.46480411e-01
7.15867341e-01 1.09852098e-01 -2.87752241e-01 -7.69287586e-01
-1.20750463e+00 2.21882433e-01 3.86887103e-01 -1.74136415e-01
2.71346927e-01 -1.47628748e+00 -5.73473334e-01 1.23443291e-01
-3.53244275e-01 -2.48275101e-01 -1.35097384e-01 7.37324119e-01
-4.24481481e-02 4.43183154e-01 2.21124381e-01 -6.08893692e-01
-6.13632798e-01 2.37763599e-01 6.70993268e-01 -4.06699091e-01
-4.93400127e-01 6.07487917e-01 4.93591204e-02 -5.12978435e-01
1.11391917e-01 -3.70714217e-01 4.36555356e-01 -4.56423432e-01
6.73411012e-01 6.59571052e-01 -8.32388937e-01 -3.13206837e-02
4.47876640e-02 1.89221904e-01 -8.28284100e-02 -7.19741881e-01
1.00339270e+00 -2.74608344e-01 3.03030759e-01 7.08670020e-01
8.11494589e-01 -3.48310679e-01 -2.11014462e+00 -3.28754485e-01
2.00953707e-01 -7.75561094e-01 1.17605537e-01 -6.65391684e-01
-3.06964397e-01 6.09809697e-01 4.30037141e-01 2.60457367e-01
3.09096664e-01 1.83982641e-01 5.56453705e-01 2.24153310e-01
5.29362381e-01 -1.29956090e+00 -9.62042585e-02 5.56065202e-01
3.93210113e-01 -5.74478447e-01 1.38462484e-01 -3.99541259e-02
-3.21580052e-01 7.00044930e-01 2.92430490e-01 -4.36577946e-01
8.83953989e-01 4.83009934e-01 -4.51608986e-01 1.40119970e-01
-1.08038425e+00 -7.76874050e-02 1.94956407e-01 9.31328297e-01
-1.32878229e-01 3.26342374e-01 -1.38586581e-01 3.91679615e-01
-1.59073636e-01 9.67052579e-02 3.97903174e-01 6.83224559e-01
-1.02735567e+00 -1.14360666e+00 -7.37715065e-01 7.93055415e-01
-3.40210795e-01 2.51731575e-01 3.61194193e-01 4.61256236e-01
-5.22588968e-01 9.21374381e-01 2.60156661e-01 1.04207136e-01
-7.62161091e-02 2.55347967e-01 8.24167371e-01 -4.24881667e-01
4.36917782e-01 6.13956153e-01 2.23232925e-01 -4.48367149e-01
3.61221209e-02 -1.07943344e+00 -9.58121598e-01 -6.64567113e-01
-6.37075603e-01 3.45921338e-01 3.12268734e-01 1.21076536e+00
-6.90729693e-02 2.09640458e-01 5.03442228e-01 -1.01794231e+00
-1.83396208e+00 -1.15268004e+00 -8.80796552e-01 -5.09147942e-02
4.81436044e-01 -9.99256432e-01 -6.17410719e-01 -6.28118291e-02] | [6.737189292907715, 3.8612890243530273] |
4082be2e-c2ea-4b54-9849-4367be7200ba | sergiojimenez-at-semeval-2016-task-1 | null | null | https://aclanthology.org/S16-1116 | https://aclanthology.org/S16-1116.pdf | SERGIOJIMENEZ at SemEval-2016 Task 1: Effectively Combining Paraphrase Database, String Matching, WordNet, and Word Embedding for Semantic Textual Similarity | null | ['Sergio Jimenez'] | 2016-06-01 | null | null | null | semeval-2016-6 | ['negation-detection'] | ['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] | [-7.449975967407227, 3.6470084190368652] |
4f5644de-92c3-46b5-a151-072e26b90b4a | graphglow-universal-and-generalizable | 2306.11264 | null | https://arxiv.org/abs/2306.11264v1 | https://arxiv.org/pdf/2306.11264v1.pdf | GraphGLOW: Universal and Generalizable Structure Learning for Graph Neural Networks | Graph structure learning is a well-established problem that aims at optimizing graph structures adaptive to specific graph datasets to help message passing neural networks (i.e., GNNs) to yield effective and robust node embeddings. However, the common limitation of existing models lies in the underlying \textit{closed-world assumption}: the testing graph is the same as the training graph. This premise requires independently training the structure learning model from scratch for each graph dataset, which leads to prohibitive computation costs and potential risks for serious over-fitting. To mitigate these issues, this paper explores a new direction that moves forward to learn a universal structure learning model that can generalize across graph datasets in an open world. We first introduce the mathematical definition of this novel problem setting, and describe the model formulation from a probabilistic data-generative aspect. Then we devise a general framework that coordinates a single graph-shared structure learner and multiple graph-specific GNNs to capture the generalizable patterns of optimal message-passing topology across datasets. The well-trained structure learner can directly produce adaptive structures for unseen target graphs without any fine-tuning. Across diverse datasets and various challenging cross-graph generalization protocols, our experiments show that even without training on target graphs, the proposed model i) significantly outperforms expressive GNNs trained on input (non-optimized) topology, and ii) surprisingly performs on par with state-of-the-art models that independently optimize adaptive structures for specific target graphs, with notably orders-of-magnitude acceleration for training on the target graph. | ['Junchi Yan', 'Chenxiao Yang', 'Qitian Wu', 'Wentao Zhao'] | 2023-06-20 | null | null | null | null | ['graph-structure-learning'] | ['graphs'] | [ 2.07390517e-01 4.72080827e-01 -3.44980896e-01 -4.11586612e-01
-5.76604247e-01 -7.38797843e-01 5.10304272e-01 3.91093493e-01
-1.06552176e-01 5.55992007e-01 -1.02845922e-01 -4.62727547e-01
-4.72634971e-01 -1.22669923e+00 -1.08382380e+00 -6.69598043e-01
-5.00593662e-01 9.55928624e-01 1.57086700e-01 -3.25655341e-01
7.39080459e-02 6.75451815e-01 -1.04162097e+00 -2.99765766e-01
8.17959487e-01 5.74284494e-01 1.14672936e-01 1.06288600e+00
-1.85013413e-01 4.46196914e-01 -4.07842070e-01 -5.37227035e-01
3.15235108e-01 -1.88001186e-01 -9.05907750e-01 7.85688162e-02
6.70334578e-01 1.70740053e-01 -9.05839145e-01 1.11935735e+00
3.78452897e-01 -5.27328253e-03 5.52647829e-01 -1.41949415e+00
-1.12612116e+00 9.02708113e-01 -2.22182050e-01 2.86548853e-01
7.94798061e-02 1.88965812e-01 1.34196532e+00 -2.88128048e-01
6.38411641e-01 1.19143426e+00 8.58997822e-01 6.63585007e-01
-1.64588177e+00 -3.70630771e-01 3.33176702e-01 -1.34024501e-01
-1.64660680e+00 -9.27339792e-02 6.86151266e-01 -2.83364626e-03
8.36103320e-01 2.45178670e-01 5.56902111e-01 1.20766604e+00
4.17391062e-01 5.64778149e-01 5.72685301e-01 -5.38807586e-02
1.59606546e-01 -2.22847402e-01 1.31811708e-01 9.76779759e-01
6.54776692e-01 1.73297018e-01 -1.78689644e-01 -2.04861462e-01
6.50381625e-01 -7.41935298e-02 -2.41988108e-01 -8.62980604e-01
-8.73574436e-01 8.88570428e-01 9.76689100e-01 3.46697927e-01
4.94503416e-02 6.11093283e-01 3.04970324e-01 5.56938708e-01
2.96719313e-01 4.85745251e-01 -5.79455137e-01 1.76327348e-01
-4.75055963e-01 2.02344030e-01 1.11852539e+00 1.22610223e+00
1.10160935e+00 1.80131003e-01 2.95603901e-01 4.29839194e-01
4.82851774e-01 5.07764041e-01 6.36451170e-02 -1.20830692e-01
6.28200531e-01 8.64520609e-01 -7.31356680e-01 -1.20168924e+00
-5.61097980e-01 -8.71894181e-01 -1.01734245e+00 -3.07329655e-01
4.81379062e-01 -4.63319868e-02 -1.18926275e+00 2.07390404e+00
3.26842517e-01 5.91419220e-01 -1.03792911e-02 3.43745053e-01
7.08144188e-01 5.80968320e-01 1.38207123e-01 4.01981950e-01
8.90142977e-01 -8.27358603e-01 -7.14755431e-02 -4.28171724e-01
1.05653346e+00 -3.37596655e-01 1.22920465e+00 -5.83993159e-02
-7.18954206e-01 -2.52515346e-01 -1.17201984e+00 -4.43433411e-02
-5.80295026e-01 -5.97003222e-01 8.30014706e-01 8.73399377e-01
-1.54290342e+00 7.14194417e-01 -8.56964588e-01 -8.00370455e-01
5.03923595e-01 5.53158700e-01 -5.61094582e-01 -4.00370926e-01
-8.94010007e-01 6.69038236e-01 6.81655347e-01 -2.26795182e-01
-1.24852252e+00 -8.56623709e-01 -1.19370365e+00 2.72768557e-01
4.20828074e-01 -1.08492565e+00 7.01837659e-01 -8.03682208e-01
-1.27376878e+00 6.62007153e-01 2.71151423e-01 -6.70886099e-01
8.49646144e-03 1.69350654e-01 -3.70901287e-01 1.25465915e-01
-1.44857690e-01 4.36808974e-01 8.17087352e-01 -1.32171214e+00
-5.14955930e-02 -3.72141749e-01 2.91242361e-01 2.00926155e-01
-4.00405169e-01 -4.35682595e-01 -4.69683349e-01 -6.76505923e-01
1.20589010e-01 -9.21605468e-01 -4.49080110e-01 -1.51282623e-01
-5.86288869e-01 -2.55822688e-01 6.89239204e-01 -1.36536464e-01
1.22832060e+00 -1.82616282e+00 3.02752256e-01 6.41105950e-01
5.91172993e-01 1.39538899e-01 -6.07468665e-01 6.90215468e-01
-1.92210525e-01 3.35522503e-01 -3.27103347e-01 -2.07972392e-01
5.83083108e-02 7.04609692e-01 -2.21718892e-01 6.77589715e-01
2.19403729e-01 1.30081582e+00 -1.06008387e+00 -3.03855330e-01
1.87552303e-01 3.93163055e-01 -7.41659582e-01 2.74291962e-01
-3.96778256e-01 1.28674135e-01 -4.91995037e-01 3.93952668e-01
8.15963864e-01 -8.64860237e-01 6.34642005e-01 -3.04678734e-02
6.61997974e-01 2.38884389e-01 -1.23741090e+00 1.72529542e+00
-4.08902436e-01 1.21635459e-01 1.64333116e-02 -1.42933726e+00
8.65689695e-01 -7.04880655e-02 2.07885742e-01 -3.82934272e-01
8.03661048e-02 1.64083973e-01 1.26422867e-01 -1.89129651e-01
1.48077324e-01 -3.83596607e-02 -3.02184612e-01 5.33239245e-01
6.35322273e-01 -1.48217633e-01 9.86094549e-02 6.12660170e-01
1.48362565e+00 -2.84847677e-01 1.97249800e-01 -7.11168349e-01
4.19081599e-01 -3.57002676e-01 2.19689831e-01 9.98696983e-01
4.37370241e-02 5.44026196e-01 5.69616854e-01 -6.98249936e-01
-9.16582227e-01 -1.45251012e+00 1.34428024e-01 1.34308624e+00
4.08768386e-01 -6.73622549e-01 -6.62693501e-01 -1.03234339e+00
1.58822745e-01 4.41517919e-01 -8.30108821e-01 -4.37471539e-01
-6.91895902e-01 -8.95155907e-01 7.29193211e-01 4.33538944e-01
1.16126828e-01 -6.26517117e-01 4.25916970e-01 5.22589125e-02
3.30250084e-01 -1.04677045e+00 -6.50441885e-01 1.99710261e-02
-9.69439626e-01 -1.37536931e+00 -1.93169788e-01 -1.03023124e+00
1.05143464e+00 2.01415449e-01 1.46432269e+00 7.20825672e-01
-1.36532903e-01 6.95578337e-01 -6.40065372e-02 2.83481218e-02
-6.23294234e-01 7.35141814e-01 -7.05345497e-02 -1.85566931e-03
-6.20356053e-02 -9.79964554e-01 -3.63211036e-01 1.57003760e-01
-1.10518289e+00 -1.79386869e-01 5.97659409e-01 7.18105376e-01
4.50467169e-01 3.38211618e-02 5.28141797e-01 -1.23062611e+00
6.96265876e-01 -7.39001513e-01 -6.65772736e-01 5.81990063e-01
-7.56654501e-01 5.77836037e-01 1.03852749e+00 -3.42832118e-01
-4.11800712e-01 -2.71773338e-01 -1.35034561e-01 -2.10780323e-01
5.72550744e-02 6.11707389e-01 -2.57486314e-01 -6.49341822e-01
8.71819615e-01 4.45029408e-01 1.71823010e-01 -1.72322467e-01
6.50608540e-01 -4.75796759e-02 3.11470240e-01 -9.52309549e-01
1.40930510e+00 3.59180540e-01 4.05385852e-01 -7.70482421e-01
-8.13220978e-01 -2.07291871e-01 -4.93183970e-01 1.65885970e-01
4.48652744e-01 -6.46295607e-01 -5.73798060e-01 3.31759781e-01
-8.12834918e-01 -5.77874541e-01 -2.53320277e-01 1.47799179e-01
-5.56221843e-01 5.44651747e-01 -7.21607089e-01 -1.43875882e-01
-4.15665120e-01 -9.84615386e-01 8.37685823e-01 1.68747380e-01
1.66179925e-01 -1.81580889e+00 2.03658238e-01 -7.48517513e-02
6.15583122e-01 1.92549661e-01 1.17868733e+00 -1.10469782e+00
-1.08300531e+00 -2.56310463e-01 -2.90057957e-01 2.08890080e-01
5.41163608e-02 -1.70346141e-01 -5.61917484e-01 -6.97237492e-01
-3.57927769e-01 -3.43317419e-01 6.47348106e-01 1.52211174e-01
1.21936560e+00 -4.35785413e-01 -5.39962888e-01 1.10051394e+00
1.64498246e+00 -6.69586062e-01 4.96723026e-01 -9.06182826e-02
9.96950388e-01 8.27375948e-02 -1.34580851e-01 -1.11618765e-01
7.90560901e-01 2.84309298e-01 6.24481857e-01 -9.71799064e-03
-2.19854668e-01 -8.24589849e-01 1.78946421e-01 8.79981339e-01
3.02375734e-01 -6.93286955e-01 -1.00757051e+00 6.31826401e-01
-1.58716416e+00 -5.78912675e-01 1.23783648e-01 2.10783434e+00
4.92065877e-01 1.20654002e-01 1.89311672e-02 -2.56563187e-01
7.90864170e-01 4.00979280e-01 -5.23206711e-01 -2.36801803e-01
-1.49789527e-01 6.57434762e-01 7.67447472e-01 7.11798787e-01
-9.23606753e-01 1.02229011e+00 6.62532425e+00 8.82687211e-01
-1.06481767e+00 -4.61631678e-02 5.69266379e-01 3.43164444e-01
-6.87889397e-01 2.93135613e-01 -6.63621783e-01 2.07447156e-01
1.26353967e+00 -5.37766337e-01 9.31266665e-01 7.87147641e-01
-4.77900058e-01 5.77501178e-01 -1.27186763e+00 7.81274915e-01
1.77590027e-01 -1.61759579e+00 4.91048485e-01 2.48425081e-01
7.55084932e-01 5.19956410e-01 6.42197430e-02 5.43472707e-01
8.44368219e-01 -1.39596438e+00 1.55777529e-01 1.99513003e-01
5.90074420e-01 -5.01047373e-01 4.46856350e-01 3.27385396e-01
-1.54725516e+00 5.82822561e-02 -5.41173875e-01 2.24829480e-01
5.97043596e-02 3.94748747e-01 -9.93915498e-01 9.54819262e-01
3.46298873e-01 7.47806013e-01 -7.70330966e-01 9.59937990e-01
-3.27010065e-01 6.64175153e-01 -6.92420721e-01 -1.14492156e-01
3.84016752e-01 -1.83102101e-01 6.18895292e-01 1.00390160e+00
2.89346814e-01 -1.98880926e-01 4.26347136e-01 9.19305623e-01
-5.51702321e-01 6.28470033e-02 -9.43236053e-01 -3.01931620e-01
4.79960740e-01 1.27380586e+00 -7.90250897e-01 -1.64413527e-01
-2.69476384e-01 6.78577781e-01 9.82777297e-01 5.25329709e-01
-8.39364469e-01 -3.17307055e-01 5.97324967e-01 1.81760177e-01
4.14026439e-01 -4.00108010e-01 -1.86606735e-01 -1.18311965e+00
-9.21974182e-02 -7.54301846e-01 8.67481470e-01 -1.56215459e-01
-1.61985862e+00 5.70434153e-01 -8.52662325e-02 -6.91616058e-01
6.23550154e-02 -7.26758778e-01 -8.07345748e-01 5.79526901e-01
-1.47698224e+00 -1.45061731e+00 -1.70152605e-01 7.67559111e-01
-8.35123733e-02 2.45471001e-02 8.41471612e-01 2.72134990e-01
-4.66244966e-01 1.02663064e+00 4.32174169e-02 3.10683042e-01
3.24530602e-01 -1.45056570e+00 8.32226336e-01 9.42908585e-01
4.36425984e-01 9.13139403e-01 5.84546447e-01 -5.57226121e-01
-1.86287653e+00 -1.40685248e+00 6.64773643e-01 -6.22454286e-01
1.06654358e+00 -7.86924541e-01 -1.01314950e+00 1.07764947e+00
-9.01059210e-02 6.65766537e-01 5.16954780e-01 4.69520152e-01
-5.92474937e-01 -1.71048641e-01 -1.22734141e+00 6.65134788e-01
1.56387651e+00 -5.65114796e-01 -3.62521976e-01 5.20946562e-01
9.01397288e-01 -5.36403358e-01 -1.07155263e+00 3.24027419e-01
2.68435255e-02 -5.93390882e-01 1.04793298e+00 -1.07915735e+00
-1.79411292e-01 -3.12450081e-01 -1.56530529e-01 -1.41279995e+00
-3.06988031e-01 -9.65695083e-01 -3.01640809e-01 8.89151394e-01
6.09364808e-01 -1.07008028e+00 9.37929809e-01 3.50604028e-01
-2.07005352e-01 -9.18854952e-01 -9.02750969e-01 -9.02314901e-01
3.32572430e-01 -4.39382374e-01 8.51605654e-01 8.49377811e-01
-1.73712417e-01 5.33600450e-01 -2.18898011e-03 7.57156610e-01
8.05995345e-01 1.46453932e-01 1.14568663e+00 -1.23628068e+00
-3.55121702e-01 -3.98933768e-01 -9.47951257e-01 -1.38140070e+00
4.07976955e-01 -1.63132000e+00 -2.79319704e-01 -1.53237867e+00
1.90294936e-01 -7.82734156e-01 -3.20712090e-01 1.96284503e-01
-2.54400700e-01 -4.44287732e-02 1.38991624e-02 -3.31633151e-01
-8.80257964e-01 6.88015580e-01 1.26401055e+00 -1.48075834e-01
3.41763422e-02 8.99203420e-02 -1.09799242e+00 2.42234841e-01
7.85208941e-01 -5.71354091e-01 -8.86866152e-01 -5.82467318e-01
4.18657005e-01 -3.70710909e-01 5.37073791e-01 -9.74016666e-01
3.48866910e-01 -1.23279192e-01 -3.71752754e-02 -1.16318092e-01
-1.34079814e-01 -6.41968369e-01 1.85840771e-01 3.45752448e-01
-1.50964320e-01 2.28139892e-01 2.20653325e-01 1.10682237e+00
2.62105852e-01 -1.52371367e-02 7.61671424e-01 3.95958051e-02
-4.28294480e-01 1.03420389e+00 2.21885875e-01 6.89879954e-01
9.40096259e-01 -2.47000292e-01 -5.51168740e-01 -4.48093712e-01
-7.30692565e-01 3.35554719e-01 6.88164830e-01 3.52639973e-01
5.19362092e-01 -1.19285309e+00 -6.48702681e-01 3.41408223e-01
2.42652178e-01 2.34717786e-01 2.51390129e-01 5.79650402e-01
-6.13022268e-01 1.15575664e-01 2.43431896e-01 -7.38058805e-01
-6.70912445e-01 7.46954203e-01 5.92604756e-01 -6.71849668e-01
-7.56122947e-01 1.02214730e+00 3.69062603e-01 -1.11760044e+00
-2.07993332e-02 -4.03627604e-01 4.01318640e-01 -8.14103842e-01
1.89341396e-01 1.59693975e-02 2.91455965e-02 -5.21552742e-01
-1.75882831e-01 4.97545451e-01 -1.85034469e-01 5.09352624e-01
1.42379570e+00 -3.53528336e-02 -1.64703056e-01 -4.04466002e-04
1.36256492e+00 -1.91502422e-01 -1.11675858e+00 -4.43532169e-01
1.62594125e-01 -2.46243462e-01 -3.45313400e-01 -3.18857133e-01
-1.19202626e+00 6.60143793e-01 1.44054413e-01 4.41435993e-01
6.73011124e-01 4.38410431e-01 7.31760323e-01 5.86672306e-01
5.49624324e-01 -6.59640491e-01 1.71442688e-01 6.77514374e-01
6.42931223e-01 -9.84396696e-01 -2.84037814e-02 -5.53256571e-01
-1.23701349e-01 1.10782254e+00 6.28372014e-01 -5.84065616e-01
9.83537555e-01 7.78050050e-02 -3.04750115e-01 -5.81135452e-01
-7.03748345e-01 -3.89313698e-02 3.58158916e-01 8.84675145e-01
1.11955106e-01 1.98675498e-01 2.88904250e-01 2.95846105e-01
-3.01274598e-01 -4.57898408e-01 3.27014416e-01 5.54283679e-01
-3.12661141e-01 -1.30577815e+00 1.94039598e-01 6.68827355e-01
-1.54869184e-01 -1.57614723e-01 -2.21911639e-01 1.13379133e+00
-3.92037421e-01 5.72583139e-01 -1.13831937e-01 -4.08013046e-01
1.37059450e-01 -1.50110111e-01 6.67065561e-01 -8.86276007e-01
-3.96098942e-01 -7.54781127e-01 -2.92457361e-02 -6.54241323e-01
6.64508864e-02 -7.83299878e-02 -1.18947053e+00 -6.88449502e-01
-3.53729069e-01 1.03554040e-01 2.28927121e-01 8.18361461e-01
7.59502351e-01 3.05767536e-01 4.55460727e-01 -6.10636890e-01
-8.06787431e-01 -6.97119355e-01 -5.80737174e-01 5.90797067e-01
2.62787908e-01 -5.08538365e-01 -5.88885069e-01 -4.56935138e-01] | [6.969025135040283, 6.207122802734375] |
154c2a7f-8dbe-430e-9411-0bb0d3938dbb | completely-unsupervised-phoneme-recognition-1 | 1904.04100 | null | https://arxiv.org/abs/1904.04100v3 | https://arxiv.org/pdf/1904.04100v3.pdf | Completely Unsupervised Speech Recognition By A Generative Adversarial Network Harmonized With Iteratively Refined Hidden Markov Models | Producing a large annotated speech corpus for training ASR systems remains difficult for more than 95% of languages all over the world which are low-resourced, but collecting a relatively big unlabeled data set for such languages is more achievable. This is why some initial effort have been reported on completely unsupervised speech recognition learned from unlabeled data only, although with relatively high error rates. In this paper, we develop a Generative Adversarial Network (GAN) to achieve this purpose, in which a Generator and a Discriminator learn from each other iteratively to improve the performance. We further use a set of Hidden Markov Models (HMMs) iteratively refined from the machine generated labels to work in harmony with the GAN. The initial experiments on TIMIT data set achieve an phone error rate of 33.1%, which is 8.5% lower than the previous state-of-the-art. | ['Lin-shan Lee', 'Hung-Yi Lee', 'Da-Rong Liu', 'Che-Ping Tsai', 'Kuan-Yu Chen'] | 2019-04-08 | null | null | null | null | ['unsupervised-speech-recognition'] | ['speech'] | [ 3.59475464e-01 5.11835575e-01 5.99032082e-02 -4.71394002e-01
-1.15823710e+00 -5.18058717e-01 6.52059972e-01 -5.57196558e-01
-2.57717401e-01 9.59643483e-01 2.36935258e-01 -4.18918401e-01
8.22995365e-01 -5.73290229e-01 -4.10535246e-01 -8.57743740e-01
3.35335046e-01 7.68116832e-01 6.75141159e-03 -2.15331167e-01
-2.44090557e-01 2.62062788e-01 -1.15522265e+00 1.66632131e-01
1.02305615e+00 5.24619460e-01 3.19212615e-01 7.77375877e-01
-2.44288728e-01 9.54569936e-01 -1.04975450e+00 -5.02641082e-01
1.59898177e-01 -8.67846966e-01 -8.22837591e-01 1.86776251e-01
-1.15035795e-01 -2.09411412e-01 -4.61740464e-01 1.02971017e+00
6.39305711e-01 2.53910154e-01 5.12231648e-01 -1.11731982e+00
-8.85792673e-01 8.42373431e-01 -1.57150060e-01 -1.33799374e-01
7.68682361e-02 1.38657555e-01 6.81950390e-01 -6.80662334e-01
3.30036432e-01 1.08891511e+00 1.84236765e-01 1.24750841e+00
-8.34536433e-01 -8.12544227e-01 -5.68043105e-02 -2.78832525e-01
-1.60339117e+00 -8.04159880e-01 8.05011511e-01 -2.01217979e-01
1.06263149e+00 3.15574825e-01 3.11359286e-01 1.44607747e+00
-4.17868197e-01 7.43050396e-01 1.17501760e+00 -6.11813068e-01
2.34935582e-01 2.07075104e-01 -3.98371935e-01 4.83107269e-01
-2.20593423e-01 1.19659863e-01 -3.63493681e-01 1.66725904e-01
9.04596388e-01 -9.71109197e-02 -1.38153628e-01 2.75345057e-01
-9.96211171e-01 9.48990047e-01 1.60539165e-01 4.74128217e-01
-2.77729243e-01 -1.23345107e-01 1.14038587e-01 3.16309929e-01
5.95425308e-01 3.03849220e-01 -2.88597137e-01 -5.39044857e-01
-9.97448325e-01 -2.32599884e-01 7.14325130e-01 1.12301123e+00
5.43611765e-01 8.56357515e-01 1.02972157e-01 1.27681661e+00
4.68701124e-01 6.75100029e-01 7.21339047e-01 -6.88274741e-01
6.57682419e-01 3.84787649e-01 -2.04936549e-01 -2.50224471e-01
2.30775088e-01 -3.86820495e-01 -1.18565762e+00 3.33164334e-02
2.67590314e-01 -3.97016466e-01 -1.57893598e+00 1.68835521e+00
8.65993183e-03 2.45749071e-01 4.82813627e-01 7.72779584e-01
6.94928944e-01 1.06914902e+00 1.27869517e-01 -2.68664807e-01
9.30162907e-01 -1.25723398e+00 -1.00865281e+00 -4.50256258e-01
5.02624452e-01 -8.98495376e-01 1.12367237e+00 3.08245212e-01
-1.05402589e+00 -6.21344388e-01 -1.06324100e+00 1.80919707e-01
-3.55649233e-01 1.79580167e-01 4.58693266e-01 1.09745002e+00
-1.22307217e+00 2.36163996e-02 -1.07809246e+00 -2.46238664e-01
3.06747347e-01 3.48294199e-01 -2.89509922e-01 2.09961176e-01
-1.22521710e+00 8.31937909e-01 4.79763597e-01 7.04639331e-02
-1.17238843e+00 -1.71608210e-01 -8.63101065e-01 -1.61949009e-01
2.12914526e-01 -2.22324952e-01 1.37446666e+00 -1.17414844e+00
-2.23344398e+00 8.30605686e-01 -2.01390773e-01 -5.22393703e-01
3.64341378e-01 -7.55164996e-02 -8.62628400e-01 -2.76727617e-01
-3.31890494e-01 7.59049177e-01 7.68938363e-01 -1.31601846e+00
-4.85594988e-01 -4.27397229e-02 -1.34149224e-01 1.89531043e-01
-2.92445213e-01 2.91078538e-01 -5.61318576e-01 -9.86575067e-01
-7.26596043e-02 -1.17932749e+00 -3.16904962e-01 -8.05879056e-01
-4.16845500e-01 -2.81070828e-01 9.22266603e-01 -1.03807271e+00
1.20679951e+00 -2.08863854e+00 5.63401431e-02 -8.14085007e-02
-2.23023534e-01 7.59876013e-01 -6.12091310e-02 4.23995852e-01
2.49998271e-02 3.16549897e-01 -5.47845125e-01 -6.52656019e-01
-1.53840691e-01 5.00037968e-01 -5.30213714e-01 -3.08460407e-02
3.82718116e-01 8.07297468e-01 -8.12454164e-01 -1.07636653e-01
1.76346838e-01 5.94617844e-01 -3.82342696e-01 7.10718274e-01
-2.38590196e-01 8.78461659e-01 -3.48151594e-01 6.72284484e-01
3.44790339e-01 3.03048305e-02 1.11831941e-01 4.18442070e-01
1.63941458e-01 6.88866973e-01 -8.92005086e-01 1.85885727e+00
-5.74969411e-01 4.37217921e-01 -6.89470172e-02 -1.00628245e+00
1.39162397e+00 8.46778631e-01 5.97983226e-02 -3.88124138e-01
-5.13704531e-02 3.37873846e-01 1.77963018e-01 -1.12844154e-01
4.78699893e-01 -4.21527892e-01 -1.49280414e-01 3.01774234e-01
3.29564452e-01 -3.57627451e-01 -1.35362118e-01 6.18994683e-02
1.06397998e+00 -6.63252641e-03 1.12126671e-01 1.43716082e-01
5.24125874e-01 -2.26593077e-01 7.27422237e-01 5.08146465e-01
-9.30528194e-02 9.72777784e-01 3.94866616e-02 -1.71007246e-01
-1.44724500e+00 -1.03600729e+00 1.41915008e-01 6.08730435e-01
-3.97346914e-01 -2.56538421e-01 -9.58496153e-01 -8.32886755e-01
-7.32972324e-01 1.00237417e+00 -6.66034669e-02 -5.62095493e-02
-6.55179203e-01 -5.44374108e-01 9.29731548e-01 6.15609407e-01
6.67392015e-01 -1.38282347e+00 2.33036190e-01 3.00242186e-01
-1.60624966e-01 -1.30889750e+00 -4.38917816e-01 2.04308584e-01
-5.67891121e-01 -3.19962919e-01 -8.53524804e-01 -9.42163348e-01
7.19115853e-01 -2.77143389e-01 1.04603410e+00 -1.26753211e-01
2.56058425e-01 -1.09738037e-01 -6.50322616e-01 -4.83520329e-01
-1.26163864e+00 3.93533677e-01 1.33908063e-01 6.51139766e-02
2.89128095e-01 -4.43497986e-01 -1.25865052e-02 4.29270446e-01
-9.00614381e-01 1.79634005e-01 6.12901747e-01 9.58597064e-01
4.81561989e-01 1.04735211e-01 9.70328331e-01 -9.46282029e-01
3.40168536e-01 -3.16216081e-01 -4.96209353e-01 2.67592430e-01
-3.74296725e-01 6.84313849e-02 8.76690269e-01 -5.40043294e-01
-1.44535017e+00 1.72370806e-01 -8.10059130e-01 -4.66132641e-01
-4.83353943e-01 2.75900573e-01 -6.08913720e-01 3.39413643e-01
3.75490695e-01 3.04978192e-01 -1.20544404e-01 -4.44341928e-01
3.27707380e-01 1.37423360e+00 5.21858811e-01 -4.03979391e-01
8.72232139e-01 -2.91528013e-02 -6.19686127e-01 -9.62866366e-01
-7.84517348e-01 -2.23695964e-01 -5.26189506e-01 4.31590527e-02
8.64976168e-01 -1.05739152e+00 6.21558353e-03 7.17963934e-01
-1.07998276e+00 -6.81045115e-01 -3.06718528e-01 6.92106843e-01
-4.46795434e-01 2.25950684e-02 -7.04647005e-01 -9.82043684e-01
-4.02361333e-01 -1.28052664e+00 9.42188501e-01 2.54545927e-01
-1.67103544e-01 -9.68590200e-01 2.76411027e-01 7.29495704e-01
5.78440368e-01 -1.18412562e-01 5.45969784e-01 -1.00952125e+00
-5.25590777e-01 -1.21757306e-01 2.85050720e-01 9.40104365e-01
4.84514892e-01 -4.62918617e-02 -1.20011854e+00 -4.47183430e-01
1.53703004e-01 -3.92825902e-01 4.61475015e-01 -1.05273046e-01
1.04922366e+00 -3.79576832e-01 -1.53209344e-01 2.82408535e-01
9.98405397e-01 7.94792771e-01 1.03800952e+00 -1.31422445e-01
8.81772339e-01 2.35595360e-01 3.67772549e-01 4.30139527e-02
7.09474133e-03 7.04929233e-01 6.02115132e-02 -1.59425721e-01
-3.62834811e-01 -6.34634495e-01 7.04503179e-01 1.75234950e+00
-1.94642860e-02 -6.61155343e-01 -1.03544390e+00 6.25208914e-01
-1.38020122e+00 -8.45340014e-01 3.81344050e-01 2.23906279e+00
1.22532940e+00 1.23659946e-01 2.62857657e-02 5.72857484e-02
9.07316923e-01 1.84165552e-01 -3.26157212e-01 -3.57968569e-01
-1.35136887e-01 5.18617570e-01 2.22599030e-01 6.79861069e-01
-9.33081269e-01 1.36515307e+00 6.43702555e+00 1.07756054e+00
-1.35051525e+00 2.07442313e-01 8.82335186e-01 2.63923913e-01
-4.05105501e-01 1.07385121e-01 -8.88411462e-01 5.79628766e-01
1.49874651e+00 1.88882932e-01 6.64690435e-01 8.59938562e-01
1.50569513e-01 4.65601712e-01 -6.93787038e-01 9.15388465e-01
2.23240107e-01 -9.17622924e-01 1.28684446e-01 1.39385432e-01
1.05672896e+00 1.60619751e-01 -5.80102354e-02 5.13164043e-01
9.65875089e-01 -1.35787582e+00 5.02339840e-01 1.52110115e-01
1.09183276e+00 -9.36127305e-01 7.52174973e-01 6.17307663e-01
-8.98503661e-01 4.41234976e-01 -3.70428890e-01 1.57254532e-01
5.17696917e-01 5.27404428e-01 -1.21995270e+00 4.80526537e-01
3.10227573e-01 2.50814617e-01 -2.31797084e-01 5.99140465e-01
-6.39991879e-01 1.23957849e+00 -3.07976812e-01 -3.62028852e-02
2.48946965e-01 -2.44561225e-01 4.21747178e-01 1.01168990e+00
5.47968984e-01 6.51465654e-02 1.48562402e-01 6.28948331e-01
-4.40753549e-01 2.99632680e-02 -6.95653617e-01 -4.88113970e-01
6.41265094e-01 1.15098500e+00 -5.47461629e-01 -4.20136988e-01
-4.98172730e-01 1.20571601e+00 4.27910268e-01 3.81090134e-01
-8.62884700e-01 -2.64814883e-01 4.52000558e-01 -1.31934911e-01
1.75658524e-01 -5.25391638e-01 1.70204956e-02 -1.37668955e+00
-6.39650449e-02 -1.30674469e+00 5.36530185e-03 -7.20650554e-01
-1.27043009e+00 1.15198624e+00 -4.03038681e-01 -1.08907843e+00
-7.20378757e-01 -3.55800420e-01 -5.45934677e-01 1.15099573e+00
-1.07393396e+00 -1.21473229e+00 -3.09379604e-02 4.55503404e-01
1.03734481e+00 -6.75920963e-01 1.31737852e+00 3.03732693e-01
-6.46586537e-01 7.83668160e-01 4.13791388e-02 5.27695715e-01
5.28486967e-01 -1.15632188e+00 9.15810585e-01 1.14091718e+00
6.46774828e-01 5.37507713e-01 4.26733136e-01 -6.99241877e-01
-1.17566407e+00 -1.11760771e+00 9.37748432e-01 -4.25822645e-01
5.57168841e-01 -6.89948499e-01 -1.02413273e+00 8.67009878e-01
5.73221505e-01 -7.66144246e-02 7.30023682e-01 -1.29370525e-01
-2.09153458e-01 1.16489477e-01 -1.04611158e+00 6.16351783e-01
9.50031638e-01 -7.39308715e-01 -5.63710868e-01 3.18160281e-02
9.51062262e-01 -4.29744244e-01 -8.86738181e-01 5.00804007e-01
8.74808431e-03 -5.13519347e-01 5.31745911e-01 -5.32126784e-01
7.53662959e-02 -4.78509754e-01 -2.28114218e-01 -1.77848065e+00
9.29846838e-02 -1.01269829e+00 3.47817987e-02 1.92873323e+00
6.66049838e-01 -6.85130179e-01 8.83859038e-01 4.36216205e-01
-4.82613474e-01 -4.37459618e-01 -9.27385807e-01 -9.21921313e-01
2.94652842e-02 -3.82499307e-01 6.70623362e-01 1.09538651e+00
-2.63701588e-01 6.81683302e-01 -5.34996569e-01 1.42377421e-01
2.87329733e-01 -3.77988309e-01 7.41436899e-01 -6.62676513e-01
-3.64656597e-01 -2.33720448e-02 -2.76068300e-01 -1.22105527e+00
3.88245821e-01 -8.65206659e-01 2.93905437e-01 -1.24168754e+00
-5.35208769e-02 -6.89858019e-01 -1.66778892e-01 6.56991243e-01
-1.45824000e-01 1.61449701e-01 1.96198057e-02 1.48316264e-01
-1.81771696e-01 8.26426983e-01 1.13541555e+00 -2.36441866e-01
-2.08795488e-01 6.77514523e-02 -4.89925504e-01 5.98438323e-01
1.11973631e+00 -5.07856131e-01 -6.11591101e-01 -5.11282086e-01
-5.01759887e-01 1.26000866e-01 -2.00887054e-01 -1.05120933e+00
5.08958362e-02 -1.55602932e-01 7.27545619e-02 -4.09300894e-01
4.88740087e-01 -5.90456903e-01 3.84107858e-01 1.04543172e-01
-3.56242567e-01 -1.53906569e-01 4.63536307e-02 2.01362938e-01
-6.48458302e-01 -2.89703816e-01 8.79409373e-01 -1.27365649e-01
-5.47717392e-01 2.68294036e-01 -3.22728395e-01 2.21328452e-01
7.73014307e-01 1.62583783e-01 -1.96889833e-01 -6.18346810e-01
-6.88534677e-01 -2.29365528e-01 4.04011667e-01 6.10441089e-01
4.90432084e-01 -1.48200595e+00 -9.39082384e-01 5.44718027e-01
-7.78351575e-02 1.20709069e-01 4.72775139e-02 1.31528288e-01
-3.31815600e-01 4.16964918e-01 5.19269668e-02 -3.47159266e-01
-1.02846181e+00 3.56652468e-01 1.74615577e-01 -4.09938633e-01
-3.24750841e-01 7.63898551e-01 2.45516282e-02 -7.23866224e-01
1.89901412e-01 8.91682357e-02 -4.60632071e-02 -4.81931716e-01
3.49485368e-01 5.98592497e-02 1.27279341e-01 -8.44954014e-01
-1.03457138e-01 4.38893475e-02 -2.36513000e-02 -4.63667780e-01
1.21246648e+00 2.79611126e-02 1.30739436e-01 4.00108844e-01
1.05433154e+00 2.15625674e-01 -1.24775040e+00 -1.46226853e-01
-2.01356143e-01 -2.48881593e-01 -1.12041958e-01 -8.03170204e-01
-1.15370619e+00 1.01003742e+00 3.89146179e-01 3.67225230e-01
9.38909650e-01 1.49463996e-01 9.81852829e-01 1.62145987e-01
4.72745359e-01 -1.00221515e+00 -1.45303253e-02 6.40425503e-01
7.98113465e-01 -1.07344663e+00 -6.09124362e-01 -3.90062302e-01
-9.51170325e-01 6.62234128e-01 6.79956496e-01 4.44300994e-02
4.14404839e-01 5.89078009e-01 3.69942874e-01 3.66945326e-01
-6.77200377e-01 -2.01152116e-01 9.44165234e-03 7.44941711e-01
6.30175412e-01 2.46835351e-01 2.27267016e-03 6.13245964e-01
-4.65198100e-01 -3.17505032e-01 3.15668404e-01 7.29937375e-01
-3.21204782e-01 -1.62500346e+00 -2.86460876e-01 1.47199169e-01
-6.72533214e-01 -3.06952864e-01 -4.18121248e-01 4.66264576e-01
-2.21970886e-01 1.17254806e+00 1.03118606e-01 -5.35032690e-01
1.48994222e-01 3.93089384e-01 1.07816227e-01 -8.85991156e-01
-3.78502697e-01 3.09168816e-01 3.57095450e-01 -6.90301582e-02
-3.52192521e-01 -4.93326128e-01 -1.29612768e+00 1.57679673e-02
-3.54804486e-01 3.98697674e-01 7.32024610e-01 8.46431434e-01
1.86137855e-01 4.62289184e-01 8.71397018e-01 -4.06965524e-01
-5.46576977e-01 -1.34917867e+00 -5.07498682e-01 3.98831785e-01
6.08915417e-03 -3.22346151e-01 -4.86206830e-01 3.70883733e-01] | [14.5980806350708, 6.639204978942871] |
8ef5839e-0264-4859-acd2-abeff1bc73c7 | mv-han-a-hybrid-attentive-networks-based | 2210.07660 | null | https://arxiv.org/abs/2210.07660v1 | https://arxiv.org/pdf/2210.07660v1.pdf | MV-HAN: A Hybrid Attentive Networks based Multi-View Learning Model for Large-scale Contents Recommendation | Industrial recommender systems usually employ multi-source data to improve the recommendation quality, while effectively sharing information between different data sources remain a challenge. In this paper, we introduce a novel Multi-View Approach with Hybrid Attentive Networks (MV-HAN) for contents retrieval at the matching stage of recommender systems. The proposed model enables high-order feature interaction from various input features while effectively transferring knowledge between different types. By employing a well-placed parameters sharing strategy, the MV-HAN substantially improves the retrieval performance in sparse types. The designed MV-HAN inherits the efficiency advantages in the online service from the two-tower model, by mapping users and contents of different types into the same features space. This enables fast retrieval of similar contents with an approximate nearest neighbor algorithm. We conduct offline experiments on several industrial datasets, demonstrating that the proposed MV-HAN significantly outperforms baselines on the content retrieval tasks. Importantly, the MV-HAN is deployed in a real-world matching system. Online A/B test results show that the proposed method can significantly improve the quality of recommendations. | ['Junyang Chen', 'Kai Wang', 'Chaoyun Zhang', 'Ge Fan'] | 2022-10-14 | null | null | null | null | ['multi-view-learning'] | ['computer-vision'] | [-1.94912910e-01 -5.88379622e-01 -6.68993652e-01 -2.25718632e-01
-1.02211702e+00 -6.88480496e-01 3.70928437e-01 -5.57668395e-02
3.13999385e-01 2.18540937e-01 5.12185156e-01 2.28316665e-01
-7.43998945e-01 -8.47507596e-01 -6.82869732e-01 -5.14606118e-01
1.46572486e-01 6.22585714e-01 1.11049071e-01 -6.16054118e-01
3.35060269e-01 2.11011782e-01 -1.80401754e+00 7.62061775e-01
8.03830385e-01 1.36842060e+00 4.59097683e-01 6.40843809e-02
-2.95946617e-02 5.95776081e-01 -1.56555146e-01 -4.95425761e-01
8.42930377e-01 1.87463060e-01 -5.93804538e-01 -2.20492363e-01
7.88460076e-01 -7.31143832e-01 -5.61249495e-01 7.15457439e-01
6.39150679e-01 5.48181236e-01 4.89731550e-01 -1.19710517e+00
-1.04746675e+00 1.04026532e+00 -3.76326919e-01 5.47456779e-02
6.36504114e-01 -4.44240987e-01 1.55027914e+00 -1.27608001e+00
5.51873624e-01 1.32614517e+00 6.02698624e-01 1.41845837e-01
-9.56922650e-01 -7.25568712e-01 5.62831581e-01 4.02324498e-01
-1.38433075e+00 -2.03527197e-01 6.24806464e-01 -9.70151499e-02
7.12536871e-01 4.92843896e-01 4.70991582e-01 9.73431468e-01
-9.93145108e-02 1.00109696e+00 4.71779227e-01 1.61074996e-01
-7.32655153e-02 5.22026658e-01 3.42294484e-01 2.27509812e-01
3.09268743e-01 -1.48586437e-01 -9.36324239e-01 -5.00063479e-01
7.08582580e-01 8.70661676e-01 -4.84545887e-01 -6.09310210e-01
-1.15262592e+00 6.26438379e-01 5.57966352e-01 2.80375123e-01
-6.10531271e-01 -2.09044814e-01 3.66529971e-01 6.14465654e-01
3.59297574e-01 5.04041553e-01 -5.16455233e-01 2.52462149e-01
-7.50796854e-01 2.39301860e-01 9.13869321e-01 1.67747009e+00
5.30839086e-01 -3.24355155e-01 -4.31297153e-01 1.11170137e+00
3.90196204e-01 8.91568065e-01 4.90150005e-01 -1.01462340e+00
6.94495618e-01 5.31734347e-01 2.60221720e-01 -1.28587663e+00
-1.48961684e-02 -7.43507802e-01 -7.43515551e-01 -6.00666702e-01
-1.75686125e-02 2.67238110e-01 -1.47335216e-01 1.36366951e+00
4.40403581e-01 1.57663599e-01 8.45529959e-02 1.12595916e+00
1.00203538e+00 7.63628542e-01 -5.13219953e-01 -1.47895232e-01
1.27188444e+00 -1.45877409e+00 -6.76739991e-01 3.16112578e-01
3.76134038e-01 -9.16300297e-01 8.53471994e-01 6.09567881e-01
-1.21728110e+00 -8.02563548e-01 -9.55768585e-01 -1.25387711e-02
-3.37823510e-01 1.07038215e-01 6.88179135e-01 3.69677275e-01
-7.80900896e-01 8.89138341e-01 -1.39244139e-01 -3.16747159e-01
1.82065010e-01 4.69268650e-01 -2.39537343e-01 -6.42737091e-01
-1.33834612e+00 2.77347833e-01 -5.02186790e-02 1.15909487e-01
-7.81250894e-01 -1.07191515e+00 -4.20024484e-01 5.07330716e-01
6.17951810e-01 -9.72915411e-01 1.13189793e+00 -6.56976342e-01
-1.38441932e+00 -1.91318821e-02 2.89428998e-02 -6.64286539e-02
5.10891676e-02 -7.26385117e-01 -8.66496503e-01 2.23461360e-01
-1.00815465e-04 1.43140629e-01 8.53236318e-01 -1.52284551e+00
-9.10803616e-01 -4.65590417e-01 3.13648582e-01 4.34218049e-01
-8.97854090e-01 -3.13709408e-01 -9.03251648e-01 -7.15690672e-01
4.71805334e-02 -9.08955753e-01 -9.61248856e-03 -5.32729104e-02
-1.63122788e-01 -2.63711661e-01 5.77230036e-01 -4.04261619e-01
1.34330237e+00 -2.05435848e+00 2.68189371e-01 6.77844644e-01
3.11769545e-01 1.13757804e-01 -5.93868673e-01 9.07579899e-01
2.64712542e-01 -4.60877299e-01 7.10707188e-01 -6.54913187e-02
3.84507418e-01 2.53365077e-02 -6.08196795e-01 2.05761552e-01
-6.15584493e-01 8.85313272e-01 -1.01818573e+00 -1.02237925e-01
-5.33365570e-02 4.50492650e-01 -8.07418823e-01 3.65086317e-01
5.21578752e-02 -4.51591909e-02 -7.16205120e-01 7.59226978e-01
6.43736780e-01 -6.54738009e-01 3.44057471e-01 -7.07507432e-01
3.22323978e-01 2.13614464e-01 -1.30043590e+00 1.95965827e+00
-7.88193941e-01 -1.89664111e-01 2.21881777e-01 -6.73674464e-01
7.24377930e-01 3.06257099e-01 7.73952186e-01 -8.85668814e-01
-1.05034411e-01 2.22412080e-01 -2.91892081e-01 -1.90189332e-01
9.09947157e-01 6.32943332e-01 -2.87693813e-02 5.70971906e-01
2.82594897e-02 4.59938109e-01 5.62707819e-02 5.91257393e-01
9.50967014e-01 1.56770386e-02 -1.17140241e-01 -2.56034136e-01
4.78400856e-01 -2.93417871e-01 4.72190917e-01 1.10501647e+00
4.21599954e-01 3.56509060e-01 -3.73564839e-01 -4.23998296e-01
-5.97526133e-01 -9.77433681e-01 5.24378866e-02 1.56432950e+00
7.13436067e-01 -8.38395238e-01 -3.55445504e-01 -6.43844962e-01
4.79604036e-01 4.00532961e-01 -2.55008250e-01 -2.65214860e-01
-2.94985324e-01 -7.72033110e-02 -2.00634673e-01 5.60878098e-01
1.00345768e-01 -7.57001162e-01 1.49968535e-01 2.15646848e-01
-3.26638967e-01 -9.13600266e-01 -9.25380170e-01 -3.16648215e-01
-8.21934521e-01 -8.75788093e-01 -8.05933475e-01 -5.73297083e-01
4.76443052e-01 1.33274758e+00 1.27546358e+00 1.67145342e-01
1.15166195e-01 8.11941922e-01 -7.94582069e-01 -4.77189012e-03
2.98897363e-02 3.19133550e-01 3.70358288e-01 8.59902054e-02
3.14128131e-01 -7.81585813e-01 -1.02444208e+00 9.39888000e-01
-8.42428863e-01 -4.54413354e-01 5.95989525e-01 7.32182801e-01
6.45842373e-01 -1.30242646e-01 7.87739217e-01 -1.11388803e+00
6.65885210e-01 -9.37987566e-01 -3.99027079e-01 4.79117751e-01
-1.16914868e+00 -3.40871751e-01 8.80128384e-01 -5.14266551e-01
-9.26279128e-01 -3.45964253e-01 1.73085779e-01 -7.47807801e-01
2.04131201e-01 4.68079060e-01 -2.08862707e-01 -2.36532241e-02
3.72445345e-01 8.78486112e-02 -2.54484922e-01 -9.17936683e-01
8.01478982e-01 9.11205947e-01 3.33484530e-01 -7.11881399e-01
9.20448720e-01 3.29146564e-01 -3.10582995e-01 -2.49494448e-01
-1.07028174e+00 -1.09808671e+00 -3.01800847e-01 -1.93926632e-01
4.29128930e-02 -1.42712939e+00 -6.89024329e-01 3.47778969e-03
-6.64750040e-01 2.33494297e-01 -3.85499686e-01 4.31919605e-01
-3.64688069e-01 4.07808363e-01 -7.90175259e-01 -2.95618266e-01
-8.62108290e-01 -9.91557539e-01 1.15286875e+00 1.21518433e-01
9.32819769e-02 -5.84647477e-01 -1.68018416e-01 7.50006497e-01
7.22781241e-01 -8.20259213e-01 8.13614011e-01 -1.00887120e+00
-8.46729279e-01 -3.11731607e-01 -2.82166451e-01 3.21862027e-02
2.79127091e-01 -5.42839289e-01 -7.88820744e-01 -6.81270957e-01
-3.60386103e-01 -1.56705290e-01 7.41865396e-01 5.63775972e-02
1.18761981e+00 -2.99481839e-01 -4.47647452e-01 3.85004669e-01
1.19370437e+00 7.17521459e-02 2.65236139e-01 2.13654131e-01
8.98554087e-01 5.30124545e-01 1.10859883e+00 7.79374301e-01
5.60957551e-01 9.61596966e-01 4.70233709e-01 1.64242461e-02
7.96348527e-02 -3.67658675e-01 4.67520714e-01 1.59332275e+00
9.24180523e-02 -4.27484095e-01 6.00332059e-02 5.22004545e-01
-2.22925544e+00 -1.04398608e+00 5.87391146e-02 2.29577303e+00
5.13739288e-01 -2.88272977e-01 2.50817925e-01 -2.81713367e-01
6.32642388e-01 -1.04919806e-01 -6.31158888e-01 6.67593554e-02
1.39761612e-01 -1.95642963e-01 2.91517466e-01 -1.52688520e-02
-8.80724847e-01 5.37202358e-01 5.93704462e+00 1.08905888e+00
-5.71848631e-01 1.35151207e-01 1.65739022e-02 -4.90556151e-01
-6.59876347e-01 -2.92314589e-01 -1.05380559e+00 4.03429568e-01
7.26332366e-01 -3.36304218e-01 9.05328512e-01 1.04465771e+00
-7.78110400e-02 4.95636910e-01 -1.28128219e+00 1.10830677e+00
2.91210771e-01 -1.39359534e+00 5.52625418e-01 4.78637479e-02
9.17030275e-01 9.36344489e-02 3.94694835e-01 4.95125651e-01
2.74364501e-01 -2.54190087e-01 4.78292972e-01 7.46902049e-01
3.95670772e-01 -7.00921357e-01 7.13465393e-01 1.72387347e-01
-1.52180707e+00 -4.77377355e-01 -7.02433825e-01 4.09381479e-01
2.05765158e-01 6.13465190e-01 -1.87833115e-01 1.25691450e+00
8.90453696e-01 1.05887842e+00 -1.45030364e-01 1.00063026e+00
5.29857874e-01 2.02962413e-01 -2.07799986e-01 7.74539039e-02
-1.10683300e-01 -4.76702929e-01 4.09170091e-01 8.03465545e-01
5.45056224e-01 5.10898158e-02 4.25951183e-01 6.89248979e-01
-3.51027042e-01 5.77698827e-01 -6.34148359e-01 1.21317849e-01
7.10956275e-01 1.66690004e+00 -1.53306603e-01 -4.38569248e-01
-6.40728712e-01 8.29873025e-01 1.58923924e-01 3.33440453e-01
-6.65111303e-01 -1.80405334e-01 7.64816403e-01 7.46008009e-02
7.12982178e-01 4.60088402e-01 4.99776632e-01 -1.21049142e+00
7.47794732e-02 -1.41050351e+00 5.38190484e-01 -7.95056105e-01
-2.00614667e+00 6.24167860e-01 -4.19146232e-02 -1.74378181e+00
-1.88307315e-02 -2.35446513e-01 -4.52405065e-02 6.20497644e-01
-1.51230192e+00 -1.17697847e+00 -2.78583407e-01 7.92516470e-01
6.54399753e-01 -5.15481472e-01 8.05576861e-01 1.04964018e+00
-4.65563446e-01 9.55915749e-01 6.85104370e-01 -5.31588495e-01
1.12143171e+00 -9.17837799e-01 7.40089789e-02 2.02189863e-01
4.50321347e-01 1.18016768e+00 1.47424534e-01 -4.27154273e-01
-1.98432183e+00 -1.21678698e+00 2.46251807e-01 -2.02088863e-01
5.74229658e-01 -1.77304164e-01 -8.28450322e-01 4.68839705e-01
3.41362506e-01 -4.37220838e-03 1.03648138e+00 4.92401510e-01
-7.45694339e-01 -6.36567831e-01 -1.06780744e+00 4.08637226e-01
1.20817196e+00 -5.89045584e-01 -2.49715313e-01 6.12769961e-01
8.65604699e-01 -2.08544627e-01 -1.59791541e+00 3.27639431e-01
9.93550181e-01 -7.19284892e-01 1.38105333e+00 -6.75056756e-01
4.20576721e-01 -2.51493365e-01 -6.10094726e-01 -1.28364742e+00
-8.61637235e-01 -6.62521243e-01 -5.29523730e-01 1.39077377e+00
2.98837990e-01 -6.39555037e-01 3.91096324e-01 4.41323519e-01
2.75996123e-02 -7.10064769e-01 -2.48654440e-01 -6.98972762e-01
-4.08991218e-01 6.96979687e-02 9.40611124e-01 9.30312335e-01
1.35792315e-01 4.09984976e-01 -7.11234212e-01 2.38604277e-01
5.40045202e-01 8.73329043e-01 8.53594601e-01 -1.48551118e+00
-7.43931532e-01 -9.28427652e-02 9.78021100e-02 -1.52775943e+00
2.73923352e-02 -1.11548889e+00 -3.56800437e-01 -1.29613113e+00
5.97029209e-01 -6.80829227e-01 -9.54047740e-01 2.19011098e-01
-1.27796724e-01 2.14569867e-01 3.11619371e-01 5.87911069e-01
-1.14505291e+00 5.45812011e-01 1.31899822e+00 -2.09599569e-01
2.79017128e-02 3.50812197e-01 -9.67498481e-01 3.39709312e-01
4.88268167e-01 -3.90079230e-01 -7.66500175e-01 -3.60360831e-01
5.71123958e-01 5.34495600e-02 -2.48693362e-01 -6.41358674e-01
3.75399530e-01 1.83081210e-01 1.86916158e-01 -8.19838464e-01
4.67993885e-01 -1.38891149e+00 3.53118956e-01 -1.97899163e-01
-5.36343098e-01 3.30907851e-01 1.94874071e-02 1.01670015e+00
-5.13091758e-02 -1.15606546e-01 6.98378682e-03 -3.17678265e-02
-5.84612608e-01 4.80736881e-01 -1.26735149e-02 -3.17282081e-01
6.22769535e-01 2.04845563e-01 -4.51634020e-01 -6.94910467e-01
-5.69896579e-01 4.88617688e-01 3.07386011e-01 8.20802808e-01
5.48172832e-01 -1.72392726e+00 -3.34524572e-01 1.43098623e-01
4.69260901e-01 -4.72020894e-01 5.38967848e-01 7.91188300e-01
1.86627805e-01 4.10394162e-01 -4.37955409e-02 -3.61049026e-01
-1.22678471e+00 9.80193675e-01 -6.01855107e-02 -4.69957799e-01
-6.40248895e-01 5.88296115e-01 3.26712370e-01 -5.72585583e-01
4.29883331e-01 7.81640038e-02 -4.88534391e-01 1.32329494e-01
6.93298042e-01 7.54166424e-01 1.40117452e-01 -5.31043887e-01
1.10485449e-01 5.53949416e-01 -6.82908535e-01 4.01037157e-01
1.43962622e+00 -5.24607539e-01 -1.84070408e-01 3.27107847e-01
1.09449339e+00 1.83665067e-01 -5.96295774e-01 -8.77174735e-01
-4.49261695e-01 -6.77805007e-01 2.90212125e-01 -8.00609291e-01
-1.59484100e+00 2.86563367e-01 6.06806278e-01 3.14913988e-01
1.15385377e+00 -1.76776662e-01 1.33506489e+00 8.66979241e-01
7.70043314e-01 -1.08587098e+00 7.11955177e-03 1.41380727e-01
8.57474864e-01 -1.03627825e+00 2.18472168e-01 -7.33889997e-01
-5.10724306e-01 8.63765180e-01 6.52135611e-01 -1.17986098e-01
8.06028008e-01 3.24700885e-02 4.40484434e-02 -4.22167450e-01
-1.13954985e+00 -3.16292457e-02 6.25428438e-01 2.55622625e-01
1.68193325e-01 -1.18189052e-01 1.77601576e-02 8.80483210e-01
3.27683389e-01 -1.61966071e-01 4.73718643e-02 8.39224279e-01
-1.93385437e-01 -1.34212518e+00 -1.55399993e-01 8.66287231e-01
-4.59434569e-01 -2.55542904e-01 -1.80860877e-01 3.50169718e-01
-3.62311900e-01 1.18695402e+00 -1.22990310e-01 -6.85275316e-01
7.00204968e-01 -3.09879363e-01 3.78850102e-01 -4.83788550e-01
-1.16011977e+00 4.14513826e-01 -3.35698873e-02 -1.17785537e+00
-4.82727081e-01 -5.68947852e-01 -7.01755941e-01 -3.03142130e-01
-8.75275552e-01 4.65533435e-01 3.12561184e-01 6.43566847e-01
1.03063107e+00 6.02034867e-01 1.08292818e+00 -8.77343595e-01
-1.08342505e+00 -7.65800178e-01 -9.68034804e-01 6.38152599e-01
2.60255132e-02 -7.57573843e-01 -1.72746271e-01 -3.14874470e-01] | [10.143126487731934, 5.5727081298828125] |
940a5c2a-1f7d-4ea1-b6bd-0d8d1c8d1c5e | awesome-typography-statistics-based-text | 1611.09026 | null | http://arxiv.org/abs/1611.09026v2 | http://arxiv.org/pdf/1611.09026v2.pdf | Awesome Typography: Statistics-Based Text Effects Transfer | In this work, we explore the problem of generating fantastic special-effects
for the typography. It is quite challenging due to the model diversities to
illustrate varied text effects for different characters. To address this issue,
our key idea is to exploit the analytics on the high regularity of the spatial
distribution for text effects to guide the synthesis process. Specifically, we
characterize the stylized patches by their normalized positions and the optimal
scales to depict their style elements. Our method first estimates these two
features and derives their correlation statistically. They are then converted
into soft constraints for texture transfer to accomplish adaptive multi-scale
texture synthesis and to make style element distribution uniform. It allows our
algorithm to produce artistic typography that fits for both local texture
patterns and the global spatial distribution in the example. Experimental
results demonstrate the superiority of our method for various text effects over
conventional style transfer methods. In addition, we validate the effectiveness
of our algorithm with extensive artistic typography library generation. | ['Jiaying Liu', 'Shuai Yang', 'Zongming Guo', 'Zhouhui Lian'] | 2016-11-28 | awesome-typography-statistics-based-text-1 | http://openaccess.thecvf.com/content_cvpr_2017/html/Yang_Awesome_Typography_Statistics-Based_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Yang_Awesome_Typography_Statistics-Based_CVPR_2017_paper.pdf | cvpr-2017-7 | ['text-effects-transfer'] | ['natural-language-processing'] | [ 1.19130999e-01 -3.70298177e-01 -3.41996588e-02 4.21199389e-02
-2.00141832e-01 -7.71454632e-01 5.26711881e-01 -4.60819900e-01
3.27444613e-01 8.28665018e-01 3.13065797e-01 1.39743164e-01
-1.03287205e-01 -8.87376368e-01 -5.06836414e-01 -6.50499940e-01
3.95557761e-01 3.22211593e-01 2.74955720e-01 -4.14554030e-01
7.04001725e-01 6.98104143e-01 -1.51054168e+00 1.62729770e-01
1.08397603e+00 7.71682441e-01 4.28036422e-01 5.68437159e-01
-2.97216535e-01 3.93849939e-01 -7.48552501e-01 -3.20394576e-01
2.64737010e-01 -7.02643275e-01 -2.53438503e-01 2.52320170e-01
3.68275374e-01 -1.64621741e-01 -8.88179615e-02 1.12643361e+00
4.60002810e-01 1.30539969e-01 1.10569739e+00 -8.73110235e-01
-9.82437551e-01 3.13464165e-01 -9.04486895e-01 -1.73060104e-01
3.31721157e-01 1.05016813e-01 8.53797853e-01 -7.75977910e-01
8.44784796e-01 1.49082303e+00 4.73948509e-01 4.74653810e-01
-1.40876591e+00 -7.22599745e-01 -1.48338497e-01 -2.04404503e-01
-1.18998933e+00 -1.53467566e-01 1.04138780e+00 -3.58684897e-01
1.22619588e-02 4.41367686e-01 7.90081024e-01 1.03895712e+00
5.02160132e-01 7.08960176e-01 1.51020944e+00 -5.66270411e-01
2.40665108e-01 1.40591413e-01 -4.21026140e-01 5.60965836e-01
9.46731344e-02 -3.26029509e-02 -5.49366891e-01 1.42216444e-01
1.59909415e+00 -2.03697905e-01 -1.20853290e-01 -1.08128846e-01
-1.20864999e+00 5.61329663e-01 2.16793627e-01 3.08376700e-01
-9.91051719e-02 1.77771509e-01 6.81412816e-02 2.10615739e-01
5.72813749e-01 9.01001155e-01 -1.52274758e-01 -2.68470019e-01
-8.20475876e-01 5.34244299e-01 6.56293213e-01 1.09101963e+00
7.36985445e-01 8.48960876e-02 -2.96822935e-01 1.28857553e+00
-2.14219436e-01 8.42828155e-01 2.35859871e-01 -7.43707120e-01
7.28023946e-02 3.79195392e-01 6.86318800e-02 -1.27486646e+00
1.08482130e-02 -1.38072968e-01 -9.90468681e-01 3.68435115e-01
6.14954948e-01 -1.15964800e-01 -7.39461482e-01 1.39760518e+00
1.90169901e-01 -1.13270983e-01 -4.49051857e-01 8.73552084e-01
4.45577949e-01 7.17427671e-01 -1.20868564e-01 8.07125121e-03
1.28057277e+00 -6.07206225e-01 -1.03812051e+00 1.87586501e-01
2.11908817e-01 -1.21270370e+00 1.70099318e+00 1.30788073e-01
-1.18331540e+00 -5.53574145e-01 -8.61612022e-01 6.89979568e-02
1.64007545e-02 3.45260531e-01 5.82017720e-01 4.89016593e-01
-7.04955757e-01 7.19140291e-01 -2.65581876e-01 -1.90712407e-01
3.07307601e-01 -2.36324407e-02 1.35319769e-01 5.14872611e-01
-9.42346334e-01 4.39714640e-01 -3.21573168e-02 -1.24948993e-01
-1.19197600e-01 -8.35741818e-01 -3.82370532e-01 -4.26262729e-02
1.50479555e-01 -7.21507549e-01 9.05579090e-01 -1.24750900e+00
-2.12570477e+00 6.83055460e-01 -8.84637311e-02 3.36143821e-01
5.72075665e-01 9.03996229e-02 -1.87646136e-01 -1.92370474e-01
2.64650017e-01 6.18893027e-01 1.00998318e+00 -1.63790500e+00
-6.20061815e-01 -2.23082416e-02 -1.86260656e-01 2.68607080e-01
-4.56913680e-01 -1.91743270e-01 -7.15138495e-01 -1.36497366e+00
1.20075703e-01 -9.18816328e-01 -1.22084357e-01 -7.93944858e-03
-6.22737169e-01 6.22373149e-02 6.93298042e-01 -4.14944917e-01
1.24437869e+00 -2.18858767e+00 3.40507776e-01 6.65834844e-01
1.52214900e-01 -3.57626200e-01 -8.24990310e-03 3.23190004e-01
2.30364755e-01 1.81691721e-01 1.05087437e-01 -9.62617919e-02
8.07553530e-02 -6.94464445e-02 -6.10468686e-01 2.99888700e-02
1.25654429e-01 8.73899817e-01 -6.94377244e-01 -5.27854204e-01
1.00786589e-01 1.51088879e-01 -5.80018163e-01 2.48620033e-01
-3.95875007e-01 7.84855008e-01 -7.97564983e-01 6.52088940e-01
8.25159788e-01 3.70601378e-02 1.64316073e-01 -3.06729853e-01
-1.59445614e-01 -1.81655809e-01 -1.23059213e+00 1.21178257e+00
-5.88555038e-01 6.45979285e-01 -2.24661261e-01 -1.72703087e-01
1.37587368e+00 -2.02348843e-01 2.20628276e-01 -8.22247446e-01
1.76287368e-01 3.60423744e-01 -2.39974082e-01 -4.06205982e-01
8.07556808e-01 -4.31129098e-01 -2.22402960e-01 5.86528361e-01
-4.83892977e-01 -6.83741927e-01 1.40554234e-01 -1.32265985e-01
4.45742160e-01 1.94679067e-01 -5.49658164e-02 -7.47346401e-01
3.08602780e-01 -1.11896396e-01 3.76950383e-01 6.00100279e-01
4.54246640e-01 8.86937857e-01 8.85344803e-01 -4.89015937e-01
-1.46231532e+00 -1.05074942e+00 -2.17966750e-01 8.57664049e-01
4.68650222e-01 -2.83872604e-01 -9.90457237e-01 -3.59677881e-01
7.07120588e-03 4.99147296e-01 -9.65990245e-01 5.40934242e-02
-7.00377762e-01 -5.40115058e-01 3.09940875e-01 4.77345735e-01
6.02356374e-01 -1.18961334e+00 -1.66861489e-01 -3.79010141e-02
-3.81493159e-02 -6.56421423e-01 -8.31599951e-01 -3.36116225e-01
-5.92603147e-01 -6.09049976e-01 -1.07307720e+00 -7.87496030e-01
7.79486716e-01 5.83778955e-02 9.38341498e-01 4.11589630e-02
-2.64802307e-01 -1.97340790e-02 -2.45557770e-01 -1.99148044e-01
-5.49848497e-01 1.35971770e-01 2.21375618e-02 1.39757603e-01
-2.56846309e-01 -7.29298890e-01 -6.89557493e-01 7.47434258e-01
-9.54845786e-01 4.41053420e-01 5.27104199e-01 6.71039045e-01
7.12407649e-01 1.89007446e-01 4.83764023e-01 -8.96841764e-01
1.04898751e+00 -1.60333395e-01 -5.76330662e-01 2.44621694e-01
-1.93795115e-01 1.92093655e-01 8.83040190e-01 -8.03478122e-01
-1.30778396e+00 -3.01908374e-01 2.70065814e-01 -3.10993016e-01
-2.34976802e-02 1.34206846e-01 -2.78346628e-01 -4.39416058e-02
6.51876628e-01 1.83118850e-01 -8.17145482e-02 -3.46952200e-01
4.70234632e-01 4.82611239e-01 3.59186679e-01 -1.08803332e+00
9.54186380e-01 4.79145527e-01 1.71225563e-01 -9.30281878e-01
-5.19788086e-01 3.31033468e-01 -6.55623257e-01 -4.07237589e-01
8.14935565e-01 -3.01206261e-01 -7.25933969e-01 6.07655704e-01
-1.02987123e+00 -6.89149380e-01 -4.07042980e-01 7.84133598e-02
-7.01736450e-01 2.51286328e-01 -5.41361332e-01 -6.74275398e-01
2.23597549e-02 -1.14865339e+00 1.21249783e+00 3.25705320e-01
-4.74909306e-01 -1.14758897e+00 2.14229375e-01 -1.76478818e-01
4.19742137e-01 2.92445749e-01 1.14643025e+00 4.80285108e-01
-6.62275136e-01 8.98055062e-02 -4.63594794e-01 -1.48353189e-01
5.87052345e-01 3.98717403e-01 -6.56189501e-01 4.17552069e-02
-3.58264089e-01 1.13188801e-02 5.86762369e-01 5.42032003e-01
1.39036298e+00 -1.57588840e-01 -2.41512313e-01 8.93515885e-01
1.25449300e+00 2.04648525e-01 8.27806830e-01 4.19590950e-01
9.74384665e-01 7.25417256e-01 6.74986064e-01 5.27315974e-01
-1.35047793e-01 9.94282782e-01 -2.53867269e-01 -3.05570871e-01
-4.72339451e-01 -4.68211025e-01 1.91092696e-02 7.91738749e-01
-4.49378610e-01 -3.30866575e-01 -6.35527790e-01 3.31934184e-01
-1.77443647e+00 -8.84197056e-01 -5.35930768e-02 2.00966620e+00
9.67405796e-01 -6.48667896e-03 2.21039638e-01 -2.45408744e-01
9.68523681e-01 5.53907873e-03 -3.17054600e-01 -5.44487119e-01
-4.31145608e-01 3.10168594e-01 5.67349017e-01 4.00045842e-01
-7.10546792e-01 1.26086140e+00 6.86586285e+00 1.52151930e+00
-1.17897773e+00 -4.20469701e-01 9.32946384e-01 1.17175085e-02
-7.58880436e-01 -1.40566483e-01 -6.85136974e-01 6.66380286e-01
1.10068992e-01 -2.34612256e-01 6.84176922e-01 4.44843143e-01
3.91451389e-01 -2.48429663e-02 -7.19185233e-01 1.02776206e+00
-1.64056495e-01 -1.45068359e+00 3.89094114e-01 -5.68106547e-02
1.19104946e+00 -1.04730475e+00 4.99206007e-01 -1.49051741e-01
3.64942670e-01 -1.10765409e+00 9.99825478e-01 7.57613599e-01
1.38828707e+00 -8.09421182e-01 1.52550444e-01 -1.37067869e-01
-1.27130771e+00 1.40636548e-01 -4.78602946e-01 -4.42919694e-02
2.32182257e-02 5.11641145e-01 -5.02386093e-01 3.17725092e-01
5.24992943e-01 7.29756832e-01 -4.49795842e-01 7.11239517e-01
-3.06395024e-01 3.95166576e-01 -2.00872466e-01 -5.57705462e-01
1.56452674e-02 -6.46356285e-01 5.06906152e-01 1.18831015e+00
6.63521171e-01 -4.01197150e-02 -1.36446953e-01 1.31100166e+00
-1.46135231e-02 4.26590711e-01 -5.30690074e-01 9.95295718e-02
5.07754922e-01 1.19590652e+00 -1.16113329e+00 -1.64866135e-01
1.25589266e-01 1.07089365e+00 2.89568305e-01 5.54988623e-01
-8.18588018e-01 -5.18698990e-01 3.66957456e-01 5.08642137e-01
1.78446144e-01 -2.20509365e-01 -1.07130110e+00 -1.07652152e+00
-4.70835678e-02 -9.71908331e-01 -2.85959303e-01 -8.73064697e-01
-1.45816588e+00 4.79223937e-01 -6.29023984e-02 -1.38886487e+00
2.34669745e-01 -4.60391194e-01 -9.86017466e-01 9.80285943e-01
-7.79182613e-01 -1.29123533e+00 -3.32362145e-01 4.09756780e-01
5.84323525e-01 -2.26602480e-01 5.41519642e-01 -1.42744640e-02
-4.86435086e-01 6.98757470e-01 3.79045367e-01 -2.46128723e-01
9.50688183e-01 -1.32938623e+00 4.65894341e-01 4.73732203e-01
-2.63420224e-01 5.62169909e-01 7.79464066e-01 -9.99240875e-01
-1.21240270e+00 -8.22694123e-01 4.14918154e-01 -3.73466820e-01
8.18685830e-01 -5.14610112e-01 -6.19512081e-01 2.94462591e-01
6.89215809e-02 -6.56428576e-01 3.37309748e-01 1.88253358e-01
-1.32006273e-01 -8.04617777e-02 -9.06642675e-01 1.29994392e+00
1.11726415e+00 -1.73959002e-01 -2.22128302e-01 1.56775296e-01
2.32851818e-01 -3.62834007e-01 -7.15665221e-01 1.27006963e-01
8.39649618e-01 -9.16288853e-01 7.15255797e-01 -1.36258975e-01
9.46396410e-01 -2.96471626e-01 1.20181881e-01 -1.59274030e+00
-5.62275827e-01 -8.68915677e-01 5.91930091e-01 1.37684405e+00
3.52264166e-01 -5.51674306e-01 6.75710022e-01 3.33218455e-01
1.46120220e-01 -6.08186781e-01 -4.51058567e-01 -5.91199815e-01
1.62829861e-01 -1.14913452e-02 9.39886153e-01 1.01560009e+00
-1.53687224e-01 1.67471960e-01 -5.93798637e-01 -2.14821026e-01
5.48705399e-01 5.71990728e-01 1.09793675e+00 -1.02352440e+00
-5.28633177e-01 -8.67471337e-01 -1.08770423e-01 -1.25100863e+00
6.11388721e-02 -6.04323506e-01 4.76680771e-02 -1.14025247e+00
2.92701066e-01 -1.05051553e+00 1.40146896e-01 1.24099916e-02
-3.38872343e-01 4.90920067e-01 9.53490660e-02 4.77293551e-01
-1.37918100e-01 6.15592360e-01 2.12876987e+00 9.53744650e-02
-5.57114840e-01 -9.68344733e-02 -7.44816422e-01 6.43153787e-01
8.34509492e-01 -1.32580355e-01 -3.55192751e-01 -2.46457815e-01
3.66559774e-01 -1.62674040e-02 1.77008793e-01 -8.19666803e-01
-3.05142611e-01 -5.80641270e-01 6.18161738e-01 -4.29172963e-01
3.50953713e-02 -5.14927566e-01 2.42568806e-01 7.26472810e-02
-2.86303133e-01 -4.36202772e-02 2.03590229e-01 3.30057472e-01
-2.06565410e-02 9.64772925e-02 9.25349951e-01 1.26875401e-01
-3.07541609e-01 1.73668608e-01 -3.51006299e-01 -1.50429420e-02
8.35098326e-01 -3.71253580e-01 -1.99272737e-01 -5.54956436e-01
-5.04583359e-01 -1.66521832e-01 9.53194082e-01 2.83910066e-01
4.63768363e-01 -1.69052315e+00 -4.94897574e-01 4.90497291e-01
-7.30318651e-02 -1.61229238e-01 2.39208832e-01 5.30281186e-01
-8.94221604e-01 -1.04788326e-01 -6.45099759e-01 -5.32684326e-01
-1.11937284e+00 3.05124938e-01 1.41395003e-01 -5.38389990e-03
-5.92885733e-01 6.05983615e-01 8.22774589e-01 -1.59591541e-01
-1.60296559e-01 -2.89684653e-01 -9.54161808e-02 -1.34362593e-01
3.53434741e-01 3.55039924e-01 -4.33836490e-01 -2.74612308e-01
2.19605535e-01 1.25593126e+00 1.23469621e-01 -2.42432967e-01
1.18578315e+00 -1.69284120e-01 -3.11024040e-01 4.78617102e-01
7.18998969e-01 6.87584639e-01 -1.59921765e+00 1.06958769e-01
-3.87988389e-01 -8.73692930e-01 -2.84582019e-01 -5.79974592e-01
-9.63250875e-01 5.63087642e-01 2.31903642e-01 4.27148193e-01
1.05036354e+00 -1.60008341e-01 8.83172691e-01 -3.57914604e-02
1.56917825e-01 -1.27835453e+00 3.77137721e-01 4.31028247e-01
1.02494001e+00 -5.81573308e-01 -1.62743077e-01 -6.98493659e-01
-7.38366544e-01 1.34102488e+00 5.08815885e-01 -4.93247092e-01
3.36081415e-01 5.65783441e-01 2.49422774e-01 -1.24697194e-01
-2.37778232e-01 1.86205775e-01 4.82061148e-01 3.28910798e-01
3.88958067e-01 1.72832817e-01 -3.55842859e-01 2.93285847e-01
-5.74927092e-01 -1.83113277e-01 4.45162565e-01 5.32921672e-01
-4.31324184e-01 -1.19736600e+00 -6.20774806e-01 2.23439261e-01
-2.98647881e-01 -9.77672487e-02 -5.77259362e-01 6.95387006e-01
6.64975569e-02 4.90078568e-01 1.41698688e-01 -3.75166714e-01
4.47020769e-01 -3.74298960e-01 7.78575718e-01 -2.53837019e-01
-3.34678054e-01 5.51320612e-01 -1.66221261e-01 -2.48189479e-01
2.15300340e-02 -5.02360284e-01 -9.21805203e-01 -7.03141093e-01
-1.64342955e-01 -2.17883855e-01 4.62038994e-01 5.10407925e-01
1.51516542e-01 7.40826309e-01 9.22986984e-01 -1.10593963e+00
-4.53666598e-02 -7.95786560e-01 -1.00236452e+00 8.26926053e-01
-1.72645763e-01 -8.67046177e-01 -3.67828846e-01 2.09675357e-01] | [11.70462703704834, -0.5701958537101746] |
bcc504e9-bd88-4783-9a0a-b23d57de3399 | cross-lingual-retrieval-augmented-prompt-for | 2212.09651 | null | https://arxiv.org/abs/2212.09651v4 | https://arxiv.org/pdf/2212.09651v4.pdf | Cross-Lingual Retrieval Augmented Prompt for Low-Resource Languages | Multilingual Pretrained Language Models (MPLMs) have shown their strong multilinguality in recent empirical cross-lingual transfer studies. In this paper, we propose the Prompts Augmented by Retrieval Crosslingually (PARC) pipeline to improve the zero-shot performance on low-resource languages (LRLs) by augmenting the context with semantically similar sentences retrieved from a high-resource language (HRL) as prompts. PARC improves the zero-shot performance on three downstream tasks (binary sentiment classification, topic categorization and natural language inference) with multilingual parallel test sets across 10 LRLs covering 6 language families in both unlabeled settings (+5.1%) and labeled settings (+16.3%). PARC-labeled also outperforms the finetuning baseline by 3.7%. We find a significant positive correlation between cross-lingual transfer performance on one side, and the similarity between the high- and low-resource languages as well as the amount of low-resource pretraining data on the other side. A robustness analysis suggests that PARC has the potential to achieve even stronger performance with more powerful MPLMs. | ['Hinrich Schütze', 'Helmut Schmid', 'Sheng Liang', 'Ercong Nie'] | 2022-12-19 | null | null | null | null | ['cross-lingual-transfer'] | ['natural-language-processing'] | [-2.35658720e-01 -9.66974646e-02 -4.97607172e-01 -5.82064867e-01
-1.78929210e+00 -8.32553148e-01 8.46781611e-01 2.33432099e-01
-1.04801309e+00 7.68544912e-01 4.66873646e-01 -4.03555900e-01
2.85680443e-01 -4.02212977e-01 -9.39686477e-01 -2.56472856e-01
1.23627983e-01 5.84182918e-01 1.24304280e-01 -3.41185838e-01
-1.66216001e-01 -1.70947481e-02 -9.92900312e-01 7.25071549e-01
1.11686492e+00 4.21026617e-01 4.04831231e-01 1.10023133e-01
-2.18404174e-01 5.16767144e-01 -1.74029335e-01 -7.93746829e-01
2.08074674e-01 -7.77815282e-02 -7.45740652e-01 -4.67861861e-01
7.86258280e-01 -6.89659789e-02 2.24798977e-01 7.42886245e-01
5.07370234e-01 9.29286797e-03 6.71349525e-01 -8.42365384e-01
-1.05128205e+00 1.12363160e+00 -7.38724411e-01 6.81437850e-02
1.53900996e-01 1.39480203e-01 1.13599539e+00 -1.28241241e+00
9.85268593e-01 1.55971313e+00 6.03284478e-01 4.77153152e-01
-1.31984663e+00 -9.69189167e-01 3.02235156e-01 -9.83945876e-02
-1.36213088e+00 -6.77418768e-01 2.81389028e-01 -4.72245961e-01
1.46573043e+00 -2.74188697e-01 1.69173703e-01 1.16099179e+00
4.48315181e-02 9.19247329e-01 1.54937100e+00 -8.18346143e-01
-1.29506558e-01 6.19065702e-01 2.13732660e-01 4.74081874e-01
4.86385599e-02 -7.90937338e-03 -8.45577657e-01 1.80428252e-01
1.40235960e-01 -5.10734677e-01 -1.83300674e-01 7.56862387e-02
-1.21517515e+00 9.33550715e-01 2.47562706e-01 5.24556816e-01
-1.18269406e-01 -3.28795724e-02 7.96952009e-01 5.86990595e-01
9.94037628e-01 6.19418085e-01 -1.11241639e+00 7.78944194e-02
-6.99947000e-01 -1.72129735e-01 6.05840564e-01 1.04981089e+00
8.43113363e-01 -2.34524328e-02 -2.18261689e-01 1.42835975e+00
2.43562981e-01 9.44618404e-01 8.33942056e-01 -3.88192654e-01
8.48621786e-01 3.14944535e-01 -3.97669375e-01 -3.04677427e-01
-2.51132064e-03 -3.88799191e-01 -2.27813169e-01 -3.54389787e-01
3.85407478e-01 -3.90528321e-01 -6.62407935e-01 2.23774767e+00
1.40027672e-01 -3.44047487e-01 4.88325447e-01 3.79620284e-01
5.32366216e-01 9.09875512e-01 6.05407655e-01 -5.70245720e-02
1.53000653e+00 -9.26924884e-01 -3.71127129e-01 -4.55669612e-01
1.29636681e+00 -1.08704984e+00 1.70849550e+00 1.49495173e-02
-8.11037362e-01 -5.28303742e-01 -8.84922802e-01 -4.36648041e-01
-7.20447779e-01 3.83636922e-01 4.60096449e-01 5.59369087e-01
-1.33549988e+00 6.22790903e-02 -5.57476044e-01 -6.08130515e-01
1.02813490e-01 4.59480919e-02 -5.38100541e-01 -7.80738056e-01
-1.51753974e+00 1.08131933e+00 5.15729725e-01 -4.35078233e-01
-8.25351596e-01 -1.17642379e+00 -9.87724662e-01 -1.58105239e-01
1.09588981e-01 -1.12337135e-01 1.07349610e+00 -9.42205012e-01
-1.17704272e+00 1.29551911e+00 2.83899773e-02 -2.90709734e-01
3.60054523e-01 -6.73781037e-01 -3.61782730e-01 -2.40151659e-01
7.00767577e-01 1.20530283e+00 2.26133510e-01 -7.04742610e-01
-7.42273688e-01 -3.43646049e-01 -3.41218621e-01 5.48359871e-01
-5.81535816e-01 5.62503159e-01 -5.60236931e-01 -6.47271991e-01
-4.56074297e-01 -9.83014703e-01 1.90913141e-01 -4.96523917e-01
1.24337032e-01 -7.94887960e-01 4.22709584e-01 -9.51368093e-01
7.96449423e-01 -2.01901984e+00 -2.07865372e-01 -1.86948493e-01
-6.49061978e-01 2.63135046e-01 -6.15133882e-01 5.15600622e-01
-9.44855139e-02 2.69098639e-01 1.43930569e-01 -5.38011849e-01
3.02407388e-02 1.14789344e-02 -3.57480764e-01 2.99982339e-01
6.02246881e-01 1.00689590e+00 -8.86897027e-01 -3.53254914e-01
-1.46518171e-01 6.02656186e-01 -6.17051780e-01 -9.69955176e-02
-5.78999147e-02 3.95753413e-01 4.85930406e-02 4.12874788e-01
3.27583283e-01 8.01313519e-02 2.88312405e-01 1.17592104e-01
-4.07736540e-01 8.66172850e-01 -4.71114784e-01 2.04113317e+00
-9.84550059e-01 6.32397830e-01 -1.09866127e-01 -5.10342240e-01
7.19685674e-01 3.88241053e-01 1.48313381e-02 -1.15965104e+00
-3.21626514e-01 6.89791918e-01 1.12143710e-01 -1.80607632e-01
6.00438356e-01 -3.67921025e-01 -5.13130426e-01 6.83889270e-01
4.95858371e-01 -1.18060438e-02 1.66248471e-01 3.84145260e-01
7.41803527e-01 3.26416284e-01 2.13651463e-01 -8.14938664e-01
4.89906043e-01 1.10957719e-01 5.04901886e-01 6.25341415e-01
-1.23533033e-01 1.00590818e-01 1.71067029e-01 1.72394916e-01
-9.04121697e-01 -1.03265035e+00 -4.75969791e-01 1.85898125e+00
-5.59151530e-01 -3.18000883e-01 -6.87350631e-01 -8.50591898e-01
1.02476493e-01 9.52272177e-01 -3.72164011e-01 -9.63024497e-02
-6.72313571e-01 -6.42357469e-01 7.55649626e-01 4.61205482e-01
2.12894842e-01 -1.06877041e+00 1.94935009e-01 6.21370450e-02
-2.58833468e-01 -1.39620185e+00 -6.73009098e-01 1.75988048e-01
-3.88073415e-01 -4.04207736e-01 -8.71053219e-01 -9.71628249e-01
3.30551863e-01 2.42129967e-01 1.43949699e+00 -3.90203953e-01
6.91335648e-02 4.45976555e-01 -4.72516477e-01 -3.35613400e-01
-4.78779167e-01 4.67010021e-01 4.35526937e-01 -3.31437975e-01
6.40528321e-01 -9.47357640e-02 -2.66912300e-02 7.36197606e-02
-4.83923793e-01 5.79885095e-02 5.81029654e-01 6.85512543e-01
6.71790659e-01 -4.17647511e-01 9.85849142e-01 -9.79808986e-01
4.49197769e-01 -5.78377843e-01 -6.09486699e-01 5.65161288e-01
-5.43077409e-01 1.90426797e-01 5.05397320e-01 -4.65005100e-01
-1.41035628e+00 -3.65233779e-01 8.41308609e-02 -6.52821288e-02
-2.28794053e-01 6.67174101e-01 -2.24577948e-01 3.53001773e-01
5.91863573e-01 -1.15227468e-01 -4.81636137e-01 -7.45404482e-01
7.31102347e-01 7.67669976e-01 3.79779965e-01 -1.05079198e+00
4.45282966e-01 1.79374777e-03 -7.66313672e-01 -7.74027348e-01
-1.22196066e+00 -5.15325487e-01 -8.11975241e-01 -8.44708737e-03
1.06270146e+00 -1.63574409e+00 -5.52319400e-02 3.15401852e-01
-1.08960044e+00 -7.33473778e-01 -5.66456653e-02 8.68307054e-01
-1.37132779e-01 -2.31900021e-01 -8.65661681e-01 -4.40191150e-01
-5.76792121e-01 -1.18714809e+00 1.28249335e+00 -2.67263114e-01
-1.82017490e-01 -1.22520673e+00 3.22721094e-01 3.64217281e-01
3.02370578e-01 -4.22802448e-01 1.12297368e+00 -1.03875899e+00
-2.29626730e-01 4.97257002e-02 -3.30707818e-01 4.79706287e-01
1.90886874e-02 -2.80838281e-01 -1.05662262e+00 -6.11019790e-01
-3.21642518e-01 -8.44309568e-01 7.73431420e-01 2.40966246e-01
2.77496725e-01 -8.59928504e-03 -1.48955271e-01 3.59918416e-01
1.54490161e+00 -9.66727212e-02 1.51260763e-01 2.23931193e-01
7.07593918e-01 1.03063095e+00 6.99514627e-01 -4.01027977e-01
6.60351634e-01 6.33744299e-01 -5.60198128e-01 -1.13461971e-01
-5.05251467e-01 -5.74900925e-01 1.14863253e+00 1.53615594e+00
6.27750218e-01 -5.43158203e-02 -1.27085114e+00 7.47701824e-01
-1.43548596e+00 -4.14018989e-01 6.30120859e-02 2.21706295e+00
1.34165680e+00 3.06106522e-03 -2.14699894e-01 -6.79876864e-01
6.58300579e-01 -1.46759033e-01 -3.22650939e-01 -4.12436932e-01
-3.82142305e-01 3.54317278e-01 4.26980317e-01 6.92558229e-01
-9.68339920e-01 1.77062547e+00 5.61110210e+00 1.11322784e+00
-1.14200890e+00 5.34126341e-01 7.12609947e-01 -1.84398979e-01
-3.39605004e-01 -1.76357478e-02 -1.51585364e+00 4.35843647e-01
1.40929615e+00 -2.23183110e-01 2.53956497e-01 7.58950472e-01
-1.48384869e-01 -6.82550445e-02 -1.21825314e+00 6.15964293e-01
3.19336802e-01 -8.45758617e-01 4.53360379e-02 -1.24710619e-01
9.71029401e-01 9.90337610e-01 2.87630111e-01 9.26560402e-01
7.76130438e-01 -6.34192646e-01 6.80198908e-01 6.46354854e-02
1.28164005e+00 -7.55267620e-01 6.55254066e-01 3.60232532e-01
-1.02801108e+00 4.12866175e-01 -2.68240482e-01 1.91954672e-01
1.54823244e-01 2.42500797e-01 -1.06024516e+00 3.18897337e-01
8.53395224e-01 7.77597368e-01 -9.14882720e-01 3.64118099e-01
-4.36002314e-01 8.55020583e-01 -3.94354939e-01 1.38918251e-01
3.78797740e-01 1.48703288e-02 2.22966745e-01 1.70358908e+00
9.36457291e-02 -3.72983307e-01 2.50989050e-01 4.15641367e-01
-4.81589258e-01 8.60614777e-01 -7.63542295e-01 -1.56725302e-01
4.54561144e-01 1.17834270e+00 -5.22728801e-01 -4.98177558e-01
-7.40752757e-01 7.50605106e-01 9.65604484e-01 2.54355907e-01
-4.29995596e-01 -2.01724082e-01 4.64712471e-01 -1.21379316e-01
1.92136050e-03 -1.44299522e-01 3.01177781e-02 -1.32006419e+00
-1.29139468e-01 -9.26509619e-01 5.35491347e-01 -7.18799353e-01
-1.77817357e+00 6.12852097e-01 -9.00102034e-02 -8.13443959e-01
-3.42285424e-01 -7.20716357e-01 -7.63407946e-02 1.30223191e+00
-1.75142074e+00 -1.59579527e+00 5.64401388e-01 6.50594831e-01
8.33932161e-01 -4.09938723e-01 8.22867155e-01 5.64654291e-01
-6.22518659e-01 1.01207197e+00 8.56857225e-02 1.86202422e-01
1.52157581e+00 -1.06989515e+00 3.08563173e-01 1.03386998e+00
2.36674592e-01 8.32685649e-01 1.32091314e-01 -7.59249449e-01
-1.03978205e+00 -1.38022614e+00 1.35596967e+00 -6.60520732e-01
1.05619526e+00 -7.29791284e-01 -1.10291100e+00 1.05980217e+00
6.39795303e-01 -3.28521043e-01 9.81311679e-01 6.48896277e-01
-9.27042365e-01 -1.22635365e-01 -8.30259323e-01 6.40946686e-01
7.05849826e-01 -1.19486678e+00 -6.00111246e-01 4.85186905e-01
1.21227145e+00 1.29494116e-01 -1.01306057e+00 3.62893939e-01
3.88017237e-01 -3.70759100e-01 6.28156185e-01 -7.01632977e-01
4.54692692e-01 1.93275392e-01 -5.48057914e-01 -1.42583740e+00
-9.24045742e-02 -1.90997526e-01 7.33934760e-01 1.84497023e+00
8.32866430e-01 -7.12077677e-01 -5.90617731e-02 3.48488539e-01
-3.39006543e-01 -4.95151728e-01 -8.21347713e-01 -8.43978584e-01
8.56251061e-01 -6.60900593e-01 1.31108850e-01 1.32916868e+00
2.32298393e-02 1.10875559e+00 -1.42509416e-02 -1.29123433e-02
3.63920510e-01 -1.54034644e-01 5.07320821e-01 -9.47493076e-01
-2.77045906e-01 -3.18671823e-01 1.54968917e-01 -6.57548428e-01
8.23987424e-01 -1.70592380e+00 1.38987824e-01 -1.11787236e+00
4.47859973e-01 -7.54104555e-01 -4.70030129e-01 7.73386478e-01
-3.66042763e-01 1.73194081e-01 2.11842075e-01 1.65441498e-01
-7.26988614e-01 4.62311387e-01 6.83435559e-01 2.12692201e-01
-2.22714424e-01 -4.71278995e-01 -7.78472960e-01 7.43042886e-01
6.52290225e-01 -5.92403173e-01 -2.24173620e-01 -1.04925144e+00
1.85392693e-01 -3.11030328e-01 -1.65527031e-01 -5.31274199e-01
-6.80077001e-02 2.81141371e-01 1.92134827e-01 -5.67348599e-01
1.07922584e-01 -4.08198178e-01 -5.12287319e-01 2.54678726e-01
-6.94261968e-01 2.87915051e-01 6.38916075e-01 8.40787813e-02
-5.12118079e-02 -1.46004692e-01 7.60077178e-01 -9.27750468e-02
-7.29041815e-01 6.37916476e-02 -2.38042966e-01 5.73863268e-01
5.07945240e-01 6.39613330e-01 -5.96257389e-01 -5.56276515e-02
-3.30749452e-01 3.25328171e-01 4.56622511e-01 9.18327272e-01
-2.61446256e-02 -1.30978048e+00 -1.08741724e+00 2.09942013e-01
4.84777063e-01 -5.16525149e-01 4.45654131e-02 7.43434310e-01
-4.97133248e-02 8.78244579e-01 -2.21293345e-02 -6.09699547e-01
-9.94555414e-01 4.37159628e-01 8.43135417e-02 -4.98650730e-01
-1.38478473e-01 1.02315736e+00 5.07559001e-01 -9.65528250e-01
3.38038951e-02 5.34106679e-02 4.09282409e-02 2.46045917e-01
3.81452858e-01 2.12060392e-01 8.53111967e-02 -9.17936385e-01
-4.08014238e-01 5.24880230e-01 -6.87557578e-01 -6.79420948e-01
1.18603027e+00 -1.47286505e-01 -2.63164073e-01 7.17573345e-01
1.50980914e+00 4.23050582e-01 -7.78154731e-01 -6.17394090e-01
4.08875942e-01 9.42876935e-02 8.17842335e-02 -1.09685886e+00
-7.61194766e-01 9.12274301e-01 4.91322994e-01 -4.91126776e-01
7.07597852e-01 3.53668004e-01 5.87606370e-01 2.92875260e-01
5.21248639e-01 -1.28107059e+00 -9.88978520e-02 8.72516394e-01
8.51897955e-01 -1.48452950e+00 -2.40351617e-01 -1.09045133e-01
-8.47324431e-01 2.98298150e-01 7.79459178e-01 2.19988842e-02
6.29641116e-01 2.07030848e-01 2.64455289e-01 5.29890358e-02
-1.07055438e+00 -2.87717253e-01 4.54671741e-01 3.87727350e-01
1.12481868e+00 2.29961947e-01 -4.60914314e-01 5.36703348e-01
-4.05494332e-01 -5.59990287e-01 -6.97513223e-02 4.96610641e-01
-1.46329865e-01 -1.24859488e+00 -1.00652575e-01 6.94553927e-02
-7.68492639e-01 -8.35470736e-01 -2.04590470e-01 9.40927625e-01
1.38374895e-01 8.88666391e-01 3.01129013e-01 9.94805172e-02
1.37653843e-01 5.24426162e-01 3.28580111e-01 -1.08319044e+00
-6.83889985e-01 4.52325761e-01 2.68643379e-01 -3.54433805e-01
-2.22011864e-01 -9.24078166e-01 -1.00627089e+00 1.72353446e-01
-2.62977034e-01 1.35877639e-01 6.95808530e-01 8.10164809e-01
4.49735612e-01 1.18319310e-01 3.25517446e-01 -3.53505820e-01
-3.13239962e-01 -1.29446805e+00 -2.73276299e-01 3.16502273e-01
-4.19170707e-02 -3.73551399e-01 -2.14485496e-01 1.49311414e-02] | [10.923150062561035, 9.91702651977539] |
168ee490-bfee-4411-8df9-768fcaadbfa8 | ai-safety-gridworlds | 1711.09883 | null | http://arxiv.org/abs/1711.09883v2 | http://arxiv.org/pdf/1711.09883v2.pdf | AI Safety Gridworlds | We present a suite of reinforcement learning environments illustrating
various safety properties of intelligent agents. These problems include safe
interruptibility, avoiding side effects, absent supervisor, reward gaming, safe
exploration, as well as robustness to self-modification, distributional shift,
and adversaries. To measure compliance with the intended safe behavior, we
equip each environment with a performance function that is hidden from the
agent. This allows us to categorize AI safety problems into robustness and
specification problems, depending on whether the performance function
corresponds to the observed reward function. We evaluate A2C and Rainbow, two
recent deep reinforcement learning agents, on our environments and show that
they are not able to solve them satisfactorily. | ['Jan Leike', 'Andrew Lefrancq', 'Victoria Krakovna', 'Tom Everitt', 'Shane Legg', 'Miljan Martic', 'Laurent Orseau', 'Pedro A. Ortega'] | 2017-11-27 | null | null | null | null | ['safe-exploration'] | ['robots'] | [-1.85857803e-01 4.64692056e-01 -1.39006332e-01 6.23720605e-03
-2.92330176e-01 -1.13623405e+00 7.56759107e-01 1.30860746e-01
-6.65471315e-01 9.89310861e-01 -1.01679087e-01 -5.33920646e-01
-1.66222259e-01 -9.01969910e-01 -8.51152062e-01 -7.35251963e-01
-8.68801892e-01 3.39674622e-01 4.23118651e-01 -5.15158415e-01
2.13888027e-02 3.78327549e-01 -1.48876858e+00 -3.29054058e-01
6.73053145e-01 5.91663778e-01 -6.36883855e-01 9.43005085e-01
7.00718224e-01 1.16642988e+00 -8.62482309e-01 -1.45089746e-01
4.57608998e-01 -2.15766400e-01 -8.52459013e-01 -3.13696533e-01
-1.66982144e-01 -9.02122557e-01 -3.71910632e-01 1.20560598e+00
1.66874766e-01 1.31370753e-01 3.95788372e-01 -2.20531559e+00
-5.29861331e-01 1.10451949e+00 -1.83473736e-01 -1.32748067e-01
5.07513285e-01 8.88206124e-01 8.05899739e-01 6.33627951e-01
3.67393345e-01 1.21077943e+00 3.27174366e-01 1.11459577e+00
-1.20720732e+00 -5.71042836e-01 1.10416792e-01 -2.31696233e-01
-9.98329401e-01 -3.33983779e-01 6.91743940e-02 -1.99823737e-01
1.25137103e+00 3.04545015e-01 5.26561558e-01 1.46539831e+00
5.49160182e-01 5.57320774e-01 1.08804190e+00 -1.52514383e-01
8.75305712e-01 5.15392348e-02 5.21071553e-02 5.86897016e-01
2.38073409e-01 1.31196499e+00 5.54087274e-02 -5.70618808e-01
4.14602816e-01 -4.13729131e-01 -5.45316562e-02 -5.06981432e-01
-9.47775662e-01 7.18135536e-01 3.10461633e-02 -1.74571112e-01
-2.92224705e-01 7.54627109e-01 7.45158255e-01 7.84185648e-01
-5.47678828e-01 8.96531403e-01 -4.07282233e-01 -3.79558504e-01
1.52056232e-01 6.79895401e-01 9.48428810e-01 1.05563688e+00
3.47613633e-01 5.20581007e-01 -1.90049842e-01 -1.54910177e-01
2.69677430e-01 7.82323420e-01 4.38051045e-01 -1.51178050e+00
-8.91696066e-02 2.34720454e-01 4.68789250e-01 -3.95391881e-01
-6.44218266e-01 -1.51065424e-01 -2.25275621e-01 9.94955003e-01
1.96438804e-01 -7.63097703e-01 -5.46373367e-01 2.31783366e+00
2.37107098e-01 1.67623043e-01 7.31303811e-01 5.48995912e-01
1.10173307e-01 4.36414242e-01 1.97943956e-01 -3.53663415e-01
1.01625764e+00 -5.59555411e-01 -4.45261419e-01 -1.88164145e-01
7.46234059e-01 1.91135764e-01 8.32108915e-01 5.24698734e-01
-1.10760510e+00 1.28645748e-01 -1.34034622e+00 5.28328300e-01
-1.32767439e-01 -9.15849805e-01 5.36595643e-01 8.74512136e-01
-1.26152265e+00 6.02742970e-01 -1.08923566e+00 -4.28063907e-02
1.18177734e-01 7.42502987e-01 -2.71016777e-01 7.16041505e-01
-1.42737091e+00 1.01123714e+00 6.37567461e-01 -6.56635642e-01
-1.82544708e+00 -2.40675911e-01 -1.03540874e+00 4.87793162e-02
7.01274693e-01 -5.18948555e-01 1.67780185e+00 -8.76499891e-01
-1.87662578e+00 4.93082136e-01 9.30563867e-01 -1.04711187e+00
6.15175188e-01 1.11601882e-01 -4.23709422e-01 -3.80988582e-04
4.35573794e-02 7.71718919e-01 8.41636360e-01 -1.06080484e+00
-9.69036162e-01 -1.89582273e-01 8.02398801e-01 2.42210642e-01
-1.69318691e-01 -3.16744263e-04 5.08582056e-01 -1.84470475e-01
-1.13184071e+00 -1.32211840e+00 -2.92829275e-01 -4.04229045e-01
-3.30854833e-01 -1.85828879e-01 6.32598102e-01 3.63278016e-02
6.30730033e-01 -2.42385244e+00 -2.40821451e-01 3.06645870e-01
8.43413994e-02 2.49911915e-03 -6.50037646e-01 2.52281070e-01
-9.16317254e-02 2.14440599e-01 -7.53307045e-02 3.53903621e-01
7.68420756e-01 4.15301681e-01 -4.44422901e-01 9.14640427e-01
5.28102145e-02 8.25672448e-01 -1.06778038e+00 -1.49821669e-01
-5.14545254e-02 -6.82104453e-02 -7.89979935e-01 5.19819379e-01
-5.55615187e-01 1.21761516e-01 -4.43071067e-01 1.74111918e-01
2.81481177e-01 5.21327257e-01 2.45765164e-01 8.04147482e-01
-2.59259373e-01 2.88516164e-01 -1.10549581e+00 8.63640368e-01
-9.15883929e-02 1.32929772e-01 3.11378479e-01 -5.63418686e-01
5.62760711e-01 4.34908330e-01 4.09052461e-01 -9.15790617e-01
2.17815652e-01 -5.62870177e-03 3.26144725e-01 -2.59494781e-01
4.86473739e-01 -1.61106944e-01 -5.81594288e-01 8.61446440e-01
-1.80100337e-01 -1.57715678e-01 1.43396959e-01 7.37104639e-02
1.52390599e+00 1.39134809e-01 3.30617219e-01 -6.50425434e-01
1.57992005e-01 4.12463956e-02 6.34975076e-01 1.30686772e+00
-7.18527138e-01 -2.56636053e-01 1.01393056e+00 -2.90622473e-01
-1.00121295e+00 -1.07629871e+00 2.20916852e-01 1.05397964e+00
3.80361766e-01 -3.25468570e-01 -9.96149838e-01 -9.35751855e-01
2.30982393e-01 1.13759053e+00 -8.46487164e-01 -9.07447696e-01
-4.48555976e-01 -1.10558823e-01 1.28798175e+00 5.15698731e-01
4.19853777e-01 -1.28494334e+00 -1.60156631e+00 2.44706962e-03
4.79654461e-01 -4.84067172e-01 -6.30821764e-01 5.96494973e-01
-2.60059178e-01 -1.07839084e+00 2.19495133e-01 -5.48095584e-01
2.64220774e-01 -9.28417873e-03 8.13711882e-01 2.69677758e-01
7.49816597e-02 8.46985519e-01 -1.20338358e-01 -2.82055438e-01
-9.35773969e-01 -2.00821340e-01 5.65092564e-01 -7.97100246e-01
3.84715758e-02 -3.54058862e-01 -2.20551461e-01 1.90253377e-01
-1.12874508e+00 -5.66950023e-01 -7.59479553e-02 7.70233512e-01
-2.33446565e-02 4.41643834e-01 5.52950323e-01 -4.07167077e-01
1.11247826e+00 -4.42314267e-01 -1.31433427e+00 3.43361288e-01
-5.27287006e-01 4.67651576e-01 1.11096668e+00 -5.10690689e-01
-7.06269860e-01 -6.40368089e-02 2.28944980e-02 6.47616684e-02
-4.54528034e-01 -3.31484415e-02 -3.65159303e-01 -2.93237325e-02
9.40092564e-01 1.28760129e-01 4.42239612e-01 4.61589754e-01
4.21703905e-01 4.54115808e-01 5.42743981e-01 -1.07510531e+00
8.42956007e-01 -4.75469828e-02 6.19595908e-02 -3.12376440e-01
1.58930551e-02 4.39868331e-01 2.23430544e-01 -1.64184161e-02
6.01659715e-01 -6.39199674e-01 -1.91224360e+00 4.18851048e-01
-6.60327733e-01 -8.82808089e-01 -6.87679172e-01 9.66354311e-02
-1.17169058e+00 3.95017803e-01 -6.21997476e-01 -1.05696774e+00
-2.79848456e-01 -1.57794321e+00 5.40897071e-01 3.84035856e-01
-2.96894848e-01 -6.85534477e-01 3.90112787e-01 -4.70921129e-01
5.26902497e-01 3.21042776e-01 9.67343867e-01 -1.05164814e+00
-3.16472828e-01 1.03134930e-01 6.43482387e-01 1.15395017e-01
-7.37452134e-02 1.73121631e-01 -5.89124382e-01 -6.06385410e-01
7.19767958e-02 -8.32450747e-01 1.20052449e-01 1.37020707e-01
8.24706614e-01 -1.09984529e+00 -5.11379093e-02 4.01140779e-01
9.95562553e-01 8.75474930e-01 5.80861807e-01 1.00079560e+00
-1.67508230e-01 3.68573666e-01 6.19906723e-01 7.27079749e-01
2.34306306e-01 4.63886082e-01 1.06282222e+00 4.86996025e-01
7.01773942e-01 -1.57503426e-01 1.00088215e+00 -5.56954920e-01
3.23421955e-01 -1.78622901e-01 -7.17737317e-01 2.79872388e-01
-1.99030077e+00 -1.21501887e+00 5.97476125e-01 2.27930140e+00
1.05209327e+00 5.11706352e-01 7.90953815e-01 -1.73628107e-01
3.71427000e-01 -2.57159114e-01 -9.36517119e-01 -1.07711148e+00
9.26628113e-02 -1.29138872e-01 8.67081821e-01 7.39923716e-01
-1.05523944e+00 9.73608911e-01 7.31847143e+00 3.44467729e-01
-8.47760618e-01 -2.71309614e-01 3.79107058e-01 -1.10044025e-01
-4.26089793e-01 -2.94240117e-02 -3.82465303e-01 3.16880375e-01
1.14097929e+00 -7.03355908e-01 1.00671637e+00 1.17812955e+00
-1.37467206e-01 1.26436539e-02 -1.48039508e+00 1.55740559e-01
-4.37427670e-01 -7.47070432e-01 -4.74206507e-01 -4.51711789e-02
4.89793807e-01 -8.32327176e-03 3.60616356e-01 8.29435468e-01
1.40174472e+00 -1.26188064e+00 9.32085454e-01 -1.61597729e-01
4.95376587e-01 -1.24210882e+00 4.68836457e-01 3.78480732e-01
-5.80745816e-01 -4.58188057e-01 3.30983512e-02 -3.11030686e-01
-3.03703189e-01 -5.76511145e-01 -6.88953996e-01 7.07751736e-02
5.98612309e-01 1.29881293e-01 -3.06550294e-01 7.25896299e-01
-3.24809045e-01 2.03073889e-01 -4.56745088e-01 -2.07976073e-01
4.47185487e-01 -4.17277217e-02 6.22322202e-01 6.16229832e-01
4.71067354e-02 1.26891255e-01 3.60988677e-01 9.34153914e-01
2.07330704e-01 -7.69470572e-01 -8.52892458e-01 -8.01329836e-02
5.72619319e-01 9.07986820e-01 -3.57391715e-01 -1.38948530e-01
6.96543157e-02 4.94719684e-01 2.58002758e-01 2.75052875e-01
-1.23527324e+00 -4.18231457e-01 1.23224652e+00 -5.11150956e-01
-1.70822114e-01 -1.53620720e-01 -8.06659833e-02 -7.91550100e-01
-4.12000626e-01 -1.55071449e+00 6.49451077e-01 -3.77393901e-01
-1.00850904e+00 6.14752054e-01 9.34040472e-02 -8.52372229e-01
-6.83895826e-01 -3.94750178e-01 -5.53405881e-01 2.56207019e-01
-1.06001878e+00 -4.57173258e-01 2.52090752e-01 8.02663088e-01
-1.71233132e-01 -3.55626196e-01 1.03169107e+00 -3.18614513e-01
-7.01087117e-01 9.68136013e-01 -5.19566946e-02 -1.40880272e-01
4.16062325e-01 -1.48778498e+00 5.37308872e-01 8.58481765e-01
-5.38906395e-01 5.75328350e-01 1.12398243e+00 -4.67338115e-01
-1.77463627e+00 -1.06735635e+00 -1.99111193e-01 -4.45987731e-01
8.67871106e-01 -1.14677638e-01 -4.73382980e-01 1.02182889e+00
4.75026995e-01 -1.54080153e-01 1.84699848e-01 -1.23347297e-01
-6.11099660e-01 -4.92218025e-02 -1.41132379e+00 1.23387694e+00
9.66218174e-01 -3.11578602e-01 -6.65993690e-01 4.97375429e-02
1.05604219e+00 -3.94656867e-01 -7.68312454e-01 2.90434122e-01
3.52646768e-01 -8.79396856e-01 6.05004072e-01 -1.03900719e+00
1.76836580e-01 -3.89460236e-01 1.93619616e-02 -1.55409575e+00
-4.67800796e-01 -1.19157267e+00 -1.47133470e-01 8.45434308e-01
2.60645360e-01 -8.79818678e-01 5.66782951e-01 1.15375996e+00
-1.30229801e-01 9.30730179e-02 -1.09646213e+00 -1.32193828e+00
5.15587747e-01 -1.31017774e-01 1.00972283e+00 7.85123885e-01
9.69750345e-01 1.00000232e-01 -4.74049717e-01 3.49242359e-01
6.46738827e-01 -8.64333063e-02 5.76581597e-01 -5.71745813e-01
-5.66808522e-01 -8.49898160e-01 -1.47084802e-01 -3.92721593e-01
6.38231754e-01 -5.55853486e-01 4.71155703e-01 -8.43456984e-01
-5.86930998e-02 -3.91184121e-01 -1.34188518e-01 1.07048178e+00
2.33674973e-01 -3.60664904e-01 1.45258754e-01 -3.43587875e-01
-1.06402373e+00 6.33928180e-01 7.35627115e-01 -3.32423866e-01
-3.95689696e-01 -8.88424367e-02 -8.06667149e-01 4.29998100e-01
1.15374756e+00 -3.66090238e-01 -5.52987039e-01 -2.05592275e-01
4.21065062e-01 3.57267678e-01 2.91980475e-01 -8.21165919e-01
8.47557336e-02 -9.98940945e-01 -3.14059556e-01 4.27943140e-01
-1.72844604e-01 -1.03617167e+00 2.98285663e-01 1.16936505e+00
-9.71195936e-01 3.88346523e-01 4.09363687e-01 2.70575374e-01
2.80682862e-01 -1.82280064e-01 8.95888269e-01 5.73849566e-02
-5.60991287e-01 2.08277181e-01 -9.10151064e-01 1.21033750e-01
1.61415255e+00 1.99554980e-01 -8.99228394e-01 -5.34812808e-01
-2.40064397e-01 8.47182333e-01 8.96913528e-01 4.56648260e-01
4.77686077e-01 -1.08894157e+00 -4.92819637e-01 2.86638618e-01
3.27196836e-01 -2.79272348e-01 -3.44044045e-02 3.61218929e-01
-2.76170552e-01 6.32337704e-02 -7.75304258e-01 2.56091096e-02
-1.08647156e+00 1.10980844e+00 7.85747707e-01 4.01850417e-02
-5.73339164e-01 3.13301265e-01 2.14016214e-01 -4.44657207e-01
6.27035260e-01 -2.61231691e-01 -8.72170106e-02 -7.28615701e-01
6.53547525e-01 2.95973629e-01 -2.72957951e-01 -1.28606156e-01
-6.22485578e-01 -1.91015333e-01 -8.59332010e-02 -4.81786430e-01
9.44473684e-01 1.69584081e-01 3.02380901e-02 -2.04539984e-01
6.09163940e-01 -3.14556807e-01 -1.52076781e+00 3.14719200e-01
5.19473851e-03 -4.27649468e-02 -2.06365600e-01 -9.04172897e-01
-7.22173035e-01 2.17289180e-01 5.21838009e-01 8.04708183e-01
9.72243905e-01 -4.46486384e-01 3.49230379e-01 6.94793224e-01
6.57987058e-01 -1.32157946e+00 -5.91947660e-02 9.02441025e-01
7.06532240e-01 -8.30555558e-01 -4.01643753e-01 4.73867118e-01
-9.19474900e-01 8.08242917e-01 1.13525414e+00 -2.76865840e-01
2.35179961e-01 8.51077378e-01 -1.74625918e-01 2.11463138e-01
-1.20354068e+00 -1.30322888e-01 -5.78307033e-01 1.12769771e+00
-4.33111370e-01 3.82943153e-01 5.94080836e-02 4.51923847e-01
-3.29646111e-01 -3.05215657e-01 1.17085361e+00 1.17139745e+00
-5.52106380e-01 -1.10596073e+00 -3.75033349e-01 -5.58883511e-02
-4.19382423e-01 2.87976652e-01 -4.43162769e-01 8.24859202e-01
-2.71861613e-01 1.19931638e+00 5.69023266e-02 -5.00204802e-01
3.04325163e-01 -2.76094496e-01 5.09891212e-01 -2.70472109e-01
-8.59894574e-01 -3.21595341e-01 2.66655296e-01 -7.64642596e-01
-7.75473379e-03 -4.90912497e-01 -1.68485200e+00 -5.70803344e-01
1.11728357e-02 3.26848805e-01 2.06064910e-01 8.53656650e-01
1.59366891e-01 4.18192208e-01 8.98498297e-01 -1.06348179e-01
-1.52964389e+00 -2.93857396e-01 -6.33097529e-01 3.79459083e-01
8.99274290e-01 -3.41001421e-01 -4.49880272e-01 -4.50225592e-01] | [4.363470077514648, 2.1012823581695557] |
73447b2b-08b8-435d-b58a-8b89377454d8 | low-resource-adaptation-for-personalized-co | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Ahuja_Low-Resource_Adaptation_for_Personalized_Co-Speech_Gesture_Generation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Ahuja_Low-Resource_Adaptation_for_Personalized_Co-Speech_Gesture_Generation_CVPR_2022_paper.pdf | Low-Resource Adaptation for Personalized Co-Speech Gesture Generation | Personalizing an avatar for co-speech gesture generation from spoken language requires learning the idiosyncrasies of a person's gesture style from a small amount of data. Previous methods in gesture generation require large amounts of data for each speaker, which is often infeasible. We propose an approach, named DiffGAN, that efficiently personalizes co-speech gesture generation models of a high-resource source speaker to target speaker with just 2 minutes of target training data. A unique characteristic of DiffGAN is its ability to account for the crossmodal grounding shift, while also addressing the distribution shift in the output domain. We substantiate the effectiveness of our approach a large scale publicly available dataset through quantitative, qualitative and user studies, which show that our proposed methodology significantly outperforms prior approaches for low-resource adaptation of gesture generation. Code and videos can be found at https://chahuja.com/diffgan | ['Louis-Philippe Morency', 'Dong Won Lee', 'Chaitanya Ahuja'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['gesture-generation'] | ['robots'] | [ 1.32061049e-01 2.18935922e-01 -1.47969455e-01 -4.59734321e-01
-1.08319545e+00 -8.12987804e-01 9.59623933e-01 -6.90083802e-01
-2.96050310e-01 4.24935967e-01 7.93426991e-01 9.87584665e-02
4.35610592e-01 -3.89451087e-01 -5.77261388e-01 -5.12241781e-01
1.73881352e-01 7.02793121e-01 -2.24719808e-01 -1.92064002e-01
-2.01076403e-01 1.86361820e-01 -1.31215167e+00 2.43849054e-01
3.82523119e-01 6.07947707e-01 2.97341011e-02 1.04352343e+00
-1.75736427e-01 3.42503935e-01 -8.03798497e-01 -2.69998491e-01
4.06065136e-01 -9.71293151e-01 -8.07121754e-01 1.13095557e-02
4.80151445e-01 -5.85420966e-01 -2.02815488e-01 6.70168996e-01
9.94973242e-01 4.57099587e-01 6.86646640e-01 -1.38438797e+00
-7.03498483e-01 9.78769839e-01 -5.89721538e-02 -4.80853021e-02
5.87172866e-01 5.88154137e-01 9.63172019e-01 -7.67236590e-01
8.05716276e-01 1.34088480e+00 3.81172985e-01 1.43256414e+00
-1.11508417e+00 -9.76367414e-01 1.40021250e-01 -3.01074892e-01
-1.37271130e+00 -8.47461998e-01 7.46280968e-01 -2.59376407e-01
7.71442592e-01 5.21889508e-01 7.58081794e-01 1.96395242e+00
-8.16386044e-01 1.06976497e+00 9.50386107e-01 -3.89070094e-01
2.43070960e-01 -3.71683612e-02 -1.92940995e-01 4.49333578e-01
-3.34072948e-01 2.92287260e-01 -9.64010358e-01 -1.82117656e-01
8.11931491e-01 -3.60892832e-01 -1.21663742e-01 -1.69137791e-02
-1.26903582e+00 6.10014617e-01 1.85146611e-02 3.12823683e-01
-1.76722333e-01 3.38123173e-01 2.71855325e-01 2.48305902e-01
2.65700847e-01 1.47286490e-01 -4.12379295e-01 -8.25105906e-01
-7.41652369e-01 5.27390003e-01 9.44284081e-01 1.38437223e+00
2.74080008e-01 1.52562842e-01 -3.77465844e-01 6.69356585e-01
4.83242512e-01 1.04284179e+00 9.32315707e-01 -8.28943193e-01
5.66192150e-01 3.05567414e-01 1.35749029e-02 -4.06799495e-01
-2.05873877e-01 1.42362162e-01 -4.80860740e-01 1.68580469e-02
8.40058684e-01 -5.89579582e-01 -9.45326686e-01 2.04955363e+00
4.03402597e-01 2.67776072e-01 2.47939944e-01 1.00183094e+00
1.03832114e+00 4.27820504e-01 3.71507913e-01 1.04584441e-01
1.20989931e+00 -8.24205160e-01 -4.20833528e-01 -2.68813688e-02
3.63786638e-01 -6.70621753e-01 1.69482172e+00 -5.36571890e-02
-9.88101125e-01 -3.34296227e-01 -5.07326663e-01 3.99896428e-02
-5.45211434e-02 1.88127384e-01 7.51952767e-01 9.95583475e-01
-9.10942256e-01 1.65885776e-01 -1.00240850e+00 -7.54641593e-01
3.91440302e-01 4.15285110e-01 -2.41554067e-01 3.34155947e-01
-8.94800007e-01 4.88838166e-01 1.25065520e-01 -1.14106394e-01
-8.73129249e-01 -7.80981600e-01 -7.25266278e-01 -1.30419701e-01
2.57940501e-01 -5.36323249e-01 1.76469183e+00 -1.32295072e+00
-2.35030150e+00 5.56863606e-01 -9.34925377e-02 -1.91467121e-01
7.44213998e-01 -2.62565106e-01 -2.99346119e-01 -9.27136540e-02
-3.43369603e-01 9.36365366e-01 9.17051971e-01 -1.10705709e+00
-5.68630040e-01 -2.48145685e-01 -9.85453576e-02 5.74204803e-01
-3.65745336e-01 1.20616987e-01 -4.35371578e-01 -8.23770344e-01
-3.48352045e-01 -1.10917079e+00 2.75263805e-02 -3.66572142e-01
-3.08534563e-01 -3.06835830e-01 8.52314770e-01 -7.34481752e-01
9.88745034e-01 -2.21964812e+00 3.09043050e-01 2.89660990e-01
-1.16861895e-01 1.28226593e-01 -5.16213238e-01 3.67117971e-01
3.61281008e-01 1.31054670e-01 -3.79650630e-02 -5.49775898e-01
4.10531551e-01 8.83916393e-02 -3.03991348e-01 1.72818914e-01
-1.39316395e-01 1.15482640e+00 -8.88315678e-01 -2.84516245e-01
-3.04886810e-02 6.28404617e-01 -6.79920793e-01 5.53088307e-01
-4.77514744e-01 9.60364938e-01 -4.17361081e-01 5.88737667e-01
2.00918585e-01 -8.25037211e-02 3.64390820e-01 5.56522682e-02
1.61789328e-01 2.50242919e-01 -1.08336389e+00 2.07474661e+00
-4.37682182e-01 5.57598412e-01 -2.13131625e-02 -2.66688943e-01
6.97699964e-01 6.51100755e-01 4.71083164e-01 -4.64032024e-01
2.34027192e-01 2.22166404e-01 1.27889100e-04 -6.08630419e-01
3.93573314e-01 -3.12917411e-01 -4.14330482e-01 1.06508064e+00
2.66142577e-01 -1.33129686e-01 -3.57616730e-02 9.43704173e-02
8.75993431e-01 3.26712728e-01 1.31492481e-01 1.10926293e-01
-5.33878617e-02 -2.01565012e-01 3.17521006e-01 6.84352517e-01
-2.36365482e-01 6.59543455e-01 2.31717341e-02 -9.62947458e-02
-9.33636010e-01 -1.04055583e+00 5.67223072e-01 1.48509479e+00
-2.95234025e-01 -2.67292738e-01 -1.14335728e+00 -7.86906242e-01
-2.03085393e-01 5.61433792e-01 -5.91459036e-01 1.44960150e-01
-7.77427316e-01 -4.06194508e-01 1.16088998e+00 6.75971329e-01
3.42607319e-01 -1.38886774e+00 -7.26665080e-01 -1.74416095e-01
-3.97818863e-01 -1.02555966e+00 -1.12902403e+00 -3.97356123e-01
-5.49409151e-01 -7.93056369e-01 -9.96951401e-01 -5.55534542e-01
4.95161682e-01 -9.86547396e-02 9.03615296e-01 -3.59309763e-02
-5.54761253e-02 7.25020945e-01 -6.45261467e-01 -4.34149683e-01
-6.78566575e-01 4.54096675e-01 1.83990493e-01 3.25245559e-02
5.27521551e-01 -4.91356969e-01 -3.16661328e-01 1.72860846e-01
-7.91284800e-01 1.45181432e-01 4.40571308e-01 7.01207817e-01
1.24733202e-01 -5.59971690e-01 4.53865230e-01 -7.61000693e-01
7.26861656e-01 -3.69707257e-01 -3.33992630e-01 1.03001378e-01
-3.17737460e-01 5.66743016e-02 1.87600285e-01 -9.38347638e-01
-1.26962650e+00 3.75159264e-01 -7.36894161e-02 -2.17820942e-01
-5.24163425e-01 2.25811273e-01 -4.94073600e-01 3.07146132e-01
6.18553936e-01 2.31017292e-01 2.14234650e-01 -5.48044741e-01
7.06864655e-01 8.86595666e-01 5.62224984e-01 -8.45120609e-01
8.58846009e-01 3.73637468e-01 -4.56950277e-01 -9.60744143e-01
-8.81149545e-02 -2.80630022e-01 -6.76782489e-01 -2.36914009e-01
8.00414383e-01 -7.69143462e-01 -8.30228865e-01 6.92953229e-01
-9.82083917e-01 -1.14668739e+00 -4.90437210e-01 5.78303635e-01
-7.24243522e-01 3.67542207e-02 -3.07827234e-01 -8.18632543e-01
-4.01473463e-01 -9.17595029e-01 1.17437387e+00 3.15022141e-01
-8.35611045e-01 -1.15388072e+00 2.43960723e-01 4.26818222e-01
4.62613761e-01 1.38144568e-01 4.39014703e-01 -9.04875100e-01
-4.65186208e-01 1.04967300e-02 2.69051731e-01 -8.49541463e-03
3.84723186e-01 -7.73034096e-02 -1.04070008e+00 -2.64299959e-01
-6.53435528e-01 -4.67030972e-01 2.73614734e-01 1.73641324e-01
5.31121731e-01 -6.55403316e-01 -9.98372808e-02 6.19493604e-01
7.85961986e-01 9.26756114e-02 2.92920828e-01 -6.37582364e-03
9.90858734e-01 6.11749470e-01 3.59053642e-01 5.90582252e-01
5.83404124e-01 9.98889685e-01 -8.62726644e-02 9.08377841e-02
-5.33995926e-01 -6.42710447e-01 7.49491513e-01 6.07986510e-01
-4.05804902e-01 -4.37188148e-01 -8.15605640e-01 5.52047133e-01
-1.76713383e+00 -1.15497661e+00 4.15222287e-01 1.92366970e+00
1.01355052e+00 -6.71459913e-01 9.01049852e-01 -2.60322183e-01
3.97497714e-01 -2.53476072e-02 -4.60973769e-01 -1.04333766e-01
-7.24268705e-02 3.41945052e-01 3.24853390e-01 6.65767252e-01
-8.89863491e-01 1.32817638e+00 6.77396250e+00 5.55162191e-01
-1.23262727e+00 3.64939302e-01 2.09460378e-01 -5.71534038e-01
-5.08824468e-01 -2.77225465e-01 -7.53258049e-01 5.12953103e-01
1.05059719e+00 -1.82877388e-02 8.04441094e-01 5.80717564e-01
2.81162977e-01 4.82788682e-01 -1.28143811e+00 9.95527089e-01
3.06645125e-01 -1.07720172e+00 1.20994881e-01 2.54804254e-01
6.78773999e-01 -1.14279157e-02 2.58372247e-01 2.31988996e-01
8.60502481e-01 -1.11611426e+00 9.96691287e-01 2.83325642e-01
1.17274296e+00 -4.66265410e-01 2.86688477e-01 2.79820174e-01
-1.03366542e+00 5.55317663e-02 3.97359431e-01 4.28424915e-03
2.92244136e-01 -4.16889459e-01 -1.25076175e+00 4.84848842e-02
5.40125430e-01 2.90582389e-01 -3.84176940e-01 3.09725314e-01
-2.93572187e-01 7.97164023e-01 -5.24132967e-01 -1.74375027e-01
-1.83210021e-03 1.73014432e-01 7.17951357e-01 1.31065547e+00
5.54583251e-01 2.34197736e-01 8.02033320e-02 7.70520329e-01
-4.20334488e-02 1.05131619e-01 -5.12761772e-01 -4.95692611e-01
6.61302090e-01 9.67916131e-01 -4.47790712e-01 -3.02361459e-01
-2.52722949e-01 1.32423222e+00 -2.67609097e-02 5.55607915e-01
-6.25116944e-01 7.18966126e-02 8.67810428e-01 -1.78004026e-01
2.61050999e-01 -2.94514209e-01 -2.65086591e-01 -1.21420264e+00
-6.19209372e-02 -1.43238604e+00 4.90451962e-01 -3.87680143e-01
-1.11095428e+00 5.71548820e-01 1.75437242e-01 -1.12126255e+00
-9.88579690e-01 -3.82141113e-01 -4.97987390e-01 9.05277371e-01
-7.47781396e-01 -1.88029516e+00 -4.56726402e-01 8.30879271e-01
7.59653866e-01 -7.33598530e-01 1.10029459e+00 1.30193904e-01
-2.59925872e-01 1.15940499e+00 -3.73437852e-01 4.53272909e-01
7.59994447e-01 -1.10324395e+00 7.72147000e-01 7.96886265e-01
5.14008164e-01 5.95368147e-01 7.13107646e-01 -6.52499735e-01
-1.46330059e+00 -8.28281164e-01 7.93340981e-01 -8.24737906e-01
4.48504895e-01 -6.42678618e-01 -4.42791641e-01 9.77193892e-01
2.69291580e-01 -1.80711091e-01 9.94553685e-01 1.20774433e-01
-5.14735460e-01 3.84040952e-01 -1.13411283e+00 8.96484554e-01
1.48676658e+00 -5.63587964e-01 -3.22087526e-01 3.03055020e-03
4.91439015e-01 -7.47829974e-01 -6.84402049e-01 -9.28766355e-02
1.08120096e+00 -5.11765242e-01 7.52000332e-01 -7.83853769e-01
1.29467681e-01 -1.57404080e-01 -4.05053824e-01 -1.52373838e+00
1.03868358e-01 -1.08894312e+00 -2.13621169e-01 1.46020961e+00
5.05637825e-01 -4.05916840e-01 9.82536614e-01 9.47306156e-01
2.86426023e-02 -2.13002600e-02 -8.56678307e-01 -7.18892097e-01
1.57057315e-01 -5.22619843e-01 1.19699228e+00 8.43058705e-01
1.68113202e-01 3.73649508e-01 -4.14827198e-01 1.93652585e-01
4.76914495e-01 -1.55783266e-01 1.43955040e+00 -9.13639426e-01
-6.87253594e-01 -5.97194374e-01 -1.45714581e-01 -1.03770590e+00
4.51704025e-01 -9.76494312e-01 2.30207518e-01 -1.10750914e+00
8.04196671e-02 -5.93563139e-01 1.75423667e-01 6.87185228e-01
-2.27105450e-02 4.35606062e-01 4.60438639e-01 3.40016872e-01
-3.73356134e-01 5.57043195e-01 1.16493142e+00 -9.80404913e-02
-6.34870946e-01 1.26723722e-01 -6.84109330e-01 5.65462112e-01
8.31067502e-01 -2.57444948e-01 -5.04196167e-01 -5.87119460e-01
-1.55185491e-01 -1.98841497e-01 3.31503451e-01 -6.64634526e-01
-3.49012688e-02 -5.22184372e-01 6.29541874e-02 1.05473772e-01
3.90413582e-01 -6.67304873e-01 3.03290457e-01 1.30411536e-01
-5.88254392e-01 -4.92164902e-02 1.84401050e-01 3.53284240e-01
1.52542233e-01 1.32076666e-01 3.36468548e-01 -1.36955410e-01
-7.71339297e-01 2.39327729e-01 -3.17211956e-01 1.78542465e-01
7.72431254e-01 -7.53304288e-02 -2.22939879e-01 -8.29843402e-01
-7.32484400e-01 -1.00781947e-01 4.64175195e-01 7.59631574e-01
4.17718679e-01 -1.56388569e+00 -7.46617138e-01 4.23572928e-01
1.59171909e-01 -2.02800438e-01 1.30468324e-01 5.16625464e-01
-2.29092807e-01 2.01902047e-01 -3.15480381e-01 -3.48706096e-01
-1.63942075e+00 -4.03870083e-02 1.72686577e-01 1.38723567e-01
-6.85137451e-01 1.05092204e+00 -7.05419481e-02 -6.88915730e-01
1.96724236e-01 -2.90389150e-01 1.13631018e-01 -1.12279631e-01
6.63551271e-01 3.92089188e-01 -3.82379919e-01 -1.10180843e+00
-2.81229824e-01 3.04023594e-01 1.94403291e-01 -1.10279775e+00
1.19317496e+00 -1.12362437e-01 4.49283630e-01 6.04644597e-01
8.66913199e-01 2.40573689e-01 -1.69170022e+00 -1.79758608e-01
-2.81717449e-01 -4.20154184e-01 -4.14684027e-01 -9.48774934e-01
-8.70470405e-01 5.63590884e-01 6.62602305e-01 -2.30962619e-01
9.34629679e-01 2.56169856e-01 9.71646011e-01 2.80231386e-01
3.44726592e-01 -1.13407743e+00 2.62723058e-01 4.71799880e-01
8.97630513e-01 -1.25054276e+00 -4.85665917e-01 -3.31052989e-02
-1.08724451e+00 7.54695296e-01 5.49017727e-01 1.94396883e-01
4.63869572e-01 4.36809927e-01 7.44134426e-01 5.11332490e-02
-4.52488333e-01 -1.90275148e-01 3.54667544e-01 1.01760280e+00
5.71330488e-01 5.44402301e-01 1.59973260e-02 8.13011527e-01
-8.21661115e-01 -8.13360438e-02 2.31119245e-01 7.67454624e-01
2.27560997e-01 -1.39905357e+00 -3.44931334e-01 -1.44056857e-01
-2.81394154e-01 -6.30768985e-02 -7.48822391e-01 8.11236918e-01
-8.97383839e-02 8.59106839e-01 1.95350587e-01 -6.23584270e-01
1.91958323e-01 5.17670810e-01 6.74353600e-01 -7.60721445e-01
-6.67542577e-01 1.46343276e-01 6.82844967e-02 -4.92018491e-01
-5.73433101e-01 -1.03322148e+00 -1.46102369e+00 -3.36617321e-01
1.99090526e-01 -1.13874309e-01 6.31767988e-01 9.78793442e-01
4.98020560e-01 2.04573974e-01 1.57545730e-01 -1.14421844e+00
-5.82637489e-01 -1.23782516e+00 -3.66630465e-01 7.37498283e-01
3.00631136e-01 -4.71073598e-01 -2.61438847e-01 5.60047150e-01] | [5.618532657623291, -0.11750481277704239] |
22751fdd-0f4d-4d23-abba-38b14a903c76 | toward-real-flare-removal-a-comprehensive | 2306.15884 | null | https://arxiv.org/abs/2306.15884v1 | https://arxiv.org/pdf/2306.15884v1.pdf | Toward Real Flare Removal: A Comprehensive Pipeline and A New Benchmark | Photographing in the under-illuminated scenes, the presence of complex light sources often leave strong flare artifacts in images, where the intensity, the spectrum, the reflection, and the aberration altogether contribute the deterioration. Besides the image quality, it also influence the performance of down-stream visual applications. Thus, removing the lens flare and ghosts is a challenge issue especially in low-light environment. However, existing methods for flare removal mainly restricted to the problems of inadequate simulation and real-world capture, where the categories of scattered flares are singular and the reflected ghosts are unavailable. Therefore, a comprehensive deterioration procedure is crucial for constructing the dataset of flare removal. Based on the theoretical analysis and real-world evaluation, we propose a well-developed methodology for generating the data-pairs with flare deterioration. The procedure is comprehensive, where the similarity of scattered flares and the symmetric effect of reflected ghosts are realized. Moreover, we also construct a real-shot pipeline that respectively processes the effects of scattering and reflective flares, aiming to directly generate the data for end-to-end methods. Experimental results show that the proposed methodology add diversity to the existing flare datasets and construct a comprehensive mapping procedure for flare data pairs. And our method facilities the data-driven model to realize better restoration in flare images and proposes a better evaluation system based on real shots, resulting promote progress in the area of real flare removal. | ['Yueting Chen', 'Zhihai Xu', 'Huajun Feng', 'Shiqi Chen', 'Zheyan Jin'] | 2023-06-28 | null | null | null | null | ['flare-removal'] | ['computer-vision'] | [ 3.36131722e-01 -7.13486791e-01 6.61297321e-01 -2.07903162e-01
-4.06990945e-01 -7.71095097e-01 5.42159796e-01 -3.41752172e-01
2.29529843e-01 6.63618624e-01 3.29234540e-01 2.19833061e-01
-4.55573380e-01 -8.22912276e-01 -3.57287377e-01 -9.31387365e-01
1.70282610e-02 5.71312420e-02 2.73220837e-01 -5.53194702e-01
1.67191789e-01 6.51293933e-01 -2.15597534e+00 2.03032538e-01
1.21358109e+00 7.86745131e-01 5.24902046e-01 7.18730748e-01
-1.55016005e-01 9.54447091e-01 -7.54177511e-01 -1.32389024e-01
5.04277349e-01 -5.54608345e-01 5.06384075e-02 2.69458264e-01
7.13104784e-01 -6.05882347e-01 -1.50152951e-01 1.24596012e+00
8.76836240e-01 9.38861296e-02 4.72091615e-01 -1.13268995e+00
-5.28035402e-01 -1.17882967e-01 -8.04661393e-01 1.73813269e-01
5.21739781e-01 7.63993084e-01 3.42041790e-01 -7.94142306e-01
4.35079485e-01 1.03565037e+00 9.37120140e-01 -4.66030203e-02
-7.26948977e-01 -4.50710803e-01 -3.29192191e-01 2.88191766e-01
-1.08277714e+00 -6.41276300e-01 8.84455562e-01 -5.22340953e-01
6.42695308e-01 4.34597313e-01 8.00694644e-01 5.74625850e-01
1.31432429e-01 3.28062266e-01 1.33452690e+00 -4.08662558e-01
-2.24057421e-01 -1.33266598e-01 4.18383837e-01 4.25123066e-01
5.14350832e-01 5.89967906e-01 -2.57637948e-01 1.78276837e-01
4.44431573e-01 1.35307595e-01 -9.83991623e-01 -9.06658918e-02
-7.26897359e-01 3.38576555e-01 3.89967203e-01 2.56379932e-01
-2.81737000e-01 -2.22237766e-01 3.87622714e-02 2.28327349e-01
3.51950526e-01 1.56785429e-01 -1.67120472e-01 2.36045659e-01
-1.06589055e+00 3.98632646e-01 3.78093630e-01 7.27067947e-01
9.49910820e-01 3.71312797e-01 -2.84847975e-01 5.51437557e-01
2.25249082e-01 1.23898411e+00 -5.91263408e-03 -6.06953681e-01
1.81000233e-01 6.90943956e-01 4.24877733e-01 -1.27679729e+00
-6.12214208e-01 -6.20496511e-01 -9.89537776e-01 6.92575634e-01
2.28222087e-01 -1.55051947e-01 -8.13881874e-01 1.28301418e+00
1.97264373e-01 5.01909256e-01 2.45207623e-01 1.43281722e+00
1.06334579e+00 6.59455717e-01 -4.88283902e-01 -7.25707293e-01
1.15017343e+00 -9.29220617e-01 -1.45621336e+00 -2.89156824e-01
3.36042762e-01 -1.26348853e+00 1.01816928e+00 5.95873535e-01
-1.24644470e+00 -7.77112961e-01 -9.32324409e-01 2.50859987e-02
-1.64497122e-02 1.94863454e-01 6.55734420e-01 4.51937646e-01
-9.62561011e-01 3.17186713e-01 -1.97764501e-01 -2.14164987e-01
1.97743073e-01 -3.78730834e-01 -6.59173727e-02 -2.37406716e-01
-1.07358778e+00 9.33853745e-01 5.22024892e-02 7.22147405e-01
-8.32929552e-01 -9.20783997e-01 -3.59817773e-01 3.28560546e-02
1.78697586e-01 -8.25675428e-01 7.06603587e-01 -1.12380457e+00
-1.08619094e+00 6.62691355e-01 4.05189209e-02 -2.37689152e-01
5.00774920e-01 -1.73233569e-01 -8.25989723e-01 2.38607213e-01
-1.62011027e-01 -2.16742903e-01 9.62475777e-01 -1.62732017e+00
-7.81517625e-01 -2.90531367e-01 -1.96121354e-03 4.85989481e-01
2.80518234e-01 5.73622771e-02 -1.49462968e-01 -3.52976978e-01
-1.25332355e-01 -4.87478793e-01 6.42293170e-02 -1.19303115e-01
2.57202070e-02 4.93347794e-01 1.00522280e+00 -6.77831292e-01
1.43631768e+00 -2.44826460e+00 -2.59955674e-01 -2.18286261e-01
8.04440081e-02 6.30808890e-01 2.25738622e-02 6.09406888e-01
-1.95293710e-01 -4.10786211e-01 -3.77913445e-01 -1.66914612e-02
-2.99470901e-01 -4.92690653e-02 -6.13277316e-01 4.75510925e-01
4.38174829e-02 4.31515336e-01 -6.49228275e-01 -2.16139227e-01
5.15560150e-01 4.79054451e-01 -2.92518318e-01 5.02773583e-01
-1.81758538e-01 5.49621761e-01 -2.90096253e-01 5.27116895e-01
1.59076977e+00 4.63269651e-01 -5.26258945e-01 -6.65461779e-01
-5.72046518e-01 -4.51342374e-01 -1.62120605e+00 1.44156122e+00
-3.48581493e-01 3.93208236e-01 3.95757526e-01 -1.14057861e-01
1.15711176e+00 2.92399130e-03 4.99147400e-02 -7.84324229e-01
2.09299266e-01 2.06543162e-01 -2.55278856e-01 -1.09223056e+00
4.47220922e-01 -3.18465203e-01 4.95174080e-01 5.70080206e-02
-4.27060783e-01 -5.16984522e-01 1.58565015e-01 4.64873277e-02
1.09854662e+00 3.59202027e-01 -3.26327607e-02 -1.34923771e-01
7.04139650e-01 2.78190792e-01 3.81829619e-01 5.12635708e-01
-3.52055393e-02 9.72323477e-01 -8.16066414e-02 -4.04146314e-01
-6.14601970e-01 -1.06852758e+00 -1.23128772e-01 3.88819724e-01
4.30281907e-01 -5.04485294e-02 -5.99228084e-01 -3.73945609e-02
-2.75764585e-01 9.16053295e-01 -7.90169090e-02 -1.31913349e-01
-2.59980708e-01 -1.01580560e+00 2.66735882e-01 -9.72110629e-02
8.81560028e-01 -1.28469336e+00 -9.07142460e-01 -3.91510539e-02
-2.36191258e-01 -8.65857244e-01 -1.75633617e-02 -1.86875805e-01
-5.81984818e-01 -1.49462640e+00 -1.11832172e-01 -4.37543958e-01
3.82356673e-01 1.04008019e+00 1.13613272e+00 4.85740364e-01
-7.72192717e-01 1.97883785e-01 -6.41069233e-01 -4.53840256e-01
-1.62916198e-01 -9.83608842e-01 -3.55755538e-01 1.98841989e-01
1.48833916e-01 -5.82454979e-01 -8.58036458e-01 6.38367310e-02
-1.19525909e+00 9.89119932e-02 3.54806930e-01 5.04891217e-01
4.02079850e-01 6.20829761e-01 3.68956238e-01 -6.96960986e-01
5.27311623e-01 -1.64510906e-01 -8.40939820e-01 5.89449555e-02
-3.57412010e-01 -4.52066302e-01 7.83114970e-01 -2.27063335e-02
-1.85175431e+00 -3.52246463e-02 3.99741046e-02 -4.79428768e-01
-3.36322516e-01 1.45715162e-01 -4.45718437e-01 -7.13365292e-03
8.12542737e-01 1.33333609e-01 -1.44955620e-01 -5.79883099e-01
3.56063157e-01 6.74649835e-01 7.26781785e-01 1.57234430e-01
1.04281664e+00 9.42589939e-01 2.30712160e-01 -8.76905978e-01
-9.11749244e-01 -5.24581432e-01 -2.21021488e-01 -6.34680927e-01
7.39830911e-01 -8.57319236e-01 -4.92862463e-01 9.08417881e-01
-1.18120003e+00 -1.90891642e-02 -2.96476930e-01 5.17424405e-01
-4.31495100e-01 6.36407971e-01 -5.89226671e-02 -1.17866981e+00
-5.15466630e-01 -1.11973071e+00 1.10099947e+00 6.20290816e-01
5.83069026e-01 -7.09766448e-01 3.33664417e-01 2.89810300e-01
3.85850817e-01 2.99480796e-01 5.69855094e-01 3.30826074e-01
-1.01153386e+00 2.68349759e-02 -4.01248515e-01 3.32274228e-01
1.79582372e-01 4.35779899e-01 -1.20068645e+00 -2.55931228e-01
5.49045205e-01 2.97484342e-02 8.24383080e-01 5.48513412e-01
4.83559132e-01 1.60797387e-01 -1.36893820e-02 1.13484228e+00
2.13293171e+00 2.05193162e-01 1.13160408e+00 3.38577867e-01
5.51219463e-01 6.94997370e-01 1.25447392e+00 6.84600532e-01
-1.31544903e-01 4.71670777e-01 8.03012252e-01 -6.32636011e-01
-7.63425529e-01 1.79812387e-01 2.68376023e-01 5.06094277e-01
-8.89323130e-02 -5.79761028e-01 -5.41211784e-01 4.79173243e-01
-1.68130696e+00 -1.39180672e+00 -1.39366639e+00 2.27691984e+00
4.48100448e-01 -3.08228791e-01 -5.05187333e-01 2.25651830e-01
6.24444187e-01 2.91525126e-01 -9.78066623e-02 -1.41483381e-01
-8.37033153e-01 2.70874679e-01 5.56053400e-01 7.89968193e-01
-7.05707490e-01 7.31961548e-01 5.80910349e+00 7.32879043e-01
-1.00791132e+00 6.86085075e-02 -2.37540424e-01 -1.25190243e-01
-5.43826640e-01 2.59822369e-01 -8.17663014e-01 5.40357709e-01
4.80839193e-01 -4.90577817e-02 7.06006765e-01 8.61347616e-02
8.46481323e-01 -6.17263079e-01 -4.22858506e-01 1.15271759e+00
4.55318660e-01 -8.53719831e-01 1.83882415e-02 -3.97791028e-01
7.73761034e-01 -2.25718901e-01 -3.27677518e-01 -2.29566157e-01
1.33873835e-01 -7.05908179e-01 4.59808350e-01 1.40008032e+00
4.88434136e-01 -7.96058774e-01 7.03767836e-01 4.90309417e-01
-9.94567275e-01 1.54849226e-02 -5.10359347e-01 -2.50794649e-01
6.54585361e-01 1.28487241e+00 -4.10686404e-01 1.26708782e+00
7.51502693e-01 6.77936494e-01 -7.94242561e-01 1.39557207e+00
-3.59933525e-01 3.63228858e-01 -3.58464748e-01 6.89042628e-01
-4.12091799e-02 -8.14195693e-01 7.17545331e-01 1.04889226e+00
5.36435783e-01 2.11549401e-01 -2.61010360e-02 6.32173300e-01
6.49971485e-01 1.04554012e-01 -8.27546954e-01 3.61289948e-01
1.73273727e-01 1.63451529e+00 -1.94053128e-01 -1.67674795e-02
-4.14197594e-01 7.71298826e-01 -2.32605547e-01 3.84329945e-01
-9.64813232e-01 -3.54674101e-01 4.63602811e-01 3.60170484e-01
-5.96154016e-03 2.71506682e-02 -5.77751733e-02 -8.41720402e-01
1.94074050e-01 -9.21490073e-01 2.31473088e-01 -1.69572151e+00
-1.26115644e+00 6.12883508e-01 -1.71895772e-01 -1.69516778e+00
1.20057158e-01 -4.09450978e-01 -7.40014732e-01 1.08507788e+00
-2.03920889e+00 -1.38392019e+00 -1.10641479e+00 8.66599619e-01
3.23671758e-01 -4.58306894e-02 5.91419816e-01 6.15218520e-01
-3.13963771e-01 -3.08952540e-01 2.69719005e-01 -6.78735912e-01
1.03956521e+00 -7.96845078e-01 -3.21803510e-01 1.37409270e+00
-2.18443021e-01 -2.41771550e-03 1.11813939e+00 -7.75213838e-01
-1.51637995e+00 -1.05747676e+00 5.23151755e-01 -2.89934576e-01
6.63877845e-01 3.47983427e-02 -1.05254364e+00 2.78867006e-01
2.88689405e-01 -7.13457838e-02 1.07205220e-01 -4.52621549e-01
2.73128867e-01 -4.57899183e-01 -1.07366669e+00 4.83755350e-01
1.13312674e+00 -1.60997495e-01 -7.42757082e-01 5.60599744e-01
4.00853962e-01 -2.62679815e-01 -3.19032788e-01 7.97426045e-01
1.06896080e-01 -1.89828932e+00 8.93528283e-01 1.47237316e-01
4.30132091e-01 -9.45275784e-01 -2.45084856e-02 -1.41725636e+00
-4.22421873e-01 -5.76369584e-01 1.39242411e-01 1.70406747e+00
-1.66506082e-01 -3.84146631e-01 5.73950671e-02 -2.37711638e-01
-4.68655765e-01 1.10463575e-01 -3.29832584e-01 -3.42296034e-01
-6.39162123e-01 -1.94606617e-01 6.38650119e-01 8.42201471e-01
-8.50188196e-01 3.45487952e-01 -5.55548608e-01 7.44441509e-01
8.11633110e-01 5.39209187e-01 1.09908867e+00 -1.38800085e+00
-2.18928009e-01 -6.34431466e-02 -1.48787960e-01 -6.35286212e-01
-3.08434397e-01 -5.09064615e-01 2.27785006e-01 -1.91101778e+00
6.36464208e-02 -2.92029470e-01 2.00338066e-01 -1.28419891e-01
-4.46096539e-01 2.15947434e-01 1.64383352e-01 4.04297769e-01
-2.60016918e-01 5.59336126e-01 1.43657970e+00 2.73115724e-01
-9.17322487e-02 -1.93456188e-01 -3.21651846e-01 9.91031229e-01
4.35098588e-01 -2.11770833e-01 -6.63649321e-01 -7.61313140e-01
4.10519272e-01 1.75996706e-01 5.25421917e-01 -1.27894640e+00
1.89900175e-01 -1.26048505e-01 1.88902095e-01 -1.00034928e+00
2.68554389e-01 -1.25442755e+00 7.49242902e-01 1.47301108e-01
5.77149570e-01 -2.77347237e-01 1.31206945e-01 3.68855923e-01
-3.38860452e-01 -5.67772329e-01 9.57855523e-01 -1.85582846e-01
-6.41213179e-01 2.14986280e-02 -1.45015828e-02 -1.45486509e-02
1.05844617e+00 -4.17811394e-01 -7.98133612e-01 -2.83409446e-01
-3.86472076e-01 1.59446299e-01 6.76525593e-01 8.02307576e-02
4.06506717e-01 -9.56276357e-01 -1.01503074e+00 6.72576010e-01
3.91575433e-02 -8.38071480e-02 9.78377104e-01 8.48562419e-01
-8.59021604e-01 -1.33471325e-01 -3.90686721e-01 -3.76035094e-01
-1.40034103e+00 6.29993916e-01 4.93438005e-01 -4.49228883e-02
-7.56522477e-01 4.11193132e-01 7.75399923e-01 4.50181551e-02
-1.51338339e-01 1.65337957e-02 -6.12183928e-01 2.80494213e-01
8.83342385e-01 5.54769099e-01 3.57116938e-01 -6.17376149e-01
-8.91396590e-03 8.99342239e-01 5.49114108e-01 3.66956592e-01
1.35954034e+00 -5.84502876e-01 -2.85078138e-01 3.37350696e-01
5.34438491e-01 6.98939145e-01 -1.17521012e+00 1.48458183e-01
-7.97497571e-01 -9.73419905e-01 3.93844754e-01 -1.12623048e+00
-1.19400549e+00 8.72289002e-01 8.99224937e-01 2.94418126e-01
1.88751554e+00 -6.15640700e-01 4.11955118e-01 -4.16235439e-02
3.64058495e-01 -6.64194465e-01 -2.63483882e-01 3.35711986e-01
1.02359927e+00 -8.50693583e-01 -4.84976359e-03 -6.69801772e-01
-5.23972750e-01 1.04882777e+00 4.30745512e-01 -1.44473389e-01
4.23802167e-01 7.28852868e-01 3.15730453e-01 -5.08734465e-01
-5.57691813e-01 -6.93486094e-01 -1.21437050e-01 8.26699972e-01
2.16628890e-02 -3.84994596e-01 -5.09560883e-01 6.95015013e-01
-1.32661715e-01 2.44532853e-01 8.21635246e-01 6.84839904e-01
-7.91407466e-01 -6.61018372e-01 -9.78606343e-01 1.89859912e-01
-7.21602216e-02 -3.25630695e-01 -1.57214567e-01 5.50610065e-01
5.63299239e-01 1.30683112e+00 -3.08764242e-02 1.06436051e-01
8.42452824e-01 -2.24582121e-01 4.83652979e-01 -3.25782090e-01
-7.93681622e-01 3.62714738e-01 1.13564856e-01 -4.54701066e-01
-7.18269467e-01 -2.98795015e-01 -1.03791702e+00 -3.23109776e-01
-5.40927708e-01 -9.96255875e-02 4.43846822e-01 5.39730847e-01
2.24244177e-01 6.98168516e-01 8.86656165e-01 -6.79966390e-01
-9.31692123e-02 -9.97607410e-01 -9.95720088e-01 8.00071836e-01
3.55402708e-01 -8.01034033e-01 -1.11219919e+00 2.54793525e-01] | [10.721453666687012, -3.053920030593872] |
761b7c42-3f6d-41a9-8836-424197dcc7d7 | prompt-based-zero-shot-relation | 2112.04539 | null | https://arxiv.org/abs/2112.04539v2 | https://arxiv.org/pdf/2112.04539v2.pdf | Prompt-based Zero-shot Relation Extraction with Semantic Knowledge Augmentation | In relation triplet extraction (RTE), recognizing unseen (new) relations for which there are no training instances is a challenging task. Efforts have been made to recognize unseen relations based on question-answering models or relation descriptions. However, these approaches miss the semantic information about connections between seen and unseen relations. In this paper, We propose a prompt-based model with semantic knowledge augmentation (ZS-SKA) to recognize unseen relations under the zero-shot setting. We present a new word-level analogy-based sentence translation rule and generate augmented instances with unseen relations from instances with seen relations using that new rule. We design prompts with weighted virtual label construction based on an external knowledge graph to integrate semantic knowledge information learned from seen relations. Instead of using the actual label sets in the prompt template, we construct weighted virtual label words. We learn the representations of both seen and unseen relations with augmented instances and prompts. We then calculate the distance between the generated representations using prototypical networks to predict unseen relations. Extensive experiments conducted on three public datasets FewRel, Wiki-ZSL, and NYT, show that ZS-SKA outperforms state-of-the-art methods under the zero-shot scenarios. Our experimental results also demonstrate the effectiveness and robustness of ZS-SKA. | ['Hoda Eldardiry', 'Jiaying Gong'] | 2021-12-08 | null | null | null | null | ['relation-classification'] | ['natural-language-processing'] | [ 3.74843329e-01 9.09813404e-01 -2.10099593e-01 -6.47490919e-01
-4.94793117e-01 -3.30214560e-01 7.07937479e-01 1.45499676e-01
-3.75882089e-02 8.83003652e-01 2.68946528e-01 -3.08464080e-01
-3.91261399e-01 -1.20584595e+00 -6.27148807e-01 -1.23497352e-01
3.16202223e-01 9.90249336e-01 2.13656753e-01 -8.79743814e-01
-3.70035291e-01 2.05963239e-01 -1.31568444e+00 6.57524407e-01
9.47074175e-01 8.35540235e-01 -2.68980056e-01 2.56276727e-01
-5.31349063e-01 1.16105187e+00 -7.39488900e-01 -9.13593292e-01
2.93009758e-01 -6.20880306e-01 -1.52583206e+00 -1.19676128e-01
2.87115604e-01 -1.65220480e-02 -7.26464570e-01 7.37942934e-01
2.19696119e-01 6.04412854e-01 5.97443640e-01 -1.39551580e+00
-1.38862789e+00 9.06087816e-01 -1.50026932e-01 3.31999838e-01
7.11899400e-01 -3.77164185e-01 1.23299086e+00 -1.05910408e+00
1.00304019e+00 1.45283139e+00 2.46458650e-01 7.91073382e-01
-1.30647337e+00 -7.11398780e-01 2.47450426e-01 8.04836392e-01
-1.41380203e+00 -5.13069451e-01 6.84262216e-01 -3.76510248e-02
1.58536589e+00 4.45315391e-01 3.33356202e-01 1.31291842e+00
-3.51093858e-01 4.57565755e-01 8.12109590e-01 -8.36451828e-01
1.00160157e-02 3.20840895e-01 5.61093152e-01 4.79195535e-01
2.37868115e-01 -3.61108035e-02 -6.92898750e-01 -1.09105147e-01
5.14937758e-01 -1.92080185e-01 -4.21546519e-01 -1.68480039e-01
-1.11185384e+00 5.03548324e-01 7.60712445e-01 4.70002711e-01
-1.36287764e-01 -3.78251880e-01 1.92334458e-01 6.44340873e-01
6.94740772e-01 8.56269419e-01 -6.32615983e-01 3.36591631e-01
1.23310294e-02 -4.89911661e-02 1.10967493e+00 1.46821856e+00
9.57336187e-01 -4.34947312e-01 -5.33449829e-01 1.15785825e+00
-1.66407481e-01 2.13071615e-01 4.56955701e-01 -6.28706217e-01
8.41897905e-01 1.16182458e+00 -1.24264583e-01 -9.87607956e-01
-3.45589846e-01 -4.65587229e-01 -6.91021800e-01 -5.89209020e-01
5.12939431e-02 -6.46317527e-02 -8.40964973e-01 1.74427915e+00
5.39642751e-01 5.06810188e-01 8.84725809e-01 7.37905383e-01
1.50407529e+00 4.96480495e-01 6.86130626e-03 -1.53146341e-01
1.38337970e+00 -1.18361807e+00 -9.92352009e-01 -4.85507995e-01
1.00001764e+00 -4.29761916e-01 8.84379923e-01 -2.75098383e-01
-4.60346103e-01 -5.15239477e-01 -9.93845224e-01 -3.40872377e-01
-7.83406734e-01 -1.45992577e-01 7.25290298e-01 1.07733548e-01
-7.23791301e-01 5.61647296e-01 -2.99329579e-01 -8.01168621e-01
2.78383344e-01 1.70435652e-01 -5.44467568e-01 -3.44007760e-01
-1.79849911e+00 1.40336752e+00 7.58801222e-01 8.36132765e-02
-4.09905553e-01 -6.60995305e-01 -1.28159034e+00 2.24554643e-01
1.09281158e+00 -8.27179551e-01 1.05129564e+00 -4.26908642e-01
-1.05209005e+00 8.85099113e-01 -3.49895597e-01 -3.60576004e-01
-2.31149063e-01 -1.35882050e-01 -9.06882584e-01 7.67867342e-02
2.75411755e-01 4.86347139e-01 1.26204357e-01 -1.52627373e+00
-3.61576170e-01 -2.85133600e-01 4.66485620e-01 5.27404130e-01
-4.86952364e-01 -4.18237224e-02 -1.09173441e-02 -3.93367946e-01
4.59448397e-01 -6.02495492e-01 1.40313059e-01 -4.55529928e-01
-7.11594939e-01 -8.78634512e-01 6.42682135e-01 -4.24385190e-01
9.96816218e-01 -1.75337148e+00 -2.10744459e-02 1.35516867e-01
3.98692131e-01 4.06525671e-01 -6.69498920e-01 5.68316460e-01
-4.94732082e-01 -4.26461473e-02 1.19900867e-01 1.27883598e-01
-2.67109483e-01 6.79858804e-01 -4.43684697e-01 -4.90732074e-01
5.28292298e-01 1.29469585e+00 -1.53144586e+00 -4.00970787e-01
-6.01891940e-03 1.33120432e-01 1.48023039e-01 5.87285161e-01
-4.44122665e-02 1.95882946e-01 -3.84270042e-01 4.91658717e-01
2.07328171e-01 -5.54668128e-01 4.49118525e-01 -4.67906207e-01
8.18323612e-01 6.81429744e-01 -7.69020379e-01 1.68912339e+00
-6.07634723e-01 3.67147565e-01 -8.74830604e-01 -1.06521261e+00
1.44204390e+00 5.29702127e-01 -6.45600483e-02 -7.00861752e-01
9.70343649e-02 1.06961861e-01 1.64612636e-01 -9.03469980e-01
4.42230463e-01 -5.29218554e-01 1.58498436e-01 2.45137751e-01
7.02996314e-01 -3.79140265e-02 3.35877419e-01 7.00037599e-01
1.52492249e+00 2.41992310e-01 5.02979159e-01 2.34794110e-01
3.85940254e-01 1.73776254e-01 6.42267406e-01 5.74487150e-01
1.96476564e-01 4.32312310e-01 5.53265512e-01 -4.27408665e-01
-4.67956156e-01 -1.07813501e+00 1.07245460e-01 9.72352207e-01
4.41854656e-01 -6.81795359e-01 -2.61706620e-01 -1.17822373e+00
-1.73747465e-01 1.17476213e+00 -6.60531223e-01 -7.94036984e-01
-4.43120331e-01 -4.76186305e-01 3.27336818e-01 5.32182813e-01
4.10169661e-01 -1.34985936e+00 1.79645434e-01 3.24460685e-01
-6.05511844e-01 -1.67604947e+00 -3.45601961e-02 2.17783272e-01
-4.80491340e-01 -1.25298536e+00 -1.09843850e-01 -9.86884952e-01
9.36653316e-01 5.04051149e-01 1.46411800e+00 8.62430707e-02
-1.65748596e-01 2.05896318e-01 -9.11428869e-01 -8.67721811e-02
-3.94630015e-01 2.20465824e-01 3.47443074e-02 1.09928399e-01
1.05628991e+00 -7.29191303e-01 -1.24904804e-01 3.20895165e-01
-6.72815621e-01 2.65106142e-01 3.44868451e-01 9.71129477e-01
2.79945791e-01 -1.26248613e-01 9.77493465e-01 -1.38403368e+00
8.27874482e-01 -7.43681848e-01 6.80722147e-02 9.39037085e-01
-6.70357168e-01 2.84692526e-01 5.80846071e-01 -5.15854180e-01
-1.38976324e+00 -3.78092289e-01 2.92960972e-01 -4.18093294e-01
-1.85105681e-01 4.88766581e-01 -2.68435508e-01 2.81675905e-01
1.09246349e+00 -1.75598755e-01 -4.84311908e-01 -3.16423237e-01
7.50070691e-01 7.50850320e-01 6.97820246e-01 -9.28377748e-01
9.29386437e-01 1.22115590e-01 -1.01992853e-01 -2.72845507e-01
-1.81067920e+00 -4.71256077e-01 -6.08245671e-01 2.18307395e-02
7.23103046e-01 -7.33522177e-01 -4.46333945e-01 -2.53921121e-01
-1.53322840e+00 -4.33550999e-02 -8.19396079e-01 2.99088418e-01
-1.67916730e-01 -2.10287306e-03 -6.55609965e-01 -7.15488851e-01
-4.43627000e-01 -4.91189152e-01 8.37345958e-01 3.30960125e-01
-5.23374140e-01 -9.76020873e-01 8.77622142e-02 6.18315876e-01
3.79792564e-02 2.17673257e-01 1.24948335e+00 -1.37013733e+00
-3.39241624e-01 -1.96519405e-01 -5.69149673e-01 3.50065500e-01
5.41575432e-01 -6.01441324e-01 -1.01296878e+00 2.37732574e-01
-2.19986975e-01 -6.71577156e-01 4.97885883e-01 -5.00690281e-01
6.90973222e-01 -3.52924168e-01 -4.80595767e-01 2.03496084e-01
1.06887388e+00 2.12026015e-01 5.22512317e-01 8.12410116e-02
9.93567169e-01 1.08465922e+00 8.29576612e-01 -1.07323058e-01
7.09439754e-01 5.96201301e-01 -3.81475926e-04 9.85173061e-02
-4.97140765e-01 -5.83865881e-01 -1.15458868e-01 1.05951202e+00
-9.28072408e-02 -3.19339812e-01 -9.41956937e-01 6.65819466e-01
-2.08262491e+00 -7.86422372e-01 -1.84996843e-01 1.79294419e+00
1.15415394e+00 2.23910436e-02 -6.96591973e-01 3.46302912e-02
9.07219172e-01 -2.10446969e-01 -5.57758451e-01 -3.31915081e-01
-2.98497379e-01 6.70854747e-01 -2.45420232e-01 5.88837147e-01
-6.56809449e-01 1.46097910e+00 4.84756041e+00 7.54423797e-01
-1.91703424e-01 1.99276298e-01 3.98216963e-01 8.19623247e-02
-3.48388314e-01 3.67382944e-01 -7.82520354e-01 -2.99008667e-01
8.02455425e-01 -3.52353215e-01 4.24813479e-01 6.35123909e-01
-4.85131234e-01 1.74281850e-01 -1.56713045e+00 8.78017485e-01
3.71570468e-01 -1.23865402e+00 3.83595645e-01 -3.45929295e-01
6.75141811e-01 -4.10460055e-01 -5.59999764e-01 8.10872138e-01
6.40234172e-01 -8.77703488e-01 7.29232747e-03 5.16880393e-01
7.62579679e-01 -4.03896064e-01 1.03717494e+00 1.79972410e-01
-1.16108418e+00 9.94622186e-02 -5.55055380e-01 -3.06012958e-01
7.72154257e-02 5.61419547e-01 -1.10545266e+00 1.21133780e+00
4.53037918e-01 8.05861831e-01 -5.99075615e-01 3.85351628e-01
-1.01631844e+00 2.40985349e-01 -5.70819676e-02 3.19592212e-03
-1.68399930e-01 -1.01106986e-01 2.12824553e-01 6.89414084e-01
2.01528966e-01 8.01873863e-01 2.82481853e-02 9.18839157e-01
-4.43538636e-01 1.27022594e-01 -8.11997533e-01 8.67618844e-02
8.55370283e-01 1.46006739e+00 -4.06165928e-01 -6.56963170e-01
-5.57078302e-01 1.03225720e+00 1.02841353e+00 6.23960137e-01
-3.74942183e-01 -5.75737000e-01 4.35601741e-01 -2.40661398e-01
-2.93361276e-01 3.31642658e-01 3.16421594e-03 -1.47059953e+00
2.55819976e-01 -5.65611184e-01 6.17439330e-01 -1.22331488e+00
-1.89026380e+00 1.03856421e+00 1.36573106e-01 -9.55587149e-01
-2.79339939e-01 -3.61017019e-01 -4.36293513e-01 9.41234708e-01
-1.40905201e+00 -1.15003169e+00 -5.34283161e-01 4.28505152e-01
4.96854186e-01 -1.96817979e-01 1.29905891e+00 1.12085804e-01
-6.70118272e-01 6.21688187e-01 -6.42744541e-01 4.93094295e-01
7.39640236e-01 -1.11785972e+00 6.04695082e-01 5.82691729e-01
5.88445842e-01 5.86877644e-01 5.22954524e-01 -9.00887728e-01
-8.06095779e-01 -1.21573043e+00 1.52940214e+00 -7.36290932e-01
8.09888005e-01 -4.16641414e-01 -1.23534179e+00 1.04580510e+00
3.74746442e-01 4.10925210e-01 9.67922032e-01 5.45108616e-01
-7.82106638e-01 -1.40048295e-01 -1.00907290e+00 6.45200074e-01
1.73057294e+00 -5.78303158e-01 -1.40857899e+00 6.91898704e-01
1.26403832e+00 -3.88600349e-01 -9.88054812e-01 6.96621835e-01
4.71460931e-02 -3.88699949e-01 8.90522897e-01 -1.27257717e+00
5.22018433e-01 -1.13073371e-01 -5.41498773e-02 -1.60100842e+00
-3.04717064e-01 -3.58063340e-01 -5.15062511e-01 1.53004456e+00
8.64644051e-01 -7.37270772e-01 6.02865577e-01 9.65897441e-01
-1.07191391e-01 -8.99220049e-01 -8.59412193e-01 -1.16787469e+00
-3.70234579e-01 -1.47425551e-02 1.00024891e+00 1.45144057e+00
5.18390119e-01 1.35830998e+00 -5.63147105e-02 1.78027272e-01
2.98894972e-01 2.23688111e-01 5.10117650e-01 -1.28605008e+00
-2.91506588e-01 2.79370338e-01 -4.51524764e-01 -6.30997717e-01
5.92877150e-01 -1.27260065e+00 -1.46470144e-01 -1.88011813e+00
3.66121739e-01 -5.26462853e-01 -4.54284549e-01 1.01874304e+00
-6.40387952e-01 4.21703719e-02 -1.07808925e-01 -5.48980720e-02
-5.94293416e-01 7.37761199e-01 1.36824298e+00 -1.85600579e-01
-5.46304323e-02 -3.97543609e-01 -8.40275288e-01 4.65087235e-01
7.49147117e-01 -6.73914373e-01 -9.46077108e-01 -5.49803436e-01
3.80201250e-01 4.83060218e-02 2.42071897e-01 -6.68156862e-01
2.31033236e-01 -1.44056603e-01 7.23802224e-02 -1.25251800e-01
5.50374091e-01 -7.75232077e-01 -1.41844943e-01 -1.32938236e-01
-7.67250359e-01 -2.79554248e-01 -1.47240564e-01 5.91484308e-01
-2.40517572e-01 -3.03939939e-01 7.96978176e-02 1.48379961e-02
-6.49975419e-01 1.07648909e-01 4.13834453e-01 5.00169396e-01
9.73387539e-01 5.93942851e-02 -8.84475708e-01 -3.28048348e-01
-1.06396329e+00 3.72842848e-01 -1.62342280e-01 9.31433558e-01
8.92326295e-01 -1.62961662e+00 -7.56645739e-01 -1.37148544e-01
9.37027812e-01 2.49760225e-01 1.16693288e-01 2.73907870e-01
1.05616145e-01 2.02220693e-01 1.89375475e-01 2.10280977e-02
-1.16669369e+00 8.57081056e-01 2.49252796e-01 -4.29965436e-01
-6.43063784e-01 1.03129673e+00 1.45830929e-01 -7.43650079e-01
-7.92301446e-02 -1.34903565e-01 -4.17002618e-01 -6.92212433e-02
4.39745694e-01 4.49966118e-02 2.49667540e-01 -4.35802877e-01
-2.18059957e-01 2.31272042e-01 -3.97882044e-01 6.35247976e-02
1.23406172e+00 3.49659957e-02 -2.83326775e-01 4.93656754e-01
9.58480895e-01 -6.13801181e-01 -4.52306002e-01 -9.98059034e-01
4.38624263e-01 -4.06582415e-01 -4.97099936e-01 -1.03093326e+00
-8.64913404e-01 5.85671782e-01 -2.59336736e-02 1.64246246e-01
8.47072124e-01 6.21679485e-01 9.64145124e-01 8.13116610e-01
5.20209193e-01 -7.63677955e-01 1.83876529e-01 5.62092125e-01
9.78659689e-01 -1.27030790e+00 -3.11218947e-01 -1.30598199e+00
-5.62727809e-01 8.43083918e-01 1.37647283e+00 1.09233454e-01
2.24993035e-01 -4.72855791e-02 1.82882771e-01 -3.97006929e-01
-1.07042980e+00 -5.99760592e-01 4.47061568e-01 9.44426537e-01
3.84605110e-01 3.96611877e-02 -2.02386826e-01 8.10233891e-01
-2.68974334e-01 -2.72393197e-01 3.60866904e-01 9.51298177e-01
-2.01330185e-01 -1.36626244e+00 1.26582477e-02 7.32585490e-01
2.82667100e-01 -5.46416879e-01 -8.02946985e-01 4.82581049e-01
1.44686475e-01 1.51068616e+00 -1.49393931e-01 -7.64886975e-01
6.78118408e-01 5.55681527e-01 5.82757294e-01 -1.41695499e+00
-5.64841866e-01 -8.19545031e-01 8.14231694e-01 -3.74078602e-01
-2.69732773e-01 4.99447621e-02 -1.38883018e+00 5.17166294e-02
-7.24522531e-01 4.27961171e-01 -6.50548050e-03 1.39202368e+00
4.66476530e-01 9.26420867e-01 4.23944235e-01 1.42392721e-02
-3.99652332e-01 -1.21442485e+00 -3.86132807e-01 9.05775070e-01
-1.98307142e-01 -8.52905273e-01 -1.97567448e-01 -1.74389392e-01] | [9.33158016204834, 8.439952850341797] |
0b9be94e-0fa4-4098-a89c-e5e84ce60237 | offline-online-associated-camera-aware | 2201.05820 | null | https://arxiv.org/abs/2201.05820v2 | https://arxiv.org/pdf/2201.05820v2.pdf | Offline-Online Associated Camera-Aware Proxies for Unsupervised Person Re-identification | Recently, unsupervised person re-identification (Re-ID) has received increasing research attention due to its potential for label-free applications. A promising way to address unsupervised Re-ID is clustering-based, which generates pseudo labels by clustering and uses the pseudo labels to train a Re-ID model iteratively. However, most clustering-based methods take each cluster as a pseudo identity class, neglecting the intra-cluster variance mainly caused by the change of cameras. To address this issue, we propose to split each single cluster into multiple proxies according to camera views. The camera-aware proxies explicitly capture local structures within clusters, by which the intra-ID variance and inter-ID similarity can be better tackled. Assisted with the camera-aware proxies, we design two proxy-level contrastive learning losses that are, respectively, based on offline and online association results. The offline association directly associates proxies according to the clustering and splitting results, while the online strategy dynamically associates proxies in terms of up-to-date features to reduce the noise caused by the delayed update of pseudo labels. The combination of two losses enables us to train a desirable Re-ID model. Extensive experiments on three person Re-ID datasets and one vehicle Re-ID dataset show that our proposed approach demonstrates competitive performance with state-of-the-art methods. Code will be available at: https://github.com/Terminator8758/O2CAP. | ['Xian-Sheng Hua', 'Xiaojin Gong', 'Baisheng Lai', 'Jiachen Li', 'Menglin Wang'] | 2022-01-15 | null | null | null | null | ['unsupervised-person-re-identification'] | ['computer-vision'] | [-1.83568045e-01 -1.62213206e-01 -2.05880612e-01 -5.56275249e-01
-6.48864865e-01 -4.86486822e-01 6.48646116e-01 -7.39158988e-02
-5.48863947e-01 4.68955576e-01 1.37548715e-01 3.92673612e-01
-1.09390736e-01 -6.50287271e-01 -5.88221729e-01 -7.29461372e-01
1.53304681e-01 8.34690154e-01 1.09712876e-01 3.89335066e-01
-5.14417887e-02 3.26541692e-01 -1.76475251e+00 8.02659690e-02
9.91878450e-01 7.53734171e-01 1.31225556e-01 2.40242764e-01
-6.29394427e-02 4.83678073e-01 -3.11058730e-01 -5.99177003e-01
4.63891268e-01 -4.02535260e-01 -5.50917625e-01 3.32326561e-01
3.11177343e-01 -2.52888680e-01 -3.52849662e-01 1.21213770e+00
5.53989887e-01 3.26097816e-01 4.57249433e-01 -1.53912568e+00
-4.74344999e-01 5.65313756e-01 -5.54013729e-01 -1.60161793e-01
2.86945820e-01 -8.18576757e-03 8.06495011e-01 -8.79373431e-01
4.30278689e-01 1.26136363e+00 8.14049721e-01 6.96072102e-01
-1.41667414e+00 -9.62821424e-01 3.55326861e-01 6.61836863e-01
-1.81573153e+00 -4.46567595e-01 8.82922769e-01 -5.03805935e-01
3.24768960e-01 7.93408975e-02 4.14073169e-01 1.03071642e+00
-6.50081694e-01 7.95501292e-01 8.92474890e-01 -3.36324573e-01
1.44908115e-01 3.91335100e-01 3.10370088e-01 4.66031432e-01
1.63346931e-01 1.52373016e-01 -2.40468323e-01 -1.48457542e-01
4.26692486e-01 3.49551320e-01 -1.22787662e-01 -5.73159218e-01
-1.17176211e+00 6.36301816e-01 3.79839033e-01 7.97142237e-02
-2.08884269e-01 7.16202781e-02 3.67609113e-01 -8.49661902e-02
2.53192425e-01 -1.27318397e-01 -1.49709284e-01 1.61427762e-02
-8.57589066e-01 4.87168580e-02 4.61874038e-01 1.02959740e+00
1.05897617e+00 -3.77233386e-01 -1.35393202e-01 1.09308934e+00
3.28207731e-01 5.59754729e-01 4.92647469e-01 -1.02366674e+00
3.50761145e-01 5.60883939e-01 2.87950009e-01 -1.07373822e+00
-3.58860344e-01 -4.24322605e-01 -1.06491292e+00 -1.16312392e-01
5.37339807e-01 -4.52103838e-03 -6.46266460e-01 1.96953011e+00
4.49290663e-01 7.05963552e-01 -2.42310405e-01 9.49708819e-01
5.54439306e-01 3.69536936e-01 9.38356370e-02 -1.32752329e-01
1.35264087e+00 -1.04522288e+00 -5.26703656e-01 -7.39090219e-02
4.29338634e-01 -3.70301664e-01 7.18754470e-01 2.10091218e-01
-6.84742033e-01 -9.05618966e-01 -8.52487147e-01 3.00040096e-01
-2.39739791e-01 4.83751535e-01 1.00052610e-01 6.16902173e-01
-1.02253258e+00 4.36512083e-01 -6.34655356e-01 -4.71758693e-01
2.22034171e-01 4.63619351e-01 -3.35906267e-01 -1.43259138e-01
-1.12654305e+00 4.74871397e-01 3.68897706e-01 7.18446076e-02
-8.20931017e-01 -5.55436134e-01 -5.72197258e-01 4.54858467e-02
4.60388333e-01 -5.70181549e-01 9.26933408e-01 -1.09090245e+00
-1.32901275e+00 1.01209414e+00 -4.28689212e-01 -2.44168326e-01
7.28270888e-01 3.37556601e-02 -6.03031635e-01 2.18606099e-01
4.63021696e-01 6.37052536e-01 9.59295809e-01 -1.60329902e+00
-9.65328276e-01 -4.58856076e-01 -1.33598536e-01 2.12596297e-01
-5.41869283e-01 -1.32289052e-01 -9.86745954e-01 -6.13874733e-01
-2.35634353e-02 -1.27600205e+00 -6.97042122e-02 -1.54399201e-01
-4.93945301e-01 -3.38573247e-01 6.55781984e-01 -5.47020555e-01
1.12701392e+00 -2.29617643e+00 6.32805452e-02 3.74753445e-01
2.17743024e-01 2.44826898e-01 -1.30733937e-01 5.86437136e-02
-1.96659893e-01 -2.04165839e-02 -4.48914506e-02 -8.57375443e-01
1.25518382e-01 -1.04612196e-02 1.14340328e-01 5.38964391e-01
-3.73558313e-01 5.16829371e-01 -9.40902710e-01 -5.79603553e-01
4.47877169e-01 5.25486052e-01 -2.74694592e-01 3.04460168e-01
2.24512443e-01 6.43583775e-01 -2.92777449e-01 5.00724375e-01
8.37927699e-01 -2.26367533e-01 2.71629274e-01 -4.85452741e-01
-2.08325125e-02 -2.65766442e-01 -1.41100633e+00 1.33203733e+00
-2.28618309e-01 1.28414676e-01 8.65523666e-02 -1.08057106e+00
8.14953566e-01 2.11399943e-01 7.63898492e-01 -4.73001152e-01
8.38254094e-02 8.03637058e-02 -4.91865128e-01 -1.21625990e-01
1.99615851e-01 1.35507479e-01 -9.01549384e-02 4.62653279e-01
-1.60917506e-01 8.62476408e-01 1.91793695e-01 1.67909145e-01
6.95256352e-01 -1.76304244e-02 -7.47658312e-02 -9.84000564e-02
9.05143201e-01 -2.14499101e-01 8.55831206e-01 7.20160186e-01
-4.89052802e-01 6.68610275e-01 -1.12104386e-01 -4.35396940e-01
-8.97101641e-01 -1.14084864e+00 -1.23441093e-01 9.53611076e-01
6.32951498e-01 -2.78363138e-01 -8.45231354e-01 -8.05059671e-01
1.24336846e-01 5.50451815e-01 -5.11540949e-01 -2.21448377e-01
-5.13438225e-01 -8.69842052e-01 4.30670381e-01 3.76261920e-01
8.95070732e-01 -7.08113670e-01 -1.37763740e-02 2.52390087e-01
-6.41737044e-01 -1.20163882e+00 -7.67189682e-01 -1.94630712e-01
-5.62965870e-01 -1.06799829e+00 -8.42884660e-01 -8.08040023e-01
9.39538002e-01 6.36169493e-01 7.60632753e-01 8.67797881e-02
-2.96597816e-02 6.84360743e-01 -4.62847978e-01 1.12261005e-01
-2.57286698e-01 8.59357640e-02 4.89100873e-01 6.61151171e-01
7.29633093e-01 -5.79233646e-01 -6.88293338e-01 7.21794188e-01
-4.65975732e-01 -7.69011527e-02 3.22181970e-01 7.34321237e-01
7.83383846e-01 3.05809170e-01 5.47211289e-01 -7.72533059e-01
1.22779503e-01 -5.86688101e-01 -5.66693842e-01 3.35838944e-01
-8.21060538e-01 -2.85402648e-02 7.82200634e-01 -5.68299174e-01
-1.11991644e+00 4.00827438e-01 1.69527471e-01 -5.35950780e-01
-3.20330620e-01 4.87813652e-02 -6.21690929e-01 9.12364349e-02
1.98726580e-01 1.50245383e-01 -8.34387913e-02 -5.66716850e-01
4.12422180e-01 8.36496174e-01 7.95874894e-01 -6.32478774e-01
1.01641726e+00 7.04162598e-01 -3.95580500e-01 -5.23797452e-01
-7.19044745e-01 -8.62875402e-01 -7.53394663e-01 -4.04176772e-01
9.67897177e-01 -1.14792192e+00 -8.56690109e-01 7.31151938e-01
-8.45451057e-01 -1.97666958e-01 -1.78485483e-01 5.01412690e-01
-3.03087831e-01 6.66401684e-01 -4.13742185e-01 -6.84697270e-01
-1.55273274e-01 -1.13360977e+00 8.08384061e-01 4.28094625e-01
1.71499252e-02 -8.43388379e-01 -5.30420989e-02 4.55112994e-01
3.11240758e-04 1.60827041e-02 4.35736030e-01 -6.92483664e-01
-5.66513598e-01 -2.92424142e-01 -4.19530690e-01 3.12662482e-01
2.95577019e-01 -3.90688330e-01 -1.07397139e+00 -5.46890497e-01
-3.33555847e-01 4.05705683e-02 7.95533955e-01 2.50259608e-01
1.15104842e+00 -3.28286380e-01 -7.04045355e-01 7.50439346e-01
1.38041663e+00 1.25652298e-01 4.24821615e-01 5.35279334e-01
1.03945589e+00 6.11301899e-01 5.67290425e-01 7.19388127e-01
8.96069050e-01 1.06799614e+00 2.31818005e-01 1.19564142e-02
-2.07714021e-01 -4.29188043e-01 2.90266484e-01 6.48760080e-01
-1.25373647e-01 -1.23645730e-01 -7.67220736e-01 5.69212794e-01
-2.10342908e+00 -1.18594003e+00 -8.58998969e-02 2.46680331e+00
5.33692062e-01 -1.35229692e-01 5.20833075e-01 -6.79896697e-02
1.40451217e+00 -1.54316947e-01 -6.72259450e-01 3.61021966e-01
-2.05230322e-02 -5.84076881e-01 6.47141218e-01 4.24868375e-01
-1.28548634e+00 7.57408738e-01 4.91892099e+00 8.61822069e-01
-6.81668520e-01 2.51247346e-01 5.95099449e-01 7.12668300e-02
3.60364318e-02 2.20684689e-02 -1.18551695e+00 1.02176404e+00
7.30426192e-01 -2.00904220e-01 6.85624003e-01 9.04284835e-01
2.12893710e-01 9.43360105e-02 -1.26875424e+00 1.42574859e+00
1.92899197e-01 -9.51937556e-01 -8.45809728e-02 1.52526021e-01
7.08579361e-01 -1.66752383e-01 -3.61024439e-02 1.75260633e-01
4.83233303e-01 -5.36659658e-01 8.66245449e-01 6.26848400e-01
8.79144549e-01 -9.13119793e-01 8.40147913e-01 2.85607725e-01
-1.59778655e+00 -4.03410077e-01 -3.29021126e-01 2.74232894e-01
1.62878528e-01 5.98740518e-01 -5.41851938e-01 6.15842819e-01
1.06723809e+00 9.40924048e-01 -7.82002568e-01 9.84384179e-01
-1.03493102e-01 4.44671780e-01 -3.17093968e-01 5.14733732e-01
-2.77461767e-01 -4.36853766e-01 4.67174262e-01 1.10162961e+00
2.90020317e-01 2.79690474e-02 4.53506410e-01 6.86711431e-01
-7.85850137e-02 -2.23156080e-01 -1.26275152e-01 5.27379274e-01
9.71759319e-01 1.25376225e+00 -6.40012324e-01 -4.78834450e-01
-4.05418992e-01 1.23455560e+00 3.16357136e-01 5.06541312e-01
-1.03239655e+00 -1.68046579e-01 7.13847637e-01 2.87724197e-01
3.38796943e-01 -3.85510772e-02 2.81367507e-02 -1.31967628e+00
3.10294796e-02 -6.96328044e-01 6.64115012e-01 -3.32295209e-01
-1.62483752e+00 4.74763423e-01 2.45523289e-01 -1.53970170e+00
-1.38014287e-01 -9.88907516e-02 -4.02186543e-01 4.50194210e-01
-1.42134881e+00 -1.28383017e+00 -7.46672809e-01 7.73162961e-01
2.50775486e-01 -2.66126722e-01 4.96153325e-01 7.72624910e-01
-8.75360250e-01 1.12389803e+00 4.08952892e-01 3.78158778e-01
1.08587945e+00 -1.10726011e+00 5.29891960e-02 8.97102356e-01
-5.92263043e-02 4.16304201e-01 3.63435864e-01 -5.36522210e-01
-8.93894851e-01 -1.35767376e+00 8.98431540e-01 -3.94208103e-01
2.69784003e-01 -3.90233964e-01 -7.81617522e-01 6.76495254e-01
-3.81094158e-01 7.10033393e-03 6.38116479e-01 -3.84386443e-02
-3.73838454e-01 -7.60832906e-01 -1.22980297e+00 4.22518998e-01
1.21211648e+00 -4.41805393e-01 -2.09317431e-01 2.32917160e-01
3.81352574e-01 5.62184341e-02 -7.80026257e-01 2.98543662e-01
3.94531220e-01 -8.96723688e-01 1.23722136e+00 1.07405208e-01
-2.24622488e-01 -7.87389457e-01 1.43481612e-01 -1.08154345e+00
-6.94904208e-01 -3.21031600e-01 9.21075344e-02 1.95325089e+00
-9.96437483e-03 -7.50622690e-01 8.46869588e-01 9.52867687e-01
1.46799952e-01 -3.50902998e-03 -9.27566111e-01 -9.79266644e-01
-3.47779244e-01 -1.68914869e-01 7.77446091e-01 9.96901393e-01
-2.50936329e-01 7.17017576e-02 -6.10465050e-01 3.49363089e-01
1.20795524e+00 8.87102336e-02 9.49808419e-01 -1.41053987e+00
-1.76652998e-01 -3.76602709e-01 -3.74012560e-01 -1.04725778e+00
4.01499778e-01 -9.97669816e-01 1.05490595e-01 -1.26017749e+00
6.14232600e-01 -9.45471764e-01 -3.29132318e-01 5.85717201e-01
-2.88885295e-01 3.28714907e-01 3.72450888e-01 7.52392769e-01
-9.73255515e-01 5.79395771e-01 5.67622542e-01 -3.15142125e-01
-2.69238234e-01 8.49771127e-02 -5.26614487e-01 6.90740705e-01
8.35176289e-01 -6.00710750e-01 -3.29058647e-01 -3.42015058e-01
-2.70017743e-01 -3.38729322e-01 5.55549920e-01 -1.26498139e+00
5.81584573e-01 5.27523831e-02 4.00622904e-01 -4.99521583e-01
1.81513533e-01 -1.02490520e+00 5.87053776e-01 2.14710265e-01
-3.03776413e-01 -1.33428127e-01 -3.11823636e-01 8.53257477e-01
-1.88689947e-01 -1.78652287e-01 9.45842922e-01 -1.12156883e-01
-8.18898857e-01 5.70956826e-01 -2.63200372e-01 -3.88520397e-02
1.14471185e+00 -3.74347895e-01 9.34983604e-03 -4.07512337e-01
-7.42310643e-01 5.17033458e-01 7.38806129e-01 4.35088575e-01
3.40118945e-01 -1.57272744e+00 -7.38770127e-01 8.68101269e-02
2.38186359e-01 -4.29496169e-02 4.83096033e-01 6.82932794e-01
-7.56792724e-02 9.86417383e-02 -1.90841984e-02 -7.44713843e-01
-1.31741273e+00 8.86569500e-01 3.17253083e-01 -3.04604977e-01
-5.72905838e-01 5.95917284e-01 4.00155991e-01 -5.10434568e-01
5.12541890e-01 4.18017864e-01 -2.38302365e-01 9.01629403e-02
6.74309373e-01 7.46842265e-01 -2.93453753e-01 -1.01211166e+00
-5.96139193e-01 8.83979380e-01 -6.34310469e-02 4.69749905e-02
1.02274203e+00 -6.84906781e-01 3.27432319e-03 2.24727139e-01
1.27120221e+00 -1.91636488e-01 -1.45080066e+00 -4.93532687e-01
1.58229575e-01 -4.86820370e-01 -2.88569987e-01 -5.96183538e-01
-1.17321968e+00 4.92667675e-01 8.94477606e-01 -1.72440797e-01
1.19452655e+00 7.98204467e-02 9.62189853e-01 1.78602964e-01
5.35924315e-01 -1.37838471e+00 1.47687972e-01 1.49786428e-01
3.05443972e-01 -1.36918187e+00 -8.31162184e-02 -3.75689566e-01
-5.91878414e-01 7.70909548e-01 6.75662220e-01 5.15708840e-03
7.17252672e-01 -1.30287856e-01 -1.57538541e-02 3.25940520e-01
-6.84363469e-02 -3.32671911e-01 1.24055944e-01 8.17588866e-01
-2.33192652e-01 2.57777184e-01 -5.26213832e-02 8.72697413e-01
4.57370244e-02 -3.57740000e-02 1.97774976e-01 3.62360448e-01
-1.19872317e-01 -1.31984067e+00 -6.12755477e-01 1.89558268e-01
-1.19282588e-01 2.01057658e-01 -1.81663007e-01 5.27546942e-01
5.66128790e-01 1.17626488e+00 1.17361575e-01 -6.83183432e-01
3.00989181e-01 -1.11452416e-01 1.22628108e-01 -3.26466084e-01
-3.45457405e-01 -2.79062670e-02 -7.89631307e-02 -5.27324915e-01
-6.67937398e-01 -8.05979788e-01 -1.25485647e+00 -5.44766009e-01
-2.74954855e-01 2.76286811e-01 2.67531991e-01 7.18808293e-01
5.74307501e-01 8.94116834e-02 1.01658070e+00 -9.29773629e-01
-3.99246186e-01 -6.65032446e-01 -5.58520198e-01 8.88035595e-01
1.80459365e-01 -8.31490576e-01 -6.21240795e-01 3.78961295e-01] | [14.814632415771484, 1.0826942920684814] |
f00d321e-b94f-42fa-aa83-8490e16ce0dc | weight-agnostic-neural-networks | 1906.04358 | null | https://arxiv.org/abs/1906.04358v2 | https://arxiv.org/pdf/1906.04358v2.pdf | Weight Agnostic Neural Networks | Not all neural network architectures are created equal, some perform much better than others for certain tasks. But how important are the weight parameters of a neural network compared to its architecture? In this work, we question to what extent neural network architectures alone, without learning any weight parameters, can encode solutions for a given task. We propose a search method for neural network architectures that can already perform a task without any explicit weight training. To evaluate these networks, we populate the connections with a single shared weight parameter sampled from a uniform random distribution, and measure the expected performance. We demonstrate that our method can find minimal neural network architectures that can perform several reinforcement learning tasks without weight training. On a supervised learning domain, we find network architectures that achieve much higher than chance accuracy on MNIST using random weights. Interactive version of this paper at https://weightagnostic.github.io/ | ['David Ha', 'Adam Gaier'] | 2019-06-11 | weight-agnostic-neural-networks-1 | http://papers.nips.cc/paper/8777-weight-agnostic-neural-networks | http://papers.nips.cc/paper/8777-weight-agnostic-neural-networks.pdf | neurips-2019-12 | ['carracing-v0'] | ['playing-games'] | [ 2.80831277e-01 3.31149369e-01 -1.71247929e-01 -5.13970256e-01
-2.82017946e-01 -5.67753077e-01 3.14208657e-01 -3.19292277e-01
-8.47906947e-01 6.85197711e-01 -3.69443223e-02 -4.87808526e-01
-4.03917342e-01 -7.72000492e-01 -7.23334610e-01 -6.20793164e-01
8.29077139e-02 8.60777617e-01 2.61969298e-01 -1.87281549e-01
2.59489983e-01 3.47544849e-01 -1.29131889e+00 3.75592560e-01
6.44013226e-01 9.31133986e-01 4.83819872e-01 7.59913564e-01
8.60417634e-03 6.92965329e-01 -6.38797164e-01 -5.60215592e-01
3.66024494e-01 -8.72573331e-02 -8.97844076e-01 -3.86231512e-01
3.61276120e-01 2.55102217e-02 -2.47325808e-01 1.00138187e+00
5.64216554e-01 7.03191459e-02 8.81758988e-01 -1.18922377e+00
-6.99630976e-01 1.03813362e+00 -8.15598667e-02 3.51539850e-01
-3.29000145e-01 4.49069113e-01 1.28478301e+00 -6.57945871e-01
3.12171251e-01 9.64510560e-01 6.78585887e-01 8.88141513e-01
-1.58556044e+00 -9.07739043e-01 2.24320680e-01 8.48476067e-02
-1.09279335e+00 -5.46702266e-01 7.00652003e-01 -3.27760726e-01
1.23094106e+00 -3.76910716e-02 5.31262994e-01 1.14813316e+00
1.13226958e-02 7.68372834e-01 5.94043314e-01 -3.73432636e-01
2.34463334e-01 -1.32286670e-02 2.16816157e-01 7.21930921e-01
4.78437901e-01 1.04983374e-01 -3.45817804e-01 -1.15491562e-01
6.97579801e-01 -1.23329140e-01 -2.49871314e-01 -4.23889339e-01
-1.08036864e+00 9.90024447e-01 5.87706029e-01 4.42288011e-01
-3.39416862e-01 7.86959827e-01 1.80013791e-01 7.13926196e-01
9.68153924e-02 1.28683448e+00 -9.63434041e-01 1.44654527e-01
-6.84829116e-01 2.64922917e-01 7.21041024e-01 5.58945656e-01
7.93678582e-01 5.25164425e-01 -9.88480151e-02 1.06711662e+00
1.20230466e-01 2.40141869e-01 9.34339821e-01 -1.16125011e+00
4.29815084e-01 3.88339520e-01 -2.16839239e-01 -5.77992022e-01
-6.39487505e-01 -6.77390635e-01 -6.87537909e-01 3.22877586e-01
7.60276854e-01 -7.33320713e-01 -7.48970687e-01 2.04757953e+00
-1.98377341e-01 6.69336393e-02 1.76271096e-01 7.08596945e-01
7.49033988e-01 4.24735814e-01 9.98315886e-02 3.08129758e-01
1.07275593e+00 -9.73325133e-01 -1.18703254e-01 -8.12264562e-01
7.68641770e-01 -5.66457212e-01 9.66828823e-01 5.03971696e-01
-1.34363472e+00 -5.61125457e-01 -1.21113062e+00 2.35101327e-01
-2.26150855e-01 2.81210184e-01 7.72050917e-01 6.85184956e-01
-1.21807015e+00 9.49362218e-01 -5.77452123e-01 5.19386269e-02
5.53340614e-01 6.95978999e-01 -2.41194099e-01 1.80546522e-01
-1.20643866e+00 1.28738678e+00 8.43568325e-01 -2.38998368e-01
-9.89392281e-01 -6.26911640e-01 -4.91892487e-01 4.32810754e-01
2.74410367e-01 -7.75715590e-01 1.48699594e+00 -1.23739552e+00
-1.24930024e+00 6.30854070e-01 6.45756200e-02 -8.94054115e-01
1.28339604e-01 1.15995057e-01 -1.43556058e-01 -4.91640717e-02
-5.25957644e-01 1.10714102e+00 9.23177421e-01 -1.11095643e+00
-5.58906198e-01 4.57040146e-02 7.22956732e-02 4.43523154e-02
-7.16996312e-01 -1.78537548e-01 -2.73452878e-01 -5.80421686e-01
6.28910884e-02 -9.92872894e-01 -3.67737919e-01 1.94943305e-02
-2.46859640e-01 -3.74235332e-01 6.81576207e-02 -1.29917309e-01
1.05031931e+00 -1.98854196e+00 3.18349212e-01 3.90832126e-01
3.06562603e-01 2.94998348e-01 -6.99609578e-01 -6.63275793e-02
-3.03088307e-01 2.06699371e-01 -1.53584152e-01 -5.20171225e-02
1.30211890e-01 3.78893524e-01 -1.03211440e-01 8.80098492e-02
3.04758728e-01 1.07997954e+00 -6.91943049e-01 -3.18300352e-02
-3.14752877e-01 3.53512913e-01 -7.85075843e-01 4.45628576e-02
-3.67922962e-01 -1.49512783e-01 -2.04228014e-01 1.36796623e-01
2.54543036e-01 -4.41051155e-01 3.60411674e-01 -4.88475375e-02
4.99637842e-01 3.59380513e-01 -1.20638692e+00 1.50246668e+00
-4.17962372e-01 6.02011561e-01 -1.79743305e-01 -1.21436381e+00
1.06640065e+00 3.14857185e-01 1.75740808e-01 -5.75583577e-01
2.42459550e-01 3.22068989e-01 9.33064342e-01 -1.27717704e-01
2.15246722e-01 -5.96054755e-02 1.84940740e-01 8.41369152e-01
6.67752743e-01 -8.64068717e-02 2.37626746e-01 -1.19440891e-01
1.43173742e+00 -2.97178656e-01 -6.68758452e-02 -2.96909958e-01
3.86531372e-03 -1.59812212e-01 5.37835121e-01 8.56196463e-01
-1.83839686e-02 6.45113289e-01 6.64228559e-01 -6.26307189e-01
-1.31562698e+00 -9.80814874e-01 -6.77403361e-02 1.36238539e+00
-4.34991777e-01 -9.89578813e-02 -6.80912673e-01 -5.24012387e-01
-1.85711615e-04 7.30681002e-01 -8.82314622e-01 -5.74087977e-01
-5.83449125e-01 -8.65853310e-01 8.81055951e-01 4.85976070e-01
2.49201044e-01 -1.45563841e+00 -6.98892236e-01 2.02536434e-01
1.24070555e-01 -6.19144499e-01 -1.77998975e-01 9.10425365e-01
-1.19117916e+00 -9.48634624e-01 -9.75261390e-01 -8.23442817e-01
8.06131780e-01 -3.74617614e-02 1.40798306e+00 4.54861671e-01
-1.23784557e-01 -5.64984791e-03 3.75741273e-02 -4.71012145e-01
-3.04236889e-01 7.65718400e-01 1.34744011e-02 -4.09918517e-01
3.43594432e-01 -6.44334555e-01 -7.09176600e-01 3.84953022e-01
-7.77433813e-01 8.61235186e-02 7.21772075e-01 1.03900707e+00
1.18589409e-01 7.96082895e-03 8.34138989e-01 -1.09506786e+00
1.03287244e+00 -6.29281938e-01 -4.19913948e-01 2.79990047e-01
-8.23331356e-01 5.95344186e-01 5.04775345e-01 -8.31951320e-01
-6.24585032e-01 1.47888348e-01 -2.30180815e-01 -3.82841587e-01
-3.95309217e-02 5.00793636e-01 2.88677990e-01 9.95079130e-02
1.29978013e+00 -3.14719253e-03 1.21197335e-01 -3.86760980e-01
3.40512991e-01 -7.16287196e-02 1.36024356e-01 -9.04487371e-01
7.36176550e-01 -9.00281593e-02 -2.87622660e-01 -4.27434444e-02
-7.68124223e-01 5.38145155e-02 -3.14433575e-01 8.72642249e-02
5.48826635e-01 -5.53099215e-01 -6.39607608e-01 1.94939688e-01
-1.12452829e+00 -1.04707396e+00 -3.97846162e-01 5.16620874e-01
-5.37577748e-01 -4.59327281e-01 -4.51288074e-01 -4.60353822e-01
-3.73533726e-01 -1.16640604e+00 9.93355885e-02 9.05497223e-02
-5.74829698e-01 -1.00376797e+00 7.20914453e-02 -7.75087476e-02
8.28281641e-01 -3.51159990e-01 1.12541962e+00 -9.54489529e-01
-1.13247968e-01 -1.82627402e-02 -2.58845657e-01 5.32533765e-01
-1.01434097e-01 -1.16884254e-01 -1.05489409e+00 -2.29762167e-01
-1.54680565e-01 -6.96157217e-01 1.20330071e+00 4.62563097e-01
1.55933762e+00 -4.13708925e-01 -1.47306949e-01 6.96612120e-01
1.36021721e+00 2.18554705e-01 6.42200172e-01 4.19547319e-01
3.88235450e-01 5.36324024e-01 -2.84350187e-01 2.02899665e-01
-1.48086157e-02 3.75004381e-01 5.45896351e-01 2.16232389e-01
-1.04139715e-01 6.98988140e-02 1.44575462e-01 4.76402313e-01
-3.13477099e-01 -2.77673841e-01 -1.09628451e+00 6.47446930e-01
-1.68510568e+00 -9.53836501e-01 2.90454358e-01 1.91879725e+00
1.12622011e+00 6.18758500e-01 5.08788899e-02 3.21764499e-01
4.79672849e-01 5.90601191e-02 -7.79506624e-01 -5.79817355e-01
-9.03996080e-02 7.36554623e-01 7.07830369e-01 3.57420146e-01
-7.15319872e-01 8.93589437e-01 7.24339342e+00 8.32277775e-01
-9.37158048e-01 1.64692461e-01 8.25079799e-01 -4.71525341e-01
-4.92053539e-01 -3.33192587e-01 -7.81987369e-01 3.03850025e-01
1.25497806e+00 -3.68202478e-02 7.63134360e-01 9.16694522e-01
-3.77497464e-01 4.28775132e-01 -1.14954388e+00 7.09642112e-01
-2.09290162e-01 -1.51930594e+00 -1.91173404e-02 -7.15270787e-02
8.80559266e-01 4.93358463e-01 4.64808524e-01 4.77351993e-01
1.03896964e+00 -1.43286288e+00 5.83481908e-01 4.76658255e-01
5.00215590e-01 -7.79398561e-01 5.39126158e-01 2.78040230e-01
-5.59263289e-01 -3.63272488e-01 -8.62534404e-01 -1.94372032e-02
-3.89277935e-01 4.19049770e-01 -8.08074951e-01 -2.95087367e-01
4.02045310e-01 2.18510598e-01 -7.40425229e-01 1.30719268e+00
-4.26084310e-01 8.74959350e-01 -3.16362917e-01 -3.43914002e-01
3.36464822e-01 2.70749539e-01 -6.24927552e-03 1.05793953e+00
5.53773940e-01 -1.20160706e-01 -2.49629065e-01 8.42932582e-01
-5.50834715e-01 -1.87247097e-01 -5.93977392e-01 4.91481647e-02
5.30127585e-01 1.15806139e+00 -6.33844852e-01 -1.68868780e-01
-6.90153763e-02 7.28511751e-01 8.53979111e-01 5.31641006e-01
-6.60078168e-01 -4.83748287e-01 6.38098896e-01 -1.33960629e-02
3.62846553e-01 -1.84950933e-01 -5.18759787e-01 -9.06347513e-01
-2.06290692e-01 -8.94624233e-01 3.81342381e-01 -8.56542408e-01
-1.30817115e+00 6.79259241e-01 -3.10072750e-01 -8.85344505e-01
-4.78014439e-01 -1.07857394e+00 -9.09770906e-01 7.70627379e-01
-1.23075056e+00 -4.39592898e-01 1.37784511e-01 4.35642809e-01
3.13424349e-01 -7.59266913e-01 1.02075529e+00 1.13581389e-01
-4.93605554e-01 8.69249225e-01 2.62384731e-02 3.40168178e-01
5.43896735e-01 -1.29483914e+00 4.69490349e-01 3.15782189e-01
4.78588551e-01 6.34519815e-01 6.53814852e-01 -1.37541622e-01
-1.06298792e+00 -8.89973819e-01 7.26953745e-01 -3.16179484e-01
5.97113848e-01 -1.33792460e-01 -9.41043496e-01 8.90188396e-01
3.96902859e-01 4.95960191e-02 6.74431324e-01 4.71464545e-01
-6.42951369e-01 -6.09641857e-02 -1.08640718e+00 7.29363978e-01
1.08722937e+00 -1.16971746e-01 -8.03976178e-01 3.58425170e-01
7.48172998e-01 -1.12730123e-01 -7.14933038e-01 2.29545176e-01
5.46119869e-01 -7.20382452e-01 1.14630806e+00 -9.69801009e-01
7.81102121e-01 9.68093798e-02 -1.20992273e-01 -1.87336719e+00
-5.99377513e-01 -1.93672806e-01 -2.31280535e-01 7.38167763e-01
1.20182538e+00 -9.59318042e-01 1.11652792e+00 7.80611634e-01
-2.84882411e-02 -9.72259343e-01 -7.96980798e-01 -7.97508478e-01
6.02119088e-01 -6.50164962e-01 6.63244724e-01 8.30357492e-01
-2.13036597e-01 4.66347039e-01 -9.81544927e-02 -2.13444501e-01
3.94034445e-01 -2.24887073e-01 2.73129195e-01 -1.48116136e+00
-7.36000180e-01 -1.01673067e+00 -1.56773537e-01 -7.75376201e-01
3.66323680e-01 -1.25012732e+00 -1.59158304e-01 -1.51510322e+00
1.03640743e-02 -7.77647555e-01 -7.76798487e-01 9.13558185e-01
2.86484547e-02 3.97789329e-01 2.17878431e-01 1.52266935e-01
-4.16237175e-01 2.36952275e-01 1.00728965e+00 -3.31147969e-01
8.72091278e-02 1.90712158e-02 -1.09092450e+00 7.84924090e-01
1.40447152e+00 -7.72888362e-01 -8.03198516e-01 -1.06219208e+00
6.30791187e-01 -3.43504757e-01 4.64046523e-02 -1.34003365e+00
2.62315691e-01 -1.70672163e-01 8.62845242e-01 8.02341402e-02
4.63461608e-01 -7.02688992e-01 8.28658044e-03 8.01412761e-01
-9.53021169e-01 2.31937259e-01 3.25297415e-01 8.41457099e-02
1.36654571e-01 -1.09989393e+00 8.96871567e-01 -3.29610974e-01
-4.34499323e-01 2.56155789e-01 -4.81011182e-01 2.33973190e-01
6.18588269e-01 3.55035588e-02 -4.57952499e-01 -1.83651879e-01
-9.61101711e-01 1.57908469e-01 7.01444224e-02 2.32837826e-01
6.77141249e-01 -1.35207343e+00 -7.81445682e-01 1.25112265e-01
-3.31283994e-02 -3.05462331e-01 -1.69174224e-01 3.04510385e-01
-3.41097653e-01 3.36314052e-01 -5.21991909e-01 -2.72958744e-02
-8.88979435e-01 2.84190685e-01 6.63667142e-01 -2.99074173e-01
-2.02378124e-01 1.22226548e+00 -1.79198366e-02 -8.16837907e-01
3.56640399e-01 -2.61435539e-01 -2.26354182e-01 1.05846100e-01
5.27471423e-01 -1.17502138e-02 9.88199934e-02 -4.40278761e-02
2.67915037e-02 2.35932872e-01 -2.49222219e-01 -2.62304485e-01
1.53529179e+00 5.94449401e-01 3.86276066e-01 2.73670375e-01
1.08771658e+00 -7.19268858e-01 -1.30165911e+00 -4.41045940e-01
2.61598378e-02 -1.14207402e-01 9.42230448e-02 -9.32101905e-01
-1.65048778e+00 9.98088181e-01 6.07542813e-01 2.44558454e-01
8.41600895e-01 -1.86000779e-01 4.91289198e-01 1.06753933e+00
2.09869340e-01 -1.11982477e+00 4.43282574e-01 8.06056261e-01
8.85621786e-01 -1.10461617e+00 -5.97846434e-02 3.20954263e-01
-6.98014557e-01 1.13315356e+00 8.36085200e-01 -6.65913224e-01
6.15345120e-01 2.66362101e-01 -1.66066065e-01 -1.33298010e-01
-1.37419176e+00 -2.31899872e-01 2.52063006e-01 6.62910223e-01
5.46427190e-01 -1.58362407e-02 6.03921264e-02 6.56304181e-01
-4.22599703e-01 -1.94118302e-02 2.87873745e-01 4.80593950e-01
-7.83384383e-01 -1.07224643e+00 6.55715493e-03 8.79237235e-01
-5.20461798e-01 -4.20958281e-01 -2.60172337e-01 3.95073771e-01
-8.49203318e-02 6.09149754e-01 2.27109879e-01 -5.73252797e-01
1.45595759e-01 5.72729051e-01 6.46237373e-01 -6.09105468e-01
-8.70385885e-01 -4.37628716e-01 2.09735259e-01 -2.26721480e-01
1.41351461e-01 -2.59216070e-01 -1.31409025e+00 -4.26971674e-01
-6.55934736e-02 4.62280028e-02 7.40924239e-01 7.06998765e-01
1.39361426e-01 8.14283311e-01 2.46075630e-01 -8.35377753e-01
-8.41443956e-01 -1.07203555e+00 -2.32301682e-01 1.83128104e-01
-3.81021127e-02 -5.23121476e-01 -3.99223566e-01 -1.78300932e-01] | [8.58804702758789, 3.2812137603759766] |
57249e9f-42ef-4f3e-937c-e4de6c2c109d | video-question-answering-using-clip-guided | 2303.03131 | null | https://arxiv.org/abs/2303.03131v2 | https://arxiv.org/pdf/2303.03131v2.pdf | Video Question Answering Using CLIP-Guided Visual-Text Attention | Cross-modal learning of video and text plays a key role in Video Question Answering (VideoQA). In this paper, we propose a visual-text attention mechanism to utilize the Contrastive Language-Image Pre-training (CLIP) trained on lots of general domain language-image pairs to guide the cross-modal learning for VideoQA. Specifically, we first extract video features using a TimeSformer and text features using a BERT from the target application domain, and utilize CLIP to extract a pair of visual-text features from the general-knowledge domain through the domain-specific learning. We then propose a Cross-domain Learning to extract the attention information between visual and linguistic features across the target domain and general domain. The set of CLIP-guided visual-text features are integrated to predict the answer. The proposed method is evaluated on MSVD-QA and MSRVTT-QA datasets, and outperforms state-of-the-art methods. | ['Xudong Jiang', 'Jianfeng Ren', 'Chenglin Yao', 'Weikai Kong', 'Shuhong Ye'] | 2023-03-06 | null | null | null | null | ['video-question-answering', 'general-knowledge'] | ['computer-vision', 'miscellaneous'] | [ 3.77432220e-02 -6.44387960e-01 -2.20087931e-01 -5.15776694e-01
-1.35608017e+00 -6.09997332e-01 6.46399021e-01 -3.68637174e-01
-5.55333972e-01 4.49796289e-01 3.79475892e-01 -1.90021709e-01
1.52898401e-01 -3.39720845e-01 -7.47409940e-01 -5.13305604e-01
4.36405450e-01 3.46460968e-01 6.08965397e-01 -1.62566200e-01
2.71823794e-01 -4.69714515e-02 -1.37695873e+00 1.08884263e+00
7.50724256e-01 1.33286452e+00 7.18729734e-01 8.60619247e-01
-7.13772893e-01 1.30123246e+00 -5.61968207e-01 -3.73269141e-01
-6.13964722e-02 -6.81126654e-01 -9.77794588e-01 3.95217806e-01
7.99049199e-01 -7.00203836e-01 -7.19981134e-01 7.83935368e-01
4.17320728e-01 3.37111235e-01 8.33390653e-01 -1.35999310e+00
-9.76689041e-01 -6.83665127e-02 -8.11699629e-01 7.32180655e-01
5.76600194e-01 4.83562827e-01 9.13357794e-01 -1.03254437e+00
8.08499932e-01 1.63129067e+00 1.28093513e-03 5.81780314e-01
-6.11751020e-01 -7.35219955e-01 5.70042193e-01 8.53633523e-01
-1.14286661e+00 -4.44387138e-01 1.10679376e+00 -6.01040304e-01
6.75159156e-01 -1.62648752e-01 3.93735707e-01 1.24982369e+00
2.14524761e-01 1.42044294e+00 7.16766298e-01 -1.63274229e-01
-2.11962145e-02 1.50870711e-01 -7.01689273e-02 9.36870754e-01
-5.03664613e-01 -1.49665222e-01 -7.14106858e-01 3.65388431e-02
5.97384572e-01 1.54425636e-01 -5.37586510e-01 -5.06465614e-01
-1.28559983e+00 8.09434593e-01 1.96554288e-01 4.25994515e-01
-3.80713224e-01 1.07383691e-02 7.62126982e-01 4.58930492e-01
2.26405367e-01 -1.59449562e-01 -5.66374838e-01 -1.63864762e-01
-8.23238671e-01 -2.99705062e-02 3.38123471e-01 1.07068765e+00
7.81085372e-01 1.55212209e-01 -6.77558959e-01 7.89800406e-01
4.37872559e-01 7.65473187e-01 5.61273575e-01 -9.35284853e-01
1.10002697e+00 5.60432136e-01 1.08020352e-02 -1.04036033e+00
7.08132312e-02 1.28998816e-01 -5.88635504e-01 -7.96849579e-02
3.82525831e-01 -1.19668096e-01 -1.00228679e+00 1.39680398e+00
2.22642675e-01 2.20488906e-01 4.40811038e-01 1.31874609e+00
1.33749330e+00 1.25428271e+00 2.57767886e-01 -2.73283899e-01
1.43708801e+00 -1.29299271e+00 -8.08617949e-01 -4.23899025e-01
4.20016408e-01 -7.47859418e-01 1.42419469e+00 2.99020648e-01
-1.10840774e+00 -9.98310685e-01 -7.60971546e-01 -4.36587214e-01
-2.18868762e-01 4.13031071e-01 -5.08588227e-03 1.21296972e-01
-7.18330383e-01 -2.34328628e-01 -3.07472140e-01 -2.76023149e-01
5.65142572e-01 1.22694835e-01 -4.03460622e-01 -7.45542169e-01
-1.38020146e+00 5.47408879e-01 4.24128950e-01 -2.52804011e-02
-1.47132552e+00 -4.70388710e-01 -1.02078748e+00 -1.88849524e-01
6.15499139e-01 -8.66176188e-01 1.17047179e+00 -1.62501574e+00
-1.38155997e+00 9.55691516e-01 -4.41729248e-01 -1.20226115e-01
2.40278140e-01 -1.29702285e-01 -6.45289600e-01 1.01710153e+00
1.25164688e-01 7.38707185e-01 1.46931410e+00 -1.31480277e+00
-8.07247996e-01 -3.91253024e-01 2.95723915e-01 5.61098993e-01
-3.63283664e-01 8.57640058e-03 -1.04938662e+00 -3.97730172e-01
-3.24652314e-01 -3.18957269e-01 3.43448102e-01 -5.97601235e-02
1.24772303e-01 -3.35401922e-01 1.31249440e+00 -1.14600873e+00
1.03899968e+00 -2.24425650e+00 3.30933809e-01 -3.67325872e-01
2.00972512e-01 4.17584658e-01 -6.72669888e-01 1.93451896e-01
4.61832657e-02 -4.24824834e-01 5.36922850e-02 2.48588528e-02
-4.27454740e-01 1.56893402e-01 -3.33950788e-01 3.29119682e-01
5.71620882e-01 1.15695584e+00 -1.05649638e+00 -1.05359805e+00
3.01229298e-01 2.46413603e-01 -5.15717924e-01 6.60166085e-01
-5.86273015e-01 3.82403642e-01 -9.40635741e-01 6.77531183e-01
5.95042825e-01 -3.21801275e-01 -5.60037829e-02 -5.93813539e-01
2.40632102e-01 -1.23373732e-01 -7.48789430e-01 2.14475346e+00
-4.34509188e-01 8.09650600e-01 2.19162226e-01 -1.25937498e+00
7.67852783e-01 2.13673279e-01 2.32537061e-01 -1.07960629e+00
2.94188172e-01 -8.63417685e-02 -2.20475808e-01 -1.29175448e+00
3.27278674e-01 9.95265469e-02 -9.05461423e-03 3.01437765e-01
6.40191555e-01 5.07078692e-03 9.70323384e-02 5.02220392e-01
7.78727651e-01 2.61844546e-01 1.13801487e-01 7.32210204e-02
1.19273460e+00 2.24383280e-01 2.82353759e-01 2.87961006e-01
-4.81594622e-01 4.75867391e-01 4.98757780e-01 -1.38067886e-01
-8.00856769e-01 -9.87534106e-01 5.55368841e-01 1.49063170e+00
3.68092984e-01 -1.32938087e-01 -5.77861428e-01 -1.25061870e+00
-7.26703778e-02 4.81248528e-01 -5.53197622e-01 -2.20464438e-01
-4.74286854e-01 2.76082784e-01 2.10493833e-01 5.07275283e-01
8.17871332e-01 -1.13115931e+00 -3.10298651e-01 -7.71069974e-02
-5.49269974e-01 -1.32370961e+00 -9.16774571e-01 -2.98232913e-01
-5.12717307e-01 -1.10278976e+00 -1.05401552e+00 -1.31722116e+00
3.56456488e-01 6.99907660e-01 1.22807455e+00 -2.04014704e-02
9.97462049e-02 1.00257337e+00 -6.04735911e-01 1.05051644e-01
-1.27659142e-01 -2.77021736e-01 -3.57444644e-01 4.22901601e-01
5.71972966e-01 -1.24537602e-01 -5.92333853e-01 2.99440086e-01
-9.20099497e-01 -2.89577302e-02 7.03579128e-01 1.07666636e+00
6.37235105e-01 -3.38175416e-01 6.19176686e-01 -4.34238642e-01
5.46519458e-01 -6.59245133e-01 -4.57759708e-01 6.05588198e-01
2.78479338e-01 3.61956842e-02 6.02825820e-01 -6.24483824e-01
-1.31183732e+00 1.47361830e-02 1.48002461e-01 -1.21255922e+00
-2.11603060e-01 4.70460981e-01 -5.84706247e-01 1.86772615e-01
2.86576539e-01 5.53615868e-01 -4.12211530e-02 -1.47093713e-01
6.01371467e-01 8.27925980e-01 6.27601564e-01 -5.66711187e-01
6.27935708e-01 4.01931852e-01 -4.93269801e-01 -8.34149420e-01
-8.26684415e-01 -8.01601768e-01 -5.89098811e-01 -5.21283388e-01
1.49714231e+00 -1.17769063e+00 -5.39245546e-01 3.20811152e-01
-1.43828893e+00 -2.19566539e-01 1.37451455e-01 4.81686234e-01
-6.28589869e-01 6.11501336e-01 -3.89924347e-01 -7.69209862e-01
-1.91314578e-01 -1.25111890e+00 1.30173111e+00 1.92153364e-01
5.09590030e-01 -9.28679705e-01 -2.86777377e-01 7.87304819e-01
-2.11298287e-01 -3.38383794e-01 7.52232134e-01 -6.84780777e-01
-7.73791850e-01 2.93590933e-01 -7.06753671e-01 4.78411108e-01
-3.20386887e-02 -2.53085822e-01 -7.04561114e-01 -2.66198397e-01
1.30931154e-01 -8.07410061e-01 9.99130249e-01 3.12458932e-01
1.30781376e+00 -1.67195275e-01 -2.49231502e-01 3.67356658e-01
1.33860123e+00 4.25150663e-01 5.85800707e-01 3.70268635e-02
8.26382935e-01 5.83858669e-01 1.22727382e+00 1.17745005e-01
6.52834058e-01 4.62959379e-01 3.75597209e-01 -2.08165050e-02
-1.03577957e-01 -3.36498618e-01 6.86304569e-01 8.74472380e-01
3.55871409e-01 -3.58944148e-01 -9.05736089e-01 8.55403781e-01
-1.82515383e+00 -1.24979615e+00 1.16324916e-01 1.62505555e+00
6.44040048e-01 -1.26137108e-01 2.27819547e-01 -2.30871379e-01
8.29544485e-01 3.16457599e-01 -6.56064570e-01 -6.25839457e-02
4.00739796e-02 -2.21433342e-01 3.25273722e-02 2.81873941e-01
-1.18895090e+00 1.12417090e+00 5.30380964e+00 1.10097611e+00
-1.11886239e+00 3.20083469e-01 5.50479770e-01 1.62075192e-03
-3.59248191e-01 -3.62227350e-01 -5.88586509e-01 4.11334544e-01
7.07321346e-01 -1.28302619e-01 2.47740254e-01 7.94800878e-01
1.43980548e-01 -4.54932414e-02 -9.58206356e-01 1.38990998e+00
8.04697156e-01 -1.36456728e+00 4.96272147e-01 -5.08012772e-01
6.32941484e-01 -2.09753215e-01 1.70324162e-01 6.28472626e-01
-1.15317971e-01 -6.51729643e-01 5.50306618e-01 5.97610533e-01
9.38325286e-01 -6.81905150e-01 5.35539746e-01 2.58109093e-01
-1.49147630e+00 -3.81054044e-01 -3.60392362e-01 2.89323062e-01
2.83437580e-01 3.27837886e-03 -4.17971194e-01 8.31508756e-01
9.45194900e-01 1.17078578e+00 -3.96549463e-01 5.83958089e-01
1.39566079e-01 5.66361845e-01 3.76190871e-01 2.05242448e-02
5.48628569e-01 -1.46705508e-01 3.24087441e-01 1.18541324e+00
1.21515676e-01 2.79206306e-01 2.26045907e-01 6.11817241e-01
-4.47752103e-02 2.24435985e-01 -4.98875022e-01 -2.96240300e-01
2.03712389e-01 1.02179456e+00 -1.46096304e-01 -6.21531546e-01
-1.06800377e+00 1.37589943e+00 1.60577506e-01 7.73244202e-01
-9.78204668e-01 -3.55682760e-01 4.39286411e-01 -2.32372165e-01
7.31847525e-01 -8.67914632e-02 3.49192411e-01 -1.43474925e+00
2.13820357e-02 -1.20334804e+00 8.26579154e-01 -1.45437753e+00
-1.55779922e+00 5.54513693e-01 -2.17280183e-02 -1.58856583e+00
-1.65373042e-01 -8.08290243e-01 -4.60130304e-01 8.68387163e-01
-1.80955839e+00 -1.40293252e+00 -4.82256532e-01 1.48526835e+00
1.20680785e+00 -6.38605833e-01 1.28047138e-01 4.51388299e-01
-2.06835791e-01 5.55817187e-01 7.72801321e-03 3.85260582e-01
9.88213658e-01 -7.14212358e-01 -1.29456997e-01 7.15624452e-01
2.12014094e-01 3.09073448e-01 1.61265522e-01 -5.27462542e-01
-1.56438184e+00 -1.20573854e+00 4.05383170e-01 -3.75993937e-01
7.41282225e-01 -1.48010239e-01 -9.15498614e-01 6.75406516e-01
6.48175538e-01 -4.90216957e-03 2.89468795e-01 -4.98397797e-01
-5.05109370e-01 -2.88803369e-01 -8.80575657e-01 3.89240980e-01
8.15943182e-01 -9.84496117e-01 -1.18915021e+00 1.24504231e-01
1.11743569e+00 -3.31511855e-01 -5.23501575e-01 1.80851117e-01
3.95667553e-01 -6.80861354e-01 1.07053304e+00 -9.10923421e-01
7.05788553e-01 -3.68861079e-01 -3.99966121e-01 -1.05760741e+00
-8.05284232e-02 -2.23313704e-01 -5.38826399e-02 1.41569281e+00
5.87162413e-02 9.03147180e-03 4.84553218e-01 -1.74366191e-01
-2.07568198e-01 -3.42683643e-01 -8.61663640e-01 -4.23293173e-01
1.29396424e-01 -3.74663740e-01 1.62471637e-01 1.02291715e+00
-1.25607967e-01 8.10751975e-01 -5.70563376e-01 -1.75459031e-02
2.38860920e-01 2.15454400e-01 8.43335450e-01 -8.97957802e-01
-2.35828280e-01 -1.60825759e-01 -3.53301346e-01 -1.67442787e+00
5.34568071e-01 -7.23221481e-01 4.46024211e-03 -1.50109303e+00
4.34290320e-01 3.36584687e-01 -3.99487346e-01 7.01188073e-02
-3.31098884e-01 9.12420377e-02 2.90143073e-01 -3.12869698e-02
-1.26334226e+00 7.70978272e-01 1.75870907e+00 -5.65410554e-01
-1.03628241e-01 -4.93544847e-01 -3.49025607e-01 5.10568857e-01
3.96700799e-01 -1.32795691e-01 -7.38502324e-01 -7.31355369e-01
-2.19953075e-01 5.87924898e-01 4.21608776e-01 -8.00027013e-01
1.27770260e-01 -3.94200414e-01 5.08908153e-01 -1.07460916e+00
5.83488286e-01 -9.38373446e-01 -6.87800169e-01 1.99234515e-01
-4.32529658e-01 8.00697953e-02 3.06434989e-01 9.27653432e-01
-6.52902603e-01 7.59281814e-02 5.47302246e-01 -2.26173013e-01
-1.41715598e+00 4.91713524e-01 -2.92691797e-01 5.87531686e-01
1.13810027e+00 -6.85780048e-02 -6.20242417e-01 -6.68689907e-01
-5.52593589e-01 7.60052025e-01 1.20660439e-01 7.75683999e-01
1.13775122e+00 -1.42788792e+00 -6.79801166e-01 1.81260016e-02
5.75445771e-01 -3.15592617e-01 8.46382558e-01 7.80968249e-01
-3.70183915e-01 5.00004232e-01 -4.57325190e-01 -9.65793788e-01
-1.45083714e+00 1.00566161e+00 3.83075595e-01 -8.64982381e-02
-2.88513213e-01 9.28623497e-01 7.34187722e-01 -3.72680053e-02
2.85938948e-01 -8.32600743e-02 -5.04679322e-01 1.25973850e-01
7.24416137e-01 -4.71512042e-02 -4.26935226e-01 -8.80970538e-01
-4.43717003e-01 8.06646883e-01 -1.29955545e-01 -2.16088623e-01
8.87809217e-01 -5.73785901e-01 1.55507028e-01 4.75217223e-01
1.55602372e+00 -2.42751271e-01 -1.36692309e+00 -5.16161799e-01
-3.41317534e-01 -6.14431500e-01 -4.03066017e-02 -7.63249636e-01
-1.28327334e+00 1.47042096e+00 6.74889147e-01 -3.50294262e-01
1.36223865e+00 3.64862949e-01 8.89182389e-01 2.26839766e-01
-2.28672773e-02 -1.08406281e+00 8.70669901e-01 4.94092911e-01
1.08711970e+00 -1.50767565e+00 -2.79094040e-01 -3.60393971e-02
-1.27824533e+00 1.02049804e+00 1.03963804e+00 -4.41346169e-02
4.69401628e-01 -4.47032958e-01 3.28107536e-01 -1.36579007e-01
-9.13483143e-01 -5.43857813e-01 7.09927619e-01 9.36544657e-01
6.51825145e-02 -4.68444049e-01 -1.63421109e-01 7.75631607e-01
5.44835448e-01 1.59153879e-01 1.37096763e-01 6.93972111e-01
-5.26600242e-01 -8.65716934e-01 -3.54504079e-01 3.16258192e-01
-1.98482215e-01 -1.02289923e-01 -3.29042017e-01 7.80876875e-01
2.07311615e-01 1.19331288e+00 2.35672742e-01 -6.25290573e-01
2.06403151e-01 2.30771378e-01 4.36149359e-01 -2.97176331e-01
-3.23773146e-01 1.64913639e-01 -2.37141717e-02 -6.83844447e-01
-7.47033834e-01 -6.30744278e-01 -1.11584234e+00 4.57986295e-02
-1.49017833e-02 2.61555612e-01 1.71096325e-01 1.16042829e+00
3.38179171e-01 5.53697944e-01 5.53004742e-01 -6.23594999e-01
-6.81544542e-02 -6.85691893e-01 -3.86766136e-01 5.47501504e-01
6.94453359e-01 -7.65948951e-01 -2.71412581e-01 4.15299118e-01] | [10.372069358825684, 1.0361772775650024] |
b52bb1de-b64f-4e30-8cfd-3eae986eea43 | cu-ud-text-mining-drug-and-chemical-protein | 2112.03004 | null | https://arxiv.org/abs/2112.03004v1 | https://arxiv.org/pdf/2112.03004v1.pdf | CU-UD: text-mining drug and chemical-protein interactions with ensembles of BERT-based models | Identifying the relations between chemicals and proteins is an important text mining task. BioCreative VII track 1 DrugProt task aims to promote the development and evaluation of systems that can automatically detect relations between chemical compounds/drugs and genes/proteins in PubMed abstracts. In this paper, we describe our submission, which is an ensemble system, including multiple BERT-based language models. We combine the outputs of individual models using majority voting and multilayer perceptron. Our system obtained 0.7708 in precision and 0.7770 in recall, for an F1 score of 0.7739, demonstrating the effectiveness of using ensembles of BERT-based language models for automatically detecting relations between chemicals and proteins. Our code is available at https://github.com/bionlplab/drugprot_bcvii. | ['Yifan Peng', 'K. Vijay-Shanker', 'Mehmet Efruz Karabulut'] | 2021-11-11 | null | null | null | null | ['drugprot'] | ['natural-language-processing'] | [ 1.48409024e-01 4.71893661e-02 -4.48361337e-01 -2.17036650e-01
-6.37370348e-01 -6.25339508e-01 5.59231281e-01 9.06962454e-01
-1.42669469e-01 1.34146309e+00 5.99151477e-02 -6.03674710e-01
-2.71471351e-01 -4.86629784e-01 -7.98673153e-01 -7.73221433e-01
-7.57041797e-02 5.77339113e-01 -2.51483858e-01 4.89952825e-02
4.75295261e-03 5.46337605e-01 -9.67150152e-01 8.22088242e-01
8.73642802e-01 4.60339814e-01 3.55513617e-02 6.79485917e-01
-2.96762474e-02 9.27405715e-01 -5.06715536e-01 -4.59335566e-01
-3.14705402e-01 -2.99684405e-01 -8.16219985e-01 -7.18011498e-01
5.80827072e-02 3.81154746e-01 -2.93186814e-01 9.79319453e-01
5.94111502e-01 -4.71905202e-01 9.20435667e-01 -9.57607448e-01
-5.29292464e-01 8.77201319e-01 -4.20101523e-01 7.57946894e-02
7.16955185e-01 3.59650105e-02 1.07649088e+00 -8.86671305e-01
7.54123509e-01 1.03153729e+00 5.14862716e-01 4.65522617e-01
-1.27681923e+00 -8.48025858e-01 -3.19391280e-01 2.07996160e-01
-1.49524713e+00 -3.66681308e-01 9.45662484e-02 -7.78569520e-01
1.71146905e+00 5.23899913e-01 3.46074611e-01 7.66855001e-01
9.26293015e-01 5.59443235e-01 9.39889252e-01 -3.84168148e-01
9.24588144e-02 2.59888530e-01 3.90397042e-01 7.49293029e-01
6.10202432e-01 -1.16506703e-01 -5.52489400e-01 -7.86737442e-01
6.83250418e-03 -2.12496258e-02 -2.70078570e-01 3.24412435e-01
-1.04339004e+00 5.99731565e-01 4.10447568e-01 4.68180478e-01
-4.93800998e-01 -1.59494936e-01 3.73335212e-01 1.08513132e-01
5.54106712e-01 1.06470764e+00 -9.09317493e-01 3.64509583e-01
-3.48821610e-01 3.33024524e-02 1.09943736e+00 8.56935740e-01
2.66402543e-01 -7.82127619e-01 -3.54178511e-02 9.11406755e-01
3.10480416e-01 1.47822902e-01 5.37083864e-01 -3.51022333e-01
2.48504896e-02 8.18708360e-01 -2.93249190e-02 -7.13505924e-01
-7.86132812e-01 -1.20272376e-01 -5.93870759e-01 -3.16761672e-01
7.94602111e-02 -3.66319031e-01 -6.78348601e-01 1.29819858e+00
2.15870604e-01 -8.41754824e-02 3.70887846e-01 3.22224736e-01
1.37258089e+00 3.93424511e-01 9.66693580e-01 -2.11164907e-01
1.50711787e+00 -7.27741241e-01 -8.98290813e-01 2.34957471e-01
1.03045011e+00 -8.60222578e-01 7.09555671e-02 4.86015469e-01
-8.61900926e-01 -2.08496973e-01 -1.12356496e+00 -1.31997749e-01
-9.96110737e-01 3.92572939e-01 8.79094899e-01 1.36861309e-01
-6.71456516e-01 9.09426987e-01 -7.08491206e-01 -3.91347170e-01
3.83765221e-01 7.45797098e-01 -5.83424449e-01 1.03600360e-01
-1.39383531e+00 1.20491827e+00 4.77142900e-01 -2.59415358e-01
-2.83272386e-01 -8.62486303e-01 -4.69414085e-01 -1.53641298e-01
-2.06116810e-01 -7.03026712e-01 1.05745780e+00 -2.33585745e-01
-8.81955147e-01 8.67531359e-01 -4.16660964e-01 -3.72408211e-01
6.42312095e-02 -3.29365730e-01 -5.69829226e-01 -3.77222598e-02
2.12584987e-01 6.88086390e-01 -4.98683453e-01 -5.65919280e-01
-6.88262403e-01 -4.64481056e-01 -2.37487629e-01 8.78197551e-02
-1.38069123e-01 2.80148685e-01 -2.53096968e-01 -8.80262628e-02
-2.98865110e-01 -9.20570314e-01 -3.66736382e-01 -3.53033066e-01
-8.75671089e-01 -6.36346877e-01 1.64330289e-01 -7.28911161e-01
1.00227475e+00 -1.60875094e+00 -1.22586243e-01 6.93023801e-02
4.60776210e-01 2.06650123e-01 -3.99886742e-02 8.08476865e-01
-7.36093700e-01 3.22173923e-01 -7.08669871e-02 4.14050192e-01
-3.66321683e-01 -2.77126878e-01 4.00399715e-02 4.73711103e-01
6.06989145e-01 9.65015233e-01 -1.03610063e+00 -2.42244005e-01
8.68523940e-02 7.03293860e-01 4.66928706e-02 -3.29178981e-02
-3.85347873e-01 3.24745834e-01 -5.59676647e-01 9.06284511e-01
4.48391348e-01 -5.02133191e-01 7.65123963e-01 -2.11291686e-01
-5.75693138e-02 4.98604327e-01 -5.00292659e-01 1.16089630e+00
2.06426293e-01 3.42850626e-01 -5.97174883e-01 -7.43883193e-01
8.96096885e-01 6.51597559e-01 9.13784325e-01 -2.33066738e-01
1.78883597e-01 2.40746975e-01 2.65785903e-01 -6.90777659e-01
-2.37667799e-01 1.13069035e-01 3.09373558e-01 -1.48830637e-01
2.16127843e-01 1.81755140e-01 5.05993783e-01 7.10768551e-02
1.63773119e+00 1.98274940e-01 6.80319428e-01 -2.86135316e-01
6.80256128e-01 3.53481323e-01 4.57136780e-01 3.54427457e-01
1.16982140e-01 -6.64303377e-02 6.23695672e-01 -4.56445873e-01
-6.37359262e-01 -4.10716027e-01 -7.00474322e-01 6.29170299e-01
-2.46362299e-01 -8.19492936e-01 -4.20200616e-01 -7.11337566e-01
2.16749579e-01 6.56310558e-01 -5.19742548e-01 -1.15448736e-01
-1.28644168e-01 -1.49025249e+00 8.00845325e-01 9.51035544e-02
-1.57719240e-01 -8.16606939e-01 1.45131782e-01 5.06524920e-01
-1.69495377e-03 -9.99830961e-01 1.45852771e-02 8.56048226e-01
-6.53368056e-01 -1.59587741e+00 -4.35838073e-01 -7.11183906e-01
5.45692801e-01 -1.95787787e-01 9.49821174e-01 -3.06777637e-02
-6.69312716e-01 -4.31442738e-01 -5.19935600e-02 -1.06877244e+00
-5.08098722e-01 -9.79163200e-02 7.80603141e-02 -6.60758734e-01
1.05176175e+00 -3.98871511e-01 -4.97946411e-01 7.66687691e-02
-4.93140817e-01 1.45533755e-01 7.94632614e-01 6.92256808e-01
6.60650373e-01 -1.56404167e-01 6.61591709e-01 -1.09417784e+00
6.78379714e-01 -9.51589882e-01 -5.46283901e-01 4.46242005e-01
-8.59567881e-01 5.82984164e-02 3.02190661e-01 -2.83386230e-01
-3.67186993e-01 6.93310440e-01 -3.52722943e-01 4.10691410e-01
-4.27744091e-01 8.87836993e-01 -1.46335438e-01 -6.30933465e-03
6.97376072e-01 -1.56185985e-01 -2.24473059e-01 -5.25848269e-01
1.06607772e-01 9.56574619e-01 5.89890331e-02 -2.62083054e-01
-8.07030797e-02 -1.21227659e-01 2.26991042e-01 -6.19732320e-01
-5.02623200e-01 -7.20781446e-01 -4.27277446e-01 2.80045122e-01
7.96956837e-01 -1.02544069e+00 -1.16326559e+00 6.47762939e-02
-1.19787288e+00 6.40137494e-02 4.95430857e-01 8.10129881e-01
-1.07965365e-01 2.11108312e-01 -1.03674650e+00 -3.94743025e-01
-1.01119518e+00 -1.18504179e+00 8.13564599e-01 9.69646033e-03
-7.80090332e-01 -7.60735273e-01 5.52075326e-01 4.06121463e-01
-3.68805587e-01 3.59626919e-01 1.09822965e+00 -1.25526214e+00
-2.10281685e-02 -4.37525332e-01 -1.40001267e-01 -1.81604773e-01
4.68156397e-01 3.93997103e-01 -9.29121435e-01 1.28450498e-01
-6.89808786e-01 -1.58887640e-01 7.83697128e-01 4.31908906e-01
9.64886010e-01 -3.78870368e-01 -1.04606032e+00 1.62634894e-01
1.32709074e+00 8.43532860e-01 4.73306865e-01 1.38192475e-01
5.81596076e-01 6.08804405e-01 5.40493608e-01 1.00497954e-01
-6.80126920e-02 4.54574168e-01 -3.46233919e-02 -2.52627045e-01
1.54730961e-01 1.59184989e-02 4.72880267e-02 2.03513384e-01
-1.63730919e-01 -3.67344081e-01 -1.24783587e+00 2.81637341e-01
-1.85644269e+00 -6.19004846e-01 -6.39673769e-01 1.59127617e+00
1.46242809e+00 -8.11486468e-02 -9.42023396e-02 -2.15593934e-01
6.67010725e-01 -7.93768406e-01 -6.05499744e-01 -6.29463673e-01
-7.23319575e-02 4.70847100e-01 6.13736987e-01 3.81775647e-01
-1.32895660e+00 8.46499920e-01 6.19890642e+00 7.78059721e-01
-9.16844249e-01 -3.76772761e-01 8.39069963e-01 7.42917284e-02
1.30034983e-01 -2.48321861e-01 -9.52853203e-01 3.81882668e-01
1.56947207e+00 -2.78603584e-01 -2.26918280e-01 6.76683068e-01
3.47545773e-01 2.56147534e-02 -1.29269469e+00 5.34010112e-01
-3.97245914e-01 -1.65680242e+00 -6.60401881e-02 3.09636861e-01
4.84206408e-01 3.97479862e-01 -3.32021058e-01 -5.22818556e-03
6.73880637e-01 -1.28636861e+00 -2.98765637e-02 5.83242178e-01
5.82849979e-01 -4.91758645e-01 9.75369155e-01 8.19629431e-02
-7.14357436e-01 2.41883248e-01 -3.48597497e-01 1.50543720e-01
-3.57129157e-01 1.22079444e+00 -1.59854519e+00 6.95601344e-01
6.51399195e-01 9.67009127e-01 -4.36478227e-01 1.20327783e+00
-3.34098041e-01 4.10768420e-01 2.19476633e-02 -4.11712378e-01
-1.91686139e-01 5.75133748e-02 1.85794860e-01 1.58650184e+00
-6.48095310e-02 4.12982821e-01 1.68492943e-01 5.32899261e-01
-2.02262506e-01 6.01468563e-01 -5.80934167e-01 -5.33556879e-01
4.96718347e-01 1.37138224e+00 -4.35022444e-01 -4.34270144e-01
-2.84755707e-01 4.85966802e-01 1.26111746e-01 1.04081452e-01
-6.01777256e-01 -5.19485474e-01 7.24338949e-01 -1.22777119e-01
-2.84362942e-01 2.73982942e-01 -3.24079841e-01 -7.33034253e-01
-5.13607800e-01 -9.58235145e-01 6.05298936e-01 -8.17717373e-01
-1.45901191e+00 6.72279000e-01 -2.75813073e-01 -1.06820083e+00
5.33515289e-02 -9.10867095e-01 -6.28711516e-03 9.32930887e-01
-9.53740358e-01 -1.05859852e+00 3.31082970e-01 -6.94235712e-02
1.50066763e-01 -3.30353878e-03 1.56365967e+00 4.14100826e-01
-8.18497539e-01 3.49664986e-01 3.48508567e-01 1.01980835e-01
1.15444827e+00 -1.24169123e+00 3.91051590e-01 -1.27628952e-01
-1.10407509e-01 1.12969673e+00 7.53420532e-01 -1.03812778e+00
-1.12724054e+00 -1.25237548e+00 1.47972119e+00 -6.20601296e-01
7.76645482e-01 -1.90964743e-01 -7.52399027e-01 5.09542584e-01
2.68694371e-01 -5.08275747e-01 1.52369535e+00 1.74638987e-01
-2.19387591e-01 2.64167309e-01 -1.22284210e+00 5.15073061e-01
6.32026494e-01 -1.36525840e-01 -2.94507593e-01 9.97429311e-01
5.55953324e-01 -3.21408391e-01 -1.66349483e+00 6.81058407e-01
6.99173689e-01 -9.22880992e-02 9.67479646e-01 -1.00349283e+00
6.53077006e-01 -3.93748105e-01 1.12559229e-01 -9.94933903e-01
-5.61510324e-01 -2.02094734e-01 -8.26266408e-02 7.06651747e-01
1.30082548e+00 -6.24679387e-01 3.80623639e-01 6.34402215e-01
3.26053314e-02 -1.07245803e+00 -5.34867406e-01 -3.02111268e-01
1.50695801e-01 6.35945648e-02 4.41767097e-01 1.26339424e+00
7.77418852e-01 7.82053649e-01 1.44853666e-02 8.81939828e-02
1.48397803e-01 -1.73674837e-01 1.87374204e-01 -1.36174297e+00
-2.89027959e-01 -3.71659189e-01 -4.33464885e-01 -1.56599700e-01
2.83817295e-02 -1.09748757e+00 -1.78507000e-01 -1.60721636e+00
7.06003249e-01 -5.53912073e-02 -6.76128566e-01 1.14291418e+00
-2.33097509e-01 1.48963019e-01 -5.08949578e-01 2.21607834e-01
-4.09588069e-01 -2.08519727e-01 7.72123218e-01 -5.12035191e-01
-2.06692845e-01 -1.49759769e-01 -1.11187863e+00 5.94109178e-01
1.24983811e+00 -6.85637414e-01 2.01070029e-02 1.41937017e-01
2.27414772e-01 -2.07947284e-01 -6.19043708e-02 -6.06723368e-01
3.16586733e-01 -2.14645147e-01 8.93614054e-01 -6.35441005e-01
2.77649432e-01 -2.56216437e-01 7.33324707e-01 1.12222958e+00
-6.38400376e-01 -1.76900730e-01 6.54050589e-01 3.85693252e-01
1.30259488e-02 -3.63535993e-02 3.44880849e-01 -2.38870472e-01
-1.57378688e-01 8.81132558e-02 -6.44492388e-01 -8.44626307e-01
8.40291321e-01 1.81999937e-01 -6.41846776e-01 2.27958500e-01
-8.83975983e-01 1.68262914e-01 -5.81032336e-02 3.47358644e-01
4.14140821e-01 -7.37139642e-01 -8.45165551e-01 -3.57617229e-01
4.60900933e-01 -7.16040492e-01 -1.54625848e-01 1.04392111e+00
-5.74750125e-01 1.07263267e+00 8.81840810e-02 -3.15598041e-01
-1.96097195e+00 6.55207038e-01 4.78109509e-01 -2.36665130e-01
-1.95705369e-01 8.82128537e-01 1.03026792e-01 -4.57811892e-01
1.06268138e-01 -1.83534324e-02 -5.95487297e-01 -1.09330632e-01
6.32721126e-01 6.23546615e-02 4.95131046e-01 -2.50563473e-01
-8.21195304e-01 1.76481098e-01 -4.33390945e-01 4.40726906e-01
1.47035635e+00 6.90832317e-01 -6.69425011e-01 1.56693712e-01
1.29260695e+00 -8.85911509e-02 8.08623359e-02 9.32234228e-02
5.39386332e-01 3.10598761e-01 -1.61519468e-01 -1.56470394e+00
-4.12508637e-01 2.10820049e-01 6.58364534e-01 -1.55762464e-01
7.18820155e-01 1.69624403e-01 3.11337680e-01 7.97578037e-01
-1.17861414e-02 -7.10608959e-01 -7.22369254e-01 3.68450195e-01
8.04172695e-01 -1.13212359e+00 5.37263691e-01 -7.60394454e-01
-2.47466803e-01 1.09257460e+00 4.05484110e-01 2.57604390e-01
5.73988795e-01 4.19480771e-01 -1.26370545e-02 -6.85370803e-01
-1.23651600e+00 5.40273264e-02 5.18192828e-01 2.81226844e-01
1.45742989e+00 4.11614895e-01 -1.21768761e+00 6.90927088e-01
6.56819865e-02 1.99426547e-01 1.41211063e-01 1.03859913e+00
-2.40609050e-01 -1.59309900e+00 -1.30588561e-01 6.26543999e-01
-9.36036468e-01 -5.82237661e-01 -1.35900652e+00 4.86054242e-01
1.04154460e-01 1.22038352e+00 -4.78191674e-01 -5.65927505e-01
2.58719981e-01 3.39349657e-01 2.23652437e-01 -7.17136323e-01
-6.84917688e-01 3.35651100e-01 6.51216447e-01 -3.97555172e-01
-4.35243100e-01 -3.59521270e-01 -1.52035224e+00 -6.68868348e-02
-5.80113709e-01 6.08134806e-01 8.17048252e-01 6.55920625e-01
8.20244551e-01 6.96515143e-01 1.85432106e-01 -5.07504828e-02
-2.49871854e-02 -1.19764113e+00 -3.78878921e-01 -2.03909487e-01
-9.26545337e-02 -3.62440526e-01 3.91090401e-02 2.48912424e-01] | [8.41549301147461, 8.711759567260742] |
5986625f-8b49-4a56-a021-4bc40b9577d3 | simplifying-full-waveform-inversion-via | 2305.13314 | null | https://arxiv.org/abs/2305.13314v1 | https://arxiv.org/pdf/2305.13314v1.pdf | Simplifying Full Waveform Inversion via Domain-Independent Self-Supervised Learning | Geophysics has witnessed success in applying deep learning to one of its core problems: full waveform inversion (FWI) to predict subsurface velocity maps from seismic data. It is treated as an image-to-image translation problem, jointly training an encoder for seismic data and a decoder for the velocity map from seismic-velocity pairs. In this paper, we report a surprising phenomenon: when training an encoder and decoder separately in their own domains via self-supervised learning, a linear relationship is observed across domains in the latent spaces. Moreover, this phenomenon connects multiple FWI datasets in an elegant manner: these datasets can share the self-learned encoder and decoder with different linear mappings. Based on these findings, we develop SimFWI, a new paradigm that includes two steps: (a) learning a seismic encoder and a velocity decoder separately by masked image modeling over multiple datasets; (b) learning a linear mapping per dataset. Experimental results show that SimFWI can achieve comparable results to a jointly trained model from the supervision of paired seismic data and velocity maps. | ['Youzuo Lin', 'Zicheng Liu', 'Shihang Feng', 'Peng Jin', 'Yinpeng Chen', 'Yinan Feng'] | 2023-04-27 | null | null | null | null | ['image-to-image-translation', 'geophysics', 'image-to-image-translation'] | ['computer-vision', 'miscellaneous', 'miscellaneous'] | [ 6.07823491e-01 3.40936840e-01 1.23202972e-01 -4.67526853e-01
-1.23352456e+00 -2.79643625e-01 8.11575532e-01 -3.44871223e-01
-3.75841618e-01 4.05190378e-01 4.57828581e-01 -2.62772679e-01
1.18353158e-01 -8.32335293e-01 -1.29539490e+00 -9.98415411e-01
-2.21669227e-01 7.04122126e-01 3.09711307e-01 -8.08096230e-02
3.24842691e-01 -9.51286703e-02 -1.11663103e+00 3.66418153e-01
6.29115820e-01 8.25529575e-01 4.91622299e-01 8.75561357e-01
-1.37258545e-01 1.00480485e+00 -2.29220405e-01 8.94790664e-02
3.58598679e-01 -5.97259402e-01 -9.14357305e-01 -2.11108848e-01
2.62565553e-01 -4.48017120e-01 -6.60149097e-01 8.73215854e-01
2.19571263e-01 -3.80266428e-01 1.05239844e+00 -1.22584283e+00
-7.76988745e-01 6.78672791e-01 -8.73789310e-01 1.34055421e-01
1.70298234e-01 -1.35838062e-01 1.09949100e+00 -6.99722886e-01
5.26230156e-01 9.70437407e-01 7.48931468e-01 2.79027104e-01
-1.54244912e+00 -5.50564766e-01 -2.72902787e-01 1.21883564e-01
-1.37055802e+00 -5.07144034e-01 8.43984902e-01 -8.56374502e-01
9.20215249e-01 -1.26892015e-01 3.48450273e-01 1.05035281e+00
-1.57238748e-02 7.75216222e-01 8.46705139e-01 -4.20758963e-01
2.39194632e-02 3.94172501e-03 -3.23072821e-01 6.45744979e-01
-8.70242044e-02 1.95370391e-01 -1.00924134e+00 1.05537258e-01
9.28332806e-01 -4.57413733e-01 -3.94839585e-01 -1.48989856e-01
-1.52367246e+00 7.52111673e-01 1.57771572e-01 1.34829164e-01
-1.38264418e-01 4.15823817e-01 2.52655447e-01 6.30894780e-01
4.95379835e-01 3.49932075e-01 -2.52676666e-01 -1.11142337e-01
-1.13792884e+00 2.55141318e-01 6.83523118e-01 7.74716258e-01
1.22427309e+00 3.49124074e-01 3.75645399e-01 4.49922144e-01
5.50989389e-01 9.39797461e-01 8.45599353e-01 -8.75280976e-01
7.69136131e-01 -5.57329878e-02 2.24636272e-02 -9.42203641e-01
-3.43665600e-01 4.22624014e-02 -7.49189138e-01 1.49267614e-01
4.97966349e-01 -2.53406674e-01 -8.16800117e-01 2.02623177e+00
-1.94945574e-01 7.77723551e-01 4.99689847e-01 7.70265043e-01
5.99999309e-01 1.13058031e+00 -2.72335559e-01 1.92014292e-01
8.63525271e-01 -7.99041152e-01 -5.03308594e-01 -6.37817025e-01
8.23035777e-01 -5.34962535e-01 9.23862219e-01 4.73485477e-02
-1.17987013e+00 -6.31161511e-01 -1.37173975e+00 -1.25615492e-01
-9.89337265e-02 -8.20203125e-03 1.31291896e-01 1.21868692e-01
-1.18260837e+00 6.96778059e-01 -1.04301012e+00 8.60378444e-02
4.16065045e-02 1.80915952e-01 -4.85984743e-01 2.97639996e-01
-1.34633696e+00 6.37896836e-01 3.08984041e-01 1.41456991e-01
-1.13029480e+00 -8.02495718e-01 -1.14936638e+00 -2.92707724e-03
-1.29177645e-01 -4.94938195e-01 1.20286262e+00 -8.84424031e-01
-1.41320801e+00 8.74859095e-01 -3.05798590e-01 -5.30075014e-01
6.38202906e-01 -1.68456793e-01 -2.62083441e-01 1.34131849e-01
4.93974864e-01 5.12435555e-01 1.10595095e+00 -1.34900534e+00
-5.50873578e-01 -3.08684736e-01 -5.20439744e-01 -1.02519756e-02
-2.26327911e-01 -4.66341138e-01 -2.42484793e-01 -4.42690074e-01
5.37306488e-01 -5.03384292e-01 4.61222939e-02 -9.81045142e-02
-2.57414252e-01 2.14893773e-01 7.56626546e-01 -9.26044345e-01
9.65395629e-01 -2.15966797e+00 5.85357785e-01 3.02504212e-01
2.31336400e-01 -6.14081509e-02 -3.00052345e-01 3.77341717e-01
-2.69933015e-01 -1.13512486e-01 -7.89719224e-01 -4.83125716e-01
-1.40195966e-01 4.27749664e-01 -5.79600215e-01 6.95789576e-01
4.58668411e-01 1.05270028e+00 -1.00922871e+00 -3.61906409e-01
-7.51733035e-02 2.20799357e-01 -4.85108078e-01 4.91363585e-01
-6.45666271e-02 8.61278176e-01 -7.87732080e-02 1.60818994e-01
9.46084380e-01 -3.99222463e-01 2.09066719e-01 -2.32844323e-01
-3.10305536e-01 2.08772838e-01 -1.09687889e+00 1.89455354e+00
-5.42967737e-01 9.77552176e-01 -1.16031066e-01 -1.64559305e+00
1.09260893e+00 5.84559441e-01 5.96112669e-01 -9.98500109e-01
-2.62733847e-01 6.55742645e-01 -2.88424551e-01 -9.36657310e-01
2.88555145e-01 -4.62261856e-01 -2.17437029e-01 6.19801939e-01
4.18124855e-01 -1.47017241e-01 -1.73892528e-01 1.06193103e-01
1.00349021e+00 2.58437216e-01 -2.63179123e-01 -1.13469809e-01
4.42091823e-01 -7.64722303e-02 3.71951699e-01 7.35597789e-01
3.53024215e-01 1.03616691e+00 6.18348897e-01 -2.74116307e-01
-1.66451895e+00 -1.38764882e+00 -1.68055370e-01 6.52995706e-01
1.43482476e-01 -1.55633595e-02 -4.72889125e-01 -3.14817913e-02
-2.14319322e-02 2.74646223e-01 -7.08469152e-01 -2.71319717e-01
-7.62785017e-01 -6.58393204e-01 8.75922143e-01 8.80611062e-01
5.84314764e-01 -7.32552588e-01 -6.13899946e-01 2.65669674e-01
-3.98877114e-01 -1.23166645e+00 -3.35440725e-01 3.49838912e-01
-7.38426447e-01 -5.08128285e-01 -1.03193760e+00 -1.05078661e+00
4.90937203e-01 2.55869120e-01 8.98336172e-01 -1.95234329e-01
7.69314021e-02 8.40310901e-02 -1.67939454e-01 -2.25146055e-01
-7.49700129e-01 1.72719195e-01 -1.99347228e-01 3.65173370e-01
-7.11636245e-02 -8.30769837e-01 -4.26582694e-01 1.56495199e-01
-1.00573909e+00 4.42704111e-01 3.63312155e-01 9.38876629e-01
1.77326486e-01 -4.03501451e-01 5.80455065e-01 -7.63752639e-01
2.53608078e-01 -7.42053270e-01 -7.41585732e-01 2.38693669e-01
-2.84501553e-01 6.04038596e-01 3.12643707e-01 -1.35348216e-01
-1.38264132e+00 1.99151617e-02 -4.53553945e-01 -2.68589437e-01
2.09369153e-01 8.39857876e-01 1.23142600e-01 5.75674102e-02
6.58517420e-01 6.22297406e-01 3.17762315e-01 -5.53882658e-01
2.77419537e-01 8.62938225e-01 1.11694014e+00 -3.81563932e-01
9.92263675e-01 8.05307209e-01 -3.48699391e-02 -9.34766948e-01
-7.30886519e-01 -6.01824641e-01 -1.00031924e+00 -1.37138069e-01
1.22639656e+00 -1.18213558e+00 -2.57649481e-01 7.26609111e-01
-1.27109158e+00 -5.39696872e-01 -3.18091989e-01 5.73839128e-01
-7.19072402e-01 4.86586839e-01 -5.02613485e-01 -4.38073725e-01
6.06448129e-02 -1.28830862e+00 1.44470537e+00 3.72479321e-03
-4.22226712e-02 -1.23281837e+00 4.50153708e-01 1.61041878e-02
2.43975118e-01 2.19924912e-01 8.26193154e-01 -3.31151485e-01
-5.73235512e-01 -7.57779926e-02 -4.40089971e-01 4.36808318e-01
7.27124959e-02 -4.13732648e-01 -1.12366784e+00 -1.81432456e-01
2.05177054e-01 -3.44529927e-01 1.21211672e+00 4.23002422e-01
7.92567551e-01 -9.37651321e-02 -1.86874792e-01 1.19594967e+00
1.52613163e+00 1.62696511e-01 7.09397197e-01 5.46936631e-01
8.14599216e-01 5.33089757e-01 -7.19555169e-02 4.16860074e-01
4.68169600e-01 3.94703925e-01 2.08956122e-01 -3.53678733e-01
-5.78808077e-02 -5.41929543e-01 3.69378507e-01 9.93703008e-01
-1.61883906e-01 -2.25432105e-02 -1.20684934e+00 6.30939603e-01
-1.90766048e+00 -9.92600203e-01 -1.80489570e-01 2.15825224e+00
6.96337402e-01 -8.49945354e-05 -2.57349253e-01 1.85290501e-01
2.60011673e-01 4.96143997e-01 -3.81966859e-01 -5.48307672e-02
-4.12279367e-01 1.60896733e-01 8.20925057e-01 6.86105788e-01
-1.11125302e+00 7.49048769e-01 6.68497801e+00 4.33271766e-01
-1.47630119e+00 3.50425869e-01 4.87233371e-01 3.98493856e-01
-7.91329384e-01 2.53624655e-02 -4.33768332e-01 3.20083857e-01
1.09622240e+00 -1.52203441e-01 2.35960379e-01 2.72768915e-01
5.37134707e-02 4.56215255e-02 -1.00408077e+00 8.05857837e-01
1.55705847e-02 -1.82857931e+00 -2.13349815e-02 3.55916694e-02
7.54363537e-01 2.77302474e-01 3.93813476e-02 -1.09669469e-01
9.12395716e-02 -9.79596138e-01 1.12329793e+00 6.61540926e-01
1.08112204e+00 -2.72825360e-01 5.38771510e-01 3.35653454e-01
-1.22572732e+00 1.22061387e-01 -2.53831387e-01 -1.20593287e-01
6.25777125e-01 1.06752574e-01 -6.66362286e-01 7.06377983e-01
6.05603993e-01 9.44657385e-01 -3.35446857e-02 6.97686076e-01
-1.93462700e-01 7.29340911e-01 -1.43227845e-01 6.13443494e-01
4.35867339e-01 -4.02891994e-01 3.58753175e-01 1.21113467e+00
6.50399387e-01 -3.22516829e-01 -1.31417722e-01 1.04216075e+00
2.15345994e-01 -3.06790382e-01 -8.56708884e-01 9.19386670e-02
2.20919341e-01 6.46049201e-01 -4.76392359e-01 -4.36344296e-01
-8.76629651e-01 1.00825274e+00 2.43478626e-01 5.14085293e-01
-9.82084572e-01 -9.99509692e-02 5.29749095e-01 2.69295126e-01
3.81171376e-01 -5.08864105e-01 -4.07771051e-01 -1.18415439e+00
-1.36443779e-01 -1.88527524e-01 9.70547944e-02 -9.07246590e-01
-9.60323393e-01 4.22159016e-01 1.12448983e-01 -1.51285946e+00
-4.39172447e-01 -6.20828867e-01 -6.03766322e-01 1.06317770e+00
-1.94228065e+00 -1.07818425e+00 -3.96000773e-01 4.83182281e-01
3.43701303e-01 -1.73075587e-01 5.93738496e-01 4.11366671e-01
-2.30856553e-01 4.53502327e-01 6.13921106e-01 4.66926634e-01
6.44945562e-01 -1.26454425e+00 7.26330459e-01 8.11037719e-01
4.07514960e-01 -7.12303957e-03 7.13330686e-01 -4.27441657e-01
-1.41095138e+00 -9.42837417e-01 8.41861963e-01 -9.70043838e-02
8.99510801e-01 -3.71323347e-01 -1.46179032e+00 1.05536675e+00
2.13818356e-01 -2.20539588e-02 4.16771948e-01 -4.03532267e-01
-3.05449963e-01 1.02514783e-02 -4.84030962e-01 2.64755756e-01
8.50349545e-01 -8.86305034e-01 -6.49750590e-01 3.24494451e-01
6.31179094e-01 -6.04881942e-01 -6.98385000e-01 2.54633158e-01
4.00368094e-01 -8.52434397e-01 1.21293366e+00 -2.80828178e-01
1.10890555e+00 -2.42147774e-01 -2.00578138e-01 -1.50262451e+00
-8.53224844e-02 -4.94119674e-01 1.32430568e-01 8.01042974e-01
7.43615925e-01 -6.56567216e-01 6.75988674e-01 1.98083475e-01
-3.89177442e-01 -3.20858747e-01 -1.03063869e+00 -8.02021027e-01
3.84149939e-01 -5.65633714e-01 5.56296945e-01 7.57250488e-01
3.98161262e-02 2.76200056e-01 -7.50350118e-01 6.44688308e-01
6.52173579e-01 3.49365994e-02 6.50292039e-01 -9.50682580e-01
-4.29929733e-01 -2.53714740e-01 -3.87484342e-01 -1.49472940e+00
1.17933780e-01 -1.07190073e+00 4.56632644e-01 -1.30277061e+00
4.23652232e-02 -1.40089169e-01 3.56914736e-02 3.02208453e-01
1.30431980e-01 2.55964160e-01 -1.52225122e-01 6.09962463e-01
-2.00506039e-02 6.18710577e-01 1.19590139e+00 -2.70745903e-01
5.76567389e-02 -2.57058412e-01 -3.32786411e-01 7.70721436e-01
4.75039631e-01 -4.44749981e-01 -4.32552695e-01 -1.05981255e+00
3.67434293e-01 7.38303006e-01 5.00542343e-01 -1.11931694e+00
5.18271983e-01 6.46173283e-02 1.33252040e-01 -4.20033187e-01
2.96239316e-01 -5.08583844e-01 1.14827193e-01 4.20329630e-01
-3.42237055e-01 -6.86306646e-03 1.07746974e-01 6.17887080e-01
-7.33176589e-01 -3.81991327e-01 5.47075033e-01 -2.37005234e-01
-1.16682422e+00 1.71485066e-01 -4.78425652e-01 1.64643481e-01
6.32831514e-01 -4.14418191e-01 -3.73324752e-02 -4.66755450e-01
-8.47176433e-01 6.20734170e-02 3.79703566e-02 1.54771820e-01
7.65150964e-01 -1.12251985e+00 -9.31442559e-01 8.23211014e-01
2.35644683e-01 3.89975786e-01 3.65424693e-01 8.83014619e-01
-7.50896573e-01 1.46657318e-01 -1.78242758e-01 -1.12759733e+00
-8.12864065e-01 4.20760028e-02 5.48105001e-01 -1.31263405e-01
-9.40520942e-01 1.19008994e+00 5.35077214e-01 -4.11312610e-01
-3.29333986e-03 -2.73695290e-01 1.02249309e-01 1.92709733e-02
5.06515384e-01 2.46802960e-02 -1.17288314e-01 -8.17271411e-01
4.21836227e-03 8.16069603e-01 2.36592144e-01 -6.67205572e-01
1.68116951e+00 -1.87818155e-01 9.34174880e-02 6.83679521e-01
1.76878023e+00 -5.08551240e-01 -1.74555886e+00 -7.14004636e-01
1.36874735e-01 -4.54050630e-01 7.19967484e-02 -9.46179852e-02
-1.17049873e+00 1.20304275e+00 2.76053905e-01 8.11417624e-02
9.68810201e-01 2.94399142e-01 7.78726757e-01 5.23461401e-02
1.06069222e-01 -8.61729145e-01 4.34418887e-01 6.44117951e-01
8.36822391e-01 -1.36650920e+00 -4.27899301e-01 -6.31807256e-04
-7.36499786e-01 1.15525770e+00 1.55126289e-01 -1.72966406e-01
8.39242041e-01 8.14241529e-01 2.10348979e-01 -2.58557290e-01
-6.00483537e-01 3.00008580e-02 1.64699197e-01 5.07090330e-01
3.97744566e-01 -1.98599070e-01 2.68597901e-01 2.49831215e-01
-3.27071637e-01 7.81268626e-02 5.45784414e-01 9.62442338e-01
-4.96421635e-01 -9.52123821e-01 -4.34804350e-01 2.22427800e-01
8.53351951e-02 1.26711940e-02 3.20765436e-01 5.66555262e-01
-1.01694003e-01 3.01168352e-01 4.89092946e-01 -4.71287400e-01
1.45794317e-01 5.59928231e-02 4.46396112e-01 -3.57077181e-01
1.79395780e-01 2.49127671e-02 -3.15099090e-01 -3.08304846e-01
-4.13274914e-01 -8.82289290e-01 -1.47719717e+00 -5.66146430e-03
8.87286514e-02 1.05313383e-01 5.99183202e-01 1.27000141e+00
-1.00851163e-01 7.04674363e-01 6.41809165e-01 -1.14945412e+00
-6.66423917e-01 -7.78843760e-01 -8.60519469e-01 5.62923133e-01
7.28929698e-01 -6.06858552e-01 -5.51189184e-01 5.44528425e-01] | [6.871970176696777, 2.520144462585449] |
6a4ac16b-2880-4b85-b7fc-08c56ab7ae2d | rank-based-causal-discovery-for-post | 2302.12341 | null | https://arxiv.org/abs/2302.12341v1 | https://arxiv.org/pdf/2302.12341v1.pdf | Rank-Based Causal Discovery for Post-Nonlinear Models | Learning causal relationships from empirical observations is a central task in scientific research. A common method is to employ structural causal models that postulate noisy functional relations among a set of interacting variables. To ensure unique identifiability of causal directions, researchers consider restricted subclasses of structural causal models. Post-nonlinear (PNL) causal models constitute one of the most flexible options for such restricted subclasses, containing in particular the popular additive noise models as a further subclass. However, learning PNL models is not well studied beyond the bivariate case. The existing methods learn non-linear functional relations by minimizing residual dependencies and subsequently test independence from residuals to determine causal orientations. However, these methods can be prone to overfitting and, thus, difficult to tune appropriately in practice. As an alternative, we propose a new approach for PNL causal discovery that uses rank-based methods to estimate the functional parameters. This new approach exploits natural invariances of PNL models and disentangles the estimation of the non-linear functions from the independence tests used to find causal orientations. We prove consistency of our method and validate our results in numerical experiments. | ['Mathias Drton', 'David Strieder', 'Grigor Keropyan'] | 2023-02-23 | null | null | null | null | ['causal-discovery'] | ['knowledge-base'] | [ 3.01205993e-01 -2.29688942e-01 -4.09656703e-01 -3.44068736e-01
-4.04036492e-01 -7.15103149e-01 5.58084965e-01 1.07045121e-01
-9.48779956e-02 1.14035714e+00 2.28664815e-01 -5.19659042e-01
-1.14199078e+00 -7.87571073e-01 -8.64775479e-01 -9.18209255e-01
-3.31793219e-01 4.39902484e-01 5.16884290e-02 2.86113098e-02
2.43731052e-01 4.95432496e-01 -1.31885529e+00 -2.19645262e-01
1.35114372e+00 4.06997263e-01 -3.20335627e-02 4.27353382e-01
1.53564528e-01 4.93232489e-01 -2.13972792e-01 -1.44098476e-01
1.58992440e-01 -5.51487505e-01 -6.16587818e-01 -1.71025500e-01
5.12512922e-02 1.63238067e-02 -8.64526778e-02 1.11054885e+00
2.55025655e-01 1.47193670e-01 8.81191909e-01 -1.33027077e+00
-5.53547382e-01 7.23139763e-01 -5.95454872e-01 1.15553319e-01
3.96183729e-01 -1.77122593e-01 1.25716352e+00 -7.49027967e-01
2.04142824e-01 1.62580812e+00 4.36002523e-01 1.16828881e-01
-1.81022215e+00 -9.16925073e-01 2.31438801e-01 3.46410006e-01
-1.37282300e+00 -3.05967808e-01 6.92871392e-01 -6.01786375e-01
1.67344734e-01 5.39626658e-01 3.39948386e-01 1.32513106e+00
2.72397727e-01 3.92036885e-01 1.57300866e+00 -3.54051977e-01
3.10999572e-01 -1.86805159e-01 3.46514195e-01 5.63698888e-01
6.99793994e-01 3.38171929e-01 -5.25158763e-01 -6.72030866e-01
1.00273538e+00 -1.67255387e-01 -4.04735655e-01 -4.49466556e-01
-1.11043072e+00 1.04614615e+00 1.63127750e-01 2.46329203e-01
-3.40768278e-01 1.33739963e-01 -5.18626831e-02 2.67204255e-01
3.55196357e-01 5.40900826e-01 -4.10592645e-01 1.71201482e-01
-4.78857279e-01 4.46799576e-01 7.13703632e-01 5.75611591e-01
6.01155460e-01 -1.77540824e-01 -9.05685574e-02 7.11036801e-01
5.32916605e-01 5.34623086e-01 -3.10598928e-02 -6.94326937e-01
1.54371887e-01 5.11105359e-01 9.35396850e-02 -1.10689259e+00
-5.02640963e-01 -4.92294282e-01 -1.09703386e+00 -6.05558679e-02
8.40418637e-01 -1.58871531e-01 -6.52113020e-01 2.03485179e+00
4.90275204e-01 4.95633751e-01 -3.77274245e-01 8.78834784e-01
3.30283582e-01 3.89625967e-01 2.19816297e-01 -5.90126574e-01
9.67246473e-01 -1.65916197e-02 -6.68309450e-01 1.81791142e-01
3.52589130e-01 -6.92995608e-01 9.10917282e-01 4.77658600e-01
-7.76669860e-01 -2.08365113e-01 -6.48443997e-01 3.51415366e-01
-5.31541277e-03 -1.51692033e-01 8.99574220e-01 5.03299236e-01
-7.70646632e-01 5.68331361e-01 -8.27990234e-01 -1.26279220e-01
1.12199172e-01 5.32421231e-01 -2.90028811e-01 -1.19610734e-01
-1.45162868e+00 5.55053115e-01 1.79852396e-01 3.21943372e-01
-8.71533632e-01 -9.14354682e-01 -5.20884752e-01 1.31914154e-01
8.17329943e-01 -6.32683516e-01 7.55290270e-01 -5.64168930e-01
-1.15546775e+00 1.42883748e-01 -1.25569105e-01 -4.37247716e-02
3.62507313e-01 -2.13512659e-01 -2.23688260e-01 -2.54961312e-01
1.02557480e-01 -2.89980322e-02 8.39857996e-01 -1.29543543e+00
-3.95393938e-01 -3.79504204e-01 1.74818456e-01 -1.57243937e-01
-1.39626443e-01 -1.00072101e-01 -1.67915151e-02 -6.05950177e-01
4.00730759e-01 -9.41126585e-01 -4.60518777e-01 -3.95181030e-01
-8.47981095e-01 -3.51963580e-01 4.38673556e-01 -3.17326933e-01
1.30158937e+00 -1.90378284e+00 3.88178706e-01 6.14640474e-01
2.37074792e-01 -3.54936421e-01 -1.68140933e-01 3.72129858e-01
-4.84200716e-01 3.58608335e-01 -3.43862146e-01 2.50718832e-01
-4.92794365e-02 2.66820759e-01 -2.83602715e-01 7.18158245e-01
2.39558607e-01 5.00715554e-01 -8.21354032e-01 -3.21592420e-01
2.73150802e-02 3.54801714e-01 -6.99659228e-01 2.56878763e-01
-1.12805255e-01 8.57177615e-01 -6.07738733e-01 3.46184164e-01
5.67120492e-01 -2.37744480e-01 4.91491020e-01 -1.83587953e-01
-3.19604963e-01 2.68893272e-01 -1.55898738e+00 9.47297454e-01
-1.71527252e-01 1.78460255e-01 -9.15602818e-02 -1.60211349e+00
7.36622393e-01 3.88854146e-01 6.69281125e-01 -2.10758537e-01
1.60491224e-02 2.89521813e-01 4.64474350e-01 -7.65948594e-01
-1.33935735e-01 -3.62675786e-01 -1.66200966e-01 6.54376596e-02
-2.10773140e-01 4.54712331e-01 3.71967733e-01 -9.36458334e-02
1.29765964e+00 1.97536983e-02 7.34673381e-01 -5.96958935e-01
3.72225761e-01 -2.99303651e-01 9.62605774e-01 9.25627470e-01
2.81749576e-01 2.23077387e-01 1.05844641e+00 -9.60127339e-02
-7.32602000e-01 -1.26865482e+00 -4.22902226e-01 6.93367243e-01
-7.04066157e-02 -1.95900574e-01 -2.49258444e-01 -5.86742640e-01
1.94925815e-01 6.49630666e-01 -6.98763847e-01 -1.19544834e-01
-5.10097802e-01 -1.22162449e+00 3.98506194e-01 2.92667031e-01
-1.10384040e-02 -5.41311085e-01 -1.61485896e-01 6.40710443e-02
-1.08532593e-01 -6.56552911e-01 -8.03792179e-02 4.15963978e-01
-8.20388973e-01 -1.52059484e+00 -4.28479135e-01 -1.78342670e-01
6.98105454e-01 2.52950024e-02 7.51705945e-01 -1.59474239e-01
-5.89761026e-02 1.97426453e-01 -1.71710461e-01 -8.09649304e-02
-3.20183992e-01 -1.79640472e-01 3.45560700e-01 2.24859536e-01
1.11307539e-01 -8.93282473e-01 -3.08359742e-01 5.05270362e-01
-9.38368857e-01 -1.84996203e-01 7.91665852e-01 1.04678428e+00
5.14328778e-01 2.68071085e-01 7.71885693e-01 -9.89615798e-01
6.45648897e-01 -8.37883174e-01 -1.05431020e+00 2.95004278e-01
-6.64863288e-01 4.13225383e-01 6.47750556e-01 -7.26987422e-01
-1.18529248e+00 8.47492740e-03 2.33686924e-01 -2.71447062e-01
-2.50559837e-01 9.46166575e-01 -4.96972501e-01 1.70726687e-01
4.70625281e-01 -2.94608623e-01 -4.31720585e-01 -4.93117750e-01
2.44706437e-01 9.96512920e-02 3.02795082e-01 -8.63183022e-01
9.73253310e-01 2.39798173e-01 6.01841450e-01 -8.17991972e-01
-8.13273907e-01 -2.82996058e-01 -4.98616189e-01 -8.79406109e-02
7.02032864e-01 -4.49009150e-01 -1.14862216e+00 -1.97151288e-01
-1.06906402e+00 -1.37425900e-01 4.14499678e-02 8.67463529e-01
-3.56175184e-01 2.16172829e-01 -3.12065512e-01 -1.11770201e+00
4.97646958e-01 -1.08559465e+00 7.51213372e-01 4.98272143e-02
-4.19291914e-01 -1.18808985e+00 3.48397523e-01 -7.39697739e-02
-1.33541580e-02 2.71205068e-01 1.40720224e+00 -5.49815297e-01
-5.51322341e-01 -1.79783795e-02 -2.31481582e-01 -1.85617492e-01
2.23473549e-01 1.62992090e-01 -6.34745538e-01 -5.62904850e-02
-2.98874043e-02 7.13083223e-02 7.77605534e-01 9.61744547e-01
1.12051284e+00 -4.50820088e-01 -5.10338247e-01 3.92286390e-01
1.20440614e+00 2.85450786e-01 3.92957360e-01 -5.95229603e-02
8.20517123e-01 9.21567917e-01 4.88894284e-01 4.98176605e-01
8.99490714e-02 7.15642750e-01 3.94363791e-01 -3.54825631e-02
4.38719004e-01 -2.03628838e-01 2.09465876e-01 6.31641924e-01
-3.34444612e-01 -3.58320236e-01 -7.09249258e-01 2.28216842e-01
-1.99907458e+00 -9.91378188e-01 -8.09845984e-01 2.49416399e+00
9.85352218e-01 -3.48571874e-02 1.38212383e-01 1.64701030e-01
7.57370532e-01 -3.13179672e-01 -4.60263729e-01 1.07559338e-01
-1.65604144e-01 1.26192749e-01 5.84102809e-01 5.87520123e-01
-1.08210742e+00 4.48208898e-01 6.27242565e+00 6.05675876e-01
-5.88840246e-01 2.25513466e-02 4.80646849e-01 8.51030201e-02
-5.23019552e-01 4.85018343e-01 -7.46926665e-01 3.94560695e-01
7.25290000e-01 -1.60161674e-01 2.46580794e-01 3.75454307e-01
8.34165096e-01 -2.12169454e-01 -1.25641656e+00 6.91724062e-01
-5.28588533e-01 -9.18195128e-01 -3.15616876e-01 3.93272310e-01
8.86260927e-01 -6.57564521e-01 -1.72759816e-02 1.42493995e-03
6.08131111e-01 -1.18847573e+00 3.51006329e-01 8.37714016e-01
5.32512367e-01 -7.36480296e-01 4.92604673e-01 2.64318734e-01
-8.96325290e-01 -2.64113843e-01 -4.52274650e-01 -2.04072312e-01
1.64108276e-01 1.12195766e+00 -4.83787090e-01 5.20407557e-01
5.59723675e-01 5.84817529e-01 -2.61349618e-01 1.21123087e+00
-4.56469625e-01 1.09570909e+00 -4.00041372e-01 2.00338196e-02
-1.66868895e-01 -4.14649665e-01 8.20765376e-01 8.90918434e-01
4.77899224e-01 1.62841842e-01 1.27651691e-01 1.19588327e+00
5.35618961e-02 1.25558376e-01 -6.46820724e-01 -7.05419900e-03
4.64610487e-01 1.00036979e+00 -7.09220171e-01 7.84446001e-02
-3.39066148e-01 3.65961909e-01 2.00100705e-01 6.06380463e-01
-7.92926013e-01 2.20126003e-01 6.57769680e-01 1.53033584e-01
-1.42297924e-01 -2.70527840e-01 -3.59267592e-01 -1.16194952e+00
-1.55633345e-01 -1.06356907e+00 5.60655415e-01 -2.32003167e-01
-1.64495528e+00 -1.43137679e-01 6.50061905e-01 -1.08704484e+00
-2.31229112e-01 -4.65563953e-01 -3.84116203e-01 8.64251435e-01
-1.10419667e+00 -7.68413126e-01 -5.00004999e-02 4.96406645e-01
3.00908446e-01 3.34870040e-01 6.74683750e-01 3.25273305e-01
-8.02591145e-01 2.27893427e-01 1.79853261e-01 -3.46412838e-01
8.37474763e-01 -1.43623960e+00 -3.10918182e-01 8.45055282e-01
-5.86206373e-03 9.92076933e-01 1.00480092e+00 -9.61635172e-01
-1.44728506e+00 -7.05237687e-01 7.57215381e-01 -2.20638290e-01
1.06758487e+00 -4.30918247e-01 -1.09070098e+00 5.82867980e-01
-1.32009476e-01 -1.11004233e-01 6.62963212e-01 6.55218780e-01
-3.45248520e-01 -1.55154765e-01 -6.13960028e-01 7.89026022e-01
1.02084851e+00 -6.92244545e-02 -2.69741386e-01 1.52176380e-01
5.26702404e-01 1.27382442e-01 -9.41191137e-01 6.64190948e-01
5.98206460e-01 -8.67480755e-01 1.01673973e+00 -6.94817185e-01
4.78416711e-01 -4.68384236e-01 -3.25263254e-02 -1.35940564e+00
-4.15262520e-01 -5.94713628e-01 1.25521660e-01 1.36639810e+00
4.38593358e-01 -7.67304420e-01 4.47188884e-01 5.22569358e-01
2.49130324e-01 -4.37011242e-01 -8.72942865e-01 -8.74341846e-01
-2.72140075e-02 -5.46907485e-01 3.07800710e-01 1.27592385e+00
-5.47212474e-02 5.81720710e-01 -6.14334881e-01 4.78356242e-01
9.44026232e-01 8.99407789e-02 6.79017782e-01 -1.72899520e+00
-6.09360337e-01 -5.03076673e-01 -2.01849163e-01 -7.90156543e-01
3.64002198e-01 -7.06759393e-01 1.54590040e-01 -1.10366011e+00
3.81838650e-01 -7.16009319e-01 -1.84453070e-01 4.44465399e-01
-3.02433968e-01 -1.30408466e-01 -3.13329905e-01 1.22526594e-01
1.89834274e-03 4.88113254e-01 1.07717025e+00 3.51509526e-02
-4.70506400e-01 3.70340705e-01 -7.11921096e-01 8.65087271e-01
6.02158487e-01 -7.79505610e-01 -7.82256424e-01 1.41531974e-01
5.67520022e-01 3.54929686e-01 7.57631361e-01 -4.30344373e-01
4.26717103e-03 -7.95086920e-01 2.63925105e-01 -4.53806281e-01
-1.57054320e-01 -7.81339288e-01 5.10696232e-01 2.53625393e-01
-5.94174504e-01 -9.18567628e-02 -4.61083017e-02 8.08326125e-01
2.16304185e-03 -2.31165022e-01 4.81725454e-01 2.79402167e-01
-1.31639764e-01 -1.27265695e-02 -4.50874627e-01 -1.49085581e-01
6.47590458e-01 2.54608482e-01 -2.22042620e-01 -3.99120301e-01
-7.59275913e-01 1.09291367e-01 -1.19981721e-01 2.09210336e-01
4.97358203e-01 -1.26035857e+00 -7.93308258e-01 -1.01531066e-01
-1.68266326e-01 -2.04421833e-01 1.98303699e-01 1.43589914e+00
2.02117071e-01 5.24930656e-01 2.31559455e-01 -5.51171660e-01
-1.24217641e+00 6.14806175e-01 1.38222247e-01 -3.57181579e-01
-2.02122122e-01 4.91535306e-01 8.10611010e-01 -1.98365122e-01
-4.75955494e-02 -2.20933884e-01 -3.43615592e-01 5.00831529e-02
2.16869310e-01 6.03940547e-01 -2.74727255e-01 -3.05088371e-01
-3.54522169e-01 4.11257118e-01 2.88601100e-01 -1.60896644e-01
1.29803622e+00 -1.40585020e-01 -4.01919127e-01 7.22084582e-01
1.03802478e+00 2.59914845e-01 -1.15695739e+00 -1.06311537e-01
4.34208721e-01 -3.73086274e-01 -1.32505730e-01 -5.86379647e-01
-7.09020436e-01 5.10848165e-01 3.33800256e-01 5.95727265e-01
1.20567107e+00 8.02301541e-02 -1.61325783e-01 2.15481833e-01
3.06828804e-02 -6.70650959e-01 -1.04102269e-01 1.97994620e-01
9.61964548e-01 -1.02224076e+00 6.32701591e-02 -8.40601146e-01
1.54679537e-01 1.10747826e+00 3.71596277e-01 -1.44680738e-01
7.62494147e-01 4.00420457e-01 -3.80412608e-01 -1.97813854e-01
-6.72702730e-01 -2.73385853e-01 7.05213606e-01 3.27873170e-01
7.50860214e-01 2.25462735e-01 -9.26072836e-01 5.79198062e-01
3.88616621e-02 -3.06656480e-01 4.14686412e-01 4.79564846e-01
3.21588479e-02 -1.23087108e+00 -8.19140196e-01 5.17884254e-01
-4.65018570e-01 -8.54860768e-02 -2.84206808e-01 9.41036105e-01
-7.06527382e-02 1.28321266e+00 -3.15554261e-01 -1.78386003e-01
4.15409267e-01 -6.83276877e-02 4.05766070e-01 -4.30030167e-01
-6.39079735e-02 6.34891748e-01 -2.99176909e-02 -5.21432221e-01
-6.23143554e-01 -1.00347865e+00 -9.22900975e-01 -2.32355818e-01
-5.94213367e-01 1.29119858e-01 2.70271152e-01 1.28396320e+00
-7.73527846e-02 7.10830510e-01 6.87920988e-01 -4.50647324e-01
-6.05723858e-01 -8.03283572e-01 -6.24731183e-01 2.35740811e-01
2.09259123e-01 -1.18459523e+00 -5.23376882e-01 4.59057483e-04] | [7.790450096130371, 5.27681827545166] |
e58e4e8c-7ab3-4a8a-a3b1-46684800b7fd | probing-visual-audio-representation-for-video | 2206.10157 | null | https://arxiv.org/abs/2206.10157v1 | https://arxiv.org/pdf/2206.10157v1.pdf | Probing Visual-Audio Representation for Video Highlight Detection via Hard-Pairs Guided Contrastive Learning | Video highlight detection is a crucial yet challenging problem that aims to identify the interesting moments in untrimmed videos. The key to this task lies in effective video representations that jointly pursue two goals, \textit{i.e.}, cross-modal representation learning and fine-grained feature discrimination. In this paper, these two challenges are tackled by not only enriching intra-modality and cross-modality relations for representation modeling but also shaping the features in a discriminative manner. Our proposed method mainly leverages the intra-modality encoding and cross-modality co-occurrence encoding for fully representation modeling. Specifically, intra-modality encoding augments the modality-wise features and dampens irrelevant modality via within-modality relation learning in both audio and visual signals. Meanwhile, cross-modality co-occurrence encoding focuses on the co-occurrence inter-modality relations and selectively captures effective information among multi-modality. The multi-modal representation is further enhanced by the global information abstracted from the local context. In addition, we enlarge the discriminative power of feature embedding with a hard-pairs guided contrastive learning (HPCL) scheme. A hard-pairs sampling strategy is further employed to mine the hard samples for improving feature discrimination in HPCL. Extensive experiments conducted on two benchmarks demonstrate the effectiveness and superiority of our proposed methods compared to other state-of-the-art methods. | ['Shuai Yi', 'Jun Hou', 'Shinan Liu', 'Lingbo Liu', 'Kunlin Yang', 'Feng Zhang', 'Shuaicheng Li'] | 2022-06-21 | null | null | null | null | ['highlight-detection'] | ['computer-vision'] | [ 3.32710475e-01 -5.12718499e-01 -3.71157199e-01 -1.37962177e-01
-1.02801776e+00 -4.70579088e-01 6.65895283e-01 1.48313999e-01
-1.72254026e-01 3.36357951e-01 6.23301327e-01 3.58871400e-01
-5.55851877e-01 -4.55959976e-01 -5.44885099e-01 -9.70032394e-01
-2.10405186e-01 -4.57469076e-01 1.24180280e-01 -1.78946555e-01
2.35086918e-01 1.28813028e-01 -1.83245742e+00 5.31901300e-01
7.59140074e-01 1.31933939e+00 1.36162415e-01 3.16545397e-01
3.61461379e-02 7.63609231e-01 -9.97297317e-02 1.35625273e-01
1.25766009e-01 -3.99045676e-01 -2.98275888e-01 1.96104929e-01
3.18762064e-01 -1.27636921e-02 -5.13731897e-01 1.11897719e+00
5.03016293e-01 4.58846122e-01 5.25605142e-01 -1.31971753e+00
-3.75793189e-01 3.06094110e-01 -9.83481765e-01 4.41758066e-01
4.98309582e-01 3.37175608e-01 1.22512639e+00 -1.22183990e+00
2.91647792e-01 1.29257631e+00 4.12301451e-01 1.68201879e-01
-1.06497121e+00 -7.24232316e-01 4.37488049e-01 5.43689907e-01
-1.54261744e+00 -4.73544329e-01 1.36069322e+00 -6.11287653e-01
4.12871718e-01 4.34554398e-01 6.00199282e-01 9.63040769e-01
-1.61823034e-01 8.40023696e-01 1.09928656e+00 -3.08920383e-01
-1.45699576e-01 -6.63314462e-02 1.17572909e-02 6.38007402e-01
-2.73026645e-01 2.13669926e-01 -1.04712486e+00 -1.56267285e-01
7.49742568e-01 4.58620965e-01 -6.22499287e-01 -3.07887614e-01
-1.50673258e+00 8.54748487e-01 2.53969252e-01 3.89714152e-01
-5.05525529e-01 -1.42214358e-01 7.89173067e-01 1.22196063e-01
3.88713092e-01 1.56105369e-01 -3.07148278e-01 -2.42031649e-01
-1.01167166e+00 4.55412082e-02 5.36672808e-02 7.25292504e-01
9.56899285e-01 3.47043872e-02 -6.72373354e-01 1.05452168e+00
3.10815543e-01 3.64865243e-01 5.64261734e-01 -7.27786839e-01
6.08695507e-01 6.94209099e-01 -3.49940330e-01 -1.30857408e+00
-1.77116245e-01 -3.78271371e-01 -9.15854156e-01 -2.23684758e-01
-7.69055560e-02 1.58167139e-01 -6.65442884e-01 1.87923992e+00
4.71898973e-01 6.65430188e-01 -2.27528766e-01 9.99585509e-01
9.80227351e-01 8.81411970e-01 2.70819753e-01 -4.33302015e-01
1.50482881e+00 -1.03280509e+00 -7.36381054e-01 2.15463519e-01
1.18382409e-01 -8.06374848e-01 1.01649821e+00 1.52076989e-01
-9.29038048e-01 -8.98618758e-01 -9.00611281e-01 -3.88064142e-03
-1.43827587e-01 1.59117132e-01 4.78278130e-01 6.00202531e-02
-2.95912534e-01 2.45511934e-01 -3.77246499e-01 6.41174912e-02
5.24723828e-01 2.37825379e-01 -5.88988423e-01 -2.78399587e-01
-1.29584992e+00 1.33221999e-01 5.36918521e-01 1.13316849e-01
-9.01226342e-01 -8.58636796e-01 -1.07056963e+00 7.35606998e-02
6.56352997e-01 -3.02048534e-01 6.31447136e-01 -8.63033831e-01
-1.15637136e+00 5.45515239e-01 -3.22648346e-01 1.35949776e-02
1.39462784e-01 -1.66328251e-01 -7.40104616e-01 6.81764364e-01
1.19777292e-01 5.71036279e-01 1.38146889e+00 -1.20548451e+00
-8.36729586e-01 -3.28878343e-01 3.80513221e-02 4.37455982e-01
-7.12943614e-01 -1.44373074e-01 -7.76310027e-01 -1.23257053e+00
1.65660545e-01 -6.25856042e-01 1.96715429e-01 1.26458198e-01
-2.79662907e-01 -1.67973384e-01 1.07157254e+00 -5.74979067e-01
1.57887423e+00 -2.54331589e+00 3.41748834e-01 2.48791724e-01
4.17782575e-01 1.25016123e-01 -4.25155312e-01 5.07647812e-01
-5.38607538e-02 -2.22555116e-01 3.41083817e-02 -1.48946941e-01
1.50962248e-01 -4.20985930e-02 -3.91180485e-01 5.06122887e-01
4.75220233e-01 6.48683071e-01 -9.25808489e-01 -8.58785510e-01
5.28425932e-01 7.79424131e-01 -4.62354928e-01 2.18124941e-01
2.27539927e-01 6.22307003e-01 -5.35039544e-01 9.42068517e-01
7.01326191e-01 -4.46271412e-02 -1.57665059e-01 -7.68946350e-01
-9.22484472e-02 2.05780496e-03 -1.26756346e+00 1.82238424e+00
-3.82268548e-01 3.29111993e-01 8.26999620e-02 -1.13748121e+00
8.13108861e-01 2.30772495e-01 7.10156798e-01 -8.34805548e-01
4.20379965e-03 9.35105756e-02 -1.17869854e-01 -8.22186410e-01
4.26693052e-01 -1.29860342e-01 -1.14621840e-01 1.31207258e-01
3.00206691e-01 4.38897610e-01 1.09403722e-01 1.62982091e-01
7.33377457e-01 1.43065751e-01 2.69769520e-01 -1.45816118e-01
9.40789640e-01 -6.29382908e-01 7.84198463e-01 3.81000489e-01
-4.06020254e-01 6.04521394e-01 2.37548038e-01 -6.26076013e-02
-5.07419527e-01 -1.03056562e+00 -1.23695157e-01 1.41761160e+00
5.01967728e-01 -6.43035412e-01 -3.00223082e-01 -7.34659731e-01
9.18782353e-02 2.14023232e-01 -9.03042138e-01 -4.63178962e-01
-4.23402369e-01 -3.94775480e-01 2.83815861e-01 5.44490576e-01
3.27325135e-01 -9.39442575e-01 -4.10773724e-01 6.66591451e-02
-3.68665606e-01 -8.63301754e-01 -7.77227402e-01 3.82398926e-02
-5.24819434e-01 -9.74127233e-01 -8.10502946e-01 -7.67050862e-01
4.20995653e-01 6.65964127e-01 7.22845972e-01 -1.07961088e-01
-3.23606968e-01 6.33228004e-01 -6.00830495e-01 1.36615664e-01
2.69088924e-01 -2.04810992e-01 -5.34702316e-02 6.40646517e-01
4.73588586e-01 -7.19033778e-01 -8.42070758e-01 2.49689549e-01
-1.09502733e+00 -3.31981331e-02 7.54498363e-01 1.29335892e+00
9.19436157e-01 3.29582691e-01 5.63214660e-01 -3.94276559e-01
3.48124593e-01 -6.91477537e-01 -1.18525256e-03 1.21329874e-01
-1.65893808e-01 -4.30785194e-02 5.01005888e-01 -7.87366986e-01
-1.02061975e+00 -1.28876582e-01 2.53898472e-01 -9.78799045e-01
-2.65254378e-02 6.74535096e-01 -4.65845555e-01 -1.85474068e-01
1.18207164e-01 6.10707462e-01 -1.76999599e-01 -3.82286936e-01
3.22154403e-01 3.89515340e-01 5.12219429e-01 -7.04456449e-01
8.19835663e-01 4.57682312e-01 8.03003907e-02 -8.99486065e-01
-6.32851124e-01 -7.72258461e-01 -4.08170670e-01 -4.37728614e-01
6.14918053e-01 -1.14794481e+00 -5.76601028e-01 3.04863632e-01
-5.56613386e-01 3.24518651e-01 -4.14785981e-01 5.65947533e-01
-3.38139653e-01 5.67633092e-01 -4.12177622e-01 -8.19644868e-01
-1.44455388e-01 -1.18104362e+00 1.40331638e+00 3.69814128e-01
1.05540216e-01 -8.79046023e-01 1.42165599e-02 4.04772729e-01
2.03663781e-01 4.37043250e-01 7.72147298e-01 -3.42752934e-01
-4.23120618e-01 -1.60996452e-01 -4.35711563e-01 2.91422397e-01
2.49232396e-01 -8.85570645e-02 -1.04246235e+00 -3.46142799e-01
-2.61545092e-01 -4.45199966e-01 1.11838567e+00 2.04693168e-01
1.44271433e+00 -2.03895867e-01 -2.28546709e-01 5.36333859e-01
1.26616681e+00 -6.04859032e-02 3.29227537e-01 2.22793087e-01
1.07511103e+00 6.33883119e-01 1.05154502e+00 8.55878949e-01
4.32824105e-01 7.65331566e-01 3.67559165e-01 3.80510874e-02
-2.88348570e-02 -4.52445924e-01 2.41194546e-01 9.74178374e-01
-4.68392298e-02 3.04925561e-01 -4.68383461e-01 5.65433860e-01
-1.98566461e+00 -1.23553860e+00 1.15154818e-01 2.17337632e+00
9.31131184e-01 -8.76892731e-02 2.23904803e-01 4.73376602e-01
8.26640248e-01 5.47903478e-01 -4.28804070e-01 2.26285398e-01
-3.04295957e-01 3.19663733e-02 -4.31477539e-02 1.03301637e-01
-1.51224613e+00 3.61237258e-01 4.12769842e+00 1.38540912e+00
-1.19110501e+00 1.89876273e-01 4.09013927e-01 -1.93421096e-01
-3.73647302e-01 -1.97575241e-02 -4.77013350e-01 8.61778617e-01
3.94114226e-01 4.69236188e-02 2.97458559e-01 5.63393354e-01
1.17976457e-01 -9.82058118e-04 -9.39927995e-01 1.37924814e+00
2.97591805e-01 -1.18326950e+00 6.76416382e-02 -8.18016306e-02
7.07688570e-01 -5.10293365e-01 2.76748985e-01 5.51540434e-01
-6.52716637e-01 -7.93456376e-01 9.42512810e-01 7.82325804e-01
9.06435966e-01 -1.01575732e+00 4.12001073e-01 9.68207791e-02
-1.92106247e+00 -3.49989474e-01 -6.14107586e-02 2.34536558e-01
-4.36255634e-02 5.13664603e-01 -1.11141959e-02 8.55306149e-01
7.62507260e-01 1.19204557e+00 -5.72453678e-01 1.08494246e+00
1.17011340e-02 5.40343225e-01 -7.45956227e-02 3.47350806e-01
3.68922889e-01 -9.26929489e-02 7.13031948e-01 1.45254016e+00
9.46204364e-02 7.93609172e-02 4.61825520e-01 5.49103677e-01
7.34268352e-02 1.43754870e-01 -2.27462023e-01 -1.11912526e-01
6.25718832e-01 1.37195635e+00 -2.88812429e-01 -1.68325022e-01
-4.96672750e-01 7.82724082e-01 1.68514416e-01 4.19489682e-01
-8.65948796e-01 -5.01901209e-01 6.04000032e-01 -1.18390828e-01
6.61452353e-01 4.77714138e-03 1.54701516e-01 -1.36604261e+00
1.58165589e-01 -1.02979457e+00 6.64288938e-01 -4.79840398e-01
-1.49599993e+00 4.00718510e-01 9.48098078e-02 -1.81658030e+00
-7.41029382e-02 -2.78564662e-01 -6.82915568e-01 8.67539048e-01
-1.96074867e+00 -1.67575181e+00 -5.42394876e-01 1.13231301e+00
6.01866782e-01 -7.50528798e-02 5.76264083e-01 6.36106670e-01
-7.03968823e-01 9.17208314e-01 7.17093498e-02 -2.57816315e-02
7.58919895e-01 -8.80366623e-01 -6.12025082e-01 7.77798772e-01
9.56990570e-02 6.69479907e-01 3.38361561e-01 -3.22608590e-01
-1.71627808e+00 -1.20355797e+00 2.49051765e-01 1.83187172e-01
7.93966532e-01 -2.41741315e-01 -1.05781758e+00 1.41478643e-01
5.79793891e-03 4.73881394e-01 9.92779613e-01 1.03172556e-01
-7.93499708e-01 -3.97163957e-01 -7.77176142e-01 3.31983298e-01
8.33989859e-01 -1.11190665e+00 -5.26647985e-01 1.62010882e-02
6.01034582e-01 -2.10117012e-01 -9.73627985e-01 6.18267357e-01
6.53116763e-01 -8.89876962e-01 1.07803214e+00 -4.16029006e-01
6.17076457e-01 -5.99455297e-01 -5.10704219e-01 -9.81705487e-01
-4.92231637e-01 -6.45781934e-01 -6.09229684e-01 1.65486312e+00
-1.60287932e-01 -3.00691664e-01 2.84953982e-01 3.71312499e-02
-1.36142701e-01 -1.05810761e+00 -1.08552587e+00 -6.02575839e-01
-3.31573606e-01 -4.26346093e-01 4.58307892e-01 1.18050313e+00
9.45448317e-03 2.14815617e-01 -6.40525818e-01 1.96743608e-01
7.93940127e-01 5.10632753e-01 4.72663373e-01 -9.12786245e-01
-5.18759489e-01 -4.78033930e-01 -5.96077979e-01 -8.88502300e-01
1.05931565e-01 -6.53566837e-01 -1.07208092e-03 -9.71395373e-01
5.75776994e-01 -3.27958524e-01 -9.52536285e-01 3.76580596e-01
-5.26472569e-01 4.46426898e-01 3.52114052e-01 3.81737977e-01
-1.00493371e+00 9.24581707e-01 1.18062305e+00 -2.42025301e-01
-8.47200751e-02 -4.47448462e-01 -6.62869215e-01 5.96713066e-01
3.27230096e-01 -2.12515518e-01 -4.80093092e-01 -1.15104578e-01
-2.12852746e-01 -2.65536923e-02 7.21775532e-01 -1.07276380e+00
8.84301886e-02 -1.98961154e-01 5.15713155e-01 -7.71998644e-01
6.30867243e-01 -8.79717946e-01 -1.07066110e-01 -1.05065089e-02
-4.64250624e-01 -3.97019655e-01 1.87662140e-01 9.57924187e-01
-7.37356365e-01 4.56044637e-02 7.18720376e-01 1.42873287e-01
-1.08468568e+00 4.93409455e-01 1.24445193e-01 2.02571094e-01
9.90035653e-01 -2.86827981e-01 -1.57288015e-01 -2.26947591e-01
-6.48496926e-01 2.63736695e-01 1.98109210e-01 5.84964871e-01
8.15895736e-01 -1.81529844e+00 -5.29255927e-01 3.17865014e-01
6.00307286e-01 -3.27855706e-01 9.91273046e-01 1.20686269e+00
3.17839772e-01 8.06657821e-02 -2.98044443e-01 -8.10515225e-01
-1.29765642e+00 5.39505899e-01 -1.75080206e-02 -1.66746810e-01
-4.13026065e-01 9.66686547e-01 5.25350749e-01 4.63396423e-02
2.40087762e-01 7.80934021e-02 -5.69126725e-01 4.89617974e-01
8.20911109e-01 4.34999138e-01 -3.06125730e-01 -1.21251452e+00
-4.81125474e-01 8.80973637e-01 -1.40572071e-01 1.04031406e-01
1.27203441e+00 -2.60489643e-01 -7.72944167e-02 5.77771842e-01
1.54836714e+00 1.69422552e-01 -1.46807349e+00 -6.42056167e-01
-4.72845733e-01 -8.28119695e-01 3.65690827e-01 -5.58985531e-01
-1.17808437e+00 1.10122347e+00 6.42945349e-01 1.27057016e-01
1.78881979e+00 -1.47535026e-01 6.79490447e-01 -3.30246426e-02
5.10050394e-02 -1.04057670e+00 5.77516973e-01 2.04672068e-01
8.89542162e-01 -1.39780390e+00 1.01137891e-01 -4.94148135e-01
-7.38920033e-01 1.12130892e+00 5.39717436e-01 -6.64138868e-02
7.81325281e-01 -1.81384057e-01 -4.04087275e-01 -2.72475213e-01
-4.94375318e-01 -2.59896487e-01 9.13008571e-01 4.84874368e-01
2.42615208e-01 3.94831076e-02 -3.91805209e-02 9.55129385e-01
4.26385522e-01 -1.78659126e-01 -1.55554160e-01 1.02817059e+00
-2.85995424e-01 -7.86684752e-01 -4.30990934e-01 4.07244265e-01
-2.74080783e-01 -1.93542287e-01 -4.90459502e-02 5.87628365e-01
5.28985381e-01 9.15247738e-01 -1.78811818e-01 -7.51330912e-01
1.61549896e-01 -2.22687796e-01 2.82747120e-01 -2.02209249e-01
-5.21666825e-01 6.71752810e-01 -2.43734568e-01 -6.15313947e-01
-6.30260527e-01 -7.81957209e-01 -1.07226360e+00 8.47223178e-02
-3.08386266e-01 9.77180153e-02 1.21525921e-01 7.99870133e-01
6.34990096e-01 7.21593440e-01 1.13482320e+00 -1.19542575e+00
-4.47485238e-01 -8.31059396e-01 -7.00026870e-01 5.81145823e-01
5.90071976e-01 -1.08972538e+00 -4.55764562e-01 1.09918699e-01] | [13.205392837524414, 4.816136360168457] |
66fed3f0-f060-4ab7-a372-a330a941a5f2 | direct-segmentation-of-brain-white-matter | 2307.02223 | null | https://arxiv.org/abs/2307.02223v1 | https://arxiv.org/pdf/2307.02223v1.pdf | Direct segmentation of brain white matter tracts in diffusion MRI | The brain white matter consists of a set of tracts that connect distinct regions of the brain. Segmentation of these tracts is often needed for clinical and research studies. Diffusion-weighted MRI offers unique contrast to delineate these tracts. However, existing segmentation methods rely on intermediate computations such as tractography or estimation of fiber orientation density. These intermediate computations, in turn, entail complex computations that can result in unnecessary errors. Moreover, these intermediate computations often require dense multi-shell measurements that are unavailable in many clinical and research applications. As a result, current methods suffer from low accuracy and poor generalizability. Here, we propose a new deep learning method that segments these tracts directly from the diffusion MRI data, thereby sidestepping the intermediate computation errors. Our experiments show that this method can achieve segmentation accuracy that is on par with the state of the art methods (mean Dice Similarity Coefficient of 0.826). Compared with the state of the art, our method offers far superior generalizability to undersampled data that are typical of clinical studies and to data obtained with different acquisition protocols. Moreover, we propose a new method for detecting inaccurate segmentations and show that it is more accurate than standard methods that are based on estimation uncertainty quantification. The new methods can serve many critically important clinical and scientific applications that require accurate and reliable non-invasive segmentation of white matter tracts. | ['Davood Karimi', 'Meritxell Bach Cuadra', 'Ali Gholipour', 'Hamza Kebiri'] | 2023-07-05 | null | null | null | null | ['brain-segmentation'] | ['medical'] | [-2.56665915e-01 -3.56655031e-01 9.07321572e-02 -2.81171799e-01
-5.03817856e-01 -5.40257215e-01 1.49526045e-01 5.35867736e-02
-6.91559911e-01 7.92949677e-01 1.21704265e-01 -3.53588313e-01
-9.23860744e-02 -6.15626454e-01 -2.37379700e-01 -7.05209732e-01
-3.46387625e-01 7.43842423e-01 5.11145413e-01 1.85866416e-01
2.05665499e-01 7.26955056e-01 -6.55488431e-01 -1.53474510e-01
1.27988064e+00 7.72496581e-01 2.20888719e-01 9.99843031e-02
-2.49001160e-01 5.73047437e-02 -2.64837265e-01 -3.14210147e-01
3.65816057e-01 -5.23944795e-01 -9.16398883e-01 -2.95635462e-02
2.38459617e-01 -6.23222530e-01 -7.62151778e-02 1.37421775e+00
6.08664215e-01 1.06224671e-01 7.39761889e-01 -7.85573721e-01
-6.30623758e-01 6.25073731e-01 -7.10600436e-01 7.03464627e-01
-1.41409025e-01 2.87862301e-01 5.21532357e-01 -6.37087405e-01
6.39012575e-01 7.47833490e-01 9.92362380e-01 5.37141681e-01
-1.36202502e+00 -6.42648101e-01 -6.27691671e-03 2.52699316e-01
-1.22200108e+00 -1.72052786e-01 4.27588850e-01 -8.35407257e-01
7.21895218e-01 4.29776013e-02 9.61998045e-01 6.76594555e-01
4.95755821e-01 4.90827352e-01 1.50073314e+00 2.91210383e-01
2.38698810e-01 -4.06635910e-01 1.62714660e-01 6.11763179e-01
5.56388378e-01 -3.73440771e-03 2.42126822e-01 -8.08799341e-02
1.06891966e+00 1.46383807e-01 -5.32296598e-01 -2.98407555e-01
-1.77125311e+00 7.62348771e-01 6.34049118e-01 5.47937751e-01
-6.21774137e-01 3.37399356e-03 3.76713634e-01 -6.18853942e-02
4.85083252e-01 3.08651626e-01 -1.29079580e-01 -5.74114732e-02
-1.70549023e+00 1.19419627e-01 4.56013083e-01 3.71107727e-01
1.98123053e-01 -3.41660120e-02 -8.78306106e-02 8.51668537e-01
4.24417257e-01 4.83618081e-01 7.73277879e-01 -9.62276697e-01
1.76828265e-01 1.76443845e-01 -1.15767241e-01 -8.31778824e-01
-1.01661623e+00 -6.24798059e-01 -1.16926813e+00 4.10041928e-01
7.86347091e-01 -1.31537676e-01 -1.03643048e+00 1.38496768e+00
3.71706724e-01 -6.52783513e-02 -6.53396249e-01 1.33025229e+00
5.08996129e-01 3.74389836e-03 4.13144939e-03 -3.35426271e-01
1.18090463e+00 -9.09770191e-01 -7.73019850e-01 5.36388755e-02
5.25639057e-01 -7.56277800e-01 8.70353281e-01 2.20084473e-01
-1.33371806e+00 -4.41866703e-02 -7.50257552e-01 5.12190908e-02
-1.07707478e-01 -7.09149688e-02 6.02233469e-01 6.66913450e-01
-1.13126194e+00 8.98014188e-01 -1.39520669e+00 -1.75253630e-01
7.48583496e-01 3.18729013e-01 -2.85698563e-01 5.40538169e-02
-8.37990403e-01 1.19819856e+00 1.07067160e-01 2.86094964e-01
-6.07150972e-01 -1.01440990e+00 -4.80092674e-01 -1.40049815e-01
5.35832271e-02 -6.48381948e-01 9.04455185e-01 -3.42079818e-01
-1.31196070e+00 7.87482142e-01 -9.91493184e-03 -4.56583560e-01
9.30165887e-01 8.00032541e-02 -1.50887921e-01 6.38613760e-01
2.66927838e-01 5.17856359e-01 6.58749104e-01 -8.60486329e-01
9.93662626e-02 -6.72597289e-01 -3.42860967e-01 -8.83298591e-02
7.34133124e-02 1.72063828e-01 -3.32185030e-02 -7.61477709e-01
7.59936810e-01 -1.04822779e+00 -4.76395100e-01 3.77631426e-01
-4.85697120e-01 1.98884636e-01 3.57054204e-01 -1.12143767e+00
9.54037726e-01 -1.55954969e+00 1.14898860e-01 3.15518081e-01
9.27805424e-01 2.99850851e-01 1.86828539e-01 -6.80781603e-02
-4.69306856e-02 2.43709087e-01 -8.04626703e-01 -1.22654559e-02
-2.65203863e-01 -1.30811796e-01 2.09242985e-01 1.12583172e+00
-3.42676818e-01 8.95743728e-01 -8.68014395e-01 -5.56572795e-01
1.75139546e-01 5.41308463e-01 -4.30661142e-01 5.00038639e-02
4.37192470e-01 9.19508755e-01 -3.15472156e-01 3.22011977e-01
9.10195351e-01 -3.72741580e-01 7.18493238e-02 -3.66719067e-01
-1.58620298e-01 1.77224100e-01 -8.44475985e-01 1.66643560e+00
-9.65183824e-02 3.84587914e-01 1.96318343e-01 -1.27561533e+00
5.31927943e-01 4.05935109e-01 8.96713674e-01 -3.72634977e-01
3.46249670e-01 5.31820357e-01 6.76667929e-01 -5.72964311e-01
-2.66092211e-01 -6.02503061e-01 6.95875585e-01 9.11545753e-01
-3.34419534e-02 -3.50663453e-01 3.12529415e-01 3.42538133e-02
9.45676923e-01 -3.98559831e-02 1.24629281e-01 -6.53446794e-01
3.60725075e-01 -2.31758043e-01 6.35634601e-01 3.92886072e-01
-8.09587419e-01 7.68149018e-01 3.37874502e-01 -4.31731224e-01
-1.12809527e+00 -1.33891129e+00 -5.77436924e-01 1.72179282e-01
-3.90148833e-02 8.26260373e-02 -1.15990591e+00 -6.39307201e-01
-5.50801083e-02 2.76401997e-01 -3.93424630e-01 4.97921519e-02
-5.60564637e-01 -1.26175189e+00 4.59099799e-01 5.67146003e-01
5.65526128e-01 -7.74140954e-01 -5.80285370e-01 3.78898352e-01
-3.83797318e-01 -1.35836279e+00 -6.37542963e-01 -3.35505813e-01
-1.43292487e+00 -1.22864449e+00 -1.35626328e+00 -3.80483896e-01
8.38339388e-01 2.19141454e-01 8.99204433e-01 3.60213518e-01
-4.04700726e-01 -5.51640950e-02 -1.27570987e-01 1.43492565e-01
-1.73063442e-01 4.93915230e-02 2.69322008e-01 -2.22013429e-01
1.99174419e-01 -9.19631422e-01 -1.13493001e+00 3.00231159e-01
-7.88487434e-01 6.92967772e-02 5.54163218e-01 6.80068731e-01
5.77211261e-01 -1.84536293e-01 6.45416677e-01 -6.76165164e-01
7.10749686e-01 -6.02803349e-01 -5.62887251e-01 -6.29564822e-02
-8.15798342e-01 -5.16213737e-02 5.44398606e-01 -3.88390988e-01
-9.25046504e-01 -3.76433492e-01 -3.89141858e-01 -1.60398737e-01
-1.98714763e-01 3.87833029e-01 5.11137009e-01 -4.00695622e-01
5.79186440e-01 8.32813457e-02 4.73465234e-01 -6.18276358e-01
8.29790980e-02 3.53810728e-01 3.32469076e-01 -4.09658015e-01
5.48116446e-01 8.16404581e-01 6.73065856e-02 -5.84229767e-01
-4.55026180e-01 -3.50159079e-01 -1.12801945e+00 -2.95253217e-01
1.00931609e+00 -3.38542789e-01 -4.75868732e-01 5.03158510e-01
-1.00167191e+00 -4.28391069e-01 4.27740365e-02 1.13272941e+00
-5.28260410e-01 8.36047590e-01 -1.10351133e+00 -4.33591664e-01
-6.34228587e-01 -1.73538709e+00 6.23451650e-01 -7.18246251e-02
-2.27978840e-01 -1.19684291e+00 -2.73566078e-02 4.43230391e-01
7.29954481e-01 3.13583702e-01 8.61767590e-01 -4.98599082e-01
-2.89918214e-01 1.64630860e-01 -4.39449340e-01 2.74053603e-01
1.67415172e-01 -2.52116412e-01 -4.11380410e-01 -3.37330133e-01
1.88377440e-01 8.27710778e-02 7.47325063e-01 8.64771426e-01
1.29180706e+00 1.49136975e-01 -3.40645701e-01 7.16891348e-01
1.28325748e+00 1.35788158e-01 4.23215419e-01 1.80974558e-01
6.41940951e-01 5.81620693e-01 2.46045459e-02 1.47711365e-02
3.67995620e-01 4.21264976e-01 1.93225831e-01 -2.33393595e-01
-5.28953254e-01 5.52151382e-01 -1.53032511e-01 1.25113654e+00
-5.11345863e-01 4.90936995e-01 -1.23126733e+00 7.10377574e-01
-1.54654014e+00 -9.15386200e-01 -6.41647339e-01 2.16079426e+00
1.08353949e+00 -3.17188427e-02 3.09968591e-01 1.86157674e-02
7.99212515e-01 -1.89500570e-01 -7.29150593e-01 -1.53841808e-01
2.52822548e-01 2.03408301e-01 5.26509285e-01 4.32398230e-01
-7.25326538e-01 6.04471266e-01 7.42941952e+00 4.39459771e-01
-1.23674273e+00 6.57989025e-01 5.09315550e-01 -1.96144149e-01
-2.35374674e-01 -3.62283736e-01 -4.02849406e-01 7.11166918e-01
7.22529531e-01 -4.34590466e-02 6.25565052e-01 4.09582108e-01
4.83530313e-01 -2.13453606e-01 -7.65110731e-01 9.48248446e-01
-2.05696717e-01 -1.15072858e+00 -4.04823691e-01 3.35087478e-01
7.52228558e-01 6.13871455e-01 -1.35505259e-01 -3.32888246e-01
1.45680636e-01 -9.97261763e-01 4.68291730e-01 5.87607324e-01
7.18156815e-01 -3.83122921e-01 7.69986987e-01 2.22561046e-01
-7.67940223e-01 2.32760429e-01 -4.13276047e-01 6.26206696e-02
5.92695236e-01 1.41195273e+00 -5.11041224e-01 1.76253691e-01
7.20481932e-01 4.55681235e-01 -1.90988973e-01 1.61438572e+00
-2.48643115e-01 6.13239348e-01 -3.46317887e-01 2.93060809e-01
2.63582468e-01 -6.97297037e-01 6.72375619e-01 1.06679761e+00
4.08244848e-01 1.26421943e-01 -2.18140744e-02 1.43795240e+00
7.06264935e-03 1.99103311e-01 -3.74418676e-01 2.87210979e-02
2.13772714e-01 1.39029038e+00 -1.33101737e+00 -4.67700720e-01
-5.34687877e-01 8.24466586e-01 4.44746017e-01 4.00417417e-01
-7.30778158e-01 -7.06008449e-02 7.35630870e-01 2.24728957e-01
-2.28848606e-02 -7.10308254e-01 -5.74332535e-01 -1.41507995e+00
8.52710009e-02 -5.06727874e-01 -3.68382037e-02 -3.52647781e-01
-1.51737380e+00 7.51480758e-01 2.13540792e-02 -8.89999807e-01
-3.62654217e-02 -5.13716400e-01 -6.13418996e-01 9.62042809e-01
-1.56772101e+00 -6.21708035e-01 -9.72828418e-02 4.81625676e-01
1.83115125e-01 1.93766475e-01 5.78336596e-01 5.04975080e-01
-6.33250117e-01 1.91207826e-01 2.13865399e-01 3.32504481e-01
4.83161718e-01 -1.27133954e+00 2.54036725e-01 9.23815012e-01
-2.14138448e-01 8.42408955e-01 4.26863343e-01 -7.78987944e-01
-7.64149606e-01 -8.95126104e-01 5.16192794e-01 3.30530554e-02
9.22678947e-01 5.42617440e-02 -1.09578848e+00 6.82899952e-01
-7.38843950e-03 4.78906393e-01 8.89799237e-01 -1.20298862e-01
-5.50477020e-02 2.07234308e-01 -1.46383929e+00 4.74419296e-01
1.06706977e+00 -2.30143115e-01 -8.60102415e-01 5.10923505e-01
2.06762850e-01 -3.67497414e-01 -1.29332423e+00 2.27559760e-01
6.79914176e-01 -1.02978384e+00 1.18803155e+00 -3.83471549e-01
2.66800940e-01 -1.17067881e-01 4.26808089e-01 -1.67560077e+00
-4.37347591e-01 -1.49162620e-01 2.50700340e-02 6.10842645e-01
2.22658738e-01 -8.28234851e-01 4.90056127e-01 7.88172007e-01
-1.58813640e-01 -8.17523956e-01 -1.00117350e+00 -1.01027060e+00
7.48963296e-01 -2.71067500e-01 5.91577530e-01 1.18499815e+00
8.10216814e-02 -3.13556641e-01 3.14940006e-01 -1.21676005e-01
1.21977222e+00 2.25556344e-01 -1.31260544e-01 -1.35743904e+00
2.01926306e-01 -9.11807656e-01 -3.74917746e-01 -7.08290815e-01
1.35604113e-01 -1.19640756e+00 -3.31840850e-02 -1.90489721e+00
2.16026023e-01 -7.13520646e-01 -1.82429567e-01 3.21740240e-01
-1.83758900e-01 6.20575249e-01 5.62787019e-02 4.73885655e-01
3.47759053e-02 3.78799051e-01 1.81462097e+00 8.12033564e-03
-9.03016701e-02 -2.61362493e-01 -3.85984719e-01 1.09522021e+00
1.07482088e+00 -4.11167741e-01 -2.61677712e-01 -5.40607154e-01
-1.32279813e-01 -5.61022647e-02 3.59707087e-01 -9.77854729e-01
1.52300015e-01 -7.18600973e-02 4.30694550e-01 -4.01688069e-01
-1.89492218e-02 -6.80781782e-01 -1.73775405e-01 7.97980964e-01
-3.15578543e-02 4.07536291e-02 -8.19532052e-02 7.22625628e-02
9.62562561e-02 -3.89949232e-01 1.14226711e+00 -4.72406089e-01
-1.43829957e-01 6.74247146e-01 -5.44232130e-01 2.69735008e-01
8.72057796e-01 1.03933010e-02 -9.22055095e-02 -1.13205254e-01
-1.06854713e+00 -3.73018347e-02 2.37675026e-01 -1.40810356e-01
6.46388292e-01 -1.31883991e+00 -7.32711136e-01 -1.51937343e-02
-5.49451172e-01 -2.38011748e-01 2.67339736e-01 1.90878367e+00
-8.92913699e-01 3.46218526e-01 -5.31483829e-01 -6.69926286e-01
-7.89878249e-01 5.00001550e-01 5.91093183e-01 -2.59592474e-01
-1.12978852e+00 5.55201292e-01 -1.11313518e-02 -4.62587774e-01
-3.10894668e-01 -6.28501713e-01 -1.61847532e-01 5.85708283e-02
6.82341397e-01 6.16273880e-01 2.42417768e-01 -7.57691562e-01
-3.97193670e-01 6.88687503e-01 3.24429758e-02 -1.38871193e-01
1.34388459e+00 -3.60529810e-01 -5.96716583e-01 1.34624541e-01
1.16832817e+00 -1.71369791e-01 -1.03633690e+00 -9.57568511e-02
-9.53843296e-02 -4.74588811e-01 4.99116004e-01 -8.14257205e-01
-1.65517092e+00 1.22915983e+00 7.32264280e-01 1.53542846e-01
8.52803946e-01 -2.01884389e-01 1.39667809e+00 -1.30095631e-01
4.81270462e-01 -1.10459328e+00 -3.85885030e-01 1.02169104e-01
7.93512404e-01 -1.37390614e+00 1.18300781e-01 -4.59949881e-01
-3.12589675e-01 1.22046208e+00 3.92016232e-01 -2.87423402e-01
8.96885753e-01 3.59938234e-01 2.10607007e-01 -3.70806277e-01
7.03036785e-02 -2.59221554e-01 4.52983946e-01 7.26537704e-01
5.97127080e-01 2.12611839e-01 -1.12993729e+00 4.24656719e-01
-1.78913668e-01 1.00602433e-01 4.23325241e-01 5.21942556e-01
-4.44190115e-01 -9.15945113e-01 -4.90836203e-01 9.00676131e-01
-7.47052491e-01 -2.71510452e-01 2.14005053e-01 5.89880645e-01
-9.99491885e-02 7.08329678e-01 3.66404057e-02 1.53139368e-01
-1.38201132e-01 -7.46604986e-03 7.55401194e-01 -4.43695426e-01
-4.75654542e-01 1.67357236e-01 -2.93414861e-01 -7.76971757e-01
-5.01356244e-01 -7.54931152e-01 -1.73961473e+00 -4.55125898e-01
-2.07386002e-01 2.60080323e-02 7.12507963e-01 1.25927210e+00
1.83966413e-01 5.47493935e-01 1.21035121e-01 -8.80366862e-01
-4.98291135e-01 -1.01342309e+00 -7.96171606e-01 5.42442262e-01
2.51705319e-01 -1.02978897e+00 -2.68316120e-01 -1.55797720e-01] | [14.07314395904541, -2.244398832321167] |
d1b6cb07-aaaa-4793-bfdc-11eb8795ee91 | actionformer-localizing-moments-of-actions | 2202.07925 | null | https://arxiv.org/abs/2202.07925v2 | https://arxiv.org/pdf/2202.07925v2.pdf | ActionFormer: Localizing Moments of Actions with Transformers | Self-attention based Transformer models have demonstrated impressive results for image classification and object detection, and more recently for video understanding. Inspired by this success, we investigate the application of Transformer networks for temporal action localization in videos. To this end, we present ActionFormer -- a simple yet powerful model to identify actions in time and recognize their categories in a single shot, without using action proposals or relying on pre-defined anchor windows. ActionFormer combines a multiscale feature representation with local self-attention, and uses a light-weighted decoder to classify every moment in time and estimate the corresponding action boundaries. We show that this orchestrated design results in major improvements upon prior works. Without bells and whistles, ActionFormer achieves 71.0% mAP at tIoU=0.5 on THUMOS14, outperforming the best prior model by 14.1 absolute percentage points. Further, ActionFormer demonstrates strong results on ActivityNet 1.3 (36.6% average mAP) and EPIC-Kitchens 100 (+13.5% average mAP over prior works). Our code is available at http://github.com/happyharrycn/actionformer_release. | ['Yin Li', 'Jianxin Wu', 'Chenlin Zhang'] | 2022-02-16 | null | null | null | null | ['action-localization'] | ['computer-vision'] | [ 2.78677732e-01 -7.70239383e-02 -6.03236973e-01 -7.18592182e-02
-8.56397390e-01 -4.92025226e-01 6.79748297e-01 -3.93283665e-01
-3.74374688e-01 3.67954642e-01 6.79926991e-01 1.54884264e-01
9.99288633e-02 -3.69876802e-01 -6.80716574e-01 -5.62720001e-01
-4.13932115e-01 -5.47238812e-02 5.65190196e-01 8.59385133e-02
2.97503799e-01 1.98251694e-01 -1.44716585e+00 8.02521288e-01
1.86620489e-01 1.23635352e+00 1.95861846e-01 9.44847047e-01
4.37352449e-01 1.33120871e+00 -5.55890203e-01 -1.30177960e-01
1.01814874e-01 -5.60276389e-01 -6.88420057e-01 1.03236884e-01
5.55914760e-01 -4.87605393e-01 -7.92660058e-01 7.01419294e-01
4.73572403e-01 1.91629559e-01 4.70705271e-01 -1.31230235e+00
-3.69193494e-01 6.65657520e-01 -5.41476250e-01 8.48113716e-01
4.70760643e-01 3.49696696e-01 1.12742269e+00 -1.10043585e+00
6.41652763e-01 1.00329363e+00 7.98990369e-01 4.01492357e-01
-9.41270411e-01 -7.73592770e-01 1.50915682e-01 8.36493671e-01
-1.40176857e+00 -9.84355330e-01 4.00554478e-01 -3.43996972e-01
1.29398787e+00 2.00895160e-01 7.45161474e-01 1.42822230e+00
3.13472003e-01 1.19837487e+00 6.88817739e-01 -1.02847852e-01
2.49101728e-01 -3.11473459e-01 -3.51504572e-02 6.40050113e-01
-3.59022111e-01 -2.25417376e-01 -1.00282371e+00 3.49317908e-01
8.84257972e-01 1.49683774e-01 -2.77562827e-01 5.07300999e-03
-1.57255328e+00 4.33401525e-01 3.59475464e-01 5.47014654e-01
-5.28726339e-01 8.73478651e-01 4.36777264e-01 1.15662210e-01
5.58904111e-01 2.63122052e-01 -3.62203032e-01 -7.80151129e-01
-1.10970557e+00 -1.14030905e-01 3.14923525e-01 9.25857663e-01
2.62127995e-01 2.07215518e-01 -5.02481222e-01 6.64024532e-01
1.08225793e-01 3.81496936e-01 4.02105957e-01 -1.58086932e+00
2.48208284e-01 1.79300681e-01 1.13324761e-01 -9.37200785e-01
-3.03903788e-01 -6.60157979e-01 -4.90907162e-01 4.06630859e-02
4.71140414e-01 1.40515845e-02 -8.62013757e-01 1.69126487e+00
-1.93180189e-01 7.04912424e-01 -2.40066811e-01 8.28513503e-01
5.54507256e-01 5.64518750e-01 1.24843337e-01 -1.23141766e-01
1.35317445e+00 -1.06114352e+00 -7.03201056e-01 -3.31808746e-01
5.34535944e-01 -4.63688076e-01 8.90999019e-01 5.76462567e-01
-1.28161788e+00 -6.72630906e-01 -9.65305150e-01 -1.88927110e-02
-1.56936362e-01 2.68737286e-01 4.94227737e-01 4.09101218e-01
-1.34580696e+00 6.85777247e-01 -1.15447474e+00 -7.13789701e-01
8.05229783e-01 1.53112128e-01 -2.75323808e-01 1.01541415e-01
-8.70206594e-01 6.78753316e-01 1.45207688e-01 -1.50513306e-01
-1.41055453e+00 -6.07570350e-01 -5.49227118e-01 6.98870644e-02
5.20077109e-01 -3.85401845e-01 1.54420543e+00 -1.04455698e+00
-1.44100344e+00 6.17790937e-01 -4.17838246e-01 -1.00600958e+00
4.83353108e-01 -5.17116129e-01 -5.23065925e-01 5.38531423e-01
2.62809396e-01 9.18842494e-01 6.86627805e-01 -5.50637305e-01
-8.49127829e-01 -4.02910486e-02 2.85416663e-01 2.10080072e-01
-4.20257539e-01 3.06647122e-01 -8.14980626e-01 -6.79532349e-01
2.91621946e-02 -6.55653298e-01 -3.79835963e-02 1.41751423e-01
-7.33697563e-02 -2.26048857e-01 7.02182710e-01 -7.46250212e-01
1.31740105e+00 -2.19975805e+00 -1.76457036e-02 -2.35834330e-01
3.87192905e-01 2.64567081e-02 -3.67713958e-01 4.22422618e-01
-8.94692913e-02 2.81226337e-02 1.60987243e-01 -3.16871017e-01
1.09837521e-02 -8.67195278e-02 -1.30997390e-01 7.24538922e-01
9.90354195e-02 9.97969449e-01 -9.62612033e-01 -4.01760638e-01
5.34892499e-01 5.67341030e-01 -6.91817522e-01 -7.69168064e-02
5.07083628e-03 1.78003132e-01 -2.17202947e-01 8.49646628e-01
5.56403510e-02 -4.13898349e-01 3.58133018e-02 -4.80302900e-01
-2.69145250e-01 1.74159631e-01 -1.01754773e+00 2.00872517e+00
-2.96207488e-01 1.19575262e+00 -1.84859574e-01 -9.58128273e-01
3.85049582e-01 4.69572216e-01 9.82864797e-01 -8.63293886e-01
2.17971921e-01 -1.20692551e-01 9.14781727e-03 -5.03544092e-01
2.72380054e-01 3.29601020e-01 1.33649364e-01 3.45098853e-01
4.69501674e-01 5.67834139e-01 5.33211648e-01 4.19386417e-01
1.76667249e+00 2.94792414e-01 2.43640900e-01 -1.19448565e-01
2.25476861e-01 -2.83962071e-01 5.38532197e-01 8.92182946e-01
-6.10314131e-01 8.11856747e-01 4.58024561e-01 -3.83015037e-01
-6.48538649e-01 -1.13771093e+00 1.48798391e-01 1.54312086e+00
9.46659967e-02 -7.79422820e-01 -7.08867192e-01 -7.44523108e-01
-3.37169141e-01 7.58016884e-01 -8.15151632e-01 -2.13433608e-01
-4.40268576e-01 -4.62746918e-01 6.66057408e-01 9.48131979e-01
7.39077091e-01 -1.15732658e+00 -9.94017422e-01 2.63954520e-01
-5.01782656e-01 -1.24285984e+00 -5.66040516e-01 1.35553554e-01
-6.46053493e-01 -1.08824396e+00 -8.64767194e-01 -3.65248561e-01
1.87978521e-01 3.26123655e-01 1.15743852e+00 -4.02781814e-01
-1.63957536e-01 7.34817386e-01 -5.14862776e-01 -7.17392936e-02
1.09122470e-01 -1.35640010e-01 1.04288511e-01 1.97394922e-01
5.08593559e-01 -7.11549222e-01 -9.59486544e-01 6.16798043e-01
-3.35122466e-01 1.13179848e-01 6.24696195e-01 4.13173556e-01
3.18813235e-01 -2.99211085e-01 4.48668748e-01 -7.10467249e-02
4.75090463e-03 -5.98870516e-01 -2.24731848e-01 1.18236251e-01
-2.33756706e-01 -2.74153173e-01 1.62053660e-01 -4.31669384e-01
-8.25682700e-01 3.66144814e-02 -6.01943992e-02 -6.44107401e-01
-1.88609257e-01 7.09403157e-02 1.59801483e-01 9.11052302e-02
7.17194855e-01 3.68227392e-01 -2.36218587e-01 -3.50126833e-01
3.04735124e-01 5.77469468e-01 6.74398601e-01 -4.69893776e-02
1.14345856e-01 7.57751465e-01 -4.21191394e-01 -7.94040442e-01
-8.46505821e-01 -6.64823592e-01 -5.24606407e-01 -7.14543819e-01
1.21095991e+00 -1.11863649e+00 -7.91152239e-01 3.80994201e-01
-9.93298113e-01 -7.25260198e-01 -2.71770865e-01 5.73618531e-01
-8.76660109e-01 2.19393238e-01 -7.04385459e-01 -7.37930655e-01
-1.09775662e-01 -7.91985869e-01 1.04942834e+00 1.71245337e-02
-5.22310376e-01 -6.74556553e-01 -9.13735703e-02 4.88568604e-01
4.71132934e-01 7.47856647e-02 -2.92490363e-01 -4.60751057e-01
-6.65664196e-01 -1.14641502e-01 -2.11052805e-01 3.03768277e-01
-2.07860651e-03 -1.64044842e-01 -1.03803527e+00 -1.41968086e-01
-1.14419647e-01 -2.93678671e-01 1.07907259e+00 6.68135703e-01
1.35446453e+00 -1.23754159e-01 -4.26230818e-01 5.79836845e-01
1.00182915e+00 4.00107116e-01 7.96402454e-01 2.55703479e-01
3.87738705e-01 -3.98078039e-02 5.52058756e-01 7.32052386e-01
2.04439506e-01 8.76339257e-01 4.99915898e-01 6.88622817e-02
-5.90319633e-01 -3.83825377e-02 8.22023153e-01 5.39918244e-01
-4.70532954e-01 -5.35238624e-01 -9.56220925e-01 6.86688185e-01
-1.90197754e+00 -1.47536361e+00 2.46204901e-02 1.92243552e+00
4.51850086e-01 3.63887131e-01 2.18356356e-01 -4.54767346e-02
5.09269774e-01 4.22196835e-01 -5.33344865e-01 1.28192395e-01
-1.41441122e-01 1.67978927e-01 6.08966053e-01 2.67913282e-01
-1.38513577e+00 9.36050773e-01 6.56155825e+00 9.56645131e-01
-7.58017361e-01 5.09345114e-01 6.51666880e-01 -6.32588685e-01
4.25785452e-01 -2.15946466e-01 -5.03298938e-01 6.32999957e-01
1.12896609e+00 -1.15181521e-01 3.14059943e-01 6.18197560e-01
6.49600565e-01 -4.86181527e-01 -1.16967690e+00 1.22653973e+00
3.12170565e-01 -1.43386674e+00 -3.42048556e-01 -4.58737835e-02
5.86030304e-01 4.49214637e-01 -9.74407140e-03 2.99606323e-01
2.55784541e-01 -8.11868370e-01 9.76158738e-01 7.40935028e-01
7.99656034e-01 -4.60482925e-01 5.73060155e-01 1.77172087e-02
-1.48062432e+00 -2.90765405e-01 -1.82418469e-02 -2.36940116e-01
3.15503061e-01 2.57793993e-01 -5.83137929e-01 1.66426986e-01
9.49469030e-01 1.45958972e+00 -5.28306067e-01 1.21310270e+00
-2.03921050e-01 9.76384401e-01 -2.80114233e-01 1.06758952e-01
2.94092834e-01 3.02575201e-01 5.49740732e-01 1.52559555e+00
4.31338876e-01 2.21105650e-01 1.50463253e-01 4.50839758e-01
-2.77141891e-02 -2.72806376e-01 -3.55412632e-01 -3.05345152e-02
2.91349471e-01 1.01870191e+00 -8.97222817e-01 -6.45773232e-01
-4.19246048e-01 1.26202214e+00 -6.23083040e-02 5.36063850e-01
-1.43595004e+00 -3.19375873e-01 7.31393933e-01 1.64316028e-01
5.76541960e-01 -2.59367675e-01 -3.61362323e-02 -1.18161917e+00
-1.27982959e-01 -7.47434795e-01 4.58167225e-01 -1.06099164e+00
-6.72860563e-01 4.29131925e-01 9.11027845e-03 -1.40938079e+00
-2.03690469e-01 -4.13578898e-01 -5.60889184e-01 2.30649233e-01
-8.56784761e-01 -9.15463984e-01 -3.82920563e-01 5.39012849e-01
1.15214241e+00 -6.84195161e-02 4.59846139e-01 5.34785926e-01
-6.59157395e-01 4.84321713e-01 -1.18325718e-01 2.54610926e-01
6.58124983e-01 -1.12100267e+00 3.53521824e-01 1.06023645e+00
5.45560062e-01 1.38185039e-01 7.73381829e-01 -4.63797897e-01
-1.26634967e+00 -9.61224139e-01 7.46949255e-01 -7.65038788e-01
9.63461399e-01 -3.56240928e-01 -4.68181610e-01 8.11434865e-01
5.60608149e-01 1.55317932e-01 4.40626949e-01 -3.48746143e-02
-3.03499848e-01 -2.45564014e-01 -7.80387461e-01 6.87099814e-01
1.66711199e+00 -4.91326600e-01 -4.49943602e-01 5.36842108e-01
3.02729905e-01 -1.44322440e-01 -8.70097160e-01 2.83667326e-01
7.76192605e-01 -1.30930448e+00 9.76082981e-01 -3.08708847e-01
3.93746495e-01 -2.34983876e-01 -3.16792309e-01 -8.21348965e-01
-7.28551567e-01 -6.72772884e-01 -5.31276584e-01 8.63191843e-01
1.71577752e-01 -3.31210524e-01 9.66638744e-01 1.71530455e-01
-3.98395926e-01 -6.61381602e-01 -1.03740621e+00 -9.60940778e-01
-5.61012626e-01 -9.44080114e-01 -1.17512248e-01 5.83990335e-01
2.02225357e-01 1.01879857e-01 -5.63128471e-01 -9.45965797e-02
5.18399298e-01 -3.53446245e-01 5.85860610e-01 -5.35355389e-01
-3.63920867e-01 -6.63421690e-01 -7.30515540e-01 -1.43310797e+00
-2.09320411e-01 -5.61804116e-01 8.64161998e-02 -1.58346784e+00
2.83066928e-01 1.80651143e-01 -8.28782260e-01 8.30470562e-01
3.26172322e-01 7.90621102e-01 2.89217591e-01 8.33117813e-02
-1.43285155e+00 5.68407893e-01 7.44588137e-01 -1.54625133e-01
6.29787892e-02 -1.13652341e-01 -4.32697266e-01 8.09452236e-01
1.15879917e+00 -3.47651601e-01 -3.74233603e-01 -3.89164776e-01
-1.78231969e-01 -1.02511242e-01 5.65505385e-01 -1.76923835e+00
3.62845808e-01 -1.72952376e-02 4.30009633e-01 -6.55784786e-01
6.49561167e-01 -5.30104458e-01 5.61562330e-02 5.56772709e-01
-4.14472044e-01 1.96632594e-02 1.72627941e-01 6.41923785e-01
-1.00299008e-01 2.14666992e-01 5.75573027e-01 -1.60790458e-01
-1.34519148e+00 1.70695245e-01 -8.12148690e-01 9.49704573e-02
1.09306157e+00 -4.92852718e-01 -2.61673808e-01 -5.94421029e-01
-1.01197696e+00 1.58205464e-01 1.14442974e-01 5.53377509e-01
5.53765655e-01 -1.34229970e+00 -6.83628440e-01 -1.95311401e-02
7.76988789e-02 -8.71976435e-01 4.80381936e-01 1.36525059e+00
-1.85053647e-01 4.55807239e-01 -3.51234943e-01 -7.52232432e-01
-1.30953932e+00 2.92649388e-01 4.85418141e-01 7.18769655e-02
-5.29503763e-01 1.00913846e+00 1.81493759e-01 5.04669487e-01
5.04274845e-01 -4.27668512e-01 -1.89603284e-01 1.57205552e-01
7.86935270e-01 6.36893332e-01 -3.12192410e-01 -6.03565037e-01
-6.42786384e-01 4.74540323e-01 1.26983657e-01 -2.96602905e-01
1.23435247e+00 -9.21197087e-02 3.42276841e-01 5.08299947e-01
1.09247684e+00 -2.33327925e-01 -1.79627645e+00 -4.95295860e-02
2.12910920e-02 -4.25402552e-01 1.61153257e-01 -9.42309380e-01
-1.13598061e+00 7.48119593e-01 7.72384226e-01 8.89303461e-02
1.21907878e+00 2.98473477e-01 4.89321232e-01 2.46642470e-01
3.57897222e-01 -1.13070059e+00 6.02072775e-01 4.68641251e-01
1.05857754e+00 -1.05010247e+00 -1.13389157e-01 -1.13836512e-01
-5.07091045e-01 9.27886963e-01 6.59704268e-01 -1.52960911e-01
4.82704341e-01 2.19398841e-01 -1.67356744e-01 -1.85794502e-01
-1.05092502e+00 -3.90567392e-01 3.08266759e-01 4.81247514e-01
3.76694173e-01 -1.35992661e-01 -4.79896329e-02 4.24250573e-01
3.29259574e-01 3.73195648e-01 4.42347735e-01 8.02960992e-01
-6.16575122e-01 -5.03286779e-01 -1.38101682e-01 4.73137140e-01
-6.02443397e-01 -1.23699963e-01 -1.86923966e-01 5.52738965e-01
2.25078955e-01 1.10848475e+00 3.19942236e-01 -6.42042458e-01
1.51937291e-01 -1.39580593e-02 4.05226320e-01 -3.88619959e-01
-3.73809457e-01 3.63834441e-01 3.66123766e-01 -1.28161395e+00
-6.12714052e-01 -8.65498662e-01 -9.87196863e-01 -1.38768598e-01
5.86365201e-02 -1.68624282e-01 2.25254834e-01 6.05899692e-01
6.80298984e-01 6.99108064e-01 3.60508263e-01 -1.05131984e+00
-1.14803992e-01 -1.13800418e+00 -2.59229779e-01 9.13754478e-02
1.24796495e-01 -6.99514449e-01 -3.52439851e-01 2.95702636e-01] | [8.402515411376953, 0.4454311728477478] |
bd34cd16-6814-4882-a277-dbe0da379af7 | building-a-computer-mahjong-player-via-deep | 1906.02146 | null | https://arxiv.org/abs/1906.02146v2 | https://arxiv.org/pdf/1906.02146v2.pdf | Building a Computer Mahjong Player via Deep Convolutional Neural Networks | The evaluation function for imperfect information games is always hard to define but owns a significant impact on the playing strength of a program. Deep learning has made great achievements these years, and already exceeded the top human players' level even in the game of Go. In this paper, we introduce a new data model to represent the available imperfect information on the game table, and construct a well-designed convolutional neural network for game record training. We choose the accuracy of tile discarding which is also called as the agreement rate as the benchmark for this study. Our accuracy on test data reaches 70.44%, while the state-of-art baseline is 62.1% reported by Mizukami and Tsuruoka (2015), and is significantly higher than previous trials using deep learning, which shows the promising potential of our new model. For the AI program building, besides the tile discarding strategy, we adopt similar predicting strategies for other actions such as stealing (pon, chi, and kan) and riichi. With the simple combination of these several predicting networks and without any knowledge about the concrete rules of the game, a strength evaluation is made for the resulting program on the largest Japanese Mahjong site `Tenhou'. The program has achieved a rating of around 1850, which is significantly higher than that of an average human player and of programs among past studies. | ['Yoshihiro Kawahara', 'Shiqi Gao', 'Yoshimasa Tsuruoka', 'Fuminori Okuya'] | 2019-06-05 | null | null | null | null | ['game-of-go'] | ['playing-games'] | [-2.65364856e-01 5.94709106e-02 -3.09696853e-01 1.31232783e-01
-7.61736512e-01 -4.85047042e-01 2.61505663e-01 -1.57744259e-01
-7.59663880e-01 6.24717772e-01 3.44591528e-01 -4.27100450e-01
-1.34497449e-01 -1.19705355e+00 -9.38181996e-01 -3.16498011e-01
-6.85638115e-02 5.74113727e-01 6.42920375e-01 -9.06902254e-01
3.20401251e-01 -1.81295469e-01 -1.71115577e+00 5.22777498e-01
7.23369002e-01 9.09746647e-01 2.96525359e-01 7.63033450e-01
2.51605988e-01 1.53167593e+00 -9.28737640e-01 -1.02299476e+00
4.09315914e-01 -2.03476116e-01 -9.67236698e-01 -5.56594253e-01
1.56906977e-01 -9.37891066e-01 -6.85252190e-01 7.01974511e-01
6.98785722e-01 9.65180155e-03 1.79596364e-01 -8.86283278e-01
-4.38225359e-01 1.35250032e+00 -4.89415318e-01 1.74439013e-01
1.73979104e-01 3.79149616e-01 1.39918518e+00 -2.85656989e-01
4.83134717e-01 4.22388822e-01 1.03945279e+00 3.68568927e-01
-7.74328828e-01 -9.31056976e-01 -2.10377321e-01 4.30764258e-01
-1.47064936e+00 -1.49761394e-01 4.51280594e-01 -4.38561738e-01
9.88390863e-01 9.62176323e-02 9.43499506e-01 8.56999576e-01
2.01668829e-01 9.94558930e-01 7.84077406e-01 -3.00466865e-01
1.69576317e-01 -4.60582435e-01 -1.18465737e-01 7.66503572e-01
2.21043244e-01 3.47922117e-01 -7.30552018e-01 1.27614826e-01
8.13492537e-01 -3.86249989e-01 1.39749244e-01 -9.23366472e-02
-8.26688051e-01 9.28562641e-01 4.75825697e-01 4.54405934e-01
-2.34804511e-01 4.94137138e-01 5.33393562e-01 3.26837122e-01
2.50627309e-01 6.11420035e-01 -3.49999845e-01 -1.01877403e+00
-1.01149607e+00 6.68943644e-01 8.77358437e-01 6.53832853e-01
4.26609695e-01 1.38950601e-01 -2.04825252e-02 7.85077155e-01
-2.19978273e-01 2.03203291e-01 2.80761689e-01 -1.18943775e+00
6.52765870e-01 6.56188548e-01 8.67750347e-02 -1.02845633e+00
-5.44667363e-01 -4.66073364e-01 -5.70230544e-01 4.27673012e-01
8.85641456e-01 -2.36693159e-01 -7.01478362e-01 1.72038269e+00
-2.93436080e-01 1.21407673e-01 -2.62665600e-01 9.22260046e-01
1.00628495e+00 3.90030503e-01 -1.30881801e-01 4.23791438e-01
1.08808661e+00 -7.14739561e-01 -4.40808773e-01 -4.43597525e-01
1.06156433e+00 -2.74595112e-01 1.35359311e+00 8.01648438e-01
-1.18387890e+00 -5.01557946e-01 -1.41885126e+00 3.41532528e-02
-2.00620860e-01 7.24909082e-02 1.22038317e+00 1.03111446e+00
-1.23057199e+00 7.42047071e-01 -8.84879410e-01 -4.27563824e-02
5.03799021e-01 5.58759332e-01 -3.62504691e-01 -4.95240055e-02
-1.47228193e+00 1.12524211e+00 6.34370625e-01 8.60217139e-02
-9.67072070e-01 -6.20006561e-01 -7.42662311e-01 1.71397597e-01
8.69003236e-01 -2.45348468e-01 1.46045923e+00 -5.78930616e-01
-1.69510376e+00 1.07386410e+00 5.65362275e-01 -4.09867495e-01
4.85596806e-01 -2.96030730e-01 -4.95159738e-02 -4.69665855e-01
7.83590898e-02 1.96806118e-01 -1.77703440e-01 -8.47677648e-01
-7.74833202e-01 -1.73527822e-01 8.55424404e-01 3.69099230e-01
-2.12012812e-01 1.41012236e-01 -7.03774571e-01 -3.11353147e-01
-1.10975072e-01 -8.48710954e-01 -3.22424382e-01 -6.12027943e-01
-1.02515131e-01 -2.77174890e-01 -2.43511707e-01 -9.44740653e-01
1.95960736e+00 -2.00224710e+00 5.66308349e-02 7.15868622e-02
3.31695795e-01 1.77185118e-01 -2.91624712e-03 4.10017252e-01
1.44303158e-01 3.65384221e-01 -4.33718413e-02 -3.04035214e-03
1.14364147e-01 2.30121031e-01 2.21060619e-01 2.58652717e-01
-5.64574413e-02 1.01133764e+00 -9.21841979e-01 -1.12311423e-01
-1.30513422e-02 -4.48610753e-01 -1.07892656e+00 -3.75418887e-02
-8.46633613e-02 -1.77998498e-01 -1.69000253e-01 4.91379231e-01
4.40557361e-01 5.64425401e-02 3.82335544e-01 3.28606606e-01
-1.94765285e-01 5.48874795e-01 -1.20984912e+00 1.84607017e+00
-3.21751148e-01 5.91823757e-01 -1.98415339e-01 -8.34787667e-01
7.40237534e-01 2.51804981e-02 3.22193801e-01 -1.01676238e+00
2.50148445e-01 2.54015476e-01 6.83261216e-01 -3.49101454e-01
9.64572787e-01 -3.22522372e-02 -7.97632277e-01 4.01527584e-01
5.40353358e-02 -3.78374666e-01 4.36780006e-01 3.18004042e-01
1.67216718e+00 2.34883815e-01 2.92373657e-01 -3.29193771e-01
4.60685380e-02 2.50338286e-01 7.47348666e-01 1.26325953e+00
-2.47050658e-01 6.83187187e-01 1.04427052e+00 -6.13694727e-01
-1.13996363e+00 -8.46448123e-01 2.86850423e-01 1.49662948e+00
1.96404591e-01 -8.06491077e-01 -8.09898078e-01 -4.78676975e-01
-2.51322448e-01 5.57270527e-01 -7.47599483e-01 -3.70768785e-01
-7.57970095e-01 -5.82231820e-01 1.06105399e+00 6.32837355e-01
8.38199854e-01 -1.16224039e+00 -3.88890535e-01 3.31896901e-01
-6.47212565e-01 -7.31211782e-01 -1.09982520e-01 2.46926665e-01
-3.23440373e-01 -1.17547393e+00 -1.80565044e-01 -5.14779389e-01
-3.08002770e-01 -7.82148317e-02 1.66631806e+00 5.56905866e-01
2.61158973e-01 -3.50080609e-01 -4.29495931e-01 -5.04405499e-01
-3.71429652e-01 4.99072880e-01 -3.83508205e-02 -8.09620142e-01
3.99909556e-01 -4.23122227e-01 -3.36127341e-01 3.15468192e-01
-7.66170263e-01 9.58874077e-02 5.35250783e-01 1.07952976e+00
1.21661322e-02 4.79322910e-01 1.88157931e-01 -9.31962192e-01
5.37167430e-01 -4.64901298e-01 -4.92844224e-01 -1.47866145e-01
-2.64520288e-01 -1.48527816e-01 4.75731730e-01 -2.77530581e-01
-8.59594285e-01 -1.95475921e-01 -5.15667558e-01 2.15228617e-01
1.39149159e-01 8.70470703e-01 -2.67443895e-01 -2.28486601e-02
9.66371953e-01 -9.28219706e-02 -3.18251133e-01 -2.10538521e-01
1.14041932e-01 6.12110615e-01 6.48900330e-01 -8.42877567e-01
6.62898839e-01 4.45669293e-02 -4.47185099e-01 -3.32987368e-01
-6.98365927e-01 -1.32003918e-01 -5.14090478e-01 -4.41500932e-01
6.53457284e-01 -9.31318939e-01 -1.25252390e+00 9.48207200e-01
-8.65028441e-01 -8.94352913e-01 -1.08378492e-01 4.03353423e-01
-6.87086105e-01 -1.82759035e-02 -1.02631211e+00 -6.29781544e-01
1.48214877e-01 -1.13160539e+00 6.06377840e-01 1.62565336e-01
-2.44824365e-01 -6.17349565e-01 1.31509319e-01 8.07272315e-01
3.54825735e-01 2.17867464e-01 7.38850236e-01 -6.09018683e-01
-5.30110538e-01 -2.87278295e-01 9.86319128e-03 2.08257347e-01
-4.54480678e-01 -1.77915156e-01 -7.25242019e-01 1.03822127e-01
-1.27298668e-01 -6.30269229e-01 7.44733393e-01 4.23343837e-01
9.38137233e-01 1.57620698e-01 3.97263877e-02 4.16350335e-01
1.40224755e+00 3.54280531e-01 1.19505978e+00 9.95848060e-01
5.84478319e-01 3.32482278e-01 6.42658770e-01 5.99440396e-01
7.76523232e-01 9.38441217e-01 6.11562669e-01 1.05152227e-01
5.66363372e-02 -4.35121864e-01 4.83955085e-01 7.94111609e-01
-6.53586268e-01 -3.26221138e-01 -1.24400413e+00 5.15725493e-01
-1.83046091e+00 -1.19586718e+00 -1.00407362e-01 2.09998298e+00
8.26018512e-01 8.10535252e-01 3.56381834e-01 2.63770729e-01
4.76352036e-01 3.08002502e-01 -2.08254501e-01 -6.32593036e-01
-1.98101223e-01 5.46741724e-01 7.60252833e-01 3.22624952e-01
-1.11794829e+00 1.32269704e+00 6.97181511e+00 1.36648870e+00
-5.37630260e-01 1.03833981e-01 5.83090186e-01 -2.83838779e-01
-1.91115644e-02 -3.45520675e-02 -4.28952575e-01 4.78609085e-01
1.08336997e+00 -1.13699637e-01 7.64721751e-01 7.46236384e-01
-8.34468305e-02 -6.16997063e-01 -7.42201865e-01 8.20252538e-01
5.98045625e-02 -1.48180175e+00 -5.50776899e-01 1.84592485e-01
6.73321009e-01 2.33074889e-01 -4.04725447e-02 1.00613928e+00
1.21874797e+00 -1.14933205e+00 1.08018994e+00 2.65581965e-01
6.95073843e-01 -1.09286618e+00 1.19293201e+00 6.80328131e-01
-1.06096649e+00 -3.79951328e-01 -2.81826168e-01 -9.94528830e-01
-1.73822403e-01 2.54126430e-01 -4.72019911e-01 5.28493583e-01
8.84736061e-01 4.50244814e-01 -5.46072125e-01 9.48324442e-01
-3.08137506e-01 1.01191843e+00 -4.89117533e-01 -1.72193736e-01
3.61337274e-01 8.29848573e-02 1.09058619e-01 6.47714019e-01
1.42127782e-01 4.17490810e-01 6.01161979e-02 5.78774333e-01
-3.78878683e-01 -1.49477810e-01 -7.30495453e-01 4.16133106e-02
3.64274621e-01 8.32774401e-01 -6.69554889e-01 -1.58105660e-02
-3.84192735e-01 5.52807450e-01 5.65394223e-01 -2.39372641e-01
-1.06251228e+00 -2.52441943e-01 5.87441623e-01 9.71425325e-02
2.44210824e-01 -1.91087186e-01 -6.48810327e-01 -8.58206570e-01
1.47002593e-01 -1.10110319e+00 3.57882649e-01 -5.20545363e-01
-7.71326303e-01 4.06773269e-01 -1.12193257e-01 -1.28005433e+00
-3.41467083e-01 -6.82987213e-01 -7.04505742e-01 6.35962903e-01
-7.32122958e-01 -9.50663447e-01 -9.07056779e-02 2.42673680e-01
2.30423898e-01 -3.03781748e-01 5.87301433e-01 4.29656565e-01
-5.13992131e-01 9.17647064e-01 8.83696526e-02 5.61831474e-01
2.45395511e-01 -1.18418515e+00 5.71545660e-01 8.01795721e-01
-4.01537232e-02 2.36999378e-01 5.46812534e-01 -6.84899211e-01
-1.16703439e+00 -2.89346546e-01 5.70473850e-01 -7.95395374e-01
8.76788914e-01 -5.20434082e-01 -6.65731549e-01 5.97591281e-01
-7.14233005e-03 -5.69785058e-01 6.57876849e-01 6.09066963e-01
-1.22709826e-01 4.44057323e-02 -8.38976383e-01 7.88685977e-01
1.43039536e+00 -4.15204525e-01 -3.02462667e-01 -3.70217599e-02
5.36433458e-01 -9.05043185e-01 -7.43359327e-01 3.59358579e-01
8.02256107e-01 -1.45126915e+00 7.72623062e-01 -6.47865534e-01
1.05624139e+00 -2.28427090e-02 -1.41690537e-01 -1.33818257e+00
-7.19623327e-01 -2.17625707e-01 2.87436426e-01 1.09133005e+00
4.89922374e-01 -3.07805873e-02 1.56262016e+00 7.48286843e-01
-4.44542229e-01 -7.61011720e-01 -9.30516601e-01 -6.66782916e-01
3.96873683e-01 -9.86697197e-01 7.05045938e-01 8.09089601e-01
3.62816870e-01 7.22383037e-02 -8.99021029e-01 -1.35687932e-01
2.37991586e-01 -2.64155686e-01 1.06066215e+00 -1.05809200e+00
-7.06561983e-01 -7.91270733e-01 -7.04207122e-01 -1.00289619e+00
3.48134153e-02 -8.98442090e-01 2.35431388e-01 -1.45241380e+00
3.49139512e-01 -4.63799149e-01 -2.41741225e-01 5.56172550e-01
-1.78072974e-01 4.23149914e-01 3.44160378e-01 -5.73509187e-02
-6.60286725e-01 2.21483335e-01 1.21834528e+00 -1.74990162e-01
-4.56074812e-02 1.63885415e-01 -1.10417140e+00 7.54211307e-01
7.79148877e-01 -3.98984522e-01 -1.97393242e-02 -6.41341567e-01
6.98351681e-01 2.86357343e-01 -1.28326252e-01 -1.22343659e+00
2.62082100e-01 -2.93833077e-01 8.01930353e-02 -3.33044797e-01
1.69274524e-01 -5.47232866e-01 3.50152463e-01 6.72500134e-01
-1.49631694e-01 -3.10095511e-02 3.68720978e-01 8.48217085e-02
-1.46261707e-01 -4.54684734e-01 2.77048528e-01 -1.96040884e-01
-1.20032024e+00 2.85372306e-02 -6.98412299e-01 1.28977835e-01
7.23239779e-01 -4.32432324e-01 -4.52564567e-01 -6.44118071e-01
-4.41810012e-01 2.73189843e-01 4.70355064e-01 2.73573726e-01
2.74544060e-01 -1.35793149e+00 -8.61693382e-01 -1.08699717e-01
1.06420785e-01 -1.16314195e-01 4.58474696e-01 5.05996525e-01
-1.06250870e+00 -1.15211569e-01 -4.56444353e-01 -1.54675171e-01
-8.06636810e-01 7.15325251e-02 3.49629283e-01 -9.42548633e-01
-3.88330758e-01 8.18531692e-01 -1.15315430e-01 -6.74363375e-01
4.78872769e-02 -8.33673105e-02 -3.97113055e-01 -2.21031323e-01
4.52566892e-01 4.18671489e-01 3.51152211e-01 -4.05344009e-01
-2.36169741e-01 5.89717403e-02 -4.78593335e-02 -7.57366419e-02
1.48429275e+00 4.60066259e-01 4.53222543e-02 4.17782933e-01
5.58716357e-01 5.12370728e-02 -1.07414520e+00 -1.37986466e-01
1.78285733e-01 -5.77150762e-01 4.76036780e-02 -9.13359106e-01
-1.23072839e+00 5.70036173e-01 3.18679094e-01 1.46122783e-01
9.56873953e-01 -1.01876855e-01 6.90965354e-01 4.26554263e-01
9.38659430e-01 -1.29450381e+00 1.03308499e-01 1.18017805e+00
5.59322596e-01 -1.27805579e+00 -8.26840028e-02 3.16109462e-03
-7.37768471e-01 6.82525635e-01 9.98675585e-01 -2.64915645e-01
3.27257007e-01 7.49464393e-01 -2.13325799e-01 -1.83621004e-01
-8.27915549e-01 -4.56968188e-01 -1.65427089e-01 7.38330424e-01
3.73340219e-01 2.87772328e-01 -4.67436969e-01 1.30762064e+00
-8.39445174e-01 3.20948005e-01 9.10279930e-01 9.52635169e-01
-4.79287654e-01 -8.69767666e-01 -1.85415924e-01 8.14172149e-01
-6.97420001e-01 -3.07225257e-01 -2.61113822e-01 1.02217078e+00
3.61425072e-01 9.99707758e-01 1.06244415e-01 -1.27464688e+00
7.79739559e-01 -3.34803402e-01 4.00069416e-01 -6.26298130e-01
-1.08977413e+00 -4.26914990e-01 5.86822927e-01 -6.78508937e-01
1.38523141e-02 -4.08460855e-01 -1.01267803e+00 -1.17331529e+00
-5.79481602e-01 8.11451674e-02 2.23234802e-01 9.11157787e-01
-2.49638379e-01 5.50630152e-01 2.31378287e-01 -8.70512486e-01
-2.93540210e-01 -1.02201128e+00 -9.42522824e-01 3.82964045e-01
-4.77994204e-01 -9.00846660e-01 -9.46601406e-02 -5.21844566e-01] | [3.471459150314331, 1.4274100065231323] |
1bbd09bd-1bca-4b3e-affd-f07425bfe4fc | contrastive-learning-of-sentence-embeddings | 2305.15077 | null | https://arxiv.org/abs/2305.15077v1 | https://arxiv.org/pdf/2305.15077v1.pdf | Contrastive Learning of Sentence Embeddings from Scratch | Contrastive learning has been the dominant approach to train state-of-the-art sentence embeddings. Previous studies have typically learned sentence embeddings either through the use of human-annotated natural language inference (NLI) data or via large-scale unlabeled sentences in an unsupervised manner. However, even in the case of unlabeled data, their acquisition presents challenges in certain domains due to various reasons. To address these issues, we present SynCSE, a contrastive learning framework that trains sentence embeddings with synthesized data. Specifically, we explore utilizing large language models to synthesize the required data samples for contrastive learning, including (1) producing positive and negative annotations given unlabeled sentences (SynCSE-partial), and (2) generating sentences along with their corresponding annotations from scratch (SynCSE-scratch). Experimental results on sentence similarity and reranking tasks indicate that both SynCSE-partial and SynCSE-scratch greatly outperform unsupervised baselines, and SynCSE-partial even achieves comparable performance to the supervised models in most settings. | ['Junxian He', 'Zhenzhong Lan', 'Junlei Zhang'] | 2023-05-24 | null | null | null | null | ['sentence-embeddings', 'sentence-embeddings'] | ['methodology', 'natural-language-processing'] | [ 5.93507111e-01 1.60372257e-01 -2.40277514e-01 -6.76954508e-01
-9.69254494e-01 -4.80740726e-01 9.57558453e-01 5.34063816e-01
-8.82203043e-01 8.36585283e-01 6.55790150e-01 -2.65612572e-01
3.57278287e-01 -6.09829009e-01 -6.36259854e-01 -2.99152732e-01
3.49858642e-01 5.68275034e-01 5.19189537e-02 -3.98756623e-01
2.58605659e-01 3.05459984e-02 -1.36907673e+00 2.40239069e-01
9.48450804e-01 5.36331654e-01 2.35366315e-01 7.06562340e-01
-6.36718333e-01 6.29986823e-01 -5.92182994e-01 -5.81463635e-01
1.72692016e-02 -3.61247927e-01 -7.85038173e-01 -4.51824367e-02
5.34456849e-01 -2.87260979e-01 -2.96984822e-01 9.82830167e-01
5.18623948e-01 2.74758041e-01 6.96226120e-01 -1.31864309e+00
-1.21843731e+00 6.82133377e-01 -2.38215193e-01 3.34999770e-01
4.61625665e-01 9.83711183e-02 1.39979911e+00 -1.55592847e+00
6.19835019e-01 1.26493478e+00 4.80342329e-01 8.49562943e-01
-1.36195552e+00 -4.00013268e-01 8.63874182e-02 -1.26640365e-01
-1.25281906e+00 -5.27969241e-01 7.67032564e-01 -3.67220014e-01
1.20617521e+00 8.81899595e-02 7.65073821e-02 1.52202296e+00
-6.67150393e-02 1.05862212e+00 9.07174706e-01 -6.99754179e-01
3.15566450e-01 3.61839324e-01 4.66834188e-01 4.25688118e-01
2.62949318e-01 -1.20361000e-01 -4.42856878e-01 -1.92384049e-01
3.39973420e-01 1.65302619e-01 -1.60620973e-01 -7.56285340e-02
-1.37178278e+00 1.05467713e+00 2.62832820e-01 4.06932175e-01
-2.34503701e-01 -2.23356992e-01 8.42739880e-01 4.25288796e-01
8.27784240e-01 8.56627762e-01 -4.82506841e-01 -9.03826728e-02
-9.45567906e-01 2.07427204e-01 6.23910546e-01 1.08132946e+00
7.44322300e-01 -3.49727389e-03 -4.03521031e-01 1.02662647e+00
1.55940995e-01 3.34153235e-01 8.37602496e-01 -4.20549452e-01
9.71455097e-01 6.14842355e-01 1.66781113e-01 -6.06678307e-01
1.12498859e-02 -1.42027378e-01 -7.73367941e-01 -2.87402630e-01
2.47399881e-02 -2.48537689e-01 -7.28699803e-01 1.72762465e+00
-3.61752659e-02 1.86911061e-01 5.01297474e-01 6.54695749e-01
1.08331752e+00 7.88956761e-01 2.18536377e-01 4.29569259e-02
1.10219657e+00 -1.34835529e+00 -7.14493155e-01 -6.25838220e-01
9.42626536e-01 -5.39664686e-01 1.50124323e+00 6.66960282e-03
-8.21076512e-01 -7.64189065e-01 -1.10790694e+00 -2.95607448e-01
-5.25247812e-01 2.97515929e-01 2.78374463e-01 3.68735790e-01
-1.04171228e+00 3.98919225e-01 -6.44267857e-01 -4.28240865e-01
3.02692115e-01 -2.86788102e-02 -5.32623708e-01 -3.43933582e-01
-1.48633194e+00 1.01417267e+00 5.98965883e-01 3.81370075e-03
-6.80125237e-01 -8.23246360e-01 -1.49614811e+00 2.57080793e-01
2.28137910e-01 -4.09577072e-01 1.20934808e+00 -6.94242418e-01
-1.25013971e+00 9.20292020e-01 -3.28051537e-01 -5.84245443e-01
2.15846837e-01 -3.81641805e-01 -4.82922822e-01 5.13171367e-02
3.56741726e-01 8.26759517e-01 5.51205873e-01 -1.13417852e+00
-2.28105411e-01 -3.56799923e-02 1.88929081e-01 1.61085591e-01
-9.57774401e-01 4.03992422e-02 -1.12276614e-01 -7.14374661e-01
-4.22255844e-01 -6.15522265e-01 -3.84322107e-01 4.03332636e-02
-4.42136347e-01 -8.82168651e-01 7.42791533e-01 -4.44513768e-01
1.21140289e+00 -2.17709851e+00 -5.88430166e-02 -3.43733281e-01
1.21970162e-01 7.32533514e-01 -7.05802739e-01 8.05890858e-01
-1.11890554e-01 3.13364178e-01 -3.41404855e-01 -8.08542967e-01
2.18445912e-01 4.12687331e-01 -5.72711766e-01 -1.02863260e-01
8.74563038e-01 9.34756398e-01 -1.54559004e+00 -7.06474423e-01
2.59875774e-01 1.28670007e-01 -5.43250799e-01 4.83948410e-01
-1.38124347e-01 5.31488024e-02 -3.88318032e-01 2.38437638e-01
3.29110533e-01 -3.18950266e-01 2.37386435e-01 -7.43718520e-02
1.95216060e-01 5.87257981e-01 -8.49534154e-01 1.70591879e+00
-8.43364716e-01 6.76397085e-01 -2.54233569e-01 -1.19962597e+00
1.06535661e+00 4.61320609e-01 8.69600475e-02 -4.91869390e-01
-7.30590969e-02 2.43675128e-01 -2.87410915e-01 -6.51403189e-01
8.53088319e-01 -2.36881703e-01 -3.24141294e-01 7.06220448e-01
4.24233019e-01 -3.24902058e-01 6.19117498e-01 5.63231528e-01
1.14813173e+00 -8.27533230e-02 3.63536030e-01 -1.00082979e-01
7.86966920e-01 -1.29893139e-01 4.82828885e-01 7.99613059e-01
-2.32699648e-01 6.95208073e-01 2.71498173e-01 -2.28082955e-01
-1.15051937e+00 -1.27732360e+00 -1.59205273e-02 1.04496443e+00
-3.91110443e-02 -6.34318590e-01 -5.38898945e-01 -1.08927417e+00
-1.03153072e-01 1.08272409e+00 -5.34444928e-01 -2.29764417e-01
-4.49926138e-01 -4.47162926e-01 4.67248887e-01 8.82740259e-01
1.42514512e-01 -1.20021284e+00 -7.01714903e-02 2.98027694e-01
-1.19116798e-01 -1.24165869e+00 -7.78923273e-01 1.53191954e-01
-6.29126251e-01 -7.79435456e-01 -4.94800955e-01 -1.06464851e+00
1.09216964e+00 3.47345442e-01 1.33721030e+00 1.24738991e-01
-2.87805140e-01 2.78999835e-01 -6.26857460e-01 -2.94632465e-01
-5.70351362e-01 7.86130130e-02 3.77882957e-01 -9.18786675e-02
7.43366957e-01 -3.17779034e-01 -2.18180791e-01 -2.05055475e-01
-1.26326907e+00 -2.83658057e-01 5.18868804e-01 1.26466608e+00
3.34186077e-01 -3.78015697e-01 1.04259384e+00 -1.23021078e+00
1.19858718e+00 -4.75330889e-01 -2.12441608e-01 5.87556422e-01
-5.91568232e-01 2.46822298e-01 9.53606367e-01 -4.38222319e-01
-9.08947706e-01 -2.65790373e-01 -3.96117091e-01 -3.76821697e-01
-1.75593123e-01 5.77834785e-01 7.06019029e-02 5.56096554e-01
7.08500028e-01 3.92149240e-01 -9.14704502e-02 -3.89255285e-01
5.50761163e-01 1.17440903e+00 4.52407777e-01 -5.92785358e-01
9.19545949e-01 -5.81791326e-02 -6.93724692e-01 -1.01515174e+00
-1.35759485e+00 -4.85996991e-01 -7.95049846e-01 1.75620526e-01
8.97411823e-01 -1.00795364e+00 1.16752125e-01 1.76667627e-02
-1.39769471e+00 -7.50851557e-02 -5.42899787e-01 5.12102544e-01
-1.75643787e-02 5.47573805e-01 -5.43740571e-01 -7.55488753e-01
-5.52454293e-01 -9.84352887e-01 1.17764986e+00 1.32765040e-01
-5.71935654e-01 -1.32958150e+00 3.10649127e-01 3.78672332e-01
3.88992012e-01 -1.32467628e-01 8.72549355e-01 -1.31103396e+00
3.12720723e-02 -4.38325435e-01 -1.84048355e-01 9.38040435e-01
4.55457062e-01 -1.15155168e-02 -9.35914874e-01 -3.87622446e-01
-2.81861246e-01 -9.62290108e-01 5.89470983e-01 -2.61544973e-01
1.12835729e+00 -2.77289659e-01 -1.27332613e-01 -9.34811234e-02
1.26344311e+00 -2.84773827e-01 3.06900203e-01 5.76605089e-02
4.34679598e-01 4.78685051e-01 6.82656825e-01 3.17881912e-01
3.21388245e-01 2.56064206e-01 -6.21012114e-02 -1.03884891e-01
-7.79571086e-02 -6.36519253e-01 3.91413391e-01 1.59591353e+00
6.73895240e-01 -2.74800330e-01 -6.98117971e-01 9.13915098e-01
-1.59192026e+00 -8.50830495e-01 9.55510587e-02 2.04613471e+00
1.25954950e+00 4.39609170e-01 -3.70863378e-01 1.37769237e-01
7.10079253e-01 4.19366091e-01 -3.73775631e-01 -5.90602458e-01
-1.03495242e-02 5.33495426e-01 7.90294781e-02 5.94385862e-01
-1.02026236e+00 8.97050917e-01 6.12040138e+00 6.74693167e-01
-9.19709623e-01 2.38830671e-01 3.56657475e-01 7.61681944e-02
-7.06854820e-01 5.39944954e-02 -9.50703979e-01 6.46197975e-01
9.10865724e-01 -2.61748403e-01 9.84221697e-03 9.79185104e-01
2.30268780e-02 2.19683945e-01 -1.40963411e+00 7.51973212e-01
4.96036619e-01 -1.40865707e+00 2.61719286e-01 -5.21706045e-01
9.36239839e-01 9.98659730e-02 -3.32278937e-01 8.94373715e-01
4.09343123e-01 -8.00470531e-01 3.74186158e-01 2.76102334e-01
8.75526011e-01 -2.70155102e-01 1.07712090e+00 4.74338055e-01
-9.49870527e-01 2.43662164e-01 -4.71456558e-01 -1.75896421e-01
3.40131104e-01 5.92076659e-01 -1.04577518e+00 5.87948680e-01
2.52532601e-01 9.58559394e-01 -7.91255713e-01 5.48299074e-01
-6.62528813e-01 7.91257441e-01 -8.06657597e-02 -5.54953098e-01
4.18798089e-01 -1.45005450e-01 1.08504474e-01 1.38281977e+00
3.10509279e-02 -3.21117371e-01 4.57091063e-01 8.75340044e-01
-5.11453152e-01 1.11950815e-01 -7.51477718e-01 -5.69234848e-01
8.44483316e-01 1.20516622e+00 -2.47166961e-01 -7.36214459e-01
-7.75651157e-01 9.97133493e-01 6.94544077e-01 3.01644266e-01
-4.42086309e-01 -8.66031468e-01 4.69444156e-01 -2.29686558e-01
2.61781365e-01 -1.47622108e-01 -8.29567388e-02 -1.37533724e+00
1.85907304e-01 -5.53150952e-01 1.05764344e-01 -6.71663344e-01
-2.01085830e+00 7.98757553e-01 -1.49343640e-01 -1.40189874e+00
-4.37471211e-01 -5.86116791e-01 -9.71635878e-01 7.95243800e-01
-1.79607606e+00 -8.58255804e-01 -8.36508051e-02 2.29042172e-01
1.01729023e+00 -2.74640232e-01 1.19919145e+00 2.86392272e-01
-6.59412205e-01 7.50129879e-01 2.81898052e-01 5.24760962e-01
8.38259101e-01 -1.34814322e+00 7.10642040e-01 8.60857069e-01
3.97175461e-01 8.45616400e-01 5.42847514e-01 -3.42657089e-01
-1.04359186e+00 -1.14700127e+00 1.63219428e+00 -4.97871339e-01
9.10362840e-01 -7.94134617e-01 -9.97378647e-01 5.78349948e-01
5.61658978e-01 3.52176458e-01 1.05496061e+00 8.49857032e-02
-4.29425091e-01 -2.11469363e-02 -1.03584397e+00 7.11521387e-01
8.00315440e-01 -9.48510468e-01 -1.30151832e+00 5.08423984e-01
1.08727670e+00 3.63460206e-03 -7.96665549e-01 2.09375337e-01
-4.61398922e-02 -3.93593580e-01 7.47581184e-01 -9.60331500e-01
8.39142144e-01 -1.32727027e-01 -6.30874187e-02 -1.50136936e+00
1.13639183e-01 -3.68168026e-01 1.16471998e-01 1.55642974e+00
6.74088776e-01 -6.62228107e-01 5.19616485e-01 4.82804209e-01
-2.46257797e-01 -7.85129130e-01 -6.86856329e-01 -9.49611127e-01
1.97571918e-01 -1.76420823e-01 4.55648005e-01 1.03994536e+00
1.80211246e-01 7.72193849e-01 -1.29455626e-01 -9.56227109e-02
3.50915909e-01 3.38124149e-02 7.53275633e-01 -1.07405090e+00
-6.02260605e-02 1.68225635e-02 -2.05712304e-01 -1.20269823e+00
6.93915904e-01 -1.23687994e+00 4.86073166e-01 -1.81446815e+00
1.55985102e-01 -2.49426663e-01 -2.77027458e-01 4.05734748e-01
-6.75253272e-01 9.54258814e-02 2.91749537e-02 -1.10618517e-01
-6.87595963e-01 9.09770012e-01 9.66873229e-01 -4.15391058e-01
1.70081750e-01 -3.28259438e-01 -6.07733428e-01 4.44163173e-01
6.36896670e-01 -5.09947240e-01 -6.63357794e-01 -6.19828165e-01
7.72429928e-02 -2.40085244e-01 3.48937921e-02 -6.24571383e-01
1.23438977e-01 -1.43279936e-02 -5.59616415e-03 -4.82704848e-01
2.54844904e-01 -4.15773451e-01 -9.25783873e-01 2.33764902e-01
-9.85249102e-01 1.70256466e-01 -5.18020168e-02 6.45703852e-01
-6.48893654e-01 -8.03231537e-01 4.25545961e-01 3.19739990e-03
-6.77068055e-01 7.48136267e-02 -4.65626985e-01 7.26460814e-01
7.34911442e-01 1.31585106e-01 -3.79565388e-01 -1.49918526e-01
-4.26081151e-01 4.52876002e-01 1.72080472e-01 7.91049421e-01
9.53259349e-01 -1.53567123e+00 -8.62046182e-01 4.07347053e-01
6.54951215e-01 1.16473220e-01 -6.96861148e-02 3.28091085e-01
-2.30787128e-01 4.84548837e-01 1.81510657e-01 -3.75110149e-01
-1.06626189e+00 6.15412414e-01 -3.76049250e-01 -4.64999050e-01
-2.89770991e-01 9.80623484e-01 3.44138406e-02 -1.04270303e+00
2.03210846e-01 -3.91461194e-01 -1.74278438e-01 7.55378744e-03
5.55111349e-01 -1.07957296e-01 7.37431198e-02 -3.00652742e-01
-3.26509595e-01 1.63220651e-02 -4.64197814e-01 -1.66205633e-02
1.37010145e+00 7.38124028e-02 -8.25419202e-02 5.06269336e-01
1.53178012e+00 -1.34585962e-01 -9.12632585e-01 -7.29356349e-01
4.06208277e-01 -4.31284994e-01 -2.15578958e-01 -3.57002050e-01
-4.89014745e-01 1.11709476e+00 4.42820713e-02 2.78400909e-02
5.68902016e-01 -1.43095881e-01 1.07521749e+00 7.07332432e-01
1.94755718e-01 -1.17030895e+00 4.71785337e-01 6.16815627e-01
7.75705934e-01 -1.50825202e+00 -1.63924888e-01 -1.72845304e-01
-8.80620241e-01 1.08881783e+00 7.88758934e-01 -3.15620989e-01
5.03890872e-01 1.33002922e-01 1.06164858e-01 8.10386539e-02
-8.66221488e-01 2.83288974e-02 2.16581792e-01 3.58626425e-01
7.55913734e-01 -1.23079978e-01 -4.04626846e-01 6.54685676e-01
-3.00394278e-02 -7.24055916e-02 5.77286661e-01 1.19152439e+00
-2.26339027e-01 -1.41909230e+00 2.41902635e-01 6.66108549e-01
-1.43250540e-01 -5.14936626e-01 -4.61106509e-01 5.33954978e-01
-2.80392855e-01 1.04287291e+00 1.48227140e-01 -2.60465026e-01
3.78468454e-01 2.59930462e-01 1.71825856e-01 -1.45026505e+00
-4.94265616e-01 -5.75171947e-01 3.36322516e-01 -6.16281591e-02
-3.45961004e-01 -4.40978795e-01 -1.11871958e+00 3.76774788e-01
-4.32822466e-01 4.97084767e-01 4.30527598e-01 1.23152781e+00
3.00057650e-01 5.04762709e-01 8.49469960e-01 -7.19919920e-01
-1.12421584e+00 -1.37990594e+00 -3.72885972e-01 7.69404709e-01
3.30972314e-01 -4.01080877e-01 -6.03622377e-01 -2.94440072e-02] | [10.84504508972168, 8.569464683532715] |
c256604f-05bc-492b-be7c-efef1e6637bf | deep-visual-anomaly-detection-with-negative | 2105.11058 | null | https://arxiv.org/abs/2105.11058v1 | https://arxiv.org/pdf/2105.11058v1.pdf | Deep Visual Anomaly detection with Negative Learning | With the increase in the learning capability of deep convolution-based architectures, various applications of such models have been proposed over time. In the field of anomaly detection, improvements in deep learning opened new prospects of exploration for the researchers whom tried to automate the labor-intensive features of data collection. First, in terms of data collection, it is impossible to anticipate all the anomalies that might exist in a given environment. Second, assuming we limit the possibilities of anomalies, it will still be hard to record all these scenarios for the sake of training a model. Third, even if we manage to record a significant amount of abnormal data, it's laborious to annotate this data on pixel or even frame level. Various approaches address the problem by proposing one-class classification using generative models trained on only normal data. In such methods, only the normal data is used, which is abundantly available and doesn't require significant human input. However, these are trained with only normal data and at the test time, given abnormal data as input, may often generate normal-looking output. This happens due to the hallucination characteristic of generative models. Next, these systems are designed to not use abnormal examples during the training. In this paper, we propose anomaly detection with negative learning (ADNL), which employs the negative learning concept for the enhancement of anomaly detection by utilizing a very small number of labeled anomaly data as compared with the normal data during training. The idea is to limit the reconstruction capability of a generative model using the given a small amount of anomaly examples. This way, the network not only learns to reconstruct normal data but also encloses the normal distribution far from the possible distribution of anomalies. | ['Seung-Ik Lee', 'Muhammad Zaigham Zaheer', 'Marcella Astrid', 'Jin-ha Lee'] | 2021-05-24 | null | null | null | null | ['one-class-classification'] | ['miscellaneous'] | [ 3.44066530e-01 8.87928158e-02 3.21423352e-01 -3.07904035e-01
-7.12141246e-02 -3.78381550e-01 5.56305826e-01 2.22570449e-01
-3.51898968e-01 4.94295895e-01 -4.19816852e-01 -3.12105447e-01
7.10997060e-02 -1.15813673e+00 -5.65082848e-01 -8.40561748e-01
-7.64271468e-02 4.39104557e-01 2.33670190e-01 -3.02296519e-01
1.66684031e-01 7.01697946e-01 -1.86416745e+00 3.75607610e-01
1.00571561e+00 1.09298122e+00 2.53274858e-01 3.70306164e-01
-5.42241335e-01 6.01431906e-01 -9.72127259e-01 -5.22620343e-02
2.69257993e-01 -5.90247750e-01 -3.06866765e-01 3.70630831e-01
1.04021773e-01 -4.11484718e-01 -2.90239453e-01 1.32390952e+00
3.60861540e-01 5.61609454e-02 4.23324078e-01 -1.35050702e+00
-4.39361840e-01 1.65688723e-01 -3.86385381e-01 5.13015449e-01
2.29803339e-01 3.24361682e-01 5.20529270e-01 -6.82753682e-01
3.05925667e-01 8.94670308e-01 2.91746587e-01 7.69172370e-01
-9.56790507e-01 -5.04393518e-01 2.35037073e-01 2.79210895e-01
-1.15207899e+00 -1.96065083e-01 1.01292264e+00 -2.77270555e-01
6.47874951e-01 3.62551332e-01 9.79766309e-01 1.30781841e+00
2.08813045e-02 7.36442924e-01 7.15814531e-01 -4.23424602e-01
3.48074704e-01 2.35681579e-01 -3.27253900e-02 4.62134928e-01
3.64236414e-01 1.05304115e-01 -2.49774694e-01 -5.23051322e-02
4.65315640e-01 4.43248004e-01 -3.86609256e-01 -1.41254395e-01
-7.42446005e-01 6.54331982e-01 2.49448776e-01 5.63051403e-01
-5.16294956e-01 -3.29565406e-01 3.88666272e-01 4.29807961e-01
4.19239938e-01 5.01892805e-01 -3.05043459e-01 -4.98297736e-02
-8.09878170e-01 4.33033854e-02 5.70566773e-01 5.66074729e-01
6.19076252e-01 5.28733313e-01 1.82522103e-01 6.92815542e-01
2.07569987e-01 3.72597516e-01 9.00433421e-01 -2.19875842e-01
3.96202415e-01 1.00389290e+00 -1.66585729e-01 -1.04202652e+00
-2.06954956e-01 -6.45671725e-01 -1.13886428e+00 5.14355898e-01
5.70753992e-01 -1.32231042e-01 -1.19435441e+00 1.62557960e+00
1.72438860e-01 2.43884265e-01 2.71839350e-01 7.12749958e-01
4.20765638e-01 7.98802912e-01 -1.17722027e-01 -3.96148592e-01
7.70254612e-01 -4.35044497e-01 -7.78777242e-01 -2.76999682e-01
7.97661901e-01 -6.34452343e-01 1.16196692e+00 7.74604321e-01
-7.08204150e-01 -5.12691975e-01 -1.00717223e+00 6.59919143e-01
-6.52120113e-01 -5.12418598e-02 5.39400816e-01 5.10641098e-01
-6.77319944e-01 6.15899742e-01 -7.54063845e-01 -3.66065472e-01
4.54675049e-01 2.61388928e-01 -3.51344526e-01 -1.61103755e-01
-1.13064313e+00 7.00874329e-01 5.98478675e-01 3.81591260e-01
-7.84713089e-01 -3.02228630e-01 -6.63180590e-01 2.73808062e-01
3.39525968e-01 -8.83603618e-02 8.89707863e-01 -1.40753222e+00
-8.39700401e-01 3.66832107e-01 3.01315803e-02 -3.79491538e-01
7.38461316e-01 -1.89814478e-01 -7.84118474e-01 -8.81351829e-02
-1.23918943e-01 1.17146045e-01 9.21428502e-01 -1.31669378e+00
-4.46950793e-01 -3.37930560e-01 8.67308490e-03 -4.69864942e-02
-6.33926332e-01 -4.37923402e-01 -8.72523487e-02 -6.76436901e-01
5.97138703e-01 -6.82502031e-01 -4.71065432e-01 9.92316082e-02
-3.90076756e-01 -9.96883735e-02 1.30628550e+00 -3.57153147e-01
1.10503614e+00 -2.46976066e+00 -3.67396683e-01 6.67596519e-01
9.65492502e-02 6.21105671e-01 7.00997189e-02 2.13798255e-01
-4.22045946e-01 3.60828787e-02 -5.13270915e-01 -6.40884042e-02
-4.08451021e-01 5.60442030e-01 -5.90762675e-01 1.45845398e-01
3.77398729e-01 3.70506883e-01 -8.38079274e-01 -3.60526264e-01
2.76131451e-01 1.76778644e-01 -5.94733357e-01 5.20094275e-01
-2.17966363e-01 5.95130384e-01 -4.80441630e-01 7.60485530e-01
6.72841549e-01 2.06085555e-02 -1.38064787e-01 4.36962917e-02
5.01906760e-02 -1.52945057e-01 -1.21368062e+00 1.20607448e+00
-2.76221573e-01 4.45222557e-01 -5.38339436e-01 -1.48651934e+00
1.00938058e+00 5.97510219e-01 4.42477852e-01 -7.26239860e-01
-4.89266403e-02 3.66638869e-01 4.61104125e-01 -8.50790501e-01
-4.43126727e-03 -2.47277677e-01 3.22769761e-01 5.05768716e-01
5.27120866e-02 2.12998018e-01 1.99415818e-01 -5.25465943e-02
1.33390689e+00 -1.64565831e-01 1.76263869e-01 3.63065600e-01
6.46346986e-01 -1.54047385e-01 7.80324280e-01 5.89278340e-01
5.54881990e-02 5.76720417e-01 7.23886967e-01 -7.95172989e-01
-1.04538810e+00 -1.02179277e+00 -1.27302453e-01 5.11812031e-01
-7.08332881e-02 -9.31629092e-02 -4.13475960e-01 -9.24825609e-01
-3.99136186e-01 8.16563249e-01 -5.09556413e-01 -4.12462175e-01
-4.51727778e-01 -1.15367436e+00 4.62490052e-01 4.06611472e-01
6.43607438e-01 -1.48924851e+00 -7.19126523e-01 1.78278953e-01
3.18398811e-02 -7.38367677e-01 2.52142519e-01 4.23011661e-01
-1.07316852e+00 -1.12387884e+00 -3.34679395e-01 -4.56569552e-01
1.19977629e+00 -9.73442942e-02 8.08667839e-01 4.48158592e-01
-3.25631976e-01 -1.78557187e-02 -5.78184128e-01 -6.25770271e-01
-5.11708796e-01 -3.68869603e-01 1.59359183e-02 3.09874058e-01
4.18334872e-01 -8.47750247e-01 -2.36079961e-01 4.12098430e-02
-1.26014996e+00 -2.71230012e-01 8.20616961e-01 9.31624711e-01
4.52622950e-01 4.49025452e-01 8.01848352e-01 -9.68086779e-01
4.39510673e-01 -7.32067764e-01 -4.53125417e-01 1.28832370e-01
-5.89325607e-01 2.83768903e-02 1.06233788e+00 -8.02176297e-01
-9.89057183e-01 -1.52320890e-02 -5.25475144e-01 -6.72373056e-01
-5.63771069e-01 3.75560820e-01 -3.39212060e-01 1.85978040e-01
7.92636454e-01 6.36507273e-01 2.52393205e-02 -3.38311076e-01
-2.44113013e-01 5.50595105e-01 2.96905607e-01 -1.97451666e-01
8.42641175e-01 3.50870162e-01 1.17578119e-01 -8.87180746e-01
-7.52436399e-01 -1.52179196e-01 -4.87415522e-01 -3.48440945e-01
5.66152155e-01 -2.77883589e-01 -2.58147389e-01 5.95755160e-01
-1.05904555e+00 1.23434342e-01 -5.72410762e-01 5.08971453e-01
-2.72660345e-01 5.30435145e-01 -5.68391867e-02 -1.06090772e+00
-1.05661474e-01 -1.01876307e+00 3.94698143e-01 1.06836312e-01
-9.30595249e-02 -7.96135247e-01 -1.34222507e-01 -2.47288212e-01
4.42989856e-01 3.59371245e-01 1.13024783e+00 -1.27568662e+00
-5.93141913e-01 -7.18607545e-01 1.77885994e-01 7.93264985e-01
2.30932489e-01 -3.68775241e-02 -1.06758428e+00 -9.29782242e-02
4.10216212e-01 -8.61438513e-02 6.53599560e-01 -6.45138472e-02
1.73879313e+00 -4.50614154e-01 -8.44797045e-02 5.54510951e-01
1.21644700e+00 6.82916582e-01 8.45452189e-01 2.88017958e-01
6.34933949e-01 6.59138083e-01 5.45801222e-01 5.30246317e-01
-2.93768913e-01 2.55944520e-01 9.46359932e-01 -2.81629711e-01
3.21189284e-01 -6.60561472e-02 1.73859388e-01 3.99391323e-01
-1.64528891e-01 -4.44609344e-01 -8.82694006e-01 5.70977688e-01
-1.66699433e+00 -1.22667885e+00 -8.31840187e-02 2.44069934e+00
5.13604164e-01 3.48487705e-01 -2.58812070e-01 6.95777535e-01
6.66080356e-01 6.77219480e-02 -6.85841918e-01 -4.23571020e-01
-1.29363224e-01 1.60285011e-01 2.39734650e-02 -6.72367215e-02
-9.21137452e-01 3.67822647e-01 5.18571806e+00 6.42210722e-01
-1.44442511e+00 -1.79553643e-01 5.53151667e-01 -7.94079453e-02
-3.09389502e-01 -1.17581859e-01 -3.36271673e-01 8.55120420e-01
8.60773861e-01 1.49569690e-01 1.51459783e-01 1.02444875e+00
9.74774882e-02 -1.52194247e-01 -1.15590072e+00 9.09812570e-01
1.21072806e-01 -8.83974135e-01 2.19606861e-01 6.16049506e-02
3.25531453e-01 -2.34645277e-01 1.00455694e-02 2.98178375e-01
-3.97496909e-01 -9.92054403e-01 3.59971821e-01 6.35169148e-01
2.89650053e-01 -8.43484223e-01 1.13797987e+00 8.71393025e-01
-5.41480720e-01 -3.95856827e-01 -4.67399091e-01 -2.07676576e-03
1.34686148e-02 7.96892107e-01 -8.19701612e-01 3.23905647e-01
7.25338936e-01 3.45182747e-01 -5.30562043e-01 1.21676528e+00
-1.22888148e-01 6.74581468e-01 -4.92327571e-01 3.31055731e-01
1.80655748e-01 -1.41932696e-01 6.27646446e-01 7.79778898e-01
8.37501705e-01 8.99889916e-02 1.46111503e-01 7.89966941e-01
1.35963842e-01 9.14921761e-02 -1.10498762e+00 -9.90924761e-02
8.17793310e-02 1.02102864e+00 -6.68330729e-01 -4.03485328e-01
-4.66649383e-01 9.07187164e-01 9.17498488e-03 3.03284943e-01
-5.18108666e-01 -2.43548036e-01 2.37535521e-01 3.69827360e-01
-1.58115447e-01 1.04963958e-01 -3.21988642e-01 -1.09235144e+00
3.22991520e-01 -7.55842447e-01 3.70012730e-01 -5.43006241e-01
-1.28313291e+00 7.74842560e-01 -1.43993348e-01 -1.62452817e+00
-5.76248109e-01 -4.08782423e-01 -1.02835810e+00 8.18473637e-01
-1.28616118e+00 -8.08703840e-01 -3.78954411e-01 6.95670843e-01
6.46675825e-01 -4.45096195e-01 9.24707413e-01 4.07800853e-01
-4.56699759e-01 5.05192995e-01 -1.14516765e-01 2.48141631e-01
5.05772233e-01 -1.13249850e+00 2.18232777e-02 1.14964771e+00
2.29870945e-01 3.93524081e-01 7.71173835e-01 -5.96836030e-01
-7.27232933e-01 -1.09340394e+00 7.90966213e-01 -1.35882288e-01
4.07763362e-01 -1.54324502e-01 -1.59631979e+00 5.89834154e-01
-2.14572042e-01 4.92545158e-01 6.83917284e-01 -1.06783867e-01
1.94413308e-02 -1.69361413e-01 -1.40451419e+00 5.00362515e-01
6.89652622e-01 -1.19378030e-01 -7.13509679e-01 3.48191530e-01
2.28104249e-01 -1.86951533e-01 -2.26751551e-01 5.20570576e-01
2.04584435e-01 -1.14017141e+00 5.79951048e-01 -5.37669241e-01
2.55131155e-01 -4.65752304e-01 -7.59557188e-02 -1.37877667e+00
2.23414108e-01 -3.71195376e-02 -2.99788386e-01 1.09326458e+00
3.99361700e-01 -7.98573792e-01 7.83380687e-01 4.37322766e-01
-3.68319839e-01 -9.99123335e-01 -9.23742056e-01 -6.98963940e-01
-3.44877034e-01 -5.73460102e-01 5.99572957e-01 1.02066314e+00
-2.99615085e-01 -1.92631736e-01 -3.52566928e-01 4.72889930e-01
3.77192527e-01 -7.71740973e-02 4.84317422e-01 -1.43203557e+00
-2.66686052e-01 -2.09735334e-02 -8.33910286e-01 -5.72192490e-01
-2.63830507e-03 -6.29215240e-01 -3.21710855e-03 -1.21015859e+00
-2.15799287e-01 -5.94984055e-01 -4.32969123e-01 5.58116913e-01
1.18465144e-02 1.66575611e-01 -1.89227629e-02 2.49575123e-01
-1.61566257e-01 5.04511952e-01 1.10268176e+00 -2.78305653e-02
-2.59340912e-01 4.78288889e-01 -3.71037424e-01 1.12558210e+00
9.00089741e-01 -4.32118624e-01 -6.93283558e-01 -1.77635565e-01
1.85142994e-01 -3.92283164e-02 4.69704092e-01 -1.19764864e+00
1.44673884e-01 -1.46570295e-01 8.74377728e-01 -7.47809112e-01
2.58559048e-01 -1.14137447e+00 3.55162844e-02 4.43001032e-01
-8.92975554e-02 2.88011938e-01 -5.65355644e-02 6.06851995e-01
-4.21978265e-01 -5.48762977e-01 7.79632568e-01 -3.60387981e-01
-8.73960495e-01 2.99242020e-01 -4.92361039e-01 -2.18748480e-01
1.19502306e+00 -4.62932378e-01 -8.00709575e-02 -3.42827886e-01
-9.92901206e-01 2.86808293e-02 2.69213706e-01 3.15981686e-01
8.77663910e-01 -1.21273351e+00 -2.27459684e-01 7.85090923e-01
2.08046988e-01 3.52451622e-01 4.71396565e-01 6.36557817e-01
-3.82941276e-01 -7.76814371e-02 -3.79045427e-01 -5.84178507e-01
-9.44000900e-01 7.19232917e-01 3.25370789e-01 -1.42440408e-01
-7.66154110e-01 5.01395166e-01 2.66336143e-01 -1.69345513e-01
1.78978264e-01 -9.24339816e-02 -4.81651425e-01 1.12334691e-01
5.49846649e-01 4.13398221e-02 1.88142374e-01 -3.23523939e-01
2.12137587e-03 7.07921237e-02 -9.61096436e-02 1.89454362e-01
1.37423980e+00 2.24975795e-01 -1.07796833e-01 6.42506421e-01
8.15918505e-01 -1.55769095e-01 -9.64273334e-01 -7.06216991e-02
4.90558445e-02 -6.54105127e-01 -2.38342404e-01 -6.42175615e-01
-1.23644352e+00 1.10278785e+00 9.17029500e-01 5.78310013e-01
1.48150945e+00 -2.53949910e-01 6.13584518e-01 6.41633034e-01
1.12622537e-01 -1.04624009e+00 2.04191506e-01 3.13833773e-01
6.17161453e-01 -1.40854001e+00 -3.71067524e-01 -1.12933248e-01
-5.48710763e-01 1.27073729e+00 9.94797707e-01 -1.37271866e-01
5.87385595e-01 2.10741952e-01 7.19778836e-02 -1.84445485e-01
-6.77575707e-01 -1.84701338e-01 6.28120229e-02 7.18942881e-01
2.22852200e-01 -2.61871278e-01 -2.25511834e-01 3.12844515e-01
-1.48145914e-01 -3.43938529e-01 7.78663099e-01 8.66777778e-01
-6.42046988e-01 -1.19776714e+00 -4.76079673e-01 8.45950842e-01
-5.87807417e-01 3.23505290e-02 -4.44566719e-02 7.54121661e-01
5.73801994e-01 7.89069653e-01 3.28601837e-01 -1.79303810e-01
3.04973304e-01 3.93432796e-01 1.87744275e-02 -7.73811162e-01
-8.42741430e-02 -1.38031200e-01 -3.47800970e-01 -4.29416686e-01
-1.46062881e-01 -5.09013593e-01 -1.30556262e+00 -4.89045717e-02
-2.76750505e-01 1.35786384e-01 5.54918051e-01 1.21836567e+00
-7.71305636e-02 6.74905360e-01 6.42765462e-01 -4.69187528e-01
-4.84797329e-01 -9.51530993e-01 -5.97711205e-01 5.07786214e-01
5.54620802e-01 -6.14758551e-01 -6.37099922e-01 -6.57962784e-02] | [7.609884262084961, 2.248992443084717] |
9e7d3312-cbe1-46ed-9759-c82be6e10db2 | molecular-mechanics-driven-graph-neural | 2011.07457 | null | https://arxiv.org/abs/2011.07457v1 | https://arxiv.org/pdf/2011.07457v1.pdf | Molecular Mechanics-Driven Graph Neural Network with Multiplex Graph for Molecular Structures | The prediction of physicochemical properties from molecular structures is a crucial task for artificial intelligence aided molecular design. A growing number of Graph Neural Networks (GNNs) have been proposed to address this challenge. These models improve their expressive power by incorporating auxiliary information in molecules while inevitably increase their computational complexity. In this work, we aim to design a GNN which is both powerful and efficient for molecule structures. To achieve such goal, we propose a molecular mechanics-driven approach by first representing each molecule as a two-layer multiplex graph, where one layer contains only local connections that mainly capture the covalent interactions and another layer contains global connections that can simulate non-covalent interactions. Then for each layer, a corresponding message passing module is proposed to balance the trade-off of expression power and computational complexity. Based on these two modules, we build Multiplex Molecular Graph Neural Network (MXMNet). When validated by the QM9 dataset for small molecules and PDBBind dataset for large protein-ligand complexes, MXMNet achieves superior results to the existing state-of-the-art models under restricted resources. | ['Lei Xie', 'Yang Liu', 'Shuo Zhang'] | 2020-11-15 | null | null | null | null | ['formation-energy'] | ['miscellaneous'] | [ 3.31363112e-01 2.06536204e-01 -4.05889601e-01 -2.85581708e-01
-6.27819970e-02 -2.33868256e-01 3.41953158e-01 5.93155801e-01
-3.25844884e-01 1.19559300e+00 -2.02257425e-01 -5.52523911e-01
-1.53409004e-01 -1.12143910e+00 -9.34830606e-01 -7.13490546e-01
-2.23200485e-01 3.84292185e-01 3.58729869e-01 -3.64360780e-01
1.46518052e-01 6.70374930e-01 -1.09075272e+00 2.94473052e-01
1.17045534e+00 6.31249607e-01 1.86089873e-01 2.38848642e-01
-1.96155772e-01 9.19415593e-01 -1.07799619e-01 -4.86468881e-01
-1.09745964e-01 -5.04268169e-01 -5.29238880e-01 -4.79463905e-01
1.31407276e-01 1.82675913e-01 -3.97616565e-01 9.17140961e-01
8.20817173e-01 1.70547649e-01 7.30965257e-01 -9.07196045e-01
-4.11976427e-01 7.49779522e-01 -1.75851554e-01 -1.85167983e-01
1.47722393e-01 2.79473841e-01 1.18330109e+00 -6.57215774e-01
8.44658554e-01 1.15836084e+00 5.00429213e-01 5.52115500e-01
-1.47233212e+00 -7.90372849e-01 2.28611857e-01 3.34056348e-01
-1.28665102e+00 -2.05560908e-01 7.03752637e-01 -2.45994866e-01
1.33279800e+00 6.20879084e-02 6.39782548e-01 1.01755691e+00
5.75250328e-01 3.23084980e-01 6.41817570e-01 2.67050285e-02
4.49399203e-01 -1.70362025e-01 1.51381746e-01 8.84734571e-01
3.91676635e-01 -6.43782392e-02 -5.43157220e-01 -4.55514610e-01
5.40766060e-01 1.49600133e-01 -2.19741195e-01 -5.70916414e-01
-7.68619418e-01 8.86840403e-01 9.17438388e-01 7.61266500e-02
-4.53908622e-01 2.31237248e-01 3.14877748e-01 1.20758109e-01
1.39642790e-01 6.19381785e-01 -4.09318030e-01 4.49341625e-01
-4.96413589e-01 8.62864554e-02 9.41243052e-01 6.25578165e-01
8.47416401e-01 -1.17263205e-01 6.24673963e-02 6.18384957e-01
5.56392312e-01 -1.69720426e-01 2.10192397e-01 -2.79937554e-02
3.61015558e-01 1.13955438e+00 -2.26661652e-01 -1.06210315e+00
-9.75588441e-01 -5.67361414e-01 -1.14970016e+00 2.35637021e-03
-4.74401005e-02 -2.26678252e-02 -8.12857389e-01 1.72251523e+00
5.79340279e-01 4.26188000e-02 1.15054809e-01 6.60787165e-01
1.17274368e+00 8.32654774e-01 6.94586754e-01 -3.19343388e-01
1.14076221e+00 -9.68115866e-01 -5.43062150e-01 1.18859105e-01
9.65596497e-01 -4.03904051e-01 5.29926121e-01 2.86995620e-01
-8.63461971e-01 -3.99834663e-01 -1.36972475e+00 4.22634333e-02
-6.19497895e-01 4.45173681e-02 1.16540945e+00 4.27485287e-01
-7.14050353e-01 1.08195508e+00 -7.73000002e-01 -4.54529375e-02
4.60529298e-01 1.05317342e+00 -5.51614344e-01 8.86903778e-02
-1.43663299e+00 8.06119323e-01 7.87274659e-01 2.13771597e-01
-6.92920148e-01 -6.10066473e-01 -6.91809654e-01 1.38405487e-01
2.82085299e-01 -8.78001690e-01 5.56651294e-01 -6.89785480e-01
-1.74321091e+00 2.89888173e-01 1.67729691e-01 -3.40842336e-01
1.09601893e-01 4.04112607e-01 -3.17067355e-01 1.49061494e-02
-3.65475714e-01 6.45297647e-01 3.75260681e-01 -9.27798271e-01
-2.88446024e-02 -4.06546444e-01 2.40961716e-01 1.93012670e-01
-2.69005179e-01 -2.54530698e-01 -2.75513142e-01 -3.41703117e-01
-3.74216624e-02 -8.42390239e-01 -6.40163898e-01 -7.40901828e-02
-5.87378263e-01 -2.98270524e-01 2.62611926e-01 -1.96508393e-01
1.35967743e+00 -1.55420530e+00 7.82294571e-01 5.01356959e-01
6.71349466e-01 5.60176611e-01 -3.27306539e-01 8.17940652e-01
-1.99295074e-01 -1.72102898e-02 -2.17676356e-01 -1.81734283e-02
-1.06722079e-01 1.50036588e-01 2.90623397e-01 4.15848732e-01
3.76232535e-01 1.09198260e+00 -7.13542700e-01 -4.00479674e-01
8.76703486e-02 6.90454483e-01 -8.36262584e-01 2.50958055e-01
-8.01477134e-01 4.29578424e-01 -7.43090212e-01 5.66462398e-01
9.06353116e-01 -6.56162918e-01 9.43764031e-01 -4.32728589e-01
-1.13774529e-02 2.78731883e-01 -9.99290705e-01 1.64448524e+00
-1.02382071e-01 -2.47211054e-01 -1.11181408e-01 -1.01250851e+00
1.03073812e+00 1.90727204e-01 3.46765727e-01 -7.01332808e-01
2.55649388e-01 1.34805128e-01 4.53463107e-01 -3.10044229e-01
-1.19295688e-02 -2.33318537e-01 5.03172755e-01 1.58075675e-01
1.57042757e-01 2.72097975e-01 3.18358600e-01 1.59404650e-01
1.18117154e+00 3.15584332e-01 2.90076643e-01 -1.53466240e-01
9.67099845e-01 -1.19471408e-01 5.90159774e-01 4.00492102e-01
2.13275060e-01 1.49361454e-02 7.28108168e-01 -7.44547129e-01
-8.31271529e-01 -5.98167121e-01 7.74005353e-02 1.15650523e+00
4.72740717e-02 -8.79939795e-01 -7.00191021e-01 -6.02785170e-01
-8.84231329e-02 3.83620337e-02 -5.03998220e-01 -4.68749791e-01
-4.56329703e-01 -1.23650312e+00 3.88270050e-01 1.36171386e-01
2.22712845e-01 -1.06820607e+00 7.87983388e-02 6.62268877e-01
3.00741881e-01 -7.74492562e-01 -1.76371917e-01 6.32638633e-01
-8.38262320e-01 -1.17746460e+00 -3.13074678e-01 -6.19471729e-01
7.48025298e-01 -1.20558910e-01 9.53392565e-01 2.83877313e-01
-2.29355142e-01 -6.81342483e-01 -1.80553764e-01 -2.84181863e-01
-4.04057235e-01 5.32169104e-01 3.71642690e-03 1.10545412e-01
2.05874681e-01 -9.87756014e-01 -7.87090361e-01 1.01875864e-01
-1.15421569e+00 3.53618890e-01 9.59493697e-01 9.28137183e-01
6.86095774e-01 -2.87952542e-01 7.88783371e-01 -1.10086429e+00
6.85039043e-01 -5.97020864e-01 -7.10404098e-01 3.87399137e-01
-6.08374655e-01 5.12280881e-01 1.07965767e+00 -5.75127840e-01
-6.47641242e-01 4.48671132e-01 -3.48183751e-01 2.33025476e-02
2.14271471e-01 1.00023818e+00 -4.84818608e-01 -5.61192334e-01
6.00627720e-01 1.70078501e-01 -7.36855790e-02 -4.55992848e-01
2.97151059e-01 4.00798082e-01 -1.19107075e-01 -6.54780746e-01
4.44449157e-01 -6.03162497e-02 7.37765789e-01 -8.46343100e-01
-4.70633060e-01 -4.16308790e-02 -6.39331996e-01 1.37899831e-01
7.72989213e-01 -6.54669225e-01 -1.39520681e+00 1.99817538e-01
-1.17004669e+00 -1.37680486e-01 4.59009200e-01 5.28432310e-01
-2.81815261e-01 4.86631811e-01 -6.16783738e-01 -4.92030919e-01
-6.13697529e-01 -1.45701206e+00 6.80803776e-01 7.53375292e-02
7.51302466e-02 -8.58522177e-01 2.50258714e-01 8.70774388e-02
2.45956585e-01 4.59223002e-01 1.46156943e+00 -7.06587493e-01
-8.39896798e-01 -5.47147617e-02 -3.64403635e-01 2.11340692e-02
1.58273026e-01 -1.66812360e-01 -6.28964245e-01 -2.28880093e-01
-6.52288377e-01 -3.39011401e-01 9.12509382e-01 1.80389851e-01
9.16027844e-01 -3.79843593e-01 -5.48673093e-01 7.56753027e-01
1.44303691e+00 2.95020908e-01 7.45720446e-01 2.35505968e-01
1.08579898e+00 3.82202417e-01 2.13845782e-02 2.54329622e-01
2.42116734e-01 6.09357953e-01 7.19057381e-01 -2.66456783e-01
2.95727581e-01 -3.44207138e-01 2.37811401e-01 9.33075905e-01
-2.73871392e-01 -6.54682159e-01 -7.87827849e-01 -2.55212098e-01
-2.13432455e+00 -5.32982349e-01 -2.37526447e-01 1.99251354e+00
1.02715409e+00 2.25460842e-01 2.90407054e-02 -1.23868629e-01
3.99839461e-01 9.82878581e-02 -7.21218586e-01 -3.86043876e-01
-2.11907506e-01 4.95362014e-01 2.89759815e-01 4.98243600e-01
-8.10718775e-01 1.13291347e+00 5.74695873e+00 9.12259817e-01
-1.08742571e+00 -3.04845721e-01 4.92259532e-01 1.34484917e-01
-3.05206358e-01 1.29687846e-01 -7.89074361e-01 3.74528766e-01
1.13500297e+00 2.51419127e-01 3.24378788e-01 5.12434006e-01
1.84297487e-01 2.49872893e-01 -1.18408287e+00 6.63546264e-01
-3.45516354e-01 -1.91040266e+00 6.97266459e-01 1.13683246e-01
5.41074872e-01 -5.69887236e-02 -2.92962700e-01 6.21104389e-02
6.23735301e-02 -1.40744758e+00 4.79330635e-03 6.79745793e-01
5.54057777e-01 -9.27360356e-01 6.27582729e-01 3.11902851e-01
-1.22422802e+00 2.49986559e-01 -5.67326903e-01 -6.57821745e-02
-1.09636597e-01 3.75076532e-01 -6.37981653e-01 8.70744824e-01
1.92635395e-02 7.04318821e-01 -4.09690887e-01 8.36916268e-01
-1.35239393e-01 2.77382135e-01 -2.35503718e-01 -4.90722597e-01
3.82478744e-01 -6.95237696e-01 5.08270524e-02 1.13304329e+00
-2.35161722e-01 3.39180112e-01 3.41383278e-01 9.54379976e-01
-5.00710964e-01 5.71191788e-01 -5.00291169e-01 -3.22597861e-01
2.56542057e-01 1.34221935e+00 -5.88520885e-01 -1.51582167e-01
-4.36031818e-01 7.88626909e-01 8.10030162e-01 3.22132349e-01
-8.65310192e-01 -4.89373893e-01 6.87054038e-01 1.10080317e-01
7.13645518e-02 -2.22154945e-01 3.17753613e-01 -1.01339960e+00
-2.81581283e-01 -9.04569626e-01 1.85812995e-01 -5.02618432e-01
-1.15561795e+00 6.89999223e-01 -5.01734078e-01 -9.06785369e-01
3.67522240e-01 -9.97114182e-01 -4.27026659e-01 8.56105208e-01
-1.64329553e+00 -1.22952092e+00 -2.95114387e-02 3.84742349e-01
-1.89823024e-02 -2.91433721e-03 9.98186231e-01 4.92191195e-01
-1.01347840e+00 5.84184825e-01 1.04163550e-01 -3.79825115e-01
5.72319746e-01 -8.62387896e-01 3.05146247e-01 2.52274632e-01
-2.90745795e-01 9.68902588e-01 4.74917561e-01 -8.43556285e-01
-1.84557939e+00 -1.18646538e+00 7.10353374e-01 -2.42553368e-01
6.90759540e-01 -4.82777297e-01 -1.06944561e+00 2.94171035e-01
-2.10281789e-01 -1.98656265e-02 8.79510403e-01 -6.51244521e-02
-3.42750877e-01 5.70836887e-02 -9.06041563e-01 6.00118220e-01
1.20423508e+00 -3.05519909e-01 -8.13560858e-02 5.38657904e-01
8.85231733e-01 -3.27076674e-01 -1.17199051e+00 4.84454542e-01
4.30613190e-01 -6.52277231e-01 1.07182837e+00 -1.10288334e+00
4.88572001e-01 -5.01097143e-01 5.53430878e-02 -1.08746099e+00
-4.73528266e-01 -7.06120014e-01 -2.55849123e-01 6.42241478e-01
7.48340666e-01 -6.87954962e-01 9.26279962e-01 3.81731361e-01
-1.02541983e-01 -1.16341770e+00 -7.69482493e-01 -5.44834793e-01
-6.92999363e-02 1.19005870e-02 6.76901698e-01 8.30894172e-01
1.98810562e-01 9.91405547e-01 -5.30645847e-01 -5.69795407e-02
4.36327398e-01 -9.31324661e-02 6.92336261e-01 -1.48629320e+00
-3.39434624e-01 -3.42002541e-01 -5.46855927e-01 -1.01649380e+00
2.17909530e-01 -1.24564326e+00 -5.54430366e-01 -1.56854892e+00
4.33633119e-01 -2.46273756e-01 -5.04407406e-01 5.25248706e-01
-1.20342158e-01 -8.32148343e-02 -1.17280118e-01 2.66697761e-02
-7.35024929e-01 8.21635902e-01 1.46908140e+00 -3.25683266e-01
-4.04035330e-01 -3.02431524e-01 -5.57914495e-01 3.37422699e-01
7.04435885e-01 -4.71265435e-01 -4.07149643e-01 -1.22964084e-02
7.11839676e-01 1.06995419e-01 3.49936597e-02 -8.19968641e-01
3.59427810e-01 -3.14131588e-01 4.05010164e-01 -4.79759663e-01
3.21601331e-01 -7.21292436e-01 5.88274419e-01 7.11682439e-01
-2.46997744e-01 -8.02139044e-02 1.69437841e-01 7.49820948e-01
-5.04300371e-02 -2.55447552e-02 7.39443302e-01 -1.17966205e-01
-1.77371800e-01 8.73676956e-01 -2.70749480e-01 -6.05348825e-01
7.65507936e-01 -1.33349791e-01 -3.69276375e-01 1.12852700e-01
-7.56595016e-01 2.28518292e-01 2.78398544e-01 6.69650063e-02
5.99534750e-01 -1.15835416e+00 -3.90609920e-01 1.47853374e-01
2.57067770e-01 -1.36846498e-01 1.85069591e-01 7.23436177e-01
-7.25042045e-01 6.15587771e-01 -2.67525911e-01 -3.15288723e-01
-1.24252450e+00 7.40102649e-01 5.04875243e-01 -5.46275675e-01
-9.36766863e-02 7.03445196e-01 3.98663163e-01 -6.33549213e-01
1.39645070e-01 -1.24262556e-01 -4.46720630e-01 -5.91419621e-07
1.72018990e-01 1.76267862e-01 9.61232632e-02 -4.86951858e-01
-3.81832689e-01 4.99791175e-01 -2.48047709e-01 6.81074023e-01
1.66217947e+00 4.43460256e-01 -5.11484504e-01 -8.27073604e-02
1.23005152e+00 -4.42447275e-01 -9.63809490e-01 -2.63096333e-01
-1.78767294e-02 2.63321102e-01 -9.25903991e-02 -7.02581942e-01
-8.36868286e-01 7.72255003e-01 4.92912889e-01 -3.02803069e-01
8.44560027e-01 -2.44496807e-01 6.50997162e-01 8.61371160e-01
2.51112252e-01 -8.93007815e-01 9.56790522e-02 5.59420764e-01
6.02973342e-01 -1.09993672e+00 3.20206761e-01 -5.82085788e-01
-1.88860163e-01 1.30390608e+00 6.56466544e-01 6.97250962e-02
4.68156964e-01 -2.06355333e-01 -5.35337329e-01 -7.01974332e-01
-8.52260351e-01 -5.88051528e-02 4.21739370e-01 3.68477464e-01
7.42354155e-01 1.05076984e-01 -6.77783728e-01 6.90636992e-01
2.55288392e-01 2.77078636e-02 8.38578418e-02 8.88105392e-01
-4.90760833e-01 -1.67948318e+00 2.14499101e-01 2.98195839e-01
-3.02142739e-01 -3.91746730e-01 -8.99084806e-01 6.49865150e-01
2.00651363e-01 6.63299203e-01 -6.04896009e-01 -6.04906261e-01
3.74540865e-01 -3.92780341e-02 5.93566835e-01 -5.03652751e-01
-8.41388583e-01 1.75126607e-03 1.73576117e-01 -5.18858075e-01
-4.38973069e-01 2.15001181e-01 -1.45300758e+00 -4.61256504e-01
-6.63078606e-01 4.26496744e-01 5.70836544e-01 6.72756255e-01
6.78115010e-01 7.37383723e-01 5.92523992e-01 -8.09829950e-01
-2.57599860e-01 -8.55123162e-01 -4.99406219e-01 5.85465617e-02
9.30777416e-02 -4.85342920e-01 4.01282459e-02 -2.79935747e-01] | [5.128928184509277, 5.862573146820068] |
8a328659-082e-49fb-ab9e-9fd6a7f99688 | big-learning-a-universal-machine-learning | 2207.03899 | null | https://arxiv.org/abs/2207.03899v4 | https://arxiv.org/pdf/2207.03899v4.pdf | Big Learning | Recent advances in big/foundation models reveal a promising path for deep learning, where the roadmap steadily moves from big data to big models to (the newly-introduced) big learning. Specifically, the big learning exhaustively exploits the information inherent in its large-scale complete/incomplete training data, by simultaneously modeling many/all joint/conditional/marginal data distributions across potentially diverse domains, with one universal foundation model. We reveal that big learning ($i$) underlies most existing foundation models, ($ii$) is equipped with extraordinary flexibilities for complete/incomplete training data and trustworthy data tasks, ($iii$) is capable of delivering all joint/conditional/marginal data capabilities with one universal model, and ($iv$) unifies conventional machine learning paradigms and enables their flexible cooperations, manifested as a universal learning paradigm. Diverse experiments are carried out to validate the effectiveness of the presented big learning. | ['Miaoyun Zhao', 'Yulai Cong'] | 2022-07-08 | null | null | null | null | ['self-learning'] | ['natural-language-processing'] | [-5.42929828e-01 2.32937232e-01 -3.63107532e-01 -3.86305630e-01
-9.83625174e-01 -1.76633745e-01 7.25741804e-01 -2.66604662e-01
-1.19366691e-01 9.18985903e-01 -1.89980194e-01 -2.56462097e-01
-6.19340599e-01 -8.32045794e-01 -9.16473866e-01 -9.20840144e-01
-2.06829965e-01 7.05080688e-01 -4.42253985e-02 -1.28350064e-01
-1.95165426e-01 2.48940095e-01 -1.41841424e+00 3.06327701e-01
9.05777931e-01 1.37449539e+00 5.51381335e-02 2.08422795e-01
-1.12285025e-01 1.13034666e+00 -4.17457312e-01 -8.90815616e-01
1.02898926e-01 3.04357290e-01 -6.63431108e-01 -2.69016534e-01
3.62643063e-01 -3.51254731e-01 -2.30064556e-01 8.97912621e-01
3.10194045e-01 -3.58830988e-01 5.86783290e-01 -1.34377503e+00
-9.61487055e-01 7.93415725e-01 -2.82540351e-01 -1.63044691e-01
-1.11928530e-01 2.40698308e-01 1.16013408e+00 -9.04304445e-01
4.91678536e-01 1.13357842e+00 6.44929826e-01 5.55412292e-01
-7.32289314e-01 -7.28595972e-01 -2.71939486e-03 1.90116599e-01
-1.40794873e+00 -3.35815072e-01 7.40909576e-01 -6.55913055e-01
9.18792069e-01 -4.11086297e-03 2.04490185e-01 1.48833323e+00
3.42276096e-01 1.25472403e+00 1.10991526e+00 -9.71461385e-02
4.21698272e-01 2.70061642e-01 4.09421474e-01 5.75518787e-01
3.56696963e-01 3.40944946e-01 -5.26245296e-01 -2.46841945e-02
2.72388726e-01 4.10935253e-01 1.33642450e-01 -3.52100313e-01
-1.05594158e+00 6.03290439e-01 2.81998247e-01 3.25041443e-01
-2.79852033e-01 2.55200446e-01 3.66575301e-01 6.56115353e-01
3.21221709e-01 1.25905544e-01 -6.95912838e-01 -1.73809916e-01
-7.47830153e-01 2.69404739e-01 7.76175141e-01 1.18853557e+00
1.15304577e+00 5.93507230e-01 1.65625826e-01 6.15203440e-01
1.90224290e-01 6.58181131e-01 5.62120676e-01 -9.07503307e-01
6.31229758e-01 8.40288281e-01 1.37873664e-01 -7.55537212e-01
-4.59503025e-01 -1.01221347e+00 -1.38789022e+00 6.54295534e-02
3.34232956e-01 -1.00249574e-01 -4.51455086e-01 1.62028658e+00
2.93093383e-01 5.95161282e-02 8.21254775e-02 4.79384243e-01
9.48665261e-01 2.76712477e-01 -6.75353929e-02 2.38672346e-02
1.02172160e+00 -8.41315091e-01 -4.09266323e-01 -3.03865254e-01
6.48896217e-01 -2.67834485e-01 1.27603149e+00 8.41108084e-01
-9.36961532e-01 -8.44959676e-01 -8.69274855e-01 1.99371483e-02
-4.86526012e-01 -3.69212143e-02 9.43912923e-01 6.63733304e-01
-9.89831090e-01 5.48519373e-01 -6.66674018e-01 2.51505584e-01
6.44590437e-01 4.82479483e-01 -7.69351006e-01 -5.24661064e-01
-1.26439905e+00 5.40045619e-01 4.77942556e-01 2.29858831e-01
-1.55145967e+00 -6.70943499e-01 -6.39854610e-01 3.22834961e-02
3.96243900e-01 -7.54380465e-01 8.88017654e-01 -6.64225519e-01
-1.28033805e+00 7.47319221e-01 3.45030665e-01 -4.66586620e-01
5.17114758e-01 -5.30103087e-01 -5.21136463e-01 -2.37488046e-01
-2.88129926e-01 2.55257159e-01 1.04759896e+00 -1.19619501e+00
-5.29046118e-01 -7.28217304e-01 1.47507815e-02 -4.05915260e-01
-7.37171590e-01 -2.19991654e-01 -1.50170997e-01 -5.61785996e-01
-2.75909096e-01 -6.22334003e-01 -1.14362486e-01 -1.74226165e-01
-4.87176389e-01 -4.70034301e-01 8.01250219e-01 -1.47444054e-01
1.22815919e+00 -2.07774067e+00 1.96054071e-01 2.48808879e-03
7.31306970e-01 4.94252473e-01 -2.42184028e-01 7.89761901e-01
3.60176563e-01 5.97740747e-02 -9.77297872e-02 -7.60704815e-01
3.87108207e-01 4.70604867e-01 -3.56956989e-01 3.34036499e-01
-2.79012192e-02 1.20566082e+00 -7.16736317e-01 -3.30293596e-01
1.43712610e-01 5.23997426e-01 -4.05800313e-01 4.65512633e-01
-2.40536630e-01 2.02740759e-01 -8.64520788e-01 9.14887249e-01
6.99213624e-01 -5.25720179e-01 -6.33484870e-02 8.73050168e-02
1.93066806e-01 -1.11036010e-01 -1.18555772e+00 1.54544497e+00
-3.95272553e-01 5.85890971e-02 4.77119237e-01 -1.42701054e+00
1.42625952e+00 3.80684763e-01 4.53051150e-01 -6.44844115e-01
4.40214062e-03 6.74822271e-01 -3.61986011e-01 -4.10336167e-01
2.06572205e-01 -2.01048046e-01 -2.45601550e-01 5.31329274e-01
5.42665839e-01 5.32018468e-02 -1.87995046e-01 1.47998214e-01
1.17365384e+00 4.09989096e-02 -2.40540672e-02 -3.79534483e-01
4.59210128e-01 -3.89151245e-01 7.42559969e-01 8.71193767e-01
-4.20652628e-01 4.45614845e-01 6.61122978e-01 -1.00490010e+00
-8.97816896e-01 -1.09401143e+00 -4.66389544e-02 1.13405681e+00
-1.80058166e-01 -3.44199449e-01 -5.23802519e-01 -7.18693256e-01
2.03656346e-01 5.17720699e-01 -6.92166686e-01 -2.31571153e-01
-2.59341300e-01 -6.05036378e-01 4.76419747e-01 5.26121616e-01
5.77954352e-01 -1.09962916e+00 -2.64997900e-01 -7.51595199e-02
2.54906386e-01 -1.02702475e+00 2.71162182e-01 4.21427220e-01
-8.67047310e-01 -9.18223917e-01 -3.79438102e-01 -5.14406443e-01
6.50483295e-02 4.05527689e-02 1.51315176e+00 3.22969258e-02
-1.70037821e-01 1.96203187e-01 -4.18329298e-01 -2.89363533e-01
-5.15940070e-01 9.07017142e-02 2.61410534e-01 1.07217431e-01
4.20169681e-01 -8.16188812e-01 -5.99672973e-01 3.78651977e-01
-8.39376867e-01 -1.67736039e-01 8.61905098e-01 1.06220734e+00
5.10096669e-01 2.56536782e-01 1.16938627e+00 -1.20080817e+00
4.89246309e-01 -8.08697462e-01 -4.49952245e-01 3.69614780e-01
-9.36508477e-01 -1.27521202e-01 1.12578583e+00 8.50751474e-02
-9.82304752e-01 -4.81693447e-01 -5.17237961e-01 -5.38469255e-01
-3.07726264e-01 4.21557307e-01 -6.20309949e-01 1.17461458e-01
6.34966314e-01 2.32357025e-01 -1.74313530e-01 -8.57451737e-01
5.33435762e-01 9.37318325e-01 4.62822318e-01 -8.53213191e-01
7.32080758e-01 3.67152631e-01 -2.52067178e-01 -5.19346893e-01
-7.65974283e-01 -2.43134230e-01 -9.40592825e-01 2.34553009e-01
3.89246076e-01 -1.05785060e+00 -8.96254599e-01 7.87374854e-01
-7.16600597e-01 -2.60425985e-01 -5.52910388e-01 4.01001781e-01
-4.89430785e-01 3.04747194e-01 -6.62837267e-01 -6.10874772e-01
-6.45624101e-01 -1.11066937e+00 1.21650994e+00 1.05911642e-01
3.26882005e-01 -1.23613787e+00 2.40532197e-02 6.73624873e-01
5.79575837e-01 3.27488244e-01 8.87392938e-01 -9.94522393e-01
-6.30103469e-01 -9.07069325e-01 -2.56889641e-01 8.68750989e-01
-1.76011652e-01 -3.05033512e-02 -1.28723323e+00 -3.57742906e-01
4.55393121e-02 -8.41641724e-01 9.09613431e-01 2.46680826e-02
1.16688454e+00 -2.82341927e-01 -1.32030591e-01 7.05697894e-01
1.20245683e+00 -1.52777687e-01 5.30162692e-01 1.61543176e-01
8.34582090e-01 3.22207540e-01 2.59095103e-01 6.75119758e-01
6.38168812e-01 2.72023112e-01 7.78048396e-01 9.27595943e-02
1.42968848e-01 -3.44832331e-01 4.53502208e-01 8.07193458e-01
-1.25180796e-01 4.90193442e-02 -8.47084582e-01 4.06703800e-01
-1.75844264e+00 -1.03705013e+00 6.56407326e-02 2.15039587e+00
6.30265474e-01 3.82468373e-01 -2.11554430e-02 3.51029485e-02
1.73502252e-01 4.84133184e-01 -7.23851621e-01 -2.90372729e-01
-1.92703724e-01 2.04266459e-01 -1.39226168e-01 -1.97248738e-02
-9.92850602e-01 9.54109371e-01 6.70348692e+00 1.38305533e+00
-8.81076038e-01 3.20150882e-01 7.06484675e-01 3.73600167e-03
-7.66785324e-01 -1.74199253e-01 -1.06119907e+00 2.80370414e-01
9.41937327e-01 1.75787807e-01 1.80662140e-01 1.33777642e+00
-2.65389144e-01 3.88745725e-01 -1.09832156e+00 7.51289248e-01
-2.89737552e-01 -1.56787968e+00 1.69897467e-01 2.41867825e-01
7.33853340e-01 5.60976505e-01 1.82199448e-01 5.83753884e-01
5.75755715e-01 -1.04867280e+00 5.50928354e-01 4.77563232e-01
8.25394213e-01 -6.44678354e-01 1.00511718e+00 9.63363230e-01
-8.09864461e-01 -5.80165148e-01 -6.94109917e-01 -1.51445970e-01
-2.28210673e-01 9.07564640e-01 -4.53678578e-01 7.76176155e-01
8.42253983e-01 9.51361537e-01 -6.06220901e-01 5.42191565e-01
-1.40607789e-01 6.48027837e-01 -1.68701828e-01 9.32654738e-02
2.28115737e-01 -2.29637951e-01 2.75625169e-01 1.05709326e+00
2.80974329e-01 -4.96352434e-01 1.69186562e-01 1.04756200e+00
-2.31760576e-01 -1.71304300e-01 -6.83904171e-01 -2.10396454e-01
5.19477963e-01 1.50646400e+00 4.50685769e-02 -3.55433732e-01
-4.62751657e-01 3.59830618e-01 6.36633933e-01 3.16095233e-01
-5.19827425e-01 -1.50813356e-01 7.11023271e-01 -1.74194165e-02
1.79302156e-01 -6.35426566e-02 -2.24995524e-01 -1.43235481e+00
2.16031298e-02 -9.70923960e-01 5.40189207e-01 -5.00537395e-01
-1.56684566e+00 9.42009330e-01 -2.96158642e-01 -1.07764757e+00
-2.33545989e-01 -8.02836776e-01 -6.23302937e-01 5.40117919e-01
-1.76834607e+00 -1.79467440e+00 -1.94194630e-01 1.31958342e+00
3.16302687e-01 -7.62385488e-01 1.09746337e+00 2.83452213e-01
-8.23637486e-01 8.17878544e-01 7.14848280e-01 4.00758386e-02
4.83765125e-01 -1.31679237e+00 1.81537732e-01 4.56019551e-01
3.06027830e-01 6.78815603e-01 3.52656096e-01 -3.05753767e-01
-1.87495816e+00 -8.52346122e-01 6.06383085e-01 -5.70255220e-01
8.20052683e-01 -4.56404507e-01 -8.92238200e-01 7.47735918e-01
-1.58801779e-01 4.56991464e-01 9.79151070e-01 5.33475935e-01
-8.53294790e-01 -8.69920671e-01 -1.16730940e+00 -2.96589714e-02
7.58614182e-01 -5.95135748e-01 -4.11422849e-01 4.14535522e-01
9.07505035e-01 7.75308385e-02 -1.15561116e+00 4.56206888e-01
5.95804334e-01 -1.42650759e+00 9.50311482e-01 -7.46709108e-01
4.86083537e-01 2.29966134e-01 -5.55107296e-01 -1.02068520e+00
-2.54484326e-01 -8.65475535e-01 -5.23534000e-01 1.15463960e+00
1.43994540e-01 -8.19862425e-01 7.76843250e-01 8.40839267e-01
-5.06428719e-01 -7.55117595e-01 -1.23848641e+00 -9.12422836e-01
4.79345381e-01 -8.65502775e-01 9.11396265e-01 7.40628064e-01
-1.18901275e-01 -6.15356788e-02 -8.27456236e-01 -9.24419761e-02
1.05883241e+00 4.10327524e-01 9.67319727e-01 -1.36265349e+00
-5.55602431e-01 -3.60421240e-01 -3.51706624e-01 -1.00497472e+00
1.71455428e-01 -7.97749162e-01 -3.54340643e-01 -1.23586321e+00
1.68402091e-01 -8.29205632e-01 -7.87647367e-01 6.28879964e-01
2.45298892e-02 1.58155695e-01 7.39111155e-02 4.15628970e-01
-8.90926182e-01 6.76779687e-01 1.26838636e+00 -7.20182210e-02
1.33209124e-01 3.20310235e-01 -1.04729807e+00 7.70941794e-01
3.59296143e-01 -1.79485023e-01 -4.79320467e-01 -5.28879046e-01
4.31379318e-01 3.24924067e-02 4.02925432e-01 -9.59071934e-01
1.76226154e-01 -2.01144591e-01 1.58065036e-01 -1.87646195e-01
1.80896282e-01 -9.24164295e-01 1.05529204e-02 2.65765041e-01
-1.66428357e-01 -3.62543792e-01 -1.78897232e-01 6.40941560e-01
-5.17554104e-01 -6.84251785e-02 7.96840966e-01 -2.12131158e-01
-7.78940141e-01 7.30501652e-01 3.33699256e-01 3.22362959e-01
1.02169895e+00 2.15744856e-03 -5.00044286e-01 -1.42525360e-01
-6.56178594e-01 2.33533263e-01 2.34757587e-01 4.47625935e-01
6.69368267e-01 -1.23869479e+00 -6.42165601e-01 5.68993151e-01
1.20259613e-01 2.51898408e-01 5.76946378e-01 6.84850752e-01
-5.96772842e-02 4.86163080e-01 -1.56013891e-01 -5.64298630e-01
-6.05652809e-01 8.21232438e-01 2.03698203e-01 -5.07196903e-01
-6.26518130e-01 1.01748896e+00 3.36510152e-01 -7.70182490e-01
3.43876868e-01 7.92510584e-02 1.11934625e-01 -8.07066411e-02
6.46736920e-01 4.03917879e-01 1.21297039e-01 -2.81798691e-01
-2.64028549e-01 5.73946834e-01 -1.95577130e-01 5.41103482e-01
1.81715703e+00 2.50518639e-02 -1.13903329e-01 4.43976611e-01
1.19565630e+00 -2.24618077e-01 -1.49302936e+00 -4.96498376e-01
4.68339212e-03 -2.99767017e-01 -8.29254240e-02 -9.78897810e-01
-1.22233737e+00 1.70364726e+00 9.80583131e-02 1.73637673e-01
1.02211547e+00 1.76053867e-01 9.76825237e-01 6.16322994e-01
9.69223261e-01 -1.12514055e+00 2.74208069e-01 4.50686038e-01
6.36801422e-01 -1.54312027e+00 -2.17792884e-01 4.08554792e-01
-7.95323312e-01 1.03904450e+00 5.72143853e-01 -1.70365036e-01
1.01495457e+00 4.77079630e-01 -1.74770653e-01 -3.35484207e-01
-1.13722765e+00 1.20319128e-01 6.32256567e-02 7.67555654e-01
7.47728944e-02 2.03681260e-01 2.58648008e-01 1.26317155e+00
2.26209946e-02 4.50241178e-01 6.61403388e-02 4.16174531e-01
-7.52076566e-01 -1.17500091e+00 -6.14484511e-02 3.27752322e-01
-3.50266635e-01 -4.35930453e-02 -1.11333199e-01 8.07275891e-01
3.76983732e-01 7.38503516e-01 -3.43861818e-01 -7.42936850e-01
8.46230984e-02 7.86583200e-02 -1.62225723e-01 -3.35719973e-01
-6.16561830e-01 6.60472512e-02 -3.03485334e-01 -7.03022420e-01
-7.86860846e-03 -1.40653625e-01 -1.11156785e+00 -5.38547158e-01
-4.53691818e-02 3.15195024e-01 5.50747871e-01 1.01519799e+00
6.81453764e-01 4.06492082e-03 8.01132619e-01 -5.14227092e-01
-1.08427882e+00 -8.11620355e-01 -1.15256500e+00 2.27781013e-01
3.55874181e-01 -6.03217721e-01 -6.63738847e-01 -2.63463408e-01] | [9.281145095825195, 3.6130502223968506] |
c80e6bd4-799c-417c-9af2-eb9bea30ae24 | pixel-wise-agricultural-image-time-series | 2303.12533 | null | https://arxiv.org/abs/2303.12533v1 | https://arxiv.org/pdf/2303.12533v1.pdf | Pixel-wise Agricultural Image Time Series Classification: Comparisons and a Deformable Prototype-based Approach | Improvements in Earth observation by satellites allow for imagery of ever higher temporal and spatial resolution. Leveraging this data for agricultural monitoring is key for addressing environmental and economic challenges. Current methods for crop segmentation using temporal data either rely on annotated data or are heavily engineered to compensate the lack of supervision. In this paper, we present and compare datasets and methods for both supervised and unsupervised pixel-wise segmentation of satellite image time series (SITS). We also introduce an approach to add invariance to spectral deformations and temporal shifts to classical prototype-based methods such as K-means and Nearest Centroid Classifier (NCC). We show this simple and highly interpretable method leads to meaningful results in both the supervised and unsupervised settings and significantly improves the state of the art for unsupervised classification of agricultural time series on four recent SITS datasets. | ['Mathieu Aubry', 'Jean Ponce', 'Elliot Vincent'] | 2023-03-22 | null | null | null | null | ['time-series-classification'] | ['time-series'] | [ 6.37810767e-01 -2.46902794e-01 -3.47880393e-01 -6.14185214e-01
-3.89953673e-01 -1.12934887e+00 4.84896392e-01 4.94447887e-01
-2.83929884e-01 3.85371923e-01 -3.51120293e-01 -6.92328215e-01
-4.18376684e-01 -9.49394763e-01 -4.56864744e-01 -8.35550606e-01
-6.35171294e-01 2.36028641e-01 2.94267923e-01 -3.79858911e-01
-6.84024990e-02 7.05796778e-01 -1.61282051e+00 -6.30307868e-02
1.17425001e+00 9.18478787e-01 4.34375674e-01 7.14571714e-01
-5.39194979e-03 -1.81125790e-01 -1.52543321e-01 4.09276545e-01
5.66446185e-01 -2.21768677e-01 -8.20037663e-01 4.54942286e-01
2.69843638e-01 -1.90666169e-01 2.30927750e-01 9.52827156e-01
8.61317758e-03 1.28784597e-01 7.35108495e-01 -1.16090369e+00
-3.07200283e-01 5.92115700e-01 -8.51383746e-01 7.39096925e-02
-1.55288741e-01 -1.02040015e-01 5.79551578e-01 -9.57759619e-02
6.27294481e-01 8.24862063e-01 1.14872527e+00 -9.46848691e-02
-1.66792095e+00 -2.49829456e-01 2.83496350e-01 -1.15643039e-01
-1.16466463e+00 -1.97310209e-01 3.60962451e-01 -6.64975405e-01
8.95303190e-01 5.53508341e-01 1.05007708e+00 2.63685375e-01
-1.45799085e-01 8.13607275e-01 1.09390020e+00 -5.21565974e-01
2.03732863e-01 -3.64448160e-01 2.85030097e-01 1.47450373e-01
1.71853065e-01 2.73256779e-01 3.00241113e-01 -1.29728526e-01
8.36969912e-01 1.58539265e-01 -7.92182386e-02 -5.37491143e-01
-1.34737206e+00 8.12140942e-01 5.91091037e-01 4.15479153e-01
-5.54535270e-01 -1.51581496e-01 3.89163196e-01 3.06329012e-01
8.78627479e-01 6.82633579e-01 -1.14569294e+00 3.48087132e-01
-1.69943774e+00 3.62798236e-02 4.76064920e-01 8.57335925e-01
9.25299287e-01 3.39944959e-02 3.72098267e-01 5.32424986e-01
1.05379432e-01 8.64849746e-01 4.04210359e-01 -1.03325295e+00
-1.23143978e-01 6.35092616e-01 2.28272900e-01 -9.83648598e-01
-7.17009544e-01 -1.44990265e-01 -7.16966391e-01 1.05666801e-01
4.04920936e-01 -5.85460998e-02 -1.54038930e+00 1.20940340e+00
3.63875151e-01 8.00291300e-02 1.92623854e-01 5.69945335e-01
2.59962231e-01 7.91142702e-01 2.72641987e-01 -4.62758511e-01
1.23826575e+00 -5.58369458e-01 -6.02226615e-01 -1.83859423e-01
1.02625835e+00 -7.42035568e-01 7.76837349e-01 1.24345943e-01
-3.24318618e-01 -3.28698367e-01 -9.80633616e-01 4.52289402e-01
-1.01718163e+00 3.63008648e-01 1.04903817e+00 7.00289667e-01
-9.69695330e-01 7.91802645e-01 -1.27858675e+00 -9.27655578e-01
4.17498320e-01 2.25290537e-01 -3.73481750e-01 2.15889513e-01
-7.86397159e-01 6.77659094e-01 9.14539635e-01 3.56072485e-01
-4.30454671e-01 -7.42436826e-01 -9.31495011e-01 -2.89613247e-01
2.57267565e-01 2.19085351e-01 1.12454855e+00 -1.15129066e+00
-1.24551964e+00 1.16925049e+00 -1.39424860e-01 -9.22076464e-01
2.00923547e-01 -1.42652258e-01 -4.60575581e-01 2.54694462e-01
3.82900953e-01 8.31828058e-01 6.75168753e-01 -1.11622941e+00
-8.12489212e-01 -4.31795359e-01 -4.35188204e-01 -1.21197283e-01
-3.11368585e-01 -2.75329202e-02 1.77348226e-01 -7.94338942e-01
7.79365420e-01 -1.17111063e+00 -5.70946693e-01 -1.62875596e-02
-1.21489070e-01 3.01329583e-01 1.22669113e+00 -1.01209879e+00
9.43864942e-01 -2.03952956e+00 -1.14303634e-01 3.62533420e-01
-4.22117651e-01 3.94042969e-01 -1.28293827e-01 3.29382151e-01
-3.26066315e-01 2.47711390e-01 -1.02082503e+00 3.33390921e-01
-2.57128567e-01 5.95510781e-01 -4.40428793e-01 6.29113555e-01
2.84246802e-01 8.94325376e-01 -1.06389594e+00 -3.08256835e-01
5.16472638e-01 -3.78020667e-02 2.27688048e-02 -4.73970622e-02
-4.62483764e-01 6.48177087e-01 -2.77237743e-01 8.59037697e-01
1.05665803e+00 2.29813922e-02 3.57253015e-01 -2.01657429e-01
-6.52289629e-01 -2.58451343e-01 -1.11827576e+00 1.53203785e+00
-1.55715300e-02 7.87442863e-01 1.38374925e-01 -1.34803891e+00
8.03505301e-01 2.34667659e-01 7.29816556e-01 -4.13587004e-01
-2.68679857e-01 3.78131419e-01 -4.05337811e-01 -4.51694489e-01
7.72060692e-01 2.41699010e-01 -6.70811832e-02 1.61748961e-01
-2.28083394e-02 -5.49653888e-01 3.41828704e-01 -1.92295581e-01
3.19289804e-01 8.28058481e-01 4.19012159e-01 -8.41685593e-01
1.34650305e-01 9.66759443e-01 4.14089382e-01 5.17411530e-01
-2.31715128e-01 4.10685003e-01 1.90971553e-01 -6.37054145e-01
-1.20039928e+00 -7.50912011e-01 -5.43885291e-01 1.22366953e+00
-3.21548022e-02 8.70676264e-02 -6.97171569e-01 -4.70933497e-01
3.05991918e-01 5.58669746e-01 -6.11409664e-01 3.46288949e-01
-2.86710799e-01 -1.43634760e+00 6.59299791e-01 5.84521174e-01
5.27697563e-01 -8.16660881e-01 -9.65231717e-01 3.88869315e-01
-2.90183842e-01 -1.23181212e+00 -2.61514075e-02 6.60989463e-01
-1.43247116e+00 -9.64388967e-01 -7.34399259e-01 -6.51894987e-01
7.02553988e-01 5.86909890e-01 9.02700961e-01 -4.05421317e-01
-4.11492616e-01 1.56319842e-01 -6.91319227e-01 -4.73933429e-01
-2.50889122e-01 4.43678796e-01 -1.55271918e-01 -2.23801583e-01
5.27056277e-01 -4.85366076e-01 -4.85047311e-01 5.07165194e-01
-1.15533030e+00 -6.68944791e-02 3.19532335e-01 6.35969698e-01
7.41480589e-01 2.80993164e-01 1.30605578e-01 -7.27846980e-01
-3.10103893e-02 -4.37969774e-01 -1.09934771e+00 4.95894790e-01
-6.76155448e-01 1.57170072e-02 1.84098437e-01 -3.64397019e-01
-9.67268527e-01 8.24684620e-01 3.26944619e-01 2.28032038e-01
-7.91615367e-01 1.13517070e+00 2.89958358e-01 -3.18765312e-01
8.81933093e-01 1.20360397e-01 -1.38587326e-01 -3.83664370e-01
4.23357427e-01 7.62162566e-01 8.20185244e-01 -2.85350740e-01
8.57659042e-01 8.33417475e-01 9.28579569e-02 -1.32383335e+00
-6.20766044e-01 -8.67640257e-01 -1.49910498e+00 -5.62693663e-02
9.27498221e-01 -9.92596507e-01 8.54184106e-02 7.68504620e-01
-7.12605715e-01 -8.13115776e-01 -3.65972251e-01 5.88192582e-01
-4.61616427e-01 5.08534908e-01 -1.10434741e-01 -8.57495368e-01
-1.24735184e-01 -6.88814402e-01 1.22216177e+00 2.71312147e-01
-3.12844180e-02 -1.13011193e+00 -7.41310343e-02 -8.57627019e-02
3.84586066e-01 9.40324843e-01 5.66290498e-01 -3.13544691e-01
-1.76902771e-01 -3.52795988e-01 -2.76915133e-01 2.04250425e-01
4.67490047e-01 5.26857257e-01 -8.97149801e-01 -2.62059271e-01
-2.56030142e-01 5.50103411e-02 1.02971840e+00 1.01963425e+00
7.09327221e-01 7.93217942e-02 -5.50476551e-01 8.79533172e-01
1.50934541e+00 2.22041056e-01 6.12520456e-01 6.44971073e-01
4.31922078e-01 1.03043306e+00 1.21234453e+00 3.05670112e-01
1.84564948e-01 3.52739692e-01 4.73692536e-01 -6.20709598e-01
6.66906357e-01 1.65226400e-01 5.31269871e-02 3.73553574e-01
-2.58733600e-01 1.23220421e-01 -1.43774319e+00 1.15151322e+00
-2.12275553e+00 -9.36704516e-01 -5.98104358e-01 2.19471598e+00
8.23897958e-01 -2.87566990e-01 1.01804756e-01 3.22778165e-01
7.28746176e-01 1.33388877e-01 -4.31532145e-01 -1.69135913e-01
-5.69134057e-01 1.00732140e-01 1.56784761e+00 2.24925384e-01
-1.92314446e+00 1.39240205e+00 7.17589951e+00 5.03120899e-01
-1.17533052e+00 -2.29533076e-01 4.11576390e-01 5.37013173e-01
8.35457891e-02 2.38686979e-01 -4.81385171e-01 -1.22847902e-02
9.98384774e-01 1.39302507e-01 3.70193958e-01 6.32984281e-01
4.63945150e-01 -4.66758996e-01 -5.92573583e-01 3.87528807e-01
-4.92764622e-01 -1.09940398e+00 -2.99562693e-01 7.01293722e-03
1.02195632e+00 3.00119340e-01 -1.25392556e-01 -3.49627793e-01
5.55433214e-01 -8.25547397e-01 5.46686113e-01 3.64932209e-01
7.56090403e-01 -4.00619239e-01 5.77220798e-01 1.89980596e-01
-1.61533380e+00 -3.77960093e-02 -1.97921097e-01 -2.21875161e-02
2.35126037e-02 4.87908155e-01 -5.03738105e-01 9.20693576e-01
9.78917301e-01 1.10249102e+00 -5.84951282e-01 1.09785378e+00
3.85097750e-02 7.62366772e-01 -9.26034331e-01 6.08110547e-01
7.05943882e-01 -5.31179309e-01 4.01872694e-01 1.31293595e+00
4.28266197e-01 1.37355641e-01 3.84987622e-01 2.67784178e-01
7.47366071e-01 4.13680822e-03 -6.07451022e-01 -4.61766034e-01
3.78723621e-01 1.24061072e+00 -1.35029793e+00 -2.38508880e-01
-6.50128722e-02 9.33691800e-01 -5.05574107e-01 4.82646108e-01
-4.52802598e-01 -4.00590003e-01 4.72407818e-01 2.77729612e-02
5.58261275e-01 -8.59477520e-01 -2.66330689e-01 -8.50362182e-01
-1.42341331e-01 -5.92481673e-01 5.52354276e-01 -7.72420824e-01
-8.48380566e-01 2.44334921e-01 5.34183562e-01 -1.40273333e+00
-2.52330601e-01 -7.69590259e-01 -2.66230464e-01 7.42052197e-01
-2.03471208e+00 -1.60196948e+00 -6.21040523e-01 2.53441334e-01
2.79808402e-01 1.10804014e-01 1.18447554e+00 -2.17642948e-01
-2.11773321e-01 -1.55146316e-01 7.04076469e-01 -5.00131361e-02
6.61818981e-01 -1.45476329e+00 5.60987890e-01 1.18196309e+00
-1.26620710e-01 1.10004887e-01 8.51507783e-01 -6.79704487e-01
-1.07844484e+00 -1.48479211e+00 6.11723483e-01 -8.58815461e-02
8.42523873e-01 1.90216780e-01 -8.98893774e-01 8.38176370e-01
-6.30801246e-02 5.80031797e-02 6.85020924e-01 -1.21506676e-01
-2.15064496e-01 -1.90822512e-01 -1.23842072e+00 2.05749884e-01
7.10337222e-01 -2.92624176e-01 -2.90302902e-01 5.69335461e-01
6.09724224e-01 -2.44221985e-01 -1.16713917e+00 5.84955513e-01
6.27191782e-01 -5.44555426e-01 8.79586756e-01 -5.41451871e-01
1.14284061e-01 -5.50752997e-01 -3.11263144e-01 -1.37239945e+00
-3.20728779e-01 -6.19252563e-01 4.90740895e-01 9.30126667e-01
3.35712910e-01 -6.25402987e-01 6.20646298e-01 8.73588845e-02
-1.28231138e-01 1.90165430e-01 -5.30309200e-01 -1.11652386e+00
2.05708921e-01 -3.04369181e-01 7.70743132e-01 1.36000323e+00
-2.44963571e-01 -5.30889750e-01 3.78285833e-02 7.99954414e-01
7.60377645e-01 4.98393446e-01 5.40330768e-01 -1.73580265e+00
3.97945881e-01 -3.88486922e-01 -6.16762698e-01 -7.91831493e-01
-3.91985252e-02 -5.98419666e-01 1.40254885e-01 -1.48320735e+00
-3.44584912e-01 -5.92613041e-01 1.17747992e-01 8.65343809e-01
-2.08601132e-02 3.32568765e-01 -3.77555370e-01 3.94315302e-01
-1.77298952e-02 5.41508012e-02 6.59477949e-01 -2.03259796e-01
-7.25184262e-01 -7.21610710e-02 -1.52425855e-01 6.84534311e-01
1.15316010e+00 -3.76921386e-01 -2.01376945e-01 -5.67042172e-01
7.31628090e-02 -2.32642144e-01 5.05498111e-01 -7.45535493e-01
-9.12116617e-02 -6.50648117e-01 2.54156619e-01 -9.23091233e-01
-3.55490804e-01 -1.10955906e+00 2.22840846e-01 5.26915967e-01
-1.48428470e-01 -7.20243976e-02 6.91781461e-01 5.35448313e-01
-3.49576265e-01 -1.07371897e-01 7.95687795e-01 -2.73456007e-01
-1.23958647e+00 1.20334215e-01 -5.18202364e-01 -4.70672190e-01
1.11971259e+00 -5.02211511e-01 -2.11698711e-01 1.05166852e-01
-9.56820786e-01 6.15915418e-01 6.49745405e-01 3.30304861e-01
2.57661510e-02 -1.04117548e+00 -8.60403836e-01 2.77411729e-01
4.15835500e-01 2.22549643e-02 -3.77207845e-02 1.00821304e+00
-1.07663608e+00 6.05545998e-01 -4.18399662e-01 -1.04871082e+00
-1.30708671e+00 3.36498499e-01 4.82728869e-01 -1.48821160e-01
-3.73414099e-01 4.27654833e-01 -2.91814566e-01 -8.18734705e-01
-1.36889160e-01 -7.56787002e-01 -4.14478987e-01 5.31285346e-01
9.73819792e-02 8.36707577e-02 1.39309913e-02 -8.75772893e-01
-3.30646098e-01 6.25581801e-01 3.60542357e-01 -7.27736577e-02
1.72332191e+00 -2.03024089e-01 -3.14870566e-01 5.12463450e-01
7.88380623e-01 -6.50078297e-01 -1.44993532e+00 -2.31641814e-01
6.14961743e-01 -3.99182707e-01 2.61903495e-01 -6.80226028e-01
-8.85980189e-01 6.90486312e-01 1.06336534e+00 8.57955217e-01
1.41299224e+00 -3.67267281e-01 5.01008511e-01 4.87603307e-01
1.77646548e-01 -1.37741780e+00 -8.04075003e-01 3.23652506e-01
4.65687901e-01 -1.59676564e+00 2.67894685e-01 -5.51835775e-01
-3.79438996e-01 1.30761445e+00 -4.44314703e-02 1.04186013e-01
8.50572288e-01 2.75790840e-01 3.05403173e-01 -6.68950602e-02
2.58947313e-02 -7.52544105e-01 1.83276191e-01 1.11912596e+00
4.25818652e-01 6.55825555e-01 -4.71317358e-02 -2.47399863e-02
-1.45750754e-02 6.31534904e-02 2.43626356e-01 1.39193654e+00
-6.41400337e-01 -9.96392608e-01 -6.42211080e-01 5.76113641e-01
-1.66589513e-01 -1.25919417e-01 -1.16596103e-01 6.74502492e-01
6.75701573e-02 8.56845558e-01 1.64081499e-01 5.92755824e-02
1.29147783e-01 4.15079594e-02 3.38939965e-01 -3.94349366e-01
-3.78968298e-01 4.30432588e-01 -8.75316337e-02 -2.41728500e-01
-1.23178887e+00 -1.20197392e+00 -1.13117218e+00 -1.35202203e-02
-7.50594616e-01 2.04706546e-02 8.88287723e-01 8.40619743e-01
2.07830474e-01 1.99660748e-01 6.32035792e-01 -1.22282434e+00
-2.64670968e-01 -8.13028634e-01 -7.84884036e-01 6.03950545e-02
4.13148195e-01 -2.61566490e-01 -1.82076842e-01 7.95485139e-01] | [9.458680152893066, -1.5409806966781616] |
f7939174-302e-4bed-a660-05babb32b185 | holistic-deep-reinforcement-learning-based | 2302.02921 | null | https://arxiv.org/abs/2302.02921v1 | https://arxiv.org/pdf/2302.02921v1.pdf | Holistic Deep-Reinforcement-Learning-based Training of Autonomous Navigation Systems | In recent years, Deep Reinforcement Learning emerged as a promising approach for autonomous navigation of ground vehicles and has been utilized in various areas of navigation such as cruise control, lane changing, or obstacle avoidance. However, most research works either focus on providing an end-to-end solution training the whole system using Deep Reinforcement Learning or focus on one specific aspect such as local motion planning. This however, comes along with a number of problems such as catastrophic forgetfulness, inefficient navigation behavior, and non-optimal synchronization between different entities of the navigation stack. In this paper, we propose a holistic Deep Reinforcement Learning training approach in which the training procedure is involving all entities of the navigation stack. This should enhance the synchronization between- and understanding of all entities of the navigation stack and as a result, improve navigational performance. We trained several agents with a number of different observation spaces to study the impact of different input on the navigation behavior of the agent. In profound evaluations against multiple learning-based and classic model-based navigation approaches, our proposed agent could outperform the baselines in terms of efficiency and safety attaining shorter path lengths, less roundabout paths, and less collisions. | ['Jens Lambrecht', 'Teham Bhuiyan', 'Marvin Meusel', 'Linh Kästner'] | 2023-02-06 | null | null | null | null | ['motion-planning'] | ['robots'] | [-3.92957538e-01 3.65343802e-02 -4.61786687e-02 -8.78421739e-02
-2.76592344e-01 -5.07334292e-01 5.24807274e-01 2.39601940e-01
-8.89696360e-01 7.81759083e-01 -1.39509693e-01 -6.84780955e-01
-3.26074630e-01 -1.21849084e+00 -8.05314541e-01 -7.38982260e-01
-2.66588241e-01 4.43608075e-01 7.21224785e-01 -9.72688675e-01
3.06441426e-01 4.01369989e-01 -1.40535367e+00 -4.97467518e-01
9.78493929e-01 6.63947582e-01 2.06564948e-01 5.58835030e-01
1.57478914e-01 7.17157781e-01 -3.18950415e-01 -1.89492375e-01
2.95962006e-01 -2.57742345e-01 -6.63463116e-01 -3.90751481e-01
-9.84795094e-02 -3.72975409e-01 -5.58649242e-01 8.33029389e-01
4.99812156e-01 5.15462399e-01 3.80669355e-01 -1.32299805e+00
1.48086190e-01 3.63602102e-01 -4.39195067e-01 1.71962559e-01
-7.18617737e-02 4.87663627e-01 7.65991032e-01 -2.65913993e-01
5.16004145e-01 1.08095026e+00 6.79091573e-01 3.77705872e-01
-6.35951757e-01 -5.91044843e-01 2.96829998e-01 6.31795645e-01
-1.01513565e+00 -2.23525062e-01 4.07369405e-01 -1.79060727e-01
1.01250744e+00 -4.16404724e-01 5.77501059e-01 8.95740628e-01
8.25514734e-01 5.30140340e-01 6.17385983e-01 -4.25166041e-02
5.44444978e-01 -1.76359579e-01 9.53640565e-02 9.08432066e-01
4.10799533e-01 6.83949172e-01 -2.82028448e-02 1.85601130e-01
3.55967492e-01 -2.69630283e-01 9.40544903e-02 -7.64189780e-01
-9.28697586e-01 9.86761212e-01 7.73312330e-01 -5.58809228e-02
-5.49202919e-01 3.08861494e-01 4.94961739e-01 2.34702647e-01
-3.41048926e-01 5.85106552e-01 -4.48324084e-01 -2.79329091e-01
-3.30568552e-01 6.86887324e-01 6.92066491e-01 6.38536572e-01
8.63527715e-01 3.38556111e-01 1.53964341e-01 3.22060436e-01
1.66315407e-01 5.23575962e-01 3.46064210e-01 -1.15129435e+00
7.51514733e-01 5.15656769e-01 3.99587423e-01 -1.12351537e+00
-1.11700177e+00 -5.47527313e-01 -7.05806017e-01 8.01353276e-01
3.95169854e-01 -7.14034736e-01 -8.15376043e-01 1.94538915e+00
4.72563207e-01 5.68979681e-02 3.73935580e-01 8.61333251e-01
2.61620879e-01 6.10717773e-01 1.51702359e-01 2.89158553e-01
1.10823739e+00 -1.42123663e+00 -5.52907348e-01 -6.37935758e-01
1.09041929e+00 -2.74443477e-01 7.06724107e-01 1.98104694e-01
-9.54743147e-01 -6.32366061e-01 -1.29636681e+00 2.60664165e-01
-6.26836360e-01 -2.45629370e-01 4.13385212e-01 5.28047800e-01
-9.36377704e-01 5.73392630e-01 -9.72423196e-01 -3.80331039e-01
1.33239836e-01 4.81794804e-01 -2.40931317e-01 -1.02046594e-01
-1.50297332e+00 1.36126721e+00 3.35016906e-01 7.13503584e-02
-1.11462069e+00 -2.20571756e-01 -9.22959924e-01 9.55641046e-02
7.44819820e-01 -8.75581264e-01 1.25942516e+00 -2.68403560e-01
-1.61326468e+00 -3.00360341e-02 1.31228924e-01 -8.51524174e-01
5.88905036e-01 -3.99862736e-01 -3.42362076e-01 -2.62397259e-01
1.47988275e-01 6.95892453e-01 4.31938022e-01 -1.17727518e+00
-1.27295363e+00 -2.30399683e-01 5.15671611e-01 6.21992052e-01
1.40264943e-01 -8.14067066e-01 -2.26901248e-01 6.33998599e-04
2.73086261e-02 -1.32725549e+00 -7.25172698e-01 -4.72522348e-01
-1.05181217e-01 -9.27869678e-02 7.39053965e-01 -3.63422453e-01
8.66933644e-01 -2.01245022e+00 2.97753483e-01 1.54604346e-01
-7.03339577e-02 2.88473636e-01 -2.63438791e-01 6.40279770e-01
4.26759839e-01 -4.63883430e-02 6.21565580e-02 -1.97838917e-01
-1.10566996e-01 3.15321863e-01 -2.21278027e-01 4.26594049e-01
-1.56356737e-01 8.07144642e-01 -1.08770740e+00 -6.88747838e-02
3.94746840e-01 2.96265095e-01 -7.74203181e-01 1.79126244e-02
-1.35286450e-01 6.68787122e-01 -4.43656206e-01 2.03047633e-01
4.64728057e-01 2.49249995e-01 6.83170408e-02 2.42492318e-01
-4.52578276e-01 7.73111358e-02 -1.04592562e+00 1.55197835e+00
-6.20808721e-01 5.10929108e-01 -1.16984479e-01 -9.54684734e-01
6.21284664e-01 -6.37333989e-02 3.71228755e-01 -1.46422589e+00
1.83971584e-01 1.87685654e-01 3.55264515e-01 -5.20629227e-01
9.68898833e-01 1.26715451e-01 -3.06815743e-01 1.87764242e-01
-2.61279494e-01 5.82036488e-02 2.58639544e-01 1.49752432e-03
1.19265139e+00 8.78175050e-02 1.31092951e-01 -5.80817126e-02
5.93588293e-01 3.40131044e-01 4.89225149e-01 1.01653385e+00
-4.87769723e-01 -5.99073097e-02 3.64185214e-01 -5.09695828e-01
-9.24364567e-01 -8.32265139e-01 3.97011369e-01 9.38655257e-01
9.30672586e-01 -3.10074180e-01 -6.85180128e-01 -5.05740941e-01
-2.13594392e-01 9.69970703e-01 -6.44179165e-01 -5.59687197e-01
-9.90164340e-01 -5.55012226e-01 5.78828037e-01 3.42578560e-01
8.28893006e-01 -1.05717897e+00 -1.20293951e+00 3.73871833e-01
-1.67188317e-01 -1.18004251e+00 -8.46230388e-02 1.80591926e-01
-6.53366506e-01 -1.23115492e+00 -2.83069760e-01 -7.39793897e-01
4.17935163e-01 5.59410751e-01 6.48811996e-01 1.70209602e-01
2.67379403e-01 2.19749957e-01 -3.60416174e-01 -2.48024300e-01
-4.38601881e-01 6.10488355e-01 1.37843221e-01 -3.90893698e-01
-1.26746893e-01 -2.42928967e-01 -8.49825144e-01 5.94528973e-01
-4.92962003e-01 -5.40585443e-02 7.03546464e-01 8.34245503e-01
1.28486961e-01 5.67059994e-01 5.54310262e-01 -5.21768570e-01
6.91049397e-01 -6.56394422e-01 -8.80810857e-01 -2.60940026e-02
-8.60348761e-01 3.21547151e-01 9.44532752e-01 -7.93565810e-03
-9.03857768e-01 -1.53886810e-01 -3.76889527e-01 2.27238074e-01
-4.72829789e-02 6.74452066e-01 -7.26242214e-02 -7.43701532e-02
5.68741560e-01 2.54640579e-01 2.07985997e-01 1.54509425e-01
2.92976320e-01 -5.15990239e-03 2.72213459e-01 -2.96864748e-01
8.99675310e-01 3.74965012e-01 3.60130280e-01 -6.25516474e-01
-3.58353913e-01 -1.90317392e-01 -1.91766232e-01 -2.66554296e-01
6.70841813e-01 -8.54052722e-01 -9.41984117e-01 5.83822489e-01
-8.95304978e-01 -6.93101346e-01 -8.00664723e-02 2.46827781e-01
-4.93635684e-01 2.98195839e-01 -2.38397747e-01 -6.61313593e-01
-1.20507680e-01 -1.41917467e+00 5.07356107e-01 7.19457388e-01
1.03083804e-01 -9.99575734e-01 3.94920796e-01 6.49497733e-02
6.39621317e-01 1.62036419e-02 8.21734607e-01 -3.19251269e-01
-6.66364729e-01 -2.61161178e-01 2.69744545e-02 -1.34856090e-01
-1.80636093e-01 -4.52493548e-01 -4.24757451e-01 -5.06271601e-01
-4.14235473e-01 -1.71852469e-01 7.12329626e-01 3.43356341e-01
4.14570332e-01 -1.35644704e-01 -6.38514996e-01 4.95583057e-01
1.42742360e+00 6.57639146e-01 7.35698164e-01 1.12258434e+00
3.90048295e-01 8.29544604e-01 9.65028822e-01 2.54129201e-01
9.39854980e-01 6.74261630e-01 1.12683499e+00 9.31290537e-02
3.51309299e-01 -2.09583387e-01 5.04381001e-01 3.12856197e-01
5.66915162e-02 -4.31726724e-01 -9.92143869e-01 5.82484186e-01
-2.18367219e+00 -8.71798635e-01 1.34510279e-01 2.13984942e+00
-4.64912429e-02 5.48200488e-01 1.83748230e-01 -4.82516959e-02
3.44031960e-01 2.28618961e-02 -7.02535212e-01 -3.97199392e-01
5.09026460e-02 -4.00774956e-01 9.17803943e-01 7.93037176e-01
-9.72654164e-01 1.15627933e+00 5.77596617e+00 4.63832319e-01
-1.13664317e+00 -5.91281578e-02 1.54735312e-01 1.59963474e-01
1.48933932e-01 9.27157477e-02 -9.33016241e-01 4.29249763e-01
1.02079535e+00 -2.41747648e-02 4.56691355e-01 7.93131411e-01
3.95072967e-01 -4.97267485e-01 -5.73207617e-01 4.35932904e-01
-4.79280800e-01 -1.24169481e+00 -1.67035267e-01 1.91095755e-01
6.66826487e-01 3.44276726e-01 1.04741156e-01 8.85896206e-01
4.07720089e-01 -8.01496804e-01 7.51029074e-01 2.53322482e-01
-1.71932444e-01 -1.19585979e+00 1.09886551e+00 5.69953144e-01
-1.18852150e+00 -4.85998213e-01 -1.32723615e-01 -1.97887897e-01
5.36436141e-01 -1.61040321e-01 -7.41110623e-01 8.72955501e-01
6.27767861e-01 2.44141653e-01 -5.21213353e-01 1.30699289e+00
-2.48007968e-01 2.07073018e-01 -2.53138304e-01 -2.38657340e-01
8.03084612e-01 -2.92085409e-01 4.57711935e-01 6.01766765e-01
3.46250325e-01 2.15886440e-02 3.16563904e-01 1.69530585e-01
4.22486067e-01 -1.79943085e-01 -7.35711396e-01 3.98753583e-01
3.51219803e-01 9.80793417e-01 -7.28054225e-01 -3.96998636e-02
-3.55605692e-01 5.26555777e-01 3.43677014e-01 3.77795398e-01
-1.24202609e+00 -5.49436212e-01 8.54605794e-01 3.90006863e-02
3.96902114e-01 -6.21634901e-01 -2.29942352e-01 -5.89716375e-01
-2.65190423e-01 -6.08188450e-01 9.41284597e-02 -3.65242720e-01
-4.16571647e-01 7.06997752e-01 -1.59573019e-01 -1.23873627e+00
-4.17160422e-01 -4.87249792e-01 -6.88748181e-01 5.05181372e-01
-1.76481664e+00 -7.08342850e-01 -5.52194595e-01 3.83840710e-01
4.64873165e-01 -4.10505295e-01 5.49648166e-01 4.36323643e-01
-7.26071596e-01 5.03414810e-01 1.04006961e-01 -1.88519850e-01
3.62690240e-01 -8.85468423e-01 6.41629398e-01 8.49705100e-01
-4.14292544e-01 2.85749912e-01 1.03543091e+00 -4.44901943e-01
-1.34297156e+00 -9.66409743e-01 3.40581477e-01 -1.90360889e-01
5.13389766e-01 1.41282991e-01 -5.90234816e-01 4.75758016e-01
2.33376294e-01 -4.05081362e-01 8.39141384e-02 1.10346504e-01
3.50185484e-01 -1.50311083e-01 -7.76919127e-01 1.24118781e+00
8.39356244e-01 1.08053200e-01 -1.24866582e-01 -1.72799170e-01
6.48243189e-01 -6.74317420e-01 -2.86466330e-01 4.83606219e-01
5.84114850e-01 -1.06110632e+00 9.20578957e-01 -5.65875709e-01
2.41820708e-01 -4.94421065e-01 3.34038362e-02 -1.53244483e+00
-4.05526876e-01 -3.59651148e-01 -5.51594906e-02 6.42695546e-01
5.59525669e-01 -7.31916010e-01 1.00451899e+00 2.53405124e-01
-4.48389113e-01 -8.33346903e-01 -1.01070404e+00 -7.12987185e-01
1.30930901e-01 -2.75320649e-01 4.68372285e-01 3.02466542e-01
-2.32010990e-01 6.01609468e-01 -5.45216560e-01 6.33124828e-01
4.27536219e-01 -1.32278413e-01 1.06059945e+00 -7.61128604e-01
1.14141800e-01 -5.96106231e-01 -3.14455301e-01 -1.16323125e+00
1.34664476e-01 -4.68005180e-01 4.43748832e-01 -1.75884545e+00
-5.15103579e-01 -5.54968953e-01 -3.14067751e-01 2.47930259e-01
-7.88294245e-03 -3.89438450e-01 6.73988760e-02 -5.08437455e-02
-9.53698575e-01 8.21617007e-01 1.25145257e+00 -1.47877723e-01
-5.13078868e-01 3.03196996e-01 -5.82178354e-01 5.41685998e-01
1.10090256e+00 -3.46416265e-01 -6.57095671e-01 -6.33644164e-01
4.71375674e-01 3.65441233e-01 2.27620229e-01 -1.48967445e+00
7.94978261e-01 -1.15781792e-01 -1.66275173e-01 -6.52347922e-01
3.49458605e-01 -8.52195978e-01 -5.63334152e-02 1.03379667e+00
1.10536143e-01 4.95993614e-01 4.78307366e-01 7.50444889e-01
-2.18947843e-01 -2.00432166e-01 7.50321269e-01 -9.64039266e-02
-1.25673318e+00 1.99662849e-01 -8.81380558e-01 -6.27458021e-02
1.36123013e+00 -3.09390336e-01 -3.88579458e-01 -4.84509856e-01
-3.59341949e-01 9.57500577e-01 4.27307129e-01 5.63218057e-01
4.76070374e-01 -8.48590612e-01 -3.51951659e-01 1.44035697e-01
-5.46521768e-02 8.25947598e-02 5.37288666e-01 8.70999157e-01
-8.21988344e-01 6.59183204e-01 -6.80211365e-01 -4.81892377e-01
-6.91630542e-01 3.78045917e-01 6.53548181e-01 -5.25758803e-01
-5.06016016e-01 3.83457989e-01 2.01212943e-01 -6.99772716e-01
2.54101008e-01 -9.68229547e-02 -4.67834949e-01 -7.50520229e-02
1.62184879e-01 5.05279064e-01 -9.89687163e-03 -4.53850627e-01
-4.23222035e-01 7.03810751e-01 -1.98580563e-01 -2.44524777e-01
9.20273006e-01 -5.92362225e-01 5.80545008e-01 1.38473306e-02
6.13755822e-01 -2.39701420e-01 -1.55039930e+00 1.15933619e-01
1.65884405e-01 5.06622419e-02 7.98108708e-03 -6.68587148e-01
-9.18087304e-01 6.15710020e-01 7.52187669e-01 2.25385427e-01
6.75904870e-01 -5.72237313e-01 1.00755274e+00 7.98455536e-01
7.94729829e-01 -1.12857044e+00 1.68543860e-01 1.08436525e+00
4.11225975e-01 -1.63928974e+00 -2.20951021e-01 1.53499812e-01
-7.27481544e-01 9.96302307e-01 1.14427018e+00 -1.95316941e-01
5.37786901e-01 -3.68180387e-02 1.41175225e-01 -8.55752155e-02
-7.61550069e-01 -3.28776836e-01 -2.90453821e-01 5.53217292e-01
-2.33674139e-01 -1.81181341e-01 -1.66547000e-01 3.60796660e-01
-2.66171932e-01 -2.67051607e-01 6.93838179e-01 1.19316411e+00
-7.16125309e-01 -1.03668165e+00 6.25733137e-02 8.20629522e-02
1.02440707e-01 2.92091966e-01 2.96779692e-01 1.25613809e+00
2.05071896e-01 9.63693798e-01 -1.68999687e-01 -5.52900016e-01
7.08259881e-01 -3.36397856e-01 2.51005739e-01 -2.71435291e-01
-3.66937518e-01 -4.54506159e-01 2.34777838e-01 -6.75244749e-01
-8.74539018e-02 -5.87533057e-01 -1.64314508e+00 -5.90585530e-01
-2.15798765e-01 2.01147914e-01 6.80316925e-01 1.18056667e+00
5.52448273e-01 7.79213011e-01 5.14766395e-01 -8.44973683e-01
-4.89893138e-01 -4.77795988e-01 -1.59059286e-01 -3.04882787e-02
5.41528344e-01 -9.78401840e-01 -7.88571686e-02 -7.50316978e-01] | [4.928910732269287, 1.2260489463806152] |
890d4d1b-4b73-4e97-8a37-276de73049b7 | asmr-learning-attribute-based-person-search | 2108.04533 | null | https://arxiv.org/abs/2108.04533v1 | https://arxiv.org/pdf/2108.04533v1.pdf | ASMR: Learning Attribute-Based Person Search with Adaptive Semantic Margin Regularizer | Attribute-based person search is the task of finding person images that are best matched with a set of text attributes given as query. The main challenge of this task is the large modality gap between attributes and images. To reduce the gap, we present a new loss for learning cross-modal embeddings in the context of attribute-based person search. We regard a set of attributes as a category of people sharing the same traits. In a joint embedding space of the two modalities, our loss pulls images close to their person categories for modality alignment. More importantly, it pushes apart a pair of person categories by a margin determined adaptively by their semantic distance, where the distance metric is learned end-to-end so that the loss considers importance of each attribute when relating person categories. Our loss guided by the adaptive semantic margin leads to more discriminative and semantically well-arranged distributions of person images. As a consequence, it enables a simple embedding model to achieve state-of-the-art records on public benchmarks without bells and whistles. | ['Suha Kwak', 'Jicheol Park', 'Boseung Jeong'] | 2021-08-10 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Jeong_ASMR_Learning_Attribute-Based_Person_Search_With_Adaptive_Semantic_Margin_Regularizer_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Jeong_ASMR_Learning_Attribute-Based_Person_Search_With_Adaptive_Semantic_Margin_Regularizer_ICCV_2021_paper.pdf | iccv-2021-1 | ['person-search'] | ['computer-vision'] | [ 1.16269477e-01 -2.78583057e-02 -1.62240103e-01 -8.02872062e-01
-8.60957384e-01 -6.23185754e-01 8.77306163e-01 4.48787421e-01
-7.87344217e-01 2.80766159e-01 5.36666751e-01 3.02552879e-01
-3.85516435e-01 -6.63216352e-01 -6.45367265e-01 -6.10652745e-01
2.19275787e-01 8.39377761e-01 -1.26437515e-01 5.31574301e-02
-8.10168684e-02 1.02343462e-01 -1.53178334e+00 2.41426632e-01
5.92671156e-01 1.20089591e+00 -6.25822991e-02 1.93585426e-01
4.49216850e-02 1.84797212e-01 -1.36072293e-01 -1.10681176e+00
3.99846256e-01 -4.33876842e-01 -8.69281292e-01 5.83383292e-02
1.17510617e+00 -1.60137229e-02 -6.65439427e-01 1.12196231e+00
8.04069281e-01 2.73102611e-01 1.01668549e+00 -1.50324392e+00
-9.95191157e-01 3.68472934e-01 -5.82912564e-01 4.29946706e-02
5.27489245e-01 4.11863178e-02 1.40031648e+00 -9.18705046e-01
4.38898355e-01 1.59293854e+00 5.61534345e-01 8.23898733e-01
-1.71071947e+00 -4.44423974e-01 1.66860849e-01 4.38386619e-01
-1.70404661e+00 -4.08715814e-01 7.22070515e-01 -4.82789934e-01
4.59719241e-01 3.48410934e-01 5.02917171e-01 1.22107506e+00
-1.79787129e-01 5.83453417e-01 9.40121293e-01 -4.26113814e-01
6.63554594e-02 5.34890771e-01 1.56201972e-02 5.38648546e-01
1.61200926e-01 -1.35471895e-01 -8.06888640e-01 -3.72087240e-01
2.58981973e-01 1.76100552e-01 -1.86767399e-01 -1.03969955e+00
-1.40899765e+00 8.10473502e-01 6.26319230e-01 7.45079620e-03
-2.96830893e-01 1.70212701e-01 4.43668485e-01 4.98159602e-02
2.42528409e-01 4.15917248e-01 -1.03711262e-01 2.56826758e-01
-4.97185707e-01 4.93931144e-01 4.09560353e-01 9.11398947e-01
7.76538551e-01 -7.66174436e-01 -5.84113777e-01 1.09671366e+00
3.81808519e-01 6.13076448e-01 1.23536237e-01 -8.48538041e-01
5.81461072e-01 6.93381250e-01 6.22678287e-02 -1.03168738e+00
-1.08282417e-02 -1.82142824e-01 -8.53004813e-01 -6.73220158e-02
6.76565707e-01 3.77433598e-01 -6.88821554e-01 2.25138950e+00
4.70322967e-01 -2.20061928e-01 -5.89865446e-02 1.16359603e+00
5.44858813e-01 2.50494689e-01 4.35122937e-01 3.08676064e-01
2.02146339e+00 -7.49146521e-01 -4.56027478e-01 -5.63888788e-01
1.04141615e-01 -4.92014974e-01 1.39750826e+00 -2.46006429e-01
-1.04217672e+00 -4.94875014e-01 -7.18241572e-01 -3.29323262e-01
-5.59025288e-01 -1.00491062e-01 3.80010843e-01 5.63591242e-01
-8.46696138e-01 3.80647302e-01 -2.52721012e-01 -6.91283166e-01
4.56976831e-01 2.25737393e-01 -6.96573913e-01 -2.30410516e-01
-1.31038165e+00 6.58525884e-01 1.85272470e-01 -2.63629556e-01
-5.11334717e-01 -8.62440407e-01 -9.94480968e-01 1.00362606e-01
2.03402460e-01 -1.22417808e+00 7.29526818e-01 -8.67989838e-01
-7.10766435e-01 1.70766437e+00 -2.91811854e-01 -4.12320465e-01
7.93377459e-01 -2.01068789e-01 -3.97808701e-01 2.77433991e-01
5.10946155e-01 9.75115836e-01 9.84680951e-01 -1.26231432e+00
-7.40695894e-01 -9.16139305e-01 -7.00732619e-02 3.85186017e-01
-7.39659488e-01 -1.94701646e-02 -8.22855294e-01 -5.95995545e-01
-2.92503666e-02 -9.73267794e-01 1.41354889e-01 4.56033975e-01
-4.79575127e-01 -6.17354393e-01 3.10629278e-01 -5.75125456e-01
1.02385306e+00 -2.24025106e+00 4.59382981e-01 3.41693521e-01
3.20967317e-01 -4.72795337e-01 -1.87388361e-01 2.31781363e-01
-4.29724380e-02 -4.55659330e-02 -9.15854797e-02 -6.60769939e-01
4.39880311e-01 -1.01931997e-01 -2.88619369e-01 5.63235521e-01
1.97791867e-02 8.96560311e-01 -9.28457379e-01 -7.65822947e-01
-1.27825081e-01 4.87066567e-01 -3.28573912e-01 2.99027532e-01
1.81762397e-01 3.10645014e-01 -3.35553735e-01 7.23813057e-01
5.09195149e-01 -4.59402874e-02 -1.89799815e-01 -5.86563349e-01
5.22012115e-01 -1.13663472e-01 -1.03179049e+00 1.81338906e+00
-1.64043158e-01 3.08157772e-01 -1.90275013e-01 -6.96291506e-01
8.87221515e-01 -2.15892051e-03 4.60496753e-01 -7.85631597e-01
-8.91854241e-02 8.96451250e-03 -4.37093139e-01 -3.79333645e-01
3.51972967e-01 -1.12922125e-01 -5.03164768e-01 3.39082181e-01
8.21246803e-02 2.76028156e-01 4.10164669e-02 2.62110084e-01
5.51682889e-01 -1.24922819e-01 -2.62184031e-02 -2.41560832e-01
6.39598012e-01 -4.01081443e-01 4.36820507e-01 8.77967238e-01
-3.76710862e-01 6.90568030e-01 3.53449136e-01 -5.56130409e-01
-1.45498109e+00 -1.43684399e+00 -2.36005679e-01 1.48379755e+00
3.83882374e-01 -2.26310834e-01 -6.25213027e-01 -8.54681969e-01
5.16142547e-01 5.17584860e-01 -9.98775244e-01 -4.38523382e-01
-2.72957712e-01 -3.68309528e-01 5.00646412e-01 4.61472422e-01
4.33585942e-01 -8.09723973e-01 -2.71772444e-01 -1.96783841e-01
-4.08707380e-01 -1.15099239e+00 -1.16167045e+00 -2.06431583e-01
-2.99372405e-01 -1.00715828e+00 -1.05184364e+00 -9.55265880e-01
7.65658736e-01 -2.12637987e-02 1.41100883e+00 -2.29346082e-01
-3.02900881e-01 8.05360317e-01 -3.92521359e-02 -1.57953590e-01
-6.52266368e-02 -4.85986024e-02 2.66022742e-01 6.27786994e-01
8.33968818e-01 -2.88764805e-01 -9.51645553e-01 4.63561207e-01
-6.55943930e-01 -2.80628145e-01 3.34056973e-01 8.04480135e-01
7.86755204e-01 -2.10814506e-01 1.75860196e-01 -4.15846556e-01
5.05752563e-01 -4.73953247e-01 -2.17453584e-01 6.43280149e-01
-6.72807455e-01 2.56061018e-01 3.53045791e-01 -5.40984511e-01
-7.45105684e-01 1.15379959e-01 3.89031500e-01 -3.73331308e-01
-2.57500082e-01 -5.77475876e-02 -4.31478262e-01 2.00753734e-01
5.20891786e-01 1.83788121e-01 7.47053400e-02 -4.66915727e-01
5.62846363e-01 6.56930566e-01 9.44853723e-01 -7.80531287e-01
7.18548238e-01 6.63540959e-01 4.21866886e-02 -4.30894762e-01
-8.74562919e-01 -6.12387359e-01 -6.61723137e-01 -9.84072760e-02
1.10993159e+00 -9.44351912e-01 -8.29181969e-01 3.83590341e-01
-8.53267729e-01 1.73301995e-01 -3.97630423e-01 3.95582080e-01
-6.37061417e-01 4.16250467e-01 -2.98617959e-01 -6.36200249e-01
-2.91099131e-01 -8.34235370e-01 1.25138104e+00 2.96126276e-01
-3.74661982e-01 -8.73630226e-01 1.37939183e-02 5.05855680e-01
8.28199312e-02 2.15185612e-01 1.09191883e+00 -7.45148420e-01
-3.31057131e-01 -4.51867104e-01 -3.49100322e-01 2.41704658e-03
6.62957951e-02 -5.99658072e-01 -9.54208255e-01 -4.86740083e-01
-4.57152665e-01 -4.48968798e-01 1.11013877e+00 8.79566073e-02
1.28437126e+00 -3.23420703e-01 -5.00789702e-01 7.18117654e-01
1.44706273e+00 -3.24108273e-01 4.60512221e-01 4.71745700e-01
7.59105146e-01 8.70371401e-01 4.85736638e-01 3.26714009e-01
7.14759469e-01 1.04832518e+00 2.89603472e-01 -6.92869946e-02
-5.52331507e-02 -6.31440401e-01 -8.72382149e-02 -1.53003216e-01
2.24835247e-01 -1.90975368e-01 -8.50421011e-01 8.16102624e-01
-1.84680557e+00 -1.12516665e+00 3.65973622e-01 2.71397758e+00
1.01320326e+00 -1.58746064e-01 6.09297633e-01 -2.66255766e-01
1.00635505e+00 8.48281756e-02 -6.96453154e-01 -1.40145430e-02
-2.86902755e-01 -4.32601094e-01 4.01341677e-01 3.10935974e-01
-1.45228863e+00 5.99435091e-01 5.27578592e+00 8.09282839e-01
-4.35558438e-01 -1.84890884e-03 7.49167383e-01 -1.61959291e-01
-5.02378821e-01 -2.57938087e-01 -1.06838286e+00 8.01337540e-01
4.99955565e-01 -2.20563218e-01 3.81614685e-01 6.85859919e-01
-1.67696819e-01 2.61467993e-01 -1.70543861e+00 1.19048393e+00
2.44986087e-01 -9.99524951e-01 2.84990698e-01 1.88397273e-01
3.86975706e-01 -4.86491770e-01 4.42216635e-01 1.10174462e-01
9.07049179e-02 -1.15626824e+00 1.08647370e+00 7.69991875e-01
8.81967664e-01 -7.82612622e-01 5.61506748e-01 3.90790263e-03
-1.20854175e+00 -3.37663680e-01 -4.77588981e-01 4.64437366e-01
1.58985615e-01 2.49468490e-01 -4.70438778e-01 1.54735535e-01
9.87777174e-01 4.09509420e-01 -8.54261994e-01 1.02000523e+00
2.37093791e-01 -5.04314005e-02 -1.98510721e-01 1.50239348e-01
2.29940098e-02 -2.61013895e-01 5.47606230e-01 1.26063526e+00
1.24517843e-01 -2.25159138e-01 2.49005452e-01 1.04919016e+00
-2.68160671e-01 -7.46324752e-03 -5.67160606e-01 2.55653650e-01
8.73038709e-01 1.03415573e+00 -1.88213259e-01 -2.52202183e-01
-3.02600980e-01 1.23601758e+00 5.30386150e-01 3.86847794e-01
-8.02908599e-01 -2.56984383e-01 1.00903201e+00 3.47632200e-01
1.06808856e-01 3.34621131e-01 -2.41163388e-01 -8.14981639e-01
3.09875131e-01 -6.13534689e-01 9.28125679e-01 -6.75449967e-01
-1.99510920e+00 5.17362833e-01 8.50385278e-02 -1.13395822e+00
-1.54966593e-01 -3.54669839e-01 -2.11683765e-01 1.08650517e+00
-1.32643056e+00 -1.42163634e+00 -4.46712106e-01 8.53847384e-01
2.29791224e-01 -4.18125480e-01 8.58957410e-01 3.85765374e-01
-2.37319946e-01 1.18930340e+00 -4.29450497e-02 3.77412856e-01
1.18698108e+00 -1.55154657e+00 3.10309142e-01 5.48754096e-01
1.06850304e-01 6.66116476e-01 8.93604815e-01 -3.63068759e-01
-1.07673597e+00 -9.82852101e-01 1.28948629e+00 -7.51908004e-01
4.81593728e-01 -5.59934855e-01 -9.23858643e-01 4.59214240e-01
2.10166931e-01 1.56189308e-01 8.26195002e-01 2.72830158e-01
-9.59956646e-01 -5.09604156e-01 -1.20454884e+00 5.20132065e-01
1.09722388e+00 -8.46234977e-01 -4.99001563e-01 3.83850694e-01
6.02782488e-01 -8.70065689e-02 -9.42115188e-01 1.05653256e-01
6.92593038e-01 -7.50012875e-01 1.57916152e+00 -9.21019852e-01
3.25433999e-01 -2.59273261e-01 -3.78545076e-01 -1.23434913e+00
-6.03306293e-01 -2.21852109e-01 1.22881025e-01 1.56239855e+00
3.46516639e-01 -4.59759235e-01 6.81019783e-01 1.04991686e+00
4.58584875e-01 -6.72461510e-01 -1.03844011e+00 -8.11900020e-01
1.32240489e-01 1.66698769e-01 7.01756895e-01 7.70574391e-01
-2.54286200e-01 3.96981299e-01 -4.21105802e-01 1.27715543e-01
1.33461821e+00 1.75900340e-01 5.35298705e-01 -1.33130527e+00
-2.11795479e-01 -5.79664946e-01 -6.49552822e-01 -8.25737417e-01
2.50449777e-01 -8.87359917e-01 -1.46310881e-01 -1.24682009e+00
9.08701003e-01 -5.25573432e-01 -4.95244354e-01 3.85878533e-01
-4.82935011e-01 3.47244561e-01 3.21681857e-01 2.95911640e-01
-7.02916920e-01 6.21878684e-01 7.72130489e-01 -5.28943956e-01
1.23300023e-01 -4.53468412e-02 -1.01329887e+00 4.78322774e-01
6.34783745e-01 -4.06484544e-01 -3.02459598e-01 -5.50214112e-01
2.66975194e-01 -3.62539589e-01 1.00110090e+00 -6.28350556e-01
3.47042471e-01 1.12489343e-03 6.59368634e-01 -3.66862208e-01
5.87139904e-01 -1.00762415e+00 -6.88833147e-02 4.08827290e-02
-8.41308534e-01 1.19007416e-01 -2.83615708e-01 9.37810063e-01
-1.56303629e-01 7.26287216e-02 9.67601359e-01 1.03550786e-02
-8.50007534e-01 5.72231352e-01 4.63200212e-01 3.87763023e-01
1.07628596e+00 -3.29512745e-01 -1.18780300e-01 -3.69129509e-01
-7.95744181e-01 4.73994374e-01 6.36395037e-01 6.19719565e-01
6.11208260e-01 -1.87154508e+00 -8.65812600e-01 1.40591547e-01
8.03994238e-01 -4.36757624e-01 2.50525206e-01 6.02794230e-01
1.07634597e-01 2.29661390e-01 -2.25051567e-01 -7.63360083e-01
-1.50451529e+00 8.07420015e-01 5.80613792e-01 3.48218270e-02
-4.74451840e-01 1.15924096e+00 4.83887225e-01 -5.40207684e-01
5.37392557e-01 4.72147912e-01 -1.26432553e-01 2.52108902e-01
6.31710529e-01 1.56025618e-01 -2.42016077e-01 -9.19390142e-01
-6.90180182e-01 7.33904183e-01 -7.79418200e-02 -2.74343848e-01
1.07935381e+00 -3.86802524e-01 -8.45959038e-03 3.15080047e-01
1.39386272e+00 -1.09696850e-01 -1.40373218e+00 -6.87352002e-01
-1.45642847e-01 -8.80166173e-01 -3.91728163e-01 -8.79954576e-01
-8.24656248e-01 6.48530483e-01 1.01078713e+00 6.29158691e-02
1.03128493e+00 6.80146456e-01 6.76829398e-01 9.66791585e-02
1.18406534e-01 -1.21702385e+00 1.99342638e-01 7.40984380e-02
8.30111384e-01 -1.34266639e+00 -1.95164949e-01 -2.11306050e-01
-9.04643834e-01 6.64686918e-01 6.07579291e-01 1.84404641e-01
2.80690581e-01 -4.45935726e-01 -7.66134709e-02 -1.20818131e-01
-3.93972963e-01 -3.33841264e-01 7.40501046e-01 8.09774220e-01
7.49088675e-02 2.22433567e-01 -3.41398455e-02 5.82497418e-01
-1.12578601e-01 -5.62903941e-01 -2.42260888e-01 3.77517700e-01
-3.23549747e-01 -9.62235272e-01 -5.26435912e-01 2.99446672e-01
-2.29027510e-01 -4.30855714e-02 -4.35140848e-01 5.34689546e-01
1.84334964e-01 7.23333597e-01 3.32766145e-01 -1.39519796e-01
6.68491781e-01 1.94939926e-01 5.08666396e-01 -1.40711606e-01
-1.58342376e-01 -2.93948174e-01 -1.39883876e-01 -4.26171064e-01
-1.27701744e-01 -8.35222900e-01 -7.09099412e-01 -4.31831688e-01
2.65950888e-01 1.51802287e-01 3.03412020e-01 6.16267383e-01
4.43431228e-01 -1.22482374e-01 6.83760583e-01 -4.57362980e-01
-9.29772377e-01 -6.57376766e-01 -6.54533446e-01 1.29682636e+00
3.31345767e-01 -6.64896190e-01 -1.58688575e-01 1.21594995e-01] | [14.608053207397461, 0.9663406014442444] |
d282504a-d9f9-4d96-bfd3-c6f91226d2b5 | are-neural-topic-models-broken | 2210.16162 | null | https://arxiv.org/abs/2210.16162v1 | https://arxiv.org/pdf/2210.16162v1.pdf | Are Neural Topic Models Broken? | Recently, the relationship between automated and human evaluation of topic models has been called into question. Method developers have staked the efficacy of new topic model variants on automated measures, and their failure to approximate human preferences places these models on uncertain ground. Moreover, existing evaluation paradigms are often divorced from real-world use. Motivated by content analysis as a dominant real-world use case for topic modeling, we analyze two related aspects of topic models that affect their effectiveness and trustworthiness in practice for that purpose: the stability of their estimates and the extent to which the model's discovered categories align with human-determined categories in the data. We find that neural topic models fare worse in both respects compared to an established classical method. We take a step toward addressing both issues in tandem by demonstrating that a straightforward ensembling method can reliably outperform the members of the ensemble. | ['Philip Resnik', 'Rupak Sarkar', 'Pranav Goel', 'Alexander Hoyle'] | 2022-10-28 | null | null | null | null | ['topic-models'] | ['natural-language-processing'] | [-1.6126262e-01 5.6450057e-01 -2.9829136e-01 -6.0076147e-01
-8.9492655e-01 -7.7497792e-01 9.4676155e-01 3.2853007e-01
-4.4515684e-01 5.5934888e-01 3.0957034e-01 -3.5195374e-01
-2.0883317e-01 -6.6705376e-01 -4.0934622e-01 -3.2931140e-01
1.2947989e-01 8.0632508e-01 2.5981030e-01 -1.8123053e-03
5.0322700e-01 -2.2302018e-01 -1.6963806e+00 1.6043557e-01
9.7706497e-01 8.5803741e-01 -1.8093705e-01 2.6233292e-01
-2.3541149e-01 5.1097631e-01 -7.1615243e-01 -8.4939784e-01
-8.1914082e-02 -6.0155299e-02 -7.6779932e-01 -5.0591260e-02
3.8358596e-01 -4.1071504e-01 1.6957158e-01 1.2268283e+00
2.5827470e-01 -2.7940720e-02 9.7648138e-01 -1.5090035e+00
-5.2303338e-01 8.4828001e-01 -2.9236570e-01 1.7299624e-02
1.4826669e-01 -1.9843498e-02 1.2444330e+00 -7.6702583e-01
6.6732019e-01 1.1072230e+00 9.9273789e-01 1.4432353e-01
-1.3450496e+00 -6.6640139e-01 2.8381690e-01 -1.2677944e-01
-1.4864591e+00 -5.8428669e-01 5.3533286e-01 -8.7641704e-01
7.3916423e-01 1.7894253e-01 6.0329479e-01 1.1078851e+00
1.3382359e-01 4.4681877e-01 1.2623467e+00 -3.3631095e-01
6.5335906e-01 8.7946224e-01 5.2029806e-01 7.4296840e-02
8.4402806e-01 -6.0782943e-02 -5.2102190e-01 -6.4580327e-01
3.5529789e-01 -4.2469028e-01 -2.5637323e-01 -5.5605501e-01
-1.0128542e+00 1.0996252e+00 -4.3218117e-02 4.7265232e-01
-4.7855598e-01 2.0485258e-02 3.2257918e-01 -6.0308255e-02
9.6640944e-01 8.7411195e-01 -4.8277739e-01 -4.5816496e-01
-1.3463827e+00 6.4055014e-01 9.4720566e-01 7.9268342e-01
5.0890875e-01 -2.0897530e-01 7.3763127e-03 7.7062744e-01
4.5507786e-01 1.8422903e-01 7.7623951e-01 -9.1956306e-01
-8.0627739e-02 5.2332217e-01 3.9424431e-01 -1.1637100e+00
-2.4848813e-01 -4.2963761e-01 -2.0988660e-01 2.4310562e-01
4.8653245e-01 -5.1201917e-02 -6.3135523e-01 1.8076705e+00
3.3905932e-01 -2.5077736e-01 -2.7020156e-01 6.0076743e-01
3.7810814e-01 4.9033126e-01 6.0606784e-01 -1.7786828e-01
1.3572526e+00 -4.0801173e-01 -6.2241369e-01 -1.9681452e-01
5.3207958e-01 -5.8641160e-01 1.1496401e+00 5.3183872e-01
-1.0771035e+00 -3.6100060e-01 -1.0982078e+00 1.7726555e-01
-4.6143192e-01 -1.3285643e-01 8.0346942e-01 1.1392759e+00
-1.0165869e+00 7.0999521e-01 -7.3318672e-01 -5.5218840e-01
2.6609495e-01 2.8771052e-01 -1.6722925e-01 3.5237002e-01
-1.1283073e+00 1.2305304e+00 5.3156042e-01 -1.7232822e-01
-6.0985452e-01 -6.2560648e-01 -5.3471071e-01 6.4177372e-02
2.5962463e-01 -5.3820133e-01 1.6715750e+00 -1.0511522e+00
-1.1000222e+00 8.7159669e-01 2.9825158e-02 -4.9075037e-01
5.1107877e-01 -1.7473036e-01 -4.9020943e-01 -1.8369536e-01
2.8351247e-01 5.5826777e-01 7.5835359e-01 -1.3027978e+00
-8.2556856e-01 -3.9044294e-01 6.0315866e-02 2.1077114e-01
-5.5781001e-01 -1.8764727e-03 -2.1914527e-02 -5.6591851e-01
6.7917667e-02 -8.0107486e-01 -2.9236324e-02 -1.5380265e-01
-3.2984543e-01 -5.0666350e-01 4.8394099e-01 -3.8532799e-01
1.4429961e+00 -1.7102226e+00 -4.0595439e-01 2.9987559e-01
5.1972955e-01 6.0059663e-02 3.1219360e-01 4.3081033e-01
7.4984685e-02 6.2750053e-01 9.0995468e-02 -3.4803241e-01
4.2598870e-01 -3.1258139e-01 -6.1442322e-01 4.6965268e-01
-1.2619750e-01 6.3033468e-01 -8.5158908e-01 -7.1967715e-01
-4.1801456e-02 4.3949172e-01 -4.8107946e-01 7.6097928e-02
-2.8947365e-01 -2.0606910e-01 -4.8532623e-01 4.9114394e-01
4.3104711e-01 -4.0522024e-01 3.1781873e-01 -1.8103704e-02
-9.2120424e-02 7.7857548e-01 -9.6495223e-01 1.3375241e+00
-3.7525162e-02 7.1715534e-01 -2.5875354e-01 -5.6134653e-01
6.8993461e-01 4.7155565e-01 2.7470562e-01 -2.0910136e-01
1.4988942e-01 2.2813140e-01 1.1718450e-01 -2.9752725e-01
9.2314070e-01 -4.7662574e-01 1.8904649e-01 8.9430904e-01
2.8859788e-01 -3.0823961e-01 -2.0170826e-01 1.8742892e-01
6.8955988e-01 1.5031099e-01 4.6984750e-01 -6.8812782e-01
-3.2679352e-01 4.5205492e-01 3.5708541e-01 9.0331042e-01
-2.4627085e-01 6.7380053e-01 4.4591618e-01 -3.9692417e-01
-1.0828873e+00 -8.1578732e-01 -5.5571520e-01 1.1395723e+00
-3.6323439e-02 -6.0590750e-01 -1.0217671e+00 -7.9396725e-01
-9.1871142e-02 1.3391722e+00 -9.3218529e-01 -1.3056326e-01
2.0726347e-01 -7.6120287e-01 5.1437515e-01 4.3815145e-01
-6.6260359e-04 -6.6131657e-01 -9.8014939e-01 3.3192098e-01
-2.0621140e-01 -7.1452308e-01 -6.9772080e-02 3.0857800e-02
-8.6158472e-01 -8.6405075e-01 -7.4155730e-01 -3.3505518e-02
3.9193058e-01 1.9085053e-01 1.6294398e+00 -1.9910714e-02
3.3070031e-01 4.7447222e-01 -2.6577380e-01 -8.8173825e-01
-5.9995526e-01 3.4651518e-01 1.3908084e-01 -3.7247217e-01
1.0370054e+00 -6.4001936e-01 -5.2941597e-01 5.2201587e-01
-8.2371944e-01 -1.8326281e-01 3.2331371e-01 5.5329818e-01
-1.4574023e-01 2.1803308e-01 6.7674983e-01 -1.2982411e+00
1.0736032e+00 -7.0881063e-01 -3.8683030e-01 2.7849746e-01
-1.2945307e+00 -1.2975085e-01 -2.5757429e-01 -4.4284812e-01
-1.0200678e+00 -4.8260966e-01 1.5503633e-01 -7.5581945e-02
-1.9908491e-01 7.1342927e-01 1.9676492e-01 3.2167125e-01
1.0051268e+00 -2.3376587e-01 7.8874588e-02 -2.9389870e-01
1.9803303e-01 7.9906553e-01 2.5774181e-01 -7.3924047e-01
6.6229898e-01 2.8083345e-01 -9.1927147e-01 -5.7516867e-01
-8.2720035e-01 -3.5335013e-01 -3.7856275e-01 -1.9309697e-01
5.2293342e-01 -1.0938675e+00 -3.9339882e-01 1.3690281e-01
-1.2378362e+00 -2.5614572e-01 -2.2390887e-01 4.1103035e-01
-4.7358796e-01 1.4428867e-01 -2.2362351e-01 -1.1662290e+00
-2.2197273e-01 -1.0510939e+00 1.0560570e+00 -7.1344741e-02
-1.1773926e+00 -1.0283688e+00 3.8981450e-01 1.4281841e-01
7.5317776e-01 -1.9456217e-02 1.0435928e+00 -1.2673873e+00
-1.1159499e-01 -6.0111797e-01 -6.3538991e-02 1.2608580e-01
-1.6190672e-01 2.0938115e-01 -1.4556136e+00 -9.6139334e-02
2.3174606e-01 -2.8039238e-01 5.7348478e-01 4.8769632e-01
5.7435459e-01 -1.7931160e-01 -5.7711947e-01 -1.0085671e-01
1.2044530e+00 1.0750973e-01 4.3601406e-01 4.8918927e-01
4.1567922e-02 9.9327123e-01 4.7315022e-01 3.4479484e-01
3.8759232e-01 7.7094173e-01 4.8966616e-02 2.2771081e-01
3.2645869e-01 -4.6853036e-01 1.3006350e-01 5.0126028e-01
-5.9750974e-02 -3.0021873e-01 -1.2055641e+00 6.4695907e-01
-1.8048187e+00 -7.9741013e-01 1.6643406e-01 2.5680823e+00
7.5927401e-01 6.0391480e-01 4.5236459e-01 -6.2440775e-02
6.7689973e-01 -5.3325709e-02 -2.8835353e-01 -1.6635486e-01
1.8846074e-01 -2.7384821e-01 1.4514747e-01 2.2985770e-01
-9.9261457e-01 7.1608973e-01 7.8901200e+00 7.5603074e-01
-8.4102887e-01 3.0461416e-01 9.5591789e-01 6.5852195e-02
-6.3756031e-01 1.2948456e-01 -6.2529135e-01 5.6834227e-01
1.1528119e+00 -5.4826033e-01 -1.3421412e-01 1.3619012e+00
-3.5341837e-02 -2.2419977e-01 -1.3588760e+00 5.7836634e-01
3.5156090e-02 -1.0307157e+00 -1.3170268e-01 4.7302070e-01
6.2103164e-01 -6.9664955e-02 2.2833690e-01 5.4689270e-01
6.0149276e-01 -1.1549836e+00 1.0912197e+00 1.7678387e-01
5.5255765e-01 -3.8924390e-01 8.6667329e-01 3.2153720e-01
-4.0768099e-01 1.9173971e-01 -4.1213709e-01 -8.9552894e-02
9.6253954e-02 6.7884290e-01 -1.0540849e+00 -1.8151203e-01
5.7236671e-01 1.5119044e-01 -6.7742163e-01 9.9891865e-01
1.0635810e-01 9.1322547e-01 -4.5103928e-01 -1.3056310e-01
1.7932446e-01 3.3045446e-03 5.9961194e-01 1.0520185e+00
3.0762321e-01 -1.4253750e-01 -2.1317072e-01 1.2125958e+00
2.8603268e-01 4.3436557e-02 -8.8701355e-01 -3.4753436e-01
5.8796668e-01 1.1337004e+00 -1.0335261e+00 -4.5134446e-01
-1.8121816e-01 3.2608768e-01 1.6288370e-01 3.2115334e-01
-7.3775208e-01 -3.2724224e-03 6.3727039e-01 3.1233361e-01
2.6956083e-02 4.6085265e-02 -6.7279404e-01 -1.0949290e+00
1.3037388e-02 -1.0982392e+00 2.4001004e-01 -8.8861752e-01
-1.2384233e+00 8.9792740e-01 2.8162113e-01 -1.1454192e+00
-5.1150978e-01 -4.7957525e-01 -5.8238494e-01 7.4679190e-01
-8.3197176e-01 -1.0136482e+00 -9.5605113e-02 -4.9313623e-02
4.2536607e-01 -5.3637635e-02 9.8984063e-01 -1.4692464e-01
-2.1029635e-01 4.7552130e-01 3.6903648e-03 -3.1389248e-01
7.9149920e-01 -1.3548430e+00 6.1229527e-01 5.1369983e-01
3.3212432e-01 1.1140004e+00 1.2883719e+00 -7.4000400e-01
-6.9341165e-01 -5.6569839e-01 9.0820032e-01 -1.2117357e+00
7.0778579e-01 -4.0371025e-01 -9.7705597e-01 7.9261422e-01
2.8145704e-01 -6.8929553e-01 9.3033284e-01 1.0246273e+00
-6.0659289e-01 3.9317954e-01 -1.0701531e+00 4.8360461e-01
6.2186539e-01 -5.7631391e-01 -8.5829318e-01 1.9533218e-01
6.3408083e-01 -4.8873503e-02 -7.9398233e-01 2.4023604e-01
1.0235801e+00 -1.0778002e+00 6.3897735e-01 -9.7372085e-01
6.5918589e-01 1.3235807e-01 -2.9987550e-01 -1.3501981e+00
-2.7255103e-01 -4.4749188e-01 1.1267129e-01 1.4407696e+00
7.5289428e-01 -6.0766464e-01 8.5641021e-01 1.4338423e+00
2.0242283e-01 -6.4855492e-01 -5.9107631e-01 -4.6225879e-01
4.3512139e-01 -7.9851943e-01 6.4142025e-01 1.1102310e+00
5.0487161e-01 4.0291178e-01 -4.8729838e-03 -5.7803817e-02
5.5152154e-01 -4.7082923e-02 7.4323863e-01 -1.6774877e+00
-1.7297290e-01 -6.3832855e-01 -5.3644323e-01 -6.2035215e-01
3.3827517e-02 -2.5309587e-01 1.1416740e-01 -1.1867652e+00
4.5838296e-01 -5.9573954e-01 -3.6040425e-01 5.7116482e-02
-3.5655686e-01 1.1588989e-01 1.4310058e-02 4.4254407e-01
-6.1141747e-01 3.0837530e-01 3.2056066e-01 5.5766225e-02
-1.1874850e-01 7.3748499e-02 -1.2711434e+00 1.0988389e+00
6.0030329e-01 -5.2128285e-01 -6.5137190e-01 -2.5777304e-01
6.1884421e-01 -4.3576461e-01 2.3228051e-01 -1.0079062e+00
3.0352557e-01 -1.1803320e-02 2.7723756e-01 -3.7022290e-01
3.1120086e-01 -8.1610698e-01 3.1437680e-01 -8.5781693e-02
-7.6107734e-01 1.2138892e-01 1.8359482e-01 6.1286032e-01
-1.3443184e-01 -4.4836137e-01 3.5541162e-01 -1.4464435e-01
-4.7352841e-01 -5.6320466e-02 -4.4903687e-01 8.0906384e-02
7.5179464e-01 -4.4488096e-01 -3.9809278e-01 -7.1106714e-01
-4.7988871e-01 -1.0225261e-01 7.1976048e-01 3.7188023e-01
2.1375285e-02 -1.0420536e+00 -6.0095775e-01 -2.0841162e-01
3.6802721e-01 -3.2155481e-01 3.9667264e-02 8.5795909e-01
-3.3109676e-02 6.6553426e-01 5.6226101e-02 -5.1113200e-01
-8.3541679e-01 4.7553059e-01 2.5311381e-01 -1.5835637e-01
-3.5697472e-01 6.2834662e-01 4.3090084e-01 -3.1305641e-01
3.3921272e-01 -1.5039334e-01 -5.7081684e-02 2.0132390e-01
4.9389184e-01 3.3137324e-01 1.6776510e-01 -4.2112750e-01
2.7387772e-02 4.3710726e-03 -4.3040931e-01 -5.9739643e-01
1.2169774e+00 -4.6976779e-02 1.1018462e-01 9.4331342e-01
7.7656204e-01 -1.7160229e-01 -9.5166111e-01 -1.2680492e-01
2.0773666e-01 -3.9128491e-01 2.4724746e-01 -1.0531552e+00
-3.3365306e-01 7.7244723e-01 4.9879846e-01 8.6864656e-01
5.6918913e-01 -2.6210204e-02 2.7241084e-01 3.3376971e-01
4.6264455e-01 -1.3096491e+00 -4.6431601e-01 -1.3323650e-01
7.0848083e-01 -1.4645149e+00 2.6164806e-01 -4.0102389e-01
-7.6770264e-01 8.8925225e-01 5.1396090e-01 2.1569577e-01
7.4297851e-01 -9.0139983e-03 2.1206078e-01 -5.2210736e-01
-1.0350373e+00 2.1123303e-01 3.4936914e-01 6.5630072e-01
7.8144014e-01 1.5800130e-01 -3.8174027e-01 9.2346269e-01
-4.8264313e-01 5.8050565e-03 4.3766701e-01 6.6211170e-01
-6.2499291e-01 -6.4509505e-01 -2.5117511e-01 6.5421277e-01
-6.3103706e-01 -1.3411671e-01 -5.2212298e-01 1.0984236e+00
-1.4788057e-01 9.3914819e-01 9.0112649e-02 -4.8488048e-01
-4.4211145e-02 5.0367564e-01 -9.2598125e-02 -7.5295478e-01
-5.3780919e-01 1.2167194e-01 4.1103643e-01 -3.5875338e-01
-1.8251553e-01 -9.0271354e-01 -3.7364402e-01 -1.8714140e-01
-7.7874148e-01 6.2736505e-01 8.8083643e-01 9.7793597e-01
4.4796443e-01 -1.0990140e-01 2.0006362e-01 -5.9834331e-01
-1.1261487e+00 -1.3209008e+00 -6.6494876e-01 3.6662433e-01
-1.4821495e-01 -8.8030392e-01 -4.5731971e-01 9.3492068e-02] | [10.377471923828125, 7.138607501983643] |
9a3a2eb8-a662-468c-9120-74828d40ccc0 | automated-seismic-source-characterisation | null | null | https://doi.org/10.31223/osf.io/nbmzt | https://doi.org/10.31223/osf.io/nbmzt | Automated Seismic Source Characterisation Using Deep Graph Neural Networks | Most seismological analysis methods require knowledge of the geographic location of the stations comprising a seismic network. However, common machine learning tools used in seismology do not account for this spatial information, and so there is an underutilised potential for improving the performance of machine learning models. In this work, we propose a Graph Neural Network (GNN) approach that explicitly incorporates and leverages spatial information for the task of seismic source characterisation (specifically, location and magnitude estimation), based on multi-station waveform recordings. Even using a modestly-sized GNN, we achieve model prediction accuracy that outperforms methods that are agnostic to station locations. Moreover, the proposed method is flexible to the number of seismic stations included in the analysis, and is invariant to the order in which the stations are arranged, which opens up new applications in the automation of seismological tasks and in earthquake early warning systems. | ['Jean-Paul Ampuero', 'Martijn van den Ende'] | 2020-09-16 | null | null | null | null | ['seismic-source-localization'] | ['time-series'] | [ 5.78520559e-02 -1.86377808e-01 2.57859588e-01 1.37304552e-02
-4.92134184e-01 -6.44457459e-01 4.39429551e-01 7.60458529e-01
-4.50821340e-01 4.63113576e-01 2.33491048e-01 -5.52207172e-01
-6.21501207e-01 -1.30801678e+00 -5.30709386e-01 -8.45538378e-01
-8.02655816e-01 4.57070947e-01 3.70299041e-01 -1.85153186e-01
4.51131403e-01 8.90780449e-01 -8.11900795e-01 -1.84899092e-01
2.37207234e-01 7.68682420e-01 1.60412431e-01 8.12075615e-01
2.92937547e-01 8.38774979e-01 -6.48106873e-01 7.71316588e-02
-1.32417366e-01 -1.35271519e-01 -6.68799400e-01 -5.33699334e-01
-1.53875068e-01 -8.07356536e-02 -5.53329647e-01 5.87233067e-01
7.11282134e-01 2.90486038e-01 8.15473199e-01 -8.26393008e-01
-8.11756328e-02 7.81489313e-01 -4.40280974e-01 6.71951354e-01
6.87571391e-02 -3.36604506e-01 1.09173250e+00 -6.25967562e-01
1.95966214e-02 6.83241844e-01 1.30466723e+00 -2.97129601e-01
-9.71147597e-01 -2.79962331e-01 -1.35864377e-01 2.50635177e-01
-1.51623034e+00 -4.10773516e-01 1.15701342e+00 -6.87452137e-01
7.51418114e-01 1.48132592e-01 4.99479145e-01 5.29775858e-01
9.94858220e-02 2.15213329e-01 4.71866757e-01 -4.31743890e-01
6.14000976e-01 -6.36844993e-01 1.19911090e-01 2.54948735e-01
3.75962228e-01 -2.88166165e-01 -6.50541782e-01 -5.81827402e-01
1.02796924e+00 5.90423420e-02 -2.65117019e-01 8.32281560e-02
-1.18824971e+00 1.11102939e+00 4.30919886e-01 4.71465260e-01
-6.99193358e-01 3.89371753e-01 6.35248005e-01 -5.75147606e-02
6.45171165e-01 7.33601034e-01 -2.49767795e-01 -1.87627703e-01
-1.37284517e+00 1.77880570e-01 8.30467284e-01 2.53401607e-01
6.60626054e-01 4.90602702e-01 4.85111207e-01 7.87323952e-01
1.14630966e-03 4.02293921e-01 1.89907476e-01 -7.58586645e-01
4.42339480e-01 4.72106099e-01 -2.08866224e-01 -1.70395887e+00
-8.61909986e-01 -5.15281320e-01 -1.01719284e+00 -1.68586433e-01
4.54267561e-01 -5.85374832e-01 -3.36677432e-01 1.43756258e+00
1.95637092e-01 7.88468480e-01 -9.63481888e-02 6.28694654e-01
5.52181423e-01 5.51803946e-01 -2.01481923e-01 8.57551470e-02
1.15419412e+00 -2.29970187e-01 -2.86265701e-01 -3.19055438e-01
7.07694054e-01 -3.91644478e-01 2.70765245e-01 2.38240242e-01
-7.95731544e-01 -4.11410667e-02 -7.14838207e-01 4.53453541e-01
-3.12063575e-01 -1.30546346e-01 6.93167984e-01 4.89751101e-01
-1.15720987e+00 7.11625218e-01 -9.75019276e-01 1.47058919e-01
1.63064703e-01 4.03284132e-01 -3.51209104e-01 6.36988282e-02
-1.52050495e+00 8.74388456e-01 5.75393200e-01 7.01088488e-01
-5.77144921e-01 -6.70150340e-01 -8.82288277e-01 4.00807619e-01
2.53997773e-01 -1.95575356e-01 8.32236350e-01 -9.75204557e-02
-8.74105990e-01 6.83851689e-02 1.59975991e-01 -4.90142941e-01
2.68146366e-01 8.00124630e-02 -2.29173601e-01 4.26879883e-01
2.03605950e-01 -1.12455808e-01 3.80791694e-01 -9.01480496e-01
-3.87518615e-01 -8.30100700e-02 -2.94978265e-02 -1.27883255e-01
-3.60115349e-01 -8.24282393e-02 -8.99203569e-02 -8.31973433e-01
3.01345557e-01 -6.49488807e-01 -5.66780806e-01 -4.24622774e-01
-5.11900187e-01 -5.84029481e-02 5.89301348e-01 -8.76707792e-01
1.43020344e+00 -2.21816802e+00 -2.08126474e-02 8.78732264e-01
1.87279105e-01 -5.33722267e-02 1.92137659e-01 1.15860522e+00
3.61109413e-02 -1.12378746e-01 -6.44511044e-01 2.91712526e-02
-2.64519155e-01 2.82740086e-01 -2.92400807e-01 6.01609528e-01
1.07862502e-01 4.47471142e-01 -1.06597710e+00 -3.02477807e-01
1.85008913e-01 4.18288171e-01 -4.95989293e-01 -1.77317951e-02
2.56867975e-01 4.94349867e-01 -6.56268179e-01 4.55199003e-01
2.61525065e-01 -2.78369248e-01 3.50855947e-01 1.20596059e-01
-2.31315404e-01 2.37021014e-01 -1.30138171e+00 1.17441022e+00
-4.27361518e-01 8.64168465e-01 -2.34495759e-01 -1.34794569e+00
1.05706441e+00 6.63819134e-01 8.71776700e-01 -2.18112946e-01
-3.10280472e-01 2.15066224e-01 -1.23184964e-01 -6.18601143e-01
4.36314285e-01 2.95029208e-02 -2.41144374e-01 5.35817087e-01
-2.52854615e-01 5.94061948e-02 1.22909799e-01 3.87592055e-02
1.44906330e+00 -3.63472193e-01 3.85417193e-01 -2.14956254e-01
1.63634852e-01 -5.01723448e-03 4.46014017e-01 9.11993146e-01
3.94773424e-01 7.18073905e-01 8.17134142e-01 -5.43852568e-01
-1.06683946e+00 -6.60891235e-01 -1.30127043e-01 9.76386786e-01
-2.69544303e-01 -6.85585216e-02 -5.27916729e-01 -6.34048879e-02
-8.89123231e-03 4.15804535e-01 -6.47985160e-01 5.49831912e-02
-9.95288134e-01 -9.90968287e-01 1.14095330e+00 7.49241114e-01
-3.30512077e-02 -9.33135152e-01 -7.53551185e-01 7.06062257e-01
-2.08183885e-01 -1.07660997e+00 4.20141630e-02 2.01516971e-01
-9.29578066e-01 -1.05628633e+00 -7.69087136e-01 -6.30945683e-01
4.40163493e-01 4.98420671e-02 7.99305797e-01 9.87809524e-02
-2.76415772e-03 2.49929979e-01 -3.76186639e-01 -2.35834152e-01
-1.16901882e-01 3.79689336e-01 -1.81101680e-01 2.03152195e-01
-1.83238789e-01 -1.04106760e+00 -6.31717801e-01 4.38903049e-02
-8.98597360e-01 -3.53933007e-01 4.86922830e-01 5.43724298e-01
4.88511920e-02 2.73740202e-01 1.07954407e+00 -6.88658893e-01
5.18547535e-01 -1.20657480e+00 -4.46016163e-01 8.25982764e-02
-1.27162308e-01 -1.53766021e-01 6.53183520e-01 -3.57688725e-01
-6.64713979e-01 -5.83062656e-02 -9.47670043e-02 4.35539931e-02
-1.37476608e-01 1.49264884e+00 3.47085774e-01 -3.28032762e-01
4.93715346e-01 1.10919908e-01 -3.23937863e-01 -6.16117895e-01
-2.64378097e-02 4.94744480e-01 7.41765797e-01 -3.14329803e-01
5.48037767e-01 4.34991330e-01 6.01389349e-01 -1.18781793e+00
-3.25653106e-01 -6.77948594e-01 -8.32544148e-01 -4.44977224e-01
6.18637800e-01 -6.54389858e-01 -6.12174928e-01 5.01125932e-01
-1.29497683e+00 -1.09777503e-01 2.44529620e-01 7.97399342e-01
-2.05003649e-01 5.07634759e-01 -5.80636799e-01 -1.07475424e+00
-9.22054723e-02 -6.27237916e-01 7.91644037e-01 -1.55206308e-01
-2.85010040e-01 -1.68270254e+00 2.84066856e-01 -1.32309705e-01
3.77943218e-01 6.55099869e-01 1.12576246e+00 -9.60883558e-01
-3.52124244e-01 -5.39409161e-01 -1.13759913e-01 -7.69397691e-02
1.98756620e-01 -8.03080276e-02 -8.94343495e-01 -1.03623889e-01
-3.08280252e-02 3.02995384e-01 8.32328439e-01 7.14976549e-01
7.78554857e-01 -5.20193934e-01 -1.78684279e-01 2.74551868e-01
1.34231830e+00 -6.89029247e-02 3.97504359e-01 5.47379911e-01
8.96102309e-01 9.09608543e-01 -1.26593038e-01 7.21191049e-01
4.31082428e-01 4.79078770e-01 5.97942770e-01 -2.04514101e-01
2.74312079e-01 1.11316510e-01 -1.71290606e-01 9.56654906e-01
-4.93175447e-01 -2.59620577e-01 -1.56022227e+00 1.06130373e+00
-1.84407473e+00 -1.16324973e+00 -4.53480661e-01 2.09293842e+00
6.11792862e-01 -1.07521303e-02 1.06316067e-01 6.04996145e-01
8.63159835e-01 4.88789767e-01 -8.19450021e-02 -1.33766085e-01
-1.78780988e-01 1.20052591e-01 8.22956085e-01 2.67856836e-01
-1.02786756e+00 3.48886639e-01 6.73929691e+00 6.57167196e-01
-1.15787196e+00 -1.10681161e-01 2.58987337e-01 1.93508387e-01
-2.58610755e-01 -8.97032488e-03 -2.11617231e-01 4.53410208e-01
1.06003833e+00 1.80611670e-01 1.75561309e-01 5.36220372e-01
6.08117640e-01 -2.07180783e-01 -9.12140369e-01 3.46457511e-01
-8.29712749e-02 -1.64016998e+00 -1.96535796e-01 -1.01623954e-02
3.81882846e-01 2.28579164e-01 -2.35802323e-01 -3.61124873e-01
3.86399627e-02 -8.51161063e-01 6.47466481e-01 6.42607033e-01
2.30568796e-01 -8.62487018e-01 9.51461852e-01 4.24077988e-01
-1.20204413e+00 -2.06882417e-01 -8.81559029e-02 -4.67843235e-01
5.40937424e-01 7.59082615e-01 -1.21141338e+00 8.26363266e-01
6.22524261e-01 5.54164112e-01 -4.21010762e-01 1.58319259e+00
-1.46074072e-01 1.26375306e+00 -5.07439315e-01 1.01114295e-01
5.25387108e-01 4.22304831e-02 5.32360792e-01 1.35922647e+00
4.92965221e-01 -3.34627205e-03 1.01057306e-01 1.80221811e-01
1.87070653e-01 -5.35087958e-02 -6.63748324e-01 3.11019212e-01
8.73631001e-01 8.22698295e-01 -9.02509630e-01 -2.63974932e-03
-4.02644247e-01 -2.82968543e-02 2.02858374e-01 5.89662075e-01
-3.56529832e-01 -5.39554298e-01 5.75704947e-02 2.93166071e-01
4.59867835e-01 -6.61569536e-01 -5.99150836e-01 -6.46223009e-01
3.18179205e-02 -3.15848947e-01 4.75662529e-01 -6.48613095e-01
-1.23346281e+00 4.09280419e-01 2.30759114e-01 -1.16065621e+00
-6.16262972e-01 -2.77284563e-01 -1.21389139e+00 7.16688275e-01
-1.60774255e+00 -1.10555959e+00 1.28893256e-01 3.19628716e-01
6.57441691e-02 -5.85040404e-03 7.89149344e-01 5.01012743e-01
-4.27698791e-01 7.93741271e-03 4.20588881e-01 7.38775790e-01
2.07324460e-01 -9.47006047e-01 5.75998247e-01 8.96673739e-01
1.39028966e-01 3.36402953e-01 8.38673472e-01 -7.32794821e-01
-1.00182641e+00 -1.10550022e+00 1.15207529e+00 -6.44205660e-02
1.23935497e+00 -1.06880866e-01 -1.22504258e+00 5.23080885e-01
-1.52482465e-01 -9.11308974e-02 8.64351869e-01 1.98606074e-01
-8.52124691e-02 -2.44858954e-02 -6.57221138e-01 3.41865778e-01
3.89437437e-01 -5.13333499e-01 -5.61149478e-01 3.27476591e-01
2.75896519e-01 -2.02763360e-02 -1.10641718e+00 2.98808306e-01
2.96172857e-01 -4.46229219e-01 1.20817637e+00 -2.84353971e-01
2.47418985e-01 -1.50799900e-01 -5.08554559e-03 -1.19549656e+00
-5.53284109e-01 -4.10575837e-01 -9.05747153e-03 9.59651113e-01
4.45584327e-01 -8.05251062e-01 6.51549935e-01 3.93137217e-01
-3.11311722e-01 -5.44361174e-01 -1.39887440e+00 -6.40490949e-01
8.21900293e-02 -5.50884545e-01 7.20817626e-01 1.15691924e+00
-2.34525055e-02 -5.15569150e-02 -5.64379096e-01 8.70536149e-01
6.70712948e-01 -1.03509642e-01 1.74245864e-01 -1.58238029e+00
-3.09927136e-01 -5.91819108e-01 -7.07398117e-01 -5.43918192e-01
-4.30888217e-03 -4.58680600e-01 -5.02402075e-02 -1.45726764e+00
-2.89447218e-01 -8.17004740e-01 -3.97409499e-01 5.30961454e-01
-9.53479186e-02 4.73638535e-01 -4.19856131e-01 5.50974429e-01
-6.79085925e-02 7.88441226e-02 4.82855141e-01 7.24472404e-02
-8.70021135e-02 3.88471752e-01 -1.52835995e-01 9.54583287e-01
9.28777874e-01 -7.55215824e-01 -3.45164910e-02 -8.76119018e-01
6.48002863e-01 4.47602361e-01 7.50620484e-01 -1.02145112e+00
7.97336280e-01 -3.17950636e-01 1.54314563e-01 -5.13445199e-01
1.86609298e-01 -5.39662421e-01 2.69659966e-01 2.13224337e-01
-2.64048904e-01 3.52061898e-01 2.74195732e-03 7.03027308e-01
-3.69806021e-01 -6.06714845e-01 1.57069489e-01 1.32504534e-02
-6.05411470e-01 1.04611836e-01 -6.88664377e-01 -1.21818520e-01
5.58644950e-01 -3.12392503e-01 -7.60041848e-02 -4.33600336e-01
-5.84836841e-01 -9.64049995e-03 -9.18205827e-03 -1.89097136e-01
5.37734210e-01 -9.82044101e-01 -8.86171937e-01 -5.51118068e-02
-1.27066404e-01 1.82159796e-01 4.41417158e-01 8.95482004e-01
-9.43903565e-01 1.66539863e-01 9.90506783e-02 -4.50175136e-01
-6.65722668e-01 6.79625124e-02 1.90428406e-01 -2.04375312e-01
-5.32324493e-01 5.68177044e-01 -1.85648799e-01 -1.13774166e-01
9.00373911e-04 1.45658106e-02 -6.05635524e-01 2.67297596e-01
3.09490681e-01 8.76021266e-01 2.29998127e-01 -6.63028002e-01
-3.37039292e-01 4.30935651e-01 5.18253207e-01 -2.25677833e-01
1.94255853e+00 9.46967602e-02 -2.55342394e-01 5.90466917e-01
7.36840665e-01 5.46256527e-02 -1.08164084e+00 -3.80021244e-01
5.20693362e-01 -7.20345750e-02 1.92307651e-01 -2.55296022e-01
-1.10689521e+00 9.70581651e-01 -3.16420734e-01 6.62866950e-01
9.71192241e-01 -9.35457647e-02 5.56102455e-01 6.70089662e-01
2.04463392e-01 -8.64668310e-01 -2.08739623e-01 6.00818634e-01
5.50928712e-01 -6.97101414e-01 2.49023885e-02 -1.58824325e-01
-2.62375116e-01 1.34402585e+00 -1.74930319e-01 -2.92050987e-01
8.26465666e-01 3.54615241e-01 -5.60737774e-02 -3.50971252e-01
-3.01041991e-01 2.23075971e-01 1.96092322e-01 7.29482591e-01
1.87575698e-01 -1.09242834e-01 1.71167597e-01 2.79861420e-01
-1.11832149e-01 -4.57113147e-01 5.80041409e-01 1.09584403e+00
-6.80435419e-01 -7.44580626e-01 -6.38506055e-01 4.16985095e-01
-6.36711776e-01 -4.54284459e-01 9.77874454e-03 5.74365675e-01
-2.92085886e-01 1.01535714e+00 1.17387421e-01 -2.77815126e-02
1.80236608e-01 -2.87668645e-01 -1.10806786e-01 -5.52980781e-01
-6.08645558e-01 -2.31912881e-02 9.48046669e-02 1.42894119e-01
-3.22389543e-01 -7.47296333e-01 -1.23440218e+00 -3.86591882e-01
-5.08506954e-01 5.72843969e-01 7.90355682e-01 1.19219697e+00
1.59179419e-01 5.56164980e-01 8.54339123e-01 -1.18619466e+00
-1.37935132e-01 -9.77604508e-01 -8.00914705e-01 -9.68404934e-02
6.15471244e-01 -5.66027462e-01 -3.59210342e-01 3.90325822e-02] | [6.870316505432129, 2.7198286056518555] |
9a48dbef-fa92-4451-a2e7-6cd04dcf0c86 | w2vv-fully-deep-learning-for-ad-hoc-video | null | null | https://dl.acm.org/doi/pdf/10.1145/3343031.3350906?download=true | http://lixirong.net/pub/mm2019-w2vvpp.pdf | W2VV++: Fully Deep Learning for Ad-hoc Video Search | Ad-hoc video search (AVS) is an important yet challenging problem in multimedia retrieval. Different from previous concept-based methods, we propose an end-to-end deep learning method for query representation learning. The proposed method requires no concept modeling, matching and selection. The backbone of our method is the proposed W2VV++ model, a super version of Word2VisualVec (W2VV) previously developed for visual-to-text matching. W2VV++ is obtained by tweaking W2VV with a better sentence encoding strategy and an improved triplet ranking loss. With these simple changes, W2VV++ brings in a substantial improvement in performance. As our participation in the TRECVID 2018 AVS task and retrospective experiments on the TRECVID 2016 and 2017 data show, our best single model, with an overall inferred average precision (infAP) of 0.157, outperforms the state-of-the-art. The performance can be further boosted by model ensemble using late average fusion, reaching a higher infAP of 0.163. With W2VV++, we establish a new baseline for ad-hoc video search. | ['Xirong Li; Chaoxi Xu; Gang Yang; Zhineng Chen; Jianfeng Dong'] | 2019-10-21 | null | null | null | acm-multimedia-2019-2019-10-1 | ['ad-hoc-video-search'] | ['computer-vision'] | [ 2.31340788e-02 -2.78148502e-01 -2.96367943e-01 -2.73927271e-01
-1.41769326e+00 -5.17197251e-01 9.26550329e-01 5.36144555e-01
-7.92603076e-01 2.18951568e-01 4.84022975e-01 6.96143433e-02
-1.55777559e-01 -3.89885962e-01 -8.57355595e-01 -3.20062459e-01
3.51813855e-03 4.34681416e-01 4.70623791e-01 -4.81005043e-01
1.37410656e-01 9.63717699e-02 -1.89199126e+00 7.49278545e-01
4.63569045e-01 1.64328706e+00 1.23647392e-01 5.54652631e-01
-1.20102949e-01 6.71291769e-01 -3.14126819e-01 -7.78563678e-01
2.90448099e-01 8.46773908e-02 -8.16812038e-01 -2.63665408e-01
1.10872090e+00 -3.93431008e-01 -8.21401417e-01 5.69113910e-01
9.29665446e-01 2.10768983e-01 7.74111867e-01 -1.12424791e+00
-6.27647340e-01 4.94138330e-01 -4.57465321e-01 1.10301092e-01
6.38509810e-01 -2.21480373e-02 1.50077534e+00 -1.16216993e+00
8.62230062e-01 1.29424083e+00 4.36858952e-01 3.25100809e-01
-1.00340307e+00 -5.83069384e-01 9.89831612e-02 9.08536077e-01
-1.89255548e+00 -3.16511154e-01 5.63114822e-01 -3.46522957e-01
9.97518480e-01 4.71610218e-01 6.33336484e-01 1.30860209e+00
-1.13484845e-01 1.39109254e+00 4.09064204e-01 -5.03459752e-01
4.32541184e-02 2.49199048e-01 9.39373076e-02 5.70119560e-01
-4.85322289e-02 -1.67943984e-01 -7.81789482e-01 -2.33172908e-01
-2.21849401e-02 -6.68942854e-02 -3.45844030e-01 -5.64722061e-01
-1.17136908e+00 8.76409233e-01 6.65644944e-01 3.33483785e-01
-3.73593539e-01 5.55322170e-01 7.32528508e-01 4.32970911e-01
4.42859143e-01 3.30261409e-01 -2.71083355e-01 3.32742371e-02
-1.40233481e+00 7.52578914e-01 4.02785629e-01 1.02251434e+00
4.23380643e-01 -2.02509224e-01 -1.06662869e+00 9.45261061e-01
3.66861224e-01 6.39251769e-01 3.96284550e-01 -9.26691175e-01
3.93912166e-01 3.35635900e-01 1.67763740e-01 -1.08667350e+00
-2.97140509e-01 -6.16881907e-01 -4.40717936e-01 -3.20556819e-01
-5.53615727e-02 3.69381338e-01 -1.04706287e+00 1.58197141e+00
1.38024669e-02 1.17770575e-01 -2.06403658e-01 9.65041161e-01
1.27614319e+00 8.00912499e-01 2.51394808e-01 -5.76214939e-02
1.56100523e+00 -1.03861403e+00 -6.95180357e-01 1.57270089e-01
6.20251536e-01 -7.98239470e-01 9.20225441e-01 3.71320188e-01
-1.08992815e+00 -5.35437346e-01 -1.08685374e+00 -2.68159330e-01
-4.64765608e-01 -5.74312638e-03 6.35774016e-01 4.95135903e-01
-1.66017818e+00 2.23013178e-01 -2.92603105e-01 -5.77409387e-01
3.73866320e-01 2.83293694e-01 -4.38448668e-01 -4.35014158e-01
-1.40059948e+00 8.73108566e-01 2.73875415e-01 -1.98790222e-01
-9.33878779e-01 -8.44840765e-01 -7.97386765e-01 2.16634408e-01
5.25398016e-01 -1.01300251e+00 1.17402947e+00 -5.86432099e-01
-9.45306361e-01 8.41824472e-01 -3.52242500e-01 -7.35120833e-01
6.36223435e-01 -2.69188583e-01 -3.28829348e-01 4.96715069e-01
-3.85856889e-02 1.16353929e+00 9.53982472e-01 -1.35821533e+00
-4.06026602e-01 -1.68595240e-01 -5.73130585e-02 1.27901152e-01
-7.84414887e-01 2.47207314e-01 -1.17931366e+00 -6.76059842e-01
-2.32374534e-01 -8.49205256e-01 5.43412231e-02 2.30004609e-01
-5.33056855e-02 -7.07759976e-01 4.71704125e-01 -7.00414062e-01
1.51688218e+00 -1.93788993e+00 3.25384498e-01 2.97627985e-01
3.94364059e-01 4.88887399e-01 -7.54345715e-01 7.69970357e-01
1.50646701e-01 -1.73137859e-02 7.61021227e-02 -6.77659810e-01
2.24962786e-01 -1.99874267e-01 -1.98368683e-01 2.44709969e-01
2.10100189e-01 1.23244274e+00 -8.57316375e-01 -8.48845541e-01
1.23997085e-01 7.35655308e-01 -7.01862395e-01 3.16615254e-02
-3.73072028e-01 -2.20243514e-01 -3.05163980e-01 6.65835440e-01
6.27523959e-01 -1.86842889e-01 -1.04557641e-01 -4.49221522e-01
1.27287373e-01 -1.04045197e-01 -8.59804809e-01 2.20859051e+00
-2.83478111e-01 8.63810837e-01 -2.15003937e-01 -9.38519299e-01
6.26694798e-01 3.96624357e-01 8.46661747e-01 -1.21518290e+00
1.00873172e-01 9.15985405e-02 -5.41464269e-01 -7.78460443e-01
8.99543524e-01 5.18502235e-01 -7.07937330e-02 4.28921729e-02
4.11872596e-01 4.89222556e-02 3.80537987e-01 8.18579376e-01
1.07086062e+00 3.64032164e-02 -2.52596319e-01 -9.23266485e-02
6.36965990e-01 -1.65457591e-01 1.32039204e-01 9.78851855e-01
-7.02876505e-03 8.25788975e-01 1.99507996e-01 -3.24395210e-01
-1.06305027e+00 -8.67106676e-01 -7.89547712e-02 1.33998573e+00
6.18683957e-02 -8.40917110e-01 -5.57461560e-01 -5.43700755e-01
1.49771169e-01 6.01959109e-01 -6.49453521e-01 -2.55776107e-01
-4.41796392e-01 -3.54291916e-01 5.87703824e-01 3.53154242e-01
2.66238749e-01 -7.17441380e-01 -1.30535975e-01 4.45755161e-02
-4.16090280e-01 -1.29974377e+00 -6.58701777e-01 -4.86131430e-01
-3.36289316e-01 -7.58335114e-01 -1.30676997e+00 -7.53448486e-01
-2.20756494e-02 3.77799243e-01 1.39861035e+00 1.51297852e-01
-3.77733737e-01 9.29178596e-01 -7.87319958e-01 -4.06601220e-01
8.64600316e-02 -3.07638780e-03 -2.29415316e-02 2.64827073e-01
3.55365157e-01 -2.51781046e-02 -9.47725534e-01 -6.29364550e-02
-1.04566193e+00 -3.46959740e-01 3.67570579e-01 8.41248333e-01
5.79143345e-01 -5.56453288e-01 2.21783102e-01 -2.44644523e-01
5.85790217e-01 -4.19413686e-01 -3.73492569e-01 6.08571708e-01
-1.00212741e+00 9.66914445e-02 3.02163628e-03 -4.02967244e-01
-5.41418731e-01 -1.29858002e-01 -2.53456175e-01 -9.05793786e-01
2.36162096e-01 7.10620761e-01 1.97332054e-01 -1.47756025e-01
5.35365939e-01 2.64708072e-01 -1.89011097e-01 -5.52698255e-01
5.13650656e-01 5.44586420e-01 2.68890351e-01 -1.43221229e-01
5.80935001e-01 2.82006234e-01 -1.74692962e-02 -7.56744266e-01
-5.70483744e-01 -8.19075048e-01 -2.35581532e-01 -3.81196856e-01
8.63095343e-01 -1.40620124e+00 -7.35012531e-01 2.43136927e-01
-1.26254165e+00 1.68507963e-01 3.97878885e-02 3.72128040e-01
-2.54199743e-01 6.37205064e-01 -2.82668889e-01 -8.29751134e-01
-6.22346878e-01 -1.00872636e+00 1.49356604e+00 -2.60069996e-01
-8.27566832e-02 -7.13215292e-01 1.36050180e-01 6.28939211e-01
6.62573636e-01 -2.62833275e-02 7.14552701e-01 -8.08052182e-01
-5.10457933e-01 -5.95323265e-01 -4.06743586e-01 1.53464615e-01
-6.23210609e-01 -1.53614685e-01 -1.01396370e+00 -5.47678113e-01
-6.13071918e-01 -5.74833274e-01 1.63203645e+00 3.18672508e-01
1.32416058e+00 -1.31048024e-01 -2.78581113e-01 4.32137012e-01
1.49164963e+00 -1.01071283e-01 7.07599819e-01 3.66954982e-01
6.09524131e-01 4.77225333e-01 6.12136781e-01 5.25670588e-01
7.40017056e-01 1.32432425e+00 6.25982165e-01 3.58251520e-02
-4.72684860e-01 -1.71430036e-01 2.68883735e-01 6.45522773e-01
-5.87846860e-02 -6.26398265e-01 -8.53674710e-01 5.97260773e-01
-1.98267555e+00 -1.11013305e+00 -1.62469987e-02 2.20902634e+00
5.83967388e-01 -2.14814171e-01 1.41836375e-01 2.88592670e-02
3.22030216e-01 5.04334748e-01 -2.18464345e-01 -2.76022941e-01
-2.82501727e-01 4.11076427e-01 3.42046559e-01 4.92881447e-01
-1.22289038e+00 9.74275470e-01 5.95586920e+00 1.35196388e+00
-9.47759271e-01 2.91158587e-01 1.06563263e-01 -4.51661825e-01
-7.06221879e-01 -2.26652458e-01 -6.18446410e-01 3.58704925e-01
8.62134516e-01 -6.45309165e-02 2.90108979e-01 4.79923844e-01
-6.77680299e-02 2.00437173e-01 -1.13217056e+00 1.48441136e+00
7.31458664e-01 -1.60204887e+00 5.50656378e-01 -2.97553271e-01
4.58102524e-01 2.20298737e-01 2.45763227e-01 6.32806361e-01
-3.12829316e-01 -8.47207665e-01 8.30792546e-01 6.45381033e-01
9.46481466e-01 -5.46938181e-01 8.34157288e-01 -2.22247094e-01
-1.26769269e+00 -1.61411270e-01 -3.19742113e-01 6.35050476e-01
1.83047086e-01 3.01754087e-01 -4.66255099e-01 7.67440617e-01
7.96275198e-01 8.27342629e-01 -1.00919831e+00 1.29098213e+00
3.30922931e-01 3.87552172e-01 -1.10923499e-01 -3.20338547e-01
3.71254444e-01 2.54794747e-01 6.67366862e-01 1.58379567e+00
3.27544242e-01 -2.66014725e-01 2.31514703e-02 4.15367484e-01
-4.46575254e-01 5.12945890e-01 -4.40780520e-01 -6.17456883e-02
4.00392741e-01 1.12099314e+00 -2.76006833e-02 -5.15707850e-01
-3.67199481e-01 1.09960938e+00 1.98181435e-01 5.15854418e-01
-8.78848255e-01 -4.15383369e-01 5.82975745e-01 1.29056215e-01
6.48929358e-01 7.61342645e-02 3.69557589e-01 -1.22297263e+00
2.40075439e-01 -8.04219961e-01 5.88517308e-01 -9.48886752e-01
-1.13567173e+00 7.62450516e-01 1.64834768e-01 -1.47874773e+00
-3.52453768e-01 -4.67454284e-01 -1.27450213e-01 5.29782712e-01
-1.66344821e+00 -1.42031944e+00 -3.28440249e-01 6.95065856e-01
6.36820138e-01 -4.47340190e-01 8.07878971e-01 6.73245072e-01
-3.39414626e-02 1.23875570e+00 2.22326607e-01 -3.59676890e-02
9.54945803e-01 -7.85612345e-01 1.55289888e-01 5.00246346e-01
4.28247631e-01 3.56393576e-01 7.92867541e-01 -1.78077087e-01
-1.68308687e+00 -1.09555221e+00 1.24123132e+00 -5.80601335e-01
6.35743856e-01 -5.00584006e-01 -7.00995386e-01 1.81604996e-01
6.39704466e-01 -5.71978278e-02 6.04030490e-01 -1.59675464e-01
-8.86589527e-01 -3.61359566e-01 -9.93347526e-01 5.28907239e-01
1.12081230e+00 -7.05820739e-01 -3.40635478e-01 6.43168330e-01
1.15110970e+00 -1.20797738e-01 -8.92321467e-01 5.64624548e-01
8.35312843e-01 -7.56142020e-01 1.44714510e+00 -7.79319286e-01
3.66039842e-01 4.44982499e-02 -4.88313377e-01 -9.13101256e-01
-3.05859029e-01 -3.58582258e-01 -2.51980841e-01 1.08343756e+00
3.99036288e-01 -1.30962372e-01 4.96969938e-01 4.92520005e-01
-7.85543397e-02 -7.24292219e-01 -1.13536561e+00 -8.77303779e-01
1.12807311e-01 -7.89631724e-01 3.04040253e-01 6.81632817e-01
-9.19515118e-02 1.91525102e-01 -5.39520502e-01 -2.79380769e-01
7.29658782e-01 -1.12831786e-01 5.73541462e-01 -1.14888823e+00
-1.07065447e-01 -5.34584641e-01 -7.83630490e-01 -1.09366751e+00
1.72173858e-01 -1.04580569e+00 -2.87447602e-01 -1.72931731e+00
5.86440623e-01 -3.66630522e-03 -6.44289136e-01 2.71963030e-01
-1.07018873e-01 5.95030785e-01 5.97794771e-01 1.70759633e-01
-1.23609877e+00 8.69076908e-01 7.76648462e-01 -7.59487271e-01
1.57260016e-01 -3.99226815e-01 -4.42562073e-01 1.21176898e-01
3.00636441e-01 -2.58883864e-01 -2.85355568e-01 -7.98603594e-01
4.22068745e-01 -6.36805743e-02 4.56515759e-01 -8.47623646e-01
3.42483014e-01 4.27548587e-01 2.96037555e-01 -8.48213494e-01
8.25543642e-01 -8.17435503e-01 -1.44603938e-01 3.71079445e-01
-7.00366020e-01 2.21570611e-01 1.50502428e-01 7.29091048e-01
-5.42771578e-01 -7.94010535e-02 1.25545785e-01 1.14316419e-01
-9.95201349e-01 5.65718293e-01 -4.48321670e-01 -7.12392572e-03
7.09665477e-01 7.92564750e-02 -3.24640572e-01 -7.94509530e-01
-5.36241710e-01 6.87427461e-01 1.14903077e-01 8.55410933e-01
1.05991924e+00 -1.59527636e+00 -1.11374938e+00 -2.29379311e-01
7.75550723e-01 -5.82590938e-01 3.73715580e-01 8.21804047e-01
-2.02258885e-01 8.82458270e-01 2.16108024e-01 -6.93045795e-01
-1.56591129e+00 7.73012340e-01 -6.12138547e-02 -2.29208976e-01
-2.40198702e-01 1.15132821e+00 -1.79200411e-01 1.21658714e-02
6.77830160e-01 1.18306242e-01 -5.54822445e-01 4.31921214e-01
6.65062428e-01 2.84959942e-01 2.72943228e-01 -6.57891691e-01
-5.36824107e-01 7.58126974e-01 -2.63773382e-01 -3.06261063e-01
1.25721002e+00 -9.35878754e-02 1.46529004e-02 1.37565896e-01
1.76917076e+00 -2.49769911e-01 -5.06800294e-01 -4.62398052e-01
-1.73705332e-02 -5.18695831e-01 3.00859243e-01 -6.85732722e-01
-1.08008528e+00 7.72143781e-01 1.07652545e+00 -1.81278765e-01
1.08948863e+00 2.61180520e-01 9.00824487e-01 6.02970779e-01
3.90739650e-01 -1.04547679e+00 2.24308088e-01 3.99917454e-01
1.15467048e+00 -1.50363195e+00 7.19554946e-02 3.88524123e-02
-7.55260587e-01 7.58170128e-01 1.98069647e-01 4.48677056e-02
5.96066773e-01 -3.77716482e-01 -6.41631261e-02 -2.75863588e-01
-1.03572905e+00 -4.87598717e-01 9.59046066e-01 3.67934853e-01
4.40127134e-01 -3.61540496e-01 -5.49119174e-01 3.56038570e-01
2.92329133e-01 6.69162199e-02 -2.90556625e-02 7.33753920e-01
-2.30150804e-01 -1.10706258e+00 -1.03866495e-01 3.58883917e-01
-4.59940255e-01 -5.29068112e-01 -3.73730958e-01 5.00546634e-01
-2.15267137e-01 1.05241406e+00 1.80251017e-01 -8.36573541e-01
4.33203220e-01 1.00981131e-01 5.22801399e-01 -7.86787570e-02
-6.89056635e-01 -7.77770132e-02 1.28899008e-01 -1.10124302e+00
-6.00672066e-01 -5.78061461e-01 -7.23557413e-01 -1.93177640e-01
-1.88476548e-01 1.15877964e-01 8.42001438e-01 6.12569451e-01
5.74919999e-01 3.30298394e-01 5.30049562e-01 -7.97864079e-01
-4.60127413e-01 -9.60190117e-01 -2.18249589e-01 6.39147758e-01
4.76219803e-01 -5.91510832e-01 -1.34838089e-01 -1.38663411e-01] | [10.472843170166016, 0.9085996150970459] |
5f45e28c-3078-44e2-a189-e5681fa95e96 | brain-like-object-recognition-with-high | 1909.06161 | null | https://arxiv.org/abs/1909.06161v2 | https://arxiv.org/pdf/1909.06161v2.pdf | Brain-Like Object Recognition with High-Performing Shallow Recurrent ANNs | Deep convolutional artificial neural networks (ANNs) are the leading class of candidate models of the mechanisms of visual processing in the primate ventral stream. While initially inspired by brain anatomy, over the past years, these ANNs have evolved from a simple eight-layer architecture in AlexNet to extremely deep and branching architectures, demonstrating increasingly better object categorization performance, yet bringing into question how brain-like they still are. In particular, typical deep models from the machine learning community are often hard to map onto the brain's anatomy due to their vast number of layers and missing biologically-important connections, such as recurrence. Here we demonstrate that better anatomical alignment to the brain and high performance on machine learning as well as neuroscience measures do not have to be in contradiction. We developed CORnet-S, a shallow ANN with four anatomically mapped areas and recurrent connectivity, guided by Brain-Score, a new large-scale composite of neural and behavioral benchmarks for quantifying the functional fidelity of models of the primate ventral visual stream. Despite being significantly shallower than most models, CORnet-S is the top model on Brain-Score and outperforms similarly compact models on ImageNet. Moreover, our extensive analyses of CORnet-S circuitry variants reveal that recurrence is the main predictive factor of both Brain-Score and ImageNet top-1 performance. Finally, we report that the temporal evolution of the CORnet-S "IT" neural population resembles the actual monkey IT population dynamics. Taken together, these results establish CORnet-S, a compact, recurrent ANN, as the current best model of the primate ventral visual stream. | ['Daniel L. K. Yamins', 'Jonathan Prescott-Roy', 'Kohitij Kar', 'Rishi Rajalingham', 'Najib J. Majaj', 'Jonas Kubilius', 'Daniel Bear', 'Pouya Bashivan', 'Kailyn Schmidt', 'Ha Hong', 'Elias B. Issa', 'Aran Nayebi', 'Martin Schrimpf', 'James J. DiCarlo'] | 2019-09-13 | brain-like-object-recognition-with-high-1 | http://papers.nips.cc/paper/9441-brain-like-object-recognition-with-high-performing-shallow-recurrent-anns | http://papers.nips.cc/paper/9441-brain-like-object-recognition-with-high-performing-shallow-recurrent-anns.pdf | neurips-2019-12 | ['object-categorization'] | ['computer-vision'] | [ 6.28587902e-02 -1.79204475e-02 1.31338621e-02 -1.26735613e-01
2.96930343e-01 -5.50433218e-01 7.42444575e-01 1.26939163e-01
-7.66248524e-01 2.64702320e-01 2.11298928e-01 -4.66045678e-01
-2.61592925e-01 -3.88075948e-01 -7.42682874e-01 -2.37162709e-01
-4.34494734e-01 3.40395212e-01 3.16492200e-01 -4.38036233e-01
2.68953562e-01 7.39587426e-01 -1.43183351e+00 2.40618214e-01
6.06710672e-01 1.16551924e+00 1.39150068e-01 6.22073174e-01
1.42192230e-01 6.10671341e-01 -1.55748546e-01 -4.35081333e-01
3.94047439e-01 -6.78647220e-01 -7.27907181e-01 -7.72794187e-01
8.48034620e-01 -2.87219509e-02 -4.62913364e-01 8.37717175e-01
3.81139427e-01 -3.00458372e-01 7.94139028e-01 -9.58914876e-01
-4.25204813e-01 6.30731046e-01 -3.43535572e-01 9.19899225e-01
-1.08341336e-01 7.17954159e-01 1.18055844e+00 -7.07503140e-01
7.72707999e-01 1.17132270e+00 1.15849042e+00 5.05337358e-01
-1.70953405e+00 -5.97431302e-01 7.88173676e-02 6.91148862e-02
-1.13558292e+00 -5.98471224e-01 2.17421040e-01 -8.39133680e-01
1.29716873e+00 -1.44458130e-01 1.52879131e+00 1.07742274e+00
6.56509399e-01 2.59942263e-01 1.16668928e+00 1.26951009e-01
2.87972033e-01 -3.95267069e-01 4.27114181e-02 8.00606966e-01
3.68155003e-01 3.47189665e-01 -6.05836809e-01 -7.63393641e-02
1.08836448e+00 -3.59888785e-02 -1.77242234e-01 -1.67701498e-01
-1.53020859e+00 5.77414989e-01 1.12568808e+00 4.27207321e-01
-4.95974630e-01 5.90921581e-01 4.95997965e-01 2.04107016e-01
-2.54835375e-02 7.14987874e-01 -3.77725929e-01 -3.70903797e-02
-1.13278842e+00 1.45037413e-01 5.71767449e-01 5.38540423e-01
4.22132015e-01 2.83394665e-01 1.54556753e-02 7.86743045e-01
2.94619769e-01 1.88637804e-02 6.90540850e-01 -1.15456903e+00
-4.48196903e-02 8.14861894e-01 -6.97848558e-01 -8.17693114e-01
-8.10200393e-01 -8.56823087e-01 -1.00970948e+00 5.96984088e-01
5.52177072e-01 2.07685769e-01 -8.41801822e-01 1.99713230e+00
-3.81926447e-01 -1.82436556e-01 -4.48210448e-01 9.76683736e-01
6.66408062e-01 2.19560131e-01 2.40830839e-01 1.01296090e-01
1.17384386e+00 -6.01295710e-01 2.31070861e-01 -8.20133746e-01
4.42570478e-01 -1.23734325e-01 8.24572146e-01 1.58943668e-01
-1.24900115e+00 -3.92206699e-01 -1.18849814e+00 -1.01445004e-01
-4.31571722e-01 -2.09415182e-01 9.08084393e-01 5.94888508e-01
-1.52042389e+00 7.97834039e-01 -6.55189574e-01 -7.32984543e-01
1.04723835e+00 3.78014237e-01 -7.18145490e-01 3.57637882e-01
-5.64774513e-01 1.28952420e+00 3.32835555e-01 1.30813345e-01
-1.32711470e+00 -9.54248488e-01 -3.57514352e-01 2.50906616e-01
-3.46650869e-01 -1.09305203e+00 1.09950888e+00 -1.02921069e+00
-9.80032623e-01 1.09920013e+00 -1.47079051e-01 -8.58382046e-01
3.62357646e-01 3.90519530e-01 3.33903357e-02 2.77976364e-01
-1.39514759e-01 1.18166447e+00 6.40208423e-01 -1.12659919e+00
-3.03053409e-01 -5.64250112e-01 -1.87234730e-01 -1.27138928e-01
-2.19778642e-02 -1.73690193e-03 7.51300007e-02 -4.71231163e-01
2.82623500e-01 -6.17835820e-01 -2.89264798e-01 5.75813830e-01
1.54667825e-01 -2.74278726e-02 1.01618275e-01 -5.32018840e-01
9.38896596e-01 -2.07769203e+00 2.09279120e-01 1.55051693e-01
9.11850572e-01 2.14665100e-01 -4.52907860e-01 1.91545188e-01
-4.10803139e-01 4.83951926e-01 -5.63910365e-01 -3.78367752e-02
-2.63181031e-01 9.96614769e-02 -1.35334939e-01 7.54848778e-01
3.54362786e-01 1.34868395e+00 -7.43983388e-01 -3.60103458e-01
-1.46253005e-01 2.22627416e-01 -8.96608889e-01 -2.89326549e-01
-9.30285454e-02 3.45873863e-01 2.57300496e-01 3.85503918e-01
3.28594446e-01 -1.81399748e-01 1.41789606e-02 -3.86584215e-02
-2.41657779e-01 1.02726601e-01 -2.48492241e-01 1.73102450e+00
-1.43343925e-01 1.10473931e+00 1.67676389e-01 -1.08940458e+00
5.27949452e-01 1.27890389e-02 2.39378184e-01 -1.10420048e+00
5.46205521e-01 3.58683050e-01 1.09332228e+00 -5.76061830e-02
1.11169061e-02 -2.17664868e-01 2.91274726e-01 3.44949305e-01
8.70787084e-01 -2.11683586e-02 1.81532368e-01 2.66269863e-01
1.33306217e+00 4.37138863e-02 2.67622024e-01 -6.46131217e-01
-5.01159653e-02 -1.55023821e-02 4.84421730e-01 7.61917591e-01
-3.95178467e-01 8.28181088e-01 9.75102007e-01 -3.82910937e-01
-1.39574099e+00 -1.17454290e+00 -4.24097955e-01 9.66327727e-01
-3.05421472e-01 -3.52804095e-01 -9.60056543e-01 -1.08793102e-01
5.36051206e-02 2.66708136e-01 -1.12139821e+00 -4.72980678e-01
-4.55705583e-01 -5.88926971e-01 1.16918063e+00 3.93469006e-01
3.21065664e-01 -1.18566024e+00 -1.03874063e+00 2.33858600e-01
3.65760595e-01 -8.33726287e-01 3.12678516e-02 4.93335038e-01
-9.65277791e-01 -1.22312677e+00 -7.34254241e-01 -8.50222647e-01
5.55557251e-01 -3.51026878e-02 1.17346263e+00 3.93190801e-01
-7.19753325e-01 1.61196470e-01 8.38464051e-02 -3.50069046e-01
-2.86667287e-01 1.88245311e-01 1.82163954e-01 -2.79846549e-01
3.17102596e-02 -1.00733960e+00 -9.65348423e-01 2.93272793e-01
-8.51940095e-01 -7.85801485e-02 8.29649150e-01 8.71333063e-01
3.49007159e-01 -4.43538219e-01 6.76930189e-01 -4.36915427e-01
5.83389163e-01 -7.24163532e-01 -3.56513411e-01 -6.88292012e-02
-4.31142479e-01 1.40411295e-02 7.09377706e-01 -3.37032527e-01
-4.22362149e-01 -3.08325052e-01 -3.80647689e-01 -1.90579951e-01
-2.24498749e-01 5.39840341e-01 4.94960427e-01 -4.96167481e-01
1.04178393e+00 3.67928892e-01 1.15986057e-01 -3.37020069e-01
7.11916313e-02 -5.72604910e-02 6.59173548e-01 -3.52816790e-01
6.12306774e-01 4.53432262e-01 5.04016876e-02 -1.20575809e+00
-2.95795739e-01 -1.17843442e-01 -8.66539419e-01 -2.87376642e-01
9.15588200e-01 -5.92460752e-01 -1.14557600e+00 4.99492198e-01
-1.32273662e+00 -6.16453886e-01 -2.82203048e-01 3.73801291e-01
-6.45881355e-01 2.60527302e-02 -5.24679124e-01 -5.04028320e-01
-2.05988556e-01 -9.09684658e-01 5.34448802e-01 1.36821270e-01
-4.86016929e-01 -9.21598256e-01 2.16852650e-01 -1.08210616e-01
7.97325552e-01 1.53898492e-01 1.44721639e+00 -5.14409959e-01
-3.20234239e-01 5.22965491e-02 -4.99035180e-01 2.22889632e-01
-5.67251146e-01 -3.21669355e-02 -9.50562477e-01 -1.82347566e-01
-1.34664446e-01 -3.44551176e-01 1.37566042e+00 5.16913712e-01
9.84786630e-01 -3.20712030e-01 -2.17322573e-01 1.06711257e+00
1.50095928e+00 3.10900752e-02 6.13650918e-01 2.67656893e-01
1.88841179e-01 5.89617908e-01 -5.19350052e-01 -3.69942933e-02
8.86179730e-02 2.21228749e-01 7.45280504e-01 -2.65783556e-02
-3.75726253e-01 -3.83972108e-01 3.42826426e-01 7.46345758e-01
-3.29529315e-01 1.75089836e-01 -1.27310503e+00 6.84247136e-01
-1.51008594e+00 -1.13363659e+00 2.78367903e-02 2.07182264e+00
6.70461118e-01 3.35204095e-01 3.55440706e-01 -1.87304839e-01
2.50899523e-01 3.36463638e-02 -8.57592642e-01 -5.51689565e-01
-5.97457707e-01 4.38320577e-01 3.77355486e-01 5.01179434e-02
-4.58735883e-01 8.28628361e-01 7.83775854e+00 4.19353992e-01
-1.20129228e+00 2.08675474e-01 6.05196834e-01 -2.85464734e-01
-9.94457677e-02 6.68706894e-02 -5.59503913e-01 3.57884884e-01
1.16505957e+00 -1.78053200e-01 9.05649126e-01 4.35695142e-01
-3.75293978e-02 -1.06053278e-01 -1.46466768e+00 9.08515990e-01
5.69016598e-02 -1.62839139e+00 1.68103755e-01 3.97710860e-01
4.22439903e-01 8.68288517e-01 2.61453748e-01 1.83377028e-01
9.78982747e-02 -1.72117209e+00 1.17330229e+00 5.75417817e-01
8.09892952e-01 -2.41151825e-01 3.20828825e-01 4.60154563e-01
-9.20630336e-01 -3.62031788e-01 -5.32979906e-01 -1.40242755e-01
-1.38503984e-01 -2.74750646e-02 -3.05582881e-01 -3.20700943e-01
9.81079876e-01 7.30815291e-01 -1.10286009e+00 1.63920033e+00
3.73593718e-01 6.82157397e-01 -4.33964401e-01 -1.39876515e-01
2.75861055e-01 1.07497670e-01 6.62735403e-01 1.30612719e+00
1.64015889e-01 1.03866450e-01 -5.18413246e-01 1.42444253e+00
-2.34447509e-01 -4.48191427e-02 -8.42384458e-01 -3.54831338e-01
1.84608966e-01 1.31820798e+00 -8.82047772e-01 -1.23030618e-02
-2.39784569e-01 5.99631727e-01 6.20156646e-01 3.41230035e-01
-5.33558011e-01 -1.14490554e-01 8.68129969e-01 3.47862661e-01
7.97135159e-02 -5.37512004e-01 -6.03438675e-01 -8.64676714e-01
-1.86743677e-01 -6.59169197e-01 -1.35233149e-01 -8.51483345e-01
-1.18625593e+00 8.59154344e-01 -4.14427251e-01 -8.65583897e-01
7.13599697e-02 -1.04326630e+00 -6.16807044e-01 8.58205855e-01
-1.18938267e+00 -8.51084650e-01 -1.99215978e-01 5.94684005e-01
3.13378364e-01 -4.19548362e-01 7.59272218e-01 4.09594662e-02
-4.21269953e-01 5.82257748e-01 -6.20526914e-03 1.62291989e-01
1.29740670e-01 -9.65898514e-01 6.79009914e-01 7.20899224e-01
3.60051781e-01 1.01239944e+00 4.54485446e-01 -3.35474670e-01
-1.21101213e+00 -6.56831443e-01 3.01689446e-01 -3.57720643e-01
8.57373178e-01 -5.65185249e-01 -7.87711024e-01 6.67362213e-01
4.46570158e-01 1.00699238e-01 4.08030927e-01 -4.90717068e-02
-8.02939951e-01 -1.24897443e-01 -9.66890872e-01 6.41800523e-01
1.49600899e+00 -5.64558387e-01 -7.34349966e-01 -1.46913277e-02
3.15331161e-01 2.52833337e-01 -9.03642356e-01 1.80014193e-01
9.56407070e-01 -1.21914971e+00 8.54204834e-01 -7.62825012e-01
7.00128436e-01 -1.46442965e-01 2.90351380e-02 -1.57848918e+00
-7.98615336e-01 -3.17523777e-01 2.89084673e-01 6.68453693e-01
6.18446469e-01 -9.09515381e-01 6.23082042e-01 2.14538023e-01
-3.45903814e-01 -7.60421515e-01 -1.37095201e+00 -6.84044003e-01
4.96589571e-01 -3.21403801e-01 6.46213740e-02 6.06548309e-01
7.04102069e-02 4.68124092e-01 2.60742366e-01 -5.12240469e-01
6.96997583e-01 -4.80251342e-01 2.25590825e-01 -1.65114737e+00
-1.25725865e-01 -1.47121775e+00 -7.87706375e-01 -6.92881763e-01
7.44821504e-02 -1.47382379e+00 -1.09967180e-01 -1.63632238e+00
3.38653952e-01 -3.31385821e-01 -2.47775584e-01 6.64900422e-01
4.50635433e-01 4.96370882e-01 2.35084534e-01 3.29328448e-01
-2.42411032e-01 3.32464039e-01 1.03593981e+00 -8.58766492e-03
2.93123186e-01 -4.95281905e-01 -8.27746570e-01 9.34693873e-01
7.87496328e-01 -3.60525817e-01 -1.97015479e-02 -6.57534301e-01
5.52485704e-01 -2.35980541e-01 7.91780293e-01 -1.47138536e+00
4.40515995e-01 1.86537221e-01 8.72367740e-01 5.45769446e-02
3.28582466e-01 -8.48043144e-01 -2.74742171e-02 8.36997569e-01
-5.28821409e-01 2.98820794e-01 4.26861346e-01 3.81459713e-01
6.94918782e-02 4.76908823e-03 1.22118235e+00 -4.62977290e-01
-7.21007645e-01 5.51308513e-01 -7.83213794e-01 2.13732317e-01
6.95431054e-01 -7.04952538e-01 -8.06850851e-01 7.29031935e-02
-5.94469249e-01 -3.75488251e-02 4.95684505e-01 3.92254651e-01
7.15814769e-01 -1.00069475e+00 -8.02337468e-01 3.58883411e-01
1.09519877e-01 -4.83112782e-01 1.42455280e-01 1.04312801e+00
-8.83616865e-01 4.83174205e-01 -9.27098632e-01 -6.76075339e-01
-6.31959438e-01 2.82683462e-01 8.44358265e-01 3.56025279e-01
-4.95929211e-01 8.77450526e-01 3.85848552e-01 -3.19817573e-01
-2.12776549e-02 -3.46299320e-01 -2.93550968e-01 2.16709629e-01
1.62789434e-01 9.06347632e-02 -3.48303765e-02 -7.83442974e-01
-3.30625772e-01 4.45660174e-01 1.43164456e-01 -5.45608103e-02
1.58038151e+00 1.62901342e-01 -6.11306429e-01 5.13195038e-01
1.03166747e+00 -4.60344315e-01 -1.04172885e+00 5.88008463e-02
6.99769662e-05 -1.08122431e-01 -1.62792057e-01 -1.01553440e+00
-1.09217322e+00 1.30899727e+00 6.10336185e-01 5.38641773e-02
7.28937626e-01 4.43954021e-02 4.42360967e-01 3.32296371e-01
3.88860196e-01 -8.32198620e-01 6.37785271e-02 7.59596050e-01
1.20271254e+00 -7.84281075e-01 -2.17975810e-01 4.09088701e-01
-3.24490666e-01 9.27464008e-01 8.79954815e-01 -4.51310903e-01
6.53689086e-01 9.88145098e-02 -2.62730718e-01 -2.91747868e-01
-1.01560521e+00 -1.74114242e-01 1.10761888e-01 6.59081757e-01
4.32170331e-01 -2.02090293e-01 -6.00919090e-02 5.56564331e-01
-6.72463655e-01 -2.01800942e-01 2.94216454e-01 3.87245357e-01
-7.11321235e-01 -4.10016686e-01 8.68135095e-02 8.96768391e-01
-3.34273636e-01 -3.34155709e-01 -5.91225803e-01 6.58963382e-01
1.72490656e-01 3.03087980e-01 2.08867237e-01 -4.01286244e-01
2.50135779e-01 1.10284679e-01 7.58123577e-01 -4.99413341e-01
-1.11170781e+00 -3.25706303e-01 -2.81186253e-01 -5.83946526e-01
5.73357046e-02 -4.52859104e-01 -1.00840461e+00 -5.71711838e-01
1.39992252e-01 -2.61755466e-01 7.30488360e-01 7.17035353e-01
3.82086843e-01 4.36750889e-01 6.20817356e-02 -1.19484115e+00
-2.73149967e-01 -9.73448575e-01 -5.15814185e-01 1.14420220e-01
4.67940718e-01 -4.03515637e-01 -3.00201833e-01 -2.82362521e-01] | [9.559746742248535, 2.5049867630004883] |
2004be93-ae6e-49b3-80b7-a15f9185dbf3 | detecting-deception-in-political-debates | 1910.01990 | null | https://arxiv.org/abs/1910.01990v1 | https://arxiv.org/pdf/1910.01990v1.pdf | Detecting Deception in Political Debates Using Acoustic and Textual Features | We present work on deception detection, where, given a spoken claim, we aim to predict its factuality. While previous work in the speech community has relied on recordings from staged setups where people were asked to tell the truth or to lie and their statements were recorded, here we use real-world political debates. Thanks to the efforts of fact-checking organizations, it is possible to obtain annotations for statements in the context of a political discourse as true, half-true, or false. Starting with such data from the CLEF-2018 CheckThat! Lab, which was limited to text, we performed alignment to the corresponding videos, thus producing a multimodal dataset. We further developed a multimodal deep-learning architecture for the task of deception detection, which yielded sizable improvements over the state of the art for the CLEF-2018 Lab task 2. Our experiments show that the use of the acoustic signal consistently helped to improve the performance compared to using textual and metadata features only, based on several different evaluation measures. We release the new dataset to the research community, hoping to help advance the overall field of multimodal deception detection. | ['Daniel Kopev', 'Preslav Nakov', 'Ivan Koychev', 'Ahmed Ali'] | 2019-10-04 | null | null | null | null | ['deception-detection'] | ['miscellaneous'] | [ 2.08918959e-01 1.47165880e-01 2.10804343e-01 -5.34139276e-01
-1.48932779e+00 -8.27036858e-01 1.15907598e+00 1.76093400e-01
-4.19304192e-01 5.57962835e-01 7.90006638e-01 -3.38076979e-01
3.83095920e-01 -1.58948079e-01 -6.12981558e-01 -4.65273499e-01
1.79242343e-01 3.45226675e-01 -1.06737442e-01 -2.18682379e-01
6.12961471e-01 3.35577689e-02 -1.34553862e+00 1.04200029e+00
4.66499865e-01 1.08646262e+00 -5.10773599e-01 8.68565321e-01
1.10302813e-01 1.27903247e+00 -1.00324070e+00 -1.16470599e+00
-2.06230402e-01 -4.98069435e-01 -1.30506051e+00 9.88316238e-02
9.94741023e-01 -4.25884128e-01 -4.28104192e-01 9.81290221e-01
3.88430953e-01 -2.24039882e-01 6.33378446e-01 -1.08163416e+00
-6.10928118e-01 5.87704420e-01 -1.58520266e-01 1.92694932e-01
8.20269883e-01 1.02154173e-01 1.02724040e+00 -8.79096031e-01
7.83321857e-01 1.21336877e+00 6.87923312e-01 5.70095181e-01
-9.42650855e-01 -4.24545497e-01 -2.76655197e-01 5.66017091e-01
-1.05998898e+00 -1.20235145e+00 7.73424387e-01 -5.77462673e-01
8.31517279e-01 4.10168141e-01 3.41149598e-01 1.90348458e+00
-2.02621862e-01 1.02869701e+00 1.22405982e+00 -5.10370076e-01
6.83014188e-03 6.44587055e-02 1.15147203e-01 7.39945173e-01
-3.57381195e-01 -2.39668354e-01 -9.57980812e-01 -3.86044145e-01
-7.91568682e-02 -7.60252714e-01 -6.25518978e-01 -1.62805170e-01
-1.53250241e+00 9.16035950e-01 1.09268665e-01 4.70896304e-01
-1.54467523e-01 4.08260077e-02 8.80816519e-01 6.01523995e-01
5.85551918e-01 3.81191075e-01 -1.23592289e-02 -7.80053735e-01
-1.23189819e+00 3.09822053e-01 1.12704015e+00 3.47739875e-01
4.45870571e-02 -2.00850949e-01 -2.39886999e-01 8.87665451e-01
7.64328018e-02 2.49470428e-01 3.64130288e-01 -1.05691683e+00
1.00664747e+00 3.41171384e-01 5.37027776e-01 -1.36752677e+00
-3.63524407e-01 -3.54639404e-02 -4.98562932e-01 -1.41815811e-01
8.36397767e-01 -1.59520328e-01 -6.59986913e-01 1.44365549e+00
-1.62772667e-02 -2.34237790e-01 2.02882692e-01 1.16623592e+00
8.66918206e-01 5.54197192e-01 -3.60261947e-01 -2.76648998e-01
1.38569891e+00 -7.36768484e-01 -9.02425587e-01 -1.00107171e-01
8.02037716e-01 -9.05617774e-01 9.80336189e-01 7.28829265e-01
-9.87155378e-01 -3.62986214e-02 -1.00459540e+00 -2.17097044e-01
-1.73862681e-01 2.53787398e-01 1.72905654e-01 7.89370418e-01
-7.72916853e-01 3.05763870e-01 -6.27750397e-01 -4.24066901e-01
5.62756479e-01 -3.43912095e-01 -6.78515196e-01 -2.06256106e-01
-1.20499361e+00 1.16376376e+00 1.20841406e-01 4.56744403e-01
-1.01140344e+00 -9.34106782e-02 -9.70066845e-01 -1.67443305e-01
5.08626819e-01 -3.05008382e-01 1.24141681e+00 -1.09488571e+00
-1.40750086e+00 1.44442308e+00 -5.56199588e-02 -6.24511480e-01
8.74500215e-01 -1.78800389e-01 -6.36049509e-01 4.20994848e-01
-9.58851427e-02 3.82240057e-01 1.05874455e+00 -1.19884861e+00
-2.75184005e-01 -3.60844433e-01 2.23725870e-01 -4.65032160e-02
-2.96222717e-01 4.40905035e-01 2.74063777e-02 -3.84687066e-01
-1.86672494e-01 -9.00191844e-01 6.05762005e-01 -4.83436823e-01
-6.77863061e-01 -2.66876221e-01 8.76882017e-01 -1.31033969e+00
1.11382139e+00 -2.30364227e+00 2.54993796e-01 -2.57147789e-01
2.65671492e-01 2.58916974e-01 -1.96721125e-03 8.84870946e-01
-5.17374184e-03 2.84845054e-01 -3.46339136e-01 -6.30648494e-01
3.05139929e-01 -2.82318871e-02 -5.78307927e-01 8.99426877e-01
1.80331320e-01 7.38362432e-01 -7.36403644e-01 -3.98980439e-01
-1.42160669e-01 3.12605768e-01 -2.53886431e-01 1.14346363e-01
-1.10555626e-01 3.47545117e-01 -1.92768835e-02 7.08516002e-01
4.42807317e-01 7.84630179e-02 1.36016294e-01 -3.34901154e-01
-9.37733278e-02 6.61399961e-01 -5.74195683e-01 1.64633739e+00
-3.29556048e-01 1.27690697e+00 5.72555125e-01 -1.19428682e+00
6.19140625e-01 6.02166831e-01 -2.71607302e-02 -4.58586127e-01
2.53360718e-01 3.68425518e-01 -6.36637807e-02 -9.28656518e-01
6.64867342e-01 -2.04975963e-01 -4.34799343e-01 3.93596590e-01
1.92183591e-02 -3.17340702e-01 6.71266541e-02 4.92727846e-01
1.14083016e+00 -1.28552034e-01 -6.08792007e-02 3.83697934e-02
5.38467705e-01 1.57281339e-01 8.92787706e-03 6.87345684e-01
-5.33584893e-01 8.23136747e-01 1.11089826e+00 -4.13283974e-01
-8.65365148e-01 -6.51263297e-01 -3.83629948e-02 7.90809035e-01
-2.93801636e-01 -5.11160195e-01 -8.07074666e-01 -1.07299483e+00
-2.60045975e-01 1.00701785e+00 -6.53049469e-01 3.67202014e-02
-5.90998411e-01 -4.34284180e-01 9.29446757e-01 -1.08221307e-01
5.91196358e-01 -8.27882171e-01 -5.91094792e-01 -1.50723025e-01
-7.70938575e-01 -1.33205724e+00 -4.97338355e-01 -2.13835686e-01
-9.99717191e-02 -1.39536738e+00 -6.02169335e-01 -4.63207603e-01
7.44978487e-02 -1.15023501e-01 9.73940074e-01 1.62730500e-01
5.03662191e-02 6.42656028e-01 -4.97420043e-01 -1.99130177e-01
-8.16030920e-01 -1.76826775e-01 -1.64266720e-01 3.64278972e-01
3.48368585e-02 -1.80401996e-01 -2.86128819e-01 2.59864181e-02
-8.79911363e-01 -2.16342196e-01 1.44013569e-01 1.08106291e+00
-4.16152507e-01 -4.06489819e-01 4.76403505e-01 -6.01187825e-01
9.33799803e-01 -3.39512497e-01 -1.80489376e-01 1.15344331e-01
-5.02429977e-02 -1.59224451e-01 3.98876667e-01 -1.05267346e-01
-9.09118950e-01 -4.34230268e-01 -3.51658493e-01 -3.11650425e-01
-3.84330392e-01 6.02670372e-01 6.44707680e-02 6.97671100e-02
5.18855870e-01 2.52431154e-01 2.20439397e-02 -2.14852542e-01
4.05804694e-01 1.10329008e+00 8.45350325e-01 -3.52679163e-01
3.80449891e-01 3.96137953e-01 -3.15582812e-01 -8.44685972e-01
-1.20180535e+00 -3.84770721e-01 -5.37356436e-01 -2.88749993e-01
8.28001499e-01 -6.99240506e-01 -9.98047590e-01 5.74992418e-01
-1.52987456e+00 3.39844152e-02 3.81237328e-01 1.63740337e-01
-5.55544078e-01 9.51455295e-01 -7.20000923e-01 -1.11406553e+00
-4.59700562e-02 -1.03218186e+00 1.15272415e+00 -5.74560702e-01
-3.80740851e-01 -9.00743783e-01 -1.54419169e-02 1.30969918e+00
3.23470414e-01 4.96019423e-01 7.51455009e-01 -1.03505111e+00
-1.19177112e-02 -3.95188808e-01 -8.89269188e-02 4.75633591e-01
-2.89663762e-01 -7.24533722e-02 -1.10238063e+00 -1.48211330e-01
2.57964343e-01 -1.03112376e+00 1.12760746e+00 -3.14646572e-01
8.69416058e-01 -8.20706606e-01 -2.10706424e-02 3.72038409e-02
7.73902118e-01 -4.07512695e-01 6.69537842e-01 3.34527284e-01
4.78477776e-01 8.32023263e-01 3.28475684e-01 2.30013832e-01
6.33416593e-01 9.93400037e-01 6.59967244e-01 3.75709891e-01
-8.88642669e-02 -1.92934811e-01 5.95669866e-01 5.66686153e-01
1.91822246e-01 -5.56955516e-01 -1.13525140e+00 8.39099407e-01
-1.78120315e+00 -1.35441363e+00 -4.31750178e-01 2.07811594e+00
8.16281557e-01 2.06419658e-02 2.43721604e-01 2.15901062e-01
5.06515503e-01 4.46832001e-01 1.90822631e-01 -7.79000461e-01
-2.50459939e-01 -4.99145955e-01 -1.72628537e-02 6.13788962e-01
-1.33063090e+00 6.48367226e-01 6.20672321e+00 7.45766640e-01
-1.17142498e+00 2.96457469e-01 6.94745898e-01 -2.83606857e-01
-1.20826796e-01 -2.27426887e-01 -2.37237588e-01 6.46740854e-01
1.24547291e+00 3.40878367e-01 3.85627836e-01 2.62368679e-01
2.26555556e-01 -3.03224742e-01 -1.26897502e+00 1.12244260e+00
7.47879446e-01 -1.26534808e+00 -1.95811883e-01 -5.51302247e-02
4.60379004e-01 1.01037055e-01 -5.78136370e-03 2.76161551e-01
-1.88544486e-02 -1.24447477e+00 1.14061999e+00 5.17223299e-01
5.60397029e-01 -5.50831556e-01 1.02807581e+00 6.79531753e-01
-9.26221758e-02 -6.92025945e-02 1.85991243e-01 -9.91959497e-02
4.64695036e-01 5.59759617e-01 -1.02802217e+00 5.81887424e-01
4.93295640e-01 4.93104130e-01 -4.69605416e-01 5.13361394e-01
-2.69169062e-01 9.93633211e-01 -8.05671588e-02 -2.43762229e-02
3.97397876e-01 1.66290075e-01 8.40225160e-01 1.50675535e+00
1.38165534e-01 -1.53427511e-01 -1.48717668e-02 7.13026762e-01
-4.14561510e-01 -1.15668654e-01 -5.11470973e-01 -3.89213204e-01
1.11197665e-01 1.30499184e+00 -1.92921441e-02 -2.66752630e-01
-3.56423497e-01 1.36531186e+00 4.72395927e-01 8.49299207e-02
-7.66224682e-01 -2.13191643e-01 2.10553080e-01 2.62452159e-02
3.10158372e-01 -3.94620299e-01 -2.14169500e-03 -1.50340891e+00
2.75747538e-01 -1.32142401e+00 4.49757904e-01 -8.87129009e-01
-1.25392699e+00 5.86710632e-01 -2.11161539e-01 -7.48485625e-01
-4.32420075e-01 -5.16836047e-01 -4.56025183e-01 6.28381550e-01
-1.29270160e+00 -1.36629426e+00 -3.41440327e-02 3.24232817e-01
4.71782476e-01 -1.01329878e-01 8.25962961e-01 1.06264658e-01
-4.65565294e-01 4.41503376e-01 -2.38848358e-01 4.05555099e-01
9.79122460e-01 -9.94295657e-01 1.93408169e-02 8.59468997e-01
5.55128694e-01 4.53914583e-01 1.06270373e+00 -3.58717144e-01
-1.38189721e+00 -3.79562110e-01 1.32258403e+00 -1.02926099e+00
1.17286372e+00 -6.25902176e-01 -8.89091551e-01 5.71240246e-01
6.94379926e-01 -1.81064948e-01 4.71489280e-01 1.48117885e-01
-6.67979479e-01 3.07635009e-01 -1.16352761e+00 2.21648842e-01
7.70204186e-01 -9.27688420e-01 -9.33468878e-01 5.50190866e-01
1.77109852e-01 -5.62759995e-01 -5.44142306e-01 -4.79138363e-03
5.53650975e-01 -1.19671524e+00 5.25161684e-01 -7.75863349e-01
8.97886693e-01 -3.93802263e-02 -2.76646346e-01 -1.39211094e+00
2.20138744e-01 -5.87234557e-01 -2.15931088e-01 1.42968333e+00
4.14645404e-01 -3.25044692e-01 3.88706923e-01 4.78428781e-01
-2.38088042e-01 -3.20480227e-01 -1.53983843e+00 -3.44210953e-01
7.06766397e-02 -4.83878940e-01 -5.21495054e-03 1.12091315e+00
4.89729464e-01 4.86832976e-01 -7.58646667e-01 9.40517411e-02
2.53003538e-01 8.01994577e-02 7.26701081e-01 -7.83042431e-01
-7.05486611e-02 -3.64529490e-01 -4.45478827e-01 -9.81245041e-01
5.05262196e-01 -7.18692124e-01 1.27400368e-01 -1.29667854e+00
2.73490220e-01 2.56205440e-01 2.77038127e-01 5.36086857e-01
-5.98063655e-02 5.33313692e-01 2.90806860e-01 3.25134397e-01
-8.63571405e-01 5.36337018e-01 8.21370721e-01 -4.57831115e-01
4.76291120e-01 -1.97072595e-01 -5.49948037e-01 6.43789768e-01
4.47291046e-01 -2.58049071e-01 2.62829095e-01 -6.16759002e-01
4.37759012e-01 2.60092556e-01 9.50252414e-01 -6.22192442e-01
1.09772645e-01 2.67833352e-01 6.94780722e-02 -4.50909525e-01
7.42629111e-01 -6.05618954e-01 -3.45288604e-01 1.92910284e-01
-5.60673654e-01 -2.67345995e-01 3.09829414e-02 5.56801438e-01
-5.28275728e-01 -3.75474483e-01 5.86005092e-01 -7.09377825e-02
-3.62023532e-01 -4.91841584e-01 -5.34393072e-01 9.15748179e-02
6.45445466e-01 1.46246001e-01 -7.52516568e-01 -1.05054688e+00
-6.15373492e-01 5.02444170e-02 5.08879840e-01 3.46507728e-01
5.42993069e-01 -1.11828697e+00 -1.17475712e+00 -3.62596840e-01
2.11683571e-01 -7.62100816e-01 8.87842327e-02 1.30806804e+00
-2.96086788e-01 5.59213996e-01 1.97423309e-01 -4.61104095e-01
-1.48695397e+00 3.87744218e-01 3.67148548e-01 -2.32306309e-02
-3.54528397e-01 7.25270033e-01 -4.05800939e-01 -3.50932211e-01
9.66595188e-02 9.78469197e-03 -1.72428891e-01 3.92934322e-01
6.55472398e-01 3.60880315e-01 3.65392298e-01 -1.01170862e+00
-4.65756595e-01 -8.29191282e-02 1.45426067e-02 -5.10873079e-01
1.23334837e+00 -2.98972487e-01 -1.95148528e-01 6.36343718e-01
1.28956568e+00 6.08798206e-01 -8.34637403e-01 1.56651393e-01
6.25253394e-02 -6.16207123e-01 8.51009861e-02 -1.22055709e+00
-5.82171142e-01 8.97216260e-01 1.10542387e-01 7.91616559e-01
5.88895619e-01 2.47210816e-01 6.87745869e-01 4.20182526e-01
2.17797935e-01 -9.97632027e-01 1.71431139e-01 6.74895346e-01
1.51051605e+00 -1.56130230e+00 -1.91161722e-01 -5.70389107e-02
-9.48841512e-01 1.36643505e+00 -1.97396744e-02 2.13144332e-01
-1.05682045e-01 -1.17516495e-01 2.21652448e-01 -5.61091721e-01
-7.21965075e-01 2.77106911e-01 3.60835135e-01 4.25845414e-01
5.89765906e-01 -4.96277167e-03 -3.10754806e-01 5.21134913e-01
-5.00851035e-01 -2.49316826e-01 8.33004296e-01 5.70892930e-01
-2.72985160e-01 -7.97661901e-01 -5.86805403e-01 2.46377602e-01
-8.51557910e-01 -9.28871483e-02 -1.09256434e+00 7.19675601e-01
-1.43111169e-01 1.48467183e+00 -1.65505007e-01 -1.88764453e-01
2.79253513e-01 3.87898892e-01 5.92368543e-01 -2.94639945e-01
-6.38294876e-01 -3.46081138e-01 1.11380255e+00 -5.82241595e-01
-5.85962534e-01 -9.99989331e-01 -5.63087344e-01 -3.96400034e-01
-1.74452379e-01 2.39233121e-01 7.92484820e-01 1.25395179e+00
1.65538386e-01 6.05076877e-03 3.67568105e-01 -7.41422832e-01
-6.91556334e-01 -1.26559162e+00 -3.83609422e-02 6.30753279e-01
7.80310869e-01 -3.34446371e-01 -6.62123859e-01 2.09810480e-01] | [8.264176368713379, 10.383140563964844] |
61da76a9-ef8d-4cd9-ae66-a08c11470de4 | enhanced-biologically-inspired-model-for | 1710.10188 | null | http://arxiv.org/abs/1710.10188v1 | http://arxiv.org/pdf/1710.10188v1.pdf | Enhanced Biologically Inspired Model for Image Recognition Based on a Novel Patch Selection Method with Moment | Biologically inspired model (BIM) for image recognition is a robust
computational architecture, which has attracted widespread attention. BIM can
be described as a four-layer structure based on the mechanisms of the visual
cortex. Although the performance of BIM for image recognition is robust, it
takes the randomly selected ways for the patch selection, which is sightless,
and results in heavy computing burden. To address this issue, we propose a
novel patch selection method with oriented Gaussian-Hermite moment (PSGHM), and
we enhanced the BIM based on the proposed PSGHM, named as PBIM. In contrast to
the conventional BIM which adopts the random method to select patches within
the feature representation layers processed by multi-scale Gabor filter banks,
the proposed PBIM takes the PSGHM way to extract a small number of
representation features while offering promising distinctiveness. To show the
effectiveness of the proposed PBIM, experimental studies on object
categorization are conducted on the CalTech05, TU Darmstadt (TUD), and GRAZ01
databases. Experimental results demonstrate that the performance of PBIM is a
significant improvement on that of the conventional BIM. | ['Li-Hao Jia', 'Yan-Feng Lu', 'Yi Li', 'Hong Qaio'] | 2017-10-27 | null | null | null | null | ['object-categorization'] | ['computer-vision'] | [ 2.67722130e-01 -3.28678370e-01 2.08209023e-01 -1.71404570e-01
-2.25447059e-01 2.14682236e-01 5.68834364e-01 -2.34941557e-01
-4.77116227e-01 4.70559776e-01 -2.33864740e-01 -5.37375221e-03
-4.32452649e-01 -7.92364895e-01 -5.84611297e-01 -1.05174661e+00
-6.10332191e-02 -3.28538328e-01 5.18376946e-01 -1.50122374e-01
7.38063633e-01 6.93131864e-01 -2.07860661e+00 5.58032930e-01
5.72861671e-01 1.34614301e+00 4.94057566e-01 1.79050714e-01
-1.11060672e-01 5.73752463e-01 -4.35460418e-01 1.57363132e-01
1.23533837e-01 -3.75062793e-01 -6.89854741e-01 1.88654840e-01
1.34344429e-01 5.35105588e-03 -1.52527973e-01 1.09057152e+00
6.25771463e-01 1.76885456e-01 9.10786450e-01 -1.02248991e+00
-7.27125347e-01 3.16182226e-01 -5.64834654e-01 3.36077839e-01
4.23138961e-02 -4.06570226e-01 4.36938703e-01 -1.24835479e+00
2.80211002e-01 1.37631226e+00 5.33890724e-01 3.76425654e-01
-1.00924969e+00 -6.45754337e-01 2.97364622e-01 5.38986981e-01
-1.79165637e+00 -2.70476639e-01 7.95194626e-01 -4.03598934e-01
9.64769900e-01 3.05990338e-01 5.87324679e-01 4.87936765e-01
4.40905035e-01 7.40518630e-01 1.32713866e+00 -7.32511938e-01
3.24973345e-01 1.43869370e-01 4.65449750e-01 8.33144724e-01
2.77468890e-01 9.17931050e-02 -5.21117032e-01 -1.94451362e-01
1.01651752e+00 4.44767773e-02 -2.30682194e-01 -2.65342891e-01
-1.04871988e+00 5.92386007e-01 7.94059396e-01 6.07663333e-01
-4.62058336e-01 -2.25118436e-02 1.66617870e-01 7.31778219e-02
1.75575227e-01 4.94980738e-02 7.00485110e-02 4.72956002e-01
-7.55571187e-01 8.40450078e-02 4.07673895e-01 6.98246479e-01
7.67204225e-01 2.23779067e-01 -3.08919400e-02 1.00104177e+00
5.63083470e-01 4.29280758e-01 8.40817809e-01 -4.72139686e-01
-7.33309016e-02 7.74440944e-01 -4.76510637e-02 -1.38267553e+00
-4.33340639e-01 -3.48782629e-01 -1.06310487e+00 3.53867948e-01
1.70767814e-01 3.24286968e-01 -1.04295170e+00 1.20233536e+00
1.92146525e-01 2.64009200e-02 -4.73671816e-02 1.16235995e+00
1.15833044e+00 8.20449114e-01 8.28691944e-02 -3.48592997e-02
1.40263760e+00 -7.41272867e-01 -3.52496892e-01 4.29601930e-02
2.65925318e-01 -7.22077310e-01 6.45424604e-01 4.95366603e-01
-7.41757274e-01 -9.44829762e-01 -1.25015080e+00 4.71617103e-01
-5.92611551e-01 3.98910582e-01 7.07650781e-01 5.31572461e-01
-9.96541262e-01 3.92683923e-01 -6.46579981e-01 -5.33876181e-01
3.98913413e-01 3.35182607e-01 -5.94356596e-01 -8.08258820e-03
-8.90207827e-01 7.68752038e-01 6.17268920e-01 3.41240883e-01
-4.92599815e-01 -1.40865088e-01 -6.13318086e-01 2.06804261e-01
-2.47361675e-01 -3.15203249e-01 6.97521389e-01 -1.07613599e+00
-1.49489164e+00 6.82287097e-01 -3.41996282e-01 -4.05445784e-01
-2.05017198e-02 2.55322099e-01 -2.65950412e-01 5.54948449e-01
-2.28555799e-01 7.90747166e-01 1.01151049e+00 -1.05483878e+00
-6.38847709e-01 -3.51820081e-01 -2.23588347e-01 5.64787425e-02
-5.18158436e-01 1.49607360e-01 -1.12236649e-01 -6.51734591e-01
7.49809742e-01 -6.22994304e-01 -1.64380744e-01 -1.14548676e-01
-5.90558797e-02 -3.99446070e-01 6.94533348e-01 -5.09063482e-01
9.15557981e-01 -2.41727042e+00 -2.01465979e-01 2.19479203e-01
-2.17917591e-01 3.99466783e-01 -1.11952521e-01 3.50423425e-01
-9.60838050e-02 -1.38941510e-02 -1.71466529e-01 2.60209590e-01
-1.87599033e-01 -7.21746981e-02 -1.81643710e-01 5.06268620e-01
3.82041395e-01 5.51791489e-01 -3.32181573e-01 -5.41311502e-01
2.54706472e-01 4.16655838e-01 -4.03172374e-01 -1.09191120e-01
3.04859251e-01 2.88243413e-01 -6.70960426e-01 7.64094114e-01
9.68053877e-01 -1.05682895e-01 -1.52814537e-01 -3.11407715e-01
-1.99404910e-01 -9.95627716e-02 -1.16954708e+00 1.29431963e+00
-1.52542889e-01 3.79983217e-01 -9.18964446e-02 -1.38616168e+00
1.44480181e+00 1.95424750e-01 2.53150880e-01 -7.98570633e-01
3.30537856e-01 1.82213500e-01 2.47993007e-01 -3.65571350e-01
1.02734584e-02 -7.39435926e-02 3.00302088e-01 -3.23976241e-02
1.07675418e-01 3.71524781e-01 -6.94956258e-03 -2.36425415e-01
7.89760411e-01 1.10637300e-01 3.85560423e-01 -7.42810130e-01
9.23668444e-01 -4.35175858e-02 6.61605716e-01 6.45063937e-01
-2.27202356e-01 5.98447025e-01 -9.42927748e-02 -5.40575087e-01
-6.48744702e-01 -9.26837862e-01 -5.20667374e-01 6.35357201e-01
7.23339736e-01 -9.68156829e-02 -6.96865499e-01 -3.25011104e-01
-5.20629585e-02 2.69905955e-01 -5.65821767e-01 -3.64580125e-01
-3.56234789e-01 -1.18927336e+00 2.93659389e-01 5.13740182e-01
1.25259936e+00 -1.47512805e+00 -8.14009666e-01 3.08664858e-01
7.47673139e-02 -7.19017506e-01 1.90239593e-01 9.03965086e-02
-9.08816040e-01 -1.00468147e+00 -7.06200838e-01 -1.26539731e+00
8.10399354e-01 5.87661326e-01 4.22267556e-01 1.66922107e-01
-7.20211148e-01 2.32821107e-01 -6.38792098e-01 -5.48913777e-01
7.90039152e-02 -3.49152833e-01 -1.73277766e-01 4.78508294e-01
5.10945141e-01 -4.28278953e-01 -8.60272050e-01 5.57460427e-01
-7.87965000e-01 1.19334618e-02 8.58174562e-01 1.17214179e+00
6.70415044e-01 4.06829536e-01 9.91334021e-01 -3.10606867e-01
2.63434649e-01 -1.85034007e-01 -4.32042509e-01 7.04490468e-02
-4.70612526e-01 -3.65699440e-01 5.08865058e-01 -4.35105920e-01
-1.23695028e+00 6.42198175e-02 -2.29901195e-01 -3.56154367e-02
-3.72654140e-01 4.64647889e-01 -2.61944979e-01 -7.24944949e-01
3.79497021e-01 7.02727914e-01 5.22643030e-02 -6.70340598e-01
1.87336449e-02 9.61977422e-01 2.90867597e-01 -2.92640954e-01
2.93312818e-01 4.76759583e-01 -9.57947150e-02 -1.20159829e+00
-9.80660096e-02 -4.95544642e-01 -4.96529728e-01 -9.74735841e-02
6.76792741e-01 -6.79255486e-01 -6.59136236e-01 6.36508644e-01
-1.09253299e+00 3.46132815e-01 2.29249358e-01 8.20395768e-01
-4.91220266e-01 4.22153324e-01 -6.55868053e-01 -1.10326707e+00
-4.91988391e-01 -1.00995970e+00 6.97704196e-01 4.82609302e-01
1.26698151e-01 -4.70687270e-01 -4.67640698e-01 1.21999159e-01
3.90207708e-01 7.39947781e-02 1.16959810e+00 -4.84492153e-01
-5.61621785e-01 -3.61776769e-01 -5.19077718e-01 4.30932075e-01
3.99399269e-03 -1.02748759e-01 -1.01935351e+00 -2.54566252e-01
2.30307311e-01 -1.27733991e-01 1.09786808e+00 4.25626040e-01
1.24810851e+00 1.52205117e-02 -4.66055751e-01 5.58579087e-01
1.67003202e+00 8.14442396e-01 9.34348047e-01 8.32267344e-01
1.41268328e-01 4.83091056e-01 7.99391747e-01 2.82174081e-01
-8.00503045e-03 3.97471070e-01 5.00729442e-01 -2.34717578e-01
-6.98170066e-02 1.01519711e-01 1.23394750e-01 9.19446647e-01
-2.15267047e-01 3.81272733e-02 -4.65593576e-01 5.56775928e-01
-1.91026103e+00 -8.96832824e-01 9.67401359e-03 2.13771319e+00
3.37804258e-01 6.31039143e-02 -2.05805421e-01 4.56312269e-01
8.58337939e-01 -1.77938566e-01 -2.43798524e-01 -5.34349620e-01
-2.47017235e-01 3.73090565e-01 1.39003843e-01 -2.12192954e-03
-1.21374011e+00 7.79990673e-01 6.03940678e+00 1.11097276e+00
-1.36050951e+00 5.44426925e-02 4.35184807e-01 4.55653697e-01
5.00433266e-01 -7.23635852e-02 -9.40789819e-01 6.46461010e-01
5.42697787e-01 2.41322219e-02 1.90761477e-01 8.31506789e-01
6.82851970e-02 -3.43169481e-01 -7.01347530e-01 1.37125969e+00
1.48293704e-01 -1.08134007e+00 3.57106358e-01 -8.74671265e-02
2.40771770e-01 -3.07032675e-01 -6.42406195e-02 2.37823725e-01
-4.86845493e-01 -9.54234421e-01 6.38603747e-01 7.22065806e-01
2.79702783e-01 -7.90646315e-01 8.29054713e-01 5.64912379e-01
-1.17472637e+00 -2.49934420e-01 -1.14675832e+00 -1.31379724e-01
-5.30379534e-01 4.44624245e-01 -5.35744429e-01 6.73844874e-01
9.23875093e-01 4.04978454e-01 -6.53789818e-01 1.54416513e+00
1.69833794e-01 6.25326216e-01 -1.14555091e-01 -3.62605631e-01
1.55565485e-01 -2.04010129e-01 4.33891714e-01 1.22098005e+00
4.62517381e-01 3.09062470e-02 9.69519839e-02 7.17552602e-01
3.82072985e-01 4.90755677e-01 -5.39402664e-01 2.10361749e-01
4.19069141e-01 1.39036250e+00 -9.32630539e-01 -4.34434354e-01
-2.08623111e-01 8.41870487e-01 1.56866625e-01 3.42008889e-01
-6.05081737e-01 -7.50156105e-01 1.27634957e-01 4.32307050e-02
6.73130810e-01 -1.62605003e-01 -1.11944854e-01 -8.44695032e-01
4.35006358e-02 -7.55208015e-01 2.15748876e-01 -7.04885960e-01
-1.29431319e+00 9.92975831e-01 -1.68317810e-01 -1.41365123e+00
3.30284387e-01 -9.37740207e-01 -5.77182412e-01 1.00300705e+00
-1.46885967e+00 -1.03330874e+00 -3.33601564e-01 5.75603127e-01
5.96493602e-01 -3.89721751e-01 8.07129920e-01 1.53199464e-01
-5.42490959e-01 4.72720683e-01 9.89206210e-02 9.81028285e-03
4.84151781e-01 -6.89095676e-01 -1.48974016e-01 6.33291304e-01
1.51564837e-01 7.21696556e-01 1.44994408e-01 -2.92048454e-01
-1.20984745e+00 -1.06896687e+00 7.30032444e-01 3.74546230e-01
2.12466568e-01 -1.73962429e-01 -1.02596068e+00 2.07324609e-01
1.85657427e-01 -6.70989230e-03 4.99965012e-01 -5.15216351e-01
-1.74936205e-01 -4.15097386e-01 -1.26024199e+00 4.71397310e-01
8.26222897e-01 1.01355957e-02 -9.13473129e-01 3.13657582e-01
6.58078045e-02 2.74753213e-01 -8.13220680e-01 7.53747344e-01
6.62065089e-01 -1.16067076e+00 9.97918844e-01 -1.91979110e-01
-3.92740518e-02 -6.82747960e-01 -2.48209924e-01 -1.11588013e+00
-9.88719225e-01 1.32391751e-01 4.10179377e-01 1.30444717e+00
7.77250454e-02 -1.08827174e+00 5.46765029e-01 2.72733136e-03
-4.42299880e-02 -8.96597266e-01 -9.64765012e-01 -1.10229230e+00
-1.81803986e-01 -1.00803137e-01 4.56075013e-01 4.58835632e-01
1.20499246e-01 1.83296174e-01 4.89371866e-02 2.03301370e-01
5.82479656e-01 4.67844129e-01 3.26903760e-01 -1.25221825e+00
-1.82128608e-01 -5.51158428e-01 -1.01838100e+00 -9.16710615e-01
-1.41064942e-01 -8.79170120e-01 2.69831508e-01 -1.65761900e+00
2.76137114e-01 -4.95947301e-01 -7.42177725e-01 3.01907510e-01
-4.76778485e-02 4.90979224e-01 5.02119139e-02 4.09680396e-01
-3.10998023e-01 7.47134328e-01 1.14618421e+00 -2.44612142e-01
1.21837631e-01 -2.54675522e-02 -4.60009068e-01 9.44972813e-01
7.88328767e-01 -1.63718149e-01 -3.71528774e-01 -1.82852179e-01
-5.49052715e-01 -1.86557561e-01 5.86341381e-01 -1.36060119e+00
2.39275306e-01 2.30259728e-02 8.80064607e-01 -7.44591475e-01
5.06148100e-01 -7.09053397e-01 -6.12599701e-02 7.33608961e-01
5.61454408e-02 -2.29277462e-01 3.22368264e-01 5.80606639e-01
-5.80308735e-01 -3.73461843e-01 1.06589651e+00 -1.67219713e-01
-1.13069487e+00 1.07676074e-01 -6.84645593e-01 -7.69395828e-01
1.09632683e+00 -5.81450582e-01 -2.17432678e-01 2.32521132e-01
-6.21691048e-01 -1.33905753e-01 9.88398865e-02 1.56253517e-01
1.08673036e+00 -1.35720885e+00 -5.69312871e-01 5.25578499e-01
1.83307439e-01 -4.01815623e-01 2.94624895e-01 9.32918072e-01
-6.41697347e-01 4.96779948e-01 -5.54512620e-01 -7.03045547e-01
-1.20838177e+00 6.78615391e-01 3.05197388e-01 1.91787839e-01
-7.39049494e-01 9.58880484e-01 6.66950583e-01 -1.26373649e-01
1.14250652e-01 -1.75484210e-01 -6.73556149e-01 -2.43105143e-01
6.84543312e-01 9.74610746e-02 2.69721389e-01 -5.83663404e-01
-6.45061612e-01 7.29738057e-01 -5.05783921e-03 7.92558566e-02
1.07254016e+00 5.52121084e-03 -4.19233739e-01 1.60048649e-01
9.74372149e-01 -3.40443820e-01 -8.15840304e-01 -2.16185078e-01
1.25657052e-01 -4.46608394e-01 5.97943924e-02 -5.79162121e-01
-8.67482841e-01 9.24226344e-01 9.73283708e-01 1.54245615e-01
1.32694483e+00 -5.61111152e-01 6.80765450e-01 4.00529772e-01
8.33179533e-01 -8.64509463e-01 -1.43345103e-01 3.15800816e-01
1.21812558e+00 -9.46683466e-01 -1.35632619e-01 -6.78718090e-01
-2.60010868e-01 1.26955450e+00 7.89939940e-01 -6.66550279e-01
9.18309093e-01 -7.94523358e-02 -2.71466598e-02 7.31296688e-02
-6.06248856e-01 -2.40685895e-01 4.22076613e-01 4.76150185e-01
3.31665784e-01 -1.07913703e-01 -9.99448240e-01 7.86782026e-01
1.15153812e-01 -4.99675833e-02 2.75834892e-02 1.12075293e+00
-8.86923075e-01 -8.49743128e-01 -7.12527633e-01 4.28734034e-01
-3.05611670e-01 -2.24254932e-02 -2.45185763e-01 7.36210644e-01
3.48418027e-01 9.82695878e-01 1.12522030e-02 -6.37091160e-01
1.72108874e-01 1.07654609e-01 6.90418363e-01 -4.78745162e-01
-3.00856531e-01 2.96664834e-01 -1.79514259e-01 -3.10597450e-01
-4.64550793e-01 -2.31947944e-01 -1.34785604e+00 2.54307359e-01
-4.38550293e-01 3.32016438e-01 8.20827663e-01 9.27268565e-01
4.09796804e-01 3.59407395e-01 7.08737075e-01 -1.03213263e+00
-5.19360960e-01 -1.18534946e+00 -9.50000405e-01 1.90899521e-01
-1.33743778e-01 -1.13175261e+00 -2.58715659e-01 1.37269244e-01] | [10.245092391967773, -0.18476247787475586] |
36062072-65d8-444f-a3c0-064de4fed17f | online-deception-detection-refueled-by-real | 1707.09406 | null | http://arxiv.org/abs/1707.09406v1 | http://arxiv.org/pdf/1707.09406v1.pdf | Online Deception Detection Refueled by Real World Data Collection | The lack of large realistic datasets presents a bottleneck in online
deception detection studies. In this paper, we apply a data collection method
based on social network analysis to quickly identify high-quality deceptive and
truthful online reviews from Amazon. The dataset contains more than 10,000
deceptive reviews and is diverse in product domains and reviewers. Using this
dataset, we explore effective general features for online deception detection
that perform well across domains. We demonstrate that with generalized features
- advertising speak and writing complexity scores - deception detection
performance can be further improved by adding additional deceptive reviews from
assorted domains in training. Finally, reviewer level evaluation gives an
interesting insight into different deceptive reviewers' writing styles. | ['James Caverlee', 'Ruihong Huang', 'Zeyu Dai', 'Wenlin Yao'] | 2017-07-28 | online-deception-detection-refueled-by-real-1 | https://aclanthology.org/R17-1102 | https://aclanthology.org/R17-1102.pdf | ranlp-2017-9 | ['deception-detection'] | ['miscellaneous'] | [-3.14969361e-01 -2.33204082e-01 -2.74115026e-01 -8.15728009e-01
-4.63013113e-01 -9.72643018e-01 7.02185869e-01 2.14398541e-02
-1.60799176e-01 6.07619762e-01 -1.80281520e-01 -2.94092983e-01
-1.78475708e-01 -3.01333278e-01 -2.32660711e-01 1.35936914e-02
-8.29124302e-02 3.32131863e-01 -2.60747299e-02 -5.47482789e-01
8.64544928e-01 5.11825800e-01 -1.14512146e+00 3.12605381e-01
1.16763747e+00 9.54405487e-01 -8.38325739e-01 7.96841621e-01
2.86458343e-01 6.89766943e-01 -1.25726342e+00 -1.60834825e+00
6.01825416e-01 -3.93357307e-01 -4.44029003e-01 2.21748501e-01
1.05434322e+00 -7.09722161e-01 -3.72344196e-01 1.20688665e+00
2.80116856e-01 -7.98303708e-02 6.75479949e-01 -1.46129537e+00
-1.39188600e+00 2.24934638e-01 -5.82179189e-01 7.00885773e-01
3.78874332e-01 4.80407089e-01 1.09078193e+00 -8.93029034e-01
3.42010885e-01 1.37116563e+00 3.45588833e-01 6.64254367e-01
-1.00242603e+00 -9.19369161e-01 -2.03530028e-01 4.31707203e-01
-1.08314252e+00 -7.40468860e-01 8.10548782e-01 -3.24793935e-01
3.51663172e-01 3.03068161e-01 4.48146790e-01 1.61954868e+00
3.15298736e-01 6.24301732e-01 1.26726437e+00 -9.46405604e-02
2.71120280e-01 6.56477749e-01 6.99113429e-01 5.73139191e-01
9.39974070e-01 3.91208529e-01 -8.31393063e-01 -8.89914989e-01
3.70993651e-02 -2.09218636e-01 1.40978962e-01 -1.74299151e-01
-2.84882337e-01 1.41655278e+00 2.22202107e-01 -2.18570441e-01
-2.39156149e-02 -1.96872056e-01 6.07302606e-01 8.20688009e-01
7.72878885e-01 1.13783109e+00 -1.14548199e-01 -1.83845788e-01
-8.54896069e-01 2.89530814e-01 1.43894327e+00 6.47852302e-01
3.50959241e-01 3.13639700e-01 2.20079362e-01 1.02708733e+00
3.27189788e-02 6.25763893e-01 5.48279405e-01 -1.22981310e+00
1.87236458e-01 4.80756909e-01 4.61456150e-01 -1.77407074e+00
5.51251061e-02 -4.59590942e-01 -4.02513504e-01 2.85425305e-01
4.14663404e-01 -7.70349614e-03 -5.21098018e-01 9.32793438e-01
6.27744943e-02 -3.50861549e-01 -2.61227041e-01 1.25647914e+00
3.27069700e-01 8.08742419e-02 -2.25196734e-01 -8.04168209e-02
1.38397801e+00 -9.34822142e-01 -5.76014817e-01 -6.19159400e-01
3.60111654e-01 -7.28777051e-01 1.11628723e+00 1.14928591e+00
-1.22283137e+00 -1.83732733e-01 -1.34499276e+00 -4.01030220e-02
-5.35256863e-01 -2.61580348e-01 9.66094971e-01 1.43401718e+00
-7.06195831e-01 7.36348510e-01 -8.04897398e-02 2.88266465e-02
8.83684993e-01 1.46934614e-01 -2.47394964e-01 -3.63941997e-01
-1.24669814e+00 1.18946278e+00 -4.71687526e-01 -8.73979852e-02
-1.27706456e+00 -5.18814981e-01 -5.44113934e-01 -1.09932214e-01
3.54815900e-01 -4.40208137e-01 1.32376552e+00 -1.48328292e+00
-1.39149725e+00 9.66953814e-01 1.66741777e-02 -4.33346570e-01
6.99878097e-01 -2.14201584e-01 -1.07386041e+00 2.94523716e-01
-8.11944306e-02 -9.21894684e-02 1.50657606e+00 -1.31859481e+00
3.37741151e-02 -8.70191276e-01 -5.56243435e-02 -7.22540841e-02
-7.32186615e-01 4.22870457e-01 5.04256308e-01 -3.61532480e-01
-4.40744966e-01 -7.11878121e-01 1.41907990e-01 1.84428114e-02
-2.79312283e-01 5.39015234e-02 6.61532104e-01 -5.95715046e-01
1.12804222e+00 -2.05507922e+00 -4.10495162e-01 4.06138480e-01
8.87500584e-01 7.68477201e-01 -2.75504649e-01 2.94988632e-01
2.34335274e-01 6.22938812e-01 2.44848847e-01 -4.04247820e-01
2.56491929e-01 8.02909359e-02 -1.94877461e-01 6.49929881e-01
2.15332806e-01 9.91998553e-01 -8.53413403e-01 -1.53359607e-01
-3.88798952e-01 -7.54963309e-02 -2.91899025e-01 -5.77481128e-02
3.31137925e-01 -4.05840486e-01 -1.93049669e-01 8.44863474e-01
9.50501144e-01 -2.11047620e-01 -2.91297827e-02 3.43372315e-01
5.02716601e-01 4.27479416e-01 -4.57534373e-01 7.88844049e-01
-2.46061683e-01 8.37500751e-01 5.21923661e-01 -5.28891385e-01
1.18481779e+00 -2.79254526e-01 -3.90646040e-01 -5.50866485e-01
3.37504417e-01 4.54499781e-01 3.23762089e-01 -4.24328119e-01
8.52829933e-01 -3.08775783e-01 -1.68614492e-01 8.28933060e-01
-3.36810909e-02 -1.79106995e-01 5.65289855e-02 6.84527278e-01
1.19947851e+00 -1.17986298e+00 1.69380680e-01 -2.19131514e-01
3.84945422e-01 1.96953461e-01 1.41600490e-01 1.00802243e+00
-1.06262922e+00 2.86994547e-01 8.42592895e-01 -5.08021712e-01
-1.01456904e+00 -1.17138052e+00 7.38280639e-02 1.10378385e+00
-1.79011170e-02 -2.51059890e-01 -3.79816145e-01 -1.21743071e+00
6.54276073e-01 8.52319658e-01 -6.91232085e-01 -8.74240816e-01
-1.31590456e-01 -7.56377876e-01 8.05699944e-01 5.48837595e-02
3.24791849e-01 -4.22212183e-01 1.76232889e-01 -3.97041589e-01
2.93612897e-01 -9.29692388e-01 -7.05843627e-01 -1.65900812e-01
-7.50177562e-01 -1.11098921e+00 -2.93093711e-01 -3.43884885e-01
4.05301034e-01 6.51123822e-01 1.38269246e+00 2.57615983e-01
-3.16774726e-01 1.65417150e-01 -3.67550820e-01 -5.18544793e-01
-7.53622651e-01 -4.61227223e-02 4.44693029e-01 4.18467559e-02
1.02471244e+00 -3.25565696e-01 -6.20449305e-01 8.11589479e-01
-5.42198420e-01 -1.05352652e+00 4.95856762e-01 7.11297572e-01
-6.09128773e-01 -2.21505776e-01 1.02516305e+00 -9.60956812e-01
1.57413781e+00 -6.96650922e-01 -3.64539444e-01 -8.72511417e-02
-1.06664360e+00 -3.60159487e-01 9.59783614e-01 -7.16409743e-01
-8.11754644e-01 -6.39458477e-01 4.37260062e-01 -3.04359555e-01
1.19848743e-01 6.13624193e-02 1.86263636e-01 -5.66445649e-01
1.30966461e+00 9.11929384e-02 5.61887085e-01 -2.90159762e-01
1.57306448e-01 9.88430321e-01 1.26288712e-01 2.01207353e-04
8.01488698e-01 2.25121662e-01 -4.23103303e-01 -5.79751849e-01
-1.03524435e+00 -3.48630905e-01 -3.06064636e-01 -7.13268295e-02
-1.18771933e-01 -5.86929083e-01 -1.15712631e+00 4.08397198e-01
-9.21011984e-01 3.10578018e-01 1.00544944e-01 2.50982076e-01
9.94687080e-02 7.12225080e-01 -1.00593662e+00 -1.00158131e+00
-3.83287549e-01 -6.29299641e-01 6.02285087e-01 -1.18936270e-01
-3.96335989e-01 -9.82372880e-01 2.76606288e-02 9.81095433e-01
6.52363718e-01 -2.68381625e-01 4.15997714e-01 -1.34396195e+00
-2.08998293e-01 -7.72040308e-01 -3.61518204e-01 1.16981030e+00
-8.09428990e-02 6.93638325e-02 -7.01581717e-01 -2.07086578e-01
2.87464321e-01 -7.92337060e-01 7.56936252e-01 -5.00733927e-02
8.47667336e-01 -7.28395998e-01 5.62631637e-02 6.13263473e-02
9.79215503e-01 -1.61315143e-01 5.82431495e-01 -5.55401854e-02
5.00376165e-01 7.03254700e-01 5.08240283e-01 6.24601662e-01
2.38936067e-01 2.85779476e-01 5.64676821e-01 5.11963010e-01
3.71649891e-01 -1.04839325e-01 6.33544207e-01 6.19661331e-01
4.32605565e-01 -3.54839236e-01 -7.00365305e-01 3.56658190e-01
-1.26286542e+00 -1.04578042e+00 -4.49234635e-01 2.11412430e+00
7.53366470e-01 3.33199859e-01 6.14781439e-01 6.34743348e-02
8.25075924e-01 2.75681227e-01 -8.08878005e-01 -1.37077343e+00
-1.25345185e-01 1.02212159e-02 5.78973472e-01 6.45555437e-01
-6.18320584e-01 7.22080708e-01 7.70678473e+00 6.84240818e-01
-5.40144205e-01 1.91591367e-01 7.74406493e-01 -5.95232368e-01
-3.35305393e-01 -3.20184499e-01 -7.66276181e-01 7.55439043e-01
1.39254344e+00 -5.61403155e-01 5.64649880e-01 1.05095649e+00
4.35717642e-01 -1.35106355e-01 -9.62929010e-01 9.10770357e-01
8.73619914e-01 -8.96392822e-01 1.58393294e-01 2.73950577e-01
6.47685945e-01 -3.19786072e-01 4.43056285e-01 3.08755577e-01
7.98522115e-01 -1.12106383e+00 2.25269988e-01 3.29888463e-01
2.21795917e-01 -9.26319480e-01 7.62744606e-01 4.25602078e-01
2.98842371e-01 -3.99449766e-01 -8.01282227e-01 -5.15386820e-01
2.98928414e-02 1.06380665e+00 -9.55398440e-01 -4.42723744e-02
3.26831847e-01 7.72843659e-01 -1.08083248e+00 7.36782372e-01
-2.03195274e-01 7.18856215e-01 1.17388882e-01 -6.23867333e-01
1.09795481e-01 -1.05767362e-01 4.30447310e-01 8.93274963e-01
-1.57825664e-01 -4.51887697e-02 -4.37153637e-01 1.01906109e+00
-4.44451332e-01 -1.35651901e-01 -7.46676683e-01 -3.56563598e-01
4.57955092e-01 1.58677197e+00 8.84102881e-02 -3.08494449e-01
-6.02030046e-02 1.29517674e+00 3.71298581e-01 3.00850030e-02
-7.70326853e-01 -4.30944115e-01 9.26309645e-01 2.94990540e-01
-1.20730422e-01 -3.87297362e-01 -4.57057118e-01 -1.20164478e+00
-1.04217266e-03 -1.33556759e+00 2.05123752e-01 -7.53532290e-01
-2.29150319e+00 2.74871886e-01 -4.07618642e-01 -7.46072054e-01
9.77110863e-03 -9.61596906e-01 -5.50589204e-01 7.95486331e-01
-1.39162683e+00 -7.00898767e-01 -3.73210192e-01 4.89931375e-01
3.64752591e-01 -4.40134317e-01 3.69736344e-01 2.92499930e-01
-5.70259392e-01 1.01572502e+00 1.12844901e-02 3.81021351e-02
1.01204479e+00 -1.19126797e+00 5.36960304e-01 5.46307862e-01
-1.28315181e-01 8.13781738e-01 5.81160009e-01 -6.18470490e-01
-1.16068172e+00 -5.95748603e-01 9.52810884e-01 -1.06868243e+00
1.22156143e+00 -7.25678325e-01 -9.13367629e-01 4.18755293e-01
2.01443151e-01 5.32201566e-02 1.03621113e+00 2.78954893e-01
-9.12480056e-01 -2.24685982e-01 -1.40909088e+00 3.16279113e-01
1.13748956e+00 -6.95820868e-01 -5.16618907e-01 5.71271360e-01
2.02895999e-01 2.74064928e-01 -6.24651551e-01 -4.58592862e-01
6.55125737e-01 -1.27892208e+00 7.05796719e-01 -1.10835922e+00
9.08403933e-01 4.30779904e-01 2.46971399e-01 -1.64482248e+00
-4.68415588e-01 -7.27001786e-01 -2.83016384e-01 9.04696405e-01
4.41808194e-01 -9.22956586e-01 9.87126946e-01 8.80968034e-01
1.48730710e-01 -2.25311488e-01 -6.43466294e-01 -1.17193973e+00
2.79365450e-01 -2.95420121e-02 2.20358148e-01 1.18303192e+00
4.26586866e-01 6.66680098e-01 -6.33440912e-01 -1.46310359e-01
6.53607488e-01 -1.93956569e-01 7.08156347e-01 -1.45105195e+00
-1.90572113e-01 -3.37118477e-01 -5.41401923e-01 -9.57841575e-01
4.39818978e-01 -6.93991303e-01 -4.32240844e-01 -6.51880503e-01
2.67653018e-01 -5.63938953e-02 1.67631030e-01 6.08236268e-02
-1.24389365e-01 4.53524500e-01 -1.67215802e-02 3.15632522e-01
-8.21110427e-01 3.61582667e-01 1.42215800e+00 -1.11136980e-01
2.42003307e-01 9.55268070e-02 -1.33665371e+00 4.25098479e-01
9.46937621e-01 -6.70406342e-01 -2.36364782e-01 -3.17629188e-01
4.79463279e-01 -1.98973000e-01 6.62347019e-01 -3.69772762e-01
2.36208782e-01 -3.14657507e-03 3.45135182e-01 -8.73382483e-03
5.12781084e-01 -5.56174934e-01 -6.49573386e-01 3.00399661e-01
-4.47849244e-01 4.47336644e-01 -1.54684082e-01 9.32149827e-01
-4.47024517e-02 -5.47339380e-01 8.52635801e-01 -4.22014356e-01
-1.15463868e-01 -1.06471881e-01 -4.61882263e-01 2.95770139e-01
8.79654169e-01 -4.00899768e-01 -1.05926108e+00 -9.03360248e-01
-6.01410568e-01 4.46469113e-02 6.52441740e-01 5.73310852e-01
6.59648240e-01 -1.12161934e+00 -9.72940922e-01 1.58411279e-01
2.90300637e-01 -1.22006989e+00 -7.36929849e-02 6.46978855e-01
-3.53759617e-01 9.61309001e-02 -3.16838205e-01 -1.62678078e-01
-1.28789353e+00 7.42094100e-01 4.34024274e-01 -4.36873548e-03
4.57778603e-01 7.10308254e-01 -4.46426004e-01 -2.21800402e-01
-3.29798132e-01 5.72884202e-01 2.13574588e-01 1.19842239e-01
7.62334347e-01 9.39529717e-01 3.74184370e-01 -5.56539178e-01
-4.30024356e-01 -3.14130783e-01 -7.27207184e-01 1.32169826e-02
8.56849611e-01 -2.41983607e-01 2.45855935e-02 2.61856139e-01
1.36913943e+00 3.44746023e-01 -7.25178301e-01 2.78246671e-01
4.14668731e-02 -1.32480121e+00 5.44567853e-02 -1.28404033e+00
-1.03421676e+00 7.03367472e-01 1.76133886e-01 7.71834493e-01
5.35079956e-01 -2.84108341e-01 6.02164090e-01 2.26344585e-01
3.15246165e-01 -1.19035435e+00 7.94881463e-01 2.89367110e-01
9.41065192e-01 -1.68792057e+00 4.29845750e-01 -4.84013647e-01
-9.82682645e-01 8.09571087e-01 7.03302920e-01 -5.49215674e-01
5.68594098e-01 -3.00001591e-01 5.83705157e-02 -6.82390749e-01
-8.31347585e-01 5.90009153e-01 2.71851599e-01 6.78580999e-01
-6.51130304e-02 1.22836363e-02 -7.13664234e-01 9.08810675e-01
-3.26774538e-01 -3.51514727e-01 1.24565196e+00 4.60503012e-01
-4.60982651e-01 -9.35683906e-01 -2.55401045e-01 9.20301557e-01
-5.94492793e-01 -2.49241173e-01 -1.54245555e+00 7.05248773e-01
-5.49383283e-01 1.48041105e+00 -7.20162168e-02 -5.77488959e-01
4.04094875e-01 -7.76751190e-02 3.30453217e-01 -6.31529868e-01
-1.09102368e+00 -7.35657156e-01 5.87864995e-01 -6.52623534e-01
1.92681178e-01 -6.45756602e-01 -1.64005518e-01 -1.30604529e+00
-7.94109762e-01 2.22992957e-01 6.41784668e-01 4.83371615e-01
8.84296954e-01 -3.61475110e-01 1.04001570e+00 -4.68768030e-01
-1.48902845e+00 -9.31043625e-01 -1.10506678e+00 7.47925878e-01
3.01355898e-01 -5.38750708e-01 -1.46891785e+00 -5.18950999e-01] | [8.093703269958496, 10.190067291259766] |
0f41e931-1b47-45b9-987d-a541ac669814 | destress-deep-learning-for-unsupervised | 1911.13213 | null | https://arxiv.org/abs/1911.13213v1 | https://arxiv.org/pdf/1911.13213v1.pdf | DeStress: Deep Learning for Unsupervised Identification of Mental Stress in Firefighters from Heart-rate Variability (HRV) Data | In this work we perform a study of various unsupervised methods to identify mental stress in firefighter trainees based on unlabeled heart rate variability data. We collect RR interval time series data from nearly 100 firefighter trainees that participated in a drill. We explore and compare three methods in order to perform unsupervised stress detection: 1) traditional K-Means clustering with engineered time and frequency domain features 2) convolutional autoencoders and 3) long short-term memory (LSTM) autoencoders, both trained on the raw RRI measurements combined with DBSCAN clustering and K-Nearest-Neighbors classification. We demonstrate that K-Means combined with engineered features is unable to capture meaningful structure within the data. On the other hand, convolutional and LSTM autoencoders tend to extract varying structure from the data pointing to different clusters with different sizes of clusters. We attempt at identifying the true stressed and normal clusters using the HRV markers of mental stress reported in the literature. We demonstrate that the clusters produced by the convolutional autoencoders consistently and successfully stratify stressed versus normal samples, as validated by several established physiological stress markers such as RMSSD, Max-HR, Mean-HR and LF-HF ratio. | ['María Rodríguez Martínez', 'Ali Oskooei', 'Arvind Sridhar', 'Sophie Mai Chau', 'Jonas Weiss', 'Bruno Michel'] | 2019-11-18 | null | null | null | null | ['heart-rate-variability'] | ['medical'] | [-9.27717313e-02 -5.72502501e-02 1.89706370e-01 -6.48772418e-01
-1.80725604e-02 -1.85775086e-01 -2.22425982e-01 4.30331796e-01
-5.08484602e-01 5.06066799e-01 4.22556669e-01 -1.02691904e-01
-7.92930424e-01 -5.47716856e-01 6.50365204e-02 -7.49302983e-01
-7.07443833e-01 1.63301796e-01 -5.78661680e-01 -4.83116090e-01
3.44940454e-01 5.08572400e-01 -1.78130901e+00 2.43611887e-01
6.27680480e-01 1.05678904e+00 -4.24548030e-01 9.85080242e-01
1.59627452e-01 7.67319858e-01 -1.12292814e+00 3.78807127e-01
-4.30084690e-02 -6.53337955e-01 -7.20549285e-01 -3.30691129e-01
2.87095279e-01 -4.65428047e-02 -3.18639964e-01 4.22258437e-01
1.12735665e+00 7.42925882e-01 3.13201338e-01 -9.37409401e-01
-4.98099685e-01 6.51593089e-01 -1.53686870e-02 9.12806392e-01
1.57014742e-01 -1.73795238e-01 1.95651770e-01 -4.37439978e-01
1.54296041e-01 7.58282959e-01 1.31090212e+00 4.33889896e-01
-1.19676232e+00 -4.04599994e-01 -3.66616249e-01 3.55874717e-01
-1.44158232e+00 -1.36799172e-01 8.51175725e-01 -6.38615966e-01
1.20115376e+00 2.53759921e-01 9.28619146e-01 1.27420306e+00
6.02325976e-01 -2.62324452e-01 1.22210705e+00 -2.09979773e-01
4.64064211e-01 2.00730786e-01 6.25085235e-01 4.42288131e-01
-1.46018341e-01 4.26917106e-01 -6.90919161e-01 -2.76871264e-01
5.10384321e-01 3.20141226e-01 -1.61695749e-01 3.08200926e-01
-1.33181417e+00 6.91502094e-01 3.26406121e-01 7.60947347e-01
-7.82333314e-01 -5.42450920e-02 1.00342202e+00 5.76226652e-01
3.92450243e-01 6.75680995e-01 -5.74616194e-01 -2.56766886e-01
-1.23362052e+00 -9.97064188e-02 6.32252455e-01 1.71914458e-01
4.56771731e-01 3.83188635e-01 -1.28484502e-01 8.57687712e-01
-2.03223005e-01 1.09493941e-01 1.29815543e+00 -1.07850981e+00
-4.47096556e-01 4.58096504e-01 -3.18846107e-01 -1.23658228e+00
-1.04395878e+00 -6.47479296e-01 -1.00975120e+00 6.23475090e-02
1.39308155e-01 -6.23971045e-01 -6.64358079e-01 1.47598815e+00
1.81944370e-01 3.12726915e-01 3.38124186e-01 1.04539418e+00
1.03591311e+00 2.84826428e-01 1.58335492e-01 -4.90967512e-01
1.36706626e+00 -9.48367640e-02 -1.06412899e+00 1.65299892e-01
4.34293538e-01 -1.97790295e-01 8.13750327e-01 4.90183741e-01
-7.15614617e-01 -1.27137434e+00 -1.07399273e+00 2.42479056e-01
-6.47128284e-01 1.05483895e-02 3.06469649e-01 9.56666708e-01
-1.15881658e+00 1.30722868e+00 -8.32331538e-01 -4.01930928e-01
6.35382459e-02 3.09747428e-01 -2.57267922e-01 6.15810990e-01
-1.58259499e+00 7.51624227e-01 3.63273978e-01 4.78807777e-01
-6.63886547e-01 -9.43897069e-01 -6.25227034e-01 2.03461617e-01
-2.43401870e-01 -5.74078023e-01 7.78813064e-01 -7.74508059e-01
-1.36707962e+00 7.73817301e-01 2.49211401e-01 -6.45731390e-01
-1.11678481e-01 -2.69887507e-01 -6.84839964e-01 4.04255331e-01
-4.35036004e-01 2.94443935e-01 6.60828948e-01 -8.16074491e-01
-2.10071523e-02 -6.17254138e-01 -6.72094643e-01 -1.29709631e-01
-5.82082033e-01 -4.84586805e-02 1.02830720e+00 -4.32484210e-01
2.14308470e-01 -8.90909612e-01 -2.40672737e-01 -9.25798476e-01
-2.98657656e-01 -9.51144174e-02 7.05388367e-01 -7.36272395e-01
1.39501381e+00 -2.20571446e+00 -1.63146570e-01 4.93844807e-01
5.33785403e-01 1.34728342e-01 3.97836030e-01 3.62925053e-01
-8.15010905e-01 8.18276033e-02 -9.91869066e-03 8.03335458e-02
7.74147436e-02 3.35306793e-01 -3.25545996e-01 6.67559803e-01
-9.62100178e-02 4.98589069e-01 -7.22037315e-01 -5.56824386e-01
3.02814752e-01 8.88574302e-01 -1.14173573e-02 2.97323763e-01
6.80254459e-01 5.45136034e-01 1.14251927e-01 2.87792593e-01
2.85128087e-01 3.17490637e-01 1.52257100e-01 -2.66121715e-01
-2.44567320e-01 8.60243887e-02 -9.13919151e-01 1.50063181e+00
-3.11462972e-02 9.67572689e-01 -5.40033519e-01 -1.38116813e+00
1.56402421e+00 8.03765595e-01 9.66812313e-01 -4.41366047e-01
4.67319816e-01 -1.95086971e-02 1.80663690e-01 -8.31570208e-01
4.24313724e-01 -6.88825548e-01 -2.77684689e-01 3.48591030e-01
5.65615177e-01 4.59139973e-01 -6.16864190e-02 -1.70584053e-01
1.04899704e+00 -4.70076680e-01 2.67825313e-02 -6.20945454e-01
2.71264225e-01 -2.19638452e-01 6.99673831e-01 6.47027314e-01
-7.13341713e-01 5.42252302e-01 6.25536919e-01 -9.01582956e-01
-5.95839739e-01 -9.68694866e-01 -3.47303867e-01 1.15357471e+00
-6.44405484e-01 -8.41288567e-02 -7.47665226e-01 -1.59732923e-01
-1.02381676e-01 5.51946819e-01 -1.13899255e+00 -7.32924998e-01
-4.68225658e-01 -7.55737007e-01 1.10435855e+00 8.82036984e-01
1.35397613e-01 -1.35224473e+00 -1.34903431e+00 2.23699436e-01
-2.34028712e-01 -6.09337509e-01 1.08155683e-02 9.82006133e-01
-1.14278126e+00 -7.37829626e-01 -4.90063995e-01 -7.59973109e-01
8.65745246e-02 -3.78466934e-01 1.22563255e+00 -5.94255924e-02
-7.68824637e-01 5.19843459e-01 -3.63503933e-01 -6.22568071e-01
-7.35392869e-02 -4.10218164e-02 6.03956163e-01 -1.86053291e-01
7.74994910e-01 -1.15131676e+00 -7.89270341e-01 7.45314360e-02
-5.32177925e-01 -1.08144164e+00 1.76115841e-01 5.23832083e-01
6.35121167e-01 1.83812067e-01 1.04861200e+00 -4.87206161e-01
1.07204759e+00 -4.96252537e-01 3.46824467e-01 -1.44165233e-01
-8.97472739e-01 -3.18156481e-01 4.61988837e-01 -4.16068614e-01
-5.84399581e-01 -2.18045637e-01 -2.67370939e-02 -9.01311278e-01
-8.46073508e-01 5.85226774e-01 6.20496571e-01 3.01743865e-01
1.46479774e+00 3.22939247e-01 2.97901720e-01 -1.04879744e-01
-2.40463186e-02 6.51558042e-01 9.70418990e-01 -3.15448672e-01
1.65506676e-01 2.16825187e-01 -5.49013168e-02 -1.04598200e+00
-6.77161634e-01 -6.08520567e-01 -1.02787757e+00 -5.34491301e-01
1.31119907e+00 -7.94972956e-01 -1.00094545e+00 9.25665200e-02
-5.70720255e-01 -3.73207122e-01 -8.07574332e-01 8.41803849e-01
-6.57562971e-01 2.13821694e-01 -8.38161469e-01 -1.07167339e+00
-7.59628892e-01 -3.61149251e-01 5.29692709e-01 4.36662883e-01
-6.17515028e-01 -1.13687170e+00 8.31821799e-01 5.61668687e-02
6.45918787e-01 9.99628067e-01 6.61090672e-01 -9.38384414e-01
9.48386490e-01 -2.90079176e-01 5.93221724e-01 7.07494736e-01
2.27904432e-02 1.22202620e-01 -1.27343893e+00 -1.04481526e-01
5.41971624e-01 -3.00467491e-01 7.99486756e-01 7.25939214e-01
1.24187517e+00 -4.69982512e-02 7.63515830e-02 5.08537591e-01
1.16624308e+00 3.54110360e-01 8.61634910e-01 -7.94799849e-02
5.10845363e-01 8.78216147e-01 2.09434599e-01 5.93516350e-01
1.43656418e-01 -3.49340707e-01 -3.69226076e-02 -2.10644871e-01
4.34177130e-01 3.63488883e-01 3.47959280e-01 1.07317162e+00
-6.63193345e-01 3.50654870e-01 -1.13444078e+00 5.00340998e-01
-1.39965522e+00 -1.44617319e+00 -2.01656878e-01 2.02810168e+00
6.57344282e-01 -8.19631293e-02 5.03856540e-01 7.98493981e-01
7.65551329e-01 3.19197476e-02 -3.95338446e-01 -1.13053548e+00
1.20275602e-01 7.56317496e-01 3.88994105e-02 -4.09686007e-02
-1.07073236e+00 1.57917872e-01 6.90679121e+00 -2.63939232e-01
-1.34562123e+00 -6.98244125e-02 5.93441784e-01 -3.74507189e-01
3.93018514e-01 -5.59052706e-01 -9.80019271e-02 3.01309139e-01
2.06567597e+00 -1.00466674e-02 9.46110711e-02 9.60748017e-01
2.80750424e-01 1.52635932e-01 -9.48560476e-01 9.73677933e-01
1.74081519e-01 -8.04757595e-01 -9.63810861e-01 3.75454873e-02
3.96857947e-01 -3.66311669e-02 1.73617266e-02 7.16652930e-01
-3.43393236e-01 -1.24287200e+00 8.87734219e-02 1.22692204e+00
6.87194586e-01 -1.12614691e+00 9.70223784e-01 -4.38633412e-02
-7.83884466e-01 -4.53808218e-01 -4.81331855e-01 -3.20731103e-01
-3.71836245e-01 1.00704432e+00 -7.03792989e-01 2.68793195e-01
1.03926873e+00 5.68466902e-01 -4.96974766e-01 6.49454534e-01
3.77280831e-01 7.82132983e-01 6.15021912e-03 2.42500916e-01
-1.38958171e-01 1.42912701e-01 2.72886544e-01 1.38041377e+00
2.22852573e-01 2.79924363e-01 1.50228888e-01 8.88565660e-01
5.85725427e-01 -8.07466581e-02 -5.02355933e-01 -1.16890080e-01
1.66106313e-01 1.38095772e+00 -7.59660780e-01 -4.54084784e-01
8.36330503e-02 5.20380080e-01 -1.84323907e-01 3.45314026e-01
-7.41394162e-01 -1.03298271e+00 8.69401455e-01 4.61387001e-02
5.97031228e-03 -3.02794129e-01 -3.80851924e-01 -6.52309895e-01
-4.90367621e-01 -6.59925163e-01 6.81697488e-01 -6.65796280e-01
-1.47921920e+00 8.80409479e-01 -1.40854299e-01 -1.01189959e+00
-3.60148609e-01 -2.63780951e-01 -9.79816258e-01 9.18252051e-01
-1.01092494e+00 6.19581752e-02 -6.75453484e-01 8.94528031e-01
6.00620508e-02 -5.17880358e-02 1.30679119e+00 2.79803813e-01
-8.36318791e-01 4.32544947e-01 -2.01550946e-01 2.83584416e-01
8.60812128e-01 -1.72759795e+00 -2.85659909e-01 2.96662778e-01
-3.38892519e-01 9.50791359e-01 7.20482528e-01 -4.64091718e-01
-8.02906215e-01 -9.70359385e-01 9.20922339e-01 -2.79860377e-01
2.21580282e-01 -9.71669797e-03 -1.19450533e+00 3.94565791e-01
3.21954519e-01 2.09684551e-01 1.63150167e+00 5.36293268e-01
1.24255866e-01 -2.59872764e-01 -1.13165975e+00 -2.19880596e-01
3.93158495e-01 -7.23438978e-01 -1.12760699e+00 2.44536832e-01
4.69254881e-01 -2.84325004e-01 -1.96573734e+00 3.20874333e-01
6.58047616e-01 -1.46092880e+00 9.22586977e-01 -7.25601614e-01
5.04097879e-01 -6.58647940e-02 3.40843014e-02 -1.58065593e+00
-4.97699022e-01 -4.04073089e-01 -1.02963962e-01 8.99291098e-01
-5.32432683e-02 -5.14711976e-01 5.15522063e-01 5.63916981e-01
-6.05504632e-01 -7.41280079e-01 -6.24008179e-01 -4.91907924e-01
1.20745525e-01 -2.08363876e-01 3.08055699e-01 1.43526268e+00
6.75057054e-01 1.62618369e-01 -4.99268575e-03 -3.30219828e-02
4.39101607e-01 1.49847284e-01 1.29338235e-01 -1.56407368e+00
-1.13921408e-02 -2.02957258e-01 -7.34638870e-01 2.90429622e-01
1.62131459e-01 -7.54370809e-01 1.74068972e-01 -1.13779771e+00
-3.24987501e-01 -9.95532423e-02 -1.12245524e+00 4.90629256e-01
-3.01835798e-02 4.95097376e-02 -4.93687272e-01 -2.62064070e-01
-1.73920408e-01 3.64845067e-01 6.45525396e-01 3.93129945e-01
-7.96435773e-01 -3.01681191e-01 -6.44358754e-01 4.76884007e-01
1.27350593e+00 -4.37846661e-01 -5.36178291e-01 2.74307519e-01
8.05316120e-02 3.69182885e-01 5.28564513e-01 -1.67871869e+00
-7.68453302e-03 5.27957939e-02 1.13801014e+00 -4.88912255e-01
-5.55570722e-02 -3.67046028e-01 -5.16373217e-02 5.95125616e-01
-6.49676859e-01 3.95185411e-01 2.84783959e-01 3.24709386e-01
-2.74706542e-01 1.22101910e-01 6.43585443e-01 -3.70661080e-01
-3.71594459e-01 -1.34002998e-01 -8.63992870e-01 -4.93840016e-02
6.76237702e-01 -3.44541550e-01 -3.03274721e-01 -1.10473923e-01
-1.56906748e+00 -1.33319646e-01 -3.03024352e-01 2.59518027e-01
7.76305914e-01 -1.26737070e+00 -5.65521955e-01 3.09310436e-01
-1.04052462e-01 -5.02307296e-01 9.49438930e-01 1.38849425e+00
-3.18903834e-01 3.36267263e-01 -7.89349139e-01 -7.52282321e-01
-1.06053829e+00 7.99685955e-01 8.11768711e-01 2.26082742e-01
-6.28951788e-01 6.35248899e-01 -5.84451973e-01 -4.16371047e-01
3.03811401e-01 -5.23793995e-01 -6.74476683e-01 5.94729900e-01
6.44838870e-01 7.24284351e-01 3.52103651e-01 -3.48485976e-01
-4.87458080e-01 2.39624307e-01 4.44605857e-01 3.23084295e-01
1.35806572e+00 -7.22321197e-02 -1.56456858e-01 9.92907941e-01
1.24716401e+00 -4.92099464e-01 -6.46787405e-01 1.73295096e-01
-3.20970267e-02 7.67753571e-02 3.58925015e-01 -6.87956870e-01
-1.15034628e+00 1.04412556e+00 1.48533821e+00 6.44729435e-01
1.31511128e+00 -5.15218556e-01 8.20265651e-01 5.95533967e-01
-3.85387540e-01 -1.55395341e+00 3.19022834e-01 2.70876437e-01
5.96336305e-01 -8.02704751e-01 -3.11757058e-01 5.79267442e-01
-9.87943888e-01 1.50233316e+00 4.22125518e-01 -5.28698146e-01
9.15983856e-01 9.25451815e-02 5.60890377e-01 -6.71064079e-01
-7.55895615e-01 -2.22451136e-01 1.07398905e-01 7.72600055e-01
8.13643456e-01 1.32533923e-01 -1.36149690e-01 8.53204429e-01
-7.50813842e-01 1.40223637e-01 7.25330293e-01 8.15290391e-01
-7.23577142e-01 -4.12846416e-01 -5.81804574e-01 6.27848864e-01
-5.73729575e-01 1.93479970e-01 -3.72005373e-01 3.71804327e-01
4.55047339e-01 1.07523537e+00 4.21032906e-01 -1.00106668e+00
5.93054593e-01 9.29268181e-01 1.69608742e-01 -4.06210035e-01
-1.22007358e+00 1.70055851e-02 -1.60118997e-01 -4.80193287e-01
-5.49369156e-01 -6.69023037e-01 -1.36563897e+00 -3.06031883e-01
4.44260687e-02 2.27430105e-01 6.39808178e-01 7.23870158e-01
4.68347400e-01 1.26007104e+00 7.33795345e-01 -6.88475132e-01
-3.49368036e-01 -1.30722523e+00 -1.00974417e+00 2.80860513e-01
6.17107928e-01 -5.80044031e-01 -4.94915694e-01 3.62875193e-01] | [13.81567096710205, 3.1222331523895264] |
7fda7103-54d0-4479-afde-1557130a00c9 | a-comparative-analysis-of-portfolio | 2305.17523 | null | https://arxiv.org/abs/2305.17523v1 | https://arxiv.org/pdf/2305.17523v1.pdf | A Comparative Analysis of Portfolio Optimization Using Mean-Variance, Hierarchical Risk Parity, and Reinforcement Learning Approaches on the Indian Stock Market | This paper presents a comparative analysis of the performances of three portfolio optimization approaches. Three approaches of portfolio optimization that are considered in this work are the mean-variance portfolio (MVP), hierarchical risk parity (HRP) portfolio, and reinforcement learning-based portfolio. The portfolios are trained and tested over several stock data and their performances are compared on their annual returns, annual risks, and Sharpe ratios. In the reinforcement learning-based portfolio design approach, the deep Q learning technique has been utilized. Due to the large number of possible states, the construction of the Q-table is done using a deep neural network. The historical prices of the 50 premier stocks from the Indian stock market, known as the NIFTY50 stocks, and several stocks from 10 important sectors of the Indian stock market are used to create the environment for training the agent. | ['Soubhik Maji', 'Manas Kumar Sarkar', 'Kushagra Kumar', 'Atish Kumar Majee', 'Anshuman Pathak', 'Aditya Jaiswal', 'Jaydip Sen'] | 2023-05-27 | null | null | null | null | ['q-learning', 'portfolio-optimization'] | ['methodology', 'time-series'] | [-5.83735943e-01 -8.73849392e-02 -1.75752386e-01 -1.25287369e-01
-5.59541583e-01 -5.73364556e-01 5.11641145e-01 -7.58597404e-02
-3.33409309e-01 1.08481455e+00 2.89964318e-01 -5.37052453e-01
-7.29323924e-01 -1.22129476e+00 -3.08692038e-01 -6.85283482e-01
-3.04650456e-01 4.75435406e-01 2.20027104e-01 -3.98054153e-01
9.07914340e-01 4.58521605e-01 -1.11050868e+00 -6.70573562e-02
4.94320244e-01 1.42982924e+00 -1.25332057e-01 1.49507508e-01
-1.37359768e-01 9.19596910e-01 -7.19457507e-01 -7.94157863e-01
7.52027035e-01 -3.87560397e-01 -2.73246430e-02 -4.24916416e-01
-3.75011116e-01 -5.46284080e-01 -1.23581514e-01 1.02725303e+00
5.18240154e-01 2.22807989e-01 6.61101580e-01 -9.27733839e-01
-3.04601699e-01 9.20118093e-01 -4.66602892e-01 5.17690539e-01
-1.96995109e-01 -2.08463129e-02 1.37331998e+00 -5.69121540e-01
1.62160084e-01 1.01718473e+00 5.64243138e-01 1.11801282e-01
-6.25258267e-01 -8.10445845e-01 -2.95777559e-01 2.04770401e-01
-7.11283505e-01 1.16985522e-01 6.74060643e-01 -5.40052056e-01
1.07626379e+00 1.13042193e-02 1.03297913e+00 3.75052154e-01
1.03187907e+00 3.05748552e-01 1.28858244e+00 -3.14297199e-01
5.85423410e-01 1.38829038e-01 -1.00258082e-01 9.05789733e-02
6.57392502e-01 1.03465748e+00 -3.01428378e-01 -4.19921309e-01
7.67731488e-01 -1.34074494e-01 3.46444249e-01 2.14325450e-02
-5.44017255e-01 1.16761768e+00 2.33259946e-01 1.59324110e-01
-8.20693552e-01 1.67458177e-01 2.47655407e-01 6.10964477e-01
3.29333901e-01 6.05262160e-01 -5.28182626e-01 8.63627903e-03
-8.66256475e-01 7.49634206e-01 9.18800771e-01 3.24650377e-01
1.39289796e-01 5.60444713e-01 -2.46606484e-01 2.57877648e-01
6.48892701e-01 5.07730544e-01 5.80171168e-01 -1.15801501e+00
6.57767236e-01 3.90141994e-01 4.00113106e-01 -9.76307809e-01
-2.76554585e-01 -5.73929250e-01 -3.21142942e-01 7.14981675e-01
2.64158070e-01 -5.99849045e-01 -3.61996830e-01 1.20003760e+00
-8.64566565e-02 -1.69444129e-01 4.72473949e-01 2.43735641e-01
1.88903674e-01 8.89362812e-01 -2.29976833e-01 -5.10883689e-01
8.97359014e-01 -7.24837720e-01 -6.24754548e-01 7.98738524e-02
-1.52523890e-01 -3.67315292e-01 1.24096096e-01 4.74114299e-01
-1.32333589e+00 -3.38838398e-01 -1.12200391e+00 1.12290728e+00
-3.11425805e-01 -4.47084486e-01 5.15807927e-01 8.05463374e-01
-7.97933996e-01 1.08811295e+00 -4.50485498e-01 6.38267219e-01
1.95346504e-01 4.66568291e-01 3.26615334e-01 6.59459233e-01
-1.41499555e+00 1.17872059e+00 6.88143373e-01 -4.66513671e-02
-9.33228850e-01 -3.52407277e-01 -3.43809932e-01 4.10222769e-01
2.80860037e-01 -9.29868072e-02 1.16836548e+00 -1.11280406e+00
-1.73009503e+00 1.24319732e-01 1.00665331e+00 -8.93468857e-01
5.43717206e-01 -2.64805108e-02 -2.39172325e-01 9.45927948e-03
-3.08101177e-01 -2.99892426e-02 6.27618909e-01 -5.95349789e-01
-9.26707029e-01 -2.07266454e-02 9.49866138e-03 2.46438593e-01
-1.84912369e-01 4.32204217e-01 4.94152308e-01 -1.13446009e+00
-3.42489898e-01 -6.71804070e-01 -3.43639433e-01 -8.62606466e-01
7.01943710e-02 -5.53468280e-02 -7.65545517e-02 -9.39939976e-01
1.26874864e+00 -1.85118592e+00 -6.96674064e-02 6.41440034e-01
-4.94636089e-01 -4.24734131e-02 2.16149583e-01 7.40641296e-01
-3.19888383e-01 -1.59795824e-02 -1.21338591e-01 4.12258565e-01
3.41300011e-01 2.99447738e-02 -4.46842968e-01 1.59381121e-01
9.52775329e-02 8.00189435e-01 -5.12196064e-01 -1.90042146e-02
-7.11984187e-02 -2.89396554e-01 -3.39714438e-01 4.27312136e-01
-3.15648317e-01 -2.02829707e-02 -4.66785014e-01 5.28898299e-01
4.08823788e-01 1.98590636e-01 1.22142099e-01 3.26448888e-01
-3.72755766e-01 2.96775967e-01 -1.33211803e+00 3.26865137e-01
2.97811683e-02 8.53981450e-02 -5.17952204e-01 -8.39412212e-01
1.29705822e+00 4.40082967e-01 5.74148357e-01 -8.82181287e-01
3.55773494e-02 4.38099384e-01 3.64523977e-01 -5.20512871e-02
5.66867888e-01 -7.80978978e-01 -1.13873467e-01 8.53618443e-01
-1.45456702e-01 -7.71500394e-02 3.92582506e-01 -4.05543715e-01
8.02732170e-01 1.73016950e-01 4.28911746e-01 -4.18765873e-01
3.81147951e-01 -1.70334667e-01 9.31882083e-01 4.02477682e-01
-9.89485383e-02 8.82190317e-02 1.09919786e+00 -5.92062652e-01
-9.65900242e-01 -1.08447993e+00 2.81117260e-02 6.32084966e-01
-3.70723069e-01 3.16454023e-01 -4.85759050e-01 -4.15263116e-01
3.56677085e-01 8.59117985e-01 -5.86777031e-01 -1.90837875e-01
-4.81242925e-01 -1.21828794e+00 2.02872921e-02 6.36488497e-01
5.14155149e-01 -1.54695785e+00 -1.01312244e+00 5.96535742e-01
6.82689548e-01 -1.36426747e-01 -8.35284516e-02 1.89086825e-01
-8.74473333e-01 -9.95768309e-01 -8.40087712e-01 -1.93244934e-01
8.64112228e-02 -6.37690842e-01 1.09915555e+00 -4.74800229e-01
3.16234708e-01 -8.11746791e-02 -2.31153280e-01 -9.29584503e-01
-4.81149763e-01 -3.63900125e-01 -2.56342553e-02 -1.44981556e-02
-8.99302363e-02 -2.11575732e-01 -5.60398102e-01 9.52468216e-02
-5.44035971e-01 -5.60356498e-01 6.87169969e-01 8.78461182e-01
4.72115010e-01 6.48567140e-01 1.26676631e+00 -5.65014720e-01
8.78777564e-01 -6.34158015e-01 -1.64348114e+00 3.95607352e-01
-9.74520683e-01 1.22997396e-01 1.40602604e-01 -1.94793373e-01
-1.33477199e+00 -4.21996891e-01 -7.79725686e-02 -7.55621120e-02
7.47080863e-01 8.04966450e-01 -2.21908614e-02 4.35974374e-02
5.14099151e-02 7.40874261e-02 4.90479283e-02 -3.67795229e-01
-2.34441653e-01 3.69235367e-01 1.47373185e-01 -2.11388007e-01
7.44993508e-01 -2.47200325e-01 2.79350840e-02 1.16466187e-01
-5.88570058e-01 2.11347118e-01 -1.06541216e-01 -4.17924672e-01
6.19452119e-01 -5.63892245e-01 -6.38989925e-01 7.75992155e-01
-5.51199198e-01 -3.13372523e-01 -4.89930123e-01 5.53016305e-01
-7.60241389e-01 -3.50122809e-01 -7.06569910e-01 -1.21964467e+00
-7.28520334e-01 -1.12894189e+00 1.88714359e-02 4.94780660e-01
3.62087563e-02 -1.07080197e+00 4.75753993e-01 -9.01811384e-03
5.95340729e-01 7.32646763e-01 1.13093579e+00 -1.02361917e+00
-5.44270813e-01 -2.44065374e-01 4.84788464e-03 6.60124421e-01
-3.78659852e-02 1.31591469e-01 -4.39047247e-01 -3.18855941e-01
4.34388101e-01 -6.48264065e-02 4.65224952e-01 5.49405754e-01
5.07030129e-01 -4.43311453e-01 3.81179720e-01 2.72299170e-01
1.72989392e+00 1.23244357e+00 5.80546856e-01 9.88347173e-01
-1.42227113e-01 8.48754704e-01 8.93130124e-01 8.09938192e-01
1.21215351e-01 1.66221857e-01 5.10020673e-01 5.97845137e-01
7.47555137e-01 2.25647688e-01 6.58491790e-01 6.89988613e-01
-1.62850916e-01 -4.53696912e-03 -9.28956330e-01 2.94132501e-01
-1.51775050e+00 -1.24035859e+00 4.54563349e-01 2.16771650e+00
6.03267670e-01 8.49159479e-01 5.15936732e-01 2.57692188e-01
4.33688730e-01 1.71195835e-01 -5.57563365e-01 -7.49188781e-01
-1.05009131e-01 4.23270971e-01 8.90734375e-01 2.63840020e-01
-9.42295909e-01 3.71660769e-01 7.12523842e+00 4.32461560e-01
-8.59595239e-01 -3.26876730e-01 9.11225319e-01 -7.45825991e-02
-3.31465691e-01 -4.33437936e-02 -8.61388803e-01 7.55222082e-01
1.35165775e+00 -7.67348289e-01 2.35800520e-01 6.45741761e-01
1.31176665e-01 -1.66798860e-01 -4.03717071e-01 3.03965330e-01
-4.05375421e-01 -1.56606400e+00 -1.99694052e-01 2.92138785e-01
7.96616375e-01 -9.01708081e-02 3.54978055e-01 3.83381367e-01
5.75942993e-01 -8.10824752e-01 1.06479180e+00 1.10552287e+00
9.44136679e-02 -1.56838584e+00 1.25589728e+00 8.29873458e-02
-7.05651700e-01 -7.02815890e-01 -3.66034716e-01 -1.57553762e-01
8.04287195e-02 4.29046631e-01 -1.77285686e-01 5.59920371e-01
6.83394313e-01 3.40367407e-01 -2.16503039e-01 1.14238596e+00
2.01713722e-02 6.12686932e-01 -1.52500033e-01 -3.04115176e-01
4.18534607e-01 -6.96894705e-01 2.49735132e-01 3.52347046e-01
4.38080996e-01 6.07886277e-02 -2.33640745e-01 7.52699137e-01
2.02034935e-02 -6.57487754e-03 -3.68255615e-01 -1.97566181e-01
3.53010952e-01 7.79604435e-01 -5.09089649e-01 -5.29428460e-02
-4.14276123e-01 6.58774227e-02 -2.66460210e-01 -3.92472520e-02
-6.00114405e-01 -5.90930939e-01 3.85204822e-01 -1.76286444e-01
4.78508741e-01 4.73082028e-02 -3.62855017e-01 -5.18094897e-01
-2.07311034e-01 -1.08815837e+00 4.66069043e-01 -3.55841786e-01
-1.32141399e+00 3.20482463e-01 4.41204041e-01 -1.03068960e+00
-7.23162591e-01 -7.33147323e-01 -1.03643680e+00 1.14281869e+00
-1.61208105e+00 -1.44168764e-01 5.41855633e-01 2.76759528e-02
1.51905194e-01 -1.17035091e+00 4.00525510e-01 -6.89604655e-02
-6.87897086e-01 2.91730642e-01 8.71606648e-01 1.28564313e-01
-7.30555132e-02 -1.47241032e+00 6.65739655e-01 5.22804320e-01
-2.06200987e-01 9.47362557e-02 5.33432007e-01 -9.87626255e-01
-9.54198718e-01 -7.14729607e-01 6.23058140e-01 -4.80237417e-02
1.13729870e+00 4.29170668e-01 -6.64334118e-01 4.56999928e-01
4.87043738e-01 -6.38518274e-01 6.42607272e-01 -6.39491200e-01
3.59844714e-01 -4.56729501e-01 -1.13316739e+00 3.64290327e-01
-6.91118166e-02 4.19008918e-02 -9.75325644e-01 3.11877262e-02
4.93140429e-01 -1.31448284e-01 -1.17969286e+00 2.65067309e-01
7.18400717e-01 -1.21524727e+00 8.80661964e-01 -4.28406954e-01
2.87127912e-01 3.52174580e-01 -8.41737241e-02 -1.24409926e+00
-3.37017804e-01 -6.91140354e-01 -4.83631268e-02 1.21194959e+00
4.44234490e-01 -1.02339327e+00 8.94039929e-01 6.29137874e-01
3.49319100e-01 -9.74645972e-01 -1.05165267e+00 -8.69244337e-01
4.33357805e-01 1.90238103e-01 1.02437198e+00 4.57666218e-01
-2.87181824e-01 -1.62749514e-01 -2.90928364e-01 -3.11836064e-01
9.10194159e-01 4.08022165e-01 -3.06994584e-03 -1.02987075e+00
-6.28797472e-01 -8.23113263e-01 -1.69262290e-02 2.12155476e-01
6.42732903e-02 -5.16864836e-01 -3.68613303e-01 -1.08360517e+00
-9.75217521e-02 -3.37637484e-01 -9.38123822e-01 2.00970590e-01
6.14685304e-02 -4.28338736e-01 4.69269514e-01 6.96348995e-02
2.30237052e-01 6.46652162e-01 8.72231901e-01 3.89230698e-02
-7.38776028e-02 7.71854222e-01 -5.80837965e-01 7.30662227e-01
1.36441469e+00 -5.67689836e-01 -2.19642282e-01 1.37233272e-01
6.16904020e-01 6.50184393e-01 -2.24337980e-01 -8.69970858e-01
-1.81252301e-01 -6.79500103e-01 4.02357310e-01 -9.15023029e-01
1.29351276e-03 -5.51781118e-01 4.86219943e-01 1.04638672e+00
-3.98058176e-01 9.50475335e-01 1.19888149e-01 3.08306783e-01
-4.10129488e-01 -8.48215938e-01 8.66608739e-01 -4.36232150e-01
-4.19628084e-01 1.16258845e-01 -3.96365523e-01 -2.81032287e-02
1.24021339e+00 -6.49634600e-02 7.97476433e-03 -3.06571960e-01
-4.16482538e-01 3.22481066e-01 1.15171790e-01 1.79719687e-01
7.04428017e-01 -1.30593181e+00 -8.11816990e-01 -5.11113778e-02
-4.99809891e-01 -5.63678443e-01 -2.87043750e-01 2.02802673e-01
-9.34801042e-01 4.10476148e-01 -7.96307027e-01 4.87365514e-01
-4.85417783e-01 3.39786828e-01 9.70469892e-01 -9.03680265e-01
-2.71240592e-01 6.24248862e-01 -1.11559190e-01 -5.03196269e-02
2.37218127e-01 -3.59639376e-02 -7.03804076e-01 6.00694180e-01
6.54227972e-01 6.91513002e-01 -6.48387149e-02 -3.15033704e-01
-1.34943098e-01 4.01251763e-01 2.54207671e-01 -6.43480480e-01
1.75107598e+00 3.56674910e-01 -1.63656488e-01 4.16352481e-01
6.40051126e-01 -2.06981167e-01 -1.41824019e+00 3.73265482e-02
8.65826011e-01 -2.56342977e-01 1.55159412e-02 -8.16962719e-01
-1.40195251e+00 6.33640110e-01 5.95403194e-01 2.75956690e-01
1.12406337e+00 -7.94272780e-01 6.04669929e-01 2.44052783e-01
3.15994442e-01 -1.39224398e+00 1.05262615e-01 5.81546485e-01
1.01103890e+00 -7.23501265e-01 2.01775104e-01 5.98175526e-01
-8.63340318e-01 1.17705739e+00 2.01510534e-01 -6.19719863e-01
1.07495856e+00 4.16022092e-01 -2.89740646e-03 4.87424843e-02
-8.86793375e-01 1.44335970e-01 1.34268284e-01 2.37209424e-01
4.53139395e-02 6.67513022e-03 -5.58605373e-01 8.18866909e-01
-3.35888237e-01 -2.54536241e-01 5.93902230e-01 1.03347266e+00
-6.24777019e-01 -1.16288316e+00 -5.08544028e-01 6.99343681e-01
-9.72896457e-01 -4.36859019e-02 -1.67896688e-01 7.35523403e-01
-2.62868941e-01 4.01023835e-01 2.44997174e-01 -1.29529461e-01
4.79065686e-01 -1.45410016e-01 2.81310529e-01 -2.79504418e-01
-1.25858378e+00 6.00703396e-02 3.51305977e-02 -2.40430951e-01
-2.16115162e-01 -9.55451488e-01 -1.13384604e+00 -2.15901777e-01
-2.06143454e-01 3.42182130e-01 5.04806280e-01 8.44283283e-01
-3.71050596e-01 4.10129040e-01 1.12364590e+00 -6.72477305e-01
-1.45672452e+00 -7.39921689e-01 -1.07538104e+00 -1.85595304e-01
7.61840120e-02 -7.30135620e-01 -3.53196114e-01 -5.92860699e-01] | [4.530313968658447, 3.990506649017334] |
06d92e1c-35d4-4ad3-855a-de530f397990 | code-to-comment-translation-a-comparative | 2106.08415 | null | https://arxiv.org/abs/2106.08415v1 | https://arxiv.org/pdf/2106.08415v1.pdf | Code to Comment Translation: A Comparative Study on Model Effectiveness & Errors | Automated source code summarization is a popular software engineering research topic wherein machine translation models are employed to "translate" code snippets into relevant natural language descriptions. Most evaluations of such models are conducted using automatic reference-based metrics. However, given the relatively large semantic gap between programming languages and natural language, we argue that this line of research would benefit from a qualitative investigation into the various error modes of current state-of-the-art models. Therefore, in this work, we perform both a quantitative and qualitative comparison of three recently proposed source code summarization models. In our quantitative evaluation, we compare the models based on the smoothed BLEU-4, METEOR, and ROUGE-L machine translation metrics, and in our qualitative evaluation, we perform a manual open-coding of the most common errors committed by the models when compared to ground truth captions. Our investigation reveals new insights into the relationship between metric-based performance and model prediction errors grounded in an empirically derived error taxonomy that can be used to drive future research efforts | ['Kevin Moran', 'Antonios Anastasopoulos', 'Raihan Islam Arnob', 'Fahim Faisal', 'Junayed Mahmud'] | 2021-06-15 | null | https://aclanthology.org/2021.nlp4prog-1.1 | https://aclanthology.org/2021.nlp4prog-1.1.pdf | acl-nlp4prog-2021-8 | ['code-summarization'] | ['computer-code'] | [ 2.99627751e-01 3.20148826e-01 -4.83543217e-01 -3.27852517e-01
-1.30126941e+00 -6.46439493e-01 4.91165787e-01 7.93330431e-01
-8.08057375e-03 4.14355606e-01 6.14510238e-01 -6.41170084e-01
-6.03342839e-02 -3.27559173e-01 -6.18563771e-01 2.18758211e-01
2.56421685e-01 8.77769068e-02 3.69800366e-02 -3.44330907e-01
1.03338325e+00 -7.92351812e-02 -1.34163678e+00 4.74063873e-01
1.41721511e+00 4.11829263e-01 -5.24959303e-02 5.83406806e-01
-3.76788139e-01 1.13795412e+00 -6.29416227e-01 -8.83356631e-01
-1.25712082e-01 -6.95118666e-01 -1.15187275e+00 -1.80165768e-01
3.14299703e-01 1.72241151e-01 2.09518224e-01 1.35647881e+00
1.63171068e-01 -4.64373529e-01 4.27745014e-01 -1.26762831e+00
-1.06313407e+00 8.22430551e-01 -4.74960834e-01 2.39679039e-01
9.61831748e-01 1.89264953e-01 1.25578868e+00 -8.62627387e-01
9.08227205e-01 7.32400894e-01 9.59856927e-01 4.70741421e-01
-1.26242554e+00 -2.18352139e-01 -1.92517743e-01 4.22856882e-02
-1.08749413e+00 -4.12061751e-01 7.11265266e-01 -1.06012094e+00
1.32735944e+00 3.51040363e-01 3.21219653e-01 8.21295321e-01
4.64279145e-01 5.00734448e-01 8.63270104e-01 -8.13215375e-01
1.79067865e-01 6.34025633e-01 3.74858886e-01 7.80381143e-01
7.53392100e-01 -3.51625323e-01 -4.24045205e-01 -4.80493903e-01
-4.50536832e-02 -2.69322723e-01 -3.10089171e-01 -2.88531750e-01
-1.15984738e+00 8.19113493e-01 2.94427224e-03 5.66750646e-01
-2.11127758e-01 1.54879317e-01 6.48057640e-01 3.87093484e-01
7.32691884e-01 1.01759672e+00 -3.73057872e-01 -7.44152725e-01
-1.38853014e+00 2.94809699e-01 1.09548211e+00 1.27184844e+00
9.55921590e-01 -2.36669481e-01 -1.15959972e-01 7.48546064e-01
6.14228785e-01 3.03895384e-01 5.52132368e-01 -9.11522806e-01
7.10248232e-01 1.24473214e+00 3.73076290e-01 -1.16382039e+00
3.49552669e-02 -3.05495709e-01 5.83548360e-02 -5.33752181e-02
-5.46411909e-02 2.37097330e-02 -2.34632477e-01 1.24964452e+00
-3.95786494e-01 -6.08827233e-01 6.40240088e-02 5.06773114e-01
5.48281610e-01 2.92108506e-01 -6.14635199e-02 -3.03802192e-01
1.04641354e+00 -1.02730584e+00 -4.64765370e-01 -2.72347599e-01
1.13030255e+00 -1.07982612e+00 1.37926948e+00 -1.12766780e-01
-1.10430598e+00 -2.77500123e-01 -1.09660375e+00 -1.34917889e-02
-1.20716207e-01 3.00250262e-01 3.77979904e-01 6.96503580e-01
-1.17330992e+00 5.64527512e-01 -8.48133683e-01 -8.01534534e-01
2.98706107e-02 -3.02214921e-01 -1.56892799e-02 8.04395750e-02
-5.60307384e-01 1.20123649e+00 -7.10813552e-02 -3.61357719e-01
-5.18634439e-01 -7.44974852e-01 -7.65021563e-01 4.06144671e-02
6.86967596e-02 -5.25598109e-01 1.66469872e+00 -1.12817514e+00
-1.08461106e+00 9.37980771e-01 -2.97459990e-01 -2.42632210e-01
3.15641999e-01 -1.45469129e-01 -3.07539791e-01 -1.41723260e-01
4.33394223e-01 1.22650370e-01 8.21852535e-02 -1.48302245e+00
-6.36622488e-01 -6.11002259e-02 1.77465588e-01 -2.21775874e-01
-3.46845746e-01 5.52538812e-01 -2.17572730e-02 -5.60867310e-01
-1.51785687e-01 -8.01236570e-01 -1.34698927e-01 -3.47493857e-01
-2.19361514e-01 -1.17782302e-01 1.06206335e-01 -8.90119135e-01
1.89947057e+00 -1.94351661e+00 1.98217139e-01 -9.84425694e-02
2.57814050e-01 -1.80429220e-02 -2.23760605e-01 9.58648384e-01
9.18566138e-02 6.97807610e-01 -5.63352048e-01 -2.59964347e-01
2.14135244e-01 -4.95412618e-01 -5.27894557e-01 2.40387350e-01
2.49666214e-01 8.97099733e-01 -1.21285224e+00 -5.61698616e-01
-2.73778737e-01 1.20261632e-01 -7.71864951e-01 2.40898550e-01
-2.62963504e-01 -1.05758175e-01 -3.79709929e-01 7.76845992e-01
1.61318228e-01 -2.39869446e-01 -2.98768468e-02 1.32093668e-01
-5.13767302e-01 7.13050246e-01 -3.60794753e-01 1.83710980e+00
-5.86036265e-01 1.08270693e+00 -4.50314969e-01 -4.56546783e-01
1.06609380e+00 2.65750825e-01 2.87992001e-01 -4.56964642e-01
9.05647967e-03 4.73677635e-01 9.98765230e-03 -7.75969148e-01
9.96488273e-01 2.10946381e-01 -2.45459452e-01 7.61764646e-01
-2.28773102e-01 -2.66385317e-01 5.74419677e-01 2.02139154e-01
1.30748892e+00 3.94421756e-01 4.88142699e-01 -4.34498101e-01
4.49595094e-01 6.95527613e-01 3.82539541e-01 6.84585810e-01
-1.93914428e-01 6.61939502e-01 7.24657357e-01 -3.32653791e-01
-1.20695829e+00 -6.83696866e-01 -2.54536364e-02 8.54896903e-01
6.91567957e-02 -8.77253115e-01 -1.26696634e+00 -7.62517869e-01
-2.39804044e-01 1.15537095e+00 -5.37597775e-01 -2.58611023e-01
-4.47847545e-01 -6.05318308e-01 8.34783852e-01 3.24909538e-01
-1.34045500e-02 -9.17779267e-01 -9.53937650e-01 3.24878633e-01
-4.78107452e-01 -7.00550675e-01 -4.35552806e-01 -3.14875752e-01
-9.20795679e-01 -1.13426125e+00 -3.37717593e-01 -6.40529096e-01
6.56539023e-01 2.90004998e-01 1.53779399e+00 5.78495920e-01
1.24072451e-02 5.06406128e-01 -9.76236105e-01 -4.17831242e-01
-1.34101439e+00 3.41902673e-01 -2.96487898e-01 -5.94926476e-01
9.32376564e-01 -5.07918894e-01 -2.17708722e-01 1.28965706e-01
-8.65520537e-01 1.52064711e-01 6.65836751e-01 5.60251951e-01
-4.23135310e-02 -6.54689491e-01 5.55196404e-01 -8.96688342e-01
1.15842557e+00 -6.83117747e-01 -5.13450086e-01 7.60333300e-01
-1.15424657e+00 2.71408051e-01 4.79274690e-01 -5.06916903e-02
-9.96133804e-01 -4.87137944e-01 -2.97469683e-02 2.41835654e-01
3.41290906e-02 1.06969190e+00 5.73467016e-01 -1.64982334e-01
1.16140831e+00 3.21694255e-01 -2.80149192e-01 -3.82610470e-01
2.65171707e-01 1.10880840e+00 2.05389991e-01 -6.99458539e-01
7.64975190e-01 -6.47081807e-02 -8.47221434e-01 -4.90515918e-01
-6.37050927e-01 -4.15925682e-01 -5.27628005e-01 -1.70794085e-01
5.94583154e-01 -6.84886396e-01 5.79035953e-02 -5.17954901e-02
-1.48342311e+00 -9.56189539e-03 -2.09883153e-01 2.78508723e-01
-7.54825473e-01 5.38289070e-01 -4.93684292e-01 -6.38159811e-01
-4.10083890e-01 -1.57099032e+00 1.08620346e+00 5.04588261e-02
-1.04710472e+00 -9.33617473e-01 6.37939572e-01 5.62105238e-01
7.66336739e-01 2.21220464e-01 1.26717019e+00 -5.52471638e-01
-4.26834434e-01 -3.45126450e-01 -2.64394462e-01 1.41719967e-01
1.52123675e-01 5.08320928e-01 -6.08680665e-01 -2.27801248e-01
-6.61391914e-02 -1.37118727e-01 3.16064149e-01 -1.29433393e-01
5.84368050e-01 -4.26585197e-01 -1.96815595e-01 1.86958164e-01
1.75350106e+00 1.04506023e-01 5.28449118e-01 6.01381600e-01
5.34922600e-01 6.72163606e-01 6.15680337e-01 4.72948819e-01
6.39520943e-01 4.25379574e-01 1.28258854e-01 4.43158507e-01
-1.45330876e-01 -4.18290645e-01 5.25297284e-01 1.53123939e+00
1.00521050e-01 -9.33980756e-03 -1.47393298e+00 9.41932082e-01
-1.80886674e+00 -7.43435919e-01 -2.77509183e-01 2.13427615e+00
9.00664330e-01 1.40333191e-01 -5.65453693e-02 -2.13443384e-01
5.79310417e-01 -1.61575973e-02 -1.18487343e-01 -6.15522265e-01
3.27907622e-01 -2.38900363e-01 2.51517177e-01 2.60294020e-01
-3.85817736e-01 5.74947774e-01 6.84918404e+00 3.51897091e-01
-9.68062937e-01 1.85390294e-01 2.13415548e-01 2.13907972e-01
-8.66333008e-01 6.01359606e-01 -3.10656190e-01 5.12303531e-01
1.31947923e+00 -8.32945883e-01 4.07110959e-01 1.04445064e+00
2.51424015e-01 -2.06918329e-01 -1.65379095e+00 6.28997564e-01
3.19025517e-01 -1.52110553e+00 4.79287729e-02 -1.40855521e-01
9.92175817e-01 1.76271409e-01 -3.40492845e-01 3.88770849e-01
3.07668477e-01 -7.67297864e-01 1.05440879e+00 6.01656497e-01
5.80542028e-01 -2.10878983e-01 7.91986465e-01 3.05106521e-01
-7.28617013e-01 -7.00042769e-02 -2.43086666e-01 -2.02348486e-01
-1.04089588e-01 4.35506314e-01 -8.89459193e-01 5.65783918e-01
4.76989388e-01 7.28414476e-01 -8.96499097e-01 1.07453322e+00
-6.23724833e-02 7.77428985e-01 4.43901807e-01 -3.08359325e-01
-1.48731666e-02 -1.18642054e-01 5.62473655e-01 1.66348147e+00
4.39032763e-01 -4.31094468e-01 -1.16874315e-02 1.54883277e+00
-9.91263017e-02 3.23081493e-01 -6.33872151e-01 -4.71194148e-01
6.01159036e-01 1.04479361e+00 -6.82759941e-01 -3.00337434e-01
-7.83663452e-01 6.57674491e-01 3.44931722e-01 2.15252459e-01
-7.88522303e-01 -7.08732605e-01 6.07105970e-01 1.69853389e-01
-1.26694962e-01 -1.50843546e-01 -7.99141824e-01 -1.49462330e+00
5.26736200e-01 -1.10044670e+00 -2.31824398e-01 -8.59770060e-01
-9.49881434e-01 8.33184361e-01 4.39673588e-02 -1.43152249e+00
-2.60482490e-01 -7.14263171e-02 -7.44711995e-01 6.63032770e-01
-1.39839637e+00 -7.07402170e-01 -3.13966542e-01 -4.55742717e-01
7.13506639e-01 -1.98196262e-01 7.17879534e-01 3.10377955e-01
-5.57346702e-01 6.05523050e-01 1.92849696e-01 1.10526890e-01
5.80592692e-01 -1.19563282e+00 8.79522860e-01 1.19351268e+00
7.40496144e-02 1.12019598e+00 1.14960182e+00 -8.02914083e-01
-1.25045323e+00 -1.04054844e+00 1.34344506e+00 -9.18780744e-01
8.60313177e-01 -5.59232570e-03 -9.09883440e-01 6.49825931e-01
3.65583479e-01 -5.12984157e-01 8.65721166e-01 3.79546694e-02
-4.53082770e-01 3.02461892e-01 -1.04255152e+00 4.70657796e-01
7.68516660e-01 -8.13131928e-01 -9.39409018e-01 1.70964971e-01
7.87832677e-01 -1.72596693e-01 -1.12233210e+00 8.21179971e-02
4.74151611e-01 -1.09359097e+00 1.62816077e-01 -5.04881501e-01
1.20734704e+00 -3.34418327e-01 -2.83098489e-01 -1.33755612e+00
-1.28953725e-01 -5.40115952e-01 3.06102276e-01 1.34912312e+00
8.85455489e-01 -3.16831589e-01 4.07841831e-01 7.80682921e-01
-3.30661446e-01 -7.28760183e-01 -5.35277784e-01 -5.59888303e-01
3.11684430e-01 -3.97182435e-01 5.80470979e-01 1.01411176e+00
7.74455786e-01 5.27257733e-02 1.48703486e-01 -1.28219962e-01
3.52201819e-01 2.04109266e-01 8.42485845e-01 -1.19127131e+00
-1.95302457e-01 -6.11738741e-01 -4.26460803e-01 -6.84727728e-01
1.97522014e-01 -1.01167929e+00 2.31512129e-01 -1.62633371e+00
7.38095045e-01 -1.18989080e-01 8.08361918e-02 3.24444175e-01
-1.08738780e-01 -3.14065651e-03 1.43145934e-01 5.23929477e-01
-7.05221474e-01 3.23479861e-01 3.80021930e-01 -1.43919811e-01
-1.46954089e-01 -2.99051166e-01 -1.14667404e+00 7.18733430e-01
5.57200968e-01 -8.68225932e-01 -2.35150129e-01 -7.82163203e-01
8.38292181e-01 2.42475644e-01 2.38619111e-02 -9.69626963e-01
1.62401363e-01 -3.12905163e-01 -6.07993782e-01 2.58187149e-02
-6.35137439e-01 -5.95359445e-01 5.04741818e-02 4.74412769e-01
-8.07755470e-01 5.35064340e-01 1.08421855e-01 2.03714639e-01
-2.99336135e-01 -7.88878202e-01 5.37321329e-01 -1.14233859e-01
-4.55751628e-01 -3.60848635e-01 -5.71607471e-01 3.07108462e-01
8.78605723e-01 -4.65267509e-01 -5.50114155e-01 -1.68858215e-01
6.21249750e-02 -3.82863022e-02 1.19880271e+00 6.97743058e-01
4.02691573e-01 -1.13337457e+00 -9.14164603e-01 -1.99429587e-01
7.84721494e-01 -5.17115474e-01 -3.70251060e-01 9.12098467e-01
-8.54992568e-01 4.62079942e-01 -1.83807999e-01 -3.82418990e-01
-8.84084821e-01 4.00892913e-01 1.82187185e-01 -2.02301815e-01
-2.19167262e-01 2.12844357e-01 -4.33515966e-01 -4.49993283e-01
-6.06447533e-02 -6.01932168e-01 5.11883274e-02 -2.81049520e-01
4.13818628e-01 6.25338852e-01 3.03447276e-01 -6.14843130e-01
-3.97543550e-01 3.27742934e-01 -6.06835335e-02 -6.76633567e-02
1.30118561e+00 -2.24923745e-01 -5.82850695e-01 6.86088800e-01
1.08003283e+00 3.37143481e-01 -6.88092887e-01 -1.81088999e-01
8.81144166e-01 -5.75780332e-01 -3.29221219e-01 -8.97428453e-01
-4.41151589e-01 6.26767039e-01 2.49206245e-01 4.44306880e-01
8.23863566e-01 -6.45744056e-02 4.35318738e-01 4.90699649e-01
5.38374424e-01 -9.53499138e-01 -1.60720244e-01 4.37414914e-01
8.64972711e-01 -1.26743901e+00 -2.71922667e-02 -1.99039549e-01
-4.97669905e-01 1.21629894e+00 6.16500854e-01 3.71317789e-02
6.95244148e-02 2.37130776e-01 1.84660196e-01 -2.96154052e-01
-8.66947889e-01 2.84721822e-01 2.53621250e-01 3.43389034e-01
1.21245074e+00 -8.14128295e-02 -7.35135019e-01 5.21455705e-01
-3.58867586e-01 1.82786971e-01 1.32077348e+00 1.11335111e+00
-6.01767004e-01 -1.17339790e+00 -2.04407439e-01 5.49178660e-01
-5.70924878e-01 -3.40153396e-01 -7.41868019e-01 6.08096600e-01
-3.38631421e-01 1.24873173e+00 -2.81125069e-01 -5.46308875e-01
4.77928609e-01 9.98358727e-02 2.75591761e-01 -1.04992843e+00
-1.01697135e+00 -5.63583612e-01 1.18796147e-01 -3.99835110e-01
-4.61481065e-01 -7.31338799e-01 -9.94979560e-01 -2.86255717e-01
-4.96205837e-01 4.80266213e-01 8.93057585e-01 9.90764916e-01
6.34506941e-01 4.33563828e-01 3.80383551e-01 -3.97885650e-01
-7.94194639e-01 -1.05086493e+00 1.42091021e-01 4.81703907e-01
3.26813698e-01 -2.81214356e-01 -3.78449619e-01 3.33729684e-01] | [7.686118125915527, 7.906425476074219] |
4b6eedd9-ebc8-4dba-a677-189abc11a6f6 | unsupervised-visual-time-series | 2111.10309 | null | https://arxiv.org/abs/2111.10309v1 | https://arxiv.org/pdf/2111.10309v1.pdf | Unsupervised Visual Time-Series Representation Learning and Clustering | Time-series data is generated ubiquitously from Internet-of-Things (IoT) infrastructure, connected and wearable devices, remote sensing, autonomous driving research and, audio-video communications, in enormous volumes. This paper investigates the potential of unsupervised representation learning for these time-series. In this paper, we use a novel data transformation along with novel unsupervised learning regime to transfer the learning from other domains to time-series where the former have extensive models heavily trained on very large labelled datasets. We conduct extensive experiments to demonstrate the potential of the proposed approach through time-series clustering. | ['Richi Nayak', 'Gaurangi Anand'] | 2021-11-19 | null | null | null | null | ['time-series-clustering'] | ['time-series'] | [ 4.80087727e-01 -1.58441365e-01 -1.61729142e-01 -5.52985013e-01
-4.71010983e-01 -4.59044874e-01 6.67263985e-01 1.16586752e-01
-2.33746886e-01 7.04288781e-01 2.51173198e-01 -4.14087027e-01
-5.30715883e-01 -8.16820562e-01 -3.24108034e-01 -5.14697433e-01
-6.97201192e-01 1.60623506e-01 2.58137137e-02 -1.03639670e-01
-8.82732403e-03 5.04921377e-01 -1.88738704e+00 1.10929452e-01
6.44350052e-01 1.31672764e+00 -2.64217317e-01 6.08922541e-01
-2.00602323e-01 7.83253729e-01 -4.27024186e-01 3.04726034e-01
4.08479154e-01 -1.82736181e-02 -3.31534803e-01 1.72987550e-01
-3.42205584e-01 1.99070618e-01 -1.28740385e-01 6.83058381e-01
4.67386723e-01 2.59170085e-01 5.36970615e-01 -1.62586403e+00
-3.63836050e-01 5.71466446e-01 -4.75593895e-01 4.71404582e-01
4.64981735e-01 -1.14894427e-01 3.96657556e-01 -3.90144914e-01
3.36346149e-01 9.94062662e-01 9.58018303e-01 1.16723590e-01
-8.89921010e-01 -6.61212862e-01 -1.56545296e-01 2.90910959e-01
-1.16867876e+00 -2.47599915e-01 1.29978836e+00 -6.31410182e-01
9.91579652e-01 2.12634966e-01 6.39878690e-01 1.17490172e+00
2.79223144e-01 4.08526897e-01 1.30398881e+00 -4.30127531e-01
4.40232337e-01 1.81840640e-02 9.71545354e-02 -1.57621965e-01
-1.21569023e-01 1.05557553e-01 -9.07949731e-03 -1.24669254e-01
3.44120383e-01 7.35297322e-01 4.99871552e-01 -9.30182338e-02
-1.14986920e+00 4.09945160e-01 1.06692621e-02 6.87292099e-01
-6.23503566e-01 2.12019645e-02 6.41230524e-01 8.24499607e-01
5.56181073e-01 -1.22633956e-01 -8.80097806e-01 -5.31740069e-01
-7.97461689e-01 -3.09785753e-01 7.43911445e-01 1.11115313e+00
6.68564796e-01 3.67801577e-01 2.23712236e-01 5.82106829e-01
1.51298344e-01 4.88504887e-01 1.14557683e+00 -4.02345151e-01
4.66382056e-01 7.48381495e-01 -1.27920434e-01 -8.54975402e-01
-6.51540518e-01 -5.14831468e-02 -1.23615730e+00 -3.72776836e-01
-9.78881642e-02 -5.51612079e-01 -8.90776992e-01 1.29608333e+00
5.02682149e-01 5.96729755e-01 1.79199487e-01 2.63395369e-01
5.76601028e-01 9.04449463e-01 3.03158551e-01 -6.22889638e-01
1.03177011e+00 -3.27215672e-01 -8.89663160e-01 3.30886722e-01
5.67906022e-01 -5.19943655e-01 7.36341238e-01 5.65890610e-01
-6.28080189e-01 -9.54248786e-01 -8.42607200e-01 3.28694850e-01
-8.91046524e-01 4.78350893e-02 7.61676669e-01 6.93987191e-01
-7.25307524e-01 7.08332241e-01 -1.01071560e+00 -9.10674214e-01
4.98860300e-01 5.41405678e-01 -1.87892616e-01 2.69057512e-01
-1.02385306e+00 3.75846446e-01 5.13798237e-01 -3.17462534e-01
-4.84355778e-01 -4.94576484e-01 -6.19456470e-01 -3.16101015e-01
-1.19986413e-02 -2.68354118e-01 7.94110417e-01 -8.30998361e-01
-1.58875239e+00 5.71003437e-01 1.77130908e-01 -8.52099359e-01
2.87177324e-01 -6.42399639e-02 -1.28383231e+00 -6.74235970e-02
3.72661278e-02 2.01572418e-01 1.12908566e+00 -5.64200878e-01
-5.92587709e-01 -5.06428957e-01 -3.65014464e-01 -5.88268876e-01
-8.03977251e-01 5.30001074e-02 4.23036098e-01 -7.41949737e-01
1.39200374e-01 -1.02742124e+00 -4.27090317e-01 -5.08135915e-01
-1.71879828e-01 -6.69614077e-01 1.55968869e+00 -3.91794413e-01
1.12778282e+00 -2.32314730e+00 -2.87358731e-01 4.08732295e-01
-1.83807209e-01 -2.95394994e-02 1.23324938e-01 8.49234343e-01
-3.73372644e-01 -3.06551512e-02 -1.49001524e-01 -1.97937116e-02
-2.71830056e-02 4.89629239e-01 -5.22515893e-01 4.31709945e-01
2.08371148e-01 7.54645467e-01 -7.69056141e-01 -1.72387436e-01
6.27339244e-01 4.86347228e-01 -2.68383473e-01 5.59059083e-02
-5.16494503e-03 1.01290441e+00 -6.56963706e-01 5.68667352e-01
3.50508243e-01 -2.00834095e-01 -6.19416349e-02 -3.40985745e-01
-4.13986653e-01 1.01721451e-01 -1.01839435e+00 1.74863791e+00
-3.94426227e-01 4.08927798e-01 -7.01572895e-01 -1.58657610e+00
1.13978791e+00 5.69862306e-01 1.36284614e+00 -9.08612847e-01
4.05576468e-01 9.61845554e-03 -1.53782114e-01 -1.01309001e+00
4.50289212e-02 -1.29480526e-01 -2.91123182e-01 5.76544940e-01
1.98174864e-01 2.36409009e-02 2.04226017e-01 -3.07870597e-01
1.28000736e+00 9.87355858e-02 4.61006373e-01 -2.33450443e-01
6.11271799e-01 -1.46046832e-01 2.39395425e-01 2.49062672e-01
-9.04134959e-02 8.92886519e-02 -2.19766751e-01 -7.78742135e-01
-1.14094150e+00 -1.02220595e+00 -7.32435957e-02 1.23758376e+00
-3.38185042e-01 -3.47558647e-01 -1.71479478e-01 -3.62709850e-01
-2.57031452e-02 5.51516294e-01 -4.69616592e-01 -1.88066840e-01
-3.99000198e-01 -6.76147044e-01 4.64031905e-01 6.39141262e-01
1.45931825e-01 -1.12983346e+00 -6.93682671e-01 4.18524712e-01
2.73070335e-01 -1.23051429e+00 6.57738596e-02 4.85512704e-01
-1.32521570e+00 -5.97763956e-01 -2.41768450e-01 -8.47697616e-01
3.49303812e-01 9.81507078e-02 8.26439559e-01 -5.17157733e-01
-4.35645849e-01 8.11698735e-01 -7.30025589e-01 -1.00483251e+00
-1.86086059e-01 4.55293953e-02 4.50711489e-01 4.61541593e-01
7.15305567e-01 -1.48240399e+00 -4.31860834e-01 2.98913330e-01
-1.06162095e+00 -4.72051799e-01 4.69303846e-01 1.31749883e-01
5.70132792e-01 5.70151508e-01 1.09570897e+00 -5.97252309e-01
6.38740540e-01 -9.84576225e-01 -2.73961186e-01 -2.86311537e-01
-4.64510798e-01 -2.41905838e-01 8.77666652e-01 -8.80858004e-01
-5.95715046e-01 3.00486326e-01 -4.56857607e-02 -5.08676887e-01
-6.11635447e-01 5.49133837e-01 4.40092385e-03 1.80379063e-01
7.75133967e-01 3.55009884e-01 -1.51744321e-01 -6.43956125e-01
4.77905661e-01 1.09817672e+00 5.69262266e-01 -5.11998653e-01
1.01742673e+00 7.64466107e-01 -1.62783399e-01 -1.18406034e+00
-3.26248497e-01 -8.25870574e-01 -8.22812796e-01 -2.27800682e-01
6.59763157e-01 -1.02189934e+00 -3.31176788e-01 1.61848575e-01
-7.61292577e-01 -3.29492800e-02 -7.84756660e-01 8.89464080e-01
-6.14944160e-01 1.12970695e-01 -2.23181620e-01 -1.07727695e+00
-2.80527741e-01 -3.72447938e-01 1.16458786e+00 1.11278161e-01
-3.41202974e-01 -1.15160918e+00 3.71191680e-01 -9.72967222e-02
3.92893851e-01 7.54517436e-01 4.79533881e-01 -9.48738515e-01
2.10993979e-02 -5.70233047e-01 1.35749355e-01 4.60210949e-01
7.12885022e-01 -1.14803508e-01 -9.93447006e-01 -3.38332236e-01
4.02326703e-01 -2.03817204e-01 3.11643541e-01 2.47923285e-01
1.48909509e+00 -2.87333488e-01 -3.11011463e-01 4.65600640e-01
1.32164538e+00 6.48443818e-01 7.47818887e-01 9.57706645e-02
7.75928617e-01 6.73513293e-01 4.09276903e-01 8.44784141e-01
4.76051301e-01 1.77109256e-01 7.15821981e-02 3.54969688e-02
4.13133353e-01 1.27121257e-02 3.86352271e-01 1.56454563e+00
-4.83155787e-01 1.09781362e-01 -8.94660115e-01 6.65120602e-01
-1.93748415e+00 -1.19020510e+00 -1.64865866e-01 2.05475807e+00
5.07685900e-01 1.66790277e-01 4.81903672e-01 9.20328498e-01
6.19442821e-01 -1.23942494e-01 -6.16482437e-01 -3.44132721e-01
1.19524434e-01 4.15618420e-01 3.31327319e-01 -3.13392639e-01
-1.34218884e+00 3.41843426e-01 6.65271664e+00 4.25156444e-01
-1.36110532e+00 3.74248475e-01 2.55732238e-01 1.65266484e-01
-9.52151567e-02 -2.43198365e-01 -6.48565516e-02 7.17007995e-01
1.51799512e+00 -4.85625446e-01 3.84538680e-01 9.48510587e-01
5.85878134e-01 6.46579802e-01 -1.23410249e+00 1.47654629e+00
-1.82661638e-01 -8.30170214e-01 -1.98984653e-01 6.32420406e-02
9.09568965e-01 3.59438539e-01 -6.69375509e-02 3.71982902e-01
5.62716797e-02 -7.56598175e-01 3.67441088e-01 5.55506647e-01
6.15000188e-01 -5.22273540e-01 4.21089530e-01 3.06162000e-01
-1.58441198e+00 -4.44316894e-01 -2.06924245e-01 -5.33893347e-01
1.95195183e-01 9.42376554e-01 -6.06626749e-01 4.98269767e-01
7.94099331e-01 1.21577227e+00 -4.69132662e-01 9.08017755e-01
3.27732444e-01 7.96225905e-01 -6.42418623e-01 4.48219180e-02
8.39321688e-02 -1.70739904e-01 3.27267379e-01 1.28512657e+00
6.97856069e-01 1.67959318e-01 3.79870415e-01 3.29003125e-01
2.04268947e-01 1.91677436e-01 -1.08183134e+00 -3.02934051e-01
2.73835748e-01 1.28286850e+00 -7.97836363e-01 -5.19849718e-01
-5.72350919e-01 4.16249216e-01 -3.03942919e-01 3.31907213e-01
-8.49839568e-01 -3.28644186e-01 1.96597427e-01 1.85122862e-01
2.83590108e-01 -4.38959062e-01 5.27636781e-02 -1.11200321e+00
1.80309817e-01 -4.76986915e-01 4.06041086e-01 -7.48259425e-01
-1.90462422e+00 4.59211498e-01 1.72072887e-01 -2.12748742e+00
-5.07694542e-01 -4.76949900e-01 -7.06544757e-01 3.38734202e-02
-1.35875356e+00 -1.12907863e+00 -2.66937286e-01 1.34526932e+00
4.83491033e-01 -4.93466944e-01 5.98474205e-01 7.67249286e-01
-9.58538279e-02 8.69559273e-02 1.76549613e-01 -4.19019349e-02
4.35035586e-01 -1.08870184e+00 3.77666444e-01 5.41499138e-01
1.44605830e-01 4.35883760e-01 4.50281948e-01 -4.77917403e-01
-1.80204368e+00 -1.60762608e+00 6.07576251e-01 -3.55231792e-01
1.09180701e+00 -3.25484842e-01 -7.15251267e-01 7.72485614e-01
7.42176101e-02 2.56568134e-01 1.11861122e+00 -1.68607086e-02
-3.53418559e-01 -6.75717354e-01 -1.33198166e+00 1.96951747e-01
9.71822321e-01 -7.44648039e-01 -9.61547077e-01 5.87046802e-01
7.00438023e-01 3.83254886e-01 -1.36323190e+00 5.26021123e-01
4.90690261e-01 -5.48125148e-01 1.02580917e+00 -6.04122281e-01
-4.11437936e-02 -3.39511573e-01 -4.49160814e-01 -1.04238236e+00
-2.71187186e-01 -9.16940928e-01 -4.41899747e-01 1.33935666e+00
4.66501676e-02 -7.83182263e-01 5.00332594e-01 7.02936798e-02
-8.09762776e-02 -4.75400425e-02 -8.93557727e-01 -7.84006774e-01
-2.87815988e-01 -9.00159001e-01 8.31089318e-01 1.34588063e+00
1.77116573e-01 4.16794598e-01 -4.82840419e-01 4.94766142e-03
6.48295939e-01 -1.18359916e-01 7.77431965e-01 -1.90176332e+00
2.15801988e-02 5.37490696e-02 -1.01294911e+00 -5.98130822e-01
-2.84470588e-01 -6.32690668e-01 -1.92095608e-01 -1.12729585e+00
-4.30522174e-01 -4.78789747e-01 -8.19417179e-01 3.95828843e-01
4.74722713e-01 3.48645449e-01 -1.61336780e-01 3.09119225e-01
-8.34618211e-01 3.61877114e-01 7.30421126e-01 -2.90170550e-01
-5.14203489e-01 1.97018459e-01 -3.13655943e-01 8.17911267e-01
1.13741541e+00 -4.95259345e-01 -9.55293715e-01 -1.72903627e-01
9.39938873e-02 -5.06565683e-02 1.91500962e-01 -1.42430687e+00
3.85847420e-01 -2.16795087e-01 4.21209931e-01 -6.81001604e-01
-7.45027661e-02 -1.56937349e+00 5.57635248e-01 2.83457965e-01
-6.08368181e-02 3.09353769e-01 2.48544231e-01 7.32346714e-01
-1.61672965e-01 4.94214028e-01 3.67927462e-01 7.08443895e-02
-8.08162332e-01 4.29690927e-01 -4.21356589e-01 -2.36205399e-01
1.27436197e+00 -4.37457919e-01 2.03526944e-01 -3.24180961e-01
-1.10851085e+00 -1.98244914e-01 -3.47727835e-02 8.94581854e-01
3.80833209e-01 -1.72117293e+00 -3.53470236e-01 5.68664432e-01
2.95907438e-01 -3.36278796e-01 4.77233939e-02 9.13964808e-01
-2.16861051e-02 2.48440117e-01 -5.22937417e-01 -9.94895279e-01
-7.44055092e-01 9.76805091e-01 -3.22477013e-01 4.20922088e-03
-7.75469244e-01 -3.77288647e-02 -4.99945253e-01 -6.67305052e-01
1.92418069e-01 -7.44712353e-01 -4.59122151e-01 1.29240215e-01
5.86443424e-01 6.02173269e-01 1.75902262e-01 -6.84212983e-01
-4.38883215e-01 9.17204976e-01 4.04657543e-01 1.07262030e-01
1.65638173e+00 -2.57854402e-01 -1.09866753e-01 1.07690811e+00
1.50476158e+00 -1.99206606e-01 -8.09713900e-01 -1.33184388e-01
4.35252547e-01 1.30060226e-01 -3.88471216e-01 -2.08389610e-01
-8.39670956e-01 5.79360545e-01 1.03634596e+00 1.03072178e+00
1.36189699e+00 -1.25557020e-01 7.94718146e-01 7.44540632e-01
8.74155879e-01 -1.21666956e+00 -2.32168268e-02 3.10212016e-01
4.24472213e-01 -1.11191380e+00 -7.85735697e-02 -8.78004916e-03
-2.12444603e-01 9.73466158e-01 1.55385584e-02 -4.88370270e-01
1.35065424e+00 4.92009908e-01 7.48447888e-03 -2.33394671e-02
-8.18457842e-01 -4.44042891e-01 1.56480476e-01 1.07418525e+00
3.68748367e-01 6.71833009e-02 -1.98677033e-01 5.42756319e-01
-4.26356643e-01 6.19061708e-01 1.97497532e-01 1.28978968e+00
-2.21368656e-01 -1.06305790e+00 -3.50145519e-01 7.34789729e-01
-4.72760648e-01 3.47519755e-01 2.05359813e-02 5.24949670e-01
4.41176116e-01 1.35790384e+00 3.31185222e-01 -7.14688241e-01
4.68138009e-01 2.02534899e-01 1.40520677e-01 -3.63697946e-01
-4.76304799e-01 1.51422635e-01 -2.94208467e-01 -5.93245149e-01
-1.11980867e+00 -4.91434216e-01 -1.28926420e+00 7.84240663e-03
1.68096483e-01 1.29555106e-01 7.69913435e-01 1.09752584e+00
5.47310710e-01 7.52033412e-01 1.16478503e+00 -1.00122702e+00
-5.48189618e-02 -9.74444032e-01 -5.94996095e-01 6.49066687e-01
3.73834997e-01 -4.05844390e-01 -3.35838109e-01 6.62845552e-01] | [7.218612194061279, 2.858477830886841] |
03a99316-16cb-4173-8cd3-bbb51a71ac1c | deep-keyphrase-completion | 2111.01910 | null | https://arxiv.org/abs/2111.01910v1 | https://arxiv.org/pdf/2111.01910v1.pdf | Deep Keyphrase Completion | Keyphrase provides accurate information of document content that is highly compact, concise, full of meanings, and widely used for discourse comprehension, organization, and text retrieval. Though previous studies have made substantial efforts for automated keyphrase extraction and generation, surprisingly, few studies have been made for \textit{keyphrase completion} (KPC). KPC aims to generate more keyphrases for document (e.g. scientific publication) taking advantage of document content along with a very limited number of known keyphrases, which can be applied to improve text indexing system, etc. In this paper, we propose a novel KPC method with an encoder-decoder framework. We name it \textit{deep keyphrase completion} (DKPC) since it attempts to capture the deep semantic meaning of the document content together with known keyphrases via a deep learning framework. Specifically, the encoder and the decoder in DKPC play different roles to make full use of the known keyphrases. The former considers the keyphrase-guiding factors, which aggregates information of known keyphrases into context. On the contrary, the latter considers the keyphrase-inhibited factor to inhibit semantically repeated keyphrase generation. Extensive experiments on benchmark datasets demonstrate the efficacy of our proposed model. | ['Ji Liu', 'Xiaojie Wang', 'Qing Li', 'Fuzhen Zhuang', 'Huali Feng', 'Jia Song', 'Yu Zhao'] | 2021-10-29 | null | null | null | null | ['keyphrase-generation'] | ['natural-language-processing'] | [ 1.17143989e-01 8.25541094e-02 -2.40947381e-01 3.92712027e-01
-5.91582477e-01 -6.08140826e-01 1.07447898e+00 6.22091770e-01
-4.58512843e-01 7.72711277e-01 1.05029392e+00 -2.22436562e-01
-3.09682965e-01 -9.23201501e-01 -7.78300941e-01 -6.95021808e-01
4.87731814e-01 7.37215206e-02 1.28715634e-01 -3.63836974e-01
7.05998242e-01 2.53000021e-01 -1.35060775e+00 5.73407114e-01
7.15938091e-01 7.21622288e-01 7.98111200e-01 4.54312027e-01
-6.91386044e-01 1.21681392e+00 -9.55537558e-01 -1.56408653e-01
-2.11216763e-01 -4.00406539e-01 -8.96445394e-01 -7.84708001e-03
1.98759094e-01 -6.90810025e-01 -9.55241621e-01 9.76734698e-01
3.92352849e-01 3.20824683e-02 5.46615124e-01 -1.04461873e+00
-8.71050358e-01 1.20998096e+00 -3.57860088e-01 4.04082596e-01
3.26361090e-01 -3.69873047e-01 1.08712411e+00 -8.05623770e-01
7.08651364e-01 1.20479476e+00 -4.45027947e-02 9.77007225e-02
-5.24204373e-01 -5.89771926e-01 3.15452740e-02 2.73230940e-01
-1.46173441e+00 -1.94071040e-01 9.46011484e-01 -1.70252100e-01
8.90151501e-01 3.63806933e-01 7.45770156e-01 8.93528819e-01
1.44886985e-01 1.35058308e+00 8.08086336e-01 -4.28750277e-01
-1.10322148e-01 8.87561440e-02 2.68171817e-01 5.78198969e-01
1.18464991e-01 -3.41623932e-01 -6.35215223e-01 -5.21994429e-03
8.15287590e-01 7.12477863e-02 -5.07423460e-01 1.87549472e-01
-1.65483892e+00 6.14353299e-01 2.42997900e-01 4.88251626e-01
-5.75178683e-01 -4.86545376e-02 4.80885118e-01 2.87491113e-01
3.54279540e-02 6.94936931e-01 -2.80967474e-01 -2.10245684e-01
-8.06437016e-01 7.14460611e-01 7.48659790e-01 1.05720568e+00
5.19262016e-01 -3.87886286e-01 -8.21006119e-01 7.50363767e-01
-2.09941976e-02 4.90741730e-01 7.03643978e-01 -6.98125601e-01
7.51509666e-01 8.86305928e-01 4.73852083e-02 -1.40935504e+00
-1.91635609e-01 -4.10947025e-01 -9.59341168e-01 -7.20297873e-01
-1.82547852e-01 -2.86840405e-02 -5.75886607e-01 1.27670646e+00
1.23779871e-01 -1.33522138e-01 3.23167622e-01 5.70606232e-01
1.10249591e+00 1.25896907e+00 -3.43089998e-01 -3.31596136e-01
1.47593415e+00 -1.07693934e+00 -1.07525647e+00 1.75789744e-02
4.66767043e-01 -1.10328043e+00 9.14503336e-01 2.95420736e-01
-7.00122058e-01 -4.95706677e-01 -1.05339909e+00 -5.60176909e-01
-5.91444671e-01 5.44874966e-01 5.08441389e-01 1.43986018e-02
-6.51546478e-01 3.19914967e-01 -2.61593431e-01 6.88346177e-02
3.87748957e-01 -2.40804136e-01 -9.42899063e-02 -1.01485901e-01
-1.33411133e+00 5.31172037e-01 9.40441668e-01 -2.02619240e-01
-8.11781406e-01 -7.69510925e-01 -5.90592384e-01 1.82781518e-01
8.68613243e-01 -5.36531687e-01 1.12457371e+00 -3.40029597e-01
-1.12857676e+00 5.43198109e-01 -9.83442590e-02 -4.03048575e-01
1.18267231e-01 -5.93760669e-01 -3.98836374e-01 5.34504950e-01
3.21473300e-01 6.61652923e-01 9.88789260e-01 -9.01732147e-01
-7.25740612e-01 -3.76810692e-02 3.74253064e-01 6.11081541e-01
-9.10031974e-01 7.91017190e-02 -7.12564826e-01 -1.29963064e+00
6.76689222e-02 -6.10691547e-01 3.98453951e-01 -2.99581766e-01
-7.83205688e-01 -5.18880248e-01 8.27378690e-01 -1.08397448e+00
1.81164813e+00 -1.99732792e+00 3.97270441e-01 -1.09890059e-01
4.95314211e-01 1.95016950e-01 6.98646232e-02 1.01336646e+00
1.70818806e-01 1.89561918e-01 -5.31183463e-03 3.92350644e-01
-3.15965116e-02 3.97073925e-02 -9.82275486e-01 -1.99211821e-01
-9.22626704e-02 9.89857197e-01 -9.85310495e-01 -6.58801675e-01
1.11879252e-01 9.96512100e-02 -2.59197325e-01 3.02311927e-01
-5.34376621e-01 -1.10486396e-01 -9.60963309e-01 4.47057486e-01
3.26421499e-01 -4.26007181e-01 -2.50503153e-01 -5.49353063e-01
-1.85340002e-01 4.74367768e-01 -9.75502014e-01 1.42344773e+00
-3.95682573e-01 6.93002939e-01 -2.58268237e-01 -8.38819563e-01
7.37242579e-01 3.72759908e-01 4.71458763e-01 -6.35789156e-01
9.25267860e-02 1.23409539e-01 -4.37514246e-01 -4.48988974e-01
1.06957316e+00 1.99393630e-01 -9.16571021e-02 6.68168008e-01
-8.61722603e-02 -2.30554447e-01 3.27864826e-01 8.13655138e-01
8.70597303e-01 -1.78400412e-01 3.65257174e-01 -1.48751616e-01
6.89342260e-01 2.86615249e-02 -4.05524187e-02 8.05128992e-01
4.18448567e-01 2.94246852e-01 5.87019086e-01 -2.13664547e-01
-1.02793932e+00 -5.68809688e-01 1.39005437e-01 9.59226489e-01
4.25864786e-01 -9.84247684e-01 -6.36680186e-01 -6.15798712e-01
-1.17960535e-01 8.77295613e-01 -4.63895559e-01 -4.24573720e-01
-7.04321027e-01 -4.77835894e-01 8.32878888e-01 4.56424087e-01
7.23599732e-01 -1.15088725e+00 -3.47073019e-01 3.51479471e-01
-5.04915118e-01 -1.18019044e+00 -6.32716775e-01 -2.18251169e-01
-4.26114708e-01 -1.01639152e+00 -1.18672085e+00 -7.92876065e-01
6.40616536e-01 6.77838206e-01 9.01941299e-01 1.26963750e-01
-3.48236226e-02 4.04679060e-01 -8.11866879e-01 -4.46024686e-01
-4.40588981e-01 1.37818143e-01 -3.42484564e-01 -1.50039509e-01
2.90783167e-01 -1.84344053e-01 -4.22214568e-01 -1.61612868e-01
-1.42720652e+00 6.23412967e-01 8.27322304e-01 8.39269102e-01
5.74626684e-01 6.16004825e-01 3.74188304e-01 -6.16986454e-01
1.24983203e+00 -4.35838580e-01 -2.98637956e-01 4.87028003e-01
-5.54928184e-01 3.40422064e-01 8.56948078e-01 -3.10213298e-01
-1.06169224e+00 -7.77286887e-01 7.30660558e-02 -3.65728259e-01
3.55413675e-01 7.77886987e-01 -1.54791325e-01 5.13740182e-01
2.37395793e-01 1.04724383e+00 -3.69496167e-01 -6.41909540e-01
5.29164553e-01 8.04532766e-01 4.91493553e-01 -7.95761883e-01
5.45653582e-01 3.40124696e-01 -1.86604723e-01 -8.83624852e-01
-9.92662311e-01 -5.18753886e-01 -2.51785547e-01 -1.65474355e-01
7.54839897e-01 -9.42681015e-01 -3.93084675e-01 6.40499592e-01
-1.25980473e+00 3.62741798e-01 -2.64867097e-01 2.65974998e-01
-1.61377743e-01 6.69431269e-01 -5.52935898e-01 -3.82155508e-01
-6.31317675e-01 -8.77710700e-01 1.26330304e+00 4.96735841e-01
-1.54197391e-03 -6.96902454e-01 -2.98790991e-01 3.48234385e-01
5.75249903e-02 -1.06208868e-01 1.31350255e+00 -7.91648507e-01
-7.38081694e-01 -2.62111664e-01 -4.80562329e-01 3.11113119e-01
3.52665365e-01 -3.41942221e-01 -6.00746095e-01 -2.08905473e-01
-3.17112841e-02 -5.28046608e-01 1.16909397e+00 -4.25193459e-02
1.58115768e+00 -7.84089446e-01 -3.23536843e-01 1.34407550e-01
9.49043751e-01 2.55691081e-01 5.87510526e-01 4.84851956e-01
1.01015878e+00 5.46896636e-01 4.84354407e-01 5.77985108e-01
6.67029679e-01 5.03753662e-01 6.43201992e-02 2.06641644e-01
-2.62925267e-01 -6.98114753e-01 2.13257670e-01 1.41124797e+00
1.60938770e-01 -3.58804464e-01 -7.06490397e-01 5.82605600e-01
-1.62994432e+00 -8.02639782e-01 2.64516771e-01 1.66556621e+00
1.16825986e+00 1.97620004e-01 -4.52197105e-01 1.85525745e-01
5.31085193e-01 6.46658838e-01 -3.76521856e-01 2.06501499e-01
-4.02824521e-01 3.05499285e-02 3.39124769e-01 1.74860746e-01
-9.50763166e-01 1.12124395e+00 4.46122837e+00 1.29880834e+00
-7.46901453e-01 -3.21681798e-01 3.42576414e-01 2.34997645e-01
-5.59987545e-01 7.88356811e-02 -1.03095210e+00 4.20567393e-01
1.93800926e-01 -6.11939251e-01 4.44244385e-01 7.91912198e-01
-2.64034607e-02 -4.18192483e-02 -8.22882533e-01 1.04928541e+00
1.81075603e-01 -1.72377086e+00 7.76360393e-01 -9.01954472e-02
6.51302457e-01 -5.95409155e-01 -1.45949006e-01 2.74490595e-01
8.15189853e-02 -6.72324002e-01 7.97872066e-01 7.89895654e-01
3.77872914e-01 -6.91703320e-01 5.94094038e-01 4.62675601e-01
-1.09525001e+00 -5.55407554e-02 -4.65469509e-01 1.62766457e-01
-2.25738555e-01 6.91753030e-01 -7.52561629e-01 7.07222760e-01
3.12464029e-01 8.58423531e-01 -5.20430148e-01 5.58640957e-01
-3.53895634e-01 4.60420936e-01 -1.35362759e-01 -3.76154661e-01
5.90870798e-01 1.38896823e-01 6.90795600e-01 1.27677381e+00
3.79372329e-01 3.81184757e-01 3.75413060e-01 7.19203115e-01
-4.28075254e-01 4.75592285e-01 -2.83025324e-01 -8.49817574e-01
8.25418830e-01 1.08460903e+00 -9.04941738e-01 -8.69475543e-01
-1.56285003e-01 1.03784728e+00 8.09332728e-02 2.90136456e-01
-2.11390808e-01 -6.52197540e-01 3.02209556e-01 -1.81227803e-01
2.52605438e-01 -2.30533883e-01 1.00133844e-01 -1.30418777e+00
3.44043225e-01 -1.12989879e+00 3.89651418e-01 -9.46920156e-01
-9.22903001e-01 2.28995651e-01 4.39191550e-01 -9.13948238e-01
-1.28713652e-01 -4.08307761e-01 -3.75014305e-01 8.32718551e-01
-1.45146286e+00 -1.14315903e+00 -1.48650393e-01 4.65388477e-01
9.96136844e-01 -4.30164218e-01 6.54211819e-01 -7.82820731e-02
-3.44338745e-01 2.67135143e-01 3.32037687e-01 3.23716283e-01
3.35091442e-01 -1.17431486e+00 2.32442096e-01 8.20012808e-01
2.16475144e-01 9.83488739e-01 5.07540643e-01 -8.42876792e-01
-1.71120560e+00 -8.89659762e-01 9.69060540e-01 -1.34368032e-01
6.03063345e-01 -7.35995770e-02 -9.21292067e-01 2.56072283e-01
3.26670110e-01 -6.91006184e-01 4.67596263e-01 -3.71778756e-01
-2.93946058e-01 -1.33882239e-01 -3.13657433e-01 1.02999008e+00
5.72861373e-01 -8.04057956e-01 -1.12952006e+00 6.15980923e-01
1.31761348e+00 -4.62677896e-01 -5.82042217e-01 1.97371677e-01
4.33685422e-01 -4.63573843e-01 1.05099618e+00 -3.43807787e-01
8.90441775e-01 -2.70233542e-01 -1.35628343e-01 -1.16776848e+00
-1.58205867e-01 -6.19189382e-01 -6.96913004e-01 1.20590794e+00
-8.83189738e-02 -1.83097392e-01 2.60679901e-01 9.90206748e-02
-2.12937444e-01 -6.94130182e-01 -5.72929919e-01 -4.13236231e-01
-4.76589575e-02 -2.59168863e-01 9.09001291e-01 9.95398819e-01
-5.11330727e-04 6.15105689e-01 -5.08216977e-01 -1.33554712e-01
1.70886233e-01 4.33408022e-01 6.65621221e-01 -7.56200492e-01
-1.73428252e-01 -6.91380799e-01 1.01754300e-01 -1.62713802e+00
-1.32144690e-01 -9.59977090e-01 -3.76184434e-01 -1.76134336e+00
3.82964522e-01 -1.13268182e-01 -2.47090459e-01 4.90937233e-01
-5.72916508e-01 -5.01075149e-01 2.96419948e-01 5.66502750e-01
-4.77540076e-01 7.89612412e-01 1.77418494e+00 -4.99081343e-01
-1.98951721e-01 -2.97723204e-01 -1.08082747e+00 5.49116671e-01
5.61210394e-01 -3.34466100e-01 -6.39042854e-01 -3.72196943e-01
6.08908176e-01 2.05841720e-01 3.87454212e-01 -6.34382129e-01
4.97166902e-01 -3.10563534e-01 4.49486047e-01 -9.14036930e-01
8.85495767e-02 -3.98029447e-01 -4.08633649e-01 2.15372682e-01
-7.68072724e-01 1.00708790e-01 2.21171364e-01 4.77401793e-01
-3.70759189e-01 -3.70057851e-01 6.69250414e-02 -4.83934164e-01
-7.71355331e-01 3.48295607e-02 -4.55804050e-01 2.78900146e-01
4.01971608e-01 2.17727602e-01 -3.62637013e-01 -3.63534808e-01
-6.46742508e-02 1.77982271e-01 -1.10724252e-02 8.17010880e-01
9.49824810e-01 -1.27393508e+00 -9.13447380e-01 8.41660798e-02
3.13050389e-01 1.55354619e-01 1.88317299e-02 2.72872239e-01
-5.11513412e-01 9.42669392e-01 5.55188060e-02 6.04723692e-02
-1.04931676e+00 6.09024704e-01 -2.90688962e-01 -4.79497403e-01
-9.91588056e-01 7.91495502e-01 3.68329853e-01 -1.34237692e-01
2.70545781e-01 -5.64074695e-01 -7.46254623e-01 4.21268433e-01
1.04598308e+00 2.60198593e-01 -3.58665586e-02 -5.39020896e-01
2.56934106e-01 1.80416331e-01 -7.55700707e-01 -4.15201597e-02
9.89095151e-01 -2.01338887e-01 -5.06335437e-01 3.44787747e-01
1.23873031e+00 8.87797698e-02 -7.17958927e-01 -6.78357124e-01
-4.84956708e-03 -2.14809820e-01 4.65201885e-01 -7.91637063e-01
-6.55403137e-01 7.06506431e-01 5.87430485e-02 1.28742427e-01
1.15275550e+00 -4.74718250e-02 1.28045964e+00 9.04864430e-01
6.25749901e-02 -1.25171471e+00 4.66302425e-01 7.32677400e-01
1.12380087e+00 -7.46172547e-01 2.97697186e-01 -1.98197529e-01
-5.37320197e-01 1.30804086e+00 4.71500635e-01 3.39378595e-01
4.73020434e-01 -6.95233271e-02 -1.91417545e-01 -4.47392166e-01
-5.02720356e-01 1.00828204e-02 5.00474811e-01 6.32231757e-02
8.54242966e-02 -1.02742705e-02 -4.83335823e-01 8.78644943e-01
-5.31677961e-01 -5.12920208e-02 4.87251163e-01 1.10431516e+00
-7.17898488e-01 -8.34174097e-01 -3.69621277e-01 7.58220196e-01
-5.03929019e-01 -5.71617782e-01 -6.39176786e-01 4.04851496e-01
-2.02933978e-02 6.83750868e-01 -1.61571860e-01 -2.64956594e-01
-7.02679157e-02 6.86618313e-02 2.03855634e-01 -5.17182171e-01
-3.28400880e-01 1.96057320e-01 -6.51577637e-02 -5.85816130e-02
-3.04185688e-01 -3.21218431e-01 -1.25198841e+00 -1.50513262e-01
-2.34004825e-01 3.92003894e-01 4.46984559e-01 1.19734240e+00
2.83973426e-01 8.79132092e-01 6.30498230e-01 -1.82465240e-01
-4.53907251e-01 -1.10290051e+00 -3.85288477e-01 3.65141660e-01
2.52852768e-01 -3.49868268e-01 1.56082511e-01 2.71080852e-01] | [12.31525993347168, 8.892391204833984] |
1b0a549d-1d16-4e06-9ed7-9a5acbdb66d9 | zeronlg-aligning-and-autoencoding-domains-for | 2303.06458 | null | https://arxiv.org/abs/2303.06458v1 | https://arxiv.org/pdf/2303.06458v1.pdf | ZeroNLG: Aligning and Autoencoding Domains for Zero-Shot Multimodal and Multilingual Natural Language Generation | Natural Language Generation (NLG) accepts input data in the form of images, videos, or text and generates corresponding natural language text as output. Existing NLG methods mainly adopt a supervised approach and rely heavily on coupled data-to-text pairs. However, for many targeted scenarios and for non-English languages, sufficient quantities of labeled data are often not available. To relax the dependency on labeled data of downstream tasks, we propose an intuitive and effective zero-shot learning framework, ZeroNLG, which can deal with multiple NLG tasks, including image-to-text (image captioning), video-to-text (video captioning), and text-to-text (neural machine translation), across English, Chinese, German, and French within a unified framework. ZeroNLG does not require any labeled downstream pairs for training. During training, ZeroNLG (i) projects different domains (across modalities and languages) to corresponding coordinates in a shared common latent space; (ii) bridges different domains by aligning their corresponding coordinates in this space; and (iii) builds an unsupervised multilingual auto-encoder to learn to generate text by reconstructing the input text given its coordinate in shared latent space. Consequently, during inference, based on the data-to-text pipeline, ZeroNLG can generate target sentences across different languages given the coordinate of input data in the common space. Within this unified framework, given visual (imaging or video) data as input, ZeroNLG can perform zero-shot visual captioning; given textual sentences as input, ZeroNLG can perform zero-shot machine translation. We present the results of extensive experiments on twelve NLG tasks, showing that, without using any labeled downstream pairs for training, ZeroNLG generates high-quality and believable outputs and significantly outperforms existing zero-shot methods. | ['David A. Clifton', 'YaoWei Wang', 'Xian Wu', 'Yuexian Zou', 'Fenglin Liu', 'Bang Yang'] | 2023-03-11 | null | null | null | null | ['zero-shot-machine-translation'] | ['natural-language-processing'] | [ 4.90535915e-01 1.74025521e-01 -2.32709467e-01 -1.89356774e-01
-1.20398462e+00 -7.05643952e-01 9.92701590e-01 -4.28402454e-01
-1.29732609e-01 8.64437938e-01 3.16577941e-01 -4.16019112e-01
5.70520520e-01 -8.35593462e-01 -1.18931603e+00 -6.35860026e-01
7.57499337e-01 5.46039879e-01 -1.50934473e-01 1.51327088e-01
-1.68780714e-01 -5.89127727e-02 -1.23317111e+00 4.89638835e-01
7.44447351e-01 6.95733786e-01 5.86072028e-01 6.78064227e-01
-2.61809170e-01 6.36679113e-01 -4.13190454e-01 -4.48982984e-01
2.80369639e-01 -9.58520234e-01 -7.22363114e-01 3.33393425e-01
4.14569020e-01 -4.48700190e-01 -2.73027927e-01 9.33389127e-01
4.26656693e-01 -2.42399480e-02 8.00789535e-01 -1.44619322e+00
-1.14305294e+00 4.89600033e-01 -2.87344337e-01 -5.23723304e-01
4.83131677e-01 5.84289610e-01 8.63581955e-01 -1.25976384e+00
1.22599816e+00 1.20176339e+00 1.21291608e-01 8.86131048e-01
-1.49648607e+00 -4.98450696e-01 -1.41150832e-01 -2.26455331e-01
-1.22490740e+00 -5.21314085e-01 5.10716677e-01 -5.97162604e-01
7.20147491e-01 -3.73270921e-02 4.82971519e-01 1.73728526e+00
6.52275458e-02 7.41042674e-01 8.80403161e-01 -5.88167250e-01
2.11626127e-01 9.65012088e-02 -5.00992537e-01 6.76318586e-01
-1.03738263e-01 1.22879200e-01 -7.77186394e-01 5.93218803e-02
8.30618203e-01 -2.31639996e-01 -3.51777583e-01 -3.59894663e-01
-1.88067162e+00 8.76340210e-01 1.10553771e-01 2.76741743e-01
-9.79569405e-02 2.06743479e-01 2.19466940e-01 3.07507545e-01
3.91270816e-01 3.27023864e-01 5.19774528e-03 1.93091296e-02
-9.51332748e-01 1.02349063e-02 6.17132008e-01 1.43264461e+00
9.56215143e-01 1.93308279e-01 -5.19674301e-01 5.78457475e-01
2.79486626e-01 9.11660194e-01 6.95380092e-01 -8.81895840e-01
9.03920889e-01 2.83660144e-01 1.68386772e-01 -6.62395000e-01
2.14319304e-01 1.81814447e-01 -9.26623583e-01 -7.30416551e-02
2.63385892e-01 -4.17176515e-01 -1.10840464e+00 2.03963804e+00
1.94141865e-01 2.66148508e-01 6.73106492e-01 9.01262760e-01
9.33940053e-01 1.20527720e+00 -4.44809683e-02 -2.33978108e-01
1.29923558e+00 -1.06418514e+00 -6.16890073e-01 -3.44179928e-01
7.12137103e-01 -7.43541777e-01 1.46933997e+00 -5.52447401e-02
-9.95838106e-01 -7.94542372e-01 -8.61492515e-01 -4.34074581e-01
-2.58372813e-01 2.72035033e-01 -1.34369908e-02 1.84901878e-01
-9.96766746e-01 1.00975320e-01 -6.03600919e-01 -5.52239835e-01
1.14560500e-01 -7.83065036e-02 -6.59708381e-01 -4.76516038e-01
-1.31113338e+00 7.22815692e-01 4.00507599e-01 -1.20931983e-01
-1.24743879e+00 -4.72744167e-01 -1.23354757e+00 -1.06709026e-01
2.45695487e-01 -1.14208770e+00 1.22356713e+00 -1.14118862e+00
-1.56527972e+00 9.61807132e-01 -2.59247720e-01 -2.60477066e-01
7.04171598e-01 1.76079795e-02 -2.57397711e-01 2.52252460e-01
4.48681265e-01 1.25786388e+00 1.15032339e+00 -1.26181948e+00
-2.27634296e-01 -1.26797423e-01 -4.06454168e-02 3.57945412e-01
-1.74787164e-01 -1.43504724e-01 -6.53110981e-01 -7.65439630e-01
-1.80244789e-01 -9.99874771e-01 1.11372033e-02 2.53365695e-01
-3.82738233e-01 -1.28834084e-01 8.72988582e-01 -6.98028982e-01
6.39663875e-01 -2.26148295e+00 6.02636755e-01 -3.12943995e-01
2.20476706e-02 2.92018671e-02 -5.30083299e-01 5.22734284e-01
1.70668215e-02 -1.24809749e-01 -4.49128062e-01 -5.99850118e-01
1.67458933e-02 3.47396851e-01 -5.77512383e-01 2.27267399e-01
2.22040296e-01 1.37312365e+00 -1.11883581e+00 -6.74293518e-01
3.85578513e-01 3.32072765e-01 -3.33352894e-01 5.93067288e-01
-5.36775529e-01 5.41869879e-01 -2.30157822e-01 3.75189990e-01
2.34102711e-01 -4.11209017e-01 1.13143295e-01 -1.97431117e-01
7.97891319e-02 -1.73502058e-01 -7.53779054e-01 2.22329926e+00
-7.55340278e-01 8.34724963e-01 -2.55501390e-01 -1.01696908e+00
9.83091950e-01 7.89861858e-01 2.20746070e-01 -4.56893057e-01
-4.14160714e-02 2.77551800e-01 -4.84884828e-01 -6.61159635e-01
1.94059938e-01 -4.21853215e-01 -4.65475798e-01 8.28166783e-01
6.10704064e-01 -3.98118883e-01 1.31422907e-01 4.31624591e-01
7.05534041e-01 5.05885303e-01 1.05222806e-01 2.21531957e-01
2.76122004e-01 9.40300077e-02 2.50406623e-01 5.66258371e-01
2.10841760e-01 9.55037355e-01 4.00850981e-01 -2.06523925e-01
-1.46716630e+00 -1.39782715e+00 2.75509655e-01 8.50147665e-01
4.53359224e-02 -2.15073630e-01 -9.19785857e-01 -5.36011219e-01
-3.07306916e-01 1.03020251e+00 -5.75984180e-01 -1.77743882e-01
-1.56835869e-01 -2.48087853e-01 6.99092448e-01 4.43337679e-01
4.25364882e-01 -1.24746919e+00 -4.31396157e-01 8.87442157e-02
-6.66475534e-01 -1.52637684e+00 -7.64922857e-01 -1.42869905e-01
-6.43700778e-01 -7.32662380e-01 -1.21872294e+00 -9.99307871e-01
9.28971350e-01 4.22766089e-01 8.98711860e-01 -4.20209169e-01
-1.63207844e-01 5.19339979e-01 -3.40253264e-01 -1.93247423e-02
-8.34207773e-01 -1.92894444e-01 9.45171714e-02 2.40309745e-01
1.14727072e-01 -3.13681334e-01 -1.07066430e-01 2.23744929e-01
-1.21803880e+00 8.82255733e-01 5.54575145e-01 9.50267851e-01
5.01450658e-01 -7.07666039e-01 5.14158130e-01 -7.28097379e-01
5.53191185e-01 -5.04767120e-01 -5.90579510e-01 5.66913307e-01
-1.89983830e-01 1.83264613e-01 8.59239519e-01 -6.74864173e-01
-1.10374951e+00 2.04466522e-01 2.03766093e-01 -1.02886081e+00
-2.00290486e-01 4.13252503e-01 -3.74268621e-01 5.35889447e-01
7.13238060e-01 6.10751390e-01 9.53556299e-02 -1.16146915e-01
9.02535737e-01 8.50432515e-01 8.05294394e-01 -6.12625778e-01
1.02470374e+00 3.43377948e-01 -2.56487310e-01 -8.30661178e-01
-6.91833973e-01 -4.26540338e-02 -7.60218382e-01 -2.01768547e-01
1.40912688e+00 -1.20044482e+00 1.75676029e-02 4.74203497e-01
-1.47248614e+00 -5.23345530e-01 -3.56136888e-01 5.66325366e-01
-9.52831864e-01 2.59509563e-01 -4.93592292e-01 -4.24126416e-01
-3.24842185e-01 -1.17345726e+00 1.36490130e+00 -9.89240929e-02
-2.07384109e-01 -1.05448985e+00 -5.00459634e-02 4.84195590e-01
-1.19068399e-02 2.42910951e-01 9.42461252e-01 -3.36100489e-01
-5.89586496e-01 4.67461385e-02 -2.97340691e-01 3.79054993e-01
1.95516780e-01 -1.14706725e-01 -7.41356015e-01 -2.33064890e-01
-1.20620668e-01 -7.62660205e-01 4.55949605e-01 2.70550191e-01
7.72723913e-01 -3.44274700e-01 -2.40791336e-01 7.10231304e-01
1.33558691e+00 -1.95300151e-02 5.93443632e-01 -1.75582319e-01
7.69510508e-01 6.99639678e-01 5.59035420e-01 5.57123683e-02
3.69270623e-01 6.49048984e-01 9.65625048e-03 -3.21837962e-01
-2.78886557e-01 -7.94424355e-01 7.21870542e-01 7.74407446e-01
1.96019545e-01 -4.24063325e-01 -6.57001019e-01 6.57483160e-01
-1.93112445e+00 -9.83530521e-01 8.27119723e-02 2.01136231e+00
1.01188159e+00 -2.30939224e-01 -2.34382361e-01 -5.88894725e-01
9.60652053e-01 1.04947664e-01 -6.34406567e-01 -2.51787007e-01
-1.55961782e-01 8.45788419e-03 1.23545192e-01 5.14219761e-01
-6.85231149e-01 1.28278673e+00 5.47164917e+00 6.77343369e-01
-1.26553333e+00 3.57998997e-01 6.76110327e-01 -1.41047940e-01
-6.10320210e-01 1.43305272e-01 -6.25218093e-01 5.18020928e-01
8.88882160e-01 -4.17635649e-01 4.83840913e-01 7.19144404e-01
2.35157534e-01 1.88085705e-01 -1.35246623e+00 1.22764933e+00
3.36043537e-01 -1.60506058e+00 4.25966978e-01 -1.10348105e-01
1.02495873e+00 -2.49433592e-02 -1.18396021e-01 3.64311248e-01
4.72512573e-01 -1.01164341e+00 8.88660133e-01 5.41934311e-01
1.47479081e+00 -1.60362646e-01 3.70431930e-01 6.05145574e-01
-8.95317733e-01 3.52513731e-01 -4.19180632e-01 1.84720993e-01
5.73023796e-01 3.04452747e-01 -7.46346891e-01 6.57147646e-01
2.33315572e-01 7.16682911e-01 -1.68439776e-01 2.01747373e-01
-6.18470430e-01 3.73615026e-01 3.00349556e-02 1.22437559e-01
3.24739963e-01 -3.24177802e-01 4.18518215e-01 1.06982875e+00
8.44411254e-01 2.22073570e-02 2.64896899e-01 1.27085924e+00
-2.36734614e-01 3.97944041e-02 -1.06777394e+00 -3.67073447e-01
3.28856170e-01 1.09467804e+00 -4.94901776e-01 -7.30199933e-01
-6.30619049e-01 1.48102927e+00 6.35996163e-02 7.69273520e-01
-9.58094597e-01 -3.38045239e-01 3.90339255e-01 7.17919469e-02
6.34944579e-03 -2.09451020e-01 -5.80043122e-02 -1.61493731e+00
2.32504867e-02 -8.02283764e-01 5.68834245e-02 -1.49349833e+00
-1.07156563e+00 7.63211250e-01 1.39354765e-01 -1.61906695e+00
-7.27638602e-01 -4.23739433e-01 -4.69482183e-01 1.01623356e+00
-1.00331795e+00 -1.50616729e+00 -3.34363163e-01 9.73283410e-01
8.86579573e-01 -3.97034109e-01 1.01024115e+00 -5.22865281e-02
-1.90710261e-01 3.85509044e-01 9.18465666e-03 3.17463636e-01
9.66693401e-01 -8.57403100e-01 6.31645143e-01 1.00128794e+00
4.83688116e-01 3.69982421e-01 4.53527957e-01 -6.22893631e-01
-1.37528110e+00 -1.34838605e+00 9.58982766e-01 -3.07915717e-01
6.60216868e-01 -6.53444707e-01 -6.17128074e-01 9.22487795e-01
4.20181990e-01 1.30597338e-01 6.01370394e-01 -5.39887547e-01
-4.21466768e-01 1.52921006e-01 -8.48881245e-01 7.61314750e-01
9.31198537e-01 -8.34579825e-01 -7.53641009e-01 5.73030651e-01
1.14416540e+00 -4.28838700e-01 -5.23359716e-01 -2.83698756e-02
3.38764369e-01 -5.73142588e-01 6.00944996e-01 -6.99574947e-01
1.07323420e+00 -3.79581779e-01 -3.07896793e-01 -1.42066145e+00
3.11824009e-02 -5.85867584e-01 5.10865971e-02 1.13902676e+00
5.29090405e-01 -4.22001690e-01 3.82404834e-01 3.89127463e-01
-1.34187952e-01 -4.04653460e-01 -8.95438671e-01 -6.71630144e-01
4.66703661e-02 -4.77403432e-01 4.17924553e-01 9.92242932e-01
-1.33360863e-01 8.13260078e-01 -8.14469814e-01 -8.35798755e-02
4.89770025e-01 3.72116119e-01 1.18672168e+00 -7.86659122e-01
-4.26008850e-01 -4.05470841e-03 -7.88060948e-02 -1.24860334e+00
3.00526381e-01 -1.13385725e+00 4.00128663e-01 -1.90383255e+00
2.13497058e-01 5.44829480e-02 3.11101526e-01 7.40314543e-01
-2.06537209e-02 4.36199307e-01 4.41865236e-01 3.31248879e-01
-2.97724217e-01 6.58887863e-01 1.44943547e+00 -2.81754911e-01
3.88683043e-02 -3.56707573e-01 -6.14062428e-01 4.38178867e-01
3.20025653e-01 -3.16017419e-01 -5.40356159e-01 -7.96266735e-01
3.72518189e-02 5.75544834e-01 4.51430827e-01 -8.14919412e-01
3.50021869e-02 -4.37494785e-01 3.93627375e-01 -1.45812184e-01
3.45936298e-01 -6.13726795e-01 2.65099347e-01 3.05736780e-01
-3.17248106e-01 -1.49555340e-01 -1.48135558e-01 4.26957995e-01
-2.99983472e-01 -2.70321637e-01 7.24682391e-01 -1.64515391e-01
-6.19282603e-01 3.09768558e-01 -3.66124332e-01 5.31053506e-02
1.29027581e+00 -2.47561097e-01 -2.64047772e-01 -6.42702639e-01
-8.42107773e-01 1.56417668e-01 6.84511304e-01 7.62440801e-01
7.94788480e-01 -1.64020634e+00 -8.92696202e-01 4.01979357e-01
3.98100644e-01 9.43912417e-02 1.64260089e-01 5.23560882e-01
-3.31404507e-01 4.64924604e-01 -2.36661956e-01 -8.18291306e-01
-1.03038311e+00 6.51598930e-01 2.16857181e-03 -3.71483453e-02
-6.16604090e-01 5.94273627e-01 6.89222157e-01 -6.00300729e-01
-1.60758898e-01 -1.27974004e-01 3.42437506e-01 -1.58337560e-02
4.30860400e-01 -2.40632445e-01 -3.43155682e-01 -7.42586911e-01
1.29526258e-01 5.67102492e-01 3.34791243e-01 -7.52144814e-01
1.13784325e+00 -1.20722175e-01 5.94588481e-02 6.46552145e-01
1.28012812e+00 -2.26147130e-01 -1.43977082e+00 -2.44203791e-01
-4.13269460e-01 -3.33530843e-01 -3.56139898e-01 -4.78203475e-01
-8.56202304e-01 1.23373842e+00 1.79159865e-01 -3.65935385e-01
1.01173604e+00 1.85577676e-01 9.41481590e-01 2.78837979e-01
4.48784083e-01 -9.45216119e-01 4.37043607e-01 4.80573416e-01
8.64698350e-01 -1.21799850e+00 -4.39049035e-01 -1.17231615e-01
-9.82321322e-01 1.04942083e+00 6.39411509e-01 1.91043079e-01
1.74210161e-01 5.32651730e-02 2.21102774e-01 2.36463532e-01
-9.48685050e-01 3.61094773e-02 2.21390858e-01 6.86265528e-01
2.02104300e-01 3.43803875e-02 -8.71526375e-02 4.15062487e-01
-1.77286059e-01 1.94164172e-01 4.43096668e-01 6.66097283e-01
-1.88500449e-01 -1.19818151e+00 -4.13547665e-01 9.36616883e-02
8.03214610e-02 -2.60532707e-01 -2.51781881e-01 5.28770745e-01
1.46343157e-01 7.54724979e-01 2.17308253e-01 -2.02945739e-01
-1.10083379e-01 2.66339093e-01 5.23093879e-01 -9.80696678e-01
-4.07767259e-02 2.18207806e-01 -1.99876592e-01 -2.80631304e-01
-3.82313132e-01 -4.58290279e-01 -1.22530329e+00 5.37098050e-02
-4.39501256e-02 2.12585241e-01 5.33496916e-01 1.12643170e+00
4.22797292e-01 2.82026708e-01 4.76778537e-01 -8.53943706e-01
-3.22191507e-01 -1.03695393e+00 -2.66236812e-01 6.37348711e-01
1.31648466e-01 -3.77500772e-01 -2.07969323e-01 8.19427788e-01] | [11.290642738342285, 1.032114028930664] |
6b514fee-c17f-4ffe-9ceb-869f0b524a2c | sliced-at-semeval-2022-task-11-bigger-better | null | null | https://aclanthology.org/2022.semeval-1.205 | https://aclanthology.org/2022.semeval-1.205.pdf | Sliced at SemEval-2022 Task 11: Bigger, Better? Massively Multilingual LMs for Multilingual Complex NER on an Academic GPU Budget | Massively multilingual language models (MMLMs) have become a widely-used representation method, and multiple large MMLMs were proposed in recent years. A trend is to train MMLMs on larger text corpora or with more layers. In this paper we set out to test recent popular MMLMs on detecting semantically ambiguous and complex named entities with an academic GPU budget. Our submission of a single model for 11 languages on the SemEval Task 11 MultiCoNER shows that a vanilla transformer-CRF with XLM-R_{large} outperforms the more recent RemBERT, ranking 9th from 26 submissions in the multilingual track. Compared to RemBERT, the XLM-R model has the additional advantage to fit on a slice of a multi-instance GPU. As contrary to expectations and recent findings, we found RemBERT to not be the best MMLM, we further set out to investigate this discrepancy with additional experiments on multilingual Wikipedia NER data. While we expected RemBERT to have an edge on that dataset as it is closer to its pre-training data, surprisingly, our results show that this is not the case, suggesting that text domain match does not explain the discrepancy. | ['Barbara Plank'] | null | null | null | null | semeval-naacl-2022-7 | ['xlm-r'] | ['natural-language-processing'] | [-5.37863970e-01 9.99385789e-02 1.19876832e-01 -3.49695891e-01
-1.12928617e+00 -6.35542750e-01 7.68013597e-01 2.35810727e-01
-9.83720541e-01 1.02779198e+00 3.79567802e-01 -6.09736979e-01
3.16893220e-01 -5.83898306e-01 -9.36051846e-01 -5.68169989e-02
1.36925206e-01 9.67142105e-01 1.33297130e-01 -2.01978907e-01
1.04964897e-01 2.57672101e-01 -1.08260119e+00 2.62405813e-01
9.27160919e-01 9.47428644e-02 5.47520876e-01 5.04072189e-01
-3.17960829e-01 1.03579891e+00 -5.64759552e-01 -7.60102987e-01
1.46578237e-01 -1.59729943e-02 -1.15159833e+00 -5.50258517e-01
7.69129992e-01 9.76756513e-02 -4.03407514e-02 8.04359555e-01
7.36242235e-01 1.17695138e-01 4.50834572e-01 -8.52670670e-01
-4.72409368e-01 1.00388920e+00 -4.83664304e-01 3.10844898e-01
1.70425341e-01 -4.73711081e-02 9.76692796e-01 -1.05155706e+00
1.25125110e+00 1.26618624e+00 1.03916335e+00 2.24497333e-01
-1.00994158e+00 -5.91076016e-01 2.68483069e-02 8.62606913e-02
-1.55886710e+00 -4.30963844e-01 -3.26687321e-02 -2.41655394e-01
1.81886148e+00 1.96933076e-02 2.41526276e-01 9.24904466e-01
2.80846804e-01 8.24331522e-01 1.55788755e+00 -5.18009722e-01
-1.34302676e-01 3.20352644e-01 -4.37029935e-02 6.90961957e-01
4.71661747e-01 -2.01598287e-01 -3.03050637e-01 -2.79770285e-01
3.15722466e-01 -6.42571747e-01 -1.06851682e-01 6.34290278e-02
-1.38217342e+00 7.20563233e-01 2.88472742e-01 6.08712435e-01
-3.86490315e-01 1.46756880e-02 6.05250537e-01 2.71453083e-01
5.25318563e-01 7.54304111e-01 -7.78973758e-01 -4.55195844e-01
-1.14024007e+00 6.48005158e-02 1.05689573e+00 8.39106262e-01
6.04350686e-01 -9.47485641e-02 8.83623883e-02 8.39830518e-01
1.30705327e-01 3.42055827e-01 4.82703507e-01 -6.67624235e-01
7.32136130e-01 4.29662079e-01 -1.25209242e-01 -6.42186522e-01
-7.04353571e-01 -5.47962070e-01 -4.72204208e-01 -8.13744031e-03
6.40930355e-01 -3.30851346e-01 -5.32326877e-01 1.67923760e+00
9.19233486e-02 -1.42981485e-01 1.93519101e-01 8.28627467e-01
6.41151369e-01 7.81238794e-01 4.77281123e-01 1.61347345e-01
1.28107345e+00 -9.19794679e-01 -3.61084074e-01 -3.96235943e-01
1.21009076e+00 -1.28930581e+00 8.99825990e-01 3.46477419e-01
-9.19895232e-01 -4.26846623e-01 -8.93049359e-01 -3.89507741e-01
-7.16543257e-01 3.28679413e-01 8.59950840e-01 4.93193835e-01
-1.30764806e+00 4.59120274e-01 -9.06264305e-01 -8.20508003e-01
-1.34185106e-01 2.49288693e-01 -5.59478939e-01 -1.14255875e-01
-1.24076068e+00 1.56238139e+00 5.78733087e-01 3.70685644e-02
-5.69412589e-01 -4.72271174e-01 -7.58145690e-01 -1.21207610e-01
-4.70369048e-02 -4.59261268e-01 1.11812150e+00 -7.18846619e-01
-1.06113064e+00 1.24085474e+00 -2.08058327e-01 -6.12670958e-01
5.24508536e-01 -1.90547943e-01 -5.63122451e-01 -3.38956445e-01
3.15911978e-01 7.79394805e-01 -8.28243122e-02 -9.83437121e-01
-6.18296623e-01 -2.50930756e-01 -4.64125443e-03 2.00748220e-01
-3.14730629e-02 5.11719942e-01 -1.50554344e-01 -4.60559428e-01
-2.10957929e-01 -9.25862968e-01 -2.67236650e-01 -8.52936029e-01
-3.66904885e-01 -4.19794232e-01 2.62582779e-01 -9.14852321e-01
1.08874643e+00 -1.69654095e+00 -1.91628501e-01 -3.61281484e-02
-1.14872735e-02 4.62299436e-01 -1.38733253e-01 7.10554898e-01
-6.92385435e-02 1.81415454e-01 -2.94730365e-02 -6.36828005e-01
1.12848707e-01 9.13153738e-02 -1.75597951e-01 5.33999443e-01
9.35754403e-02 8.16878676e-01 -8.31244886e-01 -7.00798094e-01
-8.18670094e-02 7.38468707e-01 -3.74830335e-01 -7.14408010e-02
-1.78834468e-01 3.35045278e-01 -1.30686164e-01 3.27832848e-01
6.85198843e-01 -3.73567432e-01 4.65387523e-01 4.13525552e-02
-6.65554583e-01 7.43128896e-01 -1.09988320e+00 1.76485848e+00
-7.00995982e-01 8.74737322e-01 -3.38411443e-02 -5.22106826e-01
7.54226565e-01 4.84654576e-01 3.05310469e-02 -1.02292633e+00
-2.48658136e-01 8.71647954e-01 6.16771318e-02 -2.22456977e-01
9.82251823e-01 -6.85930857e-03 -2.11186588e-01 4.70434576e-01
2.32979819e-01 1.65267393e-01 4.18637037e-01 2.96858549e-01
1.10727644e+00 4.91683275e-01 3.07300180e-01 -4.78914112e-01
3.34904522e-01 3.96407157e-01 5.56156993e-01 7.07364559e-01
1.08145453e-01 5.41125119e-01 9.75775421e-02 -4.88097668e-01
-1.20834613e+00 -8.01726282e-01 -2.97890127e-01 1.24547362e+00
-3.47842366e-01 -5.27860165e-01 -7.69777119e-01 -6.71790302e-01
-4.29190457e-01 1.01159883e+00 -4.67897803e-02 4.92464960e-01
-9.57749844e-01 -1.02178681e+00 1.00757492e+00 3.63753289e-01
3.92868996e-01 -1.17212665e+00 -4.74093378e-01 4.75774616e-01
-3.00578028e-01 -1.26056826e+00 -2.50750154e-01 4.89685684e-01
-5.23186028e-01 -7.60271370e-01 -6.59823656e-01 -1.05041778e+00
5.24131119e-01 -2.45020077e-01 1.61911523e+00 -1.81913987e-01
-9.42289680e-02 1.40815541e-01 -3.07370126e-01 -2.51620442e-01
-7.22677827e-01 5.61495900e-01 2.11895872e-02 -7.59185970e-01
7.21917033e-01 -4.33272243e-01 -4.07052487e-01 3.78824323e-02
-6.24660969e-01 6.84116110e-02 6.74352109e-01 6.96008563e-01
4.92890298e-01 -3.85079384e-01 4.61920410e-01 -1.01007843e+00
4.59651142e-01 -5.24562538e-01 -7.02424705e-01 3.80158246e-01
-6.30067766e-01 2.02934906e-01 7.52328157e-01 -1.08332105e-01
-1.08434010e+00 -1.43000960e-01 -5.03534794e-01 1.41781703e-01
-1.83056504e-01 7.05711603e-01 2.67367698e-02 9.40878391e-02
6.98539257e-01 1.33342519e-01 -5.85472286e-01 -7.07781136e-01
2.83062249e-01 7.44139314e-01 3.78770083e-01 -7.90234387e-01
2.99639016e-01 1.61201626e-01 -3.93889993e-01 -8.08911026e-01
-7.54003525e-01 -3.86107892e-01 -4.49124575e-01 1.79063603e-01
1.02892363e+00 -1.31957853e+00 -3.86840373e-01 1.56711817e-01
-1.45628262e+00 -5.62718332e-01 6.61840960e-02 6.31635308e-01
-2.48722434e-01 2.34916821e-01 -1.08595347e+00 -5.51484525e-01
-4.38353300e-01 -1.22344494e+00 9.97084916e-01 -5.38865738e-02
-3.39212120e-01 -1.20186770e+00 4.96649384e-01 3.76142114e-01
5.92641056e-01 -3.00741252e-02 8.22810650e-01 -1.07804978e+00
-3.64139318e-01 -3.44110206e-02 -2.89259374e-01 1.51810572e-01
-3.82759422e-01 -2.59363681e-01 -8.46868753e-01 -3.96022201e-01
-2.41783038e-01 -4.39232618e-01 7.52856791e-01 6.33628946e-03
4.65607762e-01 -5.79487085e-02 -3.19909513e-01 4.11721498e-01
1.62915289e+00 -1.15839899e-01 5.65728843e-01 8.47806692e-01
6.75505698e-01 2.97550112e-01 4.44960684e-01 -2.74424311e-02
9.40725744e-01 7.50577211e-01 1.15412720e-01 -3.51382405e-01
-2.03178525e-01 -3.32172930e-01 6.94437087e-01 1.29238939e+00
-2.75512170e-02 -3.20232034e-01 -1.36443079e+00 6.33201063e-01
-1.54261637e+00 -6.41089261e-01 -4.18788940e-01 2.23997521e+00
7.98825026e-01 1.84703752e-01 -8.73742476e-02 -6.95805907e-01
6.87018991e-01 -1.39369788e-02 -4.58592549e-02 -7.03984618e-01
-5.46938539e-01 3.20293128e-01 7.36666322e-01 4.29324001e-01
-8.70372295e-01 1.19930208e+00 6.01685476e+00 8.51557851e-01
-1.12085211e+00 5.27300417e-01 5.93270421e-01 2.69130468e-02
-1.84828758e-01 4.08030808e-01 -1.26419854e+00 4.03836548e-01
1.33256459e+00 4.43878509e-02 2.81038493e-01 5.89018881e-01
2.01979186e-02 -1.55032128e-01 -9.49521065e-01 5.91811657e-01
-6.70760050e-02 -1.01856852e+00 -1.07335292e-01 2.07759842e-01
8.11147213e-01 1.02246714e+00 -3.46475512e-01 7.84309626e-01
8.27498019e-01 -1.20850003e+00 6.02732182e-01 4.22362119e-01
5.53548634e-01 -6.76879585e-01 9.10274386e-01 5.15964746e-01
-1.08225024e+00 4.62840974e-01 -6.34981871e-01 -2.58927979e-02
3.62858683e-01 4.43876237e-01 -9.69847500e-01 5.72613657e-01
5.88655412e-01 2.58713424e-01 -8.37652683e-01 1.12388706e+00
-7.11760223e-02 6.77169859e-01 -4.32959914e-01 9.98186991e-02
5.17287314e-01 2.03042068e-02 4.89303082e-01 1.81701040e+00
3.88547868e-01 -4.41510141e-01 1.80116832e-01 5.25112033e-01
-2.54603416e-01 4.99768823e-01 -4.95022327e-01 -8.92082080e-02
4.18910116e-01 1.42363107e+00 -6.92157090e-01 -3.38209569e-01
-7.35675693e-01 1.01916170e+00 9.47850227e-01 5.41478768e-02
-7.68882930e-01 -1.52550772e-01 2.64246404e-01 6.07428029e-02
2.75774330e-01 -4.30133879e-01 3.66601311e-02 -1.24661863e+00
-4.18092608e-02 -1.06314015e+00 5.62594414e-01 -7.26112843e-01
-1.19770646e+00 9.71944451e-01 -3.96100283e-01 -8.88959289e-01
-2.76299804e-01 -6.58020616e-01 -1.86937392e-01 1.07831645e+00
-1.63358140e+00 -1.24560893e+00 2.81933337e-01 2.87387401e-01
4.35330451e-01 -1.17850177e-01 1.06611407e+00 6.50811911e-01
-4.14956927e-01 6.22763991e-01 2.69990206e-01 2.84458339e-01
1.06108022e+00 -1.36237848e+00 6.62666440e-01 8.51083815e-01
4.47916955e-01 9.07072961e-01 6.74529374e-01 -7.15806127e-01
-1.16205013e+00 -1.02418399e+00 1.67188776e+00 -7.24976659e-01
8.25694323e-01 -3.84076118e-01 -8.36832345e-01 1.10985923e+00
8.06376576e-01 -2.36925334e-01 5.17167985e-01 3.14198464e-01
-3.09510469e-01 3.85355294e-01 -8.75231564e-01 4.16868031e-01
9.36165929e-01 -5.14669716e-01 -6.47319078e-01 4.98286784e-01
4.73043710e-01 -6.38787210e-01 -1.09494650e+00 3.08008194e-01
1.52953416e-01 -8.39768052e-01 5.79322040e-01 -4.38182175e-01
4.03903693e-01 -2.22925797e-01 -2.35397667e-01 -1.35225022e+00
-8.44765455e-02 -3.80185783e-01 3.46524090e-01 1.58972919e+00
8.36214006e-01 -7.43145764e-01 4.53911006e-01 3.48279089e-01
-2.22663656e-01 -5.24753690e-01 -9.90940750e-01 -7.83404052e-01
5.33180416e-01 -5.55499017e-01 3.71830910e-01 1.22803736e+00
-4.38851044e-02 5.29555976e-01 -2.52554029e-01 1.33791253e-01
4.26826954e-01 -1.43063277e-01 5.65937579e-01 -9.69176710e-01
-3.80451083e-01 -3.26934010e-01 -1.98023677e-01 -8.73277366e-01
4.17719245e-01 -1.39787543e+00 2.25249957e-02 -1.77082491e+00
4.26939428e-01 -6.56698406e-01 -1.82332590e-01 6.19490504e-01
-1.34507567e-01 2.38992304e-01 1.64672360e-01 2.30343029e-01
-8.79896224e-01 9.02607366e-02 7.84799635e-01 8.29491913e-02
4.81522270e-02 -3.78630042e-01 -5.23442447e-01 6.51008010e-01
5.65391719e-01 -5.44445693e-01 2.33161896e-01 -8.47250044e-01
7.88071752e-01 -1.08734801e-01 1.33278947e-02 -1.01630974e+00
3.53527516e-01 4.73674119e-01 3.73078734e-01 -4.65348303e-01
5.77064678e-02 -4.90846068e-01 2.02971295e-01 2.43903473e-01
-1.28719106e-01 5.66898823e-01 5.00421524e-01 -1.03379890e-01
-3.09868991e-01 -2.61690229e-01 4.56154644e-01 -4.46004421e-01
-8.96186769e-01 -1.02307945e-01 -6.00503743e-01 3.65979940e-01
4.30706054e-01 1.76916897e-01 -5.08871853e-01 4.16698232e-02
-6.11257017e-01 2.64447361e-01 5.75107574e-01 5.05231023e-01
-9.16957483e-02 -9.16979134e-01 -1.05731833e+00 -3.83893013e-01
3.51520739e-02 -2.17098683e-01 -1.42784708e-03 9.49791610e-01
-6.71012640e-01 9.11630690e-01 7.00034797e-02 -3.93401831e-01
-9.68045712e-01 2.77790934e-01 4.27767158e-01 -9.48176384e-01
-5.80306411e-01 8.28350186e-01 -7.77468160e-02 -1.11514461e+00
-1.50504395e-01 7.86636025e-02 -3.53283323e-02 3.07102017e-02
2.84796417e-01 3.84092748e-01 5.04251063e-01 -8.46055329e-01
-5.40331244e-01 1.37186885e-01 -3.82773280e-01 -1.81264818e-01
1.48642027e+00 -9.85902622e-02 -3.17535102e-01 4.21221048e-01
1.13632989e+00 4.08879369e-01 -7.56246388e-01 -1.36157081e-01
4.52826560e-01 1.81284845e-01 -9.29835960e-02 -1.07887983e+00
-7.67456412e-01 6.08410895e-01 5.10005057e-01 -1.17331713e-01
7.37784743e-01 1.78435014e-03 8.06760788e-01 4.07463551e-01
8.36533904e-01 -1.21313441e+00 -6.10750556e-01 9.71604943e-01
5.33088148e-01 -1.18939686e+00 -9.03471187e-02 9.42085311e-02
-6.40948832e-01 8.84416461e-01 6.50540411e-01 5.85783906e-02
1.75624534e-01 4.81877863e-01 1.94750607e-01 -2.87042052e-01
-7.52098143e-01 -2.08440855e-01 -4.92608324e-02 2.12792158e-01
1.17384160e+00 9.97287333e-02 -6.56824350e-01 2.78835624e-01
-5.20987689e-01 -7.28437528e-02 4.19699281e-01 7.91260242e-01
-1.59335554e-01 -1.44489050e+00 -2.67273486e-01 2.89380252e-01
-8.14248025e-01 -6.78384483e-01 -2.20375299e-01 1.08567345e+00
1.61154896e-01 7.95907199e-01 4.03790250e-02 3.60875949e-02
2.68792450e-01 2.97881097e-01 5.01889467e-01 -6.56176209e-01
-1.12083018e+00 -1.99474886e-01 5.09627223e-01 -3.51817936e-01
-2.05809042e-01 -7.90136814e-01 -1.36003661e+00 -3.70358527e-01
-2.28984922e-01 3.74252826e-01 9.56820548e-01 1.03902233e+00
5.05833328e-01 1.53995141e-01 1.14660896e-02 -6.08210623e-01
-3.26217085e-01 -1.16354084e+00 -3.34698290e-01 1.68787360e-01
-2.18650535e-01 -2.29886740e-01 -2.71772534e-01 -3.40249330e-01] | [10.368569374084473, 9.887960433959961] |
5faa19e8-d610-45c7-bebf-1ed44ba0935a | machine-learning-for-faster-and-smarter | 2008.02320 | null | https://arxiv.org/abs/2008.02320v1 | https://arxiv.org/pdf/2008.02320v1.pdf | Machine learning for faster and smarter fluorescence lifetime imaging microscopy | Fluorescence lifetime imaging microscopy (FLIM) is a powerful technique in biomedical research that uses the fluorophore decay rate to provide additional contrast in fluorescence microscopy. However, at present, the calculation, analysis, and interpretation of FLIM is a complex, slow, and computationally expensive process. Machine learning (ML) techniques are well suited to extract and interpret measurements from multi-dimensional FLIM data sets with substantial improvement in speed over conventional methods. In this topical review, we first discuss the basics of FILM and ML. Second, we provide a summary of lifetime extraction strategies using ML and its applications in classifying and segmenting FILM images with higher accuracy compared to conventional methods. Finally, we discuss two potential directions to improve FLIM with ML with proof of concept demonstrations. | ['Xiao-Tong Yuan', 'Scott S. Howard', 'Varun Mannam', 'Yide Zhang', 'Cara Ravasio'] | 2020-08-05 | null | null | null | null | ['lifetime-image-denoising'] | ['medical'] | [ 8.57044220e-01 -8.72968435e-01 -5.29090278e-02 -3.92005980e-01
-1.00265872e+00 -7.50672162e-01 1.47856683e-01 4.62123692e-01
-8.13965440e-01 1.17565322e+00 -4.63252246e-01 -1.08671300e-01
1.77591577e-01 -2.85755306e-01 -3.07546347e-01 -1.38482881e+00
1.13482349e-01 2.26578698e-01 3.32337201e-01 6.56444728e-01
5.74679077e-01 1.18978786e+00 -1.32438922e+00 4.92508829e-01
1.97994933e-01 8.90082598e-01 2.63921022e-01 1.06068206e+00
-5.82509488e-02 6.71810091e-01 -8.54362369e-01 2.42256999e-01
-5.51390111e-01 -5.67476749e-01 -8.49356771e-01 -1.26887754e-01
2.44421139e-01 -3.10119927e-01 1.66843515e-02 6.32447064e-01
9.37613785e-01 7.82434344e-02 1.00704777e+00 -8.84456694e-01
-7.19811097e-02 -7.47809112e-02 -7.54724920e-01 1.07034743e+00
4.77394462e-01 5.96381575e-02 9.17159691e-02 -6.99087560e-01
8.13035071e-01 8.54804337e-01 4.26375806e-01 8.24051321e-01
-1.44796610e+00 -5.03360450e-01 -3.19798917e-01 1.99068874e-01
-9.08189118e-01 -4.57308948e-01 4.52834457e-01 -6.39776587e-01
1.16919041e+00 1.90277696e-01 6.85084701e-01 6.77281857e-01
8.39368463e-01 7.04750896e-01 1.81944680e+00 -7.08507836e-01
1.28730088e-01 -1.24290712e-01 4.74906385e-01 6.08360052e-01
-1.67818293e-01 1.26001030e-01 -3.75759631e-01 -2.96319336e-01
6.48524225e-01 1.30293537e-02 -2.03628406e-01 2.36708611e-01
-1.18263614e+00 5.94742239e-01 -3.71972263e-01 4.09678876e-01
-1.03004381e-01 1.06274411e-01 3.54219943e-01 3.44841480e-02
5.11511624e-01 4.18988913e-01 -2.34307915e-01 -1.14812493e-01
-1.21469486e+00 -2.59358194e-02 2.49171048e-01 1.76387757e-01
5.28986216e-01 -1.91453263e-01 -8.02560523e-02 4.35795933e-01
4.30977978e-02 6.42166138e-01 5.44579506e-01 -1.28574657e+00
-7.34487653e-01 1.15082018e-01 3.40207636e-01 -1.18943334e-01
-7.16498911e-01 3.50557715e-01 -2.42812842e-01 5.71936429e-01
7.11695015e-01 -1.54720992e-01 -9.43111479e-01 1.13996077e+00
2.37869158e-01 -5.65621518e-02 -1.74036577e-01 5.04669607e-01
7.81068325e-01 6.99608982e-01 1.82774052e-01 -9.97067571e-01
1.27665031e+00 -5.81217825e-01 -1.10652483e+00 4.02574092e-01
4.65745509e-01 -7.14842558e-01 6.61268890e-01 5.55004895e-01
-1.21256828e+00 -7.18921199e-02 -1.04986989e+00 -3.11011821e-01
-4.90312874e-01 -3.93693112e-02 8.02278936e-01 7.31648028e-01
-8.18953395e-01 6.93295181e-01 -1.14187801e+00 -2.01694578e-01
5.73476017e-01 6.05605364e-01 -3.99099767e-01 -1.24700174e-01
-4.91432399e-01 7.87631869e-01 -9.88722499e-03 -3.00604939e-01
-7.31439829e-01 -8.66037607e-01 -4.32300210e-01 -7.86613762e-01
-4.93094862e-01 -3.34012777e-01 1.27472675e+00 -3.49211693e-01
-1.82394528e+00 1.26432002e+00 -9.69250858e-01 -2.81437546e-01
7.40673095e-02 1.44664913e-01 -2.63155878e-01 8.69682074e-01
-1.81053311e-01 8.22933197e-01 3.86924148e-01 -8.50431561e-01
-6.78456783e-01 -4.31254864e-01 -3.06059867e-01 -3.89601678e-01
1.47877425e-01 4.17280734e-01 2.04421416e-01 1.06078461e-01
-2.33848721e-01 -5.47325194e-01 1.94588259e-01 2.94805914e-01
6.26316853e-03 -1.32661358e-01 1.17286372e+00 -3.78821492e-01
8.44840944e-01 -1.97930598e+00 -2.65205443e-01 -2.82577753e-01
3.13689053e-01 3.76412034e-01 1.01820350e-01 4.77919608e-01
3.05126552e-02 -6.41752481e-02 -1.64073721e-01 -2.06061214e-01
-6.66187644e-01 -4.53323349e-02 2.45652925e-02 1.12387621e+00
-1.40169814e-01 9.12408352e-01 -1.09729671e+00 -7.10016012e-01
5.98701119e-01 7.42020130e-01 3.75493556e-01 4.55361344e-02
-1.09959215e-01 1.16002727e+00 -8.65464211e-02 1.08252275e+00
6.80385768e-01 -5.91318607e-01 6.92549795e-02 -6.53333008e-01
-6.35078967e-01 -3.19235116e-01 -1.63061693e-01 1.26644993e+00
-1.14368461e-01 1.05849922e+00 -1.72453523e-01 -8.89433444e-01
5.33422470e-01 3.19437146e-01 9.72412109e-01 -6.91061199e-01
3.86803031e-01 3.40736926e-01 -3.49907875e-01 -8.56620669e-01
-2.27377594e-01 -7.34971344e-01 6.84687972e-01 8.02587748e-01
-1.45754427e-01 -3.51529792e-02 4.59644198e-01 -3.07771266e-01
9.98789549e-01 9.85910371e-03 2.66769797e-01 -2.89320081e-01
6.07365966e-01 -1.26282007e-01 4.67625745e-02 4.48373407e-01
-5.74640930e-01 4.06377047e-01 1.57337070e-01 -6.01610720e-01
-8.87682140e-01 -9.52738523e-01 -4.98832554e-01 6.38424754e-01
1.92699596e-01 2.44631290e-01 -8.66942763e-01 -4.22735065e-01
-1.54866695e-01 1.32695228e-01 -5.11349678e-01 4.69983608e-01
-4.25604731e-01 -1.34377027e+00 4.91593331e-01 3.26577574e-01
7.03964457e-02 -9.95419919e-01 -1.07324445e+00 2.88794130e-01
-2.06053540e-01 -1.38181114e+00 -1.00938573e-01 2.46088922e-01
-9.55132484e-01 -1.23382556e+00 -9.84481990e-01 -6.32844031e-01
7.36026883e-01 5.18075049e-01 5.33724368e-01 8.71113166e-02
-9.79884863e-01 5.11146128e-01 -1.20885231e-01 -6.30895019e-01
-1.83446914e-01 -4.26146716e-01 -1.42529324e-01 -3.55325460e-01
6.50746107e-01 -6.05607748e-01 -9.77258623e-01 1.11581862e-01
-9.50498700e-01 -2.16228485e-01 4.09119368e-01 6.99614346e-01
1.33141577e+00 -1.58968531e-02 3.97656918e-01 -9.53371108e-01
2.52407730e-01 5.91067821e-02 -4.96926874e-01 1.60341367e-01
-5.47045112e-01 -2.92631894e-01 3.15908074e-01 -6.88383162e-01
-1.09575033e+00 6.65186569e-02 -5.34613840e-02 5.96860461e-02
-3.28517854e-01 2.19822228e-01 2.90491998e-01 -6.74946666e-01
6.36393189e-01 1.27641976e-01 6.35000885e-01 -1.85517043e-01
-1.33815244e-01 7.14472830e-01 3.64067405e-01 -4.01158571e-01
-6.75706118e-02 1.48163748e+00 3.04747850e-01 -1.30203605e+00
-8.59742105e-01 -7.43994236e-01 -5.96822917e-01 -4.51938450e-01
7.82826304e-01 -4.64523166e-01 -9.59748089e-01 9.54385281e-01
-9.97754514e-01 -5.32478988e-01 1.24539016e-02 7.13726103e-01
-6.58255398e-01 5.81942260e-01 -1.18388844e+00 -9.72676158e-01
-6.47380054e-01 -1.20390069e+00 1.20286608e+00 6.80702448e-01
-1.44382924e-01 -1.32060194e+00 3.37061554e-01 5.49537778e-01
2.69727021e-01 9.12406862e-01 9.68367517e-01 1.02706917e-01
-3.14203113e-01 -3.19598645e-01 -3.10100228e-01 -3.53920841e-05
4.16857034e-01 4.49348390e-01 -1.43887341e+00 -5.86546600e-01
1.94794938e-01 -3.51284534e-01 8.46418619e-01 1.10206234e+00
1.30554116e+00 4.27646309e-01 -9.04730439e-01 6.12918496e-01
1.57498837e+00 7.29303837e-01 7.01728344e-01 1.11284941e-01
6.92692176e-02 5.98880351e-01 6.04306102e-01 3.05687308e-01
-4.19315159e-01 3.03039223e-01 -7.84813091e-02 -2.74748772e-01
-2.41122484e-01 4.47063625e-01 4.34517682e-01 4.99907374e-01
-1.29303008e-01 -4.57256049e-01 -2.93802679e-01 2.45735481e-01
-1.20514059e+00 -9.87505317e-01 -3.18524957e-01 1.94749677e+00
7.70879447e-01 -1.74062297e-01 3.31389844e-01 2.25789964e-01
7.63612866e-01 -2.50884712e-01 -7.32165992e-01 -2.30159163e-01
-4.36034650e-01 4.61051166e-01 4.08040911e-01 4.97218460e-01
-8.87914419e-01 5.48466206e-01 8.37285614e+00 6.87185407e-01
-1.61307096e+00 8.05213824e-02 5.86300015e-01 -1.40127286e-01
-1.19026013e-01 -1.93841219e-01 -9.45545018e-01 5.78863442e-01
1.22163475e+00 6.20956607e-02 1.86684951e-01 -2.13346541e-01
6.13505781e-01 -7.53385305e-01 -1.12756288e+00 1.20949209e+00
-4.85100336e-02 -1.67915392e+00 -9.02471691e-02 1.40741721e-01
3.29766810e-01 -1.24615373e-03 9.23124328e-02 -6.01895690e-01
-5.05004227e-01 -1.05625331e+00 2.19660312e-01 7.38438427e-01
1.33382845e+00 -5.29393673e-01 7.86623061e-01 1.50201783e-01
-7.86486924e-01 1.22418001e-01 -4.47857469e-01 5.65227456e-02
5.48360050e-01 1.03344512e+00 -4.49559122e-01 1.02272719e-01
5.81161320e-01 5.39286733e-01 -2.09043369e-01 1.17031991e+00
2.55425274e-01 3.69764179e-01 -2.29586884e-01 -7.05265328e-02
-7.70320073e-02 -1.63035259e-01 1.72239602e-01 1.57039893e+00
2.23055482e-01 2.79327899e-01 -2.58593380e-01 6.29770577e-01
3.67907137e-01 -2.94774979e-01 -2.95523018e-01 -6.48150861e-01
4.28012282e-01 1.49262083e+00 -1.23360085e+00 -1.82367295e-01
-5.10606289e-01 7.95777678e-01 -1.40242308e-01 3.49397838e-01
-6.08560205e-01 -5.98911107e-01 3.21251273e-01 3.71603936e-01
-3.43346089e-01 -2.08566085e-01 3.71653080e-01 -5.34713387e-01
-4.38062757e-01 -2.02857003e-01 3.54908496e-01 -8.07296395e-01
-1.40736759e+00 8.81145373e-02 9.79328007e-02 -6.67851925e-01
2.25659132e-01 -1.09163761e+00 -4.57392335e-01 6.91196918e-01
-1.92464912e+00 -1.05123878e+00 -2.87821233e-01 4.17988986e-01
3.85856837e-01 3.38584393e-01 8.93512189e-01 1.03538416e-01
-3.39020759e-01 1.58793688e-01 6.40503705e-01 -2.50463128e-01
9.43882287e-01 -1.27222812e+00 -2.05831394e-01 2.19282553e-01
-2.18394279e-01 5.31484365e-01 5.59841812e-01 -3.63472760e-01
-1.39968991e+00 -5.94109833e-01 4.67705399e-01 -5.86870372e-01
4.40525979e-01 1.22479506e-01 -7.58671522e-01 4.09690499e-01
7.96301067e-02 2.14334846e-01 1.53921759e+00 -8.07309866e-01
2.31793359e-01 1.93037093e-01 -1.60104859e+00 1.29062876e-01
3.23226154e-01 -6.99436128e-01 -1.13685457e-02 6.97405696e-01
1.28676057e-01 -4.77932036e-01 -1.16276002e+00 1.48326889e-01
1.27023602e+00 -9.09198284e-01 8.70936453e-01 -1.88870922e-01
-3.11849862e-01 -3.34191769e-01 2.21295923e-01 -6.34839118e-01
-1.92758758e-02 -7.24558771e-01 -1.51649907e-01 6.81091785e-01
1.64706707e-01 -5.48980713e-01 1.11888850e+00 3.09875101e-01
9.08749104e-02 -7.40164578e-01 -7.72707045e-01 -7.53208578e-01
1.45426258e-01 1.05308607e-01 7.78149068e-02 4.17236328e-01
2.88299173e-01 1.83341295e-01 1.34447962e-01 -2.33960301e-01
1.00589657e+00 1.87038481e-01 5.94582781e-02 -1.20650578e+00
2.08957598e-01 -1.35823742e-01 -3.60831380e-01 -8.91520619e-01
2.99957573e-01 -5.73975980e-01 -5.53963929e-02 -1.40234137e+00
6.65464640e-01 2.11940452e-01 -3.86128813e-01 -3.66609246e-02
-2.15570685e-02 4.95914191e-01 -4.66766119e-01 3.54966074e-01
-4.84742641e-01 -2.87649095e-01 1.37486005e+00 -2.79693663e-01
1.74471308e-02 -6.82417676e-03 -1.20091558e-01 4.08286452e-01
7.62804389e-01 -4.29167062e-01 -4.56225187e-01 -1.83597565e-01
-1.39418989e-01 8.33511427e-02 2.58988917e-01 -8.95931721e-01
2.36380845e-01 -1.77672252e-01 8.80675673e-01 -7.16563225e-01
4.21767801e-01 -3.94964546e-01 2.20976502e-01 4.04177964e-01
-6.68839216e-02 1.82489175e-02 2.44833082e-01 4.65652555e-01
-3.49106267e-02 -4.58634198e-01 1.50696611e+00 -5.21760881e-01
-5.25201499e-01 3.51861030e-01 -1.20947754e+00 -3.40024143e-01
1.09011543e+00 -7.24481523e-01 -7.16868222e-01 5.93944937e-02
-8.92206430e-01 -8.76733959e-02 7.11966634e-01 -5.19431412e-01
9.00241315e-01 -7.82428563e-01 -1.07660972e-01 -3.00021637e-02
-1.23271942e-01 -3.66882622e-01 6.17801309e-01 1.39812076e+00
-8.50870073e-01 5.44509411e-01 -4.16684687e-01 -6.80146158e-01
-1.64377463e+00 5.88909924e-01 6.14829957e-01 -1.55805275e-01
-6.07365847e-01 8.15862477e-01 1.22550607e-01 3.31161439e-01
-2.58151352e-01 -9.62299258e-02 -4.33485568e-01 -6.41863942e-02
1.20406902e+00 7.16580391e-01 1.18174486e-01 -4.48562026e-01
-4.87226635e-01 1.06557465e+00 -1.33727133e-01 -2.59915948e-01
1.08204818e+00 -4.42469001e-01 -4.38547403e-01 7.67398775e-01
1.49709487e+00 -2.16335401e-01 -1.26321042e+00 2.73667425e-01
-1.22850008e-01 -1.10548884e-01 1.82504520e-01 -6.85514033e-01
-7.86383271e-01 1.39711714e+00 9.51733172e-01 -2.19475508e-01
1.04070175e+00 -1.13605671e-01 1.04965055e+00 1.18622504e-01
3.66930157e-01 -9.94788766e-01 2.28317514e-01 -1.50009200e-01
1.12747304e-01 -9.96608794e-01 2.94007689e-01 -4.72371101e-01
-3.52538303e-02 1.52154517e+00 1.95828304e-01 3.68322313e-01
6.76295221e-01 7.25186229e-01 1.71195999e-01 -5.46456993e-01
-7.23600745e-01 2.01892510e-01 -4.28051323e-01 1.01998603e+00
8.12892258e-01 -2.61289328e-01 -4.27135527e-01 -5.11236265e-02
6.37858331e-01 4.11503196e-01 6.76302195e-01 1.48234510e+00
-7.97342002e-01 -1.15223885e+00 -1.79356858e-01 5.56117654e-01
-1.03362620e+00 3.71435702e-01 -2.72228181e-01 5.49926579e-01
1.36229163e-02 1.02153087e+00 -5.53588420e-02 1.56879008e-01
-6.24855645e-02 9.95643660e-02 1.27128005e+00 -4.07718688e-01
3.33624519e-02 3.46413374e-01 -3.64227355e-01 -5.70304573e-01
-1.34047019e+00 -7.01837063e-01 -1.71954310e+00 -2.35387102e-01
-5.60976565e-01 2.91502625e-01 7.16970682e-01 1.13397872e+00
5.80217913e-02 2.63110012e-01 3.55355084e-01 -8.20482135e-01
2.94405729e-01 -5.38467288e-01 -1.06390047e+00 -4.25691120e-02
6.92376852e-01 -8.22418094e-01 -6.94762945e-01 6.48382187e-01] | [14.286603927612305, -3.165252208709717] |
ff0bbea7-c037-477d-98d0-b4147ab0ab2e | unlocking-the-power-of-representations-in | 2305.01521 | null | https://arxiv.org/abs/2305.01521v1 | https://arxiv.org/pdf/2305.01521v1.pdf | Unlocking the Power of Representations in Long-term Novelty-based Exploration | We introduce Robust Exploration via Clustering-based Online Density Estimation (RECODE), a non-parametric method for novelty-based exploration that estimates visitation counts for clusters of states based on their similarity in a chosen embedding space. By adapting classical clustering to the nonstationary setting of Deep RL, RECODE can efficiently track state visitation counts over thousands of episodes. We further propose a novel generalization of the inverse dynamics loss, which leverages masked transformer architectures for multi-step prediction; which in conjunction with RECODE achieves a new state-of-the-art in a suite of challenging 3D-exploration tasks in DM-Hard-8. RECODE also sets new state-of-the-art in hard exploration Atari games, and is the first agent to reach the end screen in "Pitfall!". | ['Bilal Piot', 'Michal Valko', 'Oliver Groth', 'Leopoldo Sarra', 'Pablo Sprechmann', 'Charles Blundell', 'Daniele Calandriello', 'Steven Kapturowski', 'Alaa Saade'] | 2023-05-02 | null | null | null | null | ['atari-games'] | ['playing-games'] | [-5.00379920e-01 8.65348950e-02 -3.16885382e-01 6.32575247e-04
-1.22064149e+00 -4.31369424e-01 6.84892952e-01 -2.85548661e-02
-4.94377583e-01 6.97688520e-01 2.97834456e-01 -3.82790983e-01
-4.68768388e-01 -4.74080235e-01 -8.37070346e-01 -6.26276255e-01
-9.99036610e-01 1.11151135e+00 -2.62632698e-01 -2.81624962e-02
4.51893881e-02 1.38701469e-01 -1.27194750e+00 -1.20509237e-01
7.22685516e-01 1.06299758e+00 2.34051630e-01 6.20144188e-01
2.38018036e-01 7.86339819e-01 -2.62578070e-01 1.87363356e-01
3.15239847e-01 -4.33960378e-01 -5.80826104e-01 -2.04274178e-01
-1.85887694e-01 -4.53469813e-01 -6.48034334e-01 8.61965477e-01
5.43366194e-01 4.94720191e-01 6.31933212e-01 -1.15868509e+00
-3.49307865e-01 9.85830307e-01 -6.36562049e-01 4.81559396e-01
1.51271269e-01 5.85172951e-01 9.32466865e-01 -4.71234560e-01
6.05745494e-01 1.41226447e+00 4.21397924e-01 3.70200425e-01
-1.75863504e+00 -8.27694654e-01 6.31545544e-01 1.37850836e-01
-1.14596128e+00 -3.50720644e-01 4.84725147e-01 -2.55398124e-01
1.04406333e+00 -1.00480013e-01 1.00457191e+00 1.67860436e+00
6.42500445e-03 1.06939650e+00 1.35629630e+00 3.48200239e-02
9.42715764e-01 -1.55184105e-01 -1.13058135e-01 7.76737332e-01
-3.28713941e-04 7.00646281e-01 -8.02626312e-01 -3.32675844e-01
8.77100348e-01 3.49854305e-02 1.40008479e-01 -8.35120082e-01
-1.17601395e+00 1.04987621e+00 2.10811064e-01 -2.65096575e-01
-3.59148920e-01 7.02003121e-01 3.36177558e-01 3.22013199e-01
4.71899986e-01 5.73753357e-01 -3.81427258e-01 -1.07741487e+00
-1.03097808e+00 4.45508868e-01 9.35139596e-01 8.30774307e-01
6.22011840e-01 2.02005908e-01 -3.62177938e-01 1.81940272e-01
3.23853016e-01 5.52356422e-01 4.96575356e-01 -1.30973053e+00
3.42254311e-01 3.26526493e-01 3.36684942e-01 -2.93399304e-01
-4.11029190e-01 -9.17046964e-01 -5.83717644e-01 4.83932585e-01
9.96072963e-02 -5.11969507e-01 -1.02254546e+00 2.24153399e+00
3.66567016e-01 3.76460791e-01 -1.92863960e-03 7.99773276e-01
-3.27047557e-02 6.68657184e-01 -1.78515330e-01 -1.34040132e-01
6.62724555e-01 -8.02836120e-01 -3.71558636e-01 -2.76202530e-01
3.73411536e-01 4.23979253e-01 1.22771859e+00 7.59022355e-01
-1.34033453e+00 -1.79017335e-01 -8.06080699e-01 3.53441775e-01
-1.66891754e-01 -1.93005964e-01 9.71058607e-01 2.28140622e-01
-1.21761799e+00 9.21712220e-01 -1.77729666e+00 -9.37747732e-02
6.66320920e-01 3.15088421e-01 1.38625890e-01 3.17912012e-01
-7.37145543e-01 7.50695586e-01 3.38233411e-01 -1.01002842e-01
-2.03893542e+00 -6.81538463e-01 -7.67632604e-01 1.10227086e-01
8.21040630e-01 -6.10534608e-01 1.33355772e+00 -3.35915148e-01
-1.74807465e+00 2.67719001e-01 -1.54185504e-01 -9.42319989e-01
5.95847547e-01 -4.97142524e-01 -1.71802893e-01 -1.32337272e-01
1.13218658e-01 5.82849622e-01 6.98833644e-01 -1.06403148e+00
-4.31211233e-01 -4.10659015e-01 -2.09009722e-01 5.02344549e-01
-1.04060128e-01 -6.53695881e-01 -5.31400323e-01 -3.55487585e-01
-1.49584368e-01 -9.86686707e-01 -5.62025130e-01 -2.71722168e-01
-5.69032848e-01 -3.68542045e-01 4.97021765e-01 -2.80599177e-01
1.28302097e+00 -1.85847723e+00 8.25474620e-01 4.53120679e-01
3.43204618e-01 -3.16802859e-01 -1.01665244e-01 5.88957787e-01
3.12844098e-01 -1.28087208e-01 -1.93194702e-01 -1.00098133e+00
5.65734565e-01 9.42466632e-02 -4.49366719e-01 4.92368013e-01
-3.16527009e-01 1.18398023e+00 -1.15584874e+00 -2.06707954e-01
2.27770016e-01 1.13390721e-01 -5.28824389e-01 1.83306202e-01
-6.93242311e-01 4.07916665e-01 -5.68774343e-01 5.65324903e-01
1.60057813e-01 -5.19877315e-01 2.04621270e-01 5.18299401e-01
-7.39123598e-02 4.08157706e-01 -1.12608552e+00 2.20026612e+00
-4.31178033e-01 3.83739114e-01 1.68180957e-01 -8.45555961e-01
5.67827642e-01 -1.84447914e-01 5.64942062e-01 -5.63629508e-01
1.27175733e-01 -1.39980659e-01 -3.58586490e-01 4.03906265e-03
5.15860856e-01 4.11781579e-01 -4.33626413e-01 8.76254857e-01
3.43976095e-02 1.65711507e-01 7.77800083e-02 5.19623637e-01
1.51591694e+00 4.33822304e-01 7.69505464e-03 -4.71196324e-01
-4.98905271e-01 -5.75108342e-02 3.98839951e-01 1.18648493e+00
-8.94011557e-02 -6.78957030e-02 7.08958387e-01 -2.55277812e-01
-8.24842751e-01 -1.77560246e+00 1.35278538e-01 1.11211109e+00
8.68214667e-02 -6.49869263e-01 -4.23570871e-01 -6.11384630e-01
2.87194163e-01 7.14119017e-01 -9.39436316e-01 -4.53076959e-01
-1.86889291e-01 -5.54496646e-01 3.72788608e-01 3.50631684e-01
3.19489479e-01 -1.15472114e+00 -8.77327085e-01 4.04033035e-01
6.48180097e-02 -5.21957040e-01 -4.82226312e-01 7.45526612e-01
-8.86712551e-01 -8.34481716e-01 -4.13443238e-01 -3.65490824e-01
2.64398724e-01 -1.68615609e-01 1.22111058e+00 -6.55578434e-01
-2.42638856e-01 5.20644665e-01 -1.16578087e-01 -1.96748257e-01
-1.49739489e-01 4.29972708e-01 5.28197706e-01 -4.10356462e-01
3.07954490e-01 -1.00576723e+00 -7.92771935e-01 1.09202430e-01
-3.79560083e-01 -2.22711951e-01 6.18143201e-01 6.71445549e-01
9.30768847e-01 6.11656252e-03 3.65682065e-01 -3.90995622e-01
7.43480623e-01 -8.60305071e-01 -1.13511026e+00 -1.08169951e-01
-7.56640673e-01 5.96679509e-01 2.40627170e-01 -7.90455401e-01
-6.19749069e-01 -1.92732215e-01 3.78555626e-01 -1.05455017e+00
1.17882818e-01 5.37992775e-01 3.19512308e-01 2.70966798e-01
5.00395775e-01 4.04795498e-01 1.76058933e-01 -5.81561863e-01
7.03438103e-01 3.78019176e-02 7.06477463e-01 -6.52072430e-01
8.00223768e-01 5.14199734e-01 1.10524878e-01 -3.22688192e-01
-8.45761776e-01 -1.82676226e-01 9.31642763e-03 9.67486575e-02
6.86775386e-01 -1.48710072e+00 -1.47018945e+00 2.20322922e-01
-2.92980373e-01 -1.21513641e+00 -8.41017246e-01 4.63317275e-01
-9.59254503e-01 7.26643056e-02 -5.76031744e-01 -1.21605241e+00
-2.94529051e-01 -1.02103186e+00 9.90184069e-01 1.63272813e-01
-6.91869482e-02 -9.24760163e-01 6.26264930e-01 -1.52109116e-01
4.66560602e-01 3.80657315e-01 6.66482270e-01 -3.56756806e-01
-8.66624713e-01 1.61628515e-01 3.33396465e-01 -1.71366602e-01
-2.61412621e-01 -5.72432756e-01 -6.40803874e-01 -7.43808031e-01
-1.64909795e-01 -7.21108675e-01 1.32785451e+00 8.54433596e-01
1.45554209e+00 -3.69669735e-01 -6.70448661e-01 8.85663867e-01
1.02903116e+00 -3.27911563e-02 2.96343416e-01 3.10485661e-01
2.59500831e-01 4.82742582e-03 6.68163121e-01 9.46122110e-01
7.20939994e-01 6.86096847e-01 8.19489002e-01 1.96188197e-01
4.63203371e-01 -9.30779338e-01 7.57421851e-01 3.73516709e-01
3.02204072e-01 -1.77600071e-01 -6.36637151e-01 7.78043747e-01
-2.13180208e+00 -1.18616176e+00 7.79434383e-01 2.30642343e+00
9.43408430e-01 4.46001858e-01 7.47610390e-01 -6.33045673e-01
4.26206321e-01 3.11210573e-01 -1.33276224e+00 -2.39068300e-01
9.09230113e-02 3.36151391e-01 5.59241056e-01 6.03991091e-01
-1.04658186e+00 1.17926848e+00 6.55181742e+00 9.30178225e-01
-6.09297514e-01 2.44205043e-01 8.03005397e-01 -9.07410204e-01
-3.05781096e-01 1.44684464e-01 -8.74625087e-01 5.90310931e-01
1.15570033e+00 3.15160826e-02 1.18341613e+00 1.06022000e+00
1.34431764e-01 -4.76012915e-01 -1.24739254e+00 1.04122889e+00
-2.74088800e-01 -1.41217446e+00 -5.34918010e-01 7.21267462e-01
8.83102596e-01 7.20583856e-01 7.29556739e-01 7.38277674e-01
1.40732992e+00 -1.31604826e+00 5.86625218e-01 6.10568464e-01
8.35789442e-01 -1.03704929e+00 -4.17006649e-02 4.82882738e-01
-1.02635193e+00 -4.65974301e-01 -2.64003187e-01 -2.94081774e-02
4.94060934e-01 5.43367267e-01 -7.31858552e-01 -8.64654034e-02
8.08842182e-01 7.46312439e-01 -3.43850851e-01 7.60264874e-01
-1.12220131e-01 6.50754809e-01 -7.40646303e-01 -2.21131325e-01
5.78055024e-01 -2.91049987e-01 9.70422447e-01 6.74887896e-01
3.44844997e-01 -2.98204690e-01 3.26257616e-01 1.59094512e+00
9.06720664e-03 -7.65428066e-01 -6.41013682e-01 -3.37069929e-01
8.93355072e-01 1.09346414e+00 -3.50780427e-01 -8.89230296e-02
4.97911781e-01 9.46127653e-01 7.82453656e-01 4.04790878e-01
-7.69186318e-01 -4.05731127e-02 9.54425812e-01 -2.40065064e-02
8.26573491e-01 -5.48800409e-01 2.56311387e-01 -1.16773629e+00
-1.83982387e-01 -7.97410131e-01 3.51847321e-01 -3.67998213e-01
-1.06553614e+00 4.57833976e-01 2.45820016e-01 -9.46119308e-01
-7.22904682e-01 -7.26835057e-02 -8.01393032e-01 4.05044556e-01
-1.21642971e+00 -7.60485530e-01 7.40527883e-02 5.95971525e-01
4.28563535e-01 -3.54682565e-01 7.42664814e-01 -4.21887636e-01
-6.64097488e-01 6.69233263e-01 7.07577825e-01 -3.81010950e-01
2.21269786e-01 -1.67520702e+00 9.95920658e-01 6.13461971e-01
4.42421585e-01 4.31776851e-01 6.82253957e-01 -8.17523241e-01
-1.77164352e+00 -1.01503575e+00 -1.50818914e-01 -7.70066917e-01
7.59608924e-01 -8.14339161e-01 -3.81371588e-01 9.09079969e-01
-7.84870312e-02 -1.42591313e-01 3.20444942e-01 5.73899031e-01
-1.80948421e-01 1.56503543e-01 -7.87965655e-01 7.52284884e-01
1.31059158e+00 -4.85280722e-01 -2.99477190e-01 2.74920374e-01
9.46276009e-01 -4.85935032e-01 -7.38552272e-01 -1.54678136e-01
2.86081433e-01 -9.26066101e-01 1.00130105e+00 -6.98442936e-01
5.17262146e-02 6.86635599e-02 -1.94360688e-01 -1.56052434e+00
-3.43755573e-01 -1.46659410e+00 -1.12227225e+00 5.72702765e-01
4.92957830e-01 -4.31234986e-01 1.12051308e+00 6.14697263e-02
-6.55016005e-02 -1.04582119e+00 -1.23383081e+00 -9.10950541e-01
8.49211589e-02 -2.89109051e-01 4.40774918e-01 3.25020403e-01
2.95277268e-01 1.24066889e-01 -5.72857738e-01 5.22326154e-04
1.05637908e+00 2.71382451e-01 8.64153385e-01 -1.07168519e+00
-9.82052207e-01 -2.48819038e-01 9.38673839e-02 -1.32832372e+00
2.72920012e-01 -7.12972403e-01 3.31672132e-02 -1.26723158e+00
2.98462093e-01 -4.34635937e-01 -4.00422513e-01 2.85646319e-01
2.05962688e-01 -3.15600514e-01 1.69547915e-01 1.14621378e-01
-1.60036910e+00 1.30639517e+00 8.56673121e-01 1.88259572e-01
-7.47461975e-01 9.74610448e-02 -7.56655037e-01 3.58390212e-01
6.29600167e-01 -4.01307881e-01 -5.12379408e-01 -2.40585402e-01
4.10687387e-01 3.22763860e-01 4.56026763e-01 -1.09612560e+00
1.48301914e-01 -8.41203332e-02 3.49815041e-01 -8.90228629e-01
8.88308764e-01 -4.35313553e-01 2.08139673e-01 5.37248313e-01
-7.04916298e-01 3.37626874e-01 2.48890013e-01 1.27482581e+00
4.43608373e-01 4.63577867e-01 4.13992643e-01 -2.27016017e-01
-5.15573025e-01 6.14867985e-01 -4.19442862e-01 5.55227458e-01
9.02816594e-01 -1.14986815e-01 -2.32618406e-01 -7.09258139e-01
-9.50792193e-01 9.36588645e-01 4.82917011e-01 1.56925038e-01
5.69682896e-01 -1.33871663e+00 -4.18239057e-01 -4.46459278e-02
-8.38164911e-02 3.27701807e-01 2.14543000e-01 8.67421627e-01
1.06215753e-01 1.19405158e-01 1.39186382e-01 -6.22584224e-01
-5.70847213e-01 5.21181822e-01 2.57025003e-01 -7.75597155e-01
-8.48282933e-01 9.24846828e-01 -3.62892896e-02 -3.42035145e-01
5.88429868e-01 -3.70537907e-01 3.02722603e-01 -3.25020641e-01
3.53247017e-01 5.63232064e-01 -5.07951021e-01 1.86172992e-01
-2.06095099e-01 1.04559310e-01 -2.77598709e-01 -7.74773300e-01
1.35231781e+00 -3.08975935e-01 4.44319576e-01 8.72852981e-01
1.05521679e+00 -6.01342916e-01 -2.05081987e+00 -2.42649913e-01
-3.20713550e-01 -2.01326072e-01 4.74821329e-01 -1.06612873e+00
-7.01204360e-01 6.22717559e-01 8.70940983e-01 -1.83565691e-01
7.27338135e-01 3.98236126e-01 7.61989117e-01 8.71697724e-01
6.07057691e-01 -1.27309430e+00 4.40146178e-01 5.36398947e-01
5.27480960e-01 -1.21293175e+00 -1.18822120e-01 7.15877831e-01
-9.63310957e-01 3.50228310e-01 3.97002071e-01 -3.05805206e-01
7.69931793e-01 2.53032088e-01 -6.25572503e-01 -5.63503385e-01
-1.35770142e+00 -2.38259092e-01 -5.33475950e-02 6.54807031e-01
-5.51136136e-01 1.56773418e-01 5.29827237e-01 6.66878581e-01
-2.97546655e-01 -1.75796747e-01 2.74941832e-01 8.77064049e-01
-6.02060616e-01 -6.59725189e-01 2.27196962e-01 6.82713270e-01
-1.60346031e-01 -7.33855665e-02 1.65078267e-02 6.31272793e-01
-5.55419028e-01 7.66102910e-01 3.58776718e-01 -5.21395028e-01
-1.77919909e-01 -2.90622443e-01 4.62847292e-01 -4.25436556e-01
-3.83359998e-01 9.66994986e-02 -1.24286398e-01 -1.27192366e+00
2.38361001e-01 -9.04131591e-01 -9.54886198e-01 -5.14241219e-01
-1.17168598e-01 4.04372990e-01 5.87183654e-01 6.62530482e-01
9.15445685e-01 2.46444792e-01 7.30811477e-01 -1.02747273e+00
-1.26393592e+00 -9.14001286e-01 -7.52946258e-01 2.55553722e-02
5.86820126e-01 -9.69628096e-01 -5.19019723e-01 -7.95009792e-01] | [4.137279033660889, 2.041069269180298] |
de46cc87-5f61-4f2a-abcb-6fb090bcf1b6 | neural-multi-task-learning-in-automated | 1801.06830 | null | http://arxiv.org/abs/1801.06830v1 | http://arxiv.org/pdf/1801.06830v1.pdf | Neural Multi-task Learning in Automated Assessment | Grammatical error detection and automated essay scoring are two tasks in the
area of automated assessment. Traditionally these tasks have been treated
independently with different machine learning models and features used for each
task. In this paper, we develop a multi-task neural network model that jointly
optimises for both tasks, and in particular we show that neural automated essay
scoring can be significantly improved. We show that while the essay score
provides little evidence to inform grammatical error detection, the essay score
is highly influenced by error detection. | ['Ronan Cummins', 'Marek Rei'] | 2018-01-21 | null | null | null | null | ['automated-essay-scoring', 'grammatical-error-detection'] | ['natural-language-processing', 'natural-language-processing'] | [ 7.72048011e-02 3.03849727e-01 1.49400875e-01 -8.38114500e-01
-1.09756708e+00 -4.42812324e-01 2.20848233e-01 6.08857512e-01
-7.47719586e-01 9.58511114e-01 1.37671173e-01 -2.30010584e-01
-2.25944251e-01 -5.30415952e-01 -2.51215756e-01 -1.48004487e-01
3.97388548e-01 7.14612722e-01 -1.04946136e-01 -2.09661141e-01
8.15572917e-01 -1.36842176e-01 -1.31154418e+00 2.29831830e-01
1.12934697e+00 6.30535901e-01 -7.78970793e-02 1.42099667e+00
-2.07597271e-01 1.11147881e+00 -1.14647663e+00 -6.50118351e-01
-1.49895340e-01 -4.20335472e-01 -1.08342254e+00 -4.61147904e-01
6.81391120e-01 -3.94194216e-01 7.96609223e-02 1.06722641e+00
4.39937502e-01 1.27500117e-01 9.09115613e-01 -9.63196099e-01
-7.82803059e-01 5.93692839e-01 -1.04654646e-02 2.04616874e-01
4.66853946e-01 -1.92182049e-01 1.39717090e+00 -9.33390737e-01
9.93707851e-02 1.03531802e+00 9.36155021e-01 5.48518002e-01
-8.64456952e-01 -4.58327562e-01 -1.23689003e-01 2.14287028e-01
-6.81908846e-01 -4.38987821e-01 6.27569735e-01 -5.06792963e-01
1.02965510e+00 7.48447850e-02 3.52087408e-01 7.37214506e-01
5.11749268e-01 9.08283710e-01 1.13513052e+00 -8.15022707e-01
-1.26195863e-01 2.09528968e-01 1.03003013e+00 1.01301134e+00
3.03221315e-01 3.53812464e-02 -7.32438803e-01 -9.97207910e-02
4.03471589e-01 -3.98880422e-01 4.75238217e-03 4.76563334e-01
-6.09036326e-01 1.12495136e+00 -1.98886335e-01 4.38997239e-01
-1.31045565e-01 4.15926486e-01 5.55348635e-01 1.03354442e+00
6.08978748e-01 6.85267091e-01 -6.88025951e-01 -3.60500932e-01
-1.26795208e+00 4.48988408e-01 1.12432802e+00 3.31951678e-01
5.87943196e-01 2.22806320e-01 -4.00783002e-01 1.16153157e+00
5.37390232e-01 2.16984689e-01 7.60298491e-01 -1.02758253e+00
4.10347223e-01 5.75513184e-01 1.45592943e-01 -8.78860831e-01
-7.44996667e-01 -3.09825063e-01 -4.84785855e-01 3.49379897e-01
8.03386688e-01 -3.64089042e-01 -4.50065762e-01 1.66955400e+00
-2.48540118e-01 -5.60991049e-01 -1.60018906e-01 5.61830997e-01
9.52740312e-01 3.67395550e-01 2.53618717e-01 -5.72569948e-03
1.11542726e+00 -1.07715750e+00 -9.27392066e-01 -4.21987504e-01
1.22896039e+00 -7.14272261e-01 9.78247225e-01 7.12548852e-01
-1.62784374e+00 -5.73439658e-01 -1.14705527e+00 -4.14822161e-01
-8.16592574e-02 4.07214135e-01 3.64505082e-01 8.02803278e-01
-1.25162578e+00 8.09558868e-01 -4.01290596e-01 9.85064432e-02
1.72670290e-01 4.14053887e-01 -1.47459283e-01 2.72060096e-01
-1.26593232e+00 1.51853132e+00 2.07209319e-01 -1.39358882e-02
-3.42871159e-01 -3.34702998e-01 -6.96712136e-01 2.08987102e-01
-2.14610592e-01 -5.97966492e-01 1.95838177e+00 -1.02410340e+00
-1.56203187e+00 1.09415710e+00 -2.65406877e-01 -3.43327254e-01
2.53945768e-01 -2.89392591e-01 -2.35617533e-02 -2.64819652e-01
1.25942841e-01 2.94218659e-01 6.89453602e-01 -6.34881198e-01
-8.88788283e-01 -7.20924020e-01 -8.85360688e-02 2.98245430e-01
-5.68533242e-01 5.23742855e-01 3.38801652e-01 -6.15911305e-01
-4.08849150e-01 -4.80277628e-01 1.10745896e-02 -9.10951197e-02
-4.18606512e-02 -8.72047246e-01 2.38247484e-01 -1.01161575e+00
1.45205998e+00 -1.52139735e+00 1.92301661e-01 7.74440775e-03
4.60533142e-01 2.66271889e-01 -1.75736338e-01 -1.11691482e-01
1.77118495e-01 2.30395406e-01 -2.03872606e-01 -5.95473707e-01
1.92530513e-01 2.41520219e-02 3.84846419e-01 2.05250695e-01
3.10964018e-01 1.04112744e+00 -7.18964338e-01 -3.59316409e-01
-3.63431782e-01 -2.74016671e-02 -3.43396276e-01 4.09098923e-01
6.48619905e-02 -1.28274500e-01 -2.61460990e-01 5.46093404e-01
1.14780061e-01 -2.33709380e-01 -4.93708849e-02 5.73968530e-01
1.39221037e-02 6.68906033e-01 -8.92776728e-01 1.27924693e+00
-5.92008412e-01 9.20639217e-01 1.83649287e-01 -1.00295436e+00
1.29932284e+00 4.06848520e-01 -2.84987807e-01 -7.38031089e-01
2.87676543e-01 5.57874084e-01 2.37893373e-01 -5.69628239e-01
6.72797382e-01 -3.67987186e-01 -2.43367180e-01 1.24142230e+00
2.98029870e-01 -3.25200200e-01 2.42288932e-01 6.95657727e-05
1.25300074e+00 -1.09887622e-01 5.81999063e-01 -1.94366500e-01
6.35460675e-01 -5.15310019e-02 3.78882289e-01 9.15453553e-01
-5.48501730e-01 4.43347782e-01 5.11245668e-01 -5.13797581e-01
-1.21312523e+00 -5.45924187e-01 -2.45193735e-01 1.61959028e+00
-5.58379769e-01 -1.78962827e-01 -7.54694819e-01 -9.68431830e-01
6.76023215e-02 8.43674898e-01 -7.05701113e-01 -2.25783780e-01
-6.96905851e-01 -9.22971368e-01 9.01311100e-01 6.47388816e-01
2.33600046e-02 -1.31381261e+00 -6.32903993e-01 5.24095595e-01
-6.31964281e-02 -4.91613925e-01 -4.95128453e-01 6.61048412e-01
-8.83978903e-01 -1.20017970e+00 -5.99374592e-01 -8.90496850e-01
3.81379664e-01 -2.67469555e-01 1.61712766e+00 7.42142856e-01
5.25882728e-02 4.67467606e-01 -3.02003384e-01 -7.54486561e-01
-7.96988428e-01 3.51170689e-01 -1.17593825e-01 -5.11590242e-01
7.33658433e-01 -4.25244778e-01 1.08199023e-01 -3.20930295e-02
-5.29330969e-01 -3.95822674e-01 3.14761519e-01 1.26450098e+00
-1.73807576e-01 -6.85278177e-01 1.09810901e+00 -1.06987429e+00
1.42738736e+00 -2.50511855e-01 -1.90508217e-01 3.71993870e-01
-9.64809418e-01 2.33345926e-01 4.66105372e-01 4.38318364e-02
-7.73190439e-01 -2.54464835e-01 -3.43115002e-01 3.05012614e-01
-2.63706863e-01 7.72613764e-01 3.47449273e-01 -4.74130780e-01
7.16717541e-01 -5.06216101e-02 6.29589036e-02 -4.34862256e-01
-2.36480013e-01 7.94037402e-01 2.99714476e-01 -5.00423968e-01
2.36469626e-01 -4.87779319e-01 -1.13654710e-01 -4.16720539e-01
-1.18060184e+00 -1.90377504e-01 -9.54245448e-01 -4.56309050e-01
4.39536810e-01 -6.13485932e-01 -4.41124737e-01 6.08553350e-01
-1.52692127e+00 -3.04228783e-01 -6.85439631e-02 1.18014738e-01
-6.10968351e-01 5.38374901e-01 -7.55411327e-01 -1.06532383e+00
-7.96350837e-01 -9.09197628e-01 8.96975935e-01 2.32521594e-01
-7.31878877e-01 -1.48154736e+00 5.62289357e-01 4.28334624e-01
3.14573437e-01 -3.41830760e-01 1.07315385e+00 -1.31706929e+00
7.04482645e-02 -4.42849129e-01 -2.94951826e-01 5.38340449e-01
-2.77401626e-01 -4.60998677e-02 -9.46189880e-01 7.81249702e-02
1.45033270e-01 -8.56701612e-01 1.04263020e+00 4.49229956e-01
1.17231882e+00 -2.51441091e-01 3.55268389e-01 2.08167031e-01
1.13938844e+00 3.63023840e-02 2.52514690e-01 5.68495810e-01
4.61391300e-01 8.11741054e-01 4.33961987e-01 2.62191117e-01
5.12755871e-01 5.89293301e-01 -2.62884982e-03 4.83156800e-01
4.17945869e-02 3.58709872e-01 4.74242568e-01 1.13478315e+00
1.80780850e-02 -2.94487298e-01 -1.25423384e+00 7.26937413e-01
-2.05234456e+00 -6.59108639e-01 -7.84253895e-01 1.87986481e+00
9.97310400e-01 4.40432727e-02 1.97290882e-01 5.75656712e-01
6.04256868e-01 -3.16230953e-02 -1.32556558e-01 -1.23866236e+00
-1.21978484e-01 7.27647841e-01 -1.03665367e-02 9.78523552e-01
-9.50767577e-01 5.10564744e-01 8.06694508e+00 3.62246633e-01
-2.95187145e-01 4.61629272e-01 5.68583965e-01 -3.04398030e-01
-1.92830995e-01 -5.55869162e-01 -9.08724308e-01 5.27497113e-01
1.41185904e+00 -3.39644939e-01 -4.08856533e-02 5.97521186e-01
6.10423088e-02 -2.12790236e-01 -1.12754691e+00 4.03897375e-01
3.94214720e-01 -7.04774320e-01 -7.03583956e-02 -2.42171705e-01
4.99974281e-01 -3.29570681e-01 1.94511786e-02 6.50197744e-01
5.85473716e-01 -1.29222441e+00 6.08554602e-01 6.96713865e-01
4.50361967e-01 -8.94992232e-01 1.22255003e+00 4.41807330e-01
-4.38513398e-01 -2.78739065e-01 -3.93430829e-01 -6.68490231e-01
3.39449104e-03 7.65948176e-01 -5.09953141e-01 3.82259116e-02
2.98826993e-01 5.75457692e-01 -8.00819039e-01 1.10854292e+00
-6.89086199e-01 7.48599231e-01 -2.97048613e-02 -4.92987126e-01
2.93370366e-01 -4.18564817e-03 4.28608298e-01 1.32344270e+00
3.03056180e-01 -6.42158538e-02 1.72320027e-02 6.64059460e-01
-2.80132800e-01 2.81059951e-01 -4.34958845e-01 1.66124269e-01
1.66929126e-01 1.10103059e+00 -8.43555480e-02 -3.75334322e-01
-4.54106301e-01 9.47824538e-01 9.91736591e-01 -1.51077807e-01
-4.60607231e-01 -7.24093080e-01 3.60402286e-01 -3.42155069e-01
-9.82471183e-02 -6.95560575e-02 -1.19860315e+00 -1.08283246e+00
-3.89292240e-02 -6.01348698e-01 4.22983438e-01 -6.50696814e-01
-1.42111516e+00 2.26949915e-01 -6.00744009e-01 -3.85219961e-01
-6.09158576e-01 -1.04367411e+00 -1.02337670e+00 1.19305146e+00
-1.62282526e+00 -7.00685561e-01 -1.15383081e-01 -3.02466098e-02
7.86192119e-01 -4.11557376e-01 1.11624622e+00 2.46427581e-01
-6.88995600e-01 1.11391139e+00 3.69853109e-01 2.47203410e-01
9.79331911e-01 -1.82148588e+00 3.11690837e-01 4.18506026e-01
7.74039254e-02 1.58115476e-01 4.86230403e-01 -5.26323199e-01
-5.21485925e-01 -9.72971022e-01 1.83308136e+00 -8.76023531e-01
6.75139070e-01 2.07626805e-01 -1.17759490e+00 6.28521442e-01
1.50227025e-01 -5.63840806e-01 8.85038435e-01 7.44703531e-01
-1.91131026e-01 4.17116255e-01 -1.05029881e+00 2.52214726e-02
4.84736592e-01 -6.65301144e-01 -9.37375546e-01 4.67490315e-01
1.05783351e-01 -2.62573630e-01 -8.82199049e-01 -8.86723306e-03
6.95266545e-01 -9.76913095e-01 5.58463991e-01 -1.05174482e+00
9.89676952e-01 4.05277759e-01 2.11619034e-01 -1.71331239e+00
-5.93742192e-01 2.32431218e-02 -5.21578193e-01 9.50839520e-01
7.35597014e-01 -3.08827132e-01 8.03702712e-01 7.91568816e-01
-2.51720756e-01 -7.45371461e-01 -1.01635027e+00 -4.97653127e-01
7.41543591e-01 -3.37083459e-01 1.61382973e-01 7.47324228e-01
3.33058715e-01 6.22114778e-01 -3.26208651e-01 -3.38877171e-01
5.29099941e-01 -3.03042561e-01 2.36472413e-01 -1.86274791e+00
-1.03714749e-01 -1.02130771e+00 -2.35046163e-01 -6.56028152e-01
4.76970553e-01 -9.45938110e-01 3.25832874e-01 -1.37045431e+00
3.77165616e-01 -9.12749022e-03 -2.06858039e-01 5.91258645e-01
-7.13166893e-01 3.58808368e-01 3.74306478e-02 -6.22551925e-02
-7.15032578e-01 4.22503978e-01 7.35282660e-01 -1.75582662e-01
4.13916297e-02 2.09997877e-01 -7.80974030e-01 7.69387841e-01
1.08086264e+00 -7.60608435e-01 2.67746001e-01 -7.78653741e-01
8.42585623e-01 4.32869140e-03 2.39401132e-01 -8.32183301e-01
3.57907802e-01 1.47972152e-01 3.46314847e-01 -2.47282073e-01
-1.39240161e-01 -1.36348426e-01 -8.39009702e-01 2.84195095e-01
-8.66382241e-01 4.64727789e-01 1.29010201e-01 3.08023840e-01
-2.57085592e-01 -1.38273787e+00 8.17350447e-01 -4.19476688e-01
-1.02227017e-01 -1.28424600e-01 -8.26598823e-01 1.02656983e-01
4.03412730e-01 -1.96294740e-01 -1.26545757e-01 -4.89633948e-01
-9.32094216e-01 4.30250764e-01 1.98967569e-02 2.61861771e-01
6.06789112e-01 -1.13906527e+00 -1.19135392e+00 -1.57574413e-03
-1.24665536e-01 -3.17610472e-01 -7.32619017e-02 6.55639470e-01
-2.53468037e-01 7.71672487e-01 -2.17020303e-01 -2.65534997e-01
-1.58095336e+00 -3.03406537e-01 5.09425402e-01 -7.87877321e-01
1.21836469e-01 1.05181432e+00 -4.05658633e-01 -8.42655718e-01
3.73594433e-01 3.20475325e-02 -6.72289789e-01 2.41760209e-01
8.06241691e-01 6.11770928e-01 4.59867865e-01 -1.86884493e-01
2.75442511e-01 3.61803144e-01 -4.69489664e-01 -7.74071813e-02
1.44076812e+00 -1.16039708e-01 -2.99710989e-01 8.88226748e-01
5.86498141e-01 -5.15567243e-01 -5.08620441e-01 -8.95861462e-02
3.98719370e-01 -2.70666271e-01 4.06298518e-01 -1.42037475e+00
-4.75524217e-01 1.27747715e+00 1.24072932e-01 3.63245547e-01
7.31960118e-01 -4.83074576e-01 6.45886838e-01 7.92547762e-01
1.18419990e-01 -1.50425994e+00 2.11182773e-01 1.21857822e+00
7.20958233e-01 -1.45367920e+00 -4.72247489e-02 1.21713065e-01
-3.37550730e-01 1.55580807e+00 9.79902983e-01 -2.81165719e-01
3.28092575e-01 1.92900509e-01 2.40399353e-02 -4.20580894e-01
-9.86028671e-01 2.05395132e-01 4.12493259e-01 5.04622459e-01
1.00561726e+00 -1.94309384e-01 -6.49838865e-01 1.13170612e+00
-4.03596282e-01 -8.39956384e-03 8.94734144e-01 6.83031917e-01
-9.09540474e-01 -1.19073057e+00 -6.06203854e-01 1.03485227e+00
-6.77237391e-01 -2.11851045e-01 -8.88713837e-01 3.94679487e-01
-3.00349802e-01 8.70728731e-01 -1.47277974e-02 -1.68392062e-01
5.69163337e-02 9.62677300e-01 6.85939968e-01 -8.51354063e-01
-1.34638548e+00 -7.76890576e-01 5.04537702e-01 -2.94474587e-02
-2.65006244e-01 -8.27590227e-01 -8.54765177e-01 -3.56356114e-01
-6.81571662e-01 2.34895825e-01 6.15577340e-01 1.26088572e+00
-3.34928669e-02 7.04719841e-01 3.47410232e-01 -4.29182708e-01
-9.61938918e-01 -1.35078549e+00 -6.45361483e-01 -1.26607031e-01
3.87802035e-01 -5.00624359e-01 -4.93527144e-01 -3.56964797e-01] | [11.2818021774292, 9.402926445007324] |
21033483-09da-42de-823d-a25c5591ee8a | optimal-precoder-design-for-mimo-ofdm-based | 2109.12452 | null | https://arxiv.org/abs/2109.12452v1 | https://arxiv.org/pdf/2109.12452v1.pdf | Optimal Precoder Design for MIMO-OFDM-based Joint Automotive Radar-Communication Networks | Large-scale deployment of connected vehicles with cooperative awareness technologies increases the demand for vehicle-to-everything (V2X) communication spectrum in 5.9 GHz that is mainly allocated for the exchange of safety messages. To supplement V2X communication and support the high data rates needed by broadband applications, the millimeter-wave (mmWave) automotive radar spectrum at 76-81 GHz can be utilized. For this purpose, joint radar-communication systems have been proposed in the literature to perform both functions using the same waveform and hardware. While multiple-input and multiple-output (MIMO) communication with multiple users enables independent data streaming for high throughput, MIMO radar processing provides high-resolution imaging that is crucial for safety-critical systems. However, employing conventional precoding methods designed for communication generates directional beams that impair MIMO radar imaging and target tracking capabilities during data streaming. In this paper, we propose a MIMO joint automotive radar-communication (JARC) framework based on orthogonal frequency division multiplexing (OFDM) waveform. First, we show that the MIMO-OFDM preamble can be exploited for both MIMO radar processing and estimation of the communication channel. Then, we propose an optimal precoder design method that enables high accuracy target tracking while transmitting independent data streams to multiple receivers. The proposed methods provide high-resolution radar imaging and high throughput capabilities for MIMO JARC networks. Finally, we evaluate the efficacy of the proposed methods through numerical simulations. | ['Onur Altintas', 'Chang-Heng Wang', 'Eylem Ekici', 'Ceyhun D. Ozkaptan'] | 2021-09-25 | null | null | null | null | ['joint-radar-communication'] | ['robots'] | [ 3.14919531e-01 1.66004315e-01 -2.10686862e-01 -3.12312365e-01
-6.79096937e-01 -2.76190668e-01 6.85035765e-01 -3.40545505e-01
-3.03439617e-01 7.20384598e-01 -1.93908185e-01 -7.90987432e-01
-4.99808848e-01 -9.24527705e-01 -2.01618224e-01 -8.21918845e-01
-5.79315424e-01 -3.08483411e-02 5.26071005e-02 -1.41859114e-01
-1.59883574e-02 7.20229685e-01 -1.40902090e+00 -1.15591355e-01
7.99901783e-01 1.35682738e+00 5.72074413e-01 6.71832800e-01
1.56441227e-01 1.41249463e-01 -8.30524445e-01 -2.66705096e-01
5.41964829e-01 -1.23854131e-01 7.01494515e-01 -1.80174246e-01
4.17116284e-01 -5.05273163e-01 -7.87353039e-01 9.50543702e-01
3.64147544e-01 -4.13030535e-01 7.69087791e-01 -1.43095076e+00
-1.65545940e-01 2.81272769e-01 -5.52084565e-01 1.38061777e-01
-4.44911420e-01 1.97789967e-02 2.19746217e-01 -7.78914213e-01
3.27048957e-01 9.32085454e-01 2.45306149e-01 3.19624454e-01
-4.54809606e-01 -1.43950272e+00 -3.81871581e-01 5.88738739e-01
-1.33128989e+00 -6.57448232e-01 5.07987440e-01 -2.67722487e-01
5.00864506e-01 5.46940088e-01 6.55879676e-01 5.82659960e-01
8.39581251e-01 6.89501390e-02 9.23333883e-01 -1.76312685e-01
5.46126738e-02 5.25076538e-02 1.42071202e-01 4.22398061e-01
1.11034596e+00 8.18960786e-01 -3.62950742e-01 -5.43356389e-02
3.59936893e-01 -2.09737673e-01 -4.14629579e-01 1.76928807e-02
-9.93017197e-01 8.65646780e-01 9.61440131e-02 2.23358020e-01
-6.38969660e-01 5.03984988e-01 -1.34060010e-01 6.69511557e-01
3.26204970e-02 1.67191997e-01 1.95756704e-01 4.14440930e-01
-9.06699717e-01 -2.65000924e-03 5.59179187e-01 1.22399831e+00
4.77925777e-01 6.19545817e-01 -1.96050093e-01 2.26146430e-01
7.61667132e-01 1.69794559e+00 -3.44222963e-01 -9.12187576e-01
6.18215084e-01 -2.97082573e-01 6.81872964e-02 -1.07755387e+00
-5.14648318e-01 -1.22943819e+00 -1.15559840e+00 4.79951441e-01
-1.47305563e-01 -5.95692933e-01 -7.60555983e-01 1.37327147e+00
2.76920289e-01 3.18133026e-01 8.58018875e-01 7.62838483e-01
6.95088208e-01 7.35707879e-01 -2.43844673e-01 -7.20039368e-01
1.44306040e+00 -2.43876517e-01 -1.13255394e+00 -4.23372149e-01
2.36297548e-01 -8.64443243e-01 -4.22839850e-01 3.98326516e-01
-7.67570257e-01 -3.82177830e-01 -1.63716900e+00 9.77994680e-01
2.57776856e-01 2.22855151e-01 3.76997828e-01 1.34593153e+00
-5.25238633e-01 -4.73231196e-01 -3.95572513e-01 1.11580595e-01
4.48126644e-01 1.71293437e-01 5.25264479e-02 -5.17993689e-01
-1.22558641e+00 9.13109541e-01 -1.30736589e-01 1.44181266e-01
-6.73019826e-01 -1.06795299e+00 -6.70541644e-01 -1.06818505e-01
1.42216295e-01 -4.12322938e-01 7.79205322e-01 -1.36734083e-01
-1.13130414e+00 -1.62896097e-01 -8.60033464e-03 -8.88173401e-01
-2.25715078e-02 1.28575759e-02 -9.24810946e-01 4.98425364e-01
-1.76745102e-01 4.59110945e-01 8.60391080e-01 -1.34263098e+00
-9.12883103e-01 -3.70761096e-01 -3.61386865e-01 -2.75561661e-01
-9.82778240e-03 -1.79563478e-01 9.12476778e-02 -5.29990852e-01
1.86860055e-01 -8.37647200e-01 -2.96694875e-01 -1.81466043e-01
-1.55062955e-02 2.90760159e-01 1.32761180e+00 -4.12556082e-01
1.06814873e+00 -2.29261208e+00 -5.00105858e-01 4.27342981e-01
2.56715655e-01 4.37333107e-01 -2.79728144e-01 5.60601771e-01
4.04040933e-01 -6.20828748e-01 1.71594515e-01 4.36640233e-02
-2.60227084e-01 -2.51194462e-02 -4.58825737e-01 8.58004391e-01
-8.22841600e-02 5.81356883e-01 -3.28435689e-01 -2.01989233e-01
1.28408432e-01 4.53648776e-01 -2.80269355e-01 -1.79415286e-01
1.92341805e-01 3.50617886e-01 -5.69952726e-01 7.60365069e-01
1.42072392e+00 3.97486836e-01 2.99426407e-01 -5.51992178e-01
-4.97823596e-01 -5.70810080e-01 -8.67282510e-01 6.63703263e-01
-5.57913542e-01 9.88809347e-01 7.84797251e-01 -1.20743418e+00
1.19117737e+00 4.33782190e-01 5.87384164e-01 -1.10993385e+00
1.80312589e-01 4.13515955e-01 3.17191482e-01 -3.01052809e-01
2.43807763e-01 -2.52919793e-01 -1.58226639e-01 4.09482569e-02
-1.78953752e-01 -6.58381954e-02 -3.45830359e-02 2.71233827e-01
1.26623428e+00 -8.41348946e-01 2.32693538e-01 -1.59030125e-01
6.92275345e-01 2.58858930e-02 5.44997215e-01 6.54870450e-01
6.06028251e-02 -3.56491625e-01 -3.63576353e-01 2.68881172e-01
-5.67273319e-01 -1.15833616e+00 -2.85896659e-01 -2.42906287e-02
6.55281126e-01 -3.88553999e-02 9.49686486e-03 -2.20564604e-01
3.87707561e-01 9.80266690e-01 1.22870483e-01 -2.41946116e-01
-4.95585561e-01 -3.72792840e-01 5.23005724e-01 -4.01907973e-02
6.62526965e-01 -9.92064998e-02 -9.27738726e-01 4.42092627e-01
1.98487863e-01 -1.49354041e+00 -9.18708295e-02 3.91492508e-02
-4.12264079e-01 -1.06300962e+00 -3.68061602e-01 -4.97180104e-01
3.86991918e-01 1.18644559e+00 3.34313750e-01 -2.28071049e-01
-3.35153997e-01 8.22604775e-01 -3.55546415e-01 -8.37224603e-01
-2.74219155e-01 -7.01314688e-01 2.41407499e-01 1.18508868e-01
1.98787585e-01 -5.15947640e-01 -4.54296649e-01 3.18226188e-01
-4.20387030e-01 -6.91408142e-02 1.18009698e+00 4.32917923e-01
-4.58160788e-02 1.12820588e-01 1.27955163e+00 -6.89877644e-02
2.32168302e-01 -6.01920366e-01 -1.06994057e+00 6.92523643e-02
-3.59655380e-01 -3.35846663e-01 3.74691039e-01 -1.18376642e-01
-1.04947686e+00 -1.17756337e-01 3.19421887e-02 -1.58440113e-01
2.38741532e-01 5.63212156e-01 -1.91331178e-01 -7.38603234e-01
4.14538801e-01 2.87594527e-01 4.07293856e-01 1.71608940e-01
3.87821436e-01 1.08938563e+00 6.22730374e-01 3.05405259e-01
1.50615883e+00 6.16258979e-01 5.35081029e-01 -1.49544907e+00
-2.15384603e-01 -4.30703789e-01 1.03251770e-01 -8.82087409e-01
8.18720341e-01 -1.36592185e+00 -6.85411036e-01 -1.64979398e-02
-1.31558335e+00 2.03506440e-01 2.77963281e-01 1.38925588e+00
-1.87713712e-01 4.15508389e-01 1.41699776e-01 -1.38345253e+00
-3.17059845e-01 -8.10134232e-01 6.30356550e-01 2.70326361e-02
2.67160922e-01 -3.12536359e-01 -2.97419876e-01 4.25556988e-01
8.52498293e-01 1.70564592e-01 6.21471524e-01 4.47938079e-03
-1.32054603e+00 -2.66049087e-01 -3.04512382e-01 -1.13470472e-01
-7.03593865e-02 -8.76902163e-01 -3.94154429e-01 -4.39743161e-01
8.52696747e-02 1.94811687e-01 8.71356726e-01 6.10423148e-01
4.37441140e-01 -4.63775471e-02 -8.61852527e-01 6.16236627e-01
1.45876241e+00 6.73630834e-01 7.83288240e-01 -1.56408414e-01
-9.66229364e-02 5.29425085e-01 1.12806320e+00 5.00152171e-01
2.22283155e-01 6.68025017e-01 8.23340535e-01 3.90629649e-01
-1.83311224e-01 5.24021149e-01 4.16131347e-01 5.60098231e-01
4.25007075e-01 -4.89085734e-01 -2.62205362e-01 2.88103938e-01
-1.44013619e+00 -1.20026350e+00 -6.38555110e-01 2.23707628e+00
1.90939531e-01 5.25792455e-03 -6.13211691e-01 -8.28432515e-02
7.14593589e-01 2.08457485e-02 -2.23860204e-01 3.58008184e-02
-3.47151235e-02 2.30928600e-01 1.29764628e+00 6.62636220e-01
-7.69176722e-01 1.04461052e-01 4.90951538e+00 1.16008615e+00
-1.07442224e+00 3.00217986e-01 -2.36506492e-01 -1.22334175e-02
-5.90532899e-01 -3.38120937e-01 -9.78492320e-01 2.50380158e-01
1.10097098e+00 -4.80069846e-01 -2.86890209e-01 3.27608466e-01
4.33743328e-01 -4.13822532e-01 -5.02577364e-01 1.13337457e+00
1.41305938e-01 -1.66530049e+00 -3.30359668e-01 5.95504224e-01
4.39703614e-01 -1.43275693e-01 7.69330934e-02 3.29171717e-01
-5.22685833e-02 -6.86044633e-01 3.82912129e-01 6.30285680e-01
7.87169814e-01 -8.95894587e-01 6.32815599e-01 5.30951083e-01
-1.36148727e+00 -5.16954124e-01 -2.30388954e-01 1.34189904e-01
6.53308690e-01 1.33475327e+00 -6.87908113e-01 1.01707375e+00
6.51097894e-02 -1.83196440e-02 1.31938487e-01 1.26759183e+00
9.99439135e-02 5.10685802e-01 -4.01477784e-01 -3.23592126e-01
1.05167121e-01 -2.83084273e-01 1.11399019e+00 1.07101870e+00
1.25955451e+00 5.70206523e-01 9.59161147e-02 1.90100133e-01
3.57840210e-01 -2.83309460e-01 -1.11467242e+00 1.97360784e-01
9.99839902e-01 1.34951532e+00 -2.21828163e-01 -1.37363136e-01
-6.52926207e-01 2.97372732e-02 -6.66071951e-01 4.78411049e-01
-1.01452875e+00 -7.16154575e-01 1.01665771e+00 1.37369722e-01
6.01829767e-01 -1.16429937e+00 -2.75038660e-01 -3.17060411e-01
-2.29620874e-01 -4.59331453e-01 -2.14457378e-01 -2.26028025e-01
-8.58913660e-01 3.25281471e-01 1.14035212e-01 -1.70339310e+00
-9.87739936e-02 -3.04798543e-01 -3.52818668e-01 6.60566866e-01
-2.01856875e+00 -9.63080943e-01 -5.39751709e-01 3.51919979e-01
2.62670726e-01 -7.96856642e-01 3.84679675e-01 6.59972727e-01
1.51036933e-01 4.05409604e-01 3.18461299e-01 -3.68784219e-01
4.91093993e-01 -3.04423779e-01 -2.26669207e-01 8.65394831e-01
-3.33054215e-01 1.78848848e-01 7.40455151e-01 -1.01230645e+00
-2.17035890e+00 -1.55658793e+00 5.67308724e-01 4.88919854e-01
3.42805713e-01 -2.82702297e-01 -1.17235512e-01 3.60077210e-02
5.43200731e-01 -2.19510734e-01 8.87858331e-01 -6.72165632e-01
-6.22514077e-02 -3.94079059e-01 -1.08265579e+00 4.16552424e-01
7.08961666e-01 2.67460704e-01 -1.58486471e-01 2.36031592e-01
4.13947701e-01 -5.40530160e-02 -4.53651309e-01 7.44980574e-01
5.99564016e-01 -5.78579187e-01 9.85012889e-01 1.52390778e-01
-2.68229157e-01 -6.15088165e-01 -8.20552230e-01 -1.19814110e+00
-3.70624423e-01 -5.68278432e-01 -1.33974582e-01 7.58329153e-01
2.68742949e-01 -9.75966454e-01 7.41385996e-01 -5.28960705e-01
-3.36209774e-01 -3.07810634e-01 -1.46499491e+00 -1.15561724e+00
-4.50850695e-01 -8.74219000e-01 1.12028629e-01 1.97618693e-01
-7.86608756e-02 4.35804486e-01 -5.27977049e-01 1.15206409e+00
1.48073053e+00 4.44441400e-02 9.68601465e-01 -1.47674870e+00
-1.62177429e-01 -9.37999338e-02 -7.63563097e-01 -1.45751047e+00
1.32609919e-01 -9.34252977e-01 3.89384516e-02 -1.49715626e+00
-5.54838896e-01 -6.18023694e-01 6.27493024e-01 -4.24721867e-01
4.45991099e-01 4.98245090e-01 4.33857411e-01 -1.28193542e-01
-9.39094089e-03 8.46503437e-01 1.09970391e+00 -2.54138112e-01
2.55199075e-01 4.31115180e-01 -4.73677009e-01 1.66852638e-01
7.89243579e-01 -1.62300065e-01 -5.11661112e-01 -1.90790579e-01
-1.30041286e-01 7.15829194e-01 4.82973844e-01 -1.41728961e+00
6.77327096e-01 -2.60969967e-01 2.60143369e-01 -9.74384189e-01
7.62366056e-01 -1.30397928e+00 4.28490013e-01 9.82588351e-01
4.92077023e-01 -6.76133931e-01 5.60959838e-02 1.01451468e+00
-1.33582652e-01 -1.29026100e-01 8.58591557e-01 5.77007174e-01
-5.66640973e-01 1.96610421e-01 -1.11260629e+00 -6.33541167e-01
1.69794953e+00 -3.06197405e-01 -7.53079951e-01 -8.48561943e-01
-2.89833605e-01 6.86604440e-01 -5.59590101e-01 3.37073326e-01
9.34947550e-01 -1.38437772e+00 -1.08391976e+00 2.76851028e-01
5.89746889e-03 -9.73622620e-01 4.79105264e-01 1.01770937e+00
9.70749184e-03 1.07845485e+00 -2.32466429e-01 -5.90645373e-01
-1.62648749e+00 2.74063721e-02 8.67944956e-03 3.46469164e-01
-4.18437749e-01 2.36706316e-01 1.37939289e-01 1.48413956e-01
7.13420585e-02 2.84451008e-01 -3.96177053e-01 6.98073432e-02
7.47745514e-01 3.37091327e-01 4.77687940e-02 -4.27662075e-01
-3.61966878e-01 5.80369413e-01 3.14682841e-01 -2.97121614e-01
1.03825355e+00 -3.13056678e-01 1.76340342e-01 -5.05446911e-01
8.91421616e-01 5.31930089e-01 -7.75994003e-01 4.78516668e-02
-3.72619808e-01 -5.62279999e-01 5.35450280e-01 -3.24953824e-01
-1.08933496e+00 5.65385282e-01 8.11747313e-01 2.62130946e-01
9.90258336e-01 -1.70603380e-01 7.90423810e-01 6.22843325e-01
9.58072007e-01 -7.82258451e-01 -6.90014213e-02 4.01525229e-01
7.13511229e-01 -6.40317202e-01 1.52859807e-01 -9.57022965e-01
-1.81263953e-01 1.29129386e+00 4.95390147e-01 1.81372344e-01
1.03931475e+00 8.38011086e-01 1.79411694e-01 -4.62440163e-01
-8.28933597e-01 -5.12663305e-01 -1.51027381e-01 1.20087802e+00
-1.87154680e-01 5.06140813e-02 -7.23327100e-01 5.10511816e-01
-1.40965044e-01 -3.79977405e-01 7.14569628e-01 8.62322807e-01
-1.18597460e+00 -1.08037829e+00 -8.83125722e-01 8.04124415e-01
8.13944861e-02 1.75738230e-01 3.16046715e-01 6.69156373e-01
6.30576238e-02 1.54066312e+00 9.85757411e-02 -4.69111562e-01
3.33193481e-01 -4.31279063e-01 6.02028608e-01 -2.06709117e-01
2.70063281e-01 4.27247435e-02 4.91509438e-01 -7.19926432e-02
-2.69979984e-01 -5.20361841e-01 -1.07763672e+00 -3.82755660e-02
-3.76235753e-01 3.98868710e-01 9.39908028e-01 9.09076035e-01
4.38131928e-01 7.28599787e-01 1.11671948e+00 -5.45065880e-01
-6.56101942e-01 -6.09928250e-01 -7.07297206e-01 -9.20035899e-01
4.50435162e-01 -8.59170198e-01 -2.67960191e-01 -3.77070785e-01] | [6.364284992218018, 1.2205051183700562] |
e522df09-63da-4a15-b8f7-5c07dfa88a19 | automated-writing-support-using-deep | null | null | https://aclanthology.org/2020.lrec-1.46 | https://aclanthology.org/2020.lrec-1.46.pdf | Automated Writing Support Using Deep Linguistic Parsers | This paper introduces a new web system that integrates English Grammatical Error Detection (GED) and course-specific stylistic guidelines to automatically review and provide feedback on student assignments. The system is being developed as a pedagogical tool for English Scientific Writing. It uses both general NLP methods and high precision parsers to check student assignments before they are submitted for grading. Instead of generalized error detection, our system aims to identify, with high precision, specific classes of problems that are known to be common among engineering students. Rather than correct the errors, our system generates constructive feedback to help students identify and correct them on their own. A preliminary evaluation of the system{'}s in-class performance has shown measurable improvements in the quality of student assignments. | ['Joseph MacKinnon', "Lu{\\'\\i}s Morgado da Costa", 'Roger V P Winder', 'Benedict Christopher Lin Tzer Liang', 'Shu Yun Li', 'Francis Bond'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['grammatical-error-detection'] | ['natural-language-processing'] | [-8.22085664e-02 2.96044827e-01 1.29501805e-01 -4.16001856e-01
-1.08839667e+00 -8.74456823e-01 -1.23918401e-02 9.72100377e-01
-3.71203721e-01 1.03110993e+00 -1.35994166e-01 -1.03581440e+00
-3.65000993e-01 -8.95318270e-01 -5.43581486e-01 2.58160561e-01
6.87068343e-01 3.40659797e-01 4.84237969e-01 -4.91584331e-01
1.09500837e+00 7.16682017e-01 -1.56502461e+00 1.12269178e-01
1.58319306e+00 3.41619998e-02 3.74907166e-01 1.12101710e+00
-6.77715838e-01 1.20010436e+00 -1.29263854e+00 -7.20571339e-01
-4.94568169e-01 -6.37781382e-01 -1.28605711e+00 -1.67770281e-01
8.27983320e-01 -7.85190985e-03 1.97837099e-01 1.26794696e+00
3.97758186e-01 3.63476276e-01 2.71885663e-01 -6.79427922e-01
-1.00724339e+00 5.39249897e-01 -1.08758621e-02 6.08335972e-01
9.52665627e-01 -3.43954384e-01 7.98367083e-01 -8.96616697e-01
6.35529816e-01 8.86326015e-01 7.49766827e-01 6.93421721e-01
-8.05491567e-01 -6.65069282e-01 -1.69117048e-01 3.30075622e-01
-9.90595281e-01 -1.42241806e-01 4.67186749e-01 -6.78880751e-01
9.32231903e-01 2.75550544e-01 7.30125427e-01 4.56109017e-01
5.12835979e-01 5.04653215e-01 1.03365231e+00 -1.15654016e+00
7.47259930e-02 7.00026512e-01 6.71684861e-01 1.14176118e+00
6.10086858e-01 -4.90794897e-01 -7.27210104e-01 1.72925353e-01
2.52876014e-01 -6.79809153e-01 -2.58988976e-01 2.72232860e-01
-5.73033929e-01 4.78701502e-01 -4.52376604e-01 3.18418443e-01
1.78159416e-01 -1.51065648e-01 1.69028416e-01 5.00346124e-01
9.65106580e-03 8.05561066e-01 -7.92376101e-01 -8.18583429e-01
-8.13933551e-01 3.47428530e-01 1.12099564e+00 1.14333820e+00
2.17127964e-01 -9.37933568e-03 -4.74948771e-02 9.83333230e-01
5.88261366e-01 2.13960141e-01 5.96763432e-01 -8.97975624e-01
5.22642612e-01 9.39661622e-01 4.35150117e-02 -8.69683981e-01
2.03741983e-01 -3.97237569e-01 2.06122518e-01 6.33978188e-01
6.81056261e-01 -1.27218559e-01 -6.31100893e-01 1.35902703e+00
1.06401779e-02 -2.40032345e-01 2.47358531e-02 1.52907684e-01
1.34129739e+00 4.21971709e-01 4.44519848e-01 3.71196046e-02
1.37036061e+00 -6.42496049e-01 -1.01703131e+00 -1.76483653e-02
1.10594106e+00 -1.31835735e+00 1.25293112e+00 1.04880202e+00
-1.53659999e+00 -6.74175143e-01 -8.94722223e-01 -1.87313065e-01
-4.80550170e-01 2.84958899e-01 4.03483398e-02 1.26964867e+00
-1.01302576e+00 9.13133025e-01 -2.84408629e-01 -2.24251509e-01
8.27104785e-03 7.12023750e-02 -1.99572355e-01 5.84131144e-02
-8.31393480e-01 1.01249957e+00 1.58729181e-01 -4.40256357e-01
-5.09857684e-02 -1.04684222e+00 -9.81704116e-01 1.72471777e-01
7.70098120e-02 -2.16731638e-01 1.91769290e+00 -3.57823223e-01
-1.91063201e+00 1.19228101e+00 -2.26238891e-01 3.91374439e-01
3.31018358e-01 -3.62298906e-01 -3.08986068e-01 -7.49718174e-02
2.70221084e-01 7.13770017e-02 -1.20790802e-01 -8.31399500e-01
-1.04231000e+00 -7.23498166e-02 -6.34958744e-02 2.42860749e-01
-2.58017778e-01 5.61445415e-01 -1.97304934e-01 -5.96757352e-01
2.55414695e-01 -4.35132682e-01 2.00201616e-01 -3.31206828e-01
8.50376561e-02 -9.02991176e-01 4.40756679e-01 -9.83976841e-01
1.70375681e+00 -1.57995522e+00 -2.05117643e-01 3.41353267e-01
1.98248699e-01 4.57860202e-01 2.25359410e-01 2.69974321e-01
-2.74367064e-01 5.98196268e-01 3.06839794e-01 1.03911445e-01
1.69084147e-01 -8.78222585e-02 1.56591669e-01 -2.79513717e-01
1.28150312e-02 4.25493062e-01 -1.42195618e+00 -6.28563702e-01
2.17918321e-01 -1.50825843e-01 -3.74603063e-01 3.69664401e-01
1.19782455e-01 -3.71087226e-03 -2.90065169e-01 6.46924853e-01
2.08169982e-01 5.94343096e-02 5.81479669e-02 1.02527332e+00
-2.53297180e-01 1.00731134e+00 -1.44090080e+00 1.43708658e+00
-7.59089291e-01 8.78914654e-01 -2.65608817e-01 -6.03149056e-01
1.24052906e+00 3.70972306e-01 -9.81836244e-02 -4.48592722e-01
-1.81341067e-01 5.30775726e-01 -9.96867642e-02 -8.76295984e-01
8.19842935e-01 1.62981808e-01 -3.43654156e-02 5.58013618e-01
3.93765122e-01 -4.90669608e-01 8.88899446e-01 6.07722521e-01
1.14144278e+00 4.46896434e-01 2.67548889e-01 -4.21147645e-01
6.84024215e-01 1.13248803e-01 4.72787231e-01 7.90671706e-01
-1.79766029e-01 2.07480803e-01 3.68766338e-01 -1.45154759e-01
-6.75953925e-01 -9.02821481e-01 -1.16920114e-01 1.27277935e+00
-3.45304847e-01 -9.20772254e-01 -8.71401310e-01 -8.89076173e-01
-2.52300113e-01 1.09996331e+00 -1.49306133e-02 -7.84004182e-02
-2.65485734e-01 -3.45275581e-01 1.96844608e-01 4.01985645e-01
2.15219915e-01 -1.22137773e+00 -6.09695435e-01 4.54254150e-01
-3.08302104e-01 -7.40826607e-01 -2.41881937e-01 2.27992877e-01
-7.77760386e-01 -1.09219754e+00 -1.94872737e-01 -1.28221822e+00
1.08201277e+00 -1.11811057e-01 1.51980996e+00 7.04898417e-01
-2.59504676e-01 7.20836580e-01 -5.38686574e-01 -8.20389390e-01
-9.25629258e-01 -1.16906799e-01 -2.14342713e-01 -1.36001563e+00
9.66703475e-01 -1.71466723e-01 1.18346654e-01 -2.11331546e-01
-4.38023448e-01 -2.18407899e-01 3.19269538e-01 6.04056001e-01
1.15089275e-01 2.73841411e-01 6.23962820e-01 -1.30367768e+00
1.31951141e+00 -1.31487949e-02 -6.49487913e-01 6.17285728e-01
-8.91109884e-01 -2.54638270e-02 5.25813043e-01 1.43193975e-01
-1.14061904e+00 -2.98395872e-01 -6.10157609e-01 6.40356779e-01
-6.37817085e-01 6.98491275e-01 -8.69712085e-02 -7.11535573e-01
9.06732142e-01 -1.40050068e-01 -5.76867223e-01 -4.23554331e-01
-5.22026479e-01 8.29199314e-01 4.81387049e-01 -1.06965327e+00
5.57180524e-01 -1.06977916e+00 -1.96055979e-01 -6.33526921e-01
-1.06113183e+00 -4.69068527e-01 -5.99830329e-01 -5.90884626e-01
3.65603834e-01 -6.95419550e-01 -8.47116649e-01 2.06607774e-01
-1.21246231e+00 -3.65303248e-01 -4.61101770e-01 4.53811437e-01
4.32779565e-02 4.71681386e-01 -8.90921295e-01 -7.71517456e-01
2.82993447e-02 -1.01429486e+00 4.52857703e-01 9.78901207e-01
-6.83437347e-01 -1.08768499e+00 2.64682025e-01 6.92825615e-01
-7.39237815e-02 -2.08595052e-01 1.02482545e+00 -6.67233229e-01
-1.65590733e-01 -2.41057128e-01 2.23926619e-01 4.59177017e-01
-8.80880356e-02 4.96401578e-01 -5.85868239e-01 2.64821649e-01
-3.44830513e-01 -1.46549210e-01 4.19136062e-02 -1.28257573e-01
1.29576802e+00 -1.74995467e-01 5.31831793e-02 4.60217446e-02
1.35487795e+00 4.05280530e-01 4.33113545e-01 8.23736727e-01
3.42327118e-01 7.02727854e-01 5.17987490e-01 7.22369924e-02
3.55029047e-01 3.49240899e-01 -2.44162127e-01 6.93683267e-01
-2.04241395e-01 -3.43452662e-01 4.76193160e-01 1.41099906e+00
4.77807261e-02 -1.86837643e-01 -1.28777456e+00 8.52000117e-01
-1.30360925e+00 -7.64186144e-01 -9.42477763e-01 2.04355073e+00
1.20448053e+00 3.78601342e-01 -1.72269374e-01 2.97270268e-01
4.51311827e-01 -7.44130254e-01 3.92006695e-01 -1.40273428e+00
4.50257361e-01 1.10330379e+00 2.67981678e-01 1.00463212e+00
-3.27417225e-01 9.44641292e-01 7.05171108e+00 4.61829990e-01
-4.88052666e-01 -1.34716168e-01 2.91592807e-01 4.45784032e-01
-5.40845931e-01 -1.63761824e-01 -1.24426961e+00 4.25360292e-01
1.24621272e+00 -4.96492267e-01 -1.00340508e-01 8.53229523e-01
2.17816502e-01 -2.46512547e-01 -6.29087508e-01 4.04892117e-01
5.16757779e-02 -1.27960348e+00 -2.00018913e-01 -3.60078514e-01
8.15059364e-01 -8.27572227e-01 -3.89634937e-01 5.42058468e-01
1.01254666e+00 -8.81220162e-01 6.66803062e-01 5.73534906e-01
3.41225117e-01 -1.04883933e+00 7.78198302e-01 2.69607633e-01
-5.54444492e-01 1.36420250e-01 -4.11277682e-01 -4.52475995e-01
-4.69986677e-01 5.10735393e-01 -1.03167975e+00 2.82931030e-01
8.70003223e-01 1.97464198e-01 -1.00552332e+00 1.42788994e+00
-1.05443513e+00 8.95938933e-01 6.84022754e-02 -6.16306663e-01
-2.51826823e-01 -1.54165238e-01 4.06409800e-01 1.41889751e+00
6.80430651e-01 5.49586535e-01 8.30052793e-02 7.86884606e-01
-8.37967768e-02 6.37844324e-01 -2.86829084e-01 1.19484872e-01
7.12769210e-01 1.18721223e+00 -6.12070799e-01 -3.79236966e-01
-1.16758540e-01 8.33692431e-01 2.97250986e-01 7.76417777e-02
-1.73591256e-01 -1.35743904e+00 3.73981565e-01 1.87243387e-01
-3.50064486e-02 -3.17050338e-01 -7.31093228e-01 -8.52723718e-01
1.23846889e-01 -1.00683272e+00 1.76216021e-01 -9.53993499e-01
-8.82770300e-01 1.65307134e-01 -4.56696689e-01 -8.07983041e-01
-1.40046537e-01 -9.99337316e-01 -1.05963862e+00 1.26157856e+00
-1.27595532e+00 -8.35541636e-02 -5.40867269e-01 9.89192352e-02
5.96127093e-01 -3.47813666e-01 9.39584196e-01 1.69562414e-01
-6.74407065e-01 9.45050001e-01 1.07547166e-02 9.59133804e-02
9.15104806e-01 -2.08359027e+00 1.44991443e-01 1.06610560e+00
2.38013610e-01 6.98411644e-01 8.48593414e-01 -8.07871222e-01
-7.45192707e-01 -5.73491693e-01 1.97452414e+00 -7.03258038e-01
5.40730715e-01 1.75130606e-01 -1.07340384e+00 3.92418236e-01
2.83062518e-01 -5.16171873e-01 1.08604467e+00 2.46082842e-01
1.12408763e-02 1.20251048e-02 -1.10184050e+00 3.61104310e-01
6.29710197e-01 -3.71304870e-01 -1.26155376e+00 6.63948715e-01
1.51660219e-01 -9.62775648e-01 -1.15438294e+00 -5.62802516e-02
9.09456760e-02 -5.29523015e-01 4.66186732e-01 -7.53608882e-01
9.78984237e-01 -2.70951360e-01 6.73147917e-01 -1.46695352e+00
-2.80660272e-01 -7.09805787e-01 2.51366287e-01 1.48054314e+00
6.38323486e-01 -8.27387348e-02 9.19926286e-01 1.01393735e+00
-5.45889735e-01 -3.34922880e-01 -3.69734347e-01 -4.66824412e-01
3.22352201e-01 -3.63469452e-01 1.54746100e-01 1.17690074e+00
6.72055960e-01 1.87762812e-01 4.09763575e-01 1.29998296e-01
3.68215919e-01 -2.82663822e-01 5.04121900e-01 -1.67249358e+00
-4.40265350e-02 -6.60448194e-01 -4.73611355e-01 -7.04759121e-01
1.97096333e-01 -8.19580615e-01 3.39843899e-01 -1.65564668e+00
-3.07100385e-01 -3.90225321e-01 -3.88243273e-02 6.30748749e-01
-5.66804767e-01 3.14358212e-02 -1.68947056e-01 -5.44552445e-01
-5.77007353e-01 -2.67099380e-01 9.37087893e-01 4.87154752e-01
-2.81526804e-01 1.05441548e-02 -1.08258355e+00 7.51723528e-01
9.08345461e-01 -5.77305794e-01 -9.21411961e-02 -3.03558260e-01
6.85316384e-01 -4.13053893e-02 -4.85382490e-02 -1.19388330e+00
4.79129523e-01 -5.30150294e-01 5.47400773e-01 -1.90596074e-01
-9.62204576e-01 -4.45854455e-01 -4.61474001e-01 1.57340959e-01
-6.07071638e-01 5.85050762e-01 7.39733279e-01 -6.79175332e-02
-2.20184222e-01 -1.37032449e+00 7.29522288e-01 -3.49871963e-01
-7.44018734e-01 -5.18928766e-01 -9.67333913e-01 1.33691251e-01
9.87179220e-01 -2.59952128e-01 -3.42184722e-01 -3.01771849e-01
-7.85903752e-01 2.57595330e-01 1.93451747e-01 3.35831314e-01
5.56173682e-01 -1.00324321e+00 -6.62560940e-01 2.85136588e-02
1.49189040e-01 -3.76084894e-02 -9.52486768e-02 1.95460081e-01
-1.28675306e+00 5.58103442e-01 -4.59781200e-01 -1.08025454e-01
-1.79203868e+00 -1.45438403e-01 3.13605040e-01 -4.79199290e-01
-3.93936753e-01 1.36597216e+00 -8.19250345e-01 -9.37931776e-01
4.68245834e-01 -2.36541033e-01 -5.57367682e-01 -2.21123993e-01
9.62056518e-01 3.63976717e-01 7.16518521e-01 3.96445543e-02
6.32539988e-02 3.32339615e-01 -2.44224295e-01 -2.28490196e-02
1.15571427e+00 -1.47814348e-01 9.34599862e-02 4.93626654e-01
3.53720069e-01 8.20489883e-01 -3.48859429e-01 -1.59609690e-02
6.27560019e-01 -5.22104740e-01 1.09424040e-01 -1.60634685e+00
-5.36650121e-01 8.05397868e-01 3.60217869e-01 1.13675967e-01
7.75181293e-01 -5.40887058e-01 4.28526163e-01 7.00785756e-01
1.30254375e-02 -1.70428002e+00 1.26274303e-02 9.10719693e-01
4.45342600e-01 -1.10346878e+00 -9.81341116e-03 -5.93166113e-01
-2.00407982e-01 1.77365434e+00 1.19830728e+00 1.46389157e-01
2.90229023e-01 1.14236087e-01 1.39218152e-01 -2.11134236e-02
-6.95366144e-01 2.43806109e-01 4.83607799e-01 4.25560445e-01
1.36030912e+00 -1.20082878e-01 -9.22627926e-01 5.82584381e-01
-5.65477014e-01 1.46640554e-01 1.34361601e+00 1.28963065e+00
-1.00291991e+00 -1.54467404e+00 -6.74491942e-01 3.43116760e-01
-8.65433574e-01 -3.40457320e-01 -8.07226300e-01 5.95742106e-01
4.28966284e-02 9.75757182e-01 -3.74780178e-01 -9.72629786e-02
7.28878438e-01 7.69970775e-01 7.12806880e-01 -1.17596614e+00
-1.16888916e+00 -7.25987732e-01 9.85012427e-02 -8.63785297e-03
5.20229824e-02 -6.60118520e-01 -1.38540101e+00 -5.62114835e-01
-3.60692382e-01 8.02759647e-01 8.65954638e-01 1.13818264e+00
3.61439288e-02 1.08223069e+00 -9.78374779e-02 2.64640361e-01
-7.63450265e-01 -8.36292863e-01 -3.81472737e-01 1.77024350e-01
-1.00953542e-01 -2.38736093e-01 -1.56151414e-01 1.99043989e-01] | [11.224066734313965, 9.475326538085938] |
3cc4cb08-a7c4-48c9-8bbc-79f61a9219f3 | does-meta-learning-help-mbert-for-few-shot | null | null | https://aclanthology.org/2022.coling-1.373 | https://aclanthology.org/2022.coling-1.373.pdf | Does Meta-learning Help mBERT for Few-shot Question Generation in a Cross-lingual Transfer Setting for Indic Languages? | Few-shot Question Generation (QG) is an important and challenging problem in the Natural Language Generation (NLG) domain. Multilingual BERT (mBERT) has been successfully used in various Natural Language Understanding (NLU) applications. However, the question of how to utilize mBERT for few-shot QG, possibly with cross-lingual transfer, remains. In this paper, we try to explore how mBERT performs in few-shot QG (cross-lingual transfer) and also whether applying meta-learning on mBERT further improves the results. In our setting, we consider mBERT as the base model and fine-tune it using a seq-to-seq language modeling framework in a cross-lingual setting. Further, we apply the model agnostic meta-learning approach to our base model. We evaluate our model for two low-resource Indian languages, Bengali and Telugu, using the TyDi QA dataset. The proposed approach consistently improves the performance of the base model in few-shot settings and even works better than some heavily parameterized models. Human evaluation also confirms the effectiveness of our approach. | ['Pawan Goyal', 'Sudeshna Sarkar', 'Amrith Krishna', 'Ashim Gupta', 'Isha Sharma', 'Rupak Kumar Thakur', 'Aniruddha Roy'] | null | null | null | null | coling-2022-10 | ['question-generation'] | ['natural-language-processing'] | [ 6.84291497e-02 1.00925177e-01 -9.86515805e-02 -1.61597118e-01
-1.58187699e+00 -6.00870728e-01 8.20163786e-01 -1.81406781e-01
-4.36516047e-01 1.11007094e+00 4.13106441e-01 -4.08126801e-01
1.82827860e-01 -8.52194667e-01 -7.83626854e-01 -3.90657037e-01
3.72467816e-01 8.26485395e-01 1.75687894e-01 -9.45605993e-01
8.69980752e-02 -2.56561220e-01 -1.23648572e+00 3.13624829e-01
1.43576658e+00 2.74849981e-01 3.91438812e-01 8.77482116e-01
-4.37255800e-01 9.80441928e-01 -6.54244781e-01 -7.48510718e-01
1.37224138e-01 -8.04613411e-01 -1.14999115e+00 -2.63127834e-01
4.26392674e-01 -2.59015169e-02 2.55347162e-01 7.25252450e-01
8.85870039e-01 4.33769524e-01 6.43842459e-01 -1.22118664e+00
-8.77563477e-01 9.24194574e-01 -3.69647682e-01 1.83366522e-01
3.55712265e-01 2.06044465e-01 1.16337383e+00 -8.60827625e-01
7.78737605e-01 1.55519807e+00 4.04512346e-01 7.62368381e-01
-1.02817643e+00 -6.61204755e-01 9.23484936e-03 3.35014284e-01
-1.19109237e+00 -5.43441832e-01 4.78284776e-01 -2.23404363e-01
1.10369313e+00 -2.22485676e-01 9.81904566e-02 1.16674566e+00
1.41673371e-01 1.05841827e+00 1.22529876e+00 -9.18257296e-01
2.45070651e-01 1.10513054e-01 3.59298736e-02 6.48528457e-01
-2.45522767e-01 -3.81812640e-02 -6.36624634e-01 -1.23415142e-01
2.25820422e-01 -7.78474212e-01 -1.25983059e-01 8.85366872e-02
-1.21406758e+00 1.19510245e+00 1.95692871e-02 4.09623742e-01
-1.09631278e-01 2.86552250e-01 3.29551160e-01 5.25232196e-01
7.40278065e-01 7.35873759e-01 -7.81530559e-01 -4.66320157e-01
-8.43475878e-01 3.56958866e-01 8.36272240e-01 1.30846322e+00
7.77521670e-01 1.95689186e-01 -7.59644091e-01 1.27832186e+00
-3.77601460e-02 3.86860967e-01 7.72088408e-01 -6.83030009e-01
7.63858020e-01 1.76754922e-01 1.39860749e-01 -1.20586112e-01
-5.95136806e-02 -1.04528412e-01 -5.97079098e-01 -2.89092898e-01
3.97546589e-01 -6.16683006e-01 -1.09591746e+00 2.04779553e+00
3.99959207e-01 1.78457886e-01 4.50040549e-01 5.59813440e-01
7.50414491e-01 1.16132259e+00 4.03731763e-01 -2.21790954e-01
1.55638111e+00 -1.32827854e+00 -6.50382757e-01 -2.74642766e-01
1.00327754e+00 -8.59227717e-01 1.54488206e+00 1.97149426e-01
-8.60871971e-01 -6.18108213e-01 -6.71922863e-01 -2.96065211e-01
-4.10389662e-01 4.15653028e-02 4.32367593e-01 7.12705910e-01
-9.39792931e-01 3.94176364e-01 -5.14513195e-01 -8.07744801e-01
-4.21740115e-02 -1.20464109e-01 -6.46256357e-02 -4.52845275e-01
-1.81808794e+00 1.08439422e+00 6.05023801e-01 -3.33648503e-01
-8.60517204e-01 -7.51184881e-01 -9.33794379e-01 -1.01257257e-01
7.35933304e-01 -8.28164697e-01 1.68556464e+00 -6.99235022e-01
-1.87210214e+00 4.70852733e-01 -2.90636599e-01 -5.06134093e-01
4.12771553e-01 -3.30909431e-01 -2.49449626e-01 -2.20447090e-02
3.73643488e-01 9.14353788e-01 6.79114103e-01 -9.93752301e-01
-6.57864571e-01 2.04153378e-02 3.84745508e-01 3.19425106e-01
1.04415240e-02 9.85377282e-02 -3.21489155e-01 -8.33489716e-01
-5.52519143e-01 -1.00240421e+00 -2.24023923e-01 -7.00334728e-01
-7.75204524e-02 -6.10141098e-01 3.38448912e-01 -7.94676542e-01
1.06249285e+00 -1.64820993e+00 6.96776137e-02 -5.18366754e-01
-6.52220547e-01 3.56706798e-01 -5.82551658e-01 9.31956530e-01
3.33663881e-01 7.49893114e-02 -4.96097267e-01 -2.65969098e-01
1.15018874e-01 2.85327673e-01 -3.17049056e-01 -2.11946368e-01
3.99661392e-01 1.41869402e+00 -1.21666896e+00 -6.77853525e-01
-9.94690433e-02 1.81498542e-01 -5.46629906e-01 4.18958575e-01
-6.84284925e-01 3.51058692e-01 -3.42248201e-01 5.65445006e-01
2.94147313e-01 -3.57757732e-02 9.65295509e-02 -5.12682088e-02
-7.13056624e-02 2.57460028e-01 -6.87715888e-01 2.06578803e+00
-9.76554394e-01 2.69321203e-01 -5.26408076e-01 -8.60832095e-01
8.56669962e-01 5.68598747e-01 -4.13042940e-02 -9.63897705e-01
-9.53484401e-02 2.42093310e-01 2.55787641e-01 -6.03765190e-01
8.77526939e-01 -6.36426687e-01 -3.57147098e-01 9.28326666e-01
5.98781288e-01 -3.64933223e-01 5.71164548e-01 3.78403574e-01
6.69798374e-01 6.01202250e-01 6.77443624e-01 -2.91572750e-01
3.81334245e-01 1.49870604e-01 2.83903450e-01 9.07323420e-01
-1.92874923e-01 6.03685200e-01 3.23991865e-01 1.90660089e-01
-9.60526288e-01 -8.43987644e-01 2.95999378e-01 1.67548442e+00
-1.07109286e-01 -3.22997540e-01 -8.58177125e-01 -8.93677294e-01
-3.48910391e-01 1.47796631e+00 -5.39979339e-01 -2.00092301e-01
-5.88529766e-01 -1.03700125e+00 6.19308233e-01 2.97676563e-01
5.07567108e-01 -1.30549288e+00 -3.36626023e-01 6.99194312e-01
-5.94789803e-01 -1.15408170e+00 -6.45625114e-01 -1.35005713e-01
-6.17248416e-01 -7.12355137e-01 -1.04605365e+00 -6.31248653e-01
1.19005330e-01 1.28757477e-01 1.34942150e+00 -3.76539886e-01
-2.69689225e-02 3.80839735e-01 -1.00379813e+00 -5.33307731e-01
-8.19502831e-01 5.60388744e-01 -2.02663630e-01 -1.33047432e-01
2.15735599e-01 -2.05435693e-01 -1.38677657e-01 1.37865558e-01
-8.55031550e-01 2.31101573e-01 5.45095384e-01 1.00534260e+00
2.82591224e-01 -4.49498743e-01 1.32615626e+00 -1.29647040e+00
9.86028671e-01 -5.61953962e-01 -4.85291988e-01 8.05201113e-01
-4.56159085e-01 4.20225173e-01 6.85660839e-01 -3.85528475e-01
-1.53450000e+00 -4.31456178e-01 -2.55254656e-01 -4.04518880e-02
7.28198588e-02 6.83130741e-01 -2.46453822e-01 3.82127136e-01
6.56688511e-01 1.94466636e-01 -3.08277071e-01 -4.27246571e-01
9.75483000e-01 7.45326459e-01 1.30039930e-01 -8.63360226e-01
6.51436388e-01 -1.85326654e-02 -4.98640031e-01 -7.49438286e-01
-1.24313462e+00 -2.91134983e-01 -5.57908177e-01 -4.48511727e-02
8.74588788e-01 -9.36853647e-01 -1.94204170e-02 3.13808858e-01
-1.31347811e+00 -7.84706652e-01 -3.21636111e-01 3.17931294e-01
-7.95777380e-01 1.10314570e-01 -4.78135943e-01 -7.84342468e-01
-7.31937051e-01 -9.97578025e-01 1.23428333e+00 8.52261409e-02
-1.17764279e-01 -1.26820993e+00 5.57436049e-01 5.66434264e-01
4.60198939e-01 -1.60189196e-01 1.20912969e+00 -6.04306161e-01
-3.83956045e-01 2.13481188e-01 -5.39973453e-02 2.84167200e-01
3.59938323e-01 -3.63650769e-01 -1.07480228e+00 -3.67408991e-01
-2.81106085e-01 -7.84867883e-01 7.29540884e-01 1.22908004e-01
6.27433836e-01 -2.76720643e-01 1.05058067e-01 2.11865067e-01
1.45138037e+00 2.02947762e-02 5.00840366e-01 2.24939689e-01
6.44778311e-01 7.26896465e-01 1.06323862e+00 1.34392470e-01
7.54481316e-01 7.41828203e-01 -2.33167902e-01 1.76550508e-01
-2.65761256e-01 -4.75034028e-01 6.76806390e-01 1.28916800e+00
8.78077932e-03 -6.22656167e-01 -1.01736617e+00 7.95776606e-01
-1.95603395e+00 -9.61977959e-01 2.85316974e-01 1.90293443e+00
1.21270895e+00 -2.50380635e-01 1.45659340e-03 -4.30291414e-01
4.95026886e-01 1.50773659e-01 -2.66031682e-01 -7.44255424e-01
-1.02395996e-01 4.99696672e-01 1.10805668e-01 5.76356292e-01
-8.53635848e-01 1.67652416e+00 5.78874016e+00 1.15593684e+00
-9.16092992e-01 6.24152005e-01 5.81515431e-01 1.15322724e-01
-2.93184966e-01 1.60421938e-01 -9.82756376e-01 2.34749362e-01
1.28755796e+00 -7.28782654e-01 3.83992910e-01 6.11072779e-01
1.93057954e-01 -1.40466645e-01 -1.05764210e+00 5.60303450e-01
4.45857167e-01 -1.05223346e+00 3.41311872e-01 -2.78147697e-01
1.16796792e+00 8.66484568e-02 -2.59496093e-01 1.09320164e+00
7.12843239e-01 -9.46203351e-01 6.16429090e-01 2.61976898e-01
8.63748729e-01 -9.10286903e-01 7.90488601e-01 6.82774603e-01
-9.57496107e-01 1.73622295e-01 -5.20276487e-01 2.21717775e-01
5.71864545e-01 4.47962731e-02 -1.29808831e+00 9.25174892e-01
3.55940521e-01 3.50761145e-01 -6.04105949e-01 6.89930916e-01
-4.73866850e-01 9.31757271e-01 2.23056912e-01 -1.68845132e-01
4.67284381e-01 -1.31478474e-01 3.89673412e-01 1.20395136e+00
4.56125796e-01 1.57894760e-01 1.37274355e-01 7.54064262e-01
-2.25306764e-01 5.41524649e-01 -5.01859844e-01 -2.03600958e-01
2.94480711e-01 1.19037175e+00 -2.57986605e-01 -7.00981736e-01
-4.07959700e-01 1.07103479e+00 5.80060601e-01 3.74806672e-01
-5.90481699e-01 -4.98239994e-01 2.80000567e-01 -1.56229004e-01
3.35551709e-01 -1.12994842e-01 3.64712715e-01 -1.34821403e+00
-4.27674025e-01 -1.19584382e+00 4.47447300e-01 -8.96191061e-01
-1.62228894e+00 7.33247042e-01 3.65296066e-01 -1.19329226e+00
-1.02987099e+00 -2.84284651e-01 -6.18318915e-01 9.36504424e-01
-1.81134987e+00 -1.57478774e+00 9.00715217e-02 3.81089985e-01
1.14430809e+00 -2.34157115e-01 8.23165238e-01 7.07072914e-02
-2.35968694e-01 7.47579813e-01 9.32873636e-02 -4.54659574e-03
9.89520669e-01 -1.30628347e+00 7.71725833e-01 9.83216763e-01
3.60178411e-01 5.19065261e-01 6.42991900e-01 -6.62011981e-01
-1.24944437e+00 -1.47014356e+00 1.13224769e+00 -4.84259188e-01
7.88596630e-01 -4.61916924e-01 -9.64739442e-01 6.43941641e-01
6.90806389e-01 -3.46415579e-01 6.64080918e-01 1.26531631e-01
-1.76058903e-01 1.43480018e-01 -9.83567655e-01 7.14951932e-01
7.24858701e-01 -4.19478208e-01 -9.27524984e-01 3.79602581e-01
1.15649307e+00 -1.11208968e-01 -6.68405950e-01 1.19118005e-01
3.51560056e-01 -5.18731236e-01 4.96611893e-01 -8.39489222e-01
5.39881825e-01 -9.92833674e-02 -1.76399589e-01 -1.93784440e+00
-3.20887491e-02 -6.94009781e-01 9.33654979e-02 1.58244097e+00
6.56598151e-01 -4.95048851e-01 4.29128379e-01 1.55781344e-01
-1.59740314e-01 -5.19803405e-01 -8.96247447e-01 -1.00902998e+00
6.85947895e-01 -5.40889680e-01 7.04764485e-01 8.81975532e-01
-1.34027600e-01 8.76863301e-01 -6.25392556e-01 -2.07235441e-01
5.30772507e-01 1.25979766e-01 1.01263416e+00 -6.37037337e-01
-5.10989785e-01 4.26093042e-02 2.99729079e-01 -8.45675111e-01
3.86962414e-01 -1.10313177e+00 4.38698411e-01 -1.68150294e+00
1.52430162e-01 -4.05718565e-01 -1.35289937e-01 3.93586040e-01
-7.44223177e-01 1.62822559e-01 4.19753790e-01 -2.05721483e-01
-5.39483368e-01 1.04170966e+00 1.33063281e+00 -1.21993653e-01
-2.19810590e-01 -2.31562406e-01 -7.25078881e-01 2.82730818e-01
8.34109128e-01 -4.86777455e-01 -6.16685271e-01 -6.19324386e-01
1.25415340e-01 1.69967473e-01 -1.62733287e-01 -5.21679401e-01
7.84839410e-03 -2.81872153e-01 -3.27139139e-01 -3.74633461e-01
3.12138110e-01 -5.22013046e-02 -2.93957263e-01 1.78232282e-01
-3.62039626e-01 -4.47202101e-02 2.53698260e-01 2.75022537e-01
-3.04847449e-01 -5.39237559e-01 7.15062022e-01 -4.59776551e-01
-7.94907570e-01 3.57896537e-01 -2.60906041e-01 7.06499457e-01
7.67457902e-01 1.73636805e-02 -4.52850074e-01 -5.58320224e-01
-2.72351980e-01 4.33157176e-01 1.47125378e-01 8.09127808e-01
3.47085118e-01 -1.29468346e+00 -1.23167276e+00 -1.20322168e-01
6.70042753e-01 -3.13257903e-01 2.53498018e-01 6.65369987e-01
-1.99286222e-01 5.71151078e-01 4.21782583e-02 -2.88422972e-01
-8.39242637e-01 2.62005776e-01 2.68664837e-01 -5.50578535e-01
-2.06786707e-01 8.23147953e-01 2.25623950e-01 -8.86003554e-01
-3.77814651e-01 -5.23888059e-02 -3.24837774e-01 3.01338762e-01
3.70478213e-01 3.25432181e-01 6.60265088e-02 -5.54906249e-01
2.25862656e-02 4.23026651e-01 -2.69828886e-01 -7.05270171e-01
1.10902321e+00 -2.14865491e-01 1.24528237e-01 8.61717582e-01
9.83027458e-01 -1.07686713e-01 -8.74287605e-01 -3.36337566e-01
2.90189832e-01 -1.26535207e-01 -2.42999643e-01 -9.56776321e-01
-5.72895110e-01 1.06495547e+00 1.94123939e-01 -2.19399467e-01
7.41523683e-01 7.79188424e-02 9.91652489e-01 4.06051844e-01
6.84321284e-01 -1.12506950e+00 9.33834314e-02 9.35390949e-01
9.66373086e-01 -1.47052538e+00 -3.16972882e-01 -5.36629632e-02
-1.07814646e+00 7.00385988e-01 8.06735933e-01 1.81824222e-01
3.17539394e-01 -2.60179579e-01 3.46664548e-01 2.12776944e-01
-1.32098389e+00 -4.10606325e-01 3.05122197e-01 6.39792264e-01
7.49881804e-01 4.08402868e-02 -4.38379139e-01 5.03287554e-01
-5.95408320e-01 1.92576230e-01 6.77230597e-01 9.01467204e-01
-3.76227736e-01 -1.41208220e+00 -2.23304018e-01 2.69486547e-01
-3.76869738e-01 -6.30367219e-01 -1.74206614e-01 8.74351859e-01
5.59394434e-02 1.06686795e+00 -1.28043368e-01 1.76501758e-02
2.02165902e-01 3.89286608e-01 6.33563280e-01 -1.05943763e+00
-7.11128473e-01 7.17061833e-02 3.37829620e-01 -1.63083926e-01
-3.94767523e-01 -5.29418409e-01 -8.74340951e-01 1.69259906e-01
-5.11223257e-01 4.32864726e-01 5.28065205e-01 1.18387163e+00
2.16922402e-01 3.90355289e-01 4.98296767e-01 -2.34855592e-01
-6.61970675e-01 -1.50536156e+00 -2.80159384e-01 1.64007425e-01
2.75148582e-02 -4.41762269e-01 -1.45556405e-01 1.32254705e-01] | [11.772818565368652, 9.007895469665527] |
05895a33-9563-4261-9ced-139bd54821b1 | future-transformer-for-long-term-action | 2205.14022 | null | https://arxiv.org/abs/2205.14022v1 | https://arxiv.org/pdf/2205.14022v1.pdf | Future Transformer for Long-term Action Anticipation | The task of predicting future actions from a video is crucial for a real-world agent interacting with others. When anticipating actions in the distant future, we humans typically consider long-term relations over the whole sequence of actions, i.e., not only observed actions in the past but also potential actions in the future. In a similar spirit, we propose an end-to-end attention model for action anticipation, dubbed Future Transformer (FUTR), that leverages global attention over all input frames and output tokens to predict a minutes-long sequence of future actions. Unlike the previous autoregressive models, the proposed method learns to predict the whole sequence of future actions in parallel decoding, enabling more accurate and fast inference for long-term anticipation. We evaluate our method on two standard benchmarks for long-term action anticipation, Breakfast and 50 Salads, achieving state-of-the-art results. | ['Minsu Cho', 'Seong Jong Ha', 'Manjin Kim', 'Joonseok Lee', 'Dayoung Gong'] | 2022-05-27 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Gong_Future_Transformer_for_Long-Term_Action_Anticipation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Gong_Future_Transformer_for_Long-Term_Action_Anticipation_CVPR_2022_paper.pdf | cvpr-2022-1 | ['action-anticipation'] | ['computer-vision'] | [ 4.03475940e-01 2.48977497e-01 -2.90360123e-01 -8.14012945e-01
-4.53969419e-01 -1.38836771e-01 1.03278148e+00 -2.47324720e-01
-4.14490312e-01 7.61792481e-01 1.02659917e+00 6.46107197e-02
2.43002713e-01 -5.25445700e-01 -1.00449193e+00 -4.41359907e-01
-5.66079617e-01 4.77434546e-01 1.55837759e-01 -6.00107349e-02
1.19205989e-01 1.53985754e-01 -1.38602495e+00 8.13386261e-01
1.40109956e-01 8.54477584e-01 3.23162705e-01 1.19907403e+00
4.18844163e-01 1.93484294e+00 -2.73204774e-01 -6.36708796e-01
3.76018062e-02 -4.29799169e-01 -1.05776882e+00 2.91122466e-01
5.78843206e-02 -9.07480955e-01 -8.58346462e-01 4.94594306e-01
7.03672245e-02 3.77395600e-01 2.74770916e-01 -1.26066709e+00
-6.70588613e-01 8.59601259e-01 -2.20192209e-01 2.40077421e-01
1.03749549e+00 8.13172877e-01 1.03991461e+00 -3.69226068e-01
6.92813873e-01 1.63065481e+00 3.15198511e-01 5.99337995e-01
-7.54634380e-01 -3.67069632e-01 9.25229073e-01 7.50983298e-01
-6.82220817e-01 -5.17022073e-01 6.26077712e-01 -3.69442999e-01
1.68158340e+00 5.34605384e-02 9.51632440e-01 1.66611707e+00
6.76742256e-01 1.45324445e+00 4.61517870e-01 7.94795752e-02
2.16046162e-02 -8.27609241e-01 -1.72314212e-01 2.91266143e-01
-7.28786469e-01 1.11476898e-01 -7.74763107e-01 1.83581874e-01
5.09912312e-01 3.49568248e-01 9.38283876e-02 1.97877258e-01
-1.90591085e+00 4.24159855e-01 2.28846908e-01 -4.41099778e-02
-1.09507716e+00 7.58167505e-01 5.02545178e-01 2.17148900e-01
6.82115257e-01 2.38298669e-01 -6.14189208e-01 -9.37202275e-01
-5.04741132e-01 5.00710726e-01 7.26440787e-01 7.73390651e-01
3.55118304e-01 -2.07551494e-01 -5.88803828e-01 3.89176831e-02
2.85836279e-01 3.35011870e-01 3.67955953e-01 -1.25116003e+00
6.46773875e-01 2.74789751e-01 6.53971195e-01 -5.78950942e-01
-3.28247637e-01 -7.44318962e-02 -4.94607449e-01 -5.20529635e-02
3.52853984e-01 -3.59167308e-01 -8.05049598e-01 1.76233029e+00
2.47503504e-01 9.70417261e-01 3.22333753e-01 7.11546063e-01
4.11836952e-01 8.85256886e-01 5.30501544e-01 -3.22971374e-01
9.24194276e-01 -1.40764916e+00 -6.70692146e-01 -6.97690368e-01
7.13739812e-01 -4.77084756e-01 5.44813514e-01 3.37118208e-01
-1.10131133e+00 -6.87754571e-01 -4.74790335e-01 -1.12762861e-01
1.71472982e-01 2.34822333e-02 1.10652339e+00 -2.70660102e-01
-9.00277436e-01 7.22020507e-01 -1.48619711e+00 -2.37162575e-01
3.01198632e-01 2.66638726e-01 -3.25043261e-01 -7.54519999e-02
-1.16361022e+00 6.71145976e-01 3.56030762e-01 3.32341105e-01
-1.35773265e+00 -5.55573821e-01 -9.83101666e-01 2.39859253e-01
4.50408846e-01 -6.17640316e-01 1.72580826e+00 -1.28860700e+00
-1.72777259e+00 4.70200151e-01 -5.95203042e-01 -1.13597572e+00
7.18815923e-01 -8.95860672e-01 -4.27411973e-01 -1.27055809e-01
-7.42400214e-02 9.51542616e-01 6.87452018e-01 -1.94774479e-01
-7.93847144e-01 -1.48955867e-01 6.82306588e-01 4.47067440e-01
2.67534196e-01 2.68039972e-01 -3.11536372e-01 -5.05233228e-01
-3.50735486e-01 -1.11487758e+00 -4.68651563e-01 -9.92554724e-02
-1.88540012e-01 -6.07350469e-01 6.71880305e-01 -6.69424415e-01
9.79947209e-01 -2.06613660e+00 2.28825837e-01 -6.61566257e-01
-3.51568833e-02 1.81543343e-02 -5.03762841e-01 6.47083521e-01
-2.81368226e-01 -3.80076319e-01 2.32169315e-01 -6.04396760e-01
2.24837705e-01 3.39049637e-01 -7.18418419e-01 3.61609638e-01
3.78016621e-01 1.28987837e+00 -1.39243960e+00 -2.55109668e-01
5.29959440e-01 4.82319266e-01 -5.18512964e-01 6.18592620e-01
-8.41720819e-01 8.55676770e-01 -7.49355257e-01 4.30042833e-01
8.93444344e-02 -4.27584052e-01 2.57184595e-01 3.50534678e-01
4.49232683e-02 5.59116602e-01 -5.70876479e-01 1.82868302e+00
-5.94568551e-01 8.51273000e-01 -5.94717562e-01 -8.52912366e-01
5.69327772e-01 6.57019734e-01 8.43782187e-01 -7.24011242e-01
-2.40427256e-01 -3.82119507e-01 -9.68986303e-02 -8.59555483e-01
5.31154096e-01 8.45022425e-02 -3.00487429e-01 5.12093782e-01
-1.54683411e-01 3.88409764e-01 2.30903670e-01 2.80863255e-01
1.37924945e+00 8.78364027e-01 5.68868756e-01 5.32124698e-01
4.35117573e-01 -2.54077852e-01 8.18333328e-01 7.64859438e-01
-3.48999351e-01 3.27495128e-01 7.63877392e-01 -1.14817500e+00
-7.67725587e-01 -7.61998951e-01 7.31646419e-01 1.49527335e+00
-8.78983140e-02 -4.76624370e-01 -3.34571958e-01 -7.40813136e-01
-2.49395341e-01 9.47341025e-01 -7.65501440e-01 1.15749026e-02
-9.85025465e-01 -2.60051131e-01 -1.15194358e-02 8.65135670e-01
4.38851684e-01 -1.74664867e+00 -1.13563347e+00 5.15497208e-01
-5.26825011e-01 -1.43794405e+00 -4.49547112e-01 -3.99199128e-01
-6.04037166e-01 -9.30657148e-01 -6.42594457e-01 -2.02203304e-01
1.35356888e-01 4.25982140e-02 1.64730322e+00 -3.75487953e-02
3.19860101e-01 5.24284720e-01 -5.68201661e-01 -2.27920115e-01
-3.50753576e-01 -4.86628920e-01 -5.69683909e-02 2.57667065e-01
5.20968318e-01 -6.25407338e-01 -6.08410776e-01 1.12170637e-01
-3.44833463e-01 5.34724116e-01 5.60021877e-01 6.46120667e-01
5.10282099e-01 -1.58469990e-01 3.69922936e-01 -7.03575611e-01
-1.85023283e-03 -6.92264438e-01 -3.95565420e-01 4.53454584e-01
2.36008808e-01 -6.81777596e-02 7.52777100e-01 -6.68687701e-01
-1.35794091e+00 4.01634544e-01 -3.10259432e-01 -2.81450897e-01
-4.44602519e-01 3.46003860e-01 -1.00287132e-01 7.13943899e-01
-1.56311840e-02 2.65431434e-01 -4.18733269e-01 -1.89756647e-01
3.62762362e-01 1.76415354e-01 6.49310827e-01 -1.66683882e-01
9.15549099e-02 4.58459318e-01 -7.26493970e-02 -5.01659751e-01
-1.05091131e+00 -2.76479840e-01 -6.91479564e-01 -5.37845194e-01
1.13858497e+00 -1.26562488e+00 -1.32613313e+00 7.68151522e-01
-1.50363505e+00 -9.95367348e-01 -1.95504025e-01 8.13941061e-01
-1.08623743e+00 2.25392938e-01 -5.32830060e-01 -1.01233816e+00
-3.63440625e-02 -1.22240877e+00 1.30269086e+00 1.01863533e-01
-5.80217659e-01 -9.68772948e-01 5.94454519e-02 3.91902238e-01
-7.26174936e-02 4.01688159e-01 1.84887215e-01 -5.05022585e-01
-1.07169044e+00 -1.80720568e-01 2.44894743e-01 2.41102017e-02
-1.72899663e-01 -4.49026003e-02 -6.06430352e-01 -1.70166403e-01
-2.53345549e-01 -4.45084512e-01 9.81790066e-01 6.23657584e-01
1.07729268e+00 -4.43941325e-01 -4.70899642e-01 4.24594253e-01
8.03048432e-01 4.30259585e-01 9.81248081e-01 1.28272340e-01
6.00108564e-01 5.26308894e-01 1.28194666e+00 8.91661942e-01
6.58746898e-01 7.39564121e-01 6.97926760e-01 3.14061463e-01
3.46560143e-02 -3.90164107e-01 9.13081169e-01 1.35989144e-01
-2.24639237e-01 -7.52927065e-01 -7.41617382e-01 7.61605203e-01
-2.29645300e+00 -1.51128113e+00 1.22834891e-01 1.80599105e+00
2.85422176e-01 2.82174885e-01 1.57597676e-01 -3.91445756e-01
4.33269650e-01 8.40528011e-01 -9.58379328e-01 -3.87266010e-01
2.88282722e-01 -3.04925442e-01 1.47588804e-01 5.53704560e-01
-1.46995556e+00 1.11436629e+00 6.91883850e+00 2.88459867e-01
-1.01446354e+00 -1.49237081e-01 9.65184033e-01 -5.20395219e-01
-1.05251096e-01 1.74759492e-01 -7.04694986e-01 5.15362978e-01
1.17110205e+00 4.06182222e-02 4.99403596e-01 8.55780363e-01
5.41035771e-01 -2.27297708e-01 -1.39416897e+00 6.22690976e-01
-2.32830606e-02 -1.30852461e+00 -2.42841810e-01 -2.29044348e-01
5.99355936e-01 3.43274146e-01 -4.35455255e-02 5.21793485e-01
8.40583086e-01 -9.90392745e-01 8.09277594e-01 1.07145071e+00
3.67799580e-01 -6.04163229e-01 5.15555084e-01 5.47307372e-01
-1.21406806e+00 -5.20905674e-01 5.31661585e-02 -8.83527756e-01
9.28902626e-01 3.62137288e-01 -7.80263484e-01 2.15502560e-01
4.92267668e-01 1.65051091e+00 -2.81117320e-01 4.33364868e-01
-6.04790986e-01 5.12438476e-01 5.07425033e-02 3.33562680e-02
4.77981478e-01 1.47877671e-02 5.10459960e-01 9.72427785e-01
3.41617703e-01 3.66067559e-01 3.27961564e-01 2.75202245e-01
1.13126084e-01 -6.31287873e-01 -5.41624963e-01 -4.32679057e-01
2.55911592e-02 8.03572893e-01 -3.73102218e-01 -6.38860285e-01
-6.57019854e-01 1.17912793e+00 3.22732896e-01 4.52997804e-01
-1.33961344e+00 3.30370814e-01 1.04304278e+00 -2.78810620e-01
5.07953227e-01 -3.04046184e-01 2.53270924e-01 -1.25253928e+00
2.01675460e-01 -7.97181845e-01 5.01676738e-01 -1.17565715e+00
-8.82384181e-01 5.46656609e-01 -9.95952934e-02 -1.33984530e+00
-1.02939844e+00 -2.42729574e-01 -9.17609870e-01 4.40002859e-01
-1.38578773e+00 -1.48532820e+00 -1.21086247e-01 3.23475122e-01
1.15157735e+00 1.14700712e-01 6.69608891e-01 -1.94720790e-01
-3.58423114e-01 -4.39494476e-02 -3.27102035e-01 4.56575342e-02
5.14188886e-01 -1.07783043e+00 1.28004265e+00 1.02662194e+00
3.35817784e-01 1.57358199e-01 8.19180548e-01 -8.20231915e-01
-1.29951501e+00 -1.18772423e+00 1.31087792e+00 -5.56790650e-01
7.22446740e-01 -2.20457166e-01 -6.12507939e-01 1.63706350e+00
3.10161084e-01 2.99346279e-02 2.34054983e-01 1.12262949e-01
3.43997106e-02 1.31319538e-01 -4.50573117e-01 7.50336528e-01
1.31070328e+00 -4.48971957e-01 -5.62782764e-01 6.87850177e-01
1.08954370e+00 -5.83261132e-01 -6.83201492e-01 1.70150504e-01
1.01460278e+00 -1.20740211e+00 1.22027004e+00 -8.90122354e-01
9.16814506e-01 1.22591309e-01 4.86553349e-02 -1.03732765e+00
-3.42557043e-01 -9.97887194e-01 -4.79739785e-01 7.79539585e-01
2.04096600e-01 -3.01739633e-01 8.61152470e-01 8.82569671e-01
-2.18390331e-01 -8.57126355e-01 -7.91629255e-01 -5.10523617e-01
-5.73808730e-01 -6.65407717e-01 8.66260588e-01 4.06217098e-01
-3.33327018e-02 9.32478607e-02 -1.21224451e+00 1.15235806e-01
1.17216378e-01 3.42653066e-01 1.00637221e+00 -6.76895738e-01
-7.04200566e-01 7.91128501e-02 -5.71506202e-01 -1.81318033e+00
7.09102750e-01 -2.29258969e-01 3.37634057e-01 -1.62484753e+00
2.87524313e-01 2.78051674e-01 -2.58656681e-01 4.46504533e-01
-3.10784638e-01 -1.37445569e-01 3.53834808e-01 9.42750368e-03
-1.40020525e+00 6.82890952e-01 1.45436764e+00 -1.22591071e-01
2.65565142e-02 3.60575259e-01 -9.50543061e-02 9.70096469e-01
3.81590307e-01 -4.16225165e-01 -4.91654277e-01 -6.85065627e-01
2.47141778e-01 8.30840588e-01 5.28450608e-01 -1.05776513e+00
1.98474899e-01 -8.38573635e-01 3.28562260e-01 -8.09351385e-01
8.74351382e-01 -5.84178030e-01 2.73892879e-01 4.73421037e-01
-8.05744469e-01 2.15201512e-01 -4.02005613e-01 9.25473928e-01
-3.19134235e-01 4.29135293e-01 1.05519891e-01 -3.03141534e-01
-1.41536689e+00 5.91539502e-01 -5.46041369e-01 -2.09470972e-01
1.31778300e+00 -1.73963550e-02 -2.46359751e-01 -1.02631736e+00
-1.07292855e+00 5.79769850e-01 2.19500110e-01 6.76920891e-01
6.38923526e-01 -1.17751420e+00 -8.51199985e-01 -4.98331487e-02
-5.33866994e-02 -2.86925107e-01 7.02243626e-01 9.60723400e-01
-1.73251137e-01 4.81441677e-01 -3.83674413e-01 -4.29212004e-01
-1.43937802e+00 8.03184628e-01 8.73770863e-02 -7.73728132e-01
-9.75467324e-01 1.07271528e+00 4.85193819e-01 9.56992209e-02
2.86295623e-01 -3.76055658e-01 -4.06055421e-01 -2.62754887e-01
9.39899325e-01 1.28613055e-01 -7.49668181e-01 -7.46304214e-01
-3.67973864e-01 2.57221870e-02 1.92930107e-03 -4.81454059e-02
1.57551730e+00 -2.78851271e-01 4.56028897e-03 5.68543375e-01
9.53943014e-01 -6.23125672e-01 -2.25243783e+00 -1.72701195e-01
-8.68694410e-02 -6.82296753e-01 -3.62454474e-01 -6.96605921e-01
-8.65017772e-01 8.34385216e-01 5.24216406e-02 -8.33591968e-02
1.15930700e+00 2.92128436e-02 1.15348613e+00 5.54757535e-01
4.11808848e-01 -8.34983706e-01 2.89519966e-01 7.47637749e-01
9.77766216e-01 -1.34985352e+00 6.02596626e-02 -3.86361666e-02
-1.18751383e+00 1.05360508e+00 4.90906000e-01 -1.95957392e-01
2.50088960e-01 3.75744589e-02 -1.04727104e-01 1.66776348e-02
-1.65999091e+00 -7.80868903e-02 1.20281428e-01 4.51651633e-01
5.74460745e-01 2.92074144e-01 1.92811504e-01 2.72365212e-01
5.69345206e-02 4.69359308e-01 5.63517630e-01 8.03592980e-01
3.64981927e-02 -8.94280672e-01 1.52159721e-01 3.02100599e-01
-4.00570005e-01 1.02044851e-01 -1.92542121e-01 3.45962435e-01
2.85778171e-03 1.11379111e+00 4.41435069e-01 -3.33720684e-01
-3.22067458e-03 9.68087092e-02 3.79203349e-01 -2.77186722e-01
-4.11772281e-01 -1.66061878e-01 4.73202109e-01 -1.40857053e+00
-8.38285267e-01 -9.72735584e-01 -1.06202531e+00 -3.08171362e-01
3.77428055e-01 -2.54060984e-01 1.33847237e-01 1.26123190e+00
4.68143195e-01 7.33667970e-01 5.41474283e-01 -1.09786570e+00
-2.54809707e-01 -1.03804386e+00 -3.05636283e-02 6.79060400e-01
5.45302987e-01 -5.22751868e-01 7.98570514e-02 3.59645873e-01] | [8.057863235473633, 0.49661901593208313] |
3b90d484-38d6-4380-afd8-8edf2ee55932 | all-neural-beamformer-for-continuous-speech | 2110.06428 | null | https://arxiv.org/abs/2110.06428v1 | https://arxiv.org/pdf/2110.06428v1.pdf | All-neural beamformer for continuous speech separation | Continuous speech separation (CSS) aims to separate overlapping voices from a continuous influx of conversational audio containing an unknown number of utterances spoken by an unknown number of speakers. A common application scenario is transcribing a meeting conversation recorded by a microphone array. Prior studies explored various deep learning models for time-frequency mask estimation, followed by a minimum variance distortionless response (MVDR) filter to improve the automatic speech recognition (ASR) accuracy. The performance of these methods is fundamentally upper-bounded by MVDR's spatial selectivity. Recently, the all deep learning MVDR (ADL-MVDR) model was proposed for neural beamforming and demonstrated superior performance in a target speech extraction task using pre-segmented input. In this paper, we further adapt ADL-MVDR to the CSS task with several enhancements to enable end-to-end neural beamforming. The proposed system achieves significant word error rate reduction over a baseline spectral masking system on the LibriCSS dataset. Moreover, the proposed neural beamformer is shown to be comparable to a state-of-the-art MVDR-based system in real meeting transcription tasks, including AMI, while showing potentials to further simplify the runtime implementation and reduce the system latency with frame-wise processing. | ['Sefik Emre Eskimez', 'Dongmei Wang', 'Xiaofei Wang', 'Zhuo Chen', 'Naoyuki Kanda', 'Takuya Yoshioka', 'Zhuohuang Zhang'] | 2021-10-13 | null | null | null | null | ['speech-extraction'] | ['speech'] | [ 5.20342827e-01 -2.12030951e-02 4.34693784e-01 -2.88162827e-01
-1.46059859e+00 -5.36452055e-01 3.40342253e-01 -3.25311780e-01
-3.53668749e-01 3.10286343e-01 7.54645705e-01 -4.11824256e-01
9.30107385e-02 4.37053777e-02 -3.31770092e-01 -8.72357130e-01
1.95102632e-01 -8.18758607e-02 -8.92545953e-02 -4.24650498e-02
-9.05192569e-02 6.38156235e-01 -1.58445930e+00 7.14801431e-01
5.98520339e-01 1.07387471e+00 5.91693044e-01 1.12601864e+00
1.37328625e-01 5.74785352e-01 -1.10757673e+00 -2.22111493e-02
2.43792102e-01 -4.31822330e-01 -3.28894019e-01 5.28456867e-02
5.83925605e-01 -3.13646436e-01 -7.25675762e-01 8.08224499e-01
1.32518196e+00 3.96333367e-01 5.84496140e-01 -6.52143717e-01
-3.13196957e-01 8.48158598e-01 -3.38942945e-01 5.24534464e-01
5.74778616e-01 -1.14029139e-01 7.66352057e-01 -1.49963331e+00
-2.15599328e-01 1.16478837e+00 5.82388937e-01 6.32482171e-01
-1.24791729e+00 -8.19930971e-01 9.47521999e-02 1.79616466e-01
-1.79654050e+00 -1.37860739e+00 7.95178890e-01 -2.05607563e-01
1.52107620e+00 7.86933959e-01 9.10240933e-02 1.17726600e+00
-1.23153381e-01 8.00591707e-01 6.40775859e-01 -4.95608985e-01
3.15467715e-01 -3.23241465e-02 2.94370409e-02 -5.02577890e-03
-3.05033714e-01 -1.62487403e-02 -1.14695728e+00 -1.04296528e-01
3.76875222e-01 -3.85121852e-01 -8.80034268e-01 3.87435913e-01
-1.00095081e+00 3.56488913e-01 -6.48439601e-02 5.68212926e-01
-4.28075343e-01 -8.29701051e-02 2.81287760e-01 3.55578482e-01
5.86683989e-01 1.46224916e-01 -3.51620972e-01 -1.82522535e-01
-1.42261887e+00 7.00159147e-02 8.27345014e-01 7.27258205e-01
-1.25048161e-01 8.09159160e-01 -4.72371161e-01 1.39404905e+00
3.24910015e-01 6.63583815e-01 6.41365886e-01 -8.46928298e-01
7.38905847e-01 -3.02638680e-01 8.66503641e-02 -7.18107760e-01
-4.24381882e-01 -8.59264672e-01 -1.06993818e+00 2.16376316e-02
1.62003174e-01 -3.62138271e-01 -8.33589435e-01 1.69196987e+00
6.80589974e-02 2.91299462e-01 2.81918585e-01 1.12883449e+00
9.83550429e-01 1.02213514e+00 -7.68692613e-01 -6.78036928e-01
1.16848910e+00 -8.52569997e-01 -1.30455399e+00 -4.28622961e-01
3.51285450e-02 -9.81575072e-01 9.38680112e-01 1.05447686e+00
-1.34060669e+00 -8.35950136e-01 -1.14491129e+00 1.23265237e-02
7.89325759e-02 2.95589685e-01 -2.42663212e-02 1.19468045e+00
-1.31445348e+00 -7.11961910e-02 -7.27868676e-01 1.89078376e-01
-4.93418239e-02 6.00370347e-01 -2.05568418e-01 6.60967305e-02
-1.11176491e+00 5.43610334e-01 -1.55423567e-01 4.73951995e-01
-8.98371696e-01 -7.20115423e-01 -8.48782718e-01 3.63004476e-01
1.93417501e-02 -3.87546957e-01 1.61031342e+00 -8.25467825e-01
-1.91830897e+00 4.47681278e-01 -8.23030770e-01 -6.90594554e-01
1.82571113e-01 -5.09543955e-01 -8.98768723e-01 2.00023130e-02
-5.39059222e-01 3.28432083e-01 1.14759684e+00 -1.09060812e+00
-4.38831836e-01 -2.45157033e-01 -5.19300759e-01 5.36483467e-01
-5.23425698e-01 1.85301572e-01 -1.72619462e-01 -1.06627846e+00
2.98037857e-01 -4.49277312e-01 5.30019663e-02 -6.73234046e-01
-2.73451924e-01 6.18838370e-02 6.10893309e-01 -1.22276354e+00
1.67547345e+00 -2.46454620e+00 1.62120044e-01 -2.92482674e-02
-1.33554682e-01 4.67487276e-01 -1.84222996e-01 4.25145894e-01
-2.73723453e-01 -3.54999930e-01 -2.83645332e-01 -7.04602122e-01
4.33994941e-02 -3.78254086e-01 -6.46788418e-01 6.85108662e-01
-2.29714494e-02 2.82142848e-01 -2.50933766e-01 6.98453113e-02
2.24647015e-01 9.00418639e-01 -5.35244167e-01 5.16064405e-01
3.68876249e-01 3.12042594e-01 6.13143206e-01 4.64213818e-01
8.07340562e-01 5.81359446e-01 -1.83616914e-02 -1.78280219e-01
-5.69847152e-02 7.71546841e-01 -1.36713767e+00 1.79195368e+00
-6.25946581e-01 1.05735683e+00 8.85630548e-01 -7.98911631e-01
1.03159583e+00 9.04180527e-01 2.09387705e-01 -4.85714525e-01
5.89178801e-02 2.78527498e-01 3.38217437e-01 -2.95725971e-01
3.50689828e-01 -1.28436476e-01 2.21332774e-01 1.02232076e-01
1.98735788e-01 -1.13667794e-01 -3.51732343e-01 -2.14363515e-01
1.16115832e+00 -6.27667308e-01 1.27730876e-01 -1.35306865e-01
7.69189239e-01 -7.42877126e-01 3.27405155e-01 6.89678133e-01
-1.10170126e-01 1.11246860e+00 -2.41375431e-01 3.14418584e-01
-4.64367509e-01 -1.32971680e+00 -1.09606676e-01 1.15901995e+00
-4.41639394e-01 -8.67405310e-02 -1.02669406e+00 4.51281527e-03
-3.75518799e-01 1.13669860e+00 6.54079542e-02 7.48735070e-02
-8.14201295e-01 -4.86361742e-01 8.57548118e-01 4.50585425e-01
1.13995701e-01 -6.60666943e-01 -1.26031041e-01 3.25595379e-01
-4.47428912e-01 -1.25645316e+00 -9.91460085e-01 4.09000427e-01
-3.54379952e-01 -2.07375094e-01 -9.76274252e-01 -1.00941563e+00
2.99068749e-01 5.50459981e-01 6.49742424e-01 -6.67178750e-01
1.75832547e-02 3.72668326e-01 -1.86602652e-01 -4.04640228e-01
-3.66536111e-01 -4.64130268e-02 6.00288987e-01 3.61919761e-01
4.23719674e-01 -6.87493682e-01 -6.43061042e-01 4.18926716e-01
-7.67369390e-01 -1.52975515e-01 4.14316863e-01 8.08412552e-01
2.76448041e-01 1.94087952e-01 9.47944820e-01 -1.13977209e-01
1.07229435e+00 -1.10898726e-01 -3.34720969e-01 -1.10625312e-01
-2.79585775e-02 -4.80548799e-01 5.28866231e-01 -7.21660256e-01
-1.32127547e+00 1.96007252e-01 -7.88820148e-01 -4.45632488e-01
-1.58743009e-01 4.00726169e-01 -7.13093400e-01 3.63236874e-01
6.00820124e-01 4.86937672e-01 -1.56433985e-01 -5.32969296e-01
2.75359839e-01 1.59127474e+00 9.10790324e-01 1.20010465e-01
4.18765247e-01 6.45147264e-02 -5.62379062e-01 -1.36384130e+00
-3.47465545e-01 -8.18118989e-01 -4.74201947e-01 -5.91162182e-02
7.24049866e-01 -1.32276869e+00 -6.55282199e-01 5.45545340e-01
-1.28242612e+00 -3.14852178e-01 -2.82006059e-03 8.50195706e-01
-3.58779192e-01 2.42256835e-01 -6.11347020e-01 -1.44450974e+00
-5.67216814e-01 -1.11139703e+00 1.03198051e+00 -1.34931237e-01
-5.38140416e-01 -5.15537441e-01 -4.14796732e-02 4.84625369e-01
5.98590255e-01 -5.00990391e-01 3.64288092e-01 -9.14834857e-01
-2.85253860e-03 -1.76628888e-01 3.58595550e-01 7.70069242e-01
3.42373163e-01 -6.01441503e-01 -1.78868997e+00 -5.22576690e-01
7.54364133e-01 2.45945394e-01 8.61044586e-01 8.93293798e-01
1.00465941e+00 -4.66861665e-01 -2.05223486e-02 4.98772144e-01
7.29550123e-01 5.92399299e-01 4.64753538e-01 -4.31834906e-01
4.45325196e-01 4.62011874e-01 2.14411795e-01 3.29201639e-01
-1.02486327e-01 7.26587057e-01 9.03917924e-02 -2.74663240e-01
-5.92178106e-01 1.92070335e-01 7.79987514e-01 1.60577977e+00
4.80754435e-01 -7.12474465e-01 -8.66720378e-01 6.44877672e-01
-1.43956351e+00 -9.67369020e-01 6.25080913e-02 2.21125150e+00
7.96703219e-01 1.18790343e-01 2.44821459e-02 7.25761116e-01
7.50075281e-01 2.67276406e-01 -5.23370147e-01 -7.19852865e-01
-3.25638473e-01 4.11782146e-01 2.55112618e-01 9.17356789e-01
-8.50019574e-01 4.07059759e-01 6.13309908e+00 1.09097910e+00
-1.05950809e+00 3.65594327e-01 5.19515097e-01 -5.84172189e-01
-3.80461738e-02 -9.19203222e-01 -7.47603774e-01 6.38002008e-02
1.46938074e+00 7.45870098e-02 8.43371749e-01 4.02755708e-01
6.97188437e-01 1.07111320e-01 -1.32806230e+00 1.44728076e+00
4.75786179e-01 -9.43367422e-01 -3.15926492e-01 -5.19832224e-02
4.23637956e-01 -1.90358013e-01 3.60536873e-01 1.81476116e-01
-2.77907848e-01 -1.30834496e+00 7.99963355e-01 3.02080363e-01
1.04768014e+00 -9.44596350e-01 6.70275807e-01 4.58957583e-01
-1.23703384e+00 -3.71243417e-01 -1.36512965e-01 -1.78826049e-01
1.80976629e-01 6.22300327e-01 -1.31352186e+00 3.23715121e-01
5.67203701e-01 -4.04169448e-02 4.78664488e-02 8.48056793e-01
1.12630419e-01 9.88333702e-01 -2.07491890e-01 2.42530525e-01
-1.69769436e-01 1.93980381e-01 8.85723948e-01 1.84351778e+00
4.91244823e-01 3.36832732e-01 -2.27514550e-01 5.42797446e-01
-2.18657836e-01 -1.75832585e-01 -3.34993660e-01 7.11469948e-02
7.61740148e-01 8.39212835e-01 -2.60643959e-01 1.05780698e-02
-4.55315441e-01 1.08395851e+00 -2.31868744e-01 7.35451996e-01
-7.21286237e-01 -7.18628585e-01 1.02786267e+00 7.86686912e-02
5.42470753e-01 -4.16696787e-01 -5.54233909e-01 -7.43970215e-01
1.08585589e-01 -1.17655623e+00 -2.28594914e-01 -7.33288884e-01
-1.02555692e+00 9.29476559e-01 -3.32687497e-01 -1.13853908e+00
-4.10869390e-01 -5.30285716e-01 -2.86765307e-01 1.44739425e+00
-1.09264767e+00 -6.66914165e-01 1.00849152e-01 5.91282845e-01
1.26840603e+00 -5.47498286e-01 9.46996152e-01 6.35108352e-01
-4.62473571e-01 9.48095441e-01 2.12655932e-01 5.58072664e-02
6.46975458e-01 -1.11723316e+00 6.67369246e-01 1.22952521e+00
3.40399414e-01 5.88358700e-01 9.67791438e-01 -2.41576821e-01
-1.49549186e+00 -9.54154313e-01 8.13283622e-01 -1.33462757e-01
2.29291081e-01 -1.03348422e+00 -9.98943508e-01 2.23089367e-01
5.97494662e-01 -3.70177686e-01 9.10928726e-01 -2.56053269e-01
-4.86403629e-02 -4.65801179e-01 -8.90065849e-01 5.98337829e-01
9.73960221e-01 -8.66587579e-01 -8.45630229e-01 -5.76677062e-02
1.03373742e+00 -6.11447394e-01 -3.45349669e-01 3.56856138e-01
4.37249392e-01 -8.60065520e-01 1.03995430e+00 9.70734432e-02
-1.28899917e-01 -4.59328145e-01 -6.66877270e-01 -1.43187940e+00
-2.40188196e-01 -1.13634479e+00 -3.70990962e-01 1.52449620e+00
5.61041713e-01 -4.26645577e-01 3.64215612e-01 3.42488885e-01
-5.88857353e-01 -3.83424968e-01 -1.28590703e+00 -5.92652261e-01
-2.53590494e-01 -9.24060404e-01 2.83209085e-01 4.42590058e-01
2.52288412e-02 4.84270275e-01 -4.81079280e-01 6.35244608e-01
2.47283667e-01 -5.36991477e-01 5.18555641e-01 -6.69752657e-01
-5.19990742e-01 -4.64058459e-01 -8.94725546e-02 -1.59229076e+00
1.19738385e-01 -7.02164233e-01 4.27978992e-01 -1.51803088e+00
-5.32145917e-01 4.02681500e-01 -4.08248782e-01 -1.30473018e-01
-4.34394740e-02 -3.96403074e-02 4.26930226e-02 -2.52980441e-01
6.99780742e-03 5.73619366e-01 5.33208132e-01 -2.68872201e-01
-5.31322122e-01 5.17260015e-01 -5.57866216e-01 5.85366428e-01
5.55042982e-01 -3.19349915e-01 -6.58776164e-01 -4.43768442e-01
-3.13893229e-01 5.10460258e-01 -1.28711447e-01 -1.17562747e+00
5.93769908e-01 4.68020916e-01 3.54153574e-01 -8.82427812e-01
1.03231883e+00 -8.57662618e-01 2.38911137e-01 1.82464540e-01
-6.89554989e-01 -1.36815801e-01 4.82589215e-01 4.45579797e-01
-4.18144464e-01 6.09456003e-03 6.53832614e-01 2.96025008e-01
-2.11937781e-02 -3.77228677e-01 -1.04270041e+00 -1.44543931e-01
4.14679915e-01 -3.04362565e-01 4.78380127e-03 -7.27577746e-01
-7.71725833e-01 -3.71015012e-01 -4.39848036e-01 2.86934912e-01
1.02911115e+00 -1.10410428e+00 -9.90214348e-01 3.39409769e-01
-5.88441133e-01 1.06353378e-02 3.32787603e-01 7.77558267e-01
-8.97982195e-02 7.10255027e-01 3.62022936e-01 -6.99293673e-01
-1.66773713e+00 4.07638818e-01 5.95409751e-01 2.20164418e-01
-4.41552341e-01 1.44762409e+00 4.99177694e-01 -1.83275953e-01
1.10516334e+00 -5.49450636e-01 -5.99229895e-02 -2.22297069e-02
9.50046480e-01 4.93483901e-01 7.27247715e-01 -6.63011134e-01
-5.13431013e-01 1.73551098e-01 1.28695861e-01 -8.25988352e-01
1.12698960e+00 -4.71386731e-01 3.13661307e-01 7.38995314e-01
1.12693799e+00 6.42211795e-01 -1.04929531e+00 -2.39392504e-01
-2.12692976e-01 -3.90264094e-01 4.62657779e-01 -9.92973149e-01
-8.10200870e-01 1.14883327e+00 9.64323997e-01 2.58397192e-01
1.68579686e+00 -4.19527650e-01 7.56519854e-01 2.19736442e-01
-2.81120203e-02 -8.41216445e-01 3.40796225e-02 3.83387178e-01
1.33149087e+00 -7.96003103e-01 -3.60686362e-01 -2.28189096e-01
-6.96483493e-01 1.07316995e+00 2.60699391e-01 2.58746594e-01
7.18785167e-01 8.48817229e-01 3.61724287e-01 4.16336894e-01
-7.29394078e-01 -9.40468814e-03 6.43403232e-01 7.49788761e-01
6.48156106e-01 -5.65403961e-02 1.39935717e-01 1.08500743e+00
-5.47001839e-01 -5.41304111e-01 3.09890419e-01 4.24595863e-01
-5.38967669e-01 -5.98595977e-01 -8.98546875e-01 1.16026841e-01
-6.99025929e-01 -4.76679415e-01 -5.22686243e-01 -2.81815324e-03
-2.49134839e-01 1.77436781e+00 2.20656008e-01 -5.19614875e-01
6.14542186e-01 2.90321141e-01 1.96861535e-01 -5.11343300e-01
-9.37634706e-01 8.67744684e-01 1.75056368e-01 -1.16684839e-01
-1.65331066e-01 -6.63814247e-01 -1.05446768e+00 1.01670541e-01
-4.83558863e-01 3.15590240e-02 7.89440989e-01 7.53773034e-01
3.65106821e-01 1.13077641e+00 6.69542611e-01 -1.04479885e+00
-5.42844653e-01 -1.31512332e+00 -4.89110321e-01 -2.28746802e-01
8.96178484e-01 -7.41561130e-02 -4.72080976e-01 1.52173311e-01] | [14.816976547241211, 5.970615863800049] |
15b74351-44a6-4add-8780-fc35ef599460 | modulation-spectral-features-for-speech | 2301.05868 | null | https://arxiv.org/abs/2301.05868v1 | https://arxiv.org/pdf/2301.05868v1.pdf | Modulation spectral features for speech emotion recognition using deep neural networks | This work explores the use of constant-Q transform based modulation spectral features (CQT-MSF) for speech emotion recognition (SER). The human perception and analysis of sound comprise of two important cognitive parts: early auditory analysis and cortex-based processing. The early auditory analysis considers spectrogram-based representation whereas cortex-based analysis includes extraction of temporal modulations from the spectrogram. This temporal modulation representation of spectrogram is called modulation spectral feature (MSF). As the constant-Q transform (CQT) provides higher resolution at emotion salient low-frequency regions of speech, we find that CQT-based spectrogram, together with its temporal modulations, provides a representation enriched with emotion-specific information. We argue that CQT-MSF when used with a 2-dimensional convolutional network can provide a time-shift invariant and deformation insensitive representation for SER. Our results show that CQT-MSF outperforms standard mel-scale based spectrogram and its modulation features on two popular SER databases, Berlin EmoDB and RAVDESS. We also show that our proposed feature outperforms the shift and deformation invariant scattering transform coefficients, hence, showing the importance of joint hand-crafted and self-learned feature extraction instead of reliance on complete hand-crafted features. Finally, we perform Grad-CAM analysis to visually inspect the contribution of constant-Q modulation features over SER. | ['Goutam Saha', 'Md Sahidullah', 'Premjeet Singh'] | 2023-01-14 | null | null | null | null | ['speech-emotion-recognition'] | ['speech'] | [ 9.95958820e-02 -3.76974523e-01 4.99866664e-01 -3.27358872e-01
-1.02118433e+00 -4.39932913e-01 5.29562593e-01 4.84965503e-01
-6.53101563e-01 3.33686382e-01 4.17621464e-01 -5.58296684e-03
-2.74194032e-01 -4.81353968e-01 -2.39377081e-01 -7.44922638e-01
-4.94036198e-01 -5.72171688e-01 3.94403189e-01 -7.45530307e-01
3.63292485e-01 5.12989283e-01 -1.94371283e+00 6.29211545e-01
3.87551725e-01 1.66731203e+00 2.69990891e-01 9.41282690e-01
2.40702946e-02 2.74513096e-01 -9.41772699e-01 6.08373433e-02
3.33936550e-02 -4.43224907e-01 -6.83694124e-01 -3.96253645e-01
7.60094300e-02 2.37163156e-01 -9.84770432e-02 7.36095071e-01
7.34910965e-01 4.21559960e-01 6.72103047e-01 -1.05825114e+00
-2.00776637e-01 2.36657113e-01 -1.69488564e-01 9.63001072e-01
5.24910331e-01 -1.23698935e-01 1.04712880e+00 -1.13500810e+00
3.91411662e-01 1.14812136e+00 7.52372086e-01 1.18341036e-01
-8.81424546e-01 -5.54713070e-01 -2.23796129e-01 5.15356302e-01
-1.28976345e+00 -5.76652408e-01 1.17050827e+00 -2.86025852e-01
1.41507137e+00 4.99797374e-01 8.24397862e-01 7.93807626e-01
4.23236042e-01 3.61520737e-01 1.32359052e+00 -6.51261985e-01
2.64392585e-01 -9.78814662e-02 1.34998346e-02 4.44120705e-01
-6.60796225e-01 4.35553551e-01 -1.07430053e+00 -2.74551451e-01
4.80378926e-01 -5.88557601e-01 -4.45846111e-01 2.97620326e-01
-9.98250663e-01 6.12107217e-01 3.24141622e-01 7.60846317e-01
-4.44736958e-01 1.95268348e-01 7.30688930e-01 6.71077073e-01
6.80018961e-01 4.50567335e-01 -6.71793401e-01 -7.19554484e-01
-9.73051906e-01 2.26668436e-02 3.61793607e-01 1.44217581e-01
6.02265477e-01 5.82440674e-01 -9.48568583e-02 1.03803456e+00
1.58712298e-01 4.35937494e-01 7.15951085e-01 -8.01321924e-01
-8.47659707e-02 -2.33961307e-02 -2.64292806e-01 -1.03038168e+00
-8.32840681e-01 -4.30184543e-01 -4.88703430e-01 9.75320041e-02
2.98059314e-01 5.03330342e-02 -5.69193721e-01 1.85686302e+00
3.13331813e-01 -1.14593588e-01 -1.36506766e-01 1.06808996e+00
8.67312014e-01 5.75994134e-01 4.57878374e-02 -3.46617997e-01
1.77266908e+00 -1.52041137e-01 -8.97213638e-01 1.85158938e-01
2.67684400e-01 -8.34911168e-01 1.01499736e+00 4.94646698e-01
-9.76944327e-01 -8.01768899e-01 -1.13795090e+00 1.19945891e-01
-6.48490906e-01 -3.76434661e-02 5.18261671e-01 8.22488308e-01
-1.08662510e+00 7.18950331e-01 -6.06254399e-01 -1.37100518e-01
1.06174931e-01 2.71325707e-01 -5.54948449e-01 8.30804527e-01
-1.55616999e+00 8.61346006e-01 2.85372827e-02 -9.35340151e-02
-3.86162788e-01 -8.74664307e-01 -8.18676531e-01 1.71740025e-01
-1.82395920e-01 -1.76084727e-01 1.22704244e+00 -9.32253122e-01
-1.73341537e+00 6.15103960e-01 -2.62971580e-01 -5.58913291e-01
-1.99758068e-01 1.32199645e-01 -1.00129139e+00 8.77780497e-01
-3.38715732e-01 2.94684500e-01 1.50951934e+00 -7.08932400e-01
-3.00784439e-01 -2.05811918e-01 -2.35259235e-01 -4.39338088e-02
-6.04571879e-01 3.76788408e-01 3.95034373e-01 -9.62110102e-01
2.80604780e-01 -4.50497657e-01 4.24121648e-01 -1.19117491e-01
1.20223701e-01 -8.80515501e-02 8.92372370e-01 -8.54472697e-01
1.35790229e+00 -2.52244520e+00 -2.29295149e-01 1.48391992e-01
2.49520555e-01 2.95292169e-01 -2.09301144e-01 6.74903631e-01
-5.15926838e-01 -7.56337196e-02 -8.89429301e-02 -6.24253042e-02
1.05281897e-01 -3.56593966e-01 -4.11228657e-01 4.99498665e-01
5.87384105e-01 7.14462399e-01 -6.78665996e-01 -1.35530800e-01
2.43298367e-01 9.56592560e-01 -5.39952040e-01 -2.29241580e-01
2.45805234e-01 4.47945923e-01 -1.17253445e-01 4.55636770e-01
5.50109446e-01 3.84243637e-01 -3.14921767e-01 -7.46453941e-01
-2.66788810e-01 5.32030404e-01 -7.68551826e-01 1.57349312e+00
-6.41368091e-01 1.07150197e+00 2.16893330e-01 -1.00885773e+00
1.27409232e+00 6.93540335e-01 5.33808410e-01 -1.16122937e+00
3.20351481e-01 3.32807720e-01 2.42018029e-01 -4.77147251e-01
3.94762337e-01 -4.19360042e-01 4.51028682e-02 3.06287140e-01
5.55231750e-01 -2.54576474e-01 -1.37717173e-01 -6.66329870e-03
1.15983415e+00 -1.58008024e-01 2.95039505e-01 -4.38093573e-01
7.71217883e-01 -7.48468399e-01 1.42582446e-01 1.73104912e-01
-6.33200407e-01 5.85297942e-01 2.93020397e-01 -2.08412543e-01
-7.07762003e-01 -1.09124815e+00 -5.20166278e-01 1.34425712e+00
-3.40536654e-01 -8.40917647e-01 -6.54473066e-01 -2.88725048e-01
-7.94565976e-02 3.98222804e-01 -7.39308953e-01 -5.74782550e-01
-4.53076482e-01 -4.96701092e-01 6.80929720e-01 3.36556047e-01
2.69850612e-01 -1.14436078e+00 -1.12365985e+00 3.80696744e-01
-2.33475596e-01 -1.05376554e+00 -4.08990443e-01 6.78086817e-01
-5.44760525e-01 -5.99863887e-01 -5.27275681e-01 -2.94129074e-01
-2.15555161e-01 2.84198020e-02 8.03646207e-01 -3.21562201e-01
-6.29081368e-01 8.85954857e-01 -7.15722740e-01 -5.58126509e-01
-2.41605103e-01 -4.33805436e-01 5.47326766e-02 2.83132374e-01
2.57527709e-01 -1.16644764e+00 -6.12874389e-01 4.51185070e-02
-8.33435893e-01 -4.49208707e-01 3.02453190e-01 7.06560731e-01
3.98739427e-01 2.11712904e-02 1.02436292e+00 -3.54894772e-02
9.34423566e-01 -9.14012417e-02 4.85242829e-02 -3.22709411e-01
5.63394688e-02 6.81208912e-03 6.26507342e-01 -6.77817762e-01
-8.82524431e-01 -3.19712967e-01 -3.63331288e-01 -4.58485633e-01
-2.18894742e-02 5.02846539e-01 2.15049580e-01 -3.60373348e-01
9.47662055e-01 2.78741270e-01 1.29799545e-03 -3.69602084e-01
3.84047329e-02 8.25070679e-01 5.84611595e-01 -4.62195516e-01
4.34677094e-01 5.05827427e-01 4.95708212e-02 -1.33422649e+00
-4.50298995e-01 -5.50391614e-01 -4.20047671e-01 -4.93931651e-01
9.25313592e-01 -6.67211115e-01 -8.55812848e-01 3.20450485e-01
-1.06182122e+00 1.13160484e-01 -4.06569719e-01 6.61823809e-01
-7.53590822e-01 2.46071190e-01 -7.22605824e-01 -1.43441451e+00
-5.31052649e-01 -7.60506034e-01 1.40028536e+00 4.38491143e-02
-4.44702119e-01 -7.05484331e-01 5.62261119e-02 1.00088745e-01
8.06933701e-01 3.72784048e-01 8.40129614e-01 -5.84643662e-01
1.47711560e-01 -1.83623597e-01 2.40399987e-01 3.18967044e-01
8.83759260e-02 -8.23184103e-02 -1.65411210e+00 -1.66173384e-01
3.58416498e-01 -2.70124167e-01 9.72362936e-01 5.26513100e-01
1.06151819e+00 1.68360054e-01 3.34202886e-01 4.93137002e-01
9.23622370e-01 3.77505660e-01 3.93060744e-01 1.33931726e-01
3.40919197e-02 5.53737879e-01 5.98507822e-01 7.77108073e-01
1.02651075e-01 7.93355584e-01 2.63726860e-01 2.58716475e-02
-2.46778667e-01 1.07884906e-01 5.60071349e-01 1.25014496e+00
-2.23180145e-01 3.06607425e-01 -5.69837809e-01 4.51446742e-01
-1.31081545e+00 -9.66273129e-01 6.51579257e-03 2.00760198e+00
7.80059397e-01 1.66887090e-01 2.89967716e-01 9.59157228e-01
3.83435518e-01 1.41746655e-01 -1.85484543e-01 -8.83382201e-01
-3.69419962e-01 1.07331502e+00 -2.62506813e-01 3.38616937e-01
-1.03606832e+00 5.07327020e-01 5.66491079e+00 1.08663952e+00
-1.50259006e+00 2.87749887e-01 2.09391341e-01 -9.55475047e-02
-1.24183372e-01 -4.08517480e-01 -1.50182648e-02 2.01809809e-01
1.45822990e+00 -1.48046762e-01 4.72126395e-01 4.66585249e-01
3.55675697e-01 -1.79928735e-01 -8.16500604e-01 1.20541060e+00
-2.55660623e-01 -1.04528677e+00 -4.38049197e-01 -1.56029671e-01
-9.61876288e-02 -9.91124511e-02 3.34937751e-01 1.88279137e-01
-7.35865414e-01 -9.91033852e-01 1.09674776e+00 5.08627534e-01
9.15450573e-01 -9.21820223e-01 4.08600688e-01 -1.25223458e-01
-1.75222373e+00 -1.46023288e-01 -1.97596222e-01 -3.70754153e-02
-2.23525017e-02 6.62962258e-01 -8.19225430e-01 5.90385735e-01
8.58328342e-01 5.15217781e-01 -2.87852108e-01 7.49682605e-01
1.91848382e-01 6.98549330e-01 -3.18742633e-01 1.81882698e-02
-7.97412992e-02 2.04733267e-01 8.08509827e-01 1.54992068e+00
3.53604168e-01 7.14176297e-02 -3.29241365e-01 7.19281197e-01
5.31096041e-01 1.60880506e-01 -1.38998926e-01 -4.61212963e-01
4.65426058e-01 1.56006372e+00 -7.36059725e-01 6.69158902e-03
-1.49101108e-01 7.37899065e-01 -7.92481005e-02 4.26875800e-01
-6.15897715e-01 -7.40885139e-01 9.38369155e-01 -2.24551465e-03
6.43832743e-01 -4.77450311e-01 1.27560318e-01 -7.92178571e-01
-7.46956170e-02 -7.36533821e-01 2.39952922e-01 -9.55323577e-01
-1.14657843e+00 1.08651793e+00 -2.49398038e-01 -1.22937381e+00
-3.16657156e-01 -6.38137221e-01 -5.69873512e-01 1.11575603e+00
-1.49139118e+00 -8.61797690e-01 4.28291969e-02 8.94367814e-01
3.68816853e-01 -4.07548361e-02 1.26525593e+00 2.10428998e-01
-3.65618393e-02 5.88604569e-01 -3.96114767e-01 -1.02880605e-01
5.00182807e-01 -1.19283760e+00 1.82547659e-01 4.57332015e-01
3.27427864e-01 5.31519353e-01 7.96523988e-01 -1.61536530e-01
-1.26058006e+00 -5.93291879e-01 8.44024360e-01 -1.83875114e-01
6.44339740e-01 -5.82938015e-01 -9.15091455e-01 -2.31640339e-01
9.83411819e-02 2.92681336e-01 8.15630615e-01 -2.22046599e-01
-5.88006735e-01 -2.81220019e-01 -1.26134992e+00 8.73529986e-02
6.49099827e-01 -1.38032520e+00 -6.77417636e-01 -1.75674297e-02
7.61501014e-01 -6.08112551e-02 -1.09434509e+00 3.21453571e-01
6.43762350e-01 -1.28046286e+00 1.02840829e+00 -3.25344592e-01
2.57674269e-02 -1.77985743e-01 -5.04108429e-01 -1.50348008e+00
-2.56904006e-01 -8.47141147e-01 -8.07833076e-02 8.50659013e-01
8.67569745e-02 -7.48879790e-01 6.69862702e-02 -1.38350546e-01
-3.24366271e-01 -6.83950186e-01 -1.45248997e+00 -7.79785991e-01
-7.61053115e-02 -1.09480417e+00 3.87211770e-01 7.49068081e-01
4.73426044e-01 2.06722617e-01 1.53045714e-01 1.91659294e-02
7.80019984e-02 4.31596953e-03 4.29728854e-04 -1.25041878e+00
-2.45233834e-01 -4.70076382e-01 -7.14885414e-01 -2.21677065e-01
1.61950774e-02 -7.08233237e-01 -1.32808089e-01 -8.00765157e-01
-4.73945647e-01 1.17647976e-01 -6.61177874e-01 2.24682540e-01
2.09170640e-01 3.39907199e-01 2.85765499e-01 -3.56581837e-01
-4.07359600e-02 7.74318874e-01 1.23982024e+00 8.12175497e-02
-1.18955009e-01 -2.36374274e-01 -1.74413696e-01 6.26261711e-01
6.19911253e-01 -2.42676213e-01 -3.08754325e-01 4.34034795e-01
1.87342286e-01 1.53332904e-01 4.23516810e-01 -1.17491853e+00
3.65568921e-02 1.93201736e-01 4.56532717e-01 -2.73537278e-01
9.52484369e-01 -6.04758859e-01 -2.13746995e-01 3.52109224e-01
-9.29326117e-02 9.19937715e-02 6.94573581e-01 3.38527441e-01
-5.88862062e-01 1.44186333e-01 8.23861957e-01 3.17101106e-02
-6.79893911e-01 -1.15901701e-01 -8.21679711e-01 -6.68002060e-03
3.30697179e-01 -4.09642965e-01 -1.49123386e-01 -6.06139302e-01
-7.98384368e-01 -8.20825040e-01 -2.99019635e-01 3.05850416e-01
7.55372107e-01 -1.22101092e+00 -5.54753363e-01 5.60288370e-01
-3.68414558e-02 -8.23057413e-01 5.91942430e-01 1.15569377e+00
3.36916782e-02 4.50166076e-01 -4.17786449e-01 -5.96989095e-01
-1.28607273e+00 1.73861817e-01 5.03679514e-01 1.17452696e-01
-5.08139789e-01 1.14758146e+00 1.63701430e-01 -6.10872060e-02
-2.57764682e-02 -7.06442773e-01 -4.06319976e-01 3.98295969e-01
6.43666983e-01 1.59315974e-01 6.75371051e-01 -8.98397684e-01
-6.86298788e-01 4.31649685e-01 3.88869584e-01 -6.95796907e-01
1.50426483e+00 -1.20115675e-01 4.30837041e-03 8.68596673e-01
1.39988708e+00 2.72629470e-01 -8.85416090e-01 -8.03491697e-02
-3.64794023e-02 -2.43885964e-01 3.91644686e-01 -8.89597297e-01
-9.34801280e-01 1.20138752e+00 7.52637208e-01 6.51421547e-01
1.72140229e+00 -2.41098791e-01 6.78754628e-01 5.62342666e-02
3.68248641e-01 -1.24949729e+00 5.02378881e-01 6.65798903e-01
1.25832808e+00 -5.65845072e-01 -2.98541903e-01 -3.94876331e-01
-4.78124112e-01 1.57865965e+00 1.40605584e-01 -1.51289953e-02
9.98823941e-01 4.34621215e-01 1.20386831e-01 -2.16554925e-01
-8.67592156e-01 -3.58011812e-01 7.86694348e-01 7.05584228e-01
6.92799807e-01 -1.92670748e-02 -2.08468869e-01 1.01964641e+00
-1.00148761e+00 -6.15526378e-01 7.03434348e-02 9.00182009e-01
-5.82569659e-01 -8.82493794e-01 -5.71827471e-01 3.57309878e-01
-4.62662220e-01 -2.05872700e-01 -3.19378287e-01 4.79926348e-01
1.38621643e-01 1.02370024e+00 2.14519456e-01 -7.68744648e-01
2.81572163e-01 6.30119920e-01 8.05236816e-01 -2.40634710e-01
-1.11686468e+00 4.82751429e-01 1.81675255e-01 -7.11595595e-01
-4.30035353e-01 -4.82764572e-01 -1.40959728e+00 1.52449295e-01
-3.41583908e-01 5.17385192e-02 9.30538297e-01 1.02166581e+00
3.77033621e-01 9.16121006e-01 6.70393765e-01 -9.28668857e-01
-2.97893703e-01 -1.21629691e+00 -8.36759627e-01 3.46966475e-01
8.83982778e-01 -7.33453631e-01 -5.28727531e-01 -4.26088832e-02] | [15.248759269714355, 5.483029842376709] |
fb88b102-acc9-4fa6-a431-e977cc26aa1b | uc-net-uncertainty-inspired-rgb-d-saliency | 2004.05763 | null | https://arxiv.org/abs/2004.05763v1 | https://arxiv.org/pdf/2004.05763v1.pdf | UC-Net: Uncertainty Inspired RGB-D Saliency Detection via Conditional Variational Autoencoders | In this paper, we propose the first framework (UCNet) to employ uncertainty for RGB-D saliency detection by learning from the data labeling process. Existing RGB-D saliency detection methods treat the saliency detection task as a point estimation problem, and produce a single saliency map following a deterministic learning pipeline. Inspired by the saliency data labeling process, we propose probabilistic RGB-D saliency detection network via conditional variational autoencoders to model human annotation uncertainty and generate multiple saliency maps for each input image by sampling in the latent space. With the proposed saliency consensus process, we are able to generate an accurate saliency map based on these multiple predictions. Quantitative and qualitative evaluations on six challenging benchmark datasets against 18 competing algorithms demonstrate the effectiveness of our approach in learning the distribution of saliency maps, leading to a new state-of-the-art in RGB-D saliency detection. | ['Nick Barnes', 'Fatemeh Sadat Saleh', 'Deng-Ping Fan', 'Jing Zhang', 'Saeed Anwar', 'Yuchao Dai', 'Tong Zhang'] | 2020-04-13 | uc-net-uncertainty-inspired-rgb-d-saliency-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Zhang_UC-Net_Uncertainty_Inspired_RGB-D_Saliency_Detection_via_Conditional_Variational_Autoencoders_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Zhang_UC-Net_Uncertainty_Inspired_RGB-D_Saliency_Detection_via_Conditional_Variational_Autoencoders_CVPR_2020_paper.pdf | cvpr-2020-6 | ['rgb-d-salient-object-detection', 'thermal-image-segmentation'] | ['computer-vision', 'computer-vision'] | [ 2.81997621e-01 5.91376185e-01 -1.73890874e-01 -3.75586063e-01
-9.50444639e-01 -3.32583606e-01 7.13865876e-01 1.16186209e-01
-7.33328015e-02 3.71134579e-01 2.52130330e-01 1.85748011e-01
2.05142513e-01 -3.24412555e-01 -9.98871565e-01 -3.70475978e-01
4.03098494e-01 4.60571080e-01 8.40622902e-01 -5.92685379e-02
5.58303118e-01 1.62744284e-01 -1.97763693e+00 1.92980289e-01
1.03098691e+00 1.22609913e+00 7.76206315e-01 4.68384266e-01
3.93363908e-02 7.43557274e-01 -3.78325194e-01 -1.90168805e-02
2.74052918e-01 -6.69694766e-02 -8.74880910e-01 3.17149997e-01
4.27190274e-01 -4.09262419e-01 1.31251160e-02 1.09085929e+00
2.18977004e-01 6.41903877e-02 7.98328638e-01 -1.47453809e+00
-7.57178187e-01 6.54536784e-01 -7.38268256e-01 -9.37703326e-02
4.38807487e-01 7.63054192e-02 1.17191386e+00 -1.22404635e+00
7.99095035e-01 1.15836167e+00 5.45562327e-01 6.88775837e-01
-1.39452934e+00 -6.96401950e-03 1.82713971e-01 7.33441710e-02
-1.29685771e+00 -1.04202271e-01 1.13918424e+00 -3.85324985e-01
5.79323113e-01 2.03344822e-01 7.80965805e-01 1.12140858e+00
-4.41802517e-02 1.71957099e+00 1.16494977e+00 -3.76290113e-01
6.48216367e-01 2.09945738e-01 -1.40689299e-01 4.98079151e-01
-7.59076094e-03 5.67909889e-03 -1.07774436e+00 6.23536967e-02
9.41134751e-01 -3.13199647e-02 3.33696604e-02 -1.18640792e+00
-1.58258462e+00 7.41421759e-01 1.04106700e+00 -5.16958386e-02
-4.16689336e-01 3.23609501e-01 -1.47780571e-02 -5.25148451e-01
7.54763722e-01 4.30797070e-01 -2.51527667e-01 3.97548527e-01
-1.56548858e+00 4.63881999e-01 3.35119247e-01 1.21855152e+00
1.07960272e+00 -5.09360433e-02 -4.62064713e-01 2.46057943e-01
7.25714028e-01 7.64389634e-01 1.93565965e-01 -1.16759622e+00
-9.09489393e-02 5.90833366e-01 6.00772142e-01 -8.08216929e-01
-3.99125278e-01 -1.57765478e-01 -3.64864320e-01 5.94904006e-01
1.07824512e-01 2.22502291e-01 -1.32509887e+00 1.50872064e+00
4.77737874e-01 5.02319515e-01 4.74183634e-02 1.39658165e+00
9.32404995e-01 2.77583599e-01 1.34739101e-01 3.61560106e-01
8.77627075e-01 -1.03930986e+00 -3.05781901e-01 -3.18780929e-01
-2.31987443e-02 -6.21856272e-01 1.10663128e+00 2.71294708e-03
-1.19513154e+00 -7.22717702e-01 -1.08265471e+00 -5.70902765e-01
-3.25831056e-01 4.21667725e-01 6.17988229e-01 2.02268168e-01
-1.44685745e+00 4.52489257e-01 -1.14885259e+00 -1.64468780e-01
6.43492579e-01 2.02628031e-01 -2.02156268e-02 4.42539096e-01
-1.06868553e+00 1.04804945e+00 5.54650962e-01 -9.55998525e-02
-1.40600550e+00 -8.41538250e-01 -9.17204022e-01 -2.92228311e-01
3.43680024e-01 -8.91667962e-01 1.37734830e+00 -1.16197395e+00
-1.37954962e+00 1.13997495e+00 -3.26768547e-01 -5.96852541e-01
5.89834690e-01 -3.76993030e-01 2.70052046e-01 3.04146707e-01
5.07404387e-01 1.58522904e+00 1.37440026e+00 -1.75946474e+00
-7.95298398e-01 -9.37333927e-02 -1.42596141e-01 2.40961194e-01
3.62172544e-01 -3.02127719e-01 -3.13171297e-01 -4.37299877e-01
4.32734400e-01 -9.56259847e-01 -2.34976217e-01 1.73689172e-01
-8.14232647e-01 -3.81425858e-01 7.64308691e-01 -4.32564586e-01
4.85663712e-01 -1.87001491e+00 7.21327484e-01 -3.01015675e-02
5.93823373e-01 -2.39848495e-01 1.98588580e-01 -4.31762815e-01
2.79315650e-01 3.12547060e-03 -5.38962722e-01 -9.79551971e-01
4.12339628e-01 -5.59739023e-02 -7.16628730e-01 3.75290394e-01
7.26025164e-01 1.29378498e+00 -1.37553668e+00 -5.96995533e-01
6.62301540e-01 5.11875093e-01 -4.13317382e-01 3.87878060e-01
-7.77507424e-01 3.28618467e-01 -5.98473787e-01 9.22151923e-01
6.58863664e-01 -3.56023073e-01 -3.86520833e-01 -1.96512416e-01
-2.42199406e-01 1.76763788e-01 -1.00909162e+00 2.48397255e+00
-2.26358920e-02 6.62259817e-01 -4.61883098e-01 -5.24929702e-01
9.04954314e-01 -4.10746261e-02 3.82580876e-01 -2.38376558e-01
9.14239213e-02 4.21493798e-01 -8.66490424e-01 1.26978412e-01
8.83264065e-01 1.85362652e-01 -2.60126352e-01 5.76417267e-01
5.04701257e-01 -9.03763056e-01 -2.90572941e-01 4.44336683e-01
5.36465585e-01 9.23886597e-01 -3.54449935e-02 -3.25051069e-01
1.30089045e-01 2.79129565e-01 2.61672854e-01 9.32234228e-01
-5.17845929e-01 1.24833786e+00 2.95689195e-01 -4.91582096e-01
-1.22720325e+00 -1.67888069e+00 2.93109799e-04 9.48171735e-01
8.20739508e-01 3.74876708e-02 -9.70530450e-01 -8.23051572e-01
1.99372619e-01 8.57201993e-01 -1.02761066e+00 -5.39125092e-02
-8.64454918e-03 -1.17259406e-01 -8.50876942e-02 4.11079317e-01
4.09009457e-01 -1.28971803e+00 -1.24015594e+00 -2.22191755e-02
-2.81227767e-01 -9.18550789e-01 -2.37297192e-01 2.57233948e-01
-6.66264892e-01 -8.23500097e-01 -9.79595542e-01 -7.60850787e-01
5.42557538e-01 5.15112400e-01 1.35168350e+00 -2.06542432e-01
-2.70022899e-01 6.59622848e-01 -2.31873944e-01 -8.07366669e-01
-2.66823649e-01 2.89172620e-01 2.34012887e-01 -1.89717248e-01
4.38999057e-01 -1.34850323e-01 -7.02565849e-01 1.54424325e-01
-8.56237710e-01 5.56887388e-01 6.08428895e-01 5.27848005e-01
1.08764446e+00 -5.53234935e-01 4.08371955e-01 -2.62006909e-01
1.12114601e-01 -5.91617942e-01 -6.64876342e-01 2.80212343e-01
-2.87629277e-01 2.96607643e-01 -2.80534118e-01 -4.41964529e-02
-1.09722710e+00 5.97719729e-01 3.70162398e-01 -9.18979526e-01
-4.61589038e-01 -2.03157105e-02 1.50552094e-01 -3.09313331e-02
6.55435979e-01 3.06614995e-01 -6.93975314e-02 -2.63113707e-01
8.36188495e-01 3.30458015e-01 7.06351221e-01 -3.49484682e-01
7.22114146e-01 6.67177498e-01 -8.25114921e-02 -4.40620720e-01
-1.34289408e+00 -3.97886962e-01 -1.08100808e+00 -4.23400611e-01
1.26569796e+00 -1.30491638e+00 -3.55200648e-01 4.48063493e-01
-1.27700961e+00 -4.58232403e-01 -4.39490378e-01 -1.31836915e-02
-1.05289638e+00 -1.26839345e-02 -7.93004259e-02 -7.14096665e-01
-1.29506692e-01 -1.39958274e+00 2.07640004e+00 4.49203283e-01
-2.38979444e-01 -8.39575291e-01 1.83812957e-02 9.58948955e-02
2.71687120e-01 4.32877302e-01 2.07083732e-01 -1.33913264e-01
-1.23106694e+00 1.46592230e-01 -4.72746700e-01 3.08592301e-02
7.19299912e-02 -6.82695955e-02 -1.33527207e+00 1.87249824e-01
-1.52745962e-01 -5.08065999e-01 1.04434943e+00 7.48539746e-01
1.12885463e+00 2.69222468e-01 -4.52857405e-01 3.80205542e-01
1.31535840e+00 -8.70437443e-01 3.34814608e-01 2.21135274e-01
8.12578678e-01 5.10214269e-01 9.24377084e-01 3.26519430e-01
6.03355944e-01 6.07635319e-01 8.78583848e-01 -1.93467572e-01
-3.45544279e-01 -6.90287769e-01 1.07647590e-01 1.39985681e-01
1.34338915e-01 8.28669965e-02 -1.04361522e+00 8.84924054e-01
-2.10815215e+00 -6.61202252e-01 -5.58629353e-03 1.91921103e+00
9.14463460e-01 3.18897218e-01 3.30126107e-01 -2.42420793e-01
7.07671881e-01 1.18266396e-01 -7.90329754e-01 1.12054795e-01
-2.81542838e-01 -5.76650798e-02 5.15377760e-01 6.03578627e-01
-1.36106300e+00 1.45097995e+00 6.63646746e+00 6.06085181e-01
-1.16617954e+00 1.02112643e-01 7.39797711e-01 -2.25682542e-01
-7.49936700e-01 8.10959097e-03 -7.51008153e-01 4.51623201e-01
4.53610718e-01 8.05116370e-02 5.78732342e-02 1.13781738e+00
2.76314393e-02 -5.67650795e-01 -1.05992925e+00 8.93549681e-01
1.52292639e-01 -1.67673731e+00 1.26855329e-01 -3.80180150e-01
1.28465497e+00 4.29466903e-01 4.92445588e-01 -1.80072680e-01
5.76993465e-01 -8.63903761e-01 1.34528506e+00 9.56261754e-01
4.55522358e-01 -5.22025585e-01 4.80693042e-01 1.28614634e-01
-7.94826090e-01 2.02322960e-01 -5.04710555e-01 1.76388964e-01
1.62573129e-01 8.07818830e-01 -1.11574733e+00 2.27370843e-01
9.80532408e-01 1.07466245e+00 -8.71471226e-01 1.04045832e+00
-7.49191880e-01 1.25922233e-01 -4.20167089e-01 -3.03271025e-01
2.87695646e-01 2.13098094e-01 8.31807852e-01 8.19964468e-01
1.69679761e-01 -3.54938298e-01 1.78709775e-01 1.69553900e+00
3.30796726e-02 -3.74570519e-01 -3.40370297e-01 4.31001425e-01
4.18031901e-01 1.18591452e+00 -8.89030159e-01 -1.83830485e-01
2.62662638e-02 1.25224388e+00 3.92173201e-01 3.13834012e-01
-8.74237776e-01 -2.67511923e-02 5.08403897e-01 -1.89959735e-01
5.11936724e-01 -1.74639940e-01 -7.98494041e-01 -1.06007469e+00
-1.29720718e-01 -2.49293178e-01 -2.17000008e-01 -1.53486907e+00
-1.03482783e+00 4.42586571e-01 1.70564398e-01 -1.32821238e+00
-2.16780663e-01 -4.87525374e-01 -3.70226026e-01 9.03976202e-01
-1.67759204e+00 -1.61848056e+00 -4.42115039e-01 3.88988554e-01
5.55714190e-01 9.48176533e-02 5.29795349e-01 -6.16886914e-01
9.82354209e-02 1.41289933e-02 -2.31031686e-01 -4.10517961e-01
5.50816178e-01 -1.92674053e+00 6.76470816e-01 1.00328803e+00
3.75065356e-01 3.50400358e-01 9.82736468e-01 -6.72166348e-01
-9.37901139e-01 -9.93532300e-01 6.76446795e-01 -1.20422912e+00
5.72313786e-01 -6.37883663e-01 -6.62023306e-01 5.80092669e-01
1.97436646e-01 1.41455159e-01 1.92484111e-01 -1.71325684e-01
-1.42550975e-01 3.18388402e-01 -1.14356577e+00 6.23606384e-01
9.45870042e-01 -8.13349664e-01 -1.05376554e+00 1.53248951e-01
1.25523722e+00 -7.92284250e-01 -5.18329978e-01 2.57752210e-01
2.78081626e-01 -1.06147003e+00 1.22059715e+00 -1.87821999e-01
8.16299200e-01 -7.74421990e-01 -1.31274477e-01 -1.37850654e+00
-1.06142960e-01 -3.96442294e-01 -4.12517011e-01 8.21643353e-01
3.26782912e-01 2.00442627e-01 9.82304573e-01 6.76168561e-01
-3.49455446e-01 -5.26776552e-01 -1.01899624e+00 3.62026971e-03
-2.42984444e-01 -4.88901079e-01 6.87580884e-01 4.27789271e-01
-2.34479055e-01 -4.05297987e-02 -3.07895690e-01 4.30698425e-01
1.25940716e+00 4.72072899e-01 7.30217755e-01 -1.35373461e+00
-3.44688771e-03 -4.85420257e-01 -4.33783203e-01 -1.19428325e+00
4.65641886e-01 -7.76620209e-01 6.85072541e-01 -1.47262943e+00
1.34774387e-01 -4.50636744e-01 -6.08531833e-01 6.01435125e-01
-4.52145427e-01 4.96942282e-01 1.80304557e-01 2.24403203e-01
-1.20357621e+00 7.50630438e-01 1.07694149e+00 -3.78246516e-01
-2.90507138e-01 -6.11094087e-02 -6.47155046e-01 7.09206462e-01
4.86311376e-01 -3.62665147e-01 -2.77892858e-01 -5.60050011e-01
2.13289708e-01 -3.05188209e-01 9.65359747e-01 -1.09486747e+00
3.25164735e-01 -1.67828605e-01 6.86898530e-01 -1.04922795e+00
4.04333472e-01 -5.72531283e-01 -3.42952877e-01 4.88754399e-02
-5.06830454e-01 -6.24446869e-01 1.88838765e-01 7.28910923e-01
-7.55420746e-03 3.70339118e-02 6.04808629e-01 -1.12673447e-01
-1.35766685e+00 2.35386223e-01 -1.48013989e-02 -1.20936997e-01
9.96854722e-01 -1.95312500e-01 -1.46310292e-02 -9.48912650e-02
-8.97143424e-01 2.74921805e-01 9.13362861e-01 8.49317431e-01
9.36130762e-01 -1.37064683e+00 -5.52163780e-01 2.55654842e-01
4.40894544e-01 5.49932122e-01 8.65043998e-02 4.99102235e-01
-2.82999307e-01 3.13453823e-01 -3.87999952e-01 -1.32285249e+00
-6.08006179e-01 5.64109564e-01 3.03698987e-01 3.56495798e-01
-2.31075943e-01 1.27151394e+00 2.89833695e-02 -4.67542976e-01
1.50724053e-01 -6.54262722e-01 4.79919910e-02 -1.49595425e-01
2.43112415e-01 7.18695596e-02 -2.70386964e-01 -6.82754457e-01
-4.21543002e-01 4.54723477e-01 2.19324216e-01 -2.30719239e-01
1.31406713e+00 -4.25468564e-01 2.46275496e-02 8.07817400e-01
7.26667523e-01 -3.82547796e-01 -2.02982402e+00 -2.12713495e-01
1.75669834e-01 -5.07513940e-01 4.29642618e-01 -6.59711480e-01
-5.84614456e-01 7.44883835e-01 8.17487240e-01 1.25482365e-01
7.82548308e-01 5.27735770e-01 4.45908576e-01 2.17161011e-02
5.78746557e-01 -1.21602035e+00 5.00404537e-01 2.10067332e-01
1.02602577e+00 -1.87715209e+00 9.20799822e-02 -2.52870739e-01
-1.01999712e+00 7.50414193e-01 4.84699219e-01 -5.23810923e-01
5.99201202e-01 -3.35807860e-01 3.61249633e-02 -1.72206745e-01
-3.73538852e-01 -5.03448367e-01 7.27642953e-01 7.95981944e-01
-8.63178074e-02 2.54068285e-01 4.80202466e-01 4.53565657e-01
-1.36709094e-01 -1.21636186e-02 5.67673981e-01 7.01473415e-01
-7.47398138e-01 -6.88273072e-01 -1.91871643e-01 2.16105312e-01
1.06758535e-01 2.07008179e-02 -5.75189114e-01 4.12832439e-01
1.31533712e-01 6.99395001e-01 1.06861621e-01 -2.22357467e-01
-4.93221171e-02 7.15114363e-03 3.34458470e-01 -7.21380770e-01
-4.09835950e-02 -6.31888956e-02 -6.36740386e-01 -7.96610057e-01
-9.03085172e-01 -9.47821379e-01 -1.33343577e+00 3.78203541e-01
-3.44725162e-01 -1.50109991e-01 8.66093636e-01 1.02755976e+00
5.84985614e-01 4.08137798e-01 5.05903125e-01 -1.62893593e+00
-3.24312806e-01 -7.68941879e-01 -5.46741724e-01 4.76105541e-01
6.91110969e-01 -1.14197850e+00 -4.01146233e-01 1.38993070e-01] | [9.845017433166504, -0.5815160870552063] |
66045a28-319c-41db-9e31-807701810e27 | bdd100k-a-diverse-driving-video-database-with | 1805.04687 | null | https://arxiv.org/abs/1805.04687v2 | https://arxiv.org/pdf/1805.04687v2.pdf | BDD100K: A Diverse Driving Dataset for Heterogeneous Multitask Learning | Datasets drive vision progress, yet existing driving datasets are impoverished in terms of visual content and supported tasks to study multitask learning for autonomous driving. Researchers are usually constrained to study a small set of problems on one dataset, while real-world computer vision applications require performing tasks of various complexities. We construct BDD100K, the largest driving video dataset with 100K videos and 10 tasks to evaluate the exciting progress of image recognition algorithms on autonomous driving. The dataset possesses geographic, environmental, and weather diversity, which is useful for training models that are less likely to be surprised by new conditions. Based on this diverse dataset, we build a benchmark for heterogeneous multitask learning and study how to solve the tasks together. Our experiments show that special training strategies are needed for existing models to perform such heterogeneous tasks. BDD100K opens the door for future studies in this important venue. | ['Yingying Chen', 'Xin Wang', 'Wenqi Xian', 'Fisher Yu', 'Fangchen Liu', 'Vashisht Madhavan', 'Trevor Darrell', 'Haofeng Chen'] | 2018-05-12 | bdd100k-a-diverse-driving-dataset-for | http://openaccess.thecvf.com/content_CVPR_2020/html/Yu_BDD100K_A_Diverse_Driving_Dataset_for_Heterogeneous_Multitask_Learning_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Yu_BDD100K_A_Diverse_Driving_Dataset_for_Heterogeneous_Multitask_Learning_CVPR_2020_paper.pdf | cvpr-2020-6 | ['semi-supervised-instance-segmentation', 'multi-object-tracking-and-segmentation', 'drivable-area-detection'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-1.05738707e-01 -4.42494214e-01 -2.56112278e-01 -5.06991148e-01
-6.11444533e-01 -5.32873690e-01 6.52463913e-01 -3.69046867e-01
-3.96520495e-01 5.77328026e-01 1.70732476e-03 -4.05008733e-01
-5.64078018e-02 -5.15136719e-01 -9.36775565e-01 -7.49975443e-01
-2.67033845e-01 3.46354663e-01 4.59680140e-01 -5.10230839e-01
1.95685521e-01 -8.66499394e-02 -2.10963559e+00 4.05321151e-01
8.15826476e-01 7.87250459e-01 6.76404595e-01 7.91845560e-01
2.73628712e-01 8.09025049e-01 -5.66886246e-01 -1.96640268e-01
5.29920518e-01 4.47092466e-02 -4.16405708e-01 -4.50390987e-02
8.70765686e-01 -2.18360081e-01 -5.50263107e-01 7.52112389e-01
5.41466534e-01 1.54545590e-01 5.95639467e-01 -1.85022414e+00
-7.22340882e-01 7.56245553e-02 -5.56820750e-01 5.85745156e-01
-9.98396873e-02 4.44965184e-01 7.75792360e-01 -6.73341215e-01
4.37586129e-01 1.01545429e+00 4.53290433e-01 6.66067660e-01
-8.12903583e-01 -8.84290040e-01 3.05949092e-01 8.20727348e-01
-9.80081201e-01 -4.79053080e-01 6.83989406e-01 -6.59248471e-01
8.72349143e-01 1.07981831e-01 3.48101139e-01 1.56259346e+00
4.95222390e-01 8.86320829e-01 1.05911660e+00 1.93798363e-01
1.57877937e-01 1.70364603e-01 2.47662902e-01 5.33979177e-01
2.64310986e-01 3.95531982e-01 -7.94136405e-01 2.92774677e-01
9.09557790e-02 -5.47110382e-03 -5.40939867e-02 -4.21974599e-01
-1.48497486e+00 8.38602006e-01 2.18839139e-01 -1.10906072e-01
1.94988828e-02 1.74589723e-01 6.17091894e-01 7.14914083e-01
3.69038910e-01 4.19129342e-01 -4.64148790e-01 -3.24937671e-01
-6.10706270e-01 6.08852446e-01 6.11390948e-01 1.00750840e+00
9.36921358e-01 3.77320088e-02 -5.01179025e-02 8.93322051e-01
-9.50566605e-02 7.65646875e-01 6.06546879e-01 -1.08173764e+00
6.21318161e-01 1.57390922e-01 2.19781790e-02 -8.01468432e-01
-4.89491016e-01 -1.33747801e-01 -6.51279628e-01 4.08433020e-01
4.47040468e-01 -3.21758240e-01 -8.64232063e-01 1.47832119e+00
4.77920622e-02 5.62913656e-01 2.99211860e-01 9.27074194e-01
1.00951135e+00 7.12229609e-01 -1.08518571e-01 2.84624219e-01
1.22299767e+00 -1.31229448e+00 -3.39445233e-01 -7.48421490e-01
8.06994438e-01 -6.30656540e-01 1.22448862e+00 3.55435103e-01
-5.59590459e-01 -9.29587662e-01 -1.09742284e+00 9.56578478e-02
-4.92690444e-01 -1.56393215e-01 7.40260243e-01 4.92709190e-01
-9.42601323e-01 -6.77218735e-02 -4.25780982e-01 -2.30714008e-01
4.31109369e-01 -3.55136059e-02 -3.37233335e-01 -5.56275785e-01
-1.20007598e+00 1.18652415e+00 2.37728328e-01 -2.77998429e-02
-1.59057045e+00 -7.16860294e-01 -9.03091788e-01 -4.34459150e-01
6.50685251e-01 -4.47319150e-01 1.02718747e+00 -6.47044122e-01
-8.10106635e-01 8.54140222e-01 -2.06716239e-01 -5.02763391e-01
6.07196212e-01 8.77526924e-02 -6.36097491e-01 -2.63454020e-01
3.60746711e-01 9.23367143e-01 1.11764407e+00 -1.24315035e+00
-1.03586006e+00 -3.63258630e-01 2.11740449e-01 1.51568070e-01
-2.82252789e-01 -2.62718737e-01 -5.15675306e-01 -3.14861089e-01
-5.60975134e-01 -1.34366632e+00 -1.76783323e-01 -3.32312524e-01
-4.75023203e-02 -4.78311539e-01 1.40626478e+00 -9.30740461e-02
6.23670220e-01 -2.43940592e+00 -1.37157934e-02 -4.14429814e-01
4.35562134e-01 1.10973090e-01 -5.70812285e-01 6.47357553e-02
2.45333821e-01 -1.88054349e-02 2.63299048e-01 -2.35009253e-01
-2.01213241e-01 4.73923653e-01 -3.75700474e-01 2.42845222e-01
2.93303967e-01 9.54118848e-01 -8.50989461e-01 -3.24293256e-01
3.05395395e-01 -3.94562818e-02 -2.42443547e-01 6.37633204e-02
-2.45387316e-01 3.58105391e-01 -5.35282791e-01 6.44663990e-01
5.84097862e-01 -4.25152108e-02 -3.85673016e-01 1.64857000e-01
-7.59328455e-02 -1.97061792e-01 -8.43233943e-01 1.58342695e+00
-4.99509096e-01 1.10147226e+00 -2.23472387e-01 -1.34040344e+00
9.97017562e-01 -1.28238544e-01 4.06769633e-01 -1.13717902e+00
-1.39900401e-01 -1.22955851e-02 1.41694680e-01 -1.05587041e+00
7.04927444e-01 1.32994726e-01 -3.28133702e-01 1.39782444e-01
-2.71301538e-01 -3.30348313e-01 3.45837206e-01 1.35541379e-01
1.03962278e+00 -9.94857922e-02 -4.04345930e-01 -3.31607819e-01
6.20243512e-02 5.43207705e-01 5.55876970e-01 8.79976928e-01
-7.31067896e-01 2.89276093e-01 3.44545692e-01 -1.00036454e+00
-1.13885999e+00 -8.13785136e-01 -2.07444295e-01 1.34074879e+00
5.14939666e-01 -1.95726231e-01 -2.91490972e-01 -5.04881620e-01
3.53983432e-01 5.46139717e-01 -8.12399447e-01 -3.01081479e-01
-4.35721636e-01 -8.92815232e-01 4.60960001e-01 5.62171638e-01
5.07196784e-01 -1.07669139e+00 -7.85801172e-01 -2.13067140e-02
-2.30789870e-01 -1.49049711e+00 -2.87113875e-01 3.78262550e-01
-4.02894497e-01 -1.15237999e+00 -6.30787075e-01 -9.79622483e-01
8.31315815e-02 9.78610635e-01 1.17262983e+00 -1.32583261e-01
-2.82233059e-01 3.56420308e-01 -1.84293449e-01 -8.39908183e-01
-3.13546211e-01 2.13195667e-01 2.37961456e-01 1.18435873e-02
6.18858635e-01 -2.60884434e-01 -2.87317306e-01 8.64884973e-01
-6.96935356e-01 1.34892672e-01 5.37760615e-01 8.27481270e-01
3.45825225e-01 1.78217903e-01 8.79115820e-01 -4.62009221e-01
3.76120597e-01 -8.68126094e-01 -5.85826457e-01 2.03345850e-01
-4.21074688e-01 -3.17242518e-02 5.46485543e-01 -6.07829213e-01
-7.70007670e-01 2.69650426e-02 3.65736008e-01 -6.17514372e-01
-3.01592827e-01 3.89961839e-01 1.05412185e-01 -1.25736654e-01
7.75896013e-01 2.92803347e-01 1.14636393e-02 8.33938569e-02
1.17273398e-01 6.29666150e-01 6.45822883e-01 -5.77615678e-01
7.40272164e-01 4.38709736e-01 -2.29298119e-02 -1.10316741e+00
-6.76016152e-01 -4.50988770e-01 -3.20390642e-01 -5.81733763e-01
8.65653753e-01 -1.41312528e+00 -7.53458083e-01 7.07991362e-01
-6.46004081e-01 -6.97540820e-01 2.68080115e-01 5.05123734e-01
-6.09355330e-01 1.05063140e-01 1.72608979e-02 -4.55436110e-01
4.28840548e-01 -1.56246948e+00 1.10935020e+00 1.14290960e-01
2.66026676e-01 -7.62504041e-01 1.62708372e-01 8.75433028e-01
4.50438023e-01 -1.53908040e-02 6.60352409e-01 -2.41267741e-01
-7.72681832e-01 5.80763929e-02 -2.71475136e-01 2.58977950e-01
-3.29404622e-02 1.89138111e-02 -1.19519114e+00 -3.22771102e-01
-1.32746190e-01 -9.24449146e-01 1.54218304e+00 4.47230875e-01
1.37461102e+00 4.04900879e-01 -4.77277070e-01 6.62727952e-01
1.19928133e+00 2.97683120e-01 5.60778499e-01 6.16868258e-01
8.49591970e-01 7.28833497e-01 7.93283761e-01 1.74849585e-01
1.06976306e+00 7.65754521e-01 7.98692644e-01 -4.30230312e-02
-1.11018308e-01 4.05220836e-02 5.44322610e-01 4.37435836e-01
-9.54244211e-02 -2.76818186e-01 -1.24954462e+00 7.38507926e-01
-2.01203060e+00 -1.04122353e+00 -1.22527279e-01 1.87687314e+00
3.29216540e-01 3.82445872e-01 2.31327802e-01 -1.36440784e-01
4.81273562e-01 4.71987754e-01 -9.38460290e-01 -1.96237773e-01
-2.97058582e-01 -4.31156129e-01 6.32234037e-01 6.94181547e-02
-1.32276869e+00 9.05405164e-01 6.78777075e+00 7.83656895e-01
-1.28055632e+00 7.39566237e-02 6.90945089e-01 -5.21417081e-01
-2.40303129e-01 -3.45251262e-01 -1.19491613e+00 6.84200048e-01
1.09677851e+00 -2.16072246e-01 3.25783491e-01 9.47646260e-01
2.93678284e-01 -3.45219702e-01 -9.61661279e-01 1.18660188e+00
3.25587541e-01 -1.25539112e+00 -2.52528608e-01 1.25957146e-01
8.70278060e-01 8.33936214e-01 3.49587619e-01 6.80157542e-01
4.81846452e-01 -1.10282290e+00 6.11572206e-01 3.18953663e-01
5.51578760e-01 -4.61760253e-01 4.88593757e-01 5.16817153e-01
-1.03397238e+00 -4.61979389e-01 -7.15276957e-01 -2.48869538e-01
-1.45000881e-02 3.81840229e-01 -5.34429491e-01 1.96349457e-01
1.13845313e+00 1.14728498e+00 -8.69070768e-01 9.00662065e-01
3.51663768e-01 5.30050218e-01 -7.89624229e-02 -1.26056239e-01
5.14249742e-01 -6.34632036e-02 2.63819635e-01 8.70496690e-01
2.97281802e-01 -4.73962367e-01 6.14088476e-01 1.78872317e-01
5.39237214e-03 -2.52706617e-01 -1.27855611e+00 1.44179851e-01
2.16896191e-01 1.34379458e+00 -4.02088344e-01 -3.01748276e-01
-9.46973264e-01 5.38287044e-01 2.83263117e-01 3.59326601e-01
-1.07977366e+00 9.16142985e-02 1.13018441e+00 -1.08705945e-01
3.81583750e-01 -5.61778009e-01 -3.38794559e-01 -1.19810998e+00
1.80901632e-01 -8.70813668e-01 1.99827284e-01 -8.68507564e-01
-1.14264035e+00 5.15706658e-01 -1.18614405e-01 -1.30900693e+00
-7.74750933e-02 -6.26686156e-01 -5.81109762e-01 4.46736157e-01
-2.24185705e+00 -1.12203813e+00 -7.62137532e-01 8.30380917e-01
1.03050625e+00 -5.27536869e-01 4.23273414e-01 3.56516093e-01
-7.45267808e-01 3.85598511e-01 1.29688114e-01 -2.13829458e-01
1.20177865e+00 -9.79951382e-01 6.12642348e-01 5.43846965e-01
1.50428236e-01 -3.30691487e-01 6.30192280e-01 -3.47339660e-01
-1.70737505e+00 -1.47979951e+00 6.01684690e-01 -9.42262113e-01
7.56231189e-01 -5.33505559e-01 -7.53931701e-01 6.60557330e-01
1.99430332e-01 2.72451073e-01 4.19754446e-01 2.94922948e-01
-4.96478349e-01 -4.25849408e-01 -6.34884298e-01 3.54841352e-01
1.17867446e+00 -4.78438556e-01 -4.51998949e-01 5.33484697e-01
6.89021409e-01 -2.39759535e-01 -5.30514240e-01 3.40683460e-01
6.00159287e-01 -9.43774760e-01 8.94853234e-01 -6.90270483e-01
6.22765422e-01 -3.25343400e-01 -3.00029039e-01 -1.58817494e+00
-3.30232799e-01 -1.51281059e-01 2.48832405e-01 6.99184060e-01
6.83475494e-01 -5.83745658e-01 7.20135331e-01 2.72170335e-01
-6.57185555e-01 -5.99829674e-01 -7.71230996e-01 -1.06824958e+00
1.94698423e-01 -7.66442180e-01 4.07849371e-01 9.11259949e-01
-4.51917559e-01 2.76697963e-01 -7.99563527e-01 2.04504281e-01
5.01012266e-01 2.40893170e-01 1.17893994e+00 -1.22205496e+00
-1.15157971e-02 -3.77847463e-01 -5.76956570e-01 -8.18931520e-01
4.03934568e-01 -7.03162730e-01 1.49456427e-01 -1.21214402e+00
2.98032582e-01 -5.83032131e-01 -2.60689765e-01 5.63883126e-01
-3.41615260e-01 4.15732712e-01 8.78580064e-02 1.97274819e-01
-8.60336661e-01 4.32251990e-01 1.34719801e+00 -5.97706497e-01
-2.39975825e-02 2.86489241e-02 -7.30260313e-01 4.24797684e-01
8.16706359e-01 -2.77721792e-01 -7.13096619e-01 -7.55597115e-01
1.36228427e-01 -1.57848462e-01 5.05717754e-01 -1.19373989e+00
2.84256309e-01 -5.01867592e-01 2.35398024e-01 -5.88903666e-01
3.97753984e-01 -6.22076273e-01 -3.53805460e-02 3.06499869e-01
-1.32916421e-01 2.30704501e-01 2.93504715e-01 5.73439896e-01
-3.55597168e-01 1.84120595e-01 6.05085313e-01 -1.13724940e-01
-1.76620650e+00 5.73548079e-01 -4.38602537e-01 3.97056371e-01
1.51385939e+00 -2.68224299e-01 -8.00850570e-01 -2.88665682e-01
-5.22161961e-01 8.24573576e-01 3.31897110e-01 1.22250211e+00
5.04639626e-01 -1.26261795e+00 -9.85675454e-01 3.20270330e-01
8.40516329e-01 -9.12907571e-02 6.21140420e-01 5.96224904e-01
-7.56780058e-02 4.35284585e-01 -5.73647618e-01 -9.61480200e-01
-1.27438343e+00 7.75168002e-01 1.93867698e-01 1.48148403e-01
-5.81603169e-01 6.78209066e-01 3.65802228e-01 -2.99350619e-01
1.05912462e-01 -1.91675231e-01 -1.94901273e-01 3.26299727e-01
4.74834621e-01 2.89069861e-01 1.62322938e-01 -5.64524889e-01
-2.97704399e-01 5.54378867e-01 -1.66996747e-01 4.12643701e-01
1.30212975e+00 -1.47074252e-01 5.40511012e-01 5.17262697e-01
1.02560639e+00 -5.43781221e-01 -1.55938470e+00 -1.61216900e-01
-1.73313454e-01 -4.97515440e-01 1.52179480e-01 -6.17723227e-01
-1.21730673e+00 9.42037940e-01 6.56084239e-01 2.68001556e-01
9.89795268e-01 1.79801643e-01 6.80026054e-01 7.10705280e-01
5.40843368e-01 -1.20270979e+00 2.66862214e-01 6.72036052e-01
7.00077295e-01 -1.88762951e+00 -5.16579926e-01 5.50663508e-02
-1.08584285e+00 7.81549752e-01 9.55031335e-01 4.94264700e-02
7.19714880e-01 4.15435761e-01 4.07003075e-01 -2.73079723e-01
-1.16531825e+00 -2.68947810e-01 7.57558048e-02 8.71367872e-01
-2.12280720e-01 2.23338544e-01 3.90714824e-01 9.39664841e-02
-2.43133292e-01 9.12483409e-03 6.58349633e-01 6.17602229e-01
-5.97682714e-01 -7.94470549e-01 -7.98832402e-02 6.48359954e-01
5.20498082e-02 1.68702573e-01 3.47453132e-02 6.45267010e-01
4.07047272e-01 1.10850537e+00 9.81796831e-02 -6.81052387e-01
3.49936068e-01 -1.41924899e-02 2.19959497e-01 -3.52138489e-01
3.41709219e-02 -5.02124548e-01 2.19648451e-01 -5.68964720e-01
-3.61954331e-01 -8.54553163e-01 -6.33475423e-01 -3.65496665e-01
-8.80875513e-02 -2.36650780e-01 9.19303298e-01 9.83088195e-01
4.83763278e-01 6.16193414e-01 8.76767159e-01 -8.63768518e-01
-2.44941548e-01 -6.87716246e-01 -5.77121377e-01 3.60002935e-01
6.17992818e-01 -1.03136837e+00 -2.99657553e-01 3.10973227e-02] | [8.0474214553833, -1.5524640083312988] |
94c3dd3e-7646-43a1-be58-b8c30db6eaa3 | demonstrating-emma-embodied-multimodal-agent | null | null | https://aclanthology.org/2022.sigdial-1.62 | https://aclanthology.org/2022.sigdial-1.62.pdf | Demonstrating EMMA: Embodied MultiModal Agent for Language-guided Action Execution in 3D Simulated Environments | We demonstrate EMMA, an embodied multimodal agent which has been developed for the Alexa Prize SimBot challenge. The agent acts within a 3D simulated environment for household tasks. EMMA is a unified and multimodal generative model aimed at solving embodied tasks. In contrast to previous work, our approach treats multiple multimodal tasks as a single multimodal conditional text generation problem, where a model learns to output text given both language and visual input. Furthermore, we showcase that a single generative agent can solve tasks with visual inputs of varying length, such as answering questions about static images, or executing actions given a sequence of previous frames and dialogue utterances. The demo system will allow users to interact conversationally with EMMA in embodied dialogues in different 3D environments from the TEACh dataset. | ['Verena Rieser', 'Oliver Lemon', 'Ioannis Konstas', 'Claudio Greco', 'Arash Eshghi', 'Amit Parekh', 'George Pantazopoulos', 'Malvina Nikandrou', 'Bhathiya Hemanthage', 'Alessandro Suglia'] | null | null | null | null | sigdial-acl-2022-9 | ['conditional-text-generation'] | ['natural-language-processing'] | [ 2.79538214e-01 7.26329029e-01 6.78379655e-01 -3.48217487e-01
-7.25547612e-01 -7.55670071e-01 1.27367508e+00 -3.80140156e-01
-2.47622773e-01 8.27036500e-01 4.77806240e-01 -3.75106037e-02
5.22105277e-01 -6.85443044e-01 -8.19866240e-01 -6.57824218e-01
5.26317358e-02 8.79506171e-01 -3.45601857e-01 -3.11378539e-01
-1.60603523e-01 1.03987709e-01 -1.55459464e+00 6.93538368e-01
3.09636086e-01 3.42140257e-01 7.82966375e-01 1.50062120e+00
-1.19773656e-01 1.45337105e+00 -8.49938571e-01 -4.61855680e-01
-7.44699687e-02 -8.08609188e-01 -1.13791168e+00 4.87443596e-01
2.29366526e-01 -7.24058807e-01 -5.97188473e-01 6.76879168e-01
6.83024406e-01 5.17163396e-01 8.46751571e-01 -1.79202664e+00
-6.24532580e-01 5.52073419e-01 1.47838905e-01 -5.09278059e-01
1.22755516e+00 6.18821859e-01 6.97646916e-01 -8.13328564e-01
9.98842716e-01 1.79184377e+00 -8.81190673e-02 1.08066273e+00
-1.44839442e+00 -4.41864952e-02 1.85115770e-01 -1.88746757e-03
-9.62521434e-01 -5.43714166e-01 3.79496604e-01 -2.73038030e-01
1.42045343e+00 2.82042593e-01 6.95953608e-01 1.86431336e+00
1.52203947e-01 1.30821633e+00 9.12535250e-01 -3.73879761e-01
2.64311105e-01 1.16638750e-01 -3.70374233e-01 9.12333906e-01
-6.84452295e-01 -7.11412951e-02 -7.58860409e-01 -5.69825992e-02
8.94458234e-01 -4.01121616e-01 -1.84533596e-01 -4.00741458e-01
-1.66994810e+00 9.19685245e-01 3.60609442e-02 2.68005002e-02
-4.38818812e-01 3.97358000e-01 2.00911164e-01 5.01719952e-01
1.32673502e-01 3.57140779e-01 -1.29776180e-01 -4.19241488e-01
-2.76824892e-01 7.58165538e-01 1.18850112e+00 1.19410813e+00
6.11227810e-01 -7.80687183e-02 -3.80947948e-01 4.29818422e-01
5.92604578e-01 9.82806981e-01 3.55484635e-02 -1.49303710e+00
4.07150775e-01 4.52387154e-01 3.29034925e-01 -5.22541463e-01
-4.26320225e-01 7.90812612e-01 -5.88133156e-01 7.60358036e-01
4.77050483e-01 -6.89906359e-01 -9.91796851e-01 1.82882929e+00
3.67501616e-01 -1.55353412e-01 6.62347257e-01 9.67769802e-01
1.48793018e+00 1.05323100e+00 3.60089570e-01 1.64088279e-01
1.20949960e+00 -9.56233263e-01 -7.52601385e-01 -2.19150797e-01
3.67809445e-01 -7.90450096e-01 8.79846990e-01 2.38234386e-01
-1.57452190e+00 -4.27531928e-01 -5.41230261e-01 -3.39442283e-01
-4.91632193e-01 3.10375709e-02 5.64179838e-01 2.43466586e-01
-1.37661767e+00 -7.96758458e-02 -7.92381227e-01 -7.69140899e-01
-1.31257460e-01 3.15254658e-01 -5.44241309e-01 4.12125029e-02
-9.63955402e-01 1.09086311e+00 3.78487706e-01 -3.14237140e-02
-1.72788918e+00 -1.78424746e-01 -1.14832425e+00 -2.53039241e-01
2.68731505e-01 -1.32458735e+00 1.81609905e+00 -9.07599390e-01
-1.99006903e+00 1.18966472e+00 -2.15091109e-01 -3.82415235e-01
6.75874949e-01 -1.65354028e-01 6.86859712e-02 3.80927235e-01
-9.80800390e-02 1.50902355e+00 6.76188529e-01 -1.66565883e+00
-4.24470872e-01 -2.29859710e-01 7.20092058e-01 8.05294812e-01
7.28674412e-01 -1.96419865e-01 -2.92675048e-01 -1.84916511e-01
-5.15582085e-01 -1.05715764e+00 -2.94939548e-01 -2.78036356e-01
-4.65573221e-01 -2.14396834e-01 8.42098892e-01 -4.44471389e-01
1.10560447e-01 -1.84841347e+00 1.12597799e+00 -1.97137162e-01
1.68532714e-01 -2.85674810e-01 -4.07920003e-01 8.51601064e-01
2.15160593e-01 -1.39745742e-01 1.04651570e-01 -8.69717598e-01
7.25391805e-01 3.29484910e-01 -1.99243441e-01 -2.07517259e-02
1.50707960e-01 1.37387276e+00 -9.62670922e-01 -5.11879206e-01
6.86002791e-01 5.62749445e-01 -1.01405405e-01 8.29779744e-01
-9.16671038e-01 6.96866155e-01 -2.80353516e-01 4.52924311e-01
2.87817985e-01 -2.22305760e-01 2.81659275e-01 2.48725966e-01
3.83495651e-02 -1.00408323e-01 -9.27682102e-01 2.39308167e+00
-5.71851790e-01 8.17928255e-01 3.91316205e-01 -6.50605083e-01
5.76207399e-01 7.43658900e-01 1.43213928e-01 -5.54798424e-01
3.08607578e-01 -3.67976725e-01 -6.07356131e-01 -9.65701401e-01
7.24243462e-01 -7.70151839e-02 -4.85178798e-01 7.70353258e-01
4.65567768e-01 -7.98681855e-01 2.33540267e-01 7.71715701e-01
1.17632842e+00 6.35488987e-01 -1.55319452e-01 3.87823611e-01
1.06528349e-01 2.53476799e-01 -5.29899657e-01 9.80991185e-01
3.20384316e-02 5.65478086e-01 4.42152739e-01 -2.41920516e-01
-1.20431006e+00 -1.57373500e+00 5.33649445e-01 1.42332876e+00
3.59111786e-01 -3.54090512e-01 -8.00065577e-01 -5.16087174e-01
-3.48193794e-01 1.12510145e+00 -6.81678712e-01 2.23079219e-01
-3.64436150e-01 -3.24204355e-01 5.53531706e-01 2.12286964e-01
4.90065485e-01 -1.68054736e+00 -1.06787539e+00 2.65387595e-02
-5.94305515e-01 -1.13909888e+00 -9.52321365e-02 7.36835748e-02
-2.06247255e-01 -9.63382840e-01 -9.38666344e-01 -8.96773517e-01
6.97421670e-01 1.42699763e-01 1.51937020e+00 -1.57945111e-01
-3.05612713e-01 1.51772797e+00 -5.15649915e-01 -2.71580368e-01
-9.03000891e-01 -1.17576472e-01 -4.69765455e-01 -4.57080573e-01
1.01551376e-01 -3.60725313e-01 -3.47942561e-01 1.34545863e-01
-9.32821870e-01 9.67179418e-01 4.03605998e-01 8.17426205e-01
-6.86334372e-02 -6.27645552e-01 2.98919469e-01 -4.57841009e-01
9.15628076e-01 -6.15460873e-01 -1.84395999e-01 4.86798346e-01
5.53122759e-01 -6.03032820e-02 8.35884064e-02 -6.05425656e-01
-1.53479028e+00 2.79246509e-01 -3.94222821e-04 -2.11056277e-01
-9.40934896e-01 1.28618717e-01 -2.34900966e-01 4.98845547e-01
4.40255135e-01 2.48661011e-01 1.25516802e-01 1.49601683e-01
8.26024532e-01 2.86115766e-01 7.74854183e-01 -9.47286129e-01
5.65710008e-01 3.16188455e-01 -1.59660533e-01 -8.64435613e-01
-4.59375195e-02 1.92941770e-01 -6.34558022e-01 -6.43246531e-01
1.48780513e+00 -1.02088976e+00 -1.53111494e+00 5.53862751e-01
-1.60695589e+00 -1.05033946e+00 -1.83928385e-01 2.06421807e-01
-1.31716919e+00 -2.06684805e-02 -5.87947965e-01 -9.87518728e-01
-1.55035689e-01 -1.19246745e+00 1.55363214e+00 5.08826911e-01
-7.23731339e-01 -1.12889624e+00 1.73703194e-01 3.52126837e-01
1.37705207e-01 6.14890635e-01 7.09146917e-01 -2.86036253e-01
-8.15333962e-01 1.91532582e-01 1.32584348e-01 -3.20983112e-01
-2.28989482e-01 -4.90865223e-02 -8.24739099e-01 -1.03630409e-01
-3.91657293e-01 -1.16896188e+00 3.79068792e-01 1.73525319e-01
3.37890625e-01 -1.58415318e-01 -2.76998311e-01 1.29975438e-01
8.79029870e-01 4.35047269e-01 6.88708723e-01 6.92750961e-02
3.29069227e-01 6.62454605e-01 3.90755773e-01 3.77513438e-01
7.56976128e-01 4.96332288e-01 6.64904594e-01 -2.43166938e-01
9.85192508e-03 -4.03550535e-01 6.29350066e-01 6.58672675e-02
-1.56823143e-01 -8.12080204e-01 -7.59576678e-01 5.18353999e-01
-2.29132557e+00 -1.27242398e+00 9.13881361e-02 1.46658897e+00
5.28928816e-01 -4.56228733e-01 1.14789307e-01 -6.77675128e-01
3.60981792e-01 1.06175095e-01 -4.12345737e-01 -4.11951095e-01
-3.62736046e-01 1.31738484e-01 -2.85085082e-01 6.42620564e-01
-9.38414156e-01 1.00777102e+00 6.81260490e+00 1.98923081e-01
-6.35010779e-01 -7.45537365e-03 4.40704703e-01 -4.18677390e-01
-2.71740377e-01 -1.16356835e-01 -3.63062620e-01 1.06389508e-01
7.31626272e-01 -1.59616336e-01 7.80097842e-01 4.57958490e-01
1.83846384e-01 -7.23659873e-01 -1.47980762e+00 9.29099679e-01
3.24905783e-01 -1.32444072e+00 7.08260864e-04 -1.49142534e-01
6.74269736e-01 -2.41432246e-02 9.23254192e-02 5.16710222e-01
1.27913916e+00 -1.51627159e+00 8.38821709e-01 8.39451373e-01
4.66968656e-01 -6.72734261e-01 4.05340821e-01 6.57663047e-01
-8.09513152e-01 3.53484273e-01 1.11589395e-01 -7.80285820e-02
7.71525562e-01 -5.61563790e-01 -1.14435053e+00 4.21897650e-01
6.27525210e-01 1.12030499e-01 -2.73132265e-01 4.32345271e-01
-4.08924609e-01 -3.91228730e-03 -1.36012763e-01 -4.42981541e-01
3.22523743e-01 -2.51182824e-01 6.98518217e-01 1.16426432e+00
3.67403567e-01 4.47877347e-01 1.83812216e-01 1.04449153e+00
3.57425138e-02 -2.96504408e-01 -9.97696996e-01 -1.27449483e-01
-1.37023136e-01 1.22846782e+00 -4.60634977e-01 -4.13094610e-01
-2.92297304e-01 1.56667149e+00 1.53259173e-01 8.32573473e-01
-8.01196218e-01 -2.73724496e-02 5.37369847e-01 -4.99737859e-01
1.02905437e-01 -5.75991094e-01 1.90769196e-01 -1.21490026e+00
-5.72166502e-01 -1.04754937e+00 1.73942633e-02 -1.74156499e+00
-9.82094646e-01 7.49655545e-01 2.90004551e-01 -7.19457746e-01
-8.85573268e-01 -5.39208472e-01 -7.29884565e-01 8.25779080e-01
-5.58120608e-01 -1.57787120e+00 -5.76771319e-01 7.96022892e-01
9.97406721e-01 -3.12431425e-01 1.29475093e+00 -3.78203750e-01
-1.09979175e-01 -1.77468464e-01 -3.63632977e-01 4.05809097e-02
6.99287176e-01 -1.49502456e+00 4.75043714e-01 1.46670386e-01
2.35481158e-01 2.84851283e-01 1.01084924e+00 -6.01692855e-01
-1.65541780e+00 -3.93413812e-01 5.71579456e-01 -7.18358219e-01
4.95705873e-01 -6.64803743e-01 -4.35591877e-01 1.08577442e+00
1.38544559e+00 -7.73132980e-01 5.71428478e-01 -3.70588183e-01
3.63802820e-01 8.48096192e-01 -1.13822436e+00 9.02268529e-01
1.11008906e+00 -5.37189603e-01 -8.04175615e-01 2.81892896e-01
6.80067062e-01 -9.28463042e-01 -6.25342071e-01 -3.28119360e-02
4.96468216e-01 -8.30442786e-01 9.08023953e-01 -8.35847318e-01
8.17362905e-01 -3.18521857e-01 -4.58149135e-01 -1.60845923e+00
3.36199492e-01 -8.74738634e-01 -1.95205435e-02 1.16028392e+00
3.21862757e-01 -2.34896932e-02 5.60257077e-01 8.77450466e-01
2.09076419e-01 -5.86496405e-02 -7.25140989e-01 1.59515552e-02
-2.11679190e-02 -2.88770080e-01 4.48501438e-01 6.07906163e-01
2.82138377e-01 7.69281447e-01 -6.24529362e-01 1.14446752e-01
4.52512175e-01 1.16903670e-02 1.56019390e+00 -7.50872135e-01
-3.04894865e-01 -6.42034635e-02 -1.18325777e-01 -1.35693741e+00
5.70947945e-01 -7.97571540e-01 4.04301167e-01 -1.85177588e+00
3.63707423e-01 3.19950283e-01 8.10116768e-01 4.99376506e-01
1.38704747e-01 -3.44339460e-02 6.75045848e-01 -2.25705847e-01
-9.22920644e-01 7.85386562e-01 1.33343458e+00 -4.28192496e-01
-1.32666528e-01 -1.68117702e-01 -9.43832248e-02 6.79776669e-01
5.09115458e-01 5.95315434e-02 -5.98843396e-01 -5.39080799e-01
4.00531262e-01 7.63125360e-01 9.38079059e-01 -5.35104871e-01
2.86511540e-01 -3.44824553e-01 6.16945803e-01 -8.40095103e-01
7.84188807e-01 -6.68088377e-01 5.78099728e-01 6.88446015e-02
-5.10947466e-01 1.14937812e-01 3.56869370e-01 2.89810032e-01
-7.12444484e-02 -1.65093288e-01 1.51585266e-01 -6.07351661e-01
-9.33578610e-01 -2.25845143e-01 -1.06162512e+00 -2.34897718e-01
1.25410378e+00 1.56646132e-01 -6.72693908e-01 -1.04906523e+00
-1.48024368e+00 7.50975132e-01 6.78176999e-01 6.09814227e-01
9.06262100e-01 -1.33073437e+00 -8.27462196e-01 -2.20937148e-01
-1.07065976e-01 9.05985162e-02 4.22677666e-01 3.30310851e-01
-5.73534250e-01 1.65588737e-01 -3.50332797e-01 -7.93917239e-01
-1.49643636e+00 3.36911738e-01 3.73523712e-01 -8.70744660e-02
-5.45766890e-01 6.64671838e-01 4.22648668e-01 -7.60600030e-01
3.28901500e-01 2.58322079e-02 2.32139472e-02 -1.83050618e-01
5.22974670e-01 1.46434769e-01 -5.25267720e-01 -6.89696372e-01
9.98975560e-02 9.49405059e-02 4.29634154e-01 -8.88533175e-01
1.33085382e+00 -5.84730148e-01 9.04762596e-02 7.01668739e-01
8.43792498e-01 -4.46997970e-01 -1.42243886e+00 6.54429868e-02
-6.98580086e-01 -4.21773195e-02 -5.06812513e-01 -1.17290211e+00
-3.56115341e-01 7.68974543e-01 4.20236647e-01 3.42438817e-01
8.21811616e-01 4.95948941e-01 4.62320387e-01 8.19670916e-01
4.28564548e-01 -9.41666245e-01 7.18321860e-01 6.33865476e-01
1.25625610e+00 -1.41312015e+00 -4.82575744e-01 1.55583322e-01
-1.28542995e+00 1.13606632e+00 6.68256402e-01 2.27870330e-01
-8.13802853e-02 6.15815461e-01 2.07367063e-01 -3.84004444e-01
-1.26336527e+00 -3.92950356e-01 1.77266728e-02 9.87583876e-01
2.51814574e-01 1.59126356e-01 4.06757295e-01 1.64128780e-01
-3.92328471e-01 -1.55892134e-01 5.47184587e-01 1.08452678e+00
-3.28511983e-01 -9.46551919e-01 -5.83057940e-01 -3.57659161e-01
1.70968607e-01 2.43154705e-01 -9.42237914e-01 9.14917111e-01
2.26851478e-02 1.17397082e+00 2.56907433e-01 -2.14217201e-01
2.38161191e-01 3.81396741e-01 9.91745889e-01 -5.28052151e-01
-5.10986924e-01 -1.79020800e-02 3.57114792e-01 -3.62271190e-01
-7.13617921e-01 -8.01470876e-01 -1.50504684e+00 -4.74501818e-01
3.46284151e-01 -8.40125754e-02 6.64790094e-01 9.23124254e-01
-1.00395819e-02 3.78632516e-01 2.88620263e-01 -1.52952695e+00
2.18182787e-01 -1.01354098e+00 -1.84657261e-01 5.82213700e-01
3.69480610e-01 -3.04577500e-01 -1.14023775e-01 6.10199094e-01] | [10.694376945495605, 1.3037956953048706] |
21edc82b-efe2-4ce0-a7b4-6f0245af7f76 | automatic-liver-segmentation-with-adversarial | 1811.11566 | null | http://arxiv.org/abs/1811.11566v1 | http://arxiv.org/pdf/1811.11566v1.pdf | Automatic Liver Segmentation with Adversarial Loss and Convolutional Neural Network | Automatic segmentation of medical images is among most demanded works in the
medical information field since it saves time of the experts in the field and
avoids human error factors. In this work, a method based on Conditional
Adversarial Networks and Fully Convolutional Networks is proposed for the
automatic segmentation of the liver MRIs. The proposed method, without any
post-processing, is achieved the second place in the SIU Liver Segmentation
Challenge 2018, data of which is provided by Dokuz Eyl\"ul University. In this
paper, some improvements for the post-processing step are also proposed and it
is shown that with these additions, the method outperforms other baseline
methods. | ['Bora Baydar', 'Savas Ozkan', 'Gozde Bozdagi Akar'] | 2018-11-28 | null | null | null | null | ['liver-segmentation'] | ['medical'] | [ 1.55310169e-01 4.63983208e-01 1.76700637e-01 -3.26389849e-01
-5.81532359e-01 -3.79019260e-01 2.60680109e-01 2.66067028e-01
-6.74657464e-01 5.88120341e-01 4.02842611e-02 -5.66427529e-01
1.74510494e-01 -5.19407153e-01 -5.97245216e-01 -7.76042461e-01
5.81364110e-02 3.28920901e-01 2.23875597e-01 2.16433052e-02
-9.20576900e-02 4.68205601e-01 -3.89335066e-01 8.32598507e-02
8.81365478e-01 9.22281682e-01 -1.51872814e-01 8.05269778e-01
3.31071645e-01 7.61200607e-01 -3.45783353e-01 -4.08053160e-01
6.03670239e-01 -4.85135525e-01 -1.15880692e+00 1.54602945e-01
1.77731395e-01 -5.39607882e-01 -4.67926085e-01 1.27091265e+00
6.75778389e-01 -3.02119460e-02 6.14590824e-01 -6.15747154e-01
-3.61563057e-01 9.66627002e-01 -4.95465487e-01 3.85831267e-01
-5.59878461e-02 -1.38131045e-02 2.59008229e-01 -5.81802726e-01
6.43452644e-01 7.29510009e-01 7.56215096e-01 5.43632746e-01
-9.18645799e-01 -6.05389595e-01 -4.25281078e-01 8.97002593e-02
-1.27411377e+00 -2.50054859e-02 7.15558350e-01 -6.33189380e-01
3.11471939e-01 2.57432818e-01 6.25333905e-01 7.41448641e-01
3.33268940e-01 9.42288458e-01 1.27922857e+00 -5.47445595e-01
8.64285231e-02 -4.27267849e-02 3.55000705e-01 1.00373363e+00
1.82944536e-01 1.43948764e-01 4.20386374e-01 4.36816923e-02
7.95609355e-01 6.99998438e-02 -4.09818530e-01 -3.14404637e-01
-1.27131093e+00 9.35636818e-01 7.76333272e-01 7.41480768e-01
-5.14423668e-01 5.27861379e-02 5.26805758e-01 3.06908786e-02
6.82150722e-01 2.96588272e-01 -3.12242001e-01 4.44341302e-01
-1.38793015e+00 -1.77473679e-01 7.41408646e-01 8.42207909e-01
1.16177030e-01 1.28418198e-02 -3.81368577e-01 3.31862003e-01
4.00520086e-01 1.09113261e-01 8.20928037e-01 -5.06107986e-01
-2.01489013e-02 6.74397051e-01 -9.99490619e-02 -7.86224306e-01
-6.65700972e-01 -5.57232857e-01 -1.25088084e+00 1.15189783e-01
4.92263436e-01 -5.16871572e-01 -1.48003161e+00 1.29096222e+00
5.03737509e-01 5.31323612e-01 1.37949109e-01 9.94448781e-01
1.13252175e+00 2.58574933e-01 3.22352290e-01 -1.06330097e-01
1.50101089e+00 -1.40316677e+00 -9.42228138e-01 3.39005142e-02
3.49981368e-01 -1.00614560e+00 3.83628666e-01 3.21048051e-01
-1.16276252e+00 -3.46855044e-01 -1.22411180e+00 1.99440215e-02
-4.22845542e-01 2.45407254e-01 7.86897361e-01 8.71854842e-01
-1.31499588e+00 6.82467997e-01 -1.20570493e+00 -2.73918569e-01
7.46073425e-01 5.51256478e-01 -2.68995762e-01 4.89151552e-02
-1.20104444e+00 8.78864765e-01 6.40002608e-01 1.69228315e-01
-9.19854641e-01 -7.19007015e-01 -9.59163606e-01 -1.55289978e-01
3.15063596e-01 -6.70624018e-01 1.28851020e+00 -9.99555171e-01
-1.54445457e+00 1.06930351e+00 4.15745407e-01 -1.01664662e+00
1.07605815e+00 2.42556408e-02 -5.55902161e-02 5.80533922e-01
-3.32139283e-01 6.13281071e-01 7.02114642e-01 -8.47515583e-01
-2.49576628e-01 -3.28009158e-01 -9.84116551e-03 -1.05188578e-01
9.22028348e-02 -5.46087371e-03 -5.46061754e-01 -1.12423849e+00
1.04452610e-01 -1.05857193e+00 -8.37565184e-01 -2.60633916e-01
-6.10827446e-01 -1.30182831e-02 6.81235611e-01 -1.22779512e+00
7.62083828e-01 -1.96683753e+00 -7.26352334e-02 2.75750220e-01
1.71532601e-01 4.49650377e-01 2.58462340e-01 -2.85150707e-01
-2.62795091e-01 2.14534566e-01 -5.33395708e-01 -2.82904029e-01
-2.46155128e-01 8.23045671e-02 3.44870180e-01 7.67502427e-01
-1.36105403e-01 1.19768000e+00 -7.36340821e-01 -6.88660622e-01
5.70759952e-01 5.89458704e-01 -3.37231368e-01 2.17207432e-01
1.14464834e-01 1.00450337e+00 -1.72442123e-01 5.38122714e-01
8.72822821e-01 -1.74224481e-01 -1.59611236e-02 -3.78413588e-01
2.07665354e-01 -4.29137200e-01 -8.14057410e-01 2.10050416e+00
-1.84376284e-01 2.57562876e-01 2.62742341e-01 -1.24399316e+00
4.61932778e-01 8.84510517e-01 9.15488064e-01 -1.80431738e-01
7.55051553e-01 1.81816712e-01 1.84681758e-01 -4.30697203e-01
1.15059286e-01 -2.04620481e-01 2.27329731e-02 1.22560069e-01
2.12112337e-01 -3.34882915e-01 3.90809268e-01 1.82929352e-01
9.01649952e-01 7.41410702e-02 5.57554305e-01 -6.61706507e-01
8.93175304e-01 9.07912701e-02 4.31727022e-01 5.75461090e-01
-8.00002158e-01 6.65072918e-01 4.86630462e-02 -3.70783865e-01
-9.91237760e-01 -7.57605195e-01 -3.26435030e-01 3.24393779e-01
1.86969995e-01 -2.90764822e-03 -1.41557693e+00 -1.18128252e+00
-3.41961890e-01 2.43694305e-01 -9.32077110e-01 8.93192217e-02
-6.79594338e-01 -8.10831606e-01 6.57153964e-01 4.68052655e-01
7.92305946e-01 -9.67979491e-01 -6.20499551e-01 1.47747159e-01
-1.89996511e-01 -1.33219719e+00 -6.51345074e-01 -1.41229108e-01
-1.00646746e+00 -1.42216396e+00 -1.33282948e+00 -1.05988967e+00
1.12111390e+00 -2.32243225e-01 1.00956285e+00 2.39471838e-01
-7.67526746e-01 1.58574104e-01 -3.86670440e-01 -5.90989769e-01
-6.22283816e-01 7.52270371e-02 -4.44657296e-01 -1.12796389e-01
1.50512502e-01 -2.07043037e-01 -1.10430634e+00 7.89776668e-02
-9.45746541e-01 2.52499759e-01 9.28439379e-01 1.17298722e+00
7.70571291e-01 -1.18009076e-01 4.50024962e-01 -1.46890116e+00
2.53456414e-01 -4.20864642e-01 -4.96364832e-01 1.16582036e-01
-5.55002511e-01 -2.28812560e-01 6.14939511e-01 -1.34355977e-01
-9.48745131e-01 3.83484751e-01 -5.74970126e-01 -3.42732519e-01
-5.01819253e-01 4.28370178e-01 2.34863415e-01 -5.64318955e-01
4.19238389e-01 8.66814982e-03 -3.30675133e-02 -4.67745692e-01
1.21803083e-01 3.99417073e-01 5.67249179e-01 7.50145987e-02
6.17663503e-01 4.71951008e-01 1.29704982e-01 -4.41922128e-01
-6.13036513e-01 -4.89834398e-01 -7.58116961e-01 -1.00832045e-01
1.22416389e+00 -7.38276601e-01 -2.90556490e-01 6.59094870e-01
-9.12393332e-01 -2.83949941e-01 -3.65998387e-01 8.06655109e-01
-4.67019051e-01 7.35772908e-01 -1.16062081e+00 -3.26534033e-01
-9.54162836e-01 -1.73270440e+00 4.03622717e-01 4.79642689e-01
1.82141498e-01 -1.18580854e+00 -2.46771500e-01 5.08775651e-01
7.89676309e-01 6.82065785e-01 5.97437978e-01 -1.14307022e+00
-3.96355659e-01 -4.56410080e-01 -1.62480831e-01 6.28339827e-01
5.95467016e-02 -4.32210296e-01 -6.25687540e-01 -6.39082909e-01
3.05034161e-01 6.53595254e-02 7.41700053e-01 7.36851335e-01
1.29159606e+00 -2.88156509e-01 -2.62712240e-01 8.03927779e-01
1.54866147e+00 2.51525819e-01 5.23521423e-01 8.41509774e-02
4.07904685e-01 1.32401526e-01 3.07096183e-01 5.83210327e-02
4.42960039e-02 1.93301454e-01 6.29561007e-01 -8.08072746e-01
-4.20323342e-01 2.03697056e-01 -2.36796364e-01 8.89106870e-01
-8.28102380e-02 1.01511724e-01 -9.69087005e-01 6.60150945e-01
-1.64337695e+00 -3.86188090e-01 -1.26920104e-01 1.79227400e+00
8.66450310e-01 -1.14930309e-01 -1.36195511e-01 6.34533241e-02
6.95293427e-01 -2.37015516e-01 -2.86369324e-01 -1.49367794e-01
2.91717052e-01 7.42961645e-01 8.90834451e-01 7.06481457e-01
-1.77238321e+00 7.31599212e-01 6.62974405e+00 6.86869383e-01
-1.12887406e+00 4.86222476e-01 1.07835650e+00 4.68851000e-01
4.14402336e-01 -3.37461442e-01 6.72247037e-02 4.13472205e-01
7.64119804e-01 5.15480340e-02 9.25882533e-02 8.02150905e-01
2.44913064e-02 -3.04480016e-01 -7.90494621e-01 6.31931424e-01
3.03829104e-01 -1.15864444e+00 -2.18497425e-01 -1.91685319e-01
8.61171246e-01 -1.71202831e-02 -2.70582978e-02 8.61686245e-02
2.88178593e-01 -1.17632806e+00 2.06174970e-01 4.61511999e-01
6.08328044e-01 -9.23184037e-01 1.53790951e+00 2.56008893e-01
-7.86969244e-01 4.16049957e-01 -2.86955722e-02 3.29617232e-01
5.55379204e-02 5.32666624e-01 -1.19454360e+00 9.91385281e-01
4.74636763e-01 2.59896457e-01 -6.64679885e-01 1.29204178e+00
-3.83769184e-01 9.40539062e-01 -1.23475030e-01 2.76246935e-01
2.53186703e-01 -1.95947945e-01 5.88063657e-01 1.62043142e+00
2.22942263e-01 2.65486270e-01 4.50596482e-01 5.12661517e-01
-2.03147963e-01 4.48333234e-01 -3.59099984e-01 1.82349503e-01
-3.91379505e-01 1.62874877e+00 -1.20281923e+00 -5.88111043e-01
-8.50134790e-02 1.16255260e+00 -3.13354254e-01 1.55322686e-01
-9.47220623e-01 -3.96330655e-01 -2.83381999e-01 -1.33316591e-01
-3.02594919e-02 -1.06117144e-01 -4.60762829e-01 -1.10554004e+00
-4.36650187e-01 -6.57341242e-01 6.34185612e-01 -2.82217771e-01
-1.01191568e+00 8.58309865e-01 -1.87065080e-01 -9.04731095e-01
-2.38120139e-01 -6.72210932e-01 -4.00947839e-01 8.74419391e-01
-1.71576965e+00 -1.42210495e+00 -3.18203866e-01 7.26856291e-01
6.75509155e-01 -1.26865774e-01 9.20519829e-01 5.92615664e-01
-4.48244721e-01 8.41766357e-01 1.05723832e-02 6.81962609e-01
5.56975603e-01 -1.44022262e+00 3.57816406e-02 1.32338977e+00
-9.11993533e-02 4.36922520e-01 7.70571053e-01 -5.52129567e-01
-9.76766884e-01 -1.09301686e+00 6.31198406e-01 1.38438538e-01
3.83785933e-01 1.45066842e-01 -6.46919489e-01 7.77545810e-01
1.01568413e+00 2.74029881e-01 8.95512581e-01 -7.78843284e-01
3.50899965e-01 3.69394481e-01 -1.84068108e+00 3.31618935e-01
1.59328893e-01 1.00180350e-01 -5.53081989e-01 6.11729503e-01
5.25469303e-01 -1.02922440e+00 -1.12037730e+00 7.52421856e-01
-3.57722640e-02 -7.35033929e-01 1.10362613e+00 -4.39612389e-01
3.22509587e-01 -2.57097602e-01 3.43984753e-01 -1.39300931e+00
-1.33117167e-02 -5.63621581e-01 -4.11228351e-02 6.66709960e-01
2.88872957e-01 -4.67846185e-01 8.41019630e-01 4.18125212e-01
-3.43718916e-01 -8.67646873e-01 -9.69912887e-01 -3.72334182e-01
2.87418008e-01 -9.47300196e-02 2.45785922e-01 9.67703640e-01
-2.56768316e-01 -1.40444055e-01 -2.06584290e-01 1.13999359e-01
8.64902914e-01 -1.07495956e-01 3.02629381e-01 -8.17793846e-01
-1.19113922e-01 -2.64917880e-01 -5.45619011e-01 -5.52805603e-01
-9.40536410e-02 -1.12524986e+00 7.26771355e-02 -1.60319614e+00
1.40374631e-01 -1.83067545e-01 -5.66436648e-01 5.10696590e-01
-2.60959685e-01 8.00457120e-01 2.34826893e-01 2.39273869e-02
-3.34549755e-01 -9.58295316e-02 1.45982563e+00 -3.83201778e-01
2.26980463e-01 1.86759710e-01 -3.23968470e-01 9.66563523e-01
9.55861449e-01 -2.04791784e-01 7.30126351e-03 -1.84947476e-01
-4.93346930e-01 4.21349853e-02 2.67423004e-01 -1.09652543e+00
3.67092252e-01 3.28288555e-01 5.36505044e-01 -5.94420552e-01
-8.17642957e-02 -1.14259803e+00 5.01822010e-02 1.07395482e+00
-2.78064251e-01 1.26159964e-02 3.12126875e-01 2.59853870e-01
-3.29220712e-01 -4.46602166e-01 1.36433780e+00 -5.58110297e-01
-2.93303043e-01 5.29767811e-01 -3.32720488e-01 -2.44432852e-01
1.49549830e+00 1.82729334e-01 1.93351254e-01 -3.68047617e-02
-1.19831359e+00 2.41135538e-01 1.47470385e-01 -1.45883620e-01
5.18048465e-01 -1.00479639e+00 -8.94171119e-01 1.26913846e-01
-5.09212017e-01 2.78898478e-01 5.78299761e-01 1.44587016e+00
-1.28463376e+00 4.00106490e-01 -1.21894635e-01 -4.54279602e-01
-1.21312618e+00 7.47001290e-01 5.23919642e-01 -8.01376820e-01
-8.36981773e-01 7.50752628e-01 -1.65789396e-01 -3.58521938e-01
2.41233498e-01 -2.61764616e-01 -4.64627683e-01 -2.17909098e-01
4.35101986e-01 1.22678973e-01 3.03468198e-01 -6.23739362e-01
-1.45338669e-01 2.01700449e-01 -2.96392143e-01 1.76857561e-02
1.19542611e+00 2.09755800e-03 -2.74765998e-01 -3.29334959e-02
9.63876307e-01 -9.33733359e-02 -8.20232272e-01 -2.13585541e-01
1.52105480e-01 -3.34715605e-01 2.57421494e-01 -1.24042308e+00
-1.68905783e+00 1.02038431e+00 1.12589633e+00 4.39801104e-02
1.29358935e+00 -3.50482583e-01 1.17248023e+00 -1.48877159e-01
1.05050199e-01 -8.66072476e-01 -3.28144878e-01 3.13123494e-01
6.90492690e-01 -1.37773800e+00 -5.29363789e-02 -5.34489989e-01
-4.75293875e-01 1.27459908e+00 5.16226947e-01 -4.22692209e-01
9.01171148e-01 3.75804126e-01 3.95159394e-01 1.81135256e-02
1.34868279e-01 -1.20358698e-01 4.33561742e-01 4.16046113e-01
4.96433556e-01 1.74761042e-01 -1.10125756e+00 7.33374953e-01
1.37625307e-01 2.61499822e-01 5.60732305e-01 8.61591935e-01
3.97651866e-02 -9.90214288e-01 -1.84302151e-01 3.90952498e-01
-1.38804948e+00 -3.06660861e-01 -1.46761864e-01 8.33033741e-01
1.61355123e-01 6.28305256e-01 -3.45949054e-01 -6.82756118e-03
1.15505010e-02 6.65126145e-02 2.86763728e-01 -3.52474064e-01
-1.01204538e+00 2.38012671e-01 -4.13466454e-01 -3.56085360e-01
-5.68465829e-01 -2.94865459e-01 -1.38699687e+00 5.48497811e-02
-3.92299175e-01 3.46356362e-01 8.67725432e-01 7.92218268e-01
1.26524597e-01 9.06695724e-01 4.77275848e-01 -7.60091484e-01
-4.87678528e-01 -1.21401334e+00 -3.16980809e-01 6.73477411e-01
1.99746385e-01 -2.76035905e-01 -1.73990220e-01 2.12894380e-01] | [14.530289649963379, -2.6272568702697754] |
fc704d58-1165-45d0-86a3-fdbb72529731 | hope-speech-detection-in-under-resourced | 2108.04616 | null | https://arxiv.org/abs/2108.04616v2 | https://arxiv.org/pdf/2108.04616v2.pdf | Hope Speech detection in under-resourced Kannada language | Numerous methods have been developed to monitor the spread of negativity in modern years by eliminating vulgar, offensive, and fierce comments from social media platforms. However, there are relatively lesser amounts of study that converges on embracing positivity, reinforcing supportive and reassuring content in online forums. Consequently, we propose creating an English-Kannada Hope speech dataset, KanHope and comparing several experiments to benchmark the dataset. The dataset consists of 6,176 user-generated comments in code mixed Kannada scraped from YouTube and manually annotated as bearing hope speech or Not-hope speech. In addition, we introduce DC-BERT4HOPE, a dual-channel model that uses the English translation of KanHope for additional training to promote hope speech detection. The approach achieves a weighted F1-score of 0.756, bettering other models. Henceforth, KanHope aims to instigate research in Kannada while broadly promoting researchers to take a pragmatic approach towards online content that encourages, positive, and supportive. | ['Bharathi Raja Chakravarthi', 'Prabakaran Chandran', 'Kingston Pal Thamburaj', 'Anbukkarasi Sampath', 'Ruba Priyadharshini', 'Adeep Hande'] | 2021-08-10 | null | null | null | null | ['hope-speech-detection'] | ['natural-language-processing'] | [-3.21422726e-01 4.72788990e-01 -4.76598024e-01 -3.89097668e-02
-5.36759675e-01 -4.25804645e-01 7.14181542e-01 2.83084750e-01
-2.00652272e-01 3.68286282e-01 8.95597100e-01 -2.59041190e-01
3.63998234e-01 -3.36969137e-01 -8.42517838e-02 -3.58434796e-01
2.65050530e-01 -4.69873697e-01 -1.90755174e-01 -8.15734625e-01
5.86579740e-01 -2.84377486e-01 -1.10149240e+00 3.84988397e-01
7.94759035e-01 2.96064407e-01 -4.46672998e-02 7.56347954e-01
4.68207002e-02 1.66907656e+00 -6.14123881e-01 -8.04379702e-01
-3.75520587e-01 -7.11611807e-01 -9.44177628e-01 -7.00001791e-02
2.17135493e-02 -9.21685547e-02 9.38620493e-02 1.02405405e+00
6.28599346e-01 -1.10116504e-01 3.34754676e-01 -1.14098167e+00
-1.16633761e+00 1.38553238e+00 -5.10376096e-01 2.27762669e-01
6.62495077e-01 1.67844057e-01 1.25129986e+00 -8.72346997e-01
9.23211217e-01 1.23282993e+00 7.21513867e-01 5.85536242e-01
-7.98202395e-01 -5.62598884e-01 -4.00670588e-01 -4.36791545e-03
-8.51908922e-01 -6.24583066e-01 1.08567536e+00 -6.33534193e-01
9.79270935e-01 4.54320878e-01 9.93271410e-01 1.77061725e+00
-2.24646311e-02 7.78109312e-01 1.02271855e+00 -3.76447558e-01
9.28871334e-02 4.13470358e-01 1.16642909e-02 5.40801287e-01
-4.78935868e-01 -5.93333781e-01 -1.01447821e+00 -4.12615389e-01
1.43946875e-02 -4.85806853e-01 -3.48435909e-01 3.08329314e-01
-1.22356200e+00 1.19013560e+00 3.06263149e-01 4.33310598e-01
-5.76706052e-01 -3.58049959e-01 8.43264937e-01 5.30342638e-01
7.83286393e-01 3.91431689e-01 -6.22486584e-02 -9.14531887e-01
-5.30143499e-01 8.65133405e-02 1.02562761e+00 5.62151611e-01
3.84556949e-01 -3.36185358e-02 -1.01153672e-01 1.42595267e+00
5.22959411e-01 6.84162676e-01 5.92168272e-01 -9.60849702e-01
3.00413847e-01 7.03352213e-01 -2.00242341e-01 -1.27181292e+00
-2.81988442e-01 -2.63972580e-01 -5.18889844e-01 -1.72933981e-01
2.04320159e-02 -5.46991229e-01 -1.67567711e-02 1.52259612e+00
1.57060802e-01 -5.83042026e-01 1.45654336e-01 6.91278934e-01
8.85868311e-01 9.31834459e-01 1.07843570e-01 -5.07084310e-01
9.33735132e-01 -1.02867186e+00 -1.04413450e+00 -2.33913749e-01
1.09878361e+00 -1.25108874e+00 1.64510071e+00 5.84030747e-01
-9.99583364e-01 1.86315596e-01 -1.09357047e+00 -3.39007497e-01
-5.37376739e-02 -7.79972076e-02 2.81952053e-01 9.75665450e-01
-1.04222620e+00 1.89290613e-01 -3.62897187e-01 -5.80140710e-01
4.10001397e-01 -3.87519717e-01 -3.43575716e-01 3.64898831e-01
-1.14188504e+00 7.04370260e-01 -2.34426588e-01 -8.54966268e-02
-6.41711950e-01 -3.99216354e-01 -7.11885452e-01 -1.05680846e-01
1.05214633e-01 -2.13223621e-02 1.35291195e+00 -1.41357863e+00
-1.57416618e+00 1.06354344e+00 8.67489651e-02 -2.55172580e-01
2.10528567e-01 -3.18005055e-01 -4.01446253e-01 -6.49837106e-02
2.95172453e-01 5.63013256e-01 8.47160578e-01 -7.69540548e-01
-1.39299139e-01 1.62763763e-02 7.33142868e-02 3.42742428e-02
-1.20519376e+00 5.88239193e-01 3.40850562e-01 -6.70165300e-01
-3.59992325e-01 -8.09071064e-01 2.64840871e-01 -3.72949868e-01
-6.67828202e-01 -5.24381578e-01 9.08680916e-01 -7.60665536e-01
1.67884374e+00 -2.25769281e+00 -1.99094206e-01 3.26371938e-02
6.17848575e-01 1.37462139e-01 -1.83865473e-01 9.25461054e-01
3.76491584e-02 5.68097293e-01 1.54698603e-02 -1.57006949e-01
-2.53782049e-02 -7.04996213e-02 -3.85510661e-02 5.80484271e-01
2.19332889e-01 6.23043120e-01 -1.14384186e+00 -2.58553296e-01
-3.45311850e-01 5.00899076e-01 -9.26335752e-01 9.46392212e-03
6.93759769e-02 2.21471582e-02 -2.39021719e-01 6.74600244e-01
3.53424132e-01 -4.16830480e-01 3.38207603e-01 6.18808448e-01
-6.61480546e-01 8.00233364e-01 -1.93274602e-01 1.12619865e+00
-3.54145437e-01 1.16778934e+00 2.23561361e-01 -3.02854449e-01
1.33355963e+00 3.19072396e-01 1.67635843e-01 -8.88195932e-01
4.57506388e-01 2.43689641e-01 1.59871489e-01 -7.64153302e-01
8.25957060e-01 -1.99397311e-01 7.21561816e-03 4.77136672e-01
-3.60588521e-01 -1.58035830e-01 1.26135647e-01 7.24150181e-01
1.19211936e+00 -3.93747032e-01 5.80442727e-01 -4.10375357e-01
5.55606484e-01 -1.32182568e-01 2.64485270e-01 3.74623209e-01
-8.17540467e-01 3.54966879e-01 1.26094329e+00 1.43819585e-01
-9.37432587e-01 -5.11757851e-01 7.08905831e-02 1.48130548e+00
-4.62289691e-01 -8.76766384e-01 -8.25187266e-01 -4.78689134e-01
-5.73749423e-01 7.51439869e-01 -6.30030751e-01 2.11692043e-02
-1.92129955e-01 -5.06153405e-01 6.69434905e-01 -2.83455968e-01
4.45033878e-01 -1.20611048e+00 -3.62603903e-01 1.55971706e-01
-8.30216467e-01 -7.77586818e-01 -3.32324147e-01 -2.09110212e-02
-3.14938635e-01 -9.18152094e-01 -7.39615738e-01 -8.42481971e-01
-6.52262121e-02 3.05908829e-01 9.91397023e-01 3.70534927e-01
4.17880744e-01 1.40547410e-01 -9.68391478e-01 -3.97219300e-01
-9.02085721e-01 3.48142147e-01 -2.75882363e-01 -1.64535671e-01
3.63752246e-01 -4.83476400e-01 -7.13731349e-02 6.73407631e-05
-5.86530387e-01 2.43732706e-02 2.05184802e-01 9.04829204e-01
-4.12876993e-01 -7.49766886e-01 1.19120538e+00 -9.39634323e-01
1.10084546e+00 -1.13893914e+00 3.04278731e-01 -2.88930267e-01
-4.59832847e-01 -6.47177935e-01 5.32851040e-01 -4.81338203e-01
-1.25757563e+00 -4.05494481e-01 -6.64337814e-01 2.49726355e-01
3.35978866e-01 8.16361785e-01 5.10141253e-01 2.59069145e-01
1.27045858e+00 -2.70345449e-01 1.60927638e-01 -7.39839226e-02
2.28675410e-01 1.34839392e+00 3.00870985e-01 -6.18479066e-02
2.84878373e-01 3.23819071e-01 -8.69731486e-01 -1.38426661e+00
-9.21689212e-01 -7.38871157e-01 6.50760308e-02 -6.74161077e-01
6.46845937e-01 -1.07569075e+00 -6.58793569e-01 5.38393497e-01
-1.14054990e+00 -6.23178244e-01 1.40984982e-01 2.32390851e-01
-2.53665537e-01 5.12983024e-01 -9.68050718e-01 -1.17125869e+00
-5.11706173e-01 -5.10060787e-01 4.61501479e-01 1.21498838e-01
-8.97925794e-01 -9.83093917e-01 5.20051479e-01 6.95210755e-01
4.23688769e-01 1.13073811e-01 7.01989114e-01 -6.76979184e-01
4.02273923e-01 1.48891926e-01 -1.39758727e-02 5.67594707e-01
-6.77403063e-02 5.96441269e-01 -1.16387200e+00 2.91686147e-01
4.16967034e-01 -1.03579676e+00 3.06239754e-01 -1.10492013e-01
2.44616643e-01 -8.71248364e-01 3.94438326e-01 -4.00099099e-01
7.56927192e-01 -1.84079558e-01 8.85337174e-01 6.67356670e-01
3.54423791e-01 6.66621745e-01 5.90908587e-01 1.16166997e+00
6.90316319e-01 -1.02674648e-01 5.49404442e-01 1.32758602e-01
3.60443704e-02 -2.55016983e-01 1.08977664e+00 1.60654390e+00
3.20192307e-01 -3.68570358e-01 -1.16761768e+00 8.75211000e-01
-1.61547542e+00 -1.11838543e+00 -1.00000489e+00 1.39247859e+00
1.18587661e+00 2.02635005e-01 3.75213951e-01 3.22122872e-01
5.78868687e-01 6.18729949e-01 2.06784487e-01 -8.74220192e-01
-4.47616726e-01 -2.98047155e-01 -2.27941856e-01 5.81129313e-01
-8.28084290e-01 6.19828999e-01 5.64141750e+00 7.43181944e-01
-1.34937108e+00 2.89036930e-01 8.09036255e-01 -1.03556275e-01
-5.65664589e-01 -2.00211391e-01 -1.44327804e-01 2.29663044e-01
1.11336958e+00 -2.57553279e-01 3.79889667e-01 1.13039088e+00
8.63327920e-01 -2.77702451e-01 -5.13132513e-01 8.78780425e-01
4.27302927e-01 -1.09960699e+00 -6.78239822e-01 -2.96154648e-01
1.12830591e+00 5.45261919e-01 3.03547174e-01 4.43615943e-01
2.44943816e-02 -6.08841240e-01 1.07286489e+00 9.98996273e-02
2.37835929e-01 -4.36973274e-01 7.40845442e-01 7.25624382e-01
-2.33266443e-01 -9.86330286e-02 -7.83630386e-02 -8.49463642e-01
-2.15703296e-03 6.95877194e-01 -1.01196229e+00 -2.11370736e-01
7.00331926e-01 1.27918303e+00 -6.58323288e-01 5.09596407e-01
-4.87015396e-01 1.17865944e+00 -7.62550533e-02 -8.85117471e-01
5.46632171e-01 1.45354107e-01 5.79370141e-01 1.42793584e+00
1.03863053e-01 -2.16053501e-01 -2.82216012e-01 7.93604553e-01
-5.56909561e-01 6.54425323e-01 -5.85641801e-01 -6.67804897e-01
4.06523377e-01 1.65391266e+00 -5.50443530e-01 -1.56992033e-01
-4.43422109e-01 6.65534914e-01 3.99345040e-01 2.41951346e-02
-8.19212317e-01 -9.61209536e-02 1.79426521e-01 1.32269874e-01
-2.19035596e-01 5.59086613e-02 -6.02050483e-01 -9.41530228e-01
-1.26549229e-01 -1.10085773e+00 -6.81768879e-02 -8.99674535e-01
-1.42176676e+00 4.59968030e-01 -6.43298030e-01 -7.09809959e-01
-5.14118792e-03 -1.79154783e-01 -7.32373655e-01 5.96927106e-01
-1.28655827e+00 -8.41094971e-01 -5.45160845e-02 3.59726548e-01
6.76475406e-01 1.68124169e-01 4.46981072e-01 3.60710084e-01
-4.91197377e-01 2.03035593e-01 -1.83768034e-01 -1.27498654e-03
9.93228853e-01 -9.30142581e-01 8.87379572e-02 6.36581719e-01
-3.24180901e-01 7.45596170e-01 7.98676431e-01 -7.33850837e-01
-1.03931534e+00 -6.23317480e-01 1.80403352e+00 -5.56494713e-01
1.41769695e+00 -3.01238626e-01 -8.06820095e-01 4.73039389e-01
9.05668378e-01 -6.99146509e-01 8.86463881e-01 5.20927787e-01
-4.92990911e-01 6.21515453e-01 -8.88453841e-01 8.17700982e-01
7.88053513e-01 -8.12413990e-01 -4.50578332e-01 5.81374884e-01
4.77783591e-01 3.94902267e-02 -7.88313568e-01 -4.74992156e-01
3.05357605e-01 -1.35344291e+00 2.23749876e-01 -1.11750454e-01
1.25591838e+00 3.09472024e-01 1.54435728e-02 -1.25933826e+00
-1.17983028e-01 -1.19359529e+00 3.50139767e-01 1.63422012e+00
6.84411764e-01 -3.80250156e-01 5.09067297e-01 1.95505634e-01
-4.29675758e-01 -5.02817333e-01 -6.71439588e-01 -1.68049365e-01
2.50299156e-01 -5.25176227e-01 -1.74169883e-01 1.59530437e+00
1.00256026e+00 9.51596260e-01 -6.73171282e-01 -5.10420144e-01
-1.13354929e-01 -7.61772692e-01 7.87719369e-01 -8.90516460e-01
1.78151891e-01 -5.45556307e-01 1.75211772e-01 -6.63603306e-01
8.08934346e-02 -8.86639416e-01 9.82750654e-02 -1.09784687e+00
5.95395625e-01 -1.11101858e-01 4.47331637e-01 6.77858829e-01
6.03714809e-02 3.17520946e-01 1.40518218e-01 2.49671265e-01
-7.48468161e-01 6.67496443e-01 1.26857185e+00 3.78303975e-02
-3.29669565e-01 -7.13432252e-01 -1.16866291e+00 7.63087928e-01
9.63580489e-01 -5.25608778e-01 -3.30643564e-01 1.81322426e-01
1.16348124e+00 -6.50092363e-02 3.47770721e-01 -5.66183686e-01
-1.26933344e-02 -1.04471244e-01 -5.24722874e-01 -2.60379732e-01
1.75589025e-02 -1.48201538e-02 -2.96186566e-01 2.72363752e-01
-6.80212677e-01 1.23690136e-01 -3.23378712e-01 1.47964954e-01
-2.35947043e-01 -4.31060314e-01 7.77877510e-01 1.05511911e-01
-2.28950575e-01 -5.16288638e-01 -1.03216505e+00 4.69722062e-01
5.60528338e-01 1.52139217e-01 -8.37790251e-01 -9.03289557e-01
-3.40284586e-01 -4.73860055e-02 4.49922949e-01 6.05846345e-01
4.95704800e-01 -1.01702690e+00 -1.02701998e+00 -2.28012413e-01
2.58232176e-01 -8.51080596e-01 3.98597419e-01 1.30997241e+00
-5.03176570e-01 -6.43070564e-02 -4.49360423e-02 -3.18949282e-01
-1.27105904e+00 2.41234004e-01 -1.47338405e-01 9.18393657e-02
-4.97461379e-01 9.48098123e-01 -3.14524770e-01 -6.40502691e-01
-2.77550489e-01 2.13699471e-02 -5.44737339e-01 4.57838863e-01
5.95805585e-01 7.06962049e-01 -3.67908292e-02 -8.88999283e-01
-9.52857137e-02 -4.76022243e-01 1.87575966e-01 -1.69369847e-01
1.23607147e+00 -4.01075304e-01 -6.17509902e-01 8.01226735e-01
1.32571292e+00 6.48534596e-01 -5.92243433e-01 2.51697153e-01
1.46946371e-01 -3.51794064e-01 -2.54503302e-02 -9.38130498e-01
-5.32638431e-01 6.12381399e-01 1.89717978e-01 8.41455936e-01
6.64561987e-01 -2.88014375e-02 7.61227429e-01 4.04937804e-01
-4.28403497e-01 -1.48870182e+00 5.53480208e-01 1.00274086e+00
1.14135945e+00 -1.27776372e+00 -4.89470452e-01 -4.29897845e-01
-1.04260528e+00 1.14439285e+00 3.30451310e-01 1.13953345e-01
4.87903982e-01 9.50212777e-02 4.99329776e-01 -3.42907846e-01
-1.13212740e+00 3.18995386e-01 -3.57660770e-01 4.82352853e-01
1.22851026e+00 6.77362382e-02 -9.43563104e-01 7.23893225e-01
-6.08702540e-01 3.41763087e-02 1.35069466e+00 5.78036249e-01
-6.14733279e-01 -6.27752483e-01 -3.13818336e-01 2.04210997e-01
-9.02694762e-01 -4.56513494e-01 -9.77420032e-01 5.90918064e-01
-2.12418735e-01 1.51123846e+00 -3.10942501e-01 -5.03955781e-01
-1.21944375e-01 6.60151709e-03 -4.69494253e-01 -4.67516214e-01
-1.26424706e+00 4.44172233e-01 9.42285061e-01 -1.12331390e-01
-7.12232888e-01 -7.68792272e-01 -1.07525957e+00 -8.32648575e-01
-3.14350784e-01 1.33749142e-01 5.11022687e-01 6.50305748e-01
7.40465596e-02 2.82545328e-01 8.43653083e-01 -3.59717160e-01
-3.79959077e-01 -1.18209445e+00 -3.90194833e-01 2.59978056e-01
2.04417199e-01 -1.34262636e-01 -6.26144350e-01 -8.73706937e-02] | [8.972841262817383, 10.673158645629883] |
c60391b8-0d12-4edb-a61a-8345d02ace5f | same-author-or-just-same-topic-towards-topic | null | null | https://openreview.net/forum?id=8sLibFf9Cr1 | https://openreview.net/pdf?id=8sLibFf9Cr1 | Same Author or Just Same Topic? Towards Topic-Independent Style Representations | Style is an integral component of language. Recent advances in the development of style representations have increasingly used training objectives from authorship verification (AV): Do two texts have the same author? The assumption underlying the AV training task (same author approximates same writing style) enables self-supervised and, thus, extensive training. However, AV usually does not or only on a coarse-grained level control for topic. The resulting representations might therefore also encode topical information instead of style alone. We introduce a variation of the AV training task that controls for topic using conversation, domain or no topic control as a topic proxy. To evaluate whether trained representations prefer style over topic information, we propose an original variation to the recent STEL framework. We find that representations trained by controlling for conversation are better than representations trained with domain or no topic control at representing style independent from topic. | ['Anonymous'] | 2022-01-16 | null | null | null | acl-arr-january-2022-1 | ['authorship-verification'] | ['natural-language-processing'] | [ 1.72165185e-01 1.74751833e-01 -3.16587418e-01 -6.44850135e-01
-3.71453613e-01 -9.06641543e-01 1.36157584e+00 2.31362402e-01
-3.07398766e-01 6.66352510e-01 6.74997747e-01 -2.64021128e-01
1.79129004e-01 -6.74413919e-01 -4.39048201e-01 -3.11332226e-01
3.98130208e-01 6.92428827e-01 -2.85221428e-01 -5.93255647e-02
4.48434591e-01 2.44089946e-01 -1.12586236e+00 1.90604836e-01
7.43738413e-01 6.81309164e-01 -8.36555362e-02 6.73884809e-01
-6.60966873e-01 6.44431233e-01 -1.16696453e+00 -4.14812267e-01
-5.25185131e-02 -6.89984083e-01 -9.53028321e-01 3.21407348e-01
9.65648830e-01 -9.10004005e-02 -1.68609619e-01 7.99908936e-01
3.98854055e-02 1.07587926e-01 1.05564642e+00 -1.01177311e+00
-1.09026968e+00 7.03853309e-01 -3.89145166e-01 1.93047956e-01
3.76834273e-01 -1.35663912e-01 1.24024725e+00 -5.39766788e-01
1.00646937e+00 1.62865436e+00 5.88736713e-01 6.89275563e-01
-1.84416556e+00 -4.15822566e-01 5.33340573e-01 -3.56618971e-01
-9.61593211e-01 -3.62564057e-01 1.07487166e+00 -5.70987344e-01
4.57663357e-01 1.59079030e-01 3.94544691e-01 1.91626227e+00
2.31216744e-01 8.36806297e-01 1.61322749e+00 -5.07438064e-01
1.05768614e-01 5.56549072e-01 5.76774836e-01 4.31213289e-01
3.29940110e-01 -2.09392682e-01 -7.47017562e-01 -3.79105061e-01
6.81205094e-01 -2.06971541e-01 -2.33388931e-01 -1.18623503e-01
-1.46619415e+00 1.01807272e+00 1.26790434e-01 4.93918359e-01
4.66319285e-02 8.49136859e-02 5.30323327e-01 8.64285231e-01
6.87358439e-01 1.16729808e+00 -3.81517529e-01 -2.60196418e-01
-1.35572255e+00 5.10925055e-01 1.23795044e+00 9.23316360e-01
8.11235487e-01 2.45967805e-01 -6.81596816e-01 9.26242530e-01
2.38634855e-01 4.47587609e-01 6.53613985e-01 -1.07772195e+00
1.37608573e-01 5.28911352e-01 1.31609604e-01 -8.78106177e-01
-6.31393865e-02 -3.71596485e-01 -5.89868724e-01 1.83744624e-01
7.31132627e-01 -8.50740150e-02 -9.37739015e-01 1.99996936e+00
-3.52106512e-01 -3.45710397e-01 -1.91154167e-01 5.39205730e-01
6.73959911e-01 4.83442724e-01 2.62356967e-01 -1.88584790e-01
1.51026821e+00 -7.24521995e-01 -8.61148894e-01 -3.24538559e-01
6.01944387e-01 -9.31781769e-01 1.41806400e+00 5.73482156e-01
-9.46369827e-01 -6.38172388e-01 -9.32163239e-01 -2.74419069e-01
-4.63222176e-01 3.27505797e-01 5.51867723e-01 7.14372694e-01
-1.15729499e+00 7.99690723e-01 -3.31361651e-01 -5.43934047e-01
1.90113753e-01 -1.15685217e-01 -3.34772527e-01 2.11662307e-01
-1.21532071e+00 1.05391967e+00 -1.74845546e-01 -6.11088455e-01
-5.50965250e-01 -8.85560930e-01 -9.00050521e-01 -7.58483782e-02
9.15218890e-02 -5.63618362e-01 1.30559695e+00 -1.56595325e+00
-1.69139814e+00 1.35626447e+00 -3.67901176e-01 -2.57918686e-01
5.41439891e-01 -4.42020208e-01 -3.78155649e-01 -5.76903112e-02
3.22419137e-01 5.41567922e-01 1.20018470e+00 -1.29833579e+00
-7.18338788e-02 -2.69530237e-01 -6.78639486e-03 -3.93802337e-02
-5.24918139e-01 3.49146612e-02 -1.12543158e-01 -1.19977915e+00
-1.87599540e-01 -9.30530369e-01 4.04193401e-01 2.81370342e-01
-4.07896101e-01 -6.28974617e-01 9.87512290e-01 -6.75145686e-01
1.26753712e+00 -2.07232428e+00 3.14865649e-01 1.59265816e-01
4.22488272e-01 -2.24538863e-01 -1.73325032e-01 3.35866004e-01
1.25014141e-01 5.32508075e-01 -6.02074564e-02 -8.28093588e-01
2.17021212e-01 2.47601599e-01 -3.65238488e-01 3.64753723e-01
3.39508027e-01 5.25266171e-01 -6.69003606e-01 -4.69572663e-01
-2.44363263e-01 3.42913777e-01 -5.66192806e-01 2.65018672e-01
-4.29617286e-01 3.36187184e-01 -1.67018846e-01 3.18434924e-01
3.07945281e-01 -3.73769075e-01 2.98002958e-01 1.00005075e-01
-1.22863606e-01 6.24199986e-01 -9.25768554e-01 1.83371079e+00
-5.46560466e-01 1.29011548e+00 1.01379713e-03 -5.35277903e-01
1.31935978e+00 2.99622059e-01 2.07793228e-02 -2.99407572e-01
4.86172475e-02 2.60519590e-02 5.05978726e-02 9.03045312e-02
6.93775833e-01 -2.29602352e-01 -4.09326464e-01 8.69769037e-01
2.94829518e-01 -2.93937206e-01 1.38032928e-01 3.62949699e-01
9.54010785e-01 2.61363119e-01 1.81264266e-01 -7.10357487e-01
4.55413997e-01 -1.94676176e-01 5.97361863e-01 1.01460779e+00
-1.65803090e-01 7.17342198e-01 1.01658309e+00 -2.24169433e-01
-1.24875319e+00 -8.91889274e-01 -3.18923652e-01 1.31997466e+00
-2.56485820e-01 -5.57253778e-01 -6.57316506e-01 -8.78817141e-01
2.36597911e-01 9.15956020e-01 -1.05005515e+00 8.76344144e-02
-4.93269026e-01 -4.41058949e-02 6.58710301e-01 4.31157738e-01
1.95550025e-01 -1.07921433e+00 -2.58451998e-01 4.20377366e-02
1.53667092e-01 -7.95365334e-01 -6.51067853e-01 6.97498992e-02
-7.99870670e-01 -6.64772868e-01 -1.06912518e+00 -3.76237601e-01
3.63099694e-01 -1.82372227e-01 1.67157543e+00 1.19632535e-01
2.81240970e-01 5.63013315e-01 -2.95082241e-01 -4.72462684e-01
-7.36607432e-01 5.11497736e-01 -3.11196856e-02 -8.49358365e-02
4.32003260e-01 -4.73059505e-01 -2.74450660e-01 7.35485107e-02
-6.94110215e-01 -1.86783373e-01 4.35695231e-01 1.06605470e+00
-1.05264887e-01 -6.11787498e-01 4.34836626e-01 -1.53793144e+00
9.61978912e-01 -4.31002945e-01 -8.67247581e-02 1.11012354e-01
-9.25413787e-01 4.98116255e-01 4.85896587e-01 -7.40465879e-01
-1.07587242e+00 -5.11142552e-01 2.34139413e-01 -5.43741465e-01
-5.98645031e-01 3.22330564e-01 -1.65030181e-01 2.95343369e-01
6.74999475e-01 2.18173955e-02 2.41260469e-01 -6.67335749e-01
1.80521309e-01 5.22149503e-01 1.41052723e-01 -1.01448226e+00
7.03055084e-01 2.03107357e-01 -2.62832761e-01 -8.75141323e-01
-7.99763739e-01 -2.15681478e-01 -5.98985910e-01 1.19474269e-02
6.73655510e-01 -7.59190917e-01 -3.45703870e-01 2.18162581e-01
-1.33259511e+00 -3.88093412e-01 -3.54332000e-01 2.01000497e-01
-5.08682013e-01 3.63903224e-01 -5.93671441e-01 -6.78589523e-01
-1.80369794e-01 -8.70410264e-01 1.23515773e+00 -4.27428596e-02
-1.19296956e+00 -1.40070724e+00 1.69948906e-01 -1.33327216e-01
5.91345847e-01 2.68346548e-01 1.15595877e+00 -7.69102395e-01
-4.66220714e-02 -5.29267378e-02 -1.29915267e-01 2.01140806e-01
2.61557996e-01 1.55844137e-01 -1.33635998e+00 -2.27124378e-01
9.29559395e-03 -3.19626063e-01 9.80229557e-01 -3.06690801e-02
1.01945591e+00 -4.17229176e-01 -8.93784910e-02 5.26096106e-01
1.14438856e+00 -1.89577863e-01 3.81701559e-01 4.24137086e-01
7.43264079e-01 7.67252386e-01 5.69467284e-02 3.69728386e-01
9.18740034e-02 6.97640479e-01 -4.30842727e-01 5.15364595e-02
-2.91007698e-01 -2.46707097e-01 3.96461368e-01 4.52029377e-01
1.06564239e-01 -3.83611470e-01 -7.88537443e-01 4.51028645e-01
-1.24962211e+00 -1.07506192e+00 -4.95061055e-02 1.94143450e+00
1.19338131e+00 3.37366134e-01 3.32567006e-01 3.35787833e-01
4.35444206e-01 4.55773592e-01 -3.27240825e-01 -8.61009896e-01
-2.27167100e-01 5.99622391e-02 1.18255310e-01 6.45498693e-01
-1.06362128e+00 9.42414463e-01 6.79519367e+00 3.43992382e-01
-1.20528638e+00 -5.36129577e-03 5.42026281e-01 1.09450348e-01
-6.42529249e-01 -5.25553599e-02 -7.97319531e-01 5.85538626e-01
9.32443142e-01 -4.49566655e-02 2.66829312e-01 9.02456820e-01
8.08976889e-02 9.91474986e-02 -1.69301605e+00 6.99413836e-01
2.90947765e-01 -1.11577034e+00 3.52953166e-01 2.07176432e-01
6.80799782e-01 -6.35257661e-01 1.70741543e-01 3.64340067e-01
3.90155584e-01 -1.02272463e+00 9.67799783e-01 7.15920508e-01
8.45881283e-01 -3.54629695e-01 4.23445165e-01 -2.17854381e-02
-6.68756962e-01 2.26648360e-01 -2.02146471e-01 -1.43076882e-01
-2.18152851e-01 5.06358325e-01 -5.22588134e-01 1.06772602e-01
3.93378466e-01 1.00451517e+00 -8.41564000e-01 3.94428879e-01
-1.88167974e-01 8.63895833e-01 1.81829482e-02 7.40034599e-03
1.36447906e-01 -6.20226935e-02 7.71171510e-01 1.56717849e+00
-1.77784581e-02 -4.80677575e-01 1.62863865e-01 1.19904375e+00
-1.68297574e-01 -1.22811747e-04 -8.90175700e-01 -3.13291460e-01
4.31996882e-01 9.20214713e-01 -5.85213780e-01 -5.33157110e-01
-4.69873697e-01 1.19410169e+00 3.73327792e-01 4.94802654e-01
-6.28230348e-02 -2.13636503e-01 1.09612036e+00 1.77387342e-01
4.56838161e-02 -3.23735178e-01 -9.02896523e-01 -1.25236762e+00
-1.72145426e-01 -1.04625571e+00 2.49343306e-01 -5.78134358e-01
-1.68207633e+00 5.51324487e-01 -1.30017474e-01 -8.66002142e-01
-3.58837932e-01 -7.75493443e-01 -9.84964192e-01 1.31000149e+00
-1.31292105e+00 -1.15609932e+00 -2.38713190e-01 3.51383150e-01
8.69838476e-01 -4.43460882e-01 1.12205887e+00 -3.69523019e-01
-3.47933918e-01 7.85260856e-01 -1.51916649e-02 3.50591332e-01
1.23212087e+00 -1.91647446e+00 2.13941380e-01 3.69821131e-01
6.75984472e-02 1.03861046e+00 9.82384324e-01 -6.98078454e-01
-1.11746633e+00 -6.82527483e-01 1.29072654e+00 -9.62359905e-01
7.61204720e-01 -6.13307297e-01 -1.05031681e+00 6.98095858e-01
6.11715615e-01 -5.79459548e-01 7.97256827e-01 7.69168735e-01
-8.19386899e-01 1.67176396e-01 -9.62352574e-01 5.02664804e-01
7.83213794e-01 -1.03126478e+00 -1.08717322e+00 1.33588508e-01
5.83502471e-01 1.24545127e-01 -9.91852880e-01 -2.51029700e-01
5.63521266e-01 -6.69028461e-01 5.27843595e-01 -6.40498698e-01
7.78532982e-01 1.63738474e-01 -2.61877198e-02 -1.42598152e+00
-4.95265663e-01 -5.96804798e-01 -1.30724818e-01 1.76319194e+00
3.63852531e-01 -4.04600471e-01 8.45930517e-01 6.52427018e-01
1.77644361e-02 -3.42684329e-01 -4.22604412e-01 -7.91557014e-01
7.21957445e-01 2.42872462e-02 4.74327654e-01 1.25108683e+00
-1.95961017e-02 6.22028232e-01 -1.32822737e-01 -3.33906084e-01
4.72016871e-01 1.62737191e-01 6.20556414e-01 -1.94887984e+00
-3.40792477e-01 -9.93465006e-01 -1.59934238e-01 -6.80733800e-01
6.82383478e-01 -7.66255617e-01 -1.55984595e-01 -1.28935730e+00
-2.53834911e-02 -3.67490649e-01 -3.79544087e-02 2.61704803e-01
-2.81594962e-01 7.84351677e-02 1.56321764e-01 5.18364668e-01
-2.19538525e-01 3.48055989e-01 1.11549091e+00 -2.86678940e-01
-3.96201834e-02 -1.42723233e-01 -9.44534659e-01 5.82269371e-01
7.75316179e-01 -1.27284512e-01 -3.79245162e-01 -4.14714187e-01
2.42334530e-01 -1.69886991e-01 1.97847009e-01 -6.44348860e-01
-3.59152615e-01 -1.23569623e-01 7.03144431e-01 -1.72006357e-02
3.36780041e-01 -5.19354105e-01 -4.09283668e-01 1.03556298e-01
-1.02574003e+00 3.22179586e-01 1.34904981e-01 5.68781674e-01
-2.76745170e-01 -2.37447947e-01 7.45304883e-01 -3.54230285e-01
-4.57576871e-01 -2.44542714e-02 -7.20615745e-01 1.64479822e-01
3.63855273e-01 -1.64686099e-01 -2.23682210e-01 -4.96429980e-01
-8.11715603e-01 -1.40445486e-01 9.28457797e-01 4.96488899e-01
1.63148135e-01 -1.32967341e+00 -6.69749379e-01 3.03554147e-01
2.52086490e-01 -5.51648080e-01 -4.66928691e-01 3.60264629e-01
-8.73285010e-02 3.77032548e-01 -1.57770902e-01 -4.98330832e-01
-1.11630797e+00 3.85435522e-01 1.42755449e-01 -1.45411834e-01
-4.65644717e-01 7.07999945e-01 3.87591839e-01 -4.35042500e-01
2.46619642e-01 -1.84869885e-01 -3.09376478e-01 3.94935578e-01
3.84412557e-01 1.69324771e-01 -1.51536331e-01 -4.99149740e-01
-5.22740968e-02 3.80933017e-01 -4.37210262e-01 -4.53260124e-01
9.84056771e-01 -1.27856165e-01 -1.60162263e-02 1.19109666e+00
1.07241511e+00 1.28290325e-01 -1.13294661e+00 -3.68717164e-01
4.10410792e-01 -4.23890978e-01 1.10927120e-01 -9.35027540e-01
-6.15513802e-01 1.06187022e+00 3.62575084e-01 4.31309074e-01
4.42914039e-01 -8.53605643e-02 1.53997734e-01 3.09357762e-01
-2.06177086e-02 -1.17826116e+00 2.33934566e-01 7.36151814e-01
1.11147380e+00 -1.20449197e+00 4.44429815e-02 1.40431961e-02
-8.86122942e-01 1.21035516e+00 6.97044075e-01 -3.38083237e-01
6.99923456e-01 1.24551661e-01 2.99342871e-01 -1.57458395e-01
-7.78805614e-01 9.52541679e-02 6.40152335e-01 5.69025576e-01
1.27928817e+00 7.91713297e-02 -3.67817312e-01 3.97635132e-01
-5.69152772e-01 -3.03481102e-01 5.57215810e-01 6.98711038e-01
-1.77311227e-01 -1.34122741e+00 -3.60193461e-01 7.65849292e-01
-5.12812912e-01 -1.55614316e-01 -1.07892072e+00 7.85220385e-01
-1.77424297e-01 7.61140943e-01 4.87769693e-01 -2.37762868e-01
3.51227960e-03 6.60155594e-01 3.60333711e-01 -9.16137159e-01
-8.82042527e-01 -7.00628087e-02 2.33425647e-01 -2.17356682e-01
-1.84615180e-01 -9.45627987e-01 -4.43352133e-01 -6.56428456e-01
-2.39773616e-02 3.20427775e-01 5.68172097e-01 1.07284772e+00
1.87490612e-01 3.85414004e-01 5.12671351e-01 -8.21687102e-01
-5.69081306e-01 -1.45284116e+00 -6.60704851e-01 7.61438549e-01
4.73164737e-01 -6.82880759e-01 -3.93425822e-01 2.12545127e-01] | [11.321292877197266, 9.828271865844727] |
7064fad0-613a-49b2-a8e3-87313213e221 | concorde-net-cell-count-regularized | 1908.00907 | null | https://arxiv.org/abs/1908.00907v1 | https://arxiv.org/pdf/1908.00907v1.pdf | ConCORDe-Net: Cell Count Regularized Convolutional Neural Network for Cell Detection in Multiplex Immunohistochemistry Images | In digital pathology, cell detection and classification are often prerequisites to quantify cell abundance and explore tissue spatial heterogeneity. However, these tasks are particularly challenging for multiplex immunohistochemistry (mIHC) images due to high levels of variability in staining, expression intensity, and inherent noise as a result of preprocessing artefacts. We proposed a deep learning method to detect and classify cells in mIHC whole-tumour slide images of breast cancer. Inspired by inception-v3, we developed Cell COunt RegularizeD Convolutional neural Network (ConCORDe-Net) which integrates conventional dice overlap and a new cell count loss function for optimizing cell detection, followed by a multi-stage convolutional neural network for cell classification. In total, 20447 cells, belonging to five cell classes were annotated by experts from 175 patches extracted from 6 whole-tumour mIHC images. These patches were randomly split into training, validation and testing sets. Using ConCORDe-Net, we obtained a cell detection F1 score of 0.873, which is the best score compared to three state of the art methods. In particular, ConCORDe-Net excels at detecting closely located and weakly stained cells compared to other methods. Incorporating cell count loss in the objective function regularizes the network to learn weak gradient boundaries and separate weakly stained cells from background artefacts. Moreover, cell classification accuracy of 96.5% was achieved. These results support that incorporating problem-specific knowledge such as cell count into deep learning-based cell detection architectures improve the robustness of the algorithm. | ['Priya Lakshmi Narayanan', 'Teresa Marafioti', 'Yinyin Yuan', 'Yeman Brhane Hagos', 'Ayse U. Akarca'] | 2019-08-01 | null | null | null | null | ['cell-detection'] | ['computer-vision'] | [ 1.90005556e-01 -2.36624777e-01 -5.74572906e-02 -1.34866089e-01
-8.43570054e-01 -5.07703722e-01 4.11467969e-01 5.61856806e-01
-9.28495705e-01 9.07607317e-01 -3.71224850e-01 -1.13075852e-01
1.70043841e-01 -7.99598992e-01 -2.48256728e-01 -1.38488317e+00
-2.55116634e-02 5.17516673e-01 7.83032998e-02 1.54011890e-01
7.68917650e-02 9.71132934e-01 -1.09300506e+00 5.00815630e-01
7.31647253e-01 1.21131885e+00 1.56627521e-02 8.86836290e-01
-2.28629127e-01 8.03339601e-01 -5.02848506e-01 -1.97160944e-01
-1.28003478e-01 -1.33165389e-01 -5.88594556e-01 -1.23945802e-01
3.19098085e-01 -9.75384787e-02 4.56831680e-04 1.05573046e+00
6.71406627e-01 -4.99871343e-01 1.10451388e+00 -8.43950748e-01
-1.50487751e-01 6.43186495e-02 -7.31969655e-01 4.53422636e-01
-3.14932257e-01 3.17041993e-01 7.40487993e-01 -8.82400095e-01
7.85085380e-01 6.08108878e-01 9.65161800e-01 4.26198184e-01
-1.60632527e+00 -6.86275244e-01 -5.05347729e-01 -5.82097545e-02
-1.71325111e+00 -1.61649227e-01 3.30157727e-01 -7.53985047e-01
7.29480088e-01 4.37390119e-01 9.24325645e-01 5.64998865e-01
5.18454194e-01 9.11565483e-01 1.13924718e+00 -3.53653044e-01
3.26251000e-01 2.80482739e-01 -1.52983600e-02 6.37983799e-01
3.61347914e-01 -1.82570994e-01 1.11272193e-01 -1.96232833e-02
8.10255110e-01 1.90499336e-01 -2.28758603e-01 -2.65357822e-01
-1.22559500e+00 7.58813441e-01 6.99176550e-01 6.93610966e-01
-1.64568067e-01 2.30080783e-02 6.20390594e-01 -1.44420741e-02
4.42798555e-01 4.85549629e-01 -2.22664922e-01 3.21556866e-01
-9.78131175e-01 -6.15356490e-02 4.86341059e-01 1.55382931e-01
5.80743909e-01 -2.87089199e-01 -4.54999447e-01 8.99163187e-01
-7.46842697e-02 3.71988118e-01 5.17629445e-01 -7.01018512e-01
-3.68701041e-01 1.03868163e+00 -1.72974572e-01 -8.27999473e-01
-7.89465845e-01 -9.42462265e-01 -1.36225414e+00 4.07141805e-01
8.16163421e-01 4.52403948e-02 -8.67558241e-01 1.31103325e+00
3.32488298e-01 -1.18711978e-01 -1.75634488e-01 1.08127213e+00
6.50585651e-01 2.93852419e-01 1.51724860e-01 -3.19887251e-01
1.42671943e+00 -5.46076179e-01 -5.89331388e-01 8.61250609e-02
1.30260265e+00 -4.91544843e-01 8.28901947e-01 1.60114825e-01
-7.47215629e-01 -1.94652960e-01 -1.09303474e+00 -9.74821392e-03
-5.46231627e-01 5.05966902e-01 7.52598464e-01 3.93936157e-01
-9.60848331e-01 3.00646037e-01 -7.59171546e-01 -3.76521111e-01
1.19917977e+00 5.59612513e-01 -5.70478141e-01 2.66585499e-02
-7.22246945e-01 6.52021825e-01 1.41029865e-01 2.19504252e-01
-5.20481348e-01 -9.81327713e-01 -4.52947646e-01 2.98457295e-02
-1.89349756e-01 -5.97710729e-01 9.41194415e-01 -1.04510474e+00
-1.19521379e+00 1.48626447e+00 1.58872046e-02 -4.63914216e-01
5.49460292e-01 5.46252370e-01 -1.28646731e-01 3.59984845e-01
9.53849554e-02 8.08165252e-01 6.69389591e-02 -1.05425537e+00
-8.11780691e-01 -4.01632428e-01 -3.65936279e-01 -1.27431720e-01
-2.80461013e-01 -2.50875890e-01 -2.72700578e-01 -4.63773072e-01
-1.31177649e-01 -6.15208685e-01 -2.51776993e-01 5.02548695e-01
-8.10961053e-02 5.40435463e-02 6.32667482e-01 -4.03529257e-01
1.01681507e+00 -2.20215678e+00 -1.09725989e-01 3.50040793e-01
4.06555682e-01 3.04042578e-01 -8.99294242e-02 -1.28931344e-01
1.07245460e-01 1.08507192e-02 -2.32973963e-01 -5.58976196e-02
-1.00030430e-01 -1.77407801e-01 5.28504431e-01 9.55197692e-01
4.23211187e-01 1.07765985e+00 -9.72734332e-01 -7.86457300e-01
1.83812618e-01 7.80661762e-01 -3.70429009e-01 1.64454896e-02
2.11514551e-02 3.55908632e-01 4.48386148e-02 9.62006807e-01
8.01136434e-01 -4.06603932e-01 2.21949637e-01 -2.57231861e-01
7.83876553e-02 -4.46736723e-01 -6.92173421e-01 1.04207206e+00
-3.43561828e-01 9.02611196e-01 4.56030518e-01 -8.98225665e-01
7.61411309e-01 8.56562927e-02 6.27109826e-01 -7.61651278e-01
5.40513098e-01 4.98323500e-01 2.20350742e-01 -3.16824317e-01
-1.07543848e-01 -6.13305449e-01 1.48445964e-01 3.23763564e-02
-9.94863547e-03 -1.08343221e-01 3.97695273e-01 -1.44405201e-01
1.23131585e+00 -6.20825827e-01 3.04599851e-01 -5.64658225e-01
9.37578082e-01 6.65032417e-02 5.73972940e-01 4.48874325e-01
-5.85261881e-01 7.26197481e-01 9.24103975e-01 -8.09198380e-01
-8.29677045e-01 -8.01272869e-01 -5.67563832e-01 7.34948814e-01
-7.82752708e-02 2.13694885e-01 -6.51670039e-01 -6.64020658e-01
1.36451572e-01 4.19688132e-03 -9.04305279e-01 1.14340574e-01
-4.23514724e-01 -1.32078135e+00 6.82251573e-01 4.26303208e-01
3.82848233e-01 -8.87117445e-01 -4.47170198e-01 2.68192887e-01
3.03039327e-02 -8.43336344e-01 -2.90700972e-01 5.52322030e-01
-5.65161943e-01 -1.32259977e+00 -8.86962116e-01 -1.08401048e+00
1.03108370e+00 -7.42698042e-03 1.01563454e+00 4.69425350e-01
-1.04838705e+00 -1.66968524e-01 -1.89209133e-02 -3.92207801e-01
-3.73423040e-01 2.55069256e-01 -3.94415766e-01 1.27990946e-01
7.50779510e-01 -1.62635252e-01 -9.39490974e-01 1.56196788e-01
-9.04314399e-01 -1.60605073e-01 1.01940525e+00 1.31943846e+00
9.87174273e-01 -2.73320451e-02 5.08339465e-01 -9.63484347e-01
1.70207635e-01 -2.69456297e-01 -5.67549527e-01 -9.22831055e-03
-3.39035571e-01 -4.16452140e-01 6.94056749e-01 -3.62203568e-01
-6.11503839e-01 2.87482023e-01 -2.05079138e-01 4.18948494e-02
-1.66978315e-01 4.22063261e-01 1.57487318e-01 -3.26391280e-01
7.29800522e-01 5.64791635e-02 6.67771757e-01 2.52383947e-01
-3.60532969e-01 6.55110240e-01 5.11202216e-01 2.96984334e-02
1.44416481e-01 8.35992873e-01 3.12901944e-01 -8.43089104e-01
-6.33325756e-01 -8.06306124e-01 -5.23843050e-01 -2.34329417e-01
8.74646544e-01 -9.68594134e-01 -9.40875590e-01 7.96980023e-01
-8.69794190e-01 -6.99241221e-01 -1.73101604e-01 5.29520988e-01
-3.45297486e-01 5.25374711e-02 -1.06657803e+00 -5.95286548e-01
-5.45540214e-01 -9.69466746e-01 1.16395926e+00 2.27450684e-01
-2.73237437e-01 -1.17694354e+00 1.88979477e-01 1.91107288e-01
4.84607786e-01 6.42861128e-01 9.48928952e-01 -4.86695617e-01
-2.26741970e-01 -7.86539137e-01 -5.22174954e-01 2.12017506e-01
1.65855467e-01 2.37657070e-01 -1.05054259e+00 -3.97921711e-01
-3.85856062e-01 -2.99401760e-01 1.13443732e+00 7.35758305e-01
1.22227383e+00 -5.50721772e-02 -7.95409143e-01 8.86123240e-01
1.68477941e+00 1.10829645e-03 7.61179566e-01 4.47901398e-01
5.18625915e-01 6.25995755e-01 3.44918400e-01 3.05513442e-01
-6.39697015e-02 3.74664694e-01 4.16793436e-01 -8.91997695e-01
-1.00472800e-01 3.66522402e-01 -2.25689203e-01 4.19214040e-01
1.94597304e-01 -8.81537348e-02 -9.93758142e-01 7.91017354e-01
-1.43114853e+00 -8.80888700e-01 -3.40100437e-01 1.86232138e+00
1.03700829e+00 1.03763789e-01 1.47313654e-01 2.65298486e-01
8.03407431e-01 -2.14277908e-01 -5.67088902e-01 -9.92606804e-02
-4.27593827e-01 1.83056742e-01 6.31643832e-01 3.63553971e-01
-1.22816026e+00 4.90455836e-01 5.82963800e+00 1.14634001e+00
-1.43451560e+00 -1.31102249e-01 1.39654982e+00 -3.08946639e-01
1.32919580e-01 -4.86424327e-01 -6.75038934e-01 4.90954101e-01
4.52331156e-01 2.59541303e-01 -8.90106782e-02 3.35946470e-01
4.65782285e-02 -2.29307398e-01 -8.75954151e-01 9.42884564e-01
-1.28856733e-01 -1.51710880e+00 -1.12927273e-01 3.91903281e-01
7.51040697e-01 -1.01130731e-01 9.13357642e-03 3.12464267e-01
2.41584815e-02 -1.36358273e+00 1.11111403e-01 4.76044059e-01
1.08871877e+00 -8.25455785e-01 1.67425323e+00 2.80515611e-01
-8.99641335e-01 -6.82200538e-03 -4.61988389e-01 1.16777197e-01
-2.96288073e-01 9.82647061e-01 -1.01036286e+00 -2.22254787e-02
4.22917455e-01 5.30005634e-01 -6.01583421e-01 1.07312047e+00
4.79632378e-01 4.02070582e-01 -2.94381171e-01 -3.67206275e-01
2.75394708e-01 -5.46754226e-02 -4.18884121e-02 1.80125093e+00
3.26410621e-01 -6.49359450e-02 -1.78319514e-01 7.41408527e-01
-9.50904042e-02 2.18687534e-01 -2.12798730e-01 1.61541160e-02
3.53499860e-01 2.02516651e+00 -1.31840694e+00 -2.00716957e-01
-3.42801303e-01 5.60783803e-01 4.81613755e-01 2.28221446e-01
-6.87924206e-01 -4.59013790e-01 4.47782576e-01 4.88178849e-01
1.95641622e-01 3.84787023e-01 -5.61588705e-01 -5.92387140e-01
-3.87208968e-01 -5.79639494e-01 4.56045657e-01 -1.63282841e-01
-1.55650032e+00 4.34689403e-01 -5.97121537e-01 -1.22500205e+00
3.23940068e-01 -1.03835440e+00 -5.77527046e-01 6.19750798e-01
-1.57330704e+00 -1.03592265e+00 -5.51010847e-01 4.67290021e-02
1.45088121e-01 -6.62021041e-02 8.23804498e-01 4.09491688e-01
-6.13059580e-01 6.62518680e-01 2.83544481e-01 3.08272511e-01
6.83423102e-01 -1.53540576e+00 -4.41802084e-01 2.24408433e-01
-4.09932882e-01 3.72262210e-01 3.12757671e-01 -2.09035695e-01
-1.00190854e+00 -1.25887072e+00 6.60517991e-01 -1.81564510e-01
6.45560324e-01 -4.46910560e-01 -8.75298619e-01 1.29853711e-01
6.63056644e-03 7.26371586e-01 1.15529144e+00 -4.06564295e-01
8.31387937e-02 -3.44753027e-01 -1.30432904e+00 5.44505656e-01
4.03149992e-01 -3.84023905e-01 2.78200805e-01 3.23025852e-01
-1.05512246e-01 -4.38560098e-01 -1.01025927e+00 4.63430464e-01
7.59349108e-01 -9.87935424e-01 7.58795321e-01 6.89998120e-02
3.17897707e-01 -3.39916766e-01 1.31424382e-01 -9.17422414e-01
-6.14850402e-01 4.13684659e-02 2.75223285e-01 7.91119277e-01
5.40030062e-01 -6.91146255e-01 1.15401208e+00 2.42973026e-02
-1.39192730e-01 -1.20292521e+00 -1.05913973e+00 -4.54315543e-01
5.14290094e-01 2.94016212e-01 3.76282007e-01 9.45252061e-01
3.77393398e-03 2.10189298e-01 4.16359484e-01 -2.04139143e-01
6.24473691e-01 -1.17094450e-01 7.36519098e-01 -1.40496361e+00
6.27941564e-02 -1.07341123e+00 -6.92115784e-01 -5.06831884e-01
1.48918927e-01 -1.19204211e+00 -4.00127210e-02 -1.45625818e+00
5.35743713e-01 -3.93260092e-01 -5.78474641e-01 4.08227444e-01
-2.71956712e-01 8.51828933e-01 -3.00923616e-01 1.59329131e-01
-5.60723126e-01 1.50841013e-01 1.46051562e+00 -5.90934396e-01
3.64597421e-03 -3.01596820e-01 -6.52981222e-01 4.91399497e-01
7.51823604e-01 -2.08983585e-01 1.97757900e-01 7.40457997e-02
1.79472715e-01 -2.37064406e-01 5.42911828e-01 -1.21465421e+00
3.36049259e-01 1.19470797e-01 1.11375856e+00 -6.77556515e-01
4.06186506e-02 -8.02782238e-01 1.78967148e-01 8.78833234e-01
-3.00210625e-01 -6.52721524e-01 3.04310858e-01 2.63655543e-01
-4.05630082e-01 -9.76852328e-02 1.15568364e+00 -2.39374578e-01
-2.27659553e-01 2.32527941e-01 -6.66592300e-01 -3.18399072e-01
1.05816007e+00 -7.17354238e-01 -5.05227745e-01 2.40451738e-01
-4.49631721e-01 3.01244467e-01 5.38224280e-01 -4.59800214e-01
3.03521305e-01 -1.43452752e+00 -7.83089817e-01 2.11741462e-01
4.12223160e-01 1.90832928e-01 4.03913140e-01 1.36368406e+00
-1.20930648e+00 2.89737374e-01 -2.68203050e-01 -1.10705245e+00
-1.08943343e+00 2.17626974e-01 7.47048378e-01 -5.98330021e-01
-2.44363979e-01 1.05473721e+00 3.63046080e-01 -5.30753791e-01
1.30093381e-01 -3.61011177e-01 -3.81027341e-01 1.50571227e-01
4.80674654e-01 2.62802929e-01 1.11374721e-01 -5.19996285e-01
-5.36941528e-01 5.97444892e-01 -3.52747917e-01 4.37890977e-01
1.08738911e+00 2.15060636e-01 -4.15857792e-01 3.97768110e-01
1.48497188e+00 -1.99100405e-01 -1.24413240e+00 1.22442715e-01
-7.42252916e-02 -2.08069831e-01 2.92604804e-01 -8.79326284e-01
-1.26063418e+00 6.62074685e-01 8.87984812e-01 1.59927737e-02
1.08155251e+00 -1.58312365e-01 5.71362972e-01 1.04692347e-01
6.80666193e-02 -1.13779557e+00 1.59298740e-02 2.77017236e-01
4.91173774e-01 -1.39900684e+00 8.16658288e-02 -3.33436549e-01
-2.21223161e-01 1.15232432e+00 7.41594315e-01 -9.72720832e-02
4.18626845e-01 6.40074313e-01 3.72152120e-01 -3.79719824e-01
-7.08904564e-01 -1.53022870e-01 -5.47151379e-02 5.46640873e-01
7.30767250e-01 4.58366647e-02 -3.97080153e-01 6.30097806e-01
9.82314125e-02 1.68075144e-01 2.75891006e-01 7.24943757e-01
-4.08998996e-01 -3.98045868e-01 -1.33569300e-01 7.38271356e-01
-7.48315275e-01 2.54495502e-01 -4.42813516e-01 1.02664530e+00
2.03614444e-01 4.45192754e-01 4.83565748e-01 1.07342442e-02
7.59180263e-02 -3.44440579e-01 3.92253220e-01 -3.81967545e-01
-8.29607666e-01 3.30884963e-01 -3.98334950e-01 -1.43054113e-01
-3.49316746e-01 -3.95774901e-01 -1.38283372e+00 -2.17013389e-01
-4.57154006e-01 -4.87170294e-02 2.93644726e-01 6.00260437e-01
1.21826693e-01 5.90019524e-01 4.75516289e-01 -6.28562391e-01
-7.67809600e-02 -1.04735482e+00 -9.68618035e-01 2.46375054e-01
5.17563760e-01 -4.23108697e-01 -6.58333421e-01 1.04804732e-01] | [14.952441215515137, -3.0870776176452637] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.