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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
4b398a0d-1a75-476b-b518-0c607f82c031 | dialo-ap-a-dependency-parsing-based-argument | null | null | https://aclanthology.org/2022.coling-1.74 | https://aclanthology.org/2022.coling-1.74.pdf | Dialo-AP: A Dependency Parsing Based Argument Parser for Dialogues | While neural approaches to argument mining (AM) have advanced considerably, most of the recent work has been limited to parsing monologues. With an urgent interest in the use of conversational agents for broader societal applications, there is a need to advance the state-of-the-art in argument parsers for dialogues. This enables progress towards more purposeful conversations involving persuasion, debate and deliberation. This paper discusses Dialo-AP, an end-to-end argument parser that constructs argument graphs from dialogues. We formulate AM as dependency parsing of elementary and argumentative discourse units; the system is trained using extensive pre-training and curriculum learning comprising nine diverse corpora. Dialo-AP is capable of generating argument graphs from dialogues by performing all sub-tasks of AM. Compared to existing state-of-the-art baselines, Dialo-AP achieves significant improvements across all tasks, which is further validated through rigorous human evaluation. | ['Rohini K. Srihari', 'Souvik Das', 'Sougata Saha'] | null | null | null | null | coling-2022-10 | ['argument-mining'] | ['natural-language-processing'] | [ 4.96203929e-01 1.25419414e+00 -3.01092416e-01 -5.52521288e-01
-1.20247149e+00 -7.69702852e-01 1.17350984e+00 6.20223463e-01
-3.95371139e-01 1.11712313e+00 9.27219808e-01 -1.17259252e+00
2.22766086e-01 -8.81515920e-01 -5.89822173e-01 -3.34520452e-02
9.66975838e-02 9.70094800e-01 -8.08712617e-02 -6.57951057e-01
2.65203565e-01 -1.97064295e-01 -9.71617579e-01 8.41094196e-01
8.57027650e-01 4.19801891e-01 -2.81500608e-01 1.11992240e+00
-5.69026887e-01 1.36099374e+00 -1.02959728e+00 -1.02678752e+00
-4.80180770e-01 -8.99179220e-01 -1.79336858e+00 -3.98503542e-01
2.87542999e-01 -3.17333698e-01 2.30355352e-01 6.76702976e-01
5.83681464e-01 -4.12419885e-02 5.57627261e-01 -9.16797280e-01
-7.43259966e-01 1.41134012e+00 -1.04672283e-01 3.14914852e-01
5.45989335e-01 -1.37161240e-01 1.62366855e+00 -4.03605551e-01
7.81101823e-01 1.77786946e+00 6.56081975e-01 9.38198924e-01
-1.26338696e+00 -4.57334787e-01 2.61248320e-01 -1.24949336e-01
2.61309862e-01 -3.88394028e-01 1.02231634e+00 -1.84829116e-01
1.35601699e+00 1.89666271e-01 7.71080077e-01 1.61788583e+00
-1.44730031e-01 1.22073901e+00 1.03471386e+00 -6.89627707e-01
1.30442411e-01 -3.76405239e-01 6.27816141e-01 5.68760693e-01
2.10908949e-01 -3.24484080e-01 -4.20795560e-01 -3.52788121e-01
3.94605488e-01 -9.74615514e-01 8.16567391e-02 1.77806258e-01
-1.01436329e+00 1.40219271e+00 1.09733053e-01 1.68863803e-01
-3.50638568e-01 9.13994983e-02 8.87650609e-01 4.51401770e-01
6.70821965e-01 7.11565256e-01 -4.76289839e-01 -6.49512827e-01
-1.81576714e-01 6.70704484e-01 1.50626433e+00 4.80409294e-01
-1.41149253e-01 -7.58343935e-02 -1.88140124e-01 9.76189554e-01
3.03242385e-01 2.64148246e-02 2.15196311e-01 -1.00926149e+00
1.16930842e+00 7.69105196e-01 -1.11502580e-01 -4.47522551e-01
-4.67041880e-01 2.70469896e-02 -3.33061248e-01 7.77865946e-02
8.77131820e-01 -7.82288849e-01 -2.92942822e-01 1.84464777e+00
5.04979491e-01 -4.31916296e-01 8.05941403e-01 3.41763496e-01
1.32375467e+00 5.10408163e-01 5.16897976e-01 5.31460419e-02
1.65817511e+00 -1.12067878e+00 -6.92440569e-01 -5.54832458e-01
9.04897213e-01 -6.76772296e-01 1.05459833e+00 2.15134416e-02
-1.43720531e+00 -1.95795447e-01 -8.75416994e-01 -3.83474141e-01
-1.37116626e-01 -2.17463210e-01 9.63509262e-01 6.40203953e-01
-5.52905679e-01 2.55157620e-01 -5.26146173e-01 -1.92920744e-01
7.42098927e-01 1.36184305e-01 -1.05158478e-01 3.97309452e-01
-1.44448912e+00 9.88011181e-01 5.04094124e-01 4.07287031e-02
-4.65101182e-01 -4.39208984e-01 -1.13424695e+00 -4.21747081e-02
3.72691661e-01 -8.48423839e-01 1.97912586e+00 -1.00999629e+00
-1.74817741e+00 1.22261155e+00 1.74343754e-02 -9.36351478e-01
5.61423957e-01 -6.19256496e-01 -1.03412904e-01 -4.48629111e-02
3.17755580e-01 6.69400692e-01 3.07262748e-01 -9.93002355e-01
-7.14816153e-01 1.89582277e-02 6.91837609e-01 5.25102258e-01
8.24667960e-02 3.58559608e-01 2.71329105e-01 -2.94802845e-01
-3.82256627e-01 -8.02750528e-01 -4.51621898e-02 -5.48723578e-01
-7.44960666e-01 -1.11653030e+00 6.60306454e-01 -6.48577571e-01
8.40534985e-01 -1.51018405e+00 -9.89077240e-02 -4.25131023e-01
1.41930178e-01 4.35691714e-01 -2.03110680e-01 6.30692840e-01
-3.48292030e-02 2.27244407e-01 -9.95674133e-02 -3.25698406e-01
1.69748962e-01 4.30134356e-01 -3.89317989e-01 5.18022571e-03
5.47490060e-01 1.15963519e+00 -1.13591409e+00 -6.10316157e-01
4.17215638e-02 3.63030910e-01 -6.90177619e-01 4.34987634e-01
-7.68392265e-01 5.32734334e-01 -7.83472061e-01 3.31433982e-01
5.71132302e-02 -4.60657388e-01 6.31640553e-01 3.74264538e-01
4.92954627e-03 1.43010485e+00 -4.25479293e-01 1.63940251e+00
-3.56817156e-01 9.22369421e-01 1.79134429e-01 -1.29165554e+00
7.28013039e-01 6.01972044e-01 -1.35756388e-01 -6.89581096e-01
4.32401121e-01 2.93746918e-01 4.53875124e-01 -3.49440008e-01
5.14919341e-01 1.80247158e-01 -3.29887271e-01 8.64468217e-01
-1.80666089e-01 -1.24451526e-01 4.53363627e-01 3.67881149e-01
9.97396588e-01 3.45901966e-01 4.26291466e-01 -3.08097452e-01
7.44798958e-01 4.50806051e-01 1.03736721e-01 7.53406823e-01
-3.93167809e-02 1.21744484e-01 9.77166235e-01 -6.53269172e-01
-1.06845891e+00 -8.57328594e-01 -7.20966905e-02 1.28578711e+00
-3.06788832e-01 -2.77818769e-01 -1.31287467e+00 -1.17777920e+00
-1.52038187e-01 9.56119478e-01 -6.96543992e-01 6.02754712e-01
-1.37021136e+00 -7.60494769e-01 9.23682153e-01 5.38218796e-01
7.60984182e-01 -1.78358603e+00 -8.16487432e-01 6.24408007e-01
-4.74562824e-01 -1.14324307e+00 2.33148217e-01 2.71648884e-01
-7.23594010e-01 -1.62789726e+00 -3.05423200e-01 -1.04924262e+00
3.27806145e-01 -2.50261843e-01 1.74449897e+00 4.01496768e-01
9.69476998e-02 3.31208616e-01 -3.36052030e-01 -8.16411436e-01
-1.21373427e+00 3.69362146e-01 -7.97167957e-01 -7.06251025e-01
3.95647109e-01 -3.43170196e-01 -4.65405583e-01 -2.59819537e-01
-2.77791947e-01 4.69968885e-01 2.44405001e-01 1.16560268e+00
7.00918362e-02 -8.58151913e-01 9.33438659e-01 -1.58088100e+00
1.66864741e+00 -4.33225721e-01 -2.58853644e-01 6.14534914e-02
-4.97381240e-01 1.98121607e-01 5.92357337e-01 -1.98274240e-01
-1.48662055e+00 -7.28913128e-01 -6.71127617e-01 1.10437441e+00
-1.90529376e-01 4.36331511e-01 7.39775673e-02 7.33171284e-01
7.95991719e-01 -4.25117701e-01 1.88571304e-01 -1.85885817e-01
6.98145866e-01 6.80578649e-01 7.24092782e-01 -1.34164619e+00
2.84196407e-01 9.91299748e-02 -1.28802106e-01 -9.55810845e-01
-1.35550523e+00 1.71403706e-01 -2.54081339e-01 2.98376549e-02
1.08539760e+00 -6.99020684e-01 -9.92242813e-01 1.34564526e-02
-1.57341993e+00 -7.78075695e-01 -3.04915547e-01 7.67366439e-02
-5.38611948e-01 3.38148057e-01 -1.03112578e+00 -7.84399271e-01
-9.13448274e-01 -8.59079242e-01 7.52606869e-01 4.08560127e-01
-9.53526437e-01 -1.26209772e+00 3.39196891e-01 1.05046904e+00
7.79885799e-02 4.73655164e-01 1.28958356e+00 -1.19986117e+00
-3.97398621e-02 1.34348035e-01 -1.42308921e-01 1.62445113e-01
-1.84167266e-01 -3.19217652e-01 -9.21034992e-01 1.63205996e-01
-3.11326444e-01 -1.02970052e+00 4.52536762e-01 3.27975571e-01
7.09263921e-01 -4.63464707e-01 -1.04695447e-01 -4.80473787e-02
4.73087817e-01 1.29598811e-01 4.68503505e-01 9.33508754e-01
2.85248160e-01 9.52481151e-01 4.86439258e-01 -9.65736806e-02
7.94926226e-01 8.23014006e-02 4.03919667e-01 -2.87683189e-01
-2.43086919e-01 -4.24107492e-01 1.30602300e-01 5.39347291e-01
-1.14400119e-01 -5.82624614e-01 -8.76245081e-01 8.21700037e-01
-2.12448621e+00 -1.14338291e+00 -5.37718654e-01 1.33741212e+00
1.08435249e+00 4.47772980e-01 1.43495694e-01 1.62586764e-01
5.67232013e-01 5.83437562e-01 -4.45770115e-01 -1.17031348e+00
-2.22948641e-01 4.88902807e-01 -2.27634549e-01 7.09628642e-01
-1.18262112e+00 1.01207411e+00 6.19377804e+00 2.88057476e-01
-3.99871171e-01 -1.22000299e-01 1.09143865e+00 5.01947165e-01
-3.69863808e-01 1.73425794e-01 -7.34888136e-01 7.15890005e-02
9.89022076e-01 2.09644828e-02 1.51375577e-01 8.30476880e-01
-4.67373617e-02 -8.01244304e-02 -1.26070774e+00 3.29542369e-01
-1.16573766e-01 -1.75434136e+00 -8.29328299e-02 -8.20142925e-02
4.98906642e-01 1.08499907e-01 -3.27492148e-01 5.85160613e-01
1.12585616e+00 -1.14981580e+00 4.68522310e-01 -3.57498556e-01
3.57784867e-01 -6.28805876e-01 7.60022521e-01 3.49639982e-01
-8.60660851e-01 -7.55702145e-03 4.92060147e-02 -5.61428726e-01
4.81671989e-01 1.36584446e-01 -1.19255841e+00 3.56264412e-01
2.67653674e-01 2.42488727e-01 -1.19534701e-01 1.49505794e-01
-1.01610708e+00 1.14368987e+00 -1.84634849e-01 -5.78345776e-01
6.67978585e-01 -1.42491877e-01 4.40965682e-01 1.42850423e+00
-5.39729536e-01 2.65471697e-01 2.12704733e-01 5.92125893e-01
-5.13812900e-01 2.03494281e-01 -3.61858875e-01 -4.39724237e-01
5.70514798e-01 1.17421496e+00 -5.16781688e-01 -5.36139786e-01
-6.26742065e-01 5.29985309e-01 8.19311738e-01 -2.07860791e-03
-6.41338229e-01 -7.44203106e-02 5.77131987e-01 -2.55171388e-01
-7.49714524e-02 -1.58937797e-02 -4.31629300e-01 -6.24167144e-01
-1.25430301e-01 -1.35722053e+00 7.75263846e-01 -3.99266809e-01
-1.31651449e+00 6.76793814e-01 2.46918052e-02 -2.76616156e-01
-9.52473879e-01 -7.35812366e-01 -1.09826529e+00 9.03654337e-01
-1.52292573e+00 -1.44166124e+00 6.14514314e-02 1.02080218e-01
1.04058337e+00 -2.41836071e-01 1.29535246e+00 -3.00716937e-01
-2.64611244e-01 2.47720033e-01 -4.85977173e-01 8.13533187e-01
3.42331320e-01 -1.38171995e+00 1.28055787e+00 4.77603942e-01
1.45042181e-01 4.82239664e-01 6.94019020e-01 -6.68649733e-01
-1.05189240e+00 -5.62065661e-01 1.23720038e+00 -6.89145088e-01
9.03614581e-01 -2.48547971e-01 -7.61231661e-01 7.90250242e-01
9.88226116e-01 -5.19927084e-01 8.17565441e-01 7.78121054e-01
-3.31240296e-01 7.22835243e-01 -8.27953756e-01 7.90674746e-01
1.04020262e+00 -6.29607677e-01 -1.42874551e+00 3.53683323e-01
7.37083972e-01 -7.66678810e-01 -6.83207691e-01 1.13652445e-01
6.35790169e-01 -8.02473545e-01 1.03298056e+00 -1.19553661e+00
9.66816068e-01 3.34671348e-01 2.92702317e-01 -1.01746869e+00
2.79115081e-01 -9.82537091e-01 -1.73962742e-01 1.37246859e+00
7.98399568e-01 -7.16222227e-01 9.97636318e-01 6.32028818e-01
-3.44358057e-01 -7.97993600e-01 -7.29597569e-01 -1.02191038e-01
7.22312748e-01 -2.22456604e-01 4.36953276e-01 9.67152059e-01
2.94707716e-01 1.33572888e+00 2.84760073e-02 -3.80486488e-01
5.82432985e-01 4.14253771e-01 1.15179253e+00 -1.61795986e+00
-3.26622903e-01 -6.02732301e-01 5.86146653e-01 -1.27784991e+00
6.78851426e-01 -8.63782644e-01 -2.56349184e-02 -2.14402175e+00
-8.31341892e-02 -1.27223283e-01 4.12041903e-01 3.84844899e-01
-3.62519622e-01 -1.46461919e-01 -9.50869173e-02 -9.31407511e-02
-4.98239338e-01 3.52117598e-01 1.39410365e+00 -1.03472844e-01
-2.67566621e-01 -7.61886016e-02 -1.00707030e+00 9.85588789e-01
1.02260053e+00 -3.83542806e-01 -4.19972986e-01 -5.12186170e-01
3.72353315e-01 1.61263973e-01 1.42378330e-01 -3.78873765e-01
3.81724574e-02 -1.54784815e-02 1.55587673e-01 -6.39256299e-01
2.73925126e-01 5.71231581e-02 -5.86900949e-01 3.62878323e-01
-8.73881876e-01 3.72305870e-01 1.98813632e-01 3.74357551e-01
-3.14630508e-01 -3.75600904e-01 3.89218479e-01 -3.22512060e-01
-2.46319905e-01 -4.84390169e-01 -4.51926321e-01 9.41258132e-01
4.88736570e-01 -1.35707274e-01 -1.09279692e+00 -5.06644964e-01
-4.24035013e-01 4.77037579e-01 -5.77438921e-02 3.09246063e-01
1.92350060e-01 -7.08594859e-01 -1.14094722e+00 -2.19443411e-01
-3.32180858e-01 5.59518635e-01 -3.27753454e-01 3.27369094e-01
-5.32088220e-01 5.07426262e-01 3.20845284e-02 -2.56013066e-01
-1.69940472e+00 -1.73955411e-01 1.09316334e-01 -9.04076159e-01
-1.08712709e+00 9.92049634e-01 -5.08411452e-02 -6.66831732e-01
2.00907618e-01 -1.84975490e-01 -6.27803087e-01 -9.78550687e-03
4.18910712e-01 6.02190755e-02 -3.81201744e-01 -3.01345319e-01
3.26846838e-02 -6.23158216e-02 -2.18632489e-01 -3.57335895e-01
1.43386555e+00 2.96389312e-01 -3.35758226e-03 1.27149850e-01
5.79218984e-01 2.41259918e-01 -1.15407109e+00 1.08891204e-01
3.25742066e-01 3.24219167e-01 -5.18913567e-01 -1.00370193e+00
-5.42628989e-02 7.58955121e-01 -2.26283476e-01 6.60750329e-01
3.54864299e-01 2.90600598e-01 9.52434182e-01 7.40874767e-01
-1.62796110e-01 -1.29080546e+00 2.08744183e-01 1.13080263e+00
8.00673366e-01 -1.23137903e+00 -4.69180057e-03 -4.64323223e-01
-6.96845591e-01 1.18125129e+00 4.63549435e-01 -2.39938512e-01
2.63662469e-02 3.46974552e-01 2.65180141e-01 -5.31255722e-01
-8.64907384e-01 2.75469981e-02 9.06525739e-03 6.02334023e-01
1.15662491e+00 1.12934604e-01 -7.43786752e-01 6.34177983e-01
-6.63258195e-01 -4.88587797e-01 3.00438076e-01 1.05035043e+00
-3.25171322e-01 -1.59364116e+00 -1.45686612e-01 2.12842524e-01
-9.20109570e-01 -3.61713439e-01 -1.06938910e+00 1.27355528e+00
-4.89782721e-01 1.26386178e+00 1.31989241e-01 3.93399179e-01
2.19054729e-01 3.80718827e-01 5.21313846e-01 -6.25779152e-01
-1.23401868e+00 -3.15810144e-01 1.66537786e+00 -1.32715151e-01
-7.11172521e-01 -7.57887125e-01 -1.69507873e+00 -2.92718351e-01
-2.02189356e-01 6.98752403e-01 7.14356124e-01 1.09094453e+00
8.61394256e-02 7.43350029e-01 -1.39275879e-01 -5.27047873e-01
-6.70410633e-01 -1.10727310e+00 4.66508448e-01 5.75342953e-01
1.23491816e-01 -2.63946384e-01 -1.24918699e-01 -1.91460758e-01] | [10.603743553161621, 9.483752250671387] |
45be4b1c-c461-4791-9e5b-7d8d204120cd | pseudo-labels-for-single-positive-multi-label | 2306.01034 | null | https://arxiv.org/abs/2306.01034v1 | https://arxiv.org/pdf/2306.01034v1.pdf | Pseudo Labels for Single Positive Multi-Label Learning | The cost of data annotation is a substantial impediment for multi-label image classification: in every image, every category must be labeled as present or absent. Single positive multi-label (SPML) learning is a cost-effective solution, where models are trained on a single positive label per image. Thus, SPML is a more challenging domain, since it requires dealing with missing labels. In this work, we propose a method to turn single positive data into fully-labeled data: Pseudo Multi-Labels. Basically, a teacher network is trained on single positive labels. Then, we treat the teacher model's predictions on the training data as ground-truth labels to train a student network on fully-labeled images. With this simple approach, we show that the performance achieved by the student model approaches that of a model trained on the actual fully-labeled images. | ['Julio Arroyo'] | 2023-06-01 | null | null | null | null | ['multi-label-image-classification', 'multi-label-learning'] | ['computer-vision', 'methodology'] | [ 6.99524522e-01 2.44647324e-01 -7.26843894e-01 -6.72345698e-01
-9.91966009e-01 -6.20120704e-01 3.10256362e-01 1.49235040e-01
-4.50258702e-01 7.33702302e-01 -5.15151381e-01 -2.48415828e-01
4.85840976e-01 -5.10645866e-01 -6.41050696e-01 -6.55656040e-01
6.21744275e-01 6.46418989e-01 3.39902878e-01 3.51329535e-01
7.97018558e-02 1.41306207e-01 -1.61725271e+00 5.80495417e-01
5.25832057e-01 8.99884880e-01 2.36555442e-01 5.65689623e-01
-2.98282087e-01 9.78982866e-01 -5.34795702e-01 -3.72126639e-01
1.26142472e-01 -4.10215020e-01 -1.06874919e+00 5.04808426e-01
6.59043133e-01 -2.10412428e-01 4.60153371e-01 1.35547447e+00
-2.31838953e-02 -3.67836684e-01 9.34053600e-01 -1.64265156e+00
-4.53123689e-01 3.75383228e-01 -9.54893947e-01 -2.95977563e-01
2.36895755e-02 -2.25056082e-01 1.20239723e+00 -1.21768951e+00
3.78338218e-01 1.14862788e+00 5.92245162e-01 8.28083992e-01
-1.31420708e+00 -6.46892548e-01 2.50985235e-01 2.19851907e-04
-1.20650780e+00 -1.11139350e-01 6.77042127e-01 -6.91607594e-01
2.90597826e-01 3.11519325e-01 3.15861672e-01 8.92566204e-01
-7.11138695e-02 9.21870530e-01 1.43087983e+00 -8.50049973e-01
1.71259064e-02 7.05244899e-01 3.84423465e-01 6.14185512e-01
1.60341099e-01 -2.02018693e-01 -1.89879864e-01 -1.42176583e-01
4.49120998e-01 2.13831916e-01 2.37474777e-02 -4.44174051e-01
-1.13593280e+00 8.02122116e-01 1.77288696e-01 3.58445138e-01
-1.42713815e-01 8.04504007e-02 2.43586913e-01 2.14174539e-01
7.42464423e-01 2.63193130e-01 -7.48465121e-01 5.22936821e-01
-8.75228941e-01 -2.30661660e-01 4.42290038e-01 8.75225842e-01
1.27982628e+00 -3.30008596e-01 2.11936259e-03 1.01392972e+00
5.59646130e-01 4.89716828e-01 6.19258344e-01 -8.58932257e-01
7.08321333e-02 8.97373199e-01 1.50216907e-01 -7.43368626e-01
-4.98795152e-01 -3.60915393e-01 -8.55775476e-01 2.52982706e-01
3.85379076e-01 -8.80508423e-02 -1.14869022e+00 1.63305664e+00
4.94665354e-01 1.92697063e-01 1.53586045e-01 5.59264839e-01
8.68672252e-01 5.98935187e-01 4.46772546e-01 -2.84513056e-01
1.10701740e+00 -1.52624023e+00 -5.34421742e-01 -5.43717921e-01
1.11404824e+00 -7.51364946e-01 1.07246327e+00 4.40958500e-01
-6.63169563e-01 -7.05085993e-01 -6.60779595e-01 2.20083416e-01
-4.11064148e-01 7.49044180e-01 3.27302575e-01 5.80167234e-01
-1.07568896e+00 1.68145359e-01 -2.50107080e-01 -2.23296687e-01
1.93030417e-01 3.36543113e-01 -4.95323181e-01 -2.38217995e-01
-8.09496343e-01 1.01872647e+00 6.72513485e-01 -2.46649995e-01
-1.07709444e+00 -4.10718203e-01 -7.43715823e-01 -1.48443626e-02
4.35459673e-01 -3.44400965e-02 1.51927698e+00 -1.37564421e+00
-9.32560146e-01 1.53413439e+00 -2.05304086e-01 3.23156342e-02
4.14656550e-01 2.64302552e-01 -3.37533355e-01 8.54544267e-02
4.71290708e-01 1.04950178e+00 9.89876390e-01 -1.95522988e+00
-1.09407640e+00 -4.23137285e-02 1.04579248e-01 2.89559290e-02
-4.32156771e-01 1.22455291e-01 1.23702325e-02 -3.74494672e-01
3.26006383e-01 -1.10438704e+00 -3.05430114e-01 -1.74115255e-01
-5.44673920e-01 -6.44011140e-01 7.64666140e-01 -2.63163745e-01
8.49668384e-01 -2.31563783e+00 -2.39191413e-01 1.65311381e-01
2.84556657e-01 2.06338495e-01 -4.27384555e-01 4.33560908e-02
-4.74757075e-01 4.33554649e-01 -1.40336275e-01 -5.64738393e-01
-3.12000662e-01 3.52235645e-01 -1.86083019e-01 2.13410512e-01
4.71177965e-01 7.24671781e-01 -9.40295398e-01 -8.19085002e-01
7.11264610e-02 9.20486003e-02 -3.49731922e-01 3.45824927e-01
-1.99948162e-01 6.82467997e-01 -9.25624594e-02 6.10021472e-01
6.63146257e-01 -7.76291847e-01 3.96925062e-01 -1.46232337e-01
-1.15088122e-02 -2.71265179e-01 -1.18626535e+00 1.19808221e+00
-5.15508413e-01 1.07664295e-01 -1.42142251e-01 -1.11447883e+00
9.40346003e-01 6.06357098e-01 4.63954329e-01 -3.69363964e-01
6.27982691e-02 5.21797895e-01 -3.77450168e-01 -3.88574183e-01
2.31654286e-01 -4.46022689e-01 -1.35208830e-01 7.82865345e-01
3.92761201e-01 -1.55090913e-01 2.35398233e-01 9.85193104e-02
6.54917955e-01 -3.37510854e-02 4.06761169e-01 -1.51796611e-02
6.42742813e-01 1.53329626e-01 8.61352026e-01 5.97832203e-01
-2.15832263e-01 5.42620659e-01 3.30458522e-01 -5.58679461e-01
-1.07447088e+00 -6.90233231e-01 -5.18060401e-02 1.51144397e+00
1.67794213e-01 -2.45739132e-01 -6.38105273e-01 -1.22051024e+00
-2.40969419e-01 3.66219908e-01 -5.98253429e-01 1.20718805e-02
-3.59906137e-01 -6.48909807e-01 6.20519407e-02 2.58963108e-01
1.82578191e-01 -1.04784477e+00 -2.96892822e-01 9.33262184e-02
-3.45171541e-01 -1.00106287e+00 -2.56701887e-01 5.23072481e-01
-6.49163365e-01 -1.24878788e+00 -7.16279984e-01 -1.27536416e+00
1.28045142e+00 5.79122484e-01 1.09016240e+00 4.25109774e-01
1.09211408e-01 1.39781103e-01 -3.90892625e-01 -3.94209027e-01
-6.95247233e-01 3.83225679e-02 -1.49270613e-02 3.62032264e-01
2.53104746e-01 -1.89852670e-01 -1.37548417e-01 5.05432129e-01
-7.91290879e-01 5.42347848e-01 7.48245239e-01 1.11907768e+00
9.45425391e-01 7.93190971e-02 9.26254034e-01 -1.36212790e+00
8.06870759e-02 -3.99964064e-01 -4.67796504e-01 7.45817900e-01
-6.86094642e-01 -2.00702772e-01 6.42975569e-01 -7.19111919e-01
-9.81811345e-01 6.84685826e-01 -7.01896325e-02 -4.35148925e-01
-5.31000257e-01 3.08285475e-01 2.15950180e-02 -3.08054954e-01
5.34528136e-01 3.93998288e-02 -2.64143884e-01 -4.61060792e-01
3.94506454e-01 9.41937864e-01 3.90247464e-01 -3.69728863e-01
5.50467551e-01 3.08507860e-01 2.26101046e-03 -3.81629169e-01
-1.78197229e+00 -7.60512590e-01 -1.10901284e+00 -5.07398069e-01
8.03815842e-01 -1.02577841e+00 -2.79518098e-01 4.92145985e-01
-1.17233515e+00 -3.99483413e-01 -2.47785419e-01 3.87818307e-01
-4.25042212e-01 2.30174258e-01 -4.56316411e-01 -6.37097001e-01
-7.75834620e-02 -1.18925774e+00 1.01673841e+00 3.02223116e-01
-1.38144135e-01 -1.14248991e+00 8.50878730e-02 4.68664914e-01
9.27385166e-02 -9.20975059e-02 9.79123354e-01 -1.05986261e+00
-1.44752190e-01 -4.28779811e-01 -4.59477246e-01 6.73625469e-01
1.05176635e-01 -3.30991745e-02 -1.25926006e+00 -4.33521897e-01
1.34051532e-01 -1.01834929e+00 7.14636743e-01 3.85824069e-02
1.30414534e+00 -1.05456717e-01 -4.37341571e-01 2.20076069e-02
1.53785467e+00 1.71452872e-02 -1.04314022e-01 1.02917157e-01
8.90941858e-01 7.19789326e-01 8.24459374e-01 2.23905131e-01
5.62111318e-01 2.14420840e-01 6.43777788e-01 -5.36727250e-01
-7.79257864e-02 -2.09713742e-01 4.67236489e-02 8.88011813e-01
3.31931382e-01 -2.50470221e-01 -9.50398624e-01 5.48387110e-01
-1.74031091e+00 -5.89094341e-01 -4.94510621e-01 2.07310963e+00
1.04176998e+00 6.94422647e-02 -8.17827508e-02 2.64638305e-01
1.04680610e+00 -1.62168860e-01 -5.03374994e-01 4.44004536e-02
-1.28586426e-01 -2.41375919e-02 3.48818749e-01 4.50094640e-01
-1.55734038e+00 8.84919763e-01 7.28692389e+00 8.31711054e-01
-1.24250889e+00 5.18359661e-01 9.16098833e-01 2.89051771e-01
-2.18078151e-01 2.15767473e-02 -1.06614256e+00 3.40289384e-01
1.00234616e+00 1.22043900e-01 -1.28999367e-01 1.17153656e+00
-2.56405175e-01 -3.03476721e-01 -1.19240868e+00 9.83068228e-01
2.73497313e-01 -9.18491900e-01 7.27296397e-02 4.60761748e-02
1.06655765e+00 -2.31315047e-01 -3.03434152e-02 2.93888122e-01
5.83780885e-01 -7.68340528e-01 7.24138677e-01 4.97776866e-02
1.08787191e+00 -3.95195395e-01 7.58517265e-01 8.67811978e-01
-1.01267540e+00 -1.59566447e-01 -3.54872435e-01 -2.40547881e-02
5.12039475e-02 6.51922345e-01 -8.42726290e-01 3.26798886e-01
4.16982859e-01 8.36811602e-01 -7.72282004e-01 9.53996301e-01
-5.22541225e-01 6.67817354e-01 4.28504944e-02 2.64078557e-01
1.34685710e-01 2.86008324e-02 -4.38647926e-01 1.05991554e+00
2.61013538e-01 -1.18291214e-01 8.38152051e-01 3.21180254e-01
-2.51158178e-01 2.36712992e-01 -5.66412628e-01 1.89189926e-01
3.32794726e-01 1.76925290e+00 -9.15848017e-01 -7.65892148e-01
-8.20327997e-01 8.48183692e-01 4.49908614e-01 1.28580838e-01
-5.37156343e-01 1.31720141e-01 -1.61089450e-01 -2.38443837e-01
-1.76944345e-01 3.61972272e-01 -3.18204522e-01 -1.12091529e+00
-2.20247105e-01 -5.64095438e-01 4.93451178e-01 -8.19211423e-01
-1.41123676e+00 5.08679628e-01 -3.00754815e-01 -1.42649281e+00
-1.59046158e-01 -6.84513569e-01 -3.37934494e-01 7.97932982e-01
-1.83391058e+00 -1.43239963e+00 -3.56801182e-01 4.03458208e-01
4.49906558e-01 1.68242790e-02 1.17969942e+00 5.89128613e-01
-5.30510306e-01 6.00127101e-01 -4.67076316e-04 1.73238695e-01
9.79644597e-01 -1.24139321e+00 -1.91157490e-01 6.20155990e-01
2.37974793e-01 5.74120544e-02 2.81716555e-01 -5.61531901e-01
-5.54276586e-01 -1.38753593e+00 1.23572266e+00 -4.11542803e-01
5.59364200e-01 -1.33679835e-02 -1.01090240e+00 1.08198762e+00
1.39069527e-01 3.14003378e-01 1.23437691e+00 -1.26364827e-01
-3.59349042e-01 1.29240751e-01 -1.26246727e+00 7.96030313e-02
4.15957987e-01 -5.12635648e-01 -4.44579899e-01 8.17908525e-01
7.02066720e-01 -1.80118978e-01 -8.08130503e-01 3.80911678e-01
1.87173247e-01 -5.66859663e-01 6.77479267e-01 -4.00723696e-01
3.89113426e-01 -3.25118363e-01 -9.36004519e-02 -1.44387066e+00
-2.87806660e-01 2.82332361e-01 3.91965777e-01 1.20994508e+00
6.84720695e-01 -4.70525116e-01 8.27818096e-01 5.06740987e-01
-1.64640229e-02 -7.68696785e-01 -6.94817960e-01 -6.43415272e-01
8.82036686e-02 -1.02702655e-01 3.30089420e-01 1.44817162e+00
-2.32981414e-01 5.97793996e-01 -6.47382140e-01 6.26936182e-02
7.39218473e-01 2.48681575e-01 6.11197412e-01 -1.81040633e+00
8.81429911e-02 2.29577143e-02 2.84965355e-02 -8.71060252e-01
4.73601848e-01 -1.04857087e+00 2.45723218e-01 -1.43347585e+00
7.22389817e-01 -9.29528892e-01 -3.84027630e-01 9.96018648e-01
-2.82203943e-01 7.65923560e-01 1.24959610e-01 5.50230086e-01
-8.81332397e-01 8.85020718e-02 1.34884560e+00 -2.82833040e-01
2.11955875e-01 1.39552534e-01 -5.23232102e-01 1.01611996e+00
8.62524927e-01 -8.91457975e-01 -2.68524438e-01 -2.52115995e-01
2.92446584e-01 -1.93435512e-02 1.39363497e-01 -8.11127067e-01
2.68249154e-01 -5.33779740e-01 1.98922202e-01 -7.66441941e-01
3.66550893e-01 -1.02194786e+00 6.63429126e-02 4.79956478e-01
-6.10324681e-01 -1.30357593e-01 -2.91765302e-01 2.99907088e-01
-3.77891749e-01 -7.77848601e-01 1.02003026e+00 -3.78430754e-01
-9.30055261e-01 2.57112473e-01 -1.85495213e-01 -1.33846700e-01
1.36392915e+00 -3.08145024e-02 -2.73769975e-01 1.02993242e-01
-1.07441807e+00 3.53884190e-01 5.33212125e-01 1.75708473e-01
5.10688066e-01 -1.50161588e+00 -5.89649856e-01 2.51397789e-01
4.23151642e-01 1.10580355e-01 1.76507682e-01 4.81875300e-01
-3.79006304e-02 3.28813970e-01 -1.57141522e-01 -7.82075524e-01
-1.53121316e+00 7.47299254e-01 2.38512397e-01 -4.31867510e-01
-9.83095095e-02 8.37545514e-01 4.14949507e-01 -1.04450893e+00
1.96536019e-01 1.96111593e-02 -4.82725710e-01 2.64399618e-01
5.67565501e-01 -1.54302567e-01 -1.42793998e-01 -9.06834006e-01
-1.26211077e-01 7.37255096e-01 -9.92603600e-02 7.79367834e-02
1.14323628e+00 -1.18437029e-01 -4.25322384e-01 9.19417560e-01
1.21044004e+00 -3.44902724e-01 -1.04195368e+00 -5.45666397e-01
1.65837303e-01 -2.83735931e-01 -2.26125702e-01 -8.32109869e-01
-1.01964080e+00 1.01057613e+00 6.00614548e-01 2.25700274e-01
1.03811491e+00 2.46001244e-01 2.53890634e-01 2.80261755e-01
5.17065465e-01 -1.03542304e+00 5.38024902e-01 3.20640743e-01
3.98940593e-01 -1.87780726e+00 -3.29842329e-01 -6.73572242e-01
-6.88238204e-01 9.90381062e-01 9.93105114e-01 2.52986431e-01
7.72618055e-01 -1.18716411e-01 4.55553234e-01 -8.41319188e-02
-6.96806312e-01 -1.30625203e-01 1.21220767e-01 3.76085252e-01
3.99767369e-01 3.09568733e-01 -1.49622068e-01 2.07179829e-01
2.82060683e-01 4.30715270e-02 6.41042292e-01 8.39997888e-01
-6.86065972e-01 -1.45346487e+00 -4.27932084e-01 6.74893260e-01
-5.03036439e-01 1.09927848e-01 -3.32518429e-01 3.50731641e-01
4.79329914e-01 1.10397100e+00 -9.60002169e-02 -5.09081721e-01
1.07932284e-01 3.31046522e-01 1.78273201e-01 -1.27817070e+00
-4.68595237e-01 1.25488499e-02 -2.46252000e-01 -5.33669181e-02
-8.75026226e-01 -4.08442199e-01 -1.20957625e+00 9.65163186e-02
-6.23729348e-01 2.22533152e-01 6.65526032e-01 1.01241076e+00
-2.46417318e-02 3.58579218e-01 9.78833973e-01 -8.11800003e-01
-5.03524542e-01 -1.08469415e+00 -7.82649279e-01 5.62893093e-01
3.15485179e-01 -6.64516151e-01 -5.67567468e-01 3.35006118e-01] | [9.589408874511719, 4.051997184753418] |
bee1e2d7-e58c-4fec-b80d-b98f06ed340c | ct2-colorization-transformer-via-color-tokens | null | null | https://dl.acm.org/doi/abs/10.1007/978-3-031-20071-7_1 | https://ci.idm.pku.edu.cn/Weng_ECCV22b.pdf | CT2: Colorization Transformer via Color Tokens | Automatic image colorization is an ill-posed problem with multi-modal uncertainty, and there remains two main challenges with previous methods: incorrect semantic colors and under-saturation. In this paper, we propose an end-to-end transformer-based model to overcome these challenges. Benefited from the long-range context extraction of transformer and our holistic architecture, our method could colorize images with more diverse colors. Besides, we introduce color tokens into our approach and treat the colorization task as a classification problem, which increases the saturation of results. We also propose a series of modules to make image features interact with color tokens, and restrict the range of possible color candidates, which makes our results visually pleasing and reasonable. In addition, our method does not require any additional external priors, which ensures its well generalization capability. Extensive experiments and user studies demonstrate that our method achieves superior performance than previous works. | ['Boxin Shi', 'Si Li', 'Yu Li', 'Jimeng Sun', 'Shuchen Weng'] | 2022-10-23 | null | null | null | eccv-2022-10 | ['colorization'] | ['computer-vision'] | [ 5.18438630e-02 -2.58570015e-01 1.98379196e-02 -3.94986928e-01
-5.85398972e-01 -6.53388202e-01 4.00146484e-01 -3.86457205e-01
-2.82644868e-01 3.72289717e-01 9.91544724e-02 -1.86770663e-01
2.47594103e-01 -7.03548312e-01 -5.34391224e-01 -5.78060269e-01
5.40324450e-01 -2.03269586e-01 5.73839962e-01 -1.07776307e-01
2.39257231e-01 2.96446711e-01 -1.50149536e+00 4.48084831e-01
1.23170936e+00 1.07447958e+00 2.93730736e-01 3.02722603e-01
-2.71051645e-01 6.32561386e-01 -4.71208006e-01 -5.23329556e-01
1.75588712e-01 -2.24704117e-01 -6.95493817e-01 3.24997187e-01
4.70516413e-01 -4.87010747e-01 -1.48459512e-03 1.34332645e+00
3.74933809e-01 -1.55364443e-02 3.88516277e-01 -1.43553030e+00
-1.16061282e+00 4.75062609e-01 -8.86553407e-01 -4.74986136e-01
2.43511051e-01 1.29343614e-01 9.48245227e-01 -1.22029245e+00
2.46063396e-01 1.45773041e+00 4.88156170e-01 5.62801242e-01
-1.01709878e+00 -9.17220414e-01 7.75560319e-01 3.03810507e-01
-1.42081046e+00 -2.66857207e-01 9.15132642e-01 -1.03316754e-01
2.87230432e-01 4.00709718e-01 8.53013754e-01 9.28060174e-01
-4.08012062e-01 1.16027462e+00 1.47230077e+00 -3.90179068e-01
2.56857902e-01 2.45436549e-01 -3.03774744e-01 7.73770273e-01
1.15480363e-01 -2.89278738e-02 -3.86100352e-01 6.70710951e-02
9.04492080e-01 1.65793791e-01 -2.74621099e-01 -6.27388775e-01
-1.29270411e+00 3.87999803e-01 7.04677045e-01 7.12164715e-02
-3.33506204e-02 3.82506579e-01 -3.01488023e-02 -2.82678813e-01
3.41278732e-01 1.55963078e-01 -3.37366462e-01 3.82537805e-02
-8.26555669e-01 -1.05250828e-01 1.19122498e-01 1.21277153e+00
8.73942137e-01 -1.09424435e-01 -2.19748527e-01 1.05075967e+00
5.14234245e-01 7.32059836e-01 4.47308682e-02 -8.08598757e-01
2.71003127e-01 5.77246964e-01 2.02652350e-01 -1.07543445e+00
-2.19511971e-01 -3.07845145e-01 -6.43507123e-01 3.63617480e-01
3.64639103e-01 6.18235879e-02 -1.18597174e+00 1.73728657e+00
2.91158855e-01 1.83960930e-01 -1.83878303e-01 1.21856129e+00
5.33470213e-01 5.10749519e-01 3.69475514e-01 1.70114711e-01
1.50475132e+00 -1.21027815e+00 -7.53008842e-01 -3.47843587e-01
9.44349170e-02 -9.77215111e-01 1.52774119e+00 5.90084195e-01
-7.95557976e-01 -5.18430233e-01 -1.02040184e+00 -2.81027287e-01
-4.87176359e-01 3.33766311e-01 9.65063393e-01 7.77656317e-01
-9.68489110e-01 9.29907933e-02 -6.76953018e-01 -2.39457563e-01
4.68397617e-01 -5.82334287e-02 -5.48622943e-02 -2.59178609e-01
-1.12294459e+00 6.93417370e-01 4.65839624e-01 3.74402046e-01
-4.07257795e-01 -3.69872868e-01 -6.83867216e-01 2.01949123e-02
5.04829943e-01 -5.85590839e-01 1.08518410e+00 -1.17854857e+00
-1.65203774e+00 4.14058566e-01 -1.68885246e-01 3.93140048e-01
6.93691730e-01 -3.45383823e-01 -4.01993036e-01 9.69355106e-02
-7.33110076e-03 8.94907713e-01 8.37501764e-01 -1.74968481e+00
-7.60005713e-01 -1.96204245e-01 4.16528970e-01 3.45293522e-01
-7.25387573e-01 -1.72983810e-01 -1.34178782e+00 -8.41319501e-01
3.22722495e-01 -8.61836612e-01 -2.35531896e-01 3.42448354e-01
-5.06117523e-01 7.80887678e-02 6.20128572e-01 -5.05932391e-01
1.14476955e+00 -2.31982040e+00 -5.54330572e-02 3.51509720e-01
1.59496188e-01 -8.86034593e-02 -2.25480527e-01 1.31159917e-01
1.44943237e-01 3.34447235e-01 -2.65350968e-01 -2.35648334e-01
1.42550200e-01 2.83134095e-02 -3.02239895e-01 1.26822025e-01
3.16535652e-01 8.08949590e-01 -1.00244975e+00 -7.89036334e-01
4.00621474e-01 5.42504013e-01 -4.45560485e-01 1.00941062e-01
-3.40888649e-01 2.27277279e-01 -6.99457169e-01 9.23707724e-01
1.24216914e+00 -1.92977399e-01 1.75485268e-01 -7.70292342e-01
-9.01496261e-02 -2.33745158e-01 -1.35750937e+00 1.96913564e+00
-5.16097128e-01 2.30266482e-01 -1.91204727e-01 -4.05106425e-01
6.48785114e-01 -1.61722809e-01 1.53289750e-01 -9.58001673e-01
4.26968597e-02 -9.11587700e-02 -4.36642915e-01 -4.38757002e-01
6.96480811e-01 1.07049933e-02 -5.77249676e-02 2.79796988e-01
-3.45934540e-01 -8.40232670e-02 5.67694716e-02 3.96497577e-01
4.01327074e-01 7.41477370e-01 -1.98723003e-01 -1.77013665e-01
3.13120037e-01 -1.46069095e-01 7.95639455e-01 4.97933358e-01
-1.03840217e-01 9.23787534e-01 4.47134256e-01 -1.26524195e-02
-6.80260062e-01 -1.15848458e+00 -2.03934610e-02 1.13534760e+00
8.38083148e-01 -4.13089126e-01 -7.12216318e-01 -8.15109909e-01
-1.57281667e-01 5.95872641e-01 -5.55496156e-01 -1.38185099e-02
-2.88791120e-01 -8.38186383e-01 2.91298747e-01 6.64906919e-01
6.64653540e-01 -6.19672000e-01 -3.24173927e-01 -1.05463155e-01
-5.24152160e-01 -1.21054041e+00 -7.61207938e-01 -2.81344444e-01
-5.58878303e-01 -1.13646424e+00 -8.04143906e-01 -7.54809856e-01
9.71044838e-01 6.71605110e-01 8.82212639e-01 4.81363907e-02
-2.55540401e-01 3.24536532e-01 -4.90318060e-01 -3.43138188e-01
8.67355913e-02 -2.88885146e-01 -3.09318423e-01 2.06994489e-01
2.36355126e-01 -3.15483004e-01 -9.64699447e-01 2.54017800e-01
-1.15603006e+00 5.85546374e-01 8.27360272e-01 6.85422421e-01
6.66378260e-01 -1.36531005e-02 3.22933853e-01 -9.41816568e-01
5.24418354e-01 -6.98273012e-04 -6.56993747e-01 6.96566164e-01
-7.18814552e-01 1.88559607e-01 6.18110359e-01 -4.87514526e-01
-1.48233211e+00 2.87904352e-01 2.59281963e-01 -3.59182924e-01
4.15390395e-02 2.50803590e-01 -5.65103173e-01 -1.15302496e-01
3.25019658e-01 1.71840519e-01 -1.13914490e-01 -5.15045464e-01
1.05409729e+00 5.93435228e-01 4.40292031e-01 -7.35292435e-01
9.71425474e-01 6.80721641e-01 -3.78632843e-01 -4.10651952e-01
-7.92082548e-01 -1.36969045e-01 -3.78113955e-01 -4.07943964e-01
8.41834009e-01 -1.07660460e+00 -8.28831434e-01 5.63897550e-01
-9.15610373e-01 -7.52367526e-02 2.29101688e-01 2.14853793e-01
-1.21998295e-01 7.56670952e-01 -5.54982007e-01 -8.12305927e-01
-1.82413206e-01 -1.23381031e+00 1.08801639e+00 5.30630052e-01
3.48536581e-01 -6.95714653e-01 -4.83094633e-01 2.55197197e-01
5.88299453e-01 1.72668085e-01 8.29253793e-01 2.53231287e-01
-9.46395874e-01 2.15782508e-01 -8.03287208e-01 2.91496903e-01
1.97658539e-01 2.81331718e-01 -1.15344310e+00 -1.65978909e-01
-4.80741382e-01 -2.98486352e-01 9.52799857e-01 -1.00121222e-01
1.66573465e+00 3.83647494e-02 -2.75757998e-01 6.80570483e-01
1.53280544e+00 -2.20606662e-02 6.68979585e-01 4.31568474e-01
1.01315463e+00 5.11725724e-01 8.96430910e-01 3.98133844e-01
6.62826955e-01 5.14029205e-01 5.72196364e-01 -7.96548009e-01
-2.48257145e-01 -3.20523560e-01 1.70618668e-01 5.36718607e-01
8.94749165e-02 -1.69284105e-01 -5.72092414e-01 3.83778423e-01
-1.95859826e+00 -8.18051040e-01 2.80217677e-02 1.97999990e+00
9.37524259e-01 -1.00805471e-02 3.86650264e-02 -2.33012497e-01
7.26581812e-01 9.49081182e-02 -6.30997181e-01 1.64887533e-02
-1.46470368e-01 -2.64126230e-02 5.23321867e-01 3.20018977e-01
-1.10999823e+00 1.13835692e+00 6.23087120e+00 9.00864482e-01
-1.22610867e+00 -1.80742711e-01 7.03232884e-01 6.96933083e-03
-8.04940522e-01 2.04553708e-01 -2.82571375e-01 5.03735244e-01
-1.64322823e-01 2.21471339e-01 5.46348810e-01 6.89312637e-01
6.79395720e-02 -2.13762894e-01 -8.79725039e-01 1.20087194e+00
1.31857842e-01 -9.14044738e-01 2.01853454e-01 -2.65215933e-01
5.62824011e-01 -3.86902660e-01 4.07448739e-01 1.36737630e-01
4.74456370e-01 -7.78602004e-01 1.21958089e+00 4.66092914e-01
1.05442178e+00 -7.49739230e-01 1.86760291e-01 -1.45378590e-01
-1.27299154e+00 4.26758528e-02 -2.59402126e-01 3.09440345e-01
9.49049965e-02 6.92042530e-01 -3.13941270e-01 6.92797005e-01
8.55598211e-01 6.92138672e-01 -9.80764091e-01 1.01335812e+00
-5.50887585e-01 3.38264406e-01 -3.05070072e-01 -1.22088358e-01
3.35064977e-02 -2.42220029e-01 -1.28269240e-01 1.33318853e+00
2.25969508e-01 1.50709778e-01 3.65845114e-01 1.04012537e+00
6.93627447e-02 2.70848330e-02 -6.45880625e-02 2.09540814e-01
7.06009924e-01 1.54497087e+00 -7.89537549e-01 -2.60684013e-01
-5.74125707e-01 1.31084907e+00 2.58431405e-01 7.19185472e-01
-1.16316986e+00 -5.90675294e-01 5.01962364e-01 -2.37212911e-01
2.31879786e-01 -1.00476339e-01 -3.37110847e-01 -1.49564064e+00
2.63231099e-01 -7.95327723e-01 2.76666909e-01 -1.07937348e+00
-1.42300045e+00 5.83751976e-01 -1.13464948e-02 -1.59258509e+00
3.61398309e-01 -6.76831424e-01 -3.77304167e-01 8.06741178e-01
-1.89265847e+00 -1.74185097e+00 -6.34912550e-01 6.56085730e-01
2.11286753e-01 4.35870588e-01 5.78033090e-01 3.78364891e-01
-5.37286520e-01 7.30196595e-01 -4.22219224e-02 1.07275836e-01
1.07149565e+00 -1.32823706e+00 2.67002553e-01 1.15686452e+00
-8.50341991e-02 6.88515544e-01 5.25934339e-01 -4.05614883e-01
-1.45325577e+00 -1.01652539e+00 2.03814447e-01 -2.46589288e-01
4.69451368e-01 -5.88435471e-01 -6.08671308e-01 3.14779282e-01
2.72999823e-01 -9.79009084e-03 5.27407646e-01 2.69402862e-01
-8.02009642e-01 -3.09612513e-01 -8.91322017e-01 9.66912091e-01
1.13569438e+00 -5.24574876e-01 -2.73507684e-01 1.59827814e-01
6.14360869e-01 -4.62153226e-01 -4.65950012e-01 3.34674239e-01
6.95794225e-01 -1.06851411e+00 1.06237888e+00 -7.52634406e-02
2.15980619e-01 -8.91625822e-01 5.71186058e-02 -1.17701268e+00
-4.32666749e-01 -2.58903950e-01 3.41609418e-01 1.48852944e+00
4.93163973e-01 -5.40139318e-01 4.70583349e-01 8.83090496e-01
-2.80947355e-03 -5.66185772e-01 -4.48566139e-01 -5.38826406e-01
-8.81381631e-02 -4.38857615e-01 8.46850693e-01 9.36684847e-01
2.46385168e-02 9.77922902e-02 -5.88385582e-01 3.16270500e-01
6.99579656e-01 5.38356721e-01 5.83063722e-01 -8.49796832e-01
-2.42403075e-01 -4.75495130e-01 -1.07960356e-02 -1.19905829e+00
-2.47893915e-01 -5.00184476e-01 2.32970685e-01 -1.57847142e+00
4.62019712e-01 -7.72411108e-01 -6.87776506e-01 8.19769442e-01
-5.27730525e-01 4.47815120e-01 3.46948743e-01 7.44551495e-02
-8.90186787e-01 6.91713631e-01 1.36662900e+00 -3.36218357e-01
8.43790080e-03 -4.44954664e-01 -1.07514322e+00 7.16193080e-01
6.66240513e-01 -1.38162091e-01 -6.45326078e-01 -8.55524778e-01
2.41739884e-01 -3.28329176e-01 3.22085202e-01 -6.72368884e-01
8.15431252e-02 -5.28864145e-01 7.49607325e-01 -5.60980141e-01
4.62965339e-01 -9.36088979e-01 -6.59720004e-02 8.15316364e-02
-1.27010792e-01 3.14103253e-02 1.86051175e-01 5.84629774e-01
-1.79814726e-01 2.73523629e-01 8.28344166e-01 -8.52661133e-02
-9.89968240e-01 2.71332055e-01 -9.07164365e-02 -2.06922382e-01
9.88946617e-01 -2.31343359e-01 -4.32935119e-01 -2.90259272e-01
-4.30996060e-01 4.57221240e-01 8.87933075e-01 7.14034379e-01
7.88770914e-01 -1.53637493e+00 -4.04763699e-01 1.81295604e-01
4.88993496e-01 5.22326911e-03 4.39248294e-01 5.66907883e-01
-4.54306811e-01 -1.10286437e-01 -7.06207752e-02 -4.96892244e-01
-1.06245434e+00 7.72679448e-01 1.38729244e-01 3.52720499e-01
-5.39854884e-01 7.31250405e-01 4.18445081e-01 -2.72695750e-01
2.77278811e-01 -4.54201162e-01 -1.31239101e-01 3.50692775e-03
4.69652921e-01 -4.36782464e-02 -2.52679437e-01 -3.51731777e-01
-5.00031710e-01 7.80310988e-01 -2.46186525e-01 -2.38387451e-01
1.03654456e+00 -4.60495621e-01 -1.95802286e-01 2.08284020e-01
8.68269265e-01 3.23648304e-01 -1.31358039e+00 -2.99450606e-01
-2.68062413e-01 -7.49053121e-01 7.92058781e-02 -1.22667217e+00
-1.24615431e+00 8.46433222e-01 7.21784413e-01 -2.28535347e-02
1.49000025e+00 -1.96925253e-01 8.04906011e-01 1.39934439e-02
2.56790072e-01 -1.26590705e+00 2.09765375e-01 2.22502172e-01
6.66633785e-01 -1.33087695e+00 5.65762296e-02 -7.64083266e-01
-8.86099100e-01 1.13170469e+00 1.00494516e+00 2.32935533e-01
3.17356706e-01 8.44054818e-02 4.80834246e-01 1.64302677e-01
-4.16707426e-01 -4.67779398e-01 3.52928758e-01 5.48335552e-01
4.16564941e-01 3.01517863e-02 -2.93480426e-01 5.20812750e-01
1.82300523e-01 -1.31752044e-01 3.58278126e-01 7.73734450e-01
-3.62928122e-01 -1.25477052e+00 -3.49126309e-01 -5.01664504e-02
-1.13261171e-01 -3.17306817e-01 -3.20078641e-01 6.56693637e-01
2.01667383e-01 1.07127547e+00 -3.06975901e-01 -5.39407730e-01
3.43422204e-01 -2.33531281e-01 4.31552380e-01 -2.48375714e-01
-7.86672458e-02 4.04824644e-01 -1.83970064e-01 -7.44671166e-01
-4.44422364e-01 -3.19936365e-01 -1.45987213e+00 -1.03960425e-01
-6.01469278e-01 -2.44970173e-02 6.84409738e-01 6.39013767e-01
4.95173961e-01 5.82962513e-01 6.49736822e-01 -6.09675646e-01
-3.18244815e-01 -7.17688441e-01 -3.92201275e-01 7.03330576e-01
1.57037884e-01 -6.48857296e-01 -1.04098268e-01 4.67398353e-02] | [11.39420223236084, -1.030258297920227] |
2a6f7097-c0d8-4ff9-864b-0a2c220f3a08 | how-to-learn-from-unlabeled-volume-data-self | null | null | https://openreview.net/forum?id=rJlYmEE46N | https://openreview.net/pdf?id=rJlYmEE46N | How to learn from unlabeled volume data: Self-Supervised 3D Context Feature Learning | The vast majority of 3D medical images lacks detailed image-based expert annotations. The ongoing advances of deep convolutional neural networks clearly demonstrate the benefit of supervised learning to successfully extract relevant anatomical information and aid image-based analysis and interventions, but it heavily relies on labeled data. Self-supervised learning, that requires no expert labels, provides an appealing way to discover data-inherent patterns and leverage anatomical information freely available from medical images themselves.
In this work, we propose a new approach to train effective convolutional feature extractors based on a new concept of image-intrinsic spatial offset relations with an auxiliary heatmap regression loss. The learned features successfully capture semantic, anatomical information and enable state-of-the-art accuracy for a k-NN based one-shot segmentation task without any subsequent fine-tuning. | ['Anonymous'] | 2019-05-23 | null | null | null | null | ['one-shot-segmentation'] | ['computer-vision'] | [ 4.94905800e-01 4.28692073e-01 -5.37273169e-01 -6.71509147e-01
-1.16924381e+00 -4.61158037e-01 4.08265859e-01 4.91116345e-01
-6.53968930e-01 5.77239513e-01 1.92494810e-01 -6.03507906e-02
-4.78979558e-01 -5.39917767e-01 -5.03493071e-01 -8.25376332e-01
-3.23247671e-01 3.38771582e-01 2.67267942e-01 -1.36793360e-01
7.27344975e-02 6.45730853e-01 -1.17178547e+00 2.16939509e-01
7.71929145e-01 1.28719711e+00 2.25843936e-01 4.87879038e-01
-2.28095055e-01 8.27446401e-01 -1.77458525e-01 -1.24973923e-01
2.88309544e-01 -2.43929893e-01 -1.05057466e+00 -9.86405313e-02
3.50803196e-01 -2.06322283e-01 -3.86397541e-01 7.89719999e-01
6.79081619e-01 -3.38416874e-01 8.64000797e-01 -6.27326608e-01
-3.89487237e-01 3.22106749e-01 -2.66497225e-01 4.62579191e-01
-1.20986894e-01 3.91406327e-01 7.83599198e-01 -5.74604690e-01
8.50243211e-01 3.64737689e-01 9.99666393e-01 4.33476567e-01
-1.45271301e+00 -2.07642034e-01 -5.41180789e-01 -2.82294787e-02
-1.21685398e+00 -2.23969176e-01 9.67975318e-01 -8.25565815e-01
6.94931865e-01 3.16707581e-01 7.39494383e-01 9.96437609e-01
3.53945911e-01 6.11757159e-01 1.33255839e+00 -4.92903680e-01
1.61911160e-01 3.92178725e-03 8.61842483e-02 1.11406279e+00
-6.15491942e-02 2.74369419e-01 -4.07979846e-01 4.53142039e-02
8.49878728e-01 2.30429530e-01 -7.91146755e-02 -8.18166196e-01
-1.23007452e+00 7.98038065e-01 1.07283688e+00 5.00733018e-01
-4.00070518e-01 2.18473613e-01 4.66295809e-01 -6.94266185e-02
5.13392270e-01 8.08205009e-01 -3.95567626e-01 -7.70806521e-02
-1.14622843e+00 -2.55544782e-01 3.19080710e-01 5.05567193e-01
7.35561848e-01 -2.75247067e-01 -2.76359409e-01 8.04435372e-01
-2.62484048e-02 3.67484093e-01 6.03961170e-01 -7.51938343e-01
-3.96084189e-02 9.52402234e-01 -3.49438578e-01 -8.55477393e-01
-1.04512072e+00 -6.47589803e-01 -8.93144906e-01 3.95027280e-01
4.22916114e-01 1.09935112e-01 -1.33801472e+00 1.31153846e+00
3.76099050e-01 -2.37171248e-01 -3.87852371e-01 9.65964913e-01
9.41431761e-01 -1.82885557e-01 1.68470755e-01 -8.88113678e-02
1.11202323e+00 -8.45641911e-01 -4.90736723e-01 -1.36158451e-01
7.90705860e-01 -4.65512663e-01 1.16558576e+00 1.69383362e-01
-9.47768509e-01 -2.51692742e-01 -9.70100343e-01 -5.14625534e-02
-4.18818653e-01 1.81364715e-01 1.10984004e+00 6.62918448e-01
-9.62372959e-01 7.52453864e-01 -1.10595679e+00 -1.59680083e-01
1.21872878e+00 6.07807696e-01 -6.51125669e-01 -5.50551456e-04
-8.33469927e-01 1.03680456e+00 1.38786167e-01 1.56186596e-01
-1.00685048e+00 -1.07449889e+00 -8.71802449e-01 -2.19529480e-01
4.00587142e-01 -6.71490312e-01 1.04541576e+00 -9.25644577e-01
-1.44215250e+00 1.62043881e+00 2.53807366e-01 -4.94797319e-01
6.44348443e-01 -9.53409225e-02 -5.88015877e-02 6.46974623e-01
2.05040261e-01 6.03930831e-01 8.91807139e-01 -1.11795974e+00
-1.83345214e-01 -7.01033950e-01 -2.44810849e-01 -1.23109557e-01
-4.34811711e-01 -1.82968289e-01 -1.29913881e-01 -7.95833230e-01
2.69049704e-01 -8.56735826e-01 -6.14140153e-01 5.18923581e-01
-4.68014747e-01 2.59054750e-01 3.31515729e-01 -5.26770711e-01
9.64282274e-01 -1.94893575e+00 1.81765452e-01 3.38427573e-01
6.50229394e-01 1.75566152e-01 2.53213078e-01 1.02715388e-01
-8.46294612e-02 3.37082520e-02 -6.57650888e-01 -3.92377749e-02
-3.00888687e-01 1.90082595e-01 1.95376903e-01 6.13633692e-01
4.15803343e-01 1.32666528e+00 -9.98218298e-01 -7.40984738e-01
5.90892494e-01 4.19012368e-01 -3.87510508e-01 2.47998282e-01
8.04811120e-02 9.21601117e-01 -4.76539284e-01 6.23843789e-01
2.89600849e-01 -5.99863946e-01 1.63732301e-02 -4.96564746e-01
4.12428230e-02 -2.84922756e-02 -3.10663402e-01 2.57659149e+00
-6.56885087e-01 5.33817351e-01 4.39888798e-02 -1.33924818e+00
7.30978847e-01 7.95519724e-02 1.09248102e+00 -8.21076930e-01
3.18722397e-01 4.22905862e-01 -1.68223768e-01 -7.70468414e-01
-1.86773971e-01 -4.01037931e-01 -1.88954234e-01 1.08126171e-01
5.51855743e-01 -6.27904236e-02 -2.98663765e-01 1.51134962e-02
1.41882896e+00 1.45309895e-01 3.14826488e-01 -5.29097497e-01
3.23233008e-01 2.96748012e-01 2.75208831e-01 8.16413045e-01
-2.60416776e-01 8.36821735e-01 2.68023223e-01 -7.84342110e-01
-9.65108335e-01 -1.17256737e+00 -4.77510333e-01 8.17407072e-01
5.38349040e-02 -2.03247190e-01 -6.73113883e-01 -1.07441723e+00
-6.14792891e-02 1.27459047e-02 -1.15255868e+00 -3.23549718e-01
-4.95434552e-01 -7.49466479e-01 5.02156317e-01 6.92328811e-01
1.92175165e-01 -8.96626890e-01 -1.01054096e+00 2.08722800e-01
-5.32315001e-02 -1.08814192e+00 -1.55988321e-01 6.32649660e-01
-1.17542338e+00 -1.13406730e+00 -1.01317728e+00 -6.93559051e-01
1.04919624e+00 -2.10011780e-01 1.06059813e+00 4.66116332e-02
-1.24818218e+00 2.86769241e-01 -3.62899631e-01 -2.73042649e-01
-1.70496657e-01 5.10611296e-01 -3.88972133e-01 -2.11163819e-01
3.94159667e-02 -8.02195311e-01 -9.98718262e-01 9.33067724e-02
-7.21686304e-01 1.62673801e-01 1.00544989e+00 1.15894961e+00
8.94194782e-01 -3.96612525e-01 3.89487177e-01 -1.12881148e+00
1.26097500e-01 -2.05572337e-01 -1.81153551e-01 2.83493191e-01
-6.75265491e-01 3.11726928e-01 4.45490867e-01 -1.63145497e-01
-7.63630152e-01 4.78220552e-01 -2.58457005e-01 -4.11981136e-01
-4.14047420e-01 5.38483322e-01 3.22869241e-01 -5.25596499e-01
1.01140261e+00 5.51375747e-02 2.74800003e-01 -3.21551979e-01
3.24686021e-01 3.09668839e-01 8.02615881e-01 -3.37303400e-01
5.56377888e-01 9.54293966e-01 2.38785401e-01 -5.49833715e-01
-1.22207987e+00 -7.91759014e-01 -1.22122300e+00 -2.67001897e-01
1.27897561e+00 -5.70601463e-01 -4.37415183e-01 2.29852259e-01
-7.00013459e-01 -5.37493050e-01 -7.12866187e-01 5.65765202e-01
-7.48233616e-01 1.46698430e-01 -5.46686113e-01 -3.62811863e-01
-6.35433018e-01 -1.24316370e+00 1.26361334e+00 1.65448070e-01
-2.11446092e-01 -9.27038133e-01 1.20592676e-01 4.99789327e-01
6.78804696e-01 6.84525967e-01 9.63787973e-01 -5.96474886e-01
-5.24824142e-01 -3.99934560e-01 -2.91620523e-01 2.86012620e-01
4.09379572e-01 -4.28361893e-01 -9.72510695e-01 -1.34581923e-01
-2.91410357e-01 -4.37159419e-01 8.52570593e-01 7.60374963e-01
1.43433321e+00 -1.26760110e-01 -2.85029441e-01 1.00637853e+00
1.54976702e+00 -3.36265504e-01 5.95176935e-01 1.95179880e-01
8.51159692e-01 4.89357740e-01 2.27918893e-01 1.68637812e-01
1.35636538e-01 5.95558882e-01 4.72011596e-01 -7.12544501e-01
-3.06654543e-01 -8.98273811e-02 -5.01401901e-01 7.21943617e-01
-1.26563936e-01 6.29616916e-01 -1.12777066e+00 6.10569000e-01
-1.68692684e+00 -6.29643440e-01 8.63442272e-02 2.16674089e+00
9.85658467e-01 2.79352188e-01 -1.16899395e-02 2.54347455e-03
2.54805148e-01 -1.38595412e-02 -6.14478707e-01 1.48935944e-01
1.44944251e-01 7.12963164e-01 9.36284065e-01 1.25850752e-01
-1.42274654e+00 8.15396965e-01 6.94385004e+00 7.86079049e-01
-1.24576378e+00 3.16143095e-01 7.27827191e-01 -1.83155704e-02
-1.41802803e-01 -1.46253228e-01 -1.39406979e-01 2.46033922e-01
7.46439457e-01 2.31284887e-01 8.38345811e-02 8.98740590e-01
1.07397847e-01 -2.02955067e-01 -9.29916084e-01 1.13054168e+00
4.24782634e-02 -1.77174783e+00 -3.01447749e-01 6.04714453e-02
6.21669292e-01 1.15250036e-01 3.40997591e-03 -1.40265346e-01
-7.51150250e-02 -1.32143545e+00 2.76698709e-01 7.39550948e-01
1.14943290e+00 -4.83464599e-01 6.62647903e-01 -3.97535414e-02
-7.31343687e-01 -4.57186140e-02 -1.08371548e-01 3.11115682e-01
4.13781032e-02 5.76031685e-01 -8.60376000e-01 5.68573296e-01
7.59563267e-01 7.54505217e-01 -7.43377566e-01 1.21133459e+00
-4.88314070e-02 5.57483077e-01 -2.08186656e-01 2.80292809e-01
3.54498059e-01 2.97808677e-01 3.29586089e-01 1.21455038e+00
-1.66471690e-01 -1.16657401e-02 7.84752518e-02 8.45888615e-01
2.00210232e-02 2.62118638e-01 -4.32069778e-01 8.19863155e-02
-1.75442263e-01 1.62711084e+00 -1.18324804e+00 -7.54097030e-02
-1.14958353e-01 9.95152473e-01 3.56373668e-01 2.69060169e-04
-5.28228104e-01 -1.53627038e-01 1.34571746e-01 4.68922496e-01
1.84590816e-01 -2.21892625e-01 -6.62315845e-01 -8.82282734e-01
-1.51152432e-01 -3.61395121e-01 4.17791516e-01 -5.04299164e-01
-1.49578094e+00 5.21537185e-01 -2.24791884e-01 -1.16999829e+00
-1.92771256e-01 -7.42650867e-01 -4.01327372e-01 5.47895849e-01
-1.50020742e+00 -1.54686916e+00 -6.10437632e-01 4.93847907e-01
1.71567842e-01 -1.45942166e-01 1.19001198e+00 3.07379693e-01
-1.58822104e-01 5.29949427e-01 1.62151411e-01 3.68869752e-01
7.32105672e-01 -1.49980247e+00 -1.86829105e-01 3.55028808e-01
1.30645499e-01 3.49667221e-01 4.39455509e-01 -3.65670711e-01
-1.32060170e+00 -7.22927213e-01 1.62171245e-01 -4.80668783e-01
6.18898988e-01 -1.54920757e-01 -6.50933087e-01 2.43536294e-01
-1.46072760e-01 7.68212914e-01 9.83655274e-01 4.00099121e-02
-2.35800087e-01 -3.15801561e-01 -1.27267575e+00 8.12667459e-02
1.09821987e+00 -5.97026587e-01 -6.31887078e-01 4.38838989e-01
5.00278950e-01 -4.98941451e-01 -1.28610134e+00 7.04158545e-01
5.94588578e-01 -9.07275021e-01 1.20698690e+00 -7.52373815e-01
6.03140235e-01 7.28739724e-02 1.31669089e-01 -1.04509926e+00
-1.72084630e-01 -4.04279679e-01 4.97143082e-02 4.90556508e-01
5.05320132e-01 -2.26972669e-01 1.07983077e+00 6.47193015e-01
-4.02308911e-01 -1.06196702e+00 -1.07222128e+00 -4.88523960e-01
2.13691801e-01 -2.81937391e-01 1.34580163e-02 1.10373223e+00
8.85553211e-02 -2.47089677e-02 -1.67233720e-01 -1.75112903e-01
8.40227008e-01 7.25779533e-02 4.39292043e-01 -1.28169632e+00
-2.08767340e-01 -5.95682919e-01 -8.71823609e-01 -2.85492569e-01
-4.35093231e-02 -1.29094040e+00 -4.79536783e-03 -1.70185542e+00
3.01969200e-01 -7.09299982e-01 -6.87318027e-01 5.62648296e-01
-1.04237227e-02 5.84171832e-01 -2.87579745e-01 2.60405511e-01
-7.56974816e-01 3.47652137e-01 1.29238093e+00 -2.12864280e-01
-4.63555455e-02 -5.89495599e-02 -4.95288223e-01 6.46760106e-01
6.36444092e-01 -6.87972188e-01 -2.47870311e-01 -2.96021849e-01
-5.16545773e-02 -1.23646028e-01 7.86819339e-01 -1.02373052e+00
2.13399231e-01 1.25565261e-01 7.04285443e-01 -2.98840106e-01
8.53621587e-02 -9.72215235e-01 -1.82184324e-01 5.11024117e-01
-5.97255528e-01 -4.30333257e-01 2.10448533e-01 5.15038490e-01
-2.11310625e-01 -6.97044730e-02 7.50384569e-01 -3.61748129e-01
-6.30890489e-01 4.96152043e-01 -5.55702597e-02 2.43963614e-01
1.04645908e+00 -2.35438779e-01 1.04411580e-01 1.31954746e-02
-1.12690985e+00 -1.23939358e-01 2.87314087e-01 1.93423703e-01
5.86552322e-01 -1.30917060e+00 -4.92074817e-01 8.40371549e-02
3.94979119e-01 1.10674590e-01 7.18982697e-01 1.27130485e+00
-6.39419436e-01 3.19078922e-01 -5.01012385e-01 -1.08412933e+00
-8.79430950e-01 3.25665504e-01 5.59103847e-01 -6.24572158e-01
-1.01291680e+00 7.99103737e-01 9.41862091e-02 -6.13623559e-01
4.39951718e-02 -2.63975561e-01 -1.43787879e-02 -1.17413372e-01
3.22524965e-01 -1.26684919e-01 3.73818874e-01 -4.47867543e-01
-3.96583080e-01 6.01905823e-01 -4.73455042e-02 9.66633260e-02
1.60807526e+00 1.31928936e-01 1.39610261e-01 4.72394675e-01
1.51046515e+00 -4.20021772e-01 -1.28386545e+00 -2.96668231e-01
2.44326666e-01 -4.66938943e-01 3.61635625e-01 -1.20122111e+00
-1.20918596e+00 1.12235773e+00 1.07611585e+00 -2.00536519e-01
1.09110034e+00 2.61305392e-01 8.23128641e-01 1.02353401e-01
4.12899047e-01 -1.06843245e+00 2.74086088e-01 -1.25048101e-01
5.84593832e-01 -1.65121996e+00 3.68623063e-02 -3.76960248e-01
-5.74702382e-01 1.06056035e+00 4.13033634e-01 -1.34905636e-01
8.15057099e-01 4.54516530e-01 2.01572731e-01 -6.40519619e-01
-8.73981491e-02 -4.95350808e-01 6.70412123e-01 7.59828210e-01
3.60548466e-01 5.46016097e-02 -2.95423418e-01 7.37648904e-01
2.97417846e-02 1.97428927e-01 -5.64206280e-02 9.92504597e-01
-2.41288349e-01 -9.66623604e-01 1.59094527e-01 8.60797942e-01
-7.13001192e-01 -3.36452611e-02 -1.60633162e-01 6.68090224e-01
7.72738457e-03 1.80480927e-01 -1.72502711e-01 -2.67139554e-01
2.55061507e-01 -1.03494324e-01 7.87679374e-01 -5.82845747e-01
-7.21633136e-01 4.19886969e-02 -2.25343555e-01 -7.22306550e-01
-5.27716756e-01 -2.27751523e-01 -1.29410207e+00 3.91720593e-01
-2.37621516e-01 -3.18661004e-01 8.03823173e-01 1.02945805e+00
3.24479133e-01 7.76792347e-01 5.23679316e-01 -9.39572096e-01
-2.74850160e-01 -7.21581042e-01 -3.00141186e-01 6.54749870e-01
3.25973481e-01 -7.98413396e-01 3.54254013e-03 -7.06049055e-03] | [14.758092880249023, -2.3271780014038086] |
ffccc4b4-ebfe-4b1f-ad7a-ead6c67b4d4f | offboard-3d-object-detection-from-point-cloud | 2103.05073 | null | https://arxiv.org/abs/2103.05073v1 | https://arxiv.org/pdf/2103.05073v1.pdf | Offboard 3D Object Detection from Point Cloud Sequences | While current 3D object recognition research mostly focuses on the real-time, onboard scenario, there are many offboard use cases of perception that are largely under-explored, such as using machines to automatically generate high-quality 3D labels. Existing 3D object detectors fail to satisfy the high-quality requirement for offboard uses due to the limited input and speed constraints. In this paper, we propose a novel offboard 3D object detection pipeline using point cloud sequence data. Observing that different frames capture complementary views of objects, we design the offboard detector to make use of the temporal points through both multi-frame object detection and novel object-centric refinement models. Evaluated on the Waymo Open Dataset, our pipeline named 3D Auto Labeling shows significant gains compared to the state-of-the-art onboard detectors and our offboard baselines. Its performance is even on par with human labels verified through a human label study. Further experiments demonstrate the application of auto labels for semi-supervised learning and provide extensive analysis to validate various design choices. | ['Dragomir Anguelov', 'Boyang Deng', 'Khoa Vo', 'Pei Sun', 'Mahyar Najibi', 'Yin Zhou', 'Charles R. Qi'] | 2021-03-08 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Qi_Offboard_3D_Object_Detection_From_Point_Cloud_Sequences_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Qi_Offboard_3D_Object_Detection_From_Point_Cloud_Sequences_CVPR_2021_paper.pdf | cvpr-2021-1 | ['3d-object-recognition'] | ['computer-vision'] | [ 1.03296556e-01 5.27823716e-02 -1.12663664e-01 -5.13045490e-01
-7.75825918e-01 -8.86609733e-01 7.86204398e-01 1.49193287e-01
-1.64993748e-01 -1.70933351e-01 -2.56371766e-01 -2.21683338e-01
2.64987707e-01 -2.96186596e-01 -7.31860459e-01 -2.05119982e-01
1.52500972e-01 9.22526479e-01 9.34398711e-01 -4.49965596e-02
1.39036030e-01 8.88550937e-01 -1.98158491e+00 3.01657706e-01
5.15810288e-02 1.10694277e+00 2.37310588e-01 6.77483618e-01
-1.48057967e-01 4.38954085e-01 -4.00324374e-01 1.30068079e-01
8.65391910e-01 5.09431064e-02 -5.31326890e-01 6.03360236e-01
9.15243685e-01 -6.38334215e-01 -1.97134465e-01 9.03266549e-01
4.71873909e-01 3.10972910e-02 5.90263307e-01 -1.61729085e+00
-3.68329912e-01 -1.01440072e-01 -5.66363096e-01 2.34336942e-01
6.32227182e-01 2.43306175e-01 1.02032554e+00 -1.10953629e+00
7.90998399e-01 1.28475642e+00 7.13558853e-01 4.30467397e-01
-1.19938934e+00 -7.03102171e-01 3.83749843e-01 2.83363890e-02
-1.31856501e+00 -5.38658857e-01 5.72346091e-01 -6.34179831e-01
1.39610422e+00 5.05739562e-02 6.57028317e-01 8.58761847e-01
-2.63936460e-01 9.77025449e-01 1.19039834e+00 -5.00364542e-01
2.84656227e-01 1.25734895e-01 1.50550261e-01 7.46506810e-01
2.05578819e-01 4.37842101e-01 -8.60616922e-01 1.38662225e-02
7.93279469e-01 -8.86915773e-02 2.09659524e-02 -9.20940340e-01
-1.35858548e+00 3.41681212e-01 3.82467836e-01 -1.58415228e-01
-1.59802690e-01 1.42463120e-02 1.34600103e-01 6.85360134e-02
5.52112997e-01 2.15912521e-01 -6.32371068e-01 -4.22754250e-02
-9.34780777e-01 3.29163939e-01 6.11573935e-01 1.52482140e+00
7.56983042e-01 -2.83949226e-01 1.99467130e-02 4.02601153e-01
6.13041222e-01 6.57531083e-01 -2.40826190e-01 -1.14990389e+00
1.82704717e-01 8.46896291e-01 2.87932813e-01 -6.36098862e-01
-8.16731393e-01 -2.92282939e-01 -2.08768919e-02 9.22204435e-01
3.55602294e-01 3.96454722e-01 -1.30080426e+00 1.19640028e+00
8.00321519e-01 2.59601980e-01 -3.32397789e-01 1.35645497e+00
9.25384581e-01 4.04956281e-01 -8.20435882e-02 1.99857265e-01
1.40375423e+00 -9.76511180e-01 -2.58651048e-01 -2.86439538e-01
6.49108171e-01 -8.12616527e-01 7.97080398e-01 3.27072293e-01
-9.77702141e-01 -7.80946374e-01 -1.13956439e+00 -2.78772235e-01
-2.39460230e-01 1.14889957e-01 5.30577302e-01 6.30270422e-01
-8.98263514e-01 1.59383148e-01 -1.13696265e+00 -7.02751815e-01
5.72432458e-01 2.74240643e-01 -6.35685980e-01 -1.02773167e-01
-3.25883955e-01 1.16428614e+00 3.41382504e-01 -3.24584782e-01
-1.03537345e+00 -8.03542256e-01 -7.59640276e-01 -4.31954205e-01
5.34906328e-01 -7.13490546e-01 1.68387628e+00 -4.98378694e-01
-1.32549584e+00 1.56765771e+00 9.19513181e-02 -2.62383908e-01
6.46635532e-01 -3.90251368e-01 -8.60836357e-02 2.87896425e-01
2.51823068e-01 1.10191274e+00 6.61108434e-01 -1.49530375e+00
-1.23926067e+00 -3.75016183e-01 2.59177923e-01 3.24200034e-01
1.47071108e-01 3.83634344e-02 -7.06461668e-01 -1.88870311e-01
4.99300331e-01 -1.26948106e+00 -8.84352997e-02 5.86124539e-01
-3.11976343e-01 -4.07960325e-01 1.15731728e+00 -1.38931170e-01
2.37849653e-01 -2.16151118e+00 -2.34613121e-01 2.33862735e-02
8.91413316e-02 1.34468049e-01 -4.92346659e-02 1.69432580e-01
1.31078526e-01 -2.91268796e-01 3.03659514e-02 -6.33236468e-01
1.82181776e-01 1.71514228e-01 -2.16379598e-01 7.11062253e-01
3.95176917e-01 7.08675921e-01 -1.04612017e+00 -4.48297888e-01
7.03178704e-01 2.23641366e-01 -5.35092175e-01 2.57698864e-01
-4.97491270e-01 2.87551373e-01 -3.37495744e-01 1.07474494e+00
6.80174649e-01 -4.41700011e-01 -2.17314735e-01 -3.06729108e-01
-3.15448523e-01 3.23450387e-01 -1.34901190e+00 2.10142255e+00
-3.39810885e-02 6.71601772e-01 8.49343613e-02 -3.89617354e-01
9.72082138e-01 1.98802412e-01 5.16243339e-01 -5.19919395e-01
1.57855168e-01 2.20635787e-01 -2.57422358e-01 -3.34115207e-01
5.26436269e-01 -8.95865168e-03 1.03915900e-01 5.60611367e-01
1.31349981e-01 -5.42847276e-01 -1.02602750e-01 2.18616053e-01
1.35972297e+00 8.68594587e-01 3.04460824e-01 -4.18436639e-02
2.00131703e-02 6.06954396e-01 3.02070022e-01 8.26700926e-01
-5.60090005e-01 1.07647789e+00 -1.35040119e-01 -3.76420140e-01
-1.14331186e+00 -1.00694764e+00 -2.76244968e-01 1.00498974e+00
6.99430168e-01 -4.18118238e-01 -3.44037503e-01 -9.85956907e-01
2.92118818e-01 6.97261333e-01 -4.13475573e-01 2.21910551e-01
-2.22293764e-01 -1.97498217e-01 3.05181801e-01 6.41412914e-01
2.38716111e-01 -8.64002526e-01 -1.40020287e+00 6.39875755e-02
1.84380144e-01 -1.63175321e+00 -1.32890373e-01 3.85319829e-01
-7.54550278e-01 -1.14986265e+00 -2.75145531e-01 -6.00170255e-01
4.98714268e-01 9.88205671e-01 1.19672823e+00 -9.35897678e-02
-3.16489667e-01 8.05903375e-01 -5.91485023e-01 -7.73299158e-01
-3.30491602e-01 1.25451824e-02 3.50566149e-01 -3.05466443e-01
7.83361435e-01 -4.31413591e-01 -4.32318151e-01 6.20203614e-01
-5.25242805e-01 2.69740909e-01 5.71693838e-01 3.47794235e-01
7.87149966e-01 -3.64430875e-01 2.07723975e-01 -5.61902225e-01
-1.67157874e-01 -7.30258450e-02 -8.81529450e-01 -4.27202433e-02
-5.23901045e-01 -1.59679964e-01 -2.34944493e-01 -4.34829295e-01
-8.61293614e-01 6.51115358e-01 1.07750654e-01 -9.34654891e-01
-5.76394796e-01 -1.92125052e-01 -6.86283186e-02 -1.17154703e-01
8.72088730e-01 -2.88958192e-01 -8.76130089e-02 -6.13779485e-01
4.85739231e-01 6.79719865e-01 5.72081923e-01 -4.21228200e-01
1.05813646e+00 9.08269227e-01 2.75071338e-02 -5.84019363e-01
-1.23264921e+00 -9.67471242e-01 -9.41477001e-01 -4.43653852e-01
9.15856481e-01 -1.14966679e+00 -5.73717058e-01 1.77645072e-01
-1.37916267e+00 -3.39684635e-01 -4.38322991e-01 4.50891197e-01
-7.78243005e-01 4.75031771e-02 -2.63062686e-01 -9.90259349e-01
7.95005560e-02 -9.71454680e-01 1.76666331e+00 -3.70212048e-02
-3.43071699e-01 -3.26663673e-01 3.28646298e-03 3.85676593e-01
-2.00029135e-01 2.87931144e-01 3.99827808e-01 -6.19894981e-01
-9.67328668e-01 -3.79438609e-01 -4.68200535e-01 4.17922139e-02
-6.39438257e-02 -1.06387556e-01 -1.29578102e+00 -5.79022393e-02
-2.89363265e-01 -4.58843172e-01 5.94232500e-01 8.72495174e-02
6.72249675e-01 6.13291383e-01 -6.16340756e-01 4.36709315e-01
1.04237270e+00 -3.12403683e-02 1.09402381e-01 4.68569189e-01
6.45159364e-01 6.63836539e-01 9.09913242e-01 4.78452355e-01
4.49532628e-01 9.72404957e-01 7.65218139e-01 -8.44977796e-02
-5.06543696e-01 -3.64182234e-01 2.00203836e-01 4.87580180e-01
6.73931390e-02 -1.59102425e-01 -1.22452891e+00 4.82916087e-01
-1.88726664e+00 -6.73431695e-01 -3.76176566e-01 1.94299662e+00
3.81507277e-01 6.84545219e-01 2.16071084e-01 1.49050921e-01
6.00996196e-01 -2.08806157e-01 -6.65770173e-01 2.92473465e-01
-1.64905205e-01 -1.44082922e-02 6.81567192e-01 2.29408592e-01
-1.26339567e+00 8.52282345e-01 6.58046007e+00 2.87740290e-01
-7.33101249e-01 3.20363402e-01 1.03016473e-01 -2.97855407e-01
1.47567123e-01 2.27431431e-01 -1.21429014e+00 -6.40181899e-02
4.05116558e-01 4.00437683e-01 -1.33278277e-02 1.04690206e+00
1.92976803e-01 -5.57075515e-02 -1.61657321e+00 1.35962701e+00
1.58473328e-01 -1.35287321e+00 -3.50415170e-01 9.74330157e-02
8.37793648e-01 4.75496590e-01 -4.18613136e-01 1.65514454e-01
4.09862459e-01 -5.92893958e-01 1.45058548e+00 2.72971362e-01
6.16042018e-01 -2.61817247e-01 4.54195112e-01 6.43153310e-01
-1.08668828e+00 1.45727456e-01 -1.87407508e-01 -3.84949833e-01
4.34296489e-01 5.55100918e-01 -1.21095574e+00 3.57016623e-01
9.95507479e-01 8.52848589e-01 -7.17655718e-01 1.15613341e+00
-2.98358023e-01 3.27095687e-01 -7.40307987e-01 4.58764238e-03
1.61896229e-01 2.82811135e-01 7.40530849e-01 1.13468981e+00
2.15850547e-01 3.75478804e-01 6.07133031e-01 7.19493747e-01
8.58000442e-02 -3.13493818e-01 -6.89220488e-01 2.60187238e-01
5.34828067e-01 1.31824672e+00 -1.01978576e+00 -2.19711870e-01
-5.62021971e-01 8.40604961e-01 2.56503582e-01 6.48484519e-03
-6.46941423e-01 5.12280054e-02 6.95360541e-01 4.14780796e-01
4.78340209e-01 -7.26291955e-01 -4.47871178e-01 -9.65944946e-01
-3.57457693e-03 -5.33932209e-01 2.36512214e-01 -1.21032798e+00
-1.31401873e+00 3.72187406e-01 2.52723873e-01 -1.62632418e+00
6.05068691e-02 -8.78889024e-01 -2.94505004e-02 3.57217401e-01
-1.43938673e+00 -1.45083082e+00 -6.58326626e-01 2.90553987e-01
7.50476301e-01 3.30493436e-03 6.52298987e-01 2.27923781e-01
-7.35427663e-02 1.26377538e-01 -5.14682233e-01 -2.55834639e-01
6.50647759e-01 -1.02996850e+00 6.08068824e-01 8.33136857e-01
6.28166199e-01 5.98593168e-02 6.62625194e-01 -6.31041884e-01
-1.42275620e+00 -8.88546944e-01 5.64163327e-01 -1.24473953e+00
3.08634609e-01 -7.34193623e-01 -6.62589669e-01 7.27795899e-01
-1.89387470e-01 4.06979471e-01 4.05770034e-01 7.91332573e-02
-5.77696204e-01 2.56880194e-01 -1.00054288e+00 3.92590106e-01
1.78164673e+00 -4.83429313e-01 -6.86207652e-01 5.61811805e-01
8.55583131e-01 -8.07153165e-01 -5.57063103e-01 6.12389743e-01
5.95223069e-01 -9.71938670e-01 1.04842782e+00 -5.35795152e-01
1.33335784e-01 -9.92103338e-01 -5.28938651e-01 -7.58234918e-01
-3.97032708e-01 -4.24238950e-01 -3.60672235e-01 1.12029958e+00
1.15943424e-01 -3.06136087e-02 1.06049240e+00 5.41186929e-01
-5.33700466e-01 -2.30023801e-01 -9.36130583e-01 -8.86442661e-01
-7.78912842e-01 -9.77315903e-01 4.43361580e-01 8.36646080e-01
-3.59603524e-01 5.11106670e-01 -1.05866408e-02 5.55324316e-01
7.40579486e-01 3.20785165e-01 1.35087919e+00 -1.40066075e+00
-2.53593475e-01 -2.70840466e-01 -9.36515868e-01 -1.45060170e+00
6.06186464e-02 -1.02079499e+00 4.36243564e-01 -1.61609805e+00
1.02966046e-02 -7.82252073e-01 -1.46723852e-01 6.49341166e-01
2.42936000e-01 6.74039066e-01 4.58918482e-01 4.12889987e-01
-9.27366197e-01 4.18198317e-01 9.15132046e-01 -3.61270234e-02
-1.61608309e-01 -2.44296670e-01 -2.24847421e-01 8.92273545e-01
3.39809895e-01 -5.71407080e-01 -1.90387443e-01 -5.77999294e-01
-2.55726883e-03 -3.13363612e-01 6.27090752e-01 -1.24687362e+00
2.64881819e-01 -1.56436816e-01 3.74559909e-01 -1.22486496e+00
6.89622402e-01 -1.13591266e+00 9.96954218e-02 1.66829169e-01
-6.56195581e-02 -3.86587568e-02 2.86728114e-01 6.70563400e-01
1.91685781e-01 -4.56496626e-02 6.61328256e-01 -1.16185322e-01
-1.11942768e+00 3.41704547e-01 -9.45651457e-02 -1.52141750e-01
1.30334496e+00 -6.28161073e-01 -2.10800439e-01 -7.12314770e-02
-4.25396740e-01 1.82476476e-01 7.92255878e-01 7.64588058e-01
5.82985282e-01 -1.34406042e+00 -5.79744995e-01 2.82114953e-01
6.65250421e-01 3.83816808e-01 -1.13918059e-01 6.10405803e-01
-5.09507179e-01 2.89254338e-01 -1.70277894e-01 -1.40598667e+00
-1.13626516e+00 4.81174707e-01 2.53946275e-01 3.46679837e-01
-1.07353878e+00 8.44495058e-01 2.49569729e-01 -4.86712664e-01
2.84878194e-01 -4.46556270e-01 3.38693053e-01 -9.43586379e-02
3.97171527e-01 3.16324204e-01 4.26921010e-01 -5.93190074e-01
-6.93394184e-01 6.17066622e-01 1.50267750e-01 -2.57062227e-01
1.19816053e+00 -7.63772056e-02 4.68022674e-01 6.90980852e-01
8.11755717e-01 -2.54265726e-01 -1.84415555e+00 -3.17350328e-01
4.09123860e-02 -7.46770322e-01 1.94917470e-01 -6.79971755e-01
-5.22263885e-01 6.49484277e-01 9.14212644e-01 1.18955538e-01
7.43142366e-01 5.07543623e-01 4.86254245e-01 4.46754873e-01
7.86642849e-01 -9.89111722e-01 2.14925617e-01 5.11498630e-01
6.17052972e-01 -1.54129422e+00 2.56421983e-01 -6.76026046e-01
-2.82003373e-01 9.37956929e-01 7.51472831e-01 -1.79774657e-01
6.29260600e-01 3.82938504e-01 3.70632768e-01 -4.00465041e-01
-8.00673008e-01 -4.73428994e-01 3.14903796e-01 8.85714650e-01
-1.07094906e-02 -6.55765608e-02 4.47607338e-01 9.05914754e-02
-1.28379762e-01 -1.40846401e-01 3.81328404e-01 1.17236459e+00
-6.40891671e-01 -8.99049401e-01 -5.28101027e-01 3.32117528e-01
1.80899397e-01 5.03320515e-01 -4.26886588e-01 9.20660913e-01
3.39001000e-01 1.00491834e+00 2.55163431e-01 -3.98079455e-01
7.31058180e-01 8.64723697e-02 8.75956118e-01 -1.08353555e+00
-3.95976335e-01 3.51272337e-02 1.92453131e-01 -8.10076773e-01
-9.14460659e-01 -1.11832726e+00 -1.37713552e+00 1.80961430e-01
-7.13018000e-01 -5.94802141e-01 1.00204003e+00 1.03225589e+00
5.15561163e-01 2.86730111e-01 2.19086543e-01 -1.54927957e+00
-3.93988639e-01 -7.23570049e-01 -4.27127719e-01 6.94249153e-01
2.92834789e-01 -1.24322438e+00 -1.47808447e-01 3.06648076e-01] | [7.605356216430664, -2.62255597114563] |
e1b27e96-aafd-4717-899e-288ecf5493f0 | unsupervised-detection-of-lung-nodules-in | 2108.02233 | null | https://arxiv.org/abs/2108.02233v1 | https://arxiv.org/pdf/2108.02233v1.pdf | Unsupervised Detection of Lung Nodules in Chest Radiography Using Generative Adversarial Networks | Lung nodules are commonly missed in chest radiographs. We propose and evaluate P-AnoGAN, an unsupervised anomaly detection approach for lung nodules in radiographs. P-AnoGAN modifies the fast anomaly detection generative adversarial network (f-AnoGAN) by utilizing a progressive GAN and a convolutional encoder-decoder-encoder pipeline. Model training uses only unlabelled healthy lung patches extracted from the Indiana University Chest X-Ray Collection. External validation and testing are performed using healthy and unhealthy patches extracted from the ChestX-ray14 and Japanese Society for Radiological Technology datasets, respectively. Our model robustly identifies patches containing lung nodules in external validation and test data with ROC-AUC of 91.17% and 87.89%, respectively. These results show unsupervised methods may be useful in challenging tasks such as lung nodule detection in radiographs. | ['H. R. Tizhoosh', 'Christos Karanassios', 'Nedim Hodzic', 'David Ramon Prados', 'Nitish Bhatt'] | 2021-08-04 | null | null | null | null | ['lung-nodule-detection'] | ['medical'] | [ 5.26121259e-01 5.44032454e-01 1.64419860e-01 -1.35849100e-02
-1.23369932e+00 -3.13318551e-01 2.81629980e-01 -9.91844758e-02
-2.48487175e-01 5.50034761e-01 -1.27914799e-02 -4.34537888e-01
1.81030989e-01 -6.33516848e-01 -6.60370588e-01 -8.49006176e-01
4.75397967e-02 7.44439542e-01 5.73586345e-01 4.31487441e-01
-3.76558959e-01 5.04260182e-01 -8.21227491e-01 4.85656410e-01
5.12130678e-01 9.84124005e-01 -2.15751827e-01 1.47927237e+00
3.36295098e-01 1.10939026e+00 -4.35773045e-01 -4.04274046e-01
6.10478580e-01 -1.02088606e+00 -7.13229418e-01 2.00584292e-01
4.47026849e-01 -5.44984639e-01 -7.00001538e-01 8.95695150e-01
5.55141509e-01 -3.22220206e-01 9.94776428e-01 -9.35364485e-01
-3.61526221e-01 3.57543617e-01 -4.92253870e-01 8.16587389e-01
-1.68906868e-01 5.77864528e-01 6.95323706e-01 -8.88060510e-01
3.07967335e-01 3.88018906e-01 9.94706571e-01 1.00387251e+00
-9.99353886e-01 -4.23930615e-01 -7.81084239e-01 -2.39817008e-01
-1.07755578e+00 1.55125633e-01 3.20511460e-01 -5.33384085e-01
6.64299488e-01 6.65335655e-01 8.28407526e-01 1.23449743e+00
6.75929904e-01 6.65196002e-01 8.35323691e-01 -3.54844481e-01
5.35015352e-02 -2.61703849e-01 -4.60698336e-01 9.87866580e-01
5.47777772e-01 2.00346336e-01 2.57847104e-02 -5.70699394e-01
1.03083479e+00 2.79138327e-01 -1.42734870e-01 -2.34050214e-01
-1.50028837e+00 8.67851019e-01 6.18410408e-01 2.96396285e-01
-7.42171168e-01 3.81090999e-01 4.49253768e-01 2.35616341e-01
9.94733870e-02 3.45354468e-01 1.35020107e-01 1.15977414e-01
-8.65560234e-01 -1.48641869e-01 6.13046527e-01 7.26063311e-01
-1.64081365e-01 3.53906065e-01 -4.69228894e-01 6.93336785e-01
4.34273094e-01 6.00690663e-01 1.08056271e+00 -6.70837581e-01
1.53056206e-02 4.43926841e-01 -5.67678213e-01 -3.68042886e-01
-1.45017475e-01 -3.66137922e-01 -1.24326146e+00 6.59434646e-02
2.82922000e-01 -1.84028000e-02 -1.63734388e+00 1.07276762e+00
2.10972175e-01 2.84759372e-01 1.16894431e-02 6.80529177e-01
6.59749687e-01 -4.54982277e-03 2.87151068e-01 9.97372717e-02
1.43027580e+00 -9.57973838e-01 -3.41882288e-01 -1.07532464e-01
7.61398196e-01 -5.90375483e-01 1.14387155e+00 9.91089270e-02
-1.13248301e+00 -2.58113354e-01 -8.50092888e-01 2.68405199e-01
1.64329261e-01 3.11093360e-01 7.49888867e-02 6.87085867e-01
-1.03139675e+00 1.96034610e-01 -1.41286314e+00 -4.15370524e-01
1.02409792e+00 4.35553551e-01 -3.59589368e-01 -1.25548402e-02
-6.85530007e-01 5.53973019e-01 2.12919742e-01 -1.30318478e-01
-1.35304761e+00 -6.86715662e-01 -5.99958420e-01 -1.98545516e-01
4.19038951e-01 -9.69763756e-01 1.66935611e+00 -9.45433199e-01
-1.11840916e+00 1.17763972e+00 3.30582798e-01 -1.04168236e+00
7.33035862e-01 -4.78755906e-02 -5.86050630e-01 6.39337838e-01
2.68104702e-01 4.78054345e-01 1.04371178e+00 -6.49958611e-01
-4.06490922e-01 -5.99718168e-02 -8.03955793e-01 -9.00104716e-02
1.05713286e-01 -1.58176556e-01 -1.14279300e-01 -1.28056717e+00
2.28630200e-01 -1.14108884e+00 -5.78348100e-01 2.30355278e-01
-6.36057556e-01 1.87744781e-01 9.71240163e-01 -9.93345797e-01
1.02082741e+00 -1.95555055e+00 -6.07405901e-01 7.36718416e-01
6.21750593e-01 3.76146674e-01 7.48198330e-02 -2.90371805e-01
-3.55560720e-01 3.87581468e-01 -8.05331647e-01 3.58161241e-01
-3.76094222e-01 4.46098000e-01 1.82298094e-01 4.15675789e-01
6.24243915e-01 1.43281782e+00 -7.55554080e-01 -7.48319745e-01
8.27006623e-02 1.54464886e-01 -5.99174321e-01 4.92427796e-01
-1.52181253e-01 6.95785403e-01 -4.21029717e-01 1.10086977e+00
1.14014477e-01 -5.07532179e-01 -1.81514949e-01 2.57496119e-01
7.24365413e-01 6.13830723e-02 -4.32640433e-01 1.48612571e+00
-3.49719189e-02 4.17034090e-01 -1.75992712e-01 -4.88832474e-01
4.71803397e-01 6.86772168e-01 6.04100943e-01 -4.75121170e-01
-3.55550833e-02 5.93880475e-01 6.74692035e-01 -6.87560678e-01
-2.01061249e-01 -4.08080369e-01 1.64649740e-01 6.51058257e-01
1.54768765e-01 -3.17733169e-01 2.70884344e-03 3.44550103e-01
1.99983156e+00 -4.02088016e-01 5.08645415e-01 -6.28563091e-02
6.60118937e-01 1.10281996e-01 1.57284483e-01 9.86282885e-01
-5.39030790e-01 1.22649324e+00 6.18414521e-01 -3.35279018e-01
-1.38747501e+00 -1.45164227e+00 -3.44299257e-01 4.70270008e-01
-7.79150248e-01 -2.27716826e-02 -6.21501565e-01 -1.65452600e+00
-1.75370965e-02 3.86354834e-01 -9.01748538e-01 -2.38433346e-01
-7.79062152e-01 -6.85224116e-01 9.24873590e-01 9.45214093e-01
3.37362349e-01 -1.31415963e+00 -6.57509983e-01 2.13206410e-01
5.74484169e-02 -6.62429273e-01 -5.34184039e-01 2.70606041e-01
-9.68467057e-01 -1.58571231e+00 -1.03141713e+00 -6.87554777e-01
1.04705727e+00 -4.02031273e-01 1.52300298e+00 2.79079229e-01
-1.08109474e+00 8.45255733e-01 -2.78400276e-02 -5.18101692e-01
-1.14220011e+00 5.41049689e-02 -2.40501449e-01 -2.11860403e-01
3.88642430e-01 -4.71801847e-01 -8.79399955e-01 4.33799416e-01
-1.11007237e+00 -3.99469048e-01 1.28484762e+00 1.14737344e+00
1.27440774e+00 -3.31277370e-01 5.16882181e-01 -1.43032253e+00
2.80974090e-01 -8.32779109e-01 -1.62234142e-01 -2.32648738e-02
-5.18912137e-01 -1.93444163e-01 4.68814701e-01 -2.13744715e-01
-6.94573820e-01 1.45687968e-01 -4.69811082e-01 -7.17387974e-01
-4.57481563e-01 2.20757332e-02 2.93922365e-01 -1.74583927e-01
1.03492677e+00 2.42359951e-01 8.00338462e-02 -5.61410449e-02
-2.55412012e-01 3.58013451e-01 9.40814257e-01 9.97636020e-02
1.13605618e+00 4.22944695e-01 3.01649094e-01 -6.89627767e-01
-6.66148067e-01 -7.76597261e-01 -6.11230612e-01 -7.41907060e-02
1.12342966e+00 -7.56592453e-01 2.35873461e-01 2.11497039e-01
-3.13319832e-01 -2.53483742e-01 -1.20381141e+00 7.21788049e-01
-6.16869867e-01 3.73870343e-01 -6.93274498e-01 -3.91713619e-01
-7.18057930e-01 -7.18604565e-01 1.01299167e+00 -1.32382251e-02
-4.04412538e-01 -8.60973835e-01 4.25816864e-01 8.83048624e-02
6.82111561e-01 5.73725462e-01 8.56878817e-01 -1.56256485e+00
-4.55402434e-01 -7.71580100e-01 1.41979098e-01 7.98953831e-01
1.89950004e-01 -7.51838041e-03 -8.20897400e-01 -2.26074710e-01
3.20404232e-01 -4.37548488e-01 8.36594641e-01 4.22283053e-01
1.72338796e+00 -4.10600752e-01 -2.38874555e-01 7.23503888e-01
1.21223068e+00 2.18451098e-01 8.52613568e-01 1.25928506e-01
6.95484698e-01 -2.51593441e-01 2.34062403e-01 7.85794482e-02
-4.59762722e-01 -2.13068068e-01 6.38057292e-01 -2.64270842e-01
-4.73414570e-01 -2.82232702e-01 -2.53788475e-02 7.30797827e-01
-6.59461394e-02 -3.28934431e-01 -1.41956162e+00 6.39207184e-01
-1.20058513e+00 -7.56579936e-01 -2.81879097e-01 1.92084861e+00
5.90541601e-01 2.21385613e-01 1.28099382e-01 1.88223302e-01
5.92314243e-01 -1.01198070e-01 -5.94967067e-01 -5.49392365e-02
5.20732924e-02 9.41631794e-01 5.02844155e-01 -1.48014650e-01
-1.32555306e+00 5.16085774e-02 6.64406443e+00 5.07519066e-01
-8.85772705e-01 2.36614928e-01 9.46472168e-01 -8.73872116e-02
-9.69364122e-02 -3.76141518e-01 3.72871347e-02 4.88905698e-01
1.06242597e+00 -4.92186695e-02 -2.80165762e-01 9.95019376e-01
-4.58704114e-01 -8.90667140e-02 -1.05611813e+00 5.97595334e-01
1.61240339e-01 -9.18094635e-01 3.70683372e-02 3.14109698e-02
8.01672280e-01 3.96379113e-01 1.98438391e-01 2.11531371e-01
1.48223296e-01 -1.37524915e+00 -2.14432985e-01 6.25942886e-01
1.07849693e+00 -5.06018460e-01 1.21074462e+00 2.88773179e-02
-8.79656017e-01 3.53086710e-01 -7.70498663e-02 9.88338113e-01
-3.62065494e-01 3.95996094e-01 -1.79537761e+00 3.48651290e-01
8.15804124e-01 4.08192784e-01 -1.07201827e+00 1.19557881e+00
-3.11485708e-01 1.30729413e+00 -4.60726589e-01 4.32793260e-01
1.50682807e-01 2.71770149e-01 8.55691195e-01 1.17467010e+00
4.54748929e-01 3.98913771e-02 -1.27228603e-01 7.34761655e-01
-3.30190688e-01 1.16083957e-02 -7.90405810e-01 -1.34541109e-01
-7.41481930e-02 1.25628638e+00 -8.59020531e-01 -4.13969278e-01
-4.02794302e-01 1.03191745e+00 -4.44660872e-01 5.39580137e-02
-1.09563088e+00 -4.60865274e-02 -3.43430787e-01 6.00052774e-01
2.99303591e-01 5.10253429e-01 -3.96386124e-02 -9.75127518e-01
-3.44756283e-02 -1.05128562e+00 7.35546112e-01 -6.51915371e-01
-1.53274512e+00 7.89739311e-01 -3.84318054e-01 -1.54319894e+00
-6.09497309e-01 -5.51512241e-01 -1.09220028e+00 6.14720702e-01
-9.74135995e-01 -1.17800963e+00 -5.96852362e-01 6.35434687e-01
4.66047019e-01 -6.60236120e-01 1.05791509e+00 -1.94153443e-01
-3.80480975e-01 7.37191975e-01 3.35553437e-02 5.71664095e-01
6.24499440e-01 -1.77531970e+00 5.35323501e-01 1.07325733e+00
5.89885488e-02 -2.42537782e-01 3.15569267e-02 -7.81888723e-01
-7.39172101e-01 -1.69200504e+00 3.00790995e-01 -7.18078315e-01
5.20210028e-01 1.79878891e-01 -1.27473795e+00 1.00948894e+00
1.18475229e-01 7.45668471e-01 1.07272446e+00 -1.01113296e+00
-1.54683460e-02 4.82442558e-01 -1.43533051e+00 3.69158894e-01
6.49987280e-01 -4.61784214e-01 -6.30097091e-01 5.70957899e-01
1.44985199e-01 -6.14219069e-01 -1.11561787e+00 7.28922665e-01
2.24362120e-01 -7.99432456e-01 9.21201110e-01 -7.02544451e-01
5.71844220e-01 -7.55594894e-02 1.50131270e-01 -8.81916463e-01
-2.49631256e-01 -4.64211643e-01 -4.28797781e-01 4.68089849e-01
4.44938600e-01 -7.08092511e-01 1.19301748e+00 2.90567786e-01
-2.32538715e-01 -9.93166864e-01 -9.93659139e-01 -3.68471801e-01
-1.54687446e-02 -2.61742681e-01 4.05011550e-02 6.28337860e-01
-8.61955881e-01 -2.55850196e-01 3.29091728e-01 8.16904008e-02
6.11216247e-01 -5.71582675e-01 6.19707584e-01 -9.76901472e-01
-4.62894022e-01 -2.68686026e-01 -7.67159700e-01 -1.69824764e-01
-3.29210907e-01 -1.14524865e+00 -8.42065960e-02 -1.28832889e+00
2.79794991e-01 -2.34204665e-01 -6.59079313e-01 5.70248365e-01
-3.07186931e-01 8.11373353e-01 -4.11429375e-01 6.13660514e-01
-4.02262300e-01 2.77213007e-01 1.30976593e+00 -1.60649016e-01
2.22945169e-01 6.12489283e-01 -2.72603810e-01 1.00976694e+00
1.02670658e+00 -9.84350145e-01 -1.37483791e-01 1.50281176e-01
-2.68942356e-01 -7.98095539e-02 8.48058581e-01 -1.42095888e+00
-3.95318605e-02 2.61609226e-01 9.60301518e-01 -7.48355150e-01
-2.85744965e-01 -1.03611946e+00 1.68149412e-01 1.18578565e+00
-3.18671018e-01 3.47926170e-01 1.36976570e-01 1.09650958e+00
-3.50564867e-01 -1.83843359e-01 1.02499449e+00 -4.77580428e-01
-2.27935880e-01 5.35132408e-01 -6.49281919e-01 3.47728848e-01
1.46298981e+00 -3.06392074e-01 -1.01089746e-01 -1.68675467e-01
-1.07820702e+00 -2.75730658e-02 2.85953045e-01 -1.34914830e-01
8.42040896e-01 -1.51681733e+00 -1.02145243e+00 6.89697385e-01
1.75999895e-01 5.59057236e-01 5.68652861e-02 1.36474383e+00
-1.19614303e+00 1.92108676e-01 -1.54545441e-01 -1.13043940e+00
-1.17357564e+00 2.25068018e-01 8.56926560e-01 -7.79013932e-01
-8.39303553e-01 9.63124096e-01 3.20105463e-01 -7.09010437e-02
-1.10952863e-02 -4.05438304e-01 3.48672181e-01 -7.30158627e-01
6.87126368e-02 1.92225084e-01 3.70349318e-01 -3.18616688e-01
-1.75126046e-01 -1.94067601e-02 -3.59693706e-01 1.41335100e-01
8.54690850e-01 4.31504011e-01 -4.13125455e-02 2.71621346e-01
9.40029502e-01 1.27688751e-01 -7.55977929e-01 -1.68560386e-01
8.19220021e-02 -1.98348507e-01 -2.64260322e-01 -9.63019013e-01
-1.19847214e+00 6.52573824e-01 1.07202387e+00 3.57821405e-01
1.21649277e+00 1.73905462e-01 8.13475490e-01 4.76938963e-01
-5.06489038e-01 -6.97954774e-01 6.35156751e-01 1.45952478e-01
8.85549486e-01 -1.29599392e+00 -9.79011208e-02 -2.05271751e-01
-7.17917740e-01 1.11096537e+00 8.65827739e-01 -4.54005808e-01
7.03373313e-01 3.68940592e-01 3.71183127e-01 -3.68054569e-01
-7.06128001e-01 1.11755179e-02 5.25833905e-01 6.52420342e-01
5.40208697e-01 7.12964088e-02 8.57382044e-02 6.29281402e-01
-4.92970794e-02 -2.25813404e-01 3.29691887e-01 1.19694877e+00
-2.35997483e-01 -8.88393939e-01 -3.96298647e-01 1.43362451e+00
-1.07896161e+00 -1.24581873e-01 -6.32309139e-01 1.27787030e+00
-7.90383816e-02 1.71572976e-02 1.79986894e-01 -5.49056269e-02
2.35517517e-01 4.95113134e-01 2.52073675e-01 -8.24636579e-01
-9.86748219e-01 2.06046134e-01 -2.56661624e-01 -4.38977540e-01
-1.73930749e-01 -4.58180934e-01 -1.18222940e+00 3.04092079e-01
-3.51704955e-01 9.96294841e-02 1.70709237e-01 4.55595016e-01
1.54435664e-01 1.19548130e+00 5.65504611e-01 -1.81562006e-02
-6.99533165e-01 -1.09436870e+00 -5.52581429e-01 5.26530087e-01
3.91332328e-01 -1.14550464e-01 -7.13067770e-01 1.30682960e-01] | [15.061203956604004, -2.041186571121216] |
81c70b04-6efa-423d-90dd-cac9948e317e | revisiting-inferential-benchmarks-for | 2306.04814 | null | https://arxiv.org/abs/2306.04814v1 | https://arxiv.org/pdf/2306.04814v1.pdf | Revisiting Inferential Benchmarks for Knowledge Graph Completion | Knowledge Graph (KG) completion is the problem of extending an incomplete KG with missing facts. A key feature of Machine Learning approaches for KG completion is their ability to learn inference patterns, so that the predicted facts are the results of applying these patterns to the KG. Standard completion benchmarks, however, are not well-suited for evaluating models' abilities to learn patterns, because the training and test sets of these benchmarks are a random split of a given KG and hence do not capture the causality of inference patterns. We propose a novel approach for designing KG completion benchmarks based on the following principles: there is a set of logical rules so that the missing facts are the results of the rules' application; the training set includes both premises matching rule antecedents and the corresponding conclusions; the test set consists of the results of applying the rules to the training set; the negative examples are designed to discourage the models from learning rules not entailed by the rule set. We use our methodology to generate several benchmarks and evaluate a wide range of existing KG completion systems. Our results provide novel insights on the ability of existing models to induce inference patterns from incomplete KGs. | ['Egor V. Kostylev', 'Ian Horrocks', 'Bernardo Cuenca Grau', 'Shuwen Liu'] | 2023-06-07 | null | null | null | null | ['knowledge-graph-completion'] | ['knowledge-base'] | [ 3.62959951e-01 8.31600428e-01 -4.82786417e-01 -4.46077079e-01
-5.87942488e-02 -4.09549415e-01 6.83297932e-01 1.60580680e-01
1.51635513e-01 1.04841077e+00 1.33292764e-01 -4.78098929e-01
-4.56035048e-01 -1.38209319e+00 -1.25060022e+00 -2.75537759e-01
-2.81614542e-01 8.09633911e-01 4.80684608e-01 -1.66216016e-01
-7.32460022e-02 2.52916783e-01 -1.83901477e+00 6.19252324e-01
9.13000286e-01 8.19754004e-01 -2.36732468e-01 3.54593813e-01
2.33501475e-02 1.28593957e+00 -5.85250616e-01 -4.50122476e-01
1.76464066e-01 -4.99425083e-01 -1.03771973e+00 -3.65526788e-02
3.74391586e-01 -1.86739549e-01 -2.60082960e-01 1.01589704e+00
-2.09911242e-01 1.66822746e-02 6.15691423e-01 -1.73619759e+00
-4.08690214e-01 1.03319490e+00 1.85090587e-01 -2.10268497e-01
7.85198271e-01 -4.87265550e-02 1.17975545e+00 -5.58391690e-01
9.54469740e-01 1.08373916e+00 5.59995711e-01 6.44326568e-01
-1.21633482e+00 -3.47211272e-01 1.91615477e-01 6.35879517e-01
-1.21447682e+00 -5.16168296e-01 6.27392530e-01 -1.81324974e-01
1.11710846e+00 5.32391489e-01 7.05109894e-01 8.83121014e-01
2.96356063e-02 5.48723638e-01 1.04813588e+00 -7.76580930e-01
4.77595925e-01 5.28847158e-01 3.30263048e-01 8.08484077e-01
6.84995592e-01 3.30895662e-01 -6.48399591e-01 -2.02969626e-01
3.13300014e-01 -3.19896013e-01 -2.65165269e-01 -5.01180351e-01
-8.46198976e-01 4.74884957e-01 3.91975278e-03 1.31706744e-01
-2.90876329e-01 4.08856831e-02 1.91876292e-01 3.46195519e-01
-1.58892795e-01 6.12962961e-01 -7.39181638e-01 3.40583146e-01
-5.91395020e-01 7.40254223e-01 1.30305064e+00 1.05112851e+00
1.11672044e+00 -2.41371647e-01 -1.08597822e-01 2.87018865e-01
2.94131845e-01 4.35589671e-01 2.64340132e-01 -1.07017958e+00
3.51624250e-01 1.14570010e+00 2.32205555e-01 -1.00709891e+00
-9.05602202e-02 -3.61606598e-01 -3.57549429e-01 1.51125133e-01
6.17693901e-01 -2.70492192e-02 -7.53734291e-01 2.02279830e+00
3.26197147e-01 2.14339063e-01 5.03446043e-01 2.56942272e-01
8.12806606e-01 3.79364520e-01 -8.22985396e-02 -4.45482284e-01
8.10108066e-01 -4.33463067e-01 -4.57422405e-01 -2.60246933e-01
1.11849105e+00 -1.95715353e-01 9.09796536e-01 4.51314121e-01
-1.09522080e+00 -4.53948349e-01 -1.29084623e+00 3.55459362e-01
-4.85200733e-01 -2.82810718e-01 7.06638396e-01 2.49752313e-01
-7.43962407e-01 7.52041280e-01 -3.45209658e-01 -1.81346610e-02
2.33783424e-01 4.52158511e-01 -3.02786380e-01 -6.13389254e-01
-1.50443244e+00 1.03684974e+00 9.43984807e-01 3.16952099e-03
-8.86153221e-01 -8.15605640e-01 -1.00480390e+00 1.88198209e-01
8.43259335e-01 -6.52060628e-01 1.12903047e+00 -8.26576948e-01
-6.13268912e-01 8.54978681e-01 -2.45491758e-01 -4.44688976e-01
4.53205645e-01 6.99164420e-02 -7.72405565e-01 -1.94314361e-01
2.63234466e-01 4.03069437e-01 3.26672286e-01 -1.47711897e+00
-8.76994371e-01 -2.56050140e-01 3.71290922e-01 -1.07828699e-01
5.90396710e-02 -4.61489469e-01 -3.58487725e-01 -9.71415788e-02
-1.15732141e-01 -7.20219672e-01 -1.13985822e-01 -5.94723225e-01
-8.87943685e-01 -3.03102702e-01 7.89057314e-01 -3.14414382e-01
1.38021457e+00 -1.93851542e+00 -2.11841881e-01 6.99223340e-01
2.13035978e-02 1.25836447e-01 1.85779437e-01 4.44826692e-01
-2.84160286e-01 3.18400234e-01 -4.88774367e-02 2.93295234e-01
2.55384594e-01 7.35738456e-01 -5.62700868e-01 -2.21649691e-01
8.57997760e-02 7.55621791e-01 -9.25612748e-01 -4.88472521e-01
3.92990783e-02 -4.05637950e-01 -5.99673390e-01 1.90770984e-01
-1.03230095e+00 -1.99534878e-01 -4.67058986e-01 2.26254269e-01
3.80402714e-01 -2.35039979e-01 5.00795007e-01 -2.37486437e-01
4.75141317e-01 6.80059433e-01 -1.46458530e+00 1.06024849e+00
4.58553806e-02 -5.40753417e-02 -6.78841114e-01 -9.52564538e-01
8.08387578e-01 2.46613219e-01 1.17793955e-01 -4.24745172e-01
-3.32740158e-01 1.30355060e-01 1.68727815e-01 -8.01771104e-01
4.95602116e-02 -3.89325947e-01 -2.01398022e-02 4.88764286e-01
-4.49108938e-03 -1.59857109e-01 8.34820628e-01 4.82997715e-01
1.27490604e+00 1.79671049e-01 3.29348177e-01 -6.29912615e-02
5.54024279e-01 5.45946777e-01 1.12034380e+00 1.07736480e+00
4.53833073e-01 -6.75301254e-02 9.81305540e-01 -6.43482625e-01
-7.97287941e-01 -1.00099993e+00 1.20471887e-01 4.21801269e-01
-1.63077906e-01 -8.40609491e-01 -3.20698023e-01 -1.30751741e+00
1.93755656e-01 1.22069383e+00 -7.18470454e-01 -2.98263222e-01
-3.09974790e-01 -3.85593206e-01 6.47153795e-01 5.83223939e-01
3.40432286e-01 -1.28762197e+00 -3.68550837e-01 1.46718234e-01
-2.50352770e-01 -1.23770928e+00 1.84786186e-01 1.03641197e-01
-8.57223272e-01 -1.96295404e+00 5.31934381e-01 -7.21966743e-01
1.03078270e+00 -4.05924261e-01 1.33051813e+00 5.42782485e-01
4.95928265e-02 3.20688665e-01 -1.98376760e-01 -5.06998062e-01
-8.45265508e-01 -4.27024633e-01 1.23311944e-01 -1.68839261e-01
6.19287550e-01 -6.89281762e-01 5.37799448e-02 2.56974548e-01
-9.48033988e-01 2.98737943e-01 5.03460050e-01 5.44443786e-01
7.06124067e-01 6.87456787e-01 6.71910822e-01 -1.50996363e+00
4.32219267e-01 -4.54234838e-01 -4.04389560e-01 8.00301850e-01
-9.50927138e-01 4.01442558e-01 8.97709072e-01 -1.82714343e-01
-1.08667791e+00 -1.54715344e-01 2.56228119e-01 -1.02563359e-01
-3.32709730e-01 9.73839343e-01 -6.28219187e-01 5.19645512e-01
8.89189422e-01 2.18095332e-01 -2.91155249e-01 -1.82309031e-01
3.27574521e-01 -2.17780471e-02 7.05114305e-01 -1.16206801e+00
8.34271908e-01 1.44033745e-01 3.84414345e-01 -2.09850118e-01
-9.77791429e-01 3.91060114e-02 -4.14953709e-01 -9.76052806e-02
5.52999191e-02 -4.35798675e-01 -4.57187742e-01 -3.79935130e-02
-7.10125208e-01 -5.42124569e-01 -8.08750987e-01 3.59705329e-01
-5.95395088e-01 2.98681736e-01 -2.73875982e-01 -5.54463983e-01
-9.37894434e-02 -7.64698684e-01 3.70245904e-01 -5.23106121e-02
-6.19844437e-01 -1.02247179e+00 -2.01224703e-02 2.40532123e-02
-3.03931266e-01 4.57942694e-01 1.78500080e+00 -9.48870003e-01
-4.50285137e-01 -2.85156548e-01 3.11427504e-01 5.79004169e-01
3.20077479e-01 2.04258129e-01 -6.15318775e-01 4.21780385e-02
-2.18672052e-01 -4.66688037e-01 2.85258651e-01 2.61337329e-02
1.02597499e+00 -7.70842075e-01 -5.93258798e-01 1.43891005e-02
1.46068251e+00 7.15796649e-02 8.79420280e-01 1.79282516e-01
3.45637172e-01 5.49420178e-01 5.93749881e-01 5.16809663e-03
3.10094684e-01 5.87909579e-01 2.30361760e-01 2.94960082e-01
-1.16189336e-02 -7.28435159e-01 1.89569652e-01 1.97177798e-01
-8.68610889e-02 1.21747829e-01 -1.05988395e+00 6.77514136e-01
-1.96649897e+00 -1.13911057e+00 -3.45064342e-01 2.34195185e+00
1.17453444e+00 6.71232045e-01 -1.18425012e-01 7.14139879e-01
3.77167076e-01 -4.02790427e-01 -3.76057863e-01 -3.13499540e-01
-4.08335216e-02 3.52110267e-01 -2.15401486e-01 7.72869170e-01
-6.20780468e-01 6.83395326e-01 6.46926355e+00 5.53014219e-01
-5.38790882e-01 -5.94142377e-01 2.19012752e-01 -1.77022237e-02
-6.74203277e-01 4.48926240e-01 -9.50031042e-01 1.48355603e-01
7.16932416e-01 -3.80033702e-01 2.45073408e-01 8.75427663e-01
2.20874045e-02 -3.22196752e-01 -1.86470175e+00 1.59518033e-01
-7.45852068e-02 -1.47145760e+00 4.19618279e-01 3.49444449e-02
9.24051702e-01 -4.94682938e-01 -3.97419810e-01 5.06028414e-01
6.86395645e-01 -9.91357565e-01 6.61557376e-01 5.17054975e-01
5.86149275e-01 -7.01633811e-01 7.10478544e-01 7.09798396e-01
-7.15473056e-01 1.09376376e-02 -1.86159104e-01 -4.08015549e-01
-3.01857919e-01 9.84292388e-01 -1.35143149e+00 8.74751210e-01
4.23330158e-01 7.21448541e-01 -5.76907337e-01 7.66071856e-01
-9.82768118e-01 6.08264029e-01 -3.15385878e-01 1.48742467e-01
-1.61975503e-01 1.23308711e-01 3.60868216e-01 9.82743919e-01
3.69303324e-03 5.83359040e-02 2.00421602e-01 8.73719037e-01
-8.53888020e-02 -3.02685261e-01 -9.02253449e-01 2.10191578e-01
5.14733553e-01 8.44756544e-01 -2.32416406e-01 -7.96096385e-01
-4.65723425e-01 3.70294601e-02 4.00007784e-01 5.88717520e-01
-3.79032373e-01 -1.15908824e-01 3.51894736e-01 2.56368607e-01
2.26150453e-01 3.41321409e-01 -2.78869152e-01 -1.02296913e+00
3.69353354e-01 -1.11334097e+00 7.04979539e-01 -9.79836166e-01
-1.14224505e+00 -2.55822507e-03 4.20798481e-01 -6.72490001e-01
-6.43007338e-01 -5.33870935e-01 -7.18025684e-01 8.77718985e-01
-1.26499712e+00 -9.90278721e-01 -2.86461234e-01 8.36561739e-01
-1.29819319e-01 2.02411070e-01 1.18874216e+00 -1.85471147e-01
-3.74942482e-01 3.61798376e-01 -5.67075133e-01 1.70214042e-01
1.96872115e-01 -1.22117019e+00 -1.88219231e-02 1.13967311e+00
7.08937347e-02 7.52600193e-01 9.68288183e-01 -1.03764570e+00
-1.24014270e+00 -1.23903513e+00 1.27903891e+00 -4.64321703e-01
4.68518376e-01 -1.26040846e-01 -1.05712450e+00 1.26547778e+00
-3.49140078e-01 -4.36976738e-02 8.01501095e-01 5.72773695e-01
-6.32158220e-01 -4.51815516e-01 -1.10154879e+00 6.72892153e-01
1.14432216e+00 -1.56940565e-01 -1.10712802e+00 1.12797536e-01
2.50576377e-01 -2.59351820e-01 -9.09372211e-01 8.68479669e-01
4.71242279e-01 -8.11598539e-01 5.56951344e-01 -1.40772259e+00
8.69235694e-01 -5.71948528e-01 -2.32441276e-01 -1.31266749e+00
-1.82520851e-01 -1.76756978e-01 -7.19571829e-01 1.14867604e+00
8.32746327e-01 -5.29904246e-01 9.29269791e-01 8.41803730e-01
-1.43102184e-01 -1.01533973e+00 -5.76174557e-01 -7.13620722e-01
-3.75603855e-01 -6.61269009e-01 7.92571604e-01 9.68359351e-01
6.00297809e-01 3.71052563e-01 -1.17681257e-01 2.33450845e-01
7.18169630e-01 4.76793200e-01 8.84512603e-01 -1.29192185e+00
-6.73618555e-01 -2.71439590e-02 -3.09353948e-01 -5.02683520e-01
1.58655792e-01 -9.53852534e-01 -7.08325356e-02 -1.69063532e+00
2.96286732e-01 -7.65496671e-01 1.98661257e-02 1.17495584e+00
-1.52528614e-01 -3.20695549e-01 -9.49231237e-02 8.01163018e-02
-3.93253148e-01 2.96530835e-02 1.20334458e+00 -1.05928600e-01
-6.40134960e-02 1.41522840e-01 -8.53090882e-01 8.22322428e-01
5.69544673e-01 -4.92252856e-01 -9.39320982e-01 8.77973139e-02
8.08868945e-01 2.61127912e-02 5.02381384e-01 -7.77606845e-01
2.07710907e-01 -4.25697327e-01 4.34832960e-01 -2.15222999e-01
-1.15945674e-01 -7.30341852e-01 6.12977266e-01 6.18028462e-01
-4.00072575e-01 -2.92131811e-01 2.05691665e-01 4.41779137e-01
-1.16763331e-01 -3.68037403e-01 2.60172695e-01 -2.77026713e-01
-7.72267699e-01 -1.32915219e-02 -1.57888010e-01 3.06025356e-01
9.99303937e-01 -1.41320795e-01 -5.45769751e-01 -2.34736487e-01
-1.09441113e+00 4.73638326e-01 4.01015610e-01 8.81750509e-02
7.08017290e-01 -1.17157829e+00 -7.16199279e-01 2.70176500e-01
2.00537711e-01 3.62625390e-01 -2.99389064e-01 6.30488873e-01
-2.87004132e-02 3.58428419e-01 -1.39952317e-01 -2.16966376e-01
-1.02989233e+00 7.36010849e-01 4.62958723e-01 -7.03405023e-01
-4.76576239e-01 4.22982603e-01 9.67085660e-02 -3.24438006e-01
3.37838948e-01 -4.56334382e-01 -1.85793549e-01 -4.80342299e-01
5.22075951e-01 1.46076009e-01 3.95574540e-01 1.31756797e-01
-3.29990208e-01 -1.55529216e-01 -2.53746092e-01 8.71387776e-04
1.43088055e+00 5.08571982e-01 -1.48872763e-01 4.36789304e-01
4.68149602e-01 1.44986123e-01 -6.80865169e-01 -4.92423624e-01
1.14640035e-01 -3.05017889e-01 -5.46167791e-01 -1.28288758e+00
-7.26355553e-01 2.38917455e-01 -3.81805629e-01 2.86479563e-01
1.05649889e+00 1.44783258e-01 7.92631656e-02 6.17676020e-01
5.11093318e-01 -9.81704891e-01 -3.06776822e-01 3.37156862e-01
8.84542048e-01 -6.72035336e-01 6.61709160e-02 -7.88406789e-01
-3.96463752e-01 9.24411952e-01 8.58891964e-01 6.79904670e-02
2.78535187e-01 3.09333682e-01 -3.67924541e-01 -4.10357267e-01
-1.13677824e+00 -2.36419849e-02 2.13183060e-01 7.22439110e-01
-1.45395566e-02 1.01156630e-01 -1.92528740e-01 8.44619334e-01
-2.82184392e-01 4.50400621e-01 5.85688233e-01 9.28139269e-01
-3.26854289e-01 -1.35813820e+00 -2.44992673e-01 9.15592372e-01
-1.52892873e-01 1.13245256e-01 -6.50845706e-01 1.14901268e+00
4.55054402e-01 9.28295612e-01 -5.53353846e-01 -4.87900436e-01
5.96387923e-01 5.13340771e-01 7.73535132e-01 -9.84536409e-01
-1.32175654e-01 -7.14702189e-01 8.39894354e-01 -4.01612133e-01
-2.83763111e-01 -5.56569517e-01 -1.41827261e+00 -3.92029762e-01
-2.62703508e-01 6.32621288e-01 -3.74288153e-04 1.52682912e+00
1.15471885e-01 3.62731069e-01 3.78403097e-01 1.49039939e-01
-4.70462114e-01 -8.03983510e-01 -6.51142955e-01 8.80629659e-01
-1.28404200e-01 -5.14910638e-01 -2.41831943e-01 2.87996829e-01] | [9.046649932861328, 7.1761651039123535] |
e0e31223-dda9-441c-9f1a-4270b49cf6f1 | multi-scale-guided-attention-for-medical | 1906.02849 | null | https://arxiv.org/abs/1906.02849v3 | https://arxiv.org/pdf/1906.02849v3.pdf | Multi-scale self-guided attention for medical image segmentation | Even though convolutional neural networks (CNNs) are driving progress in medical image segmentation, standard models still have some drawbacks. First, the use of multi-scale approaches, i.e., encoder-decoder architectures, leads to a redundant use of information, where similar low-level features are extracted multiple times at multiple scales. Second, long-range feature dependencies are not efficiently modeled, resulting in non-optimal discriminative feature representations associated with each semantic class. In this paper we attempt to overcome these limitations with the proposed architecture, by capturing richer contextual dependencies based on the use of guided self-attention mechanisms. This approach is able to integrate local features with their corresponding global dependencies, as well as highlight interdependent channel maps in an adaptive manner. Further, the additional loss between different modules guides the attention mechanisms to neglect irrelevant information and focus on more discriminant regions of the image by emphasizing relevant feature associations. We evaluate the proposed model in the context of semantic segmentation on three different datasets: abdominal organs, cardiovascular structures and brain tumors. A series of ablation experiments support the importance of these attention modules in the proposed architecture. In addition, compared to other state-of-the-art segmentation networks our model yields better segmentation performance, increasing the accuracy of the predictions while reducing the standard deviation. This demonstrates the efficiency of our approach to generate precise and reliable automatic segmentations of medical images. Our code is made publicly available at https://github.com/sinAshish/Multi-Scale-Attention | ['Jose Dolz', 'Ashish Sinha'] | 2019-06-07 | multi-scale-guided-attention-for-medical-1 | null | null | arxiv-preprint-2019-6 | ['deep-attention', 'attentive-segmentation-networks', 'deep-attention'] | ['computer-vision', 'computer-vision', 'natural-language-processing'] | [ 2.39694774e-01 3.26500028e-01 -1.28260642e-01 -4.02666390e-01
-6.82416022e-01 -3.10374022e-01 3.43551010e-01 5.10423660e-01
-6.79023266e-01 6.79805160e-01 1.12340949e-01 7.42426738e-02
-1.64944813e-01 -7.00792551e-01 -6.57987058e-01 -6.93215251e-01
1.04991250e-01 3.22275907e-01 4.50869203e-01 -1.78429782e-01
1.95269853e-01 6.18471682e-01 -1.31152761e+00 3.76090974e-01
1.04698431e+00 9.48085368e-01 5.48613787e-01 4.41553503e-01
-3.13928038e-01 4.03707117e-01 -4.71579373e-01 -3.14707190e-01
3.05714142e-02 -5.26742280e-01 -8.65664423e-01 -8.81943014e-03
2.74803400e-01 -8.06250647e-02 -2.98747569e-02 1.13203824e+00
4.50890779e-01 -1.54210433e-01 5.34110188e-01 -7.43532360e-01
-4.96159911e-01 4.44260150e-01 -4.72686112e-01 2.97954381e-01
-7.23008513e-02 -2.74577225e-03 9.61903989e-01 -6.12840474e-01
5.16109407e-01 9.48962092e-01 5.49824297e-01 5.12509763e-01
-1.34659290e+00 -3.81823182e-01 2.51401544e-01 1.34323344e-01
-1.36552811e+00 -1.56760231e-01 7.68097341e-01 -3.94105703e-01
7.35022962e-01 1.78248405e-01 7.78276801e-01 9.73722577e-01
4.39894587e-01 7.60455787e-01 9.96398628e-01 -2.24375710e-01
7.27318078e-02 2.65632957e-01 1.70595765e-01 8.92369747e-01
3.90472412e-01 -7.58821070e-02 -1.83052063e-01 1.23062074e-01
8.49147260e-01 1.58834383e-01 -1.47835791e-01 -4.79419261e-01
-1.16265106e+00 8.85702431e-01 8.01193357e-01 9.45921659e-01
-5.14931142e-01 1.17826909e-01 3.84201497e-01 -1.52765676e-01
4.69291866e-01 4.79481608e-01 -4.96060908e-01 2.78472483e-01
-9.94320333e-01 -6.03280403e-02 4.63685393e-01 6.51576281e-01
8.67758572e-01 -1.94173813e-01 -4.17958587e-01 7.31282830e-01
2.69754589e-01 5.17359786e-02 7.43649483e-01 -7.05211580e-01
2.75456548e-01 7.89593160e-01 -3.51789147e-01 -9.88465369e-01
-8.38455141e-01 -7.64087200e-01 -9.53712881e-01 2.06618607e-01
4.56930876e-01 -1.39422059e-01 -1.00608313e+00 1.76883709e+00
2.41879359e-01 -1.20680101e-01 -7.17028975e-02 8.94178331e-01
8.35533440e-01 2.45150521e-01 4.16155279e-01 1.38503104e-01
1.44937456e+00 -1.01971197e+00 -6.89237058e-01 -3.15394998e-01
5.38852394e-01 -6.32665336e-01 7.84630179e-01 2.70823110e-02
-1.05804682e+00 -6.26740813e-01 -8.60746384e-01 -1.81680262e-01
-5.37931621e-01 2.33052298e-01 6.08778954e-01 4.24585819e-01
-1.11512744e+00 8.26175511e-01 -1.01203561e+00 -3.98022622e-01
8.17799449e-01 5.27555645e-01 -3.24574649e-01 2.80001760e-01
-1.04680967e+00 8.07298243e-01 4.90778059e-01 1.41837075e-01
-5.54632425e-01 -6.55668855e-01 -8.41433048e-01 3.19943160e-01
1.02374405e-01 -7.06297159e-01 8.74718368e-01 -1.38098764e+00
-1.38073790e+00 8.88520181e-01 -1.47996247e-01 -4.15591151e-01
4.85956669e-01 -1.79295927e-01 -1.92661192e-02 4.62120384e-01
1.22190922e-01 1.02029145e+00 5.57542145e-01 -1.07330024e+00
-5.35573661e-01 -3.74818355e-01 3.33159231e-02 1.61457628e-01
-4.30897474e-01 -1.57730475e-01 -6.44070506e-01 -7.22639024e-01
1.18344776e-01 -8.12588334e-01 -6.09543681e-01 8.70741382e-02
-3.90970379e-01 1.76477134e-01 4.50297028e-01 -5.79998672e-01
1.09164584e+00 -2.22013497e+00 4.17326272e-01 1.46298721e-01
2.06692010e-01 2.99846649e-01 -5.77986203e-02 1.19682513e-01
-6.30769134e-02 2.00618625e-01 -5.71727216e-01 -3.46281648e-01
-4.05283153e-01 2.38195747e-01 2.70912677e-01 4.08386230e-01
4.04772609e-01 1.06033957e+00 -7.21073747e-01 -7.27903664e-01
3.96145165e-01 7.20454454e-01 -5.59642196e-01 -3.30845043e-02
5.33230342e-02 8.71845603e-01 -6.41775489e-01 5.89840412e-01
4.83253211e-01 -4.46190566e-01 1.26862809e-01 -3.18937838e-01
-6.48889542e-02 6.19101897e-02 -8.31979632e-01 2.11029267e+00
-4.94431704e-01 3.84501278e-01 1.15807891e-01 -1.37473166e+00
7.77006805e-01 3.61448824e-01 6.93574727e-01 -7.62207627e-01
2.76855677e-01 3.12491030e-01 1.84865788e-01 -4.05751348e-01
1.91314057e-01 -4.22561504e-02 -4.96567637e-02 -7.19330758e-02
3.82792950e-01 1.11360736e-01 3.65482986e-01 3.32301706e-02
7.89133906e-01 1.39342859e-01 4.38798040e-01 -5.51091790e-01
7.05665529e-01 -4.70889360e-02 5.67870498e-01 6.35706306e-01
-2.67688453e-01 7.94037580e-01 4.55987900e-01 -4.55066830e-01
-7.44150758e-01 -7.40224242e-01 -4.08985406e-01 8.14318240e-01
3.29043537e-01 -2.69071311e-01 -8.77980351e-01 -8.52507532e-01
-1.71956182e-01 4.42670763e-01 -8.82446110e-01 -1.49149477e-01
-6.06152534e-01 -8.96311522e-01 4.17756200e-01 5.72210252e-01
4.31914687e-01 -1.21786857e+00 -1.06058836e+00 3.44528705e-01
-1.77943572e-01 -1.05819893e+00 -1.93912193e-01 3.69422764e-01
-1.17717278e+00 -1.08104539e+00 -9.66321111e-01 -6.87646747e-01
9.24777687e-01 -1.06991433e-01 1.10569000e+00 2.23997891e-01
-6.34396911e-01 1.25947684e-01 -2.56426990e-01 -1.79856107e-01
-2.59663343e-01 4.33850586e-01 -6.80731475e-01 2.12868601e-02
9.54693630e-02 -4.66394901e-01 -8.29995632e-01 1.92274436e-01
-9.40748453e-01 2.42394120e-01 9.73313153e-01 1.01713300e+00
7.51295149e-01 -4.52186167e-01 6.50428295e-01 -1.07843971e+00
3.24395001e-01 -4.23976570e-01 -5.19181252e-01 1.45930648e-01
-4.71249819e-01 1.41372323e-01 7.15252578e-01 -1.30000591e-01
-9.55844998e-01 2.79173583e-01 -3.99137735e-01 -1.81856334e-01
-5.12274027e-01 3.66776109e-01 5.16827479e-02 -1.43470436e-01
4.49375868e-01 2.03899682e-01 1.22105800e-01 -3.18830818e-01
2.02545047e-01 2.33038425e-01 8.83815214e-02 -3.86836946e-01
3.42523158e-01 5.02009928e-01 9.57964212e-02 -6.81033254e-01
-7.28231311e-01 -3.75018418e-01 -9.53793943e-01 8.30227975e-03
1.23436034e+00 -7.13928640e-01 -3.95179749e-01 3.21173877e-01
-1.06422329e+00 -2.47902766e-01 -4.67917860e-01 5.01107216e-01
-4.58083153e-01 3.14816266e-01 -7.84528196e-01 -4.08927828e-01
-4.25565660e-01 -1.53231525e+00 1.13848162e+00 4.53026295e-01
-2.24360496e-01 -1.13465440e+00 -1.30693316e-01 2.63985574e-01
6.34583712e-01 3.85089487e-01 8.65942538e-01 -7.96547174e-01
-4.80722159e-01 -2.09076554e-01 -3.27197731e-01 2.52176732e-01
2.46377468e-01 -1.27768695e-01 -9.04852211e-01 -1.33243963e-01
-3.20256054e-01 -1.71513706e-01 1.06309521e+00 6.34120822e-01
1.36974049e+00 5.03480844e-02 -5.46283662e-01 6.68901205e-01
1.43374217e+00 1.56262308e-01 4.78499055e-01 1.68351337e-01
6.14838958e-01 6.39106572e-01 3.35366130e-01 2.85885751e-01
3.02915812e-01 6.08059227e-01 5.27277291e-01 -6.88513398e-01
-2.56089300e-01 1.44436285e-01 -1.28498033e-01 4.96840388e-01
-1.40863881e-01 -3.92078720e-02 -7.68166840e-01 6.88947439e-01
-1.80380845e+00 -6.59504175e-01 -9.62992292e-03 2.06835699e+00
7.29456961e-01 9.04463679e-02 -9.70837623e-02 -1.21231735e-01
5.92368126e-01 1.21329822e-01 -3.65410835e-01 -4.57282245e-01
3.81185152e-02 4.14319336e-01 5.06161630e-01 4.61823314e-01
-1.34257400e+00 8.72013092e-01 5.72125196e+00 7.91073918e-01
-1.28397548e+00 2.43472204e-01 8.54647279e-01 8.83771181e-02
-1.79992720e-01 -2.59625971e-01 -6.67702258e-01 4.91229177e-01
7.83171892e-01 1.66130140e-01 -9.42774396e-03 6.71624541e-01
-5.89136817e-02 -5.96090369e-02 -8.65529299e-01 5.51275551e-01
5.59220128e-02 -1.33886302e+00 -1.02782287e-01 -1.44079793e-02
4.86903936e-01 1.43248379e-01 -5.60651831e-02 8.85694996e-02
-4.15247306e-02 -9.64895904e-01 6.05503798e-01 4.75174755e-01
6.28109217e-01 -6.09449804e-01 9.11799371e-01 1.47658229e-01
-1.13344705e+00 -1.01061746e-01 -2.42134780e-01 3.29374194e-01
1.38291076e-01 6.13113165e-01 -6.88489616e-01 7.50212491e-01
6.98012471e-01 6.19163513e-01 -7.34521925e-01 1.09094763e+00
-1.85136735e-01 2.60882646e-01 -2.73031175e-01 6.83388207e-03
3.47802401e-01 -1.17572948e-01 3.73592287e-01 1.44867039e+00
2.70659328e-01 -8.83553401e-02 4.07069884e-02 1.01109326e+00
9.50819030e-02 4.03164566e-01 -3.51990312e-01 1.81041405e-01
-1.07211396e-01 1.57396841e+00 -1.16820300e+00 -4.03859973e-01
-5.44386506e-01 1.07600021e+00 3.74893397e-01 2.69001871e-01
-9.22598481e-01 -2.83333898e-01 3.92181605e-01 4.50613312e-02
4.32519883e-01 -2.27868855e-02 -5.10785341e-01 -1.08066773e+00
-2.17839226e-01 -5.01180410e-01 4.42071557e-01 -2.39496067e-01
-9.88205016e-01 9.01544631e-01 -1.12562075e-01 -1.12207782e+00
5.21460455e-03 -4.72674519e-01 -4.44425672e-01 8.78061414e-01
-1.75232697e+00 -1.21950412e+00 -3.14671040e-01 5.45861781e-01
6.63853884e-01 1.04396231e-01 1.06333458e+00 4.85616863e-01
-5.25783956e-01 5.16108155e-01 -1.57868396e-02 6.22278675e-02
5.27442873e-01 -1.27482605e+00 6.35715276e-02 6.82732701e-01
1.73574522e-01 4.72200006e-01 4.09710467e-01 -4.67171937e-01
-6.43795609e-01 -9.59110200e-01 6.87511563e-01 -2.43404489e-02
3.70901793e-01 -2.61491895e-01 -1.00894070e+00 3.86529088e-01
2.47905970e-01 3.69979054e-01 7.43697286e-01 -3.18204015e-02
-4.92094643e-02 -9.38723609e-02 -1.19327736e+00 3.92004579e-01
7.59931207e-01 -1.48703277e-01 -4.25180674e-01 2.11150825e-01
5.04474342e-01 -3.06682020e-01 -8.63740325e-01 5.20744085e-01
5.03544033e-01 -1.20583105e+00 9.87376511e-01 -2.89974451e-01
5.71322680e-01 -1.16107613e-01 1.51356131e-01 -1.18976378e+00
-4.60057527e-01 -6.69778958e-02 2.21693277e-01 9.66575682e-01
5.78216136e-01 -6.60773754e-01 6.32635534e-01 3.90424132e-01
-3.54722977e-01 -1.04602778e+00 -1.11192775e+00 -3.11253577e-01
2.26884142e-01 7.97402635e-02 2.81158000e-01 7.60189235e-01
-1.45284146e-01 2.12766483e-01 -1.07743226e-01 4.84158844e-02
4.55989689e-01 3.54121178e-01 2.05427662e-01 -1.17456424e+00
-3.55523735e-01 -6.96434677e-01 -4.84500796e-01 -7.27444768e-01
5.24921864e-02 -9.71119702e-01 -9.15739387e-02 -1.61156154e+00
1.94467157e-01 -4.33676451e-01 -6.56060874e-01 5.73938847e-01
-2.04623595e-01 4.57281142e-01 2.71531522e-01 9.33300555e-02
-5.51334918e-01 3.88151318e-01 1.43413949e+00 -9.56941843e-02
-1.00525230e-01 -3.99465412e-02 -7.11669683e-01 7.93277323e-01
9.84955013e-01 -4.18363273e-01 -1.42939463e-01 -4.27263558e-01
-3.08345348e-01 -1.42303202e-02 3.81930977e-01 -1.05112219e+00
8.19786862e-02 1.37321800e-01 6.90260172e-01 -2.52056807e-01
2.00688884e-01 -9.88417625e-01 -4.35041972e-02 8.24781775e-01
-4.14864957e-01 -1.28649309e-01 3.88710856e-01 2.86163777e-01
-4.97276992e-01 -2.64624238e-01 1.00709331e+00 -4.15457308e-01
-5.54185867e-01 2.44080365e-01 -3.29940766e-01 -1.29870266e-01
1.07935047e+00 -1.46217152e-01 -3.53108067e-03 6.69731200e-02
-9.25324380e-01 3.29095647e-02 3.29092562e-01 2.96164721e-01
3.77643943e-01 -9.50559735e-01 -5.66893041e-01 3.67008567e-01
1.31595535e-02 1.07605159e-01 5.56462765e-01 1.20103264e+00
-6.12177253e-01 5.40785193e-01 -4.99423176e-01 -7.44476199e-01
-1.05071795e+00 5.13716102e-01 5.16134977e-01 -5.46696126e-01
-7.53129363e-01 7.27070272e-01 6.89637125e-01 -1.86066344e-01
6.37608394e-03 -5.75906992e-01 -5.19441187e-01 1.07446209e-01
2.77680546e-01 -7.57105974e-03 1.78508162e-01 -7.12136567e-01
-4.37559426e-01 7.86194444e-01 -1.15813464e-01 2.55455256e-01
1.22746396e+00 -1.48495659e-01 -4.76149693e-02 1.79585174e-01
1.08424270e+00 -1.30768731e-01 -1.21313739e+00 -5.35100289e-02
2.70332526e-02 -3.30750763e-01 8.29219166e-03 -8.58476400e-01
-1.47348511e+00 1.08090174e+00 7.69401133e-01 5.47004417e-02
1.26550591e+00 1.35043904e-01 6.49828136e-01 -1.20145239e-01
1.29810140e-01 -9.19669390e-01 -6.97988421e-02 1.82522207e-01
7.62548566e-01 -1.34489775e+00 -1.17482066e-01 -4.80146408e-01
-6.58962667e-01 1.16318071e+00 6.78902507e-01 -2.27746218e-01
4.82413322e-01 3.05540472e-01 9.78033170e-02 -2.32534632e-01
-3.76976490e-01 -4.09776121e-01 3.77330720e-01 2.79425353e-01
7.85220027e-01 1.01052225e-02 -6.36765361e-01 6.59036338e-01
2.82326102e-01 -6.47665039e-02 2.40911901e-01 7.30916739e-01
-2.29940936e-01 -1.20508158e+00 -1.33988395e-01 2.84953088e-01
-8.62887621e-01 -1.22390240e-01 -6.63708448e-02 8.62134874e-01
2.35503137e-01 4.87634301e-01 3.57660390e-02 8.43355283e-02
2.78143346e-01 4.55607362e-02 3.65284413e-01 -5.64239025e-01
-8.75108957e-01 4.14804995e-01 -2.98281819e-01 -7.21223474e-01
-4.80628401e-01 -5.08210719e-01 -1.42397344e+00 3.52218360e-01
-2.17092633e-01 -1.67608424e-03 6.52927458e-01 8.28960717e-01
5.44872046e-01 1.03148305e+00 2.33728409e-01 -9.13852990e-01
-2.58411735e-01 -8.07099700e-01 -3.57906729e-01 6.42396629e-01
2.29054064e-01 -6.24904692e-01 -8.28271657e-02 -4.85959947e-02] | [14.625965118408203, -2.570199728012085] |
a957018d-043c-4322-8860-e4dffec05d57 | speech-denoising-in-the-waveform-domain-with | 2202.07790 | null | https://arxiv.org/abs/2202.07790v3 | https://arxiv.org/pdf/2202.07790v3.pdf | Speech Denoising in the Waveform Domain with Self-Attention | In this work, we present CleanUNet, a causal speech denoising model on the raw waveform. The proposed model is based on an encoder-decoder architecture combined with several self-attention blocks to refine its bottleneck representations, which is crucial to obtain good results. The model is optimized through a set of losses defined over both waveform and multi-resolution spectrograms. The proposed method outperforms the state-of-the-art models in terms of denoised speech quality from various objective and subjective evaluation metrics. We release our code and models at https://github.com/nvidia/cleanunet. | ['Bryan Catanzaro', 'Ambrish Dantrey', 'Wei Ping', 'Zhifeng Kong'] | 2022-02-15 | null | null | null | null | ['speech-denoising'] | ['speech'] | [-1.14039898e-01 -2.76414514e-01 2.71344692e-01 -4.42123055e-01
-1.01818383e+00 -1.22116074e-01 3.73059034e-01 -1.04991153e-01
-2.23916709e-01 6.12325013e-01 6.10990226e-01 -1.57285377e-01
1.34510666e-01 -3.85273457e-01 -6.24250889e-01 -6.10539496e-01
-5.91880903e-02 -2.14266405e-01 2.33067319e-01 -3.89047503e-01
7.45153874e-02 4.77648899e-02 -1.36633229e+00 3.75389040e-01
1.12640190e+00 1.02216208e+00 5.29748201e-01 9.74153697e-01
2.55981863e-01 9.38304245e-01 -7.42317319e-01 -5.64940512e-01
-5.67448419e-03 -6.44136250e-01 -3.73771101e-01 -3.43014479e-01
1.39048949e-01 -4.53349292e-01 -6.95165455e-01 1.48413336e+00
8.27695727e-01 1.92244515e-01 2.57450759e-01 -8.09407413e-01
-9.24240351e-01 6.75219238e-01 -3.30657333e-01 6.79670095e-01
1.76588461e-01 2.05422476e-01 8.79629374e-01 -1.02689183e+00
1.00273877e-01 1.49418592e+00 5.37937462e-01 4.42815036e-01
-1.03577578e+00 -7.16126561e-01 7.61654079e-02 7.14827538e-01
-1.42664003e+00 -1.13684845e+00 8.46122861e-01 1.59553662e-02
1.17841363e+00 1.41671345e-01 2.94522196e-01 1.27200246e+00
4.19294745e-01 5.47013223e-01 8.64423990e-01 -2.77038664e-01
1.39244467e-01 -3.42504978e-01 1.11661099e-01 3.47332150e-01
-1.28107160e-01 3.93767834e-01 -8.63564372e-01 4.38437657e-03
6.89392209e-01 -5.30764759e-01 -4.96247113e-01 2.98048407e-01
-7.10690260e-01 6.86540902e-01 4.80951041e-01 2.15106964e-01
-6.90005064e-01 3.57254088e-01 5.09802639e-01 2.15293765e-01
8.19564164e-01 -6.27020597e-02 -1.34534806e-01 -4.67924446e-01
-1.16966224e+00 -3.18732671e-02 4.15733218e-01 6.78676367e-01
2.39337713e-01 8.20564508e-01 -1.68279350e-01 1.15139461e+00
4.51935798e-01 4.96372312e-01 5.38228571e-01 -1.02565968e+00
5.03324091e-01 -2.42881238e-01 -1.20638572e-01 -6.88256621e-01
-6.68517798e-02 -6.35557652e-01 -1.14328349e+00 1.97809309e-01
-4.26634997e-01 -2.55036414e-01 -1.05211163e+00 1.61110067e+00
2.85576209e-02 6.86951518e-01 7.39999563e-02 1.18688416e+00
9.31970894e-01 1.06169522e+00 -4.59651835e-02 -2.78390616e-01
1.06254482e+00 -1.29463530e+00 -1.48423052e+00 -2.75113732e-01
-2.78592497e-01 -1.08368039e+00 9.27472889e-01 7.67217219e-01
-1.51620555e+00 -9.03550744e-01 -1.38206959e+00 -3.10510129e-01
9.66004580e-02 1.14868507e-01 5.89688756e-02 6.78139865e-01
-1.20581949e+00 7.67810524e-01 -1.10984528e+00 1.41800538e-01
2.92992145e-01 1.12454496e-01 6.09504804e-02 8.19019005e-02
-1.42668664e+00 8.27861130e-01 1.78509906e-01 4.43397641e-01
-1.32587409e+00 -4.81291622e-01 -7.95854509e-01 2.84525245e-01
2.25545272e-01 -6.06288016e-01 1.56268942e+00 -7.64389277e-01
-1.96159077e+00 6.92376941e-02 -5.24551868e-01 -6.95690334e-01
2.56714582e-01 -7.54655778e-01 -9.00902510e-01 1.58360392e-01
-2.58598566e-01 2.35134766e-01 1.37252021e+00 -1.15373194e+00
-3.77577037e-01 -4.55089137e-02 -3.35413456e-01 1.28190234e-01
-2.02202588e-01 1.97847992e-01 -5.71174800e-01 -1.11757088e+00
-2.09347099e-01 -3.85011941e-01 -1.51942492e-01 -5.67893267e-01
-3.90632838e-01 2.12204114e-01 7.06151009e-01 -1.23850930e+00
1.56770575e+00 -2.25436234e+00 3.28385234e-01 -2.52196163e-01
2.38244429e-01 6.75948024e-01 -2.67379701e-01 3.71985674e-01
-2.52169877e-01 2.14832649e-01 -2.73866653e-01 -6.65323317e-01
-5.56251146e-02 -6.58407360e-02 -4.00164902e-01 4.38141048e-01
2.99452722e-01 4.37874645e-01 -6.68567479e-01 7.75887119e-03
3.37728977e-01 1.01659358e+00 -4.92356330e-01 4.96036798e-01
3.43999863e-02 4.94746059e-01 -1.86534822e-01 2.50927657e-01
9.58559811e-01 1.99957252e-01 -3.42136174e-02 -2.39743069e-01
-2.36810774e-01 8.96401882e-01 -1.03724837e+00 1.84695506e+00
-5.27946413e-01 5.84387898e-01 5.33863664e-01 -6.80359602e-01
8.81935239e-01 6.54284954e-01 -3.83691564e-02 -7.31291771e-01
2.27680132e-01 2.51617968e-01 4.88446020e-02 -5.10010302e-01
5.72627425e-01 -5.16340174e-02 3.82077724e-01 9.37109441e-02
2.42796242e-01 -1.85107663e-01 2.31223814e-02 1.41042292e-01
1.11425245e+00 -6.52078465e-02 1.63306043e-01 -2.40155399e-01
6.76940084e-01 -5.53334951e-01 7.92717218e-01 4.45588171e-01
-2.38150820e-01 8.76630366e-01 4.29329157e-01 2.45291479e-02
-1.16119468e+00 -1.32605314e+00 -1.03566453e-01 8.89734089e-01
3.48543115e-02 -6.13365293e-01 -9.50306833e-01 -2.70949975e-02
-3.22096020e-01 1.04155707e+00 -2.78488994e-01 -3.56527507e-01
-5.06426454e-01 -6.81894958e-01 7.25058794e-01 3.31600964e-01
5.16789258e-01 -9.48747575e-01 -2.37739444e-01 3.31931710e-01
-3.26571822e-01 -1.00692952e+00 -7.39490569e-01 1.91419661e-01
-7.30123758e-01 -5.92033744e-01 -5.48452497e-01 -6.00833237e-01
1.35770738e-01 2.22958699e-01 9.77111995e-01 1.89677870e-03
6.77179173e-02 -1.64200723e-01 -6.31943822e-01 -2.37382218e-01
-5.62509775e-01 8.76786411e-02 6.43892512e-02 2.25748047e-01
-1.30068168e-01 -9.70076144e-01 -6.60658479e-01 -5.89044876e-02
-9.46766019e-01 8.85671377e-02 4.10287559e-01 7.92507946e-01
5.44282556e-01 1.82469994e-01 7.81537592e-01 -3.17215741e-01
1.08397484e+00 -5.39692104e-01 -6.23155117e-01 -5.78779168e-02
-3.30835819e-01 1.02878362e-01 7.80230105e-01 -1.93036303e-01
-1.30843782e+00 -4.74341333e-01 -7.95344174e-01 -6.60154521e-01
4.16275561e-02 3.96519959e-01 -1.76916018e-01 2.44247705e-01
4.48507637e-01 2.11884186e-01 -9.94900241e-02 -9.20347929e-01
3.63184899e-01 9.97762799e-01 8.29026699e-01 -2.55244046e-01
5.25012970e-01 9.37145725e-02 -5.77939570e-01 -9.07912016e-01
-7.09610701e-01 -2.64838576e-01 -3.51197332e-01 -8.70472267e-02
7.67261088e-01 -1.21688402e+00 -3.77000093e-01 7.11092532e-01
-1.50070584e+00 -4.09606993e-01 5.65736406e-02 5.66818178e-01
-3.23889881e-01 3.35347831e-01 -9.48534906e-01 -9.39221025e-01
-7.24242449e-01 -1.38397849e+00 8.96443844e-01 1.47319868e-01
1.32888243e-01 -6.60649300e-01 1.46116227e-01 9.07164663e-02
8.33583891e-01 -2.36771390e-01 5.95219553e-01 -1.95605546e-01
-3.73265445e-01 2.39284486e-01 -1.24696657e-01 1.04511154e+00
9.91491452e-02 -4.85756546e-02 -1.14346111e+00 -3.95988584e-01
4.46589649e-01 -1.50618806e-01 1.20875132e+00 5.75425386e-01
1.35099936e+00 -1.39717415e-01 2.82202899e-01 9.67493832e-01
1.24942875e+00 3.65580857e-01 1.01504683e+00 -4.24640216e-02
5.14960945e-01 5.20889796e-02 1.36335194e-01 5.06332636e-01
2.19474152e-01 5.41088760e-01 4.70888585e-01 6.76789507e-02
-6.14773214e-01 -3.42359468e-02 5.92863142e-01 1.70454288e+00
5.79320453e-03 -4.61500406e-01 -7.66333580e-01 6.31219268e-01
-1.70413065e+00 -1.00617993e+00 -2.32206717e-01 1.91229057e+00
8.80961537e-01 2.16438428e-01 -3.01031083e-01 2.00501904e-01
8.89325917e-01 5.64796865e-01 -5.98570943e-01 -6.86080337e-01
-1.75422132e-01 6.40860915e-01 1.52359918e-01 9.29684460e-01
-1.04772234e+00 7.84901500e-01 6.30461645e+00 1.00008297e+00
-1.17711473e+00 5.96362889e-01 5.58860064e-01 -2.93050975e-01
-2.02932209e-01 -2.68575341e-01 -3.76484334e-01 5.17136455e-01
1.47902906e+00 -3.35550040e-01 9.32188153e-01 4.91692126e-01
8.09216201e-01 1.07420407e-01 -5.49260080e-01 8.33120584e-01
1.71407908e-01 -1.06620204e+00 -1.78056344e-01 -3.01700771e-01
5.40063739e-01 3.50249439e-01 1.42904311e-01 1.42831266e-01
1.92004710e-01 -9.94234204e-01 1.08742416e+00 5.25598288e-01
8.19754958e-01 -8.80499244e-01 6.40938044e-01 6.42805770e-02
-1.19067907e+00 5.48792183e-02 -3.21740955e-01 8.35972000e-03
4.92211789e-01 9.43441749e-01 -2.99531013e-01 7.22748578e-01
8.16347480e-01 7.21369803e-01 -1.74790144e-01 1.08354950e+00
-6.85464501e-01 1.09566677e+00 -8.53331313e-02 4.42768633e-01
6.78313002e-02 -9.68786255e-02 6.79987192e-01 1.28477240e+00
4.97970402e-01 2.51424968e-01 -4.02762651e-01 7.87659883e-01
-1.79149240e-01 -2.32326627e-01 3.28492839e-03 2.41610244e-01
6.28879964e-01 9.65838730e-01 -2.66076680e-02 -3.60024780e-01
-3.79841447e-01 9.95260715e-01 2.51055568e-01 4.96004850e-01
-1.19570506e+00 -6.59816563e-01 1.06284630e+00 -1.46802738e-01
6.85333490e-01 -3.42562109e-01 -3.84032637e-01 -1.22745454e+00
-7.52489828e-03 -1.11728680e+00 -1.27217367e-01 -1.05614626e+00
-1.12646294e+00 1.19580626e+00 -2.78242201e-01 -9.98288095e-01
-1.31609693e-01 -1.79693580e-01 -7.10454643e-01 1.15441453e+00
-1.67896676e+00 -6.28419638e-01 -3.65334600e-01 5.73415875e-01
7.69051909e-01 1.87006276e-02 6.90983236e-01 6.55595958e-01
-9.19742942e-01 4.82057422e-01 2.40981445e-01 -2.40327284e-01
6.95440114e-01 -9.91988361e-01 1.01611519e+00 1.42322743e+00
3.39922383e-02 4.44436520e-01 9.26434278e-01 -6.42102540e-01
-1.14099824e+00 -1.16747046e+00 7.35711515e-01 2.66790986e-01
6.76456392e-01 -3.17706436e-01 -1.15594721e+00 4.40198481e-01
1.08298135e+00 -4.08430994e-02 2.68114507e-01 -2.20360637e-01
-2.91224092e-01 -2.51746744e-01 -8.04755330e-01 4.30522561e-01
8.81003022e-01 -5.49931288e-01 -5.84004998e-01 -1.45259544e-01
9.76873219e-01 -5.53941607e-01 -7.69026995e-01 3.16815823e-01
1.93836004e-01 -1.16928971e+00 9.13587153e-01 -2.02624500e-02
4.43571299e-01 -4.29944128e-01 -2.37813711e-01 -1.87683880e+00
-4.84112114e-01 -9.99255002e-01 -3.78280938e-01 1.25523579e+00
3.46770644e-01 -5.13107359e-01 6.89293370e-02 -2.39172474e-01
-7.60512710e-01 -5.06853402e-01 -1.23619199e+00 -6.90337181e-01
-1.89640895e-01 -7.72871792e-01 6.32887363e-01 3.57786387e-01
-3.24857593e-01 4.69161958e-01 -7.82443225e-01 5.33720613e-01
7.44341195e-01 -4.57750022e-01 2.87389010e-01 -5.59777498e-01
-2.06401438e-01 -4.23104733e-01 -4.50709760e-02 -1.08985102e+00
9.37016979e-02 -5.22946000e-01 2.81779677e-01 -1.49600661e+00
-1.51050657e-01 2.01509655e-01 -4.06354994e-01 2.24052891e-01
-2.74792552e-01 8.49061310e-02 2.49167785e-01 -9.33525711e-02
-4.33039159e-01 1.29585326e+00 1.05737031e+00 -2.55721342e-02
-3.70277837e-02 -2.13637739e-01 -6.04019523e-01 5.88595688e-01
1.24060416e+00 -5.91539443e-01 -3.93517554e-01 -1.05338788e+00
-3.43260407e-01 1.42159685e-01 4.76505518e-01 -1.25440860e+00
2.66397893e-01 1.90435365e-01 4.47900407e-02 -5.32989264e-01
7.88140893e-01 -4.06552970e-01 1.40449464e-01 3.95118862e-01
-3.34055245e-01 1.95171684e-01 4.03769135e-01 3.82296443e-01
-6.36429727e-01 -5.16399667e-02 1.08723915e+00 2.42687583e-01
-5.24955034e-01 3.39092389e-02 -4.00995940e-01 1.16161197e-01
4.40311998e-01 4.48760718e-01 -5.91869473e-01 -6.67672217e-01
-6.69554293e-01 -5.03644906e-02 -5.40101640e-02 4.93801326e-01
9.68415260e-01 -1.39951408e+00 -1.06938076e+00 3.53247136e-01
-4.82629806e-01 -3.39823455e-01 5.42823970e-01 7.74966776e-01
-5.43548703e-01 3.77004862e-01 -1.21864125e-01 -2.45057747e-01
-1.09177732e+00 3.40270162e-01 4.33833778e-01 -1.26413047e-01
-5.69415390e-01 7.75504112e-01 -1.55118611e-02 7.88433924e-02
3.91722411e-01 -4.30383831e-01 3.91950496e-02 -3.14558715e-01
8.39717567e-01 5.28195381e-01 3.98259461e-01 -7.13565230e-01
-2.95425981e-01 1.26077116e-01 5.79600707e-02 -3.56494397e-01
1.38940585e+00 -4.10715014e-01 -1.79906875e-01 2.33010843e-01
1.09095395e+00 5.39905466e-02 -1.35940695e+00 1.39953392e-02
-4.41605330e-01 -4.99044508e-01 7.17320621e-01 -8.78550828e-01
-1.29527545e+00 9.88085330e-01 6.69312119e-01 2.00366914e-01
1.64722943e+00 -4.24645305e-01 1.14621365e+00 -3.18883479e-01
4.80597354e-02 -9.43798304e-01 1.80384330e-02 8.37175131e-01
1.30151510e+00 -9.04203117e-01 -1.71566054e-01 -3.88418943e-01
-5.48896849e-01 8.26030970e-01 4.02683258e-01 -3.32655758e-01
8.25615823e-01 6.48868024e-01 3.00137073e-01 1.30797639e-01
-1.14124620e+00 -3.09128851e-01 1.61358267e-01 5.52708328e-01
6.46853507e-01 1.72213539e-02 -3.66293252e-01 9.26497817e-01
-3.00074399e-01 -1.07015774e-01 4.82890695e-01 3.95438612e-01
-3.87402266e-01 -1.14018607e+00 -6.01726651e-01 2.37569422e-01
-7.27548957e-01 -5.12962818e-01 9.36707407e-02 2.51945645e-01
-4.19325531e-02 1.51782107e+00 -5.82889840e-02 -6.88339114e-01
4.73432213e-01 -8.98784027e-02 2.05078706e-01 -3.37045699e-01
-7.11452007e-01 5.97880185e-01 2.35902920e-01 -5.57479441e-01
-2.51766264e-01 -4.97301728e-01 -1.03403127e+00 -5.27458429e-01
-1.54164568e-01 1.03892595e-01 7.21183479e-01 5.24378181e-01
5.53413808e-01 1.26303172e+00 7.93080330e-01 -1.01607335e+00
-5.59283912e-01 -1.31038582e+00 -4.77652639e-01 3.47294211e-02
7.65283167e-01 -3.51179093e-01 -3.89643461e-01 1.88704893e-01] | [14.978015899658203, 5.9712748527526855] |
b5a5a284-8fa3-49f6-9837-7a9f26fb7143 | mv6d-multi-view-6d-pose-estimation-on-rgb-d | 2208.01172 | null | https://arxiv.org/abs/2208.01172v1 | https://arxiv.org/pdf/2208.01172v1.pdf | MV6D: Multi-View 6D Pose Estimation on RGB-D Frames Using a Deep Point-wise Voting Network | Estimating 6D poses of objects is an essential computer vision task. However, most conventional approaches rely on camera data from a single perspective and therefore suffer from occlusions. We overcome this issue with our novel multi-view 6D pose estimation method called MV6D which accurately predicts the 6D poses of all objects in a cluttered scene based on RGB-D images from multiple perspectives. We base our approach on the PVN3D network that uses a single RGB-D image to predict keypoints of the target objects. We extend this approach by using a combined point cloud from multiple views and fusing the images from each view with a DenseFusion layer. In contrast to current multi-view pose detection networks such as CosyPose, our MV6D can learn the fusion of multiple perspectives in an end-to-end manner and does not require multiple prediction stages or subsequent fine tuning of the prediction. Furthermore, we present three novel photorealistic datasets of cluttered scenes with heavy occlusions. All of them contain RGB-D images from multiple perspectives and the ground truth for instance semantic segmentation and 6D pose estimation. MV6D significantly outperforms the state-of-the-art in multi-view 6D pose estimation even in cases where the camera poses are known inaccurately. Furthermore, we show that our approach is robust towards dynamic camera setups and that its accuracy increases incrementally with an increasing number of perspectives. | ['Gerhard Neumann', 'Tobias Demmler', 'Fabian Duffhauss'] | 2022-08-01 | null | null | null | null | ['6d-pose-estimation-1'] | ['computer-vision'] | [-5.09974249e-02 -5.68747744e-02 1.14545777e-01 -3.23681772e-01
-7.45346844e-01 -9.68335509e-01 4.89109755e-01 -1.17422521e-01
-4.27210718e-01 -3.70372459e-02 -1.89914063e-01 1.05241172e-01
1.36438861e-01 -5.25197268e-01 -1.03504324e+00 -4.71929222e-01
5.40442705e-01 1.06582952e+00 7.55427063e-01 -3.40410769e-02
1.86237857e-01 9.08397973e-01 -1.60390294e+00 4.17693295e-02
3.96505356e-01 1.20295894e+00 3.40922862e-01 8.27393174e-01
2.05754653e-01 1.74451411e-01 -1.75728634e-01 -4.43787426e-01
7.90770769e-01 2.08681732e-01 -4.33231562e-01 6.47383630e-01
1.01530552e+00 -6.67072833e-01 -2.16538936e-01 7.90728807e-01
3.45348001e-01 -3.22900191e-02 2.54796535e-01 -1.26712966e+00
1.26254231e-01 -3.05278838e-01 -8.55130613e-01 -8.98196027e-02
6.09773338e-01 1.58598661e-01 8.30791593e-01 -1.15814734e+00
7.59503305e-01 1.27058733e+00 7.18163729e-01 2.68614769e-01
-1.29245496e+00 -3.18738848e-01 4.34278101e-01 1.99636985e-02
-1.22470832e+00 -2.83213258e-01 9.71322656e-01 -4.26290929e-01
1.09828651e+00 8.99193808e-02 8.44268501e-01 9.32644427e-01
-2.80964077e-02 8.25638890e-01 1.06548333e+00 -1.47212654e-01
1.32475019e-01 -1.66893467e-01 -2.55857557e-01 6.13698542e-01
2.41745606e-01 4.78920573e-03 -6.45089686e-01 -3.72076482e-02
9.84364629e-01 4.56486851e-01 -5.17644957e-02 -1.41093385e+00
-1.45208347e+00 5.91759801e-01 4.12889987e-01 -2.33853698e-01
-4.75805461e-01 2.67987460e-01 -2.24302057e-03 -1.24737971e-01
5.08375645e-01 1.74747765e-01 -7.96097815e-01 6.13822527e-02
-7.10451007e-01 4.02620822e-01 5.27113438e-01 1.11916852e+00
7.33942270e-01 -3.65578115e-01 4.17189628e-01 3.36532414e-01
4.69079792e-01 7.85502374e-01 -9.67882127e-02 -1.42497945e+00
7.12075949e-01 8.56484890e-01 5.07480383e-01 -9.55225885e-01
-5.21135032e-01 -3.69835258e-01 -2.69955367e-01 3.44412416e-01
4.83780473e-01 1.42389536e-01 -1.05157447e+00 1.19544017e+00
8.88064921e-01 -1.29146874e-01 -1.78879157e-01 1.13516569e+00
6.71086371e-01 2.91706532e-01 -5.74781239e-01 8.74208584e-02
1.20152378e+00 -8.96779001e-01 -1.70177668e-01 -6.19021297e-01
1.38887703e-01 -9.31836843e-01 6.09474242e-01 7.17591584e-01
-1.16430461e+00 -4.88694668e-01 -9.11012411e-01 -3.06633234e-01
-2.93252110e-01 8.51599723e-02 5.50671458e-01 4.03475434e-01
-8.28789115e-01 3.96419317e-01 -1.15061510e+00 -3.73138011e-01
4.27601844e-01 5.90606153e-01 -7.83942699e-01 -4.25624698e-01
-3.63646507e-01 1.09512758e+00 3.50819051e-01 1.85770839e-01
-8.84155631e-01 -7.26359248e-01 -9.59009171e-01 -2.24041045e-01
8.49417150e-01 -9.72258270e-01 1.28432024e+00 -6.14294469e-01
-1.29447103e+00 1.01784635e+00 -9.70169380e-02 -5.95885552e-02
8.29786599e-01 -7.26686656e-01 2.48241097e-01 4.57173854e-01
2.11100769e-03 7.47541487e-01 7.18260765e-01 -1.61653197e+00
-5.83235145e-01 -9.39977229e-01 3.14080477e-01 6.91925585e-01
2.45761439e-01 -2.74816453e-01 -1.06211030e+00 -5.79195134e-02
7.52439141e-01 -1.32525814e+00 -5.11234581e-01 4.63532507e-01
-4.09429759e-01 1.53686583e-01 9.54637349e-01 -3.43528509e-01
1.29617825e-01 -1.88231111e+00 5.94431102e-01 6.86571077e-02
3.57503921e-01 1.67341232e-01 6.42198622e-02 2.49538854e-01
1.75981402e-01 -4.09102559e-01 5.74824139e-02 -7.22161591e-01
-6.06824607e-02 4.32195067e-01 7.62253106e-02 8.01092565e-01
2.49880049e-02 7.88537204e-01 -8.19397509e-01 -3.37907583e-01
8.30495238e-01 7.70987809e-01 -5.27828634e-01 3.44931155e-01
-4.38473761e-01 6.01616859e-01 -3.71492743e-01 7.68781006e-01
9.28463817e-01 -3.30285996e-01 -1.15199611e-01 -5.62401891e-01
-8.72246623e-02 -5.05513847e-02 -1.48909009e+00 2.31547832e+00
-4.35210139e-01 1.62261143e-01 2.50538439e-01 -5.87671340e-01
6.23708546e-01 2.28156567e-01 7.35259295e-01 -2.34540626e-01
2.01064065e-01 1.82818457e-01 -3.46018761e-01 -3.23284626e-01
2.09649816e-01 -6.87329769e-02 -4.93828990e-02 1.77534923e-01
1.15442954e-01 -5.62411129e-01 -7.05455022e-05 1.29783571e-01
8.08393240e-01 6.56034291e-01 3.28664094e-01 4.08859879e-01
3.60320151e-01 5.68640307e-02 5.02563059e-01 3.19701940e-01
8.29393417e-02 1.07987881e+00 2.64789104e-01 -5.47013581e-01
-1.22709894e+00 -1.27335107e+00 -2.23584715e-02 4.59867805e-01
4.64921892e-01 -3.61648083e-01 -3.79290372e-01 -7.91518986e-01
3.19496065e-01 4.15034682e-01 -3.65879387e-01 3.52499604e-01
-5.25085986e-01 -5.09251893e-01 -1.79579586e-01 6.57652080e-01
4.32351708e-01 -3.11656803e-01 -1.17818820e+00 3.37479599e-02
-1.40790105e-01 -1.77801573e+00 -2.72559315e-01 2.76565909e-01
-9.83440697e-01 -1.32104385e+00 -6.60928190e-01 -2.18743995e-01
5.75066209e-01 7.23624229e-01 1.16455650e+00 -4.54425156e-01
-1.55718088e-01 8.99971068e-01 -2.24698916e-01 -2.60153413e-01
-9.34669822e-02 -2.23896459e-01 1.04894489e-01 5.57275340e-02
2.25351453e-01 -7.51313508e-01 -9.19893146e-01 5.39352655e-01
-5.67397177e-01 2.48836607e-01 6.87535107e-01 2.64689088e-01
1.03428602e+00 -2.26847723e-01 -3.03273201e-01 -6.52501523e-01
-4.89502758e-01 -5.70065640e-02 -1.00897396e+00 2.52031237e-02
-2.82276213e-01 -2.51110405e-01 2.74849862e-01 -2.10772008e-01
-8.58988225e-01 9.35056686e-01 1.00398816e-01 -1.21986091e+00
-3.78795236e-01 -1.47997186e-01 -3.62272739e-01 -3.18406641e-01
3.59904170e-01 -7.97124356e-02 -1.29536083e-02 -6.97622716e-01
5.86805701e-01 2.41148636e-01 4.55126315e-01 -1.57233670e-01
9.13709998e-01 8.04082453e-01 3.67545962e-01 -5.71724355e-01
-1.08357275e+00 -8.07037592e-01 -1.33736658e+00 -4.53921735e-01
1.03165412e+00 -1.27375555e+00 -9.29022789e-01 4.07774478e-01
-1.40498435e+00 1.26148030e-01 2.96880789e-02 6.41230762e-01
-7.81413794e-01 2.97909260e-01 -3.82200390e-01 -6.64722860e-01
-1.03664145e-01 -1.32585025e+00 1.83840084e+00 -7.46077225e-02
8.35935771e-02 -6.11817896e-01 -2.74520248e-01 7.90546179e-01
-2.39163458e-01 7.15346336e-01 3.22005779e-01 -2.92837352e-01
-1.05936015e+00 -3.45793635e-01 -1.86314508e-02 1.66187540e-01
-2.18728244e-01 -1.26947761e-01 -1.07509708e+00 -8.26571807e-02
1.99006841e-01 -2.47679129e-01 5.37006497e-01 4.16582167e-01
8.85188937e-01 1.20123550e-01 -3.65372658e-01 7.75449634e-01
1.69346845e+00 -1.54854983e-01 7.51145184e-02 3.01517218e-01
1.25742161e+00 6.46729589e-01 8.41129899e-01 2.28416100e-01
5.47780931e-01 1.00703442e+00 1.21210587e+00 -2.42532156e-02
1.05163269e-01 -2.27866307e-01 1.24815419e-01 4.95551765e-01
-1.98162481e-01 -1.97445899e-01 -1.07661343e+00 3.48965347e-01
-1.78307390e+00 -4.82873470e-01 -4.65660661e-01 2.36586690e+00
1.45809367e-01 2.05529153e-01 2.10811183e-01 4.60019112e-02
3.39092225e-01 1.47987053e-01 -9.71988678e-01 1.48825586e-01
8.09218585e-02 -1.58600643e-01 8.44193995e-01 4.77874130e-01
-1.08294928e+00 8.35357249e-01 5.49642897e+00 1.26098216e-01
-8.80397499e-01 8.89177024e-02 2.26691619e-01 -4.48996156e-01
-1.09814204e-01 1.01404481e-01 -8.85252774e-01 -4.84566912e-02
3.85111511e-01 5.66069722e-01 2.21327409e-01 9.86030340e-01
-6.30364195e-02 -5.17603040e-01 -1.35067248e+00 1.37152088e+00
3.93820375e-01 -1.04920888e+00 -2.39688516e-01 2.22524911e-01
6.19040489e-01 5.13105989e-01 -9.57218632e-02 -3.24354857e-01
2.40818083e-01 -5.58948874e-01 8.00973773e-01 3.23109359e-01
3.11171800e-01 -6.99415684e-01 5.33492982e-01 7.37362564e-01
-9.79878664e-01 -3.75884259e-03 -1.82168767e-01 1.43974304e-01
4.34752822e-01 5.84291160e-01 -8.55476856e-01 6.58478022e-01
8.68264258e-01 8.50403786e-01 -7.71244466e-01 9.26195621e-01
-1.89673707e-01 -1.30798072e-01 -7.44683683e-01 5.18586993e-01
8.25008005e-02 -1.80177763e-01 6.59824550e-01 4.95201409e-01
4.13611144e-01 1.32490739e-01 3.25478464e-01 5.99178851e-01
2.95692831e-01 -4.43372518e-01 -6.67340457e-01 4.77961093e-01
1.81297407e-01 1.38623953e+00 -1.00345159e+00 -9.34413224e-02
-5.37921250e-01 1.24344838e+00 1.57112896e-01 1.64142787e-01
-6.42110825e-01 3.44721228e-01 5.04386723e-01 2.40850717e-01
8.47478688e-01 -6.47126496e-01 -1.01416714e-01 -1.46912193e+00
3.44855130e-01 -5.41498840e-01 6.26366884e-02 -1.29161537e+00
-1.07556427e+00 5.03158569e-01 3.33295345e-01 -1.53092396e+00
-2.33472615e-01 -8.64315510e-01 9.73572582e-03 6.85821474e-01
-1.42634618e+00 -1.41404879e+00 -4.08408672e-01 5.38814604e-01
6.43900216e-01 3.71271819e-01 6.21773243e-01 -9.60389301e-02
-5.13191484e-02 -7.11280778e-02 -2.38680348e-01 -1.31674871e-01
5.27227819e-01 -1.33257651e+00 4.33066726e-01 7.58946300e-01
4.10508931e-01 2.53723890e-01 7.43661463e-01 -4.55861509e-01
-1.93731201e+00 -8.05182874e-01 6.22906923e-01 -1.09291363e+00
2.70009667e-01 -8.06786835e-01 -5.07842660e-01 8.66678178e-01
-1.26032218e-01 5.35041153e-01 3.86203289e-01 -2.03513335e-02
-3.33146751e-01 -2.32601896e-01 -1.02814674e+00 2.65635431e-01
1.23082244e+00 -2.91840285e-01 -4.31691676e-01 5.48824549e-01
7.39313602e-01 -1.19083035e+00 -1.03881073e+00 5.15816748e-01
5.44069827e-01 -1.31403959e+00 1.41221213e+00 -1.00036636e-01
2.26768255e-01 -6.01509154e-01 -4.28493291e-01 -1.06702602e+00
1.26112968e-01 -2.04377741e-01 -3.11092049e-01 8.01144302e-01
-5.01076430e-02 -2.54293472e-01 1.10628271e+00 6.36954308e-01
-7.29797781e-02 -6.89698398e-01 -1.00853837e+00 -6.25193655e-01
-3.66752923e-01 -6.29796028e-01 3.09984714e-01 5.41119218e-01
-6.97860837e-01 5.35378218e-01 -3.08246791e-01 6.51916385e-01
9.73452866e-01 5.74938715e-01 1.22888172e+00 -1.42571938e+00
-4.38294202e-01 3.59772071e-02 -7.57157922e-01 -1.24692523e+00
-1.22140430e-01 -5.18685520e-01 -7.81637989e-03 -1.48697770e+00
4.06155996e-02 -1.14219375e-01 2.74372667e-01 2.00176641e-01
7.17610121e-02 4.34501231e-01 3.89085859e-01 2.60397881e-01
-7.64806032e-01 4.64481354e-01 1.34637678e+00 2.88671494e-01
-8.41758624e-02 1.73230767e-01 -2.23966077e-01 1.17345238e+00
3.16343367e-01 -4.11479264e-01 -4.40335125e-01 -7.14300990e-01
3.71794134e-01 3.50137055e-01 8.47997963e-01 -9.74260688e-01
2.10789248e-01 -9.64640006e-02 8.83662224e-01 -1.24744284e+00
1.05407941e+00 -1.40270114e+00 3.01328748e-01 1.73409015e-01
1.42254561e-01 3.31047714e-01 3.01102176e-02 9.17590380e-01
8.00116360e-02 2.33329326e-01 5.39102316e-01 -5.42012930e-01
-7.10028291e-01 5.83935618e-01 1.58709645e-01 -2.70835936e-01
1.30442071e+00 -4.80293781e-01 1.48849860e-01 -1.92535430e-01
-8.90199423e-01 2.97513008e-01 9.59082067e-01 5.11182368e-01
6.34378374e-01 -1.04492950e+00 -4.09179717e-01 2.84250706e-01
2.96038359e-01 9.37876165e-01 2.02126265e-01 8.57443929e-01
-6.62449598e-01 3.59565765e-01 2.07335688e-02 -1.38557553e+00
-1.38289082e+00 6.38692200e-01 4.98053730e-01 4.32101637e-02
-7.61474013e-01 7.12462306e-01 3.10594082e-01 -6.58440113e-01
1.50432885e-01 -3.92166823e-01 -6.41049817e-03 -5.78464530e-02
2.03227445e-01 2.10192859e-01 3.52603346e-01 -1.03067470e+00
-5.09072304e-01 1.11060750e+00 -8.76805931e-02 -2.05974400e-01
1.66413224e+00 -2.93749064e-01 4.51242551e-02 6.19893730e-01
1.17295206e+00 -1.17034771e-01 -1.91109562e+00 -1.99941128e-01
-3.54514480e-01 -8.84864807e-01 1.53117061e-01 -6.20180786e-01
-1.23791230e+00 8.92596066e-01 5.84416926e-01 -3.40222567e-01
9.61454272e-01 2.66362429e-01 5.82481980e-01 3.67753237e-01
6.38391197e-01 -8.90753925e-01 1.80723876e-01 2.99960494e-01
8.66277337e-01 -1.44250071e+00 4.29753333e-01 -7.48517334e-01
-5.73922873e-01 1.13893569e+00 7.38227189e-01 -3.07296962e-01
4.59312618e-01 5.16410954e-02 4.52131182e-02 -3.86262566e-01
-5.95092952e-01 -1.23170093e-01 3.91688526e-01 6.27704620e-01
-1.16149299e-01 -2.97049135e-01 3.69509280e-01 -9.01595131e-02
-1.87181763e-03 -3.22632998e-01 3.46137375e-01 1.07511628e+00
-1.10247269e-01 -1.14032602e+00 -6.64612293e-01 9.21279192e-02
-2.53256738e-01 3.51395935e-01 -5.41424930e-01 9.49011266e-01
3.13734680e-01 7.08137691e-01 1.13806412e-01 -3.03724468e-01
6.61994457e-01 -3.10590770e-03 9.95550573e-01 -6.95692956e-01
-4.24118131e-01 4.49486494e-01 -1.62267298e-01 -1.07873154e+00
-9.38721776e-01 -1.03158748e+00 -9.07218754e-01 -3.17280777e-02
-3.89999330e-01 -5.14517605e-01 8.47991228e-01 9.73917067e-01
3.86597902e-01 1.53259501e-01 4.58478242e-01 -1.57499683e+00
-3.05991143e-01 -5.15846550e-01 -4.42409068e-01 2.46159166e-01
5.76090753e-01 -8.23100567e-01 -2.36733198e-01 -4.03583758e-02] | [7.581632137298584, -2.6093719005584717] |
187d3a6d-33c5-412f-ba53-d12f07796bf8 | few-shot-table-to-text-generation-with-prefix | 2208.10709 | null | https://arxiv.org/abs/2208.10709v1 | https://arxiv.org/pdf/2208.10709v1.pdf | Few-Shot Table-to-Text Generation with Prefix-Controlled Generator | Neural table-to-text generation approaches are data-hungry, limiting their adaptation for low-resource real-world applications. Previous works mostly resort to Pre-trained Language Models (PLMs) to generate fluent summaries of a table. However, they often contain hallucinated contents due to the uncontrolled nature of PLMs. Moreover, the topological differences between tables and sequences are rarely studied. Last but not least, fine-tuning on PLMs with a handful of instances may lead to over-fitting and catastrophic forgetting. To alleviate these problems, we propose a prompt-based approach, Prefix-Controlled Generator (i.e., PCG), for few-shot table-to-text generation. We prepend a task-specific prefix for a PLM to make the table structure better fit the pre-trained input. In addition, we generate an input-specific prefix to control the factual contents and word order of the generated text. Both automatic and human evaluations on different domains (humans, books and songs) of the Wikibio dataset show substantial improvements over baseline approaches. | ['Shilin Wang', 'Gongshen Liu', 'Menghua Lu', 'Yutao Luo'] | 2022-08-23 | null | https://aclanthology.org/2022.coling-1.565 | https://aclanthology.org/2022.coling-1.565.pdf | coling-2022-10 | ['table-to-text-generation'] | ['natural-language-processing'] | [ 3.54843348e-01 9.71968099e-02 4.46127914e-02 -6.82830438e-02
-9.30867910e-01 -6.98177159e-01 8.78206670e-01 3.06279510e-01
-4.70895953e-02 9.14052486e-01 6.81613505e-01 -8.82311985e-02
4.39418614e-01 -9.52648520e-01 -8.12682688e-01 -3.59707564e-01
4.35649604e-01 6.27844632e-01 1.20704174e-01 -3.99714410e-01
5.25143504e-01 3.90820876e-02 -1.59607804e+00 6.39620900e-01
1.36358702e+00 3.94479573e-01 6.45738959e-01 5.30871093e-01
-5.39211094e-01 7.81486213e-01 -9.54842389e-01 -5.49886525e-01
1.26099205e-02 -8.72750759e-01 -4.52358514e-01 2.33125746e-01
2.96294272e-01 -2.43427604e-01 -3.58262002e-01 8.72001827e-01
6.97779536e-01 2.59010017e-01 8.00881147e-01 -1.03994036e+00
-7.97631145e-01 1.14091039e+00 -2.89981127e-01 2.67169680e-02
4.84777361e-01 4.19092745e-01 9.23050880e-01 -1.07175636e+00
8.59386802e-01 1.22975469e+00 3.98111701e-01 5.73423743e-01
-1.29529321e+00 -4.44444120e-01 1.06307320e-01 2.43982952e-03
-1.30805743e+00 -7.36784101e-01 6.72830701e-01 -2.73947775e-01
9.48716581e-01 1.86005965e-01 5.77996552e-01 1.38715589e+00
3.03072244e-01 8.66645396e-01 6.54453814e-01 -4.64638948e-01
3.31353277e-01 3.39879960e-01 -5.09259641e-01 4.17297542e-01
4.12870288e-01 -3.11027348e-01 -9.15069878e-01 -1.61810406e-02
7.14951038e-01 -3.18333924e-01 -2.04908714e-01 -7.54075125e-02
-1.48890567e+00 7.15146780e-01 -2.27483120e-02 1.89872399e-01
-6.89271390e-01 -3.39591593e-01 5.51488936e-01 1.23464353e-02
3.10328037e-01 9.15789008e-01 -1.77376434e-01 -3.14490438e-01
-1.18966365e+00 7.41748333e-01 1.01267433e+00 1.42724872e+00
4.66614544e-01 2.23724231e-01 -1.05088103e+00 8.62611473e-01
-2.18295425e-01 3.24340671e-01 7.68458903e-01 -4.64140534e-01
9.54333186e-01 4.87454832e-01 3.51076722e-01 -1.01672542e+00
-7.13574365e-02 -5.44096589e-01 -9.73070562e-01 -3.11860502e-01
3.50820303e-01 -7.29036480e-02 -8.05375576e-01 1.70786488e+00
4.50948291e-02 -2.62162298e-01 1.67265996e-01 5.75347841e-01
7.39021838e-01 1.03118193e+00 1.92407134e-03 -3.00244689e-01
1.30451071e+00 -1.26338255e+00 -1.00610518e+00 -6.15025163e-01
4.69390422e-01 -8.12289655e-01 1.83621025e+00 4.52331603e-01
-1.25653541e+00 -5.67390323e-01 -1.03857195e+00 -1.17877655e-01
-4.17669147e-01 4.08344328e-01 1.74975008e-01 1.84092477e-01
-6.83469355e-01 6.48472905e-01 -4.75334913e-01 -3.22691232e-01
2.43792728e-01 -3.01781029e-01 3.59018594e-02 -9.48995277e-02
-1.30176139e+00 8.05275857e-01 6.02061331e-01 -1.75636873e-01
-7.90988088e-01 -7.72085905e-01 -8.18593562e-01 2.47651264e-01
6.50478661e-01 -7.73521185e-01 1.33666563e+00 -4.95493710e-01
-1.47919512e+00 4.42871153e-01 -2.07423642e-01 -4.17926162e-01
6.04999840e-01 -1.61930695e-01 -2.49327838e-01 -1.08558230e-01
2.76965767e-01 7.70645618e-01 9.13172364e-01 -1.13234365e+00
-3.85268211e-01 1.02412686e-01 -1.76880881e-01 5.51997781e-01
-4.54822421e-01 -2.75983274e-01 -5.58702767e-01 -1.19440079e+00
-1.65452644e-01 -5.96327305e-01 -2.19858527e-01 -4.58945155e-01
-9.22447801e-01 -1.07573971e-01 2.84738302e-01 -8.56592655e-01
1.84641302e+00 -1.99883008e+00 7.46372202e-03 -3.31548870e-01
-2.27103129e-01 1.90662026e-01 -4.88624960e-01 9.54385579e-01
3.29514295e-01 2.31934667e-01 -2.83790857e-01 -4.66842622e-01
3.67265865e-02 -8.10266137e-02 -5.43392837e-01 -2.47266695e-01
3.78186226e-01 8.52957606e-01 -1.12276030e+00 -6.26583755e-01
-3.93095724e-02 2.68787712e-01 -7.14598119e-01 4.30761456e-01
-7.60562003e-01 3.53702396e-01 -1.99724555e-01 3.56526673e-01
3.44582140e-01 -2.03133479e-01 -2.93096964e-04 -1.74321219e-01
-9.97409746e-02 9.00054395e-01 -1.09988356e+00 1.74237955e+00
-7.11141109e-01 3.31598938e-01 -6.28067613e-01 -2.85483658e-01
9.68958616e-01 3.48680019e-01 -1.80710062e-01 -6.96158826e-01
-4.53842543e-02 1.53184295e-01 -5.11936657e-02 -4.83605683e-01
1.02078092e+00 1.28712598e-02 -1.86285973e-01 5.48513412e-01
-4.07907106e-02 -4.55272973e-01 7.51756847e-01 5.23091614e-01
1.04696882e+00 2.22908601e-01 5.53283572e-01 -2.79311370e-02
4.23949420e-01 7.10473284e-02 3.38750392e-01 9.88561749e-01
2.25935161e-01 1.17198372e+00 6.04742706e-01 -1.14499696e-01
-1.29635417e+00 -8.62948596e-01 3.98101598e-01 1.00255632e+00
-1.56299084e-01 -9.62288201e-01 -1.03065062e+00 -5.13450265e-01
-3.32752675e-01 1.45233774e+00 -3.98224682e-01 -3.88742000e-01
-5.80064476e-01 -6.56695843e-01 4.89847273e-01 4.44754064e-01
1.73245043e-01 -1.62463486e+00 -5.97432435e-01 7.37115204e-01
-6.79874718e-01 -1.17973411e+00 -9.84584570e-01 5.35998791e-02
-8.39427233e-01 -5.61834037e-01 -7.34532714e-01 -6.17977440e-01
6.94456935e-01 4.30194922e-02 1.43640745e+00 -7.48663992e-02
8.48992094e-02 -4.68391240e-01 -4.43790346e-01 -4.89480197e-01
-7.66541421e-01 5.70330322e-01 -8.75914767e-02 -6.93751276e-02
1.65823668e-01 -5.09247184e-01 -3.69190693e-01 5.85604608e-02
-1.07974362e+00 5.97149670e-01 7.27186680e-01 8.92309785e-01
7.30422318e-01 2.17201188e-02 7.72217989e-01 -1.04400873e+00
1.22797823e+00 -3.54476184e-01 -3.24649006e-01 4.03618842e-01
-5.88706613e-01 3.39687288e-01 1.23308909e+00 -8.27639341e-01
-1.09806776e+00 -8.42906535e-02 1.01323150e-01 -2.06381813e-01
-7.42294546e-03 4.70141888e-01 -4.28004265e-01 8.74935091e-01
9.35344934e-01 6.89230978e-01 -2.82239556e-01 -3.73423874e-01
4.05108213e-01 5.89080274e-01 6.04476810e-01 -5.41046321e-01
8.61911476e-01 2.50330828e-02 -5.80952823e-01 -5.85504234e-01
-9.67125118e-01 2.12890178e-01 -4.59909588e-01 -6.71446184e-03
4.83237058e-01 -8.74027669e-01 1.14079133e-01 2.92440593e-01
-1.38664091e+00 -5.39835274e-01 -3.56717080e-01 8.17724615e-02
-6.53884530e-01 1.45389855e-01 -5.13345838e-01 -7.93350041e-01
-6.78294003e-01 -9.46133494e-01 9.00885582e-01 2.64194965e-01
-6.50294840e-01 -5.75809598e-01 2.79002171e-02 -7.02467784e-02
5.90451360e-01 1.68427587e-01 1.13430893e+00 -7.69826055e-01
-5.88266611e-01 -2.38044769e-01 -4.02409919e-02 7.17608780e-02
2.69684017e-01 1.39599750e-02 -6.93116546e-01 -1.73540600e-02
-2.27819160e-01 -3.38285238e-01 6.21869147e-01 -8.76673311e-03
1.11461675e+00 -9.95530307e-01 -1.49591297e-01 2.32811078e-01
1.11350739e+00 2.14327291e-01 7.28036046e-01 2.05429420e-01
5.90802729e-01 6.87122524e-01 7.40738809e-01 8.58738124e-01
5.16588867e-01 6.14610672e-01 4.09751153e-03 3.52026582e-01
-3.28692824e-01 -1.08252740e+00 4.13780957e-01 1.01746750e+00
3.79663438e-01 -7.41888642e-01 -6.34552419e-01 6.65206313e-01
-1.48689187e+00 -8.83146405e-01 1.08175293e-01 2.26111007e+00
1.34292376e+00 5.57236016e-01 -2.53743455e-02 1.13942809e-01
7.53899574e-01 2.74646968e-01 -5.11790454e-01 -2.08304748e-01
-2.07684770e-01 -3.62347253e-02 3.28708999e-02 2.34214514e-01
-5.91816723e-01 1.27806795e+00 5.38779831e+00 1.09841216e+00
-8.86406660e-01 -7.40560368e-02 5.45207202e-01 -2.56568789e-01
-6.06016994e-01 -1.63882412e-02 -9.73328829e-01 7.08333492e-01
1.00812829e+00 -4.68469203e-01 6.70711279e-01 6.50789857e-01
4.44482684e-01 -5.49535677e-02 -1.11519563e+00 9.07054186e-01
2.40335077e-01 -1.29170084e+00 6.15229249e-01 -2.51845866e-01
6.03008926e-01 -5.17263114e-01 -1.63925424e-01 5.97519696e-01
-2.17622332e-02 -8.45699728e-01 1.16889155e+00 4.31779236e-01
9.22649920e-01 -8.01786005e-01 4.66131300e-01 6.06710792e-01
-9.47009325e-01 3.40633214e-01 -4.88976240e-01 6.03549369e-02
3.42398286e-01 6.50958359e-01 -1.01455188e+00 2.13579521e-01
2.83968359e-01 3.72473538e-01 -7.62639821e-01 9.01006401e-01
-5.03208160e-01 6.07159317e-01 -1.08585581e-01 -3.93371105e-01
1.05542652e-01 -6.92040399e-02 5.88243365e-01 1.27143145e+00
5.66865385e-01 -4.75787520e-02 -2.00893600e-02 1.21753418e+00
-2.50822783e-01 2.98903495e-01 -7.29336202e-01 -5.33148885e-01
8.57399225e-01 1.24525928e+00 -7.74798870e-01 -5.35365760e-01
-1.12377375e-01 1.18704891e+00 3.06717247e-01 2.62253821e-01
-6.60549998e-01 -6.94607615e-01 1.69884071e-01 3.60854894e-01
2.87468791e-01 -2.32087839e-02 -6.08159363e-01 -1.17744303e+00
1.34199977e-01 -1.31539643e+00 4.61242162e-02 -8.59694779e-01
-1.19045246e+00 1.00568628e+00 -6.19545914e-02 -1.35108340e+00
-6.35039449e-01 1.66040987e-01 -7.61069894e-01 8.59584093e-01
-1.14406180e+00 -7.27671325e-01 -2.22336277e-01 1.28816128e-01
1.09550786e+00 -2.15312332e-01 6.94618881e-01 1.69741035e-01
-5.89332223e-01 8.27295721e-01 -2.42900550e-01 -6.34250715e-02
9.31220472e-01 -1.23853838e+00 9.90192592e-01 1.05145109e+00
1.22969858e-01 7.00274110e-01 1.07949126e+00 -9.71059382e-01
-1.32557213e+00 -1.26093578e+00 1.40776289e+00 -5.09869874e-01
3.57304335e-01 -8.61423671e-01 -1.18897879e+00 3.41884822e-01
4.43825483e-01 -5.12125790e-01 3.51059914e-01 -2.20229194e-01
-2.24209413e-01 2.37740520e-02 -9.03362811e-01 1.18533659e+00
1.00386119e+00 -2.55878180e-01 -5.58109462e-01 3.77912432e-01
9.29228365e-01 -6.92692876e-01 -3.18101406e-01 -2.72745527e-02
2.94562876e-01 -8.39153290e-01 6.36166573e-01 -3.68073434e-01
8.98155630e-01 -4.08815384e-01 1.92777440e-01 -1.72063625e+00
-2.11610243e-01 -9.34034228e-01 -3.55739057e-01 1.54600000e+00
5.82175851e-01 -1.14235215e-01 4.51397836e-01 3.47728938e-01
-2.51015604e-01 -6.36164725e-01 -2.83380359e-01 -7.74595082e-01
-1.47376567e-01 -6.15711771e-02 8.23727727e-01 5.71405113e-01
4.80270572e-02 8.12530696e-01 -6.22566760e-01 -2.70798683e-01
3.36916268e-01 -1.26452208e-01 8.15649033e-01 -8.09101701e-01
-2.33137056e-01 -4.49370652e-01 4.92455989e-01 -9.02261674e-01
-1.34692013e-01 -7.40712404e-01 5.20608425e-01 -1.70596230e+00
9.76904109e-02 -2.18847007e-01 1.87612578e-01 3.61330330e-01
-4.81454223e-01 -1.03688419e-01 4.45663929e-01 2.02854350e-01
-3.90470445e-01 8.09318185e-01 1.44126809e+00 4.19855416e-02
-5.83076298e-01 -3.61971930e-02 -9.45368648e-01 2.31555298e-01
9.09739733e-01 -6.49820447e-01 -5.62859416e-01 -3.01069707e-01
3.96892488e-01 3.00513506e-01 -9.03480351e-02 -1.21347010e+00
2.30213925e-01 -2.99284488e-01 2.89774656e-01 -9.36059475e-01
8.74885842e-02 -1.27332345e-01 9.62454453e-02 2.99073339e-01
-7.68304467e-01 3.67617726e-01 1.88645542e-01 3.99403155e-01
-2.08211020e-01 -4.30123299e-01 6.26222730e-01 -4.76172626e-01
-8.07819441e-02 2.03831062e-01 -5.83984673e-01 4.63513672e-01
4.41178143e-01 1.53223220e-02 -2.34480008e-01 -6.74837232e-01
3.09484992e-02 1.04707323e-01 4.75197762e-01 6.17778182e-01
5.45786321e-01 -1.22264862e+00 -8.56748521e-01 2.33214960e-01
2.88691282e-01 2.68131495e-01 1.90635383e-01 4.13924962e-01
-3.23729187e-01 5.26747823e-01 -1.84265643e-01 3.22447568e-02
-6.44756794e-01 6.92619741e-01 -1.02887943e-01 -6.14075601e-01
-7.56003141e-01 5.68781972e-01 9.36435014e-02 -3.48279893e-01
3.77955794e-01 -3.45429182e-01 -5.65971285e-02 2.31949955e-01
8.79247367e-01 1.41231582e-01 1.93213761e-01 -4.29859348e-02
7.22654611e-02 -1.19667090e-01 -4.10770863e-01 -4.59497601e-01
1.03785622e+00 -1.12654485e-01 2.13529825e-01 5.80849886e-01
6.69942379e-01 1.58206344e-01 -1.24585426e+00 -2.31943101e-01
1.28539070e-01 -1.86759204e-01 -3.94153535e-01 -1.05163372e+00
-5.97447872e-01 8.95613611e-01 -2.42223933e-01 7.77009353e-02
9.47853148e-01 -3.15087408e-01 1.10498261e+00 5.23450494e-01
2.06933260e-01 -1.13791227e+00 3.92349184e-01 6.68001831e-01
1.23972762e+00 -8.40781629e-01 -1.46846071e-01 -2.36226514e-01
-1.11429036e+00 9.69571412e-01 9.26951289e-01 1.13139145e-01
-7.11956099e-02 2.35320881e-01 -7.76278749e-02 2.82391250e-01
-1.24923980e+00 1.18123256e-01 1.88759118e-01 6.31939709e-01
5.67274868e-01 -2.04487547e-01 -4.72645283e-01 8.38641047e-01
-7.45314896e-01 1.37877893e-02 9.64410126e-01 9.08198059e-01
-3.85027170e-01 -9.89208639e-01 -2.15732083e-01 5.98697484e-01
-2.46876583e-01 -5.38569093e-01 -4.09723610e-01 4.67175096e-01
-2.40210950e-01 8.44183207e-01 5.86540550e-02 -3.30396503e-01
5.13459742e-01 3.54165286e-01 2.99714088e-01 -9.92863894e-01
-6.98687196e-01 1.62435114e-01 6.66981861e-02 -1.70753583e-01
3.60370904e-01 -6.76502049e-01 -1.08162057e+00 -1.65888518e-01
-7.39827380e-02 2.16922965e-02 4.52154249e-01 6.79119587e-01
5.07243097e-01 6.29398882e-01 4.00474995e-01 -7.41663337e-01
-8.74100864e-01 -1.50550783e+00 -3.39889467e-01 5.19626319e-01
-1.55063989e-02 -3.69764715e-01 -1.52532041e-01 1.55512154e-01] | [11.729345321655273, 8.878050804138184] |
9a40fff1-cc65-4a88-96d7-14440fe055f2 | low-rank-extended-kalman-filtering-for-online | 2305.19535 | null | https://arxiv.org/abs/2305.19535v3 | https://arxiv.org/pdf/2305.19535v3.pdf | Low-rank extended Kalman filtering for online learning of neural networks from streaming data | We propose an efficient online approximate Bayesian inference algorithm for estimating the parameters of a nonlinear function from a potentially non-stationary data stream. The method is based on the extended Kalman filter (EKF), but uses a novel low-rank plus diagonal decomposition of the posterior precision matrix, which gives a cost per step which is linear in the number of model parameters. In contrast to methods based on stochastic variational inference, our method is fully deterministic, and does not require step-size tuning. We show experimentally that this results in much faster (more sample efficient) learning, which results in more rapid adaptation to changing distributions, and faster accumulation of reward when used as part of a contextual bandit algorithm. | ['Gerardo Durán-Martín', 'Peter G. Chang', 'Kevin Murphy', 'Matt Jones', 'Alexander Y Shestopaloff'] | 2023-05-31 | null | null | null | null | ['bayesian-inference'] | ['methodology'] | [-5.24403788e-02 -2.86354274e-01 -5.91660082e-01 -1.57449499e-01
-1.11529946e+00 -6.55051053e-01 7.12679744e-01 -6.57625571e-02
-7.86493897e-01 1.23859179e+00 2.00398937e-01 -4.96066600e-01
-4.98118073e-01 -4.43985224e-01 -9.31957066e-01 -6.45042181e-01
5.33352382e-02 8.16788375e-01 2.88883239e-01 4.13809210e-01
4.03784692e-01 3.17624569e-01 -1.37484074e+00 -1.07876971e-01
3.91430616e-01 1.08976579e+00 2.34200150e-01 9.97026801e-01
2.26622805e-01 7.21231341e-01 -2.89346814e-01 -2.59698212e-01
1.40347779e-01 -3.72554094e-01 -5.64831793e-01 -1.63886230e-02
2.51918346e-01 -8.64363015e-01 -1.73928067e-01 9.50727761e-01
2.18652502e-01 5.59952676e-01 7.76648581e-01 -7.92224526e-01
-4.52177376e-02 7.73732424e-01 -2.22220585e-01 4.03768510e-01
2.25181222e-01 6.78311214e-02 1.02525747e+00 -5.74020803e-01
2.54738450e-01 1.51400518e+00 7.55952537e-01 1.82038620e-01
-1.73957801e+00 -5.64634025e-01 3.78446698e-01 4.80062328e-02
-1.08069897e+00 -7.46483743e-01 3.37640971e-01 -2.36514717e-01
8.34302187e-01 1.72735527e-01 9.40000296e-01 1.16513443e+00
1.03877537e-01 1.03185260e+00 1.30393362e+00 -3.15451771e-01
8.36723626e-01 2.58428585e-02 6.36677742e-02 5.81262529e-01
4.21432972e-01 6.78795040e-01 -8.35421443e-01 -7.13316381e-01
1.05650926e+00 9.28392410e-02 -2.32930526e-01 -4.04981703e-01
-1.03619003e+00 1.05108345e+00 -2.36982733e-01 -2.84073293e-01
-6.66193843e-01 9.64117408e-01 1.70307606e-01 1.96810678e-01
5.24517059e-01 -3.94219682e-02 -6.71158910e-01 -8.24507594e-01
-1.19713354e+00 5.90796947e-01 1.19476879e+00 6.32613599e-01
6.72167718e-01 1.34337574e-01 -2.74707615e-01 5.37223458e-01
5.67829549e-01 1.00794244e+00 4.03899282e-01 -1.78919578e+00
1.71163872e-01 -3.82516742e-01 9.33516026e-01 -4.72958177e-01
-2.57402182e-01 -5.79252362e-01 -2.92179078e-01 2.01310709e-01
6.37228370e-01 -3.92893523e-01 -6.11869216e-01 1.70347655e+00
4.13452029e-01 6.83829367e-01 -2.52130598e-01 6.16363943e-01
-3.33623201e-01 6.73731148e-01 -2.44519114e-01 -6.72390699e-01
1.00546789e+00 -7.43615806e-01 -8.52650464e-01 -3.28802109e-01
1.37043923e-01 -5.10573149e-01 7.71579266e-01 7.68278182e-01
-1.34162414e+00 -3.69967930e-02 -8.43757153e-01 2.70065933e-01
3.37628499e-02 -2.14169964e-01 1.04232466e+00 1.02599096e+00
-9.02156174e-01 7.90179610e-01 -1.30961549e+00 9.84043628e-03
2.85095602e-01 3.78857970e-01 2.50302672e-01 4.67815064e-02
-6.97431862e-01 8.03945541e-01 3.79413366e-01 -7.07089826e-02
-1.17910051e+00 -8.50167334e-01 -4.99156445e-01 1.31348297e-01
5.91293335e-01 -7.73047447e-01 1.80080998e+00 -7.98547208e-01
-2.25645280e+00 6.44646643e-04 -4.65401173e-01 -8.75223994e-01
6.42117143e-01 -4.69911784e-01 8.61033574e-02 1.23895621e-02
-3.53317469e-01 2.48768091e-01 1.34912777e+00 -8.22251678e-01
-8.30886185e-01 -1.86319426e-01 -9.93397981e-02 1.83122844e-01
-6.57488257e-02 -4.01448794e-02 -1.01666860e-01 -5.56305468e-01
-4.56488505e-02 -1.19352686e+00 -2.71152347e-01 -1.13760211e-01
1.81879714e-01 -1.16759986e-01 4.01737303e-01 -6.11434937e-01
1.38066947e+00 -1.88878751e+00 -6.15373440e-02 3.33906502e-01
-2.97276497e-01 -1.17651917e-01 1.05990864e-01 4.21269417e-01
4.20619667e-01 -1.48391694e-01 -1.94043338e-01 -2.85247654e-01
2.34836906e-01 3.87884319e-01 -7.06195891e-01 5.72578490e-01
-4.46865886e-01 6.75108254e-01 -1.01335371e+00 -1.29892096e-01
7.34682530e-02 7.36427680e-02 -8.21702778e-01 -3.94574776e-02
-4.54838753e-01 1.50255874e-01 -5.47143161e-01 1.64007023e-01
4.16173220e-01 -3.31901371e-01 2.61754602e-01 1.33680835e-01
-3.81959192e-02 2.63764083e-01 -1.65895092e+00 1.66015720e+00
-6.01779103e-01 3.60143602e-01 2.99458861e-01 -9.17666078e-01
3.33569884e-01 2.38404706e-01 3.68319720e-01 -9.29174200e-02
-2.87353974e-02 2.23431811e-01 -4.75456089e-01 -1.97723657e-01
4.83029306e-01 -1.45737424e-01 1.23239614e-01 9.39427614e-01
8.62434413e-03 -1.16719998e-01 3.28306198e-01 1.14532657e-01
9.89965975e-01 5.43653548e-01 3.07294160e-01 -3.21023554e-01
-1.36149302e-01 -2.49022737e-01 4.52215701e-01 1.45379210e+00
-1.57256722e-01 -8.56400654e-02 3.23729545e-01 -3.49287957e-01
-8.12036514e-01 -1.18409181e+00 -2.25970313e-01 1.21596479e+00
-2.01090544e-01 -2.85786569e-01 -6.24743938e-01 -2.51919180e-01
5.11734128e-01 8.63354325e-01 -4.04591650e-01 5.00875153e-02
-3.60613763e-01 -9.52730775e-01 1.93220109e-01 4.69905674e-01
1.26572222e-01 -5.96159995e-01 -7.65438437e-01 6.70779765e-01
-5.83149791e-02 -7.92502105e-01 -5.09325445e-01 2.36534059e-01
-1.37269258e+00 -7.06769347e-01 -5.10621130e-01 1.05420493e-01
2.40881786e-01 -7.40438253e-02 1.02877951e+00 -5.62159896e-01
1.74519017e-01 8.76811564e-01 8.17436352e-02 -6.25778615e-01
-2.41868690e-01 -3.45542938e-01 3.14003915e-01 -2.92752720e-02
6.93941638e-02 -4.84829843e-01 -6.90171897e-01 5.95234260e-02
-6.09333336e-01 -2.87381917e-01 4.88501757e-01 1.21490431e+00
6.50284708e-01 5.91067113e-02 4.09163445e-01 -9.00052905e-01
9.85791385e-01 -3.81904393e-01 -1.35820413e+00 2.88859233e-02
-9.40346599e-01 4.54417765e-01 2.52912670e-01 -8.52921188e-01
-1.29843259e+00 2.75999457e-01 2.37053290e-01 -4.45563138e-01
2.23656461e-01 6.34849072e-01 7.63370097e-01 4.51023169e-02
6.84879601e-01 2.78457403e-01 1.22978590e-01 -5.32981396e-01
5.48680127e-01 4.69603956e-01 2.40374729e-01 -7.98016489e-01
5.49813092e-01 4.52890098e-01 1.98105022e-01 -4.61753041e-01
-8.75950277e-01 -1.14953034e-01 -7.65880272e-02 -4.04519141e-02
2.44700059e-01 -1.08131111e+00 -1.14024568e+00 2.27490127e-01
-7.24563181e-01 -9.13057029e-01 -4.41100270e-01 1.04167426e+00
-1.04453290e+00 1.78901255e-01 -6.39573872e-01 -1.53043664e+00
-8.43324214e-02 -6.69126272e-01 8.42121542e-01 1.40847638e-02
-2.34185830e-01 -1.00858331e+00 2.77029037e-01 1.22559972e-01
4.70957339e-01 -3.01371366e-01 4.70446497e-01 -2.69601703e-01
-8.83635581e-01 -2.33929873e-01 -1.09595181e-02 2.98129261e-01
-3.03362370e-01 -2.02687923e-03 -5.75968683e-01 -6.13714755e-01
8.77120271e-02 -3.32939684e-01 9.19070840e-01 9.77283001e-01
9.34628308e-01 -6.67229593e-01 -2.31179684e-01 3.29197615e-01
1.39095128e+00 2.98278611e-02 6.74319565e-02 1.44552454e-01
1.41113043e-01 1.13367781e-01 9.16581511e-01 1.05599999e+00
2.04148129e-01 4.12445337e-01 1.64261594e-01 6.30006850e-01
2.50513434e-01 -2.93580115e-01 7.26732612e-01 5.15101135e-01
-1.14454469e-02 9.84508544e-02 -5.49838483e-01 6.18825197e-01
-2.43401027e+00 -1.10269082e+00 1.51656881e-01 2.56860781e+00
1.10227180e+00 1.43021032e-01 4.60754395e-01 -2.91114300e-01
4.00420696e-01 -1.50703099e-02 -9.63397384e-01 -3.93933058e-01
3.17882031e-01 4.35401767e-01 1.10122705e+00 7.95183957e-01
-8.16328466e-01 7.40018070e-01 8.74051857e+00 1.06665671e+00
-4.92147893e-01 2.82405436e-01 3.70171905e-01 -5.00121534e-01
-5.06783053e-02 2.37617359e-01 -9.18063343e-01 7.52649963e-01
1.49489057e+00 -2.45045081e-01 1.21461773e+00 8.00237834e-01
4.85054940e-01 -6.18091941e-01 -1.18661153e+00 1.06693721e+00
-4.53027487e-01 -1.42667484e+00 -3.69805753e-01 7.32015073e-02
7.66961515e-01 2.88656771e-01 9.07810181e-02 2.98836350e-01
1.00524795e+00 -5.63743532e-01 8.81899357e-01 8.86102021e-01
3.75460207e-01 -1.03505588e+00 3.37320417e-01 7.74000287e-01
-6.52358711e-01 -5.48989117e-01 -4.87552553e-01 -4.03814435e-01
1.62840977e-01 7.94123650e-01 -5.75984120e-01 -1.79117039e-01
7.73997784e-01 3.60020906e-01 2.31460512e-01 1.15896773e+00
-1.63509235e-01 1.05292630e+00 -9.75874960e-01 -3.65084738e-01
1.08206227e-01 -2.55651444e-01 6.10994935e-01 1.05753303e+00
3.93373370e-01 6.65725209e-03 2.62720704e-01 6.62873507e-01
8.25078264e-02 -1.84162542e-01 -2.32946947e-01 -9.61586833e-02
7.20300317e-01 8.50054502e-01 -3.38627368e-01 -5.63535452e-01
-2.04256997e-01 6.94599152e-01 2.62865931e-01 6.32386327e-01
-6.61087394e-01 7.11797997e-02 4.87539291e-01 -2.94890910e-01
7.67583847e-01 -4.22174305e-01 5.14717251e-02 -1.14227998e+00
-2.47854248e-01 -8.83204401e-01 4.47847724e-01 -5.55640996e-01
-1.21814656e+00 -2.59439170e-01 3.80856484e-01 -7.09742010e-01
-8.78936291e-01 -4.43805009e-01 -7.29179382e-02 7.58047521e-01
-1.33695567e+00 -5.00809252e-01 4.68898684e-01 6.20777369e-01
4.05426294e-01 4.14792448e-02 8.26230168e-01 -7.48194605e-02
-3.76293242e-01 3.08585733e-01 8.92944634e-01 -5.12424171e-01
4.67040539e-01 -1.35022128e+00 3.21982279e-02 7.18541801e-01
2.22117275e-01 7.66231775e-01 1.04298842e+00 -6.75276637e-01
-1.60223615e+00 -6.95714235e-01 3.08740735e-01 -4.72048432e-01
1.02428436e+00 -1.79093137e-01 -3.31735313e-01 9.56676602e-01
6.07484989e-02 -1.34639069e-01 6.66363776e-01 4.54574466e-01
-1.84358805e-01 -3.39317441e-01 -1.08780146e+00 5.32763958e-01
7.09423006e-01 -5.81350148e-01 -3.63090783e-01 4.59888756e-01
4.53212678e-01 -4.60725933e-01 -9.33081031e-01 1.38156444e-01
9.61790264e-01 -6.35883272e-01 8.92712116e-01 -5.35300255e-01
-3.27970564e-01 -1.24362364e-01 -1.67062446e-01 -1.20538270e+00
-5.77060103e-01 -1.35477698e+00 -9.17916179e-01 6.69157386e-01
3.75033319e-01 -8.53430569e-01 8.84912729e-01 7.63803780e-01
3.92060846e-01 -6.16980612e-01 -1.04939067e+00 -1.03216839e+00
-1.51701421e-01 -6.51367486e-01 3.16108406e-01 4.52395707e-01
-1.60750493e-01 1.57132328e-01 -7.05421269e-01 5.35066985e-02
1.03280497e+00 3.02397102e-01 5.89002311e-01 -1.21568239e+00
-1.06491804e+00 -3.29434872e-01 2.99445838e-01 -1.66473675e+00
7.81672299e-02 -3.63134444e-01 9.90178734e-02 -1.01152658e+00
3.52993637e-01 -2.47181565e-01 -5.62851131e-01 3.80819499e-01
-1.00349030e-02 2.43646242e-02 -1.30570471e-01 1.29574358e-01
-6.97449386e-01 5.45539796e-01 8.67399633e-01 1.41242102e-01
-2.14151725e-01 4.40203190e-01 -4.22357827e-01 7.33952761e-01
5.69960356e-01 -8.31305981e-01 -5.23213029e-01 -9.61237699e-02
4.68681335e-01 5.48474014e-01 3.24565083e-01 -5.95566034e-01
3.28260273e-01 -5.64830065e-01 2.76326776e-01 -7.32277751e-01
5.03636777e-01 -6.84137642e-01 2.21839190e-01 6.36525750e-01
-4.03861284e-01 -1.03510983e-01 1.54445320e-01 1.14328635e+00
2.43176609e-01 -5.79931796e-01 6.55173242e-01 -1.54267654e-01
-2.39048898e-01 2.70981610e-01 -7.78659165e-01 -2.15395197e-01
5.99711776e-01 3.92345972e-02 6.38961121e-02 -7.40308702e-01
-1.00957036e+00 -4.74857958e-03 1.63067907e-01 -1.06723458e-01
3.32550079e-01 -1.28581464e+00 -6.08222067e-01 -1.60458207e-01
-5.35026252e-01 -3.82369876e-01 5.47030978e-02 8.48981380e-01
-1.70099765e-01 5.19219160e-01 3.36946219e-01 -5.36613822e-01
-8.97219479e-01 4.85175073e-01 1.73332065e-01 -4.25899327e-01
-3.07864398e-01 8.06086719e-01 -3.47473413e-01 4.35068086e-03
4.09313291e-01 -3.59597117e-01 3.33064735e-01 -2.34087352e-02
6.40544534e-01 7.53287137e-01 -3.50356907e-01 1.55351862e-01
-8.84481147e-02 1.29303202e-01 2.71341689e-02 -8.10000658e-01
1.19398201e+00 -4.11665261e-01 -1.12218987e-02 8.39431465e-01
7.75010228e-01 -1.01301201e-01 -1.92245936e+00 -5.20722210e-01
-1.69513784e-02 -7.22843766e-01 6.41918659e-01 -8.19692492e-01
-5.60785830e-01 3.39601815e-01 7.41569579e-01 1.15021169e-01
7.46091723e-01 -4.19848770e-01 6.56406581e-01 9.14302528e-01
5.84894836e-01 -1.37090659e+00 -1.02923259e-01 4.32016581e-01
5.58615625e-01 -1.00736547e+00 4.21299726e-01 1.24877229e-01
-2.49954388e-01 9.90340412e-01 -2.27512926e-01 -4.02503669e-01
9.59189236e-01 3.45814914e-01 -4.13318813e-01 3.57814848e-01
-1.18990970e+00 6.80170674e-03 3.26800942e-02 5.08222282e-01
-1.72425285e-01 1.63094789e-01 -2.68320024e-01 2.99963266e-01
6.73177615e-02 2.91941017e-01 4.98247296e-01 9.43667769e-01
-7.33989000e-01 -1.14589679e+00 -3.63069832e-01 6.86467469e-01
-6.48626149e-01 -8.12940523e-02 3.73810947e-01 3.00687492e-01
-4.45151806e-01 1.02235937e+00 -2.04706471e-02 3.25216353e-01
-1.14506729e-01 1.81376874e-01 9.13519979e-01 -2.13402718e-01
-1.29295677e-01 4.14719969e-01 1.55791909e-01 -9.56182897e-01
-4.66277599e-01 -1.22601640e+00 -8.11384797e-01 -4.39182281e-01
-3.98105651e-01 1.84132114e-01 9.30862963e-01 1.00990045e+00
2.53387362e-01 9.25409123e-02 6.36309981e-01 -1.06992018e+00
-1.34281456e+00 -8.05007935e-01 -7.44340301e-01 -1.87626369e-02
5.55748045e-01 -7.45043159e-01 -3.54578108e-01 -1.92783862e-01] | [4.809502601623535, 3.1281702518463135] |
283bb83d-cde0-479c-9f25-d299f5b6a817 | unsupervised-multi-target-domain-adaptation | 1810.11547 | null | http://arxiv.org/abs/1810.11547v1 | http://arxiv.org/pdf/1810.11547v1.pdf | Unsupervised Multi-Target Domain Adaptation: An Information Theoretic Approach | Unsupervised domain adaptation (uDA) models focus on pairwise adaptation
settings where there is a single, labeled, source and a single target domain.
However, in many real-world settings one seeks to adapt to multiple, but
somewhat similar, target domains. Applying pairwise adaptation approaches to
this setting may be suboptimal, as they fail to leverage shared information
among multiple domains. In this work we propose an information theoretic
approach for domain adaptation in the novel context of multiple target domains
with unlabeled instances and one source domain with labeled instances. Our
model aims to find a shared latent space common to all domains, while
simultaneously accounting for the remaining private, domain-specific factors.
Disentanglement of shared and private information is accomplished using a
unified information-theoretic approach, which also serves to establish a
stronger link between the latent representations and the observed data. The
resulting model, accompanied by an efficient optimization algorithm, allows
simultaneous adaptation from a single source to multiple target domains. We
test our approach on three challenging publicly-available datasets, showing
that it outperforms several popular domain adaptation methods. | ['Vladimir Pavlovic', 'Pritish Sahu', 'Behnam Gholami', 'Ognjen Rudovic', 'Konstantinos Bousmalis'] | 2018-10-26 | unsupervised-multi-target-domain-adaptation-1 | https://openreview.net/forum?id=BJxLH2AcYX | https://openreview.net/pdf?id=BJxLH2AcYX | iclr-2019-5 | ['multi-target-domain-adaptation'] | ['computer-vision'] | [ 4.71677870e-01 -1.40903309e-01 -6.02391660e-01 -4.55573082e-01
-1.09045482e+00 -9.21912968e-01 7.12209046e-01 6.50886670e-02
-3.60632628e-01 1.12446201e+00 2.01515213e-01 2.47632071e-01
-2.19126225e-01 -5.28336287e-01 -6.19604111e-01 -8.65971506e-01
2.92823344e-01 9.45580423e-01 2.47370135e-02 7.07659498e-02
-3.64735648e-02 1.75309882e-01 -1.03546321e+00 1.21865179e-02
7.70552397e-01 4.36955184e-01 1.45192355e-01 4.05497819e-01
-1.89939421e-02 2.27205440e-01 -3.59075159e-01 -4.58063811e-01
4.87715662e-01 -5.60112476e-01 -7.75665641e-01 3.74118149e-01
4.80458081e-01 -9.40366611e-02 -9.51304734e-02 1.10243642e+00
2.86141723e-01 2.60118574e-01 1.09680462e+00 -1.28964281e+00
-8.24060380e-01 3.55214357e-01 -6.00603461e-01 -5.44365421e-02
2.79568672e-01 -1.10753179e-01 1.13365853e+00 -6.32631302e-01
8.85060847e-01 1.06206977e+00 4.43334728e-01 6.36826754e-01
-1.91545773e+00 -6.54019952e-01 3.52616429e-01 -2.70515848e-02
-1.25693834e+00 -5.61180234e-01 7.54066288e-01 -5.76132774e-01
5.28923333e-01 -2.26528153e-01 2.14972589e-02 1.51197696e+00
-1.41425908e-01 5.48178434e-01 1.09297383e+00 -5.83840966e-01
4.71757263e-01 5.97855806e-01 1.17933139e-01 9.96075124e-02
4.02136177e-01 2.60920860e-02 -6.64182067e-01 -6.35326624e-01
5.11352897e-01 1.61494613e-01 1.33057516e-02 -1.24466240e+00
-1.22811115e+00 1.02226543e+00 -1.98768154e-02 2.84210891e-01
-3.72726232e-01 -5.30828416e-01 2.66478658e-01 5.51491380e-01
5.18720508e-01 4.01112586e-01 -7.95949817e-01 1.72164798e-01
-6.55738711e-01 2.34048367e-01 9.39404964e-01 1.19933879e+00
1.24955666e+00 -2.83115596e-01 1.62769794e-01 1.08744502e+00
3.03236187e-01 6.37113631e-01 5.62865555e-01 -9.70085382e-01
6.20394826e-01 5.18610656e-01 4.43188667e-01 -5.85164309e-01
-1.07806131e-01 -1.46468848e-01 -7.16171622e-01 1.18226424e-01
7.40624666e-01 -2.23556980e-01 -7.17591166e-01 2.53858614e+00
4.42317128e-01 1.87873587e-01 5.19171894e-01 6.62701130e-01
3.42580825e-02 3.13511014e-01 2.01051310e-01 -2.72150129e-01
1.07485032e+00 -6.71417832e-01 -4.27600712e-01 -5.37034094e-01
5.68878889e-01 -7.61552691e-01 7.96328247e-01 1.38578296e-01
-9.08086956e-01 -3.50821972e-01 -9.79904830e-01 -3.06698531e-02
-3.88037980e-01 1.27218038e-01 2.06299827e-01 5.86185575e-01
-8.10090244e-01 3.00011575e-01 -5.50379455e-01 -9.34051156e-01
2.90699929e-01 5.52435100e-01 -8.08465004e-01 -3.29916954e-01
-1.09188080e+00 8.30658555e-01 4.11628127e-01 -7.29941189e-01
-7.62737751e-01 -5.89080513e-01 -7.98314691e-01 -5.56754693e-02
5.31621516e-01 -9.44391251e-01 1.24455583e+00 -1.26372576e+00
-1.48892057e+00 9.54973340e-01 -4.18710619e-01 -3.84304553e-01
2.87861377e-01 -2.12552786e-01 -4.11646962e-01 -6.36983737e-02
4.23989654e-01 3.45705390e-01 9.86045897e-01 -1.32338202e+00
-5.25758028e-01 -7.38424361e-01 -2.31083184e-01 4.99341875e-01
-5.29279530e-01 -1.76911265e-01 -2.40297467e-01 -6.81802511e-01
-2.88703735e-03 -1.10354674e+00 -2.35985011e-01 -1.50420457e-01
-1.42783254e-01 7.38476068e-02 7.04228938e-01 -3.33931357e-01
7.84526110e-01 -2.29377818e+00 6.33296549e-01 3.34480524e-01
4.48815785e-02 3.10604889e-02 -3.36078376e-01 5.38431585e-01
-1.93559334e-01 -2.30775625e-01 -6.57860398e-01 -6.49816155e-01
9.10632685e-02 3.98439407e-01 -3.52437735e-01 5.35545111e-01
1.57519311e-01 5.19373238e-01 -9.09708798e-01 -3.60432684e-01
3.50123062e-03 1.65044442e-01 -6.01058543e-01 3.74763846e-01
-1.60393909e-01 6.72564209e-01 -5.99153936e-01 3.47299635e-01
7.78651714e-01 -4.27366763e-01 7.50408113e-01 2.51322329e-01
2.89672017e-01 2.40142550e-02 -1.46245706e+00 2.06813884e+00
-3.07607949e-01 3.47965449e-01 1.33483276e-01 -1.14502954e+00
8.99852276e-01 4.03042346e-01 7.00366378e-01 -3.97896558e-01
-2.17619315e-01 2.45964617e-01 -3.09538573e-01 -1.82548136e-01
2.28626907e-01 -5.23862004e-01 -3.75244081e-01 8.48994076e-01
6.23164475e-01 3.16017687e-01 -1.20739020e-01 3.37648064e-01
1.05111527e+00 1.33384526e-01 8.83973658e-01 -1.27122283e-01
3.73428345e-01 -5.37762325e-03 6.79801106e-01 9.03370619e-01
-3.81825894e-01 6.26900613e-01 2.69497901e-01 -1.43794030e-01
-1.28399003e+00 -1.41862321e+00 -9.94980782e-02 1.18760669e+00
7.69358873e-02 -4.29914780e-02 -5.46204686e-01 -8.67326498e-01
1.86379462e-01 5.23978412e-01 -6.47267401e-01 -3.07197690e-01
-2.25209400e-01 -4.86424625e-01 3.49185854e-01 2.79704630e-01
1.70672819e-01 -6.16205275e-01 7.58478940e-02 2.99101025e-01
-2.76353657e-01 -1.14157271e+00 -6.44328237e-01 3.65757436e-01
-9.48040962e-01 -9.15466249e-01 -8.65836799e-01 -6.80138648e-01
6.20232880e-01 2.94187367e-01 1.02767074e+00 -5.50024211e-01
2.84493953e-01 6.79815471e-01 -1.69774443e-01 -1.02213204e-01
-5.05601287e-01 3.54904026e-01 5.02940893e-01 3.67203355e-01
8.56867731e-01 -7.26369917e-01 -8.60485286e-02 5.21291256e-01
-1.09038210e+00 -3.12966406e-01 5.77402592e-01 1.07952154e+00
4.37357724e-01 -3.11802685e-01 7.13200629e-01 -1.24113929e+00
4.05522436e-01 -1.03439701e+00 -4.74831194e-01 4.96195942e-01
-5.91983974e-01 2.16355756e-01 6.52827859e-01 -6.85912251e-01
-1.36997652e+00 3.74285549e-01 4.61114198e-01 -4.19765025e-01
-6.49414361e-01 3.10496658e-01 -6.32453620e-01 1.19017698e-01
9.06383693e-01 1.92637160e-01 1.26753151e-01 -7.00924098e-01
4.34976280e-01 6.45844221e-01 4.81782585e-01 -8.28729212e-01
1.02238059e+00 5.21472335e-01 -2.04412833e-01 -6.10948026e-01
-7.68462479e-01 -8.23947310e-01 -1.25589406e+00 4.74642038e-01
5.90710223e-01 -1.15269160e+00 -8.46987441e-02 2.83091873e-01
-8.95067811e-01 -2.29006588e-01 -5.19247949e-01 6.10732257e-01
-8.12394619e-01 5.54724395e-01 -2.38340329e-02 -5.38161755e-01
4.49917346e-01 -9.86773014e-01 8.72052491e-01 6.70337602e-02
-4.04912323e-01 -1.46945035e+00 6.41845405e-01 2.99864024e-01
1.49559528e-01 -1.22379631e-01 8.56091976e-01 -1.40171659e+00
-3.30618501e-01 -6.65703043e-02 -1.22079089e-01 3.72525454e-01
5.69653213e-01 -5.13563573e-01 -9.50319052e-01 -5.01946747e-01
9.99748856e-02 -3.43845278e-01 6.05370164e-01 2.92411119e-01
5.05859494e-01 -3.29086989e-01 -4.47800338e-01 4.59302723e-01
1.42650652e+00 8.81580934e-02 1.74403846e-01 2.35062510e-01
5.61309278e-01 6.96834087e-01 5.03704786e-01 5.46780825e-01
3.90960693e-01 8.34319055e-01 -1.00260891e-01 -3.79689671e-02
1.86193079e-01 -3.77691925e-01 4.74090993e-01 5.78881502e-01
3.21633935e-01 -2.68174082e-01 -8.45140040e-01 8.39925110e-01
-1.93135655e+00 -9.54098403e-01 3.43330622e-01 2.62802148e+00
8.62339497e-01 -2.80088365e-01 3.65030706e-01 -4.27625269e-01
8.05408299e-01 -1.36156902e-01 -1.06470942e+00 -2.37254910e-02
-3.59071940e-01 -5.78616839e-03 5.79482794e-01 5.97251594e-01
-1.22874892e+00 7.52640545e-01 6.96782255e+00 4.86380845e-01
-6.16748452e-01 1.81798115e-01 2.58688807e-01 -7.80377984e-02
-3.54825705e-01 1.79209188e-01 -7.43891478e-01 3.79211277e-01
1.09056330e+00 -5.36095440e-01 5.66653132e-01 9.22648013e-01
-1.80251509e-01 1.17672466e-01 -1.43217444e+00 6.80916488e-01
1.14841029e-01 -8.79652739e-01 7.08847791e-02 3.80406767e-01
1.01377189e+00 5.46614453e-02 2.26976782e-01 1.49726391e-01
9.31319952e-01 -5.54548085e-01 1.92882806e-01 3.02270442e-01
6.93019509e-01 -6.24072433e-01 4.55927312e-01 4.25605983e-01
-8.85531723e-01 -9.62713733e-02 -4.05228138e-01 5.51391058e-02
-1.11989431e-01 2.34808445e-01 -8.32319498e-01 7.02022433e-01
4.54777271e-01 1.03996599e+00 -2.78576314e-01 6.18012488e-01
7.91004896e-02 4.66254443e-01 -2.77588069e-01 5.50515890e-01
-8.96885842e-02 -4.09781516e-01 6.71090722e-01 1.00940728e+00
3.00358474e-01 3.70188802e-02 3.40492725e-01 6.50949061e-01
-2.01378211e-01 2.22400144e-01 -1.11251879e+00 -3.39832231e-02
6.30592465e-01 8.63454401e-01 -2.76253790e-01 -4.59015399e-01
-7.05683112e-01 1.26666772e+00 6.05046809e-01 8.08750629e-01
-4.17905986e-01 1.61128212e-02 1.20249605e+00 -2.69228816e-01
2.63960898e-01 -2.77943879e-01 -3.29708159e-01 -1.63725054e+00
-8.06599297e-03 -9.89079595e-01 9.11201298e-01 -3.50077778e-01
-1.98862052e+00 1.15596876e-01 2.54871458e-01 -1.40978765e+00
-4.81535017e-01 -4.11565334e-01 -2.02228203e-01 1.20615745e+00
-1.45639801e+00 -1.05049229e+00 1.95955291e-01 1.04071569e+00
4.77064639e-01 -5.83483398e-01 1.11028087e+00 1.57788411e-01
-4.02872652e-01 7.05197036e-01 8.33074629e-01 -1.59425274e-01
1.37637627e+00 -1.30197215e+00 2.50153989e-01 8.52810860e-01
3.29501510e-01 9.46084440e-01 5.99063396e-01 -5.43322504e-01
-1.14552939e+00 -9.33499932e-01 9.67846453e-01 -7.40919471e-01
6.92866921e-01 -4.12483335e-01 -1.31027579e+00 1.14868462e+00
1.39700100e-01 -1.50628209e-01 1.27539480e+00 5.13326824e-01
-7.22999275e-01 1.12724155e-01 -1.33048534e+00 3.44787627e-01
8.30132723e-01 -7.20105052e-01 -6.04955256e-01 3.18578333e-01
5.72397947e-01 1.80905610e-02 -9.70877767e-01 -1.01082548e-01
4.46412742e-01 -7.03225136e-01 1.11385477e+00 -1.02207291e+00
2.44501039e-01 -2.23833933e-01 -5.40589750e-01 -1.53728020e+00
-4.37292457e-01 -3.91261071e-01 3.64461951e-02 1.34351110e+00
4.61334080e-01 -9.51596975e-01 7.04174161e-01 8.98216307e-01
4.15192515e-01 1.11552887e-01 -9.94992137e-01 -9.45510089e-01
3.91424894e-01 -1.21091351e-01 6.41718745e-01 1.20925307e+00
4.64506298e-02 4.88982201e-01 -6.47433639e-01 4.58819926e-01
8.89089823e-01 1.64805382e-01 1.00754309e+00 -1.48577189e+00
-4.65965956e-01 -7.96728209e-02 -2.03306541e-01 -1.01779914e+00
4.61408734e-01 -8.75853360e-01 -2.18270291e-02 -9.44937527e-01
5.45728743e-01 -4.41429973e-01 -6.79860711e-01 4.37547177e-01
-2.17829227e-01 1.81511492e-02 9.03645307e-02 5.24764538e-01
-5.62938452e-01 4.74395454e-01 6.80104554e-01 -5.08184917e-02
-3.37451160e-01 9.91575718e-02 -1.04179990e+00 5.80136180e-01
7.30876207e-01 -8.24812651e-01 -5.76985836e-01 -4.22796756e-01
-2.39828840e-01 -6.23746626e-02 2.86940545e-01 -7.47811735e-01
2.75855422e-01 -5.88845611e-01 4.36187088e-01 -2.37610545e-02
3.26940984e-01 -1.04941666e+00 2.85043806e-01 -2.11017989e-02
-5.72527289e-01 -2.54758894e-01 -5.21793263e-03 9.84729171e-01
-1.54931560e-01 -3.01675439e-01 1.13730621e+00 -6.23220243e-02
-7.12225854e-01 2.92873889e-01 -6.23294674e-02 2.24821880e-01
1.09035099e+00 -2.45534286e-01 3.71481627e-02 -3.18732172e-01
-1.06509197e+00 2.98954919e-02 1.00238407e+00 3.81921262e-01
1.95432171e-01 -1.45201266e+00 -6.95078254e-01 4.41971600e-01
5.67003369e-01 -2.29113042e-01 2.84603447e-01 3.97735894e-01
4.48202640e-01 3.08597773e-01 -4.00750250e-01 -6.63435102e-01
-1.21107471e+00 7.04631269e-01 4.15308997e-02 -4.37639117e-01
-3.15660208e-01 5.28328001e-01 6.20294988e-01 -8.37251484e-01
-7.03669414e-02 2.87691772e-01 2.67680380e-02 1.29718274e-01
2.87171870e-01 2.49878004e-01 -1.86408207e-01 -8.06397200e-01
-2.49262348e-01 5.71306646e-01 -4.32144731e-01 -4.43960339e-01
1.19226801e+00 -7.07192600e-01 2.36634657e-01 6.72303200e-01
1.24607968e+00 -3.80972289e-02 -1.49956751e+00 -1.07321107e+00
1.63950339e-01 -6.69026136e-01 -5.17638147e-01 -7.28707850e-01
-6.15692735e-01 7.23589838e-01 4.63132977e-01 -4.89200242e-02
1.17509520e+00 1.75543591e-01 3.22294205e-01 4.01068032e-01
3.75462919e-01 -9.89776313e-01 8.62991586e-02 3.57024997e-01
4.17667150e-01 -1.62624192e+00 -1.07424669e-01 -2.12427471e-02
-9.13516700e-01 8.08359802e-01 5.24218976e-01 -5.67958988e-02
4.53674465e-01 -1.87038481e-01 9.33766961e-02 2.50346005e-01
-8.33753288e-01 -1.40053645e-01 1.67870477e-01 1.05151808e+00
1.99592710e-01 1.35916278e-01 8.63186941e-02 5.87029934e-01
2.03609303e-01 -6.95644245e-02 4.08398360e-01 8.07785273e-01
-1.52725011e-01 -1.81073856e+00 -3.79730940e-01 2.14452893e-01
-3.35050702e-01 1.17693424e-01 -5.13150275e-01 7.63598680e-01
-5.79472445e-02 7.47701108e-01 -1.22605078e-01 -1.87705401e-02
2.14488387e-01 4.72942084e-01 2.85275787e-01 -8.81831050e-01
-1.49754196e-01 1.18427925e-01 -3.61434788e-01 -3.35064679e-01
-6.15771055e-01 -1.24586320e+00 -6.70816839e-01 -1.19707644e-01
-5.15567418e-03 2.74666995e-01 4.25981075e-01 9.99258697e-01
5.19648254e-01 -1.42853320e-01 7.61517465e-01 -5.46392620e-01
-8.07069719e-01 -7.13348985e-01 -8.84104431e-01 8.11593473e-01
6.15774512e-01 -8.54530871e-01 -3.11712921e-01 4.47476029e-01] | [10.345138549804688, 3.1325161457061768] |
3ac0b527-625d-4647-aefd-6513aa82ce97 | real-time-model-free-deep-reinforcement | 2304.04911 | null | https://arxiv.org/abs/2304.04911v1 | https://arxiv.org/pdf/2304.04911v1.pdf | Real-Time Model-Free Deep Reinforcement Learning for Force Control of a Series Elastic Actuator | Many state-of-the art robotic applications utilize series elastic actuators (SEAs) with closed-loop force control to achieve complex tasks such as walking, lifting, and manipulation. Model-free PID control methods are more prone to instability due to nonlinearities in the SEA where cascaded model-based robust controllers can remove these effects to achieve stable force control. However, these model-based methods require detailed investigations to characterize the system accurately. Deep reinforcement learning (DRL) has proved to be an effective model-free method for continuous control tasks, where few works deal with hardware learning. This paper describes the training process of a DRL policy on hardware of an SEA pendulum system for tracking force control trajectories from 0.05 - 0.35 Hz at 50 N amplitude using the Proximal Policy Optimization (PPO) algorithm. Safety mechanisms are developed and utilized for training the policy for 12 hours (overnight) without an operator present within the full 21 hours training period. The tracking performance is evaluated showing improvements of $25$ N in mean absolute error when comparing the first 18 min. of training to the full 21 hours for a 50 N amplitude, 0.1 Hz sinusoid desired force trajectory. Finally, the DRL policy exhibits better tracking and stability margins when compared to a model-free PID controller for a 50 N chirp force trajectory. | ['Alexander Leonessa', 'Connor W. Herron', 'Stephen Welch', 'Aydin Gokce', 'Ruturaj Sambhus'] | 2023-04-11 | null | null | null | null | ['continuous-control'] | ['playing-games'] | [-1.36176646e-01 5.03435969e-01 -2.91237175e-01 4.92848456e-01
-1.48313537e-01 -5.16289830e-01 9.79160815e-02 -2.66221166e-01
-5.51570117e-01 1.02176297e+00 -3.29532534e-01 -3.26784313e-01
-5.89451611e-01 -2.37530589e-01 -1.10396695e+00 -8.31710279e-01
-5.10694504e-01 2.79865742e-01 2.88075417e-01 -1.04289639e+00
-4.55407649e-02 4.05653954e-01 -1.33247948e+00 -4.47671711e-01
6.56503141e-01 5.54695487e-01 3.95127803e-01 6.24805093e-01
9.30362046e-01 3.48387897e-01 -5.18208683e-01 4.70829219e-01
6.23886049e-01 5.99755496e-02 -4.37692642e-01 -1.32216029e-02
1.34357065e-01 -6.71806395e-01 -2.88669646e-01 8.65866661e-01
9.18492854e-01 5.09535015e-01 3.69508892e-01 -1.17691958e+00
-9.87471268e-02 4.71575171e-01 -3.15601021e-01 -1.78834230e-01
9.41898301e-02 5.10707974e-01 3.56372952e-01 -4.98294175e-01
4.99073893e-01 1.06645727e+00 9.90626693e-01 7.42407918e-01
-1.28561711e+00 -8.03954482e-01 -3.21375370e-01 -2.79630810e-01
-1.01775622e+00 -1.68344855e-01 4.69547778e-01 -7.02491879e-01
9.81760621e-01 -3.50695178e-02 7.67530799e-01 1.04990363e+00
9.54513013e-01 3.80319566e-01 8.28036249e-01 -8.41839612e-02
-4.21104813e-03 -8.35285932e-02 -4.63574380e-01 5.52686870e-01
4.23660159e-01 7.27012217e-01 -1.09084889e-01 1.41671494e-01
1.22687387e+00 -3.64513040e-01 -3.55291158e-01 -2.67361313e-01
-9.26095128e-01 6.69568837e-01 6.02486908e-01 -1.97041899e-01
-7.07746804e-01 4.31476712e-01 5.43975294e-01 6.01112604e-01
-1.79417863e-01 1.21541476e+00 -8.89058232e-01 -2.99898118e-01
-2.91121334e-01 8.40368986e-01 7.87085176e-01 8.84646475e-01
-1.07178036e-02 1.01895034e+00 3.68180424e-01 6.48256302e-01
7.28250518e-02 6.08580410e-01 3.48613560e-01 -1.12400043e+00
1.83357224e-01 3.51083428e-01 6.27235532e-01 -6.68114007e-01
-7.99193859e-01 -3.74706924e-01 -6.45496428e-01 7.92508423e-01
3.87051582e-01 -1.03385425e+00 -4.96862531e-01 1.42147279e+00
4.47756290e-01 -4.30927277e-01 4.22493577e-01 1.10126066e+00
-1.06670558e-01 8.83726835e-01 -2.62711227e-01 -3.23123008e-01
1.05560207e+00 -5.96918821e-01 -7.90203929e-01 -2.80769497e-01
2.04397455e-01 -5.60602844e-01 1.11772013e+00 7.86758482e-01
-1.21052897e+00 -6.32533252e-01 -1.26408434e+00 5.14593840e-01
8.16924348e-02 2.65699297e-01 -4.74578626e-02 1.95403755e-01
-6.35662079e-01 1.27514577e+00 -1.30201840e+00 8.10001940e-02
-5.54896414e-01 8.92180026e-01 -5.39691970e-02 8.08107615e-01
-1.56405544e+00 1.36631548e+00 5.41964948e-01 1.50129229e-01
-7.33852684e-01 -1.01930952e+00 -6.86252117e-01 -2.37091154e-01
6.26359582e-01 -4.28702027e-01 1.41410875e+00 -6.78265810e-01
-2.29021788e+00 -1.52715147e-01 5.55853784e-01 -7.65655279e-01
5.51395178e-01 -9.00982618e-01 -1.09830670e-01 1.55963197e-01
-3.15014273e-01 3.03243816e-01 1.06960750e+00 -1.03876507e+00
-3.38679850e-01 4.54073370e-01 -8.68194178e-02 2.50611633e-01
-2.58951008e-01 -3.86852145e-01 4.08576429e-01 -7.61961699e-01
-2.76699334e-01 -1.60427415e+00 -7.70876482e-02 7.99688622e-02
1.11312941e-01 -3.97921503e-05 1.02916694e+00 -5.44511020e-01
1.15263176e+00 -1.88412023e+00 2.71447092e-01 1.46095634e-01
-6.25548124e-01 6.70879722e-01 3.53162616e-01 7.48627603e-01
-8.04680362e-02 -1.84458882e-01 9.79614183e-02 4.77037936e-01
-1.11333750e-01 2.95819551e-01 -3.03800911e-01 3.20737660e-01
4.49765414e-01 3.51130754e-01 -8.42533231e-01 1.53526559e-01
2.82473024e-02 3.17266375e-01 -5.57412565e-01 3.55479568e-01
-2.07263783e-01 5.64917624e-01 -3.35269958e-01 4.10499424e-01
1.18715085e-01 2.14074254e-01 1.55369163e-01 -2.84977525e-01
-7.55362749e-01 -1.76727772e-01 -1.20019853e+00 1.09905910e+00
-4.32129413e-01 4.04033929e-01 9.24260855e-01 -1.01212883e+00
1.27992308e+00 6.81837261e-01 7.31570780e-01 -3.41652125e-01
4.97724354e-01 6.97472095e-01 3.62463742e-01 -7.78606236e-01
3.15978736e-01 -3.61912698e-01 1.54229105e-01 -2.01685742e-01
-9.32456329e-02 -7.43294835e-01 2.63115149e-02 -4.36912626e-01
1.01766181e+00 5.56996584e-01 -5.41225225e-02 -8.32639337e-01
5.10473251e-01 4.48797226e-01 7.63497353e-01 1.93600848e-01
-2.17664793e-01 -2.25892469e-01 5.10682464e-02 -2.64227003e-01
-1.36612058e+00 -6.31740093e-01 -8.37153569e-02 9.13683414e-01
1.92204073e-01 1.93389475e-01 -4.20251936e-01 4.58296418e-01
4.96698260e-01 5.64974725e-01 -6.43653274e-02 -7.22875774e-01
-1.07306385e+00 -1.70557842e-01 3.16845894e-01 7.73323596e-01
4.23202902e-01 -7.64392614e-01 -1.01110029e+00 6.70285702e-01
4.12862390e-01 -8.53461802e-01 -2.33499691e-01 2.04710305e-01
-1.01899648e+00 -8.55145335e-01 -5.53828835e-01 -8.05243134e-01
4.82484162e-01 -4.24664497e-01 3.14216346e-01 -6.68435693e-02
-2.10695043e-01 2.92434186e-01 -7.43947327e-02 -8.07904363e-01
-4.52169925e-01 -1.86914310e-01 9.51061726e-01 -6.88301861e-01
-5.94707668e-01 -6.63959607e-02 -5.72154403e-01 8.22877884e-01
-4.13526297e-01 -1.66626111e-01 6.30664468e-01 1.23948801e+00
4.59785908e-01 8.19098502e-02 7.79256642e-01 -1.60252228e-01
8.36219549e-01 -1.38125196e-01 -1.12533283e+00 -2.99187958e-01
-4.17265624e-01 7.54597113e-02 1.03557336e+00 -1.04088748e+00
-1.01253545e+00 4.39438432e-01 -1.81219369e-01 -7.20040858e-01
5.32999575e-01 4.60762531e-01 7.42009580e-01 -3.14932555e-01
9.24765587e-01 1.60462561e-03 9.54513967e-01 -8.60021263e-02
-2.97240138e-01 4.19619650e-01 5.02415895e-01 -7.80083477e-01
8.82799149e-01 -7.85832107e-02 5.16342998e-01 -8.73345494e-01
-3.67327601e-01 -2.84603804e-01 -2.75366813e-01 -3.67487490e-01
4.49450225e-01 -9.69132006e-01 -1.37786365e+00 3.33863497e-01
-6.57067478e-01 -9.93172944e-01 -8.60092193e-02 9.47308421e-01
-9.42280352e-01 -2.04715490e-01 -8.57175589e-01 -1.15076673e+00
-7.47399569e-01 -1.06582677e+00 7.37076581e-01 3.94851446e-01
-4.78776217e-01 -7.70907104e-01 -7.17631653e-02 -1.59427464e-01
8.00241530e-01 6.65514290e-01 5.12266040e-01 1.92451805e-01
1.37250036e-01 -4.74921584e-01 7.75873363e-01 5.97059608e-01
-3.98725830e-02 6.10389397e-04 -4.03719246e-01 -1.18146431e+00
1.33362740e-01 -5.74232996e-01 -8.12814757e-03 5.70015728e-01
3.91824394e-01 -6.17411613e-01 -2.85886109e-01 1.31635204e-01
1.57549739e+00 4.16702837e-01 2.03055188e-01 5.54566324e-01
4.08732384e-01 4.50338840e-01 1.13441098e+00 4.60749567e-01
-2.58270025e-01 4.39684927e-01 5.89370906e-01 5.94484545e-02
3.36866051e-01 5.85949756e-02 7.77900398e-01 4.98499304e-01
-3.16535294e-01 3.45390260e-01 -8.42467010e-01 5.93478858e-01
-1.55250144e+00 -6.25695646e-01 -1.00524351e-01 2.32392001e+00
1.07544148e+00 4.22926605e-01 2.10887745e-01 3.96019131e-01
8.38374972e-01 -7.12635815e-01 -8.19333017e-01 -6.40218258e-01
3.33850980e-01 1.26066580e-01 1.28639269e+00 5.05733013e-01
-1.04522014e+00 5.31225324e-01 5.48394489e+00 4.65419829e-01
-1.67561257e+00 -7.37116337e-01 -3.63556147e-01 -6.95735738e-02
6.18523955e-01 -2.25226402e-01 -8.48488331e-01 5.13079524e-01
1.21847630e+00 -4.86121148e-01 6.17541671e-01 1.22157907e+00
8.04271162e-01 -1.63210910e-02 -6.98307753e-01 1.15516752e-01
-5.67713201e-01 -1.06864572e+00 -5.08983850e-01 -1.46603823e-01
9.70974088e-01 -1.01927295e-01 -1.43002301e-01 6.86217487e-01
1.86263517e-01 -5.26074946e-01 8.83541226e-01 4.93731499e-01
6.55243993e-01 -8.53980303e-01 7.53603876e-01 7.60626793e-01
-1.07402253e+00 -6.57202780e-01 -3.28830659e-01 -6.12210870e-01
3.67404610e-01 4.61945795e-02 -8.84303570e-01 4.02271003e-01
5.28686881e-01 3.01090777e-01 3.27765763e-01 8.27762246e-01
-1.28427684e-01 6.59817398e-01 -5.86046636e-01 -5.96274376e-01
4.03219759e-01 -1.40056372e-01 1.07289374e+00 5.40694773e-01
3.13090652e-01 1.75827667e-01 4.62402761e-01 5.37799537e-01
5.87226987e-01 -3.14581484e-01 -4.99635369e-01 3.18177305e-02
2.55180150e-01 8.92393231e-01 8.72143209e-02 -5.50161414e-02
3.47831130e-01 2.55341858e-01 -3.74355108e-01 -2.66953893e-02
-8.94042552e-01 -1.06004596e+00 8.81367803e-01 4.79205072e-01
1.05712049e-01 -8.26595306e-01 1.38382986e-01 -3.94234061e-01
-2.40104832e-02 -7.00676262e-01 -1.85741574e-01 -6.25694215e-01
-1.08022678e+00 1.30032867e-01 3.35306346e-01 -1.82224691e+00
-8.46207201e-01 -8.89935434e-01 -5.15566111e-01 8.76220763e-01
-1.31801832e+00 -3.05004627e-01 7.93662146e-02 3.33979934e-01
7.10411668e-01 1.60938166e-02 8.21197748e-01 2.04982385e-01
-3.26313972e-01 1.50508638e-02 3.57868105e-01 -7.80764669e-02
7.39824831e-01 -1.12702119e+00 -1.36485219e-01 4.90986794e-01
-1.40831172e+00 7.80294061e-01 1.33608472e+00 -9.20636773e-01
-2.14449906e+00 -1.04733956e+00 2.06066221e-01 3.30463022e-01
1.02093828e+00 1.43378139e-01 -1.11765552e+00 3.59734684e-01
1.38707370e-01 5.66342939e-03 -2.32615381e-01 -7.61908174e-01
7.19463170e-01 -2.68941402e-01 -1.09599268e+00 6.86872840e-01
4.44225848e-01 1.74988806e-01 -6.36701405e-01 1.09505706e-01
8.45460534e-01 -1.10503757e+00 -1.21030354e+00 9.17174459e-01
8.22120309e-01 6.61377981e-02 8.25154781e-01 -3.91248137e-01
3.82185489e-01 -5.31884789e-01 5.82311094e-01 -1.71987367e+00
-1.25222072e-01 -1.33480752e+00 -3.13835680e-01 4.60857242e-01
3.19687277e-01 -6.95973694e-01 5.93698740e-01 2.55088776e-01
-5.09359658e-01 -1.02686000e+00 -7.69420981e-01 -1.34416378e+00
3.08978170e-01 1.66441575e-01 -1.83248281e-01 6.12665236e-01
4.40810084e-01 2.08169222e-01 -6.33652270e-01 6.97345853e-01
4.01903152e-01 -3.32559824e-01 8.95598173e-01 -1.05850494e+00
-3.56078863e-01 -1.96343467e-01 -2.50019059e-02 -5.25707006e-01
-2.39408445e-02 -1.71234131e-01 5.16711831e-01 -1.57025361e+00
-1.01680005e+00 -1.45468205e-01 2.63578862e-01 6.50320709e-01
-5.32209091e-02 -3.25965226e-01 -3.52098346e-02 2.00012743e-01
5.03329337e-01 5.17555416e-01 1.48405063e+00 2.67457380e-03
-6.95792377e-01 3.99375588e-01 2.45406762e-01 5.51364601e-01
1.17797542e+00 -2.22218782e-01 -7.39722431e-01 -3.90873462e-01
-5.30547164e-02 5.17273068e-01 4.01350744e-02 -1.54777730e+00
1.23361409e-01 -4.70348448e-01 1.50962010e-01 -1.68138295e-01
3.05991709e-01 -9.84106779e-01 3.84204209e-01 1.44373131e+00
-2.72520185e-01 1.55377522e-01 8.03863764e-01 3.14480275e-01
-1.83001563e-01 1.24971174e-01 1.09221840e+00 1.92075133e-01
-5.52974701e-01 -3.20469707e-01 -5.37766457e-01 -4.53042418e-01
1.08372903e+00 -1.88907832e-01 -7.19324723e-02 -1.80990413e-01
-8.21708739e-01 6.87927306e-01 -4.36275154e-02 4.56002891e-01
2.57614762e-01 -8.74120712e-01 -4.68240559e-01 1.12225145e-01
-5.75761437e-01 -1.08036995e-01 9.55479890e-02 8.08456540e-01
-7.89096355e-01 3.31906646e-01 -8.27472210e-01 -5.55375993e-01
-9.75420833e-01 1.84684753e-01 5.12069881e-01 6.42661303e-02
-6.28066123e-01 4.78677452e-01 -5.32143772e-01 -4.21038657e-01
7.91135132e-02 -5.89321256e-01 1.99565155e-04 -3.51706535e-01
-5.48320711e-02 5.56340158e-01 2.01271459e-01 -5.67354411e-02
-1.44598144e-03 6.98462367e-01 4.53381002e-01 -7.50088394e-02
1.40471303e+00 3.76603007e-01 4.82184440e-01 3.33724082e-01
5.62616110e-01 -4.08385038e-01 -1.98016739e+00 3.50492477e-01
-5.30784205e-02 -2.65969504e-02 3.92582221e-03 -7.11039186e-01
-6.69060588e-01 3.75480741e-01 5.38703859e-01 1.24295026e-01
9.28036094e-01 -9.13386405e-01 6.57703161e-01 6.51154816e-01
4.05790806e-01 -1.62511194e+00 2.00516492e-01 9.88875508e-01
1.41250801e+00 -6.79366589e-01 2.93037057e-01 -3.06318104e-01
-7.66528845e-01 1.43304741e+00 9.56312656e-01 -9.64754343e-01
5.20653784e-01 5.55545926e-01 2.59645041e-02 1.96548387e-01
-6.61685050e-01 1.89507842e-01 2.19414145e-01 2.02956811e-01
2.23816007e-01 1.16569493e-02 -7.90040255e-01 2.06650764e-01
-1.71902720e-02 1.64087400e-01 5.71447551e-01 1.60091650e+00
-8.16747129e-01 -6.78424895e-01 -6.59631371e-01 4.62179519e-02
-4.36262071e-01 4.97666031e-01 2.38222152e-01 1.47360229e+00
-1.70706302e-01 5.51829338e-01 9.44629908e-02 -4.02522475e-01
1.14076042e+00 -4.61901650e-02 2.37852231e-01 -4.29465950e-01
-8.54412675e-01 4.53873187e-01 2.49115527e-01 -5.27715802e-01
-1.62014782e-01 -4.02626902e-01 -1.75386167e+00 -1.23209976e-01
-5.61250031e-01 2.76974197e-02 8.94111574e-01 4.73660171e-01
2.33507589e-01 6.55001581e-01 6.03533626e-01 -1.45415068e+00
-1.42673755e+00 -1.06418800e+00 -4.00886357e-01 -3.29099968e-02
4.81967598e-01 -1.27344918e+00 -4.55158561e-01 5.95486686e-02] | [4.931324005126953, 1.7596080303192139] |
7a5c947e-4949-488d-af5d-65696cbf94f7 | fast-abc-with-joint-generative-modelling-and | 2104.08156 | null | https://arxiv.org/abs/2104.08156v1 | https://arxiv.org/pdf/2104.08156v1.pdf | Fast ABC with joint generative modelling and subset simulation | We propose a novel approach for solving inverse-problems with high-dimensional inputs and an expensive forward mapping. It leverages joint deep generative modelling to transfer the original problem spaces to a lower dimensional latent space. By jointly modelling input and output variables and endowing the latent with a prior distribution, the fitted probabilistic model indirectly gives access to the approximate conditional distributions of interest. Since model error and observational noise with unknown distributions are common in practice, we resort to likelihood-free inference with Approximate Bayesian Computation (ABC). Our method calls on ABC by Subset Simulation to explore the regions of the latent space with dissimilarities between generated and observed outputs below prescribed thresholds. We diagnose the diversity of approximate posterior solutions by monitoring the probability content of these regions as a function of the threshold. We further analyze the curvature of the resulting diagnostic curve to propose an adequate ABC threshold. When applied to a cross-borehole tomography example from geophysics, our approach delivers promising performance without using prior knowledge of the forward nor of the noise distribution. | ['Niklas Linde', 'David Ginsbourger', 'Eliane Maalouf'] | 2021-04-16 | null | null | null | null | ['geophysics'] | ['miscellaneous'] | [ 1.62128076e-01 3.63077432e-01 3.38227302e-01 -2.39822373e-01
-1.13158274e+00 -4.17810708e-01 9.20726299e-01 -4.50198144e-01
-1.11393481e-01 7.78121948e-01 3.16553503e-01 -4.24875110e-01
-5.67305982e-01 -7.44208097e-01 -6.12586617e-01 -9.05248702e-01
1.44500034e-02 9.52509284e-01 -1.67297140e-01 2.19318539e-01
1.19673394e-01 4.91438091e-01 -1.18533802e+00 1.62144601e-01
9.28162992e-01 7.30632603e-01 2.12566972e-01 5.55851161e-01
8.47637728e-02 3.06384474e-01 -4.01357383e-01 5.71752191e-02
1.90596074e-01 -5.87701082e-01 -5.70667744e-01 2.21120760e-01
-4.35890973e-01 -2.73272336e-01 -7.21738264e-02 1.05978799e+00
1.98209926e-01 8.17954317e-02 1.29075778e+00 -9.60916698e-01
-2.37335354e-01 3.55328143e-01 -3.35376978e-01 1.08762704e-01
1.25640303e-01 2.75960237e-01 6.87363327e-01 -1.02322209e+00
6.18069768e-01 1.14333534e+00 8.15583944e-01 -9.96488333e-02
-1.93479919e+00 -3.21181267e-01 -1.65226504e-01 -3.63232195e-01
-1.56446004e+00 -4.91861016e-01 6.38420582e-01 -8.16127956e-01
5.98741472e-01 1.24519756e-02 4.33149874e-01 1.11470914e+00
1.47467881e-01 1.84122354e-01 1.25472784e+00 -4.60106790e-01
6.84080005e-01 3.40648204e-01 -3.19722205e-01 4.74604279e-01
4.07372415e-02 1.96482718e-01 -4.23212439e-01 -5.14733493e-01
9.39282894e-01 -8.70427713e-02 -3.49578112e-01 -5.06259441e-01
-1.07853425e+00 9.42102492e-01 1.93254113e-01 -1.14331901e-01
-5.77280104e-01 3.13117623e-01 -3.10048789e-01 1.16260357e-01
6.54323518e-01 5.34184456e-01 -3.97663414e-01 7.33528361e-02
-1.24398720e+00 3.51686329e-01 9.69445944e-01 8.85182858e-01
8.82214665e-01 5.70868850e-02 7.57245719e-02 5.57344615e-01
7.19455481e-01 7.00385690e-01 -8.47499147e-02 -1.25632310e+00
4.25508082e-01 6.92496821e-02 5.69916010e-01 -6.13279521e-01
-1.84808657e-01 -5.26088476e-01 -5.99108756e-01 3.08449477e-01
6.79053187e-01 -2.66826063e-01 -9.93060708e-01 1.67433441e+00
2.65776366e-01 1.50284588e-01 -3.20809782e-02 7.42178679e-01
-8.40763301e-02 8.27211738e-01 -4.61633168e-02 -3.56875479e-01
1.14344072e+00 -1.76095888e-01 -5.55929363e-01 -1.29700869e-01
2.94828087e-01 -6.01454556e-01 9.56967473e-01 2.98108011e-01
-1.06974983e+00 -2.04306811e-01 -9.31595623e-01 3.17549080e-01
1.03979580e-01 5.49768731e-02 2.12809995e-01 5.70511341e-01
-9.36163664e-01 8.63997221e-01 -1.40720379e+00 -2.79460587e-02
2.05700934e-01 9.54080559e-03 2.09552962e-02 1.78284291e-02
-9.71513271e-01 8.01565289e-01 1.58739761e-01 3.01543713e-01
-1.20184672e+00 -7.64462948e-01 -6.05222344e-01 1.85923785e-01
1.64398640e-01 -7.85275102e-01 1.11296010e+00 -3.75420898e-01
-1.52195323e+00 3.02073002e-01 -1.28506884e-01 -2.47427851e-01
8.20985973e-01 -1.29522473e-01 -1.93624064e-01 2.04079315e-01
2.35403642e-01 2.04132020e-01 1.06738937e+00 -1.37266898e+00
4.87020500e-02 -1.35021046e-01 -3.96925390e-01 -1.37635143e-02
2.72771448e-01 -3.69504184e-01 -2.04605103e-01 -4.54546839e-01
9.24183488e-01 -8.75952244e-01 -5.25270700e-01 3.63273732e-02
-4.66932893e-01 4.17998731e-01 4.49194372e-01 -5.78214407e-01
7.96575785e-01 -1.95275164e+00 3.41367722e-01 7.33517528e-01
1.41508490e-01 -6.06855929e-01 2.42365047e-01 6.47176981e-01
1.17238753e-01 2.43064743e-02 -7.18537629e-01 -4.55800682e-01
2.24429637e-01 3.24333608e-01 -7.17676342e-01 8.24645102e-01
3.73099357e-01 5.74273705e-01 -8.52016687e-01 -2.04016626e-01
1.47663251e-01 4.74947602e-01 -8.34518194e-01 4.76418674e-01
-3.26314390e-01 7.72387326e-01 -7.09164619e-01 4.42526281e-01
6.62534237e-01 -3.50084037e-01 2.64208138e-01 -4.07822505e-02
-1.20340168e-01 2.74685174e-01 -1.45816338e+00 1.51835859e+00
-4.57884401e-01 2.55236417e-01 2.55086750e-01 -9.69233990e-01
9.53594863e-01 3.18577856e-01 2.64722437e-01 -1.41188323e-01
-1.16618022e-01 2.66110271e-01 -2.27627560e-01 -5.67521214e-01
2.23374158e-01 -7.20811307e-01 -8.32357183e-02 8.24317336e-01
2.51870394e-01 -6.09588802e-01 -4.44076806e-01 6.21927157e-02
1.04759526e+00 5.24472117e-01 8.72721225e-02 -6.41892672e-01
-1.25372097e-01 -1.84272841e-01 2.71464765e-01 9.60722506e-01
3.84796172e-01 8.78748298e-01 9.16165829e-01 -8.58831033e-02
-1.23680890e+00 -1.50927317e+00 -5.64294994e-01 3.80917281e-01
-1.70267180e-01 -1.38815686e-01 -6.44034445e-01 -1.42101467e-01
5.45551665e-02 1.00594318e+00 -7.98711658e-01 -2.00138018e-01
-2.84871370e-01 -1.12452734e+00 3.45515966e-01 3.45843554e-01
-3.27865705e-02 -8.44657838e-01 -7.19399989e-01 2.13769272e-01
-2.67683864e-01 -8.98903131e-01 2.42589861e-01 6.60421491e-01
-1.17233849e+00 -6.31260276e-01 -6.04395151e-01 3.64400521e-02
7.91123033e-01 -6.47536933e-01 1.08225727e+00 -4.68080193e-01
-4.82579321e-02 2.37958014e-01 1.96141660e-01 9.16472916e-03
-7.22069442e-01 -2.02948421e-01 4.72157896e-02 -4.32147868e-02
-2.07364962e-01 -9.27963018e-01 -4.06952441e-01 2.95616001e-01
-6.35102808e-01 7.83225447e-02 4.52943802e-01 9.13526356e-01
4.14158612e-01 -8.69920179e-02 3.45355541e-01 -6.91847801e-01
4.40128416e-01 -9.29335177e-01 -9.55974042e-01 3.59397605e-02
-5.04021227e-01 7.22697556e-01 1.91894487e-01 -4.74603623e-01
-1.29395640e+00 -7.14027807e-02 1.29777268e-01 -5.76034486e-01
-1.60889044e-01 7.80288041e-01 -8.63801762e-02 5.14600873e-01
7.14247406e-01 -8.31115171e-02 -5.29196076e-02 -7.80454576e-01
4.18948382e-01 3.24850798e-01 5.88024616e-01 -9.42660749e-01
7.11244166e-01 6.20965898e-01 2.22966313e-01 -5.79493284e-01
-7.20754445e-01 -1.36263013e-01 -6.42714262e-01 -1.22988924e-01
6.85990751e-01 -8.56554210e-01 -4.70629215e-01 8.71998072e-02
-9.76060152e-01 -3.20846319e-01 -5.62458813e-01 7.99831927e-01
-8.99237812e-01 2.91631483e-02 -3.82471710e-01 -1.33134329e+00
2.49585465e-01 -1.36081970e+00 1.19580936e+00 -3.27382952e-01
-4.10543352e-01 -1.01411557e+00 1.95507467e-01 -2.15362534e-01
3.14892203e-01 2.12053090e-01 9.74396408e-01 -2.57369190e-01
-7.13587821e-01 -2.31616601e-01 -1.37962982e-01 -1.06713638e-01
-8.27543214e-02 -8.27478915e-02 -1.21376789e+00 -6.31002262e-02
4.76267636e-01 -1.00088857e-01 8.98263216e-01 7.81014144e-01
9.06460166e-01 -2.08732560e-01 -4.50697482e-01 8.33393157e-01
1.35141766e+00 -1.13078974e-01 5.73507845e-01 1.61518827e-01
2.67337948e-01 6.51743591e-01 2.25417420e-01 7.87523746e-01
2.76423898e-03 5.04221201e-01 2.75399566e-01 6.12696493e-03
3.63066196e-01 -5.07985115e-01 4.87465858e-02 4.68709677e-01
-3.21269259e-02 -1.21287763e-01 -1.11119795e+00 6.08750045e-01
-1.55863380e+00 -7.80153334e-01 8.03628936e-02 2.22602820e+00
7.63207972e-01 2.18610197e-01 -2.34583259e-01 -6.92289174e-02
5.72102427e-01 -1.73771694e-01 -4.49503660e-01 -4.97068651e-03
2.07581565e-01 2.79941410e-01 4.43427444e-01 7.42210925e-01
-7.00661004e-01 3.40698332e-01 7.20676661e+00 5.47009528e-01
-7.04864979e-01 1.35869384e-01 5.24583042e-01 1.24958843e-01
-7.73479998e-01 4.36499357e-01 -5.56362092e-01 5.55697143e-01
1.14817357e+00 2.06848308e-02 3.60689342e-01 4.29914147e-01
3.55635256e-01 -4.06182766e-01 -1.04077208e+00 4.05851990e-01
-5.02237141e-01 -1.22004116e+00 -1.72717929e-01 3.80361587e-01
7.62983859e-01 8.44978988e-02 1.20662726e-01 9.45935547e-02
6.03526831e-01 -1.00619757e+00 9.39316094e-01 1.11759460e+00
1.00230205e+00 -5.55279613e-01 5.44177711e-01 6.69104874e-01
-6.68542624e-01 -3.48225273e-02 -2.96717882e-01 -1.11138083e-01
5.69945931e-01 8.40151608e-01 -9.00935650e-01 2.67095804e-01
6.28118336e-01 3.35659921e-01 5.05491421e-02 7.26161361e-01
-2.57222891e-01 9.01936710e-01 -8.81762505e-01 4.78388995e-01
1.34126619e-01 -6.15863919e-01 6.90228164e-01 9.32199955e-01
7.75341213e-01 1.56973317e-01 2.03357022e-02 1.78520620e+00
3.05667520e-01 -5.33513963e-01 -5.56598306e-01 2.82751266e-02
5.20682454e-01 9.31981385e-01 -9.14306164e-01 -2.08899066e-01
7.01948553e-02 5.22988498e-01 7.12231547e-02 8.87526214e-01
-7.14431763e-01 1.00384705e-01 3.69015008e-01 2.36817166e-01
4.88320619e-01 -3.22733164e-01 -4.57130820e-01 -1.03924978e+00
-2.96301413e-02 -3.51697236e-01 1.91805482e-01 -9.82079744e-01
-1.04737604e+00 2.94378102e-01 5.32491505e-01 -1.20077145e+00
-6.21802390e-01 -3.32632542e-01 -6.07656419e-01 1.52432036e+00
-1.04786968e+00 -8.60155642e-01 -2.93007251e-02 1.29149348e-01
1.33932725e-01 7.05568790e-02 7.94574201e-01 -2.46226728e-01
-2.13885039e-01 -5.03434287e-03 4.80074197e-01 -3.43058556e-01
1.85261607e-01 -1.22349679e+00 5.04393399e-01 6.91375673e-01
-5.20852320e-02 6.66693389e-01 1.32009089e+00 -7.80196905e-01
-1.08746243e+00 -7.60350227e-01 4.38905746e-01 -6.26315951e-01
9.64334548e-01 -4.54093248e-01 -1.02439415e+00 7.67240644e-01
-2.15396285e-01 2.23442838e-01 2.03790128e-01 -2.89076660e-02
-6.63730800e-02 4.55618888e-01 -1.11948192e+00 2.94557780e-01
6.02862597e-01 -7.25432277e-01 -5.44602990e-01 1.90700829e-01
4.01219577e-01 -3.11927259e-01 -8.14675272e-01 3.90828758e-01
4.99559492e-01 -8.22852910e-01 9.28746700e-01 -4.85121548e-01
5.64785600e-01 -3.38159740e-01 -3.77281934e-01 -1.30814111e+00
-1.80815458e-01 -6.05341315e-01 -1.07369825e-01 1.00915444e+00
4.98179913e-01 -5.33799052e-01 6.14900529e-01 6.38395548e-01
-6.51479736e-02 -6.26012862e-01 -1.03111744e+00 -4.68882948e-01
2.17828155e-01 -8.51936340e-01 4.41592008e-01 6.66684330e-01
-2.78074354e-01 -9.54475850e-02 -2.26407111e-01 7.06137538e-01
8.80343556e-01 1.87061965e-01 3.74996066e-01 -1.13885188e+00
-7.84176469e-01 -3.77612449e-02 -6.24523647e-02 -8.99263442e-01
-3.65153849e-02 -6.71860516e-01 3.59323382e-01 -1.30677795e+00
1.48874775e-01 -7.35346794e-01 2.84144990e-02 1.19032204e-01
1.26893580e-01 1.06492922e-01 -2.13928893e-01 4.84654933e-01
-1.34903546e-02 7.17479229e-01 8.67690980e-01 2.63694882e-01
-1.10801399e-01 1.19967006e-01 -1.58222154e-01 8.30941737e-01
3.98134649e-01 -8.32623184e-01 -3.97368491e-01 -2.77434736e-01
3.79478633e-01 5.83272159e-01 7.31686592e-01 -7.51616836e-01
3.17998491e-02 -2.37709701e-01 5.89178622e-01 -7.45569408e-01
5.44936657e-01 -6.89983726e-01 7.05794215e-01 2.22633511e-01
-3.68310034e-01 -2.15772629e-01 5.59752025e-02 7.03293025e-01
4.07718383e-02 -4.08941090e-01 7.53133059e-01 -9.77665484e-02
-1.07092578e-02 -2.08810046e-01 -6.53890431e-01 1.28569296e-02
5.13622820e-01 -7.99861401e-02 2.01209873e-01 -5.19808650e-01
-1.23950922e+00 3.26250941e-02 5.53301036e-01 -2.78127074e-01
4.87679899e-01 -9.87374425e-01 -7.17889547e-01 6.30348206e-01
-7.35664293e-02 9.33330227e-03 1.66851908e-01 9.03359115e-01
-2.91384429e-01 1.60945449e-02 1.29488498e-01 -9.52314436e-01
-2.86379099e-01 2.15584561e-01 7.27844179e-01 -7.98757970e-02
-7.94732213e-01 7.45547712e-01 3.35743248e-01 -6.22728705e-01
-1.62680268e-01 -5.13568938e-01 2.04737976e-01 -4.48153801e-02
1.07299194e-01 4.77584422e-01 -8.39458220e-03 -3.37936550e-01
-1.04142621e-01 2.93723613e-01 4.52990323e-01 -6.95697486e-01
1.43336093e+00 -2.63359636e-01 3.71900387e-02 7.23946095e-01
1.07875824e+00 -1.21596918e-01 -1.68021512e+00 -1.52512014e-01
1.91496715e-01 -2.85305262e-01 2.11883381e-01 -6.55422270e-01
-6.94524109e-01 9.19315696e-01 3.50473583e-01 1.27747804e-01
7.29457498e-01 3.03913891e-01 -1.04503706e-01 1.41454890e-01
1.92248136e-01 -7.98436940e-01 -1.68247268e-01 3.39282840e-01
9.18928564e-01 -8.44426751e-01 2.83999115e-01 -8.96929726e-02
-4.41508859e-01 1.00143909e+00 -1.63174458e-02 -2.16806352e-01
9.15714324e-01 6.39422297e-01 4.97226864e-02 -5.20448864e-01
-7.76999652e-01 2.39597201e-01 2.18612835e-01 3.79066229e-01
-1.87835842e-02 -4.28389832e-02 3.23726267e-01 4.99889255e-01
-2.99884826e-01 -6.93476200e-02 4.68663782e-01 7.23414779e-01
-2.53409445e-01 -6.37891531e-01 -7.18957424e-01 4.52221304e-01
-2.21997768e-01 -1.18466556e-01 2.59625047e-01 7.07869709e-01
-2.02964440e-01 5.73281467e-01 3.33261609e-01 2.51500100e-01
9.12835374e-02 2.40365013e-01 2.47756079e-01 -5.85931778e-01
1.69828534e-01 6.78569436e-01 -2.56416318e-03 -4.61902291e-01
-9.30133164e-02 -1.05887115e+00 -1.02586341e+00 7.96294883e-02
-4.63772804e-01 3.18752199e-01 6.91343546e-01 1.13874769e+00
-8.73923227e-02 4.22519773e-01 4.43105131e-01 -1.17821562e+00
-1.11251163e+00 -1.14181244e+00 -7.60893226e-01 2.56188549e-02
5.49094737e-01 -8.82038176e-01 -8.22720826e-01 -3.09651787e-03] | [6.860017776489258, 3.7596347332000732] |
bbcb7cbf-6b7a-422d-ab66-95aeee135ed1 | mitigating-artifacts-in-real-world-video | 2212.07339 | null | https://arxiv.org/abs/2212.07339v1 | https://arxiv.org/pdf/2212.07339v1.pdf | Mitigating Artifacts in Real-World Video Super-Resolution Models | The recurrent structure is a prevalent framework for the task of video super-resolution, which models the temporal dependency between frames via hidden states. When applied to real-world scenarios with unknown and complex degradations, hidden states tend to contain unpleasant artifacts and propagate them to restored frames. In this circumstance, our analyses show that such artifacts can be largely alleviated when the hidden state is replaced with a cleaner counterpart. Based on the observations, we propose a Hidden State Attention (HSA) module to mitigate artifacts in real-world video super-resolution. Specifically, we first adopt various cheap filters to produce a hidden state pool. For example, Gaussian blur filters are for smoothing artifacts while sharpening filters are for enhancing details. To aggregate a new hidden state that contains fewer artifacts from the hidden state pool, we devise a Selective Cross Attention (SCA) module, in which the attention between input features and each hidden state is calculated. Equipped with HSA, our proposed method, namely FastRealVSR, is able to achieve 2x speedup while obtaining better performance than Real-BasicVSR. Codes will be available at https://github.com/TencentARC/FastRealVSR | ['Ying Shan', 'Chao Dong', 'Jinjin Gu', 'Shuwei Shi', 'Xintao Wang', 'Liangbin Xie'] | 2022-12-14 | null | null | null | null | ['video-super-resolution'] | ['computer-vision'] | [ 2.47897908e-01 -2.53384650e-01 -4.89425808e-02 -8.83972794e-02
-6.08395815e-01 -1.14844233e-01 4.99168783e-01 -4.71442908e-01
-8.33986606e-03 7.00306237e-01 6.26102984e-01 1.35333315e-01
2.11602807e-01 -5.04473567e-01 -5.71335793e-01 -9.06331360e-01
1.74617380e-01 -3.84859920e-01 2.92599708e-01 -3.51739824e-01
1.16602056e-01 1.81523353e-01 -1.63347220e+00 4.30192411e-01
9.63500917e-01 6.02112412e-01 6.32442474e-01 5.48204958e-01
7.83094857e-03 1.18077898e+00 -5.35767436e-01 -1.57897860e-01
5.99700883e-02 -6.20916009e-01 -5.30996919e-01 1.68234229e-01
1.48235828e-01 -4.78445977e-01 -6.73335135e-01 1.44706082e+00
4.61009532e-01 1.61279798e-01 7.85301775e-02 -7.82682061e-01
-7.96952367e-01 5.90660751e-01 -7.08797991e-01 6.43076062e-01
5.02589643e-01 3.99685770e-01 7.87137628e-01 -8.78935397e-01
6.31410837e-01 1.41365445e+00 2.28481054e-01 7.72613168e-01
-1.19854414e+00 -7.53207743e-01 4.89838868e-01 5.54993749e-01
-1.37332940e+00 -8.96961689e-01 7.86291480e-01 -1.63172573e-01
6.24937713e-01 5.27493715e-01 5.19667149e-01 1.41466236e+00
2.74625063e-01 7.90210187e-01 1.09579170e+00 4.84732166e-02
1.03194639e-01 6.91251159e-02 1.88269347e-01 4.29877818e-01
1.78506091e-01 3.28883789e-02 -6.63192689e-01 5.25569059e-02
9.71177757e-01 2.42935523e-01 -8.57556283e-01 1.86107963e-01
-1.27404964e+00 4.38818544e-01 3.13774377e-01 3.71593177e-01
-5.88540435e-01 7.04820678e-02 2.80616075e-01 1.85244843e-01
5.49060106e-01 8.17739144e-02 -7.84829557e-02 1.90212708e-02
-8.94548893e-01 -1.40375821e-02 1.43271849e-01 8.23675036e-01
7.47075200e-01 3.12020630e-01 -5.50076723e-01 6.25580251e-01
7.19727501e-02 3.78168464e-01 5.80574512e-01 -1.05456209e+00
3.20145965e-01 2.70567596e-01 4.75433528e-01 -1.00988734e+00
-4.58869189e-02 -7.48433888e-01 -1.25057626e+00 2.03101411e-01
1.22776525e-02 2.26650126e-02 -8.46865952e-01 1.77746487e+00
3.49675477e-01 8.02178442e-01 1.86118081e-01 1.29246557e+00
8.09416473e-01 1.08072889e+00 -6.85155392e-02 -7.58147240e-01
1.35524297e+00 -8.92389536e-01 -1.33617437e+00 -1.01881236e-01
-1.35001671e-02 -7.98019826e-01 1.05483103e+00 5.01870692e-01
-1.36657870e+00 -8.01928699e-01 -1.04049778e+00 -2.52128214e-01
2.09207624e-01 6.90819472e-02 1.90667033e-01 -1.49260927e-03
-1.05633628e+00 7.60070324e-01 -1.05771482e+00 -7.99644589e-02
1.47933275e-01 -5.26362360e-02 -1.34732291e-01 -5.80532197e-03
-1.38096201e+00 8.10752511e-01 1.14681877e-01 5.16243041e-01
-1.09736395e+00 -4.40329969e-01 -7.68485785e-01 1.55380398e-01
5.90740860e-01 -6.21648490e-01 1.00139499e+00 -1.17308104e+00
-1.48838305e+00 3.24377805e-01 -5.78294516e-01 -2.96075374e-01
4.92855459e-01 -3.96054029e-01 -6.89616203e-01 7.79834166e-02
-6.37767538e-02 2.17655732e-04 1.42362213e+00 -1.35278487e+00
-4.39922959e-01 -1.85036436e-01 -7.84007087e-03 3.63506824e-01
-2.78502405e-01 1.36139035e-01 -6.04541719e-01 -8.93974185e-01
1.85636654e-01 -6.94025576e-01 -3.21924776e-01 -3.81454974e-01
-3.46826673e-01 1.59712493e-01 8.95461142e-01 -1.07114816e+00
1.50144494e+00 -2.30097556e+00 5.53028822e-01 -2.84258246e-01
6.00675225e-01 4.28821683e-01 -8.69201943e-02 6.12672679e-02
-3.16857994e-01 -5.66467904e-02 -9.33008641e-02 -4.11221713e-01
-4.03859615e-01 9.06695202e-02 -4.19963330e-01 4.92564112e-01
2.35291809e-01 5.93724966e-01 -9.12863016e-01 -3.56323630e-01
2.87188202e-01 7.65654922e-01 -4.72832561e-01 4.10592198e-01
3.43462895e-03 8.58177960e-01 -4.67218399e-01 1.91789806e-01
7.81194329e-01 -4.47826177e-01 -5.15599642e-03 -4.89502549e-01
-4.74380016e-01 2.75188535e-01 -1.28334105e+00 1.61946964e+00
-4.52156484e-01 6.07840240e-01 2.85839140e-01 -5.59306204e-01
8.79149377e-01 4.56330031e-01 2.32253060e-01 -5.97873509e-01
2.31891528e-01 -1.20287642e-01 -1.73925698e-01 -6.82664394e-01
6.59756124e-01 -5.08439615e-02 3.51937979e-01 2.00292319e-01
-1.45432483e-02 4.34126794e-01 -1.58452149e-02 3.59355658e-01
1.13408875e+00 1.38312295e-01 2.51217186e-01 -1.36545956e-01
6.81892276e-01 -4.69116181e-01 9.63947892e-01 3.98072749e-01
-3.10816348e-01 8.15771222e-01 3.90779942e-01 -4.59661961e-01
-1.05598140e+00 -9.63711262e-01 2.87105381e-01 8.43990803e-01
4.71351206e-01 -4.02210832e-01 -6.91930234e-01 -2.15803698e-01
-5.93887031e-01 5.61450303e-01 -5.94871879e-01 -3.15774858e-01
-6.78130329e-01 -7.00417817e-01 -4.62697558e-02 2.41343841e-01
6.94057941e-01 -1.04650486e+00 -5.15308917e-01 2.28213713e-01
-5.10732412e-01 -9.92562830e-01 -6.67573035e-01 -1.98925942e-01
-7.77647913e-01 -7.25836575e-01 -9.96620119e-01 -4.78698760e-01
4.75333512e-01 5.51772594e-01 8.79061460e-01 2.48439893e-01
-2.16974691e-02 -1.70070007e-01 -4.80043322e-01 2.57113814e-01
-4.23353881e-01 -1.62066147e-01 7.03102276e-02 5.72964668e-01
-4.07151133e-02 -6.61288798e-01 -7.30910838e-01 8.58392864e-02
-9.54847813e-01 5.88270664e-01 5.35146177e-01 8.44064772e-01
5.20707309e-01 2.06083283e-01 3.24715227e-01 -7.58966923e-01
4.08634752e-01 -4.99421179e-01 -5.01476467e-01 7.34618083e-02
-2.21079469e-01 1.77847698e-01 7.46095955e-01 -4.95575756e-01
-1.60119462e+00 -5.23036532e-02 -1.69859156e-01 -9.47457969e-01
1.30646080e-02 1.12419076e-01 -2.70619869e-01 3.66984308e-01
4.35595006e-01 5.04581988e-01 -2.22144037e-01 -6.95956469e-01
2.38087177e-01 5.30672967e-01 6.46471441e-01 -1.55414179e-01
8.72579813e-01 5.64419627e-01 -3.75170380e-01 -7.25937307e-01
-9.14009929e-01 -2.20145643e-01 -2.66737282e-01 -2.41824523e-01
1.01937759e+00 -1.25325894e+00 -3.42713833e-01 4.97316003e-01
-1.22624910e+00 -2.66789675e-01 -1.59424454e-01 4.11235422e-01
-3.38377655e-01 4.35173184e-01 -9.44343150e-01 -7.96731293e-01
-3.29701275e-01 -1.19511592e+00 8.23198080e-01 4.73487049e-01
6.60620183e-02 -6.34860814e-01 -3.13729085e-02 1.38073474e-01
4.05501157e-01 1.71863198e-01 4.40940142e-01 2.39544157e-02
-7.43036687e-01 3.16748947e-01 -1.81319267e-01 3.92014712e-01
3.77217263e-01 2.47321855e-02 -1.04961038e+00 -4.87769246e-01
4.32668000e-01 2.43678406e-01 9.57067370e-01 4.09788191e-01
1.09894955e+00 -5.47037721e-01 7.58357067e-03 7.24211514e-01
1.35902417e+00 1.49432212e-01 9.21349525e-01 5.24727181e-02
7.90032089e-01 3.09502184e-01 4.97936487e-01 6.46127105e-01
1.80104598e-01 6.48819029e-01 3.31754804e-01 -3.32649440e-01
-4.68907773e-01 -1.82087377e-01 6.94038808e-01 1.22771513e+00
-4.58499849e-01 -1.66574687e-01 -4.87784445e-01 4.81943280e-01
-1.85276306e+00 -1.22795987e+00 -1.61945358e-01 2.11288285e+00
7.70700514e-01 2.19468266e-01 -2.51683921e-01 8.39104131e-02
1.18058038e+00 5.52616954e-01 -6.76487744e-01 -3.56478407e-03
-2.07103774e-01 3.86433676e-02 2.24291578e-01 5.88931739e-01
-8.03935528e-01 9.71547663e-01 5.34031057e+00 7.47650981e-01
-1.13992524e+00 3.99538070e-01 5.54684877e-01 -2.72629887e-01
-3.88950109e-01 1.71798412e-02 -5.68339825e-01 7.83842683e-01
7.95939147e-01 -2.74849147e-01 7.42916703e-01 3.92660409e-01
8.49599779e-01 1.29331052e-01 -7.07740247e-01 1.03173506e+00
-8.10048282e-02 -1.14986849e+00 2.31592357e-01 -3.00541937e-01
7.45380938e-01 -2.94820517e-01 2.31629640e-01 7.34413788e-02
2.87467808e-01 -7.63920069e-01 7.64874101e-01 6.54786885e-01
7.86630392e-01 -6.90996945e-01 6.35862708e-01 2.87828088e-01
-1.42698610e+00 -1.41430154e-01 -3.91542733e-01 -1.16924733e-01
3.87441218e-01 6.21285975e-01 -7.18094409e-02 7.05911517e-01
9.10366654e-01 1.11429143e+00 -5.05210936e-01 7.03557730e-01
-3.88142794e-01 5.04350066e-01 2.82842636e-01 5.01972139e-01
-8.10969770e-02 -2.17998803e-01 9.16088343e-01 9.87266302e-01
4.54802245e-01 4.65412378e-01 -1.28685385e-01 9.84966040e-01
1.76962763e-01 -1.45115018e-01 -3.07928890e-01 1.95033416e-01
3.02458286e-01 1.27101803e+00 -4.24681723e-01 -8.04618239e-01
-4.70973372e-01 1.41606903e+00 1.99550942e-01 6.91093326e-01
-1.14277244e+00 -2.37039700e-01 9.21831131e-01 5.77127934e-02
3.01023751e-01 -5.61674535e-02 -4.36398722e-02 -1.83073080e+00
-2.91277710e-02 -9.60336208e-01 2.58758277e-01 -1.18077862e+00
-9.30520117e-01 9.35170710e-01 -2.71954358e-01 -1.46010089e+00
1.89331710e-01 1.01557672e-01 -7.19736278e-01 8.90694737e-01
-1.65739584e+00 -6.10824525e-01 -6.04563177e-01 9.40979540e-01
9.87596095e-01 1.71488330e-01 2.97997683e-01 5.38853765e-01
-1.02279890e+00 3.00195098e-01 -8.58707502e-02 -1.30750731e-01
7.12873995e-01 -8.34544182e-01 3.81709993e-01 1.36356783e+00
-1.12999722e-01 6.69738650e-01 1.09970808e+00 -7.00581789e-01
-1.39608359e+00 -1.33685529e+00 4.83708650e-01 -7.14538619e-03
5.79965293e-01 -1.28801778e-01 -1.39096475e+00 8.82290065e-01
3.35000783e-01 8.17114487e-02 -6.14414401e-02 -3.98227066e-01
-2.27789730e-01 -5.14001250e-02 -7.88005888e-01 7.95752347e-01
1.17011309e+00 -4.69174087e-01 -6.36589587e-01 -8.31143409e-02
1.11215973e+00 -4.63509440e-01 -7.33693302e-01 2.98884094e-01
2.14084044e-01 -1.12886238e+00 7.92470157e-01 -3.60975057e-01
5.71436286e-01 -7.12178648e-01 7.13717267e-02 -1.31621611e+00
-8.71036291e-01 -9.34436798e-01 -4.18561161e-01 1.24229264e+00
5.54117225e-02 -6.09299242e-01 4.44452226e-01 3.73308390e-01
-1.50419742e-01 -3.16065520e-01 -7.41376758e-01 -5.99959016e-01
-4.88977283e-01 3.23090702e-02 5.37044466e-01 9.34278369e-01
-3.23623270e-01 2.95856088e-01 -8.72274697e-01 6.00871205e-01
7.66930759e-01 1.35712922e-01 4.24786121e-01 -7.12808251e-01
-3.18423152e-01 -1.15146019e-01 -1.42459080e-01 -1.11499500e+00
-4.26554792e-02 -3.30573559e-01 4.01698984e-02 -1.32051182e+00
4.08876002e-01 -2.37521827e-02 -6.17131174e-01 2.34510124e-01
-5.86501479e-01 2.94698298e-01 3.73628139e-01 3.71150225e-01
-6.92751467e-01 7.86629736e-01 1.45377874e+00 2.91132834e-02
-3.56706649e-01 -9.92337912e-02 -6.75017536e-01 7.86110163e-01
7.81877637e-01 -2.76801974e-01 -2.71085888e-01 -5.76707482e-01
-2.54916959e-02 4.37168777e-01 2.99621493e-01 -9.64721084e-01
5.22149280e-02 -1.48002297e-01 2.55892575e-01 -3.87812525e-01
4.56414968e-01 -6.22770071e-01 6.26266301e-01 5.80449045e-01
-3.76624376e-01 6.05896823e-02 -1.72081999e-02 5.17875314e-01
-2.46499658e-01 1.21784523e-01 1.01897168e+00 -2.35988751e-01
-7.99524426e-01 3.74717444e-01 -4.78856981e-01 -2.25824252e-01
8.00616264e-01 5.22214845e-02 -3.66436899e-01 -5.60854673e-01
-9.87602413e-01 1.05165616e-01 5.16508579e-01 4.38826650e-01
8.05080175e-01 -1.35468006e+00 -9.51911032e-01 1.74103841e-01
-2.68823504e-01 -1.01496495e-01 9.19822276e-01 8.62718344e-01
-1.97014123e-01 -2.35929210e-02 -2.43038163e-01 -4.74644542e-01
-1.39995396e+00 9.07175064e-01 1.53081030e-01 -9.70626250e-02
-9.72427487e-01 5.84384143e-01 5.58829963e-01 5.37516475e-01
5.05941845e-02 -3.62491816e-01 -4.74503130e-01 -5.91142550e-02
1.10707390e+00 5.57693303e-01 -2.60139644e-01 -8.16351175e-01
-8.99238437e-02 4.95126724e-01 -5.36864698e-02 7.18092099e-02
1.28757155e+00 -7.48421192e-01 -8.97271186e-02 4.85060781e-01
9.80609000e-01 3.03633846e-02 -1.69095194e+00 -3.64726096e-01
-3.94451648e-01 -6.60925388e-01 2.53078908e-01 -3.12143505e-01
-1.50836372e+00 7.39198148e-01 7.58348763e-01 2.57582009e-01
1.59375489e+00 -7.87265822e-02 9.43016946e-01 -2.27432191e-01
3.07773083e-01 -6.79298103e-01 7.74252834e-03 2.60620743e-01
1.06427300e+00 -1.05929494e+00 5.55801503e-02 -3.32180053e-01
-7.81939924e-01 7.13691771e-01 5.46679199e-01 -2.94375300e-01
2.99515337e-01 2.26971865e-01 -9.37390402e-02 9.37659591e-02
-1.21582580e+00 -3.35765868e-01 2.48615444e-02 1.88440964e-01
1.57460243e-01 -7.98278823e-02 -1.47301391e-01 6.80022776e-01
7.79337361e-02 9.19958279e-02 8.44144464e-01 5.06708741e-01
-4.51982260e-01 -6.44568086e-01 -5.62501371e-01 4.87550274e-02
-5.19052327e-01 -1.13220960e-01 2.74303019e-01 2.12100074e-01
1.30729139e-01 9.39021289e-01 -5.92295602e-02 -4.53142375e-01
2.35705793e-01 -2.68952489e-01 2.15570420e-01 -4.24490601e-01
-3.52591991e-01 4.39370543e-01 -1.51303813e-01 -1.05862665e+00
-4.25963849e-01 -6.65078819e-01 -1.03993940e+00 -3.76545131e-01
-3.36245656e-01 -2.01674718e-02 1.52868807e-01 5.94084680e-01
3.88258189e-01 1.00365329e+00 7.62235582e-01 -9.10719872e-01
-3.15205485e-01 -8.77158344e-01 -5.87050200e-01 5.34288228e-01
7.78994501e-01 -6.06666446e-01 -5.14591873e-01 4.19330329e-01] | [11.073149681091309, -1.9170470237731934] |
49b6ce64-a128-45fd-b62d-e20efcc4073b | deep-learning-for-clustering-of-multivariate | null | null | https://academic.oup.com/gigascience/article/8/11/giz134/5626377 | https://academic.oup.com/gigascience/article-pdf/8/11/giz134/30797160/giz134.pdf | Deep learning for clustering of multivariate clinical patient trajectories with missing values | Background
Precision medicine requires a stratification of patients by disease presentation that is sufficiently informative to allow for selecting treatments on a per-patient basis. For many diseases, such as neurological disorders, this stratification problem translates into a complex problem of clustering multivariate and relatively short time series because (i) these diseases are multifactorial and not well described by single clinical outcome variables and (ii) disease progression needs to be monitored over time. Additionally, clinical data often additionally are hindered by the presence of many missing values, further complicating any clustering attempts.
Findings
The problem of clustering multivariate short time series with many missing values is generally not well addressed in the literature. In this work, we propose a deep learning–based method to address this issue, variational deep embedding with recurrence (VaDER). VaDER relies on a Gaussian mixture variational autoencoder framework, which is further extended to (i) model multivariate time series and (ii) directly deal with missing values. We validated VaDER by accurately recovering clusters from simulated and benchmark data with known ground truth clustering, while varying the degree of missingness. We then used VaDER to successfully stratify patients with Alzheimer disease and patients with Parkinson disease into subgroups characterized by clinically divergent disease progression profiles. Additional analyses demonstrated that these clinical differences reflected known underlying aspects of Alzheimer disease and Parkinson disease.
Conclusions
We believe our results show that VaDER can be of great value for future efforts in patient stratification, and multivariate time-series clustering in general. | ['Holger Fröhlich', 'Martin Hofmann-Apitius', 'Henri Vrooman', 'Ashar Ahmad', 'Patrice Godard', 'Meemansa Sood', 'Reagon Karki', 'Ping Wu', 'Mohammad Asif Emon', 'Johann de Jong'] | 2019-11-15 | null | null | null | gigascience-2019-11 | ['time-series-clustering'] | ['time-series'] | [ 6.01450875e-02 -3.30319762e-01 -2.73404658e-01 -3.76081407e-01
-8.99361253e-01 -4.23646361e-01 3.93904805e-01 2.51307458e-01
-3.49112242e-01 7.67817974e-01 4.28866208e-01 -1.62822962e-01
-5.80846369e-01 -4.48057681e-01 -2.02303410e-01 -9.33748722e-01
-3.69992286e-01 1.10265410e+00 -3.90487731e-01 2.66619958e-02
-2.05597416e-01 3.63894284e-01 -1.44624150e+00 3.75255078e-01
9.44860399e-01 6.06066644e-01 3.02937120e-01 4.57188040e-01
1.69175640e-01 5.47625244e-01 -5.90914667e-01 4.82855439e-02
-1.47298202e-01 -3.64128470e-01 -3.61739904e-01 2.49189377e-01
-9.16395858e-02 -4.05328155e-01 -1.23365335e-01 7.74266422e-01
6.82364047e-01 -1.79717075e-02 1.07694256e+00 -1.37004733e+00
-4.34533566e-01 4.88482088e-01 -3.77731621e-01 1.23308618e-02
-2.35151604e-01 1.94882780e-01 8.86287689e-01 -3.36714685e-01
5.59136450e-01 1.09735954e+00 7.98250556e-01 8.13728809e-01
-1.75392103e+00 -3.52177650e-01 1.32739156e-01 2.95761317e-01
-1.07476437e+00 -1.79191083e-01 7.06161499e-01 -1.05209959e+00
6.69866502e-01 2.50730872e-01 8.25306475e-01 1.44931114e+00
3.73523593e-01 2.80521125e-01 8.73522639e-01 3.19495320e-01
6.01119399e-01 -2.38927051e-01 3.72117728e-01 1.37430549e-01
2.27730006e-01 1.58854783e-01 3.89679104e-01 -5.07009923e-01
5.04429698e-01 6.34294093e-01 -2.53344238e-01 -1.95081562e-01
-1.60087800e+00 1.16673076e+00 2.37596765e-01 5.01741469e-01
-7.03030765e-01 7.57581517e-02 4.26030129e-01 5.30937836e-02
3.57879281e-01 2.09064841e-01 -5.72221041e-01 3.69962566e-02
-1.28298783e+00 6.84942082e-02 5.95690787e-01 3.90202641e-01
2.51110256e-01 2.64043491e-02 -1.35492399e-01 8.08212757e-01
4.37074095e-01 3.46595526e-01 1.00539577e+00 -1.35729277e+00
1.00877635e-01 6.59364343e-01 1.35317132e-01 -8.12772691e-01
-9.78802919e-01 -4.43249077e-01 -1.29528391e+00 2.83860922e-01
3.93489450e-01 -2.48947397e-01 -8.05601060e-01 2.11655736e+00
1.42866731e-01 1.05070822e-01 2.12731361e-01 7.64088333e-01
5.69642782e-01 3.31150979e-01 1.46336555e-01 -4.12319541e-01
1.53904486e+00 -2.85529047e-01 -8.00830543e-01 -5.47700822e-02
8.66625667e-01 -2.20704690e-01 7.80377448e-01 3.69149566e-01
-7.93979228e-01 -2.23172203e-01 -7.81570673e-01 -4.41325158e-02
-1.29990324e-01 2.29270935e-01 6.12978458e-01 5.20983040e-01
-9.84972179e-01 8.13018978e-01 -1.56200361e+00 -2.80947715e-01
7.12705970e-01 5.39323986e-01 -4.20477271e-01 8.43410939e-02
-1.08107936e+00 7.55089700e-01 1.06400914e-01 2.79017419e-01
-8.73682320e-01 -9.59402561e-01 -6.93884313e-01 -1.01102233e-01
-2.89480388e-01 -1.04385304e+00 8.86052907e-01 -7.32322037e-01
-9.23251271e-01 6.41241610e-01 -3.69955510e-01 -5.19255161e-01
5.46026289e-01 1.41665444e-01 -5.56878388e-01 1.01571083e-01
1.76265940e-01 4.38724101e-01 7.36673772e-01 -9.72491920e-01
-2.68908948e-01 -1.01607156e+00 -5.89135289e-01 -3.16421725e-02
-6.82170391e-02 -2.04870716e-01 2.01481313e-01 -6.37138247e-01
1.77315503e-01 -1.02614105e+00 -6.53306961e-01 6.82222694e-02
-2.58498311e-01 -1.22905694e-01 7.11048603e-01 -8.91728342e-01
1.13569939e+00 -2.09439230e+00 5.29919922e-01 -2.03029260e-01
8.32789242e-01 -5.56498319e-02 9.25549790e-02 3.70499700e-01
-3.18430066e-01 2.84914970e-01 -6.72971487e-01 -5.35584331e-01
-2.01147934e-03 3.75078589e-01 -1.44779226e-02 7.72133410e-01
1.72552690e-01 7.89443791e-01 -7.93855429e-01 -2.97857314e-01
2.88568735e-01 9.16516244e-01 -7.06043661e-01 -5.34288399e-02
-2.38634661e-01 7.18351424e-01 -2.71731913e-01 5.59630454e-01
4.65243369e-01 -6.28066480e-01 3.01048905e-01 -4.58586782e-01
4.97747362e-02 -1.32842183e-01 -9.74707305e-01 1.34766710e+00
-9.61252451e-02 5.87922573e-01 2.23470619e-03 -1.21565318e+00
3.99227828e-01 5.35220265e-01 1.12536657e+00 -1.85383737e-01
3.50829065e-01 2.11670637e-01 3.30305010e-01 -5.94795346e-01
-1.03185214e-01 -4.52397496e-01 1.23906277e-01 3.95801127e-01
-2.50305474e-01 3.24138701e-01 -4.74156765e-03 -6.81011975e-02
1.16789103e+00 -3.54868680e-01 7.62492651e-03 -6.76608905e-02
1.76191375e-01 2.72112906e-01 8.55125427e-01 2.43314058e-01
-6.16529226e-01 8.33625555e-01 6.78526282e-01 -1.73436210e-01
-1.10936177e+00 -1.29739964e+00 -5.09442449e-01 2.78054416e-01
-4.43605632e-01 -3.38593125e-01 -5.87988436e-01 -1.54488847e-01
-3.37122269e-02 6.91058040e-01 -8.51945996e-01 -2.85176963e-01
-3.06834877e-01 -1.47654176e+00 4.14949328e-01 6.85624540e-01
-5.44366427e-02 -7.03758895e-01 -4.44789886e-01 5.56836545e-01
-4.18688893e-01 -8.01338911e-01 -1.57015905e-01 3.44434679e-01
-1.31182659e+00 -1.34216213e+00 -9.84376132e-01 -6.91495955e-01
5.07306516e-01 -2.30613738e-01 9.27195013e-01 -3.39314014e-01
-4.39594835e-01 3.56415719e-01 -6.81863576e-02 -1.27525091e-01
-4.92786050e-01 -2.92383909e-01 3.47039610e-01 6.67337701e-02
4.29791600e-01 -8.55047166e-01 -9.17518675e-01 1.30520359e-01
-1.05299330e+00 -2.88822919e-01 5.44528425e-01 8.95247042e-01
8.61232042e-01 2.76575476e-01 6.05325103e-01 -6.01420105e-01
5.50354481e-01 -8.18823278e-01 -2.45487154e-01 -5.82722239e-02
-5.84346652e-01 1.07167341e-01 4.91465300e-01 -6.74041510e-01
-4.94053274e-01 -4.47436906e-02 -1.71035945e-01 -6.90608025e-01
-3.68364483e-01 7.54415572e-01 -2.35633463e-01 6.86193883e-01
5.21407187e-01 1.27745703e-01 6.54054403e-01 -4.50721085e-01
1.34538129e-01 8.15385282e-01 3.70205969e-01 -1.08542599e-01
1.13455907e-01 9.35814261e-01 6.98686019e-02 -9.37218428e-01
-2.74810910e-01 -2.70566732e-01 -6.85894072e-01 1.45769536e-01
1.35247087e+00 -1.06383562e+00 -7.09490359e-01 4.52632397e-01
-8.13885987e-01 -5.29235005e-01 -1.52960658e-01 8.81038785e-01
-6.50083423e-01 4.20085162e-01 -5.75601876e-01 -5.25876164e-01
-3.03591371e-01 -1.40167642e+00 1.06058824e+00 -3.96325111e-01
-6.17402017e-01 -1.43768096e+00 4.38670516e-01 4.70951170e-01
2.22690180e-01 7.78779745e-01 1.32797587e+00 -7.60285437e-01
-1.14520729e-01 -2.23241299e-01 1.86400592e-01 4.09299344e-01
4.98033971e-01 3.77383940e-02 -6.31092966e-01 -2.79673576e-01
3.58531207e-01 1.12908721e-01 7.73313403e-01 1.01681614e+00
9.12703991e-01 -3.57737169e-02 -4.58060771e-01 6.53511882e-01
1.11166203e+00 2.40570843e-01 5.52943349e-01 1.52486131e-01
7.78549731e-01 7.09801793e-01 -2.16858722e-02 5.33429682e-01
6.38284087e-01 5.85204005e-01 4.43531543e-01 8.84700269e-02
1.34914458e-01 3.88301790e-01 4.25043225e-01 6.56700373e-01
-1.49948746e-02 8.35069269e-02 -8.85751665e-01 7.01381505e-01
-1.86003792e+00 -1.27446198e+00 -5.76426983e-01 2.20733428e+00
6.94871187e-01 -3.85718346e-01 3.42094541e-01 3.01069349e-01
7.99140811e-01 -1.19015522e-01 -8.39733601e-01 1.19265020e-01
-2.64833629e-01 -7.03764856e-02 2.24226102e-01 3.50651592e-01
-9.54461098e-01 3.48815054e-01 6.13063478e+00 2.28178233e-01
-1.16532028e+00 1.78607240e-01 4.80193168e-01 -2.91444451e-01
-2.56604075e-01 -2.31368095e-01 -4.57973808e-01 8.36853206e-01
1.34105015e+00 1.05733044e-01 3.38103414e-01 3.75735044e-01
6.84430659e-01 3.75911504e-01 -1.27887130e+00 1.04186225e+00
-2.96222776e-01 -8.96226287e-01 -1.57606184e-01 3.57012898e-01
4.17275488e-01 3.45972240e-01 2.20105439e-01 1.84969679e-01
3.23048830e-01 -1.25009334e+00 2.17710599e-01 6.44004524e-01
5.39200187e-01 -4.69712079e-01 5.68982184e-01 2.68275052e-01
-6.80196464e-01 -1.40021667e-01 -2.11925626e-01 2.19334379e-01
3.34434003e-01 9.93680835e-01 -6.32110238e-01 2.62394458e-01
5.69874644e-01 1.08368373e+00 -3.52712452e-01 8.79883409e-01
1.63598910e-01 6.95802808e-01 -3.27159315e-01 4.05373454e-01
1.10015804e-02 -3.84736717e-01 5.37458897e-01 5.87980628e-01
5.17083704e-01 -6.80589154e-02 -1.00570165e-01 1.09637940e+00
4.79176521e-01 -3.13351721e-01 -2.49764234e-01 -4.74740475e-01
2.90366739e-01 1.03672576e+00 -6.58292055e-01 -2.41162136e-01
-3.86846662e-01 9.94064808e-01 -6.80855811e-02 3.71396512e-01
-7.64874578e-01 9.06966478e-02 9.43739176e-01 1.53182149e-01
2.45080501e-01 -2.40578204e-01 -2.68948406e-01 -1.49166548e+00
-6.18685633e-02 -7.91889966e-01 6.36461318e-01 -5.21710753e-01
-1.68809354e+00 7.19497859e-01 -2.06654713e-01 -1.39185369e+00
-4.30710554e-01 -5.58442891e-01 -2.90339142e-01 7.86989152e-01
-1.13603210e+00 -9.32040453e-01 -1.55693561e-01 8.87912333e-01
1.82002306e-01 -1.36883184e-01 1.00727928e+00 3.68449837e-01
-8.64358306e-01 1.11338943e-01 6.77011728e-01 -8.28328580e-02
6.46593451e-01 -1.37020469e+00 -2.36250192e-01 1.00825436e-01
-2.41094515e-01 6.68666303e-01 7.21245944e-01 -8.33316803e-01
-9.93169010e-01 -1.41562819e+00 8.71459126e-01 -5.63830018e-01
7.87744522e-01 -1.19242504e-01 -1.02424681e+00 7.73820281e-01
-2.17975393e-01 -1.93419710e-01 1.28195345e+00 2.99235769e-02
-8.44974071e-02 1.13965429e-01 -1.30809712e+00 6.56404257e-01
5.93473077e-01 -4.11638498e-01 -7.78951347e-01 5.03915310e-01
4.56479579e-01 1.25400633e-01 -1.34943664e+00 3.37757111e-01
5.32569945e-01 -7.77863026e-01 1.09500241e+00 -7.61779547e-01
4.61197048e-01 -2.59917706e-01 -4.55772281e-01 -1.37895870e+00
-5.54878652e-01 -1.55340940e-01 -2.92760968e-01 8.69720340e-01
2.64358997e-01 -6.52291119e-01 6.90537035e-01 8.84064436e-01
-4.84650284e-02 -5.86661875e-01 -8.99796903e-01 -7.44231582e-01
5.69633305e-01 -4.79905248e-01 4.34078693e-01 9.88040209e-01
-2.30920166e-01 2.10497051e-01 -1.49525777e-01 3.31387877e-01
8.04713726e-01 4.88649160e-02 1.75011918e-01 -1.75227904e+00
-2.94472754e-01 -5.84100246e-01 -6.42689943e-01 -9.64538753e-02
2.20535383e-01 -1.03677773e+00 -3.37796569e-01 -1.77818429e+00
1.57688022e-01 -2.35031635e-01 -3.19762230e-01 2.24523351e-01
-1.62829652e-01 -3.52372155e-02 -2.58846402e-01 5.77536225e-01
-7.07336068e-02 7.49580979e-01 8.79212737e-01 -3.56311798e-01
-4.81517762e-01 1.91937938e-01 -8.25664878e-01 5.68573833e-01
1.02498043e+00 -4.71517026e-01 -5.80807686e-01 -2.13596866e-01
-3.42065319e-02 1.96645811e-01 7.55244911e-01 -1.03010952e+00
-1.10730797e-01 -9.70247984e-02 5.40900826e-01 -6.60610199e-01
4.95209873e-01 -1.05113864e+00 4.59596992e-01 6.05288088e-01
-1.61288977e-01 1.42558128e-01 2.79816594e-02 7.98574507e-01
-4.04989719e-02 2.86929514e-02 6.93364978e-01 -3.21590751e-02
-4.12148416e-01 4.79198366e-01 -7.41764545e-01 5.22370711e-02
9.91046131e-01 -1.11262882e-02 -1.30465403e-01 -3.34438682e-01
-1.54995310e+00 3.09854329e-01 4.34205443e-01 1.31434232e-01
4.81561124e-01 -1.36843717e+00 -7.63003349e-01 3.77009250e-02
-5.80905424e-03 -1.21190161e-01 6.14804864e-01 1.40192914e+00
-2.60624290e-01 1.50579348e-01 -8.08427706e-02 -9.69111085e-01
-1.11284542e+00 8.20382059e-01 4.08812732e-01 -1.66093543e-01
-8.10912251e-01 1.31752893e-01 3.88343893e-02 -4.41776663e-01
2.07897380e-01 -5.51351905e-01 -4.02791053e-01 6.48680151e-01
5.84562480e-01 3.80926073e-01 4.19106483e-02 -7.57458150e-01
-3.83715391e-01 4.28102165e-01 6.35463595e-02 -2.15356544e-01
1.76887035e+00 -1.23076878e-01 -3.35474491e-01 7.61714101e-01
1.40636504e+00 -4.82316136e-01 -1.12367475e+00 2.30940238e-01
-6.35427609e-02 2.79405028e-01 2.16820329e-01 -6.88294411e-01
-1.19385219e+00 9.33934271e-01 9.35125172e-01 6.56818673e-02
1.03059649e+00 9.63944495e-02 7.56048679e-01 1.88315615e-01
1.99186290e-03 -7.60514796e-01 -3.13966841e-01 -5.69019793e-03
7.57897377e-01 -1.27149010e+00 -2.13615462e-01 2.18138546e-01
-7.16884673e-01 9.74879384e-01 -7.24426135e-02 -1.78420991e-01
1.04289436e+00 8.49281698e-02 2.55755633e-01 -4.15915996e-01
-7.89563239e-01 -1.35512948e-01 1.49830222e-01 8.17946494e-01
3.74073744e-01 3.18301380e-01 -2.71149546e-01 1.21419442e+00
-2.18952581e-01 1.85616255e-01 4.34376627e-01 7.60499716e-01
-1.41787872e-01 -1.15229404e+00 -5.41831613e-01 8.64922404e-01
-3.97954196e-01 3.91233154e-02 -1.84859514e-01 6.48024023e-01
6.69501722e-02 1.00136817e+00 2.86663294e-01 -3.18191111e-01
7.35780001e-02 1.67152181e-01 2.52550781e-01 -4.88130927e-01
-2.91240960e-01 3.37162435e-01 -2.65540630e-01 -5.16153753e-01
-5.31139970e-01 -1.15787554e+00 -1.33842206e+00 -9.83681306e-02
8.31569955e-02 -9.01729688e-02 5.68079293e-01 8.67925048e-01
6.36295795e-01 9.23297942e-01 3.21363717e-01 -9.18386161e-01
-6.02265000e-01 -9.34861422e-01 -7.93914676e-01 5.23403585e-01
8.12659085e-01 -7.28902221e-01 -7.20153153e-01 2.53367871e-01] | [7.228837013244629, 5.38394832611084] |
a4fd702c-8c26-4388-bfc0-45b232fab9f0 | expressive-variable-and-controllable-duration | 2206.14165 | null | https://arxiv.org/abs/2206.14165v1 | https://arxiv.org/pdf/2206.14165v1.pdf | Expressive, Variable, and Controllable Duration Modelling in TTS | Duration modelling has become an important research problem once more with the rise of non-attention neural text-to-speech systems. The current approaches largely fall back to relying on previous statistical parametric speech synthesis technology for duration prediction, which poorly models the expressiveness and variability in speech. In this paper, we propose two alternate approaches to improve duration modelling. First, we propose a duration model conditioned on phrasing that improves the predicted durations and provides better modelling of pauses. We show that the duration model conditioned on phrasing improves the naturalness of speech over our baseline duration model. Second, we also propose a multi-speaker duration model called Cauliflow, that uses normalising flows to predict durations that better match the complex target duration distribution. Cauliflow performs on par with our other proposed duration model in terms of naturalness, whilst providing variable durations for the same prompt and variable levels of expressiveness. Lastly, we propose to condition Cauliflow on parameters that provide an intuitive control of the pacing and pausing in the synthesised speech in a novel way. | ['Thomas Drugman', 'Elia Gatti', 'Simon Slangen', 'Ewa Muszynska', 'Sri Karlapati', 'Alexis Moinet', 'Thomas Merritt', 'Ammar Abbas'] | 2022-06-28 | null | null | null | null | ['normalising-flows'] | ['methodology'] | [-1.99421169e-03 3.02636445e-01 -3.10400188e-01 -4.04280484e-01
-5.60550332e-01 -6.77285790e-01 7.90202856e-01 -6.80956291e-03
-2.39062294e-01 5.08544505e-01 6.36002958e-01 -4.22034353e-01
5.37601560e-02 -3.47275645e-01 -1.86577141e-01 -5.18077135e-01
6.66016191e-02 3.86939824e-01 2.71853179e-01 -1.98160186e-01
1.11330837e-01 4.38470870e-01 -1.67272997e+00 1.61971524e-01
7.02976942e-01 8.70878637e-01 3.63960087e-01 1.16553426e+00
-3.49115133e-02 6.59705997e-01 -9.34625030e-01 1.72541544e-01
-2.31062815e-01 -6.38455749e-01 -5.63458323e-01 -1.83940634e-01
-2.59835236e-02 -1.85233504e-01 -4.48681712e-01 3.12521517e-01
6.03285134e-01 4.57721591e-01 6.38395250e-01 -1.25095785e+00
-4.15427059e-01 1.28071630e+00 -7.22019607e-03 4.15769815e-01
3.79783422e-01 1.77744865e-01 9.93501902e-01 -1.46444961e-01
4.40960199e-01 1.30309677e+00 4.79045331e-01 8.54761541e-01
-1.38585401e+00 -3.64889264e-01 3.33109260e-01 1.90623745e-01
-1.19283009e+00 -7.89692581e-01 9.62592661e-01 -3.27066422e-01
1.24192560e+00 5.50355554e-01 6.47840917e-01 1.41306078e+00
-5.27853370e-02 7.56287396e-01 8.61064613e-01 -6.41635478e-01
2.95919865e-01 8.54653716e-02 -4.34999205e-02 9.42545850e-03
-8.62209141e-01 6.02206767e-01 -5.13366461e-01 1.09178282e-01
5.20848215e-01 -5.21303535e-01 -3.35757494e-01 3.24855417e-01
-1.15421319e+00 7.41934717e-01 -1.74366251e-01 6.39043033e-01
-2.87726104e-01 1.96421117e-01 5.81816852e-01 2.06995130e-01
6.78080499e-01 6.18629396e-01 -6.54056072e-01 -6.98853791e-01
-1.46614683e+00 4.59216237e-01 1.05453074e+00 9.49053049e-01
2.47750163e-01 3.58136922e-01 -8.40316832e-01 1.00732195e+00
1.36511579e-01 4.82635647e-01 6.94574118e-01 -1.07628441e+00
1.47609219e-01 -9.95740518e-02 1.66369826e-01 -4.29308861e-01
-7.47559607e-01 -3.78628731e-01 -3.77210975e-01 -1.79286778e-01
5.13121843e-01 -1.66060433e-01 -9.28987741e-01 2.19467402e+00
-3.73087563e-02 -4.77820821e-02 -1.68983400e-01 6.40024304e-01
4.81120348e-01 9.01698887e-01 1.83100253e-01 -5.56829095e-01
1.21604872e+00 -1.08025324e+00 -1.26292324e+00 -1.55668572e-01
5.38421869e-01 -6.80686653e-01 1.43814373e+00 4.76056039e-01
-1.22882164e+00 -5.89605570e-01 -8.62272620e-01 4.03751917e-02
-2.69314975e-01 1.07758917e-01 4.11145031e-01 8.01849961e-01
-1.11673045e+00 1.02112293e+00 -8.55291605e-01 -1.19877666e-01
-2.36582160e-01 2.04069942e-01 2.58114398e-01 6.40457749e-01
-1.29986596e+00 1.04438162e+00 3.16483527e-01 -2.01639906e-01
-7.21907139e-01 -7.25727797e-01 -9.48979437e-01 2.53827721e-01
2.69773781e-01 -4.11376923e-01 1.89482987e+00 -9.58375037e-01
-2.20117402e+00 4.44470704e-01 -1.92135975e-01 -4.63666350e-01
3.51681828e-01 -2.95411021e-01 -5.27889609e-01 -2.40808185e-02
-4.63055879e-01 8.15228999e-01 7.07796037e-01 -9.50917840e-01
-1.95521668e-01 -8.76565278e-03 -3.05304945e-01 4.78391461e-02
-1.88322052e-01 1.13938853e-01 -3.11161995e-01 -8.07022691e-01
-4.14189667e-01 -1.09062219e+00 -8.38086829e-02 -3.33190560e-01
-4.24710006e-01 -4.74732041e-01 7.52933681e-01 -6.96419537e-01
1.98622000e+00 -1.95939994e+00 4.25038189e-01 -1.86676845e-01
-3.95239815e-02 3.23533267e-01 -3.78593057e-02 6.51560962e-01
-5.83728515e-02 2.10694388e-01 -2.84398645e-01 -1.04518604e+00
3.99866670e-01 3.05576801e-01 -3.33162874e-01 1.23127140e-01
1.13586605e-01 8.40868831e-01 -6.45293355e-01 -4.23851460e-01
2.80455112e-01 5.96631527e-01 -5.60838342e-01 5.98270953e-01
-6.21671617e-01 8.13283503e-01 2.62618661e-01 1.29486248e-01
3.25803131e-01 2.74254531e-01 5.97604439e-02 1.05104208e-01
-5.85838377e-01 6.49061024e-01 -7.98619032e-01 1.70517862e+00
-8.87367189e-01 7.22581565e-01 -6.91978782e-02 -6.54089212e-01
9.13084030e-01 6.16173327e-01 2.29565904e-01 -7.52348244e-01
4.06362146e-01 1.04505509e-01 3.65025431e-01 -4.90374237e-01
5.49668550e-01 -2.54602730e-01 -2.48573691e-01 4.30984706e-01
3.27427052e-02 -5.98290026e-01 1.63165525e-01 -1.92979455e-01
7.70046413e-01 4.09739643e-01 2.36192212e-01 -1.20892413e-01
3.83634567e-01 -6.87911391e-01 3.18313211e-01 5.69956183e-01
-4.19416398e-01 6.91285908e-01 6.15515351e-01 -4.86812601e-03
-1.19614482e+00 -7.82828450e-01 9.96767655e-02 1.47510982e+00
-4.49198276e-01 -3.96656692e-01 -1.12588823e+00 -1.77122533e-01
-4.01672781e-01 1.26521695e+00 -5.56191146e-01 -1.48004830e-01
-6.83404803e-01 -1.04191944e-01 9.45263505e-01 5.66924572e-01
-1.36171266e-01 -1.71879292e+00 -6.54505372e-01 4.67542559e-01
-4.64128584e-01 -1.22810400e+00 -7.96089292e-01 4.63852733e-01
-6.68691635e-01 -2.22955391e-01 -7.76528001e-01 -4.60702896e-01
-3.83084893e-01 -5.08962631e-01 1.00010598e+00 -1.29971519e-01
1.34759247e-01 2.05733720e-02 -5.05853176e-01 -5.55663049e-01
-8.70558977e-01 4.97041523e-01 7.70405084e-02 -2.15419620e-01
5.00878617e-02 -9.41641450e-01 -4.41214263e-01 7.75166154e-02
-9.69808221e-01 1.65326640e-01 2.56772727e-01 4.08450544e-01
1.79601550e-01 -4.54010069e-01 1.04147768e+00 -6.44813478e-01
7.50543356e-01 -4.77784395e-01 -2.26211295e-01 -7.24493042e-02
-7.59550810e-01 2.45369211e-01 9.82623696e-01 -1.07535708e+00
-1.00396049e+00 -1.46467283e-01 -6.63413465e-01 -6.37219250e-01
-4.76929635e-01 2.82949239e-01 -2.94112325e-01 7.02950418e-01
6.14969671e-01 1.83103532e-01 9.38193500e-02 -4.76619482e-01
6.59354627e-01 8.53992522e-01 4.30515736e-01 -5.08726060e-01
1.97316483e-01 -7.48965442e-02 -4.73791689e-01 -1.07929504e+00
-5.40888667e-01 -3.39008987e-01 -4.93648559e-01 -1.39287010e-01
6.07037902e-01 -3.89829278e-01 -7.40074098e-01 6.11085534e-01
-1.29848921e+00 -9.86465931e-01 -3.90561521e-01 4.59398806e-01
-1.20981336e+00 4.11143661e-01 -8.43428850e-01 -1.20949352e+00
-3.53102654e-01 -1.07460940e+00 1.07390451e+00 1.88985214e-01
-6.54000878e-01 -1.13479388e+00 2.35324427e-01 -1.99192569e-01
8.62640977e-01 1.72253594e-01 9.09417987e-01 -6.44090831e-01
3.05978283e-02 2.46993303e-01 2.09452584e-01 3.64469647e-01
-7.07752910e-03 2.12539330e-01 -1.24685836e+00 1.55378580e-01
-8.06350857e-02 4.00061645e-02 5.22495687e-01 7.98638463e-01
1.15465593e+00 -5.06903291e-01 -1.02160238e-01 5.61180830e-01
8.63785982e-01 5.04104912e-01 6.76159978e-01 4.42708917e-02
3.45204264e-01 8.55323017e-01 4.27011013e-01 7.34661877e-01
4.23062831e-01 9.93142128e-01 5.27035594e-02 1.65873561e-02
-4.04049695e-01 -2.37392575e-01 3.83848220e-01 1.03861213e+00
1.02343537e-01 -6.74137414e-01 -6.77275181e-01 7.87012577e-01
-1.73153961e+00 -1.00397122e+00 -2.00656764e-02 2.08095384e+00
1.18003869e+00 3.70181143e-01 7.21390843e-01 4.90790963e-01
4.85507637e-01 4.01265711e-01 -3.22410971e-01 -1.17239845e+00
6.70918748e-02 3.91468078e-01 1.91063076e-01 7.90194035e-01
-7.42460370e-01 1.15484369e+00 6.68692541e+00 7.94337511e-01
-1.25559080e+00 1.05898939e-02 3.56253892e-01 -1.95714757e-01
-5.34682155e-01 2.74283476e-02 -7.38650858e-01 6.93616688e-01
1.89537668e+00 -2.25756273e-01 7.95172393e-01 5.72490633e-01
8.75121772e-01 -5.36399409e-02 -1.36118615e+00 7.75487363e-01
1.34575795e-02 -9.26362455e-01 -3.85463566e-01 -6.69977516e-02
2.66997963e-01 -4.08103555e-01 2.36313492e-02 6.84472859e-01
-1.41515464e-01 -1.17323577e+00 1.19972515e+00 7.35896647e-01
9.22714293e-01 -8.33863854e-01 3.74610960e-01 7.35538661e-01
-1.10897720e+00 -8.83467942e-02 1.97296038e-01 -2.64375448e-01
6.92832470e-01 3.87452871e-01 -8.82192373e-01 1.58826753e-01
1.32380813e-01 1.43920615e-01 -2.45080322e-01 9.64217365e-01
-2.70511478e-01 9.96849477e-01 -3.34808707e-01 -1.44379050e-01
6.42130375e-02 2.48500273e-01 5.36715746e-01 1.66389847e+00
2.71922469e-01 -2.80388594e-02 9.53063816e-02 9.72328782e-01
3.40724915e-01 1.12980582e-01 -5.08410692e-01 -3.95101726e-01
8.49268436e-01 8.47340286e-01 -5.56761801e-01 -1.87477037e-01
-2.18342543e-02 9.69815791e-01 1.38249621e-01 3.40327412e-01
-1.23556530e+00 -3.52247119e-01 4.34765697e-01 -5.73581830e-03
3.36783409e-01 -4.19185251e-01 -3.22491884e-01 -6.29218638e-01
-2.27260470e-01 -8.34058046e-01 -1.01423077e-01 -6.82988346e-01
-7.40244746e-01 9.55716848e-01 2.20165625e-01 -7.69490302e-01
-6.40619636e-01 -1.43741846e-01 -6.80523813e-01 1.04417431e+00
-1.42295718e+00 -1.04630780e+00 1.24681555e-01 1.64042041e-01
1.01482582e+00 3.67963284e-01 1.01212966e+00 1.56420156e-01
-5.15090585e-01 7.18266845e-01 -3.49906474e-01 -4.68038529e-01
7.45057225e-01 -1.31708181e+00 6.00318670e-01 5.74895978e-01
-8.87869000e-02 4.52089190e-01 1.20100284e+00 -2.90807903e-01
-5.59523284e-01 -8.46171796e-01 1.22488713e+00 -4.27233964e-01
4.34294313e-01 -4.54969943e-01 -9.50475812e-01 4.65886801e-01
4.30289477e-01 -5.78969121e-01 6.91419125e-01 1.67916805e-01
-2.36140192e-01 2.16072246e-01 -7.01044559e-01 6.09259605e-01
7.69725442e-01 -6.17152870e-01 -7.02777267e-01 -8.14101566e-03
1.34675920e+00 -3.14685196e-01 -7.86249638e-01 1.79693792e-02
6.58991039e-01 -1.01774204e+00 5.93172729e-01 -2.74968237e-01
1.90821767e-01 -4.54460457e-03 -1.82002753e-01 -1.60590875e+00
-1.60442129e-01 -1.02977693e+00 -5.36346912e-01 1.41106498e+00
4.13188130e-01 -2.53413290e-01 4.67805892e-01 4.96324867e-01
-4.77763504e-01 -5.97180963e-01 -8.48494828e-01 -1.05880356e+00
1.97699547e-01 -7.72465765e-01 5.37658393e-01 4.76300180e-01
3.53893161e-01 5.00207901e-01 -5.76696396e-01 -1.38471887e-01
-2.38060132e-02 -1.27596363e-01 3.91138852e-01 -1.08515990e+00
-5.43853343e-01 -9.05021071e-01 2.05830857e-01 -1.21612084e+00
3.75455558e-01 -4.86724705e-01 5.08936763e-01 -1.25045538e+00
-2.89166957e-01 -4.13273156e-01 3.51975821e-02 3.36419672e-01
-5.43920957e-02 -2.76061565e-01 2.96568781e-01 -1.64737150e-01
-1.30264670e-01 8.39307725e-01 1.09624648e+00 3.84358346e-01
-7.50119090e-01 3.38179380e-01 -3.28742862e-01 3.57514769e-01
9.01113629e-01 -3.85163486e-01 -6.45606935e-01 -4.64108028e-02
-2.35925332e-01 7.34301805e-01 -1.47730500e-01 -9.94777679e-01
9.08481479e-02 -3.25574517e-01 -1.54122725e-01 -4.77918297e-01
5.71184993e-01 -3.63099933e-01 1.01792268e-01 1.46030769e-01
-6.88407123e-01 3.59220468e-02 5.15616119e-01 2.75918096e-01
5.77445887e-02 -1.97942615e-01 8.00798595e-01 3.50432731e-02
-2.83830434e-01 9.15994942e-02 -8.56646001e-01 5.47012053e-02
5.98975122e-01 -1.91986442e-01 -1.57519326e-01 -6.91611469e-01
-1.17173445e+00 -3.76113653e-02 2.90228635e-01 6.81456685e-01
3.25875759e-01 -9.12654459e-01 -4.59136337e-01 1.64569676e-01
-8.27657350e-04 -4.04383391e-01 4.56570089e-01 7.95542598e-01
-4.04713713e-02 7.59779990e-01 1.42576143e-01 -4.59174037e-01
-1.07067025e+00 7.62951970e-01 3.46689016e-01 -3.02504748e-01
-4.22747105e-01 7.01812625e-01 2.71411519e-02 -4.12294894e-01
5.17668068e-01 -7.41010249e-01 -3.05563152e-01 1.95432931e-01
5.88735759e-01 3.79007965e-01 -6.19221199e-03 -5.38722217e-01
-2.58551598e-01 1.18388526e-01 1.85262516e-01 -7.09897280e-01
1.04090190e+00 -4.38091964e-01 3.67334574e-01 1.07794535e+00
1.00555301e+00 1.76280349e-01 -1.54960656e+00 4.97875400e-02
2.56021023e-01 9.29308906e-02 1.18767798e-01 -1.16805518e+00
-5.18027484e-01 9.62911606e-01 2.82895565e-01 4.80889231e-01
1.12593460e+00 -6.72730207e-02 9.46301579e-01 -2.62321323e-01
-6.99344426e-02 -1.20728934e+00 1.26413271e-01 7.76190341e-01
9.33106065e-01 -7.55172491e-01 -6.24546349e-01 -8.89182836e-02
-9.39653575e-01 1.16092944e+00 4.62844223e-01 2.89036840e-01
4.69415218e-01 4.99282897e-01 2.72283912e-01 1.89993918e-01
-1.09530520e+00 -4.58302617e-01 1.23967394e-01 7.86699593e-01
7.07996190e-01 3.60899270e-01 -5.63871682e-01 7.46003747e-01
-6.72450125e-01 -1.25770360e-01 6.90236211e-01 5.73822081e-01
-5.55656552e-01 -1.31972313e+00 -3.01642865e-01 -1.01886317e-01
-4.98622954e-01 -3.43174100e-01 -1.51355401e-01 5.76926529e-01
-4.75938283e-02 1.39396429e+00 1.55198038e-01 -3.79669249e-01
4.45208549e-01 5.47154486e-01 4.22056735e-01 -7.53307521e-01
-6.66063607e-01 5.68097115e-01 2.20934272e-01 -4.97963458e-01
-1.23740733e-01 -6.07004941e-01 -1.33338237e+00 -9.55385417e-02
-4.89631325e-01 1.88178390e-01 9.40488756e-01 9.10693586e-01
1.58275306e-01 9.34152722e-01 7.27501690e-01 -9.98622596e-01
-5.92158973e-01 -1.51466477e+00 -6.57100081e-01 5.57224192e-02
6.28546655e-01 -6.36123240e-01 -6.97927237e-01 1.01189025e-01] | [14.909103393554688, 6.57694149017334] |
307cf0b0-35fc-443f-bf64-c61aa7f5a04f | autism-spectrum-disorder-classification-in | 2307.00976 | null | https://arxiv.org/abs/2307.00976v1 | https://arxiv.org/pdf/2307.00976v1.pdf | Autism Spectrum Disorder Classification in Children based on Structural MRI Features Extracted using Contrastive Variational Autoencoder | Autism spectrum disorder (ASD) is a highly disabling mental disease that brings significant impairments of social interaction ability to the patients, making early screening and intervention of ASD critical. With the development of the machine learning and neuroimaging technology, extensive research has been conducted on machine classification of ASD based on structural MRI (s-MRI). However, most studies involve with datasets where participants' age are above 5. Few studies conduct machine classification of ASD for participants below 5-year-old, but, with mediocre predictive accuracy. In this paper, we push the boundary of predictive accuracy (above 0.97) of machine classification of ASD in children (age range: 0.92-4.83 years), based on s-MRI features extracted using contrastive variational autoencoder (CVAE). 78 s-MRI, collected from Shenzhen Children's Hospital, are used for training CVAE, which consists of both ASD-specific feature channel and common shared feature channel. The ASD participants represented by ASD-specific features can be easily discriminated from TC participants represented by the common shared features, leading to high classification accuracy. In case of degraded predictive accuracy when data size is extremely small, a transfer learning strategy is proposed here as a potential solution. Finally, we conduct neuroanatomical interpretation based on the correlation between s-MRI features extracted from CVAE and surface area of different cortical regions, which discloses potential biomarkers that could help target treatments of ASD in the future. | ['Yi Pan', 'Wenhui Xi', 'Yanjie Wei', 'Jintao Meng', 'Yanlin Wang', 'Ruitao Xie', 'Ruimin Ma'] | 2023-07-03 | null | null | null | null | ['classification-1', 'transfer-learning'] | ['methodology', 'miscellaneous'] | [ 1.40080135e-02 1.56163782e-01 6.16399571e-02 -3.53929788e-01
-1.64502844e-01 5.30858636e-02 5.85329421e-02 2.90900975e-01
-2.83300906e-01 4.76450890e-01 1.36708289e-01 1.26128271e-01
-3.49325418e-01 -5.47630608e-01 -2.05888405e-01 -5.20912290e-01
-2.91913867e-01 5.36988080e-01 -4.34340090e-02 9.59374309e-02
4.78331447e-02 1.44822836e-01 -1.50468671e+00 2.42080793e-01
1.42665088e+00 7.96850860e-01 5.09692371e-01 1.57202229e-01
2.29559466e-02 2.12926745e-01 -3.98438454e-01 -5.26780002e-02
-3.53742987e-02 -4.46991891e-01 -6.99171364e-01 -1.33168742e-01
-9.56448540e-02 -7.28422880e-01 -8.29681084e-02 1.19122088e+00
5.88044226e-01 -1.24041043e-01 1.05220199e+00 -1.21805263e+00
-1.05806017e+00 5.90110540e-01 -7.51343548e-01 4.65773821e-01
3.35985005e-01 1.40976876e-01 6.91099942e-01 -8.57190788e-01
1.38872504e-01 1.04914904e+00 4.68892276e-01 9.62007463e-01
-9.45357859e-01 -1.00966704e+00 -2.47927904e-02 4.86752987e-01
-1.15787375e+00 -4.74950261e-02 5.05790889e-01 -1.23946762e+00
9.92668331e-01 -2.06591994e-01 1.30029750e+00 1.03900671e+00
1.00752927e-01 3.11684668e-01 1.05440128e+00 -5.05072810e-02
1.16129190e-01 -2.78124720e-01 -2.69686151e-02 5.41788518e-01
3.40511084e-01 -8.33674520e-02 -2.62267500e-01 -1.56921461e-01
6.88808739e-01 1.42250463e-01 -2.36964166e-01 -1.25957956e-03
-1.14200652e+00 9.30976331e-01 3.75354856e-01 4.96888518e-01
-3.96355718e-01 -4.13151771e-01 5.53651035e-01 2.54649043e-01
6.36808097e-01 1.88974679e-01 -6.25962019e-01 6.45102412e-02
-7.31122971e-01 5.31031610e-03 -1.15134142e-01 3.10173780e-01
1.91889554e-01 1.27344623e-01 -4.00223248e-02 1.32011044e+00
7.16040730e-01 3.91870558e-01 9.69374597e-01 -6.74649656e-01
3.11863452e-01 9.56014276e-01 -5.65758407e-01 -7.91198850e-01
-6.34517670e-01 1.39499186e-02 -9.14020896e-01 1.55388787e-01
-4.42708135e-02 -3.64336103e-01 -5.61161697e-01 1.91607141e+00
5.31425834e-01 1.87363103e-01 -4.49195206e-02 9.07370389e-01
9.79718626e-01 1.32864580e-01 1.71142548e-01 -4.65210259e-01
1.37481737e+00 -5.02941132e-01 -5.09736300e-01 -1.02047026e-01
6.74910128e-01 -1.37091652e-01 8.91802251e-01 3.89582634e-01
-9.81445789e-01 -1.70774728e-01 -9.00253594e-01 3.09694916e-01
-1.43815324e-01 2.47692287e-01 6.88908160e-01 4.97851193e-01
-1.06415844e+00 5.75789332e-01 -1.03151309e+00 -4.88009214e-01
7.41480589e-01 7.61167943e-01 -6.77679837e-01 2.26642460e-01
-1.12462008e+00 8.59585285e-01 2.38620758e-01 -1.34038910e-01
-7.33520389e-01 -9.62928116e-01 -7.82307446e-01 -1.31237254e-01
-3.22926491e-01 -6.98635995e-01 9.08089936e-01 -9.38858449e-01
-1.22305822e+00 1.13860917e+00 2.62492090e-01 -2.09956035e-01
1.31006449e-01 -1.49735451e-01 -4.91104126e-01 4.54827219e-01
3.48349422e-01 7.19579279e-01 9.75855768e-01 -4.56579208e-01
-4.88388062e-01 -1.17973042e+00 -2.99798876e-01 2.24296376e-02
-6.79956675e-01 5.36029458e-01 7.39828050e-01 -6.63876653e-01
3.93916160e-01 -7.88446665e-01 6.61656484e-02 2.69855738e-01
-2.57995903e-01 -5.68454146e-01 8.13127756e-01 -1.08858502e+00
9.47096050e-01 -2.14703441e+00 2.45998964e-01 -1.42787263e-01
7.25654662e-01 4.16629642e-01 1.44197777e-01 5.10780588e-02
-5.40293634e-01 1.91729039e-01 -3.93822461e-01 6.19470328e-02
-2.94297427e-01 -3.89727533e-01 4.29284811e-01 6.94826066e-01
4.63353395e-01 5.19728839e-01 -8.13278973e-01 -2.28302434e-01
4.53030653e-02 2.86432445e-01 -8.07289302e-01 4.36172724e-01
6.30128235e-02 9.98972535e-01 -6.98132932e-01 5.98659515e-01
6.12405360e-01 -7.12341741e-02 -7.88276196e-02 1.83917016e-01
1.50794461e-02 -5.16127646e-02 -5.67800939e-01 1.47060394e+00
5.77805005e-02 4.66453910e-01 2.56553829e-01 -1.32107711e+00
8.74116600e-01 5.07887006e-01 7.76764333e-01 -3.86676490e-01
4.38645005e-01 1.94066137e-01 7.39224613e-01 -9.09502447e-01
-7.02945054e-01 1.05315726e-02 4.81745332e-01 4.13998961e-01
-3.96299623e-02 8.73729140e-02 -2.92109132e-01 -2.09787756e-01
1.04806554e+00 -3.19364369e-01 2.80297786e-01 -2.08311707e-01
4.94135767e-01 -5.32358408e-01 7.08752692e-01 -7.39778653e-02
-4.75934654e-01 4.46855098e-01 4.71899480e-01 -6.88061863e-02
-9.64119494e-01 -8.92403901e-01 -5.22424638e-01 5.74089289e-01
-5.17967522e-01 -3.57759222e-02 -1.10829210e+00 -4.29904372e-01
3.84496525e-04 2.63028383e-01 -5.63264847e-01 -5.23244500e-01
-1.34615064e-01 -8.28916848e-01 6.55063808e-01 6.85034871e-01
6.06522441e-01 -1.00848293e+00 -5.89713871e-01 -6.09312765e-02
3.45030963e-01 -9.15231168e-01 -9.83689278e-02 -4.89342809e-01
-9.28805947e-01 -1.38925600e+00 -1.07995379e+00 -9.28959131e-01
7.15855062e-01 -1.27056122e-01 2.61861384e-01 1.54409781e-01
-4.60180163e-01 5.17804265e-01 -5.23823142e-01 -6.41392291e-01
-2.75659561e-01 -4.50987548e-01 6.81233406e-01 1.19654331e-02
4.11572248e-01 -9.91241574e-01 -8.52445066e-01 -1.90768301e-01
-3.98538053e-01 9.30989608e-02 4.34504539e-01 7.76625752e-01
3.58646542e-01 -1.67822585e-01 9.93351340e-01 -2.28643000e-01
4.76695269e-01 -9.28683579e-01 -1.57887667e-01 -1.09920122e-01
-7.18796492e-01 -5.53411126e-01 2.68173069e-01 -5.79003990e-01
-7.22066700e-01 -3.74623358e-01 -3.39077830e-01 -6.18193746e-01
-5.95572591e-01 5.21852136e-01 -1.07609555e-01 1.90728545e-01
2.32243463e-01 -6.72467947e-02 4.84393597e-01 -4.12606090e-01
-3.89229894e-01 1.06490290e+00 -1.21167973e-01 -2.98191309e-01
2.48755999e-02 3.23297679e-02 -1.09123863e-01 -8.68742049e-01
-5.30722916e-01 -3.35775800e-02 -5.20408273e-01 -2.20116526e-01
1.36701035e+00 -8.74064982e-01 -5.93746781e-01 9.07220900e-01
-9.31604266e-01 -2.13529825e-01 2.55396724e-01 1.10443258e+00
-4.14261580e-01 1.90364972e-01 -6.11406326e-01 -5.83613515e-01
-8.66838992e-01 -1.22216678e+00 6.93725884e-01 1.81737781e-01
-4.99523908e-01 -7.49711752e-01 2.73723096e-01 6.92236602e-01
-1.10436231e-02 3.28461617e-01 1.39955902e+00 -1.17848778e+00
2.78595060e-01 9.75456461e-02 -3.19869250e-01 5.99693537e-01
6.03062585e-02 1.97053984e-01 -9.06818748e-01 -4.29710627e-01
7.65444934e-02 -2.90379822e-01 4.12338793e-01 6.35418832e-01
1.37061107e+00 6.03619888e-02 -3.31464261e-01 3.24224651e-01
8.85012090e-01 6.37645543e-01 3.08266580e-01 -7.22321868e-02
7.74272919e-01 8.76194417e-01 4.06399399e-01 6.35908604e-01
6.15892828e-01 3.33208084e-01 4.81693447e-01 3.20103467e-01
-3.18338834e-02 1.08353935e-01 2.61447430e-01 1.14670503e+00
-2.72800982e-01 2.98845947e-01 -1.22813272e+00 7.28433371e-01
-1.46175826e+00 -7.45935857e-01 -4.74168599e-01 2.08379006e+00
5.00534356e-01 -3.13429445e-01 3.58677506e-01 2.80326992e-01
1.09013796e+00 -3.18670332e-01 -8.79755378e-01 -4.13477808e-01
8.68054479e-02 3.55220705e-01 -4.29302394e-01 -2.99301594e-01
-6.69897020e-01 4.60989743e-01 6.00952911e+00 3.35187405e-01
-1.20306170e+00 5.05100727e-01 4.83075172e-01 -2.20295563e-01
-1.44788967e-02 -3.94531071e-01 -1.23544440e-01 8.80245984e-01
8.93673420e-01 1.39405891e-01 4.64251786e-01 5.79006791e-01
3.72024298e-01 8.82168859e-02 -1.07821822e+00 1.06039894e+00
-9.08866748e-02 -6.15493596e-01 -2.75309473e-01 1.76377267e-01
5.02094328e-01 1.23277649e-01 2.94430971e-01 1.72116712e-01
-3.92488539e-01 -1.21863878e+00 3.54572684e-01 5.61786354e-01
1.30242109e+00 -6.52001798e-01 5.92137396e-01 3.79089713e-01
-8.48419487e-01 -5.42597711e-01 -1.96101680e-01 -3.17883819e-01
-4.34225917e-01 5.84086180e-01 -8.34280252e-01 -6.95070252e-02
1.08291304e+00 9.14410412e-01 -2.86354423e-01 8.24317753e-01
-1.06595449e-01 6.99750066e-01 8.89998078e-02 1.10638522e-01
-2.15057671e-01 -4.78791445e-01 6.74359381e-01 5.59045851e-01
7.27709234e-01 6.13840342e-01 -1.35080040e-01 1.23276126e+00
2.38293737e-01 4.71602082e-01 -8.10195208e-01 -4.48817581e-01
3.95396650e-01 9.40153718e-01 -3.70338410e-01 5.95758073e-02
-7.01966226e-01 6.71562970e-01 4.96222049e-01 -7.56976977e-02
-5.19950807e-01 -1.60831623e-02 6.46713078e-01 1.45718962e-01
1.07150869e-02 7.05528632e-02 -3.72252017e-01 -1.04329979e+00
-1.26680464e-01 -7.55906701e-01 1.75970659e-01 -8.02562892e-01
-1.41992962e+00 2.52405137e-01 -1.56365111e-02 -1.07410395e+00
-1.43046781e-01 -5.09499729e-01 -1.12647295e+00 7.04517841e-01
-5.77559650e-01 -9.87332404e-01 -8.18370506e-02 6.33569598e-01
4.41189587e-01 -6.27363443e-01 9.48693573e-01 3.96820873e-01
-1.21937788e+00 5.15494764e-01 -8.33710656e-02 1.87990963e-01
4.49691117e-01 -9.99515176e-01 -2.41007134e-01 2.42980883e-01
-6.02410316e-01 4.85699177e-01 2.58357555e-01 -9.25870955e-01
-6.16447985e-01 -8.47023666e-01 5.13146698e-01 4.06935997e-02
7.44624078e-01 8.43417570e-02 -1.05613804e+00 4.71381217e-01
1.14177592e-01 -7.33016208e-02 1.11639953e+00 6.61448110e-04
2.17338149e-02 1.73565358e-01 -1.67820787e+00 3.88862491e-01
1.15566063e+00 -6.49245203e-01 -7.98536837e-01 3.24401289e-01
5.70261598e-01 5.96602401e-03 -1.44575799e+00 8.49275529e-01
5.80301225e-01 -9.94522810e-01 6.70010149e-01 -7.36135066e-01
6.83711350e-01 1.92958355e-01 2.26565637e-02 -1.52576232e+00
-1.30355477e-01 1.02614485e-01 -1.24483503e-01 1.23804891e+00
5.59552871e-02 -8.55307877e-01 4.48802471e-01 6.53790474e-01
-2.11109206e-01 -1.18820286e+00 -9.29769635e-01 -3.38561147e-01
4.28869933e-01 -3.16858143e-01 7.71357894e-01 1.18509150e+00
1.78516865e-01 4.80563901e-02 3.32777053e-01 1.80970147e-01
2.96649098e-01 -3.74260157e-01 -1.38287351e-03 -1.63695192e+00
5.81174828e-02 -5.35792410e-01 -9.07885611e-01 9.77617428e-02
5.95311642e-01 -1.06826031e+00 -6.08758867e-01 -1.57790613e+00
4.22673285e-01 -3.70037884e-01 -2.81019598e-01 4.32684243e-01
-4.01458926e-02 -2.15279937e-01 -2.59632975e-01 -1.94666445e-01
9.18537974e-02 7.66859829e-01 1.39796686e+00 -3.01418632e-01
-2.07884058e-01 8.95332471e-02 -3.89580727e-01 1.19396460e+00
1.03663361e+00 -2.91644573e-01 -7.74208784e-01 -2.41233274e-01
-9.94163230e-02 1.53134927e-01 4.15849328e-01 -9.67680275e-01
-2.64870703e-01 -1.27240062e-01 7.03613579e-01 -2.33043045e-01
3.30985814e-01 -4.76486355e-01 -4.70660673e-03 5.80063403e-01
-9.32913944e-02 -1.14411473e-01 -1.43403634e-01 1.43417612e-01
7.45036975e-02 -4.02511746e-01 9.01348174e-01 7.68132359e-02
-3.80548090e-01 7.59976029e-01 -6.85303092e-01 8.33859369e-02
1.34583819e+00 -2.90386170e-01 2.54814159e-02 1.10021971e-01
-1.13953543e+00 5.60688376e-02 2.71004945e-01 3.00568819e-01
9.39372361e-01 -1.20701611e+00 -7.11338699e-01 4.10056293e-01
5.36674112e-02 -3.94250974e-02 6.93154454e-01 1.34824288e+00
-4.26013678e-01 2.02480137e-01 -6.02641106e-01 -5.11638284e-01
-1.33053350e+00 4.27326679e-01 3.38832945e-01 3.87902826e-01
-7.77276456e-01 9.23974872e-01 2.35513419e-01 -2.58984089e-01
-4.75758687e-03 -2.17644870e-01 -8.06238413e-01 3.89577448e-01
5.90978146e-01 6.58770502e-01 -5.28861172e-02 -8.36242855e-01
-4.23718929e-01 6.10281527e-01 -9.69771296e-02 7.18063712e-02
1.50482428e+00 -5.59475878e-03 -5.36689281e-01 3.77561867e-01
9.37017381e-01 -2.94990033e-01 -7.37872720e-01 2.48374984e-01
-2.91583747e-01 -1.39395684e-01 1.21974669e-01 -5.07073998e-01
-1.38277531e+00 1.14052975e+00 1.22088540e+00 1.41418055e-01
1.10430765e+00 1.92993715e-01 7.95589864e-01 -1.67340294e-01
5.50884843e-01 -8.78457606e-01 -1.14629334e-02 1.89578652e-01
1.14288843e+00 -1.27709329e+00 -2.88355798e-01 -1.56217918e-01
-5.92813849e-01 9.42135334e-01 1.08794761e+00 -3.20175916e-01
8.75201702e-01 -1.54986337e-01 -1.19498573e-01 -3.54480505e-01
-4.89740551e-01 7.50565603e-02 3.47411036e-01 1.04401517e+00
4.16323572e-01 3.69873136e-01 -5.86924136e-01 1.51339757e+00
-2.53177226e-01 -2.89941698e-01 2.37582877e-01 4.81641054e-01
-4.66130644e-01 -8.61330569e-01 -2.54325181e-01 1.30184007e+00
-6.57138526e-01 9.75245535e-02 -4.23966467e-01 4.57187086e-01
7.43884385e-01 8.34695220e-01 2.39819854e-01 -5.28693378e-01
2.61974931e-02 1.28528133e-01 4.75221246e-01 -9.24870253e-01
-3.06293219e-01 -1.96490854e-01 -1.07410453e-01 -3.42255354e-01
-4.08858091e-01 -1.05351162e+00 -1.72018933e+00 -1.16090225e-02
-2.82569021e-01 -2.35362649e-01 6.77814424e-01 1.22653925e+00
3.80576223e-01 5.81529498e-01 5.10186672e-01 -5.47346294e-01
-3.12693506e-01 -1.17229795e+00 -9.80484903e-01 -1.62853137e-01
3.84132475e-01 -1.19072568e+00 -1.35572881e-01 -4.65276062e-01] | [12.706745147705078, 3.0255184173583984] |
87d4d71d-09bd-4e60-bae5-7172a2f9f054 | watermarking-text-generated-by-black-box | 2305.08883 | null | https://arxiv.org/abs/2305.08883v1 | https://arxiv.org/pdf/2305.08883v1.pdf | Watermarking Text Generated by Black-Box Language Models | LLMs now exhibit human-like skills in various fields, leading to worries about misuse. Thus, detecting generated text is crucial. However, passive detection methods are stuck in domain specificity and limited adversarial robustness. To achieve reliable detection, a watermark-based method was proposed for white-box LLMs, allowing them to embed watermarks during text generation. The method involves randomly dividing the model vocabulary to obtain a special list and adjusting the probability distribution to promote the selection of words in the list. A detection algorithm aware of the list can identify the watermarked text. However, this method is not applicable in many real-world scenarios where only black-box language models are available. For instance, third-parties that develop API-based vertical applications cannot watermark text themselves because API providers only supply generated text and withhold probability distributions to shield their commercial interests. To allow third-parties to autonomously inject watermarks into generated text, we develop a watermarking framework for black-box language model usage scenarios. Specifically, we first define a binary encoding function to compute a random binary encoding corresponding to a word. The encodings computed for non-watermarked text conform to a Bernoulli distribution, wherein the probability of a word representing bit-1 being approximately 0.5. To inject a watermark, we alter the distribution by selectively replacing words representing bit-0 with context-based synonyms that represent bit-1. A statistical test is then used to identify the watermark. Experiments demonstrate the effectiveness of our method on both Chinese and English datasets. Furthermore, results under re-translation, polishing, word deletion, and synonym substitution attacks reveal that it is arduous to remove the watermark without compromising the original semantics. | ['Nenghai Yu', 'Han Fang', 'Jie Zhang', 'Yuang Qi', 'Chang Liu', 'Weiming Zhang', 'Kejiang Chen', 'Xi Yang'] | 2023-05-14 | null | null | null | null | ['specificity'] | ['natural-language-processing'] | [ 7.68407643e-01 -9.71669983e-03 -8.61390114e-01 2.52250999e-01
-6.50405288e-01 -1.20966208e+00 4.12147492e-01 2.60332286e-01
-3.77257645e-01 4.67479944e-01 -1.80751100e-01 -7.82170475e-01
6.01464093e-01 -9.04448390e-01 -6.66590154e-01 -3.98611248e-01
-3.77979502e-03 -3.33589375e-01 4.48594719e-01 -6.70675486e-02
3.95892024e-01 4.17751938e-01 -8.01035821e-01 2.85722792e-01
7.16740310e-01 6.41029775e-01 -6.72966428e-03 5.88101268e-01
-2.60536194e-01 2.07154244e-01 -1.16895604e+00 -8.60650480e-01
6.51187062e-01 -3.02518666e-01 -2.75980651e-01 -4.73839529e-02
-1.85030103e-01 -4.69667643e-01 -2.84038693e-01 1.50345469e+00
3.30708355e-01 -7.75016665e-01 3.90616089e-01 -1.61633313e+00
-6.13888502e-01 9.67587709e-01 -7.00175941e-01 -2.94380665e-01
5.36288559e-01 4.46244955e-01 1.09121013e+00 -3.83154929e-01
7.32208431e-01 7.84765780e-01 3.35300326e-01 6.05597973e-01
-1.14978623e+00 -1.07965100e+00 -1.10168522e-02 -2.01493502e-01
-1.37174773e+00 -2.16405392e-01 9.11910594e-01 -3.75920981e-01
5.79859436e-01 6.59552991e-01 5.39368212e-01 1.32055044e+00
5.70948601e-01 6.42171204e-01 9.93138671e-01 -7.98190236e-01
4.53391641e-01 4.61414933e-01 -1.81626573e-01 4.12530839e-01
1.10567737e+00 1.30131179e-02 -6.22331321e-01 -7.27322161e-01
3.38314772e-01 -9.48736146e-02 -4.43841398e-01 -3.96353781e-01
-1.23220563e+00 7.83764899e-01 -1.08216941e-01 3.78426313e-01
-7.31082633e-02 3.54206890e-01 4.59097236e-01 3.98287088e-01
-1.62954450e-01 4.48172510e-01 -4.55574125e-01 5.79096144e-03
-1.05690491e+00 1.22263536e-01 7.37426996e-01 1.07152593e+00
6.68424368e-01 2.16070842e-02 1.59645438e-01 5.54850027e-02
6.56999886e-01 7.34659553e-01 8.06666970e-01 -4.05632317e-01
6.31467283e-01 6.38978720e-01 1.32652447e-01 -1.15168846e+00
2.69552916e-01 -7.12254941e-02 -6.37756288e-01 1.58335522e-01
3.08068156e-01 -1.38568252e-01 -7.13072598e-01 1.61227608e+00
2.17558444e-02 -1.41610414e-01 1.48351356e-01 4.77966487e-01
5.61307408e-02 4.27152514e-01 -3.48706990e-02 -1.73463628e-01
1.75033116e+00 -2.91602343e-01 -7.79259384e-01 -3.58876437e-01
6.07144713e-01 -8.62428963e-01 1.11320531e+00 2.86323339e-01
-5.26714444e-01 -6.22184277e-02 -1.69333756e+00 4.77096021e-01
-6.05677605e-01 -1.54556379e-01 1.40353382e-01 1.59559286e+00
-4.91712093e-01 -1.56042036e-02 -5.67097127e-01 -2.39753332e-02
2.94669479e-01 2.83058435e-01 -3.97484392e-01 1.72212347e-01
-1.57280064e+00 3.69662911e-01 5.54103196e-01 -3.81136537e-01
-6.28153980e-01 -4.78301734e-01 -9.78460252e-01 -4.33161147e-02
2.70380676e-01 -3.34502071e-01 1.01859689e+00 -1.06205797e+00
-1.34703290e+00 7.37079442e-01 2.40000384e-03 -7.11038172e-01
5.38429558e-01 3.46602440e-01 -8.65423501e-01 2.44271621e-01
1.09684810e-01 3.14569801e-01 1.62408006e+00 -1.43976116e+00
-5.48851132e-01 1.37058824e-01 1.58163592e-01 -5.13209522e-01
-7.80475795e-01 2.04079956e-01 -1.42517835e-01 -1.26504040e+00
-2.27434766e-02 -8.71067405e-01 -1.27110809e-01 -5.20908944e-02
-8.48066390e-01 5.70594013e-01 1.02171016e+00 -6.89778328e-01
1.75474596e+00 -2.47527337e+00 -4.74822193e-01 7.06648290e-01
1.80082604e-01 4.46915805e-01 -7.18098655e-02 6.78616166e-01
2.16556825e-02 8.66828084e-01 -5.23629487e-01 8.80863741e-02
3.67757916e-01 -1.06339492e-01 -9.08856988e-01 5.51220238e-01
2.39347473e-01 9.11600113e-01 -7.77992904e-01 -2.95270979e-01
-1.21052116e-01 6.51229843e-02 -5.62320888e-01 -6.50457367e-02
-5.11350274e-01 -1.88581660e-01 -5.11833489e-01 9.74077165e-01
7.94196665e-01 4.21298593e-02 6.47510171e-01 -6.05304688e-02
2.09061861e-01 4.24883850e-02 -1.45725882e+00 1.16384220e+00
-3.93767089e-01 5.26787817e-01 4.67817374e-02 -3.07720631e-01
8.45940232e-01 4.00861084e-01 -4.39013820e-03 -2.94491202e-01
7.47849643e-02 4.79615927e-01 9.67994146e-03 -2.83251405e-01
6.56768799e-01 8.57618004e-02 -6.11715317e-01 7.21653819e-01
-6.04039550e-01 -2.26312652e-01 5.99688143e-02 1.92470267e-01
1.31659865e+00 -2.05152571e-01 4.69864517e-01 3.56802553e-01
6.68758690e-01 -2.91070729e-01 5.27919173e-01 7.26881862e-01
-2.02603593e-01 2.30158150e-01 6.93582952e-01 -7.75546709e-04
-8.91514361e-01 -8.48693609e-01 2.27773413e-01 5.43718398e-01
2.41332471e-01 -8.73927474e-01 -7.11708903e-01 -1.07468545e+00
1.57832652e-01 9.90480185e-01 -1.39971122e-01 -4.62926686e-01
-4.66272682e-01 -3.98496300e-01 1.23815918e+00 3.44072580e-02
5.28840840e-01 -8.74405682e-01 -8.80010247e-01 2.41363153e-01
-1.26230091e-01 -1.19067013e+00 -8.14094961e-01 1.15469228e-02
-3.21032584e-01 -1.02944505e+00 -5.84327936e-01 -5.99504471e-01
9.07589793e-01 1.00033320e-02 3.25222999e-01 1.69848755e-01
-2.22342610e-01 1.28115639e-01 -4.44938362e-01 -3.29757303e-01
-1.08539057e+00 1.58226505e-01 -1.55666813e-01 1.78565562e-01
3.06223363e-01 -1.19482473e-01 -4.47605610e-01 3.11719328e-01
-1.66352749e+00 -4.36714470e-01 4.91815478e-01 5.64153969e-01
3.99226062e-02 4.88284558e-01 4.54089135e-01 -7.32058108e-01
8.91585886e-01 -2.08938062e-01 -8.09900701e-01 2.91373998e-01
-5.72501659e-01 8.68550874e-03 6.87008858e-01 -9.92653728e-01
-4.58782494e-01 -7.77073530e-03 2.10804269e-01 -1.06327005e-01
4.53122631e-02 4.43375438e-01 -6.49858594e-01 -2.33429238e-01
4.91447985e-01 3.51360977e-01 -2.64233172e-01 3.14288400e-02
6.39614224e-01 1.06017613e+00 2.92577744e-01 -5.89537680e-01
1.58392549e+00 4.39476490e-01 -3.35655749e-01 -5.07759869e-01
2.78460920e-01 1.47134019e-02 -2.03394331e-02 5.05965129e-02
5.17193675e-01 -6.46508515e-01 -4.99509990e-01 7.33127475e-01
-1.15848875e+00 -5.36887534e-02 -1.96051806e-01 4.23907973e-02
-2.04546377e-01 9.17525828e-01 -3.03598046e-01 -7.67540753e-01
-3.01066041e-01 -1.53423333e+00 8.62758458e-01 -5.83359264e-02
-6.28829181e-01 -5.86566687e-01 -2.70900965e-01 2.21059635e-01
4.28269565e-01 1.85123682e-01 1.31912684e+00 -7.80354142e-01
-6.25829458e-01 -9.52201605e-01 1.14600785e-01 2.49127761e-01
4.81215537e-01 2.01593831e-01 -6.99846506e-01 -4.77778077e-01
-1.14838265e-01 2.24749655e-01 1.80118427e-01 -4.18106198e-01
8.85109425e-01 -7.05219507e-01 -2.97498137e-01 4.16071802e-01
1.34934390e+00 3.68880868e-01 6.21124983e-01 7.55601466e-01
2.70413965e-01 2.86158830e-01 3.49895209e-01 5.20666838e-01
1.78054441e-02 5.70316613e-01 2.97506392e-01 1.50883883e-01
1.69896588e-01 -7.63494432e-01 7.63067782e-01 4.06040609e-01
8.43591034e-01 -5.24012983e-01 -6.83909714e-01 3.80135655e-01
-1.21110308e+00 -7.85935700e-01 2.00120658e-01 2.42314148e+00
1.24291182e+00 8.12015414e-01 -9.32241678e-02 6.16391897e-01
9.54744041e-01 2.97595978e-01 -3.21815342e-01 -4.56007630e-01
-1.70397878e-01 3.12355489e-01 1.27141833e+00 3.47435266e-01
-8.78627121e-01 1.03752530e+00 5.74391985e+00 1.05751836e+00
-1.32384372e+00 5.58640924e-04 2.03254789e-01 3.96989852e-01
-8.48114967e-01 4.92596090e-01 -7.50536263e-01 9.29029167e-01
7.73506641e-01 -6.32348716e-01 3.26813966e-01 5.45413256e-01
1.75671905e-01 -1.38491811e-02 -7.47806251e-01 6.21826172e-01
-1.59335770e-02 -1.13581562e+00 2.37924516e-01 2.02198252e-01
5.11318207e-01 -6.79833770e-01 2.19589546e-01 -3.09650525e-02
1.71962589e-01 -6.19097471e-01 1.14063549e+00 3.21018919e-02
1.04060626e+00 -6.11958027e-01 5.76241970e-01 3.11101794e-01
-1.06482375e+00 -6.61918335e-03 -8.20610859e-03 3.15876871e-01
2.50630397e-02 6.07011437e-01 -9.84611452e-01 3.96788180e-01
1.52458549e-01 1.25019839e-02 -5.36652923e-01 7.09964395e-01
-7.13034689e-01 6.69570684e-01 -2.57899821e-01 -3.76331178e-03
-4.62195650e-02 1.61112174e-01 6.81854665e-01 1.28185296e+00
5.36785007e-01 -6.91185594e-01 -6.81835413e-02 7.59710252e-01
-4.65024382e-01 9.54545513e-02 -7.88923681e-01 -5.67531168e-01
8.77462566e-01 9.96944249e-01 -8.67790878e-01 -1.14664741e-01
-4.09877539e-01 1.16388822e+00 -6.43262088e-01 5.58215380e-01
-8.25098813e-01 -8.81318212e-01 6.25532091e-01 2.54773796e-01
2.73775041e-01 -5.34278035e-01 -4.48068917e-01 -1.04224133e+00
2.73080081e-01 -1.55448663e+00 5.86048588e-02 -3.79068941e-01
-9.42437768e-01 2.25939378e-01 -8.40883702e-02 -1.38056290e+00
-1.43800572e-01 -2.79216468e-01 -5.01814365e-01 8.78817856e-01
-1.50098252e+00 -1.05150783e+00 2.69474268e-01 4.26590592e-01
1.41114414e-01 -4.85418677e-01 8.26454818e-01 9.41548124e-02
-4.90307003e-01 1.16330838e+00 -1.96038708e-01 5.15292287e-01
8.61212909e-01 -8.98006260e-01 9.19114470e-01 1.34271693e+00
9.69322622e-02 1.22141063e+00 7.17762649e-01 -1.24239278e+00
-1.69990349e+00 -8.85697365e-01 9.19563949e-01 9.89067182e-03
1.08030570e+00 -7.74004340e-01 -6.74601972e-01 6.13656521e-01
1.76811621e-01 -1.27002299e-01 8.11487913e-01 -1.03141916e+00
-8.03813279e-01 2.08418652e-01 -1.54793584e+00 1.02864480e+00
4.22292203e-01 -8.99464965e-01 -4.76219565e-01 6.02354258e-02
9.94565606e-01 -1.06563896e-01 -5.43620408e-01 2.33078182e-01
6.33901179e-01 -4.58163679e-01 7.15382218e-01 -2.45310277e-01
4.36963281e-03 -8.53709817e-01 -4.17261153e-01 -7.82996893e-01
4.39336121e-01 -1.26961720e+00 -3.55479121e-01 1.46091187e+00
4.44058716e-01 -1.07578766e+00 7.10853875e-01 4.18055981e-01
6.88695669e-01 -5.29626682e-02 -1.15101743e+00 -9.95307624e-01
-1.18818715e-01 -7.59011149e-01 1.02928734e+00 9.87319410e-01
1.56722978e-01 -4.44068789e-01 -2.68002212e-01 5.44884145e-01
4.99043375e-01 -2.69666553e-01 8.67027760e-01 -5.81512451e-01
-2.95264721e-01 -5.45623779e-01 -5.56210756e-01 -7.93028712e-01
2.82675892e-01 -8.95036280e-01 -3.00331324e-01 -8.76088202e-01
-2.79254794e-01 -2.88344026e-01 -1.35623664e-01 5.29473364e-01
1.33425733e-02 3.04074287e-01 4.06144023e-01 1.41509458e-01
1.63475990e-01 5.18301912e-02 6.47320747e-01 -6.69720173e-01
-1.16186172e-01 1.12841390e-01 -6.89372838e-01 5.60528457e-01
9.43079352e-01 -7.91179121e-01 -4.37051177e-01 8.35808516e-02
6.05944037e-01 -2.38552004e-01 2.93071121e-01 -7.71071434e-01
1.24827735e-01 -2.90015608e-01 -7.59096518e-02 -2.07548946e-01
-1.98977381e-01 -1.24748814e+00 2.12889671e-01 8.57921183e-01
-2.89210737e-01 4.34939638e-02 1.04130395e-01 5.65324843e-01
8.82472470e-03 -7.90402472e-01 4.96535987e-01 1.20580129e-01
-5.18848062e-01 -2.90403813e-02 -8.46219540e-01 -1.50336251e-01
1.27508223e+00 -4.11445826e-01 -4.12238777e-01 -3.12665999e-01
-2.12448433e-01 -1.57731116e-01 1.08347762e+00 5.54356217e-01
6.23232782e-01 -1.17488885e+00 -2.70963341e-01 7.47547209e-01
3.73033643e-01 -6.38040602e-01 -4.01967138e-01 1.13721877e-01
-6.50946498e-01 -5.45730256e-02 2.52134830e-01 -1.16241872e-01
-1.24420607e+00 7.20945835e-01 5.62662072e-02 -1.71713635e-01
-2.56520957e-01 1.72623858e-01 -3.46607983e-01 -8.93704742e-02
2.28749841e-01 -5.50223529e-01 3.95155281e-01 -2.47777300e-03
6.23051524e-01 -9.65147652e-03 2.18638107e-02 -2.82675117e-01
-4.40512598e-01 4.12439942e-01 -6.63266890e-03 -4.60771054e-01
5.29213250e-01 -1.31558189e-02 -1.99237183e-01 3.85203622e-02
1.05228865e+00 9.53073323e-01 -6.01260781e-01 -5.75433373e-02
3.54030699e-01 -5.79021573e-01 -2.11913407e-01 -7.22006321e-01
-8.32729697e-01 3.94090503e-01 4.95633334e-01 5.65505803e-01
9.94030178e-01 -5.46756864e-01 1.14469004e+00 1.94773525e-01
7.06937134e-01 -8.30224335e-01 -1.71626046e-01 -7.63047906e-03
4.33295906e-01 -7.65137494e-01 -1.54922688e-02 -4.85224098e-01
-4.29755509e-01 1.13865173e+00 1.71583906e-01 3.46651256e-01
5.23067892e-01 6.93557501e-01 2.47622401e-01 3.70193303e-01
-2.33679757e-01 1.71053290e-01 -1.12559855e-01 7.70204782e-01
-2.07986534e-02 2.24845752e-01 -5.95021129e-01 6.88911378e-01
-2.58976042e-01 4.81717549e-02 9.52128232e-01 1.43731833e+00
-2.02712178e-01 -1.83607972e+00 -8.24624479e-01 1.24498509e-01
-7.11622179e-01 -2.35065177e-01 -4.71831650e-01 6.28175259e-01
2.46429771e-01 1.17435193e+00 -3.69594067e-01 -5.63191116e-01
2.09388420e-01 3.11552107e-01 -7.02764466e-02 -5.06155372e-01
-5.85159540e-01 3.30508836e-02 -4.39899378e-02 -1.67617977e-01
-5.72050288e-02 -3.43180656e-01 -1.29074788e+00 -2.87775576e-01
-3.59053165e-01 -3.20118666e-02 7.98066080e-01 4.88223046e-01
1.55795172e-01 2.67531782e-01 8.18043172e-01 -2.20948324e-01
-6.19412124e-01 -3.80123198e-01 -4.92226839e-01 2.11831152e-01
2.98252523e-01 -2.32219011e-01 -6.34852827e-01 3.16603631e-01] | [5.903992176055908, 7.767019271850586] |
fa456c7b-041d-4d8d-b158-37f8d1bb736b | speaker-conditioning-single-channel-target | 2205.13851 | null | https://arxiv.org/abs/2205.13851v1 | https://arxiv.org/pdf/2205.13851v1.pdf | Speaker-conditioning Single-channel Target Speaker Extraction using Conformer-based Architectures | Target speaker extraction aims at extracting the target speaker from a mixture of multiple speakers exploiting auxiliary information about the target speaker. In this paper, we consider a complete time-domain target speaker extraction system consisting of a speaker embedder network and a speaker separator network which are jointly trained in an end-to-end learning process. We propose two different architectures for the speaker separator network which are based on the convolutional augmented transformer (conformer). The first architecture uses stacks of conformer and external feed-forward blocks (Conformer-FFN), while the second architecture uses stacks of temporal convolutional network (TCN) and conformer blocks (TCN-Conformer). Experimental results for 2-speaker mixtures, 3-speaker mixtures, and noisy mixtures of 2-speakers show that among the proposed separator networks, the TCN-Conformer significantly improves the target speaker extraction performance compared to the Conformer-FFN and a TCN-based baseline system. | ['Simon Doclo', 'Christian Rollwage', 'Marvin Tammen', 'Ragini Sinha'] | 2022-05-27 | null | null | null | null | ['target-speaker-extraction'] | ['audio'] | [ 1.97320879e-01 3.34834099e-01 1.22812413e-01 -6.72985196e-01
-1.17685616e+00 -3.34108293e-01 8.23922276e-01 -3.41504544e-01
-1.99435845e-01 -2.44032685e-02 3.16814899e-01 -3.55129808e-01
1.74080208e-01 -7.84157515e-02 -4.72870499e-01 -8.30538452e-01
-2.51370102e-01 4.51660633e-01 2.08197609e-01 -2.09816307e-01
-4.22174096e-01 3.88991296e-01 -1.25328696e+00 4.26656574e-01
6.26115263e-01 1.17504942e+00 -6.14102483e-02 7.45162010e-01
-1.07698120e-01 7.04335868e-01 -7.10171640e-01 -1.82927996e-01
1.87416553e-01 -5.08380175e-01 -5.72340071e-01 1.19458176e-01
6.71234310e-01 -2.68011421e-01 -2.84187317e-01 8.49644542e-01
6.15854323e-01 7.53172264e-02 5.37709773e-01 -1.57810402e+00
-1.24890335e-01 1.40100300e+00 -6.31306946e-01 3.46138239e-01
-9.44503695e-02 -4.13788222e-02 6.58369660e-01 -1.08502424e+00
-3.40802222e-02 1.79476285e+00 6.06125653e-01 6.57916963e-01
-1.29524124e+00 -1.30779672e+00 6.82961583e-01 -1.64456908e-02
-1.45378411e+00 -1.14255142e+00 1.01022124e+00 -3.60917956e-01
8.58062446e-01 3.26566070e-01 5.07117212e-01 1.34083080e+00
-2.43633613e-01 1.14992797e+00 7.70082653e-01 -2.55228549e-01
2.17707708e-01 3.62010837e-01 4.60854977e-01 2.94743270e-01
-2.82379657e-01 3.16765636e-01 -8.74933660e-01 -2.24982485e-01
4.04402465e-01 -1.94295049e-01 -3.01789224e-01 -5.27946241e-02
-1.06816375e+00 6.81604564e-01 3.40066612e-01 5.58297455e-01
-3.01150084e-01 -1.08952187e-02 3.94720227e-01 3.73836696e-01
7.05519557e-01 -4.67048615e-01 -2.56196409e-01 3.64539176e-01
-1.58464491e+00 2.16402099e-01 1.02002227e+00 1.09469986e+00
3.81925374e-01 5.99870801e-01 -1.28791034e-01 8.37131560e-01
6.65225863e-01 5.72862864e-01 8.90809953e-01 -3.28636318e-01
4.38422531e-01 2.42670402e-01 -6.22921623e-02 -2.28166759e-01
-3.20882022e-01 -8.65019679e-01 -9.26755786e-01 1.89739734e-01
9.12994072e-02 -2.97030717e-01 -1.32465827e+00 1.95953846e+00
6.10386848e-01 4.06119466e-01 2.57121235e-01 7.63705313e-01
1.17624056e+00 8.09729278e-01 -8.51303339e-02 -2.48495117e-01
1.47577417e+00 -1.30520833e+00 -9.02037323e-01 -4.84727174e-01
2.12248221e-01 -8.13397348e-01 4.61640745e-01 2.23588198e-01
-1.01785505e+00 -6.40361488e-01 -1.21084154e+00 1.78177416e-01
-2.55526572e-01 3.18599790e-01 -4.50007021e-02 6.85514987e-01
-1.24408078e+00 2.69375443e-01 -1.12895429e+00 -2.16193751e-01
1.62637427e-01 7.35317945e-01 -1.83633819e-01 5.81469834e-01
-1.15467429e+00 7.59375393e-01 2.28910893e-01 4.17380661e-01
-1.43047881e+00 -7.91645944e-01 -1.07004702e+00 3.95690650e-01
4.57378589e-02 -3.77604544e-01 1.93032825e+00 -1.21638930e+00
-2.00680733e+00 5.78888535e-01 -5.34811437e-01 -8.74182343e-01
3.88564974e-01 -1.60989940e-01 -9.18809056e-01 -2.00042725e-02
-3.00983824e-02 5.81609368e-01 1.39808702e+00 -1.08179414e+00
-9.19335544e-01 -3.67472500e-01 -5.95882714e-01 -2.60480829e-02
-1.94747314e-01 4.53328997e-01 -4.52650525e-02 -6.11005366e-01
2.53041983e-01 -8.44929218e-01 1.02532707e-01 -3.55066001e-01
-8.99786353e-01 -3.76220167e-01 1.53574836e+00 -8.63326728e-01
1.20295775e+00 -2.49091959e+00 6.80261254e-02 1.00163847e-01
3.65963489e-01 3.09980541e-01 -4.33072090e-01 2.64310420e-01
-5.81591606e-01 -3.64893436e-01 -2.12479144e-01 -1.00465548e+00
2.60064930e-01 -4.58693653e-01 -4.64994580e-01 5.18985569e-01
2.40729868e-01 5.25952399e-01 -4.17069912e-01 -3.61605406e-01
1.27857223e-01 7.50904560e-01 -6.60337806e-02 3.43137801e-01
-8.42351615e-02 3.34560812e-01 -6.76751807e-02 5.86848259e-01
8.57998312e-01 2.18761384e-01 5.76383807e-02 -1.02920011e-01
-2.66526610e-01 9.01213646e-01 -1.32905805e+00 1.24896336e+00
-3.53096008e-01 8.32548678e-01 1.03742564e+00 -5.67786574e-01
1.01324618e+00 9.61145878e-01 2.16600940e-01 -3.01520079e-01
1.86652198e-01 4.56608862e-01 3.84511709e-01 -1.05873495e-01
1.69411212e-01 -4.16950792e-01 7.26205483e-03 4.14338917e-01
4.33992714e-01 1.51380062e-01 6.39557019e-02 1.77057520e-01
8.39293599e-01 -1.71388581e-01 1.01699464e-01 -4.12612915e-01
7.90789187e-01 -7.59680867e-01 8.35713029e-01 3.24890018e-01
-4.99167442e-01 1.51626408e-01 1.73238665e-01 -2.40621626e-01
-5.99696815e-01 -1.12925661e+00 4.17710207e-02 1.25429893e+00
-3.33267480e-01 -3.63898844e-01 -8.77027929e-01 -5.13213515e-01
-1.00494228e-01 7.61038661e-01 -4.66892272e-01 -7.02342615e-02
-8.28106821e-01 -4.22539979e-01 7.08472431e-01 6.18747771e-01
5.74823380e-01 -7.82064497e-01 -2.43680045e-01 3.02872360e-01
-3.21890891e-01 -1.07224655e+00 -1.23391080e+00 6.17146552e-01
-6.32120550e-01 -4.51191992e-01 -8.44573677e-01 -1.03984261e+00
1.01611629e-01 2.99124382e-02 8.49438906e-01 -7.39723980e-01
3.82200301e-01 -9.62381437e-02 1.83447197e-01 -8.25343430e-01
-6.90143466e-01 3.74255240e-01 3.01406741e-01 6.68631732e-01
6.11440897e-01 -9.18425381e-01 -2.49158949e-01 2.62651771e-01
-6.17205739e-01 3.38022225e-03 5.72182000e-01 5.21479726e-01
6.20649718e-02 -1.47227794e-01 8.20704937e-01 -4.53061879e-01
3.25451016e-01 -2.59064376e-01 -6.94921374e-01 -8.10439363e-02
-3.22865963e-01 5.23234718e-02 3.79026860e-01 -8.52011800e-01
-1.23808253e+00 4.72343564e-01 -3.04794818e-01 -7.24383295e-01
-2.73924589e-01 1.75102875e-01 -6.09512150e-01 3.30433011e-01
2.95690030e-01 5.09859025e-01 8.33175890e-03 -7.20888674e-01
1.25078753e-01 1.09553862e+00 6.16941869e-01 -1.14139400e-01
8.54115367e-01 1.51003703e-01 -5.94196618e-01 -9.81163800e-01
-7.89863646e-01 -5.73923349e-01 -6.41996920e-01 -1.05747525e-02
7.50253320e-01 -1.36763740e+00 -7.34860897e-01 8.80232394e-01
-1.26767993e+00 -3.45358372e-01 -3.47941607e-01 7.20119894e-01
-3.18671227e-01 -5.11306375e-02 -6.75460994e-01 -1.23468387e+00
-6.81320906e-01 -1.24553144e+00 1.39635813e+00 2.18155310e-01
-2.72176892e-01 -8.81607592e-01 1.08620308e-01 1.56624973e-01
6.59564555e-01 -8.08068216e-02 5.85456312e-01 -1.43265784e+00
-4.48646426e-01 -6.42405916e-03 2.46763349e-01 4.79230165e-01
3.72121632e-01 -1.35432005e-01 -1.63308620e+00 -6.34706497e-01
4.82523292e-01 1.05798915e-01 1.00767994e+00 7.03636408e-01
3.09097141e-01 -5.97094536e-01 -5.28314590e-01 4.73058552e-01
7.90215313e-01 4.61092412e-01 1.56118512e-01 -1.27768159e-01
5.59141576e-01 5.96852362e-01 -1.33098781e-01 5.53733390e-03
4.64958668e-01 6.34187639e-01 2.41162717e-01 -2.43993059e-01
-2.69592434e-01 1.13703847e-01 1.01240170e+00 1.07763660e+00
5.20313263e-01 -2.07604259e-01 -7.71438599e-01 7.66909361e-01
-1.65681708e+00 -1.01453948e+00 5.90678640e-02 2.20928502e+00
8.34934056e-01 1.26739636e-01 7.66563416e-01 3.02938938e-01
1.11741519e+00 2.80000538e-01 -6.85575604e-01 -2.83972621e-01
4.50262427e-02 -1.99426729e-02 -8.64784867e-02 7.96347260e-01
-1.19318473e+00 6.96249247e-01 5.70651960e+00 6.24184251e-01
-1.56202626e+00 3.00968558e-01 3.43116194e-01 -4.10142213e-01
9.00055021e-02 -1.57043010e-01 -1.37517798e+00 3.00857872e-01
1.63245153e+00 -1.81074202e-01 1.70609757e-01 6.76363885e-01
2.59984910e-01 5.09064436e-01 -1.75457752e+00 9.16700244e-01
1.58708811e-01 -8.25327694e-01 -8.02472159e-02 9.78776589e-02
1.99695453e-01 1.86996505e-01 9.73411500e-02 6.61934912e-01
2.97106206e-01 -7.04972386e-01 1.34697533e+00 8.55319127e-02
3.23437840e-01 -8.22856069e-01 4.64578032e-01 4.84360009e-01
-1.68260729e+00 -1.24296151e-01 3.62853527e-01 4.53416258e-01
2.95988917e-01 6.61353052e-01 -9.86775398e-01 4.18647915e-01
7.13120878e-01 5.51196992e-01 -2.51146346e-01 6.90202177e-01
1.43037528e-01 9.79430377e-01 -6.57175183e-01 4.39797372e-01
2.40437150e-01 1.62733123e-01 1.06074905e+00 1.48552275e+00
1.51146591e-01 -3.35957080e-01 1.37306020e-01 1.01019347e+00
-2.45666265e-01 -2.25013003e-01 -9.89628434e-02 -1.41027316e-01
5.80527425e-01 1.50871646e+00 -4.67472583e-01 -5.71187735e-01
-2.12109134e-01 6.95592582e-01 1.52691752e-01 4.90329623e-01
-8.31281126e-01 -5.60922384e-01 7.48544097e-01 1.59610808e-01
7.19015718e-01 1.39596939e-01 1.29209727e-01 -1.08897567e+00
9.75986794e-02 -1.16899669e+00 3.02353740e-01 -3.02922040e-01
-1.21366191e+00 1.19428813e+00 -2.65150834e-02 -1.14593160e+00
-4.73730296e-01 -3.17096084e-01 -1.04841959e+00 1.19291294e+00
-1.46969390e+00 -1.41351640e+00 1.05372900e-02 7.44588673e-01
8.66755605e-01 -3.69292557e-01 7.47481346e-01 2.68736482e-01
-8.89339387e-01 7.68519998e-01 -1.57141954e-01 3.55862975e-01
6.61863089e-01 -1.25007379e+00 6.29070342e-01 9.83499825e-01
4.86892201e-02 6.34575665e-01 5.82753062e-01 -3.35837275e-01
-8.39533150e-01 -1.47463894e+00 1.12734056e+00 -5.48770539e-02
3.21956038e-01 -9.28017497e-01 -8.83227289e-01 1.23597288e+00
7.13042438e-01 -1.18849978e-01 7.88740754e-01 7.22444877e-02
-6.48416519e-01 -4.54136938e-01 -1.02196634e+00 3.30753595e-01
2.99488932e-01 -6.64896905e-01 -6.64088607e-01 1.08685307e-01
8.83729875e-01 -4.06744182e-01 -6.48974359e-01 2.58062929e-01
4.76883411e-01 -7.89748073e-01 7.47494996e-01 -1.99143484e-01
-2.13734776e-01 -4.36983138e-01 -1.49678871e-01 -1.31414366e+00
-3.14878881e-01 -1.02981412e+00 -4.69960600e-01 1.70342553e+00
6.63269341e-01 -7.01388955e-01 4.94897217e-01 3.32995713e-01
-4.09290135e-01 -3.84787112e-01 -1.29962003e+00 -1.05414462e+00
1.45817384e-01 -2.62421310e-01 9.70660686e-01 7.59822547e-01
-1.18510708e-01 9.62009311e-01 -3.19317847e-01 5.57360590e-01
8.76495540e-01 1.19530722e-01 5.00134349e-01 -1.35785854e+00
-2.91402668e-01 -5.82097113e-01 -4.71215285e-02 -1.07845402e+00
3.25427890e-01 -8.75572443e-01 3.96182835e-01 -1.22958446e+00
7.24419430e-02 1.24492116e-01 -4.15055215e-01 6.13355100e-01
4.76604179e-02 -1.28722847e-01 8.56141001e-02 -6.09267829e-03
-2.15884194e-01 7.70579159e-01 6.10381007e-01 -6.65714920e-01
-5.74655294e-01 4.40073103e-01 -4.83335942e-01 8.99695516e-01
5.15763044e-01 -6.21812940e-01 -4.01920348e-01 -8.08561370e-02
-6.82741106e-01 1.29771665e-01 3.25103283e-01 -1.02368116e+00
4.27609921e-01 3.73688757e-01 3.81753474e-01 -1.07451212e+00
7.64126003e-01 -7.55936384e-01 3.61584523e-03 5.20267546e-01
-5.32561421e-01 -2.43912980e-01 3.57412428e-01 3.84746432e-01
-4.79362011e-01 2.04717234e-01 1.06699431e+00 2.30643451e-01
-7.40846060e-03 2.99061120e-01 -6.26197875e-01 -2.88348109e-01
7.10990310e-01 7.55202631e-03 3.36339585e-02 -4.42911059e-01
-1.09876275e+00 2.07289815e-01 -5.03220618e-01 5.43040037e-01
5.62829852e-01 -1.52182281e+00 -9.28324521e-01 4.34758723e-01
-1.39527217e-01 2.64204014e-02 1.11248605e-01 1.05291879e+00
3.64654630e-01 6.17535293e-01 3.14378053e-01 -9.12544966e-01
-1.56796885e+00 5.25486171e-01 7.94116259e-01 -3.27692837e-01
-4.82888222e-01 1.10226107e+00 6.72558427e-01 -6.99934006e-01
7.94828296e-01 -7.32009530e-01 6.02689423e-02 2.42748093e-02
7.32874572e-01 3.48679215e-01 2.53081530e-01 -9.94771481e-01
-6.73207760e-01 3.09611242e-02 -5.66565394e-01 -4.44816053e-01
1.28635752e+00 -1.27361208e-01 6.07096516e-02 7.40559876e-01
1.06971681e+00 -6.43830746e-02 -1.18304253e+00 -6.55868530e-01
1.23604611e-01 3.12351882e-01 2.88397908e-01 -6.12266064e-01
-1.13947964e+00 8.96012187e-01 8.56124699e-01 2.44220272e-01
1.27698898e+00 -2.64526643e-02 8.47307384e-01 9.36170444e-02
-1.62572116e-01 -6.93128705e-01 -2.06636190e-01 4.63114381e-01
1.00633228e+00 -9.96434629e-01 -6.63546622e-01 -3.99001837e-01
-2.15040848e-01 1.02310717e+00 6.70353889e-01 2.46253639e-01
1.20653069e+00 5.88432074e-01 2.92997122e-01 -1.44338787e-01
-9.88422871e-01 -1.19411908e-01 4.84304637e-01 2.89650530e-01
6.49077058e-01 -4.81583849e-02 6.51816964e-01 9.66776848e-01
-4.46426779e-01 -2.79363483e-01 9.32917073e-02 7.62112379e-01
-2.76901245e-01 -1.02614069e+00 -5.42466879e-01 2.67181993e-02
-3.24656934e-01 -1.46212429e-01 -5.85329235e-01 5.80532014e-01
-4.09521461e-02 1.22739685e+00 2.26752758e-01 -3.95096004e-01
3.90468359e-01 5.41638851e-01 1.38614565e-01 -6.49893880e-01
-1.20842445e+00 9.37130094e-01 2.10967008e-02 -3.30128372e-01
-5.02480626e-01 -5.98807096e-01 -1.06896031e+00 -1.10840343e-01
-7.25922585e-01 4.61595625e-01 8.39162171e-01 1.03103840e+00
3.69561583e-01 8.71476591e-01 7.20668852e-01 -1.08746004e+00
-7.54634678e-01 -1.60816205e+00 -8.61232102e-01 -1.67923048e-01
1.20650446e+00 -1.50281399e-01 -4.76639569e-01 2.81958222e-01] | [14.652992248535156, 5.973742485046387] |
18d4d60c-30bf-423e-a852-4c9c409f263a | cvsnet-a-computer-implementation-for-central | 2305.19492 | null | https://arxiv.org/abs/2305.19492v1 | https://arxiv.org/pdf/2305.19492v1.pdf | CVSNet: A Computer Implementation for Central Visual System of The Brain | In computer vision, different basic blocks are created around different matrix operations, and models based on different basic blocks have achieved good results. Good results achieved in vision tasks grants them rationality. However, these experimental-based models also make deep learning long criticized for principle and interpretability. Deep learning originated from the concept of neurons in neuroscience, but recent designs detached natural neural networks except for some simple concepts. In this paper, we build an artificial neural network, CVSNet, which can be seen as a computer implementation for central visual system of the brain. Each block in CVSNet represents the same vision information as that in brains. In CVSNet, blocks differs from each other and visual information flows through three independent pathways and five different blocks. Thus CVSNet is completely different from the design of all previous models, in which basic blocks are repeated to build model and information between channels is mixed at the outset. In ablation experiment, we show the information extracted by blocks in CVSNet and compare with previous networks, proving effectiveness and rationality of blocks in CVSNet from experiment side. And in the experiment of object recognition, CVSNet achieves comparable results to ConvNets, Vision Transformers and MLPs. | ['Zhekai Duan', 'Hao Zou', 'Ruimin Gao'] | 2023-05-31 | null | null | null | null | ['object-recognition'] | ['computer-vision'] | [-2.73625195e-01 2.89725006e-01 1.87493131e-01 -7.27467239e-02
7.58973718e-01 -2.49935403e-01 7.75432825e-01 -6.15595520e-01
-6.38684511e-01 5.54696858e-01 1.05914041e-01 -1.01148210e-01
-6.53493702e-02 -8.58539164e-01 -7.26704299e-01 -9.35023963e-01
2.38532111e-01 1.62384063e-02 4.46693867e-01 -3.05916935e-01
2.54743665e-01 6.26265049e-01 -1.22120118e+00 5.80841064e-01
3.30513477e-01 7.81701148e-01 6.53500557e-01 3.89402270e-01
-4.89654183e-01 1.03191769e+00 -6.58486247e-01 -3.02962214e-01
4.21923041e-01 -4.33885753e-01 -8.29968095e-01 -2.90200710e-01
3.89521499e-03 -2.15814546e-01 -6.32016659e-01 1.17040277e+00
5.14327347e-01 -5.32122672e-01 7.56010115e-01 -1.35274279e+00
-9.52556074e-01 1.01340783e+00 -6.22002304e-01 6.28335848e-02
-1.17983714e-01 2.27730095e-01 6.81091309e-01 -9.28690195e-01
3.69778842e-01 1.27907085e+00 8.28527927e-01 8.65763307e-01
-1.28752625e+00 -6.38170540e-01 1.84133410e-01 3.78952831e-01
-1.16116834e+00 -1.81178883e-01 4.19809967e-01 -5.31532049e-01
9.63237166e-01 8.53919834e-02 1.34072363e+00 1.25130308e+00
3.95187140e-01 8.59908700e-01 1.18013990e+00 -5.41917145e-01
1.37484848e-01 4.65713352e-01 5.02427757e-01 5.62562525e-01
4.56286877e-01 2.35474721e-01 -3.42198402e-01 4.82377291e-01
1.18505514e+00 2.24505290e-01 -4.47446138e-01 -3.59304577e-01
-1.35037005e+00 6.61830842e-01 9.44949090e-01 7.28194833e-01
-3.33634168e-01 1.71053782e-01 3.49041462e-01 4.16377872e-01
-6.32059932e-01 3.10429513e-01 -3.41531783e-01 4.73548293e-01
-5.82236826e-01 -2.80961454e-01 7.17211485e-01 8.18298221e-01
6.63420916e-01 5.42807996e-01 -2.74042934e-01 8.05578768e-01
4.61182147e-01 4.11653101e-01 1.06509280e+00 -9.11970496e-01
-7.43624121e-02 7.97567964e-01 -5.90347171e-01 -9.00971532e-01
-7.16836095e-01 -7.13647246e-01 -1.40855873e+00 6.94902420e-01
4.34315056e-01 -1.58203825e-01 -1.15297377e+00 1.66981316e+00
-4.30635095e-01 -2.34601587e-01 2.04171315e-01 9.07575607e-01
1.31752372e+00 6.84406042e-01 -4.79114987e-02 8.95786732e-02
1.32868600e+00 -1.25428557e+00 -5.67416668e-01 -3.11753303e-01
4.35132235e-01 -2.65672326e-01 8.12828600e-01 7.52805829e-01
-9.35691595e-01 -9.63741899e-01 -1.08188045e+00 3.29737342e-03
-5.92654407e-01 3.91097456e-01 8.11989129e-01 5.23373663e-01
-1.56415391e+00 2.29723632e-01 -3.43970388e-01 -4.90774006e-01
6.65395856e-01 5.85091949e-01 -4.60790247e-01 3.95924211e-01
-9.23394322e-01 1.28876209e+00 8.19627941e-01 5.68237960e-01
-1.22033608e+00 -2.56082952e-01 -4.78506595e-01 3.65721762e-01
2.81382632e-02 -1.00621581e+00 1.01674151e+00 -1.33413672e+00
-1.30955184e+00 7.74628520e-01 1.15500623e-02 -7.88502395e-01
2.16944650e-01 1.95365682e-01 -1.85301542e-01 1.93896648e-02
-2.72317976e-01 1.20600188e+00 8.19835782e-01 -1.46723270e+00
-4.38861817e-01 -1.53625950e-01 1.17749624e-01 -1.89330176e-01
-3.60626042e-01 -2.23631546e-01 -4.31444854e-01 -3.18572193e-01
2.69962102e-01 -7.05449820e-01 -2.88567454e-01 1.68926105e-01
-3.45748007e-01 -3.18271108e-02 6.50236189e-01 -3.32650721e-01
8.58754933e-01 -2.16262674e+00 -2.14244192e-03 2.17610657e-01
8.24274063e-01 6.34714782e-01 -2.73454875e-01 2.65843898e-01
-4.60411310e-01 3.22920457e-02 -3.75492498e-02 2.81004876e-01
-1.02377407e-01 4.51371223e-01 -1.90515906e-01 1.29080743e-01
-7.20835924e-02 1.18508935e+00 -4.44939286e-01 -4.36667472e-01
2.32280239e-01 3.71412575e-01 -3.17999423e-01 -2.70470023e-01
1.81714341e-01 1.54498130e-01 -3.38503152e-01 4.01605636e-01
7.87725031e-01 -2.30152950e-01 -4.64671701e-02 -5.38289011e-01
-2.66367912e-01 -2.44931757e-01 -7.98900962e-01 1.50826895e+00
-3.73369068e-01 1.13676536e+00 1.58893526e-01 -1.66194093e+00
8.66900384e-01 3.13479215e-01 1.42864790e-02 -1.07849240e+00
4.93679523e-01 -6.70623034e-02 8.44595075e-01 -6.82536185e-01
-2.69313097e-01 8.72864649e-02 5.60474336e-01 2.11350888e-01
5.33137798e-01 -1.02544278e-02 6.79697394e-02 3.44667047e-01
1.02874243e+00 -1.11225486e-01 3.87137741e-01 -2.24730983e-01
6.06212139e-01 -1.38250306e-01 4.04752225e-01 1.10115016e+00
-1.63508549e-01 5.32998800e-01 6.01210058e-01 -6.44603610e-01
-9.79380310e-01 -1.19149637e+00 -2.07346946e-01 5.64999700e-01
4.33867313e-02 -2.33590573e-01 -8.56268466e-01 -3.15013856e-01
-3.16381842e-01 4.04976577e-01 -8.04432213e-01 -2.14693680e-01
-2.94756323e-01 -7.99749255e-01 5.81008315e-01 5.54903090e-01
1.23399115e+00 -1.74459803e+00 -7.09713697e-01 8.03485364e-02
2.54010051e-01 -9.90232527e-01 4.79085773e-01 3.95124465e-01
-8.61692250e-01 -9.63695168e-01 -7.36970007e-01 -1.15824580e+00
8.26958895e-01 3.69400203e-01 9.69503403e-01 1.48505792e-01
-5.31022489e-01 1.41906664e-01 -1.84219286e-01 -9.46923614e-01
-1.84284031e-01 -1.84375271e-01 -7.71965384e-02 -1.56699538e-01
4.55786169e-01 -7.48205841e-01 -6.32648647e-01 2.40143716e-01
-8.28065336e-01 3.86666864e-01 1.23557699e+00 1.01778245e+00
5.55162467e-02 8.68599564e-02 5.46830118e-01 -6.86824381e-01
5.01347303e-01 -9.00013670e-02 -4.67198581e-01 1.96561009e-01
-4.77192193e-01 1.20034911e-01 6.47581041e-01 -2.82993883e-01
-8.79588544e-01 7.27141276e-02 -2.39197955e-01 -2.29272082e-01
-2.82880098e-01 3.87879342e-01 -8.30957144e-02 -3.42184514e-01
7.90370405e-01 6.46585107e-01 2.45782752e-02 -2.82130331e-01
2.29514420e-01 3.92728984e-01 3.95418912e-01 -5.27418591e-02
6.47877097e-01 4.08959657e-01 -1.34024054e-01 -9.46485341e-01
-4.52255487e-01 -1.44751766e-03 -6.77969575e-01 -3.16935956e-01
8.91937137e-01 -8.61474395e-01 -1.18828845e+00 8.02887857e-01
-1.46762955e+00 -2.47726947e-01 -4.30895984e-01 7.61479557e-01
-2.49534324e-01 1.99146140e-02 -7.37667024e-01 -5.89658916e-01
-2.39048332e-01 -1.07300329e+00 3.43158066e-01 3.81524354e-01
9.36301947e-02 -1.05984902e+00 -1.07446313e-01 -3.51658426e-02
4.43068147e-01 -3.40736747e-01 8.43754053e-01 -4.60279256e-01
-5.38626730e-01 4.34533022e-02 -7.07855165e-01 8.07336211e-01
-1.80623934e-01 -2.51183603e-02 -1.46191633e+00 1.38087356e-02
3.98155808e-01 -2.39893764e-01 1.46762955e+00 6.83267772e-01
1.29581869e+00 -3.46016496e-01 -5.72165489e-01 6.00659370e-01
1.68624687e+00 5.99420786e-01 1.15859258e+00 5.23922682e-01
6.22973442e-01 6.15698457e-01 -4.03278142e-01 -1.23046890e-01
-6.44767657e-02 2.10016116e-01 6.09755754e-01 -5.80694914e-01
-2.78123796e-01 2.19763502e-01 4.43367630e-01 9.41816986e-01
-3.10695440e-01 1.67369526e-02 -7.76077032e-01 4.45100337e-01
-1.73925841e+00 -1.30828440e+00 -3.41774821e-01 1.66103470e+00
6.15108073e-01 4.32996005e-01 -1.96166992e-01 1.92532599e-01
5.88669538e-01 -1.39260143e-01 -4.62055802e-01 -5.20788491e-01
-4.66057390e-01 1.08679891e-01 3.46828759e-01 -1.09754056e-02
-5.82260787e-01 8.80072474e-01 7.18208027e+00 8.51035416e-01
-1.18479311e+00 6.00964129e-02 3.72784168e-01 8.51576403e-02
8.64563435e-02 -3.56162339e-02 -7.08883166e-01 2.83355623e-01
5.54778159e-01 1.94296092e-01 3.79727989e-01 6.68442369e-01
4.10285257e-02 -2.90614873e-01 -1.13828373e+00 1.40933454e+00
1.24863628e-02 -1.49797225e+00 3.02351356e-01 3.12182736e-02
5.12386084e-01 2.78441578e-01 -2.63171475e-02 4.79984760e-01
5.12876391e-01 -1.29183292e+00 7.39243031e-01 5.84367037e-01
1.88842997e-01 -1.74739867e-01 8.30260396e-01 5.03422976e-01
-7.91959167e-01 -5.18495142e-01 -8.10741782e-01 -2.90426046e-01
-2.83656299e-01 5.93069851e-01 -6.52633667e-01 3.02894384e-01
9.67720985e-01 7.29737818e-01 -8.21930170e-01 1.33054972e+00
-3.01577002e-01 5.88958919e-01 -4.53115851e-02 -2.37331703e-01
5.23109794e-01 -3.06111455e-01 1.87838629e-01 1.32577538e+00
1.81698248e-01 -1.22854762e-01 -5.36514461e-01 1.37162888e+00
9.38341320e-02 -2.01877862e-01 -1.05833900e+00 3.15541089e-01
2.75719762e-01 1.55584371e+00 -8.47493172e-01 -4.29138184e-01
-6.63814306e-01 7.68837810e-01 1.56576782e-01 6.19057357e-01
-8.02965760e-01 -3.70841175e-01 6.83966801e-02 -1.20763429e-01
4.21556652e-01 1.28837842e-02 -5.52349925e-01 -9.64212239e-01
-1.95641592e-01 -7.16910124e-01 -2.21564174e-01 -1.11405349e+00
-9.78276968e-01 9.25132453e-01 -3.03705961e-01 -1.19356680e+00
2.23969787e-01 -1.29683423e+00 -7.97937334e-01 7.89202750e-01
-1.08086288e+00 -1.03167903e+00 -3.49645227e-01 9.18128848e-01
4.39432651e-01 -6.18670225e-01 8.53095472e-01 -1.35106612e-02
-7.70262718e-01 2.52076626e-01 4.33491617e-02 7.46917009e-01
2.78087348e-01 -1.06323719e+00 1.86361209e-01 7.53850460e-01
5.27283907e-01 8.31144214e-01 3.77388686e-01 4.27328497e-02
-9.30125952e-01 -7.25536585e-01 3.30200911e-01 -1.93495706e-01
3.89336497e-01 -5.29058695e-01 -7.50845075e-01 6.65133715e-01
7.53912032e-01 -1.35551929e-01 2.39274740e-01 4.59048152e-02
-3.66494656e-01 -4.68126714e-01 -8.71938229e-01 8.57065976e-01
1.06825531e+00 -2.23488271e-01 -9.55861151e-01 7.54355500e-03
4.42745745e-01 3.07702303e-01 -2.76880115e-01 1.42600417e-01
7.83505797e-01 -1.36195755e+00 1.01705897e+00 -3.52085382e-01
2.59745896e-01 -3.35703105e-01 1.93964362e-01 -1.56100667e+00
-6.89406335e-01 -7.40076303e-02 2.42426813e-01 8.62530708e-01
6.15073621e-01 -9.33496594e-01 6.09956443e-01 3.83566953e-02
-2.62895465e-01 -5.00956476e-01 -5.64160168e-01 -9.00761485e-01
-1.20705394e-02 -2.33885929e-01 1.38971299e-01 7.41612434e-01
-1.14310309e-01 9.44404006e-01 -2.60599293e-02 -2.84158587e-01
4.75275278e-01 -2.70521075e-01 5.07874012e-01 -1.47382784e+00
-2.52520978e-01 -8.65504205e-01 -6.75833106e-01 -9.30523217e-01
-3.20885777e-02 -8.46018255e-01 -1.83258131e-01 -2.08603978e+00
5.03017545e-01 -1.94267616e-01 -2.94889599e-01 8.41563642e-01
5.26494026e-01 3.76378089e-01 4.19579536e-01 2.22787842e-01
-1.63264543e-01 3.08638781e-01 1.57978189e+00 -5.34459770e-01
-3.02925725e-02 -1.00126512e-01 -8.71269763e-01 1.15289068e+00
9.46210384e-01 -1.95595980e-01 -7.06214249e-01 -7.02086926e-01
3.11961055e-01 -3.66143674e-01 7.69461632e-01 -1.48555076e+00
5.31672180e-01 4.28205043e-01 1.04010808e+00 -4.47938472e-01
3.58292609e-01 -1.18530583e+00 8.59616920e-02 8.62262607e-01
-1.48916379e-01 -1.91095680e-01 3.02735150e-01 1.42940402e-01
-3.60826463e-01 -5.91006875e-01 8.28036845e-01 -6.66362464e-01
-1.07410777e+00 9.15575027e-02 -8.49469066e-01 -3.43106657e-01
9.69665647e-01 -9.52926934e-01 -4.87950981e-01 -4.22035716e-02
-1.02751994e+00 1.79471038e-02 -1.56298757e-01 1.31990597e-01
9.50557947e-01 -1.18742263e+00 -5.64454198e-01 4.44835573e-01
-2.04040214e-01 -3.60612199e-02 1.25863835e-01 1.04194319e+00
-5.84645092e-01 6.81110561e-01 -9.77265239e-01 -6.40229464e-01
-9.91072237e-01 8.80952001e-01 6.30958259e-01 1.70432866e-01
-7.63651013e-01 9.73061979e-01 1.10343134e+00 -2.86237508e-01
3.80565107e-01 -5.33329248e-01 -8.33925962e-01 6.82994425e-02
5.29369056e-01 -5.51462583e-02 -5.63639924e-02 -1.26467630e-01
-1.39105439e-01 7.54893720e-01 -2.70431757e-01 -9.43976194e-02
1.29959166e+00 2.68333405e-02 -5.18562496e-01 4.68428701e-01
9.03889179e-01 -3.52595747e-01 -8.69364917e-01 -9.54978690e-02
-3.15356493e-01 8.85039195e-02 -3.17991450e-02 -1.04472756e+00
-1.52375185e+00 1.19359422e+00 7.22856939e-01 2.49601901e-01
1.36648834e+00 -1.54552475e-01 2.83011705e-01 8.23586047e-01
4.28968936e-01 -1.01850235e+00 1.29739150e-01 6.36674225e-01
1.13800693e+00 -1.05768919e+00 -2.37225235e-01 -7.72073120e-02
-6.82713509e-01 1.18450630e+00 9.95163500e-01 -2.63627291e-01
9.16739404e-01 4.96999294e-01 3.19590308e-02 -3.98247331e-01
-7.91814625e-01 -1.27632782e-01 -1.10224843e-01 9.30065036e-01
1.58160761e-01 -2.46643335e-01 -3.26590925e-01 6.89783871e-01
-2.32658446e-01 9.20322984e-02 5.22323668e-01 5.44660449e-01
-8.09210062e-01 -7.25748837e-01 -4.08755511e-01 4.21127677e-01
-8.06406066e-02 -4.12674487e-01 -4.68407184e-01 1.09404552e+00
6.22146487e-01 6.00214541e-01 1.05436370e-01 -6.19455457e-01
1.74452469e-01 -2.43033931e-01 7.11017549e-01 -3.48053426e-01
-6.57639265e-01 -8.74398723e-02 -3.17519784e-01 -4.77849126e-01
-5.50333798e-01 -2.76959799e-02 -1.37713337e+00 -3.31208527e-01
-1.97601259e-01 1.69218510e-01 6.21622503e-01 7.61845648e-01
-6.75752535e-02 8.70752752e-01 3.15941542e-01 -6.94056571e-01
-1.85841903e-01 -1.12341642e+00 -8.18129003e-01 -6.79581314e-02
3.05542767e-01 -5.35252631e-01 -2.25468144e-01 2.46059850e-01] | [9.58764362335205, 2.2519516944885254] |
e6c79336-5c81-4966-9fcd-11f83f715578 | embedding-open-domain-common-sense-knowledge | null | null | https://aclanthology.org/L16-1732 | https://aclanthology.org/L16-1732.pdf | Embedding Open-domain Common-sense Knowledge from Text | Our ability to understand language often relies on common-sense knowledge ― background information the speaker can assume is known by the reader. Similarly, our comprehension of the language used in complex domains relies on access to domain-specific knowledge. Capturing common-sense and domain-specific knowledge can be achieved by taking advantage of recent advances in open information extraction (IE) techniques and, more importantly, of knowledge embeddings, which are multi-dimensional representations of concepts and relations. Building a knowledge graph for representing common-sense knowledge in which concepts discerned from noun phrases are cast as vertices and lexicalized relations are cast as edges leads to learning the embeddings of common-sense knowledge accounting for semantic compositionality as well as implied knowledge. Common-sense knowledge is acquired from a vast collection of blogs and books as well as from WordNet. Similarly, medical knowledge is learned from two large sets of electronic health records. The evaluation results of these two forms of knowledge are promising: the same knowledge acquisition methodology based on learning knowledge embeddings works well both for common-sense knowledge and for medical knowledge Interestingly, the common-sense knowledge that we have acquired was evaluated as being less neutral than than the medical knowledge, as it often reflected the opinion of the knowledge utterer. In addition, the acquired medical knowledge was evaluated as more plausible than the common-sense knowledge, reflecting the complexity of acquiring common-sense knowledge due to the pragmatics and economicity of language. | ['a', 'Travis Goodwin', 'S Harabagiu'] | 2016-05-01 | embedding-open-domain-common-sense-knowledge-1 | https://aclanthology.org/L16-1732 | https://aclanthology.org/L16-1732.pdf | lrec-2016-5 | ['open-information-extraction'] | ['natural-language-processing'] | [-8.64297003e-02 6.74907982e-01 -4.01991934e-01 -2.53052115e-01
-2.45498285e-01 -7.52244711e-01 6.14740968e-01 9.26930726e-01
-7.31465697e-01 6.26865745e-01 7.55220413e-01 -3.73615503e-01
-4.95653808e-01 -1.04952478e+00 -5.06988943e-01 -2.84780234e-01
7.85883293e-02 4.88500446e-01 -9.57928672e-02 -6.31701171e-01
-8.09069499e-02 4.86696362e-02 -1.21107852e+00 2.91505069e-01
9.08785343e-01 6.65671825e-01 2.42810488e-01 2.97432810e-01
-8.22822571e-01 7.84915149e-01 -4.12149429e-01 -9.40092742e-01
-2.22440973e-01 2.74034571e-02 -1.42146015e+00 -3.00619215e-01
-8.97439942e-02 2.67579943e-01 -1.69709086e-01 1.11517048e+00
2.47141212e-01 7.58590251e-02 6.70954406e-01 -7.21201062e-01
-1.21808481e+00 8.72784793e-01 3.74385603e-02 1.81912199e-01
8.20544899e-01 -2.18306676e-01 1.37555945e+00 -6.28447413e-01
9.46860194e-01 1.26346481e+00 5.36474288e-01 5.92433333e-01
-9.45767224e-01 -2.69727945e-01 2.70245910e-01 3.54238302e-01
-1.43216503e+00 1.23080827e-01 6.81403220e-01 -6.89701617e-01
1.04855418e+00 1.14340007e-01 9.42980766e-01 1.16217518e+00
1.69607416e-01 6.45891249e-01 9.04215097e-01 -6.00550830e-01
1.82683766e-01 8.80069911e-01 5.87020636e-01 5.52384496e-01
5.39016068e-01 9.92499962e-02 -5.24106562e-01 -2.96276122e-01
5.81116676e-01 2.15335354e-01 -6.95681036e-01 -2.29153976e-01
-1.27886486e+00 9.33489084e-01 5.45859098e-01 6.92339420e-01
-5.86197436e-01 -3.43418926e-01 4.90940690e-01 4.70432043e-01
3.05558234e-01 9.85028327e-01 -9.18826997e-01 -9.05705169e-02
-5.16550958e-01 -3.63058485e-02 1.47698975e+00 7.76858628e-01
9.59284723e-01 -4.21542108e-01 1.24111451e-01 7.77711511e-01
4.21025962e-01 4.10927802e-01 1.01695073e+00 -4.11152065e-01
-7.92735815e-02 1.04223800e+00 -1.04370266e-01 -1.22147286e+00
-4.22376335e-01 -5.22090435e-01 -5.89001477e-01 -3.66275579e-01
2.14697897e-01 -1.35602176e-01 -9.12718475e-01 1.96866858e+00
1.75518498e-01 -1.91977303e-02 6.19616807e-01 7.02781618e-01
1.21902251e+00 1.29076868e-01 2.47392640e-01 -1.15719415e-01
2.04391742e+00 -2.52589017e-01 -1.04549885e+00 -3.53474349e-01
6.91925049e-01 -4.59457189e-01 9.75678861e-01 -4.35543470e-02
-4.51484501e-01 -1.40481248e-01 -1.00729322e+00 -7.95764774e-02
-1.12916410e+00 -5.91509104e-01 7.10591853e-01 5.53968608e-01
-9.46578622e-01 2.71286249e-01 -3.12722236e-01 -6.44331992e-01
3.46272469e-01 -3.62986997e-02 -5.97166717e-01 -3.12797666e-01
-1.91669607e+00 1.35475957e+00 7.70869136e-01 -2.59404510e-01
-3.37072402e-01 -1.11049688e+00 -1.29043722e+00 1.03148215e-01
7.12928593e-01 -1.23653138e+00 8.59039545e-01 -9.31917608e-01
-1.08475208e+00 1.18636286e+00 -2.92004598e-03 -2.03945115e-01
-3.39421406e-02 -1.80321187e-01 -9.06596482e-01 1.04889967e-01
7.06230626e-02 1.21896252e-01 4.28849518e-01 -1.21277511e+00
-3.19233924e-01 -4.81119365e-01 4.15222049e-01 2.82879859e-01
-3.67628962e-01 -2.78147668e-01 -1.09734133e-01 -5.74982762e-01
-1.20465513e-02 -4.40899938e-01 -1.45266742e-01 -1.99963123e-01
-3.89000744e-01 -1.40850991e-01 4.83067214e-01 -5.57353973e-01
1.44798040e+00 -2.11757469e+00 1.80378497e-01 4.46859032e-01
7.09766448e-01 1.98975220e-01 1.87275752e-01 6.58001840e-01
-2.37358809e-01 4.57520068e-01 -1.75612152e-01 5.19964337e-01
2.30939209e-01 7.36826360e-01 -3.98087919e-01 -8.04439262e-02
1.02367461e-01 1.04895449e+00 -1.35886800e+00 -2.94160455e-01
1.32018730e-01 5.01557827e-01 -5.00395060e-01 -1.98779479e-02
-2.77600437e-01 3.35705690e-02 -6.57508254e-01 3.66532534e-01
1.61395207e-01 -5.13334572e-01 5.01536906e-01 -3.28935415e-01
3.75384092e-01 3.69544804e-01 -1.08731890e+00 1.66692293e+00
-8.45019758e-01 3.66967231e-01 -1.76603541e-01 -8.64707947e-01
6.15637004e-01 7.00714171e-01 3.07578027e-01 -2.65340805e-01
1.61263898e-01 9.97059867e-02 -8.20093974e-02 -8.59204948e-01
4.27385926e-01 -6.75044835e-01 -1.07266970e-01 3.24239373e-01
3.31506521e-01 -2.44618446e-01 -1.16504289e-01 5.83614171e-01
9.38060343e-01 -3.00732076e-01 9.37329471e-01 -5.80260932e-01
5.19785225e-01 2.50319749e-01 4.65562373e-01 4.76722807e-01
3.84418182e-02 -8.77026394e-02 4.38424021e-01 -5.12836874e-01
-3.45351219e-01 -1.22790051e+00 -3.55541587e-01 8.52903962e-01
2.99649648e-02 -8.08977067e-01 -4.42616165e-01 -6.66443527e-01
1.51875392e-01 9.36092436e-01 -7.89997339e-01 -1.97139025e-01
-8.13193694e-02 -3.79145086e-01 4.09904152e-01 5.36023557e-01
-1.25106294e-02 -1.06219816e+00 -5.79411387e-01 2.10804492e-01
-1.50887474e-01 -1.16924107e+00 -2.75215536e-01 2.61035472e-01
-5.28908968e-01 -1.52973187e+00 -4.71402973e-01 -7.60785043e-01
6.24741375e-01 -3.69786412e-01 1.55090606e+00 8.76993984e-02
-3.45404088e-01 1.00583255e+00 -5.73037684e-01 -7.68833756e-01
-4.24845248e-01 -1.69600099e-01 1.45779610e-01 -1.24569595e-01
9.99522328e-01 -3.80491972e-01 -3.81937295e-01 -2.00474337e-01
-1.27893639e+00 -2.70567268e-01 6.47248387e-01 9.02686059e-01
4.64939564e-01 1.65695173e-03 5.89508355e-01 -1.31064343e+00
1.03170729e+00 -7.76101470e-01 1.31126344e-01 4.90384191e-01
-5.93393981e-01 4.17367667e-01 2.65043467e-01 -4.10078287e-01
-1.07453322e+00 -4.72524434e-01 -1.54836386e-01 -7.08551854e-02
-3.63685459e-01 1.16605699e+00 1.86144449e-02 2.49374598e-01
8.06594491e-01 1.16106592e-01 -5.57843708e-02 -2.86824226e-01
9.34451163e-01 7.27714896e-01 2.81279087e-01 -7.00649261e-01
4.69297796e-01 3.50982308e-01 -4.27757740e-01 -1.24983597e+00
-1.10252523e+00 -6.39210463e-01 -5.46834111e-01 3.51235569e-01
1.17731822e+00 -6.67247951e-01 -6.58387899e-01 -2.05694661e-02
-1.09314787e+00 1.50472790e-01 -7.22023368e-01 7.17612863e-01
-1.99593142e-01 3.30291152e-01 -3.39590788e-01 -4.50974822e-01
-1.82335347e-01 -7.97354937e-01 5.85869610e-01 -4.07661684e-02
-7.52279818e-01 -1.73318160e+00 1.28778458e-01 1.52664259e-01
4.04876381e-01 1.02564387e-01 1.63416004e+00 -1.17872536e+00
-1.32966772e-01 -7.27799758e-02 9.16355662e-03 2.80753970e-01
5.85365951e-01 -4.83073175e-01 -9.20229793e-01 4.31392081e-02
1.76774204e-01 -2.10709125e-01 6.51284754e-01 1.19723357e-01
6.64851069e-01 -4.02207404e-01 -4.02482361e-01 5.29705994e-02
1.53296554e+00 -4.35364582e-02 3.50771874e-01 4.92180884e-02
6.54686391e-01 8.11149776e-01 -5.25693633e-02 2.59331942e-01
8.59192550e-01 3.33963811e-01 -2.10775197e-01 1.02715991e-01
1.23805098e-01 -2.62324303e-01 5.71476258e-02 1.09001470e+00
2.17190951e-01 1.81061015e-01 -1.40579963e+00 8.10754895e-01
-1.49776649e+00 -8.64835858e-01 2.61979997e-01 1.99939179e+00
1.55579662e+00 -2.91041080e-02 -3.12379748e-01 -3.94063331e-02
3.37464750e-01 -9.21380222e-02 -2.70939082e-01 -5.66078007e-01
-6.59232885e-02 2.79703289e-01 2.00127795e-01 6.80507421e-01
-5.74628413e-01 8.58279228e-01 6.63393164e+00 5.82403958e-01
-1.03613675e+00 7.71583095e-02 -5.52869700e-02 3.39592904e-01
-1.00771844e+00 -2.24520657e-02 -3.79514366e-01 9.73381847e-02
6.86571062e-01 -5.56063533e-01 8.23385343e-02 6.33421302e-01
-4.49650675e-01 -2.21883338e-02 -1.40633404e+00 1.11050367e+00
1.32219747e-01 -1.43117201e+00 3.23355049e-01 -1.23856023e-01
4.48666275e-01 -5.90909310e-02 -1.21698357e-01 4.24295783e-01
6.67944610e-01 -1.11712873e+00 2.06440732e-01 6.99832439e-01
7.60719776e-01 -3.57400835e-01 1.02290142e+00 3.39189082e-01
-1.05549777e+00 7.71342069e-02 -1.19671457e-01 -3.15738112e-01
3.15647274e-02 7.22945929e-01 -8.38450968e-01 8.65981460e-01
4.87983167e-01 6.77975237e-01 -2.94853866e-01 4.36578274e-01
-5.72134316e-01 3.88611734e-01 -2.15918601e-01 -1.34041384e-01
1.69788927e-01 8.82441029e-02 3.22570503e-01 1.32783401e+00
-1.41963109e-01 4.08618301e-01 2.98981607e-01 9.80193973e-01
-8.94025862e-02 4.30607349e-01 -9.26401794e-01 -4.67932671e-01
3.64003122e-01 8.60250950e-01 -2.64900506e-01 -5.86992204e-01
-7.97744691e-01 5.77509105e-01 2.49257773e-01 5.47187805e-01
-7.75815323e-02 -4.14494514e-01 9.15618777e-01 1.29390145e-02
1.63436364e-02 1.50458425e-01 -3.04397568e-02 -1.36255443e+00
2.73839887e-02 -6.38683081e-01 7.52937853e-01 -4.74645019e-01
-1.78575170e+00 7.37553477e-01 1.33852333e-01 -6.64161682e-01
-2.66962200e-01 -8.18805814e-01 -2.02431738e-01 9.28548694e-01
-1.82108891e+00 -9.58683610e-01 -5.94825018e-03 7.75389969e-01
1.10176206e-01 -7.15294704e-02 1.40319002e+00 1.25038520e-01
3.69300433e-02 4.95369822e-01 -3.15557629e-01 3.77717257e-01
5.71832240e-01 -1.33313370e+00 -1.00084461e-01 1.23137824e-01
3.16026300e-01 1.28566408e+00 6.82816148e-01 -6.95862949e-01
-1.33360362e+00 -6.07280672e-01 1.28519213e+00 -8.14617813e-01
1.01386845e+00 6.96574198e-03 -1.29779577e+00 8.59057009e-01
2.98715055e-01 -2.61900965e-02 1.48618746e+00 6.14400864e-01
-7.63834834e-01 2.22108349e-01 -1.24621701e+00 5.11732459e-01
8.33912671e-01 -1.02371514e+00 -1.61401534e+00 1.46614671e-01
9.83550310e-01 -2.67061651e-01 -1.23403847e+00 1.06730349e-01
3.75803798e-01 -2.79342741e-01 8.28121305e-01 -1.04154968e+00
3.36366385e-01 -1.64784163e-01 -4.23684746e-01 -1.69156504e+00
-1.27504513e-01 -2.84588635e-01 -1.15320697e-01 7.78370976e-01
7.83299446e-01 -7.90720880e-01 2.26579547e-01 8.49691153e-01
1.05107754e-01 -8.44684482e-01 -7.80451000e-01 -6.07459307e-01
3.17625761e-01 -6.22650146e-01 6.17069006e-01 1.68798602e+00
7.05655098e-01 5.41855514e-01 9.76864174e-02 1.48299441e-01
2.67336935e-01 7.68565163e-02 9.15711522e-02 -1.53402638e+00
-2.16518089e-01 -3.14399868e-01 -8.35508645e-01 -7.11407304e-01
3.62552345e-01 -1.36375153e+00 -4.65638459e-01 -1.76947653e+00
2.28915393e-01 -2.40191162e-01 -4.57144082e-01 5.68019509e-01
-3.31929117e-01 -4.79536861e-01 1.60070751e-02 -2.85540015e-01
-2.21827060e-01 2.97254920e-01 1.35489178e+00 -2.47177750e-01
-1.15110889e-01 -4.99253631e-01 -1.23933887e+00 1.09806299e+00
3.36643487e-01 -3.49879563e-01 -6.09061241e-01 -4.68624115e-01
9.43787456e-01 2.66270433e-02 4.02285010e-01 -2.58604825e-01
3.70924532e-01 -2.23997369e-01 7.67702237e-03 3.92183699e-02
9.47924405e-02 -1.12398088e+00 2.49395706e-03 4.64869559e-01
-3.97207826e-01 -1.76179662e-01 2.91076273e-01 7.53129423e-01
-5.08636475e-01 -1.82470605e-01 3.73553455e-01 -4.94813055e-01
-1.38751090e+00 1.98863912e-02 -1.89202741e-01 7.24075317e-01
6.94337964e-01 -3.31388116e-01 -2.19178811e-01 -3.38622153e-01
-1.13878810e+00 2.49489650e-01 2.01420084e-01 7.52616882e-01
8.91458511e-01 -1.18276799e+00 -6.57348216e-01 2.69051105e-01
5.98020494e-01 -2.37040773e-01 1.87749356e-01 6.06301606e-01
-1.37281731e-01 4.35382545e-01 5.55603579e-02 -3.08081895e-01
-8.10592711e-01 8.40107918e-01 2.68188864e-01 -1.19851604e-01
-7.41651714e-01 8.44802499e-01 3.51463437e-01 -6.38538182e-01
1.07648998e-01 -6.49882197e-01 -4.77283448e-01 3.57855827e-01
5.93366981e-01 -4.13629003e-02 -1.40772521e-01 -5.50905824e-01
-7.48163402e-01 6.49515986e-01 -1.90768808e-01 -1.46716898e-02
1.08200967e+00 -3.48588601e-02 -3.51764530e-01 6.77345872e-01
1.08355331e+00 5.57637811e-02 -2.70331074e-02 -7.98393309e-01
3.30983967e-01 -2.13399351e-01 1.56776588e-02 -1.24306691e+00
-6.63744569e-01 6.51504874e-01 1.93765879e-01 -3.25205624e-02
7.04458058e-01 4.05537933e-01 5.24928272e-01 6.87005520e-01
4.32513326e-01 -9.98621166e-01 -1.64115354e-01 6.74753308e-01
8.40900421e-01 -1.30580652e+00 -1.12911738e-01 -5.40565968e-01
-9.51486170e-01 1.03939176e+00 2.35738307e-01 4.97258872e-01
1.20582104e+00 6.76239729e-02 5.78502595e-01 -8.91369581e-01
-5.30909717e-01 -3.56168926e-01 5.69126904e-01 6.84450626e-01
6.82247281e-01 3.70485336e-01 -2.97841549e-01 1.10200846e+00
-4.12616521e-01 2.54500173e-02 4.37107742e-01 7.54521191e-01
-1.85466632e-01 -9.75501180e-01 5.63176088e-02 5.05350113e-01
-5.15621006e-01 -3.54457855e-01 -5.24148226e-01 5.79321623e-01
1.88397422e-01 9.55320358e-01 -2.11416110e-01 -2.94302046e-01
3.83632600e-01 4.03247595e-01 1.89087003e-01 -1.10477221e+00
-5.71553469e-01 -6.32350981e-01 1.72910929e-01 -4.79273856e-01
-4.60831225e-01 5.06135672e-02 -1.50451767e+00 -5.47653995e-02
-2.10454121e-01 4.56006855e-01 3.76890361e-01 1.38876140e+00
2.27151141e-01 7.32711136e-01 -4.76113670e-02 1.70521110e-01
-5.47099769e-01 -6.47450447e-01 -6.90720081e-01 8.77075434e-01
3.43776256e-01 -5.54763138e-01 -1.48209795e-01 -1.50021523e-01] | [10.114314079284668, 8.718711853027344] |
647cd26d-ca23-48c7-acd0-790e2d2259f4 | tanet-robust-3d-object-detection-from-point | 1912.05163 | null | https://arxiv.org/abs/1912.05163v1 | https://arxiv.org/pdf/1912.05163v1.pdf | TANet: Robust 3D Object Detection from Point Clouds with Triple Attention | In this paper, we focus on exploring the robustness of the 3D object detection in point clouds, which has been rarely discussed in existing approaches. We observe two crucial phenomena: 1) the detection accuracy of the hard objects, e.g., Pedestrians, is unsatisfactory, 2) when adding additional noise points, the performance of existing approaches decreases rapidly. To alleviate these problems, a novel TANet is introduced in this paper, which mainly contains a Triple Attention (TA) module, and a Coarse-to-Fine Regression (CFR) module. By considering the channel-wise, point-wise and voxel-wise attention jointly, the TA module enhances the crucial information of the target while suppresses the unstable cloud points. Besides, the novel stacked TA further exploits the multi-level feature attention. In addition, the CFR module boosts the accuracy of localization without excessive computation cost. Experimental results on the validation set of KITTI dataset demonstrate that, in the challenging noisy cases, i.e., adding additional random noisy points around each object,the presented approach goes far beyond state-of-the-art approaches. Furthermore, for the 3D object detection task of the KITTI benchmark, our approach ranks the first place on Pedestrian class, by using the point clouds as the only input. The running speed is around 29 frames per second. | ['Xin Zhao', 'Yu Zhou', 'Zhe Liu', 'Ruolan Hu', 'Xiang Bai', 'Tengteng Huang'] | 2019-12-11 | null | null | null | null | ['robust-3d-object-detection'] | ['computer-vision'] | [-1.41036719e-01 -3.22428912e-01 3.41993511e-01 3.77535168e-03
-7.21813858e-01 -2.73697108e-01 5.75057924e-01 3.82355064e-01
-5.48723578e-01 4.18147117e-01 -4.18001711e-01 -9.49866176e-02
8.78744423e-02 -6.58643126e-01 -9.41317260e-01 -9.89483654e-01
4.92636934e-02 2.65331924e-01 7.08859622e-01 -2.37467214e-01
1.66965291e-01 6.51737750e-01 -1.88898790e+00 -1.62280351e-01
9.90429938e-01 1.33267224e+00 2.30794340e-01 2.72521526e-01
-1.01341501e-01 3.16914171e-01 -5.47714829e-01 -3.96699280e-01
2.87274063e-01 3.31041873e-01 -6.58926219e-02 1.38759434e-01
6.11725867e-01 -2.36339673e-01 -7.12795109e-02 1.26877022e+00
5.58367610e-01 -7.92672299e-03 3.99304241e-01 -1.23361790e+00
-3.25933605e-01 9.94125679e-02 -1.12594306e+00 3.31770986e-01
7.82935396e-02 4.46656644e-01 6.70524061e-01 -1.38308477e+00
2.07937866e-01 1.38246894e+00 6.35826051e-01 1.65996566e-01
-1.04910171e+00 -7.13823438e-01 5.78094065e-01 2.85487741e-01
-1.60285211e+00 -1.49595127e-01 8.37678611e-01 -5.56693912e-01
7.07117260e-01 2.16214433e-01 6.08877599e-01 9.02109206e-01
-8.26492682e-02 8.76756608e-01 9.02393639e-01 -1.28048629e-01
1.04657605e-01 9.54237282e-02 3.06781054e-01 3.45058858e-01
6.18883491e-01 2.15849504e-01 -1.34082720e-01 8.83639008e-02
4.91431773e-01 9.13918540e-02 -8.60698596e-02 -5.79111457e-01
-1.05347538e+00 5.20569980e-01 8.59343827e-01 2.18842328e-01
-3.65958422e-01 -4.01683673e-02 4.55724120e-01 -1.78620934e-01
6.58711970e-01 -5.90810888e-02 -2.60746837e-01 1.14295073e-01
-8.03040206e-01 3.31895083e-01 2.30973229e-01 1.01910031e+00
6.63307369e-01 -9.64702591e-02 -2.22399026e-01 5.62934637e-01
3.51165563e-01 5.91099739e-01 4.76194285e-02 -2.66796589e-01
7.97102273e-01 7.67985463e-01 4.68360215e-01 -1.07433856e+00
-3.89656037e-01 -9.19579744e-01 -8.83221567e-01 4.40080106e-01
4.94505733e-01 1.38066813e-01 -9.27127540e-01 1.40175712e+00
8.19339573e-01 3.34131747e-01 -3.18680376e-01 1.35630810e+00
1.13377976e+00 4.96032983e-01 1.13234617e-01 -9.23409760e-02
1.46957278e+00 -9.20055211e-01 -3.63432080e-01 -2.49315292e-01
3.03586423e-01 -8.13387036e-01 9.47322428e-01 3.25393081e-01
-1.01698172e+00 -9.27208245e-01 -1.01249802e+00 7.52391070e-02
-4.65149701e-01 4.18713450e-01 4.34342027e-01 5.66338420e-01
-7.03204215e-01 3.56599122e-01 -7.79567122e-01 -4.40497547e-01
6.60472751e-01 2.69706577e-01 -2.80928910e-01 -1.31966203e-01
-7.47139573e-01 7.75841236e-01 1.70751825e-01 3.60656857e-01
-7.09099650e-01 -8.10825348e-01 -6.15378320e-01 2.08894566e-01
6.63135350e-01 -5.60415685e-01 7.82347977e-01 -4.94308740e-01
-1.18022430e+00 6.36835575e-01 -1.71299219e-01 -3.35516036e-01
8.51732850e-01 -5.66265404e-01 -7.53997415e-02 -5.63334115e-02
2.89745212e-01 4.61561948e-01 9.77290034e-01 -1.55428803e+00
-9.11630869e-01 -7.18889475e-01 -4.03550230e-02 8.28669071e-02
-1.40246883e-01 3.51338349e-02 -8.27388942e-01 -5.42074084e-01
1.76365688e-01 -7.78123796e-01 -3.91439974e-01 2.87733879e-02
-4.53545690e-01 -4.28188652e-01 9.52585340e-01 -2.66684115e-01
9.35701549e-01 -2.44519711e+00 5.80872409e-02 1.64321378e-01
2.30473176e-01 4.97521192e-01 5.44976071e-02 -9.17447433e-02
-1.03199951e-01 9.51348171e-02 -7.52188414e-02 -6.08469367e-01
-1.20158708e-02 -1.70131922e-01 -2.21221432e-01 6.19505644e-01
5.98173916e-01 8.37996900e-01 -8.72289717e-01 -4.31468517e-01
6.10073686e-01 6.62746131e-01 -3.99241507e-01 -2.49342043e-02
9.91288051e-02 5.99286139e-01 -5.19126952e-01 9.13147807e-01
1.22729468e+00 -2.03753516e-01 -5.82842052e-01 -1.76503435e-01
-4.98433739e-01 -3.37670930e-03 -1.41490698e+00 1.32804310e+00
-2.13708743e-01 1.54374897e-01 1.24360196e-01 -7.81884193e-01
9.19331193e-01 -5.23310937e-02 4.40991193e-01 -6.21347249e-01
2.09960312e-01 2.53957629e-01 -5.96041605e-02 -2.42428333e-01
4.02309835e-01 1.79687113e-01 -4.41099368e-02 -1.80150896e-01
-2.10097536e-01 1.23515628e-01 1.17672011e-01 3.43260393e-02
9.37161088e-01 8.13222975e-02 1.31461009e-01 -1.24820456e-01
6.93860948e-01 -7.31860325e-02 5.35570264e-01 7.76181281e-01
-3.67654949e-01 6.58668876e-01 2.33611360e-01 -3.90920609e-01
-8.25269401e-01 -9.93549466e-01 -2.83276618e-01 7.86964536e-01
6.25121653e-01 -3.10072929e-01 -7.08411396e-01 -6.94937885e-01
3.44500959e-01 3.02542686e-01 -5.89172006e-01 2.26758402e-02
-5.37717223e-01 -9.06448781e-01 2.23531768e-01 6.75838590e-01
5.13626277e-01 -6.37280405e-01 -7.50443876e-01 2.33059689e-01
-8.72754306e-02 -1.44039798e+00 -1.80099502e-01 -2.55721528e-03
-9.20371890e-01 -9.02029097e-01 -7.74263918e-01 -3.81365538e-01
5.93885362e-01 8.25299382e-01 9.48939860e-01 2.82566190e-01
-2.59478986e-01 1.21417224e-01 -4.65904772e-01 -5.71414113e-01
2.99012125e-01 -3.28750052e-02 6.91219941e-02 2.42585197e-01
3.30864817e-01 -3.56704772e-01 -6.77819908e-01 5.42351544e-01
-3.83191019e-01 -1.95608214e-01 7.63797104e-01 8.35807860e-01
7.71453083e-01 1.65646359e-01 2.86665529e-01 -3.95075679e-01
2.15034205e-02 -4.13479894e-01 -8.65878522e-01 -1.40258491e-01
-2.88640540e-02 -3.90672207e-01 4.45883334e-01 -4.91324365e-01
-9.03585315e-01 3.74041796e-01 -2.23784268e-01 -7.00518847e-01
-3.87261152e-01 2.08788570e-02 -4.53550547e-01 -3.06257725e-01
4.81552690e-01 1.34309545e-01 -3.68674189e-01 -7.49081373e-01
1.55625820e-01 4.30783033e-01 5.21914899e-01 -4.55865055e-01
1.23334181e+00 7.26480722e-01 5.83976954e-02 -1.03970468e+00
-6.92596972e-01 -7.77803719e-01 -7.48331904e-01 -3.53408009e-01
7.46986747e-01 -1.04708910e+00 -8.97751689e-01 4.49780583e-01
-1.29497409e+00 1.61324903e-01 -2.57415891e-01 4.31451797e-01
-1.73040077e-01 5.47924042e-01 -2.60023206e-01 -1.12685788e+00
-2.33474061e-01 -1.38432050e+00 1.53540313e+00 2.32778877e-01
3.09928268e-01 -2.39188537e-01 -5.05951822e-01 3.64898592e-01
1.44137710e-01 3.38280678e-01 5.76753855e-01 -4.22477454e-01
-1.03736079e+00 -3.38760495e-01 -6.91228390e-01 1.68944940e-01
-9.31833163e-02 5.93716204e-02 -1.12776852e+00 -3.44006836e-01
-4.73076105e-02 9.25954208e-02 9.61734593e-01 3.54676247e-01
1.19760394e+00 1.89706400e-01 -4.83254880e-01 4.63026702e-01
1.24000251e+00 -1.13209046e-01 5.29388011e-01 3.52613926e-01
8.15972209e-01 4.55922186e-01 9.89954293e-01 5.11688054e-01
4.06770229e-01 9.08988655e-01 9.30944145e-01 -3.93446147e-01
-1.13057256e-01 -1.10019445e-02 1.90826610e-01 3.14873368e-01
-4.44528818e-01 8.03911034e-03 -8.19180071e-01 5.74279666e-01
-1.94299877e+00 -7.68345118e-01 -5.56575835e-01 2.35192108e+00
1.49235040e-01 4.75005984e-01 2.35909864e-01 4.72556323e-01
8.51216614e-01 -8.10831133e-03 -4.32527751e-01 3.59706432e-01
-7.75639117e-02 1.76105484e-01 6.25169516e-01 1.34531528e-01
-1.57083094e+00 9.21563387e-01 4.71794462e+00 1.04077089e+00
-1.00583625e+00 3.12365532e-01 3.16176593e-01 -3.71061563e-02
3.71453643e-01 -2.51300126e-01 -1.09692323e+00 7.55333483e-01
2.86321908e-01 3.32896084e-01 2.96601822e-04 1.10267162e+00
3.05124581e-01 -1.43133715e-01 -8.05284560e-01 1.07845366e+00
-6.74327612e-02 -8.81103754e-01 -1.32575676e-01 9.58128050e-02
3.24011922e-01 1.45309076e-01 1.10140599e-01 3.58130366e-01
-1.15240961e-01 -7.50586390e-01 1.03411150e+00 3.96575719e-01
3.71962458e-01 -8.87375832e-01 1.05093455e+00 5.05127192e-01
-1.54691386e+00 -1.78320616e-01 -5.22506535e-01 1.53444083e-02
1.59920290e-01 1.04008329e+00 -4.46015090e-01 8.07466805e-01
1.19867694e+00 5.90054035e-01 -9.06409800e-01 1.58100390e+00
-3.58160466e-01 2.77077883e-01 -6.67434275e-01 -5.62561713e-02
2.90881246e-01 -4.57363129e-02 9.71539855e-01 1.23448908e+00
3.06198329e-01 1.37551174e-01 2.36393377e-01 8.23030472e-01
1.69764161e-01 3.02637629e-02 -4.22384858e-01 6.42190933e-01
2.13343695e-01 1.38075697e+00 -8.05612504e-01 -2.93569237e-01
-3.82274210e-01 6.00265801e-01 2.62225121e-01 2.08835900e-01
-9.59463358e-01 -1.00605018e-01 7.37916112e-01 2.89334714e-01
8.04973125e-01 -3.27396303e-01 -5.33579350e-01 -1.18513942e+00
5.28377175e-01 -5.44925988e-01 7.24826381e-02 -5.68614304e-01
-1.35537052e+00 5.42140245e-01 -2.28977129e-01 -1.41146219e+00
3.87625366e-01 -6.33624017e-01 -6.36583149e-01 7.68401206e-01
-1.69475543e+00 -1.20771325e+00 -7.15886533e-01 5.04040420e-01
5.36930025e-01 1.99911311e-01 1.19280532e-01 7.15410292e-01
-8.08081210e-01 5.37447274e-01 -2.08823428e-01 -4.25441936e-02
4.77495462e-01 -1.07557499e+00 5.90059876e-01 1.01031983e+00
-1.54848129e-01 5.17865896e-01 6.87169850e-01 -6.73058987e-01
-1.33410466e+00 -1.30700719e+00 6.63297355e-01 -5.88336587e-01
4.78323042e-01 -6.00576580e-01 -1.06142044e+00 1.80500910e-01
-4.04418081e-01 4.15532768e-01 1.65593445e-01 8.38801563e-02
-1.93512142e-01 -2.07684264e-01 -1.07749093e+00 3.80046189e-01
1.17343521e+00 -5.13807917e-03 -3.37939084e-01 3.03551465e-01
6.98367238e-01 -6.07565045e-01 -6.20992661e-01 6.70899749e-01
3.12613398e-01 -1.02602804e+00 1.48929691e+00 -2.80824035e-01
1.27179876e-01 -8.81749392e-01 -1.52311429e-01 -9.98374462e-01
-3.96792024e-01 -1.51775852e-01 -2.60773927e-01 1.38953900e+00
9.60545242e-03 -3.94941837e-01 6.52335525e-01 1.93314105e-01
-2.25211501e-01 -7.27193713e-01 -1.34514976e+00 -9.64492381e-01
-1.95008859e-01 -6.37409866e-01 7.83087373e-01 4.83451605e-01
-6.10341728e-01 2.25024670e-01 -3.06087404e-01 7.01133907e-01
9.39586759e-01 7.08975941e-02 1.10052693e+00 -1.45554137e+00
-4.94776629e-02 -5.30406713e-01 -7.48211026e-01 -1.19470704e+00
-3.55815381e-01 -5.13825476e-01 1.52125105e-01 -1.18905115e+00
4.57540601e-02 -6.34516478e-01 -3.14178914e-01 1.55563742e-01
-5.91455519e-01 3.36592019e-01 6.28907561e-01 1.43420622e-01
-7.48474538e-01 7.08807886e-01 1.07117522e+00 -3.01821291e-01
-1.07106552e-01 3.22210819e-01 -5.95681190e-01 8.75108302e-01
4.65045512e-01 -5.07258177e-01 1.92920536e-01 -3.39822978e-01
-1.68571398e-01 -3.61496150e-01 9.62780535e-01 -1.24480999e+00
2.69155890e-01 1.29969180e-01 6.03850126e-01 -1.18283665e+00
6.01870537e-01 -9.59967613e-01 -1.48691654e-01 4.67740029e-01
4.72009778e-01 -1.97022259e-01 3.24088514e-01 6.88493311e-01
-1.67164579e-01 2.74633989e-02 1.03204286e+00 5.74359447e-02
-6.07673526e-01 4.02882189e-01 7.43539110e-02 -1.68189362e-01
1.13741171e+00 -2.27373049e-01 -2.34653041e-01 1.34029120e-01
-7.03994572e-01 4.15001094e-01 3.60397249e-01 4.94983077e-01
5.58251560e-01 -1.30933881e+00 -7.23688066e-01 2.36194655e-01
2.87829489e-01 4.21651483e-01 4.83686537e-01 1.08188224e+00
6.24870230e-03 2.89409637e-01 -5.40154008e-03 -1.23332012e+00
-1.27557600e+00 7.27708936e-01 1.13052577e-01 -2.74279416e-01
-7.23549962e-01 8.31545353e-01 4.80700672e-01 -1.56726375e-01
4.18618441e-01 -5.78815639e-01 -2.88870960e-01 2.48383895e-01
5.42055428e-01 4.96786177e-01 4.62262273e-01 -7.78331161e-01
-7.00436234e-01 9.93924558e-01 1.44597832e-02 3.90049040e-01
1.36836541e+00 -6.47574514e-02 2.71806747e-01 2.31824845e-01
7.23672628e-01 -8.46677348e-02 -1.33882117e+00 -1.75441712e-01
-6.34769127e-02 -8.49814355e-01 8.52533896e-03 -4.69323069e-01
-1.04789579e+00 1.17737806e+00 8.61375511e-01 1.00826219e-01
9.65151131e-01 2.82297079e-02 6.36161327e-01 3.90201993e-02
5.09205997e-01 -7.66317010e-01 -3.34029384e-02 4.65434790e-01
9.11057711e-01 -1.56888855e+00 4.12956290e-02 -7.20847309e-01
-3.44424397e-01 7.42518544e-01 7.04089105e-01 -2.08170697e-01
4.85307187e-01 1.82172731e-01 -3.21398079e-01 -7.35598058e-02
-3.62528831e-01 -6.48573756e-01 2.97930300e-01 5.67312479e-01
-1.63216174e-01 -2.50919518e-04 -1.39165431e-01 9.29343045e-01
-5.87044610e-03 -1.81778416e-01 1.36841893e-01 7.35120535e-01
-5.96595645e-01 -6.28223717e-01 -8.28214467e-01 2.52553940e-01
-2.48089135e-01 4.56519611e-02 -2.96712786e-01 9.76743519e-01
4.95666951e-01 9.89308178e-01 8.68483409e-02 -3.55061918e-01
8.57919037e-01 -3.60018015e-01 2.03990832e-01 -3.85310084e-01
-7.59512246e-01 1.53185904e-01 -3.48258436e-01 -7.01169789e-01
-3.21142942e-01 -7.44489014e-01 -9.62032795e-01 -1.90214828e-01
-6.46786213e-01 -8.36386904e-02 7.24679649e-01 8.27866018e-01
3.92946303e-01 5.68771064e-01 5.59818745e-01 -1.41416299e+00
-4.40525353e-01 -8.13788831e-01 -2.02708185e-01 3.40765506e-01
5.16210735e-01 -1.20997906e+00 -3.69385719e-01 -4.23388571e-01] | [7.798068523406982, -1.1503320932388306] |
5a2286b1-1880-4626-a3a4-02d99135adc1 | ssit-saliency-guided-self-supervised-image | 2210.10969 | null | https://arxiv.org/abs/2210.10969v4 | https://arxiv.org/pdf/2210.10969v4.pdf | SSiT: Saliency-guided Self-supervised Image Transformer for Diabetic Retinopathy Grading | Self-supervised learning (SSL) has been widely applied to learn image representations through exploiting unlabeled images. However, it has not been fully explored in the medical image analysis field. In this work, we propose Saliency-guided Self-Supervised image Transformer (SSiT) for diabetic retinopathy (DR) grading from fundus images. We novelly introduce saliency maps into SSL, with a goal of guiding self-supervised pre-training with domain-specific prior knowledge. Specifically, two saliency-guided learning tasks are employed in SSiT: (1) We conduct saliency-guided contrastive learning based on the momentum contrast, wherein we utilize fundus images' saliency maps to remove trivial patches from the input sequences of the momentum-updated key encoder. And thus, the key encoder is constrained to provide target representations focusing on salient regions, guiding the query encoder to capture salient features. (2) We train the query encoder to predict the saliency segmentation, encouraging preservation of fine-grained information in the learned representations. Extensive experiments are conducted on four publicly-accessible fundus image datasets. The proposed SSiT significantly outperforms other representative state-of-the-art SSL methods on all datasets and under various evaluation settings, establishing the effectiveness of the learned representations from SSiT. The source code is available at https://github.com/YijinHuang/SSiT. | ['Xiaoying Tang', 'Roger Tam', 'Pujin Cheng', 'Junyan Lyu', 'Yijin Huang'] | 2022-10-20 | null | null | null | null | ['diabetic-retinopathy-grading'] | ['medical'] | [ 6.00088537e-01 2.73694813e-01 -6.82442009e-01 -5.29781282e-01
-7.28539944e-01 1.18676677e-01 2.66358793e-01 1.30427647e-02
-2.73840338e-01 7.16721714e-01 6.24123752e-01 6.91237226e-02
-2.27086470e-01 -3.70262772e-01 -7.23081172e-01 -8.45947623e-01
2.21620411e-01 -1.63876206e-01 3.81850213e-01 -6.76142052e-02
4.60598379e-01 6.87699467e-02 -1.40559363e+00 2.23765761e-01
1.42312634e+00 1.00594592e+00 7.02946842e-01 2.56020874e-01
1.22857638e-01 1.24269807e+00 -3.91462632e-02 7.02435151e-02
-8.63485485e-02 -6.76037610e-01 -8.82925630e-01 3.83688986e-01
3.59429568e-01 -2.63855755e-01 -3.14908445e-01 1.40233076e+00
4.97386396e-01 6.96696863e-02 3.42042804e-01 -7.52967060e-01
-9.75633860e-01 4.30865198e-01 -7.56999075e-01 8.52403939e-01
-2.36098424e-01 2.65233427e-01 1.07294011e+00 -8.24788392e-01
5.36612451e-01 9.76935089e-01 8.05863738e-02 5.30925751e-01
-1.13162327e+00 -3.77982199e-01 3.14509600e-01 3.58624846e-01
-1.11931753e+00 -4.37377483e-01 1.05986166e+00 -3.30844432e-01
3.82517666e-01 1.05906643e-01 6.31298184e-01 6.50090337e-01
1.01846397e-01 1.33990622e+00 1.11152542e+00 -3.52191836e-01
1.41364604e-01 1.56632319e-01 2.20731691e-01 8.87602448e-01
1.13793910e-01 1.94172367e-01 -5.42759299e-01 5.55887409e-02
1.00336897e+00 2.10343182e-01 -4.54366535e-01 -5.63262343e-01
-1.37025976e+00 9.20520246e-01 1.10169518e+00 1.69245496e-01
-6.13117516e-01 -8.23597237e-02 3.04788321e-01 -6.21961392e-02
7.37882972e-01 4.19379830e-01 -2.90891647e-01 3.08230340e-01
-8.73821199e-01 6.56581149e-02 -4.68950793e-02 6.62249386e-01
8.38814080e-01 2.10550670e-02 -5.19904196e-01 9.70949709e-01
3.51704240e-01 3.14972311e-01 7.76180565e-01 -7.49460280e-01
2.01293141e-01 1.01367390e+00 -2.79601496e-02 -8.36998940e-01
-3.43782604e-01 -8.10042679e-01 -8.81316125e-01 -8.94309301e-03
3.88114457e-03 -1.35089040e-01 -1.03873193e+00 1.38254845e+00
1.60738185e-01 5.21489918e-01 9.71506611e-02 1.41703248e+00
1.14715683e+00 3.98019671e-01 2.62058258e-01 -2.42714182e-01
1.13749540e+00 -1.32322717e+00 -4.35632527e-01 -4.37753975e-01
4.15924013e-01 -4.98205513e-01 1.08190012e+00 -5.48068844e-02
-1.09452534e+00 -5.67392766e-01 -7.60308027e-01 -1.73638389e-01
1.78238958e-01 6.95350230e-01 6.20361388e-01 4.02318593e-03
-1.02350008e+00 2.81827807e-01 -8.80706429e-01 -2.04516739e-01
1.16052508e+00 2.02669322e-01 9.97620597e-02 -9.69179869e-02
-1.08896530e+00 6.62871897e-01 2.51210451e-01 6.81898370e-02
-8.82878661e-01 -7.33673394e-01 -9.11171496e-01 9.23405290e-02
3.71877670e-01 -7.76267707e-01 1.09532177e+00 -1.30696714e+00
-1.22151077e+00 1.16956782e+00 -4.53768849e-01 -5.78723192e-01
2.13786706e-01 -2.73836881e-01 -1.98285371e-01 6.37506843e-01
5.27239859e-01 9.88948107e-01 1.05253923e+00 -9.93886113e-01
-7.40366697e-01 -2.21080661e-01 -1.74880009e-02 3.61993968e-01
-2.76026636e-01 4.17356044e-02 -3.93257499e-01 -7.94353724e-01
2.49210045e-01 -7.63630986e-01 -4.39194053e-01 1.90144971e-01
-7.64066637e-01 -1.73102289e-01 3.44233930e-01 -5.39322972e-01
1.05112207e+00 -2.26584363e+00 2.08487481e-01 -1.64007545e-01
3.11287165e-01 4.88250285e-01 -1.54670075e-01 -2.29123354e-01
-2.10662171e-01 -1.89437807e-01 -4.02277559e-01 -1.42835215e-01
-5.43705225e-01 -9.44345742e-02 -3.52926105e-01 4.32084769e-01
6.17737234e-01 1.18354440e+00 -1.14491177e+00 -7.11019456e-01
3.16870302e-01 3.04097861e-01 -5.64080477e-01 1.51894256e-01
-2.09705129e-01 6.85520232e-01 -8.78296018e-01 8.80457819e-01
4.57080722e-01 -7.93809056e-01 -2.08855093e-01 -4.15889740e-01
-2.31153473e-01 3.66411954e-01 -5.48183739e-01 1.80542958e+00
-2.38435017e-03 6.14402711e-01 -5.28443635e-01 -1.23507655e+00
8.72463405e-01 -2.41634771e-02 4.33785558e-01 -8.67513537e-01
7.21312165e-02 2.55508155e-01 -1.74512506e-01 -5.93632340e-01
3.71148475e-02 6.19836934e-02 3.15267086e-01 1.94417000e-01
1.97519168e-01 2.42297500e-01 8.70647430e-02 2.32920483e-01
5.16840756e-01 1.47694051e-01 4.34171975e-01 -3.72445941e-01
6.79114580e-01 1.48862705e-01 7.23469079e-01 5.81662178e-01
-4.02645826e-01 7.80420303e-01 4.51820612e-01 -3.69584560e-01
-6.49010122e-01 -1.02134013e+00 -2.04086274e-01 9.96288657e-01
5.79634488e-01 1.69716969e-01 -4.49476361e-01 -8.13281894e-01
-1.31815165e-01 5.07150829e-01 -7.86461711e-01 -3.89772713e-01
-4.26619381e-01 -8.67209613e-01 -9.19726044e-02 4.19612348e-01
7.73399174e-01 -1.26665449e+00 -6.69132054e-01 5.21930121e-02
-1.27867922e-01 -7.33678937e-01 -7.97484338e-01 -1.81143418e-01
-1.20005417e+00 -1.26393473e+00 -1.22902286e+00 -1.30726528e+00
1.18664825e+00 7.56579161e-01 8.97361338e-01 7.49031082e-02
-4.87800688e-01 1.21472850e-01 -2.95170158e-01 -3.65815908e-01
1.32205233e-01 9.15883034e-02 -4.08570588e-01 3.73839557e-01
4.03652310e-01 -3.82678419e-01 -1.15492666e+00 2.94144213e-01
-7.23263860e-01 4.26999569e-01 9.28819478e-01 1.00383091e+00
9.49203134e-01 -2.39459351e-01 8.54229331e-01 -9.36038435e-01
2.17650577e-01 -4.44090992e-01 -5.09028256e-01 1.55793220e-01
-3.68272245e-01 1.16590085e-02 2.42141888e-01 -2.14959010e-01
-1.08581185e+00 5.39055355e-02 1.66180283e-01 -4.99624491e-01
-7.27187097e-02 5.38418770e-01 1.01172164e-01 -1.50629401e-01
5.10464668e-01 7.13877022e-01 1.45322070e-01 -4.20088559e-01
2.20093995e-01 4.97030199e-01 5.76301575e-01 -1.13208234e-01
4.23342973e-01 5.65480292e-01 -4.09865052e-01 -4.96271908e-01
-1.71312952e+00 -6.10437632e-01 -3.52113843e-01 -8.21849257e-02
8.45833898e-01 -1.13535368e+00 -2.85692215e-01 4.62191582e-01
-6.72290683e-01 -3.17256004e-01 -3.61799389e-01 7.36305535e-01
-5.64874172e-01 1.48520097e-01 -3.92379820e-01 -4.30403352e-01
-5.16204238e-01 -1.21942949e+00 1.03629518e+00 8.46240461e-01
2.00442895e-01 -9.48011041e-01 -3.75892781e-02 5.49698949e-01
3.63782585e-01 2.27505341e-02 8.30949366e-01 -3.34374338e-01
-7.74098873e-01 2.90581226e-01 -7.33239710e-01 5.37338734e-01
6.08184397e-01 -4.05354202e-01 -7.42341936e-01 -3.16198736e-01
-1.71471521e-01 -4.71948057e-01 1.36985564e+00 9.13799584e-01
1.30659258e+00 -3.74531001e-01 -3.34086031e-01 7.55658209e-01
1.32787561e+00 -3.95483449e-02 4.81020540e-01 5.12924552e-01
6.51008606e-01 5.32753348e-01 8.51777136e-01 2.37548634e-01
5.35906196e-01 3.87316108e-01 4.12533432e-01 -6.19420528e-01
-3.77820373e-01 -2.91502476e-01 -2.49258429e-02 4.78271157e-01
-2.45641563e-02 4.12472486e-01 -5.42482436e-01 9.05612290e-01
-1.98991859e+00 -7.94964731e-01 2.47408047e-01 1.94866455e+00
1.15069687e+00 8.81023183e-02 1.00863218e-01 -3.96113455e-01
8.22501481e-01 4.35307413e-01 -1.19105077e+00 2.62521803e-01
-8.01661983e-02 -5.09488918e-02 3.55834872e-01 2.35168651e-01
-1.42544711e+00 1.03079844e+00 4.76214886e+00 5.41448295e-01
-1.39916062e+00 1.63598016e-01 1.02151656e+00 -3.39707792e-01
-1.94470182e-01 -9.85763296e-02 -6.80733681e-01 7.07913041e-01
1.67071417e-01 -4.00845855e-01 7.72995800e-02 8.89463007e-01
4.96258020e-01 -1.38990670e-01 -6.34670198e-01 9.41563010e-01
2.02701151e-01 -1.60790861e+00 2.15168819e-01 -3.49205554e-01
1.11058378e+00 3.47107232e-01 3.61671865e-01 5.06338552e-02
7.14333635e-03 -8.26350451e-01 2.42329076e-01 8.21749568e-01
6.62307441e-01 -6.45473361e-01 6.96991205e-01 2.95437220e-02
-7.84323931e-01 -1.76731288e-01 -5.11040032e-01 3.27093899e-01
8.34774449e-02 8.01678002e-01 -6.15708649e-01 2.27693111e-01
6.01480544e-01 1.36683273e+00 -7.40413249e-01 1.48161483e+00
-4.43283856e-01 8.02110136e-01 2.69671291e-01 1.44424856e-01
4.32956636e-01 -1.77619100e-01 5.96307635e-01 8.88898253e-01
-3.37222591e-02 9.07088220e-02 1.79631099e-01 1.04714453e+00
-2.97575304e-03 1.97798505e-01 -1.26914699e-02 1.10003196e-01
1.51364192e-01 1.00637567e+00 -4.87129956e-01 -3.09332192e-01
-3.48828018e-01 8.38557303e-01 3.01183254e-01 4.69088882e-01
-5.73801100e-01 -2.67561972e-01 4.59970146e-01 7.51468167e-02
4.70252961e-01 3.85926574e-01 -2.86747217e-01 -1.31147766e+00
-1.17392451e-01 -5.65431237e-01 3.86148542e-01 -9.25281942e-01
-1.16004467e+00 5.18248677e-01 -2.62780517e-01 -1.49626422e+00
1.15063705e-01 -1.12437762e-01 -6.91204846e-01 8.41985881e-01
-2.39978004e+00 -1.16715252e+00 -4.44695652e-01 6.62475169e-01
6.98557377e-01 -4.34855551e-01 3.56518060e-01 1.31465510e-01
-7.96460092e-01 2.84507126e-01 3.05453688e-02 3.86826433e-02
7.25699067e-01 -1.14998674e+00 5.35547137e-02 7.87153304e-01
-9.15723145e-02 6.58993959e-01 3.84023607e-01 -5.67854226e-01
-9.00148928e-01 -1.58761907e+00 7.57731318e-01 9.70790759e-02
5.06844640e-01 3.02800685e-01 -9.97164190e-01 5.59652805e-01
-1.71281416e-02 5.28390169e-01 4.42006260e-01 -3.37113768e-01
1.60843600e-02 -2.22973868e-01 -8.53153527e-01 4.17615801e-01
8.66603673e-01 -4.67289448e-01 -7.26988673e-01 5.53702891e-01
7.92201698e-01 -4.84599084e-01 -4.59323615e-01 3.15368176e-01
1.71672598e-01 -8.14048767e-01 1.08113635e+00 -5.83726585e-01
8.24391842e-01 -4.01892960e-01 2.86839068e-01 -1.23815119e+00
-4.10184622e-01 -3.32935542e-01 1.81467866e-03 8.18464339e-01
2.43208513e-01 -7.16283679e-01 6.70928597e-01 6.91677183e-02
-3.29454124e-01 -1.19152486e+00 -6.01155281e-01 -2.71050274e-01
-2.85511732e-01 3.65284234e-01 3.05507094e-01 7.36097634e-01
-9.55185145e-02 1.58000499e-01 -4.02563155e-01 2.60973692e-01
8.32405746e-01 6.87155247e-01 2.47277126e-01 -9.17196751e-01
-1.45355865e-01 -4.07723397e-01 -4.02903080e-01 -1.16699100e+00
6.25420585e-02 -1.01572371e+00 2.18917821e-02 -1.64884341e+00
6.54155731e-01 -4.22616959e-01 -8.22004855e-01 6.83971643e-01
-7.27987349e-01 4.81165141e-01 -9.76296794e-03 5.91978788e-01
-1.00009370e+00 8.04610193e-01 1.66170788e+00 -2.90238857e-01
-3.89767170e-01 2.68211931e-01 -1.30026650e+00 7.41805434e-01
1.01924062e+00 -2.95801371e-01 -6.87221050e-01 -4.07822311e-01
-2.14046359e-01 -1.01717897e-01 7.75805950e-01 -7.11986125e-01
1.99630469e-01 -2.21861362e-01 3.05207253e-01 -3.94615412e-01
-1.21715903e-01 -3.73353772e-02 -7.01300502e-01 4.22817022e-01
-8.00020158e-01 -7.24501610e-01 2.85817664e-02 5.23634374e-01
-5.09593070e-01 -8.84091482e-02 1.07920802e+00 -1.28392994e-01
-9.56575155e-01 4.99814183e-01 1.64692923e-01 2.42047191e-01
8.24543357e-01 5.53559214e-02 -4.07595485e-01 -4.66268398e-02
-7.89801657e-01 4.24974680e-01 2.95906305e-01 3.20436180e-01
9.14227903e-01 -1.10599136e+00 -8.77661467e-01 2.76679814e-01
3.51599902e-01 1.49967745e-01 5.65686107e-01 1.09270644e+00
-2.38466829e-01 6.29040062e-01 -2.97004342e-01 -7.50064611e-01
-1.03807116e+00 4.49381471e-01 4.17050958e-01 -1.79861933e-02
-8.79965007e-01 1.01981616e+00 7.21549153e-01 -9.75249484e-02
9.89241526e-02 -4.39731359e-01 -5.62869966e-01 -2.65435994e-01
7.67783284e-01 -2.79991026e-03 -2.28098378e-01 -4.94428277e-01
-3.27439666e-01 5.54353774e-01 -5.42205453e-01 5.87666690e-01
1.48855913e+00 -2.98550576e-01 -1.03858024e-01 2.01844633e-01
1.14221680e+00 -3.81349742e-01 -1.64258778e+00 -7.57763147e-01
-1.03855699e-01 -5.24163783e-01 5.40416062e-01 -7.80816436e-01
-1.43824792e+00 6.62803233e-01 8.71707499e-01 -3.97478193e-01
1.35565841e+00 2.88602233e-01 8.10128450e-01 -1.20710172e-02
1.41817540e-01 -6.99455559e-01 3.46614718e-01 -5.19996416e-03
8.96964490e-01 -1.66079605e+00 1.19185150e-01 -4.78381872e-01
-1.08067787e+00 6.66288257e-01 6.24511719e-01 -4.65605795e-01
4.66139197e-01 -5.11384666e-01 6.22895025e-02 -2.83322394e-01
-5.78784525e-01 -5.17866194e-01 6.00359619e-01 3.68558139e-01
3.00967544e-01 -5.25154844e-02 -3.19994062e-01 6.13229036e-01
1.66018412e-01 5.02517700e-01 4.42113727e-01 7.32765377e-01
-6.30579829e-01 -5.90582669e-01 1.21493258e-01 8.42314363e-01
-3.20140719e-01 -3.77458692e-01 8.48892480e-02 1.97225958e-01
-1.53017165e-02 6.51388526e-01 6.58392757e-02 1.96017429e-01
1.60939470e-02 -4.83729511e-01 1.46979704e-01 -8.80506933e-01
-7.57484585e-02 2.08711475e-01 -2.87705183e-01 -6.23839140e-01
-7.37312496e-01 -7.92268574e-01 -1.18110871e+00 6.15269184e-01
-1.99508399e-01 1.40758622e-02 1.29160866e-01 9.10254419e-01
6.19775414e-01 6.60960197e-01 9.37479734e-01 -8.16901565e-01
-3.60125601e-01 -7.69703865e-01 -5.38602173e-01 2.64849275e-01
7.45869637e-01 -8.58128309e-01 -2.63860047e-01 2.44994015e-01] | [9.776597023010254, -0.28591957688331604] |
84ff6e33-ab42-46d3-a502-99145f9fc593 | uncertainty-based-biological-age-estimation | 2103.08491 | null | https://arxiv.org/abs/2103.08491v1 | https://arxiv.org/pdf/2103.08491v1.pdf | Uncertainty-Based Biological Age Estimation of Brain MRI Scans | Age is an essential factor in modern diagnostic procedures. However, assessment of the true biological age (BA) remains a daunting task due to the lack of reference ground-truth labels. Current BA estimation approaches are either restricted to skeletal images or rely on non-imaging modalities that yield a whole-body BA assessment. However, various organ systems may exhibit different aging characteristics due to lifestyle and genetic factors. In this initial study, we propose a new framework for organ-specific BA estimation utilizing 3D magnetic resonance image (MRI) scans. As a first step, this framework predicts the chronological age (CA) together with the corresponding patient-dependent aleatoric uncertainty. An iterative training algorithm is then utilized to segregate atypical aging patients from the given population based on the predicted uncertainty scores. In this manner, we hypothesize that training a new model on the remaining population should approximate the true BA behavior. We apply the proposed methodology on a brain MRI dataset containing healthy individuals as well as Alzheimer's patients. We demonstrate the correlation between the predicted BAs and the expected cognitive deterioration in Alzheimer's patients. | ['Bin Yang', 'Sergios Gatidis', 'Tobias Hepp', 'Wenbin Shi', 'Sherif Abdulatif', 'Karim Armanious'] | 2021-03-15 | null | null | null | null | ['age-estimation', 'age-estimation'] | ['computer-vision', 'miscellaneous'] | [ 5.13155796e-02 2.83131033e-01 -7.19382763e-02 -6.54346526e-01
-5.77691734e-01 1.35994405e-01 4.00970608e-01 2.72108048e-01
-3.64525527e-01 1.04404604e+00 9.41125974e-02 -4.42907922e-02
-1.09753564e-01 -5.92160046e-01 -4.58001107e-01 -6.65306032e-01
-3.02302808e-01 1.07877493e+00 1.80426270e-01 2.14306638e-01
7.04686269e-02 2.74190843e-01 -1.29054499e+00 -1.59821585e-01
1.39356637e+00 1.18935609e+00 2.14897364e-01 4.20658320e-01
6.07099608e-02 3.39512259e-01 -3.16481680e-01 -4.60196018e-01
1.78962231e-01 -4.71058130e-01 -5.17656147e-01 -5.64329885e-02
5.71351588e-01 -4.31002170e-01 2.40857173e-02 1.03657162e+00
5.14991343e-01 -2.61907488e-01 1.11112094e+00 -1.13107014e+00
-2.69505382e-01 7.72415459e-01 -5.92991173e-01 3.39406252e-01
2.34088209e-02 -3.92166786e-02 7.63180315e-01 -6.89160585e-01
3.99860770e-01 9.52060759e-01 7.34979331e-01 9.00590301e-01
-1.17157674e+00 -5.13731956e-01 2.75935590e-01 4.38237101e-01
-1.08437288e+00 -4.80647862e-01 7.01920807e-01 -6.96489513e-01
1.09496333e-01 -8.01967084e-02 1.01270866e+00 1.14810729e+00
6.56279981e-01 4.74736065e-01 1.53140914e+00 -3.26999009e-01
6.35472178e-01 -4.17848736e-01 4.15481359e-01 7.43099332e-01
6.69836938e-01 1.67259589e-01 -6.16450965e-01 -4.61398989e-01
4.64449555e-01 -3.43840063e-01 -1.47049665e-01 -6.52471721e-01
-1.20530820e+00 3.88977885e-01 3.83852988e-01 1.37912214e-01
-8.15399170e-01 1.95575684e-01 2.19383106e-01 -1.51151836e-01
4.68204856e-01 1.17016077e-01 -4.58877534e-01 2.78290361e-01
-1.24463177e+00 6.22397661e-01 3.83585900e-01 3.90764326e-01
2.17922553e-01 1.15743235e-01 -1.73827331e-03 7.97415018e-01
5.59813917e-01 6.50101125e-01 6.79958463e-01 -8.34063828e-01
2.20790170e-02 2.74307221e-01 8.21314082e-02 -4.18157071e-01
-7.24728823e-01 -7.60537565e-01 -8.10214162e-01 4.30346310e-01
9.71649468e-01 -2.51178804e-04 -1.24139762e+00 2.07032466e+00
4.03450221e-01 3.62131838e-03 -3.02907437e-01 8.42961669e-01
2.17907250e-01 -1.05805747e-01 5.56090415e-01 -4.12225157e-01
1.67674994e+00 -5.17632067e-01 -4.15639937e-01 -4.88723338e-01
1.50302097e-01 -1.28055334e-01 6.46577775e-01 4.42261100e-01
-1.34883535e+00 -2.10904837e-01 -1.01396155e+00 3.21658790e-01
7.16081038e-02 2.93446183e-01 5.20946681e-01 9.72979903e-01
-7.66763687e-01 5.32151043e-01 -1.42461812e+00 -4.60446298e-01
3.99076849e-01 3.00378412e-01 -8.33042338e-02 1.43196210e-01
-1.17224944e+00 1.22582304e+00 2.21926644e-01 2.44432092e-01
-1.01647186e+00 -9.83832300e-01 -4.11131173e-01 -3.50705951e-01
-7.65446350e-02 -1.25549781e+00 1.11829066e+00 -6.99968576e-01
-1.15460098e+00 8.79820466e-01 -2.28204846e-01 -7.26469696e-01
9.18116927e-01 -2.25620583e-01 -3.73942822e-01 1.73836812e-01
2.04268709e-01 6.31215692e-01 1.04871559e+00 -1.19553304e+00
-4.17127103e-01 -1.16439211e+00 -1.88483819e-01 7.41780698e-02
5.47089279e-02 -1.52850926e-01 3.13329488e-01 -6.11412644e-01
6.19427264e-01 -1.04491770e+00 -3.64971876e-01 4.22179073e-01
-2.71018118e-01 1.17557064e-01 5.29548451e-02 -1.19104528e+00
1.06711471e+00 -1.57543409e+00 3.59509796e-01 3.83513123e-01
8.18890691e-01 -4.21770513e-01 3.82587254e-01 -3.47203374e-01
-8.96484554e-02 -2.04548210e-01 -6.16228759e-01 -1.67014763e-01
-1.94773510e-01 -3.02563369e-01 2.86808729e-01 5.86201549e-01
-1.22026287e-01 4.31761384e-01 -8.39304149e-01 -7.53707349e-01
-3.42702448e-01 1.43798545e-01 -3.34374160e-01 9.12962407e-02
-1.35986954e-01 7.16635883e-01 -4.64035720e-01 9.15344357e-01
5.49231410e-01 -4.38067876e-02 2.85876662e-01 -4.25888747e-01
2.10104287e-01 -1.62198305e-01 -6.54017806e-01 1.59673166e+00
-2.31633127e-01 -1.32506222e-01 9.52697098e-02 -9.35392976e-01
9.53142881e-01 1.72543257e-01 7.21697569e-01 -3.28256667e-01
2.36278981e-01 6.65725172e-01 7.16807067e-01 -8.81253928e-02
-1.30653635e-01 -4.64873284e-01 1.50234267e-01 5.44394016e-01
-8.31937566e-02 2.07375199e-01 1.42325908e-01 -7.69441426e-02
1.08722651e+00 1.29205182e-01 4.57771599e-01 -5.94460964e-01
7.03678071e-01 -2.11418152e-01 8.02663386e-01 4.96379554e-01
-8.58313143e-01 4.75785494e-01 4.13854241e-01 -4.49345678e-01
-1.33711517e+00 -1.62365150e+00 -5.75123847e-01 3.47500116e-01
-3.10045600e-01 -8.30580816e-02 -9.09681439e-01 -5.94680011e-01
1.08298741e-01 6.50258958e-01 -8.65960479e-01 -4.86464560e-01
-4.86045003e-01 -1.23306608e+00 4.42663997e-01 7.02602625e-01
4.21900332e-01 -4.70765352e-01 -7.05421090e-01 1.20043047e-01
-2.50490785e-01 -7.28119850e-01 -1.75683856e-01 5.37927523e-02
-1.40463257e+00 -1.03098726e+00 -1.23807478e+00 -3.78403753e-01
9.07764077e-01 -4.73304749e-01 9.43021357e-01 -3.72031704e-02
-3.21762919e-01 3.18991363e-01 4.90656011e-02 -4.95978057e-01
-5.83657980e-01 2.39154145e-01 6.50230348e-01 -1.21485613e-01
2.69989878e-01 -7.43811727e-01 -1.13390744e+00 2.23547727e-01
-2.12797016e-01 2.99745072e-02 6.42921448e-01 7.02561677e-01
5.51239133e-01 -4.07336354e-01 9.40657079e-01 -5.72473943e-01
4.25598353e-01 -4.44785446e-01 -3.20729017e-01 5.70496142e-01
-1.13146937e+00 3.40208113e-01 1.58918694e-01 -5.45705080e-01
-1.23416913e+00 2.48283818e-01 6.63443992e-04 5.40015176e-02
-4.13365066e-02 5.07082403e-01 -2.04488873e-01 1.69779941e-01
7.58651733e-01 3.32386754e-02 2.40415975e-01 -3.20361882e-01
1.18325971e-01 4.33171183e-01 6.99596107e-01 -1.06696939e+00
4.03452694e-01 3.33563656e-01 2.96553910e-01 -4.51907307e-01
-8.25604439e-01 2.50504524e-01 -8.78689051e-01 -5.82884967e-01
7.33941078e-01 -8.21053803e-01 -3.21249574e-01 7.09855497e-01
-7.76495457e-01 -2.65853584e-01 4.52370085e-02 8.36329997e-01
-7.36506462e-01 5.87704778e-01 -3.49212289e-01 -7.67251074e-01
-7.85356104e-01 -1.12806714e+00 8.40451539e-01 -3.64625640e-02
-5.11531413e-01 -9.21051919e-01 5.28133735e-02 4.79804635e-01
3.07595611e-01 2.50296354e-01 1.31218350e+00 -4.38480616e-01
-1.50143430e-01 -2.69750386e-01 -9.82690081e-02 1.68628246e-01
3.84047665e-02 -1.96065545e-01 -6.02240920e-01 2.43326314e-02
1.52890861e-01 -1.19598895e-01 6.55320227e-01 8.85821581e-01
8.06545794e-01 3.83394510e-01 -2.43452191e-01 1.09672651e-01
1.02910721e+00 1.67586118e-01 4.67209518e-01 4.28413451e-01
4.28597540e-01 6.58828199e-01 5.24042547e-01 5.69974065e-01
4.38943833e-01 4.36858207e-01 4.57551092e-01 6.28229141e-01
-1.45458743e-01 8.99204165e-02 2.92522252e-01 7.39913940e-01
-5.25098085e-01 2.76202321e-01 -1.27495837e+00 4.85139489e-01
-1.47138166e+00 -6.34276330e-01 -2.30589032e-01 2.43933439e+00
8.55117500e-01 3.79624456e-01 4.05760974e-01 9.49290171e-02
9.18672144e-01 -1.74492449e-01 -9.47834551e-01 1.35703504e-01
4.24637161e-02 1.38028622e-01 4.27155584e-01 1.81277066e-01
-6.41420722e-01 2.57737458e-01 6.81470299e+00 1.71913773e-01
-1.08059859e+00 2.76888996e-01 8.07406843e-01 7.34893158e-02
-1.14486381e-01 -6.84531182e-02 -6.89316750e-01 4.42915559e-01
1.16060197e+00 -5.04542351e-01 3.17451745e-01 6.77951217e-01
1.61423147e-01 -2.90057331e-01 -1.15638840e+00 6.42055869e-01
1.14391677e-01 -5.31448245e-01 -9.15069953e-02 5.49639203e-02
2.39423841e-01 -2.49372065e-01 -1.19701914e-01 2.22535599e-02
-7.69525468e-02 -6.06415331e-01 1.00650597e+00 1.12640774e+00
6.79703355e-01 -3.31088364e-01 6.53918147e-01 1.15303278e-01
-8.91533196e-01 -2.21709996e-01 -1.15937777e-01 2.03351900e-01
3.99829477e-01 9.83402669e-01 -7.61101544e-01 3.00888538e-01
5.11110842e-01 2.54671574e-01 -8.12749922e-01 1.20420289e+00
-1.80033952e-01 5.36756814e-01 -2.61722028e-01 4.32893157e-01
-5.02534628e-01 -2.93912798e-01 4.30272102e-01 2.16058031e-01
7.50133097e-01 -1.21035397e-01 -2.30454624e-01 9.31721866e-01
1.39521703e-01 1.38222501e-01 1.30075768e-01 -1.27952723e-02
3.63299042e-01 9.93626177e-01 -9.09846902e-01 -2.62734294e-01
-2.49633089e-01 6.27971947e-01 3.96411479e-01 -5.12788109e-02
-7.10065544e-01 3.81028563e-01 3.09840560e-01 4.72263098e-01
-2.21430734e-01 -2.03980997e-01 -6.14395261e-01 -1.24057007e+00
2.05526799e-02 -7.48315692e-01 3.82228911e-01 -7.52576947e-01
-1.49333346e+00 5.65030754e-01 2.37230971e-01 -8.48938406e-01
-2.97612488e-01 -5.82420170e-01 -4.60885525e-01 8.19843888e-01
-8.75504434e-01 -1.18291223e+00 -5.05415559e-01 1.86723351e-01
2.80992299e-01 -1.77902684e-01 6.84428990e-01 3.06220859e-01
-6.05625331e-01 4.09757465e-01 -8.87357667e-02 -3.51262003e-01
7.80132949e-01 -1.56651700e+00 5.01197055e-02 7.54681289e-01
-5.99293172e-01 5.60405374e-01 1.03280973e+00 -1.16372895e+00
-7.42247283e-01 -7.98200727e-01 7.68005252e-01 -4.04364407e-01
8.80932271e-01 4.21637334e-02 -7.99540699e-01 5.94975770e-01
-4.35502440e-01 1.23773823e-02 6.08894408e-01 2.58935124e-01
-2.44177341e-01 -2.94100642e-01 -1.28326321e+00 5.80402553e-01
1.14931870e+00 -2.91285962e-01 -8.27522278e-01 -1.45984758e-02
6.27999604e-02 -2.39757761e-01 -1.25962138e+00 6.35398567e-01
1.12896359e+00 -7.42151737e-01 1.09707141e+00 -4.83228952e-01
3.86689872e-01 -1.37870476e-01 -6.50672019e-02 -1.21796072e+00
-1.92058593e-01 2.73506999e-01 -4.30109799e-01 1.09695399e+00
2.06164971e-01 -4.06010151e-01 9.95443761e-01 1.04079688e+00
-1.14593334e-01 -7.52239823e-01 -9.65007961e-01 -8.27393651e-01
3.33755493e-01 -3.51148516e-01 5.17745018e-01 3.78270805e-01
-3.50717455e-01 9.85710099e-02 -2.57055849e-01 3.93317968e-01
1.25099647e+00 -2.58435190e-01 2.35868856e-01 -1.86539054e+00
-1.08139642e-01 -2.83223748e-01 -6.40001178e-01 -1.44683346e-01
1.99318454e-01 -7.71829009e-01 1.01394998e-02 -1.39238012e+00
4.01644319e-01 -6.41644657e-01 -4.30940986e-01 -1.03522420e-01
-2.18174741e-01 -2.12863237e-02 -1.43719614e-01 3.30647707e-01
-7.99815077e-03 5.37853420e-01 1.03921092e+00 -9.53878984e-02
4.32157181e-02 2.73183227e-01 -3.72864246e-01 8.82700682e-01
9.94918466e-01 -4.11392123e-01 -5.37373662e-01 1.40874670e-03
2.35459268e-01 8.21046233e-02 4.62828308e-01 -1.31235671e+00
-9.42736119e-02 -9.58675519e-02 5.44557750e-01 -5.72742581e-01
2.31271088e-01 -6.31531835e-01 3.39618057e-01 8.86727810e-01
-1.85106516e-01 -2.02097874e-02 -3.42758656e-01 6.26937091e-01
3.18741947e-01 -5.45723855e-01 9.57259417e-01 -2.08606482e-01
-1.92055717e-01 4.71543878e-01 -3.86216164e-01 -9.27134901e-02
8.43159318e-01 -1.58769473e-01 -9.41133350e-02 -1.67767391e-01
-1.29504645e+00 -1.06207579e-01 4.44750905e-01 -5.29319979e-02
3.90876085e-01 -1.23204052e+00 -7.35156715e-01 -2.95613885e-01
1.75276265e-01 -2.68328816e-01 3.97536606e-01 1.25383258e+00
-6.38869822e-01 7.72992745e-02 -7.72658467e-01 -5.74059784e-01
-1.14426923e+00 3.89353216e-01 5.31979740e-01 -3.07614774e-01
-3.61256152e-01 4.67582494e-01 2.13016376e-01 -1.05241999e-01
1.08342335e-01 -3.64102691e-01 -7.63669237e-02 1.69761583e-01
4.78762388e-01 5.60245574e-01 6.62069255e-03 -5.86233616e-01
-3.23963255e-01 3.47913861e-01 -3.05882424e-01 -2.12461501e-01
1.16598535e+00 -4.69921589e-01 -1.57839939e-01 6.30157888e-01
3.14240932e-01 -1.66526169e-01 -1.04271770e+00 -1.10435434e-01
2.85382092e-01 -1.91790074e-01 1.46539569e-01 -8.04421306e-01
-1.07353354e+00 7.60562718e-01 1.19022954e+00 -3.70065093e-01
9.27798510e-01 2.21506760e-01 6.96164191e-01 3.13758329e-02
6.67964280e-01 -1.00513029e+00 -2.10175350e-01 -2.50531703e-01
8.65259230e-01 -9.67064321e-01 4.51875865e-01 -3.84478003e-01
-3.25543284e-01 1.08914804e+00 7.82861948e-01 -1.71412025e-02
7.73457408e-01 -2.07456738e-01 1.42356707e-02 -1.22990496e-01
-4.72706437e-01 6.07883334e-02 2.22288042e-01 7.75486052e-01
5.43063879e-01 2.10546583e-01 -1.04329479e+00 1.02511370e+00
-3.84331107e-01 1.93055168e-01 5.12242615e-01 7.40160763e-01
-4.03626859e-01 -1.20513356e+00 -4.97669935e-01 9.57849085e-01
-4.85109419e-01 1.92683414e-01 -1.22886747e-01 4.62986380e-01
1.49033725e-01 2.10343659e-01 4.57689492e-03 8.97406936e-02
-1.58988678e-05 5.33323944e-01 9.43827987e-01 -2.98612386e-01
-6.16507158e-02 -1.60417303e-01 2.21724495e-01 -5.08807041e-02
-5.08463502e-01 -1.22719979e+00 -1.29850829e+00 1.52416304e-02
-2.10288480e-01 -1.42944738e-01 8.43296111e-01 9.72389638e-01
4.00053747e-02 4.18070942e-01 1.31781220e-01 -6.95486248e-01
-6.71084225e-01 -9.79826689e-01 -7.74095297e-01 1.26492813e-01
-9.54996496e-02 -1.20344388e+00 -3.18724275e-01 1.60817087e-01] | [14.111309051513672, -1.5970097780227661] |
d06d2487-48f1-46a2-985a-23ef00e94d90 | pausespeech-natural-speech-synthesis-via-pre | 2306.07489 | null | https://arxiv.org/abs/2306.07489v1 | https://arxiv.org/pdf/2306.07489v1.pdf | PauseSpeech: Natural Speech Synthesis via Pre-trained Language Model and Pause-based Prosody Modeling | Although text-to-speech (TTS) systems have significantly improved, most TTS systems still have limitations in synthesizing speech with appropriate phrasing. For natural speech synthesis, it is important to synthesize the speech with a phrasing structure that groups words into phrases based on semantic information. In this paper, we propose PuaseSpeech, a speech synthesis system with a pre-trained language model and pause-based prosody modeling. First, we introduce a phrasing structure encoder that utilizes a context representation from the pre-trained language model. In the phrasing structure encoder, we extract a speaker-dependent syntactic representation from the context representation and then predict a pause sequence that separates the input text into phrases. Furthermore, we introduce a pause-based word encoder to model word-level prosody based on pause sequence. Experimental results show PauseSpeech outperforms previous models in terms of naturalness. Furthermore, in terms of objective evaluations, we can observe that our proposed methods help the model decrease the distance between ground-truth and synthesized speech. Audio samples are available at https://jisang93.github.io/pausespeech-demo/. | ['Seong-Whan Lee', 'Sang-Hoon Lee', 'Ji-Sang Hwang'] | 2023-06-13 | null | null | null | null | ['speech-synthesis'] | ['speech'] | [ 2.24013001e-01 2.33821779e-01 -4.22962457e-01 -4.95005071e-01
-1.13207674e+00 -4.77111489e-01 2.58001328e-01 -2.16323718e-01
5.66715561e-02 5.71607888e-01 9.07172501e-01 -6.04890823e-01
5.63201368e-01 -5.72398722e-01 -6.22913122e-01 -4.59406585e-01
5.14649451e-01 1.50516450e-01 1.52340859e-01 -2.09006518e-01
2.93711629e-02 4.78254780e-02 -1.16097844e+00 5.60233891e-01
8.59345257e-01 8.40453327e-01 8.37352216e-01 8.68310452e-01
-4.05388951e-01 8.19573522e-01 -8.34586740e-01 4.94490825e-02
4.17431630e-02 -1.10813236e+00 -7.29603171e-01 5.73328808e-02
-2.17740446e-01 -2.33654633e-01 -2.16433600e-01 8.11015189e-01
4.44458902e-01 3.00431907e-01 3.65443051e-01 -9.43221986e-01
-4.95509595e-01 1.16156936e+00 1.21722549e-01 8.21064040e-02
5.66523969e-01 8.95336941e-02 1.30700195e+00 -1.05385196e+00
3.64831716e-01 1.41929710e+00 2.85256416e-01 8.72556031e-01
-1.08960009e+00 -6.21652067e-01 1.49972335e-01 4.95060384e-02
-1.28546166e+00 -8.48293126e-01 1.08050668e+00 -1.77386850e-01
1.15975249e+00 4.04330790e-01 5.66621661e-01 1.36160791e+00
8.93848464e-02 1.03738928e+00 5.74705422e-01 -7.07569897e-01
2.46044978e-01 1.76095776e-02 -1.61795184e-01 3.48651141e-01
-7.81376064e-01 1.37588456e-01 -6.00125849e-01 1.92785397e-01
5.66740513e-01 -3.27485442e-01 -5.16652524e-01 3.72669965e-01
-1.20961559e+00 8.43420446e-01 1.15839727e-01 5.17676651e-01
-4.99547541e-01 2.44359784e-02 4.92524475e-01 1.96346775e-01
3.52162778e-01 2.98180938e-01 -1.75000399e-01 -2.79708385e-01
-1.29547405e+00 4.39983234e-02 7.59990811e-01 1.20894217e+00
3.67619395e-01 5.64017296e-01 -3.03162396e-01 1.28641665e+00
4.45201933e-01 6.87262893e-01 8.44803989e-01 -9.47095573e-01
6.82293475e-01 4.08242233e-02 -8.10407326e-02 -6.09640300e-01
1.30615592e-01 -1.54378533e-01 -5.89602470e-01 -3.64306867e-01
-1.25317514e-01 -2.19912559e-01 -8.77934694e-01 1.84746850e+00
1.53804775e-02 1.57046065e-01 5.23059130e-01 7.97283828e-01
8.56843948e-01 1.51373649e+00 -2.98566800e-02 -5.74874163e-01
1.29939842e+00 -1.50837743e+00 -1.25043464e+00 -2.69138932e-01
3.82044524e-01 -9.78633225e-01 1.56359470e+00 1.18573718e-01
-1.39203155e+00 -8.77003491e-01 -1.08626509e+00 -1.45648614e-01
-3.04247458e-02 3.13847423e-01 -2.96119034e-01 4.46848184e-01
-9.00223613e-01 2.76627034e-01 -7.40890324e-01 -1.38421953e-01
-1.84090540e-01 -2.77270023e-02 6.00248501e-02 4.15729254e-01
-1.46181154e+00 4.89069730e-01 5.79579592e-01 -2.48463288e-01
-8.79030049e-01 -5.39281785e-01 -1.00392818e+00 2.06531733e-01
3.07323337e-01 -4.76403058e-01 1.68694842e+00 -1.01126587e+00
-2.16276836e+00 3.18244964e-01 -6.17847502e-01 -6.55477107e-01
6.53198212e-02 -3.76822986e-02 -7.21117616e-01 4.33570802e-01
3.52165569e-03 7.82374263e-01 8.37913036e-01 -1.13785124e+00
-7.42200255e-01 3.30965787e-01 -3.77468944e-01 3.77058625e-01
-1.83239132e-01 3.23373824e-01 -3.61870587e-01 -1.07719350e+00
-4.83721606e-02 -9.44111109e-01 -7.51053542e-02 -5.39992929e-01
-6.61029696e-01 -2.77540475e-01 8.09709251e-01 -8.38272691e-01
1.75155830e+00 -2.34304643e+00 1.43178016e-01 -1.79689690e-01
-4.26150173e-01 3.14835519e-01 -1.95216477e-01 8.12675774e-01
-7.45262951e-02 2.56770372e-01 -3.69383216e-01 -7.08723247e-01
2.65227687e-02 4.06970769e-01 -9.25435841e-01 -2.69859135e-01
1.97082579e-01 6.80971384e-01 -7.07282722e-01 -4.50228155e-01
4.30516809e-01 3.95222783e-01 -5.77023625e-01 7.21352637e-01
-5.21704435e-01 6.61304832e-01 -2.29435861e-01 3.40037882e-01
2.46801391e-01 2.49632403e-01 1.00768462e-01 7.37451687e-02
-3.48016292e-01 1.27639091e+00 -7.99329758e-01 1.78563869e+00
-8.68842185e-01 3.56387168e-01 -8.09672326e-02 -8.04320991e-01
1.08936322e+00 9.47569191e-01 6.62363991e-02 -5.09243131e-01
1.72159821e-01 4.51915950e-01 4.32050377e-02 -3.04254979e-01
6.04402184e-01 -4.06774640e-01 -1.85210004e-01 2.61233211e-01
8.36878270e-02 -5.85488260e-01 5.58352172e-02 -2.62747593e-02
8.12243521e-01 -1.39085621e-01 3.99357021e-01 -1.82866994e-02
7.37879932e-01 -2.54505217e-01 5.34595251e-01 2.19780639e-01
-5.69551326e-02 1.08622146e+00 1.94274187e-01 1.22952878e-01
-9.66028333e-01 -1.14436924e+00 2.13334605e-01 1.08200705e+00
-1.47142187e-01 -7.45534897e-01 -1.05756223e+00 -4.34550971e-01
-5.17828465e-01 1.31928897e+00 -6.51622862e-02 -9.64774638e-02
-8.08484852e-01 1.03436723e-01 7.07487226e-01 5.53022325e-01
1.12817205e-01 -1.57384145e+00 9.31017920e-02 6.06209517e-01
-6.85476542e-01 -1.35262060e+00 -1.04140234e+00 1.28506839e-01
-5.81835926e-01 -3.78585964e-01 -7.56101072e-01 -1.08638632e+00
2.12719306e-01 2.09371522e-01 7.38680720e-01 -8.70644227e-02
4.06944185e-01 -1.97239399e-01 -6.59574509e-01 -3.68817210e-01
-1.25303960e+00 1.82679504e-01 8.31089690e-02 -7.14735091e-02
1.53804824e-01 -6.22785926e-01 -3.62918139e-01 3.44849795e-01
-7.60024726e-01 4.20451373e-01 2.92838067e-01 6.77090168e-01
7.57206678e-01 -2.27720127e-01 9.97113049e-01 -5.10446072e-01
7.89577425e-01 -3.14692408e-01 -3.32558453e-01 8.31323937e-02
-2.27019906e-01 -5.33199161e-02 1.23465550e+00 -2.98129410e-01
-1.23251128e+00 -3.82441282e-02 -9.00697589e-01 -4.73110467e-01
-2.82476574e-01 4.07890946e-01 -5.29357851e-01 8.89463127e-01
4.22439992e-01 6.31506026e-01 -8.37125182e-02 -5.65247476e-01
4.78999466e-01 1.34890735e+00 5.60357451e-01 -5.44396102e-01
4.94443446e-01 -1.18493915e-01 -7.44371891e-01 -1.31438875e+00
-8.83939862e-01 -4.15248960e-01 -3.51367682e-01 -8.69776215e-03
9.66150463e-01 -1.06193149e+00 -1.82674035e-01 1.48917511e-01
-1.40061343e+00 -4.78565454e-01 -4.73397672e-01 6.73292816e-01
-1.01686454e+00 4.26890194e-01 -8.66071761e-01 -8.38809848e-01
-4.51977879e-01 -1.22737527e+00 1.11651540e+00 1.55498430e-01
-5.01265168e-01 -8.01880240e-01 6.45137355e-02 4.77882475e-01
2.55060375e-01 -2.81410843e-01 7.98454523e-01 -7.25258470e-01
-3.60526741e-01 3.56868684e-01 3.22314799e-01 8.47104490e-01
5.30141711e-01 3.11509259e-02 -9.28822458e-01 3.07638999e-02
1.35863841e-01 -1.67114213e-01 5.50680161e-01 3.69268030e-01
1.03155768e+00 -6.90820634e-01 -3.26759508e-03 5.10683358e-01
8.49166036e-01 7.27738917e-01 5.17898917e-01 -2.42846400e-01
5.51999748e-01 5.69172025e-01 6.99966133e-01 2.74572581e-01
2.96312094e-01 5.50663948e-01 -1.18028969e-01 1.24982513e-01
-6.37249053e-01 -8.86887312e-01 8.05533886e-01 1.93159354e+00
4.58682656e-01 -6.46462500e-01 -7.26189315e-01 7.12824762e-01
-1.62362170e+00 -8.70340109e-01 7.93693215e-02 1.96266484e+00
1.38315344e+00 3.50490004e-01 1.65361181e-01 1.77745059e-01
8.55262637e-01 4.83725578e-01 -1.06496140e-01 -7.14376867e-01
3.91986929e-02 2.18690127e-01 -1.18413754e-02 9.93210912e-01
-6.32443607e-01 1.40801930e+00 5.33201933e+00 1.14891255e+00
-1.30215621e+00 1.92086138e-02 2.24073902e-01 -2.04609055e-02
-6.27878428e-01 3.64246368e-02 -8.88780296e-01 7.74555981e-01
1.61768389e+00 -4.59577411e-01 4.73034978e-01 7.98622727e-01
8.79972398e-01 4.18331087e-01 -1.15609467e+00 7.68176913e-01
-9.55511443e-03 -1.30117083e+00 1.75721407e-01 -3.29101115e-01
5.44348478e-01 -1.55743048e-01 -6.24598004e-02 4.23605889e-01
4.70778756e-02 -8.89981091e-01 1.07064784e+00 2.04649761e-01
8.33720803e-01 -8.28628063e-01 3.27817082e-01 5.61980307e-01
-1.47300267e+00 2.61979848e-01 -1.99769273e-01 1.11396067e-01
6.19559824e-01 2.67377198e-01 -1.25850523e+00 5.41176438e-01
3.14675011e-02 4.79585916e-01 1.33430138e-01 4.97631311e-01
-7.48447478e-01 1.29309750e+00 -2.27189809e-01 -9.41281989e-02
3.32321495e-01 -8.08504894e-02 6.25283420e-01 1.56645989e+00
4.86217529e-01 2.65679330e-01 4.04099464e-01 8.23809147e-01
-1.35534495e-01 4.37652409e-01 -4.73616660e-01 -4.76939112e-01
8.71549428e-01 7.56938279e-01 -4.57995087e-01 -4.37083155e-01
-4.30789590e-01 1.09021163e+00 -1.59293130e-01 4.31348324e-01
-8.52505088e-01 -5.15357673e-01 6.31969690e-01 6.34192675e-02
3.26143861e-01 -4.20333594e-01 7.79081741e-03 -9.10695791e-01
-2.44042929e-02 -1.00634718e+00 -1.55125931e-01 -9.11147892e-01
-9.24362898e-01 9.74981010e-01 -2.10441262e-01 -1.29375911e+00
-7.55450785e-01 -1.39958620e-01 -7.11226940e-01 1.09562755e+00
-1.40821111e+00 -1.02132583e+00 2.33362660e-01 2.71336824e-01
1.50367165e+00 -2.21106961e-01 9.75527763e-01 1.06219187e-01
-5.18884599e-01 4.69480932e-01 -1.60909906e-01 2.62323707e-01
6.27538323e-01 -1.10095727e+00 8.22667122e-01 9.31189179e-01
3.25866818e-01 4.84061450e-01 8.06086659e-01 -6.04660869e-01
-8.52359414e-01 -1.21095657e+00 1.32478356e+00 4.04786356e-02
6.45346224e-01 -5.62579453e-01 -1.06534958e+00 6.55876815e-01
6.15891099e-01 -1.55328467e-01 9.04302657e-01 -2.97502637e-01
-7.70321349e-04 4.85983081e-02 -6.26840234e-01 8.34468365e-01
9.34485316e-01 -9.23346043e-01 -1.11176157e+00 9.47899669e-02
1.47303629e+00 -3.45159143e-01 -5.60037374e-01 1.68035291e-02
2.87079930e-01 -6.57677889e-01 5.39475203e-01 -3.19770724e-01
4.57358629e-01 -3.70875686e-01 -5.20005167e-01 -1.54329312e+00
4.38491814e-02 -8.93363476e-01 2.78648362e-02 1.44982088e+00
6.70685947e-01 -2.25853994e-01 4.72896159e-01 -2.98018623e-02
-5.48303664e-01 -5.01364350e-01 -7.56922483e-01 -1.14755285e+00
4.35858294e-02 -6.45320237e-01 6.80167258e-01 6.00452960e-01
3.07301700e-01 9.28704381e-01 -4.64229822e-01 2.64293134e-01
6.19431026e-02 1.86899439e-01 5.55723965e-01 -5.41279495e-01
-4.87671793e-01 -2.87203312e-01 2.80391514e-01 -1.63128710e+00
4.83276695e-01 -8.36409032e-01 6.39439583e-01 -1.56558359e+00
-3.98894519e-01 -2.11067364e-01 -1.97867826e-01 3.48996609e-01
-1.11663453e-01 -3.29215735e-01 4.28171158e-01 9.37898681e-02
-2.28479072e-01 1.04359102e+00 1.12313378e+00 7.73086399e-02
-5.03073215e-01 3.14062387e-01 -3.47615361e-01 6.56763673e-01
1.08759177e+00 -5.82516193e-01 -6.76885545e-01 -1.82813883e-01
-5.17104268e-01 7.26719022e-01 -2.56985575e-01 -9.37303603e-01
2.10405692e-01 -3.72816324e-01 -3.25928837e-01 -7.35967398e-01
6.31063044e-01 -6.10080779e-01 -2.47520357e-02 3.58339489e-01
-7.34303772e-01 -1.34242266e-01 1.95805177e-01 2.17240274e-01
-6.08629405e-01 -4.40639049e-01 8.25827599e-01 1.05508568e-03
-3.44372362e-01 5.71409315e-02 -7.34088004e-01 1.73614576e-01
6.02620065e-01 -1.20465755e-01 5.61655350e-02 -5.90973198e-01
-7.79403448e-01 -7.25759612e-03 1.56276107e-01 8.22090387e-01
7.38158941e-01 -1.40018916e+00 -6.56346023e-01 3.63267511e-01
5.61669841e-02 -1.38352051e-01 1.44195601e-01 2.81228721e-01
-2.60575622e-01 6.59967184e-01 8.97622779e-02 -3.97578895e-01
-1.26274920e+00 5.10064483e-01 1.29826829e-01 8.78817961e-02
-5.36047876e-01 7.52058566e-01 3.97124290e-01 -3.89399201e-01
3.51219416e-01 -8.76418471e-01 2.49139946e-02 -1.48156404e-01
6.20957017e-01 -5.15333340e-02 -5.50762117e-02 -8.41491699e-01
-2.54688114e-01 1.57774955e-01 9.20088738e-02 -6.84969962e-01
9.79138136e-01 -4.15151894e-01 3.12852621e-01 7.84010947e-01
1.25587118e+00 5.60103476e-01 -1.23123932e+00 -1.38889715e-01
1.07816480e-01 -1.33269891e-01 -1.00824460e-01 -6.90684557e-01
-6.54888690e-01 1.10848939e+00 -8.95548016e-02 3.64092827e-01
1.13293314e+00 -5.33886366e-02 1.45965397e+00 2.09300309e-01
4.99193976e-03 -1.26584363e+00 1.67796522e-01 8.33133876e-01
1.23927665e+00 -9.39471126e-01 -7.76473284e-01 -6.09297991e-01
-8.86423767e-01 1.13520956e+00 3.36449653e-01 9.99750197e-03
6.39710009e-01 4.33625877e-01 3.38347226e-01 4.16299164e-01
-9.58752573e-01 -2.18823209e-01 1.48380488e-01 4.54109520e-01
5.92324257e-01 3.49682689e-01 -4.22606468e-01 9.03521955e-01
-8.95972550e-01 -2.16318354e-01 5.25320172e-01 5.76856911e-01
-9.26666558e-01 -1.49520028e+00 -2.20313400e-01 -3.00029486e-01
-4.62361217e-01 -4.28432107e-01 -4.02478367e-01 2.33638033e-01
-2.07225025e-01 1.26562333e+00 7.13189691e-02 -4.83589709e-01
5.01373768e-01 4.30242121e-01 3.46604362e-02 -1.13906693e+00
-4.64607984e-01 6.11946285e-01 3.35865319e-01 -3.55975002e-01
-1.05767436e-01 -4.32356179e-01 -1.73952484e+00 6.70516789e-02
-1.72110617e-01 6.56009078e-01 6.66206598e-01 9.19891000e-01
2.49275461e-01 8.35373998e-01 1.04621649e+00 -5.31692266e-01
-5.07442355e-01 -1.08313632e+00 -4.29591507e-01 -8.26200992e-02
5.29675484e-01 -2.61486378e-02 -4.37188894e-01 2.96709210e-01] | [14.796871185302734, 6.770724773406982] |
22fde00f-08b3-4266-8e11-daa426aeb10c | grounded-human-object-interaction-hotspots | 1812.04558 | null | http://arxiv.org/abs/1812.04558v2 | http://arxiv.org/pdf/1812.04558v2.pdf | Grounded Human-Object Interaction Hotspots from Video | Learning how to interact with objects is an important step towards embodied
visual intelligence, but existing techniques suffer from heavy supervision or
sensing requirements. We propose an approach to learn human-object interaction
"hotspots" directly from video. Rather than treat affordances as a manually
supervised semantic segmentation task, our approach learns about interactions
by watching videos of real human behavior and anticipating afforded actions.
Given a novel image or video, our model infers a spatial hotspot map indicating
how an object would be manipulated in a potential interaction-- even if the
object is currently at rest. Through results with both first and third person
video, we show the value of grounding affordances in real human-object
interactions. Not only are our weakly supervised hotspots competitive with
strongly supervised affordance methods, but they can also anticipate object
interaction for novel object categories. | ['Tushar Nagarajan', 'Kristen Grauman', 'Christoph Feichtenhofer'] | 2018-12-11 | grounded-human-object-interaction-hotspots-2 | http://openaccess.thecvf.com/content_ICCV_2019/html/Nagarajan_Grounded_Human-Object_Interaction_Hotspots_From_Video_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Nagarajan_Grounded_Human-Object_Interaction_Hotspots_From_Video_ICCV_2019_paper.pdf | iccv-2019-10 | ['video-to-image-affordance-grounding'] | ['computer-vision'] | [ 5.17093420e-01 5.14787138e-01 -2.75435895e-01 -4.70122218e-01
-1.44500867e-01 -6.79039359e-01 6.07217371e-01 -4.37745042e-02
-4.43613350e-01 5.40706933e-01 4.50214088e-01 9.30204019e-02
-2.37437547e-03 -4.61948097e-01 -1.04131007e+00 -1.99025437e-01
-4.56081957e-01 5.45599401e-01 3.66381049e-01 -9.96330082e-02
2.80338943e-01 5.88090420e-01 -1.69509542e+00 6.27233744e-01
4.20128196e-01 7.03529239e-01 5.24471283e-01 1.03226328e+00
2.55023271e-01 1.17128909e+00 -3.00129622e-01 5.76206632e-02
3.72419089e-01 -3.55873168e-01 -1.32616115e+00 6.26171947e-01
5.13640285e-01 -7.40340889e-01 -4.70695555e-01 8.80547881e-01
-7.29944408e-02 5.48543334e-01 5.53078413e-01 -1.58111107e+00
-6.34575129e-01 6.83838189e-01 -2.36476928e-01 2.92003870e-01
1.08765817e+00 6.39517248e-01 1.09180081e+00 -5.23630440e-01
9.81851459e-01 1.36261892e+00 1.76165462e-01 8.41730177e-01
-1.13848078e+00 4.24313964e-03 7.04530716e-01 2.59067982e-01
-9.70937788e-01 -5.83777666e-01 7.21208453e-01 -7.17097282e-01
1.29774916e+00 3.52331638e-01 1.11194789e+00 1.03937232e+00
-2.75567353e-01 1.38676298e+00 8.04688036e-01 -3.38362455e-01
3.68532628e-01 -1.84285209e-01 3.39476056e-02 9.45346117e-01
-1.24046937e-01 2.83112288e-01 -6.32253885e-01 8.20918530e-02
1.13960123e+00 3.31155956e-01 -3.41071248e-01 -7.14282513e-01
-1.79736948e+00 1.49984851e-01 7.21867025e-01 2.93953985e-01
-5.13345957e-01 5.18381059e-01 -6.74035922e-02 1.52126804e-01
1.66414976e-01 6.30086422e-01 -5.07001996e-01 -3.03795010e-01
-5.07172644e-01 6.13281906e-01 6.52971447e-01 1.29633439e+00
7.34195590e-01 -3.10605615e-01 -2.62505442e-01 -4.48476477e-03
9.25607309e-02 1.74421385e-01 -3.40704946e-03 -1.41306388e+00
4.35065866e-01 8.11416626e-01 6.32379711e-01 -8.50705504e-01
-5.01281142e-01 3.96734834e-01 -4.46684919e-02 3.70963246e-01
6.80023551e-01 -1.21596858e-01 -1.15717244e+00 1.53869915e+00
3.55430186e-01 2.11418450e-01 -1.48851459e-03 1.39567542e+00
5.03767431e-01 4.72777724e-01 3.80623400e-01 7.30881393e-02
1.34457481e+00 -1.28615344e+00 -5.62604964e-01 -6.32226229e-01
6.05199993e-01 -1.85337495e-02 1.22350967e+00 3.66026819e-01
-1.00805533e+00 -6.80607796e-01 -8.36950719e-01 -3.52588892e-01
-1.92269325e-01 -2.27412611e-01 1.20479774e+00 -3.52158919e-02
-8.32621813e-01 7.17175841e-01 -1.26383829e+00 -6.76618457e-01
5.47206700e-01 2.60801792e-01 -6.51972651e-01 7.67429844e-02
-4.80741233e-01 7.81221926e-01 5.82462609e-01 -8.36576000e-02
-1.38586259e+00 -3.96158785e-01 -1.16189981e+00 1.27281761e-02
8.20974886e-01 -6.09698117e-01 1.36592233e+00 -1.34885907e+00
-1.26126432e+00 9.13909912e-01 -2.74214059e-01 -4.27189112e-01
5.86205065e-01 -6.49692237e-01 -1.73748150e-01 4.38582748e-01
1.19675577e-01 1.09764862e+00 8.00799847e-01 -1.49564862e+00
-6.00886583e-01 -4.55211222e-01 7.07745612e-01 5.01494706e-01
1.17552899e-01 -8.62545148e-02 -2.63483912e-01 -5.06806254e-01
2.52262443e-01 -1.02987528e+00 -4.15436953e-01 4.37833875e-01
-5.20023942e-01 -4.11324084e-01 8.44263196e-01 -3.97285134e-01
5.58013797e-01 -2.06411433e+00 5.63938141e-01 -6.95684180e-02
1.02167390e-01 -1.85143515e-01 8.28723051e-03 2.22533599e-01
1.26591390e-02 4.91773263e-02 -3.83317024e-02 -1.56021029e-01
1.47417560e-01 2.65400916e-01 -2.74650186e-01 3.04756969e-01
2.11080998e-01 1.18578565e+00 -1.37282884e+00 -3.67218614e-01
5.18858254e-01 2.76644498e-01 -8.39115739e-01 6.04404688e-01
-8.60968709e-01 7.91571438e-01 -5.59634447e-01 7.94148386e-01
-4.04221043e-02 -3.26767385e-01 1.56972364e-01 -1.25564143e-01
1.51841596e-01 2.67656650e-02 -1.03392756e+00 2.28287458e+00
-8.39795992e-02 5.31156421e-01 -1.41930118e-01 -7.94255912e-01
2.22765878e-01 3.11922342e-01 3.94917041e-01 -2.93303192e-01
-1.37695903e-02 -2.46146709e-01 8.53059255e-03 -1.26658857e+00
2.87601620e-01 -1.11415453e-01 -1.30673900e-01 5.16582787e-01
3.48650515e-01 -3.16049814e-01 1.66270301e-01 2.79868186e-01
1.01129210e+00 8.81389737e-01 2.53125399e-01 -8.42232257e-02
-8.01907480e-02 3.81879896e-01 2.16574013e-01 8.29682112e-01
-4.24974918e-01 5.86085141e-01 3.23975563e-01 -8.16077352e-01
-8.72639477e-01 -1.20426035e+00 4.02270615e-01 1.39799511e+00
5.87275505e-01 -1.87407672e-01 -7.25309849e-01 -7.35661983e-01
-1.46078363e-01 6.94856882e-01 -6.85231328e-01 -3.21725709e-03
-5.42955637e-01 -3.07487398e-02 -1.01109281e-01 7.81771600e-01
3.69415015e-01 -1.57369614e+00 -1.37976956e+00 -1.47673205e-01
-2.93059319e-01 -1.28389847e+00 -6.15722954e-01 1.19505845e-01
-7.79779375e-01 -1.21429718e+00 -5.89615405e-01 -7.72915065e-01
9.99467671e-01 2.11167753e-01 1.04086936e+00 1.55393392e-01
-5.21783710e-01 9.01096404e-01 -2.94907629e-01 -1.95508093e-01
-6.65240437e-02 -2.46117502e-01 2.49572933e-01 2.96092127e-02
4.54481959e-01 -4.61759388e-01 -7.74234772e-01 8.51765797e-02
-4.73561019e-01 4.63939339e-01 3.83005261e-01 3.95687759e-01
2.72235572e-01 -1.17825218e-01 1.06216492e-02 -6.76545799e-01
1.72930732e-01 -2.76252687e-01 -2.09160343e-01 2.22381100e-01
-4.05056030e-02 5.94971608e-03 1.73495501e-01 -6.34402096e-01
-1.24242246e+00 5.00458479e-01 3.84757757e-01 -3.61523002e-01
-8.69337082e-01 9.81689170e-02 -2.27332115e-01 3.31426233e-01
7.20368445e-01 -2.39428908e-01 -3.49050581e-01 -2.84412980e-01
6.05525255e-01 1.29246682e-01 7.42516398e-01 -8.59387934e-01
5.88705420e-01 8.01432788e-01 -4.29115146e-01 -7.82405972e-01
-9.80047762e-01 -4.38964933e-01 -1.14941621e+00 -4.35984403e-01
1.35704720e+00 -9.19354677e-01 -1.16957355e+00 1.39143378e-01
-1.12483013e+00 -8.56289446e-01 -4.55425829e-01 4.97109234e-01
-9.31944728e-01 -2.85419635e-02 -4.25455004e-01 -8.75898719e-01
5.17632782e-01 -1.01871741e+00 1.33841574e+00 2.72352397e-01
-8.14421415e-01 -7.71926761e-01 -3.18666875e-01 5.77789009e-01
-2.48652846e-01 4.33458805e-01 6.03777707e-01 -4.75254208e-01
-1.02095962e+00 -1.42859057e-01 -1.08878545e-01 -1.72144145e-01
2.50350416e-01 -3.31614524e-01 -8.57205391e-01 -1.24195181e-01
-3.19916278e-01 -5.61867714e-01 5.24813950e-01 1.32815123e-01
1.11028659e+00 -6.27968490e-01 -7.24245489e-01 4.20091063e-01
1.07540643e+00 2.97164023e-01 3.75101358e-01 7.59298652e-02
1.01140094e+00 8.42683315e-01 8.52546215e-01 3.30850691e-01
4.60631549e-01 6.18834555e-01 5.23146272e-01 3.57950702e-02
1.30640894e-01 -5.85332751e-01 3.79761368e-01 -1.82074711e-01
-5.20657539e-01 -1.78280979e-01 -9.07533705e-01 5.90868056e-01
-2.10388732e+00 -1.12757409e+00 2.24407509e-01 1.80298281e+00
5.77149272e-01 2.58164585e-01 4.24162686e-01 -1.27170131e-01
3.11200976e-01 2.74338219e-02 -9.94025588e-01 -8.65150094e-02
3.06671023e-01 -1.91816971e-01 1.29963785e-01 5.50373614e-01
-1.03412735e+00 1.18554580e+00 6.68466997e+00 -1.97657391e-01
-6.58465922e-01 2.34508663e-02 2.20859736e-01 -3.71243149e-01
-4.52597588e-02 2.66318291e-01 -2.59781688e-01 -8.54967907e-03
2.97049433e-01 3.12188268e-01 6.65009618e-01 9.10658419e-01
1.65352657e-01 -4.74510878e-01 -1.90018785e+00 8.65947485e-01
2.35650778e-01 -9.70128536e-01 8.85671750e-02 -1.59575626e-01
4.99094903e-01 -2.52877384e-01 -2.27681428e-01 1.19071014e-01
4.60584372e-01 -1.15907288e+00 1.06275892e+00 7.43580222e-01
4.45948571e-01 -1.56387106e-01 2.08956767e-02 6.51455581e-01
-9.10400927e-01 -3.44965339e-01 9.66438055e-02 -7.71531224e-01
3.33369642e-01 -1.76216528e-01 -8.01984429e-01 -1.10339172e-01
9.18390870e-01 8.20821524e-01 -3.83493423e-01 6.75240159e-01
-5.05819380e-01 7.79236155e-03 -3.23715359e-01 -1.76642612e-02
1.16679095e-01 1.55694233e-02 6.65660381e-01 8.60299766e-01
-1.54574495e-02 6.23139799e-01 5.92284739e-01 1.01086283e+00
2.90618449e-01 -5.26672244e-01 -9.43240404e-01 -3.46921325e-01
6.52953908e-02 8.94113421e-01 -9.65313077e-01 -5.64513981e-01
-2.63850719e-01 1.45039201e+00 3.55535686e-01 7.24622309e-01
-6.23852432e-01 -1.36174604e-01 6.91413879e-01 2.98050433e-01
1.19514175e-01 -4.82828259e-01 1.71474591e-02 -1.34059942e+00
2.61593193e-01 -5.85099995e-01 2.20419213e-01 -1.42260170e+00
-6.99972689e-01 3.48970592e-01 3.82865191e-01 -1.15366375e+00
-1.96695626e-01 -7.19427049e-01 -3.38463336e-01 1.99570104e-01
-9.57505107e-01 -1.27094269e+00 -6.33631527e-01 6.42059803e-01
9.17274058e-01 1.71716109e-01 7.44282722e-01 -3.36208522e-01
8.03390667e-02 -8.06013793e-02 -9.02614951e-01 2.10720375e-01
1.66396797e-01 -1.35665083e+00 3.87563139e-01 6.13082767e-01
6.50552690e-01 6.41470432e-01 8.91938150e-01 -6.62027657e-01
-1.42809224e+00 -6.66734099e-01 5.19253731e-01 -1.18410051e+00
4.47656840e-01 -7.03230441e-01 -6.91999733e-01 1.43691623e+00
2.09041715e-01 2.37437725e-01 2.98565149e-01 9.69637334e-02
-2.21411437e-01 4.13908064e-01 -1.09010243e+00 1.05868852e+00
1.89259183e+00 -7.73218572e-01 -1.16425300e+00 6.84742749e-01
9.29174483e-01 -5.43784857e-01 -5.48964560e-01 4.09489751e-01
6.66118562e-01 -1.00363243e+00 1.14408052e+00 -1.22266614e+00
2.56178796e-01 -4.72821295e-01 -9.04999897e-02 -1.11253393e+00
-1.73308358e-01 -6.27136469e-01 -4.15395170e-01 6.27925158e-01
1.67281017e-01 -8.90256241e-02 8.22434664e-01 1.09367597e+00
4.47315425e-02 -2.78145313e-01 -4.01560664e-01 -5.53807974e-01
-5.94542742e-01 -5.07742524e-01 5.49697995e-01 7.59653687e-01
6.91022336e-01 2.56760627e-01 -2.28398725e-01 2.77563304e-01
5.56497872e-01 9.56887677e-02 1.01616836e+00 -1.05631649e+00
-3.93501222e-01 -2.42488250e-01 -6.05389118e-01 -1.40909231e+00
4.66817051e-01 -6.75666749e-01 4.36713129e-01 -1.50569355e+00
3.48161906e-01 -1.44687861e-01 -7.79311359e-02 8.23010504e-01
-1.17426157e-01 1.46249354e-01 3.18188518e-01 2.16118082e-01
-1.03150654e+00 1.78971395e-01 1.48476243e+00 -2.17768654e-01
-4.64233071e-01 -1.81194544e-01 -3.08425397e-01 1.07868135e+00
5.16389966e-01 -1.56606734e-01 -6.93325698e-01 -6.51132047e-01
1.33990109e-01 3.88957918e-01 1.06397474e+00 -1.10710835e+00
2.77058840e-01 -5.25766432e-01 5.23389459e-01 -2.27231532e-01
5.43170750e-01 -1.05885518e+00 5.40597886e-02 3.76380026e-01
-6.96444333e-01 -1.04480505e-01 5.18695116e-02 8.39562833e-01
7.09449723e-02 1.43786013e-01 1.10921659e-01 -6.18987203e-01
-1.34672558e+00 2.79860556e-01 -4.64987725e-01 -1.12800911e-01
1.38240659e+00 -5.22002876e-01 -7.37860203e-02 -3.04854542e-01
-1.51558506e+00 3.54613453e-01 7.30485857e-01 8.86519015e-01
7.90834785e-01 -1.04490399e+00 -1.18603341e-01 2.71693081e-01
3.85234714e-01 2.59838432e-01 1.79436535e-01 4.18183357e-01
-3.86550725e-01 1.35230348e-01 -2.21679598e-01 -8.17214310e-01
-9.84405518e-01 9.76514995e-01 3.03666502e-01 5.77479959e-01
-9.58087444e-01 9.86058712e-01 4.66794699e-01 -4.26188529e-01
5.37955761e-01 -6.85195088e-01 1.34883467e-02 -4.56793159e-01
4.45862889e-01 1.07236944e-01 -5.50698400e-01 -6.79292381e-01
-2.56653339e-01 4.42603558e-01 1.04671352e-01 -2.63642341e-01
1.11763394e+00 -2.20585078e-01 -3.92509885e-02 1.02907145e+00
9.09157813e-01 -5.78951597e-01 -1.98685646e+00 -6.42342046e-02
4.03215259e-01 -6.70981288e-01 -4.54004139e-01 -8.90169561e-01
-6.02002561e-01 8.72945487e-01 2.81123340e-01 6.82511330e-02
7.19988108e-01 5.34883022e-01 5.30451119e-01 8.83358002e-01
8.62419963e-01 -9.73127365e-01 6.92622125e-01 2.05127507e-01
9.22443628e-01 -1.50204790e+00 -3.69776376e-02 -3.06440324e-01
-9.14827228e-01 8.08062017e-01 1.00063729e+00 -2.39437938e-01
5.37341237e-01 2.07056016e-01 -8.10490176e-02 -5.30399919e-01
-6.32171214e-01 -2.66116947e-01 4.25292253e-01 7.45321095e-01
1.38609052e-01 1.54703945e-01 6.19383156e-01 2.30589062e-01
-1.38936922e-01 2.52465338e-01 3.23613673e-01 1.13939261e+00
-5.11328816e-01 -3.85718793e-01 -4.90783155e-02 3.72548044e-01
-4.02761288e-02 2.12667063e-01 -5.58865547e-01 6.92307651e-01
2.66575277e-01 8.39227617e-01 3.21888983e-01 -1.73499659e-01
4.08946335e-01 1.68176249e-01 9.43087339e-01 -1.11327589e+00
-2.06000194e-01 -1.38685107e-01 -5.28715365e-02 -1.07909465e+00
-9.22651410e-01 -8.98512065e-01 -1.80028164e+00 1.70295686e-01
1.23988383e-03 -2.40447558e-02 1.37754291e-01 1.25631404e+00
1.38822049e-01 4.58746642e-01 3.72704379e-02 -1.46609509e+00
2.31853258e-02 -7.82732546e-01 -4.17449325e-01 8.46215427e-01
7.84297228e-01 -7.65627623e-01 -2.31277734e-01 6.37513101e-01] | [5.120748519897461, 0.18832671642303467] |
403e4163-0831-487f-9ddb-1a743ce33437 | generative-language-models-for-paragraph | 2210.03992 | null | https://arxiv.org/abs/2210.03992v3 | https://arxiv.org/pdf/2210.03992v3.pdf | Generative Language Models for Paragraph-Level Question Generation | Powerful generative models have led to recent progress in question generation (QG). However, it is difficult to measure advances in QG research since there are no standardized resources that allow a uniform comparison among approaches. In this paper, we introduce QG-Bench, a multilingual and multidomain benchmark for QG that unifies existing question answering datasets by converting them to a standard QG setting. It includes general-purpose datasets such as SQuAD for English, datasets from ten domains and two styles, as well as datasets in eight different languages. Using QG-Bench as a reference, we perform an extensive analysis of the capabilities of language models for the task. First, we propose robust QG baselines based on fine-tuning generative language models. Then, we complement automatic evaluation based on standard metrics with an extensive manual evaluation, which in turn sheds light on the difficulty of evaluating QG models. Finally, we analyse both the domain adaptability of these models as well as the effectiveness of multilingual models in languages other than English. QG-Bench is released along with the fine-tuned models presented in the paper https://github.com/asahi417/lm-question-generation, which are also available as a demo https://autoqg.net/. | ['Jose Camacho-Collados', 'Fernando Alva-Manchego', 'Asahi Ushio'] | 2022-10-08 | null | null | null | null | ['question-generation'] | ['natural-language-processing'] | [-2.43798614e-01 5.94853237e-02 1.03340290e-01 -2.37565473e-01
-1.70739317e+00 -1.02524400e+00 8.95935953e-01 -2.74743959e-02
-3.74554545e-01 1.01471150e+00 5.25187254e-01 -5.11231363e-01
-1.23807006e-01 -8.32609057e-01 -5.79256773e-01 -2.35681668e-01
4.56953913e-01 8.54005873e-01 1.58536494e-01 -7.22324371e-01
4.70555983e-02 -1.40191957e-01 -1.32196867e+00 4.08609688e-01
1.10316396e+00 6.86236858e-01 -5.10007367e-02 8.68192792e-01
-1.85639992e-01 8.14494073e-01 -9.22698140e-01 -9.92730260e-01
1.23405725e-01 -6.79935336e-01 -1.20229280e+00 -2.13432163e-01
7.48080790e-01 -2.35943869e-01 -1.60101816e-01 6.66241586e-01
8.00364435e-01 1.33451179e-01 4.91098046e-01 -1.25515604e+00
-1.15927017e+00 7.43475914e-01 1.57258317e-01 2.29853854e-01
6.29666626e-01 4.04600471e-01 1.58558786e+00 -8.84026527e-01
8.00234377e-01 1.27600276e+00 4.86341625e-01 6.81093752e-01
-1.20825768e+00 -4.21263218e-01 -1.26461640e-01 3.53987396e-01
-1.12533259e+00 -5.24344385e-01 5.34114063e-01 -2.45551527e-01
1.18143988e+00 3.28302532e-01 1.82607442e-01 1.40831518e+00
2.08115466e-02 7.86422014e-01 1.27212882e+00 -6.21207356e-01
9.38511193e-02 1.00803405e-01 1.31753877e-01 3.50236177e-01
8.06087628e-02 3.59378234e-02 -5.77936411e-01 -2.97046036e-01
2.23307967e-01 -7.52482414e-01 -2.47785598e-01 -3.49482149e-02
-1.15650272e+00 9.62756276e-01 6.92271143e-02 2.26448447e-01
-1.33640572e-01 8.82920399e-02 2.78459728e-01 5.70318878e-01
4.50744629e-01 7.17292786e-01 -5.85573018e-01 -2.35717148e-01
-8.85342836e-01 9.46830809e-01 1.08248281e+00 1.15903521e+00
6.87121212e-01 -5.62915429e-02 -6.68253362e-01 1.04183435e+00
2.28655219e-01 6.51556671e-01 5.35803020e-01 -1.15477312e+00
7.44064152e-01 5.10486782e-01 1.83081001e-01 -4.84687030e-01
-1.90500185e-01 -3.09427828e-01 -3.86587203e-01 -2.97212899e-01
6.92919850e-01 -3.34565461e-01 -6.13742054e-01 1.88911438e+00
1.61488503e-01 -3.98356229e-01 2.23908961e-01 5.77951431e-01
1.33278060e+00 6.33539021e-01 1.15324363e-01 2.70447761e-01
1.66855240e+00 -1.14889240e+00 -5.93765199e-01 -2.83266455e-01
7.10329533e-01 -9.75504339e-01 1.55907655e+00 2.15603903e-01
-1.20813262e+00 -5.88408053e-01 -6.08375132e-01 -4.54081863e-01
-5.10090172e-01 3.23154517e-02 3.86999309e-01 7.57104576e-01
-1.35891867e+00 8.17191452e-02 -3.56976926e-01 -5.74973524e-01
-1.07252665e-01 -1.55664667e-01 -6.59644529e-02 -1.88446447e-01
-1.81168616e+00 9.03380871e-01 4.44151580e-01 -3.53652358e-01
-8.91606629e-01 -6.71706498e-01 -7.79230952e-01 -2.19673991e-01
4.80470210e-01 -1.03868496e+00 1.86652386e+00 -6.03142738e-01
-1.49231386e+00 9.99244690e-01 -8.09754878e-02 -3.52562070e-01
5.01160502e-01 -2.66137481e-01 -6.66729212e-01 9.21725929e-02
4.76914078e-01 7.22909570e-01 5.18212020e-01 -8.78735483e-01
-3.76372129e-01 -1.04319423e-01 5.75988531e-01 1.34315342e-01
-1.21163670e-02 1.90857723e-01 -4.72606212e-01 -7.72030056e-01
-5.32715023e-01 -8.43810201e-01 -3.86183858e-02 -6.21965945e-01
-4.18420672e-01 -5.98987579e-01 2.70509571e-01 -7.90172338e-01
1.32039893e+00 -1.78525484e+00 -2.12109387e-01 -3.17580938e-01
-5.58655113e-02 1.55019537e-01 -5.53005993e-01 1.00031650e+00
4.16251495e-02 1.57536730e-01 -3.49030852e-01 -3.73030484e-01
3.78086269e-01 1.57377049e-01 -4.89588857e-01 -1.85153827e-01
5.07157922e-01 1.45245445e+00 -1.01585162e+00 -3.80960733e-01
-2.21785039e-01 1.15174837e-01 -8.32703233e-01 3.38665009e-01
-6.67771339e-01 2.91650027e-01 -2.12018684e-01 5.69423258e-01
3.64979565e-01 -4.39135194e-01 1.67299002e-01 1.37422532e-01
8.84426311e-02 1.03236532e+00 -9.77896333e-01 1.82664001e+00
-5.54571092e-01 2.41537675e-01 -2.70533293e-01 -5.32618284e-01
8.85546625e-01 4.65467423e-01 2.02172715e-02 -9.11518633e-01
-6.27709553e-02 4.50331926e-01 -1.79304972e-01 -3.14403594e-01
8.76117408e-01 -2.29907230e-01 -3.21285069e-01 7.52988338e-01
5.65020382e-01 -5.64989269e-01 8.25053453e-01 5.04246771e-01
9.62237835e-01 3.05977643e-01 3.05970937e-01 -3.24790746e-01
5.16863942e-01 1.44180804e-01 1.36052161e-01 9.05227244e-01
-5.95977157e-02 6.97722256e-01 4.73564506e-01 1.16735972e-01
-8.37078989e-01 -1.19097221e+00 -7.77093694e-02 1.38933635e+00
-2.20240280e-01 -8.07367623e-01 -7.74452209e-01 -9.49208975e-01
-4.15911935e-02 1.10345030e+00 -5.93650043e-01 -1.15517572e-01
-4.93669748e-01 -9.69537079e-01 7.75801182e-01 3.98320138e-01
4.46813136e-01 -1.35086417e+00 -1.41174912e-01 2.81854749e-01
-5.96844614e-01 -1.23964667e+00 -4.37482804e-01 -9.60534140e-02
-6.99958801e-01 -1.03459787e+00 -6.94260299e-01 -5.31933486e-01
4.47971039e-02 3.92593555e-02 2.13286185e+00 -1.68139145e-01
2.06079766e-01 5.83229959e-01 -7.48928010e-01 -3.56602877e-01
-7.84753740e-01 5.92168331e-01 -4.18341190e-01 -4.91532505e-01
4.40207541e-01 -3.14026952e-01 -6.00558281e-01 2.15503365e-01
-1.08247411e+00 -9.65426266e-02 3.54017615e-01 7.56839871e-01
4.96858835e-01 -6.13998115e-01 1.21631145e+00 -1.20913911e+00
1.14585769e+00 -7.84484625e-01 -5.09483695e-01 4.72266465e-01
-7.31171429e-01 1.85230687e-01 6.67218864e-01 5.33946939e-02
-1.12804067e+00 -8.44724655e-01 -7.03649580e-01 3.60019445e-01
-5.20216376e-02 5.34796715e-01 -3.13412070e-01 4.60589021e-01
1.00895047e+00 2.20880538e-01 -4.12619114e-01 -5.80262780e-01
1.03757107e+00 6.46401584e-01 4.12046880e-01 -8.88935626e-01
7.65877664e-01 1.77003238e-02 -6.81004524e-01 -6.68263197e-01
-1.05809641e+00 -4.16362733e-01 -2.06704974e-01 5.88233769e-02
6.41338587e-01 -8.59388471e-01 -9.79168713e-02 3.22201967e-01
-1.07139134e+00 -5.89320183e-01 -5.16814291e-01 1.09351709e-01
-6.82265162e-01 2.99734652e-01 -6.61842465e-01 -4.19405937e-01
-6.01671398e-01 -8.81576896e-01 1.12269044e+00 5.94057478e-02
-4.20093358e-01 -1.31115198e+00 5.54293275e-01 9.11020994e-01
4.74951565e-01 -1.82116657e-01 8.31031859e-01 -7.89447188e-01
-4.93172914e-01 1.33813083e-01 -6.99659437e-02 5.56752682e-01
-4.54314835e-02 -9.24595296e-02 -1.00544822e+00 -2.61911899e-01
-1.67424813e-01 -8.77403498e-01 8.55554521e-01 6.50198981e-02
7.17916489e-01 -2.53391266e-01 2.47573912e-01 4.45750475e-01
1.31266809e+00 -1.10654853e-01 6.31478250e-01 5.15901864e-01
4.34562892e-01 5.68359911e-01 4.60873097e-01 1.14487104e-01
1.10784233e+00 6.70155346e-01 1.24418631e-01 1.27337441e-01
-4.30658281e-01 -3.41140360e-01 4.73865777e-01 1.20907438e+00
1.78292111e-01 -5.65752804e-01 -9.84922945e-01 7.66761839e-01
-1.55577612e+00 -7.52872050e-01 -1.64281934e-01 2.06510663e+00
1.20977032e+00 -1.53855443e-01 4.33744162e-01 -1.65385425e-01
2.58928835e-01 2.58316904e-01 -2.31022611e-01 -3.82636607e-01
-4.75680798e-01 6.61370993e-01 -4.34781499e-02 6.76749885e-01
-9.15422618e-01 1.00164187e+00 6.62536526e+00 8.12443435e-01
-6.87385321e-01 4.49037582e-01 4.44761187e-01 -2.96902210e-02
-8.87161076e-01 1.60010532e-01 -9.82062459e-01 3.73825788e-01
1.26376009e+00 -5.43753982e-01 2.69095480e-01 5.81212640e-01
-1.17779814e-01 1.16048709e-01 -9.78487134e-01 6.30746365e-01
2.69265890e-01 -1.07776988e+00 2.40714878e-01 -3.05613518e-01
9.28096175e-01 3.16840231e-01 -1.15687754e-02 8.57008576e-01
7.15596795e-01 -8.91230941e-01 8.17120969e-01 3.23642880e-01
8.12492073e-01 -5.06667793e-01 7.09694564e-01 3.08151931e-01
-6.94554210e-01 1.01086259e-01 -3.60500455e-01 3.77221406e-02
3.12934101e-01 4.07353520e-01 -8.45438957e-01 9.62178409e-01
6.60785913e-01 2.32896581e-01 -9.88204002e-01 7.32905805e-01
-6.67996109e-01 1.05633569e+00 -5.39425723e-02 -3.76034603e-02
2.46502727e-01 -1.64759442e-01 4.02957678e-01 1.32962394e+00
1.94278628e-01 -1.62220523e-01 -7.96340704e-02 1.00556791e+00
-3.92596483e-01 5.29673338e-01 -3.78822833e-01 -1.38887703e-01
4.14736152e-01 1.28742945e+00 -1.38321251e-01 -3.87854069e-01
-6.78840697e-01 7.54797816e-01 5.43957889e-01 4.92360622e-01
-6.78241432e-01 -3.59669358e-01 4.04439777e-01 2.37037405e-01
-8.48327950e-02 -2.00314403e-01 1.40312359e-01 -1.45999813e+00
2.85902265e-02 -1.49441171e+00 7.17610896e-01 -8.25596690e-01
-1.57438707e+00 7.37230599e-01 2.11790323e-01 -8.52264166e-01
-8.82097065e-01 -4.95069265e-01 -2.41348863e-01 1.07541513e+00
-1.64495265e+00 -1.17371333e+00 -2.58985549e-01 4.95840818e-01
4.51659262e-01 -1.14710778e-01 9.62862730e-01 3.96367878e-01
-3.58535290e-01 7.89993823e-01 1.51379496e-01 1.26757428e-01
1.15151083e+00 -1.57497072e+00 1.05208087e+00 7.93502271e-01
6.15915418e-01 6.64149582e-01 4.33195263e-01 -4.41891789e-01
-1.01003444e+00 -1.19877899e+00 1.34802926e+00 -1.19298232e+00
8.93359363e-01 -5.35619915e-01 -1.00917840e+00 8.01092267e-01
5.85157633e-01 -3.45627278e-01 7.70945966e-01 2.24963918e-01
-4.29658353e-01 -4.14674953e-02 -7.13002443e-01 4.69129533e-01
8.37024927e-01 -6.78633213e-01 -8.66529763e-01 4.67412800e-01
8.11936438e-01 -2.77913362e-01 -8.86499465e-01 2.00867161e-01
2.31988296e-01 -1.08059812e+00 7.78603792e-01 -7.54266083e-01
5.81135392e-01 -1.30078629e-01 -2.60276824e-01 -1.37816310e+00
-2.90472955e-01 -5.76332211e-01 1.91313941e-02 1.50424981e+00
7.20173419e-01 -7.95553446e-01 4.97360975e-01 2.26888850e-01
-6.89590871e-02 -7.03237116e-01 -7.69115150e-01 -8.56403172e-01
7.64044762e-01 -5.95869541e-01 7.15236604e-01 8.37957680e-01
-4.55455095e-01 7.65826702e-01 -1.10778853e-01 -3.38983446e-01
4.63483036e-01 2.91947007e-01 9.14249718e-01 -7.42878973e-01
-6.49372578e-01 -3.90981525e-01 1.82994097e-01 -1.08457780e+00
1.05393909e-01 -1.29885602e+00 -1.69759452e-01 -1.80601192e+00
1.71866655e-01 -2.53477812e-01 -1.42683610e-01 4.10768300e-01
-7.82622337e-01 4.48652983e-01 1.96817160e-01 2.67961752e-02
-7.30226874e-01 7.64999330e-01 1.25513947e+00 2.99174264e-02
1.04455397e-01 -8.63939673e-02 -1.06628251e+00 2.75442570e-01
1.09407985e+00 -3.51679951e-01 -6.78422451e-01 -7.25232780e-01
6.16439283e-01 -9.44106840e-03 4.51532364e-01 -7.14477003e-01
-2.04082519e-01 9.26037878e-02 -1.15076117e-01 -3.58028829e-01
9.82949361e-02 -2.43094899e-02 -1.49053618e-01 9.22683775e-02
-3.14644933e-01 4.03202385e-01 2.90011942e-01 1.66194811e-01
-4.48436946e-01 -3.82005244e-01 4.76558536e-01 -2.55145401e-01
-6.25510991e-01 1.75750017e-01 -1.27620369e-01 1.12304842e+00
4.53924537e-01 1.55365676e-01 -7.05886245e-01 -8.07524145e-01
-3.66762906e-01 4.19621736e-01 3.21145654e-01 7.86066413e-01
2.45557964e-01 -1.32901490e+00 -1.33542228e+00 -2.70497403e-03
5.83532512e-01 -3.08921605e-01 3.23904246e-01 5.49117208e-01
-4.39128935e-01 7.29466438e-01 4.53094095e-02 -2.53921062e-01
-8.05411875e-01 3.35684180e-01 3.49539995e-01 -6.79110467e-01
-2.36099269e-02 8.06443453e-01 1.62320703e-01 -1.01621163e+00
-3.62887859e-01 -3.22064877e-01 -2.91857868e-01 1.13651887e-01
3.61335516e-01 3.21324289e-01 3.89640957e-01 -4.18533832e-01
-3.14582407e-01 1.30267620e-01 2.04749182e-02 -4.79051203e-01
9.70101953e-01 -1.04489446e-01 -1.25856802e-01 3.90050203e-01
8.90907586e-01 2.37520516e-01 -7.21062303e-01 -2.91738659e-01
1.67816043e-01 -2.79308241e-02 -4.35081661e-01 -1.23303473e+00
-6.23220742e-01 7.47538626e-01 8.77245665e-02 3.74423623e-01
1.09280121e+00 2.28075415e-01 8.17152858e-01 3.15543085e-01
3.30024451e-01 -9.49125469e-01 1.91477939e-01 9.19608891e-01
1.21200562e+00 -1.11548495e+00 -3.01683575e-01 -4.52022962e-02
-6.17263019e-01 6.38462365e-01 6.08440995e-01 5.04227914e-02
3.06064606e-01 -1.40593410e-01 5.54886818e-01 -1.67965516e-01
-1.07080126e+00 -3.71975750e-01 4.43985045e-01 5.41072667e-01
8.69759858e-01 5.74565530e-02 -4.83082384e-01 7.40728438e-01
-9.21042204e-01 6.77493215e-02 4.13742095e-01 5.81512272e-01
-5.74196782e-03 -1.60537362e+00 -2.50671268e-01 3.79800946e-01
-6.63085878e-01 -5.29018819e-01 -5.05276382e-01 9.06844676e-01
-3.71839516e-02 1.31225109e+00 -3.75015616e-01 -8.54987949e-02
6.10245705e-01 4.03181911e-01 6.06541872e-01 -9.10681069e-01
-7.60097146e-01 -3.12831134e-01 4.82749641e-01 -4.47083384e-01
-2.59140253e-01 -5.67377508e-01 -6.12656832e-01 -1.52362183e-01
-1.43095508e-01 4.61184651e-01 2.74593353e-01 6.65188432e-01
4.72762018e-01 3.84666294e-01 8.79821926e-02 -9.46254097e-03
-8.14155281e-01 -1.37789357e+00 -3.55211884e-01 6.35347366e-01
2.81967153e-03 -2.67160058e-01 -2.55286068e-01 -1.08179376e-02] | [11.349650382995605, 8.328505516052246] |
d2a46971-4fd8-44f2-bdb9-c10d60aa98cc | artificial-intelligence-to-advance-earth | 2305.08413 | null | https://arxiv.org/abs/2305.08413v1 | https://arxiv.org/pdf/2305.08413v1.pdf | Artificial intelligence to advance Earth observation: a perspective | Earth observation (EO) is a prime instrument for monitoring land and ocean processes, studying the dynamics at work, and taking the pulse of our planet. This article gives a bird's eye view of the essential scientific tools and approaches informing and supporting the transition from raw EO data to usable EO-based information. The promises, as well as the current challenges of these developments, are highlighted under dedicated sections. Specifically, we cover the impact of (i) Computer vision; (ii) Machine learning; (iii) Advanced processing and computing; (iv) Knowledge-based AI; (v) Explainable AI and causal inference; (vi) Physics-aware models; (vii) User-centric approaches; and (viii) the much-needed discussion of ethical and societal issues related to the massive use of ML technologies in EO. | ['Rochelle Schneider', 'Bertrand Le Saux', 'Volker Markl', 'Jorge-Arnulfo Quiané-Ruiz', 'Mihai Datcu', 'Fabio Del Frate', 'Holger H. Hoos', 'Jan N. van Rijn', 'Sašo Džeroski', 'Mrinalini Kochupillai', 'Xiao Xiang Zhu', 'Gustau Camps-Valls', 'Begüm Demir', 'Konrad Schindler', 'Devis Tuia'] | 2023-05-15 | null | null | null | null | ['causal-inference', 'causal-inference'] | ['knowledge-base', 'miscellaneous'] | [ 2.02089414e-01 9.95334908e-02 -2.23157510e-01 4.10540439e-02
2.02159379e-02 -3.16476911e-01 8.73126209e-01 4.44624782e-01
-1.93319932e-01 5.76336563e-01 3.53447884e-01 -9.68038619e-01
-4.08823192e-01 -8.69845748e-01 -5.85110247e-01 -7.89058566e-01
-5.29805303e-01 4.56077367e-01 -3.09786975e-01 -2.31434673e-01
4.65001702e-01 7.28201449e-01 -1.85344529e+00 -3.40180188e-01
1.18237782e+00 8.03649902e-01 3.63607943e-01 1.20093060e+00
-9.92521346e-02 7.03967392e-01 -3.12816910e-02 -1.16434239e-01
-8.52399170e-02 -4.92206156e-01 -5.79779327e-01 -2.90100813e-01
2.51765341e-01 -1.39553979e-01 -1.96439505e-01 8.35958719e-01
1.36077940e-01 1.71674043e-01 5.21582007e-01 -1.00322855e+00
-8.12353671e-01 1.02310792e-01 -1.58560544e-01 5.73559165e-01
2.60953605e-01 6.98459864e-01 7.22925484e-01 -5.78048646e-01
7.27963388e-01 1.30552125e+00 4.23201263e-01 2.66023248e-01
-8.08584750e-01 -3.74163032e-01 9.87990275e-02 4.36191976e-01
-1.02036631e+00 -6.34405077e-01 3.74157816e-01 -7.31505454e-01
1.05776107e+00 5.41098893e-01 1.15173554e+00 5.98805547e-01
7.35639513e-01 2.91640073e-01 1.32247674e+00 -7.55229652e-01
3.84151578e-01 1.97158992e-01 4.33809869e-02 6.70316756e-01
4.03516710e-01 6.05424404e-01 -9.00643170e-01 -3.59882742e-01
4.82923090e-01 -1.67027920e-01 -1.20332837e-01 2.48191208e-01
-1.22248852e+00 5.85802495e-01 2.60791898e-01 1.49128646e-01
-7.23993480e-01 2.70700186e-01 -1.29753575e-01 1.40819788e-01
5.66756725e-01 7.40035057e-01 -6.73045397e-01 -2.04878315e-01
-8.72111201e-01 4.25130785e-01 8.85678291e-01 5.15302002e-01
6.40438139e-01 2.73622602e-01 3.85242224e-01 1.21214360e-01
9.66193140e-01 1.12070644e+00 -3.71232688e-01 -1.02809238e+00
7.21035078e-02 2.39324465e-01 3.57487440e-01 -1.05012882e+00
-5.88423967e-01 -1.99422538e-01 -5.58198512e-01 8.58728657e-04
1.90379813e-01 -4.70852196e-01 -7.72263110e-01 1.18773234e+00
4.91612136e-01 2.62930214e-01 -6.24449141e-02 7.74735212e-01
1.02994812e+00 8.00763071e-01 4.21679229e-01 -4.93260264e-01
1.43818259e+00 -4.23459500e-01 -9.49506104e-01 -5.13346732e-01
2.07898587e-01 -4.19693828e-01 7.38433778e-01 1.74873948e-01
-8.67687762e-01 -1.87135354e-01 -7.82181203e-01 -8.69000703e-03
-5.12062430e-01 -1.92164958e-01 1.13024771e+00 6.15101695e-01
-9.00697410e-01 6.89385056e-01 -1.07093728e+00 -6.98549986e-01
1.95336774e-01 4.53841910e-02 8.55589807e-02 4.01244983e-02
-1.23904169e+00 1.16746449e+00 -1.06903456e-01 4.21973169e-01
-7.38519073e-01 -1.20807886e+00 -7.93366075e-01 -6.23186026e-03
2.32107416e-01 -8.64709318e-01 6.84579611e-01 -2.73252487e-01
-1.42720509e+00 9.23870325e-01 -4.55924690e-01 -2.95789003e-01
3.07892948e-01 -5.25254011e-01 -7.69123793e-01 2.00402066e-01
-1.51402250e-01 2.91716039e-01 3.64487916e-01 -1.20382535e+00
-6.44489825e-01 -5.23506999e-01 -4.69412148e-01 2.12684012e-04
5.56530766e-02 3.71890277e-01 1.59665406e-01 -1.54207936e-02
1.31192163e-01 -1.07128036e+00 -4.08787072e-01 1.62627935e-01
-2.05249980e-01 -3.56750667e-01 6.52639091e-01 -1.03803146e+00
1.23154557e+00 -1.86745024e+00 9.92331281e-02 3.94001603e-02
3.94178420e-01 2.47833252e-01 5.12809038e-01 1.00333357e+00
2.40068674e-01 2.48121858e-01 -2.82027125e-01 7.26533234e-02
-3.17242630e-02 7.10191801e-02 -4.55335200e-01 6.26237333e-01
9.29469615e-02 9.18362558e-01 -1.24544072e+00 -2.68994987e-01
7.71256924e-01 2.40273461e-01 -1.13772914e-01 8.03093836e-02
-3.53390574e-01 9.66757536e-01 -4.61350918e-01 9.51337516e-01
3.55329424e-01 -7.30490461e-02 2.91052580e-01 1.61726341e-01
-9.09816444e-01 3.67830545e-01 -8.62859726e-01 1.02293992e+00
-3.52175534e-01 8.90722096e-01 2.41209954e-01 -5.28219998e-01
8.27288151e-01 2.74192363e-01 1.71955526e-01 -5.36102057e-01
-1.52771577e-01 1.83224455e-01 1.78304121e-01 -9.37900484e-01
3.46458137e-01 -5.28843224e-01 2.80696750e-01 3.48399401e-01
-3.20274353e-01 -2.96294272e-01 -1.74980402e-01 -6.58577681e-02
7.22217023e-01 4.35188919e-01 4.78861362e-01 -8.02616954e-01
1.02421343e-01 2.82629162e-01 3.88034940e-01 7.69811094e-01
-4.15782452e-01 -1.61458686e-01 7.55502880e-02 -8.98525298e-01
-9.10325229e-01 -9.27086890e-01 -5.23747861e-01 9.61018503e-01
1.75144538e-01 -2.50595659e-01 -2.98245102e-01 1.62320212e-01
2.77189702e-01 1.26460493e+00 -6.21360600e-01 1.22621991e-01
-2.89787263e-01 -1.19277942e+00 4.29346263e-01 3.65121126e-01
3.73012036e-01 -9.44894373e-01 -9.05791700e-01 9.40310806e-02
-2.56608248e-01 -9.47309613e-01 5.28404295e-01 5.18429913e-02
-1.12627029e+00 -9.25441265e-01 -9.57634002e-02 4.86980788e-02
4.08540338e-01 1.64253220e-01 1.03536808e+00 3.20959203e-02
-4.46301818e-01 6.00044012e-01 -9.81713738e-03 -1.10500109e+00
-5.36895931e-01 -5.30475557e-01 3.99437964e-01 -4.10085231e-01
4.80848700e-01 -7.38267601e-01 -6.72840416e-01 -1.13761671e-01
-4.83634472e-01 5.63205704e-02 5.49928248e-01 2.40985602e-01
2.83119172e-01 -1.78669900e-01 2.82828361e-01 -5.33857286e-01
1.74869552e-01 -7.51137018e-01 -6.75556839e-01 3.59274983e-01
-1.10280108e+00 -1.15145288e-01 1.72030509e-01 2.89191008e-01
-1.34497797e+00 -2.87230819e-01 6.45592585e-02 4.77545336e-02
-4.27433759e-01 7.24722505e-01 5.41728884e-02 -2.58520972e-02
8.93194556e-01 6.20873645e-02 5.08566350e-02 -6.64502919e-01
5.02912104e-01 7.46499062e-01 5.10232508e-01 -5.30710101e-01
7.96997666e-01 9.36718047e-01 4.08248186e-01 -1.61254013e+00
-6.88911676e-01 -2.84258395e-01 -7.08442569e-01 -7.09887683e-01
1.18724501e+00 -7.98478305e-01 -7.77339220e-01 4.26862627e-01
-8.86217058e-01 -6.01875365e-01 -1.90256104e-01 7.17887580e-01
-1.26276359e-01 2.24595770e-01 -4.09932852e-01 -1.29505610e+00
-4.66357499e-01 -5.99784434e-01 1.06655478e+00 6.16632938e-01
-2.05500841e-01 -1.30377638e+00 4.30421889e-01 5.42797506e-01
4.68639940e-01 6.02877140e-01 8.91339242e-01 -2.44206246e-02
-5.66757619e-01 -1.31327868e-01 -1.79759562e-01 -1.14880972e-01
-1.59136765e-02 6.74029827e-01 -1.07601523e+00 2.03181803e-01
-1.28490359e-01 -4.25206125e-02 7.76907921e-01 4.88726377e-01
6.33738756e-01 -2.80705154e-01 -5.50626338e-01 6.14739776e-01
1.26682413e+00 3.54654305e-02 7.56543517e-01 4.49166268e-01
6.88725889e-01 1.01012456e+00 6.05835378e-01 6.02423191e-01
4.84695554e-01 2.36978367e-01 6.18843138e-01 8.24974254e-02
7.76798406e-04 -2.75183380e-01 2.23030940e-01 9.22357678e-01
-7.67435610e-01 -1.13032289e-01 -1.21291661e+00 5.76093972e-01
-1.73860896e+00 -1.06791925e+00 -9.62164104e-01 2.15529943e+00
4.12400275e-01 -3.31161231e-01 -2.85400122e-01 -4.78058666e-01
4.09216404e-01 2.43883967e-01 -4.84023839e-01 -5.81245661e-01
5.19734360e-02 5.60613647e-02 7.68964112e-01 4.44804043e-01
-1.07867932e+00 8.61590385e-01 7.78765535e+00 2.58769214e-01
-1.12993169e+00 4.09801602e-02 9.61765498e-02 3.02898377e-01
-6.67528987e-01 3.87300581e-01 -7.98986793e-01 2.74454057e-01
1.27392209e+00 -2.11283952e-01 6.24552131e-01 5.74519277e-01
6.09737396e-01 -4.99683946e-01 -4.09756899e-01 3.17894489e-01
-3.24280441e-01 -1.48037577e+00 -2.75109261e-01 4.63976771e-01
8.86211812e-01 6.50250375e-01 -4.15887892e-01 -3.21587712e-01
4.93817031e-01 -1.08007693e+00 7.75544286e-01 8.96868169e-01
7.25008786e-01 -2.34460056e-01 4.51003939e-01 2.41774440e-01
-9.71480787e-01 -1.53383389e-02 -6.55934691e-01 -7.62478828e-01
1.14281267e-01 1.00520086e+00 -2.74708718e-01 8.65279377e-01
1.13479042e+00 7.06479967e-01 -1.34636477e-01 8.97174120e-01
-6.60606921e-01 1.09145713e+00 -5.14899611e-01 -1.58860441e-02
1.05686441e-01 -4.09753889e-01 9.49890077e-01 1.14133227e+00
1.46408185e-01 2.92057335e-01 -4.05537307e-01 1.05503869e+00
5.62210143e-01 -3.64192724e-01 -8.13320041e-01 -5.55803597e-01
5.28807878e-01 1.14272141e+00 -7.21950889e-01 -1.30212978e-01
-4.20302600e-01 3.38265568e-01 -2.77178943e-01 8.31101164e-02
-4.90167975e-01 2.87062931e-03 8.83826017e-01 -3.55173759e-02
-5.03363460e-02 -5.19944370e-01 -6.81384325e-01 -9.71141458e-01
-4.68734831e-01 -5.12562156e-01 3.06892931e-01 -8.22819471e-01
-9.32053745e-01 -7.60400668e-02 1.71360686e-01 -6.40404403e-01
4.68601696e-02 -5.01093209e-01 -8.41962337e-01 7.58621633e-01
-1.61916101e+00 -1.01541317e+00 -4.40336972e-01 -2.39479214e-01
6.79467013e-03 2.91872650e-01 1.02224755e+00 1.33258939e-01
-4.49096233e-01 -1.93260998e-01 6.41766548e-01 -5.42757809e-01
3.16017628e-01 -1.13238668e+00 7.63502717e-01 7.64537096e-01
-1.56382069e-01 6.92015827e-01 1.06252885e+00 -1.08683860e+00
-2.14855289e+00 -8.53647232e-01 1.19166410e+00 -8.64698350e-01
9.05892015e-01 -4.79625240e-02 -8.27434897e-01 8.71555686e-01
-5.56054637e-02 -2.07322866e-01 7.26698577e-01 3.98896992e-01
1.85894251e-01 2.24148169e-01 -9.52930450e-01 6.38425946e-01
9.73697603e-01 -3.95504415e-01 -7.46277273e-01 5.17787755e-01
4.88047153e-01 -2.24272639e-01 -1.04196119e+00 1.49764374e-01
9.90181088e-01 -7.15498030e-01 8.05520296e-01 -7.40037143e-01
7.00990081e-01 -3.12767655e-01 1.07804395e-01 -7.94126272e-01
-5.15750349e-01 -7.33949125e-01 -5.74005879e-02 8.69796991e-01
1.81346878e-01 -9.42268133e-01 1.83260143e-01 1.01810062e+00
-3.85655642e-01 -5.14606059e-01 -7.88645983e-01 -4.59911495e-01
-5.90483248e-02 -5.71000099e-01 3.11962783e-01 1.07842612e+00
-8.83943215e-02 -6.48705810e-02 -4.47178662e-01 5.79221070e-01
9.13552284e-01 1.70596659e-01 7.44571149e-01 -1.63649750e+00
-9.51435268e-02 -7.38032833e-02 -3.23816210e-01 -4.12356764e-01
-5.00058115e-01 -5.33118963e-01 3.84080745e-02 -1.83743286e+00
2.46584013e-01 -2.63545126e-01 1.20217510e-01 1.57730207e-01
-3.16309959e-01 -1.23736598e-01 1.85536191e-01 5.84805429e-01
-1.52153417e-01 5.09493470e-01 1.22145593e+00 4.16667104e-01
-2.66446173e-01 -2.83345282e-01 -8.53593946e-01 9.66498315e-01
6.20314896e-01 -5.78279495e-01 8.98545831e-02 -2.22909316e-01
9.18584883e-01 3.22100259e-02 8.45148563e-01 -6.48838222e-01
2.70776898e-01 -8.69685113e-01 2.17932478e-01 -4.50186849e-01
-1.10924974e-01 -5.21930397e-01 4.01992440e-01 8.96528661e-01
1.03240788e-01 -4.28603500e-01 5.12857020e-01 7.04733849e-01
2.59639770e-01 -6.93384707e-02 5.83126307e-01 -2.11681083e-01
-8.86286080e-01 -4.74330783e-02 -6.38702452e-01 -1.12962797e-01
7.01583624e-01 -6.47487119e-02 -6.76940560e-01 -3.45724761e-01
-5.73688805e-01 7.19483554e-01 5.33517957e-01 1.67110637e-01
2.26408392e-01 -6.10951781e-01 -8.45656574e-01 -2.54650470e-02
-4.07207124e-02 -1.61613405e-01 5.17062306e-01 1.04635155e+00
-8.01361799e-01 6.69966161e-01 5.20512424e-02 -3.77922356e-01
-1.08544087e+00 4.20740277e-01 2.78853506e-01 8.97071958e-02
-1.01320636e+00 7.40198076e-01 2.20079660e-01 -5.15357554e-01
-2.57715583e-01 -3.41458917e-02 -9.99399945e-02 -1.52239129e-01
6.47156000e-01 1.09337378e+00 -2.71429121e-01 -2.01880455e-01
-5.53967774e-01 5.80938756e-01 6.61501586e-01 -1.38047025e-01
1.59829569e+00 -4.52361166e-01 -5.29715657e-01 9.56519663e-01
3.31035733e-01 -7.31154382e-02 -1.25657213e+00 7.67174065e-02
-3.73244546e-02 -2.18898714e-01 5.21395385e-01 -9.90198493e-01
-3.77752781e-01 1.07110393e+00 4.06779140e-01 2.62031436e-01
7.83453166e-01 1.48661211e-01 6.48262799e-01 4.15042818e-01
2.43293628e-01 -1.08196759e+00 -7.10933387e-01 4.64005023e-02
9.84646618e-01 -1.05155230e+00 7.48631537e-01 -3.63232493e-01
-3.67899626e-01 1.15033412e+00 8.37002099e-02 2.86266536e-01
8.74914944e-01 4.54524308e-02 -1.18860858e-03 -8.09102356e-01
-8.57253492e-01 -4.63408083e-01 2.60132760e-01 5.42263269e-01
2.38637641e-01 3.77152026e-01 -3.12450945e-01 2.00126603e-01
-7.59774968e-02 2.22125858e-01 3.38881344e-01 8.92980933e-01
-8.34767282e-01 -6.07416153e-01 -8.78109872e-01 6.16110206e-01
-2.79190302e-01 -2.48796135e-01 -4.81111318e-01 6.74740970e-01
9.44570601e-02 1.16355765e+00 -1.43059930e-02 -1.11286990e-01
-7.66851604e-02 2.24448174e-01 2.89262980e-01 -3.74829739e-01
-3.52016419e-01 -2.40612701e-01 3.23466599e-01 -5.13495266e-01
-5.55738866e-01 -1.10692894e+00 -1.20042908e+00 -7.44729757e-01
-2.57040709e-01 2.85696816e-02 1.16457784e+00 1.27601516e+00
6.48165524e-01 5.49109519e-01 3.03248912e-01 -1.00073862e+00
-1.12682618e-01 -9.15532768e-01 -6.68053091e-01 -4.50089872e-01
1.57068357e-01 -5.52290022e-01 -8.17035258e-01 9.21613127e-02] | [6.551549911499023, 3.287259578704834] |
d3e71da1-b31e-4084-9daf-cbbdc906d243 | mixing-signals-data-augmentation-approach-for | 2204.03737 | null | https://arxiv.org/abs/2204.03737v1 | https://arxiv.org/pdf/2204.03737v1.pdf | Mixing Signals: Data Augmentation Approach for Deep Learning Based Modulation Recognition | With the rapid development of deep learning, automatic modulation recognition (AMR), as an important task in cognitive radio, has gradually transformed from traditional feature extraction and classification to automatic classification by deep learning technology. However, deep learning models are data-driven methods, which often require a large amount of data as the training support. Data augmentation, as the strategy of expanding dataset, can improve the generalization of the deep learning models and thus improve the accuracy of the models to a certain extent. In this paper, for AMR of radio signals, we propose a data augmentation strategy based on mixing signals and consider four specific methods (Random Mixing, Maximum-Similarity-Mixing, $\theta-$Similarity Mixing and n-times Random Mixing) to achieve data augmentation. Experiments show that our proposed method can improve the classification accuracy of deep learning based AMR models in the full public dataset RML2016.10a. In particular, for the case of a single signal-to-noise ratio signal set, the classification accuracy can be significantly improved, which verifies the effectiveness of the methods. | ['Xiaoniu Yang', 'Qi Xuan', 'Shilian Zheng', 'Shanqing Yu', 'Huaji Zhou', 'Dongwei Xu', 'Zhuangzhi Chen', 'Xinjie Xu'] | 2022-04-05 | null | null | null | null | ['automatic-modulation-recognition'] | ['time-series'] | [ 2.36389145e-01 -3.27601999e-01 -1.45453110e-01 -3.49315614e-01
-7.23919749e-01 -1.58688538e-02 7.02771068e-01 -3.96899909e-01
-3.27416748e-01 7.79960275e-01 2.08200395e-01 -7.59211063e-01
-2.91507155e-01 -8.83235097e-01 -2.75979757e-01 -8.20080280e-01
2.14295294e-02 6.74232692e-02 -3.73643875e-01 -3.94086748e-01
-1.54990658e-01 5.11460662e-01 -1.15831888e+00 3.35845381e-01
7.45096862e-01 1.46278167e+00 -7.80485757e-03 3.96786690e-01
-2.06934333e-01 1.13701415e+00 -1.06598318e+00 -7.92921260e-02
2.44842738e-01 -5.96638024e-01 -4.59246904e-01 2.20605768e-02
-3.91020238e-01 3.07041332e-02 -9.20278430e-01 8.20774853e-01
9.25371289e-01 -1.97784156e-01 5.10314465e-01 -1.06239021e+00
-4.09325391e-01 8.53180885e-01 -7.24813104e-01 4.11560267e-01
-2.48445064e-01 -2.30483100e-01 7.08637178e-01 -7.28382349e-01
-2.76668787e-01 9.26942706e-01 7.17105925e-01 2.99799591e-01
-8.87112141e-01 -1.30204892e+00 -1.60869993e-02 4.15495664e-01
-1.57880104e+00 -5.75201035e-01 1.03249896e+00 -1.40543833e-01
4.05805200e-01 2.94520587e-01 5.89534402e-01 1.01399684e+00
1.47611252e-03 7.37485588e-01 1.15936351e+00 -4.39861149e-01
2.53219008e-01 -1.79342732e-01 -4.49360497e-02 2.99731672e-01
1.51057253e-02 1.03867047e-01 -3.48424196e-01 -1.70390680e-01
6.78402007e-01 1.73126951e-01 -1.87117979e-01 1.87369406e-01
-1.14467978e+00 7.87411690e-01 4.10280734e-01 6.41485095e-01
-3.17519128e-01 2.31040761e-01 8.66754055e-02 5.73397696e-01
6.55131102e-01 4.40386325e-01 -4.29031193e-01 -7.44261742e-02
-9.02479172e-01 -4.99324920e-03 3.27980280e-01 5.93553662e-01
4.67279017e-01 5.82466662e-01 -3.39109719e-01 1.15201271e+00
2.33202443e-01 6.40483439e-01 8.16032231e-01 -5.67659616e-01
3.42477113e-01 4.09118295e-01 -1.55073926e-01 -9.00380731e-01
-7.94308066e-01 -1.55533493e+00 -1.56954646e+00 -1.22873701e-01
1.31730944e-01 -3.84521425e-01 -1.11272645e+00 1.85225272e+00
-2.34470040e-01 2.68971801e-01 3.40187758e-01 5.93487024e-01
8.50212157e-01 6.37276649e-01 -3.21535707e-01 -4.37041193e-01
1.24647665e+00 -6.88827157e-01 -9.50359344e-01 -3.43972892e-01
8.01204860e-01 -8.06438148e-01 6.52509391e-01 5.24908900e-01
-5.64168572e-01 -7.43498206e-01 -1.40254307e+00 4.18799520e-01
-9.41997692e-02 3.27996254e-01 8.61919224e-01 9.75050688e-01
-6.34227455e-01 -2.17237566e-02 -2.64769018e-01 3.01181465e-01
8.40225160e-01 4.27691728e-01 -9.27571356e-02 -3.13660622e-01
-1.54035592e+00 5.59884667e-01 1.84252545e-01 1.62768051e-01
-9.76999998e-01 -4.83856231e-01 -5.03907263e-01 1.42897725e-01
2.54862309e-01 -3.81221801e-01 1.21613491e+00 -9.74882126e-01
-1.38757968e+00 3.96579832e-01 9.68365967e-02 -8.54214072e-01
3.15616429e-01 -6.65929839e-02 -1.09884262e+00 -2.45721593e-01
-2.01258257e-01 2.64063507e-01 9.56162572e-01 -1.01284122e+00
-7.24293530e-01 -2.38019437e-01 7.36792758e-02 -5.20802140e-02
-4.45544511e-01 -3.04408908e-01 -6.97861910e-02 -1.09242833e+00
3.82140428e-01 -7.98716366e-01 -4.96596336e-01 -7.36500144e-01
-1.85891449e-01 1.26480952e-01 6.87645197e-01 -7.33734250e-01
1.29629922e+00 -2.45598531e+00 -1.89741641e-01 3.94754559e-01
2.38665119e-01 3.10634792e-01 -1.67266667e-01 -1.76728025e-01
-2.64241129e-01 -6.20824881e-02 -3.57156634e-01 -8.87031406e-02
-2.48860061e-01 -1.87054854e-02 -2.08252817e-01 2.97181904e-01
1.10888526e-01 8.39520812e-01 -4.74208146e-01 -1.16512505e-02
1.55692780e-03 3.96365970e-01 -4.65138167e-01 -4.86807898e-02
-5.71145229e-02 7.02013850e-01 -2.47307763e-01 5.21158695e-01
6.13698959e-01 -3.15090030e-01 6.22587018e-02 -2.43834555e-01
1.28922626e-01 2.40896106e-01 -9.38176811e-01 1.49426568e+00
-9.56302941e-01 8.44241083e-01 -2.82277495e-01 -1.46809006e+00
1.22137046e+00 3.72171998e-01 6.50756717e-01 -1.04618227e+00
3.86099517e-01 3.83138657e-01 7.41782963e-01 -6.40064776e-02
6.31793737e-02 -1.94662869e-01 -1.07021816e-01 4.46836144e-01
-1.87421128e-01 2.82005370e-01 -1.88472003e-01 -1.16081022e-01
1.02917635e+00 -7.46639431e-01 2.08888188e-01 1.73453353e-02
6.47161484e-01 -4.89318818e-01 5.76628029e-01 9.03601825e-01
2.87062645e-01 2.63692498e-01 1.23612545e-01 -2.95575559e-01
-7.96176672e-01 -5.30513525e-01 -4.51837897e-01 7.88278401e-01
-1.67886958e-01 -6.93891197e-02 -4.52338129e-01 -4.21430260e-01
-2.40561470e-01 6.89026773e-01 -3.94221395e-01 -7.57047057e-01
-4.72589910e-01 -1.62286770e+00 7.26306260e-01 3.96487445e-01
1.04180968e+00 -8.46124232e-01 1.50263142e-02 3.10683817e-01
-4.60237265e-01 -1.36243904e+00 1.15886629e-01 3.93270910e-01
-5.97952306e-01 -5.94256043e-01 -9.23462808e-01 -7.12734461e-01
4.07671809e-01 3.41892391e-01 7.41386652e-01 7.94421881e-02
-1.15419380e-01 -1.77617073e-01 -5.28116226e-01 -4.84954685e-01
-5.70612967e-01 1.70406505e-01 2.57731140e-01 4.02532607e-01
3.17466766e-01 -7.91481614e-01 -4.81813878e-01 4.72332120e-01
-7.04678059e-01 1.23624496e-01 1.16084695e+00 7.39709795e-01
3.00571263e-01 4.86652553e-01 1.14628708e+00 -5.20424187e-01
6.51862204e-01 -5.30604661e-01 -2.56191373e-01 1.10445749e-02
-5.51895678e-01 -3.15550417e-02 6.55023575e-01 -6.00506902e-01
-6.39389932e-01 -3.52252424e-01 -5.18955469e-01 -4.57456484e-02
2.00887918e-01 8.45874548e-01 -5.43836832e-01 -1.61372006e-01
6.55521452e-01 5.61159372e-01 -3.54395248e-02 -5.91560960e-01
3.20856303e-01 1.21110725e+00 3.78047615e-01 -8.72829705e-02
9.08785880e-01 4.27360564e-01 1.52263239e-01 -8.62271130e-01
-9.87863123e-01 -5.59794754e-02 -8.24928880e-02 2.63012182e-02
3.74452680e-01 -1.20680082e+00 -3.59925956e-01 6.80943012e-01
-6.08044088e-01 -2.34788030e-01 -1.89606786e-01 8.82509708e-01
-2.38579929e-01 1.50054902e-01 -3.40766102e-01 -7.79700518e-01
-3.10285568e-01 -9.74254012e-01 3.95453662e-01 -1.36642978e-02
1.09976292e-01 -5.73816776e-01 -3.29624772e-01 4.57679987e-01
8.00380468e-01 -1.11913055e-01 1.04348171e+00 -9.07392442e-01
-3.11042607e-01 -4.94933099e-01 -1.45131260e-01 5.22965372e-01
5.67392349e-01 -7.10247517e-01 -1.17987454e+00 -2.87712336e-01
3.58301669e-01 1.06977001e-01 8.60401452e-01 6.16958380e-01
1.76349854e+00 -1.51847854e-01 -1.41946375e-01 5.95525205e-01
7.39424467e-01 6.20915353e-01 8.27510774e-01 3.47190320e-01
3.94178063e-01 -1.59641430e-01 4.63913858e-01 6.16744757e-01
2.66076252e-02 7.32022345e-01 4.61761415e-01 -4.29820895e-01
-9.54869241e-02 2.75091261e-01 1.73088044e-01 8.71241152e-01
-6.06946200e-02 -6.64119944e-02 -7.49744236e-01 1.05376311e-01
-1.49596083e+00 -8.84908438e-01 8.27150717e-02 2.12347722e+00
7.97118485e-01 3.52717578e-01 1.10765085e-01 1.01472366e+00
4.71696973e-01 1.91798553e-01 -4.70793515e-01 1.97845042e-01
-4.32576001e-01 3.52616668e-01 3.48588228e-01 3.64740312e-01
-1.05188227e+00 5.41200221e-01 5.73472214e+00 1.16211891e+00
-1.33875120e+00 2.85346866e-01 9.87719059e-01 -3.77106704e-02
-1.29110917e-01 -4.60363418e-01 -5.22113860e-01 6.25757754e-01
9.91325915e-01 -5.31760119e-02 5.10496080e-01 6.58552885e-01
2.60640204e-01 3.82356793e-01 -8.90353680e-01 1.44333982e+00
-7.62235075e-02 -1.51621497e+00 1.12844035e-01 3.42538625e-01
6.02603793e-01 -8.48924965e-02 4.76515651e-01 7.89509594e-01
-2.22472847e-02 -1.29028094e+00 3.99448335e-01 3.13926667e-01
7.97133207e-01 -1.13708854e+00 1.08978283e+00 3.47938418e-01
-9.69101489e-01 -6.37706459e-01 -2.74523884e-01 -7.57568106e-02
-2.72304770e-02 1.08446002e+00 -3.76744479e-01 7.57552624e-01
4.81098205e-01 4.76947188e-01 -6.39193416e-01 1.02166748e+00
2.66585015e-02 7.83243954e-01 -1.06813781e-01 5.28901070e-02
-4.91750659e-03 2.91955415e-02 3.00580740e-01 8.83465767e-01
5.55280626e-01 -3.06657627e-02 4.86668348e-02 2.09180966e-01
-4.35290545e-01 5.29301874e-02 -2.31321782e-01 -1.59647897e-01
5.90143740e-01 1.16351712e+00 -3.96469951e-01 -2.43822485e-01
-3.27359617e-01 4.27121580e-01 -2.59358108e-01 2.59824723e-01
-1.00399840e+00 -4.51237440e-01 4.55595523e-01 1.59583762e-01
6.88965395e-02 -3.59971702e-01 -3.55995536e-01 -7.93542325e-01
-1.01614423e-01 -1.38060439e+00 1.31148085e-01 -4.17049140e-01
-9.36013699e-01 6.24386311e-01 -4.05460119e-01 -1.52665591e+00
-5.61179593e-02 -4.19290632e-01 -3.19390774e-01 6.67512119e-01
-1.35605943e+00 -8.31908762e-01 -4.73809063e-01 6.44040585e-01
3.39402735e-01 -1.00502682e+00 7.71396697e-01 1.00542903e+00
-3.40131462e-01 8.98308098e-01 2.92078912e-01 4.13914412e-01
2.06142023e-01 -6.25974953e-01 2.81482846e-01 7.68815577e-01
4.50554699e-01 2.96681166e-01 4.73280877e-01 -4.82394733e-02
-9.08196688e-01 -1.18278456e+00 2.54624993e-01 1.28779918e-01
5.60367346e-01 -5.72620988e-01 -5.61023831e-01 3.76649410e-01
-2.03121066e-01 -5.10367751e-02 1.13998532e+00 1.53003812e-01
-4.74738367e-02 -5.24908006e-01 -8.80332649e-01 6.20066762e-01
8.66070449e-01 -4.57958192e-01 -1.55665770e-01 3.70002389e-01
7.08963454e-01 -1.20757170e-01 -7.65208364e-01 8.59041274e-01
2.85650879e-01 -4.99442577e-01 1.07100272e+00 -6.12277806e-01
5.24525940e-02 -3.43116164e-01 -6.31424010e-01 -1.34934855e+00
-4.16627675e-01 -4.55640227e-01 -4.83622789e-01 9.39592600e-01
5.39304852e-01 -6.79352582e-01 6.71715379e-01 -3.84211719e-01
-1.73729047e-01 -7.79092252e-01 -1.06611514e+00 -8.40294063e-01
-5.01802005e-02 -8.04917514e-01 8.81853521e-01 1.01937199e+00
-3.98879588e-01 5.64781129e-01 -6.79330885e-01 2.63805896e-01
3.73401344e-01 -3.74296308e-02 8.28485429e-01 -1.31855869e+00
-6.09119713e-01 -4.77356523e-01 -6.10436082e-01 -1.24774945e+00
7.30717368e-03 -1.05293870e+00 -4.24093574e-01 -1.47511816e+00
-1.78351607e-02 -7.98355162e-01 -7.66991436e-01 3.42162699e-01
2.06633694e-02 3.88658285e-01 -1.09285876e-01 2.16930777e-01
-2.78137565e-01 8.97011757e-01 1.04528141e+00 -3.85199308e-01
-1.84761241e-01 7.44520724e-01 -1.14999616e+00 6.72678709e-01
1.09305799e+00 -2.97798395e-01 -3.64762604e-01 -1.79129303e-01
2.39435881e-01 -5.29981032e-02 9.99340415e-02 -1.50074387e+00
-2.44496912e-01 2.43280277e-01 7.60178983e-01 -3.68033141e-01
4.82554942e-01 -7.96031654e-01 1.06226891e-01 8.10683548e-01
-3.77288193e-01 -3.59047294e-01 1.96791753e-01 5.42470217e-01
-1.69215888e-01 6.33430481e-02 9.40677047e-01 2.46520907e-01
-3.48829955e-01 4.58610564e-01 -6.97209001e-01 2.12121345e-02
6.77634656e-01 1.48001835e-01 -2.41728067e-01 -7.28827536e-01
-8.89334738e-01 -8.42904225e-02 -4.24733609e-01 5.94582260e-01
4.97317106e-01 -1.75666738e+00 -8.04952323e-01 2.10932136e-01
3.68316174e-02 -2.59544030e-02 1.60338938e-01 1.05221701e+00
-3.15843858e-02 2.97586024e-01 -4.35717590e-02 -5.54611862e-01
-1.03515041e+00 3.81381512e-01 7.13760614e-01 -2.54949987e-01
-4.69132572e-01 6.17419362e-01 8.64875987e-02 -2.35714287e-01
3.82525682e-01 -9.11457464e-02 -3.81071925e-01 -1.11451223e-01
6.64955258e-01 1.19801663e-01 3.32803816e-01 -4.57341820e-01
-2.71393180e-01 2.36154839e-01 -1.85224727e-01 -9.38529447e-02
1.20335603e+00 1.62216365e-01 7.01749697e-04 3.20336193e-01
1.13122702e+00 2.78945386e-01 -6.43276870e-01 -6.62755549e-01
-6.17739186e-02 -2.13838860e-01 3.34419996e-01 -7.91011155e-01
-1.37174094e+00 1.01276064e+00 1.10072112e+00 3.42972219e-01
1.27175725e+00 -5.85797839e-02 7.08857715e-01 6.12956583e-01
3.00745517e-01 -9.01626706e-01 3.24259847e-01 4.34560001e-01
6.78451478e-01 -1.04245257e+00 -7.11306855e-02 4.59949449e-02
-2.17683941e-01 7.40347266e-01 3.39839429e-01 4.43738490e-01
8.89892876e-01 2.30785564e-01 1.11727573e-01 1.01621129e-01
-2.76958108e-01 -2.67501563e-01 1.40910774e-01 4.26991284e-01
4.13341492e-01 3.20336297e-02 -1.76773444e-01 1.02247715e+00
-5.10318220e-01 -1.79745331e-01 3.21479350e-01 7.67714798e-01
-6.73653305e-01 -1.07590067e+00 -3.18471611e-01 8.95349205e-01
-7.01055825e-01 -1.91345632e-01 -1.87077016e-01 5.82378983e-01
6.62386939e-02 1.24070597e+00 2.64997389e-02 -6.98844850e-01
1.95197776e-01 -1.02837756e-01 3.99395645e-01 -3.07357907e-01
3.65125723e-02 2.49236617e-02 9.42769945e-02 1.37506858e-01
-3.36805612e-01 -1.22672930e-01 -1.12383533e+00 -1.79726765e-01
-4.14877325e-01 2.35626921e-01 7.12104559e-01 1.34780395e+00
3.22510153e-01 1.15791476e+00 1.18385375e+00 -4.51559961e-01
-6.05816841e-01 -1.34407091e+00 -5.42108119e-01 1.15587085e-01
2.42159665e-01 -6.66125417e-01 -2.85247028e-01 -4.69296873e-01] | [6.488638877868652, 1.4801249504089355] |
16c926d1-9500-4e72-80c7-c37e40df9cf5 | chatgpt-steered-editing-instructor-for | 2305.02483 | null | https://arxiv.org/abs/2305.02483v1 | https://arxiv.org/pdf/2305.02483v1.pdf | ChatGPT-steered Editing Instructor for Customization of Abstractive Summarization | Tailoring outputs of large language models, such as ChatGPT, to specific user needs remains a challenge despite their impressive generation quality. In this paper, we propose a tri-agent generation pipeline consisting of a generator, an instructor, and an editor to enhance the customization of generated outputs. The generator produces an initial output, the user-specific instructor generates editing instructions, and the editor generates a revised output aligned with user preferences. The inference-only large language model (ChatGPT) serves as both the generator and the editor, while a smaller model acts as the user-specific instructor to guide the generation process toward user needs. The instructor is trained using editor-steered reinforcement learning, leveraging feedback from the large-scale editor model to optimize instruction generation. Experimental results on two abstractive summarization datasets demonstrate the effectiveness of our approach in generating outputs that better fulfill user expectations. | ['Pengcheng He', 'Giuseppe Carenini', 'Yujia Xie', 'Wen Xiao'] | 2023-05-04 | null | null | null | null | ['abstractive-text-summarization'] | ['natural-language-processing'] | [ 4.56595480e-01 4.60131824e-01 -1.95158526e-01 -4.48309720e-01
-1.10719836e+00 -7.38664985e-01 4.59448576e-01 2.93500423e-01
-1.79295808e-01 8.04371297e-01 6.56977594e-01 -4.62596118e-01
4.24106479e-01 -8.05101335e-01 -7.05466151e-01 -1.34999484e-01
4.97237533e-01 8.65717947e-01 5.59220910e-02 -4.78784382e-01
4.75857645e-01 -8.94824639e-02 -1.40016234e+00 5.85771501e-01
1.59556305e+00 5.15391469e-01 6.18888974e-01 1.06771576e+00
-2.18311936e-01 9.63606298e-01 -9.14087594e-01 -1.38801023e-01
-1.25194624e-01 -8.67120802e-01 -5.18008649e-01 -8.61007497e-02
1.72127843e-01 -5.64110458e-01 4.53657284e-02 4.57512438e-01
8.69383931e-01 4.23818976e-01 5.29539049e-01 -1.11257005e+00
-8.98941517e-01 1.31290638e+00 -1.28069535e-01 -1.39837325e-01
6.61800802e-01 5.43471038e-01 1.07929778e+00 -3.65273535e-01
6.95774972e-01 1.17834198e+00 1.73205882e-01 9.69765007e-01
-1.12658191e+00 -5.68388164e-01 2.12334037e-01 -3.59512001e-01
-7.21117735e-01 -5.80476880e-01 5.34805655e-01 -3.06785733e-01
1.10912466e+00 2.02491090e-01 4.44560826e-01 1.04727697e+00
2.70018488e-01 9.01715279e-01 5.31449974e-01 -4.87938821e-01
2.76334822e-01 2.71051317e-01 -1.55535880e-02 6.27368212e-01
-9.48731676e-02 -2.23490089e-01 -5.88143289e-01 -2.65078664e-01
4.51406062e-01 -3.79474908e-01 -2.98129976e-01 -8.10483545e-02
-9.74764824e-01 8.45413685e-01 -5.83679751e-02 -1.43128976e-01
-5.63664317e-01 1.63937598e-01 4.13399369e-01 3.34038615e-01
2.57181317e-01 1.05608749e+00 -4.84109670e-01 -5.47406137e-01
-9.67893660e-01 6.57746136e-01 1.14019656e+00 1.20564175e+00
3.88195604e-01 2.86666334e-01 -8.96158278e-01 9.80102897e-01
2.32802466e-01 4.01182562e-01 8.71868849e-01 -9.39948499e-01
8.55259180e-01 7.25663662e-01 2.53910124e-01 -2.86918819e-01
-1.02034910e-02 -4.73679185e-01 -2.26069197e-01 5.21395355e-02
1.89332873e-01 -6.12663388e-01 -7.18245029e-01 1.93917966e+00
1.57908946e-01 -2.82914847e-01 3.73446941e-01 5.75278342e-01
8.13927233e-01 9.64829743e-01 1.76270172e-01 -1.62642524e-01
1.25335097e+00 -1.37962902e+00 -6.85581207e-01 -3.96831363e-01
7.24617958e-01 -8.40418279e-01 1.21835029e+00 2.77453125e-01
-1.55072427e+00 -6.23544455e-01 -9.04710233e-01 -2.09332734e-01
8.03627968e-02 4.45801586e-01 -9.22156591e-03 2.17589065e-01
-1.09060657e+00 5.44336379e-01 -6.11494899e-01 -1.40638009e-01
4.82467636e-02 2.26960286e-01 1.95915177e-01 1.46184936e-01
-1.00103915e+00 7.38119900e-01 3.05079490e-01 -3.27924460e-01
-1.03696680e+00 -6.73969507e-01 -1.06645238e+00 3.94730866e-01
2.46045843e-01 -1.08943558e+00 2.08603120e+00 -7.67683327e-01
-2.15084743e+00 2.83999920e-01 -8.42241496e-02 -2.92650074e-01
6.46709502e-01 -1.25119165e-01 1.43579930e-01 -1.63687795e-01
1.24084629e-01 7.20729768e-01 5.83314657e-01 -1.18980277e+00
-8.51111829e-01 7.26964744e-03 1.52680635e-01 6.31463706e-01
-2.30213016e-01 -2.40006045e-01 -4.73415732e-01 -6.23386383e-01
-6.04843378e-01 -8.57050538e-01 -2.86979049e-01 -9.08459723e-01
-5.73899686e-01 -4.99315798e-01 5.35806954e-01 -7.40566373e-01
1.65905118e+00 -1.79952037e+00 4.83337075e-01 4.38028276e-02
4.30614268e-03 9.28990617e-02 -6.28679216e-01 7.35905766e-01
6.18648231e-02 -4.07056175e-02 3.40316035e-02 -5.27570724e-01
1.99657679e-01 -1.47506684e-01 -3.38324308e-01 -4.65769887e-01
3.78811747e-01 9.09326017e-01 -1.37931466e+00 -4.10715967e-01
-1.73389822e-01 8.13106969e-02 -1.09573865e+00 1.04556692e+00
-8.42425525e-01 3.38268846e-01 -7.21873403e-01 2.40609333e-01
8.69912431e-02 -2.57817924e-01 1.91421092e-01 9.42475051e-02
-1.93628624e-01 6.39884174e-01 -9.80059028e-01 1.51164961e+00
-8.90080392e-01 2.59568572e-01 1.01234667e-01 -5.61550438e-01
1.09380281e+00 3.61604929e-01 -9.65677947e-02 -4.41234559e-01
3.38660516e-02 3.41203660e-02 2.20829636e-01 -6.89886332e-01
1.03190279e+00 2.56463915e-01 -3.75361204e-01 1.36322582e+00
9.86395329e-02 -5.75763643e-01 6.68086290e-01 6.68079734e-01
1.00433636e+00 4.34312820e-01 1.77101046e-01 1.76734373e-01
2.93604702e-01 -1.35069638e-01 3.19648415e-01 9.03564453e-01
2.12534115e-01 4.76511151e-01 6.37887001e-01 9.87845436e-02
-9.77918446e-01 -6.56754851e-01 7.41027355e-01 1.66293633e+00
-4.32199627e-01 -7.57035732e-01 -1.04868686e+00 -7.96876788e-01
-1.64319482e-02 1.44028306e+00 -3.22176337e-01 -4.86785918e-01
-7.52192736e-01 -1.71270803e-01 4.07057017e-01 6.24199271e-01
1.21893361e-01 -1.43311083e+00 -8.29804838e-01 5.40174901e-01
-4.78179336e-01 -8.52756500e-01 -1.38309216e+00 1.28216013e-01
-7.15327978e-01 -5.82222223e-01 -3.29049528e-01 -7.88592339e-01
8.32597792e-01 -2.57183611e-01 1.37098646e+00 3.10787678e-01
2.14309990e-01 2.16257021e-01 -2.78514624e-01 -5.70574701e-01
-1.36192966e+00 5.52721202e-01 -1.75078019e-01 -3.03818673e-01
-2.39400640e-01 -3.29605341e-01 -4.43522513e-01 3.00037470e-02
-8.42503190e-01 6.02982581e-01 6.68329298e-01 7.48275995e-01
2.12173760e-01 -4.91087168e-01 9.42079484e-01 -1.22447455e+00
1.37308180e+00 -5.02292812e-01 -6.16653860e-01 3.35722387e-01
-6.17956698e-01 4.31827605e-01 1.22945023e+00 -5.29167950e-01
-1.23968983e+00 1.16022013e-01 -1.42156333e-01 5.07252403e-02
-2.49604713e-02 6.88699901e-01 -3.08120251e-01 7.41745949e-01
7.93635070e-01 4.37572330e-01 -9.11464691e-02 -2.29749262e-01
4.55349654e-01 9.03371930e-01 5.95215917e-01 -9.20012355e-01
6.49770856e-01 -6.60634398e-01 -8.03370953e-01 -3.05453867e-01
-6.29987061e-01 9.90179926e-02 -2.46028349e-01 -1.76755041e-02
6.96997166e-01 -8.42863142e-01 -3.91227186e-01 1.97036907e-01
-1.23373449e+00 -1.04883587e+00 -5.97826302e-01 6.25982657e-02
-7.37418175e-01 -2.87489176e-01 -7.50076354e-01 -7.30533838e-01
-8.27575207e-01 -1.57742798e+00 1.04171288e+00 7.84611166e-01
-7.20107615e-01 -9.44104075e-01 7.07792118e-02 6.94838583e-01
5.48201442e-01 -8.09423774e-02 1.16748524e+00 -1.04112875e+00
-5.88285327e-01 -1.19278081e-01 2.56233156e-01 8.85528252e-02
1.17407270e-01 2.67347187e-01 -5.68117738e-01 -1.19292326e-01
-5.10079920e-01 -6.53037190e-01 2.82295644e-01 1.42551973e-01
1.31589413e+00 -5.20417750e-01 -9.64068547e-02 3.87892067e-01
7.22153127e-01 2.65236825e-01 2.32122555e-01 8.67795572e-02
6.11135721e-01 4.53807145e-01 4.42622483e-01 5.78130662e-01
8.03835630e-01 5.12374282e-01 1.15441613e-01 2.09858552e-01
-1.53496206e-01 -7.12821245e-01 6.95329010e-01 1.00610566e+00
1.86406955e-01 -6.74331963e-01 -6.31316304e-01 3.51167709e-01
-1.80035973e+00 -8.17916632e-01 2.69171178e-01 1.96973443e+00
1.59888482e+00 1.65406302e-01 9.77189317e-02 -5.17289221e-01
3.28156859e-01 9.23177525e-02 -8.42983663e-01 -8.76096785e-01
5.26959777e-01 2.30303645e-01 -8.70356485e-02 6.90752029e-01
-3.29535455e-01 1.08887529e+00 5.75742674e+00 5.12400150e-01
-1.03187346e+00 -2.25762218e-01 6.61578774e-01 -1.51215225e-01
-7.46183813e-01 -4.19852994e-02 -1.05550408e+00 5.06677032e-01
1.15910435e+00 -6.01333618e-01 5.09177923e-01 7.15536118e-01
7.95808911e-01 1.10433556e-01 -1.27790940e+00 2.52467752e-01
1.57306075e-01 -1.36572182e+00 4.00004089e-01 -1.49395347e-01
8.13484550e-01 -3.11792374e-01 -1.29606754e-01 7.46591628e-01
8.92392635e-01 -7.76070297e-01 8.14750969e-01 4.34185505e-01
7.11607933e-01 -9.13205981e-01 3.67430121e-01 8.27433407e-01
-8.40563536e-01 -5.26789315e-02 1.61328819e-02 1.52230135e-03
3.18761408e-01 2.56496929e-02 -1.48376369e+00 2.03856409e-01
-7.85161480e-02 2.98566163e-01 -5.06404996e-01 5.75212240e-01
-5.59842288e-01 6.20953381e-01 2.35477090e-01 -2.33575508e-01
5.57964072e-02 -2.46540576e-01 4.95181143e-01 1.31106412e+00
2.21614420e-01 3.43285799e-01 6.87634349e-01 9.32536364e-01
-5.45090616e-01 2.75573492e-01 -3.35373789e-01 -5.41244507e-01
6.90890133e-01 1.62652671e+00 -2.11455539e-01 -6.91610038e-01
7.82677233e-02 1.06548500e+00 5.40977836e-01 4.27308470e-01
-5.93362212e-01 -6.25653803e-01 5.12569427e-01 1.72860786e-01
1.28469139e-01 8.07715282e-02 -1.98987767e-01 -9.37524796e-01
-2.98566610e-01 -1.39468825e+00 3.03310215e-01 -9.67329144e-01
-7.70264149e-01 5.99302769e-01 -1.08135372e-01 -8.40931237e-01
-9.15470779e-01 -7.21719787e-02 -1.23301363e+00 1.11979020e+00
-1.06080186e+00 -1.00691450e+00 -3.84007186e-01 2.87418365e-02
1.12685883e+00 -4.14131284e-01 7.57057130e-01 -1.59846291e-01
-8.82285953e-01 8.69797349e-01 -3.23399782e-01 -9.51406434e-02
8.90285015e-01 -1.60355437e+00 4.97943610e-01 6.44575596e-01
-4.27655220e-01 7.64229178e-01 1.05916786e+00 -1.04303217e+00
-1.50767648e+00 -1.22644949e+00 1.09614432e+00 -4.00916249e-01
4.69170332e-01 -3.09184968e-01 -7.91734278e-01 6.45047784e-01
6.48410439e-01 -6.96259677e-01 9.05307651e-01 -2.23855793e-01
-3.38831283e-02 2.22639069e-01 -8.96943629e-01 9.15155888e-01
6.82186782e-01 -1.64382368e-01 -6.82929337e-01 3.74215513e-01
9.23252165e-01 -1.02851439e+00 -8.32722187e-01 -1.60113111e-01
5.04353881e-01 -3.63554865e-01 4.84384984e-01 -7.10629225e-01
1.01345229e+00 -1.16145886e-01 3.76044929e-01 -2.12200499e+00
-1.08201794e-01 -9.13527250e-01 -1.46969348e-01 1.36405444e+00
7.42188513e-01 -2.60725230e-01 5.25125682e-01 7.42726505e-01
-6.43576086e-01 -7.78866351e-01 -2.18397044e-02 -1.96791336e-01
-5.40498570e-02 4.25402299e-02 7.66156852e-01 5.11481762e-01
3.51878464e-01 9.04075861e-01 -1.67344734e-01 -9.86196175e-02
1.84957176e-01 3.22327346e-01 1.17402470e+00 -7.83297360e-01
-8.31181824e-01 -5.99631488e-01 8.11559081e-01 -1.28822660e+00
3.12018424e-01 -1.02194893e+00 5.55134058e-01 -1.77752852e+00
1.72981143e-01 -4.08761084e-01 1.91435024e-01 5.84716976e-01
-7.86765516e-01 -4.11164254e-01 3.43208402e-01 -2.79377699e-01
-6.81351602e-01 6.15694344e-01 1.40635097e+00 -7.86479786e-02
-8.31557572e-01 2.97772944e-01 -1.24596524e+00 4.87669885e-01
8.26455951e-01 -2.59958595e-01 -6.93291485e-01 -6.62231684e-01
2.93120109e-02 3.89030427e-01 -4.94974166e-01 -6.12738907e-01
3.90129387e-01 -3.80312890e-01 4.04482543e-01 -3.25798571e-01
-1.41907215e-01 -1.27302960e-01 -2.94324368e-01 3.06910336e-01
-1.06775296e+00 6.47043705e-01 3.48824650e-01 4.33586240e-02
6.90940917e-02 -5.14600694e-01 5.23903251e-01 -1.17553480e-01
-7.65282363e-02 3.10596287e-01 -5.00164151e-01 3.47385526e-01
6.15676224e-01 1.75170764e-01 -5.07252097e-01 -8.53547812e-01
-3.57183218e-01 8.62057030e-01 4.05750126e-01 6.63634002e-01
6.08676136e-01 -1.07884169e+00 -1.00548673e+00 2.70083845e-01
1.51339203e-01 1.39797226e-01 -1.16515905e-01 2.42045507e-01
-1.94311306e-01 1.28099427e-01 3.15441899e-02 -2.17765242e-01
-1.28363800e+00 -7.92006627e-02 1.63297534e-01 -7.87006974e-01
-2.03786358e-01 7.95104027e-01 3.71657275e-02 -8.17934513e-01
1.41999170e-01 -4.13231492e-01 -3.06037635e-01 1.90826759e-01
7.12706864e-01 1.68925017e-01 5.65236658e-02 -1.95903510e-01
2.73694426e-01 -1.59720197e-01 -3.75956804e-01 -3.97229940e-01
1.42215991e+00 1.39828831e-01 1.62682846e-01 3.39017540e-01
7.18855739e-01 2.90364623e-01 -1.25382686e+00 2.04925914e-03
-1.41976729e-01 4.28943988e-03 -3.41005474e-01 -1.19754934e+00
-5.80429852e-01 6.20310903e-01 -1.49219096e-01 -5.79759432e-03
8.92942309e-01 -1.63526729e-01 9.32780087e-01 3.80132169e-01
2.78073642e-03 -1.11741340e+00 5.41921437e-01 8.82882535e-01
1.14819574e+00 -8.78411412e-01 -3.67414474e-01 1.01312205e-01
-1.04624760e+00 1.12553477e+00 1.39270210e+00 1.72126576e-01
-2.82955885e-01 3.13817114e-01 1.14206366e-01 3.08606893e-01
-1.59624672e+00 2.92587757e-01 2.16806829e-01 4.83507454e-01
1.01279390e+00 1.72955409e-01 -2.10344136e-01 8.44814003e-01
-8.95738780e-01 1.67357430e-01 9.45850194e-01 9.47873473e-01
-5.58325946e-01 -1.46241009e+00 -2.35988006e-01 6.83611512e-01
-9.75494683e-02 -2.15462446e-01 -3.45540732e-01 2.55216867e-01
-2.29231402e-01 9.49333727e-01 1.92881584e-01 -2.49947101e-01
7.60415077e-01 2.79774368e-01 2.60515213e-01 -1.31073678e+00
-1.37903249e+00 2.60700309e-03 3.81090462e-01 -3.30095023e-01
7.30258346e-01 -5.35975933e-01 -1.63407135e+00 5.84776141e-02
-1.74769834e-01 6.16783619e-01 5.90224326e-01 7.37574637e-01
5.93414783e-01 9.64046836e-01 6.58314943e-01 -7.18797028e-01
-1.08928680e+00 -1.28903270e+00 1.59864143e-01 3.73916715e-01
3.35082382e-01 1.33138150e-01 -4.74321248e-04 3.15190762e-01] | [11.949604034423828, 8.779806137084961] |
f4b2b3fd-d7a4-4fdc-adfb-62877d853801 | learning-to-solve-nlp-tasks-in-an-incremental | null | null | https://aclanthology.org/2021.acl-short.106 | https://aclanthology.org/2021.acl-short.106.pdf | Learning to Solve NLP Tasks in an Incremental Number of Languages | In real scenarios, a multilingual model trained to solve NLP tasks on a set of languages can be required to support new languages over time. Unfortunately, the straightforward retraining on a dataset containing annotated examples for all the languages is both expensive and time-consuming, especially when the number of target languages grows. Moreover, the original annotated material may no longer be available due to storage or business constraints. Re-training only with the new language data will inevitably result in Catastrophic Forgetting of previously acquired knowledge. We propose a Continual Learning strategy that updates a model to support new languages over time, while maintaining consistent results on previously learned languages. We define a Teacher-Student framework where the existing model {``}teaches{''} to a student model its knowledge about the languages it supports, while the student is also trained on a new language. We report an experimental evaluation in several tasks including Sentence Classification, Relational Learning and Sequence Labeling. | ['Roberto Basili', 'Danilo Croce', 'Simone Filice', 'Giuseppe Castellucci'] | 2021-08-01 | null | null | null | acl-2021-5 | ['sentence-classification'] | ['natural-language-processing'] | [ 3.24451715e-01 1.21774532e-01 -4.86595303e-01 -6.51644289e-01
-7.80732810e-01 -1.00530756e+00 6.42138198e-02 7.06732690e-01
-7.64302790e-01 1.39642560e+00 -2.37715438e-01 -5.47640562e-01
3.24362189e-01 -7.20664680e-01 -9.45622921e-01 -3.16603899e-01
1.67608429e-02 7.79479682e-01 4.62585449e-01 -1.84871733e-01
-2.31314108e-01 5.69714546e-01 -1.25816584e+00 5.01470864e-01
1.12601173e+00 4.55573082e-01 7.27356434e-01 4.57116455e-01
-6.02951109e-01 9.94522631e-01 -7.26386368e-01 -5.79766512e-01
1.20799206e-02 -2.12790251e-01 -1.22373009e+00 1.89194366e-01
5.22409499e-01 -3.68220583e-02 -3.14601570e-01 1.12680984e+00
-1.25580966e-01 3.10823947e-01 1.37203038e-02 -1.03056443e+00
-7.60978401e-01 1.19896495e+00 -3.08230549e-01 2.91358173e-01
3.48905087e-01 -1.15080915e-01 7.08948135e-01 -9.52169001e-01
9.40852642e-01 1.00412059e+00 5.38135111e-01 6.85126305e-01
-1.24653339e+00 -6.09400213e-01 5.67252874e-01 1.67678282e-01
-1.27633190e+00 -4.35445070e-01 5.22357941e-01 -1.81113645e-01
1.18192160e+00 1.13848895e-01 4.78260279e-01 7.60666907e-01
-1.76358223e-01 8.03809583e-01 8.87418747e-01 -7.69240379e-01
-7.38225132e-02 9.00795281e-01 2.72102714e-01 8.93688500e-01
2.24378943e-01 -2.90230066e-01 -7.60439873e-01 1.71370342e-01
2.46523321e-02 -2.31091425e-01 -2.34473392e-01 -4.15598363e-01
-9.75532711e-01 5.31288385e-01 1.60707191e-01 5.77173710e-01
1.42963566e-02 -2.72901297e-01 6.18455350e-01 8.31385791e-01
6.04888737e-01 5.45650065e-01 -1.16678536e+00 7.36323819e-02
-8.76173913e-01 -1.42301202e-01 8.95562947e-01 1.25153649e+00
1.15651786e+00 -3.16492319e-02 3.89376640e-01 9.20093358e-01
1.73380766e-02 2.49764636e-01 5.74052095e-01 -5.90395868e-01
6.67166710e-01 7.03203976e-01 -2.08495900e-01 -3.37051064e-01
-1.16618954e-01 -5.17122328e-01 -6.42857075e-01 -3.31601262e-01
3.38964850e-01 -2.02149153e-01 -7.90554941e-01 1.89911711e+00
4.70477670e-01 7.39990994e-02 5.58707356e-01 5.29439636e-02
7.60836601e-01 1.01811707e+00 1.47391900e-01 -7.67048895e-01
8.17103446e-01 -1.17729104e+00 -5.79697669e-01 -5.50090313e-01
1.04034233e+00 -8.13960731e-01 1.20181167e+00 4.52085018e-01
-1.05689466e+00 -7.35718071e-01 -8.85248184e-01 -1.60614118e-01
-4.63382006e-01 6.15424663e-02 6.19950593e-01 4.16525751e-01
-1.04913163e+00 4.84698445e-01 -5.91997266e-01 -5.31707883e-01
7.40196323e-03 3.70720595e-01 -5.08812606e-01 -5.68419933e-01
-1.22910011e+00 1.13728499e+00 9.85161543e-01 -5.20307906e-02
-9.58498716e-01 -6.96177721e-01 -1.03929853e+00 6.48837760e-02
5.93155324e-01 -2.68951505e-01 1.50357378e+00 -1.09948814e+00
-1.36795557e+00 8.94311607e-01 -2.19860688e-01 -3.45483214e-01
3.67678672e-01 -6.34767115e-02 -4.77618366e-01 -1.81088403e-01
2.31438041e-01 5.52170575e-01 4.68180269e-01 -1.23998153e+00
-9.06225860e-01 -2.36244842e-01 1.32280573e-01 2.92297393e-01
-6.41541541e-01 5.12973778e-02 -4.88358468e-01 -5.02741456e-01
2.11151496e-01 -7.83859551e-01 -8.24360922e-02 -2.81907707e-01
-1.21330731e-01 -4.32417005e-01 6.82548046e-01 -7.37878978e-01
1.35916603e+00 -1.91488981e+00 3.33002716e-01 5.74361384e-02
-9.54818428e-02 4.80528712e-01 -2.49390289e-01 4.20399219e-01
-1.47780702e-01 1.79150924e-02 -2.55701691e-01 -3.56784731e-01
-5.96419632e-01 5.59337437e-01 -3.96996826e-01 7.19915703e-02
5.53227253e-02 3.54315877e-01 -1.02401638e+00 -5.91140270e-01
-1.23592556e-01 5.18585891e-02 -4.77239221e-01 2.19999835e-01
-6.19892895e-01 6.41723156e-01 -1.28832519e-01 4.45990443e-01
5.14026940e-01 -3.12212318e-01 7.77968109e-01 8.53801370e-02
-1.20268464e-01 5.40922701e-01 -1.16111231e+00 2.03862262e+00
-8.39906693e-01 3.43302101e-01 -6.61232555e-03 -1.24418306e+00
9.98351097e-01 4.19961810e-01 1.55678943e-01 -3.01122695e-01
-3.93991381e-01 5.02039135e-01 2.81471461e-02 -6.60376251e-01
4.95880753e-01 -4.24528509e-01 -1.69577047e-01 5.98293722e-01
5.60142696e-01 -1.22618169e-01 5.60301125e-01 3.26316208e-01
8.53120506e-01 1.96612522e-01 3.11526656e-01 -3.80463786e-02
8.90114486e-01 2.73033291e-01 7.05596745e-01 5.53129911e-01
-1.97704099e-02 -4.17541891e-01 1.36558726e-01 -6.63824201e-01
-1.02538908e+00 -8.95864487e-01 -1.04545979e-02 1.58539307e+00
-2.21461967e-01 -3.68744582e-01 -3.18999648e-01 -1.09699321e+00
-6.87891319e-02 9.75799799e-01 -4.06070650e-02 -2.19894812e-01
-9.55933392e-01 -3.12819779e-01 4.01017755e-01 3.01830530e-01
2.97980577e-01 -1.12048388e+00 -1.90482065e-01 4.14843887e-01
-8.21694955e-02 -1.24627042e+00 -5.77402651e-01 5.14503300e-01
-8.82122040e-01 -7.39530087e-01 -2.23287538e-01 -1.53710747e+00
1.11439705e+00 2.25877929e-02 1.07545412e+00 3.26000690e-01
9.28883180e-02 2.35298097e-01 -1.31075695e-01 -1.14345454e-01
-1.01876271e+00 4.69012618e-01 2.19704822e-01 -2.29779616e-01
2.93066531e-01 -3.96613210e-01 5.37788808e-01 -2.78441042e-01
-9.98730600e-01 3.03002119e-01 3.17025870e-01 7.73150206e-01
5.92903316e-01 4.23584461e-01 7.87024796e-01 -1.46801221e+00
4.22932982e-01 -2.89071113e-01 -7.88264692e-01 1.01078558e+00
-5.20128727e-01 3.18723619e-01 7.95154393e-01 -7.27360249e-01
-1.32663703e+00 3.10836643e-01 7.52928779e-02 -1.93820477e-01
2.49439571e-02 1.14079916e+00 -2.01379776e-01 1.60335079e-01
5.78651428e-01 4.27615166e-01 -3.66671175e-01 -6.21732533e-01
2.81754881e-01 4.98384088e-01 4.74900365e-01 -8.37905884e-01
7.82341182e-01 -1.51503921e-01 -4.24100965e-01 -7.25666225e-01
-1.32925045e+00 -1.63358063e-01 -1.12292755e+00 -8.67761374e-02
3.36177498e-01 -9.29171622e-01 -2.33656809e-01 3.53293777e-01
-1.50561512e+00 -4.42518264e-01 -5.57331383e-01 4.37637419e-01
-5.82661070e-02 3.52121592e-01 -6.11941159e-01 -5.78925431e-01
-1.51410326e-01 -1.11848915e+00 3.89065653e-01 2.31737122e-01
-2.26415038e-01 -1.32436323e+00 1.48872137e-01 3.15606713e-01
1.15911417e-01 -3.35862666e-01 1.40319216e+00 -1.01818573e+00
-5.37241876e-01 -2.25786284e-01 1.94167688e-01 5.28160036e-01
2.46963456e-01 -2.39072278e-01 -6.83434308e-01 -6.31032705e-01
5.22010140e-02 -9.41031754e-01 5.07202387e-01 -4.08077866e-01
9.52761710e-01 -5.19717216e-01 -3.88486981e-01 1.55502170e-01
1.29432428e+00 3.19982767e-01 -9.12944302e-02 3.36305380e-01
6.31095290e-01 7.39899158e-01 5.06571293e-01 -1.29060119e-01
5.22914648e-01 2.03129724e-01 -2.91528583e-01 1.27037659e-01
9.85565246e-04 -5.33971250e-01 5.46439052e-01 1.25510347e+00
5.17998397e-01 -2.09028840e-01 -1.16990626e+00 5.96612215e-01
-1.51711500e+00 -6.02136135e-01 5.02046824e-01 2.40043616e+00
1.68441617e+00 1.88761130e-01 -2.50988483e-01 -1.76093221e-01
7.15228021e-01 -2.52740771e-01 -7.44278908e-01 -3.24252665e-01
-8.88584256e-02 3.33288044e-01 1.12707615e-01 1.04494810e+00
-7.43991852e-01 1.49930561e+00 6.15851259e+00 4.04073000e-01
-1.38518190e+00 2.05503628e-01 5.20656884e-01 4.77854162e-03
-4.57623154e-01 2.91759461e-01 -1.36090183e+00 7.67365545e-02
1.06259918e+00 -5.40076554e-01 4.26225394e-01 8.80921185e-01
-3.74501467e-01 -2.46503353e-01 -1.44822848e+00 6.46595895e-01
3.88944626e-01 -1.20367205e+00 2.81326979e-01 -4.40475553e-01
7.27091014e-01 2.14881614e-01 -2.78106451e-01 7.75967896e-01
5.31352699e-01 -6.48075640e-01 5.31133175e-01 3.49796802e-01
8.99549484e-01 -7.75154114e-01 4.78033930e-01 1.02387583e+00
-8.84128988e-01 1.30536869e-01 -3.43275696e-01 1.02834053e-01
4.72219884e-02 4.00204659e-01 -1.34853327e+00 3.24759096e-01
6.04412675e-01 6.67357624e-01 -8.04348648e-01 6.41498506e-01
-5.24358034e-01 4.96456176e-01 -2.44351730e-01 -4.27903235e-02
-4.40321155e-02 -1.48224635e-02 3.28880042e-01 1.08491814e+00
3.00721645e-01 -1.26105979e-01 9.04076040e-01 3.92115146e-01
-4.30404097e-01 2.92074978e-01 -7.42956400e-01 -1.35251984e-01
5.20117223e-01 1.13452995e+00 -6.41323745e-01 -8.65467370e-01
-6.18589461e-01 9.47314441e-01 9.37172234e-01 3.15012783e-01
-3.60690027e-01 -2.02075467e-01 -3.89133431e-02 -3.31838727e-02
7.04174638e-02 -3.43459606e-01 2.00866029e-01 -1.52377903e+00
1.76842570e-01 -1.09277570e+00 7.86870182e-01 -6.72647655e-01
-1.22414219e+00 8.76297951e-01 4.50925231e-02 -7.69465864e-01
-1.80628464e-01 -4.24990892e-01 5.16038993e-03 8.25749218e-01
-1.55320585e+00 -9.89055812e-01 2.56445855e-01 5.97863138e-01
9.33557689e-01 -6.32111788e-01 9.77792740e-01 3.99809211e-01
-5.67253590e-01 5.84425628e-01 -4.07148935e-02 1.18299603e-01
7.48061657e-01 -1.02110279e+00 -6.63116202e-02 1.02004194e+00
5.56458771e-01 7.57171333e-01 6.70019329e-01 -8.05145085e-01
-1.33010161e+00 -1.13542640e+00 1.79314744e+00 -3.53002757e-01
9.07088757e-01 -4.38743800e-01 -1.48902810e+00 1.26923168e+00
3.61183017e-01 -4.54635499e-03 9.82041597e-01 2.88177609e-01
-4.30871159e-01 -4.02053833e-01 -1.00233018e+00 4.05532748e-01
6.96751356e-01 -6.30812883e-01 -7.69219697e-01 7.68339932e-01
9.36379254e-01 -5.94140291e-01 -8.11630547e-01 3.89769316e-01
1.90314297e-02 -2.00922683e-01 4.97507811e-01 -1.22599101e+00
2.34878659e-02 -3.10229480e-01 -9.04068127e-02 -1.35309744e+00
5.35465777e-02 -4.38301295e-01 -3.16801667e-03 1.36450231e+00
8.85201752e-01 -5.50416291e-01 7.05708206e-01 6.45361245e-01
-1.59151793e-01 -6.08125985e-01 -1.02253771e+00 -8.37565124e-01
1.87269688e-01 -4.29593831e-01 4.22900349e-01 1.47456133e+00
2.70604223e-01 1.04599094e+00 -2.99905747e-01 3.11875641e-01
2.04701781e-01 2.34595180e-01 6.81705713e-01 -1.13752162e+00
-3.40955883e-01 1.83092192e-01 1.89573601e-01 -1.06206489e+00
5.95539331e-01 -1.53734207e+00 1.35672465e-01 -1.28742290e+00
2.84365803e-01 -8.28896999e-01 -2.42017820e-01 8.55756581e-01
-1.60828605e-01 -3.52661610e-01 6.04783408e-02 1.55746713e-01
-7.41446316e-01 2.89222330e-01 1.10104704e+00 -1.96414486e-01
-3.21928978e-01 1.20950632e-01 -5.47824323e-01 8.72126579e-01
6.94448411e-01 -7.77967930e-01 -6.81999385e-01 -8.41307461e-01
5.75051606e-01 1.96088076e-01 -4.04623359e-01 -8.09108734e-01
5.60416162e-01 -3.43311399e-01 2.94867992e-01 -5.43802798e-01
5.38004190e-02 -9.79598522e-01 -5.16947620e-02 5.70296764e-01
-6.30107403e-01 3.07772577e-01 6.02823496e-01 2.59315610e-01
-2.68168807e-01 -7.36021459e-01 9.31706905e-01 -4.65814203e-01
-7.28665650e-01 2.62462825e-01 -2.84608930e-01 1.08212985e-01
1.01598644e+00 3.05380553e-01 -4.94216792e-02 -9.72185507e-02
-1.16403115e+00 5.41349113e-01 2.50150710e-01 4.91794258e-01
5.07673085e-01 -1.10430264e+00 -6.69948578e-01 1.88304976e-01
1.62124038e-01 2.67528385e-01 4.28896844e-02 3.16660732e-01
-4.86684769e-01 3.96739215e-01 1.42089486e-01 -5.49373806e-01
-1.50422704e+00 9.96287048e-01 1.88648358e-01 -8.21700096e-01
-2.68352360e-01 1.07896614e+00 -1.66759446e-01 -1.14630282e+00
4.84095722e-01 -3.62156630e-01 -2.76923031e-01 1.69744134e-01
3.26264292e-01 -1.54786110e-01 1.74605533e-01 -4.68652576e-01
-1.29045784e-01 1.56418309e-01 -6.71371937e-01 -1.59444198e-01
1.39956605e+00 -2.31562376e-01 -4.98333454e-01 9.55158055e-01
9.82246220e-01 1.66164324e-01 -6.88629627e-01 -1.08521390e+00
5.29698193e-01 -1.55539677e-01 -4.25031275e-01 -9.96812880e-01
-8.13944817e-01 7.44818032e-01 1.76709592e-01 -1.16154134e-01
9.99475658e-01 1.91212401e-01 6.92988396e-01 1.07156849e+00
5.78248322e-01 -1.12787259e+00 2.06881166e-01 9.26668644e-01
7.57216156e-01 -1.24759805e+00 1.12683915e-01 -4.32520062e-01
-5.15804350e-01 1.20457137e+00 1.05168068e+00 5.10627329e-01
6.39497995e-01 1.46450147e-01 3.63808274e-02 1.90715596e-01
-1.05505729e+00 3.64952475e-01 1.36222765e-01 3.39972585e-01
6.89384520e-01 -4.46773060e-02 -3.24280933e-03 3.17687184e-01
-3.38780761e-01 -4.98126037e-02 5.05244434e-01 1.19114292e+00
-6.87902927e-01 -1.66219807e+00 -1.38058648e-01 1.77104905e-01
-1.58495665e-01 -3.11231375e-01 -2.88976580e-01 6.86583519e-01
4.75020409e-01 5.26599169e-01 3.26494537e-02 1.39549896e-01
3.44910413e-01 6.64905369e-01 6.94967568e-01 -1.52159345e+00
-5.01253784e-01 -2.74758670e-03 1.63474545e-01 2.38713488e-01
-4.76413727e-01 -8.01248729e-01 -1.49544060e+00 -2.54323557e-02
-2.07085133e-01 5.78789890e-01 4.00201082e-01 1.08949697e+00
-7.97228962e-02 3.35056037e-01 5.40511549e-01 1.32738724e-01
-5.95577419e-01 -9.87337351e-01 -2.86892414e-01 6.11246713e-02
2.36528635e-01 -1.65197954e-01 -1.30061917e-02 4.95975822e-01] | [10.379491806030273, 9.391749382019043] |
e5c58118-f834-4aaf-895c-b7eb2c5a696f | adaptive-prototypical-networks-with-label | 2101.03526 | null | https://arxiv.org/abs/2101.03526v1 | https://arxiv.org/pdf/2101.03526v1.pdf | Adaptive Prototypical Networks with Label Words and Joint Representation Learning for Few-Shot Relation Classification | Relation classification (RC) task is one of fundamental tasks of information extraction, aiming to detect the relation information between entity pairs in unstructured natural language text and generate structured data in the form of entity-relation triple. Although distant supervision methods can effectively alleviate the problem of lack of training data in supervised learning, they also introduce noise into the data, and still cannot fundamentally solve the long-tail distribution problem of the training instances. In order to enable the neural network to learn new knowledge through few instances like humans, this work focuses on few-shot relation classification (FSRC), where a classifier should generalize to new classes that have not been seen in the training set, given only a number of samples for each class. To make full use of the existing information and get a better feature representation for each instance, we propose to encode each class prototype in an adaptive way from two aspects. First, based on the prototypical networks, we propose an adaptive mixture mechanism to add label words to the representation of the class prototype, which, to the best of our knowledge, is the first attempt to integrate the label information into features of the support samples of each class so as to get more interactive class prototypes. Second, to more reasonably measure the distances between samples of each category, we introduce a loss function for joint representation learning to encode each support instance in an adaptive manner. Extensive experiments have been conducted on FewRel under different few-shot (FS) settings, and the results show that the proposed adaptive prototypical networks with label words and joint representation learning has not only achieved significant improvements in accuracy, but also increased the generalization ability of few-shot RC models. | ['Kuangrong Hao', 'Yaochu Jin', 'Yan Xiao'] | 2021-01-10 | null | null | null | null | ['few-shot-relation-classification', 'few-shot-relation-classification'] | ['methodology', 'natural-language-processing'] | [ 3.61682504e-01 3.97738785e-01 -4.52037424e-01 -6.72196269e-01
-3.26763988e-01 4.34412882e-02 5.04331708e-01 4.12545472e-01
-3.26055825e-01 8.55170548e-01 -2.30925474e-02 -4.67644483e-02
-3.73226345e-01 -1.19078660e+00 -4.70409542e-01 -7.76405454e-01
3.73735763e-02 6.66565299e-01 3.26804668e-01 -3.38809848e-01
-1.45284235e-01 5.87529540e-01 -1.80019593e+00 3.20704818e-01
1.14389455e+00 1.06565714e+00 1.44364625e-01 6.16598651e-02
-3.36543053e-01 7.64012396e-01 -5.79234183e-01 -4.09700453e-01
-2.05483306e-02 -5.14771044e-01 -9.54139948e-01 1.60684511e-01
-1.26992300e-01 4.13820371e-02 -3.55094165e-01 1.08614707e+00
3.33180428e-01 5.98297417e-01 1.04528165e+00 -1.15128195e+00
-8.31154644e-01 8.80337179e-01 -4.99050528e-01 1.64338216e-01
2.24763751e-01 -1.84984133e-01 9.67382491e-01 -8.98066521e-01
6.83041811e-01 1.25484335e+00 3.44259739e-01 5.30360520e-01
-1.10314763e+00 -8.54528368e-01 3.02592099e-01 6.07038498e-01
-1.61783636e+00 -3.94395798e-01 9.11684752e-01 -1.44358754e-01
6.71616018e-01 1.05928980e-01 5.00526667e-01 1.10672402e+00
-2.45200291e-01 7.73931742e-01 6.62093282e-01 -6.32814884e-01
3.93097281e-01 4.87286270e-01 4.79213327e-01 4.36476380e-01
4.17666227e-01 -1.41989097e-01 -1.14607841e-01 -1.36697710e-01
3.79831225e-01 2.47550786e-01 -4.69674796e-01 -3.92339826e-01
-7.89741218e-01 9.41982985e-01 8.12821686e-01 7.68117011e-01
-2.13846266e-01 -4.30679858e-01 3.68924260e-01 2.28414372e-01
6.98461592e-01 3.39628518e-01 -4.92873222e-01 3.38170946e-01
-3.73113900e-01 -6.69439510e-02 8.05661500e-01 1.02948964e+00
1.03795350e+00 -2.40218207e-01 -3.52771610e-01 1.12020993e+00
-2.53565330e-02 1.08962007e-01 8.35729718e-01 -3.29517812e-01
7.26480424e-01 9.69276667e-01 -1.99179143e-01 -1.09897053e+00
-3.61613572e-01 -5.09164035e-01 -1.12268627e+00 -1.36198744e-01
6.58408403e-02 -1.91242799e-01 -9.08509314e-01 1.70924056e+00
5.10597348e-01 2.23301545e-01 4.88284498e-01 6.57349944e-01
1.10330725e+00 8.19815993e-01 4.11824770e-02 -6.60373092e-01
1.23694766e+00 -7.33548999e-01 -8.29696000e-01 -2.43526533e-01
9.22790885e-01 -2.67631888e-01 8.05482984e-01 9.32269022e-02
-3.83946538e-01 -7.00363040e-01 -1.22572005e+00 2.18787000e-01
-6.43053412e-01 5.26912361e-02 7.53024101e-01 4.83240902e-01
-2.58452684e-01 7.77301967e-01 -4.88128722e-01 -3.34042907e-01
6.55782282e-01 7.81145692e-02 -5.31944513e-01 -4.23660100e-01
-1.58088708e+00 8.75687659e-01 9.88655865e-01 -2.63872393e-03
-3.27657610e-01 -3.46187025e-01 -1.02937925e+00 2.68037498e-01
9.39828455e-01 -2.67255634e-01 8.99531662e-01 -9.04631197e-01
-1.17218530e+00 4.69674259e-01 -9.07106847e-02 -3.91380936e-01
3.39943357e-02 1.17034778e-01 -7.24440277e-01 -9.18739811e-02
6.58265576e-02 3.94144475e-01 5.26732862e-01 -1.32762873e+00
-6.74144149e-01 -4.59834129e-01 -1.57593846e-01 2.12879017e-01
-5.37478387e-01 -2.26463452e-01 -1.05780261e-02 -5.53706229e-01
1.15658082e-01 -6.95456505e-01 -1.72377825e-01 -3.91152948e-01
-2.97753215e-01 -7.19348729e-01 6.90117836e-01 -1.40151277e-01
1.12931955e+00 -2.12287974e+00 -3.58789712e-02 9.53776985e-02
8.49572271e-02 6.72781706e-01 5.60099818e-02 2.27993444e-01
-3.56420189e-01 -5.93793951e-02 -2.63389736e-01 -8.90926570e-02
-3.92812639e-01 4.11330223e-01 -2.84049749e-01 1.25805378e-01
3.09943646e-01 7.21658289e-01 -1.03368211e+00 -4.46507424e-01
-8.84021595e-02 2.39667147e-01 -1.17164999e-01 2.91388720e-01
-2.74647713e-01 7.78689384e-02 -4.76716369e-01 3.26977611e-01
5.21783769e-01 -2.46815756e-01 1.97160050e-01 -2.87875980e-01
4.35997277e-01 3.87287252e-02 -1.43128264e+00 1.34888065e+00
-3.74093413e-01 1.99612960e-01 -5.97681940e-01 -1.44009101e+00
1.34910679e+00 5.48332632e-01 2.51548111e-01 -4.45920199e-01
2.31731623e-01 1.81794927e-01 3.21285546e-01 -5.44030309e-01
2.24661097e-01 -5.27372718e-01 1.39512103e-02 2.44195625e-01
3.00299793e-01 3.10345829e-01 3.61389667e-01 2.35250983e-02
9.06225502e-01 -2.45073035e-01 6.77562714e-01 2.59884685e-01
4.93177861e-01 -4.59566899e-02 9.03008997e-01 5.18656611e-01
9.34334919e-02 4.06431645e-01 3.89459401e-01 -5.92649519e-01
-6.27683580e-01 -7.98438370e-01 -3.41443121e-01 9.06379938e-01
2.55831510e-01 -2.98602134e-01 -2.91184604e-01 -1.08131039e+00
-1.22742303e-01 1.06442809e+00 -6.05412304e-01 -5.99469483e-01
-2.84231901e-01 -9.44478869e-01 2.43338883e-01 5.08528948e-01
4.58526164e-01 -1.30906653e+00 -2.35601828e-01 2.54448622e-01
-1.66437775e-01 -9.94933069e-01 -7.29027614e-02 6.69336021e-01
-7.51557648e-01 -1.09652269e+00 -5.31351984e-01 -9.96502459e-01
6.95675611e-01 1.71182469e-01 7.58605659e-01 2.26836037e-02
-3.88767093e-01 -2.93109000e-01 -7.72554755e-01 -1.89501271e-01
-3.02280545e-01 1.40455812e-01 1.56585321e-01 3.37868273e-01
7.38731205e-01 -5.39145112e-01 -3.04475892e-02 2.41027057e-01
-9.22451556e-01 -1.05513088e-01 7.32362509e-01 1.07655239e+00
4.59025860e-01 5.82566738e-01 9.77793694e-01 -1.32749057e+00
5.66513002e-01 -7.19016075e-01 -1.74417242e-01 4.36266243e-01
-7.87197828e-01 2.00611621e-01 9.01309609e-01 -6.85881495e-01
-1.33155227e+00 -1.00471221e-01 3.38580608e-02 -3.83118778e-01
-3.56354088e-01 6.79213881e-01 -4.91493374e-01 3.84326965e-01
8.78526151e-01 1.87943801e-01 -2.57266611e-02 -5.15272141e-01
4.55332786e-01 9.94665921e-01 3.44268113e-01 -5.37176132e-01
6.11395895e-01 7.07701892e-02 -9.72651877e-03 -6.30824864e-01
-1.20284176e+00 -5.51143825e-01 -8.51239800e-01 1.84907377e-01
6.86337948e-01 -7.65173316e-01 -3.93914312e-01 1.09241717e-01
-1.10945451e+00 2.49285176e-02 -5.32177567e-01 5.28221130e-01
-2.27988094e-01 2.34290987e-01 -4.29847389e-01 -7.01628804e-01
-1.26798585e-01 -7.64742374e-01 6.35238051e-01 4.96428698e-01
-5.04342169e-02 -8.47617984e-01 7.73877278e-02 9.64799747e-02
7.47913569e-02 9.56260860e-02 1.11527491e+00 -1.27182937e+00
-1.85801119e-01 -5.24808049e-01 -2.02365071e-01 3.63827407e-01
5.93389273e-01 -2.25501060e-01 -8.78127575e-01 -2.18500718e-01
1.41638801e-01 -5.06240964e-01 8.34524751e-01 -6.14343509e-02
1.15173244e+00 -3.43758941e-01 -7.61486471e-01 2.52037495e-01
1.14835203e+00 5.00563502e-01 5.92147171e-01 1.48963584e-02
7.02610791e-01 7.77640700e-01 1.05169261e+00 3.26870263e-01
2.09192172e-01 7.56081522e-01 9.99717712e-02 9.78770256e-02
-5.88255487e-02 -2.76097059e-01 -4.73769456e-02 6.30818009e-01
-6.68632165e-02 -2.51709908e-01 -6.51568472e-01 3.96154523e-01
-1.86263037e+00 -1.02624166e+00 1.66723356e-01 2.18797135e+00
1.09716058e+00 2.04343036e-01 -7.19566420e-02 2.39644781e-01
1.01169956e+00 7.18274489e-02 -4.36666101e-01 -1.14248618e-01
7.12772384e-02 1.75359055e-01 2.79304422e-02 1.74204752e-01
-1.13737822e+00 9.12675500e-01 4.64659643e+00 1.23736095e+00
-9.03573036e-01 -2.10888721e-02 7.00794280e-01 2.03118876e-01
-5.75329512e-02 9.83249843e-02 -1.10924721e+00 3.83893967e-01
8.85559976e-01 -2.51568079e-01 1.72695026e-01 1.03298187e+00
-2.82054663e-01 7.27382526e-02 -1.36051023e+00 8.61145198e-01
1.63141683e-01 -1.07892227e+00 1.19392671e-01 1.19700953e-02
7.52273679e-01 -4.23668444e-01 -3.53361726e-01 7.21049845e-01
1.97084054e-01 -1.02051270e+00 1.07318670e-01 6.18752182e-01
7.00323939e-01 -1.03583324e+00 1.14298832e+00 7.42630363e-01
-1.21031284e+00 -2.34002262e-01 -8.22321355e-01 -1.46151498e-01
-2.53996328e-02 7.71705151e-01 -1.04231906e+00 7.65557587e-01
4.25818026e-01 7.13993371e-01 -5.91346323e-01 9.77394998e-01
-2.86248624e-01 2.46898919e-01 -1.13713168e-01 -2.97775269e-01
2.04082988e-02 -3.85077260e-02 1.84237927e-01 8.65447819e-01
2.52721548e-01 4.13320541e-01 4.14508104e-01 7.37684906e-01
-1.10146686e-01 4.00560200e-01 -8.26171041e-01 1.74133591e-02
7.08696902e-01 1.21738958e+00 -8.13652933e-01 -5.64687192e-01
-3.40894550e-01 6.93951190e-01 7.88168609e-01 2.15694010e-01
-5.07339001e-01 -8.57034564e-01 2.02800274e-01 -1.15856580e-01
4.66818482e-01 3.84593487e-01 -6.81773797e-02 -1.19869578e+00
9.46950540e-02 -6.88643217e-01 5.06340384e-01 -3.86207163e-01
-1.53473973e+00 7.47534633e-01 1.29111975e-01 -1.29595029e+00
-4.04820144e-01 -3.13405335e-01 -6.63373649e-01 7.60482132e-01
-1.14392579e+00 -8.46010864e-01 -3.17340523e-01 4.26183879e-01
4.80601788e-01 -5.02020121e-01 1.09565282e+00 2.30334163e-01
-6.50579751e-01 7.11844504e-01 1.54697716e-01 3.38771790e-01
6.83823884e-01 -1.05092132e+00 4.85687479e-02 5.86264849e-01
3.52643222e-01 6.38766289e-01 5.44987321e-01 -7.18826473e-01
-6.78835630e-01 -1.19444096e+00 8.62587214e-01 -1.01080924e-01
4.37940598e-01 -2.66176522e-01 -1.35477412e+00 6.58719063e-01
-2.47225374e-01 4.58875567e-01 8.98594856e-01 4.52348948e-01
-3.55805188e-01 -3.67925853e-01 -1.20999515e+00 4.95623469e-01
9.25197423e-01 -3.41244817e-01 -1.00473976e+00 3.41232210e-01
9.80674446e-01 9.53475982e-02 -9.66177821e-01 7.65113950e-01
9.78576541e-02 -7.00915813e-01 7.76786566e-01 -6.83583915e-01
3.83290470e-01 -3.46885681e-01 -1.56461164e-01 -1.63316679e+00
-3.85698915e-01 1.04583010e-01 -3.11073631e-01 1.62366521e+00
5.05732059e-01 -6.40726805e-01 7.15409577e-01 4.85738069e-01
-3.74749899e-02 -1.05507696e+00 -8.56904924e-01 -9.11015153e-01
-9.48797315e-02 -2.09804624e-01 7.84736514e-01 1.12512481e+00
2.86401719e-01 9.51207876e-01 -3.64447266e-01 6.61311895e-02
4.80549604e-01 3.85360897e-01 5.69262743e-01 -1.58456326e+00
-4.45153415e-01 -9.47638974e-02 -6.19157910e-01 -7.90931225e-01
2.62925774e-01 -1.06931889e+00 1.79287225e-01 -1.37488139e+00
4.41508681e-01 -8.51595581e-01 -3.17032903e-01 6.43047452e-01
-4.56706703e-01 -9.02215987e-02 -3.68395895e-02 1.82805747e-01
-5.98261833e-01 9.74370897e-01 1.10932839e+00 -2.54567415e-01
-3.00309509e-01 4.36167538e-01 -9.10830081e-01 6.02903247e-01
5.04841506e-01 -6.94749177e-01 -7.00303793e-01 2.38108248e-01
-1.06487185e-01 1.70772284e-01 4.53079538e-03 -9.58769262e-01
2.04645574e-01 -1.94355384e-01 4.99250859e-01 -4.45515245e-01
2.98554212e-01 -1.01658094e+00 1.31290317e-01 3.65555346e-01
-5.08586049e-01 -7.51587927e-01 -3.23553532e-01 8.37234318e-01
-3.43666583e-01 -6.87964797e-01 7.69885123e-01 -6.43211156e-02
-7.69079983e-01 3.39512199e-01 -2.28988333e-03 4.87259030e-02
1.24306023e+00 -8.71556476e-02 -3.73553634e-01 -7.86111131e-02
-8.41070831e-01 2.30341330e-01 1.85734436e-01 5.54044843e-01
6.42075956e-01 -1.62561989e+00 -5.59511483e-01 2.21667856e-01
5.79972386e-01 4.03977633e-01 1.51201516e-01 4.21633244e-01
4.55362909e-02 3.37690055e-01 1.28861377e-02 -2.77005911e-01
-1.08743119e+00 9.99822378e-01 5.27918823e-02 -3.67628574e-01
-7.24021673e-01 7.84899414e-01 2.11224630e-01 -5.45329511e-01
3.35298002e-01 -1.14781201e-01 -8.50775778e-01 3.96675348e-01
6.86457813e-01 7.88653493e-02 9.74460095e-02 -5.89766443e-01
-2.16637135e-01 3.12790751e-01 -6.38374388e-01 3.90322298e-01
1.35220540e+00 1.15179874e-01 1.47763612e-02 7.91732907e-01
1.25308716e+00 -3.84820253e-01 -8.78869355e-01 -7.13030398e-01
2.54832774e-01 -5.35626650e-01 -2.90958971e-01 -5.73366344e-01
-9.83601749e-01 6.71389759e-01 5.56297004e-01 3.04283798e-01
8.38802874e-01 2.94593304e-01 7.12647378e-01 6.51721537e-01
4.97090548e-01 -9.16027308e-01 9.64829251e-02 4.30749714e-01
4.75843012e-01 -1.28516817e+00 -3.90102970e-03 -6.78129196e-01
-7.10929275e-01 1.17922831e+00 8.20465386e-01 3.44931781e-02
6.85321331e-01 -1.09465756e-01 -2.54743129e-01 -1.71492696e-01
-7.64480233e-01 -3.77070040e-01 2.91511476e-01 7.48426080e-01
4.21194553e-01 1.66230306e-01 -3.08016330e-01 9.21781600e-01
1.12451151e-01 -4.85689081e-02 3.44692230e-01 7.48404980e-01
-7.65034735e-01 -1.09575152e+00 -5.98212034e-02 7.60591090e-01
-1.50199428e-01 5.96949421e-02 -1.23551674e-01 5.30025005e-01
3.61957580e-01 9.72092509e-01 -9.29917768e-02 -4.71428156e-01
3.56278896e-01 2.92830110e-01 1.78319529e-01 -1.15628588e+00
-5.61298393e-02 -3.66966158e-01 1.31063238e-01 4.25261445e-03
-2.43697122e-01 -3.52609724e-01 -1.35326838e+00 8.36956128e-02
-7.96593606e-01 5.47301650e-01 1.48543701e-01 1.41278291e+00
3.51472036e-03 6.71426237e-01 7.80608237e-01 -6.94682121e-01
-6.62907422e-01 -1.25551307e+00 -9.06888127e-01 5.83531260e-01
-2.70343591e-02 -1.00173068e+00 -5.14872313e-01 -2.47576267e-01] | [9.22569751739502, 8.469858169555664] |
fcdfeeeb-a080-410f-b79b-ec58361e0e9c | coupled-training-of-sequence-to-sequence | null | null | https://ieeexplore.ieee.org/abstract/document/9052912 | https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9052912&casa_token=7z2KnkcdEkIAAAAA:VuhPm22dVjjwhCKWpAQpJnKUkmYHZHAqxd6w89UA-eecNPtW0U05m4KmPpn69GlcmejMc1c9OOZc&tag=1 | Coupled Training of Sequence-to-Sequence Models for Accented Speech Recognition | Accented speech poses significant challenges for state-of-the-art automatic speech recognition (ASR) systems. Accent is a property of speech that lasts throughout an utterance in varying degrees of strength. This makes it hard to isolate the influence of accent on individual speech sounds. We propose coupled training for encoder-decoder ASR models that acts on pairs of utterances corresponding to the same text spoken by speakers with different accents. This training regime introduces an L2 loss between the attention-weighted representations corresponding to pairs of utterances with the same text, thus acting as a regularizer and encouraging representations from the encoder to be more accent-invariant. We focus on recognizing accented English samples from the Mozilla Common Voice corpus. We obtain significant error rate reductions on accented samples from a large set of diverse accents using coupled training. We also show consistent improvements in performance on heavily accented samples (as determined by a standalone accent classifier). | ['Preethi Jyothi', 'Nitish Joshi', 'Vinit Unni'] | 2020-05-14 | null | null | null | null | ['accented-speech-recognition'] | ['speech'] | [ 3.76140386e-01 2.67675042e-01 -5.82389534e-02 -7.35703468e-01
-9.76819456e-01 -6.34161353e-01 4.71516639e-01 -2.20744491e-01
-5.18942714e-01 4.80145752e-01 7.87832201e-01 -3.41203511e-01
3.61113191e-01 -1.77354753e-01 -7.08913743e-01 -7.12463617e-01
1.34646505e-01 5.57697177e-01 -2.57724255e-01 -5.03040254e-01
-3.28728169e-01 3.91174793e-01 -1.08423924e+00 3.30217570e-01
5.14859140e-01 5.47694564e-01 5.06628633e-01 9.20432389e-01
-4.72389683e-02 8.71228874e-01 -9.49009597e-01 -3.71343583e-01
4.61398065e-02 -6.02252722e-01 -9.33995605e-01 3.90207499e-01
6.58224344e-01 1.15931913e-01 -5.34716070e-01 9.64948654e-01
6.56140924e-01 2.74991095e-01 4.95086998e-01 -5.35207868e-01
-8.15289676e-01 1.13347697e+00 -2.30547130e-01 5.30086458e-01
2.11425051e-01 -9.13890600e-02 1.22256565e+00 -9.24586475e-01
3.38664979e-01 1.40892184e+00 3.73602390e-01 7.96804845e-01
-1.38701355e+00 -5.54306030e-01 3.27086419e-01 -7.23652244e-02
-1.41326416e+00 -1.27193928e+00 8.24546039e-01 4.87997606e-02
1.15728664e+00 4.62960869e-01 1.20398141e-01 1.21894836e+00
-3.33177894e-02 1.07936966e+00 8.94320250e-01 -6.29690170e-01
1.71955183e-01 6.06392026e-02 1.04672685e-01 8.62865746e-02
-6.62563205e-01 7.59091526e-02 -7.48405397e-01 2.92549998e-01
3.06743652e-01 -4.75160509e-01 -4.79054809e-01 2.92655557e-01
-1.08469617e+00 7.96627402e-01 1.79660916e-01 3.35717767e-01
-4.85353261e-01 -2.68030077e-01 3.53871733e-01 4.89704728e-01
4.30368453e-01 4.12845105e-01 -8.00109804e-01 -2.74925262e-01
-8.69851172e-01 -5.72812892e-02 7.80932724e-01 1.00882876e+00
5.31228364e-01 5.34008861e-01 -1.01998359e-01 1.49650788e+00
2.31614843e-01 8.44665766e-01 6.69872642e-01 -5.76583385e-01
3.32011521e-01 -1.53233618e-01 -3.50071847e-01 -1.87575623e-01
-6.19119070e-02 -5.56962132e-01 -6.04944289e-01 8.39748755e-02
3.53342354e-01 -5.10431588e-01 -1.30512464e+00 2.07470870e+00
-8.71589631e-02 1.77027322e-02 5.86823225e-01 7.30910599e-01
4.67783689e-01 9.35103595e-01 1.49459541e-01 -2.73163110e-01
1.42015576e+00 -1.04718208e+00 -9.81983006e-01 -7.56177664e-01
3.23414743e-01 -1.22160649e+00 1.08742249e+00 8.92706215e-02
-1.31737673e+00 -7.18991220e-01 -8.37491691e-01 3.51234972e-02
-2.87757874e-01 -1.10303119e-01 -1.28990561e-01 6.89423680e-01
-1.06504834e+00 4.11603227e-02 -6.64419591e-01 -4.71001565e-02
-2.93550780e-03 3.54758203e-01 -3.25789511e-01 6.93412274e-02
-1.30783904e+00 1.06071210e+00 1.09268412e-01 -1.96469557e-02
-6.35882497e-01 -8.36229444e-01 -1.02322960e+00 9.58312452e-02
1.24827167e-02 9.78661701e-03 1.61812472e+00 -1.31830001e+00
-1.71165121e+00 7.74846196e-01 -5.23041129e-01 -6.05379641e-01
3.73086371e-02 -5.13350904e-01 -9.37759340e-01 -2.63004839e-01
-1.54186964e-01 6.26123905e-01 8.91289771e-01 -9.60305452e-01
-6.21135116e-01 -3.45198244e-01 -5.92719316e-01 5.17538369e-01
-9.00984928e-02 4.72438335e-01 -1.66085318e-01 -9.47333395e-01
-9.58720520e-02 -1.14711630e+00 -7.69780651e-02 -1.08940530e+00
-6.12461150e-01 -2.31791779e-01 8.87891173e-01 -8.85881245e-01
1.10444188e+00 -2.45920610e+00 2.36132815e-01 -1.40400352e-02
-3.97270650e-01 5.02549827e-01 -2.49502108e-01 1.84414163e-01
-3.26940924e-01 4.62729074e-02 -1.18941389e-01 -5.72875559e-01
4.69276011e-02 5.31451225e-01 -5.72855234e-01 1.92989245e-01
5.93598843e-01 5.68404436e-01 -6.94997728e-01 -1.48449719e-01
2.20335931e-01 8.14326167e-01 -2.80752569e-01 7.11256206e-01
-6.72538877e-02 3.85964423e-01 1.19580887e-01 2.10745037e-01
4.26573068e-01 4.28213447e-01 2.38039672e-01 1.18067980e-01
-1.11965418e-01 1.04584396e+00 -9.17576253e-01 1.36230958e+00
-8.33677709e-01 6.24169052e-01 4.43831146e-01 -7.11373389e-01
1.07349920e+00 8.90619814e-01 -2.12259233e-01 -7.46033251e-01
-6.38701022e-02 2.17462957e-01 4.36466485e-01 -3.39821503e-02
5.80410182e-01 -4.51714307e-01 -1.92231864e-01 1.81739807e-01
3.67261499e-01 -2.72302955e-01 6.01970637e-03 -1.75650448e-01
8.45166087e-01 -5.07747769e-01 7.64472038e-02 -4.06308591e-01
3.96888971e-01 -4.93673265e-01 5.79524875e-01 7.36402035e-01
-2.27355361e-01 8.63815963e-01 -7.24344775e-02 -1.85311779e-01
-1.22687519e+00 -1.29629529e+00 -4.49003071e-01 1.53313172e+00
-4.34630096e-01 -1.40864253e-01 -7.38116264e-01 -5.78432202e-01
-2.37509713e-01 1.21150529e+00 -2.96990305e-01 -4.29751985e-02
-9.39431965e-01 -4.75440264e-01 7.05670357e-01 6.12431467e-01
2.64011640e-02 -1.33055663e+00 1.85315728e-01 3.28771174e-01
-1.37678161e-01 -1.25483239e+00 -1.03383553e+00 8.74725401e-01
-2.41508991e-01 -3.42220038e-01 -8.14993501e-01 -1.11802638e+00
4.10316378e-01 -9.66046900e-02 1.25943100e+00 -2.65525103e-01
2.90553600e-01 1.29580483e-01 -3.76802891e-01 -4.39503491e-01
-1.22531831e+00 3.98250312e-01 2.98279762e-01 3.06889802e-01
5.57374954e-01 -2.82729566e-01 1.39091611e-01 4.04037386e-01
-6.22361720e-01 -2.53531724e-01 5.51767170e-01 9.55187440e-01
5.86414039e-01 -1.24941893e-01 8.84039283e-01 -8.05109739e-01
3.96548957e-01 -4.52992618e-01 -4.24539655e-01 4.50886041e-02
-1.65176317e-01 3.40369463e-01 8.52244973e-01 -4.56739873e-01
-1.16446722e+00 2.50218153e-01 -7.41115034e-01 -3.61400813e-01
-6.22822881e-01 2.63098478e-01 -4.53199416e-01 5.41011274e-01
4.81380582e-01 3.78862590e-01 3.77322324e-02 -5.72358489e-01
4.33637887e-01 1.20269477e+00 8.18539619e-01 -2.73790747e-01
4.21993703e-01 -4.24691617e-01 -6.73072278e-01 -1.53229153e+00
-9.18032646e-01 -5.31974018e-01 -4.69010830e-01 9.53441486e-02
9.22340930e-01 -9.95585322e-01 -2.23273680e-01 4.87694621e-01
-1.17430270e+00 -4.77478266e-01 -3.61140132e-01 5.70015192e-01
-3.88220161e-01 -1.15359098e-01 -6.63551748e-01 -9.25776243e-01
-1.70582011e-01 -1.33100796e+00 1.00741005e+00 1.74961090e-01
-6.74343169e-01 -1.10821915e+00 2.36028537e-01 3.92227173e-01
5.65602124e-01 -7.16330290e-01 6.62313819e-01 -1.17591631e+00
-4.23710607e-02 2.77322363e-02 4.33385909e-01 9.15114284e-01
5.23577332e-01 -7.19807968e-02 -1.40569484e+00 -3.30280811e-01
-1.72927435e-02 -2.35217378e-01 7.72050679e-01 3.68290335e-01
6.24533653e-01 -4.76179689e-01 4.73405793e-02 3.71357441e-01
7.32635081e-01 5.18222988e-01 8.17228436e-01 -8.80108476e-02
6.39373958e-01 4.36871886e-01 7.24643990e-02 -2.47810990e-01
-5.93186244e-02 7.78180838e-01 -1.83192343e-01 -2.32813120e-01
-4.64479119e-01 -1.24101527e-01 7.04236031e-01 1.46258140e+00
3.38546753e-01 -5.21324694e-01 -7.99482226e-01 1.02968860e+00
-1.15074265e+00 -9.74935055e-01 1.13319263e-01 2.30914593e+00
1.35941827e+00 1.89253971e-01 5.06628528e-02 1.06724069e-01
1.06235707e+00 4.91472781e-01 -1.61652848e-01 -8.28583062e-01
-4.23287213e-01 6.43240333e-01 3.15272748e-01 1.13490653e+00
-1.29163074e+00 1.13531530e+00 6.99638081e+00 7.43875325e-01
-1.34392524e+00 -6.71528429e-02 8.52959692e-01 -6.27135858e-02
-1.92875952e-01 -3.14638168e-01 -1.04787195e+00 5.74875414e-01
1.67911744e+00 -9.98284742e-02 5.59203148e-01 7.99307466e-01
2.16948465e-02 4.67298955e-01 -1.11235821e+00 5.16507983e-01
4.40578721e-02 -9.19382215e-01 -1.98915496e-01 -1.54981107e-01
7.54273891e-01 6.21147871e-01 2.82574922e-01 3.23571980e-01
6.00867569e-01 -1.19751382e+00 7.80667484e-01 -2.76429988e-02
6.84689343e-01 -8.91187310e-01 7.43751764e-01 1.55132599e-02
-8.20986032e-01 3.38266581e-01 -2.51684219e-01 8.95661190e-02
2.85016209e-01 3.55255604e-01 -1.31916177e+00 1.37391359e-01
2.46478781e-01 2.00503364e-01 -2.21005082e-01 4.40402091e-01
-3.38507414e-01 1.37309408e+00 -2.66582519e-01 6.02619275e-02
1.03019260e-01 5.00304587e-02 7.09214807e-01 1.70167875e+00
-3.82933915e-02 -1.90422669e-01 9.69625637e-02 5.28913736e-01
-2.60285914e-01 1.28819928e-01 -5.54752946e-01 -2.24355593e-01
6.83913946e-01 8.74498427e-01 -1.68088585e-01 -2.35466689e-01
-6.05404437e-01 1.10680234e+00 5.21878600e-01 4.70956743e-01
-3.41289490e-01 -3.46985340e-01 1.12949634e+00 -1.45060688e-01
5.33503652e-01 -2.49657497e-01 -8.56389105e-02 -7.16703832e-01
-2.79948592e-01 -1.06000447e+00 1.21325649e-01 -6.13196313e-01
-1.45738697e+00 1.13273370e+00 -5.97001135e-01 -6.02762401e-01
-6.62810326e-01 -5.49046278e-01 -5.79019070e-01 1.48575771e+00
-1.53321242e+00 -8.29705536e-01 4.67017472e-01 2.98690468e-01
1.02401674e+00 -3.97486418e-01 1.21502626e+00 2.98030704e-01
-5.47646224e-01 8.04104507e-01 4.08016592e-02 5.91768503e-01
7.97387660e-01 -1.66826069e+00 9.41982448e-01 9.83231068e-01
5.66905200e-01 5.09842575e-01 8.22750747e-01 -4.18701142e-01
-1.02920961e+00 -1.21327698e+00 1.34322715e+00 -5.34194648e-01
5.54324746e-01 -5.06077766e-01 -1.24898887e+00 1.06190634e+00
7.59222209e-01 -4.91067767e-02 7.93099642e-01 5.54367065e-01
-3.05567235e-01 -4.86193299e-02 -5.07949769e-01 6.33089006e-01
5.18054187e-01 -9.89211619e-01 -1.00749385e+00 2.96490639e-02
9.77683604e-01 -3.78085732e-01 -6.50472879e-01 7.53742903e-02
2.12932885e-01 -6.27528727e-01 7.27245033e-01 -8.24135780e-01
1.41921863e-01 -1.68397948e-01 -4.46467489e-01 -1.98645890e+00
-4.06008214e-01 -8.17137122e-01 2.78310120e-01 1.35917401e+00
8.84929538e-01 -3.72689039e-01 3.40178728e-01 3.54187250e-01
-6.86791003e-01 -1.92579716e-01 -9.39266324e-01 -8.84163141e-01
3.09881121e-01 -3.51970136e-01 4.42557395e-01 1.04986548e+00
1.16408989e-01 9.06971097e-01 -2.48061463e-01 4.93695647e-01
1.56864420e-01 -4.58964229e-01 3.73347580e-01 -7.79717207e-01
-2.89895922e-01 -4.32151765e-01 -4.26311612e-01 -1.08456051e+00
5.48669457e-01 -9.30853307e-01 5.44927239e-01 -8.54722023e-01
-3.45193595e-01 -3.96374345e-01 -4.24183160e-01 5.40632248e-01
-4.37452644e-01 9.36416909e-02 1.33324489e-01 -1.83576331e-01
-2.22064033e-01 5.75251400e-01 7.82353342e-01 -1.91586897e-01
-1.66749775e-01 1.44553259e-01 -5.03903270e-01 6.00008309e-01
8.65518093e-01 -4.00328249e-01 -3.30755085e-01 -6.01896763e-01
-4.52365994e-01 1.04865670e-01 -2.60207027e-01 -9.07516301e-01
6.63072169e-02 -1.55911222e-01 2.37898692e-01 -2.58334517e-01
4.23820496e-01 -5.10285795e-01 -9.58424583e-02 -7.19937235e-02
-7.39407539e-01 3.07891350e-02 3.28638166e-01 -1.41054010e-02
-5.09101331e-01 -2.28415325e-01 1.19795883e+00 9.58487093e-02
-4.76766437e-01 -7.79485628e-02 -6.79137468e-01 4.62460905e-01
3.70483637e-01 2.31440559e-01 -5.72618991e-02 -5.28292716e-01
-8.16414416e-01 -1.17430471e-01 2.21538737e-01 7.06610322e-01
2.56886959e-01 -1.19249785e+00 -1.21234763e+00 8.21394324e-01
-9.53705609e-02 -6.60148636e-02 1.14534959e-01 3.08257937e-01
-2.62980461e-02 4.55627471e-01 8.24755058e-02 -4.19613600e-01
-1.45013964e+00 4.06260997e-01 5.72351098e-01 1.64912611e-01
-3.17075223e-01 1.19914711e+00 4.18741435e-01 -5.94158769e-01
2.44045645e-01 -4.90274914e-02 1.29075088e-02 -1.54380381e-01
6.84362292e-01 -4.94652875e-02 3.66176963e-01 -1.17027247e+00
-4.69940394e-01 6.68370724e-03 -6.94218576e-01 -3.93922627e-01
1.04925847e+00 -2.43127376e-01 3.07765841e-01 7.24314451e-01
1.24023104e+00 6.55813217e-01 -1.26461041e+00 -1.94665864e-01
-4.64473739e-02 -1.08559459e-01 3.69753718e-01 -9.93458867e-01
-9.53600168e-01 7.63716936e-01 3.39821726e-01 5.28348565e-01
7.64259577e-01 2.81204879e-01 7.04972267e-01 1.60227448e-01
-2.83055931e-01 -1.17125249e+00 -3.03759992e-01 1.03861713e+00
9.21667337e-01 -1.16187227e+00 -8.89579833e-01 -9.43190455e-02
-9.46974277e-01 8.65828693e-01 2.84108371e-01 6.89465329e-02
5.21254241e-01 6.93822503e-01 7.87010491e-01 2.37161696e-01
-7.40862846e-01 -2.61420608e-01 3.07243586e-01 6.57295287e-01
1.01208663e+00 4.28769857e-01 2.80915082e-01 3.69258553e-01
-6.87349319e-01 -7.82769561e-01 5.66865802e-01 6.69413388e-01
-4.41033751e-01 -1.17990768e+00 -3.94608021e-01 4.81305458e-02
-7.34881520e-01 -5.40426970e-01 -4.35968727e-01 3.39254051e-01
-3.53200078e-01 1.14628613e+00 5.60914278e-01 -2.45757863e-01
2.63486326e-01 8.11923683e-01 1.33830205e-01 -7.69218862e-01
-3.59466016e-01 3.32848370e-01 4.63824779e-01 3.65361497e-02
2.64103711e-02 -8.96000981e-01 -1.23920202e+00 -3.17296176e-03
-2.75945932e-01 4.34252620e-01 7.34817684e-01 1.08378065e+00
2.59605020e-01 7.75845885e-01 9.19008255e-01 -4.89700705e-01
-6.13813996e-01 -1.29328239e+00 -8.41475487e-01 5.88067293e-01
9.02879775e-01 -1.84123755e-01 -4.61579382e-01 3.98938090e-01] | [14.357266426086426, 6.695898056030273] |
01a867b6-b509-4e3f-8a21-4abeac731a07 | semantic-role-labeling-guided-out-of | 2305.18026 | null | https://arxiv.org/abs/2305.18026v1 | https://arxiv.org/pdf/2305.18026v1.pdf | Semantic Role Labeling Guided Out-of-distribution Detection | Identifying unexpected domain-shifted instances in natural language processing is crucial in real-world applications. Previous works identify the OOD instance by leveraging a single global feature embedding to represent the sentence, which cannot characterize subtle OOD patterns well. Another major challenge current OOD methods face is learning effective low-dimensional sentence representations to identify the hard OOD instances that are semantically similar to the ID data. In this paper, we propose a new unsupervised OOD detection method, namely Semantic Role Labeling Guided Out-of-distribution Detection (SRLOOD), that separates, extracts, and learns the semantic role labeling (SRL) guided fine-grained local feature representations from different arguments of a sentence and the global feature representations of the full sentence using a margin-based contrastive loss. A novel self-supervised approach is also introduced to enhance such global-local feature learning by predicting the SRL extracted role. The resulting model achieves SOTA performance on four OOD benchmarks, indicating the effectiveness of our approach. Codes will be available upon acceptance. | ['Javen Qinfeng Shi', 'Ehsan Abbasnejad', 'Lingqiao Liu', 'Haiyao Cao', 'YuHao Lin', 'Yu Tian', 'Maihao Guo', 'Jinan Zou'] | 2023-05-29 | null | null | null | null | ['semantic-role-labeling'] | ['natural-language-processing'] | [ 4.34974343e-01 4.27889675e-01 -5.41266024e-01 -7.29710162e-01
-8.69453013e-01 -6.92927778e-01 7.59104252e-01 8.44070494e-01
1.90575402e-02 3.93552363e-01 8.16463351e-01 1.80430889e-01
-2.93247938e-01 -4.88323867e-01 -3.48852277e-01 -5.82922518e-01
-1.44577354e-01 4.16363031e-01 2.21764058e-01 -2.32464567e-01
4.53316242e-01 2.52728105e-01 -1.85225487e+00 7.72140920e-01
9.14744198e-01 1.16946089e+00 7.14751408e-02 5.24413168e-01
-5.50008416e-01 9.95159209e-01 -9.77456510e-01 1.51927218e-01
1.12822309e-01 -4.10057515e-01 -7.89945960e-01 2.92800695e-01
6.40478373e-01 -3.15075889e-02 -2.00901136e-01 9.63242650e-01
5.15749097e-01 8.78674313e-02 9.55539227e-01 -1.16810477e+00
-6.39448106e-01 3.31567198e-01 -3.73761863e-01 5.75416148e-01
5.99177420e-01 -1.23381190e-01 1.51028299e+00 -9.91685152e-01
8.91770601e-01 1.86186898e+00 1.65771082e-01 5.78622162e-01
-1.29833460e+00 -3.08558911e-01 2.77450532e-01 7.24487752e-02
-6.69859827e-01 -2.67860383e-01 1.20962143e+00 -4.02158499e-01
1.15897322e+00 2.53293179e-02 1.30945519e-01 1.04822266e+00
2.45577097e-01 1.23872900e+00 1.05730820e+00 -6.51241422e-01
4.26502228e-01 5.19507267e-02 8.12944233e-01 5.57725251e-01
3.05988491e-01 -3.80945474e-01 -9.33003962e-01 -5.27252913e-01
1.07475080e-01 5.50135560e-02 -1.55298412e-02 -6.19512439e-01
-9.61587667e-01 9.26663041e-01 3.00911635e-01 5.01204729e-01
-2.48552337e-01 -9.37060937e-02 7.43649244e-01 6.87640846e-01
9.60654318e-01 6.32305443e-01 -7.21318841e-01 -1.30555391e-01
-3.85799050e-01 3.55307490e-01 7.08041608e-01 6.71683371e-01
9.03739572e-01 -5.76434374e-01 -4.94284719e-01 1.01944232e+00
4.35615838e-01 7.82626569e-02 7.45640576e-01 -6.97421849e-01
6.98677301e-01 1.42334592e+00 -1.05040498e-01 -1.19422460e+00
-3.41531366e-01 -5.41800670e-02 -4.62444454e-01 -1.11582782e-02
2.23109037e-01 2.99209803e-01 -5.52868545e-01 1.50410640e+00
6.26749694e-01 -2.50168800e-01 3.02931190e-01 7.14997053e-01
7.87265897e-01 6.13073289e-01 2.32320979e-01 5.90783693e-02
1.63737559e+00 -7.48125613e-01 -9.16939735e-01 -5.81000149e-01
8.90083969e-01 -5.06145239e-01 1.35342932e+00 -2.32952069e-02
-5.23501456e-01 -3.79439145e-01 -1.26901829e+00 -2.40771487e-01
-3.70075405e-01 -7.03652501e-02 5.34457803e-01 2.11790845e-01
-4.49870884e-01 5.61516941e-01 -3.78692627e-01 -2.68726677e-01
5.94058454e-01 2.53078997e-01 -4.42851484e-01 -4.18823510e-01
-1.53582180e+00 5.34799039e-01 5.07780671e-01 -3.24027538e-01
-8.61758053e-01 -7.81894624e-01 -1.36510873e+00 4.44049686e-02
3.22136581e-01 -2.14721784e-01 1.03512800e+00 -1.02675855e+00
-9.23647583e-01 1.36073971e+00 -5.59466422e-01 -4.37710196e-01
1.26202956e-01 -3.79248053e-01 -2.73589641e-01 4.58483279e-01
6.55511916e-01 3.22795749e-01 1.46479726e+00 -1.28029180e+00
-3.50856900e-01 -7.51031518e-01 1.57251596e-01 3.66194576e-01
-6.61269724e-01 2.56403834e-01 1.52056059e-02 -8.08788240e-01
3.28349441e-01 -4.49904382e-01 -4.62631974e-03 5.27217565e-03
-5.14218926e-01 -1.06884325e+00 9.73591387e-01 -2.74232894e-01
1.22655630e+00 -2.53150988e+00 -7.91851152e-03 1.29501254e-03
5.33741951e-01 -8.17523152e-03 -2.38713205e-01 4.10303891e-01
-2.36720964e-01 7.97368214e-02 -3.85767847e-01 -1.89181730e-01
3.49611908e-01 1.29735216e-01 -4.74993467e-01 4.97524649e-01
7.27212071e-01 4.80634958e-01 -1.14415121e+00 -5.07149398e-01
-3.86035629e-02 6.43713921e-02 -3.21291178e-01 4.99600917e-01
-3.80852103e-01 -2.12487727e-02 -7.12325335e-01 7.25803912e-01
6.17604852e-01 -2.03419670e-01 3.64024907e-01 -1.54119693e-02
2.89379507e-01 8.05686712e-01 -1.20163584e+00 1.44787836e+00
-3.06473881e-01 5.04374504e-01 -1.93706974e-02 -1.17171133e+00
1.19574392e+00 2.34043136e-01 1.88544899e-01 -7.13151932e-01
-5.98537624e-02 3.37350905e-01 -1.64758086e-01 -4.75328058e-01
2.66457766e-01 -1.71762332e-01 -7.10507214e-01 6.66733861e-01
1.98638618e-01 -5.37977591e-02 1.57119453e-01 2.38604665e-01
1.34871531e+00 -1.92555085e-01 8.24946880e-01 -7.21119046e-01
7.02503085e-01 9.09886509e-02 6.49529457e-01 7.82446384e-01
-5.11324108e-01 4.52525765e-01 1.19968307e+00 -5.63285232e-01
-5.95737576e-01 -9.09111083e-01 -2.96893626e-01 1.25957727e+00
3.46752942e-01 -4.49027359e-01 -6.36120677e-01 -1.41819811e+00
4.55096811e-01 6.90458417e-01 -6.33332551e-01 -4.27857459e-01
-6.92543268e-01 -6.39039040e-01 3.56189400e-01 2.81708360e-01
1.71483904e-01 -1.27034736e+00 -2.50117302e-01 2.52379447e-01
1.04935311e-01 -9.65810418e-01 -4.23269868e-01 6.34331405e-01
-8.59968126e-01 -1.11268425e+00 -3.49399030e-01 -1.12192833e+00
7.20099628e-01 3.48212063e-01 1.12348330e+00 -1.16625905e-01
-4.24832612e-01 2.02363417e-01 -3.74621004e-01 -2.70286739e-01
-6.48514867e-01 -4.32636067e-02 2.11055249e-01 2.50819325e-01
9.35221076e-01 -1.57534629e-01 -3.80778372e-01 -9.20859165e-03
-8.99874508e-01 -6.72651172e-01 5.04000843e-01 9.69358444e-01
6.40229523e-01 3.44935089e-01 7.71326184e-01 -1.20485866e+00
1.03696620e+00 -6.81486487e-01 -8.23576376e-03 4.91141155e-02
-5.47844231e-01 5.33540368e-01 7.54418433e-01 -1.99000835e-01
-1.21252596e+00 -1.42702714e-01 2.42263943e-01 -2.15925932e-01
-5.52959681e-01 1.91774204e-01 -5.48347414e-01 6.18654251e-01
7.11568773e-01 2.15640023e-01 3.78241725e-02 -7.05275059e-01
8.79731551e-02 8.44361067e-01 -1.87029466e-02 -7.28994668e-01
7.66633391e-01 4.15508211e-01 -3.12590182e-01 -1.06957579e+00
-1.56354797e+00 -8.91339302e-01 -5.72844744e-01 2.58986980e-01
7.65897393e-01 -1.11414123e+00 -1.83694899e-01 4.44618732e-01
-1.22127366e+00 1.43203260e-02 -3.81696135e-01 1.22512870e-01
-4.38079655e-01 4.96522218e-01 -4.38020974e-01 -6.02287352e-01
-1.00531094e-01 -6.10769212e-01 1.54298592e+00 3.76998447e-02
-5.82894564e-01 -9.81387138e-01 1.97239578e-01 4.74476188e-01
-1.76296458e-01 1.46512628e-01 1.46472871e+00 -1.23835313e+00
-3.05315644e-01 -3.05522144e-01 -2.91373402e-01 5.14803231e-01
3.76094848e-01 -6.09222770e-01 -9.90344763e-01 -1.06284969e-01
2.35138953e-01 -6.48260593e-01 8.66364121e-01 1.53745666e-01
1.11232615e+00 -2.05537543e-01 -2.44494155e-01 -6.58004731e-02
1.13409758e+00 -1.95313901e-01 2.45155945e-01 4.13770378e-01
5.43044388e-01 9.84009683e-01 1.18408203e+00 4.75622982e-01
1.66552827e-01 5.58668256e-01 2.90246606e-01 1.23992033e-01
-1.43401951e-01 -2.70232558e-01 6.93778455e-01 4.19846535e-01
5.86912632e-01 -1.99361235e-01 -8.34707797e-01 6.72458887e-01
-1.81851578e+00 -8.49673510e-01 -7.77799785e-02 1.70321250e+00
9.66934800e-01 5.81104696e-01 -1.52683228e-01 4.34066623e-01
8.78674030e-01 6.43264115e-01 -6.85781598e-01 -5.64315736e-01
-1.02345802e-01 9.78760570e-02 -1.57445669e-01 3.22640121e-01
-1.55336607e+00 8.28305244e-01 5.20829201e+00 7.73553669e-01
-4.96346176e-01 1.41598850e-01 5.49952567e-01 2.57053077e-01
-4.85405564e-01 -1.91175461e-01 -1.11848342e+00 3.43843818e-01
6.33537412e-01 -5.04266098e-02 -2.40763947e-01 1.31799400e+00
3.20702255e-01 1.03398368e-01 -1.13474953e+00 7.44522989e-01
3.80455554e-01 -1.09643865e+00 1.65446341e-01 2.80105360e-02
6.06260836e-01 -4.30863589e-01 -2.05562934e-01 2.95159787e-01
1.80169684e-03 -7.46711075e-01 6.53264642e-01 2.21075583e-02
4.56640601e-01 -6.23589516e-01 7.86013484e-01 3.66482407e-01
-1.10220897e+00 -3.62385482e-01 -5.45928895e-01 -1.77146584e-01
-5.10938644e-01 9.99295235e-01 -1.09939444e+00 1.06384456e-01
6.46814525e-01 1.14858747e+00 -6.94738328e-01 3.47284377e-01
-6.14059091e-01 6.49806082e-01 8.68359730e-02 -3.88116062e-01
1.89839169e-01 3.09863001e-01 8.62748802e-01 1.10275710e+00
-1.90636739e-01 -2.55043924e-01 2.40767777e-01 8.55000079e-01
-3.03421557e-01 1.85471296e-01 -8.90390098e-01 -4.10955817e-01
4.83280957e-01 1.07296145e+00 -5.21703422e-01 -1.76415563e-01
-2.54263610e-01 1.10532749e+00 4.80255723e-01 -1.74770057e-01
-3.38912904e-01 -7.32356608e-01 1.18086815e+00 1.41589150e-01
2.52624571e-01 6.82926551e-02 -1.67022884e-01 -1.39755857e+00
3.02000076e-01 -9.27606106e-01 7.20056057e-01 -1.17605425e-01
-1.92821944e+00 3.40896428e-01 -1.11699715e-01 -1.40352142e+00
-2.23550782e-01 -8.45294118e-01 -6.49657190e-01 5.02575159e-01
-1.81982422e+00 -9.28745389e-01 -1.27538875e-01 3.79018366e-01
1.05133939e+00 -2.08018824e-01 8.17713916e-01 -2.38065273e-02
-4.03260082e-01 5.76663435e-01 2.47504897e-02 2.99650043e-01
8.41198444e-01 -1.72718036e+00 1.32012680e-01 5.80910623e-01
-1.65508240e-01 7.47843504e-01 6.58860505e-01 -5.91683388e-01
-1.24701452e+00 -1.13962710e+00 1.68124270e+00 -5.97966611e-01
6.40876710e-01 -7.82005548e-01 -9.50671554e-01 2.42622927e-01
-2.66620591e-02 1.75069705e-01 7.71283388e-01 1.99526817e-01
-5.85978091e-01 -1.19221605e-01 -1.27372265e+00 2.46996820e-01
1.08723640e+00 -8.14076722e-01 -1.45180428e+00 4.62885529e-01
1.01523781e+00 1.59205515e-02 -7.58459985e-01 1.00380495e-01
-6.65999427e-02 -9.04419839e-01 1.06574297e+00 -7.92145669e-01
7.43171692e-01 -3.02353919e-01 -3.42711389e-01 -1.05828118e+00
-1.89724192e-01 -4.56927299e-01 -4.49264526e-01 1.44573569e+00
1.87861413e-01 -5.80601752e-01 6.69774652e-01 3.38055879e-01
-1.47457033e-01 -6.75760329e-01 -9.31529284e-01 -7.90735245e-01
-1.34652644e-01 -2.29859158e-01 4.76747870e-01 9.70181882e-01
1.37178674e-01 7.66299307e-01 -2.46812869e-02 2.14462668e-01
8.37017655e-01 4.47290540e-01 4.18772697e-01 -1.49479020e+00
-2.17420682e-01 -1.11268699e-01 -7.04425037e-01 -1.08503771e+00
7.28883564e-01 -1.09215307e+00 1.09739885e-01 -1.33694923e+00
2.55074739e-01 -2.75528133e-01 -4.29860502e-01 4.17812884e-01
-3.85132730e-01 -1.28185436e-01 -2.22018361e-02 2.54343361e-01
-8.60369861e-01 8.09131861e-01 1.10278523e+00 -5.18454552e-01
-2.20684230e-01 -1.90699622e-01 -9.92054403e-01 9.00709331e-01
5.68199396e-01 -1.01736605e+00 -4.96118665e-01 -8.81749466e-02
-2.76966691e-02 -3.44969541e-01 3.29129398e-01 -7.18693256e-01
-3.45001757e-01 -5.20997390e-04 3.85791242e-01 -5.01289666e-01
1.82266086e-01 -5.96493244e-01 -1.03976488e+00 5.57185054e-01
-8.39632869e-01 -3.90684068e-01 -1.90234944e-01 1.06500530e+00
-6.24484420e-01 -5.06238818e-01 6.24380469e-01 -1.81565642e-01
-1.07879233e+00 7.41911726e-03 -3.66942704e-01 6.87611818e-01
9.82889473e-01 -9.75204557e-02 -3.47836167e-01 -4.88998229e-03
-4.39943761e-01 1.08427227e-01 1.81240335e-01 7.18068421e-01
7.21094310e-01 -1.24480724e+00 -6.94017947e-01 4.57887352e-01
7.79162467e-01 2.72330903e-02 1.21290937e-01 2.14795172e-01
-1.22147277e-01 3.97014141e-01 2.31829613e-01 -4.49356198e-01
-1.34233022e+00 3.86331499e-01 1.00950494e-01 -5.88730752e-01
-6.10543847e-01 7.94770718e-01 4.00156081e-01 -8.35890651e-01
9.54638198e-02 -1.70838714e-01 -5.05047917e-01 5.39218724e-01
5.75224102e-01 2.87426002e-02 7.10337907e-02 -5.18051565e-01
-6.61678910e-01 3.06268573e-01 -3.00925016e-01 4.17677283e-01
1.48675287e+00 -1.66820705e-01 -3.64833802e-01 6.50234520e-01
1.68615961e+00 5.73583841e-02 -1.14030349e+00 -4.70596969e-01
6.58934295e-01 -6.58731461e-01 -1.06210776e-01 -4.39758897e-01
-2.47533500e-01 9.98940468e-01 5.28856754e-01 5.19647002e-01
8.29730093e-01 7.16069639e-01 7.08821416e-01 4.99370903e-01
-1.98401678e-02 -1.23537457e+00 7.66102374e-01 7.13112831e-01
9.42038536e-01 -1.36517251e+00 -2.19195083e-01 -5.23589015e-01
-7.80281484e-01 1.23039031e+00 7.53985524e-01 -4.20876771e-01
6.56507790e-01 -7.28038549e-02 -4.81481776e-02 -6.04121864e-01
-8.28285277e-01 -2.43234751e-03 3.80418748e-01 5.54467976e-01
4.05741781e-01 -4.64702249e-02 -2.06611037e-01 8.66701663e-01
2.64969558e-01 -7.01414049e-01 3.70336622e-01 1.15154421e+00
-6.95116580e-01 -1.17520463e+00 -1.95922092e-01 6.61147356e-01
-3.70555252e-01 -1.19386956e-01 -8.10829163e-01 5.50495923e-01
3.50826755e-02 8.15770268e-01 1.65081784e-01 -1.70660857e-02
3.50428939e-01 5.46115756e-01 6.54020756e-02 -1.27875197e+00
-4.40731823e-01 -1.21199980e-01 2.77219623e-01 -5.94628632e-01
-4.05517071e-01 -8.91273260e-01 -1.47086418e+00 3.26189458e-01
7.42716268e-02 -1.75750982e-02 2.65311092e-01 9.14403439e-01
6.86070323e-01 4.27622706e-01 1.02533901e+00 -5.68258107e-01
-1.07631743e+00 -7.73315489e-01 -7.72274137e-01 9.98077929e-01
6.31356001e-01 -7.56762385e-01 -8.50399494e-01 -2.09854186e-01] | [12.159680366516113, 7.499963283538818] |
b06a2c7f-3bf0-4568-9ef5-0b90dbf6208d | constrained-pure-exploration-multi-armed | 2211.14768 | null | https://arxiv.org/abs/2211.14768v1 | https://arxiv.org/pdf/2211.14768v1.pdf | Constrained Pure Exploration Multi-Armed Bandits with a Fixed Budget | We consider a constrained, pure exploration, stochastic multi-armed bandit formulation under a fixed budget. Each arm is associated with an unknown, possibly multi-dimensional distribution and is described by multiple attributes that are a function of this distribution. The aim is to optimize a particular attribute subject to user-defined constraints on the other attributes. This framework models applications such as financial portfolio optimization, where it is natural to perform risk-constrained maximization of mean return. We assume that the attributes can be estimated using samples from the arms' distributions and that these estimators satisfy suitable concentration inequalities. We propose an algorithm called \textsc{Constrained-SR} based on the Successive Rejects framework, which recommends an optimal arm and flags the instance as being feasible or infeasible. A key feature of this algorithm is that it is designed on the basis of an information theoretic lower bound for two-armed instances. We characterize an instance-dependent upper bound on the probability of error under \textsc{Constrained-SR}, that decays exponentially with respect to the budget. We further show that the associated decay rate is nearly optimal relative to an information theoretic lower bound in certain special cases. | ['Jayakrishnan Nair', 'Fathima Zarin Faizal'] | 2022-11-27 | null | null | null | null | ['portfolio-optimization'] | ['time-series'] | [ 2.96639919e-01 2.51332283e-01 -8.07443082e-01 -3.38768780e-01
-1.19444215e+00 -8.87818992e-01 2.18780324e-01 1.86028987e-01
-5.25885463e-01 9.53543544e-01 -1.14046685e-01 -5.48711717e-01
-9.43995178e-01 -8.03095639e-01 -7.36204088e-01 -8.98617923e-01
-8.55306089e-02 1.21463430e+00 -3.52167040e-01 3.21086466e-01
4.32821423e-01 6.32882774e-01 -1.13230419e+00 5.33451848e-02
6.22459590e-01 1.47859180e+00 -2.39940628e-01 6.42175555e-01
-6.95478767e-02 2.20180482e-01 -5.35877883e-01 -6.48385346e-01
4.81070459e-01 -4.73564148e-01 -6.34500682e-01 4.03467774e-01
-3.68183494e-01 -2.98212599e-02 3.98169696e-01 1.17066550e+00
1.44479379e-01 1.49443001e-01 1.05776548e+00 -1.22093511e+00
-2.51296833e-02 7.21199572e-01 -7.84991443e-01 2.10117459e-01
9.94562637e-04 -2.44766042e-01 1.16549289e+00 -2.99021274e-01
1.44720942e-01 9.81877506e-01 3.09580088e-01 3.22328746e-01
-1.52912271e+00 -4.75866556e-01 6.33372009e-01 -3.84810030e-01
-1.04379106e+00 -3.69520217e-01 4.14182961e-01 -4.14724231e-01
2.42903203e-01 6.48784399e-01 4.43423331e-01 8.17516565e-01
-7.49926791e-02 9.02790189e-01 1.05757046e+00 -6.80308282e-01
7.86283970e-01 4.98962849e-01 3.03174764e-01 1.95783332e-01
6.83754265e-01 2.94687241e-01 -3.37854981e-01 -5.50163448e-01
3.60514015e-01 -2.27974709e-02 -2.21608266e-01 -5.90105772e-01
-7.01305449e-01 1.19412875e+00 -2.81068504e-01 -1.62813917e-01
-5.88194609e-01 -1.04061235e-02 8.61510634e-03 3.50276589e-01
7.74274111e-01 3.03687066e-01 -5.61800361e-01 1.34111360e-01
-8.95710111e-01 3.12425524e-01 1.02182019e+00 1.02522373e+00
3.15888435e-01 -2.60532588e-01 -5.37224770e-01 5.18275559e-01
2.42079794e-01 6.40219808e-01 -9.31056738e-02 -8.96579444e-01
9.32755828e-01 2.78759301e-02 9.47416246e-01 -3.60853255e-01
-2.89261758e-01 -9.45569038e-01 -3.72101367e-01 1.15850762e-01
5.65915763e-01 -2.46622935e-01 -6.57400072e-01 1.91440761e+00
2.84574240e-01 -2.01478258e-01 -1.02812596e-01 7.39216924e-01
-3.82319927e-01 4.77525830e-01 -1.28358558e-01 -9.23530102e-01
9.64590311e-01 -5.08650959e-01 -6.49858654e-01 -2.99537510e-01
3.04441631e-01 -3.89138192e-01 6.38851643e-01 8.84528637e-01
-1.30078423e+00 4.27200705e-01 -9.73185718e-01 8.81832242e-01
7.16143698e-02 -5.51359951e-02 6.51904225e-01 1.40176404e+00
-4.59542394e-01 3.53857458e-01 -6.50992632e-01 2.26153865e-01
4.50349867e-01 4.77380335e-01 7.61092529e-02 4.31088805e-02
-6.79286778e-01 5.32040119e-01 4.15355384e-01 2.29257062e-01
-7.75254071e-01 -4.20314223e-01 -4.46928442e-01 1.90626264e-01
6.95410550e-01 -6.55231118e-01 1.35405481e+00 -1.01206160e+00
-1.38794804e+00 4.50471759e-01 -9.19517055e-02 -5.88064432e-01
8.65657866e-01 6.59147352e-02 -2.03940660e-01 3.54952402e-02
1.82651594e-01 -4.72003818e-01 9.45591748e-01 -1.08266234e+00
-9.29696202e-01 -7.30372787e-01 6.63844571e-02 1.15297481e-01
-2.10478485e-01 5.38561717e-02 -2.55893677e-01 -6.25392556e-01
2.75613576e-01 -1.07321203e+00 -5.88400841e-01 -4.25754666e-01
-6.30700648e-01 1.07549153e-01 4.68913838e-02 -1.49303079e-01
1.21186924e+00 -1.73670244e+00 6.40636608e-02 8.72338891e-01
-5.40036500e-01 -3.39002639e-01 -1.29821068e-02 2.89221734e-01
5.32208383e-02 3.36183459e-01 -3.29049796e-01 -3.68497014e-01
1.36836767e-01 -4.85455384e-03 -4.44156975e-01 1.03043461e+00
-1.10480726e-01 2.30442032e-01 -5.26744545e-01 -9.83048305e-02
-1.39628813e-01 -3.40627819e-01 -6.71705842e-01 4.71756682e-02
-4.33383584e-01 8.29991326e-02 -1.02716362e+00 6.33971453e-01
7.74549544e-01 -2.02146515e-01 3.19729507e-01 3.68978828e-01
-3.19278911e-02 -1.21701688e-01 -1.58043838e+00 9.79632914e-01
-4.41546082e-01 -8.89613628e-02 2.96409160e-01 -1.56279349e+00
4.24637824e-01 2.11545885e-01 2.80899197e-01 -1.93933044e-02
2.86826372e-01 3.19538087e-01 -3.52470726e-01 -1.24378987e-01
2.19399571e-01 -5.64267755e-01 -3.66814643e-01 8.05848718e-01
-2.39836544e-01 8.35435987e-02 4.51779924e-02 2.56167911e-02
7.99465239e-01 -2.36361980e-01 2.76777178e-01 -4.27116662e-01
1.15083210e-01 -3.29186827e-01 5.50749600e-01 1.39559281e+00
2.72880256e-01 4.49955136e-01 8.44734788e-01 -1.31367743e-01
-6.51700139e-01 -9.44718957e-01 -3.98752719e-01 9.13019538e-01
-1.18920676e-01 3.91403437e-01 -5.12796044e-01 -7.43201911e-01
4.51867312e-01 9.80708659e-01 -8.77646267e-01 1.42745778e-01
1.97823390e-01 -1.27664065e+00 -2.64341325e-01 1.06784008e-01
1.51330754e-01 -4.22343224e-01 -5.53690851e-01 4.46716249e-01
1.65418386e-01 -7.10222304e-01 -3.77621651e-01 4.31859910e-01
-8.59918594e-01 -9.78065670e-01 -8.02661538e-01 -2.78088134e-02
7.03876853e-01 -1.03883415e-01 9.29068565e-01 -4.96878147e-01
1.39302179e-01 6.61741376e-01 -2.30212823e-01 -7.68547237e-01
-7.18767196e-02 1.09798841e-01 7.26752132e-02 5.16777933e-01
9.17094648e-02 -2.20346168e-01 -5.71025491e-01 3.50581557e-01
-8.13176215e-01 -4.94602948e-01 5.00929356e-01 8.34451914e-01
7.95549691e-01 1.97256595e-01 7.40867317e-01 -1.18212557e+00
5.91371953e-01 -7.20557570e-01 -1.18295550e+00 7.10057199e-01
-7.92274594e-01 2.93873280e-01 1.71514377e-01 -6.04309201e-01
-1.09979439e+00 1.32008433e-01 6.39280677e-01 -1.21815704e-01
1.98373750e-01 7.42805421e-01 -3.63090724e-01 5.09596542e-02
2.82202631e-01 -2.98181549e-03 -1.74619108e-01 -5.47461510e-01
1.72136620e-01 8.00550878e-01 2.46329844e-01 -1.09871924e+00
4.47599947e-01 3.30689073e-01 3.48786801e-01 -3.76406163e-01
-1.39810014e+00 -3.12235296e-01 1.52227115e-02 -7.70935863e-02
1.73397914e-01 -5.41286290e-01 -1.13340938e+00 -8.92634690e-02
-7.42954612e-01 -1.02132015e-01 -4.36379641e-01 8.14988494e-01
-1.02344453e+00 9.30866227e-03 -9.15548727e-02 -2.03023767e+00
-5.03830463e-02 -9.34164464e-01 5.77179909e-01 3.58968973e-02
6.63588941e-02 -7.49030590e-01 -1.40428901e-01 4.37440604e-01
1.68538854e-01 2.14588791e-01 9.32620049e-01 -1.06149042e+00
-6.05407536e-01 -6.96259141e-01 7.03084990e-02 2.16406763e-01
-1.91319004e-01 -3.78579587e-01 -4.72382516e-01 -3.88515621e-01
4.33110058e-01 -2.05158979e-01 9.16663647e-01 9.46769774e-01
1.60210037e+00 -8.21054339e-01 -4.48328882e-01 4.53942358e-01
1.30601549e+00 4.27696824e-01 9.20202658e-02 6.64164841e-01
-3.17833185e-01 7.13978708e-01 9.03046310e-01 1.15206778e+00
-1.63118377e-01 5.49024522e-01 5.63239396e-01 4.85848606e-01
1.10610259e+00 1.10826284e-01 -7.67334327e-02 -2.27053329e-01
1.39022306e-01 -6.39003217e-01 -5.25284350e-01 6.91385269e-01
-1.90859151e+00 -9.13074374e-01 2.32339412e-01 2.93228388e+00
8.16216707e-01 4.13960963e-01 5.35241246e-01 1.99441344e-01
9.42666650e-01 -3.72379392e-01 -1.00496912e+00 -5.11318445e-01
-1.22688018e-01 3.06051373e-01 1.14391840e+00 6.46142423e-01
-9.91778672e-01 4.41177152e-02 6.41009998e+00 8.05769920e-01
-2.92429209e-01 -5.13704643e-02 1.21819615e+00 -5.75166285e-01
-5.83281398e-01 -3.59779596e-03 -9.75707352e-01 6.71122491e-01
1.10937035e+00 -4.73765939e-01 4.27052885e-01 8.76879156e-01
1.00189060e-01 -3.09285283e-01 -1.22575116e+00 4.03898716e-01
-3.17902774e-01 -1.28093827e+00 -1.35488659e-01 5.32276571e-01
8.23681712e-01 -5.64113915e-01 6.07252538e-01 5.02418540e-03
6.22103035e-01 -8.05044651e-01 9.42290127e-01 6.36655331e-01
8.65714550e-01 -1.44645536e+00 6.26986504e-01 5.54016650e-01
-5.75921535e-01 -5.41831076e-01 -1.97849616e-01 1.61556616e-01
1.30409189e-02 8.49967122e-01 -3.28643531e-01 5.36409140e-01
3.02288949e-01 -6.67684898e-02 3.59016120e-01 1.29641879e+00
3.56601812e-02 6.81138277e-01 -6.19809926e-01 -1.80656791e-01
2.95133024e-01 -5.70322216e-01 5.48442423e-01 9.40307438e-01
5.69999754e-01 3.76171708e-01 1.87260389e-01 6.55796528e-01
6.75174668e-02 1.11658901e-01 -2.15492517e-01 -7.39299357e-02
7.22577095e-01 7.72673130e-01 -7.55055606e-01 -1.84701532e-01
-2.92052269e-01 4.98130798e-01 -1.08126305e-01 5.34944475e-01
-5.78107476e-01 -2.62552410e-01 4.26700950e-01 -1.01233281e-01
5.75405955e-01 3.17462116e-01 -4.79214817e-01 -7.44642973e-01
2.35093385e-01 -6.49348855e-01 9.05023694e-01 -2.84810197e-02
-1.35013747e+00 1.96925431e-01 1.99902862e-01 -9.48229909e-01
-3.97985876e-01 -5.96663356e-01 -2.33776957e-01 1.17997050e+00
-1.28698432e+00 -5.93158960e-01 5.69315791e-01 4.85509753e-01
2.84597665e-01 -3.53846550e-01 6.97411954e-01 -3.88591021e-01
-7.04971313e-01 6.74153626e-01 7.54021108e-01 -4.43426400e-01
-1.03118882e-01 -1.30328202e+00 -2.60732710e-01 6.28243744e-01
-2.63646129e-04 3.78294021e-01 1.02568126e+00 -5.27381241e-01
-1.20178902e+00 -9.99319494e-01 3.54970753e-01 -2.76178658e-01
7.50174761e-01 1.19869243e-02 -3.30975294e-01 8.36582959e-01
-3.84562641e-01 1.11950390e-01 8.54929388e-01 3.98141801e-01
-1.02961190e-01 -2.55796134e-01 -1.55004060e+00 1.62292346e-01
6.68508649e-01 -2.53549181e-02 -1.96382076e-01 7.09464192e-01
4.14090335e-01 -1.94705218e-01 -6.32959068e-01 2.71206170e-01
5.30761778e-01 -6.97733939e-01 7.60801554e-01 -1.13694048e+00
6.33428851e-03 1.90603569e-01 -3.93461496e-01 -1.14429772e+00
-1.97766975e-01 -1.09427476e+00 -2.75286227e-01 1.02801991e+00
7.91437984e-01 -8.84049654e-01 1.16113830e+00 1.07672215e+00
3.37762207e-01 -7.45443344e-01 -1.38784790e+00 -1.08151221e+00
1.25090376e-01 -5.11185169e-01 6.25686526e-01 4.24409032e-01
-2.51283925e-02 -8.11334848e-02 -4.48150516e-01 3.71102750e-01
1.12756288e+00 5.68429172e-01 2.54396945e-01 -1.32290566e+00
-6.92661047e-01 -4.08265471e-01 1.29355028e-01 -9.63046670e-01
2.98360646e-01 -4.63521302e-01 -5.90758696e-02 -9.19250190e-01
4.41545069e-01 -7.43062615e-01 -6.79968119e-01 6.21024817e-02
1.93523869e-01 -3.56287062e-01 7.17678247e-03 -9.43072662e-02
-5.90027630e-01 3.71171445e-01 6.36182487e-01 -7.41828531e-02
-2.63038844e-01 1.22990453e+00 -9.83063936e-01 5.13845742e-01
7.74771273e-01 -5.95222592e-01 -3.91907543e-01 4.06763032e-02
4.48036492e-01 8.78331900e-01 -1.16336226e-01 -3.22965652e-01
7.30032325e-02 -7.42923677e-01 2.72329032e-01 -8.79312456e-01
3.41874182e-01 -7.80070961e-01 2.99210876e-01 5.16852260e-01
-9.36477602e-01 -2.58812249e-01 -2.13735953e-01 1.08977222e+00
2.90077001e-01 -7.94603407e-01 7.17292845e-01 -1.19303256e-01
3.71896148e-01 3.23412061e-01 -3.74970853e-01 4.59881090e-02
1.10924041e+00 1.95028931e-02 9.41960588e-02 -7.27813840e-01
-8.72606218e-01 3.39004129e-01 -3.25672068e-02 -3.27187419e-01
3.63864750e-01 -1.10676575e+00 -7.00489640e-01 -1.11155584e-02
-4.35747057e-02 -1.94059879e-01 1.77080352e-02 6.34758949e-01
2.83740252e-01 6.42966747e-01 5.03999352e-01 -2.34104753e-01
-8.16915929e-01 8.41891468e-01 2.23103642e-01 -3.23216230e-01
-1.54941417e-02 9.68990982e-01 1.04723416e-01 1.91062570e-01
3.65812629e-01 -9.62021500e-02 6.78682029e-02 1.97941378e-01
5.52047789e-01 4.52076733e-01 2.29039222e-01 -3.87558900e-02
-2.03724280e-01 2.22005084e-01 -3.23744655e-01 -7.26753950e-01
1.37888229e+00 -2.40506619e-01 -1.02993764e-03 2.85437822e-01
7.78146982e-01 7.53669143e-02 -1.24635911e+00 -4.59785551e-01
4.21649963e-01 -7.87102938e-01 2.57624328e-01 -1.01777959e+00
-1.09131956e+00 2.72548825e-01 4.61686492e-01 5.44998348e-01
1.06127894e+00 6.92788465e-03 1.41339470e-02 3.59630585e-01
6.10904396e-01 -1.16497207e+00 -3.86757880e-01 5.32030612e-02
7.76496649e-01 -8.63129914e-01 1.64846018e-01 -1.24896340e-01
-4.76776510e-01 1.00563455e+00 -4.47711535e-02 1.23341963e-01
6.36670470e-01 2.73358911e-01 -6.36969566e-01 3.01713139e-01
-7.91317999e-01 -1.80421367e-01 2.97085285e-01 3.94146442e-01
-3.71765159e-02 3.31683099e-01 -5.23934662e-01 1.03762078e+00
-6.93340227e-02 -8.12819302e-02 5.16275525e-01 6.72625363e-01
-6.06338263e-01 -1.03077435e+00 -6.35592341e-01 8.88974071e-01
-1.02687526e+00 1.40286595e-01 -6.43676147e-02 6.06939971e-01
-4.80989695e-01 1.04714978e+00 2.99537890e-02 3.89189035e-01
3.23074311e-01 -1.59949511e-01 5.76149940e-01 -5.98497570e-01
-3.60062085e-02 2.67779380e-01 3.36670041e-01 -1.75456926e-01
-2.09921435e-01 -1.08564878e+00 -3.69684279e-01 -8.81828964e-02
-6.69555008e-01 7.02291012e-01 7.39118934e-01 9.21162903e-01
-6.60503209e-02 5.62706813e-02 1.20168686e+00 -5.03650188e-01
-1.48457491e+00 -6.41523898e-01 -1.05795050e+00 -1.44618228e-01
4.99530703e-01 -7.55701959e-01 -8.12784851e-01 -4.92181182e-01] | [4.551568984985352, 3.3016715049743652] |
68cb2912-adce-4784-83db-1bfda72818ca | end-to-end-learning-of-multi-scale | 1906.10399 | null | https://arxiv.org/abs/1906.10399v1 | https://arxiv.org/pdf/1906.10399v1.pdf | End-to-End Learning of Multi-scale Convolutional Neural Network for Stereo Matching | Deep neural networks have shown excellent performance in stereo matching task. Recently CNN-based methods have shown that stereo matching can be formulated as a supervised learning task. However, less attention is paid on the fusion of contextual semantic information and details. To tackle this problem, we propose a network for disparity estimation based on abundant contextual details and semantic information, called Multi-scale Features Network (MSFNet). First, we design a new structure to encode rich semantic information and fine-grained details by fusing multi-scale features. And we combine the advantages of element-wise addition and concatenation, which is conducive to merge semantic information with details. Second, a guidance mechanism is introduced to guide the network to automatically focus more on the unreliable regions. Third, we formulate the consistency check as an error map, obtained by the low stage features with fine-grained details. Finally, we adopt the consistency checking between the left feature and the synthetic left feature to refine the initial disparity. Experiments on Scene Flow and KITTI 2015 benchmark demonstrated that the proposed method can achieve the state-of-the-art performance. | ['Quanhong Wang', 'Li Zhang', 'Yong Zhao', 'Haihua Lu'] | 2019-06-25 | null | null | null | null | ['stereo-matching'] | ['computer-vision'] | [ 6.22184426e-02 -4.68035311e-01 7.06323907e-02 -7.86707401e-01
-3.60700577e-01 7.88329393e-02 4.93472844e-01 -1.59488484e-01
-5.28398156e-01 7.04192162e-01 5.40442109e-01 1.03886411e-01
-1.46216035e-01 -1.01446748e+00 -5.83752453e-01 -7.15117753e-01
3.96204084e-01 -1.09263696e-01 5.54656148e-01 -3.55537236e-01
4.87507313e-01 2.87070990e-01 -1.80420327e+00 2.00264946e-01
1.17151964e+00 1.30196238e+00 3.77306640e-01 1.08891807e-01
-3.16255897e-01 7.10879385e-01 -1.91543922e-01 -1.96379259e-01
4.10870910e-01 -3.28520030e-01 -5.85649908e-01 1.29213974e-01
6.42983198e-01 -5.68187773e-01 -5.34899116e-01 1.25551426e+00
7.13678777e-01 7.84617364e-02 2.16205716e-01 -1.20093644e+00
-2.21558735e-01 1.92001969e-01 -6.30531013e-01 4.02130067e-01
1.63758412e-01 3.47992003e-01 6.94753945e-01 -8.59634697e-01
4.68328714e-01 1.47198570e+00 4.59638596e-01 1.94456518e-01
-7.42207050e-01 -8.31311285e-01 3.76815706e-01 4.28158075e-01
-1.25182259e+00 -4.97336835e-01 1.00003076e+00 -4.36865449e-01
5.59748292e-01 -7.37256482e-02 7.32508898e-01 5.32884061e-01
7.83311799e-02 7.34620512e-01 1.10684788e+00 -1.10488005e-01
-8.57969671e-02 -1.35345876e-01 -3.31268385e-02 8.68862331e-01
2.98920900e-01 4.30101454e-01 -4.99540508e-01 3.49241942e-01
9.42430675e-01 2.35231429e-01 -4.44171876e-01 -2.34590918e-01
-1.24680412e+00 7.43567288e-01 9.99343574e-01 2.88108498e-01
-3.72638822e-01 1.32184952e-01 4.98805732e-01 -3.20139043e-02
4.49622691e-01 5.83308376e-02 -3.66702616e-01 1.56093743e-02
-7.78868437e-01 9.14653689e-02 1.26469553e-01 8.37924004e-01
1.38851237e+00 -7.33581707e-02 -3.33573699e-01 9.32070673e-01
3.32388788e-01 3.86002004e-01 3.80101919e-01 -1.06065929e+00
8.10663223e-01 8.11463714e-01 -1.06225513e-01 -1.33327878e+00
-3.42668861e-01 -5.17341435e-01 -1.17263019e+00 3.23747218e-01
1.87557623e-01 -9.71769616e-02 -9.33468282e-01 1.71770179e+00
3.66678745e-01 4.18280631e-01 -9.35542360e-02 1.27604795e+00
1.11380553e+00 6.08722806e-01 -3.10864542e-02 1.09409299e-02
1.14241481e+00 -1.06462741e+00 -5.73503852e-01 -2.05716670e-01
2.35978737e-01 -7.84201264e-01 6.06924236e-01 2.17210054e-02
-1.01997757e+00 -1.03495800e+00 -1.16397893e+00 -3.17975372e-01
-2.93171108e-01 -4.70359959e-02 7.84821033e-01 2.04118177e-01
-9.68900919e-01 6.21777236e-01 -4.06127602e-01 -1.17624357e-01
6.13256037e-01 2.92152673e-01 -4.42456841e-01 -3.47447157e-01
-1.35368383e+00 7.52668977e-01 6.12557769e-01 6.38314068e-01
-3.64715576e-01 -5.26818216e-01 -1.24238503e+00 -4.96309549e-02
1.62832275e-01 -1.00991845e+00 8.02049339e-01 -9.13417697e-01
-1.44440508e+00 7.52156913e-01 -3.46951187e-01 -1.07173100e-01
5.64990759e-01 -6.24316148e-02 -2.66151309e-01 2.18147874e-01
4.41350043e-01 1.07829666e+00 5.94977379e-01 -1.12550628e+00
-1.35557342e+00 -3.86134714e-01 1.29776910e-01 4.90963936e-01
-2.21338257e-01 -2.13305861e-01 -6.85088277e-01 -5.82583010e-01
5.61348081e-01 -4.36766356e-01 -3.01572770e-01 1.76566437e-01
-3.36881399e-01 -1.57272816e-01 6.15651488e-01 -4.92680758e-01
1.09516287e+00 -2.07248664e+00 1.51916876e-01 2.59195060e-01
3.28361601e-01 3.06470364e-01 -1.82329398e-02 -1.49726734e-01
5.95061034e-02 -2.52629459e-01 -2.89546937e-01 -3.31290185e-01
-1.94180980e-01 1.44567683e-01 -3.76265422e-02 3.20510387e-01
3.48553270e-01 8.32358718e-01 -1.00495434e+00 -7.48127937e-01
7.03879535e-01 3.94472033e-01 -6.46186113e-01 2.52205461e-01
1.59797624e-01 6.83243930e-01 -6.87800467e-01 5.75834572e-01
1.19580066e+00 -6.34604618e-02 -5.76228142e-01 -5.84658921e-01
-3.80156517e-01 1.70205876e-01 -1.45174384e+00 1.99291837e+00
-5.19098938e-01 5.81397176e-01 -3.32256556e-02 -8.49621534e-01
1.18635225e+00 -2.28678688e-01 3.02281260e-01 -1.09937906e+00
2.71691799e-01 5.15254498e-01 -1.51126876e-01 -4.78355676e-01
4.28318888e-01 -9.03988481e-02 7.60996640e-02 -1.05016783e-01
-3.92310582e-02 -2.29810953e-01 8.02637786e-02 -7.99975768e-02
6.22935236e-01 2.20863238e-01 3.36056761e-02 -2.66592950e-01
1.15643585e+00 -2.54864872e-01 1.11711204e+00 3.71241719e-01
-3.42278987e-01 7.89897978e-01 1.72845826e-01 -6.27233982e-01
-9.23195541e-01 -6.96107447e-01 -1.81599051e-01 5.62704504e-01
8.77998590e-01 -1.76970467e-01 -4.96968836e-01 -5.20139098e-01
1.39827011e-02 -9.48875919e-02 -5.79104960e-01 -2.40758911e-01
-7.92423189e-01 -3.87888610e-01 1.00565881e-01 6.64338946e-01
1.44883573e+00 -1.07374859e+00 -5.72133362e-01 4.12321299e-01
-3.63944769e-01 -1.13749945e+00 -6.63679600e-01 -1.42526895e-01
-8.37502182e-01 -1.05046034e+00 -8.33707392e-01 -1.05523717e+00
6.29817009e-01 4.94657278e-01 9.37746584e-01 3.92992139e-01
-2.11819276e-01 -3.29004586e-01 -9.39761996e-02 -1.11020477e-02
2.09531650e-01 -7.36947879e-02 -3.78975928e-01 2.72732317e-01
2.04022020e-01 -6.87595606e-01 -1.07676148e+00 4.43382770e-01
-8.79506409e-01 6.88717306e-01 6.88240170e-01 9.63727415e-01
4.54813629e-01 2.35550515e-02 3.97110313e-01 -4.37477767e-01
2.18829259e-01 -7.73438960e-02 -6.17928803e-01 4.20930664e-05
-4.50462252e-01 2.28979319e-01 5.31508744e-01 3.40748653e-02
-1.39733160e+00 -1.00469589e-02 -4.26049650e-01 -4.41660672e-01
-1.40278757e-01 2.80721068e-01 -4.23222780e-01 -2.33472258e-01
1.16962858e-01 4.03988510e-01 -1.57639116e-01 -3.29372138e-01
2.41572157e-01 6.99524999e-01 7.39299834e-01 -6.30193055e-01
6.77322268e-01 6.84358835e-01 2.28363007e-01 -2.54276931e-01
-9.12462234e-01 -4.75052208e-01 -6.15844071e-01 -2.77852476e-01
1.01077509e+00 -1.12011933e+00 -8.23400021e-01 9.28865135e-01
-1.19386792e+00 5.17752916e-02 -7.58321350e-03 5.81770480e-01
-5.32764554e-01 4.96555150e-01 -5.34142435e-01 -3.32421005e-01
-3.69738936e-01 -1.36381245e+00 1.24484813e+00 7.97147870e-01
4.13685054e-01 -8.36196363e-01 -1.32209525e-01 1.91798300e-01
4.66241509e-01 3.06804091e-01 5.22152841e-01 -3.97019014e-02
-9.32081461e-01 2.65102863e-01 -9.19497073e-01 3.64737511e-01
2.45348111e-01 -8.94469172e-02 -9.48258817e-01 -7.22616389e-02
-1.16269514e-01 -1.39755487e-01 1.32263410e+00 4.39228743e-01
1.17395842e+00 8.83181244e-02 -2.37245396e-01 1.21747184e+00
1.60742939e+00 1.44392803e-01 7.73522556e-01 6.92643166e-01
9.26915228e-01 6.82011664e-01 8.90799284e-01 3.75723094e-01
6.19382858e-01 5.37184894e-01 4.01672453e-01 -3.95195574e-01
-3.30449194e-01 -4.19355392e-01 -2.01193780e-01 6.14132583e-01
-1.58541016e-02 2.96986163e-01 -5.75219035e-01 4.19086039e-01
-2.07264304e+00 -9.41050351e-01 -8.28725249e-02 1.96292567e+00
7.22385108e-01 4.25145060e-01 -2.62177587e-01 7.04088435e-02
1.06931055e+00 3.69159490e-01 -4.78760302e-01 9.03127715e-02
-3.69467765e-01 -6.45056600e-03 4.05966014e-01 5.88100851e-01
-1.20841765e+00 9.69762802e-01 4.80598402e+00 9.71474588e-01
-1.29995310e+00 -2.45601311e-01 8.39489222e-01 4.16726887e-01
-4.31956112e-01 7.03965873e-02 -8.26043725e-01 8.07193756e-01
2.15739012e-02 -7.61080161e-02 7.58732334e-02 5.38399518e-01
3.50031763e-01 -3.51527303e-01 -8.84034991e-01 1.29324079e+00
-1.38204778e-02 -1.52314043e+00 1.22922875e-01 -7.56489486e-02
9.07541931e-01 -1.21855974e-01 -9.14493576e-02 1.49371341e-01
1.70856481e-03 -7.69498408e-01 6.67928994e-01 7.30207920e-01
6.87727511e-01 -7.46287584e-01 1.06959021e+00 1.90360010e-01
-1.66066611e+00 -1.18572235e-01 -5.33618033e-01 -1.45972192e-01
2.05853865e-01 7.09565818e-01 -5.69491871e-02 9.69473183e-01
7.12788641e-01 1.42122853e+00 -5.32039583e-01 1.50147295e+00
-2.22501650e-01 -2.06440136e-01 -2.39918888e-01 2.95143455e-01
4.52982992e-01 -4.22108054e-01 1.43382117e-01 1.08834994e+00
2.76525080e-01 1.56417087e-01 2.16142714e-01 7.79111326e-01
1.23890005e-01 -3.01908818e-03 -3.44455987e-01 7.61773527e-01
4.38808262e-01 1.19836664e+00 -5.50322831e-01 -6.39636219e-01
-4.86486614e-01 8.44226360e-01 3.98872137e-01 2.04121828e-01
-6.25968575e-01 -6.07003570e-01 7.07080781e-01 -2.20255464e-01
2.92674929e-01 6.52579367e-02 -4.92588371e-01 -1.32131183e+00
2.26585999e-01 -5.26306689e-01 3.68754715e-01 -9.19828951e-01
-1.16942012e+00 6.96317017e-01 -2.90874809e-01 -1.61516190e+00
3.77629604e-03 -4.49802339e-01 -8.31117213e-01 1.22076988e+00
-2.24650693e+00 -9.42384899e-01 -1.00405252e+00 8.69354904e-01
6.09134674e-01 8.41799900e-02 2.02427372e-01 7.05069900e-01
-5.96197665e-01 2.70808458e-01 -2.49047667e-01 3.19278985e-01
7.61653423e-01 -9.18322861e-01 2.02913821e-01 8.96336794e-01
-5.21050811e-01 3.44031543e-01 4.38701987e-01 -5.27820587e-01
-7.07182229e-01 -1.25841010e+00 8.26551914e-01 2.16253266e-01
2.10600108e-01 7.23161846e-02 -8.80490839e-01 2.20506966e-01
-2.18030508e-03 4.12888974e-01 -1.29503101e-01 -2.68527865e-01
-2.22985804e-01 -5.86938202e-01 -1.09786510e+00 3.66246402e-01
1.35792398e+00 -4.47252810e-01 -6.88899875e-01 -2.40726098e-02
7.88389504e-01 -6.06523156e-01 -6.08317733e-01 8.54666471e-01
4.52098191e-01 -1.53208077e+00 9.37607288e-01 -1.13153592e-01
7.37150609e-01 -7.25867510e-01 -1.36407137e-01 -1.18939960e+00
-3.73616636e-01 -2.98634440e-01 4.41677630e-01 1.31888974e+00
-1.46871628e-02 -7.19816744e-01 7.30801344e-01 3.66418391e-01
-4.00350213e-01 -7.00504780e-01 -9.30056274e-01 -4.88995075e-01
-1.75823063e-01 -7.32270703e-02 8.51696908e-01 8.41040373e-01
-2.76724905e-01 2.13120162e-01 -3.10199112e-01 -2.71997117e-02
6.21756792e-01 3.77261430e-01 8.16135943e-01 -1.26584423e+00
1.11724459e-01 -6.28882945e-01 -7.94670403e-01 -1.40942204e+00
2.08266884e-01 -5.72794259e-01 2.77146488e-01 -1.70285749e+00
3.47365439e-01 -7.19102859e-01 -4.63575721e-01 2.07032636e-01
-6.65519059e-01 2.63075083e-01 9.93393436e-02 1.20323434e-01
-5.68357229e-01 8.93481910e-01 1.62790143e+00 -6.03640527e-02
-1.27432063e-01 -1.11656569e-01 -5.77202797e-01 6.94570303e-01
6.49641275e-01 -1.03716753e-01 -2.45595634e-01 -5.58835566e-01
-7.68875107e-02 1.18795216e-01 4.33505595e-01 -1.24435222e+00
5.28325021e-01 -2.14561760e-01 5.69381356e-01 -8.96483183e-01
3.03510219e-01 -8.03268731e-01 -1.48690015e-01 4.33919013e-01
-1.07339434e-01 5.71468053e-03 4.26164418e-02 4.12742436e-01
-7.51734197e-01 2.15072170e-01 8.61253262e-01 -1.21053800e-01
-1.33000863e+00 7.58886933e-01 3.19741756e-01 1.77452594e-01
7.52680123e-01 -4.77804840e-01 -3.81668389e-01 -1.21448249e-01
-3.63362879e-01 6.23135030e-01 3.74972135e-01 5.22211194e-01
8.86486411e-01 -1.53014934e+00 -7.55660415e-01 5.24934649e-01
2.17523530e-01 3.59296232e-01 5.73148131e-01 6.41698956e-01
-6.15912139e-01 3.79651368e-01 -6.55580401e-01 -9.15534556e-01
-1.03193557e+00 3.72238815e-01 5.73209107e-01 -4.08474654e-02
-5.59540391e-01 9.32657659e-01 5.38463950e-01 -1.71091244e-01
1.06930949e-01 -4.74769205e-01 -3.68057579e-01 -3.03820465e-02
4.64365780e-01 1.54894084e-01 6.70661917e-04 -7.31202245e-01
-3.74788761e-01 1.21667516e+00 1.33599728e-01 8.39188993e-02
1.23866415e+00 -5.32901406e-01 -1.99100211e-01 1.04381919e-01
1.53001213e+00 -2.45168626e-01 -1.63109887e+00 -6.21602058e-01
-3.45860183e-01 -1.00993776e+00 2.70145506e-01 -3.89356732e-01
-1.49979162e+00 1.10003972e+00 6.31125331e-01 -3.00132006e-01
1.30360758e+00 -3.22994411e-01 9.54082012e-01 1.76403075e-02
3.07074040e-01 -1.02692497e+00 1.85524095e-02 4.96260285e-01
6.24836028e-01 -1.63480794e+00 -5.21944799e-02 -6.75384402e-01
-4.02152866e-01 1.22096193e+00 1.12987244e+00 -3.14538836e-01
7.45510876e-01 -1.02208368e-01 1.54689685e-01 -4.71440959e-04
-4.46508795e-01 -6.16480708e-01 3.94120067e-01 4.58981186e-01
4.40899320e-02 -2.76343852e-01 -2.50639409e-01 2.97222286e-01
-1.01553462e-01 2.00399011e-01 1.87736109e-01 8.21622133e-01
-7.24357009e-01 -8.30759287e-01 -2.30905250e-01 2.58772224e-01
-4.45736200e-02 -2.86960512e-01 2.78607070e-01 5.53443015e-01
6.19991720e-01 9.92020965e-01 3.73877645e-01 -5.19627273e-01
3.76840621e-01 -5.43296278e-01 2.67008156e-01 -3.51503104e-01
-3.43974471e-01 -2.93512717e-02 -4.52179424e-02 -7.96187460e-01
-6.90990388e-01 -3.05483609e-01 -1.10830665e+00 -3.20281744e-01
-1.91878125e-01 -2.61218008e-02 4.55750465e-01 1.10147393e+00
2.56825209e-01 6.23838305e-01 9.65966940e-01 -1.03737462e+00
-1.68972742e-02 -7.47294664e-01 -4.30459350e-01 5.73220074e-01
5.72479308e-01 -7.27472544e-01 -4.09323990e-01 -2.71948367e-01] | [8.935369491577148, -2.2934250831604004] |
aa68c2d4-aa30-4491-80a3-763d5f715de6 | merlot-multimodal-neural-script-knowledge | 2106.02636 | null | https://arxiv.org/abs/2106.02636v3 | https://arxiv.org/pdf/2106.02636v3.pdf | MERLOT: Multimodal Neural Script Knowledge Models | As humans, we understand events in the visual world contextually, performing multimodal reasoning across time to make inferences about the past, present, and future. We introduce MERLOT, a model that learns multimodal script knowledge by watching millions of YouTube videos with transcribed speech -- in an entirely label-free, self-supervised manner. By pretraining with a mix of both frame-level (spatial) and video-level (temporal) objectives, our model not only learns to match images to temporally corresponding words, but also to contextualize what is happening globally over time. As a result, MERLOT exhibits strong out-of-the-box representations of temporal commonsense, and achieves state-of-the-art performance on 12 different video QA datasets when finetuned. It also transfers well to the world of static images, allowing models to reason about the dynamic context behind visual scenes. On Visual Commonsense Reasoning, MERLOT answers questions correctly with 80.6% accuracy, outperforming state-of-the-art models of similar size by over 3%, even those that make heavy use of auxiliary supervised data (like object bounding boxes). Ablation analyses demonstrate the complementary importance of: 1) training on videos versus static images; 2) scaling the magnitude and diversity of the pretraining video corpus; and 3) using diverse objectives that encourage full-stack multimodal reasoning, from the recognition to cognition level. | ['Yejin Choi', 'Ali Farhadi', 'Jize Cao', 'Jae Sung Park', 'Youngjae Yu', 'Jack Hessel', 'Ximing Lu', 'Rowan Zellers'] | 2021-06-04 | null | http://proceedings.neurips.cc/paper/2021/hash/c6d4eb15f1e84a36eff58eca3627c82e-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/c6d4eb15f1e84a36eff58eca3627c82e-Paper.pdf | neurips-2021-12 | ['visual-commonsense-reasoning'] | ['reasoning'] | [ 9.88898277e-02 -6.19471632e-02 -2.00185925e-01 -4.67952073e-01
-7.26872087e-01 -9.61534381e-01 9.63844478e-01 -6.64291531e-02
-4.50533241e-01 5.00582039e-01 6.75458550e-01 -2.91358620e-01
1.92610204e-01 -5.12995303e-01 -9.77058053e-01 -2.73997694e-01
-8.14441442e-02 3.08811367e-01 3.16783726e-01 -2.80429929e-01
1.03355339e-02 1.26992568e-01 -1.63759220e+00 9.33069587e-01
3.79332632e-01 1.01239288e+00 7.20638186e-02 9.34695661e-01
-1.12986200e-01 1.69135797e+00 -5.12630522e-01 -5.60095251e-01
-2.09017441e-01 -5.85910261e-01 -9.80464041e-01 2.68438280e-01
8.97952199e-01 -6.97324514e-01 -8.79058540e-01 6.60283566e-01
-8.36421847e-02 3.54035616e-01 4.41269606e-01 -1.33257246e+00
-1.22269487e+00 3.62279177e-01 -5.90274967e-02 5.18007815e-01
7.47841775e-01 7.45129526e-01 1.03677320e+00 -5.35874903e-01
1.01194310e+00 1.37286508e+00 3.41395497e-01 6.05522335e-01
-1.13709033e+00 -4.84171838e-01 3.94624710e-01 5.96128941e-01
-9.72493231e-01 -5.30762017e-01 4.31969285e-01 -5.94958842e-01
1.32123291e+00 2.12380499e-01 7.93966234e-01 1.55203521e+00
1.37803838e-01 8.05407882e-01 1.02091956e+00 1.49606320e-03
2.18872309e-01 -2.82548279e-01 -1.79510176e-01 8.54759693e-01
-4.14561152e-01 -4.76632006e-02 -9.58822370e-01 8.95347074e-02
7.50276566e-01 -4.33950573e-02 -2.37994209e-01 -2.75468737e-01
-1.76482916e+00 5.52163541e-01 4.60865796e-01 2.50220716e-01
-2.46154591e-01 7.35445619e-01 5.84165096e-01 2.81724513e-01
1.22725390e-01 4.61632043e-01 -1.57040924e-01 -5.49469769e-01
-8.93618643e-01 1.66352674e-01 6.43909037e-01 8.35776269e-01
5.89421332e-01 1.29041374e-01 -3.45863014e-01 4.29877669e-01
8.78074914e-02 7.44378150e-01 4.48235571e-01 -1.50267994e+00
6.28632724e-01 4.42303628e-01 9.79430303e-02 -9.94979560e-01
-2.20401958e-01 6.36592209e-02 -2.11071059e-01 -4.43221815e-02
5.90154946e-01 -9.75946411e-02 -1.14042401e+00 2.08721495e+00
6.96220845e-02 2.62318999e-01 1.90807506e-01 1.14410555e+00
8.99438262e-01 1.06824517e+00 6.15874588e-01 -5.60072018e-03
1.57679009e+00 -7.79404461e-01 -6.84607983e-01 -5.20987332e-01
4.41947877e-01 -3.53521138e-01 1.41243780e+00 2.80678600e-01
-1.13117278e+00 -4.87525672e-01 -7.53685296e-01 -5.32416642e-01
-4.93060350e-01 -7.39274621e-02 6.09327137e-01 1.83141261e-01
-1.20269668e+00 2.46400371e-01 -7.88897216e-01 -5.22895277e-01
4.91980374e-01 -2.26431251e-01 -5.68883538e-01 -1.72824457e-01
-1.27657008e+00 1.12112880e+00 2.66703963e-01 -1.36464193e-01
-1.50153589e+00 -5.90710819e-01 -1.01896322e+00 1.14383563e-01
4.58418250e-01 -7.58675158e-01 1.48289669e+00 -1.44749093e+00
-1.18257499e+00 9.60293829e-01 -3.84166777e-01 -4.70461965e-01
5.17139971e-01 -3.03276926e-01 -4.66087669e-01 8.91646564e-01
1.73612416e-01 1.06763041e+00 8.72389793e-01 -1.24336803e+00
-3.65294546e-01 -2.04521284e-01 4.96746033e-01 3.60952824e-01
-2.01048598e-01 -1.12864286e-01 -5.60353279e-01 -4.97357994e-01
-1.83189392e-01 -7.43243754e-01 2.46860445e-01 3.03929001e-01
8.14086720e-02 -7.31284022e-02 9.67540205e-01 -1.01471245e+00
9.94778514e-01 -2.19011426e+00 4.35066462e-01 -3.16631526e-01
1.02383688e-01 -1.68663591e-01 -2.54194438e-01 5.38974702e-01
-1.12427689e-01 6.92759734e-03 -7.98579864e-03 -3.96893620e-01
3.96551527e-02 3.83495539e-01 -6.51008368e-01 3.49187464e-01
3.10914725e-01 1.21232045e+00 -1.18506825e+00 -6.82642281e-01
2.87759751e-01 4.45717990e-01 -4.31706786e-01 5.26778065e-02
-7.74984300e-01 3.45833242e-01 -1.43308401e-01 7.14104056e-01
-1.27865240e-01 -6.02167845e-01 2.45345637e-01 -4.06906158e-01
1.64929554e-01 4.81716581e-02 -5.63122571e-01 1.97768247e+00
-4.24166232e-01 1.21821260e+00 -2.62519449e-01 -8.29183936e-01
2.00906202e-01 4.95138645e-01 2.00920001e-01 -9.86431718e-01
-5.08444309e-02 -2.73949504e-01 -2.44235113e-01 -1.07063031e+00
3.90621781e-01 -2.20963001e-01 -2.51869321e-01 4.26678509e-01
4.28662330e-01 -1.53206572e-01 3.18845868e-01 7.67050266e-01
1.01739228e+00 4.23218727e-01 -2.03701411e-03 2.31791243e-01
1.67870849e-01 3.27855766e-01 7.95335770e-02 6.85436904e-01
-3.29513699e-01 4.80973363e-01 7.97525465e-01 -4.67034191e-01
-1.09249246e+00 -1.32779896e+00 3.92914861e-01 1.46787226e+00
2.09975719e-01 -3.88669968e-01 -6.58930600e-01 -5.72502792e-01
3.47939543e-02 1.16715097e+00 -9.24311697e-01 -2.14814439e-01
-4.99141783e-01 -7.38548907e-03 7.45391130e-01 7.40454197e-01
5.66595733e-01 -1.25377119e+00 -1.04427612e+00 -9.73646119e-02
-6.89547062e-01 -1.64427257e+00 -4.10556883e-01 -1.38470039e-01
-6.49432003e-01 -1.11075580e+00 -4.53813076e-01 -4.46459323e-01
3.44627529e-01 2.42175713e-01 1.27703130e+00 -4.12113667e-02
-2.50444055e-01 1.17943048e+00 -3.93728733e-01 5.58469370e-02
-3.43436956e-01 -6.79876924e-01 -2.34415919e-01 4.47907895e-02
1.21300608e-01 -3.77646893e-01 -4.39381897e-01 1.73084050e-01
-9.87532377e-01 2.83368051e-01 2.63836354e-01 7.01488435e-01
2.33953893e-01 -2.15283290e-01 3.20599228e-01 -3.58132899e-01
2.84394503e-01 -6.01189911e-01 -3.08235049e-01 5.53319156e-01
9.33583230e-02 3.59913744e-02 3.97778004e-01 -7.88092136e-01
-1.28045607e+00 -1.20495297e-01 3.81949633e-01 -8.91033471e-01
-2.34244585e-01 2.81762868e-01 1.56840429e-01 2.89680034e-01
7.42239654e-01 3.99250358e-01 5.07794209e-02 1.74005702e-01
9.30418849e-01 2.05973089e-01 1.01752901e+00 -8.25462222e-01
5.22618651e-01 8.40940237e-01 -2.54681557e-01 -7.57302165e-01
-9.65446651e-01 -3.04602146e-01 -4.67717141e-01 -6.67306781e-01
1.50953817e+00 -1.14984488e+00 -1.06710291e+00 1.66698664e-01
-1.17470610e+00 -8.03175807e-01 -1.97375849e-01 2.92244673e-01
-9.62834835e-01 3.16560566e-01 -7.47147679e-01 -6.61096752e-01
1.57351747e-01 -1.02772260e+00 1.06846821e+00 -4.27998463e-03
-4.44552660e-01 -9.92476642e-01 -2.38973528e-01 8.26239884e-01
2.51496673e-01 3.35530788e-01 1.01803267e+00 -2.73749769e-01
-7.91639328e-01 1.54035881e-01 -4.92096484e-01 2.19320118e-01
-1.62332043e-01 -9.54309665e-03 -9.57826495e-01 7.39435554e-02
-2.01095641e-01 -8.62284958e-01 9.17978227e-01 1.41191751e-01
1.13119423e+00 -3.35161865e-01 -1.27197262e-02 3.51831675e-01
1.21401846e+00 1.31630555e-01 7.19890773e-01 3.20877016e-01
5.91582716e-01 6.09991670e-01 4.54239219e-01 2.95116991e-01
8.57344568e-01 4.64292824e-01 7.09978282e-01 1.91551641e-01
-3.42625558e-01 -4.08371359e-01 6.32406592e-01 1.73405781e-01
2.26686858e-02 -3.22116613e-01 -1.19914389e+00 7.86613166e-01
-2.05195808e+00 -1.72099555e+00 2.45735377e-01 1.66685510e+00
9.22003448e-01 -2.83634942e-02 1.79657400e-01 -4.27379251e-01
3.49009842e-01 5.07407665e-01 -7.34136283e-01 -3.91229421e-01
-2.83207417e-01 -3.05756360e-01 2.13850856e-01 4.42456335e-01
-1.04307961e+00 1.06649172e+00 6.75602293e+00 4.64602381e-01
-1.20868242e+00 1.97790965e-01 5.65647542e-01 -4.67204481e-01
-3.72762173e-01 8.59060511e-02 -2.51718581e-01 3.70651692e-01
1.03478241e+00 1.94195896e-01 9.24342215e-01 4.79124159e-01
1.15771912e-01 -4.00229961e-01 -1.31002188e+00 1.16489005e+00
6.28336847e-01 -1.58894420e+00 2.95809448e-01 -3.24216098e-01
6.16441607e-01 -7.56175350e-03 1.02058023e-01 4.28153515e-01
2.11616457e-01 -1.25560832e+00 1.31339216e+00 7.43558764e-01
8.60325813e-01 -1.92763969e-01 1.41703069e-01 3.40736419e-01
-9.35793579e-01 -4.17980105e-01 1.77587256e-01 -1.06762156e-01
4.08960462e-01 -3.33447084e-02 -5.79610467e-01 4.56065387e-02
9.14967477e-01 7.08249152e-01 -7.38018334e-01 3.63850534e-01
-3.21755886e-01 4.56998646e-01 -7.11042732e-02 -2.23967736e-03
5.06276727e-01 2.49638826e-01 2.96169043e-01 1.50423002e+00
6.81048408e-02 5.51988959e-01 -1.17645808e-01 8.74085248e-01
-3.60771567e-02 -4.28843230e-01 -5.57282388e-01 -5.30934274e-01
3.38451177e-01 7.81541109e-01 -4.73107815e-01 -5.59765399e-01
-5.35552919e-01 1.13913310e+00 3.49528372e-01 8.14000666e-01
-1.16250432e+00 1.01859815e-01 5.61765730e-01 -6.59487117e-03
4.14795220e-01 -5.68232417e-01 2.12016180e-02 -1.29671049e+00
-1.09567307e-01 -1.08027899e+00 6.16396844e-01 -1.76154959e+00
-1.11054277e+00 4.10478234e-01 3.29110950e-01 -9.25099969e-01
-3.78080368e-01 -8.11523676e-01 -3.66610408e-01 3.58949870e-01
-1.19199443e+00 -1.23935699e+00 -5.49015820e-01 8.99953127e-01
7.53047526e-01 1.16098948e-01 6.05750263e-01 -1.13288999e-01
-1.14475787e-01 6.07977770e-02 -3.50165874e-01 2.22455338e-01
8.42198968e-01 -1.11285400e+00 1.72750637e-01 6.09800696e-01
3.60552698e-01 6.24130666e-01 6.98744774e-01 -6.15358710e-01
-1.60329127e+00 -6.04582906e-01 6.83042407e-01 -9.57722068e-01
1.08226466e+00 -3.31275046e-01 -7.53343821e-01 1.07206547e+00
5.30644655e-01 -1.83322541e-02 6.23305023e-01 3.68734300e-02
-1.01806402e+00 2.62198057e-02 -9.64427412e-01 9.09984589e-01
1.08617091e+00 -1.26125717e+00 -1.00588381e+00 4.37010169e-01
9.17460203e-01 -5.67912698e-01 -8.39135468e-01 4.61709872e-02
8.34288776e-01 -9.88036573e-01 1.07019520e+00 -1.09613144e+00
9.81764376e-01 -2.76682109e-01 -5.75890601e-01 -9.49254632e-01
-9.62321535e-02 -4.59314883e-01 -3.55493009e-01 9.18169796e-01
1.84627339e-01 -1.24160230e-01 4.35644507e-01 6.55724883e-01
3.21069956e-02 -5.06423175e-01 -9.28028464e-01 -5.25105476e-01
-1.90466002e-01 -6.31823957e-01 1.99394256e-01 1.01633310e+00
3.01940799e-01 3.60922575e-01 -3.36239547e-01 1.33801907e-01
2.31557801e-01 1.07847285e-02 5.71981192e-01 -5.84981620e-01
-2.93226689e-01 -3.93797696e-01 -4.46322709e-01 -1.02480495e+00
3.99174392e-01 -5.12019873e-01 -5.19497730e-02 -1.62224698e+00
3.78072679e-01 2.52717346e-01 7.92358350e-03 7.41771042e-01
8.11919495e-02 2.58841753e-01 5.00563622e-01 1.72116458e-01
-1.11907721e+00 3.71515065e-01 1.37507784e+00 -1.69906482e-01
2.38153517e-01 -8.70495975e-01 -4.55840588e-01 6.02291167e-01
4.68820632e-01 2.11362150e-02 -6.96719289e-01 -8.69820356e-01
3.66276145e-01 4.77638781e-01 1.10228789e+00 -8.45424533e-01
2.15052828e-01 -4.96247947e-01 5.72929442e-01 -4.13314968e-01
8.71122539e-01 -6.82801783e-01 7.76743218e-02 1.02964483e-01
-6.94393575e-01 3.00907373e-01 5.04971981e-01 7.58778512e-01
-2.45881200e-01 1.97942168e-01 5.25772154e-01 -3.41602623e-01
-1.37903833e+00 -8.05814266e-02 -4.00547922e-01 3.18599671e-01
1.05353522e+00 -2.08681539e-01 -1.02161777e+00 -1.01547933e+00
-8.99468064e-01 3.64324123e-01 5.44003904e-01 5.88285327e-01
6.44858658e-01 -1.10358346e+00 -3.25826675e-01 -1.78213447e-01
3.46705347e-01 -4.54913914e-01 6.64017260e-01 6.93453729e-01
-5.59081435e-01 5.39435625e-01 -2.60371536e-01 -8.20281208e-01
-9.61154878e-01 7.39085495e-01 4.21412170e-01 3.08067381e-01
-5.53966641e-01 8.11201334e-01 3.44328225e-01 -5.82647137e-02
2.65315115e-01 -2.72963732e-01 7.03671351e-02 1.04025431e-01
5.74986041e-01 5.37459925e-02 -3.93343031e-01 -8.11262310e-01
-4.24204141e-01 4.75832760e-01 3.16902548e-01 -5.26184261e-01
9.08257961e-01 -2.75731593e-01 -1.89277511e-02 7.18903422e-01
1.15108144e+00 -3.42800587e-01 -1.66472042e+00 -9.03171450e-02
-3.17846924e-01 -3.36688459e-01 -9.82029885e-02 -1.28431439e+00
-5.73099315e-01 9.78461444e-01 2.92782366e-01 2.11506095e-02
1.03926146e+00 3.96968126e-01 6.54703915e-01 4.85337645e-01
1.51865631e-01 -9.89436626e-01 8.42345178e-01 5.17368317e-01
9.90355194e-01 -1.27139640e+00 -8.64771232e-02 1.63978130e-01
-1.29522264e+00 1.15395761e+00 5.66998482e-01 8.23154673e-02
1.30711928e-01 -3.63586806e-02 1.68033302e-01 -3.40888500e-01
-1.25151610e+00 -1.59502849e-01 3.52883726e-01 5.51176906e-01
2.40798175e-01 -8.13034177e-02 4.10312027e-01 1.48046017e-01
1.80102896e-03 -1.67269781e-02 3.19550455e-01 8.04800510e-01
-2.93399036e-01 -3.03289145e-01 -3.44256222e-01 1.31012693e-01
-9.10348669e-02 -6.82682768e-02 -3.78514379e-01 9.79076922e-01
7.13866204e-02 1.05916071e+00 3.88839602e-01 -1.42013013e-01
1.96543097e-01 3.18440139e-01 6.74841583e-01 -3.47966522e-01
-3.75279993e-01 -2.46958539e-01 3.71149063e-01 -1.06675112e+00
-5.03970563e-01 -7.90672541e-01 -1.36874545e+00 -2.13917315e-01
3.36760730e-01 -2.86381602e-01 5.36185205e-01 1.15545845e+00
2.68728137e-01 5.12816131e-01 2.17367988e-02 -8.61633301e-01
-3.91445190e-01 -5.21988332e-01 -8.35183542e-03 7.87071705e-01
6.59755051e-01 -5.69671988e-01 -4.00002271e-01 6.13446951e-01] | [10.505581855773926, 1.1220802068710327] |
a189452d-e3a5-4dd7-97b8-ae50a66efe29 | full-body-awareness-from-partial-observations | 2008.06046 | null | https://arxiv.org/abs/2008.06046v1 | https://arxiv.org/pdf/2008.06046v1.pdf | Full-Body Awareness from Partial Observations | There has been great progress in human 3D mesh recovery and great interest in learning about the world from consumer video data. Unfortunately current methods for 3D human mesh recovery work rather poorly on consumer video data, since on the Internet, unusual camera viewpoints and aggressive truncations are the norm rather than a rarity. We study this problem and make a number of contributions to address it: (i) we propose a simple but highly effective self-training framework that adapts human 3D mesh recovery systems to consumer videos and demonstrate its application to two recent systems; (ii) we introduce evaluation protocols and keypoint annotations for 13K frames across four consumer video datasets for studying this task, including evaluations on out-of-image keypoints; and (iii) we show that our method substantially improves PCK and human-subject judgments compared to baselines, both on test videos from the dataset it was trained on, as well as on three other datasets without further adaptation. Project website: https://crockwell.github.io/partial_humans | ['Chris Rockwell', 'David F. Fouhey'] | 2020-08-13 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2820_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123620511.pdf | eccv-2020-8 | ['human-mesh-recovery'] | ['computer-vision'] | [-3.78924161e-02 -2.26695508e-01 -3.68419766e-01 -1.05313919e-01
-1.11715853e+00 -4.77003247e-01 2.62581080e-01 -2.97214866e-01
-3.54078799e-01 3.46176416e-01 5.64697742e-01 1.55793540e-02
2.53542274e-01 -1.38498753e-01 -1.09876323e+00 -2.72760183e-01
-4.04428899e-01 6.95990503e-01 5.04236937e-01 -2.80999452e-01
1.95658118e-01 4.51634645e-01 -1.65986776e+00 4.57477510e-01
2.21949685e-02 1.00876987e+00 -2.65162855e-01 8.49919558e-01
6.22400224e-01 8.83694708e-01 -4.02658612e-01 -4.63509381e-01
7.52667308e-01 -1.11085273e-01 -1.19375718e+00 4.93525833e-01
1.32843792e+00 -7.77547896e-01 -5.64638734e-01 7.80491412e-01
7.44730234e-01 2.20520094e-01 5.78848720e-01 -1.25955880e+00
-3.06123197e-01 1.15104325e-01 -7.51557648e-01 2.52876222e-01
1.03058326e+00 3.00540954e-01 7.19118357e-01 -1.08171570e+00
1.23450112e+00 1.32535887e+00 1.10215700e+00 5.60922265e-01
-1.02045214e+00 -8.68497193e-01 -1.46051511e-01 1.95347533e-01
-1.49603975e+00 -8.30229938e-01 4.56167847e-01 -4.86380577e-01
9.06057417e-01 5.85340075e-02 1.14651740e+00 1.37596130e+00
2.52114445e-01 1.04003704e+00 6.62837446e-01 -1.70565501e-01
-1.17132790e-01 -1.96175992e-01 -4.33913499e-01 8.55131269e-01
-2.02674232e-02 2.12298289e-01 -7.18001187e-01 -2.76894838e-01
1.10091400e+00 -3.13121110e-01 -3.55353802e-01 -8.69330525e-01
-1.61765742e+00 5.40686429e-01 1.05559476e-01 -3.15583199e-01
-3.68339717e-01 2.96118200e-01 6.89694881e-01 4.69679385e-01
6.16862357e-01 -2.51779146e-03 -4.89076108e-01 -3.08824599e-01
-9.85850811e-01 8.17823350e-01 1.12044561e+00 1.28159332e+00
3.31867099e-01 -5.78925423e-02 4.35559541e-01 6.64279401e-01
-1.35877347e-02 7.71030426e-01 2.14135036e-01 -1.55579901e+00
4.27740097e-01 -9.41450000e-02 2.53045559e-01 -1.24053550e+00
-3.38839769e-01 2.16233835e-01 -5.28934002e-01 4.20488954e-01
5.41404366e-01 -2.95783747e-02 -9.43170846e-01 1.41585493e+00
5.55936813e-01 2.98432797e-01 -4.05694395e-01 1.13666427e+00
1.15141177e+00 2.06654474e-01 1.53518599e-02 3.40964645e-02
1.26202023e+00 -1.07813728e+00 -4.07416105e-01 1.51517056e-02
4.48821992e-01 -1.08544970e+00 1.00728011e+00 7.67397165e-01
-1.46990824e+00 -6.32578671e-01 -8.58150065e-01 -1.75531194e-01
-4.19966057e-02 -1.61075771e-01 6.80724025e-01 2.14332208e-01
-1.19118774e+00 6.89549863e-01 -7.49150038e-01 -6.96800709e-01
3.29223782e-01 4.04078811e-01 -8.25426042e-01 -1.08195931e-01
-9.10996497e-01 8.81582201e-01 3.57789882e-02 -2.68345326e-01
-9.53642607e-01 -7.01161265e-01 -9.37808573e-01 -5.43563843e-01
6.51010990e-01 -9.47428823e-01 1.60246027e+00 -1.07747662e+00
-1.43358362e+00 1.44079304e+00 7.57483467e-02 -2.73787946e-01
8.28720748e-01 -6.75034404e-01 -2.28912145e-01 6.70571804e-01
2.59020001e-01 1.05370152e+00 1.14134526e+00 -1.07281530e+00
-6.41577244e-01 4.18052040e-02 -2.21975055e-02 4.69510674e-01
1.47352248e-01 8.70375112e-02 -7.64101088e-01 -1.06754541e+00
-1.52560070e-01 -1.36507273e+00 8.94953907e-02 5.05778432e-01
-2.32083187e-01 -2.69539714e-01 5.81596434e-01 -8.19760263e-01
6.19114339e-01 -2.04420733e+00 4.30748403e-01 1.50163472e-01
3.79243165e-01 4.06198986e-02 -6.25365973e-02 3.56617957e-01
-2.06490651e-01 -2.83342093e-01 2.41669297e-01 -6.04636848e-01
-1.53607652e-01 9.73582640e-03 -2.26717636e-01 8.32733691e-01
-2.46381581e-01 6.81776822e-01 -9.17714953e-01 -7.99891174e-01
4.55542296e-01 5.40159166e-01 -7.88586199e-01 9.84486118e-02
-7.03351200e-02 4.92109269e-01 -2.06543673e-02 1.10373342e+00
4.91174489e-01 -4.37596500e-01 -5.23216017e-02 -5.07270336e-01
3.26274514e-01 -1.57744899e-01 -1.44702101e+00 2.16978145e+00
7.09960163e-02 4.53483313e-01 3.58485609e-01 -6.11917973e-01
7.98450410e-02 6.26520097e-01 8.81729662e-01 -3.37134212e-01
2.76250601e-01 2.29870215e-01 -5.56952596e-01 -7.03547180e-01
5.61656654e-01 3.50909494e-02 2.10379828e-02 3.91844630e-01
1.33851558e-01 -2.50256777e-01 1.25654057e-01 3.17032278e-01
1.18857145e+00 5.10348797e-01 4.24291402e-01 -2.64272720e-01
1.55783534e-01 1.90068081e-01 3.93437922e-01 4.06535596e-01
-4.45831716e-01 9.72792566e-01 2.06983179e-01 -9.07371104e-01
-1.09820127e+00 -1.09593034e+00 4.93551791e-02 1.11591482e+00
3.88614237e-01 -7.52018392e-01 -6.95103586e-01 -6.19787693e-01
1.84535280e-01 -9.90372300e-02 -6.24642611e-01 2.51812309e-01
-7.29973137e-01 -7.93698430e-02 5.53969383e-01 5.85454166e-01
3.52789313e-01 -1.04760921e+00 -5.37881136e-01 -1.45256400e-01
-4.65659738e-01 -1.36149430e+00 -9.00887668e-01 -3.89183551e-01
-7.80103624e-01 -1.55792129e+00 -9.71698225e-01 -9.33050096e-01
4.62515086e-01 4.97397751e-01 1.61360908e+00 3.03705007e-01
-4.12538409e-01 9.23048913e-01 -4.01684433e-01 -1.51576072e-01
-2.91961104e-01 1.90323759e-02 4.87802684e-01 -4.47821379e-01
1.66251004e-01 -5.16433477e-01 -8.59281182e-01 6.29548192e-01
-5.82512021e-01 -8.85469094e-02 3.66274416e-01 5.49048007e-01
7.61377871e-01 -1.70088693e-01 9.77135003e-02 -9.25082386e-01
1.45727590e-01 -4.44434851e-01 -3.47973019e-01 -1.18424892e-01
-3.58833253e-01 -5.60876846e-01 1.65420189e-01 -6.03920519e-01
-6.64581716e-01 3.24812293e-01 -3.23338181e-01 -9.72253144e-01
-1.05599642e-01 -1.43660471e-01 1.05757147e-01 -4.29226428e-01
8.29032838e-01 -3.62273097e-01 1.99251682e-01 -4.48571563e-01
2.29048073e-01 3.39182854e-01 8.51515651e-01 -7.80119777e-01
8.16181302e-01 6.74456537e-01 -5.37556857e-02 -9.10578966e-01
-6.09944522e-01 -6.57516003e-01 -6.94989324e-01 -2.51009196e-01
9.41626251e-01 -1.36104715e+00 -9.05764699e-01 4.63429332e-01
-8.81417930e-01 -5.50614238e-01 -3.85327578e-01 5.34191549e-01
-9.74305630e-01 5.64534068e-01 -1.00169134e+00 -1.79212153e-01
-2.97326118e-01 -1.29080021e+00 1.45568645e+00 -2.39100277e-01
-6.60646677e-01 -8.48444402e-01 6.21223152e-02 5.09267509e-01
5.82529008e-02 4.78671134e-01 3.72939557e-01 -1.84300214e-01
-5.78889012e-01 -2.87220895e-01 -1.57179330e-02 2.26456091e-01
-1.38291016e-01 -7.35231042e-02 -6.87322915e-01 -5.23344576e-01
-2.74686664e-01 -7.70958781e-01 5.54266036e-01 4.91383284e-01
9.36455905e-01 -4.06045467e-02 -3.52077425e-01 8.30125749e-01
9.51474488e-01 -3.78516376e-01 6.47753775e-01 3.71360600e-01
8.93380105e-01 4.48268682e-01 5.99367201e-01 3.33967000e-01
5.81250250e-01 9.62652922e-01 1.20681919e-01 -2.05873415e-01
-5.03535151e-01 -2.38653675e-01 3.59518051e-01 6.78844333e-01
-6.46545291e-01 -1.11089222e-01 -6.60444796e-01 4.04134959e-01
-1.78225851e+00 -1.11587226e+00 1.27559796e-01 2.19579864e+00
6.52164161e-01 2.96116360e-02 6.74815238e-01 -4.77319993e-02
5.76101840e-01 1.36335939e-01 -4.66966689e-01 2.21885264e-01
-2.23070234e-02 3.07269990e-01 4.85650748e-01 3.85289729e-01
-1.43804038e+00 9.34034407e-01 7.44675922e+00 7.16449440e-01
-8.98054183e-01 9.71898958e-02 2.57384926e-01 -4.08981264e-01
3.50613236e-01 -1.67862147e-01 -5.03215253e-01 2.34506220e-01
5.82981229e-01 1.53116286e-01 3.89028102e-01 1.01545155e+00
-1.03422038e-01 -4.30207253e-02 -1.44019544e+00 1.46361566e+00
4.79750186e-01 -1.23220348e+00 2.70343106e-02 2.67188787e-01
6.80146694e-01 1.82495609e-01 -1.93704385e-02 1.10564835e-01
1.45695046e-01 -8.64316285e-01 1.00406063e+00 3.62803668e-01
1.02364600e+00 -5.55728078e-01 5.63146591e-01 4.99066524e-02
-1.27810335e+00 3.64720941e-01 -2.22235784e-01 6.06379062e-02
3.98308337e-01 1.53326228e-01 -5.93186259e-01 2.09910080e-01
1.18120444e+00 1.10940254e+00 -5.60219586e-01 1.03124130e+00
-8.94173086e-02 2.23598212e-01 -3.63314092e-01 6.54569566e-01
-3.22590142e-01 3.44128937e-01 7.48409510e-01 1.23522341e+00
3.86132523e-02 2.84897953e-01 5.36573887e-01 -4.57242085e-03
-3.16007048e-01 1.56664565e-01 -7.09350646e-01 2.92595029e-01
3.39608073e-01 9.70858514e-01 -9.24739122e-01 -5.05749941e-01
-5.44703543e-01 1.24375379e+00 1.27891406e-01 2.25365743e-01
-8.40983629e-01 -1.22200631e-01 7.17976391e-01 6.76929176e-01
3.39094162e-01 -2.10122108e-01 3.29097152e-01 -1.38988197e+00
-1.90374590e-02 -1.40086198e+00 6.44081652e-01 -9.59493518e-01
-1.41755164e+00 4.07297909e-01 4.19783831e-01 -1.61504281e+00
-4.45513666e-01 -4.61686671e-01 -5.87641783e-02 2.05945656e-01
-8.52178216e-01 -1.19727528e+00 -4.62610096e-01 9.07162964e-01
7.90990472e-01 -1.25871658e-01 8.47261071e-01 6.23581767e-01
1.04017388e-02 7.19014883e-01 -4.09435958e-01 1.42485633e-01
1.33967793e+00 -9.31974947e-01 6.91292524e-01 5.26377380e-01
1.85894221e-01 4.50204641e-01 7.55318940e-01 -9.03852820e-01
-1.71014094e+00 -5.79218388e-01 4.11203265e-01 -9.92337048e-01
4.84137148e-01 -2.45739266e-01 -6.60029650e-01 1.12375510e+00
2.14453697e-01 1.25215635e-01 4.59719688e-01 6.51211739e-02
-5.08716166e-01 2.78248906e-01 -1.17211974e+00 4.88601625e-01
1.58681381e+00 -4.75812018e-01 -5.66712320e-01 6.57522559e-01
5.74509144e-01 -1.08159530e+00 -1.18681872e+00 5.24978638e-01
9.51014042e-01 -1.08241475e+00 1.34767735e+00 -3.16443563e-01
5.10290265e-01 -1.68530524e-01 -3.77635360e-01 -9.74256277e-01
-3.16673160e-01 -9.00754392e-01 -3.85293424e-01 6.01242065e-01
-8.77369046e-02 -1.02440640e-01 1.11632717e+00 6.29420400e-01
-3.97513881e-02 -6.95115328e-01 -7.29234278e-01 -5.94444275e-01
-1.15132794e-01 -3.57577711e-01 2.16620550e-01 9.19433475e-01
-1.27474219e-01 3.21967810e-01 -9.05460238e-01 -1.03362881e-01
8.98796618e-01 -2.37421781e-01 1.40671837e+00 -9.88352597e-01
-4.34279472e-01 -5.27047403e-02 -7.44618952e-01 -1.28491414e+00
2.10465431e-01 -6.07002199e-01 8.40053335e-03 -1.12077868e+00
1.99264318e-01 -2.22448841e-01 2.97401190e-01 3.43815953e-01
1.44715905e-01 8.75099719e-01 1.66199684e-01 6.12592578e-01
-9.43452477e-01 5.53182140e-02 1.26172924e+00 -2.16140598e-02
4.74656112e-02 -5.58981821e-02 -3.47653508e-01 9.85525072e-01
2.60174215e-01 -2.50810534e-01 -4.37956691e-01 -5.91334164e-01
1.79312050e-01 2.98752576e-01 6.33111119e-01 -1.23537922e+00
2.30544940e-01 9.45942700e-02 7.19237864e-01 -6.38114035e-01
5.14566362e-01 -8.24424088e-01 4.66228068e-01 2.69625396e-01
-6.25528023e-02 4.37886685e-01 1.46199733e-01 6.71151996e-01
1.32322580e-01 2.13728309e-01 7.55909681e-01 -4.46737796e-01
-1.12601471e+00 5.06282508e-01 -1.98395908e-01 5.39960563e-01
9.07858431e-01 -3.43305588e-01 5.40783890e-02 -6.70885444e-01
-8.35257232e-01 5.98460510e-02 1.10023987e+00 6.15029931e-01
7.25380838e-01 -1.44225204e+00 -6.46189809e-01 2.78163612e-01
9.29330513e-02 -4.39316109e-02 1.85842633e-01 8.73146057e-01
-9.65028763e-01 -7.45604038e-02 -2.52512693e-01 -7.18257785e-01
-1.55639780e+00 5.79112649e-01 1.66042015e-01 1.92417786e-01
-1.24122596e+00 9.07331824e-01 9.11827385e-02 -3.27441394e-01
6.62594557e-01 -2.64400125e-01 2.69042462e-01 -1.53245628e-01
6.49660468e-01 5.22751689e-01 4.72675860e-02 -9.79222059e-01
-3.71977001e-01 1.11972129e+00 -1.22036844e-01 1.10336766e-02
1.21020532e+00 -2.17207819e-01 2.99514353e-01 2.39743203e-01
1.35976458e+00 2.13417083e-01 -1.26390946e+00 -2.18579322e-01
-4.12541509e-01 -7.49441981e-01 -3.97075444e-01 -3.66971195e-01
-1.10954642e+00 3.16482604e-01 6.06410563e-01 -1.99728683e-01
9.25399363e-01 2.67126650e-01 1.41180468e+00 3.81431997e-01
7.65452802e-01 -1.10692143e+00 2.64967859e-01 2.33271897e-01
8.13188136e-01 -1.36973298e+00 5.27696967e-01 -5.42886972e-01
-6.52795851e-01 9.13779199e-01 3.53328317e-01 -5.46631336e-01
9.45936143e-01 1.82999492e-01 2.21572369e-01 -5.62184811e-01
-5.67691147e-01 1.20599158e-01 3.75371784e-01 7.10823894e-01
3.12130302e-01 -1.71405986e-01 1.92685157e-01 5.45002632e-02
-4.14577544e-01 3.76711249e-01 3.69833678e-01 1.07033801e+00
-1.36340246e-01 -8.56576264e-01 -3.96901399e-01 3.28675032e-01
-6.10192597e-01 1.36699423e-01 -2.10806727e-01 1.17579401e+00
1.16755582e-01 7.35666335e-01 -1.16793305e-01 -6.42483473e-01
7.47188330e-01 -2.07592860e-01 9.09649432e-01 -6.24155283e-01
-4.83832389e-01 2.91588157e-01 2.83198416e-01 -1.20865142e+00
-6.56496823e-01 -7.57223606e-01 -1.10002363e+00 -9.31966662e-01
-2.55973563e-02 -1.04313605e-01 2.40295008e-01 5.92496097e-01
2.98727542e-01 9.34237381e-04 1.15136415e-01 -1.72114122e+00
-3.27835649e-01 -6.05721533e-01 -6.12503529e-01 9.53914583e-01
3.57918888e-01 -9.85598683e-01 -5.58005035e-01 5.48195243e-01] | [7.172818660736084, -0.8580766320228577] |
4a2770ac-bf51-45f2-9b7a-ee99a62826a8 | towards-targeted-change-detection-with | 2203.00049 | null | https://arxiv.org/abs/2203.00049v2 | https://arxiv.org/pdf/2203.00049v2.pdf | Towards Targeted Change Detection with Heterogeneous Remote Sensing Images for Forest Mortality Mapping | Several generic methods have recently been developed for change detection in heterogeneous remote sensing data, such as images from synthetic aperture radar (SAR) and multispectral radiometers. However, these are not well suited to detect weak signatures of certain disturbances of ecological systems. To resolve this problem we propose a new approach based on image-to-image translation and one-class classification (OCC). We aim to map forest mortality caused by an outbreak of geometrid moths in a sparsely forested forest-tundra ecotone using multisource satellite images. The images preceding and following the event are collected by Landsat-5 and RADARSAT-2, respectively. Using a recent deep learning method for change-aware image translation, we compute difference images in both satellites' respective domains. These differences are stacked with the original pre- and post-event images and passed to an OCC trained on a small sample from the targeted change class. The classifier produces a credible map of the complex pattern of forest mortality. | ['Jane Uhd Jepsen', 'Stian Normann Anfinsen', 'Luigi T. Luppino', 'Jørgen A. Agersborg'] | 2022-02-28 | null | null | null | null | ['one-class-classification'] | ['miscellaneous'] | [ 9.41661060e-01 -4.42039043e-01 1.48655698e-01 -3.68972152e-01
-5.85546136e-01 -7.11446404e-01 1.05467451e+00 -4.26725037e-02
-6.14624918e-01 9.31735456e-01 -3.98643203e-02 -3.35347563e-01
-1.73109204e-01 -1.14197123e+00 -6.84964001e-01 -9.87392664e-01
-4.91645426e-01 4.05883878e-01 1.78524390e-01 -5.01469195e-01
-9.48514193e-02 8.67432714e-01 -1.67148173e+00 3.12277168e-01
7.59144485e-01 7.60890484e-01 6.23493731e-01 8.86679709e-01
3.13243061e-01 1.58327520e-01 -3.35506588e-01 3.09780478e-01
4.97102141e-01 -3.43091488e-01 -5.55412054e-01 3.49751711e-01
8.44074726e-01 -4.20358062e-01 2.09401995e-02 1.36919653e+00
3.05635422e-01 -3.47520947e-01 6.94682539e-01 -7.88095653e-01
-2.98580140e-01 5.18674374e-01 -8.20178747e-01 6.17215037e-01
-2.64186319e-02 2.89248258e-01 6.79444849e-01 -5.91414571e-01
7.78451681e-01 1.34342778e+00 8.04821670e-01 -8.44044387e-02
-1.71811664e+00 -5.47591627e-01 1.13021307e-01 1.50441259e-01
-1.26461554e+00 -4.71686929e-01 4.49528426e-01 -7.34112024e-01
7.49195993e-01 5.32388210e-01 9.47885454e-01 8.26688707e-01
4.36395913e-01 3.52937579e-01 1.62365448e+00 -2.30082095e-01
1.98160812e-01 -4.07663822e-01 -4.94772822e-01 5.03117025e-01
3.49871069e-01 7.73204863e-01 -9.04036984e-02 -6.77691549e-02
6.78769350e-01 3.96430790e-01 -5.23716152e-01 -2.30150864e-01
-1.25253069e+00 8.06624591e-01 1.02089119e+00 3.81001353e-01
-8.03874671e-01 -1.82603031e-01 1.19169064e-01 5.69618881e-01
8.92547607e-01 2.96192944e-01 -5.96630514e-01 5.89058518e-01
-1.46504271e+00 3.84211302e-01 3.02673340e-01 2.86755357e-02
9.28187966e-01 8.94623846e-02 3.32060784e-01 5.92176676e-01
3.26076269e-01 1.23363185e+00 3.82505447e-01 -3.87754440e-01
-3.28003876e-02 5.11509001e-01 1.54787883e-01 -1.16587520e+00
-5.58243036e-01 -7.50966311e-01 -1.29470861e+00 6.23124659e-01
1.69401169e-01 -1.98149532e-02 -1.13857889e+00 1.32216477e+00
3.07515711e-01 -3.23087722e-02 -8.46350193e-03 9.49619055e-01
2.09381983e-01 8.30068529e-01 3.35834399e-02 -5.15240550e-01
1.19121730e+00 -2.88957328e-01 -2.64319837e-01 -6.48889422e-01
1.93045110e-01 -4.16506976e-01 6.83462083e-01 -2.11773627e-02
-2.41320193e-01 -4.88640726e-01 -9.19460833e-01 7.40057409e-01
-7.56609917e-01 2.53037691e-01 2.71328807e-01 2.21397400e-01
-1.03316331e+00 5.49228191e-01 -7.60381520e-01 -7.00183094e-01
6.24041915e-01 6.44069016e-02 -4.05383050e-01 5.55667579e-02
-9.46830451e-01 8.47804785e-01 2.21503198e-01 4.24953163e-01
-1.27579665e+00 -5.14966667e-01 -6.88121855e-01 -1.25414208e-01
6.29587770e-02 -2.83278942e-01 7.73375511e-01 -1.71324873e+00
-9.71265376e-01 1.31655633e+00 3.43177259e-01 -6.61736906e-01
4.64476645e-01 -5.00455797e-02 -5.33653498e-01 6.12989515e-02
5.05757332e-01 5.78136802e-01 1.25957072e+00 -1.52330232e+00
-1.12102377e+00 -9.86209333e-01 -2.94461221e-01 1.63494542e-01
2.98322707e-01 9.68598351e-02 7.35486805e-01 -7.91035235e-01
2.85335362e-01 -9.24481690e-01 -2.67662883e-01 1.37763366e-01
-1.05423316e-01 6.67208374e-01 9.56152439e-01 -7.77796149e-01
5.57789087e-01 -2.05737519e+00 2.79301465e-01 -6.77195787e-02
-1.78686067e-01 5.53426981e-01 -4.84316856e-01 2.11136401e-01
-1.98219076e-01 1.03382580e-01 -1.00259781e+00 4.56028432e-01
-5.34838378e-01 4.07372266e-01 -5.83894908e-01 7.38337040e-01
3.69940430e-01 5.41751564e-01 -7.72734046e-01 6.18203264e-03
3.25761676e-01 1.68720007e-01 -9.66135263e-02 5.14106378e-02
-3.20309550e-01 6.99859142e-01 -3.82890478e-02 7.26275861e-01
1.24646568e+00 3.86740416e-01 2.92149514e-01 -2.85809841e-02
-6.99644625e-01 -1.44737080e-01 -9.09678221e-01 1.02891850e+00
-2.19589576e-01 6.20152593e-01 4.94831771e-01 -1.09815872e+00
9.37813282e-01 -2.18670126e-02 1.55841440e-01 -6.55961871e-01
-2.95895785e-01 2.17934817e-01 6.78284615e-02 -5.36657155e-01
4.31174815e-01 -6.40584230e-01 1.93123162e-01 5.32365218e-02
-1.35848269e-01 -3.55109930e-01 -1.04364552e-01 -5.43696105e-01
1.01883662e+00 7.92581141e-02 6.20114565e-01 -3.72438759e-01
5.59687972e-01 6.07133925e-01 4.48158503e-01 7.24776983e-01
-3.31135392e-01 4.45015848e-01 1.37743428e-01 -9.37309444e-01
-9.44950044e-01 -1.08981681e+00 -4.17340308e-01 1.03159428e+00
-1.88484415e-01 5.85841954e-01 -2.22281530e-01 -3.54491711e-01
2.75313854e-01 4.87623096e-01 -6.23960972e-01 -2.87677851e-02
-1.77023947e-01 -1.41238797e+00 4.71714854e-01 -1.28496826e-01
1.06236660e+00 -1.12249708e+00 -8.54744792e-01 3.09813768e-01
-8.13154727e-02 -6.83434129e-01 2.87335277e-01 3.64459366e-01
-1.11486602e+00 -1.08282566e+00 -6.60811305e-01 -7.07390308e-01
4.87554073e-01 5.69853544e-01 9.46181059e-01 -5.63315988e-01
-6.57728851e-01 1.80137515e-01 -3.96185637e-01 -1.00987807e-01
-5.36718011e-01 -4.65729088e-02 8.25669840e-02 3.74701858e-01
1.14574261e-01 -8.24324787e-01 -3.69027525e-01 1.56399518e-01
-1.07391763e+00 -1.32738858e-01 1.04970789e+00 8.53429615e-01
5.79652727e-01 3.44309121e-01 2.68197030e-01 -4.83313471e-01
1.81410789e-01 -4.86426562e-01 -9.12175655e-01 7.70115852e-02
-4.06385183e-01 -8.90712738e-02 3.65141362e-01 -2.37157598e-01
-1.02570307e+00 5.53839386e-01 3.64594549e-01 1.28892004e-01
-7.47250259e-01 9.84331369e-01 -1.80499703e-02 -1.22836113e-01
9.81945693e-01 5.30210555e-01 -8.04108474e-03 -4.50342417e-01
1.78382754e-01 8.67825747e-01 9.75137234e-01 1.97455317e-01
1.10592127e+00 9.01313007e-01 -8.73790756e-02 -1.31806910e+00
-5.96173406e-01 -3.97585630e-01 -9.56141949e-01 -3.42662275e-01
6.81551456e-01 -1.21745181e+00 1.05967656e-01 9.02280092e-01
-9.73034680e-01 -4.33279157e-01 -2.80234247e-01 5.16381741e-01
-4.28135008e-01 1.41888648e-01 9.15972963e-02 -6.80775821e-01
-3.55614394e-01 -6.04245186e-01 1.22063410e+00 1.65351987e-01
3.63727570e-01 -6.87107861e-01 5.98932445e-01 -1.04876757e-02
6.03082597e-01 5.40059507e-01 8.36345613e-01 -1.04040504e-01
-4.01201338e-01 -6.49836883e-02 -4.03310239e-01 3.83471996e-01
6.32903457e-01 1.45851359e-01 -8.25162232e-01 -4.29158002e-01
3.74373719e-02 4.68313992e-02 1.27392972e+00 7.85644710e-01
5.34288645e-01 -3.79296333e-01 -4.30239201e-01 8.96165907e-01
1.80187273e+00 1.87998131e-01 5.31574786e-01 6.28792048e-01
2.43916675e-01 4.38276768e-01 5.59599578e-01 5.76238990e-01
-5.21790683e-02 6.07515037e-01 8.26133490e-01 -4.18255299e-01
3.02312933e-02 2.67333895e-01 7.21802831e-01 -5.78072183e-02
-2.65996784e-01 2.03983814e-01 -1.13637137e+00 7.24793673e-01
-1.48482764e+00 -1.36026073e+00 -3.50067377e-01 2.24916935e+00
7.19129741e-01 -1.87097758e-01 -2.11137027e-01 -1.24880582e-01
8.64998758e-01 7.17137158e-01 -5.61897814e-01 1.56718478e-01
-6.56540096e-01 1.64501488e-01 1.04360497e+00 6.00280225e-01
-1.69857538e+00 1.03352523e+00 6.01815605e+00 2.41044357e-01
-1.51128578e+00 -5.04557490e-02 2.55504757e-01 1.60941109e-01
-3.69631834e-02 1.37903750e-01 -3.96255165e-01 9.75637361e-02
8.37173820e-01 1.41693726e-02 5.96484542e-01 5.05071461e-01
5.21173954e-01 -5.98729968e-01 -3.87061357e-01 6.38328612e-01
-5.13977334e-02 -1.07731342e+00 2.12278590e-01 5.92563115e-02
6.56729817e-01 6.51417673e-01 -8.73882137e-03 -1.06608547e-01
4.60242361e-01 -6.65626884e-01 4.35593516e-01 7.86899626e-01
9.35517251e-01 -3.36807877e-01 6.36380494e-01 2.54355162e-01
-9.59307492e-01 -1.26135185e-01 -4.36205029e-01 -4.16515738e-01
-2.26103917e-01 7.56054282e-01 -7.66520977e-01 4.46610391e-01
8.44128251e-01 1.02130234e+00 -9.05473053e-01 8.36647213e-01
-8.63668397e-02 7.10648358e-01 -3.35860521e-01 6.11992717e-01
3.31982493e-01 -3.23107630e-01 9.93949711e-01 9.07400072e-01
4.92372692e-01 -5.45095615e-02 2.43450120e-01 8.32868397e-01
2.48417839e-01 -3.45553637e-01 -1.12747896e+00 -3.13804969e-02
1.40563417e-02 1.42823100e+00 -5.02120137e-01 -2.86941707e-01
-3.43166962e-02 1.08056307e+00 -2.27143571e-01 3.21437567e-01
-3.92990798e-01 -4.86240238e-02 8.04889441e-01 2.60206044e-01
4.87149924e-01 -2.89652586e-01 1.82561725e-01 -1.19544303e+00
-2.30285570e-01 -9.36954260e-01 1.56650081e-01 -9.28299129e-01
-1.05155349e+00 5.57964146e-01 1.06015429e-01 -1.27450955e+00
-5.72012365e-02 -5.04604876e-01 -5.56256413e-01 8.40693891e-01
-1.95546436e+00 -1.40090084e+00 -6.79485917e-01 3.93761247e-01
3.34830940e-01 -3.50516647e-01 9.58913982e-01 -1.09741360e-01
-3.53851467e-01 -5.43096721e-01 7.37188101e-01 -1.94747925e-01
5.90881944e-01 -1.11118519e+00 3.03509951e-01 1.23391879e+00
-1.18771762e-01 -3.98520380e-01 7.57059157e-01 -9.26847100e-01
-1.04445732e+00 -1.63609540e+00 7.29884148e-01 3.71234357e-01
7.17677414e-01 1.38789743e-01 -7.47569382e-01 7.63347983e-01
-1.43230557e-01 1.11564353e-01 2.43363753e-01 -3.32347184e-01
-1.71352684e-01 -6.91235721e-01 -1.27753747e+00 3.50636572e-01
7.19932795e-01 -3.90089124e-01 -6.37050331e-01 5.97819030e-01
3.68386856e-03 2.43860543e-01 -4.61027026e-01 3.96024436e-01
5.03701925e-01 -8.64590168e-01 8.83734703e-01 -5.03794670e-01
5.34433663e-01 -5.06799817e-01 -7.10346878e-01 -1.84676516e+00
-7.97116876e-01 -1.07511371e-01 9.61251795e-01 8.83871138e-01
-2.87881382e-02 -5.97302020e-01 8.40769485e-02 -5.51449060e-01
1.23719722e-01 5.35910785e-01 -1.13564920e+00 -8.27446878e-01
7.73534831e-03 1.29771873e-01 2.37827644e-01 7.83656418e-01
-8.82783055e-01 1.80955514e-01 -2.77014792e-01 9.23203111e-01
6.22818351e-01 3.45169455e-01 7.79963076e-01 -1.88126922e+00
2.49192595e-01 -3.14797878e-01 -5.24146020e-01 -1.76118866e-01
2.38560047e-02 -9.14617956e-01 1.89138263e-01 -1.27298689e+00
1.09599970e-01 -6.69542030e-02 1.35918409e-01 7.42917478e-01
6.54639974e-02 3.60309988e-01 -1.00043952e-01 6.08983040e-01
2.48063922e-01 7.83628881e-01 8.19853961e-01 -5.87839663e-01
-9.62569267e-02 3.12642381e-02 -1.46013901e-01 7.29986370e-01
8.29596341e-01 -8.09602201e-01 3.56538296e-01 -3.28877240e-01
-7.99683109e-02 4.46373373e-02 1.08244801e+00 -1.41455913e+00
-3.68540674e-01 -5.71056187e-01 4.35358018e-01 -7.06682682e-01
-5.25100827e-02 -1.02057254e+00 5.33593297e-01 1.03993237e+00
-1.58626214e-01 -1.70936882e-01 2.74380028e-01 6.83729112e-01
-2.69815266e-01 -2.62438133e-02 1.25564933e+00 -4.10496056e-01
-9.23628390e-01 2.84928292e-01 -1.03767943e+00 -4.50176418e-01
7.63913631e-01 2.30732635e-01 -4.36534584e-01 -1.72003433e-01
-6.47637844e-01 -2.13781744e-01 5.36955476e-01 2.85831451e-01
4.42012787e-01 -1.07134390e+00 -1.31496203e+00 7.29419887e-01
1.79972291e-01 -2.61576682e-01 1.90066844e-01 8.09027195e-01
-8.81103516e-01 2.25744769e-01 -9.47700560e-01 -8.88868988e-01
-1.41380060e+00 3.70398194e-01 8.09121609e-01 -1.86047092e-01
-4.73928124e-01 4.47693199e-01 -5.88965975e-02 -7.84827888e-01
-6.55748725e-01 -5.26458681e-01 -1.92491978e-01 4.94913965e-01
5.67586124e-01 -1.62565455e-01 -2.67434940e-02 -1.02247441e+00
-5.20280480e-01 4.68082607e-01 1.62553117e-01 -2.65931368e-01
1.83940125e+00 -2.66811401e-01 -4.76441324e-01 5.08793116e-01
7.18348324e-01 -3.47583055e-01 -1.27029479e+00 -4.13753897e-01
3.35986912e-02 -6.07874513e-01 5.46944022e-01 -1.00930119e+00
-1.05688131e+00 6.83084548e-01 1.42410553e+00 3.26702654e-01
1.36729467e+00 -2.00780943e-01 -6.08152747e-02 4.65122968e-01
2.00052336e-01 -8.78505349e-01 -6.51907802e-01 3.80944580e-01
1.37459731e+00 -1.42182994e+00 1.88565716e-01 2.68700063e-01
-2.49526903e-01 1.21503544e+00 1.71464141e-02 -1.65494546e-01
6.46882653e-01 5.65197989e-02 1.91686213e-01 -2.21481308e-01
-4.62430418e-01 -8.05164933e-01 -1.72356442e-01 1.03266132e+00
-1.13622649e-02 6.78536832e-01 -6.39524832e-02 -2.82394588e-01
-3.13874260e-02 1.73512608e-01 5.51298559e-01 7.89345086e-01
-9.24070477e-01 -5.10298252e-01 -1.01985002e+00 6.08810961e-01
1.85688064e-01 -1.96552336e-01 -7.27265298e-01 5.89155734e-01
2.06901461e-01 6.98157370e-01 4.24879998e-01 -3.09710950e-01
7.63671920e-02 -1.48603901e-01 2.06046984e-01 -4.49217469e-01
-4.63303238e-01 7.23843873e-02 -2.44613051e-01 -2.70177066e-01
-9.68420863e-01 -1.09513366e+00 -4.54970777e-01 -2.67423421e-01
-1.10311404e-01 -3.42733532e-01 7.13649929e-01 8.43487442e-01
1.85704440e-01 3.39436293e-01 9.12545741e-01 -1.28700256e+00
-4.74202991e-01 -1.36289680e+00 -1.08401787e+00 3.47588621e-02
7.11784959e-01 -5.12978256e-01 -5.82740188e-01 6.08929731e-02] | [9.662151336669922, -1.6545783281326294] |
e67bf2e0-733f-4bbb-8ac4-d6d80752a6dd | twitter-bot-detection-using-diversity | null | null | https://aclanthology.org/W19-7401 | https://aclanthology.org/W19-7401.pdf | Twitter Bot Detection using Diversity Measures | null | ['Dijana Kosmajac', 'Vlado Keselj'] | 2019-09-01 | null | null | null | ws-2019-9 | ['twitter-bot-detection'] | ['miscellaneous'] | [-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] | [-7.326053142547607, 3.56752347946167] |
363bbd52-1190-443c-bceb-a408dcc023fa | learning-a-probabilistic-latent-space-of | 1610.07584 | null | http://arxiv.org/abs/1610.07584v2 | http://arxiv.org/pdf/1610.07584v2.pdf | Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modeling | We study the problem of 3D object generation. We propose a novel framework,
namely 3D Generative Adversarial Network (3D-GAN), which generates 3D objects
from a probabilistic space by leveraging recent advances in volumetric
convolutional networks and generative adversarial nets. The benefits of our
model are three-fold: first, the use of an adversarial criterion, instead of
traditional heuristic criteria, enables the generator to capture object
structure implicitly and to synthesize high-quality 3D objects; second, the
generator establishes a mapping from a low-dimensional probabilistic space to
the space of 3D objects, so that we can sample objects without a reference
image or CAD models, and explore the 3D object manifold; third, the adversarial
discriminator provides a powerful 3D shape descriptor which, learned without
supervision, has wide applications in 3D object recognition. Experiments
demonstrate that our method generates high-quality 3D objects, and our
unsupervisedly learned features achieve impressive performance on 3D object
recognition, comparable with those of supervised learning methods. | ['Chengkai Zhang', 'Jiajun Wu', 'William T. Freeman', 'Tianfan Xue', 'Joshua B. Tenenbaum'] | 2016-10-24 | learning-a-probabilistic-latent-space-of-1 | http://papers.nips.cc/paper/6096-learning-a-probabilistic-latent-space-of-object-shapes-via-3d-generative-adversarial-modeling | http://papers.nips.cc/paper/6096-learning-a-probabilistic-latent-space-of-object-shapes-via-3d-generative-adversarial-modeling.pdf | neurips-2016-12 | ['3d-point-cloud-linear-classification', '3d-object-recognition', 'unsupervised-3d-point-cloud-linear-evaluation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 2.76430254e-03 3.36083382e-01 1.39632933e-02 -1.37088194e-01
-7.58712471e-01 -7.15640128e-01 8.45518291e-01 -6.72299206e-01
3.26408327e-01 5.51095247e-01 1.32398427e-01 -5.74953035e-02
1.42452091e-01 -1.17385876e+00 -9.87425864e-01 -6.97061956e-01
1.69410303e-01 9.17224646e-01 1.16797522e-01 1.83907434e-01
-2.62953155e-02 1.25770354e+00 -1.14962149e+00 -2.28854418e-01
7.55362868e-01 1.24735010e+00 -1.25825420e-01 4.49480802e-01
-2.92686820e-01 3.98005426e-01 -8.16392362e-01 -2.67457783e-01
5.96886814e-01 -3.82328957e-01 -5.86298406e-01 4.86542553e-01
2.10930854e-01 -4.50963080e-01 -6.82047129e-01 8.48235071e-01
3.93780857e-01 -9.61255133e-02 1.42592788e+00 -1.37953758e+00
-1.39799917e+00 1.92438260e-01 -1.83912084e-01 -4.55498278e-01
3.83841157e-01 4.03673500e-01 5.55306852e-01 -1.08204472e+00
7.93887913e-01 1.56836629e+00 5.78530192e-01 7.98354447e-01
-1.44050896e+00 -5.53478539e-01 -1.55598328e-01 -4.84863579e-01
-1.40479767e+00 -7.87969306e-02 1.28397787e+00 -7.23487258e-01
5.58806777e-01 -1.36190765e-02 9.59171593e-01 1.40498424e+00
2.72443473e-01 9.49048281e-01 1.05043209e+00 -2.23827213e-01
3.25414121e-01 4.80083562e-02 -5.07430494e-01 7.63101220e-01
1.46482453e-01 3.49954933e-01 -1.10328563e-01 -1.63871139e-01
1.42222250e+00 1.72738656e-01 -1.20268315e-01 -9.74694788e-01
-1.15319312e+00 8.97691071e-01 6.81792438e-01 8.27191696e-02
-4.75605577e-01 2.91920900e-01 -7.52938539e-02 -1.01373494e-01
5.55786014e-01 5.21400273e-01 1.75952405e-01 3.00467104e-01
-5.59829056e-01 5.23276567e-01 6.73425496e-01 1.42094731e+00
6.73875749e-01 5.12004435e-01 -4.78007555e-01 6.03393793e-01
5.65306306e-01 1.12056708e+00 1.82208553e-01 -9.39518094e-01
1.97016299e-01 7.07121074e-01 1.16085067e-01 -7.38417268e-01
9.02706683e-02 -2.48183981e-01 -1.11361158e+00 5.24720669e-01
1.53010860e-01 1.03774294e-01 -1.30279136e+00 1.59179568e+00
4.19998467e-01 -7.44144316e-04 1.44651309e-01 9.51726913e-01
1.06628692e+00 5.07597625e-01 -2.10140467e-01 2.97770292e-01
8.28388035e-01 -4.58357453e-01 -2.32995689e-01 4.60581146e-02
-1.52211457e-01 -5.16534567e-01 8.80779207e-01 -2.10802689e-01
-1.24349988e+00 -5.87781012e-01 -1.11535013e+00 1.65792853e-01
-3.33962351e-01 -3.27567279e-01 6.29534543e-01 6.35013402e-01
-8.60665262e-01 3.95731717e-01 -7.66628444e-01 2.41416276e-01
1.15302169e+00 2.09083736e-01 -3.58438402e-01 -1.53538316e-01
-8.96554172e-01 7.75453508e-01 3.78045559e-01 -1.06483832e-01
-1.56262064e+00 -6.88948631e-01 -1.16867697e+00 -2.02823803e-01
-2.81978678e-03 -1.12127411e+00 9.60806906e-01 -4.44123268e-01
-1.67802262e+00 1.20479417e+00 1.78246424e-01 1.62031315e-03
5.86747110e-01 7.36379921e-02 4.59897891e-02 1.32562548e-01
1.54406786e-01 8.13860893e-01 1.32344222e+00 -1.62773919e+00
8.03594068e-02 -4.37197208e-01 6.02043718e-02 1.18833534e-01
1.00799307e-01 -6.28228247e-01 -2.77928352e-01 -9.51772034e-01
2.91378647e-01 -9.47135508e-01 -3.83771330e-01 3.34269583e-01
-7.69233346e-01 -2.71572202e-01 9.74556506e-01 2.57395040e-02
-2.03870307e-03 -2.05082846e+00 3.91485393e-01 3.39974284e-01
5.15091717e-01 1.17305592e-01 -2.09278092e-01 8.16006660e-02
6.56636059e-02 3.55924189e-01 -4.60464150e-01 -2.01486200e-01
2.68305004e-01 1.75262198e-01 -5.37367821e-01 3.53290915e-01
8.45234036e-01 1.67320633e+00 -1.01924574e+00 -3.83399904e-01
3.02057445e-01 6.06027126e-01 -4.47218925e-01 6.74673975e-01
-3.95402253e-01 6.32833064e-01 -9.69072640e-01 8.65887523e-01
7.55445540e-01 -2.23385647e-01 -5.44125259e-01 1.76088661e-02
5.12657046e-01 1.03657007e-01 -9.18087900e-01 1.79477036e+00
-2.90882736e-01 2.18464255e-01 -3.72347862e-01 -8.96068692e-01
1.44366133e+00 1.88378945e-01 3.85983199e-01 -2.00310335e-01
1.22387163e-01 1.96682304e-01 -6.13039374e-01 -1.53958350e-01
1.03550866e-01 -2.90678084e-01 -3.33110929e-01 6.38899565e-01
3.39115560e-01 -1.24821675e+00 -5.15393078e-01 6.39868900e-02
9.11158800e-01 4.44520563e-01 -8.56632218e-02 -1.21575341e-01
1.91096112e-01 -2.78408170e-01 3.91802907e-01 8.21181059e-01
-6.38418719e-02 8.13222647e-01 4.64913905e-01 -4.13080961e-01
-1.18930852e+00 -1.90391719e+00 -1.71935335e-01 -6.38315231e-02
3.38095367e-01 2.93256402e-01 -5.20272970e-01 -1.12059295e+00
5.34732461e-01 5.27919292e-01 -7.82982707e-01 -5.04266083e-01
-5.09830475e-01 -2.68771112e-01 4.61176455e-01 6.78707898e-01
3.54236156e-01 -1.20874870e+00 -1.59347534e-01 1.08385079e-01
2.49095574e-01 -9.63547409e-01 -6.14969909e-01 4.97667305e-03
-1.04091442e+00 -8.64476085e-01 -1.11224651e+00 -8.48005831e-01
9.26787019e-01 9.49126333e-02 1.36623871e+00 -2.64104366e-01
-3.63977939e-01 6.45402849e-01 -1.38493240e-01 -6.70655608e-01
-7.95016229e-01 -7.84264505e-02 1.23969175e-01 -1.87886685e-01
1.70111701e-01 -9.68768775e-01 -4.58570272e-01 2.33371571e-01
-9.45523918e-01 -1.07549168e-02 7.15571344e-01 1.02151644e+00
1.02406204e+00 -1.53938949e-01 9.15652931e-01 -7.22664356e-01
2.66359895e-01 -3.71636271e-01 -6.88828826e-01 -9.68249291e-02
-3.11720520e-01 2.45068014e-01 5.07276714e-01 -6.53874576e-01
-8.11625361e-01 1.29888043e-01 -2.46449709e-01 -1.08977425e+00
-4.28181261e-01 -2.10955441e-01 -6.06218696e-01 -8.79925638e-02
6.28688395e-01 5.58100402e-01 2.47971937e-01 -3.96646500e-01
7.89283514e-01 3.31803173e-01 6.04913652e-01 -8.25074434e-01
1.59745836e+00 5.73848367e-01 2.69929647e-01 -4.98066157e-01
-7.88412333e-01 1.67355642e-01 -9.30154026e-01 -1.38104558e-01
9.13920283e-01 -8.02807093e-01 -5.23697138e-01 6.20929599e-01
-1.22034943e+00 -2.66009986e-01 -9.58616734e-01 2.41417661e-01
-1.09056854e+00 -3.90250199e-02 -4.63451147e-01 -7.57366002e-01
-3.93756241e-01 -1.19236875e+00 1.38726294e+00 1.95528582e-01
-1.94915514e-02 -8.78800094e-01 -2.35736936e-01 7.11033940e-02
2.46967807e-01 9.03373539e-01 1.17263293e+00 -4.13292438e-01
-1.15720034e+00 -3.68093491e-01 -9.01190192e-02 4.95081037e-01
4.99208063e-01 -3.53260010e-01 -9.41135108e-01 -6.55513629e-02
3.24907452e-02 -6.75770760e-01 4.47717398e-01 3.58536065e-01
1.47821677e+00 -2.93655246e-01 -4.10216779e-01 7.75068343e-01
1.15776956e+00 8.34904686e-02 6.11807048e-01 -2.09189281e-01
8.92801106e-01 1.49642244e-01 1.06430881e-01 2.14631110e-01
1.52686819e-01 4.98056769e-01 7.08295166e-01 -1.97154194e-01
-4.45907235e-01 -7.58641243e-01 -4.89827804e-02 5.43020666e-01
1.45809157e-02 -6.66560978e-02 -6.45606458e-01 5.04714370e-01
-1.29221082e+00 -7.20666051e-01 4.38639909e-01 1.99752736e+00
7.87575603e-01 3.13655615e-01 1.44485131e-01 9.53506585e-03
5.67992508e-01 8.23035836e-02 -9.12796140e-01 9.00702924e-02
-8.38457868e-02 4.91577208e-01 9.75383595e-02 6.91915229e-02
-9.57112670e-01 8.44727278e-01 6.75644922e+00 7.12725759e-01
-9.86386359e-01 -2.15838075e-01 6.20748401e-01 1.08767100e-01
-7.38288164e-01 -4.77765977e-01 -7.67822862e-01 3.82192612e-01
2.34477907e-01 -3.41324657e-01 1.58673376e-01 1.13127601e+00
-3.20127219e-01 6.76225424e-01 -1.50605333e+00 1.03399575e+00
2.09734023e-01 -1.69712794e+00 6.70536458e-01 3.29417467e-01
9.73939002e-01 -2.63961107e-01 2.26947233e-01 2.59558290e-01
6.37567997e-01 -1.33306098e+00 1.08127952e+00 8.59294832e-01
1.07839704e+00 -9.83099997e-01 4.09500271e-01 3.30573499e-01
-8.48792970e-01 5.05667448e-01 -3.67543042e-01 6.13568902e-01
2.95731723e-01 7.22311914e-01 -9.42080915e-01 5.65151870e-01
5.22578180e-01 7.59792626e-01 -1.62909105e-01 8.59184861e-01
-4.66527849e-01 2.82185376e-01 -3.13024372e-01 -1.20714456e-01
2.33997405e-01 -1.38201684e-01 1.03939867e+00 7.72135258e-01
3.58997852e-01 1.30581483e-01 2.51825988e-01 1.72305727e+00
-5.68239808e-01 -3.31650794e-01 -1.25750411e+00 -1.74284652e-01
4.92142320e-01 9.06155944e-01 -5.25387466e-01 -1.16376281e-01
2.55294126e-02 9.15615678e-01 2.62997031e-01 2.56948829e-01
-7.08690882e-01 -4.63659197e-01 6.29709125e-01 2.46824883e-02
5.58841050e-01 -3.23069155e-01 -3.82431388e-01 -9.93258238e-01
1.57317758e-01 -6.47105336e-01 -3.01967412e-01 -8.47505510e-01
-1.83038986e+00 6.31598234e-01 -2.31190860e-01 -1.41920245e+00
-3.53894919e-01 -5.93107879e-01 -4.62190479e-01 1.01793122e+00
-1.34043086e+00 -1.39748406e+00 -3.01802397e-01 5.73251963e-01
2.78632641e-01 -4.11443770e-01 1.03088903e+00 -1.07983291e-01
9.09338519e-02 5.96594274e-01 -1.63911536e-01 3.62510532e-01
2.22911268e-01 -1.35147309e+00 6.81383193e-01 4.13700730e-01
2.96474755e-01 1.50022849e-01 1.02963835e-01 -5.67366898e-01
-1.58026588e+00 -1.51107848e+00 2.29845107e-01 -1.04390180e+00
2.49960467e-01 -6.67186558e-01 -8.94806325e-01 5.39668500e-01
-3.43711644e-01 3.17290246e-01 5.02714097e-01 -3.01906973e-01
-5.29892027e-01 5.20970933e-02 -1.45934010e+00 5.59145629e-01
1.37052536e+00 -5.35630584e-01 -8.28901947e-01 4.24677223e-01
1.03052354e+00 -7.88795114e-01 -1.25914502e+00 5.26638210e-01
3.96301210e-01 -6.19128704e-01 1.41656256e+00 -7.29581177e-01
6.80991769e-01 -2.00021833e-01 -1.78199515e-01 -1.49158323e+00
-3.94953758e-01 -5.58713019e-01 -5.28678119e-01 1.16629684e+00
1.80088297e-01 -5.53292155e-01 9.61602449e-01 3.68865788e-01
-4.79217023e-01 -8.87556732e-01 -9.01375532e-01 -1.14880037e+00
4.75979149e-01 -3.48110795e-01 1.20798123e+00 6.41804695e-01
-8.97795975e-01 1.54160634e-01 -1.74141765e-01 1.69332698e-01
1.02407169e+00 5.56851029e-01 9.74738479e-01 -1.20846844e+00
-1.53825849e-01 -5.41541576e-01 -8.34209800e-01 -1.45636594e+00
4.76700097e-01 -1.25413609e+00 2.37758949e-01 -1.42350864e+00
-2.30818614e-03 -9.69830096e-01 -6.86549023e-02 3.14654261e-01
3.77906230e-03 4.93141770e-01 8.40997919e-02 2.94043571e-01
-1.84846759e-01 1.21760535e+00 2.00904703e+00 -5.54656148e-01
-1.66551545e-01 1.12610623e-01 -7.38902688e-01 5.07131100e-01
2.73557782e-01 -3.99529457e-01 -2.96671003e-01 -4.27735329e-01
-4.87525612e-01 -5.42345922e-03 6.47309899e-01 -6.51455164e-01
-4.21637386e-01 -3.06892008e-01 1.01874065e+00 -8.28434467e-01
5.07301688e-01 -7.35759854e-01 1.07470572e-01 2.59338081e-01
7.12536797e-02 -6.03468657e-01 5.62389940e-02 6.60967171e-01
-3.13272849e-02 -2.58462206e-02 1.09131563e+00 -2.63924927e-01
-2.18600303e-01 9.38021779e-01 -6.21112287e-02 2.20560700e-01
1.17007768e+00 -2.84709871e-01 -1.46542601e-02 -1.66644499e-01
-7.08670139e-01 4.20614444e-02 7.83561110e-01 6.07285142e-01
9.73511279e-01 -2.21242356e+00 -7.27656543e-01 4.99900103e-01
1.92333713e-01 7.26155400e-01 -8.05556923e-02 -8.63560513e-02
-2.92157412e-01 3.08003753e-01 -2.20511183e-01 -1.13709831e+00
-2.62325078e-01 6.36471570e-01 3.68632764e-01 2.24143676e-02
-1.00745070e+00 8.38758707e-01 4.70680624e-01 -6.71082556e-01
8.20609629e-02 -9.65999067e-02 2.69906878e-01 -5.01711905e-01
1.64925277e-01 -5.40227294e-02 -2.28338465e-01 -4.43660021e-01
-2.35323399e-01 5.94960630e-01 2.51555592e-01 1.58177897e-01
1.59754729e+00 3.03728670e-01 2.15140238e-01 3.92459005e-01
1.21143150e+00 -2.60413051e-01 -1.60816598e+00 -3.69673073e-01
-3.63455087e-01 -6.97749257e-01 -1.82923973e-01 -4.26706970e-01
-1.13926387e+00 6.10304236e-01 3.25913668e-01 2.34400839e-01
7.80515313e-01 6.65161669e-01 8.13733995e-01 3.08866084e-01
4.85861748e-01 -3.97307217e-01 5.70028722e-01 5.43513596e-01
1.33618891e+00 -1.07746053e+00 -2.89596111e-01 -4.30105984e-01
-2.55286336e-01 8.52916896e-01 5.63032627e-01 -6.61217630e-01
7.65304625e-01 1.94192246e-01 -6.54366985e-02 -3.31749648e-01
-3.60519856e-01 -3.10289171e-02 6.34779811e-01 1.26758051e+00
-1.44913986e-01 2.38272727e-01 6.02534473e-01 6.27173543e-01
-3.38689476e-01 -1.29548237e-01 2.21906409e-01 6.02284431e-01
4.67763236e-03 -1.09395969e+00 -3.28945249e-01 4.67089564e-01
-3.13309878e-02 2.79780716e-01 -3.96532565e-01 8.27354789e-01
1.75275840e-02 2.09313974e-01 1.56731009e-01 -1.63661882e-01
5.55082917e-01 1.09344922e-01 9.08690453e-01 -7.77964294e-01
2.00050533e-01 -2.22733632e-01 -5.28062642e-01 -4.24630165e-01
-1.93962529e-01 -5.48765957e-01 -9.32466328e-01 2.59021632e-02
-1.80192530e-01 -1.20976102e-03 6.15623534e-01 8.19560945e-01
4.30934995e-01 4.72311288e-01 1.16068232e+00 -1.36532903e+00
-6.03847921e-01 -7.70246983e-01 -7.35081553e-01 5.22631407e-01
2.96351701e-01 -1.04365385e+00 -3.97198647e-01 1.54049143e-01] | [8.879555702209473, -3.653654098510742] |
68ad000e-91c5-4b7f-9859-91cc2b74135e | key-instance-selection-for-unsupervised-video | 1906.07851 | null | https://arxiv.org/abs/1906.07851v2 | https://arxiv.org/pdf/1906.07851v2.pdf | Key Instance Selection for Unsupervised Video Object Segmentation | This paper proposes key instance selection based on video saliency covering objectness and dynamics for unsupervised video object segmentation (UVOS). Our method takes frames sequentially and extracts object proposals with corresponding masks for each frame. We link objects according to their similarity until the M-th frame and then assign them unique IDs (i.e., instances). Similarity measure takes into account multiple properties such as ReID descriptor, expected trajectory, and semantic co-segmentation result. After M-th frame, we select K IDs based on video saliency and frequency of appearance; then only these key IDs are tracked through the remaining frames. Thanks to these technical contributions, our results are ranked third on the leaderboard of UVOS DAVIS challenge. | ['Donghyeon Cho', 'Sungil Kang', 'Sungeun Hong', 'Jiwon Kim'] | 2019-06-18 | null | null | null | null | ['unsupervised-video-object-segmentation'] | ['computer-vision'] | [ 2.28217468e-01 -4.74569723e-02 -7.85573781e-01 3.35632786e-02
-2.34000549e-01 -6.42722785e-01 3.53961915e-01 3.53211671e-01
-5.75383782e-01 5.16537368e-01 1.21720187e-01 3.77405405e-01
-1.72828436e-01 -2.89582670e-01 -6.63826704e-01 -3.99044156e-01
-4.42146510e-01 2.21250653e-01 1.06965232e+00 1.65851519e-01
5.66842735e-01 6.45426810e-01 -1.66204202e+00 5.43284602e-02
7.90694177e-01 1.06991124e+00 5.24400949e-01 7.82329500e-01
-7.26960301e-02 6.30120337e-01 -3.15824568e-01 -9.01760757e-02
2.54292548e-01 -3.97093654e-01 -1.07619762e+00 4.68015313e-01
7.20305979e-01 -4.89119440e-01 -6.52348042e-01 1.13173938e+00
-1.14436604e-01 5.96633434e-01 4.85197812e-01 -1.66349900e+00
-3.00830364e-01 5.50523758e-01 -5.84832013e-01 9.33565974e-01
3.31193686e-01 -1.04159703e-02 1.07329977e+00 -9.96376812e-01
1.03514576e+00 8.70343149e-01 2.72091001e-01 5.20158708e-01
-9.52321649e-01 -3.15146387e-01 7.71878660e-01 5.31224430e-01
-1.09625077e+00 -4.42843735e-01 6.83787465e-01 -4.72208232e-01
5.00442624e-01 2.93262392e-01 1.01653039e+00 5.54706275e-01
-1.69651166e-01 1.38290167e+00 4.92115468e-01 -8.37314799e-02
3.46718848e-01 -5.67753613e-02 4.50547308e-01 6.93632483e-01
1.77504420e-01 -2.06746683e-01 -6.52711451e-01 3.17780413e-02
7.11432397e-01 1.28595203e-01 -1.15492091e-01 -4.17909324e-01
-1.41509092e+00 5.83539844e-01 3.31228673e-01 2.09450096e-01
-5.25918126e-01 2.25134388e-01 1.18702658e-01 -9.93680134e-02
3.34849238e-01 1.93453878e-01 -3.01370382e-01 -1.87307168e-02
-1.26418209e+00 4.39880252e-01 2.09771723e-01 1.28799522e+00
1.00144112e+00 -1.68590248e-01 -4.79317725e-01 5.06516635e-01
2.00380385e-03 2.55080044e-01 2.82554209e-01 -1.18570006e+00
4.29076105e-02 5.46164334e-01 3.37875366e-01 -1.14511418e+00
-2.61263937e-01 -1.51510313e-01 -1.63324729e-01 -1.49103671e-01
1.25646472e-01 1.01154841e-01 -1.19073629e+00 1.50461519e+00
6.13012493e-01 8.31640124e-01 -1.56898439e-01 1.22722197e+00
9.56477106e-01 4.70081717e-01 3.24450374e-01 -5.93641937e-01
1.30712306e+00 -1.15034366e+00 -4.40621972e-01 -2.02187579e-02
7.93801248e-02 -5.73605716e-01 4.91439700e-01 5.83443865e-02
-1.27392244e+00 -6.55079722e-01 -7.05955088e-01 3.51200223e-01
-1.75080106e-01 1.48844570e-01 5.57238460e-01 1.55612737e-01
-1.20208216e+00 5.92642844e-01 -7.90157557e-01 -5.38576245e-01
5.89851975e-01 4.06505972e-01 1.26823857e-02 2.97935486e-01
-9.50320005e-01 2.87528574e-01 4.61960047e-01 -3.12522858e-01
-1.24580050e+00 -4.66602892e-01 -6.15407288e-01 -1.59068242e-01
6.21625781e-01 -5.14317930e-01 1.06246197e+00 -1.28025699e+00
-8.46403480e-01 8.35564494e-01 -5.85392058e-01 -7.82952666e-01
4.22290742e-01 -2.49687508e-01 -2.58198410e-01 8.27883780e-01
5.11333823e-01 1.20809758e+00 1.13474667e+00 -1.25318503e+00
-1.29814482e+00 -5.77261485e-02 3.09447050e-01 3.32588732e-01
-3.89862537e-01 3.34577322e-01 -1.07654202e+00 -7.26413071e-01
5.43697834e-01 -9.53875959e-01 -2.87408590e-01 -3.12879950e-01
-4.19298112e-01 -3.42551053e-01 1.11014676e+00 -6.56337261e-01
1.27794909e+00 -2.09909439e+00 3.72475892e-01 3.97031270e-02
5.03675461e-01 7.68721253e-02 2.27758866e-02 -1.24253921e-01
1.25162795e-01 1.06578708e-01 1.45233748e-02 -2.38565542e-02
-3.60727042e-01 -3.22239786e-01 -1.83506325e-01 4.00447547e-01
2.94958860e-01 8.62247884e-01 -1.16081107e+00 -9.40397084e-01
2.43636027e-01 -9.23513994e-02 -4.74311680e-01 1.02028765e-01
-4.65343088e-01 3.98770392e-01 -8.37563157e-01 7.55848110e-01
5.39587796e-01 -2.32793108e-01 -2.50714302e-01 -4.43070918e-01
-3.57086629e-01 -1.58534557e-01 -1.22896075e+00 1.38528466e+00
3.59384358e-01 7.00363934e-01 -2.37420201e-01 -7.91111171e-01
5.07684290e-01 1.15164712e-01 9.63567078e-01 -2.51463711e-01
-5.44964708e-02 -1.67937070e-01 -2.94344336e-01 -4.62147206e-01
9.71437752e-01 6.31043196e-01 2.16308713e-01 3.47328812e-01
1.74901530e-01 3.48560184e-01 6.73955798e-01 7.85977602e-01
8.52471948e-01 3.30886513e-01 1.30331109e-03 -3.03156614e-01
4.07958001e-01 1.61623433e-01 5.43176830e-01 8.32991838e-01
-7.74013162e-01 8.49630713e-01 3.88767421e-01 -4.70589161e-01
-8.97243440e-01 -1.08225191e+00 -2.34477539e-02 1.31855869e+00
9.16574001e-01 -3.18353534e-01 -1.03788197e+00 -9.92722929e-01
4.79937391e-03 3.59907627e-01 -7.23462343e-01 -1.74831554e-01
-6.03465974e-01 -5.20800576e-02 6.28901199e-02 4.52746361e-01
3.37933809e-01 -1.35066295e+00 -8.64916623e-01 2.12025195e-01
-2.63143033e-01 -1.33820581e+00 -8.38428557e-01 -1.71093494e-01
-8.36193204e-01 -1.04130399e+00 -9.37414765e-01 -1.10313404e+00
8.20550978e-01 7.85761356e-01 8.42580676e-01 1.97079092e-01
-2.82795578e-01 5.59236944e-01 -6.28388166e-01 -4.46807481e-02
1.43136665e-01 8.25172588e-02 2.73375213e-01 2.71928519e-01
1.88547112e-02 -3.53789739e-02 -8.70888293e-01 4.09853697e-01
-7.25755870e-01 9.52650905e-02 1.62971213e-01 2.44288668e-01
9.62973118e-01 6.79109851e-03 3.93299699e-01 -5.53036928e-01
7.49571598e-04 -3.94306481e-01 -5.91383100e-01 3.39961231e-01
-1.53852105e-01 -4.54259455e-01 -5.87414810e-03 -4.28939313e-01
-6.87460124e-01 1.72202855e-01 5.88467538e-01 -7.04896152e-01
-2.29878500e-01 -7.92215019e-02 2.37216316e-02 1.73702568e-01
1.31396249e-01 2.59143591e-01 -4.81360555e-01 -6.74214289e-02
2.80926436e-01 2.58230537e-01 7.64842808e-01 -3.22377264e-01
6.23397112e-01 5.82842350e-01 -4.04189557e-01 -1.01755631e+00
-8.72206330e-01 -6.78131163e-01 -7.32523322e-01 -7.41511583e-01
1.23853004e+00 -8.65061104e-01 -5.93516886e-01 4.25857991e-01
-1.01816857e+00 -4.41594282e-03 -2.49608949e-01 7.05611229e-01
-4.15334344e-01 4.00841057e-01 -4.47163910e-01 -9.19496953e-01
-1.82192057e-01 -1.18666649e+00 1.12713766e+00 7.90591776e-01
-4.42305356e-01 -5.94245434e-01 -4.98671919e-01 1.05412714e-01
-1.45729128e-02 8.59114230e-02 3.32003474e-01 -5.06533921e-01
-1.15062714e+00 -1.05955042e-02 -3.25375706e-01 -7.04423934e-02
7.33994991e-02 2.88741589e-01 -6.71942770e-01 -2.94555604e-01
-3.95683974e-01 1.09875932e-01 1.17574310e+00 9.86561120e-01
1.16873860e+00 7.70946546e-03 -8.16798270e-01 3.69503409e-01
1.14275348e+00 3.30226332e-01 2.32008591e-01 3.89909506e-01
8.46030295e-01 4.17721391e-01 1.08527827e+00 2.98108369e-01
3.49086195e-01 7.82580316e-01 6.19602501e-01 -7.54710585e-02
-2.01018974e-01 -9.43508670e-02 3.56085479e-01 2.94741720e-01
-5.73898852e-02 -2.18998134e-01 -6.15306497e-01 9.66984212e-01
-1.86934388e+00 -1.06740916e+00 -2.11243361e-01 2.18529105e+00
4.21611696e-01 4.27647829e-01 7.05492795e-01 -3.43976825e-01
1.20947564e+00 2.92772710e-01 -6.61865830e-01 3.54405195e-01
-3.38968575e-01 -1.64527103e-01 6.09864354e-01 5.32861352e-01
-1.34530342e+00 1.46619499e+00 6.08351374e+00 7.17271507e-01
-7.01393723e-01 2.02991292e-01 8.21395338e-01 -2.62759149e-01
-2.00028285e-01 3.48063201e-01 -1.02399838e+00 6.77842736e-01
2.98265874e-01 -1.66909635e-01 3.55025560e-01 7.05762327e-01
1.33619457e-01 -6.99022889e-01 -9.47753489e-01 7.18078673e-01
2.42297992e-01 -1.55739474e+00 9.09899473e-02 -2.59297282e-01
1.10496235e+00 3.77389975e-02 9.52290073e-02 -1.80393606e-01
9.41888243e-02 -5.36182880e-01 1.24682677e+00 5.44334590e-01
2.92080104e-01 -7.92735219e-01 2.78970957e-01 -8.06123763e-02
-1.45705736e+00 -8.51151869e-02 -2.85189509e-01 1.90258622e-01
3.18368912e-01 3.67626697e-01 -4.52860028e-01 3.22613627e-01
9.17958677e-01 9.22584474e-01 -5.31720877e-01 1.38552535e+00
1.07950896e-01 6.40428543e-01 -3.41076702e-01 -8.48787725e-02
3.59308004e-01 -2.42140427e-01 9.07668114e-01 9.87177551e-01
2.53204405e-01 1.67409495e-01 7.55470037e-01 5.24983704e-01
6.60513341e-02 -1.07903898e-01 -2.09514886e-01 -2.00728364e-02
7.54647136e-01 1.15790236e+00 -1.66925955e+00 -7.68369794e-01
-2.46029451e-01 1.17602396e+00 -5.21901809e-02 4.86392647e-01
-9.28000808e-01 -1.47222400e-01 7.54128754e-01 1.79349512e-01
5.93637764e-01 -1.99330553e-01 -1.15745731e-01 -9.03870344e-01
-1.48137519e-02 -3.24966550e-01 5.55076003e-01 -7.38658011e-01
-6.43421113e-01 6.20450795e-01 1.94544882e-01 -1.17624533e+00
1.64674297e-01 3.26751843e-02 -5.22138953e-01 3.15152317e-01
-1.16165578e+00 -8.07237625e-01 -3.33485216e-01 4.50159848e-01
9.64025378e-01 -7.58220628e-02 2.66047437e-02 4.61048782e-02
-4.73464221e-01 3.04413795e-01 -1.79729596e-01 1.38674513e-01
2.61757940e-01 -1.03063416e+00 5.58394015e-01 8.35223973e-01
3.49106342e-01 4.30502474e-01 8.17454934e-01 -1.02126467e+00
-1.01487243e+00 -1.14207768e+00 7.18156993e-01 -4.83773112e-01
5.18501103e-01 1.09171057e-02 -7.11213410e-01 6.43166423e-01
4.46954276e-04 5.87185286e-02 -5.67031130e-02 -3.84082466e-01
3.08232546e-01 8.71256888e-02 -9.83079791e-01 7.00795829e-01
1.33852088e+00 -2.45211795e-01 -4.15532023e-01 4.80013996e-01
1.07150471e+00 -5.62686980e-01 -3.66105378e-01 4.27117616e-01
2.76634336e-01 -7.29910254e-01 9.55620050e-01 -7.49169052e-01
3.01486284e-01 -7.74937093e-01 1.02325222e-02 -8.41453016e-01
-5.10492444e-01 -7.29039669e-01 -1.46988690e-01 1.29672444e+00
1.45000175e-01 1.37890819e-02 1.03139853e+00 2.91455984e-01
-6.16360232e-02 -6.62107706e-01 -6.93428636e-01 -5.70732296e-01
-8.31022799e-01 -4.53529477e-01 5.12853265e-01 7.15401053e-01
-3.15254480e-01 -3.17264311e-02 -3.72082829e-01 4.45289671e-01
7.62091875e-01 3.55733365e-01 5.73491573e-01 -1.13425446e+00
4.31087101e-03 -5.20733893e-01 -6.62642717e-01 -1.06392133e+00
1.28639325e-01 -6.67753518e-01 -6.72687516e-02 -1.35381603e+00
5.14601529e-01 -3.09150070e-01 -5.48344016e-01 2.41558209e-01
-5.75154245e-01 5.24040580e-01 3.65826786e-01 2.59682387e-01
-1.30165029e+00 1.32794738e-01 1.01166451e+00 -1.39073342e-01
-4.78912026e-01 3.57338786e-02 -2.91882902e-01 7.83176064e-01
6.38502479e-01 -3.56845737e-01 -2.41767779e-01 -1.35547325e-01
-4.06814784e-01 7.51695856e-02 5.83901942e-01 -1.08315873e+00
7.37814456e-02 -5.48570633e-01 4.88022536e-01 -8.37525070e-01
2.64369100e-01 -4.89497572e-01 -3.99571322e-02 5.39734781e-01
-5.07405162e-01 3.98892015e-02 -1.24697024e-02 7.10295737e-01
-5.47209904e-02 -3.52680385e-01 7.05139399e-01 -1.08822390e-01
-1.46696985e+00 7.66955137e-01 -4.20510799e-01 5.31372130e-02
1.37694085e+00 -6.09242082e-01 5.80299497e-02 -2.64271855e-01
-1.09788859e+00 4.51807320e-01 6.87840641e-01 6.48163915e-01
6.57315612e-01 -1.35037637e+00 -4.96082813e-01 8.52600262e-02
1.11542970e-01 -1.21768817e-01 4.38221306e-01 8.83852422e-01
-2.17845216e-01 6.86485469e-02 -4.65903468e-02 -9.59789634e-01
-1.45017266e+00 5.10270238e-01 4.58863378e-03 2.88876206e-01
-7.62137890e-01 1.15660894e+00 4.25124615e-01 4.83958066e-01
3.52753252e-01 -1.63620651e-01 -4.52132344e-01 3.08862031e-01
5.76374590e-01 3.97564918e-01 -2.10434869e-01 -1.14304411e+00
-5.63555598e-01 6.81564629e-01 -3.16688478e-01 -1.10546231e-01
1.03006995e+00 -3.21128070e-01 4.59105112e-02 2.64729947e-01
1.00657856e+00 -2.66409367e-01 -1.65765870e+00 -3.52944851e-01
2.32058644e-01 -7.53066778e-01 -9.22089890e-02 -2.58683294e-01
-1.23399842e+00 2.44241178e-01 7.33589649e-01 1.70592204e-01
9.74357426e-01 5.61089814e-01 8.67832780e-01 -1.97549328e-01
4.12560195e-01 -1.36520910e+00 4.76003230e-01 5.02207339e-01
4.95708704e-01 -7.91753471e-01 5.91878295e-02 -6.12466514e-01
-8.35369051e-01 6.51053190e-01 7.86918402e-01 -3.07744920e-01
4.31201816e-01 -1.25798330e-01 -1.57708466e-01 -1.31072834e-01
-4.54956383e-01 -5.03962398e-01 4.20141846e-01 6.32770479e-01
5.75415641e-02 6.89336136e-02 -3.76297593e-01 2.59538591e-01
9.95473117e-02 -1.73241004e-01 4.33007449e-01 8.24821293e-01
-9.05308902e-01 -4.39328581e-01 -3.63611728e-01 6.37577116e-01
-3.40830952e-01 -6.47313008e-03 -3.25741827e-01 4.65858102e-01
1.55306533e-01 7.78596938e-01 3.52485329e-01 -3.54758948e-01
1.48528889e-01 -1.83139920e-01 3.91203821e-01 -5.00767946e-01
-2.57660896e-01 2.73296744e-01 -1.92224339e-01 -6.36742353e-01
-5.66178322e-01 -1.08558834e+00 -1.67840981e+00 4.61866930e-02
-3.31912160e-01 2.17613608e-01 3.11834961e-01 8.89862895e-01
3.17751080e-01 4.76625472e-01 6.86217010e-01 -1.19648540e+00
1.74347624e-01 -4.90299314e-01 -5.34474909e-01 5.57007909e-01
3.93919438e-01 -9.26933110e-01 -9.31953713e-02 4.49861526e-01] | [9.305869102478027, -0.2390822321176529] |
622625e1-edcc-44ec-a2a0-8854361b5abb | graph-convolutional-policy-for-solving-tree | 1910.08371 | null | https://arxiv.org/abs/1910.08371v2 | https://arxiv.org/pdf/1910.08371v2.pdf | Graph Convolutional Policy for Solving Tree Decomposition via Reinforcement Learning Heuristics | We propose a Reinforcement Learning based approach to approximately solve the Tree Decomposition (TD) problem. TD is a combinatorial problem, which is central to the analysis of graph minor structure and computational complexity, as well as in the algorithms of probabilistic inference, register allocation, and other practical tasks. Recently, it has been shown that combinatorial problems can be successively solved by learned heuristics. However, the majority of existing works do not address the question of the generalization of learning-based solutions. Our model is based on the graph convolution neural network (GCN) for learning graph representations. We show that the agent builton GCN and trained on a single graph using an Actor-Critic method can efficiently generalize to real-world TD problem instances. We establish that our method successfully generalizes from small graphs, where TD can be found by exact algorithms, to large instances of practical interest, while still having very low time-to-solution. On the other hand, the agent-based approach surpasses all greedy heuristics by the quality of the solution. | ['Taras Khakhulin', 'Ivan Oseledets', 'Roman Schutski'] | 2019-10-18 | null | null | null | null | ['tree-decomposition'] | ['graphs'] | [-2.78801261e-03 8.04248452e-01 -3.35891128e-01 -2.18878705e-02
-8.18802536e-01 -5.37159801e-01 3.81354451e-01 3.60120296e-01
-2.52712429e-01 9.69920278e-01 -2.61007398e-01 -6.98659956e-01
-5.44618547e-01 -1.20223927e+00 -1.01217341e+00 -7.65484869e-01
-3.27522755e-01 1.15256011e+00 9.68656614e-02 -9.33172256e-02
2.56673750e-02 3.99521172e-01 -1.08347237e+00 -1.24662668e-01
6.13513470e-01 8.07367086e-01 1.58672094e-01 8.33140910e-01
-9.61760283e-02 8.51973414e-01 -3.15575600e-01 -4.55120057e-01
2.42040321e-01 -4.28808689e-01 -1.17761850e+00 2.49568775e-01
1.30090281e-01 -2.18975216e-01 -6.01236701e-01 1.15648401e+00
2.36948475e-01 1.15072183e-01 3.02673101e-01 -1.41985297e+00
-5.31360447e-01 1.12232113e+00 -5.32738030e-01 -9.48458686e-02
1.09250590e-01 -2.04644486e-01 1.26387680e+00 -6.58041658e-03
6.88131094e-01 1.26829183e+00 5.67783535e-01 6.14321709e-01
-1.35656559e+00 -2.97978878e-01 3.62266183e-01 3.58058900e-01
-9.46042717e-01 1.12832047e-01 6.96579278e-01 -3.31964940e-02
1.04059446e+00 -6.48208931e-02 7.17166126e-01 6.13183141e-01
1.71231583e-01 8.70570600e-01 9.65866446e-01 -8.20073009e-01
4.98045206e-01 -3.74433756e-01 1.24880470e-01 1.23377585e+00
6.05149984e-01 2.15085849e-01 -1.70966640e-01 -1.20221280e-01
6.15289271e-01 -2.36574158e-01 -7.11881146e-02 -7.87627757e-01
-7.53616691e-01 1.17001235e+00 3.61618042e-01 1.17212199e-01
-2.68676490e-01 9.03593421e-01 3.45721602e-01 3.91377568e-01
2.01864108e-01 4.30891842e-01 -6.09174728e-01 4.17873040e-02
-5.53585649e-01 2.06228405e-01 1.16398823e+00 8.96789312e-01
9.17557538e-01 2.54364938e-01 1.86137438e-01 3.16290170e-01
1.67446464e-01 4.26131338e-01 1.17495582e-01 -1.07350719e+00
3.75616014e-01 4.80411291e-01 4.46015820e-02 -6.82435393e-01
-6.09179795e-01 -4.05360013e-01 -6.45359755e-01 1.06638677e-01
6.24634087e-01 -3.41136575e-01 -8.37180912e-01 1.81628621e+00
3.14721823e-01 2.65515387e-01 -4.41071391e-02 4.88305479e-01
4.82383460e-01 6.15688324e-01 -1.42160684e-01 -3.75649959e-01
9.65185583e-01 -8.93824220e-01 -5.30997813e-01 -3.80321294e-01
8.97160292e-01 -3.18880118e-02 4.69682038e-01 5.91906846e-01
-1.24257302e+00 -1.04621723e-02 -9.47552979e-01 2.00281531e-01
-2.23853156e-01 -1.19668186e-01 1.14814341e+00 8.71143401e-01
-1.43336415e+00 6.67152226e-01 -1.04600215e+00 -4.29672897e-01
3.22797120e-01 7.41212189e-01 -1.18264921e-01 -3.75955671e-01
-9.01159108e-01 9.23544586e-01 6.00948155e-01 1.89339444e-01
-1.14820063e+00 -1.95844859e-01 -1.00612438e+00 3.09796453e-01
1.06316543e+00 -7.31553972e-01 1.48662043e+00 -8.21819127e-01
-1.62789118e+00 3.78733724e-01 -2.09885426e-02 -8.71021807e-01
4.67714928e-02 2.98291147e-01 -1.26837881e-03 1.19043469e-01
-2.75266796e-01 3.82705092e-01 6.61395073e-01 -1.04836166e+00
-6.96198702e-01 -3.35184574e-01 5.92098057e-01 5.15934499e-03
-8.92179385e-02 -2.96533048e-01 -3.43826741e-01 -4.64621820e-02
-3.64734307e-02 -1.16023827e+00 -7.77752578e-01 -3.95963907e-01
-4.11253899e-01 -4.89234090e-01 2.49132797e-01 -4.11090165e-01
9.21669424e-01 -1.78357720e+00 5.14336646e-01 4.61076885e-01
4.41073447e-01 9.71519649e-02 -3.60753953e-01 6.56559944e-01
2.70038694e-02 -2.53380537e-02 -1.56890735e-01 6.35202155e-02
3.69281501e-01 6.55042171e-01 -1.96435705e-01 3.80553603e-01
4.25409116e-02 1.14685571e+00 -1.20638132e+00 -3.67212981e-01
8.56253970e-03 -1.23691015e-01 -8.36516738e-01 2.54677366e-02
-8.35332394e-01 -1.70727968e-01 -5.88348627e-01 2.55453914e-01
5.43167055e-01 -4.91095483e-01 1.10309672e+00 2.23753527e-01
3.55449975e-01 1.76717743e-01 -1.25874102e+00 1.51147485e+00
-6.34644270e-01 4.53485429e-01 1.74244896e-01 -1.57463944e+00
5.21026611e-01 1.10671632e-01 4.05588388e-01 -5.45898318e-01
-4.89044599e-02 1.38351902e-01 2.71916613e-02 -2.18598768e-01
5.30713856e-01 6.47748932e-02 -2.13202313e-01 7.45559633e-01
1.45101011e-01 -1.62440449e-01 4.91765887e-01 3.10607851e-01
1.46334434e+00 1.57642052e-01 2.91207999e-01 -6.83365315e-02
1.72217831e-01 6.34995177e-02 4.44652349e-01 9.99516666e-01
1.39603958e-01 -1.67990670e-01 1.13790405e+00 -7.14558303e-01
-7.70396471e-01 -8.61961484e-01 5.51163435e-01 9.82904732e-01
-5.94702899e-04 -5.30785680e-01 -1.05577886e+00 -8.17780137e-01
-3.44676748e-02 7.36733735e-01 -6.34612143e-01 -1.52192444e-01
-6.06932998e-01 -7.92196691e-01 2.44878620e-01 4.93637443e-01
2.26262689e-01 -1.02373171e+00 -4.63590056e-01 6.21599436e-01
2.52174847e-02 -1.30374551e+00 -3.34328383e-01 3.65075141e-01
-1.09555948e+00 -1.36929750e+00 -2.81292886e-01 -8.32285047e-01
6.90271378e-01 1.38815403e-01 1.17106628e+00 3.32467139e-01
-1.85993060e-01 7.73264349e-01 -7.20546767e-02 -2.17147514e-01
-4.88630146e-01 2.89207965e-01 -2.68552303e-01 -2.80751705e-01
1.46775320e-01 -5.66017568e-01 -1.02768220e-01 -1.13711506e-01
-6.60620749e-01 -1.78875148e-01 4.77415055e-01 1.02211618e+00
5.45325994e-01 5.64607084e-01 3.53556365e-01 -1.22123051e+00
7.84851789e-01 -2.62877494e-01 -1.20137894e+00 5.65254807e-01
-7.96767414e-01 7.00324118e-01 7.59562612e-01 -9.50924978e-02
-6.60000205e-01 3.29422355e-01 2.82288253e-01 -1.61547467e-01
1.14503130e-01 7.87445128e-01 6.66344212e-03 -2.64189065e-01
4.22576576e-01 2.88074911e-01 -2.86045037e-02 -9.16222949e-03
5.83840132e-01 -1.28562441e-02 1.57460243e-01 -9.45369661e-01
8.17562938e-01 3.05681527e-01 5.03833771e-01 -5.45703650e-01
-9.28392053e-01 -5.12655824e-02 -2.94396549e-01 4.21203002e-02
5.01261234e-01 -5.77484369e-01 -1.21122134e+00 1.92002337e-02
-1.13743973e+00 -7.51116574e-01 -2.31868029e-01 3.30465108e-01
-1.02646959e+00 5.83208084e-01 -7.35894263e-01 -1.05444288e+00
-6.94151148e-02 -9.94732261e-01 8.67699564e-01 3.24030779e-02
2.19961807e-01 -1.11549211e+00 2.39472315e-01 7.45471641e-02
1.42074704e-01 1.18601508e-01 1.53207970e+00 -5.36415517e-01
-9.70078051e-01 -6.24769479e-02 -2.22975805e-01 -1.89381644e-01
-2.19737858e-01 -1.34261206e-01 -5.61773837e-01 -3.59095752e-01
-3.57330054e-01 -4.11561012e-01 9.01922762e-01 7.93746710e-01
1.26547670e+00 -3.69779676e-01 -4.00663763e-01 2.98982769e-01
1.66393805e+00 1.89008355e-01 5.98440647e-01 2.93755531e-01
3.96160752e-01 4.04307127e-01 3.47047061e-01 3.04581761e-01
6.73091769e-01 5.56692183e-01 7.28374124e-01 1.43292427e-01
1.16965674e-01 -2.87534088e-01 3.19235563e-01 5.77000678e-01
-3.48333538e-01 -3.30244064e-01 -9.35462236e-01 4.18918908e-01
-2.24638677e+00 -9.28448856e-01 1.22940361e-01 2.13844180e+00
5.25586843e-01 2.70462185e-01 3.02307606e-01 6.91726878e-02
8.33984792e-01 9.13846120e-03 -4.20716435e-01 -9.12295818e-01
3.37882459e-01 6.51508391e-01 7.29263663e-01 7.60109067e-01
-8.64831090e-01 1.21445310e+00 6.91567707e+00 7.88120627e-01
-6.28619492e-01 -3.33710317e-03 3.77787650e-01 3.06920975e-01
-3.08319449e-01 2.42720410e-01 -6.19427502e-01 -8.22175071e-02
1.24522161e+00 -3.49785179e-01 1.33652711e+00 1.05906117e+00
-1.20379195e-01 3.60443443e-02 -1.17661512e+00 6.32044613e-01
-8.96863043e-02 -1.41284835e+00 -1.84817806e-01 3.05062920e-01
7.08053350e-01 -8.84789452e-02 -1.25310540e-01 6.11120760e-01
1.11621380e+00 -9.32971179e-01 3.24365020e-01 1.27857059e-01
3.46286654e-01 -1.07723534e+00 7.11476624e-01 3.40099365e-01
-1.12941718e+00 -3.84084016e-01 -4.61097896e-01 -1.15776911e-01
-1.61837175e-01 5.53211689e-01 -1.17873096e+00 6.97526634e-01
2.72172809e-01 3.64639431e-01 -2.23644093e-01 1.04372859e+00
-4.93009567e-01 5.98944902e-01 -3.39749932e-01 -4.00284916e-01
4.06653076e-01 -2.20584884e-01 1.49039596e-01 8.07290256e-01
2.39938259e-01 5.72353676e-02 4.65054899e-01 6.94218218e-01
-1.96161777e-01 -1.94380507e-01 -7.13175237e-01 -4.64263648e-01
2.73098558e-01 1.18156850e+00 -9.37438726e-01 -3.28190550e-02
-2.98763633e-01 5.68053663e-01 8.75454247e-01 4.81758088e-01
-8.47895503e-01 -3.87267292e-01 -2.39830744e-02 -1.91044360e-01
7.61346757e-01 -5.86934090e-01 1.20077707e-01 -8.80460441e-01
-1.05663776e-01 -9.67538357e-01 6.77678227e-01 -5.00957012e-01
-9.11634505e-01 5.35849750e-01 -3.52587476e-02 -5.45275688e-01
-6.36062920e-01 -7.96325147e-01 -4.37241703e-01 2.90966898e-01
-1.61384141e+00 -1.01499009e+00 1.73585609e-01 4.85227674e-01
2.40308881e-01 -5.53973541e-02 1.03461516e+00 -1.90824881e-01
-5.73624551e-01 4.88852322e-01 2.30321676e-01 3.45941372e-02
-1.05115153e-01 -1.59938204e+00 2.10188389e-01 7.83672154e-01
4.00044024e-01 1.58576414e-01 6.64765716e-01 -4.38966185e-01
-1.98294806e+00 -9.33763027e-01 7.57277071e-01 3.01794428e-02
9.02826846e-01 -1.84367388e-01 -3.69561076e-01 1.00704801e+00
3.43065739e-01 7.98191652e-02 1.77283213e-01 5.72326779e-01
-2.69910872e-01 -2.03017965e-01 -9.69037056e-01 6.37921512e-01
1.03411496e+00 -3.68909180e-01 -2.17352346e-01 7.70743072e-01
6.94725215e-01 -5.82814395e-01 -6.39094770e-01 -1.53943881e-01
1.64882094e-01 -6.18523240e-01 7.31880009e-01 -1.02640462e+00
4.45238173e-01 -8.47237781e-02 -1.00061342e-01 -1.49795747e+00
-3.13849419e-01 -9.81630743e-01 -6.88610673e-01 6.49517357e-01
5.29210389e-01 -7.96640217e-01 1.11173511e+00 4.13756400e-01
-2.80908830e-02 -7.52873600e-01 -1.19846308e+00 -8.79417002e-01
1.16439871e-01 -3.99160713e-01 5.74181974e-01 5.89271665e-01
1.87244892e-01 3.12219232e-01 -2.95725584e-01 4.64391232e-01
8.05977345e-01 6.79064810e-01 6.56528175e-01 -1.13170409e+00
-7.71990597e-01 -3.93709213e-01 -4.22016680e-01 -8.44702005e-01
6.33790255e-01 -1.08162975e+00 3.80603448e-02 -1.71795392e+00
2.39366829e-01 -3.75545382e-01 -3.06290001e-01 6.26099169e-01
2.06676841e-01 -3.41990650e-01 1.61937699e-01 -4.55418557e-01
-9.59942043e-01 3.46625447e-01 1.20485449e+00 -3.72310460e-01
-1.65116061e-02 1.60167456e-01 -6.97760165e-01 5.16368032e-01
8.96998227e-01 -7.42672324e-01 -4.95580792e-01 -3.27331990e-01
6.95263565e-01 5.75075686e-01 2.17957631e-01 -6.85950696e-01
3.82773042e-01 -4.18524206e-01 -2.32780159e-01 -2.60629982e-01
5.89372106e-02 -7.50876665e-01 -3.01560331e-02 8.47148061e-01
-3.34394753e-01 1.84823573e-01 1.40638828e-01 9.77417648e-01
1.83222726e-01 -5.99225283e-01 3.96577001e-01 -2.47715741e-01
-5.09956419e-01 4.89646226e-01 -5.52514315e-01 1.50015473e-01
9.63172972e-01 6.66076839e-02 -2.75499523e-01 -7.78003573e-01
-8.53768229e-01 3.35705340e-01 1.03809282e-01 -2.97802329e-01
5.00961483e-01 -1.01963902e+00 -5.39308369e-01 -1.92167059e-01
-2.89712667e-01 4.68543964e-03 2.66541932e-02 6.67331398e-01
-5.15578926e-01 4.44177121e-01 4.07667942e-02 -1.99483886e-01
-9.02974844e-01 9.83211100e-01 4.95112270e-01 -9.54113841e-01
-6.51558161e-01 4.95778054e-01 -5.43358140e-02 -3.91713530e-01
3.50372374e-01 -3.98537755e-01 5.82637601e-02 -3.82414967e-01
1.16099574e-01 2.22123966e-01 1.23492191e-02 1.96345136e-01
-1.14291817e-01 5.17798901e-01 -3.16941410e-01 2.95593776e-02
1.62593699e+00 1.52411714e-01 -1.84930027e-01 -8.19688439e-02
6.45444274e-01 -1.81261823e-01 -9.34903800e-01 -3.09480160e-01
2.74979711e-01 -1.29383029e-02 8.69966373e-02 -6.27403855e-01
-1.33962965e+00 7.91615486e-01 2.49188915e-01 6.17004395e-01
1.06696475e+00 -2.65500937e-02 7.20443130e-01 1.06510270e+00
7.82629788e-01 -1.25080657e+00 5.09494580e-02 6.25844955e-01
4.92554367e-01 -8.66299689e-01 1.50056869e-01 -4.91776168e-01
-3.70071411e-01 1.37258387e+00 3.62177521e-01 -4.12247628e-01
4.34045881e-01 3.50783437e-01 -6.16084874e-01 -1.83229893e-01
-1.08873641e+00 -3.23297024e-01 -2.37739593e-01 8.56326401e-01
-9.65031013e-02 3.09233934e-01 -1.92110449e-01 5.50192535e-01
7.11788191e-03 -4.03400138e-02 8.51380110e-01 8.45791698e-01
-4.13200766e-01 -1.40975964e+00 2.05377452e-02 3.57799292e-01
-2.48756140e-01 -9.81400758e-02 -3.56792837e-01 9.40683663e-01
-3.45290124e-01 8.29720676e-01 -1.43774495e-01 -1.37815714e-01
-2.62095151e-03 -1.35141850e-01 1.07163465e+00 -4.95441794e-01
-2.44868577e-01 -2.71480680e-01 5.50391972e-01 -6.38689280e-01
-3.64538968e-01 -3.77770811e-01 -1.31310928e+00 -4.62382078e-01
-5.27377903e-01 3.98940891e-01 4.41288561e-01 1.09321392e+00
1.86435431e-01 5.60611248e-01 6.79685175e-01 -5.68109155e-01
-8.17146719e-01 -4.26728308e-01 -8.04976344e-01 -3.15089464e-01
4.57223915e-02 -5.21606028e-01 -1.75370798e-01 -3.54860395e-01] | [5.229701519012451, 3.0222702026367188] |
664e3525-63d5-41cc-a0f0-2721026f1e76 | temporal-modulation-network-for-controllable | 2104.10642 | null | https://arxiv.org/abs/2104.10642v2 | https://arxiv.org/pdf/2104.10642v2.pdf | Temporal Modulation Network for Controllable Space-Time Video Super-Resolution | Space-time video super-resolution (STVSR) aims to increase the spatial and temporal resolutions of low-resolution and low-frame-rate videos. Recently, deformable convolution based methods have achieved promising STVSR performance, but they could only infer the intermediate frame pre-defined in the training stage. Besides, these methods undervalued the short-term motion cues among adjacent frames. In this paper, we propose a Temporal Modulation Network (TMNet) to interpolate arbitrary intermediate frame(s) with accurate high-resolution reconstruction. Specifically, we propose a Temporal Modulation Block (TMB) to modulate deformable convolution kernels for controllable feature interpolation. To well exploit the temporal information, we propose a Locally-temporal Feature Comparison (LFC) module, along with the Bi-directional Deformable ConvLSTM, to extract short-term and long-term motion cues in videos. Experiments on three benchmark datasets demonstrate that our TMNet outperforms previous STVSR methods. The code is available at https://github.com/CS-GangXu/TMNet. | ['Ming-Ming Cheng', 'Xing Sun', 'Liang Wang', 'Zhen Li', 'Jun Xu', 'Gang Xu'] | 2021-04-21 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Xu_Temporal_Modulation_Network_for_Controllable_Space-Time_Video_Super-Resolution_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Xu_Temporal_Modulation_Network_for_Controllable_Space-Time_Video_Super-Resolution_CVPR_2021_paper.pdf | cvpr-2021-1 | ['space-time-video-super-resolution'] | ['computer-vision'] | [ 1.78115919e-01 -5.34125686e-01 -2.22344577e-01 -4.02440667e-01
-7.21699178e-01 -1.52577132e-01 4.70517039e-01 -7.13973701e-01
-3.40576112e-01 7.41831243e-01 3.48133832e-01 1.49351984e-01
-1.07381061e-01 -7.40182579e-01 -7.82461107e-01 -6.63661778e-01
-3.43385972e-02 -3.87258828e-01 5.49549699e-01 -2.21945718e-01
7.56425038e-02 3.37840408e-01 -1.32159531e+00 6.40218556e-01
7.97440588e-01 1.02213061e+00 6.89570248e-01 4.77028191e-01
2.14401379e-01 9.14200187e-01 2.48698727e-03 1.33552536e-01
1.35382667e-01 -4.00641888e-01 -6.89687312e-01 -6.41843453e-02
4.09183383e-01 -7.54002750e-01 -9.48429048e-01 9.32758451e-01
3.25714409e-01 4.69444931e-01 5.50152101e-02 -7.13055491e-01
-9.45230246e-01 4.90658879e-01 -6.09693408e-01 6.04886591e-01
4.35581028e-01 4.75478530e-01 5.07522285e-01 -1.24455881e+00
8.85963917e-01 1.33989263e+00 4.06061500e-01 7.42815912e-01
-1.06183696e+00 -6.87268078e-01 3.62174988e-01 4.19115990e-01
-1.55715513e+00 -5.12420356e-01 7.13999033e-01 -3.87973607e-01
5.83253324e-01 2.85861641e-01 5.07868886e-01 1.07461190e+00
1.25496164e-01 5.91272950e-01 9.18117046e-01 1.89007238e-01
-2.84713596e-01 -6.24066651e-01 -3.93145889e-01 6.90032363e-01
-2.98568219e-01 4.25892353e-01 -6.77987397e-01 3.25288653e-01
1.59984541e+00 2.47566670e-01 -7.87306011e-01 1.33682221e-01
-1.60029197e+00 4.80022103e-01 5.53847909e-01 6.66192889e-01
-3.77535403e-01 3.56106758e-01 2.73956180e-01 5.17509393e-02
6.70844436e-01 4.99227531e-02 -3.88269067e-01 6.67418391e-02
-9.89651382e-01 -1.52158522e-04 3.22459824e-03 8.56368124e-01
7.43305624e-01 3.12484890e-01 -7.35632241e-01 7.10131228e-01
1.79538593e-01 2.05771878e-01 5.27689159e-01 -1.32588911e+00
5.35470963e-01 9.31894705e-02 4.17220920e-01 -1.00218749e+00
-3.81454937e-02 -1.89020216e-01 -1.09075642e+00 -3.48399356e-02
1.90155923e-01 1.12641811e-01 -9.01348472e-01 1.51552582e+00
3.35217267e-01 9.71047640e-01 -1.01867117e-01 1.52802145e+00
1.00867307e+00 9.22797561e-01 4.22392152e-02 -4.80242223e-01
1.14995241e+00 -9.74814832e-01 -9.09667850e-01 1.58139944e-01
1.43060654e-01 -7.15404928e-01 9.39256370e-01 1.22040212e-01
-1.37558329e+00 -8.90366733e-01 -8.41983855e-01 -3.58409137e-01
1.91588312e-01 1.25823662e-01 3.67328256e-01 -1.95730776e-01
-1.01977062e+00 9.69959497e-01 -1.24891603e+00 4.18785989e-01
5.53668082e-01 2.70077676e-01 -3.01920801e-01 -3.50643933e-01
-1.51454234e+00 4.61816788e-01 3.89604181e-01 4.69455242e-01
-1.05062580e+00 -8.70109856e-01 -9.46513057e-01 -1.11866549e-01
3.14121008e-01 -7.06363440e-01 9.16261852e-01 -9.65458632e-01
-1.58332098e+00 4.26240385e-01 -4.07588184e-01 -3.34758729e-01
6.12075746e-01 -7.62280822e-02 -5.56658328e-01 3.78522992e-01
-6.95149526e-02 5.50423801e-01 1.05562901e+00 -1.06129754e+00
-7.64913857e-01 8.57923031e-02 5.60150407e-02 2.06989005e-01
-3.55400480e-02 1.58461109e-01 -6.90677702e-01 -1.04460502e+00
8.99360166e-04 -5.89361489e-01 -2.71089673e-01 1.43560395e-01
-4.46885191e-02 -9.84442532e-02 9.01828289e-01 -9.39642727e-01
1.42020750e+00 -2.21945119e+00 3.10946524e-01 -3.54641885e-01
3.45273793e-01 6.09363377e-01 -2.53185660e-01 -1.68046385e-01
-5.28032035e-02 -6.05705194e-03 -1.75481603e-01 -2.16327921e-01
-4.31656450e-01 1.74125046e-01 -3.88943911e-01 4.95249629e-01
3.02602232e-01 1.13828051e+00 -1.04933274e+00 -5.24071336e-01
7.22041130e-01 1.24843240e+00 -4.08439666e-01 1.86093003e-01
-2.21280202e-01 1.05708456e+00 -7.70319760e-01 4.22807872e-01
8.26283514e-01 -3.84121984e-01 -1.64597958e-01 -6.10637963e-01
-6.34354234e-01 2.10578129e-01 -8.83591592e-01 2.05932045e+00
-5.02072215e-01 7.14518189e-01 6.22855797e-02 -7.64831007e-01
7.22216666e-01 3.94577831e-01 6.26425922e-01 -8.92259061e-01
9.23484471e-03 1.87332690e-01 -3.61262381e-01 -6.28115773e-01
5.08051395e-01 1.40300646e-01 3.89699489e-01 -2.15750098e-01
-8.25927705e-02 4.88215834e-01 1.95815805e-02 -1.09251514e-01
7.22800910e-01 5.53044200e-01 -6.52335957e-02 -4.19900715e-02
8.12291741e-01 -4.66729820e-01 9.23182011e-01 2.42034674e-01
-1.27815038e-01 9.64769363e-01 -1.10357270e-01 -6.22045934e-01
-1.07145286e+00 -9.69581842e-01 -2.49997929e-01 8.78391981e-01
6.38226151e-01 -2.03285858e-01 -5.55077076e-01 -2.10281372e-01
-3.04262727e-01 1.94496185e-01 -5.34130275e-01 2.79700551e-02
-1.21839428e+00 -3.90618593e-01 5.76779805e-02 5.51907361e-01
8.11120152e-01 -9.95064795e-01 -4.60974693e-01 4.72100407e-01
-6.60636246e-01 -1.41552544e+00 -1.09725130e+00 -5.41233122e-01
-7.95191169e-01 -6.83853626e-01 -9.80727017e-01 -8.78839910e-01
4.54666764e-01 4.61205125e-01 7.64636934e-01 5.90294264e-02
-2.34814137e-01 -1.22586697e-01 -4.86189067e-01 5.12938201e-01
-2.04034634e-02 -2.33836353e-01 -1.68593302e-01 2.70193875e-01
4.16404940e-02 -6.86654747e-01 -1.12143123e+00 6.16337180e-01
-9.96166289e-01 5.43595493e-01 4.19813544e-01 8.63740504e-01
1.00486779e+00 -1.76316276e-01 2.78116405e-01 -5.25787890e-01
4.17700112e-02 -3.82232487e-01 -6.08256817e-01 1.45586357e-01
-1.11746207e-01 -3.22043598e-02 7.54283369e-01 -7.17482746e-01
-1.16992521e+00 2.36052256e-02 -2.07984030e-01 -1.03935957e+00
2.89975703e-02 2.70770639e-01 -2.05736667e-01 -2.36894801e-01
1.73265308e-01 6.23567462e-01 -3.52137983e-01 -5.35772741e-01
2.31192753e-01 3.51903886e-01 7.29441524e-01 -4.13286954e-01
8.60119283e-01 7.69494116e-01 -1.75472274e-01 -4.84676659e-01
-9.64724660e-01 -2.03153983e-01 -5.87955832e-01 -2.41927028e-01
1.22358572e+00 -1.28140318e+00 -6.30463898e-01 4.31001157e-01
-1.15876925e+00 -6.84005380e-01 -1.28645748e-01 7.32033849e-01
-6.04709804e-01 2.86846250e-01 -9.42988217e-01 -2.45086387e-01
-3.13027352e-01 -1.23516941e+00 1.13591838e+00 3.76679093e-01
3.54648322e-01 -8.50824356e-01 -3.22209656e-01 2.74888992e-01
5.78019798e-01 5.43229580e-01 8.71077999e-02 3.67046505e-01
-1.18522346e+00 4.48724985e-01 -4.61598396e-01 2.63851523e-01
3.47804338e-01 -8.34656805e-02 -6.27961159e-01 -4.84989136e-01
9.50901583e-03 8.79237354e-02 1.16164351e+00 6.84918284e-01
1.64672422e+00 -4.40979958e-01 -2.11098030e-01 1.29218674e+00
1.37609422e+00 1.74378842e-01 8.26385856e-01 1.42166972e-01
1.15676343e+00 1.50557503e-01 9.76507425e-01 4.99258339e-01
4.23623949e-01 9.65083659e-01 2.68675745e-01 -9.79056284e-02
-5.03327131e-01 -2.68285245e-01 4.83025461e-01 8.71431530e-01
-7.12136686e-01 -6.77112490e-02 -3.46188009e-01 3.24440479e-01
-1.96671093e+00 -1.27823305e+00 -1.31700709e-01 2.00166130e+00
9.20864403e-01 -1.83041602e-01 -1.64151639e-01 -1.82406753e-01
1.02385604e+00 5.08752167e-01 -7.38771439e-01 2.62380391e-01
-1.71758354e-01 1.61578491e-01 4.03017730e-01 7.82230556e-01
-1.17820799e+00 9.56767857e-01 4.51424122e+00 1.15389490e+00
-1.37690532e+00 4.00444061e-01 6.85803950e-01 -1.96302697e-01
-4.10114199e-01 -2.18892008e-01 -6.60409570e-01 8.69983494e-01
6.64240777e-01 -1.58797905e-01 6.24810100e-01 3.87479246e-01
6.81213439e-01 3.52819949e-01 -7.71004200e-01 1.14120758e+00
-3.09127063e-01 -1.73759043e+00 9.48000029e-02 -1.72528982e-01
8.60520601e-01 -2.40332403e-04 3.37581098e-01 -5.22697680e-02
-1.32088512e-02 -1.20144057e+00 6.88305676e-01 8.53049576e-01
1.46065557e+00 -7.77691185e-01 2.74967134e-01 -3.85033451e-02
-1.93680155e+00 6.65560365e-02 -4.60079581e-01 2.50986636e-01
3.01750213e-01 3.90837550e-01 2.04877090e-02 7.03839064e-01
8.74542058e-01 1.37865937e+00 -1.80173472e-01 6.94309056e-01
-2.46389076e-01 3.25471103e-01 3.90600562e-02 5.69741070e-01
1.96521431e-01 -1.82490766e-01 5.08078754e-01 1.08458602e+00
3.91780108e-01 6.65461361e-01 5.76522201e-02 9.02414918e-01
6.97596073e-02 -3.22749853e-01 -9.89666488e-03 1.59032762e-01
4.86164272e-01 1.26501906e+00 -3.59438896e-01 -2.49678478e-01
-5.71862519e-01 1.29135370e+00 1.60875738e-01 5.97653151e-01
-1.05147409e+00 -1.94537416e-01 8.20948422e-01 3.55493188e-01
5.89620650e-01 -4.85411912e-01 2.22532868e-01 -1.63459003e+00
9.84962434e-02 -5.73545039e-01 3.17679048e-01 -8.24417889e-01
-1.00268745e+00 6.15481615e-01 -2.53164649e-01 -1.62445045e+00
-5.67450821e-02 -1.81240156e-01 -2.98221201e-01 1.11137831e+00
-1.83165824e+00 -1.21066391e+00 -5.92714906e-01 9.81699526e-01
9.38614070e-01 2.68273652e-01 2.77099252e-01 6.84655786e-01
-4.50958878e-01 4.21585709e-01 2.37199217e-02 3.23989362e-01
7.03579187e-01 -6.93840861e-01 4.80519235e-01 9.38670278e-01
-3.42709571e-01 5.08653402e-01 4.27143157e-01 -5.08493662e-01
-1.38366663e+00 -1.66488326e+00 4.13562208e-01 -2.12648377e-01
4.35627133e-01 -9.33100879e-02 -1.37262428e+00 6.39292419e-01
-7.21853301e-02 8.23536456e-01 6.74152821e-02 -8.64557326e-01
-1.62161708e-01 -9.61237177e-02 -9.95740056e-01 4.34201956e-01
1.29449165e+00 -6.07681811e-01 -1.13726906e-01 1.28370030e-02
1.19008017e+00 -7.81957626e-01 -1.17880797e+00 6.44843221e-01
5.30536294e-01 -9.48530376e-01 1.27737975e+00 -1.79266393e-01
8.75709891e-01 -7.68669903e-01 -1.05908088e-01 -9.27761912e-01
-5.68460524e-01 -7.83166587e-01 -4.22023594e-01 1.10905743e+00
-1.78403735e-01 -5.26854038e-01 4.39776480e-01 4.83175427e-01
-3.51168364e-01 -9.32335496e-01 -1.00413477e+00 -6.61797047e-01
-1.50335073e-01 -1.32313579e-01 5.64359248e-01 1.15250826e+00
-4.95464236e-01 -1.48254365e-01 -7.53960907e-01 3.39004874e-01
6.69476569e-01 2.50303239e-01 1.96990654e-01 -6.67875886e-01
-3.29341292e-01 -2.66316533e-01 -2.69921303e-01 -1.43055475e+00
1.41666427e-01 -6.81972921e-01 4.59021144e-02 -1.44840038e+00
8.40512440e-02 -3.66099417e-01 -5.51754951e-01 2.12760642e-01
-3.73280257e-01 5.01200318e-01 7.68060088e-02 3.57128978e-01
-6.15348935e-01 6.47857308e-01 1.99945533e+00 1.06640622e-01
-2.74316818e-01 -3.33080530e-01 -1.71956241e-01 6.01079464e-01
7.23981500e-01 -8.57075229e-02 -3.00670356e-01 -6.83895290e-01
-3.04614097e-01 6.40446186e-01 6.04469240e-01 -7.77097464e-01
1.50161698e-01 -5.24734080e-01 6.78646147e-01 -5.36644399e-01
4.38178778e-01 -4.30061847e-01 4.26116019e-01 4.13765252e-01
-4.86615360e-01 -1.88639596e-01 -3.45340185e-02 5.46064436e-01
-3.70555729e-01 3.55956763e-01 1.12238419e+00 -6.67837262e-02
-1.06249416e+00 1.07876039e+00 -6.44221902e-02 5.86095965e-03
8.98514986e-01 -1.05161145e-01 -3.48612249e-01 -6.35475665e-02
-7.72762239e-01 2.63717055e-01 5.59535444e-01 6.39438272e-01
1.13761377e+00 -1.52010822e+00 -8.38647187e-01 1.23607963e-01
-2.21300498e-01 3.67409140e-01 8.89305949e-01 9.83746946e-01
-4.82656211e-01 1.98542565e-01 -2.77012140e-01 -4.83695120e-01
-1.07323670e+00 6.18626177e-01 4.99610960e-01 6.18934147e-02
-1.01031864e+00 8.92455637e-01 4.89335358e-01 2.63130039e-01
-1.30004033e-01 -3.88669610e-01 -2.96409160e-01 -3.03704321e-01
9.65109408e-01 2.72023618e-01 -4.91898924e-01 -9.10044670e-01
-2.60023326e-01 8.34649324e-01 -1.23002224e-01 5.80625311e-02
1.37108672e+00 -4.77641672e-01 1.18914815e-02 2.89054304e-01
1.33550477e+00 -2.79127210e-01 -2.00715947e+00 -4.33418661e-01
-2.99207658e-01 -9.88898396e-01 1.81966752e-01 -2.33997494e-01
-1.43369806e+00 7.21414566e-01 5.38721919e-01 -2.52203226e-01
1.40527153e+00 -8.58777165e-02 1.27397048e+00 -2.94258386e-01
4.09673631e-01 -7.63855934e-01 1.56984836e-01 3.83753002e-01
9.09544408e-01 -1.15988648e+00 -1.69496253e-01 -6.04028642e-01
-4.27886903e-01 1.22015989e+00 7.00916290e-01 -3.49010944e-01
4.34339046e-01 1.75485358e-01 -2.52599716e-01 2.23756537e-01
-9.11666512e-01 -1.87429637e-01 4.50537413e-01 4.87612247e-01
5.31347215e-01 -5.42795695e-02 -4.10829991e-01 4.23429161e-01
2.62393981e-01 3.99736494e-01 4.28250819e-01 4.63677853e-01
-3.25980633e-01 -7.60044158e-01 4.98719234e-03 1.62653923e-01
-5.37792385e-01 -2.56565392e-01 4.38444585e-01 5.14350653e-01
3.52377921e-01 7.09510505e-01 1.73948541e-01 -5.06396174e-01
1.01963110e-01 -5.91501236e-01 6.21179104e-01 -2.64029413e-01
-2.24924535e-01 3.57146084e-01 -1.96269423e-01 -1.17247462e+00
-7.70496964e-01 -5.55090964e-01 -1.40142918e+00 -3.61072004e-01
3.10356505e-02 -8.80987793e-02 1.83725044e-01 6.87612951e-01
4.97649103e-01 8.09786677e-01 7.95336127e-01 -1.22418761e+00
1.04674725e-02 -6.84421062e-01 -4.20250326e-01 3.76759797e-01
8.00347924e-01 -5.24001956e-01 -2.87474185e-01 4.58201587e-01] | [11.019201278686523, -1.8642851114273071] |
80b813ac-52c2-4f9f-9ce7-df1606d2475e | mixsim-a-hierarchical-framework-for-mixed | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Suo_MixSim_A_Hierarchical_Framework_for_Mixed_Reality_Traffic_Simulation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Suo_MixSim_A_Hierarchical_Framework_for_Mixed_Reality_Traffic_Simulation_CVPR_2023_paper.pdf | MixSim: A Hierarchical Framework for Mixed Reality Traffic Simulation | The prevailing way to test a self-driving vehicle (SDV) in simulation involves non-reactive open-loop replay of real world scenarios. However, in order to safely deploy SDVs to the real world, we need to evaluate them in closed-loop. Towards this goal, we propose to leverage the wealth of interesting scenarios captured in the real world and make them reactive and controllable to enable closed-loop SDV evaluation in what-if situations. In particular, we present MixSim, a hierarchical framework for mixed reality traffic simulation. MixSim explicitly models agent goals as routes along the road network and learns a reactive route-conditional policy. By inferring each agent's route from the original scenario, MixSim can reactively re-simulate the scenario and enable testing different autonomy systems under the same conditions. Furthermore, by varying each agent's route, we can expand the scope of testing to what-if situations with realistic variations in agent behaviors or even safety-critical interactions. Our experiments show that MixSim can serve as a realistic, reactive, and controllable digital twin of real world scenarios. For more information, please visit the project website: https://waabi.ai/research/mixsim/ | ['Raquel Urtasun', 'Sergio Casas', 'Alexander Cui', 'James Tu', 'Justin Xu', 'Kelvin Wong', 'Simon Suo'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['mixed-reality'] | ['computer-vision'] | [-4.03368682e-01 1.59698352e-01 1.30654782e-01 -3.50557566e-01
-2.41851121e-01 -8.12740922e-01 9.28157806e-01 -2.97920793e-01
-2.26730928e-01 8.75166774e-01 -2.02109829e-01 -6.83344185e-01
1.79026738e-01 -1.12652063e+00 -7.43349612e-01 -2.69615948e-01
-3.94467443e-01 9.78149176e-01 7.71250725e-01 -8.40924621e-01
-3.01016599e-01 7.39492953e-01 -1.99275684e+00 -2.05125108e-01
8.17196131e-01 8.43642503e-02 1.57647088e-01 9.71259654e-01
4.24922317e-01 8.01518977e-01 -5.97739935e-01 7.60811418e-02
4.28732693e-01 -1.51198015e-01 -3.36803555e-01 -7.32231736e-02
-6.93638921e-02 -5.15528023e-01 -6.87339485e-01 7.15316117e-01
4.09313321e-01 2.10481793e-01 2.27812335e-01 -2.09181118e+00
3.85169834e-01 5.40739357e-01 3.13069783e-02 6.50366470e-02
6.27053082e-01 1.19131434e+00 2.85075009e-01 6.29362762e-02
8.00043702e-01 1.36392403e+00 1.19146377e-01 5.91112912e-01
-1.37773967e+00 -8.65180135e-01 3.55224043e-01 1.81769505e-02
-1.34669054e+00 -7.46315420e-01 5.31292260e-01 -3.40450972e-01
7.85737216e-01 2.90968984e-01 7.13151455e-01 1.54070842e+00
1.97206661e-01 4.86035615e-01 9.82087195e-01 1.26077548e-01
5.71141064e-01 1.53170303e-01 -6.80429265e-02 6.06973767e-01
3.24655652e-01 8.49014938e-01 -1.85555313e-02 -1.13715135e-01
3.46902996e-01 -3.48769844e-01 -2.38652565e-02 -7.51839936e-01
-1.50875759e+00 3.21864992e-01 4.24234867e-02 -2.18296885e-01
-3.68795156e-01 4.38157350e-01 4.28859830e-01 5.85831761e-01
-2.72196293e-01 3.23368460e-01 -2.43521959e-01 -4.45346206e-01
-4.61918145e-01 8.12781811e-01 1.16572356e+00 1.11791921e+00
7.36918569e-01 4.12994087e-01 -4.30415235e-02 4.39743660e-02
3.66589248e-01 8.25750709e-01 -4.53158841e-02 -1.42692411e+00
2.93572545e-01 5.17362297e-01 4.95740026e-01 -6.13894343e-01
-5.07872283e-01 -1.20653763e-01 -2.25994974e-01 8.86922240e-01
3.05883378e-01 -6.63439929e-01 -7.06681848e-01 1.94453788e+00
5.90711832e-01 6.49161696e-01 3.62983406e-01 9.43392873e-01
2.88326383e-01 5.94318867e-01 9.84991491e-02 2.48359032e-02
9.48134065e-01 -5.59490979e-01 -2.36436889e-01 -3.06307226e-01
8.33297908e-01 -1.43903330e-01 9.95953262e-01 1.00859493e-01
-1.02398813e+00 -3.73847634e-01 -1.18451345e+00 7.85582125e-01
-4.22113508e-01 -6.21684968e-01 4.83441263e-01 4.96765196e-01
-1.15788090e+00 1.87206447e-01 -1.06047475e+00 -4.90455419e-01
-1.68284476e-02 4.24690217e-01 -2.06171602e-01 -1.28273591e-01
-1.42087722e+00 1.12298369e+00 2.42080063e-01 -2.79458970e-01
-1.77726984e+00 -6.19481862e-01 -9.51056123e-01 -2.76191682e-01
5.50551414e-01 -7.66238153e-01 1.54889727e+00 -4.95761395e-01
-1.62510467e+00 4.02646989e-01 5.15050106e-02 -5.02856076e-01
8.90902400e-01 5.21887302e-01 -9.05012071e-01 -2.17317432e-01
1.77635625e-01 6.56190336e-01 1.96976334e-01 -1.61916316e+00
-5.98453283e-01 -2.88450951e-03 7.22788453e-01 1.53160170e-01
6.88124955e-01 -2.76822776e-01 -1.89937711e-01 2.05708176e-01
-4.57442373e-01 -1.39402080e+00 -5.78117728e-01 -2.64451742e-01
-2.37522542e-01 1.64893925e-01 9.18517470e-01 2.56300718e-01
6.61099315e-01 -2.20959520e+00 -2.13253960e-01 4.08113658e-01
2.11634085e-01 8.93383026e-02 -4.16265875e-01 8.15575719e-01
2.77392536e-01 -5.97221367e-02 3.97871621e-02 1.72248594e-02
4.68331277e-01 4.16349024e-01 -2.19438627e-01 4.55840319e-01
-1.81685880e-01 8.60328078e-01 -1.11402047e+00 -3.62829864e-01
6.05773151e-01 2.92743236e-01 -5.07425606e-01 2.41765410e-01
-5.87180853e-01 7.35119522e-01 -6.40185893e-01 2.13087484e-01
6.41949654e-01 3.13714772e-01 3.72448951e-01 3.41873854e-01
-3.73848736e-01 2.24531993e-01 -1.38752210e+00 1.20420659e+00
-7.34274328e-01 7.42071629e-01 1.83948904e-01 -5.54861367e-01
6.36681139e-01 2.27508396e-01 3.13903302e-01 -1.27091408e+00
2.50612110e-01 5.91571406e-02 2.43842468e-01 -5.59581816e-01
4.08141822e-01 7.24820718e-02 -4.85520780e-01 6.48699105e-01
-5.52247882e-01 -4.27139640e-01 5.74773550e-01 2.61827499e-01
1.43504131e+00 1.15441032e-01 -7.95281753e-02 -1.19930290e-01
3.35219115e-01 3.94661486e-01 5.51831961e-01 8.89450371e-01
-4.86033380e-01 2.48205177e-02 6.17032468e-01 -3.76737654e-01
-9.88136947e-01 -1.50598669e+00 2.20457941e-01 7.29018867e-01
6.35841012e-01 -1.24673650e-01 -6.34947240e-01 -5.42173564e-01
1.83450893e-01 1.48449063e+00 -3.48804116e-01 -3.25059593e-01
-7.89995015e-01 -2.72839814e-01 7.70429730e-01 5.53775355e-02
6.01477385e-01 -1.00759709e+00 -8.84734452e-01 3.58303219e-01
6.16293885e-02 -1.17378414e+00 -1.93310320e-01 -2.50863254e-01
-1.68185398e-01 -1.11342943e+00 1.92928314e-01 -2.61387825e-01
3.97465944e-01 4.59122151e-01 1.03760493e+00 1.29544288e-01
1.19335778e-01 4.57139730e-01 2.72046328e-02 -1.98370859e-01
-1.08606744e+00 -5.45330578e-03 1.86074108e-01 -3.00394654e-01
5.48037477e-02 -6.71694100e-01 -4.88037169e-01 9.07132566e-01
-8.48138630e-01 1.22000612e-01 9.31084976e-02 1.87598661e-01
1.87241286e-01 1.57033727e-01 5.53346694e-01 -6.80555999e-01
8.04468811e-01 -7.74120629e-01 -1.14360225e+00 -3.12193073e-02
-5.28499603e-01 3.86249907e-02 7.77263284e-01 -4.32464272e-01
-8.89375031e-01 -2.14971319e-01 2.34485492e-02 -1.27422689e-02
-6.06495857e-01 2.35892817e-01 -4.36895609e-01 1.29442722e-01
7.19045401e-01 9.52918157e-02 2.29418769e-01 3.34622532e-01
3.92263889e-01 4.49263752e-01 3.78791451e-01 -7.78382063e-01
1.18699479e+00 4.32202160e-01 2.70040520e-03 -5.00638187e-01
4.04201485e-02 1.91157073e-01 -4.71190177e-02 -5.82492948e-01
5.00611007e-01 -9.33874726e-01 -1.36716974e+00 2.96687841e-01
-6.36135995e-01 -1.29682398e+00 -3.39213371e-01 3.90258431e-01
-7.88506210e-01 -8.77513662e-02 -1.54729739e-01 -5.82870066e-01
5.95102906e-01 -1.70174098e+00 8.01527977e-01 2.99042314e-01
-3.55597228e-01 -9.24364984e-01 4.83423948e-01 -3.77654261e-03
6.78808331e-01 3.76202613e-01 3.65715653e-01 -4.44194108e-01
-8.94916236e-01 -1.68295428e-01 1.50750905e-01 -3.38813633e-01
-1.62676573e-01 2.93461293e-01 -6.95925891e-01 -4.82811987e-01
-6.16482973e-01 -3.70810851e-02 3.54919061e-02 1.52814798e-02
5.01972675e-01 -2.51075983e-01 -5.71626902e-01 2.73212105e-01
1.07536733e+00 3.71168733e-01 7.10697711e-01 6.50987089e-01
1.40773967e-01 4.82531488e-01 7.35976756e-01 3.11781228e-01
1.15418887e+00 9.61919248e-01 6.20972037e-01 -9.20040160e-03
-3.70431133e-02 -4.58402753e-01 6.62442863e-01 1.04667924e-01
3.59753251e-01 -4.96330589e-01 -1.14033818e+00 3.92442465e-01
-1.89733839e+00 -1.31215608e+00 -1.50743723e-01 2.28211951e+00
3.53128910e-01 6.11762881e-01 3.91540408e-01 -2.76603580e-01
6.22841179e-01 3.54756117e-02 -6.85825527e-01 -5.96358180e-01
8.63619819e-02 -3.64457786e-01 6.69373572e-01 7.31040180e-01
-4.83999014e-01 1.09669566e+00 6.15671730e+00 2.95739412e-01
-1.12732112e+00 -3.40395570e-02 1.26100600e-01 -1.82536080e-01
-5.22639215e-01 4.32913065e-01 -5.39375901e-01 6.55814111e-01
1.62070525e+00 -7.45625913e-01 6.70242012e-01 6.28185928e-01
1.09203112e+00 -3.52682233e-01 -1.15377402e+00 2.71008641e-01
-4.86572087e-01 -1.25121260e+00 -3.72155577e-01 2.45088890e-01
5.71666658e-01 4.14390177e-01 -1.64580703e-01 8.74671102e-01
1.42426777e+00 -8.17117989e-01 8.46202850e-01 3.78667444e-01
6.20879412e-01 -8.43730211e-01 4.67938244e-01 5.62637031e-01
-1.06272089e+00 1.81867499e-02 2.75679946e-01 -2.29788348e-01
5.49092412e-01 4.96089868e-02 -1.15009713e+00 3.23846430e-01
2.62289256e-01 2.36155733e-01 -4.35260981e-01 1.04421592e+00
-1.18913598e-01 5.47095656e-01 -4.53980535e-01 1.82072017e-02
2.25903660e-01 -1.70260191e-01 8.47032368e-01 8.29885483e-01
5.24344519e-02 -2.93405186e-02 5.30513942e-01 7.24486709e-01
3.16514701e-01 -6.96223676e-01 -1.20564818e+00 3.10429513e-01
7.62320518e-01 9.04138088e-01 -3.85645360e-01 -3.56824696e-01
8.27892404e-03 3.86541635e-01 -6.54220134e-02 6.86433256e-01
-1.38948834e+00 -1.55441687e-01 1.35998404e+00 3.94702375e-01
-2.52194941e-01 -6.32698119e-01 1.85784847e-01 -9.14307773e-01
-2.57846147e-01 -1.03585064e+00 -8.93110931e-02 -9.36143756e-01
-6.42254889e-01 6.07467234e-01 4.18107629e-01 -1.39055312e+00
-5.30116558e-01 1.09769322e-01 -8.34047854e-01 4.20063913e-01
-1.42484319e+00 -8.82948935e-01 -4.93355453e-01 5.50228119e-01
3.97546530e-01 -2.13893861e-01 4.97522771e-01 1.80603981e-01
-7.52481759e-01 3.22170228e-01 -3.01232576e-01 -2.40923837e-01
5.44475853e-01 -8.13063622e-01 9.09121752e-01 7.46638596e-01
-5.39726973e-01 2.50627220e-01 1.36709487e+00 -6.47283614e-01
-1.65004456e+00 -1.19104874e+00 1.76083326e-01 -4.33590561e-01
8.85366976e-01 -4.22800958e-01 -5.60482323e-01 9.33219433e-01
1.71775356e-01 2.37449761e-02 -2.82095633e-02 -3.88600141e-01
-5.30183539e-02 -3.74192685e-01 -1.31059718e+00 1.40811861e+00
1.20358467e+00 -1.70501828e-01 5.37572103e-03 1.35425016e-01
8.74731123e-01 -4.83691186e-01 -4.71870035e-01 2.61416316e-01
4.58650291e-01 -9.83224511e-01 7.16694951e-01 -4.42830443e-01
-1.77972764e-01 -8.78685653e-01 -2.11654961e-01 -1.69877076e+00
6.85805455e-02 -7.90839672e-01 1.92436144e-01 9.86303806e-01
6.26682281e-01 -1.23164463e+00 5.44286430e-01 8.15680206e-01
-2.55291730e-01 -1.67880490e-01 -8.75887334e-01 -8.69242191e-01
-8.89271721e-02 -8.17135811e-01 1.14789259e+00 5.31330168e-01
6.80978075e-02 -8.59259292e-02 -1.53651331e-02 5.81003070e-01
6.49880767e-01 -2.52890348e-01 1.42863679e+00 -8.49213183e-01
-8.82725865e-02 -2.80903667e-01 -5.85901976e-01 -7.11571395e-01
6.19251013e-01 -5.84680200e-01 2.57056147e-01 -1.44360173e+00
-2.53551096e-01 -8.62527072e-01 3.10407311e-01 1.76440492e-01
4.70204949e-01 -1.49106339e-01 2.01973543e-01 -6.47287667e-02
-8.90748858e-01 4.47162390e-01 1.18254459e+00 -1.80988416e-01
-2.53247499e-01 1.88598499e-01 -3.94112140e-01 4.21497405e-01
1.21029294e+00 -3.56245548e-01 -8.75726879e-01 -2.50322729e-01
1.08693078e-01 6.34650469e-01 5.93767941e-01 -1.23347104e+00
4.46622521e-01 -8.53685141e-01 -3.55086923e-01 -3.29285473e-01
3.18241805e-01 -9.01047051e-01 7.44728208e-01 4.55638617e-01
-2.29721963e-01 3.68183017e-01 4.14842844e-01 3.75164956e-01
6.12422712e-02 3.09947997e-01 6.12571657e-01 -1.52850226e-02
-9.91041243e-01 2.86206752e-01 -1.10851002e+00 2.08889350e-01
1.79298091e+00 -3.35027397e-01 -7.32041955e-01 -5.23076475e-01
-5.54449260e-01 1.00417709e+00 1.12333620e+00 5.06145656e-01
5.60109079e-01 -1.02000260e+00 -7.45513201e-01 3.77042592e-01
2.63241976e-01 -2.36405864e-01 3.32746685e-01 6.69833362e-01
-6.66319549e-01 5.07239364e-02 -3.60447824e-01 -5.86809099e-01
-8.16406548e-01 5.31715572e-01 7.30408251e-01 5.15216291e-02
-6.21393323e-01 -2.27636933e-01 1.21448159e-01 -8.39362025e-01
-1.10893257e-01 1.29456788e-01 4.94248047e-02 -4.96194452e-01
3.51867706e-01 2.72126883e-01 -1.84525605e-02 -6.50270462e-01
-5.32903016e-01 -2.55484600e-02 2.84658462e-01 -5.87176919e-01
1.05895865e+00 -3.94791454e-01 3.35671186e-01 3.11460853e-01
6.41620755e-01 -5.67413457e-02 -1.60179484e+00 3.50219071e-01
-2.46306404e-01 -3.68402272e-01 -1.57787994e-01 -7.51769304e-01
-9.61415946e-01 1.22448407e-01 4.86490220e-01 2.43790239e-01
5.63335359e-01 -1.36096284e-01 5.21361411e-01 4.35488582e-01
9.51334000e-01 -9.06527638e-01 -3.04602146e-01 3.89881223e-01
6.47574484e-01 -1.04178190e+00 -4.15197045e-01 -9.38095599e-02
-9.50993776e-01 6.89850628e-01 1.00915253e+00 -1.91745430e-01
4.23090458e-01 8.33194375e-01 2.04771563e-01 -2.39342749e-01
-1.33059263e+00 -3.41593325e-01 -6.97331846e-01 1.02213156e+00
-2.36939088e-01 5.02367377e-01 3.00412983e-01 -2.25785568e-01
-4.03565079e-01 1.52792662e-01 1.40275872e+00 8.66427779e-01
-3.94313723e-01 -9.70018387e-01 -2.94259459e-01 6.51835650e-02
4.86005068e-01 6.30694330e-01 -3.90639668e-03 1.22860754e+00
-1.13021739e-01 1.27593160e+00 1.70317009e-01 -5.28010070e-01
7.37274647e-01 -3.39751452e-01 1.44156978e-01 -3.52370352e-01
-2.91005760e-01 -6.40571296e-01 7.18975008e-01 -8.44035923e-01
7.49917850e-02 -7.24864960e-01 -1.61049986e+00 -1.06206810e+00
3.38625401e-01 1.61678597e-01 7.04048216e-01 7.48826385e-01
5.64935625e-01 6.78383291e-01 9.47309732e-01 -9.54180896e-01
-3.83825451e-01 -2.72704899e-01 -1.15117393e-01 1.94707260e-01
4.30188805e-01 -8.72136533e-01 -4.26102757e-01 -4.63558465e-01] | [5.071260929107666, 1.325390338897705] |
ac6e41e4-e7cb-4ed6-84a3-101f830e71d5 | dynamic-path-controllable-deep-unfolding | 2306.16060 | null | https://arxiv.org/abs/2306.16060v1 | https://arxiv.org/pdf/2306.16060v1.pdf | Dynamic Path-Controllable Deep Unfolding Network for Compressive Sensing | Deep unfolding network (DUN) that unfolds the optimization algorithm into a deep neural network has achieved great success in compressive sensing (CS) due to its good interpretability and high performance. Each stage in DUN corresponds to one iteration in optimization. At the test time, all the sampling images generally need to be processed by all stages, which comes at a price of computation burden and is also unnecessary for the images whose contents are easier to restore. In this paper, we focus on CS reconstruction and propose a novel Dynamic Path-Controllable Deep Unfolding Network (DPC-DUN). DPC-DUN with our designed path-controllable selector can dynamically select a rapid and appropriate route for each image and is slimmable by regulating different performance-complexity tradeoffs. Extensive experiments show that our DPC-DUN is highly flexible and can provide excellent performance and dynamic adjustment to get a suitable tradeoff, thus addressing the main requirements to become appealing in practice. Codes are available at https://github.com/songjiechong/DPC-DUN. | ['Jian Zhang', 'Bin Chen', 'Jiechong Song'] | 2023-06-28 | null | null | null | null | ['compressive-sensing'] | ['computer-vision'] | [ 2.43364081e-01 -3.24617475e-01 -8.35633501e-02 -2.78034836e-01
-5.22086978e-01 -3.33560079e-01 1.02976747e-01 -2.21441120e-01
-1.51974946e-01 3.31480801e-01 2.40887821e-01 -4.39051062e-01
-3.01545322e-01 -7.68966556e-01 -6.08022273e-01 -8.52861583e-01
-3.18102598e-01 -1.60108283e-02 1.26291983e-04 -1.41838983e-01
8.30845758e-02 4.87505645e-01 -8.12924266e-01 6.21797703e-02
9.26366329e-01 1.18403244e+00 6.69802129e-01 3.37264657e-01
2.75362760e-01 5.99910438e-01 -1.10997349e-01 -6.97031170e-02
6.11163735e-01 -5.78134358e-01 -3.30698043e-01 5.26930839e-02
4.34541469e-03 -5.35542250e-01 -6.66675806e-01 1.11335576e+00
7.92578459e-01 1.14137516e-03 2.67542779e-01 -7.56803334e-01
-7.04187214e-01 9.53879416e-01 -8.92388582e-01 2.41935745e-01
3.99129465e-03 3.73744257e-02 8.74003947e-01 -9.89859104e-01
2.65831441e-01 9.95072782e-01 4.33679938e-01 3.39734167e-01
-1.07189357e+00 -7.89708257e-01 -1.46537405e-02 3.85913312e-01
-1.45582616e+00 -6.86529100e-01 1.05836296e+00 -2.19766840e-01
2.09274143e-01 2.00059697e-01 7.04232335e-01 8.41324508e-01
2.03619096e-02 7.78828204e-01 7.01469004e-01 -3.83602202e-01
2.95677066e-01 -4.38353866e-01 -1.28900677e-01 7.09865332e-01
1.44941866e-01 -3.43421176e-02 -4.62668449e-01 1.86787948e-01
1.21725047e+00 2.15103462e-01 -7.73236871e-01 -3.22402775e-01
-1.43668640e+00 6.68574870e-01 8.15692425e-01 2.26680532e-01
-4.32620943e-01 3.42616886e-01 2.63349265e-01 4.88484561e-01
-3.34947705e-02 5.32497704e-01 -1.43176064e-01 4.08855453e-02
-1.00964642e+00 5.01672029e-02 4.20944512e-01 8.48263562e-01
5.43371141e-01 4.29302633e-01 -1.74470544e-01 9.55647171e-01
-5.39199300e-02 4.52213019e-01 3.03623915e-01 -1.11324275e+00
4.97715503e-01 1.93391696e-01 -1.01077326e-01 -1.42973340e+00
-4.12467271e-01 -9.51661408e-01 -1.40121484e+00 -4.56815958e-02
-1.02049336e-01 -4.14088964e-01 -6.63994253e-01 1.80520034e+00
5.49857467e-02 2.34207377e-01 -2.42228448e-01 1.13354897e+00
6.56368971e-01 9.23797607e-01 -4.12933707e-01 -2.69578338e-01
1.10554039e+00 -9.30641413e-01 -4.41392064e-01 -3.63059402e-01
2.83100009e-01 -6.47606313e-01 1.02012157e+00 3.40277255e-01
-9.96549964e-01 -4.55216020e-01 -1.43203282e+00 1.88913599e-01
3.32524925e-01 4.47757125e-01 3.92631173e-01 1.75433457e-01
-1.06325352e+00 5.36910653e-01 -8.81982684e-01 8.10686350e-02
3.50298375e-01 3.90936553e-01 -6.42754743e-03 -1.66358650e-01
-1.17305112e+00 2.39811748e-01 4.71969903e-01 5.68773687e-01
-9.08506870e-01 -4.41714346e-01 -7.08034635e-01 3.50163847e-01
4.80443269e-01 -5.87429285e-01 1.14539170e+00 -8.85374844e-01
-1.61768937e+00 3.59674811e-01 -6.11813664e-02 -3.01217794e-01
5.12084544e-01 -6.58709779e-02 -3.33611637e-01 3.07407260e-01
-1.05985897e-02 4.51430410e-01 1.10725284e+00 -1.07649279e+00
-2.34834969e-01 -2.71909624e-01 2.14233752e-02 1.76705465e-01
-3.53877991e-01 -3.18225086e-01 -7.85742879e-01 -1.00143683e+00
7.33817875e-01 -9.55863297e-01 -4.81647223e-01 1.11795269e-01
-4.92473304e-01 3.09226215e-01 7.20337212e-01 -4.02261466e-01
1.57637346e+00 -2.28840852e+00 2.52145380e-01 2.04025015e-01
5.37598312e-01 3.52435082e-01 -2.35842809e-01 4.28140223e-01
-5.69811184e-03 -3.48619148e-02 -3.97033274e-01 -2.38507822e-01
-2.64276803e-01 7.78837223e-03 -1.15789637e-01 5.83862424e-01
-8.93477127e-02 6.59451067e-01 -8.85481536e-01 -1.81465119e-01
1.59993634e-01 4.17131931e-01 -6.75406575e-01 1.00625277e-01
-2.76757721e-02 5.60424149e-01 -6.89936459e-01 8.16084743e-01
9.12062943e-01 -6.36213005e-01 3.62054229e-01 -4.39619541e-01
-6.33592233e-02 -8.81772488e-03 -1.23251164e+00 1.55120564e+00
-4.87185776e-01 8.28205884e-01 4.08653587e-01 -1.27545643e+00
8.07202935e-01 1.09998554e-01 4.76772338e-01 -9.37631726e-01
1.90724641e-01 3.96319062e-01 9.27405655e-02 -3.30489606e-01
3.99770975e-01 -3.22084054e-02 3.45779583e-02 5.48567712e-01
-3.12118202e-01 2.92681038e-01 -1.97010618e-02 2.13568091e-01
1.02351880e+00 -3.48080009e-01 2.76564240e-01 -5.33731103e-01
6.10258102e-01 -3.82266849e-01 9.21046615e-01 7.55405605e-01
-2.19294637e-01 7.59652793e-01 3.58814538e-01 -4.76288527e-01
-9.61978376e-01 -9.12849724e-01 -1.91306993e-01 7.33232021e-01
5.13840556e-01 -2.64528632e-01 -5.80022037e-01 -1.22891463e-01
-3.09658676e-01 3.49832088e-01 -3.02481353e-01 -1.43308282e-01
-7.36659408e-01 -5.36600769e-01 3.06145310e-01 3.40192884e-01
8.59412730e-01 -8.95635009e-01 -6.38272107e-01 4.04274374e-01
-2.40657821e-01 -1.04064238e+00 -6.96938455e-01 7.79902339e-02
-8.41043651e-01 -8.06952655e-01 -9.07100260e-01 -9.64734852e-01
8.02431405e-01 7.68454731e-01 7.28507102e-01 1.16871625e-01
1.70840263e-01 -1.50866494e-01 -5.72974384e-01 5.92173040e-02
-5.21268807e-02 2.03485906e-01 6.96173459e-02 2.33041659e-01
-1.77570507e-01 -9.65629816e-01 -1.10637319e+00 3.73622358e-01
-1.16168499e+00 4.07228976e-01 6.26914084e-01 9.43160772e-01
6.65421724e-01 1.08276427e-01 6.01259112e-01 -6.95234239e-01
6.84161425e-01 -4.49408591e-01 -6.26317620e-01 1.18550837e-01
-5.81659079e-01 8.21159706e-02 9.16521549e-01 -2.40086362e-01
-7.29079425e-01 -3.57227288e-02 -1.08463466e-01 -6.45418406e-01
4.53487426e-01 8.15293252e-01 -1.35555372e-01 -2.38570943e-01
4.97889310e-01 3.48184973e-01 1.20097041e-01 -3.51755708e-01
4.81923558e-02 4.73065615e-01 5.05098999e-01 -4.96081054e-01
7.94691086e-01 3.47084999e-01 4.27488461e-02 -7.98443675e-01
-8.13916445e-01 -1.03749469e-01 -3.07172865e-01 -2.05264211e-01
4.40945327e-01 -1.09172821e+00 -5.24112642e-01 6.72266781e-01
-8.50236535e-01 -5.18989861e-01 1.15553230e-01 4.15803432e-01
-1.85499161e-01 4.12310749e-01 -7.29901969e-01 -3.22142571e-01
-5.24197698e-01 -1.51574397e+00 6.76442504e-01 3.43696773e-01
3.12643439e-01 -7.00682759e-01 -4.55385715e-01 1.20752091e-02
7.12104559e-01 2.40468577e-01 7.79023349e-01 8.12159106e-02
-9.39633131e-01 -4.04082909e-02 -4.51314181e-01 3.89794827e-01
8.59915689e-02 -2.07563594e-01 -4.97766674e-01 -6.69685245e-01
1.04969040e-01 -1.64674416e-01 9.44704115e-01 7.34166205e-01
1.74883163e+00 -4.66009229e-01 -5.52739203e-02 1.23268306e+00
1.72992253e+00 3.33933562e-01 6.69854820e-01 3.94650489e-01
6.43612802e-01 -8.16582330e-03 3.47872853e-01 5.82683563e-01
1.71813324e-01 6.12938762e-01 4.94005978e-01 -1.20167047e-01
8.04570913e-02 -1.27480879e-01 3.76737565e-01 1.14818776e+00
1.06900908e-01 -4.43882197e-01 -7.92684436e-01 3.38841140e-01
-1.68665922e+00 -7.57976711e-01 1.18776098e-01 2.08049655e+00
6.15438461e-01 9.01827067e-02 -3.31203699e-01 7.24530742e-02
7.69030392e-01 7.48041570e-01 -6.91736400e-01 -7.79941231e-02
-1.21107437e-01 1.84568673e-01 5.08958161e-01 5.09423137e-01
-9.00264680e-01 5.45622766e-01 5.30498505e+00 9.23506081e-01
-1.48317015e+00 3.82839218e-02 8.07809949e-01 -2.46515915e-01
-4.09231663e-01 1.35340532e-02 -2.92003036e-01 6.02444410e-01
4.06669915e-01 -1.49031818e-01 5.07055640e-01 6.16329193e-01
4.68509406e-01 3.54260206e-02 -8.30023766e-01 1.33132780e+00
-2.29347229e-01 -1.66975832e+00 2.99472325e-02 -5.25338612e-02
7.20656335e-01 1.52692005e-01 1.12619035e-01 1.42634623e-02
-9.06319022e-02 -7.95015812e-01 7.63594449e-01 2.01980025e-01
1.01461220e+00 -8.04504454e-01 4.30371016e-01 2.89174944e-01
-1.14263916e+00 -3.31090897e-01 -3.52210552e-01 1.97172523e-01
3.30877095e-01 9.01211262e-01 -4.29209441e-01 5.18718958e-01
5.74472070e-01 8.18418086e-01 -1.37074947e-01 9.81882989e-01
-1.28483072e-01 6.65268600e-01 -2.62221545e-01 6.45399839e-02
4.50483978e-01 -4.83375013e-01 7.91089714e-01 9.94340062e-01
8.09953868e-01 3.50606829e-01 2.44424924e-01 7.07801521e-01
-3.02887261e-01 -1.77164346e-01 -2.96402961e-01 1.19384915e-01
8.76567483e-01 9.87549961e-01 -6.18727982e-01 -1.49905145e-01
-1.18359275e-01 9.11161363e-01 1.03304625e-01 3.74259651e-01
-7.88656831e-01 -4.30011719e-01 4.59034473e-01 3.00367653e-01
3.76735896e-01 -5.30952632e-01 -3.70903522e-01 -1.36191225e+00
1.48925722e-01 -9.53517497e-01 3.62224393e-02 -5.70837379e-01
-8.62021983e-01 7.78099656e-01 -4.00835425e-01 -1.65182698e+00
1.87887490e-01 -3.05287689e-01 -6.11605823e-01 6.22281432e-01
-1.48079777e+00 -6.75913393e-01 -5.89066684e-01 4.90031391e-01
6.42243326e-01 -1.48664817e-01 4.16942120e-01 6.50770664e-01
-9.46545839e-01 7.93437302e-01 3.62319589e-01 1.42072603e-01
4.83637363e-01 -7.85769939e-01 2.23976657e-01 1.28012812e+00
-2.68181562e-01 7.23585546e-01 5.34841120e-01 -2.84747571e-01
-1.63549459e+00 -1.03659415e+00 2.01284438e-01 4.24247563e-01
4.45263237e-01 -1.72321051e-01 -7.39957988e-01 3.82101864e-01
1.11881293e-01 -7.65862456e-03 4.53458101e-01 -2.13697776e-01
-1.69070378e-01 -5.69649935e-01 -7.62552083e-01 6.20619535e-01
1.09412146e+00 -4.01528180e-01 4.34862264e-02 2.16592595e-01
7.14915693e-01 -6.73401773e-01 -5.50038874e-01 5.26300609e-01
4.93860930e-01 -1.25713611e+00 9.33044434e-01 1.61562577e-01
7.68033683e-01 -4.32513386e-01 -3.06038082e-01 -1.23774040e+00
-6.31134331e-01 -7.98362255e-01 -2.29752690e-01 7.08538413e-01
3.48673195e-01 -7.18997538e-01 6.61452234e-01 1.84263483e-01
-3.72671664e-01 -1.15876293e+00 -7.50702739e-01 -6.12067878e-01
-3.98453504e-01 -3.15103680e-01 8.05361569e-01 9.79481697e-01
-4.46145415e-01 2.64036536e-01 -6.37501419e-01 5.23071945e-01
7.82826126e-01 4.57102358e-01 5.28031826e-01 -7.48321235e-01
-5.91767132e-01 -4.76976842e-01 -1.77241594e-01 -1.76448083e+00
-3.75079066e-01 -7.06285894e-01 6.44840971e-02 -1.35133493e+00
8.61250237e-02 -9.04901385e-01 -3.35328192e-01 2.97295123e-01
1.24401981e-02 8.95802155e-02 3.50957185e-01 5.21848917e-01
-3.28296751e-01 7.67737091e-01 1.62174475e+00 -9.49975476e-02
-2.37749860e-01 -2.91206222e-02 -9.09067333e-01 4.51659232e-01
1.05798948e+00 -2.38190040e-01 -5.94535708e-01 -1.11052704e+00
2.57017761e-01 6.10889673e-01 1.71362817e-01 -1.25306368e+00
3.45014125e-01 -9.18230116e-02 2.91530937e-01 -5.84918857e-01
1.97657526e-01 -6.18560493e-01 3.44276100e-01 5.85752606e-01
-2.64374733e-01 4.86971028e-02 -1.67411521e-01 4.72368240e-01
-3.84050071e-01 -8.43416452e-02 1.11992037e+00 -1.28795102e-01
-8.38620067e-01 7.60523319e-01 -8.87811631e-02 -2.22145930e-01
7.76745975e-01 -1.52143911e-01 -7.05041140e-02 -7.71337867e-01
-5.29778898e-01 4.24654394e-01 4.20459837e-01 3.45042571e-02
7.91579366e-01 -1.31617618e+00 -6.57981217e-01 3.42182875e-01
-2.18598753e-01 3.34853202e-01 5.29575765e-01 8.70408714e-01
-9.54422891e-01 2.19988838e-01 3.89177836e-02 -6.28510356e-01
-6.64649725e-01 2.35033005e-01 3.75335068e-01 -1.17970869e-01
-8.42888117e-01 9.88442481e-01 1.48144037e-01 -2.62751013e-01
2.84906209e-01 -1.99485615e-01 1.07228115e-01 -1.65367231e-01
5.31658709e-01 1.79768696e-01 -1.12218633e-01 -2.90206403e-01
-1.01475030e-01 4.45740044e-01 -6.56332895e-02 2.02138752e-01
1.53884161e+00 -2.61419594e-01 -3.37927818e-01 -1.50616810e-01
1.30120051e+00 6.33719414e-02 -1.36362875e+00 -4.94385898e-01
-4.02574807e-01 -6.66260064e-01 4.44491088e-01 -3.14839423e-01
-1.76283097e+00 6.36214018e-01 5.72091043e-01 2.49392644e-01
1.59736919e+00 -3.18318963e-01 1.26716030e+00 3.79931182e-01
4.54278558e-01 -7.77110040e-01 1.29346043e-01 5.39279580e-01
9.83502269e-01 -9.31197047e-01 1.95094064e-01 -2.11736083e-01
-4.28125352e-01 1.22805047e+00 5.14823854e-01 -2.01341242e-01
8.73385251e-01 1.58970430e-01 -4.03169617e-02 -3.25727522e-01
-3.56428683e-01 2.66778380e-01 -9.84968245e-02 1.80372119e-01
1.69498563e-01 3.37512136e-01 -3.87381703e-01 2.92886913e-01
-1.19407386e-01 -1.17770277e-01 6.32407665e-01 7.58879185e-01
-5.00307918e-01 -1.11792588e+00 -1.95258737e-01 6.03969276e-01
-2.85207450e-01 -2.11217627e-01 2.81640917e-01 4.02590454e-01
-4.49231379e-02 6.71551526e-01 -1.28700152e-01 -5.61732173e-01
8.79563093e-02 -7.79937863e-01 1.66575015e-01 -4.11105394e-01
2.48893835e-02 2.17544481e-01 -9.54075977e-02 -6.71715200e-01
-3.45184654e-01 -5.14259279e-01 -1.15513563e+00 -5.33832967e-01
-2.44140625e-01 8.56746137e-02 2.57617325e-01 5.57911396e-01
6.09217703e-01 4.84703779e-01 1.22773457e+00 -8.75363111e-01
-5.71408451e-01 -6.37973785e-01 -6.05601668e-01 -1.58819005e-01
6.37324989e-01 -4.70779926e-01 -1.88720793e-01 -1.62145734e-01] | [11.127436637878418, -1.9788933992385864] |
b34204b6-ba9a-4a42-a123-d75fce151744 | exploring-in-context-learning-capabilities-of | 2305.08804 | null | https://arxiv.org/abs/2305.08804v1 | https://arxiv.org/pdf/2305.08804v1.pdf | Exploring In-Context Learning Capabilities of Foundation Models for Generating Knowledge Graphs from Text | Knowledge graphs can represent information about the real-world using entities and their relations in a structured and semantically rich manner and they enable a variety of downstream applications such as question-answering, recommendation systems, semantic search, and advanced analytics. However, at the moment, building a knowledge graph involves a lot of manual effort and thus hinders their application in some situations and the automation of this process might benefit especially for small organizations. Automatically generating structured knowledge graphs from a large volume of natural language is still a challenging task and the research on sub-tasks such as named entity extraction, relation extraction, entity and relation linking, and knowledge graph construction aims to improve the state of the art of automatic construction and completion of knowledge graphs from text. The recent advancement of foundation models with billions of parameters trained in a self-supervised manner with large volumes of training data that can be adapted to a variety of downstream tasks has helped to demonstrate high performance on a large range of Natural Language Processing (NLP) tasks. In this context, one emerging paradigm is in-context learning where a language model is used as it is with a prompt that provides instructions and some examples to perform a task without changing the parameters of the model using traditional approaches such as fine-tuning. This way, no computing resources are needed for re-training/fine-tuning the models and the engineering effort is minimal. Thus, it would be beneficial to utilize such capabilities for generating knowledge graphs from text. | ['Sven Groppe', 'Jinghua Groppe', 'Sanju Tiwari', 'Nandana Mihindukulasooriya', 'Hanieh Khorashadizadeh'] | 2023-05-15 | null | null | null | null | ['graph-construction', 'relation-extraction'] | ['graphs', 'natural-language-processing'] | [ 9.26083550e-02 5.40797234e-01 -2.77343243e-01 -3.05203080e-01
-3.13934743e-01 -7.02851653e-01 6.97764516e-01 8.06136370e-01
-2.35012382e-01 7.43964136e-01 3.79558913e-02 -5.81340909e-01
-1.98478311e-01 -1.27392757e+00 -5.20294428e-01 -7.52034411e-02
-1.17655862e-02 6.81536138e-01 3.89814973e-01 -4.29259747e-01
1.11320965e-01 3.73013705e-01 -1.56920195e+00 3.44907910e-01
1.08829725e+00 7.68335104e-01 3.44130963e-01 2.55157799e-01
-9.15306091e-01 7.66277611e-01 -5.29517174e-01 -8.23993266e-01
2.47500129e-02 -1.77675366e-01 -1.09891677e+00 -2.67809182e-01
-5.07417023e-02 1.74138859e-01 5.65129220e-02 8.03720713e-01
1.46239266e-01 1.72489733e-01 3.15848649e-01 -1.01619160e+00
-5.19042671e-01 7.57038713e-01 -1.76432297e-01 -1.04605434e-02
5.98977804e-01 -1.56559914e-01 9.55917120e-01 -7.04342246e-01
8.11875701e-01 1.14761972e+00 5.21753728e-01 2.45664939e-01
-9.19184029e-01 -4.72295135e-01 9.46533307e-02 3.07108074e-01
-1.26399267e+00 -2.47035891e-01 6.13637626e-01 -4.75426167e-01
1.30664957e+00 9.86506417e-02 6.54417515e-01 8.32737029e-01
-1.72437817e-01 4.62941289e-01 8.04691732e-01 -7.41291523e-01
1.81936145e-01 4.50473249e-01 1.56930432e-01 8.97104204e-01
5.52389860e-01 -4.05926287e-01 -5.40531814e-01 -1.34401143e-01
7.28349686e-01 -6.02158941e-02 -4.64056246e-02 -3.51535589e-01
-1.25894368e+00 7.12299287e-01 4.65024471e-01 5.72042823e-01
-4.23572659e-01 -1.61617950e-01 4.05847311e-01 3.46579283e-01
3.38138998e-01 9.02209103e-01 -6.89908803e-01 -8.75794366e-02
-6.56159282e-01 7.48149082e-02 1.31311214e+00 1.19451058e+00
1.11024690e+00 -2.21903801e-01 7.81135559e-02 7.61413991e-01
8.60559344e-02 1.49053141e-01 4.53742653e-01 -4.99188423e-01
8.82690430e-01 1.36621737e+00 1.14529610e-01 -1.20082963e+00
-4.70150173e-01 -2.97221184e-01 -5.91855645e-01 -2.66823620e-01
6.29047275e-01 -1.15546688e-01 -9.24458385e-01 1.45581388e+00
6.71781361e-01 1.10767156e-01 2.05052912e-01 5.56318343e-01
9.92735267e-01 5.20131886e-01 3.31822038e-01 -4.54251021e-02
1.62070751e+00 -6.65642262e-01 -6.85729444e-01 -5.28121293e-01
1.04872644e+00 -6.16295278e-01 1.15064561e+00 3.56230475e-02
-5.72812915e-01 -5.09084165e-01 -8.60250056e-01 -1.61931783e-01
-1.08264637e+00 -1.77671835e-01 1.12112296e+00 3.75177026e-01
-6.54068768e-01 5.47542274e-01 -6.47738695e-01 -6.34573877e-01
3.96610767e-01 3.26517612e-01 -6.81987226e-01 -3.20156455e-01
-1.53315353e+00 1.02943993e+00 9.02103782e-01 8.66813660e-02
-1.94879174e-01 -6.45576358e-01 -1.09894264e+00 2.78929621e-01
8.64285707e-01 -9.85288024e-01 9.07291651e-01 -6.98957086e-01
-1.11019850e+00 7.47931302e-01 8.64328295e-02 -2.96948075e-01
2.30794977e-02 -2.59489089e-01 -4.80126709e-01 -1.06414437e-01
1.29612848e-01 3.44295561e-01 5.13530314e-01 -9.43112850e-01
-5.94882667e-01 -3.70382458e-01 2.52129406e-01 1.66297957e-01
-4.34070528e-01 8.87703970e-02 -5.08577645e-01 -3.16418916e-01
-1.52494479e-03 -6.21037483e-01 -3.03713351e-01 -5.52403986e-01
-3.79243076e-01 -3.84724528e-01 6.67716265e-01 -6.96483076e-01
1.41022956e+00 -1.79541385e+00 -2.59264618e-01 4.46883321e-01
1.85250834e-01 5.09227931e-01 1.21740453e-01 8.01343918e-01
-1.66014303e-02 3.56738716e-01 -3.72977294e-02 2.64520854e-01
-1.01798065e-01 1.64379686e-01 -1.28638238e-01 -3.93351585e-01
3.14858228e-01 1.08357465e+00 -1.23649764e+00 -5.75886369e-01
1.52869895e-01 2.28689909e-01 -3.03414762e-01 2.95726806e-01
-5.48805237e-01 3.75520319e-01 -7.58478165e-01 4.59057808e-01
-1.66639630e-02 -5.65163612e-01 6.26037240e-01 -2.23352119e-01
1.44907981e-01 5.17007709e-01 -1.36360514e+00 1.66059208e+00
-6.53710186e-01 2.98211873e-01 -2.11355716e-01 -1.09309065e+00
1.17769432e+00 3.81108940e-01 2.44967461e-01 -6.69773161e-01
-1.06615514e-01 2.70053864e-01 -2.33665276e-02 -9.03967619e-01
5.32937229e-01 -1.40737176e-01 -1.45551369e-01 3.97933245e-01
1.38777107e-01 -7.72440359e-02 6.53662026e-01 3.95440072e-01
1.32625747e+00 2.95338452e-01 7.03608096e-01 1.09863929e-01
5.76452315e-01 4.13054585e-01 4.99672234e-01 3.70509744e-01
4.66163129e-01 -9.97890308e-02 4.47054714e-01 -4.38428611e-01
-8.75349522e-01 -6.61872089e-01 2.07597181e-01 1.02732933e+00
-2.33790159e-01 -7.95913100e-01 -6.33704782e-01 -8.41818929e-01
3.34403142e-02 7.62528658e-01 -1.05772555e-01 -8.73085260e-02
-6.03080392e-01 -3.80013853e-01 3.26873004e-01 6.00877047e-01
4.77217317e-01 -1.37159133e+00 -2.17741922e-01 3.37235063e-01
-2.20857844e-01 -1.45514417e+00 1.51080498e-02 -1.05679044e-02
-9.55638885e-01 -1.33640575e+00 -1.04512759e-01 -9.92582679e-01
8.12876225e-01 6.21075034e-02 1.33302987e+00 1.01726115e-01
-1.66083455e-01 3.19571227e-01 -4.55697745e-01 -5.59253275e-01
-4.62169021e-01 3.73643517e-01 -2.30357960e-01 -1.96449101e-01
6.59689307e-01 -5.88415444e-01 -3.01078707e-01 1.16819896e-01
-9.79905486e-01 2.34387636e-01 7.24981189e-01 5.55925488e-01
5.68336189e-01 2.49225050e-01 9.59850729e-01 -1.40976238e+00
8.30028415e-01 -5.55269718e-01 -5.97614169e-01 5.71130872e-01
-8.43647182e-01 2.67594159e-01 7.68278956e-01 -1.95219025e-01
-1.18053997e+00 6.36821687e-02 1.04913108e-01 7.31804222e-02
-2.36541644e-01 1.14577198e+00 -3.29270154e-01 2.97904253e-01
7.76388824e-01 -1.17949918e-01 -2.20300511e-01 -5.13141751e-01
7.42414355e-01 6.94695592e-01 3.68845284e-01 -6.58908010e-01
9.16356027e-01 8.36564377e-02 1.16779394e-01 -6.03937745e-01
-1.05690312e+00 -5.56611478e-01 -6.77280426e-01 6.29938841e-02
6.13457978e-01 -7.95473993e-01 -6.17606521e-01 -6.91025183e-02
-9.61786151e-01 -3.39491367e-01 -3.49432528e-01 3.66747230e-01
-1.35406524e-01 3.15164775e-01 -2.55362451e-01 -5.25813520e-01
-5.37324369e-01 -5.96424937e-01 6.88407719e-01 4.17538226e-01
-4.06300545e-01 -1.18426728e+00 -2.05397531e-01 6.87534392e-01
3.48683476e-01 2.11397216e-01 1.29951680e+00 -1.04765284e+00
-7.74521649e-01 -4.33295131e-01 -2.19299242e-01 1.28198519e-01
4.44109976e-01 -1.86052337e-01 -5.80461740e-01 1.37932524e-01
-3.63627821e-01 -2.55181491e-01 2.37829268e-01 -2.54241526e-01
1.15986776e+00 -3.82817596e-01 -5.75649858e-01 2.33593777e-01
1.21694708e+00 1.15855858e-01 5.74484229e-01 4.81930405e-01
8.88635397e-01 9.28502440e-01 7.36868382e-01 2.71908045e-02
6.74760342e-01 5.18550754e-01 8.70117173e-02 1.11344971e-01
2.98992619e-02 -6.39772594e-01 -8.12377110e-02 6.64940238e-01
-1.56394780e-01 -7.76242986e-02 -1.26532042e+00 6.99459016e-01
-1.97271121e+00 -9.54451323e-01 -1.23173021e-01 2.27152514e+00
1.15118945e+00 1.49891123e-01 -2.54195482e-01 1.79775078e-02
6.36300981e-01 -1.41497090e-01 -3.57835472e-01 -1.83119088e-01
2.39803582e-01 3.73084277e-01 1.70807078e-01 3.40338022e-01
-8.01132202e-01 1.32081151e+00 5.29200602e+00 6.51737571e-01
-9.97251570e-01 -2.58572042e-01 1.84015065e-01 4.35259134e-01
-3.11449140e-01 4.87735480e-01 -8.74517977e-01 2.18755260e-01
1.09684014e+00 -4.09977138e-01 4.76437300e-01 9.74336922e-01
1.85523823e-01 -9.19346288e-02 -1.16103566e+00 1.01861405e+00
-1.79686591e-01 -1.56043839e+00 2.61390299e-01 -5.13980649e-02
5.47747612e-01 -2.77155757e-01 -6.48680925e-01 5.19359708e-01
4.40738708e-01 -9.29232538e-01 9.16063786e-02 6.30280793e-01
6.79823399e-01 -5.44429243e-01 7.52286732e-01 5.17225027e-01
-1.34231865e+00 -1.17882781e-01 -1.46567225e-01 -9.87918079e-02
3.38013262e-01 8.67605448e-01 -1.35556066e+00 9.28241074e-01
4.18961972e-01 4.75913137e-01 -7.20286369e-01 8.83087039e-01
-6.09320819e-01 4.33895111e-01 -2.89473236e-01 -2.97155809e-02
-1.41880497e-01 -3.47748697e-01 1.63564190e-01 1.22745299e+00
3.03364515e-01 9.41846967e-02 2.86507249e-01 6.21610582e-01
-2.82901555e-01 4.75934327e-01 -9.15080905e-01 -4.58379954e-01
6.16108239e-01 1.54374492e+00 -7.45707512e-01 -5.33776402e-01
-5.76864243e-01 2.83684522e-01 5.91142893e-01 3.39365721e-01
-3.90469819e-01 -6.78267002e-01 2.04935029e-01 4.06615585e-01
1.26907811e-01 -4.17007983e-01 -1.20795190e-01 -1.10599184e+00
2.04597160e-01 -8.84601057e-01 6.80222869e-01 -8.57162416e-01
-1.21545684e+00 4.83678520e-01 2.10964248e-01 -8.31529200e-01
-6.04155660e-01 -5.14534712e-01 -5.04126251e-01 8.12127054e-01
-1.46947372e+00 -9.99497890e-01 -5.57429373e-01 5.87434709e-01
1.29893750e-01 -1.00455858e-01 9.84063447e-01 3.00355256e-01
-3.68002892e-01 1.86272964e-01 -4.60292727e-01 4.94325399e-01
6.77207649e-01 -1.25521231e+00 3.79796773e-01 7.56589472e-01
3.81689757e-01 9.51736212e-01 3.62695694e-01 -8.63038361e-01
-1.37374032e+00 -1.09823060e+00 1.31675076e+00 -4.22205389e-01
9.56314445e-01 -3.29650015e-01 -1.13586700e+00 7.71835983e-01
-7.84022436e-02 5.31599596e-02 9.93167102e-01 4.76940840e-01
-2.69735873e-01 -2.50553012e-01 -7.92784631e-01 5.67710102e-01
1.19938564e+00 -5.52074552e-01 -7.87509024e-01 4.67234522e-01
6.14452243e-01 -3.32401723e-01 -1.16917229e+00 3.38994354e-01
1.35246545e-01 -4.78876680e-01 7.75695920e-01 -8.98464978e-01
2.29604229e-01 -4.38453227e-01 1.53243035e-01 -1.28796577e+00
-6.73495382e-02 -7.55337119e-01 -4.32520896e-01 1.52643383e+00
7.07610309e-01 -7.69994080e-01 7.28376985e-01 1.03313982e+00
-1.69521421e-01 -7.74524808e-01 -3.99417132e-01 -6.61395371e-01
-4.87292826e-01 -2.62981415e-01 7.90705442e-01 1.17526996e+00
4.35477197e-01 9.34703529e-01 2.25975588e-01 -2.26826612e-02
2.07176223e-01 4.21579629e-01 9.99040544e-01 -1.60185456e+00
-6.57634810e-02 -9.61986482e-02 -3.48181993e-01 -7.87217259e-01
1.84670791e-01 -1.17959213e+00 -3.56262922e-01 -2.19479609e+00
-9.71719623e-02 -8.09827566e-01 -9.91837587e-05 8.73045325e-01
-4.02750582e-01 -5.32198370e-01 -4.17269655e-02 1.47816896e-01
-5.24146438e-01 2.35940143e-01 1.18312073e+00 6.22483864e-02
-4.11832750e-01 1.45322857e-02 -9.52981293e-01 7.68297911e-01
8.47560465e-01 -5.18583357e-01 -6.95138872e-01 -2.75837988e-01
7.26125002e-01 3.85471955e-02 1.02768734e-01 -6.59028530e-01
3.61509562e-01 -2.63611466e-01 1.30497664e-01 -1.43926814e-01
1.07913136e-01 -7.88260579e-01 2.90746689e-01 2.02065688e-02
1.81858689e-02 6.50039613e-02 6.65699095e-02 5.74506819e-01
-5.14288664e-01 -4.02487963e-01 2.47967198e-01 -4.36617434e-01
-1.05094862e+00 1.75424024e-01 8.22085217e-02 2.32755303e-01
1.05881763e+00 -2.19208702e-01 -4.50638175e-01 -1.73545644e-01
-7.21557796e-01 2.47551039e-01 2.97781646e-01 5.10710537e-01
2.87085772e-01 -1.06378424e+00 -3.56076360e-01 5.19977883e-04
3.24139327e-01 4.68862385e-01 -6.83817565e-02 5.32465219e-01
-4.20129061e-01 6.63246632e-01 -1.79134216e-03 -2.07471356e-01
-9.97787952e-01 6.98351443e-01 -4.33879495e-02 -8.39017093e-01
-6.96953833e-01 3.25442463e-01 -9.35114548e-02 -5.31911969e-01
-1.54313952e-01 -1.79800853e-01 -5.92435241e-01 1.37318507e-01
4.76143539e-01 1.03468597e-01 2.99061716e-01 -3.07158828e-01
-3.02663445e-01 1.83544219e-01 -9.72242467e-03 2.33057648e-01
1.45399618e+00 6.86429292e-02 -2.06608847e-01 2.44425744e-01
5.74041903e-01 1.41700804e-01 -7.33850241e-01 -3.94062072e-01
5.48248112e-01 -3.18808377e-01 -3.26372951e-01 -8.00609589e-01
-8.89530122e-01 6.71710551e-01 -2.27061346e-01 4.88400310e-01
8.69878650e-01 2.57370681e-01 6.73035920e-01 9.11385477e-01
5.95968246e-01 -1.21091771e+00 8.58005509e-02 6.01623118e-01
6.86354995e-01 -1.21019864e+00 5.14236353e-02 -8.50869834e-01
-6.24995410e-01 1.13838840e+00 6.13509238e-01 3.50434721e-01
6.23222589e-01 9.74512845e-02 -4.15802635e-02 -5.40045202e-01
-6.59027278e-01 -5.15166998e-01 3.34882647e-01 7.75458217e-01
7.02359021e-01 3.23327817e-03 -3.45196217e-01 5.79482555e-01
-3.31245333e-01 6.87619522e-02 2.56437480e-01 9.72927690e-01
-4.16608691e-01 -1.58470774e+00 -6.53622970e-02 8.21129322e-01
-4.12685722e-01 -2.04508260e-01 -3.81315678e-01 8.30911040e-01
-3.07412334e-02 1.02336168e+00 -4.08469021e-01 -1.19873188e-01
5.72044671e-01 5.50232470e-01 3.16159666e-01 -1.25646853e+00
-6.53971851e-01 -4.19867039e-01 6.91807330e-01 -3.51503730e-01
-4.16554242e-01 -3.11061203e-01 -1.59290588e+00 -1.55449092e-01
-4.06828493e-01 3.73484462e-01 7.30493546e-01 1.15952778e+00
7.49681830e-01 4.85846311e-01 9.61700156e-02 -3.31797808e-01
-3.30694318e-01 -9.09505367e-01 -2.51642495e-01 6.71580672e-01
-5.31118989e-01 -6.29269302e-01 1.41494498e-01 2.26288408e-01] | [9.394291877746582, 8.246756553649902] |
fd7e6914-7c56-4483-9658-1bd3d764f164 | building-blocks-for-complex-tasks-robust | 2306.09544 | null | https://arxiv.org/abs/2306.09544v1 | https://arxiv.org/pdf/2306.09544v1.pdf | Building blocks for complex tasks: Robust generative event extraction for radiology reports under domain shifts | This paper explores methods for extracting information from radiology reports that generalize across exam modalities to reduce requirements for annotated data. We demonstrate that multi-pass T5-based text-to-text generative models exhibit better generalization across exam modalities compared to approaches that employ BERT-based task-specific classification layers. We then develop methods that reduce the inference cost of the model, making large-scale corpus processing more feasible for clinical applications. Specifically, we introduce a generative technique that decomposes complex tasks into smaller subtask blocks, which improves a single-pass model when combined with multitask training. In addition, we leverage target-domain contexts during inference to enhance domain adaptation, enabling use of smaller models. Analyses offer insights into the benefits of different cost reduction strategies. | ['Mari Ostendorf', 'Meliha Yetisgen', 'Sitong Zhou'] | 2023-06-15 | null | null | null | null | ['event-extraction'] | ['natural-language-processing'] | [ 6.32318854e-01 5.82831681e-01 -3.26753378e-01 -7.85032988e-01
-1.73222351e+00 -5.57859242e-01 4.29721594e-01 3.00969362e-01
-3.55846316e-01 8.38665426e-01 6.14200592e-01 -7.11275518e-01
-2.78060913e-01 -6.34028792e-01 -7.36919403e-01 -2.90558130e-01
1.95318416e-01 6.93661153e-01 2.18969196e-01 2.16960177e-01
-1.03591517e-01 2.90743589e-01 -8.96230221e-01 8.80137205e-01
1.00239623e+00 6.22547269e-01 4.06043440e-01 8.91592026e-01
-6.00547902e-02 8.63380611e-01 -7.23766208e-01 -4.38271374e-01
-9.50496346e-02 -5.44044077e-01 -9.61380422e-01 1.72282010e-01
4.43885595e-01 -3.74020875e-01 -1.00485787e-01 4.43137914e-01
7.26620436e-01 2.75617510e-01 8.50522101e-01 -5.33123016e-01
-4.76188272e-01 8.16973805e-01 -1.97989270e-01 4.04359728e-01
3.48749071e-01 1.22709498e-01 6.30774677e-01 -3.34509134e-01
7.31121242e-01 8.80132556e-01 7.93018162e-01 7.94965982e-01
-1.38507128e+00 -5.02037585e-01 1.48474768e-01 -1.57085046e-01
-8.89639378e-01 -3.31617326e-01 4.04107004e-01 -3.71633768e-01
1.30359745e+00 3.99905413e-01 5.59976280e-01 1.31878102e+00
6.76940501e-01 9.51808751e-01 1.08097863e+00 -4.02896166e-01
6.77041709e-02 1.35815054e-01 1.24682806e-01 8.19215655e-01
4.04565603e-01 -5.00392795e-01 -5.67338467e-01 -3.18889916e-01
6.05869889e-01 -5.16939349e-02 -1.04245797e-01 -7.77977929e-02
-1.11592519e+00 7.79386044e-01 1.91902027e-01 1.42607167e-01
-5.01813591e-01 -3.77459042e-02 5.38999438e-01 3.48094814e-02
5.29799998e-01 6.69294536e-01 -2.32055411e-01 -1.78438768e-01
-1.26158202e+00 1.38042733e-01 9.38488901e-01 1.40039158e+00
2.92834044e-01 -9.58676189e-02 -5.13134718e-01 8.87334049e-01
1.01953603e-01 2.72679120e-01 7.97764778e-01 -6.95024788e-01
6.14028513e-01 4.09746259e-01 -5.34608662e-01 -1.91597134e-01
-7.07645476e-01 -7.82817185e-01 -5.10190487e-01 -4.39300537e-01
8.87815952e-02 -3.61865759e-01 -1.36710870e+00 1.59256482e+00
2.22472697e-01 -5.95801324e-02 -9.36016068e-02 3.36830348e-01
1.01397467e+00 9.06731486e-02 6.74454987e-01 -2.24695951e-02
1.78393376e+00 -8.75938296e-01 -8.01105797e-01 -4.35044795e-01
9.76206958e-01 -7.56415784e-01 9.32106793e-01 3.74078244e-01
-1.37035048e+00 -4.48422343e-01 -8.44933569e-01 -3.98748308e-01
-2.93832242e-01 1.31779686e-01 8.03491116e-01 8.69127393e-01
-8.96300673e-01 2.85936385e-01 -1.29818296e+00 -2.31087029e-01
6.39502525e-01 4.13548470e-01 -5.22871204e-02 -1.58999667e-01
-8.32168162e-01 1.18803287e+00 5.64077497e-01 -3.06972980e-01
-1.01074719e+00 -1.20717978e+00 -1.08352566e+00 1.49854377e-01
4.97098118e-01 -1.29000831e+00 1.70209181e+00 -2.16945872e-01
-1.22178125e+00 7.11322188e-01 -3.35656613e-01 -4.97722119e-01
3.98290336e-01 -8.99607409e-03 -2.22936824e-01 3.52972269e-01
1.00499727e-01 5.91163218e-01 7.47543573e-01 -9.48964655e-01
-4.54426825e-01 -2.15464264e-01 2.54803970e-02 3.07804972e-01
-3.74865055e-01 -3.51502776e-01 -4.73420590e-01 -7.72858858e-01
-4.06564213e-02 -1.01101923e+00 -4.71201807e-01 -5.89021862e-01
-5.99430859e-01 -1.90695032e-01 3.04997623e-01 -7.94944286e-01
1.36763465e+00 -1.89573133e+00 -1.16189368e-01 1.92729160e-01
5.70576966e-01 -1.66112170e-01 4.95633818e-02 1.89479247e-01
-9.65808854e-02 1.23127110e-01 -2.38879651e-01 -5.63940525e-01
-1.95628360e-01 3.61896157e-01 -8.71612690e-03 -1.06206417e-01
5.07293582e-01 1.13459623e+00 -7.06188679e-01 -9.52631474e-01
-1.09332129e-02 1.70628712e-01 -9.51308429e-01 3.62784155e-02
-2.55366206e-01 5.19043744e-01 -8.59331012e-01 5.54715574e-01
1.98476806e-01 -6.34883583e-01 5.23008943e-01 -2.83563435e-01
2.98995584e-01 8.97972703e-01 -5.19094408e-01 2.15213108e+00
-9.77947414e-01 3.41879189e-01 -1.20833395e-02 -9.58005428e-01
4.67126042e-01 4.32362407e-01 5.53717911e-01 -2.94513047e-01
7.59367049e-02 1.19840220e-01 6.97181299e-02 -7.99558043e-01
5.50166905e-01 -5.72334349e-01 -2.58521885e-01 4.67768908e-01
3.78039956e-01 -5.82576096e-01 3.35611045e-01 3.42042267e-01
1.36477780e+00 2.29026198e-01 4.45479542e-01 -8.04735199e-02
-4.00940664e-02 2.83983469e-01 2.01172799e-01 1.17430365e+00
2.22770095e-01 6.02831006e-01 2.17690423e-01 -1.34417415e-01
-8.55344296e-01 -1.06904006e+00 -3.89501005e-01 1.15132284e+00
-5.94408512e-01 -4.04952437e-01 -6.39045715e-01 -1.05978870e+00
8.64142776e-02 1.10409343e+00 -7.87167728e-01 -1.99169293e-01
-6.53091609e-01 -1.13671851e+00 6.41254187e-01 1.00498903e+00
2.01482363e-02 -7.51397073e-01 -7.45542943e-01 3.07606220e-01
-3.37338895e-01 -1.23297322e+00 -5.57230353e-01 5.93817890e-01
-1.34121799e+00 -7.48879194e-01 -7.19640374e-01 -6.11630619e-01
6.88625097e-01 -1.90355778e-01 1.28058887e+00 -3.84454817e-01
-5.34859240e-01 8.46738696e-01 -2.42060736e-01 -6.86046779e-01
-6.31206691e-01 6.56062961e-01 -5.44089198e-01 -7.40570962e-01
3.92048895e-01 -2.12013602e-01 -3.89648318e-01 -1.27219081e-01
-1.09472454e+00 1.85962319e-01 1.12110746e+00 1.09902477e+00
6.90859258e-01 -4.38117743e-01 6.53973460e-01 -1.44044197e+00
9.41950977e-01 -6.57760203e-01 4.11340930e-02 4.53118622e-01
-7.48606443e-01 3.13986063e-01 3.42426509e-01 -6.19112313e-01
-1.51784742e+00 -5.79447001e-02 -1.69129327e-01 -1.40670061e-01
-2.01167479e-01 7.78892934e-01 3.28426987e-01 2.91171018e-02
8.61726344e-01 4.00991827e-01 -1.82907321e-02 -2.70860732e-01
1.76004857e-01 5.11989415e-01 2.94628382e-01 -8.82293820e-01
4.06962752e-01 1.31645307e-01 6.93996176e-02 -5.63862622e-01
-1.13974118e+00 -3.13798815e-01 -4.87362415e-01 6.59722090e-03
1.07008564e+00 -7.77116001e-01 -4.93914306e-01 -3.94277006e-01
-8.37687075e-01 -5.28570354e-01 -5.68084359e-01 9.21134651e-01
-6.85013294e-01 2.86563814e-01 -8.51214230e-01 -4.28906947e-01
-4.30563778e-01 -1.27592266e+00 1.32037365e+00 2.69092709e-01
-6.36428058e-01 -1.36558962e+00 -4.15343009e-02 6.08755410e-01
3.12000990e-01 -3.90755385e-02 1.12101603e+00 -1.10991001e+00
-6.54201388e-01 -1.34473175e-01 2.43971851e-02 1.50144398e-01
1.17919244e-01 -5.42229712e-01 -8.23485911e-01 -9.27237496e-02
1.21583410e-01 -3.27388108e-01 8.01671207e-01 8.09752703e-01
1.50675726e+00 -9.89405140e-02 -6.83406532e-01 6.78991556e-01
1.02240133e+00 8.19290280e-02 3.70181024e-01 1.97553724e-01
5.93582809e-01 3.60042512e-01 5.35787165e-01 1.57840177e-01
5.39702237e-01 3.22524667e-01 -3.18902522e-01 -2.60212243e-01
-4.05527025e-01 -1.88152000e-01 -1.83570132e-01 9.05083835e-01
-9.46260989e-02 -1.06061094e-01 -1.09873664e+00 6.64613008e-01
-1.41842246e+00 -5.87008834e-01 2.15139315e-01 1.82240260e+00
1.30543864e+00 3.80067796e-01 -5.14422320e-02 -5.97039700e-01
2.12208908e-02 -1.52397245e-01 -5.31263173e-01 -3.05171728e-01
3.76935124e-01 7.25417733e-01 5.68285465e-01 3.05440605e-01
-8.82081568e-01 5.78489244e-01 7.66671515e+00 8.96672428e-01
-7.41880774e-01 3.61713976e-01 6.88722312e-01 -3.77037466e-01
-5.18855274e-01 -3.32628399e-01 -9.59977388e-01 7.44211376e-02
1.18164670e+00 -2.34610781e-01 -5.50056882e-02 9.42689002e-01
6.39740601e-02 -9.39079598e-02 -1.43255496e+00 5.26032388e-01
3.87030512e-01 -1.43263078e+00 1.76461533e-01 1.70073152e-01
7.64421761e-01 -9.25363898e-02 7.94981271e-02 5.58168590e-01
5.16779721e-01 -8.69665146e-01 1.79953530e-01 5.95690250e-01
8.37657630e-01 -3.52239639e-01 7.92619526e-01 1.43653199e-01
-9.13998425e-01 -3.25129530e-03 -1.45989642e-01 5.62537730e-01
2.04760492e-01 5.12516856e-01 -1.70763040e+00 9.04389024e-01
4.07473117e-01 2.90198475e-01 -6.71860456e-01 9.10190165e-01
-5.07150367e-02 7.10715950e-01 -1.74279064e-01 1.11141019e-02
8.74245018e-02 3.35184515e-01 4.78634983e-01 1.80785525e+00
3.81126195e-01 2.17518374e-01 3.98357898e-01 8.95867407e-01
-1.00098185e-01 5.18042110e-02 -6.20757580e-01 -2.11212598e-02
2.90136486e-01 1.40476024e+00 -6.48683965e-01 -7.51856327e-01
-4.76500660e-01 6.32217228e-01 1.03971489e-01 2.81474501e-01
-7.34624982e-01 -1.61422372e-01 4.08201404e-02 2.41110697e-01
2.69222409e-01 -2.29966059e-01 -6.89038992e-01 -1.03470814e+00
-2.27612615e-01 -9.39247429e-01 8.26745510e-01 -7.40595758e-01
-1.35279584e+00 5.25318384e-01 3.70764047e-01 -1.12773108e+00
-7.45052159e-01 -5.09245932e-01 -2.85581380e-01 8.94724369e-01
-1.45452738e+00 -1.40869200e+00 -2.52710193e-01 5.27397871e-01
8.32748950e-01 2.02693438e-04 1.08689821e+00 1.54521689e-01
-3.27847689e-01 8.31487596e-01 -2.70860016e-01 1.24111794e-01
8.30471218e-01 -1.45216072e+00 1.02921121e-01 6.57480419e-01
-2.54580110e-01 1.17792130e+00 2.51394242e-01 -9.62541699e-01
-1.24407923e+00 -1.05374849e+00 6.51587129e-01 -7.63661861e-01
5.81116319e-01 -2.58143961e-01 -8.13331783e-01 1.08395016e+00
1.37772843e-01 -5.34195483e-01 1.20834994e+00 6.16920292e-01
-1.79862306e-02 3.33844349e-02 -1.06609142e+00 5.44124424e-01
9.40559089e-01 -6.32969320e-01 -8.70971501e-01 4.68462706e-01
7.21124411e-01 -9.80181873e-01 -1.51019704e+00 2.80074477e-01
5.30377150e-01 -1.08327448e-01 8.40112090e-01 -8.82681787e-01
5.93105495e-01 2.90116251e-01 2.51885593e-01 -1.21051526e+00
-2.82663375e-01 -4.87948030e-01 -6.38959259e-02 6.25373244e-01
9.12728250e-01 -6.10023022e-01 7.51945674e-01 9.59548235e-01
-7.67189085e-01 -7.38090515e-01 -8.59449208e-01 -5.21167457e-01
1.93506971e-01 -4.41992193e-01 2.78030097e-01 1.01477504e+00
2.38687769e-01 3.83209556e-01 1.63000807e-01 -7.11471960e-02
4.42679167e-01 1.31613657e-01 4.32334512e-01 -8.19586217e-01
-7.02361047e-01 -2.52446532e-01 -5.78214936e-02 -1.11141300e+00
-3.07512917e-02 -1.03084767e+00 2.00082645e-01 -1.70242918e+00
5.49369156e-01 -4.47894454e-01 -2.84340709e-01 9.59882617e-01
-5.58240831e-01 1.21780120e-01 2.57623033e-03 -1.05561949e-01
-4.57074016e-01 1.27116159e-01 1.58878267e+00 -1.06607363e-01
-1.60207391e-01 1.69592336e-01 -9.45627689e-01 3.90936911e-01
6.00048006e-01 -7.01911569e-01 -6.75614893e-01 -4.32269663e-01
-1.38995349e-01 1.57579780e-01 1.28876030e-01 -8.71464789e-01
3.09578329e-01 -3.58805768e-02 7.21408188e-01 -6.04489565e-01
2.93355823e-01 -5.03969073e-01 -5.45793585e-02 4.06006426e-01
-7.45045722e-01 -9.50245708e-02 7.37511754e-01 4.79227126e-01
-5.18148988e-02 -5.26348889e-01 3.93288404e-01 -3.47768873e-01
-4.64647450e-02 -9.44771897e-03 -5.38326383e-01 1.98073074e-01
8.54218960e-01 -8.93197954e-02 -3.38351488e-01 -3.45630080e-01
-1.32538366e+00 1.51730582e-01 -7.65656158e-02 1.91343829e-01
4.69038457e-01 -8.13991129e-01 -8.39293182e-01 6.45843800e-04
-1.82655267e-02 2.46714339e-01 6.15289092e-01 1.13678133e+00
-1.75831005e-01 7.14773238e-01 9.24843103e-02 -7.54505336e-01
-1.33088565e+00 3.53455931e-01 3.27248812e-01 -9.76320148e-01
-6.77483201e-01 7.53718197e-01 3.95446926e-01 -4.13143665e-01
-1.18185595e-01 -6.33361518e-01 1.20136358e-01 -5.68687804e-02
2.69487262e-01 -9.98927373e-03 5.21681964e-01 3.67633253e-01
-2.01387882e-01 2.12606881e-02 -7.28784382e-01 -2.95092672e-01
1.42381191e+00 1.08925909e-01 4.23327953e-01 2.52054721e-01
7.16188729e-01 4.01449762e-02 -8.76677275e-01 -2.00441107e-01
-1.28744155e-01 3.36898752e-02 1.90856203e-01 -1.24591446e+00
-6.42405450e-01 6.69891179e-01 2.76807129e-01 8.47337842e-02
1.28211212e+00 2.10013300e-01 6.09774828e-01 4.72397357e-01
1.05229933e-02 -9.57805812e-01 4.88187335e-02 1.91237643e-01
5.86802959e-01 -9.49955285e-01 4.10452992e-01 -5.50904632e-01
-9.35099125e-01 1.13634789e+00 4.42934752e-01 3.33510101e-01
3.98030192e-01 7.13886440e-01 -2.54784375e-01 -3.45987469e-01
-8.58414471e-01 -4.12319340e-02 7.80194461e-01 6.78692341e-01
6.98525429e-01 1.48443148e-01 -2.48561412e-01 6.99781120e-01
-2.70871997e-01 2.45514169e-01 3.56426954e-01 1.22594583e+00
-8.83207172e-02 -1.25464594e+00 -2.16840699e-01 1.11855495e+00
-6.05293095e-01 -4.94783312e-01 2.86716670e-02 9.66657162e-01
-6.84175268e-02 6.34881198e-01 -1.71503767e-01 -8.12543854e-02
2.81718105e-01 5.67393243e-01 8.21893394e-01 -1.21901286e+00
-8.03339481e-01 4.23981071e-01 5.71441233e-01 -4.05391783e-01
-3.00532252e-01 -8.69158685e-01 -1.19428515e+00 1.33708805e-01
-2.67319947e-01 -4.96404134e-02 6.97433591e-01 1.12468112e+00
5.14668345e-01 1.45120716e+00 -5.11151552e-02 -3.23245615e-01
-6.50893033e-01 -1.29920542e+00 -2.36808583e-02 1.73114732e-01
1.44163087e-01 -5.43323696e-01 2.85831869e-01 4.41826165e-01] | [8.71165657043457, 8.632440567016602] |
35aea6cc-485b-4de5-82ef-ec9a23747c59 | genius-sketch-based-language-model-pre | 2211.10330 | null | https://arxiv.org/abs/2211.10330v1 | https://arxiv.org/pdf/2211.10330v1.pdf | GENIUS: Sketch-based Language Model Pre-training via Extreme and Selective Masking for Text Generation and Augmentation | We introduce GENIUS: a conditional text generation model using sketches as input, which can fill in the missing contexts for a given sketch (key information consisting of textual spans, phrases, or words, concatenated by mask tokens). GENIUS is pre-trained on a large-scale textual corpus with a novel reconstruction from sketch objective using an extreme and selective masking strategy, enabling it to generate diverse and high-quality texts given sketches. Comparison with other competitive conditional language models (CLMs) reveals the superiority of GENIUS's text generation quality. We further show that GENIUS can be used as a strong and ready-to-use data augmentation tool for various natural language processing (NLP) tasks. Most existing textual data augmentation methods are either too conservative, by making small changes to the original text, or too aggressive, by creating entirely new samples. With GENIUS, we propose GeniusAug, which first extracts the target-aware sketches from the original training set and then generates new samples based on the sketches. Empirical experiments on 6 text classification datasets show that GeniusAug significantly improves the models' performance in both in-distribution (ID) and out-of-distribution (OOD) settings. We also demonstrate the effectiveness of GeniusAug on named entity recognition (NER) and machine reading comprehension (MRC) tasks. (Code and models are publicly available at https://github.com/microsoft/SCGLab and https://github.com/beyondguo/genius) | ['Weizhu Chen', 'Nan Duan', 'Hailiang Huang', 'Songqiao Han', 'Yelong Shen', 'Yeyun Gong', 'Biyang Guo'] | 2022-11-18 | null | null | null | null | ['conditional-text-generation', 'sketch-to-text-generation', 'machine-reading-comprehension'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 3.64218265e-01 1.90832570e-01 -2.43658587e-01 -2.22826913e-01
-1.15382636e+00 -6.67833507e-01 8.76242280e-01 -1.03828050e-01
-3.77911478e-01 7.66209126e-01 6.26118481e-01 -6.22466624e-01
3.80159169e-01 -8.02576244e-01 -8.10718596e-01 -4.46433097e-01
4.52489197e-01 6.31466448e-01 -2.48309866e-01 -1.44507766e-01
2.69515783e-01 1.60438195e-02 -1.26342344e+00 5.54857075e-01
1.19821692e+00 4.51482534e-01 4.73197639e-01 8.51504564e-01
-7.70341039e-01 3.66394907e-01 -8.96568358e-01 -6.02371812e-01
8.85074139e-02 -3.75344574e-01 -6.24745131e-01 1.40077658e-02
4.72098917e-01 -4.49220002e-01 -3.61810565e-01 5.09406507e-01
6.93616807e-01 -3.81809808e-02 7.15390682e-01 -1.03732240e+00
-9.27277386e-01 1.13096535e+00 -3.82045090e-01 -2.14949295e-01
4.69974965e-01 1.61367357e-01 1.10910833e+00 -1.50416529e+00
5.49640954e-01 1.14045477e+00 5.22204757e-01 9.01255131e-01
-1.45028055e+00 -9.42675769e-01 1.48559466e-01 -4.62485820e-01
-1.19379652e+00 -5.66700161e-01 6.51158452e-01 -1.43718332e-01
7.83248305e-01 4.69751358e-01 3.04870158e-01 1.71069431e+00
-1.68560639e-01 1.53968143e+00 8.60986233e-01 -8.11686933e-01
5.56019172e-02 9.40046385e-02 7.04386309e-02 5.37870944e-01
9.32610855e-02 -8.69024023e-02 -7.41931021e-01 -3.21269542e-01
6.12159610e-01 -1.29235461e-01 -4.65188771e-01 1.51859865e-01
-1.37296832e+00 7.96446085e-01 -2.06087440e-01 3.23545963e-01
-2.74859369e-01 9.34335664e-02 2.82951534e-01 1.46998674e-01
7.08995402e-01 5.59738338e-01 -6.47596121e-01 -2.85179198e-01
-1.06056476e+00 4.43106532e-01 7.77616978e-01 1.36127079e+00
4.50840652e-01 1.86994329e-01 -6.63169265e-01 1.05581963e+00
-2.53520638e-01 7.64062285e-01 6.68090165e-01 -2.54952878e-01
8.94279599e-01 5.15465021e-01 8.05355459e-02 -5.36225259e-01
-1.50506815e-03 -4.29749459e-01 -8.58901858e-01 -3.49929094e-01
4.11765903e-01 -3.26114833e-01 -1.11696088e+00 1.86899793e+00
1.06493808e-01 -1.21400312e-01 1.06373271e-02 3.49042684e-01
9.18161988e-01 8.95809174e-01 1.18477255e-01 2.08593428e-01
1.41241205e+00 -8.66763771e-01 -7.41075397e-01 -4.10733610e-01
7.84952581e-01 -8.46598804e-01 1.79649031e+00 5.41247249e-01
-9.45697784e-01 -5.57364404e-01 -8.82806659e-01 -2.31404081e-01
-5.16073048e-01 7.60028660e-01 4.59882945e-01 8.74693871e-01
-7.32225180e-01 2.19073817e-01 -6.25303745e-01 -7.33496398e-02
4.25687253e-01 -1.75779521e-01 -3.61165881e-01 -3.28155726e-01
-1.19391894e+00 3.88068825e-01 3.99945736e-01 -2.91121900e-01
-6.95755363e-01 -9.94170189e-01 -1.04584455e+00 4.67955247e-02
5.55805981e-01 -5.91612101e-01 1.26745665e+00 -7.03560650e-01
-1.37642634e+00 5.08988023e-01 -4.10679042e-01 -2.56142378e-01
5.16414165e-01 -6.97238207e-01 -2.15692163e-01 1.70762502e-02
1.34508863e-01 8.03490639e-01 1.08362794e+00 -1.17326272e+00
-2.64362097e-01 -8.27511176e-02 -4.97798860e-01 1.83076918e-01
-5.78720391e-01 -1.34909034e-01 -5.98997653e-01 -1.24555922e+00
-2.09151685e-01 -5.80451906e-01 2.04926562e-02 -2.62311995e-01
-8.09606731e-01 -3.85110199e-01 7.04462111e-01 -9.31991279e-01
1.42855799e+00 -1.98651528e+00 1.07997842e-02 -1.84082277e-02
7.12480024e-02 3.28888834e-01 -4.94376540e-01 8.27742934e-01
-6.99425563e-02 5.28096497e-01 -4.50091809e-01 -8.72347593e-01
2.95847028e-01 1.51348993e-01 -7.93935061e-01 -1.98479384e-01
4.05462474e-01 1.25253201e+00 -7.39772141e-01 -2.96125710e-01
1.16364911e-01 3.52260143e-01 -4.93000060e-01 3.30122799e-01
-6.85393035e-01 1.55390024e-01 -2.37423092e-01 5.87336838e-01
5.92940092e-01 -7.27553144e-02 4.66622338e-02 1.70698017e-01
1.00660339e-01 6.21826828e-01 -1.12335634e+00 1.77830553e+00
-6.27906740e-01 5.17461419e-01 -2.16042027e-01 -4.77903128e-01
9.36932623e-01 3.88826221e-01 -1.49128914e-01 -4.40847456e-01
8.46062973e-02 9.74605605e-02 -5.78113794e-01 -2.97790527e-01
9.31681037e-01 -5.59776314e-02 -3.76015544e-01 9.57310557e-01
2.21221417e-01 -3.95681322e-01 3.36089909e-01 7.59695649e-01
1.01784813e+00 1.98052138e-01 1.46412119e-01 9.11028162e-02
1.79998115e-01 -2.04191044e-01 1.24384522e-01 9.48789954e-01
4.92693663e-01 8.42412651e-01 6.43302321e-01 2.34115943e-01
-1.33911574e+00 -1.05458677e+00 4.85192239e-02 1.19298542e+00
-5.86963117e-01 -7.40019083e-01 -8.65613878e-01 -8.42422605e-01
-5.06509189e-03 1.24665284e+00 -5.84691763e-01 -7.36890957e-02
-4.49240327e-01 -5.82314968e-01 1.09580481e+00 7.21799493e-01
4.19009835e-01 -1.23952901e+00 -3.68830822e-02 1.36663526e-01
-5.43691039e-01 -1.09930265e+00 -9.72014487e-01 -4.22624312e-02
-7.10192680e-01 -5.57302475e-01 -8.03298652e-01 -6.65419161e-01
9.55765247e-01 1.68016374e-01 1.16407943e+00 2.22451687e-01
-1.75832048e-01 4.30929333e-01 -7.04393983e-01 -5.22487760e-01
-6.05865836e-01 4.53490824e-01 -1.41145661e-01 -6.15172312e-02
2.01500356e-01 -2.89424241e-01 -1.91923708e-01 1.54504836e-01
-1.26650441e+00 4.89276707e-01 7.09293246e-01 1.19764543e+00
1.72049984e-01 -1.79083452e-01 8.54248643e-01 -1.17372262e+00
1.13223672e+00 -3.93038988e-01 -2.70742416e-01 3.52182239e-01
-4.32661057e-01 2.10986480e-01 6.68976426e-01 -8.14597368e-01
-1.12375426e+00 -1.68032587e-01 -3.91658843e-01 -2.06656486e-01
-2.42126703e-01 6.83966100e-01 -4.62306619e-01 7.13308871e-01
6.52420938e-01 7.07398832e-01 -1.64248079e-01 -6.74237728e-01
7.69448280e-01 8.92079353e-01 6.02245927e-01 -1.08915472e+00
8.67378771e-01 1.17757648e-01 -7.22307026e-01 -9.48895693e-01
-6.90015733e-01 -2.56712973e-01 -5.19838274e-01 7.95313120e-02
4.57729548e-01 -7.90970504e-01 -2.25028560e-01 6.26205027e-01
-1.07432234e+00 -6.77831829e-01 -2.54499733e-01 1.72359988e-01
-1.34386227e-01 4.37272877e-01 -5.53969562e-01 -1.03461993e+00
-7.84960747e-01 -6.22340262e-01 1.21404076e+00 2.66369674e-02
-5.13095796e-01 -8.58878255e-01 -2.08265126e-01 4.67423171e-01
2.54117250e-01 -5.34583107e-02 1.07583857e+00 -9.06717300e-01
-3.63924623e-01 -4.01350945e-01 -1.59174979e-01 4.99330312e-01
-3.20227109e-02 -8.11156854e-02 -9.11530912e-01 -2.08508536e-01
-2.99684703e-01 -5.28504193e-01 8.60764027e-01 1.59200430e-02
1.22347951e+00 -6.53192222e-01 -2.03646347e-01 3.53983194e-01
1.03439415e+00 -7.42906034e-02 7.87047684e-01 -1.62784144e-01
6.46373749e-01 6.08892024e-01 5.24957240e-01 4.14175808e-01
2.57718503e-01 2.84619093e-01 -4.42615263e-02 -1.82382911e-01
-2.81234205e-01 -9.65152919e-01 5.33428133e-01 8.55388761e-01
4.66414034e-01 -6.27585709e-01 -9.23772514e-01 8.91401827e-01
-1.49598801e+00 -9.05404568e-01 2.28799749e-02 2.17206717e+00
1.38074768e+00 1.68271989e-01 -5.67422993e-02 2.09303930e-01
6.00489795e-01 1.97147466e-02 -2.47758493e-01 -1.59399927e-01
-1.62280887e-01 7.29280531e-01 2.28672072e-01 4.73657876e-01
-8.36724162e-01 1.23902452e+00 5.46121836e+00 1.30005229e+00
-1.00919378e+00 5.89948408e-02 6.17434740e-01 -5.67932501e-02
-7.69458294e-01 -9.05411094e-02 -9.24851894e-01 4.96781915e-01
7.18622029e-01 -3.24583091e-02 3.72410744e-01 6.68493450e-01
-4.21874598e-02 -1.21800147e-01 -9.54565465e-01 7.96355009e-01
2.72299796e-01 -1.42760754e+00 4.38672006e-01 -1.70075685e-01
5.92132330e-01 -2.46987775e-01 2.28601739e-01 6.27092123e-01
3.46739084e-01 -1.10188663e+00 8.79513323e-01 2.86248237e-01
1.30691397e+00 -7.44673491e-01 4.09008920e-01 6.01088047e-01
-9.58247483e-01 1.79225489e-01 -1.75505817e-01 2.10048556e-01
1.46301463e-01 6.23607099e-01 -1.19822371e+00 5.44069350e-01
2.26363719e-01 3.32452655e-01 -6.31105661e-01 4.61830229e-01
-6.58656836e-01 1.12619293e+00 -2.62310624e-01 -3.58027965e-01
7.78024644e-02 3.13816331e-02 6.41908884e-01 1.55864799e+00
2.17948765e-01 1.32377550e-01 -2.98305191e-02 1.23851788e+00
-3.70014608e-01 2.91978925e-01 -5.32483220e-01 -5.36628187e-01
6.43660188e-01 1.26629412e+00 -4.78455424e-01 -5.39457440e-01
-2.41094217e-01 1.12132728e+00 3.02386671e-01 5.63489735e-01
-5.67977905e-01 -7.81707346e-01 2.72150159e-01 1.47598892e-01
3.62401932e-01 -3.85981292e-01 -7.56220698e-01 -1.35330725e+00
2.45020360e-01 -1.19740212e+00 2.28443190e-01 -9.41915929e-01
-1.24956572e+00 6.16243958e-01 -1.17811281e-03 -9.94703948e-01
-1.77939355e-01 -4.49623793e-01 -6.94224477e-01 1.03331661e+00
-1.05227423e+00 -1.44344914e+00 -7.83717185e-02 3.35602254e-01
9.11969781e-01 -2.82038003e-01 8.81759763e-01 1.80482462e-01
-4.39176232e-01 1.05327725e+00 -2.59306263e-02 5.17655551e-01
7.81842589e-01 -1.40350497e+00 1.06790400e+00 9.98309076e-01
2.63793647e-01 9.36938822e-01 5.26384175e-01 -1.00772226e+00
-1.31570125e+00 -1.15935886e+00 1.24574053e+00 -7.46819735e-01
5.40444314e-01 -1.17512226e+00 -1.04031622e+00 7.42788136e-01
2.51696408e-01 -6.66200578e-01 6.94193065e-01 8.31963867e-02
-4.87670064e-01 3.79741997e-01 -9.41724181e-01 1.03288400e+00
9.99193132e-01 -5.66620767e-01 -6.71642363e-01 9.73624662e-02
9.64962304e-01 -4.20674741e-01 -4.12954450e-01 5.20296469e-02
4.64259148e-01 -4.05687749e-01 7.19715118e-01 -5.40176749e-01
6.72430217e-01 1.45178079e-03 -7.64500648e-02 -1.41310501e+00
2.59360373e-01 -8.78021717e-01 1.85635276e-02 1.57779920e+00
7.01675475e-01 -7.06925809e-01 7.46040821e-01 5.42803168e-01
-1.71393543e-01 -6.38538718e-01 -6.59984171e-01 -5.84501743e-01
3.34172726e-01 -9.13334131e-01 8.10183227e-01 9.30878162e-01
1.82609931e-01 5.83478868e-01 -5.70701361e-01 -2.03219041e-01
2.77877182e-01 -1.49000809e-01 1.13473785e+00 -7.56375432e-01
-2.56999999e-01 -3.29422683e-01 4.73726124e-01 -1.37259316e+00
7.26010054e-02 -1.01255071e+00 9.42611843e-02 -1.53674340e+00
1.02353975e-01 -5.62472284e-01 4.45482373e-01 8.54331136e-01
-5.77229321e-01 -6.40379079e-03 3.55271965e-01 7.71143800e-03
-1.14281356e-01 8.58194709e-01 1.25487280e+00 -1.32868141e-01
-2.53830820e-01 1.28581464e-01 -8.54345262e-01 1.94214970e-01
9.61188138e-01 -4.00584459e-01 -3.68652105e-01 -3.37186366e-01
3.09898973e-01 7.35886544e-02 3.09046090e-01 -6.55826926e-01
-3.26926308e-03 5.03532551e-02 5.19023418e-01 -9.20725346e-01
2.45358005e-01 -2.57573992e-01 -2.64551967e-01 1.05204567e-01
-5.60080528e-01 2.14675106e-02 4.93054569e-01 2.70640135e-01
3.75398025e-02 -3.06657404e-01 3.36635292e-01 6.86943755e-02
-1.04868844e-01 1.24965996e-01 -5.18190145e-01 2.36444637e-01
3.38551819e-01 1.74385421e-02 -4.57499862e-01 -6.31801367e-01
-3.39960426e-01 8.52180123e-02 2.56299943e-01 5.78861773e-01
8.00053477e-01 -1.25518870e+00 -1.02602267e+00 5.31269312e-01
1.46704331e-01 1.32923156e-01 7.62826502e-02 2.86977768e-01
-1.29825830e-01 4.12245840e-01 2.46816725e-01 -2.14335322e-01
-1.21091557e+00 4.24867481e-01 -1.28056318e-01 -4.43748534e-01
-5.91436386e-01 7.81531572e-01 3.20425302e-01 -9.02769923e-01
3.08807522e-01 -5.13005555e-01 1.70461997e-01 -1.61088377e-01
7.03860044e-01 9.91824418e-02 -1.94060821e-02 -8.25394467e-02
2.98490733e-01 -1.44978343e-02 -2.91259050e-01 -6.18088961e-01
1.12339127e+00 6.23009242e-02 8.12578052e-02 2.75249839e-01
8.27238262e-01 5.85669398e-01 -9.83170748e-01 -4.05990690e-01
6.31834492e-02 -3.22848231e-01 -2.16209456e-01 -1.17046249e+00
-5.20176947e-01 9.23123300e-01 -1.89533010e-01 -6.87545463e-02
9.99939740e-01 5.10899350e-03 1.03316367e+00 1.48649007e-01
7.72176459e-02 -9.81215477e-01 3.68282795e-01 6.59454584e-01
1.27980328e+00 -7.65829027e-01 -1.23024009e-01 -3.69809330e-01
-9.27433193e-01 9.03739810e-01 7.31157541e-01 2.71986455e-01
3.18168938e-01 4.33665961e-01 -2.81932484e-03 9.74568501e-02
-9.74713385e-01 -5.30184917e-02 4.03811246e-01 5.40293872e-01
5.01707435e-01 1.77259877e-01 -2.67471910e-01 8.82868111e-01
-4.95546669e-01 -2.09714323e-01 4.27114785e-01 9.09066558e-01
-2.05205157e-01 -1.46220720e+00 -5.57190835e-01 5.96396446e-01
-2.86334276e-01 -6.38795137e-01 -6.14308834e-01 7.91334510e-01
-1.62106290e-01 8.25452983e-01 -5.12678251e-02 -3.50951940e-01
1.36748135e-01 6.24741495e-01 2.29088023e-01 -8.39612067e-01
-5.78961313e-01 1.69068843e-01 2.38179132e-01 4.42303866e-02
4.50740784e-01 -6.54810786e-01 -1.26510310e+00 -3.10139954e-01
-3.72853547e-01 1.29329696e-01 6.96391702e-01 7.61567354e-01
4.61960167e-01 3.59179944e-01 2.36804813e-01 -5.22937953e-01
-4.86839443e-01 -1.52401519e+00 -2.14493215e-01 3.15114766e-01
1.42993823e-01 -2.99840808e-01 -3.39653701e-01 1.33146256e-01] | [11.65408992767334, 8.884458541870117] |
e2543179-a76a-49fc-bbd1-acae9b037de4 | magnitude-attention-based-dynamic-pruning | 2306.05056 | null | https://arxiv.org/abs/2306.05056v1 | https://arxiv.org/pdf/2306.05056v1.pdf | Magnitude Attention-based Dynamic Pruning | Existing pruning methods utilize the importance of each weight based on specified criteria only when searching for a sparse structure but do not utilize it during training. In this work, we propose a novel approach - \textbf{M}agnitude \textbf{A}ttention-based Dynamic \textbf{P}runing (MAP) method, which applies the importance of weights throughout both the forward and backward paths to explore sparse model structures dynamically. Magnitude attention is defined based on the magnitude of weights as continuous real-valued numbers enabling a seamless transition from a redundant to an effective sparse network by promoting efficient exploration. Additionally, the attention mechanism ensures more effective updates for important layers within the sparse network. In later stages of training, our approach shifts from exploration to exploitation, exclusively updating the sparse model composed of crucial weights based on the explored structure, resulting in pruned models that not only achieve performance comparable to dense models but also outperform previous pruning methods on CIFAR-10/100 and ImageNet. | ['Jangho Kim', 'Namhyuk Ahn', 'Jihye Back'] | 2023-06-08 | null | null | null | null | ['efficient-exploration'] | ['methodology'] | [ 2.95856953e-01 2.40416154e-01 -3.66146773e-01 -3.99891198e-01
-1.22466087e-01 -5.57663068e-02 1.13383316e-01 1.92565992e-01
-6.73869491e-01 7.57660449e-01 1.70539439e-01 -2.19408795e-01
-5.05900264e-01 -8.64856601e-01 -5.78165054e-01 -5.97635448e-01
-3.12488377e-01 4.24442977e-01 3.49371880e-01 -1.33334249e-01
3.80128086e-01 5.89229107e-01 -1.55844963e+00 3.24370056e-01
9.91547108e-01 1.22249484e+00 5.07965863e-01 8.71463791e-02
-4.22504902e-01 8.39205921e-01 -6.05392456e-01 -1.41126901e-01
4.63527441e-01 -1.69036090e-01 -5.92501521e-01 -4.13937896e-01
4.10394907e-01 -3.03132743e-01 -2.89471090e-01 1.26704001e+00
1.54225498e-01 4.05251652e-01 1.47797748e-01 -7.40162611e-01
-2.47354373e-01 1.23466790e+00 -9.32396591e-01 9.04756367e-01
-1.27325669e-01 -1.24987233e-02 1.05319083e+00 -1.04760420e+00
6.17151380e-01 1.01213777e+00 4.94795501e-01 2.97759414e-01
-1.28744018e+00 -1.09854877e+00 9.67108190e-01 3.44900012e-01
-1.41735053e+00 -4.53698993e-01 8.70203078e-01 -8.61632824e-02
1.33392179e+00 3.30498338e-01 1.19969952e+00 5.03844261e-01
-8.81541595e-02 6.57159984e-01 6.07842565e-01 -3.71432662e-01
3.19403917e-01 6.86792582e-02 3.73582363e-01 7.85518765e-01
5.23806632e-01 1.26937881e-01 -1.02122164e+00 -4.27023023e-02
8.61909926e-01 -2.43388265e-02 -4.14033592e-01 -2.88689345e-01
-9.90165174e-01 8.30125332e-01 8.76866698e-01 4.21706855e-01
-6.85545802e-01 1.92099199e-01 1.29631668e-01 3.02321553e-01
2.21090332e-01 7.89937854e-01 -4.29195076e-01 -2.37954691e-01
-1.52062500e+00 8.32432583e-02 5.51168442e-01 7.42357373e-01
1.03984344e+00 4.80627030e-01 -2.98050642e-01 7.98331380e-01
2.69352764e-01 -1.55670330e-01 4.78009135e-01 -8.36871028e-01
8.37094367e-01 9.88393068e-01 -2.75959134e-01 -1.08267474e+00
-1.84209839e-01 -9.73481417e-01 -9.48544502e-01 1.85066342e-01
-2.57776886e-01 -9.70509499e-02 -1.06481051e+00 1.82234776e+00
3.86840582e-01 1.75417796e-01 -1.83164120e-01 5.93568265e-01
7.84658909e-01 4.84254897e-01 2.43654877e-01 -1.37816936e-01
8.81868780e-01 -1.08415723e+00 -4.20827419e-01 -4.07428712e-01
4.32174414e-01 -2.89075047e-01 1.01402330e+00 6.84777141e-01
-1.34149623e+00 -5.75635135e-01 -1.20053566e+00 5.25716022e-02
-1.87605351e-01 1.57439291e-01 8.22153211e-01 4.38224792e-01
-1.31085336e+00 9.47459579e-01 -7.90149570e-01 1.84973001e-01
8.53141606e-01 7.19560385e-01 -1.49906039e-01 -6.86040521e-02
-1.09436214e+00 7.28015542e-01 7.66570687e-01 2.23004788e-01
-1.01055348e+00 -9.11222696e-01 -8.38562191e-01 6.59354627e-01
1.64412111e-01 -4.51345235e-01 8.72527778e-01 -1.03416431e+00
-1.21150601e+00 4.07245100e-01 -1.89270943e-01 -8.70323837e-01
2.06117764e-01 -3.24287564e-01 -1.64339796e-01 3.76236677e-01
-1.33264840e-01 1.22671378e+00 1.07031178e+00 -1.11599493e+00
-7.41986632e-01 -1.05020449e-01 1.59283727e-01 4.06937540e-01
-9.20703709e-01 -2.81905890e-01 -5.09847045e-01 -9.01280642e-01
4.35791165e-01 -4.21104521e-01 -5.28774083e-01 7.09445849e-02
-3.74194086e-01 6.14845157e-02 8.28578591e-01 -2.74359673e-01
2.02580953e+00 -2.08703303e+00 3.09421569e-01 6.34385943e-01
6.43448770e-01 4.45629686e-01 -2.80681878e-01 7.55109033e-03
-1.53788373e-01 1.64046809e-01 -2.29680926e-01 -2.37063155e-01
-4.46767688e-01 3.25638533e-01 -3.69093746e-01 1.17806897e-01
1.78830177e-01 6.67609811e-01 -8.20257723e-01 -6.38292015e-01
7.84377083e-02 3.33208054e-01 -8.71960938e-01 -1.58370167e-01
5.55951372e-02 1.18738875e-01 -6.59979641e-01 7.84907520e-01
6.50206029e-01 -2.80925930e-01 1.73003092e-01 -4.27478664e-02
-1.67805985e-01 3.00123274e-01 -1.12845862e+00 1.45877707e+00
-4.05452281e-01 3.41106564e-01 3.71926039e-01 -1.04176295e+00
1.10280120e+00 -1.33377239e-01 4.51823473e-01 -7.63018608e-01
8.19563568e-02 2.26050895e-02 3.79441291e-01 4.49837036e-02
5.48487961e-01 3.15162987e-01 4.22672391e-01 3.63379210e-01
1.08033828e-01 1.48710579e-01 6.79981470e-01 3.69687259e-01
1.20785654e+00 -3.94284315e-02 1.64434716e-01 -6.24966264e-01
4.91841942e-01 -2.94358172e-02 7.59548664e-01 8.88933182e-01
8.45373943e-02 2.82161981e-01 5.74752569e-01 -4.72019047e-01
-5.41378796e-01 -8.28382134e-01 -2.19171673e-01 1.30492806e+00
1.44910723e-01 -4.92639184e-01 -4.75357801e-01 -6.24222219e-01
5.02388179e-03 7.66313314e-01 -8.89546156e-01 -4.87464994e-01
-9.98276293e-01 -6.15242481e-01 1.55712932e-01 6.92838907e-01
6.14142835e-01 -1.43820179e+00 -9.37717199e-01 3.22627753e-01
2.05810875e-01 -5.23363352e-01 -2.85454154e-01 9.16837156e-01
-1.44878876e+00 -9.07057762e-01 -6.05699241e-01 -7.12039828e-01
1.12063479e+00 1.46583319e-01 1.14040852e+00 5.38828552e-01
-2.57139504e-01 -2.13014081e-01 -2.53514647e-01 -2.15909407e-01
2.46844485e-01 4.53283221e-01 -2.60246277e-01 -2.99215138e-01
2.39189640e-01 -1.06589949e+00 -7.42168307e-01 4.58186977e-02
-6.24556065e-01 2.44454846e-01 8.38653564e-01 7.03972757e-01
9.77277339e-01 1.74355078e-02 6.11692667e-01 -8.82029295e-01
6.97534025e-01 -5.42983949e-01 -5.11617899e-01 1.01553604e-01
-7.59507298e-01 1.75491378e-01 6.89637125e-01 -7.02586174e-01
-9.51909721e-01 -9.12098810e-02 -1.92193329e-01 -8.52745593e-01
4.00522530e-01 6.92935884e-01 1.76388636e-01 -3.10470462e-01
6.61689997e-01 2.90678918e-01 -4.23076242e-01 -5.93135655e-01
1.50296027e-02 -1.68686852e-01 2.49885187e-01 -6.71617866e-01
6.89718843e-01 4.63990837e-01 -1.10296801e-01 -4.53887492e-01
-9.65092719e-01 -4.61640298e-01 -5.45864403e-01 -4.57053855e-02
2.40376204e-01 -8.71341825e-01 -3.14980745e-01 -8.83094370e-02
-6.16197109e-01 -4.62537676e-01 -9.08294678e-01 4.88780797e-01
-4.98212501e-02 8.94918665e-02 -4.89683956e-01 -5.91213107e-01
-7.33293056e-01 -1.05492270e+00 6.04527950e-01 3.33699733e-01
-3.41008931e-01 -6.47270739e-01 -1.83231518e-01 -1.96080446e-01
6.20232105e-01 -5.25865182e-02 1.10430646e+00 -4.80022848e-01
-6.47470474e-01 -1.14217117e-01 -3.24006915e-01 1.27610058e-01
-1.98731169e-01 -2.06123203e-01 -5.59011936e-01 -3.50671589e-01
8.11493546e-02 -3.30666184e-01 1.61227202e+00 4.79783356e-01
1.54177701e+00 -7.20282972e-01 -4.93744195e-01 1.06951010e+00
1.38427866e+00 1.24838993e-01 6.28803194e-01 4.73798215e-01
4.86791641e-01 3.13021719e-01 5.53176701e-01 6.93300784e-01
-3.04068206e-03 6.89945295e-02 8.12131882e-01 -4.89445440e-02
-1.27761096e-01 -2.25427419e-01 -1.35710603e-02 6.33109689e-01
2.36310065e-02 1.70819908e-02 -6.20939314e-01 6.75300419e-01
-1.48718560e+00 -9.09937203e-01 3.53076845e-01 2.19094038e+00
1.22854340e+00 6.11159682e-01 -3.87814999e-01 4.29628864e-02
5.63686192e-01 3.72121274e-01 -8.51395547e-01 -4.70997572e-01
4.85902354e-02 6.27418995e-01 2.71581680e-01 3.70502859e-01
-7.33755529e-01 1.04973447e+00 6.16661406e+00 8.98563981e-01
-1.19734764e+00 -4.11231723e-03 7.47268081e-01 -7.95570076e-01
-5.84472716e-01 7.77999088e-02 -1.21072829e+00 1.89001873e-01
4.55914259e-01 7.68540278e-02 5.03240526e-01 1.21675527e+00
-3.78602440e-03 -4.51053351e-01 -8.26488078e-01 8.10641766e-01
-1.95916548e-01 -1.54114461e+00 3.69905293e-01 -9.25968960e-02
8.73992264e-01 1.26983494e-01 2.53199011e-01 5.27190387e-01
3.06483060e-01 -1.10249233e+00 8.47224116e-01 4.77372617e-01
6.92937672e-01 -7.68411934e-01 2.55325109e-01 4.25990522e-01
-1.47374249e+00 -4.38606739e-01 -5.96106827e-01 1.54317409e-01
3.99038866e-02 8.63644540e-01 -4.13329124e-01 -3.01410705e-02
1.06980491e+00 6.06829107e-01 -5.70785522e-01 1.13436484e+00
-4.16489691e-01 6.91973090e-01 -5.64283073e-01 8.01273733e-02
3.81648839e-01 -2.04102054e-01 6.00880027e-01 1.11573505e+00
3.79725516e-01 -3.70234698e-02 7.56101683e-02 1.00555766e+00
-2.55312085e-01 5.95449703e-03 -1.10822089e-01 -2.74443626e-03
6.89390600e-01 1.33219302e+00 -9.70236242e-01 -4.38439876e-01
-1.10422159e-02 3.19613189e-01 6.96399510e-01 3.07449311e-01
-7.01732993e-01 -3.88195455e-01 4.57041830e-01 2.39849657e-01
8.23467731e-01 -6.90094978e-02 -5.33017993e-01 -9.10307169e-01
-9.03925002e-02 -8.15786600e-01 3.70730639e-01 -2.77954638e-01
-5.16295195e-01 1.01268220e+00 6.18601851e-02 -8.90614569e-01
9.70194712e-02 1.76510736e-02 -8.94841194e-01 9.23569560e-01
-1.78185642e+00 -7.62637496e-01 -3.73533875e-01 5.98935246e-01
6.50581598e-01 -3.00788641e-01 4.69282031e-01 2.00203061e-01
-6.17902637e-01 6.12981856e-01 -2.61021018e-01 -3.50066543e-01
3.52969840e-02 -9.74950790e-01 2.66934901e-01 8.51313949e-01
2.87699491e-01 1.08336759e+00 3.30497056e-01 -9.11658466e-01
-8.06365907e-01 -1.01770914e+00 5.96697271e-01 4.79953080e-01
3.69994164e-01 -1.56006917e-01 -1.15750706e+00 4.13466871e-01
-6.07747361e-02 4.89988364e-02 4.36169416e-01 3.06065500e-01
-2.92719066e-01 -2.50104159e-01 -1.08418429e+00 5.70137203e-01
1.24248278e+00 2.53165394e-01 -4.73845840e-01 -3.59547487e-03
7.22715855e-01 -3.42765391e-01 -6.42653644e-01 6.89963996e-01
3.34378839e-01 -1.03889060e+00 9.86554742e-01 -2.68067539e-01
3.34748834e-01 -1.14264846e-01 3.88053767e-02 -9.18379962e-01
-5.93077600e-01 -5.06073833e-01 -9.20836270e-01 6.91696823e-01
5.76321721e-01 -5.04987061e-01 1.15225673e+00 1.20728143e-01
-2.97964215e-01 -1.40285861e+00 -1.00269687e+00 -3.34261745e-01
-3.05988789e-01 -2.40880117e-01 7.51095772e-01 7.61158705e-01
-3.19413275e-01 -7.75774866e-02 3.60618420e-02 -7.68098459e-02
5.04386365e-01 2.41393715e-01 2.09627017e-01 -1.27302444e+00
-4.49267328e-01 -8.06028306e-01 1.03523530e-01 -1.28722334e+00
-6.41434714e-02 -8.21992636e-01 -1.64771765e-01 -1.55313957e+00
1.93512887e-01 -9.46864307e-01 -6.98320389e-01 7.97444880e-01
-1.04912594e-01 1.49378225e-01 1.66418225e-01 4.20881212e-01
-7.57218778e-01 7.20123827e-01 1.22653639e+00 -9.38887075e-02
-4.73158270e-01 2.17513144e-02 -1.02538717e+00 7.96352029e-01
7.99775958e-01 -5.80462456e-01 -6.44357800e-01 -6.97708607e-01
4.39196497e-01 -2.23009199e-01 1.09575354e-01 -1.42973256e+00
3.51308227e-01 -1.55004129e-01 7.24734843e-01 -7.64692724e-01
4.51558560e-01 -7.15898275e-01 -1.99158058e-01 6.84814692e-01
-4.75911409e-01 1.87195614e-01 4.49225068e-01 5.15942931e-01
-2.67383099e-01 -5.54735005e-01 8.74276578e-01 -4.01829571e-01
-8.23396444e-01 5.35719633e-01 -1.46509007e-01 4.18845750e-02
8.89284492e-01 -7.33955622e-01 -5.25146425e-02 1.70022063e-02
-9.99131620e-01 4.24002051e-01 7.63507485e-02 1.35276318e-01
9.26416278e-01 -9.88054335e-01 -2.99300253e-01 5.45694351e-01
-2.96424389e-01 3.42183173e-01 4.79977041e-01 7.66531527e-01
-5.85948586e-01 3.71991515e-01 -1.78606689e-01 -4.07202303e-01
-1.05888140e+00 4.78594422e-01 1.86667874e-01 -7.16196179e-01
-7.81408608e-01 1.45061624e+00 2.93804854e-01 6.57988340e-02
6.89956427e-01 -3.35550606e-01 -5.28657258e-01 6.51626661e-02
5.97298026e-01 2.35308453e-01 1.48719981e-01 -3.20009261e-01
-2.12872624e-01 2.58776635e-01 -4.54852104e-01 1.19716644e-01
1.59294116e+00 6.33479133e-02 -1.69753402e-01 -1.37983533e-02
8.57096910e-01 -7.35897496e-02 -1.58489811e+00 -2.68815607e-01
-1.30960345e-01 -5.50179839e-01 3.44120294e-01 -8.00960064e-01
-1.65108252e+00 7.55549669e-01 3.68255615e-01 -7.27556273e-02
1.36922991e+00 -2.50817150e-01 6.96419835e-01 5.68265259e-01
3.04895163e-01 -1.23376906e+00 3.67964178e-01 5.87961555e-01
1.04712331e+00 -7.05675304e-01 3.82962942e-01 -1.45127788e-01
-5.34350336e-01 9.14880574e-01 1.10731542e+00 -2.32542962e-01
8.66518974e-01 4.82162446e-01 -4.50151056e-01 -2.77602404e-01
-9.58296835e-01 -4.40336168e-02 7.94907212e-02 4.59694058e-01
1.54593274e-01 -4.22851712e-01 -1.77863359e-01 4.69785720e-01
-2.85515368e-01 -5.66224009e-02 -1.08666822e-01 1.14270067e+00
-9.59422946e-01 -8.71920943e-01 -1.29403099e-01 1.10786843e+00
-3.52708369e-01 -7.52387524e-01 5.80153707e-03 7.33582914e-01
3.74539018e-01 4.15422112e-01 1.92319736e-01 -1.19126827e-01
1.87364340e-01 -2.76944250e-01 4.18172866e-01 -9.35384929e-01
-1.06101370e+00 5.76249771e-02 -1.26160875e-01 -6.33720398e-01
-2.90311258e-02 -4.68305469e-01 -1.30098188e+00 -1.19142272e-01
-5.52723765e-01 2.97141582e-01 3.45876306e-01 6.38910711e-01
4.70352143e-01 8.58557045e-01 4.01460856e-01 -8.42303634e-01
-6.22494340e-01 -8.15715611e-01 -5.17393708e-01 -2.67712891e-01
9.69603583e-02 -7.36321807e-01 -2.00991884e-01 -3.66719395e-01] | [8.593972206115723, 3.208749532699585] |
38d7aa4b-583a-45f6-96ce-c3d45812f379 | interlayer-ferromagnetism-and-high | 2101.07684 | null | https://arxiv.org/abs/2101.07684v1 | https://arxiv.org/pdf/2101.07684v1.pdf | Interlayer Ferromagnetism and High-Temperature Quantum Anomalous Hall Effect in \textit{p}-Doped MnBi$_2$Te$_4$ Multilayers | The interlayer antiferromagnetic coupling hinders the observation of quantum anomalous Hall effect in magnetic topological insulator MnBi$_2$Te$_4$. We demonstrate that interlayer \textit{ferromagnetism} can be established by utilizing the \textit{p}-doping method in MnBi$_2$Te$_4$ multilayers. In two septuple-layers system, the interlayer ferromagnetic coupling appears by doping nonmagnetic elements (e.g., N, P, As, Na, Mg, K, and Ca), due to the redistribution of orbital occupations of Mn. We further find that Mg and Ca elements are the most suitable candidates because of their low formation energy. Although, the \textit{p}-doped two septuple layers exhibit topologically trivial band structure, the increase of layer thickness to three (four) septuple layers with Ca (Mg) dopants leads to the formation of the quantum anomalous Hall effect. Our proposed \textit{p}-doping strategy not only makes MnBi$_2$Te$_4$ become an ideal platform to realize the high-temperature quantum anomalous Hall effect without external magnetic field, but also can compensate the electrons from the intrinsic \textit{n}-type defects in MnBi$_2$Te$_4$. | ['Zhenhua Qiao', 'Xiaohong Xu', 'Shifei Qi', 'Shiyang Sun', 'Yulei Han'] | 2021-01-19 | null | null | null | null | ['formation-energy'] | ['miscellaneous'] | [ 3.77196670e-01 1.90543197e-02 -2.73624241e-01 1.24610122e-02
5.17006256e-02 2.81174421e-01 4.83095795e-01 -3.92688841e-01
-2.74243802e-01 1.09406459e+00 -1.96798474e-01 -2.19223246e-01
-2.02857137e-01 -9.50003743e-01 -7.03043103e-01 -1.50400019e+00
1.42456636e-01 4.47153687e-01 5.63253045e-01 -5.63630342e-01
5.25446534e-01 -2.76489053e-02 -1.87874711e+00 4.64080274e-01
1.13629091e+00 1.34809494e+00 3.91993582e-01 -1.72044467e-02
-8.42854101e-03 3.28390807e-01 -3.68598610e-01 4.44011539e-02
1.10031389e-01 -5.95425665e-01 -5.80874085e-01 -2.02382028e-01
3.04332018e-01 -1.53628498e-01 -4.80027467e-01 1.16122961e+00
-5.05712144e-02 3.24990079e-02 9.94283676e-01 -7.70967007e-01
-9.45529878e-01 1.02635205e+00 -9.86368060e-01 -6.62315115e-02
-1.41868778e-02 3.60761344e-01 8.19253743e-01 -5.82557321e-01
6.60598397e-01 8.54020238e-01 8.98055583e-02 1.00639856e+00
-1.00280261e+00 -1.09698701e+00 -2.04478994e-01 3.57053339e-01
-1.42517424e+00 -4.75945890e-01 8.88840914e-01 -4.41839635e-01
1.28035140e+00 6.58275902e-01 5.81641436e-01 2.14150608e-01
5.19165933e-01 5.08119315e-02 1.60096097e+00 -6.48703575e-01
-1.99364424e-01 2.24594809e-02 5.43705523e-01 4.77233887e-01
8.85472238e-01 -9.78875556e-04 -5.56084692e-01 -2.78050415e-02
4.04789537e-01 -6.14915825e-02 -4.70344037e-01 2.02235386e-01
-7.79693365e-01 -1.47336662e-01 3.86395544e-01 7.49886096e-01
-2.79059768e-01 4.60423619e-01 2.87638336e-01 2.64809191e-01
-6.36279732e-02 2.92688310e-01 -4.81603965e-02 -4.06180285e-02
-7.23673642e-01 1.55467629e-01 2.10838750e-01 8.79101098e-01
6.86865568e-01 2.63266355e-01 7.54773840e-02 7.93942213e-01
3.79023075e-01 8.45550120e-01 7.17106014e-02 -7.76947439e-01
2.39088655e-01 4.66631979e-01 3.72638047e-01 -2.68101752e-01
-1.66581914e-01 2.29880258e-01 -1.08726513e+00 -9.86241996e-02
-8.37337300e-02 3.11504990e-01 -8.06383669e-01 1.39607823e+00
1.39134824e-02 -6.21112645e-01 4.26745936e-02 6.70252621e-01
9.37589467e-01 6.18423522e-01 -5.45808524e-02 -8.51733744e-01
1.43341744e+00 -5.40111482e-01 -8.52208138e-01 2.26483673e-01
5.38779199e-01 -4.51755494e-01 9.46545184e-01 1.38595775e-01
-1.43579042e+00 -1.05880000e-01 -1.31085312e+00 3.40394139e-01
5.14166243e-02 -2.51825750e-01 4.53304470e-01 7.42609143e-01
-4.89956647e-01 9.11843538e-01 -8.05373132e-01 -9.75243822e-02
-8.45536515e-02 7.93772757e-01 -2.25803316e-01 -4.59646657e-02
-1.16844749e+00 3.09754878e-01 3.95022392e-01 3.44189078e-01
-3.23929012e-01 -3.22336674e-01 -1.78146228e-01 -4.22100425e-01
7.31194392e-02 -5.14701664e-01 5.66901028e-01 -1.50181115e-01
-1.36315417e+00 1.14580965e+00 -5.41479468e-01 -1.10176705e-01
-2.70744674e-02 4.95317519e-01 -8.90994787e-01 1.28104478e-01
4.04506117e-01 3.05483103e-01 5.53307891e-01 -1.23254609e+00
-3.17203492e-01 -3.58786047e-01 -3.66009980e-01 -2.57227272e-01
-4.95786488e-01 1.76983252e-02 3.20974380e-01 5.10844402e-04
1.16147006e+00 -9.02305841e-01 -3.83054353e-02 -1.33077502e+00
-8.17953646e-01 -2.93483704e-01 9.92088258e-01 -8.37138668e-02
1.39752781e+00 -2.03824234e+00 -2.83810824e-01 6.62963092e-01
6.31060064e-01 -2.38763373e-02 8.24617982e-01 6.00368798e-01
1.44181162e-01 5.23906589e-01 -2.58875191e-01 6.99873567e-01
1.05411693e-01 -1.52431568e-02 3.96249413e-01 5.49220264e-01
-4.20354247e-01 6.29910469e-01 -4.82671231e-01 6.13499247e-02
-1.90598872e-02 1.44091606e-01 -8.43630970e-01 -6.47805274e-01
-3.58256280e-01 7.99274683e-01 -4.21981633e-01 6.60298705e-01
1.36825001e+00 -6.51360393e-01 8.22293699e-01 -3.79230857e-01
-7.70890951e-01 8.05492759e-01 -8.69000793e-01 5.38622499e-01
4.53349859e-01 1.18190777e-02 3.42740178e-01 -8.54840755e-01
8.08963239e-01 4.02543485e-01 6.26792610e-01 -1.62899745e+00
-1.09445341e-01 9.92157400e-01 6.58247232e-01 -4.88570720e-01
6.95060253e-01 -7.30273247e-01 -1.56543836e-01 5.08194745e-01
-4.03400779e-01 -1.16436295e-01 5.73098660e-01 -1.87772021e-01
1.00591290e+00 -6.06525540e-01 -3.44992995e-01 -1.25192571e+00
9.50708389e-01 -2.49287948e-01 6.85827076e-01 9.78502214e-01
1.44442305e-01 -1.43726543e-01 4.95740503e-01 -4.65662450e-01
-1.37650895e+00 -7.17736840e-01 -9.37887490e-01 6.55646443e-01
6.09682977e-01 -5.70135176e-01 -6.93942487e-01 1.09353438e-01
-8.98837373e-02 2.40317553e-01 -4.47263792e-02 -2.21343413e-01
-8.28050137e-01 -1.68240535e+00 2.92813540e-01 2.77110249e-01
1.15268731e+00 -1.01623642e+00 -2.21711904e-01 2.90551454e-01
9.09887254e-03 -7.94911683e-01 1.80157781e-01 3.29627603e-01
-8.37203741e-01 -6.38633072e-01 -1.76125154e-01 -6.11813545e-01
8.97932112e-01 -2.59695128e-02 8.30867529e-01 6.71858191e-01
-6.62463978e-02 -2.40309946e-02 2.26546321e-02 9.29495618e-02
-4.43755567e-01 -1.08869113e-01 6.88696861e-01 -4.83210832e-01
5.86826324e-01 -8.67931902e-01 -9.12591040e-01 6.29562855e-01
-6.43774331e-01 1.98179275e-01 1.17363021e-01 8.09164464e-01
6.60750329e-01 5.20267248e-01 4.93960649e-01 -8.07049334e-01
1.18982904e-01 5.15371338e-02 -6.23640656e-01 -2.30916694e-01
-4.41232145e-01 9.86681879e-03 5.62044382e-01 1.60412431e-01
-1.09323013e+00 -1.09431946e+00 -3.47143590e-01 4.83634889e-01
-4.50747795e-02 3.99867922e-01 -4.15567368e-01 -3.22465807e-01
4.33644474e-01 1.51056916e-01 -4.67283607e-01 -4.59148586e-01
-5.02190828e-01 7.93612063e-01 9.09692347e-02 -8.31747234e-01
5.66919565e-01 6.07438266e-01 5.59021175e-01 -1.35953045e+00
-3.79576147e-01 3.54015619e-01 -3.49200338e-01 3.72265130e-02
8.87072742e-01 -5.21210432e-01 -1.28552318e+00 5.25505662e-01
-7.96886206e-01 -2.43433997e-01 -5.84126748e-02 5.48453331e-01
-6.88616559e-02 1.74762532e-01 -1.19729483e+00 -8.79817665e-01
-2.05755755e-01 -1.34009171e+00 3.05798680e-01 -9.33972225e-02
3.36521342e-02 -7.26572812e-01 -7.73491621e-01 7.93903172e-01
6.41733944e-01 1.29242511e-02 1.90225148e+00 -5.82335368e-02
-1.22385347e+00 -6.70200884e-02 -3.25260967e-01 2.94715494e-01
2.39754215e-01 -3.73058885e-01 -6.64166868e-01 -1.62314340e-01
3.14523488e-01 1.65852591e-01 1.18389559e+00 5.17563701e-01
1.01216376e+00 -2.39797056e-01 -6.32627904e-01 5.74980825e-02
1.21156192e+00 8.54587018e-01 9.80710685e-01 4.46068376e-01
1.16423380e+00 -3.97148617e-02 2.37914950e-01 1.53182372e-01
4.21997607e-01 4.28539902e-01 1.21015191e-01 3.21243107e-01
2.25400493e-01 2.73769516e-02 1.96678549e-01 1.68311691e+00
-8.31467986e-01 -3.09073497e-02 -1.07144594e+00 1.43654808e-01
-1.30109072e+00 -9.07658875e-01 -9.33875680e-01 2.58335042e+00
8.05658638e-01 4.88396227e-01 -3.94155324e-01 4.82588232e-01
9.55911756e-01 1.70920283e-01 -4.29655761e-01 -2.00923622e-01
-3.96199346e-01 2.61241794e-01 6.24178231e-01 3.74468207e-01
-5.01557052e-01 8.23922336e-01 5.27980804e+00 7.46597230e-01
-1.46643043e+00 2.36818269e-01 4.00636673e-01 -2.34881774e-01
-7.03285396e-01 4.20902371e-01 -1.03706324e+00 9.03256416e-01
5.39279103e-01 3.25147241e-01 7.53177851e-02 6.60646781e-02
3.38220596e-02 -5.35108149e-01 -1.01783395e+00 9.51756656e-01
-4.87779021e-01 -1.65643525e+00 2.17890933e-01 6.58037245e-01
1.05574071e+00 -2.41957471e-01 1.61865070e-01 -1.42600402e-01
-5.40080726e-01 -8.51919949e-01 4.98549908e-01 3.41109902e-01
1.24361265e+00 -6.84996486e-01 1.62870273e-01 1.87190905e-01
-1.13446391e+00 -4.16336618e-02 -4.35513109e-01 -4.49654728e-01
2.16113955e-01 9.70818520e-01 -1.09210826e-01 2.48698696e-01
1.08813262e+00 3.35558772e-01 2.44304240e-01 3.64551723e-01
2.70757973e-01 4.63530183e-01 -1.51598871e-01 -9.11807790e-02
2.09218204e-01 -7.59177983e-01 6.95151269e-01 3.61349106e-01
2.33349144e-01 3.03761214e-01 1.40539587e-01 9.51453328e-01
-3.98464441e-01 -1.77064061e-01 -5.09558916e-01 4.42885607e-02
4.84614104e-01 6.07661724e-01 -9.22621667e-01 -6.65236890e-01
2.72425618e-02 4.91575003e-01 -3.15548629e-01 2.19332606e-01
-3.08296084e-01 -6.45692229e-01 4.52062607e-01 1.22295368e+00
3.12189996e-01 -1.09649867e-01 -3.84913445e-01 -1.12230766e+00
4.84654933e-01 -4.41240728e-01 -2.54418522e-01 -1.30532548e-01
-8.98420513e-01 3.86751652e-01 7.45872641e-03 -6.25334680e-01
4.15563881e-01 -6.77460790e-01 -4.49984789e-01 7.77468383e-01
-1.24643981e+00 -5.79304993e-01 3.12994011e-02 3.32159065e-02
-5.66834271e-01 2.87370116e-01 4.24420416e-01 6.39486074e-01
-5.27899683e-01 3.70474547e-01 8.05688083e-01 -3.67651731e-01
2.83111393e-01 -7.80100048e-01 6.09375313e-02 2.80884773e-01
-1.02221882e+00 8.79831195e-01 8.09811592e-01 -9.29180324e-01
-1.74763238e+00 -6.16784513e-01 9.53297675e-01 -1.14509568e-01
4.42611635e-01 -5.58912098e-01 -8.31799507e-01 4.73984212e-01
1.80994272e-01 -3.48107845e-01 6.91943407e-01 -3.86044770e-01
3.08305711e-01 -3.51877660e-01 -1.52717996e+00 7.33751595e-01
1.53312671e+00 -5.30009508e-01 4.67979610e-02 4.60906446e-01
4.85793740e-01 -2.15359375e-01 -1.18196106e+00 1.04573488e+00
5.21647513e-01 -1.35731339e+00 6.01202190e-01 -6.33387491e-02
4.12384212e-01 -3.94512832e-01 -4.59529191e-01 -5.51514149e-01
-4.35486674e-01 -4.91364568e-01 3.05005401e-01 1.22756672e+00
5.75778246e-01 -1.02765238e+00 7.72852719e-01 7.10703373e-01
-3.24234724e-01 -4.02399361e-01 -1.21473944e+00 -7.29874313e-01
1.69639409e-01 -5.04432917e-02 8.78686905e-01 8.64936829e-01
4.07092929e-01 2.65481591e-01 -6.25974119e-01 -2.08399713e-01
7.01493204e-01 7.19033852e-02 1.40290514e-01 -1.54430640e+00
4.45689112e-02 -1.93486348e-01 -2.99885511e-01 -1.23371375e+00
1.94988519e-01 -1.04450440e+00 -3.10193121e-01 -1.16095173e+00
9.21851918e-02 -1.24566555e+00 -5.62346756e-01 2.92495219e-03
6.24788821e-01 4.62187856e-01 -2.03177795e-01 5.56460679e-01
-9.28521216e-01 5.07854462e-01 1.57333136e+00 -1.17196120e-01
-3.51306796e-01 -4.00041372e-01 -5.18694878e-01 7.52950370e-01
4.96549487e-01 -4.08367097e-01 2.61761278e-01 -2.92669654e-01
7.26610243e-01 3.40824991e-01 4.17500734e-02 -1.33833051e+00
-7.74429068e-02 -1.81595206e-01 2.24644169e-01 -6.33112192e-01
4.82282072e-01 -4.88905430e-01 3.73095572e-01 6.19121969e-01
5.70445657e-01 -3.09671521e-01 -1.48989424e-01 -5.17945029e-02
-7.37542063e-02 -3.08815658e-01 1.00250006e+00 -8.51254165e-01
-2.91579485e-01 2.88405530e-02 -8.44116211e-01 -2.20941678e-01
6.68616891e-01 -7.77504325e-01 -4.19742763e-01 4.11125809e-01
-6.16699040e-01 8.95877928e-02 9.59540725e-01 -2.43124083e-01
2.02276081e-01 -1.41976571e+00 9.95722190e-02 5.16289115e-01
-5.82651496e-02 -1.05651952e-01 1.12926579e+00 1.40086901e+00
-5.89907110e-01 6.69371843e-01 -3.86086047e-01 -1.08437276e+00
-9.05943394e-01 7.11107701e-02 4.39840406e-01 -1.35008171e-01
-6.85660422e-01 6.64224386e-01 3.21243465e-01 -1.45431772e-01
-4.46222872e-01 -1.72686011e-01 7.85358399e-02 -4.45263594e-01
2.55793661e-01 6.18927658e-01 2.91777611e-01 -7.06111491e-01
-3.88153374e-01 5.42930365e-01 -4.49838668e-01 9.29487646e-02
1.11827409e+00 -2.04027951e-01 -1.15598702e+00 5.19423544e-01
7.41041839e-01 1.54035315e-01 -4.81312603e-01 -7.42055534e-04
6.20516166e-02 -3.22512716e-01 -5.29655457e-01 -1.14013135e-01
-9.27771926e-01 5.07720828e-01 4.59266514e-01 4.54848140e-01
7.58158267e-01 3.91377538e-01 9.14309084e-01 2.87730783e-01
7.06748962e-01 -1.48218477e+00 -1.61069781e-01 4.42680061e-01
1.81998149e-01 -7.09974110e-01 -1.93152249e-01 -4.09838736e-01
1.00039886e-02 6.72536612e-01 1.14182568e+00 1.09851956e-01
8.29215348e-01 -2.94941477e-02 -4.75076854e-01 -8.02565217e-01
-6.93288684e-01 3.09628427e-01 -5.75687289e-02 1.77711993e-01
8.87944341e-01 4.27257836e-01 -9.04980481e-01 4.91236895e-01
-5.88487566e-01 -3.21145087e-01 7.41044939e-01 1.26802969e+00
-6.91324711e-01 -1.44218636e+00 -5.03705442e-01 1.14510071e+00
-5.31130433e-01 -2.90015161e-01 8.17663521e-02 5.79278052e-01
6.12851560e-01 6.40902042e-01 2.90285677e-01 -3.43759179e-01
1.79090261e-01 2.49127015e-01 8.03275108e-01 -1.47233412e-01
-1.26435459e-01 2.60489762e-01 -2.45759301e-02 1.67246372e-01
-4.86368477e-01 -2.06464007e-01 -1.61493587e+00 -8.05502772e-01
-4.45962101e-01 6.22869015e-01 2.18976587e-01 1.09183025e+00
1.64273620e-01 3.17949623e-01 3.96383315e-01 -8.58565271e-01
1.97954029e-01 -8.90278995e-01 -1.38622236e+00 4.48474199e-01
8.80143419e-02 -9.12134945e-01 -2.26989910e-01 -7.37276614e-01] | [5.438369274139404, 4.814970016479492] |
2fd6761a-c628-4ab9-b1d7-80f59b2e0ad5 | ai-planning-annotation-for-sample-efficient | 2203.00669 | null | https://arxiv.org/abs/2203.00669v2 | https://arxiv.org/pdf/2203.00669v2.pdf | Hierarchical Reinforcement Learning with AI Planning Models | Two common approaches to sequential decision-making are AI planning (AIP) and reinforcement learning (RL). Each has strengths and weaknesses. AIP is interpretable, easy to integrate with symbolic knowledge, and often efficient, but requires an up-front logical domain specification and is sensitive to noise; RL only requires specification of rewards and is robust to noise but is sample inefficient and not easily supplied with external knowledge. We propose an integrative approach that combines high-level planning with RL, retaining interpretability, transfer, and efficiency, while allowing for robust learning of the lower-level plan actions. Our approach defines options in hierarchical reinforcement learning (HRL) from AIP operators by establishing a correspondence between the state transition model of AI planning problem and the abstract state transition system of a Markov Decision Process (MDP). Options are learned by adding intrinsic rewards to encourage consistency between the MDP and AIP transition models. We demonstrate the benefit of our integrated approach by comparing the performance of RL and HRL algorithms in both MiniGrid and N-rooms environments, showing the advantage of our method over the existing ones. | ['Tim Klinger', 'Geraud Nangue Tasse', 'Shirin Sohrabi', 'Miao Liu', 'Don Joven Agravante', 'Michael Katz', 'JunKyu Lee'] | 2022-03-01 | null | null | null | null | ['hierarchical-reinforcement-learning'] | ['methodology'] | [ 1.67591631e-01 6.43367589e-01 -2.69033343e-01 -2.45891765e-01
-6.16482258e-01 -5.35845339e-01 7.72085309e-01 3.64789099e-01
-2.92551458e-01 9.85948503e-01 3.77967030e-01 -4.20711190e-01
-6.33353949e-01 -9.56192493e-01 -6.66136861e-01 -4.59209085e-01
-4.69271332e-01 9.67812121e-01 5.39325953e-01 -3.49247873e-01
2.57929504e-01 5.80748379e-01 -1.50463748e+00 3.14513594e-01
9.33151305e-01 9.33716238e-01 1.83962554e-01 5.13596296e-01
-2.82843918e-01 1.42752695e+00 -3.36950034e-01 4.07926053e-01
3.97463143e-01 -4.63442534e-01 -1.42532146e+00 1.13893621e-01
-3.82568538e-01 -2.78521359e-01 -9.12485830e-03 6.59874260e-01
5.11824824e-02 2.39115819e-01 5.44005275e-01 -1.48323631e+00
-2.32851446e-01 8.90445411e-01 7.09317252e-02 -2.29994863e-01
6.88570917e-01 5.62807858e-01 9.41480935e-01 -6.03207462e-02
6.48592770e-01 1.48608780e+00 2.86668807e-01 4.29938167e-01
-1.56434488e+00 -1.98428109e-01 5.66853762e-01 1.89133346e-01
-1.16718209e+00 -2.24726826e-01 3.51839781e-01 -3.90367478e-01
1.15390265e+00 1.49491847e-01 8.12911093e-01 8.21900904e-01
2.45265484e-01 9.38368142e-01 1.40656328e+00 -5.90373516e-01
8.86985958e-01 -6.18360788e-02 -7.34489933e-02 7.78985202e-01
-2.09599525e-01 5.14834046e-01 -4.94444430e-01 -2.09664464e-01
9.88799095e-01 -9.61003602e-02 1.06687225e-01 -8.92984509e-01
-1.29018259e+00 7.29724586e-01 1.86654776e-01 2.07615122e-01
-4.30029690e-01 3.30766290e-01 2.67966807e-01 5.45542300e-01
-4.30086911e-01 8.83411825e-01 -4.27825868e-01 -2.41680399e-01
-7.23394752e-01 5.21187842e-01 1.10826087e+00 1.09881771e+00
8.85649323e-01 -4.55371030e-02 -4.47954595e-01 2.00933710e-01
4.69748437e-01 2.46515095e-01 3.26695025e-01 -1.60341620e+00
3.48915070e-01 6.67063296e-01 6.07182860e-01 -4.26258594e-01
-5.64043403e-01 7.10945353e-02 -3.00031126e-01 8.04989159e-01
4.82454628e-01 -1.53837547e-01 -7.96998739e-01 1.69254100e+00
1.30381703e-01 -4.20209803e-02 3.83074373e-01 4.37193215e-01
2.67736137e-01 8.20974231e-01 2.84287810e-01 -3.96035463e-01
9.26345229e-01 -1.02834892e+00 -5.50094426e-01 -3.39520335e-01
6.40620172e-01 -1.27833143e-01 1.13350320e+00 3.95755678e-01
-1.32233334e+00 -5.15850663e-01 -8.66115510e-01 2.42349967e-01
-1.29095525e-01 -2.59082407e-01 7.74069965e-01 3.33840072e-01
-1.16836059e+00 8.80466282e-01 -1.10793602e+00 -2.11388186e-01
1.51472390e-01 6.66794837e-01 -1.52766570e-01 2.65700649e-02
-1.09170985e+00 1.08109426e+00 7.23998606e-01 -3.04399163e-01
-1.28049433e+00 -2.05702394e-01 -9.85407948e-01 2.05143616e-01
1.05955720e+00 -6.35958314e-01 1.68759155e+00 -9.65188265e-01
-2.26346254e+00 5.15784137e-02 8.60197842e-02 -5.66361368e-01
5.62613070e-01 -1.12617202e-01 -1.94043547e-01 9.98165756e-02
8.97884294e-02 7.95188546e-01 6.04101241e-01 -1.30222750e+00
-8.09951186e-01 -1.15662090e-01 5.02340257e-01 5.24259210e-01
4.81451601e-01 -2.04168007e-01 -2.36362242e-03 4.25044214e-03
1.57330468e-01 -9.04458702e-01 -7.10658073e-01 -3.09609026e-01
-1.65207237e-01 -2.22955838e-01 2.94446290e-01 -2.16939867e-01
1.08773386e+00 -2.07416892e+00 3.14161927e-01 6.13153338e-01
-2.64811575e-01 -1.57168984e-01 -2.04347804e-01 8.51787806e-01
9.25848931e-02 4.85010706e-02 -2.88132191e-01 2.19576359e-01
3.29926848e-01 6.91896856e-01 -3.07091713e-01 -3.51823904e-02
2.05975041e-01 7.45704412e-01 -1.16916013e+00 -5.38320541e-01
3.20242256e-01 -2.65358448e-01 -6.66458607e-01 3.66227865e-01
-8.45562696e-01 6.10357225e-01 -7.64383793e-01 5.59070289e-01
-1.04768798e-01 -1.30299523e-01 6.23517096e-01 5.52451015e-01
-4.28920269e-01 6.73929393e-01 -1.83992910e+00 1.59324002e+00
-3.83651644e-01 -1.80378079e-01 -1.41812405e-02 -7.38518536e-01
7.77024925e-01 3.98529887e-01 3.21057558e-01 -6.49956286e-01
-2.55814165e-01 1.04508393e-01 7.31270537e-02 -5.09328246e-01
3.82390916e-01 -8.61621425e-02 -3.70130926e-01 6.41413331e-01
5.40710129e-02 -6.05470538e-01 2.76613653e-01 -1.71611346e-02
1.26673496e+00 6.59946620e-01 7.89387167e-01 -3.05295020e-01
4.82016504e-01 2.30595008e-01 7.07314134e-01 1.14636111e+00
2.89024860e-02 -1.05646089e-01 8.85125875e-01 -6.08409703e-01
-6.46468520e-01 -1.14357901e+00 2.24803180e-01 1.13586557e+00
1.89336792e-01 -4.24436569e-01 -5.18095374e-01 -6.96483850e-01
-1.73369035e-01 1.25151730e+00 -4.82737482e-01 1.08481944e-01
-7.33329654e-01 -2.59818286e-01 3.21748823e-01 6.19786680e-01
4.37558383e-01 -1.43211257e+00 -9.66872811e-01 4.73304003e-01
1.18574295e-02 -7.72394121e-01 6.09740727e-02 6.42382443e-01
-9.91282523e-01 -9.11331952e-01 4.66809683e-02 -4.20420438e-01
5.78415930e-01 -3.93510401e-01 1.20588481e+00 -5.20549305e-02
1.78753912e-01 5.71685016e-01 -2.78108865e-01 -2.36926705e-01
-7.95453250e-01 -1.26735661e-02 -1.49583295e-02 -4.46210593e-01
-3.23876947e-01 -4.86809045e-01 -2.42534224e-02 3.26324463e-01
-8.96562517e-01 2.46458411e-01 6.26555681e-01 8.38517249e-01
5.72158515e-01 3.66921127e-01 2.57752240e-01 -7.19537258e-01
6.76866293e-01 -8.22362602e-02 -9.32191730e-01 3.46815139e-01
-6.79592788e-01 8.27922583e-01 5.74607670e-01 -5.42275794e-02
-1.23605502e+00 3.41152519e-01 1.49697319e-01 2.12549001e-01
-7.08696485e-01 4.21387583e-01 -3.19552451e-01 1.30149305e-01
5.76723754e-01 5.57876304e-02 -5.49710058e-02 -5.67047298e-02
4.39631432e-01 6.18348084e-02 2.66769707e-01 -1.14627171e+00
5.38178146e-01 1.12483129e-01 1.40710250e-01 -4.19676244e-01
-4.99121577e-01 1.39011949e-01 -7.14022934e-01 -1.09756470e-01
6.13947392e-01 -3.98602039e-01 -1.02496874e+00 1.33134365e-01
-7.71742880e-01 -1.02344239e+00 -9.21791852e-01 3.28843206e-01
-1.19827044e+00 3.02071840e-01 -5.63673675e-01 -9.80876386e-01
2.38609865e-01 -1.47236252e+00 7.34111667e-01 5.89725561e-02
-4.97869283e-01 -7.76827276e-01 3.75438072e-02 -2.93537844e-02
2.70905882e-01 2.37591088e-01 1.34418392e+00 -6.45997345e-01
-9.44882512e-01 2.16062561e-01 6.54278696e-02 -3.80743295e-02
-3.45031954e-02 -2.96178814e-02 -5.72666347e-01 -8.21112096e-02
-2.08734915e-01 -5.36047697e-01 3.28060180e-01 3.53812456e-01
1.15685177e+00 -4.85012382e-01 -2.45314479e-01 -2.14057937e-02
1.34424293e+00 5.58472872e-01 6.78660810e-01 7.65638053e-01
7.08277747e-02 8.18911433e-01 9.52634215e-01 4.46883112e-01
5.20089686e-01 7.22037554e-01 5.51159501e-01 1.73207998e-01
3.18775803e-01 -2.44298041e-01 6.10280037e-01 8.13979879e-02
-4.59693700e-01 1.18507661e-01 -1.23759353e+00 2.69861758e-01
-2.30852914e+00 -1.11570930e+00 3.04445267e-01 2.33508182e+00
9.21941876e-01 4.10629869e-01 3.80446792e-01 1.44823611e-01
1.71788380e-01 -6.90338910e-02 -4.21401083e-01 -7.04606295e-01
4.71224844e-01 2.65313089e-01 3.56044590e-01 9.15694296e-01
-8.71309698e-01 9.83263433e-01 7.40923262e+00 5.23578107e-01
-5.55158257e-01 -3.01391423e-01 3.04569304e-01 1.07789874e-01
-2.69262493e-01 3.39785367e-01 -6.74890459e-01 4.35494632e-02
9.57572460e-01 -1.84389874e-02 8.68688226e-01 9.28457797e-01
2.58451045e-01 -4.56797272e-01 -1.68110299e+00 1.24601923e-01
-6.85715497e-01 -1.40961742e+00 -6.49753511e-02 -5.16826995e-02
4.86208528e-01 -2.57541686e-01 -2.54095554e-01 6.83022916e-01
1.10533953e+00 -1.22471130e+00 8.79157782e-01 6.21360242e-01
3.14719588e-01 -9.32397902e-01 6.58243358e-01 7.47896969e-01
-1.10527110e+00 -4.99567926e-01 8.68766308e-02 -5.60682356e-01
5.42125152e-03 -1.15495406e-01 -9.83123660e-01 6.87303364e-01
4.74156171e-01 3.31882179e-01 -2.61740893e-01 6.92931771e-01
-5.55020571e-01 2.56372154e-01 -3.40413719e-01 -9.81862247e-02
6.25174105e-01 -3.01218033e-01 4.05632913e-01 9.06789899e-01
2.45216805e-02 2.17327029e-01 8.42897832e-01 8.99442017e-01
6.49374008e-01 -1.78861991e-01 -5.02469838e-01 1.38850752e-02
4.48164016e-01 7.52222955e-01 -8.06144118e-01 -4.16190475e-01
-5.65500818e-02 4.71604645e-01 3.97589535e-01 2.67170519e-01
-5.74044645e-01 3.89292017e-02 3.55217695e-01 1.99157838e-02
2.07339391e-01 -3.87688726e-01 -1.70079693e-01 -7.44711399e-01
-3.24731052e-01 -1.22746158e+00 4.55723882e-01 -6.45791709e-01
-8.07663858e-01 4.65390265e-01 5.34821570e-01 -1.04284227e+00
-7.20161378e-01 -5.17440915e-01 -4.05037850e-01 7.26916254e-01
-1.38863266e+00 -8.36346090e-01 -2.69606300e-02 6.73759997e-01
4.21833575e-01 -1.10671595e-02 1.27142572e+00 -4.78276104e-01
-3.40118855e-01 -5.74113242e-02 -2.38502830e-01 -2.79629380e-01
1.58366412e-01 -1.54089248e+00 2.17709199e-01 5.15174031e-01
-2.96521574e-01 3.81579220e-01 6.32277548e-01 -6.24177158e-01
-1.35810339e+00 -8.66194725e-01 6.02746964e-01 -2.64096200e-01
5.67762017e-01 8.66907556e-03 -6.41604006e-01 9.11860824e-01
2.49364227e-01 -5.61196387e-01 4.92232323e-01 2.65673071e-01
6.93372339e-02 -1.04938922e-02 -1.14383805e+00 8.61333072e-01
8.12526643e-01 -1.32274881e-01 -1.05324543e+00 1.69144154e-01
6.65643632e-01 -6.21050537e-01 -7.50750780e-01 2.60195345e-01
3.10241491e-01 -1.12472785e+00 9.23653424e-01 -4.89692599e-01
3.34020287e-01 -5.79939067e-01 -6.73364252e-02 -1.19445169e+00
-6.82844639e-01 -8.04528296e-01 3.05961166e-02 8.95656109e-01
5.09169877e-01 -5.49187899e-01 5.16720057e-01 8.56415331e-01
-2.32758701e-01 -5.73028684e-01 -8.10227275e-01 -8.94309282e-01
-1.42177671e-01 -4.87589568e-01 7.99034715e-01 6.40619278e-01
3.23188990e-01 2.57812202e-01 -2.42383063e-01 3.08694214e-01
3.64227653e-01 3.08883727e-01 6.63593709e-01 -1.11360753e+00
-8.10202777e-01 -3.79998088e-01 1.20375969e-01 -8.03607941e-01
1.64274275e-01 -6.66500449e-01 3.96935016e-01 -1.79638898e+00
-4.13480103e-01 -8.01947474e-01 -1.69409543e-01 9.15449858e-01
4.53783602e-01 -6.77373290e-01 4.06836718e-01 1.64089426e-01
-8.24011743e-01 4.90203857e-01 1.32557559e+00 4.87831347e-02
-6.72506273e-01 2.21043184e-01 -4.37652558e-01 9.56412554e-01
8.69931161e-01 -4.44879413e-01 -4.47838277e-01 -1.18755654e-01
2.61241943e-01 6.53879225e-01 2.42121384e-01 -1.14487135e+00
2.34351069e-01 -9.09334064e-01 1.94255918e-01 -2.25528270e-01
2.55600154e-01 -8.93182635e-01 2.78621376e-01 7.98046350e-01
-7.61493683e-01 1.48022264e-01 1.98142499e-01 5.42288601e-01
-9.96904150e-02 -4.60912794e-01 6.01225555e-01 -5.32982051e-01
-1.01312602e+00 -1.47992089e-01 -8.35921466e-01 -1.10530943e-01
1.15170276e+00 -2.53783375e-01 2.14968249e-02 -3.11426073e-01
-1.02070642e+00 5.04032731e-01 4.77415979e-01 -2.20047397e-04
3.42811346e-01 -8.90359759e-01 -2.75907606e-01 2.87288398e-01
-1.03778817e-01 2.30202377e-01 -2.83526778e-01 7.03196526e-01
-3.59189540e-01 2.90391177e-01 -5.81431925e-01 -3.77332211e-01
-8.00414741e-01 6.47887468e-01 3.19020957e-01 -8.77091229e-01
-6.81750000e-01 3.67068857e-01 -1.62726730e-01 -6.96590364e-01
4.59526867e-01 -7.80190587e-01 -2.88064331e-01 -3.00676644e-01
3.86096120e-01 3.65772158e-01 -3.08564395e-01 1.21361159e-01
-2.66976774e-01 1.97652653e-01 1.03319533e-01 -4.79320943e-01
1.28780234e+00 5.70257660e-03 -1.86833411e-01 6.61956251e-01
-1.05506808e-01 -3.04122031e-01 -1.39710629e+00 -2.31501102e-01
4.17630494e-01 -1.61098465e-01 -1.68517411e-01 -9.92038250e-01
-2.16304347e-01 4.72846806e-01 2.25679487e-01 4.67310876e-01
9.41744447e-01 -7.94499964e-02 1.64689898e-01 8.39153051e-01
7.63123930e-01 -1.42892385e+00 2.17480376e-01 8.47236574e-01
9.47136521e-01 -8.14404666e-01 5.74445836e-02 -2.98570842e-01
-1.07603002e+00 1.20399880e+00 6.20464504e-01 -1.31137371e-01
1.61587715e-01 5.57915092e-01 -1.72392860e-01 -1.08011700e-01
-9.51698840e-01 -3.36428195e-01 -9.06112045e-02 7.63681829e-01
-7.90654346e-02 1.64111212e-01 9.34238210e-02 2.84099609e-01
-5.99282682e-02 2.23767102e-01 5.18796027e-01 1.46182537e+00
-6.64991379e-01 -1.50100505e+00 -5.05528390e-01 1.37724370e-01
4.47750092e-02 9.54371765e-02 -2.08035067e-01 1.06882024e+00
7.56027400e-02 8.27169716e-01 -6.65928572e-02 4.96127978e-02
3.70120108e-01 2.89121598e-01 6.50397182e-01 -8.76620471e-01
-6.75246894e-01 -3.99794662e-03 3.81143898e-01 -9.59994137e-01
-5.13712049e-01 -9.16888893e-01 -1.60606003e+00 1.50112947e-02
1.25189245e-01 3.15956324e-01 2.84405410e-01 1.35892332e+00
1.92577481e-01 7.35849321e-01 4.39382493e-01 -8.96148801e-01
-8.71873319e-01 -3.88333261e-01 -5.61145008e-01 -1.63579006e-02
1.41817763e-01 -7.08552301e-01 9.37327743e-02 -9.18342620e-02] | [4.190720558166504, 1.4619699716567993] |
1d93a4ca-ae50-4961-b233-fd5b8b956399 | towards-hate-speech-detection-in-low-resource | 2306.00410 | null | https://arxiv.org/abs/2306.00410v1 | https://arxiv.org/pdf/2306.00410v1.pdf | Towards hate speech detection in low-resource languages: Comparing ASR to acoustic word embeddings on Wolof and Swahili | We consider hate speech detection through keyword spotting on radio broadcasts. One approach is to build an automatic speech recognition (ASR) system for the target low-resource language. We compare this to using acoustic word embedding (AWE) models that map speech segments to a space where matching words have similar vectors. We specifically use a multilingual AWE model trained on labelled data from well-resourced languages to spot keywords in data in the unseen target language. In contrast to ASR, the AWE approach only requires a few keyword exemplars. In controlled experiments on Wolof and Swahili where training and test data are from the same domain, an ASR model trained on just five minutes of data outperforms the AWE approach. But in an in-the-wild test on Swahili radio broadcasts with actual hate speech keywords, the AWE model (using one minute of template data) is more robust, giving similar performance to an ASR system trained on 30 hours of labelled data. | ['Herman Kamper', 'Bruce A. Bassett', 'Everlyn Asiko Chimoto', 'Nathanaël Carraz Rakotonirina', 'Christiaan Jacobs'] | 2023-06-01 | null | null | null | null | ['word-embeddings', 'hate-speech-detection', 'keyword-spotting', 'automatic-speech-recognition'] | ['methodology', 'natural-language-processing', 'speech', 'speech'] | [ 1.69225127e-01 2.94158578e-01 2.20599949e-01 -2.33553007e-01
-1.33653641e+00 -5.74129283e-01 9.42444265e-01 -2.98909862e-02
-7.95189619e-01 1.94485769e-01 5.20520329e-01 -4.46111053e-01
2.82602221e-01 -2.55164921e-01 -5.03586233e-01 -5.06198049e-01
7.35635730e-03 3.38481396e-01 2.27955893e-01 -5.11430323e-01
2.97447219e-02 3.63283187e-01 -1.07669973e+00 2.34388039e-01
2.62682796e-01 3.36513221e-01 3.86574090e-01 1.15027809e+00
-9.86011550e-02 7.57653892e-01 -1.25492847e+00 -4.31251079e-01
4.25548144e-02 -1.29293784e-01 -7.60159254e-01 6.58673942e-02
5.00938654e-01 -1.53244257e-01 -8.64608884e-01 8.39444160e-01
7.86280692e-01 1.20241409e-02 2.25810900e-01 -9.07186627e-01
-9.70284522e-01 5.68912387e-01 -5.75860403e-02 5.44142187e-01
7.48790324e-01 -1.41538922e-02 7.86698103e-01 -1.08520687e+00
5.53663015e-01 1.33537316e+00 6.90910220e-01 6.80879176e-01
-1.18568480e+00 -6.95785761e-01 -1.83606222e-01 3.49111971e-03
-1.69140410e+00 -7.79377401e-01 5.60734749e-01 -2.37009913e-01
1.23182428e+00 4.19689000e-01 2.96511620e-01 1.52356315e+00
-3.30432087e-01 8.87526572e-01 8.75422835e-01 -5.75984597e-01
9.03869513e-03 6.11488640e-01 -4.59118746e-02 4.05855626e-01
-1.66762099e-01 -9.14076120e-02 -7.64446855e-01 -5.94098687e-01
7.14935884e-02 -1.87680274e-01 -4.01401818e-01 2.58433610e-01
-1.23353374e+00 1.20772529e+00 -7.55513161e-02 7.63870478e-01
-3.84549469e-01 -2.18814500e-02 4.17696029e-01 4.98861521e-01
9.21971321e-01 5.75917542e-01 -2.87088573e-01 -1.75548285e-01
-1.24815428e+00 1.86976418e-01 8.69058013e-01 6.00587070e-01
6.72796369e-01 4.57226127e-01 7.97447711e-02 1.47490811e+00
2.01451749e-01 1.10158968e+00 6.81629896e-01 -3.46128523e-01
4.79246736e-01 -2.22979888e-01 4.04195860e-02 -8.70521486e-01
-8.69929343e-02 -1.97224036e-01 5.51800504e-02 -2.95804679e-01
2.43920803e-01 -2.79482454e-01 -1.33812356e+00 1.34428895e+00
1.13265976e-01 2.22421274e-01 2.53621966e-01 6.16327167e-01
6.06228828e-01 1.26320565e+00 -9.98259783e-02 -1.62164554e-01
1.38671494e+00 -7.70627558e-01 -1.00853598e+00 -4.73187894e-01
9.07982886e-01 -1.20281708e+00 1.17407727e+00 3.86845887e-01
-5.77996016e-01 -1.51220456e-01 -1.04118168e+00 6.87641948e-02
-9.72322345e-01 -4.52084690e-01 -1.36450734e-02 1.08335888e+00
-1.16583467e+00 -1.29578814e-01 -3.01135063e-01 -8.06113899e-01
-1.72801360e-01 -4.32883576e-02 -5.97800851e-01 -2.06656277e-01
-1.68664408e+00 1.19419837e+00 1.43774733e-01 -1.59272343e-01
-1.28880310e+00 -6.62979722e-01 -1.21162820e+00 -4.84715760e-01
4.06439722e-01 5.63895583e-01 1.42214847e+00 -6.59725606e-01
-1.23618972e+00 1.06760144e+00 -1.30519912e-01 -6.53096139e-01
-9.76394117e-02 -3.84398758e-01 -1.14420927e+00 2.46964991e-01
5.39792031e-02 2.66562760e-01 1.21751606e+00 -1.18850946e+00
-3.64908576e-01 -1.50060952e-01 -2.11953893e-01 -2.16196626e-01
-4.63787436e-01 8.73174071e-01 -9.22493264e-03 -1.02443719e+00
-3.85525644e-01 -8.91928911e-01 2.26184279e-01 -6.22955561e-01
-2.99734592e-01 -2.83632576e-01 1.43378878e+00 -1.36864007e+00
1.48603117e+00 -2.38422680e+00 -2.63969541e-01 2.75413096e-01
-1.20816477e-01 8.85611534e-01 -2.38846228e-01 6.63318396e-01
-1.64100025e-02 2.74772048e-01 -3.19158703e-01 -1.96234539e-01
9.49114710e-02 4.60955262e-01 -5.88209510e-01 6.98517442e-01
3.76687109e-01 4.43702489e-01 -9.31834340e-01 -2.74979949e-01
1.03216760e-01 8.19995999e-01 -1.94577947e-01 4.96706456e-01
2.38848925e-01 -2.30693832e-01 2.42716670e-02 5.84488213e-01
2.63131648e-01 6.48566484e-01 -1.69352055e-01 4.02015179e-01
-1.33198529e-01 6.41790926e-01 -8.33961546e-01 1.34679329e+00
-8.14608037e-01 1.16058385e+00 3.90195251e-01 -6.63907707e-01
1.17842686e+00 9.07579839e-01 -8.07018131e-02 -8.44679296e-01
-3.51278812e-01 4.97662216e-01 -4.88670692e-02 -5.85599661e-01
6.25441611e-01 -1.05967961e-01 -4.07098562e-01 6.07346594e-01
2.69915670e-01 -5.27559042e-01 -1.39793098e-01 3.30473691e-01
1.56644380e+00 -5.48557878e-01 -2.04872489e-02 -1.04322836e-01
4.81109917e-01 -1.18589766e-01 1.42536044e-01 8.41594279e-01
-3.63340676e-01 5.89960754e-01 7.40596130e-02 -3.69485408e-01
-1.33042014e+00 -9.52946603e-01 -2.50027627e-01 1.40733802e+00
-5.61490417e-01 -6.65631652e-01 -8.58271360e-01 -6.83218658e-01
-2.64877200e-01 1.13314366e+00 -5.08258700e-01 1.08297639e-01
-7.95351923e-01 -3.35495174e-01 1.27545881e+00 -1.90762371e-01
-9.60050896e-02 -1.26196492e+00 -2.88828492e-01 4.04166639e-01
-2.30282312e-03 -1.21697915e+00 -6.77449882e-01 3.03225696e-01
2.29447156e-01 -3.91451418e-01 -1.23778999e+00 -9.08636749e-01
1.31241173e-01 2.47826397e-01 8.33702683e-01 7.74999931e-02
-4.50934857e-01 6.27567708e-01 -7.73663461e-01 -6.26011729e-01
-9.15347159e-01 -1.30648002e-01 2.72253871e-01 1.32479891e-01
7.96824515e-01 -1.14603229e-01 1.42004276e-02 1.71890154e-01
-1.04433608e+00 -6.29759252e-01 2.74148524e-01 7.66146898e-01
-1.16856702e-01 -3.32286894e-01 4.94801313e-01 -7.41574228e-01
7.37730920e-01 -6.71522021e-01 -1.03355646e-01 2.26005763e-01
-2.60488629e-01 -1.96094602e-01 3.62631291e-01 -8.10892701e-01
-6.34976029e-01 -2.75552779e-01 -4.05723155e-01 -3.89913052e-01
-3.48727733e-01 2.08081797e-01 1.52031884e-01 1.77104816e-01
6.66336060e-01 4.47095245e-01 -1.24034911e-01 -6.47917986e-01
4.08154488e-01 1.41006935e+00 4.44285691e-01 -1.73378736e-01
1.05661297e+00 -2.73759384e-02 -1.04924011e+00 -1.84417140e+00
-5.27181029e-01 -8.56716156e-01 -2.40857050e-01 -1.90063953e-01
9.85171854e-01 -8.06304932e-01 1.77492172e-01 3.83262306e-01
-1.45115101e+00 -2.28914008e-01 6.11019880e-02 5.83326936e-01
-2.32767552e-01 2.97435015e-01 -6.17063165e-01 -1.43556678e+00
-2.06477091e-01 -9.71985459e-01 1.30281532e+00 -4.40273762e-01
-5.94714761e-01 -9.99017596e-01 5.82424462e-01 2.71163613e-01
4.85117167e-01 -3.42653871e-01 7.62244165e-01 -1.53779221e+00
3.62008989e-01 -5.11607766e-01 9.51584280e-02 6.27699614e-01
1.35026842e-01 -1.16508447e-01 -1.34738517e+00 -3.15905690e-01
8.32409710e-02 -5.45588255e-01 6.39784753e-01 -2.60082781e-01
5.59725225e-01 -5.71619451e-01 9.89197660e-03 6.69550151e-02
9.89347577e-01 4.13063258e-01 6.93480909e-01 3.90396893e-01
5.68933606e-01 7.77469337e-01 3.95656377e-01 1.45718038e-01
3.09281480e-02 7.88164258e-01 -1.52276054e-01 -1.95172012e-01
-1.59851089e-01 -3.06844115e-01 9.09787774e-01 1.29799449e+00
6.18237436e-01 -3.67199093e-01 -1.31449211e+00 1.09592843e+00
-1.22298622e+00 -1.09542966e+00 2.48263091e-01 2.07197356e+00
8.37309897e-01 6.60922825e-02 3.61990124e-01 2.15100542e-01
8.91311944e-01 7.66005397e-01 2.41255656e-01 -1.01886892e+00
-1.88341871e-01 4.59978640e-01 4.91052687e-01 9.47394371e-01
-9.88905370e-01 1.09441674e+00 6.75462103e+00 9.90618467e-01
-1.08843565e+00 5.53541064e-01 2.61128783e-01 -2.46814489e-02
-3.71029168e-01 -2.58505881e-01 -6.21181905e-01 5.07707357e-01
1.77986705e+00 3.87650654e-02 7.06736147e-01 6.64200425e-01
2.14579254e-01 2.16261640e-01 -7.91947722e-01 1.06540990e+00
6.04613066e-01 -1.00122404e+00 -2.11555570e-01 1.15935951e-01
1.76493421e-01 4.66766953e-01 2.01798201e-01 4.60111827e-01
3.17796201e-01 -1.21191800e+00 7.29271054e-01 -3.61564048e-02
6.33443594e-01 -8.26473653e-01 7.74673164e-01 3.57005626e-01
-8.38332713e-01 1.60301343e-01 -3.33077967e-01 3.55978608e-01
2.20669195e-01 1.06927142e-01 -1.47595298e+00 -5.68215065e-02
7.70906329e-01 1.36381313e-01 -6.20624542e-01 5.67370415e-01
-1.71245545e-01 1.17488134e+00 -3.19673389e-01 -2.77620584e-01
7.22036600e-01 2.27526188e-01 8.41397226e-01 2.06374002e+00
2.59934038e-01 -2.74675637e-01 -4.16170210e-02 2.46542558e-01
1.41815186e-01 3.62904549e-01 -1.22711647e+00 -6.18269265e-01
6.37250602e-01 1.05112493e+00 -1.96426064e-01 -3.18524152e-01
-5.41350007e-01 1.03351200e+00 1.84193552e-01 3.83714259e-01
-6.56247199e-01 -8.28487337e-01 7.10039020e-01 1.17070928e-01
3.76033306e-01 -4.20736581e-01 4.80769753e-01 -6.15892947e-01
-2.15336651e-01 -1.32885730e+00 8.76202211e-02 -8.82153213e-01
-1.14089429e+00 9.54255998e-01 -1.05913386e-01 -7.05957115e-01
-4.88967597e-01 -6.38577521e-01 -6.19998991e-01 9.27565038e-01
-1.10863030e+00 -1.00127244e+00 5.65987706e-01 5.99512935e-01
1.00437534e+00 -4.64572072e-01 1.19041133e+00 3.17498356e-01
-4.62025464e-01 6.56523943e-01 8.51757154e-02 5.50131440e-01
7.00117707e-01 -1.16320610e+00 7.70678461e-01 9.11362231e-01
8.87464941e-01 6.76280141e-01 1.04910684e+00 -6.16676688e-01
-1.19937372e+00 -8.72934759e-01 1.41747582e+00 -7.15062857e-01
1.32837093e+00 -1.00236177e+00 -1.36200798e+00 5.62511325e-01
6.75630331e-01 -2.20198244e-01 9.33986545e-01 1.10552378e-01
-7.69359291e-01 2.68626988e-01 -9.84331548e-01 3.92782182e-01
5.97086489e-01 -1.46295083e+00 -1.08679640e+00 7.66333342e-01
1.12435985e+00 2.46728316e-01 -7.80092537e-01 -1.44712791e-01
4.83721375e-01 -2.75857478e-01 7.11928964e-01 -7.56274223e-01
-2.37600327e-01 -1.95098072e-01 -6.13787234e-01 -1.53084886e+00
3.53930593e-02 -1.10809827e+00 2.89999276e-01 1.32168901e+00
7.39997566e-01 -4.38018292e-01 2.20632255e-01 2.16443211e-01
8.88589304e-03 -2.59164393e-01 -1.13492942e+00 -1.06024039e+00
-1.23547040e-01 -8.82396996e-01 2.90835828e-01 1.02690113e+00
-3.37684825e-02 4.49157089e-01 -7.03638196e-01 3.32358778e-01
4.06269282e-01 -8.93603384e-01 6.91509008e-01 -6.48565829e-01
-2.97907554e-02 1.29167020e-01 -6.50963843e-01 -4.97012287e-01
4.55054760e-01 -7.99547791e-01 4.02409345e-01 -1.03848088e+00
-2.36913681e-01 -1.18593641e-01 -1.35460407e-01 4.88996506e-01
5.86101152e-02 4.82022077e-01 2.56475776e-01 -8.58447924e-02
-4.07003224e-01 2.18526140e-01 2.16831103e-01 -3.45293403e-01
8.92086700e-02 -3.02316725e-01 -1.05887882e-01 5.68010271e-01
6.53436542e-01 -9.19440866e-01 -2.02218026e-01 -4.94257629e-01
-1.96393877e-01 -2.15605289e-01 1.94906607e-01 -6.53901458e-01
7.41071776e-02 1.09968549e-02 -1.18006080e-01 -3.51485610e-01
6.41908228e-01 -6.22554064e-01 -1.80277199e-01 3.63277435e-01
-4.40678746e-01 2.41350830e-01 1.58447593e-01 4.74178612e-01
-2.61027128e-01 -5.11223376e-01 7.34897733e-01 1.30429519e-02
-7.77321875e-01 -6.20294958e-02 -1.11792827e+00 1.63228333e-01
6.62194490e-01 -2.05056012e-01 -3.77822429e-01 -8.11816335e-01
-5.64199567e-01 -2.90982306e-01 2.88030118e-01 6.70166016e-01
8.23043883e-01 -1.14302301e+00 -9.58051026e-01 4.80864227e-01
4.70108628e-01 -7.24426150e-01 -3.02887797e-01 4.62844253e-01
-4.41130221e-01 6.12999678e-01 3.77181560e-01 -3.11198354e-01
-1.35581386e+00 6.96098924e-01 1.23166010e-01 1.39178887e-01
-3.77406538e-01 9.37370718e-01 7.39895105e-02 -6.64200842e-01
4.92587984e-01 5.04392862e-01 2.64932197e-02 5.62901683e-02
9.57075715e-01 1.23711064e-01 2.43346751e-01 -1.23853564e+00
-6.74262881e-01 3.62099968e-02 -3.09857816e-01 -7.56522059e-01
1.20164704e+00 2.86104791e-02 1.64006799e-01 8.65553737e-01
1.54326653e+00 6.80156112e-01 -4.80149776e-01 -4.02179420e-01
4.38727826e-01 -4.36212391e-01 2.12713748e-01 -4.41819876e-01
-2.72500515e-01 7.58421838e-01 6.17256165e-01 7.52317488e-01
4.21147406e-01 3.01010460e-01 9.69105184e-01 5.96129477e-01
2.51657665e-01 -1.51271141e+00 5.57883233e-02 6.57636940e-01
1.06556404e+00 -1.01834905e+00 -3.97247106e-01 1.92524977e-02
-7.36418962e-01 9.91200387e-01 1.15682045e-02 -6.44007847e-02
7.59184718e-01 2.31279612e-01 6.59517705e-01 -3.72071892e-01
-5.86280227e-01 -3.48249584e-01 1.27113968e-01 8.32331896e-01
3.27276736e-01 1.01973243e-01 4.08700891e-02 -6.67668656e-02
-4.59531575e-01 -9.47198510e-01 5.02599597e-01 1.01797760e+00
-6.33206725e-01 -8.71629596e-01 -6.80652261e-01 1.38371825e-01
-8.50268006e-01 -3.60250324e-01 -8.78099740e-01 9.31441605e-01
-2.33993828e-01 1.40636384e+00 7.11721256e-02 -6.34479821e-01
2.59467959e-01 5.11932313e-01 -2.15159152e-02 -1.05068827e+00
-6.34719968e-01 2.09219381e-01 5.75232923e-01 -3.67555916e-01
-2.37113178e-01 -4.55920070e-01 -6.78629339e-01 -1.53464913e-01
-3.99420798e-01 5.03727674e-01 1.12281871e+00 9.19210732e-01
-8.93166438e-02 4.47534695e-02 1.00201178e+00 -5.16750455e-01
-3.78752500e-01 -1.31432831e+00 -7.65877664e-01 4.19750392e-01
6.35819733e-01 -2.76983380e-01 -6.32228315e-01 -1.29971966e-01] | [14.317459106445312, 6.841878890991211] |
89d76270-8f35-4b78-b891-72ee52692f1c | attend-and-interact-higher-order-object | 1711.06330 | null | http://arxiv.org/abs/1711.06330v2 | http://arxiv.org/pdf/1711.06330v2.pdf | Attend and Interact: Higher-Order Object Interactions for Video Understanding | Human actions often involve complex interactions across several inter-related
objects in the scene. However, existing approaches to fine-grained video
understanding or visual relationship detection often rely on single object
representation or pairwise object relationships. Furthermore, learning
interactions across multiple objects in hundreds of frames for video is
computationally infeasible and performance may suffer since a large
combinatorial space has to be modeled. In this paper, we propose to efficiently
learn higher-order interactions between arbitrary subgroups of objects for
fine-grained video understanding. We demonstrate that modeling object
interactions significantly improves accuracy for both action recognition and
video captioning, while saving more than 3-times the computation over
traditional pairwise relationships. The proposed method is validated on two
large-scale datasets: Kinetics and ActivityNet Captions. Our SINet and
SINet-Caption achieve state-of-the-art performances on both datasets even
though the videos are sampled at a maximum of 1 FPS. To the best of our
knowledge, this is the first work modeling object interactions on open domain
large-scale video datasets, and we additionally model higher-order object
interactions which improves the performance with low computational costs. | ['Chih-Yao Ma', 'Zsolt Kira', 'Hans Peter Graf', 'Ghassan AlRegib', 'Iain Melvin', 'Asim Kadav'] | 2017-11-16 | attend-and-interact-higher-order-object-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Ma_Attend_and_Interact_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Ma_Attend_and_Interact_CVPR_2018_paper.pdf | cvpr-2018-6 | ['video-description', 'visual-relationship-detection'] | ['computer-vision', 'computer-vision'] | [ 3.10761362e-01 -1.35249481e-01 -3.15136552e-01 -3.63301933e-01
-7.59708166e-01 -6.68543518e-01 5.19433796e-01 2.38080040e-01
-3.14664334e-01 7.83173025e-01 2.25742728e-01 9.66720656e-02
-2.57118195e-01 -1.75625622e-01 -1.31095254e+00 -4.19009358e-01
-3.69632185e-01 7.32542098e-01 6.27798259e-01 1.82340607e-01
1.78573653e-01 2.89870948e-01 -1.82229745e+00 6.40123069e-01
4.82606232e-01 9.77885962e-01 1.34643584e-01 8.86899233e-01
7.87176117e-02 1.08959091e+00 -3.72940868e-01 -2.78391033e-01
2.08618984e-01 -2.23788485e-01 -8.62215459e-01 3.69494468e-01
1.06798410e+00 -5.24262428e-01 -4.61208940e-01 7.28785276e-01
1.80341285e-02 1.91405267e-01 4.15807426e-01 -1.66430855e+00
-5.61601579e-01 4.49763745e-01 -6.09720945e-01 4.23324227e-01
6.76427662e-01 2.50651926e-01 1.12467849e+00 -6.73880816e-01
6.64999068e-01 1.38982916e+00 2.54251987e-01 4.96527284e-01
-1.13210773e+00 -6.16353750e-01 6.02097809e-01 6.16854370e-01
-1.32723248e+00 -3.66967380e-01 3.49780977e-01 -6.42295241e-01
1.15076458e+00 3.75484645e-01 7.69188166e-01 1.18501806e+00
-2.42650852e-01 7.86896288e-01 7.85873353e-01 7.41924345e-02
2.42111143e-02 -3.94429088e-01 1.59436315e-01 7.07705379e-01
3.71190459e-01 -3.00546378e-01 -8.56368482e-01 -1.95756823e-01
1.10304725e+00 5.18675447e-02 -2.66373038e-01 -6.09101415e-01
-1.79929066e+00 4.91981715e-01 1.55518308e-01 1.69318825e-01
-2.21676052e-01 8.02012384e-01 4.91609216e-01 -4.85258847e-02
9.64671001e-02 2.77703196e-01 -4.56821918e-01 -6.86508358e-01
-4.05551940e-01 3.67278188e-01 7.97924936e-01 1.14397860e+00
5.61374903e-01 -4.64383572e-01 -2.41597325e-01 5.04215956e-01
1.16051316e-01 1.47246882e-01 5.15566617e-02 -1.23316848e+00
7.65668750e-01 5.43208599e-01 4.18601632e-01 -1.11125302e+00
-1.04309984e-01 2.17021957e-01 -5.69906712e-01 -3.51248719e-02
7.67722011e-01 3.23672652e-01 -7.18351007e-01 1.78161907e+00
3.30312759e-01 6.66316628e-01 -2.01409161e-01 1.18837702e+00
8.19759309e-01 6.60369873e-01 2.99214214e-01 -1.96195230e-01
1.46879256e+00 -1.37640274e+00 -7.61428773e-01 -3.00183952e-01
6.58681095e-01 -4.52368379e-01 9.00144339e-01 1.85019746e-01
-1.08391714e+00 -6.43871844e-01 -7.18285263e-01 -2.38513753e-01
-1.89743072e-01 1.24363109e-01 9.02634084e-01 1.65097043e-01
-8.21495950e-01 4.13577259e-01 -9.80688930e-01 -5.18374860e-01
7.66747594e-01 5.44280410e-01 -7.58615613e-01 -6.16702065e-02
-8.84356201e-01 6.29294693e-01 3.87166381e-01 -1.42693192e-01
-1.11534965e+00 -8.89013469e-01 -8.50525081e-01 7.89540783e-02
7.03801751e-01 -7.69628227e-01 1.21132874e+00 -9.72827435e-01
-1.03726768e+00 7.36498773e-01 -1.86513901e-01 -5.05665064e-01
5.83380222e-01 -5.73952258e-01 2.22506374e-02 3.97883594e-01
5.03702555e-03 1.01538718e+00 5.18391907e-01 -1.12654078e+00
-4.37260479e-01 -3.49180013e-01 5.99148035e-01 3.86094451e-01
-1.45116910e-01 2.83690661e-01 -8.68465006e-01 -7.68150449e-01
-2.03929931e-01 -1.08943760e+00 -1.23232342e-01 5.11082709e-01
-2.91333478e-02 -3.57798785e-01 9.42974031e-01 -4.52702492e-01
8.53491127e-01 -1.95813179e+00 3.99599850e-01 -4.12361264e-01
3.79554719e-01 1.28900543e-01 -2.33671695e-01 3.18166465e-01
-1.55585244e-01 1.21736646e-01 3.04145124e-02 -1.82093382e-01
5.23871481e-02 4.42720741e-01 -7.42839500e-02 4.66596365e-01
1.51017994e-01 1.16781938e+00 -9.86216366e-01 -6.89005136e-01
3.68904054e-01 4.68904316e-01 -6.33067012e-01 3.64280820e-01
-5.79473734e-01 5.50830424e-01 -5.60536742e-01 5.42003810e-01
4.35250521e-01 -8.03495347e-01 2.29474962e-01 -4.84078526e-01
1.51995763e-01 -1.00756496e-01 -1.08381963e+00 1.83552253e+00
7.11348578e-02 7.17778385e-01 -3.43106240e-01 -1.07522917e+00
4.59879816e-01 3.19739550e-01 7.87075698e-01 -3.16175729e-01
-2.34144572e-02 -1.33168280e-01 -2.57672705e-02 -5.94435990e-01
1.01744451e-01 2.53492177e-01 -5.66157848e-02 3.35444540e-01
-9.99081582e-02 2.44847804e-01 5.84775329e-01 3.05016428e-01
1.13789606e+00 3.81706983e-01 4.90087450e-01 6.73808083e-02
4.01362747e-01 3.30885574e-02 3.97189498e-01 7.54746556e-01
-3.09740037e-01 4.25854653e-01 6.26230359e-01 -6.78768754e-01
-9.96980011e-01 -1.00653446e+00 3.27507913e-01 1.25874317e+00
6.12871587e-01 -6.84698403e-01 -7.53702402e-01 -6.77736223e-01
-5.95543087e-02 1.19240992e-01 -7.33299136e-01 2.00274155e-01
-7.79876828e-01 -3.63813579e-01 4.11562264e-01 9.05783415e-01
4.18621600e-01 -9.70802486e-01 -6.32966161e-01 2.05456764e-02
-4.39886808e-01 -1.85623157e+00 -6.96121633e-01 -3.68025273e-01
-7.65252411e-01 -1.49184275e+00 -4.84740943e-01 -5.91254652e-01
7.18521774e-01 4.44236755e-01 1.18605328e+00 7.73088858e-02
-5.14554858e-01 5.46612024e-01 -4.26869988e-01 5.78529574e-03
5.18443324e-02 -4.99479353e-01 1.97101057e-01 1.22302324e-01
4.12537754e-01 -4.22159374e-01 -5.63205481e-01 6.64464653e-01
-6.85361743e-01 3.75179470e-01 4.51387972e-01 6.00633323e-01
7.05713391e-01 -2.20965445e-01 1.98029384e-01 -5.12131512e-01
-4.39078771e-02 -2.24855915e-01 -7.10265696e-01 3.52022767e-01
2.89452583e-01 1.04918834e-02 4.44521517e-01 -6.98611140e-01
-9.38854039e-01 3.43263984e-01 5.35672009e-01 -8.75102282e-01
-4.32724506e-01 3.85775790e-02 -7.37629384e-02 -2.35416248e-01
2.80715197e-01 9.06169564e-02 -3.13919544e-01 -3.25406253e-01
4.60169733e-01 1.91717327e-01 4.72019315e-01 -7.17679739e-01
6.77067161e-01 7.42562115e-01 1.06818751e-02 -6.63527787e-01
-9.77722108e-01 -6.51245236e-01 -8.22853684e-01 -3.46655667e-01
1.28915513e+00 -1.09358704e+00 -1.26222193e+00 3.75187010e-01
-1.29020393e+00 -2.34185606e-01 4.10603993e-02 6.21404767e-01
-9.50201452e-01 6.14784896e-01 -4.17796224e-01 -5.37357926e-01
1.74697235e-01 -1.14877141e+00 1.44190681e+00 7.78864417e-03
-2.93295711e-01 -6.37417316e-01 -1.03313491e-01 9.76609588e-01
-1.62532777e-01 3.64818007e-01 4.51107025e-01 -3.82249862e-01
-1.34225309e+00 1.28265619e-01 -4.82680947e-01 -8.37914050e-02
1.13278344e-01 -5.03812432e-02 -4.97575641e-01 -1.70598075e-01
-5.41332185e-01 -5.82702756e-01 7.09854305e-01 3.17718118e-01
1.55677140e+00 -2.55088598e-01 -5.43036938e-01 2.69123048e-01
1.03818655e+00 1.64576516e-01 5.83349526e-01 1.62053734e-01
1.11026144e+00 6.31270170e-01 1.07692552e+00 6.90468073e-01
3.45388949e-01 1.03924799e+00 5.83714426e-01 9.34687778e-02
-9.32648107e-02 -1.88851118e-01 3.41234714e-01 3.84707332e-01
-3.73530149e-01 -4.80287790e-01 -8.27205002e-01 5.93952179e-01
-2.39315271e+00 -1.21062052e+00 -1.96267724e-01 1.89999580e+00
5.97067654e-01 2.52418797e-02 2.05952078e-01 -2.69952536e-01
7.57667542e-01 1.90764263e-01 -7.36533880e-01 9.07861963e-02
1.54154589e-02 -3.17207456e-01 4.76156771e-01 2.65029877e-01
-1.27022135e+00 1.04043174e+00 6.36337566e+00 6.75064743e-01
-5.12249589e-01 -5.06486967e-02 5.66966414e-01 -4.18361664e-01
2.74941564e-01 -8.61762837e-02 -7.45820761e-01 3.05558920e-01
6.87397718e-01 6.44562170e-02 4.23587531e-01 7.39831209e-01
3.17096889e-01 -1.59598306e-01 -1.55625832e+00 1.52388418e+00
2.35824317e-01 -1.52944088e+00 3.56717229e-01 1.74840149e-02
6.90701783e-01 -1.40868500e-01 -3.01829726e-01 -8.02881196e-02
1.82723641e-01 -1.18152952e+00 6.96683764e-01 5.89111447e-01
6.09200835e-01 -3.95515442e-01 3.73679608e-01 3.16355467e-01
-1.55625808e+00 8.97223875e-03 -1.03277057e-01 -3.16134810e-01
4.21654254e-01 2.08323635e-02 -5.25036573e-01 2.52345115e-01
9.61938679e-01 1.09985435e+00 -4.75997150e-01 9.93366301e-01
1.45990789e-01 3.14438462e-01 -3.57958674e-01 -4.09119837e-02
2.32949376e-01 -5.34945801e-02 4.44718599e-01 9.65444863e-01
1.08843468e-01 5.45525074e-01 5.28309822e-01 5.22099316e-01
-1.65253907e-01 -2.15354383e-01 -4.21945512e-01 -5.05142152e-01
1.36495307e-01 9.14399147e-01 -6.26845121e-01 -6.05958641e-01
-6.40489817e-01 9.32589293e-01 5.53044677e-01 1.93113595e-01
-1.42486835e+00 2.73082703e-01 1.14326131e+00 2.34440237e-01
5.48034847e-01 -4.20889616e-01 2.22191319e-01 -1.34662223e+00
3.34345758e-01 -9.14068520e-01 5.13336897e-01 -9.87693131e-01
-1.14647770e+00 2.87796021e-01 3.69310558e-01 -1.22542870e+00
-1.54744154e-02 -6.50195360e-01 2.97813979e-03 1.50825799e-01
-1.08599544e+00 -1.18476915e+00 -7.45988250e-01 6.96949899e-01
9.24831986e-01 2.56311506e-01 6.73758507e-01 3.23009104e-01
-4.13201988e-01 2.11854473e-01 -2.13238150e-01 8.15788880e-02
6.48361206e-01 -1.03507352e+00 4.05329466e-01 4.27997470e-01
3.20283264e-01 3.72953951e-01 7.25829422e-01 -6.14436090e-01
-1.52471709e+00 -1.07708442e+00 6.06370747e-01 -8.06909263e-01
7.90192783e-01 -6.62914753e-01 -8.58004332e-01 9.23825443e-01
-7.11630881e-02 4.63126600e-01 7.21048772e-01 -7.87706971e-02
-6.44008577e-01 1.29170477e-01 -7.68618762e-01 6.55862570e-01
1.71042347e+00 -4.68708515e-01 -6.15119517e-01 7.54143536e-01
7.77122259e-01 -4.87903833e-01 -1.01552534e+00 4.49925482e-01
8.11373651e-01 -8.46444607e-01 1.23451710e+00 -1.02579904e+00
3.24004114e-01 -7.04149604e-01 -2.67921299e-01 -6.60961390e-01
-2.16585726e-01 -6.24888480e-01 -5.58170497e-01 8.41495693e-01
-3.52469049e-02 -1.53650254e-01 9.13092136e-01 6.91511452e-01
2.06756547e-01 -6.14723086e-01 -8.18315208e-01 -9.07873869e-01
-6.32882476e-01 -3.57251048e-01 2.75417656e-01 7.32190967e-01
7.17548057e-02 1.91753522e-01 -6.67498767e-01 3.85347933e-01
8.67612422e-01 3.03516865e-01 1.13466406e+00 -9.80517209e-01
-5.75484812e-01 -5.33270165e-02 -8.61760318e-01 -1.45051396e+00
2.68783808e-01 -2.74805188e-01 -5.15837334e-02 -1.48015022e+00
7.37444222e-01 -9.55074579e-02 -1.19575933e-01 5.15342653e-01
-7.00579137e-02 5.86152136e-01 3.64656091e-01 1.89084470e-01
-1.38582253e+00 4.77671057e-01 1.40853930e+00 -2.05381423e-01
1.82012290e-01 -4.40380931e-01 -4.30110812e-01 8.96368444e-01
5.99953055e-01 -3.97469103e-01 -5.16213477e-01 -5.37040353e-01
-8.86480603e-03 2.00496480e-01 7.69406617e-01 -9.50461149e-01
1.12258963e-01 -4.90536302e-01 2.38076761e-01 -6.58186972e-01
8.57521713e-01 -8.76856565e-01 5.28539419e-01 2.55235910e-01
-3.63406658e-01 -2.18232825e-01 3.13871413e-01 9.53987062e-01
-2.48747855e-01 3.13978374e-01 6.02331758e-01 -2.38894835e-01
-9.79347885e-01 6.01051390e-01 -2.93396860e-01 4.57788259e-02
1.66200447e+00 -4.08818841e-01 -4.39111948e-01 -4.00696963e-01
-8.67894888e-01 3.55873913e-01 4.34447318e-01 7.76280999e-01
4.96418804e-01 -1.30323601e+00 -5.51337600e-01 -2.89308280e-01
4.33560312e-01 -1.43127680e-01 5.55654466e-01 8.33368361e-01
-4.93322611e-01 7.27104843e-01 -4.20178890e-01 -9.00754869e-01
-1.87090647e+00 7.41491497e-01 9.16548669e-02 -9.97235999e-02
-4.62243795e-01 9.01104152e-01 8.99257421e-01 4.74293381e-02
2.15554327e-01 -5.09245932e-01 -2.19419047e-01 -1.14592955e-01
6.17072523e-01 3.86981606e-01 -5.48885286e-01 -1.06757534e+00
-5.99748731e-01 1.00600672e+00 -2.50269294e-01 2.75884539e-01
1.16869175e+00 -1.18327461e-01 -6.95126131e-02 3.45093429e-01
1.16619229e+00 -5.29793859e-01 -1.77554929e+00 -1.35974614e-02
-1.53913766e-01 -9.52810884e-01 -3.77795786e-01 -5.54307878e-01
-6.83183372e-01 6.79102421e-01 3.66794854e-01 -2.02002481e-01
8.99792492e-01 6.15361214e-01 6.90080464e-01 6.53754056e-01
5.09649396e-01 -6.76086366e-01 6.14692450e-01 2.35954732e-01
9.68804181e-01 -1.44285595e+00 -5.29007800e-02 -9.38288510e-01
-7.02011347e-01 8.56813252e-01 1.05953562e+00 -7.10136211e-03
2.56141782e-01 5.94631359e-02 -2.85825431e-01 -3.47768217e-01
-1.00780523e+00 -7.99350590e-02 3.39578331e-01 4.52008605e-01
2.07031503e-01 -9.43520367e-02 -7.94492736e-02 2.40999863e-01
3.21154416e-01 1.25179991e-01 3.99050683e-01 8.34412158e-01
-2.88848281e-01 -9.82107043e-01 -3.18417102e-01 3.76092345e-01
-4.32334781e-01 7.32146278e-02 -4.62661356e-01 7.46916771e-01
5.46299852e-02 8.66317034e-01 3.11352849e-01 -2.45290268e-02
2.23564640e-01 -2.77455360e-01 8.72117162e-01 -4.76310343e-01
-1.94530338e-01 -4.34143767e-02 2.52745152e-01 -1.14032245e+00
-1.03130209e+00 -8.53373885e-01 -1.36690104e+00 -1.55790746e-01
-1.12677321e-01 -4.32447083e-02 2.38867506e-01 9.95937169e-01
5.21430314e-01 3.16343486e-01 -1.17714396e-02 -9.60412800e-01
-3.63011248e-02 -7.45468676e-01 -2.91047335e-01 7.29773045e-01
3.12317640e-01 -1.08236408e+00 -1.36092037e-01 6.28689528e-01] | [8.9428129196167, 0.4784051477909088] |
711abb67-26d2-4b46-95ef-89b5414238bb | relational-context-learning-for-human-object | 2304.04997 | null | https://arxiv.org/abs/2304.04997v1 | https://arxiv.org/pdf/2304.04997v1.pdf | Relational Context Learning for Human-Object Interaction Detection | Recent state-of-the-art methods for HOI detection typically build on transformer architectures with two decoder branches, one for human-object pair detection and the other for interaction classification. Such disentangled transformers, however, may suffer from insufficient context exchange between the branches and lead to a lack of context information for relational reasoning, which is critical in discovering HOI instances. In this work, we propose the multiplex relation network (MUREN) that performs rich context exchange between three decoder branches using unary, pairwise, and ternary relations of human, object, and interaction tokens. The proposed method learns comprehensive relational contexts for discovering HOI instances, achieving state-of-the-art performance on two standard benchmarks for HOI detection, HICO-DET and V-COCO. | ['Minsu Cho', 'Deunsol Jung', 'Sanghyun Kim'] | 2023-04-11 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Kim_Relational_Context_Learning_for_Human-Object_Interaction_Detection_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Kim_Relational_Context_Learning_for_Human-Object_Interaction_Detection_CVPR_2023_paper.pdf | cvpr-2023-1 | ['human-object-interaction-detection', 'relational-reasoning'] | ['computer-vision', 'natural-language-processing'] | [ 1.50648594e-01 5.61686456e-01 -4.15239781e-01 -3.50398198e-02
-4.57050920e-01 -3.06475937e-01 8.10863733e-01 2.72269309e-01
1.89017132e-01 5.65484881e-01 3.14159364e-01 -2.14973062e-01
-5.20105243e-01 -1.03071463e+00 -6.43499970e-01 -4.35291469e-01
-1.57622814e-01 9.01007771e-01 4.22256559e-01 -5.10866791e-02
-3.37282956e-01 5.13151810e-02 -1.69067574e+00 7.38503098e-01
5.13627172e-01 1.08659267e+00 -7.22790778e-01 5.56880057e-01
3.56096834e-01 1.42390025e+00 -6.21106327e-01 -9.16431725e-01
5.75679094e-02 -4.37494785e-01 -9.97187555e-01 -5.09912789e-01
3.66380066e-01 -3.32825631e-01 -9.33695734e-01 7.65827954e-01
2.68559158e-01 -1.87893108e-01 6.74802125e-01 -1.77211714e+00
-5.34429014e-01 1.41645169e+00 -3.52103651e-01 1.57638669e-01
5.79479396e-01 1.47756711e-01 1.78811765e+00 -9.74832892e-01
8.72392178e-01 1.31336963e+00 4.71783519e-01 2.29109183e-01
-1.22690237e+00 -8.68885934e-01 -2.64709175e-01 7.60995388e-01
-1.47443438e+00 -2.50987202e-01 7.56842196e-01 -4.04166877e-01
1.47104430e+00 4.81570601e-01 8.99849355e-01 1.54560769e+00
-1.31527260e-01 1.09535539e+00 7.76130259e-01 -2.68899471e-01
-1.29107475e-01 1.00907378e-01 4.78413135e-01 9.10294652e-01
8.59747350e-01 3.54557931e-01 -1.25124395e+00 -1.17037505e-01
4.37014014e-01 -1.14376955e-01 1.58414990e-02 -3.82604778e-01
-1.44219840e+00 5.50295830e-01 5.47787726e-01 2.63154417e-01
8.82984474e-02 -3.95464152e-02 5.79039216e-01 1.39821321e-01
-1.63142920e-01 4.46513146e-01 1.45074040e-01 5.84357418e-02
-6.08807623e-01 5.24313509e-01 9.61677313e-01 1.33603716e+00
4.44706559e-01 -6.82131886e-01 -5.78655243e-01 2.90997118e-01
2.97726691e-01 4.05053087e-02 -6.51981402e-03 -6.24972522e-01
9.99964595e-01 1.24318957e+00 -4.37776208e-01 -9.56681728e-01
-3.71335119e-01 -5.42061150e-01 -8.34119499e-01 -2.70695567e-01
4.01416749e-01 3.49806935e-01 -4.66379791e-01 1.41806400e+00
4.42556262e-01 1.38987526e-01 5.69195487e-02 7.72023737e-01
1.27338994e+00 2.17922013e-02 -2.36235276e-01 2.87858874e-01
1.52173424e+00 -1.06787241e+00 -6.76494956e-01 -1.72838092e-01
7.86952972e-01 -3.16961288e-01 6.52190804e-01 3.78261477e-01
-1.24223101e+00 5.01818582e-02 -1.31485975e+00 -5.60809612e-01
-4.29753095e-01 3.96750420e-02 9.32135999e-01 3.04583818e-01
-3.53820860e-01 2.49905184e-01 -3.93696725e-01 -2.82318890e-01
6.86721265e-01 2.58353978e-01 -5.14719605e-01 -1.74802125e-01
-1.55132234e+00 8.73466134e-01 4.26596403e-01 3.44886094e-01
-1.22662449e+00 -6.00246668e-01 -8.79104316e-01 4.93254274e-01
7.79651642e-01 -8.01084995e-01 8.67466569e-01 2.42873639e-01
-1.02219987e+00 1.04810727e+00 1.38724148e-01 -5.35358906e-01
6.10206962e-01 -2.64418840e-01 -3.50343943e-01 2.34218478e-01
5.04417643e-02 2.00375855e-01 5.79556584e-01 -1.19874895e+00
-7.46742129e-01 -5.02330303e-01 3.46790791e-01 5.25380932e-02
-2.57967055e-01 9.89174843e-03 -2.32691035e-01 -2.30412081e-01
2.31775820e-01 -7.72697568e-01 5.81719995e-01 -3.39500576e-01
-1.11245513e+00 -6.46116912e-01 8.44570220e-01 -5.32514930e-01
1.56840539e+00 -2.15705085e+00 4.00199890e-01 4.03891325e-01
9.51599121e-01 1.02293938e-01 2.58096606e-01 3.55958045e-01
-2.10344180e-01 6.35975376e-02 3.33317906e-01 -2.42622450e-01
2.91719109e-01 3.42292935e-01 -1.68264091e-01 4.76228446e-01
1.55622005e-01 1.06675363e+00 -1.09254694e+00 -8.00344050e-01
-8.93343836e-02 2.45328799e-01 -7.20978916e-01 4.58696663e-01
-1.22560270e-01 1.35197267e-01 -7.01635284e-03 1.01053858e+00
8.04956779e-02 -4.02563721e-01 7.56033540e-01 -4.25956100e-01
2.91088521e-01 9.44771528e-01 -1.08252943e+00 1.19001174e+00
1.10969134e-01 6.46512806e-01 -2.26168245e-01 -1.01439905e+00
5.94303489e-01 6.16944730e-01 5.90700567e-01 -5.99945426e-01
2.01440826e-01 6.97580501e-02 3.02457571e-01 -5.28882623e-01
3.25680315e-01 3.48648876e-01 -1.73145458e-01 1.82105824e-01
3.34392488e-01 2.14159891e-01 4.89778906e-01 6.26897693e-01
1.54067183e+00 -1.31638870e-01 4.45276499e-01 1.70722634e-01
3.66824478e-01 -2.46766031e-01 7.45396554e-01 9.07569051e-01
-2.13875100e-01 2.38081545e-01 1.29719865e+00 -3.13931763e-01
-8.87843192e-01 -1.33372271e+00 -1.00474797e-01 9.40379024e-01
2.57548690e-01 -9.04924214e-01 -3.87282878e-01 -9.04627383e-01
1.40985981e-01 5.87567806e-01 -7.10233569e-01 -5.17293632e-01
-6.59104645e-01 -3.05233806e-01 1.14148271e+00 5.92081726e-01
4.81354326e-01 -9.95697200e-01 -6.43687963e-01 5.15580438e-02
-6.09165490e-01 -1.43470681e+00 -2.01063305e-02 4.17186141e-01
-4.56906438e-01 -1.67799735e+00 2.12587893e-01 -5.05997658e-01
3.84491265e-01 1.50300920e-01 1.45939362e+00 2.27779046e-01
-5.99246621e-01 -3.08026164e-03 -2.28569940e-01 -1.06735088e-01
-1.68036073e-01 3.11084181e-01 -1.23586148e-01 -2.77261317e-01
4.24950570e-01 -5.27877212e-01 -5.49620926e-01 7.00902581e-01
-5.86328089e-01 2.46551037e-01 5.28135002e-01 1.18948233e+00
2.32864663e-01 1.76638857e-01 -2.93216445e-02 -1.12243247e+00
1.84035614e-01 -5.95128775e-01 -3.27996790e-01 6.42608762e-01
-6.54653311e-01 7.32845217e-02 2.54876822e-01 -4.05490845e-01
-1.12571573e+00 -3.35585207e-01 5.18600583e-01 -3.97150248e-01
-7.22527225e-03 4.97576058e-01 -4.28074002e-01 1.89476803e-01
8.99423182e-01 -3.65523905e-01 -5.46585858e-01 -1.04937501e-01
4.61449087e-01 3.67203444e-01 7.98380613e-01 -8.44928503e-01
8.03599656e-01 2.93090045e-01 3.23403567e-01 -2.67497629e-01
-1.05601954e+00 -5.07526577e-01 -5.95011115e-01 1.14129353e-02
8.24131608e-01 -8.35182130e-01 -1.30743337e+00 1.85558319e-01
-1.15106356e+00 -7.49812722e-02 -1.77764267e-01 1.76324725e-01
-3.96976650e-01 5.99073134e-02 -9.73245502e-01 -6.43094599e-01
-2.96593159e-01 -1.16719866e+00 1.02027619e+00 -9.81229171e-03
-4.22890425e-01 -3.93078744e-01 -9.91322994e-02 8.99307549e-01
-1.61902413e-01 1.02390863e-01 1.36092603e+00 -7.55457759e-01
-1.01473057e+00 -3.37471813e-01 -5.47294080e-01 -2.31946409e-01
-3.84667277e-01 -8.96111503e-02 -1.04954207e+00 5.91959096e-02
-5.80117404e-01 -5.99392176e-01 8.07847440e-01 -4.01031792e-01
7.36817002e-01 -4.45455879e-01 -7.47616768e-01 5.27158320e-01
8.52226973e-01 -1.25196233e-01 8.39809299e-01 1.29977182e-01
1.11878407e+00 5.96335173e-01 5.56478620e-01 3.96967441e-01
7.95975685e-01 7.54507422e-01 5.01952767e-01 3.35021973e-01
-2.55781859e-01 -5.19698083e-01 1.87846154e-01 4.29484636e-01
-3.24595183e-01 -2.71093637e-01 -1.12279320e+00 5.47900200e-01
-2.14906287e+00 -1.07912874e+00 -3.37971807e-01 2.03221726e+00
7.84588516e-01 3.26677620e-01 2.09668338e-01 6.31673694e-01
3.81171137e-01 -8.56076851e-02 -3.88497591e-01 1.40135601e-01
-1.43478304e-01 -4.26391028e-02 2.49909595e-01 1.01747513e-01
-9.47853506e-01 7.56460607e-01 5.94374847e+00 3.98840636e-01
-4.76745218e-01 3.43704760e-01 1.12190031e-01 -4.49915767e-01
-1.91648513e-01 1.96163520e-01 -9.46857214e-01 1.08252063e-01
5.60395062e-01 1.45202681e-01 6.99622631e-01 7.31744587e-01
-6.46526814e-01 -2.40944818e-01 -1.93881977e+00 1.11478913e+00
-5.22762053e-02 -1.35303164e+00 -1.22648396e-01 2.67477334e-01
4.11641091e-01 -1.99897319e-01 -3.20428051e-02 5.53564250e-01
3.43212456e-01 -1.06175900e+00 8.77633631e-01 2.32093960e-01
6.73754215e-01 -4.89527136e-01 7.72648096e-01 1.43649668e-01
-1.37834525e+00 -3.68953139e-01 6.95227459e-02 1.07234500e-01
-1.65018409e-01 5.62943339e-01 -9.63552237e-01 6.87472582e-01
8.14093530e-01 5.90034783e-01 -6.31056607e-01 7.02202916e-01
-5.08434176e-01 5.57214916e-01 -3.83829623e-01 9.32428092e-02
-1.99657425e-01 2.15495124e-01 7.64239311e-01 1.06887567e+00
-1.29925847e-01 6.87238276e-02 1.02474719e-01 1.11577106e+00
-3.42987239e-01 -4.26618457e-01 -8.65735829e-01 -2.80411363e-01
6.15875363e-01 1.02020764e+00 -4.73559350e-01 -4.03695732e-01
-5.70023954e-01 7.73060083e-01 8.31114590e-01 2.75473744e-01
-1.04907703e+00 -4.40176129e-01 5.90195298e-01 6.07414320e-02
9.93993208e-02 -1.61699932e-02 -4.81814027e-01 -1.35098863e+00
1.50658950e-01 -9.28194821e-01 1.00795853e+00 -3.03788990e-01
-1.25663733e+00 1.63363338e-01 3.43934774e-01 -8.22987080e-01
-2.53501981e-01 -6.60896063e-01 -4.61135566e-01 3.48678529e-01
-1.06062186e+00 -1.49454999e+00 -5.28211951e-01 6.15379691e-01
-4.77801636e-02 -9.77543890e-02 6.87627375e-01 7.38846481e-01
-1.06298411e+00 1.10441446e+00 -6.90177798e-01 4.51058358e-01
5.48433065e-01 -1.31291997e+00 1.09624326e-01 6.71283007e-01
3.65750313e-01 8.99108887e-01 6.30495489e-01 -6.79295540e-01
-1.63092852e+00 -6.75037026e-01 1.14975107e+00 -8.02311957e-01
7.33456790e-01 -7.74742186e-01 -6.77231431e-01 1.10724723e+00
-3.26570831e-02 3.08351964e-01 6.58044219e-01 6.87117100e-01
-9.86687064e-01 -6.19530231e-02 -1.08907235e+00 5.92350543e-01
1.63383317e+00 -8.82795691e-01 -6.02672815e-01 4.43062216e-01
5.88460028e-01 -5.40749133e-01 -1.11145532e+00 5.59648693e-01
9.27082479e-01 -1.25710344e+00 1.17228997e+00 -6.71992540e-01
7.01096833e-01 -1.29391268e-01 -1.42706379e-01 -7.56752849e-01
-3.82625192e-01 -4.92842823e-01 -9.75275874e-01 1.16168928e+00
5.50527096e-01 -4.35913473e-01 7.35640883e-01 5.57919204e-01
1.52104929e-01 -1.02467299e+00 -7.92076051e-01 -6.58573151e-01
-5.76068401e-01 -3.83532554e-01 6.77504003e-01 9.61886704e-01
7.85368681e-01 9.61030602e-01 -3.91878486e-01 2.84040391e-01
9.59831774e-01 1.17464401e-01 7.59569228e-01 -1.28762305e+00
-6.19140625e-01 -5.16393304e-01 -6.32849038e-01 -6.69404030e-01
2.50104904e-01 -1.11942887e+00 -1.83785632e-01 -1.38414907e+00
6.93590939e-01 -3.68789047e-01 -1.86187267e-01 7.51723766e-01
-5.30447140e-02 1.97666049e-01 5.83613776e-02 2.23128438e-01
-8.13575685e-01 3.80478472e-01 8.65126312e-01 -4.55821186e-01
6.04976080e-02 -3.11829150e-01 -4.36554909e-01 5.57161450e-01
3.65341753e-01 -5.15376985e-01 -4.23767656e-01 -3.16946596e-01
7.08794534e-01 2.41132677e-01 7.33594596e-01 -1.00193071e+00
4.80154663e-01 1.67584240e-01 2.49870583e-01 -7.46047616e-01
3.73816848e-01 -7.08888233e-01 1.48139998e-01 2.78069794e-01
-4.98099118e-01 -2.39452839e-01 -4.72281873e-01 4.88184750e-01
-1.96113735e-01 9.60708633e-02 3.38891566e-01 1.08583951e-02
-2.91910946e-01 1.10630706e-01 2.54925508e-02 1.83354706e-01
8.86698425e-01 8.72916132e-02 -9.08486605e-01 -5.54658435e-02
-7.80713916e-01 3.34541529e-01 -7.26882741e-02 3.42516690e-01
7.04743207e-01 -1.32482600e+00 -4.76221621e-01 1.79424599e-01
5.04295886e-01 9.07372758e-02 -8.61986820e-03 1.15328491e+00
-1.91548228e-01 5.62033415e-01 -8.47060978e-02 -4.19993520e-01
-1.32258129e+00 7.58657515e-01 3.08282107e-01 -7.27835238e-01
-8.69723141e-01 7.72914350e-01 1.75536111e-01 -3.00863028e-01
7.14851975e-01 -4.86336857e-01 4.31169057e-03 6.30895495e-02
3.06663126e-01 6.60954535e-01 -1.88768301e-02 -6.05227947e-01
-5.44037402e-01 -1.79917708e-01 5.71658313e-02 5.64475879e-02
8.54612827e-01 1.79368570e-01 -4.09362733e-01 4.22029793e-01
9.59098101e-01 -2.75644809e-01 -4.06840861e-01 -3.96636516e-01
1.43166065e-01 -4.02559787e-01 -2.82616645e-01 -7.77334154e-01
-6.43954515e-01 7.98924088e-01 1.88661188e-01 3.30316395e-01
4.80867296e-01 5.19586980e-01 8.36597204e-01 6.44887686e-01
6.44082308e-01 -7.63880670e-01 2.94778824e-01 5.70086122e-01
6.64851904e-01 -1.12720346e+00 2.31688172e-02 -9.73648548e-01
-4.84932184e-01 8.01063955e-01 9.39188659e-01 2.74961233e-01
3.79426390e-01 1.68654636e-01 -6.79232121e-01 -7.22723246e-01
-1.01537919e+00 -2.00080737e-01 5.06931007e-01 4.56227332e-01
5.51408887e-01 3.97527575e-01 -1.85428560e-02 7.50241160e-01
-2.30317697e-01 -2.67010331e-01 8.57011005e-02 7.20981121e-01
1.92663670e-01 -7.60834694e-01 -2.61222035e-01 6.07293665e-01
-1.46869481e-01 -7.16394112e-02 -6.33929491e-01 8.05846810e-01
2.48125911e-01 8.88223708e-01 -1.22757368e-01 -6.39244497e-01
3.87372464e-01 9.95529145e-02 8.44052196e-01 -5.88207960e-01
-7.35814393e-01 -3.69196415e-01 6.13580763e-01 -9.28943753e-01
1.87819507e-02 -5.26878238e-01 -1.25278103e+00 -3.41839880e-01
-4.31295335e-01 -1.88881680e-01 5.51293930e-03 1.05376029e+00
2.68963128e-01 5.74828923e-01 4.47363779e-02 -4.59696442e-01
-3.89391154e-01 -8.79463077e-01 -4.67319131e-01 7.10487187e-01
2.13557407e-01 -1.20175552e+00 -2.02753052e-01 -3.84013057e-01] | [8.886168479919434, 7.884591102600098] |
3db4e5c8-6332-4472-b316-3dd763d17a69 | line-a-line-a-tool-for-annotating-word | null | null | https://aclanthology.org/2020.bucc-1.1 | https://aclanthology.org/2020.bucc-1.1.pdf | Line-a-line: A Tool for Annotating Word-Alignments | We here describe line-a-line, a web-based tool for manual annotation of word-alignments in sentence-aligned parallel corpora. The graphical user interface, which builds on a design template from the Jigsaw system for investigative analysis, displays the words from each sentence pair that is to be annotated as elements in two vertical lists. An alignment between two words is annotated by drag-and-drop, i.e. by dragging an element from the left-hand list and dropping it on an element in the right-hand list. The tool indicates that two words are aligned by lines that connect them and by highlighting associated words when the mouse is hovered over them. Line-a-line uses the efmaral library for producing pre-annotated alignments, on which the user can base the manual annotation. The tool is mainly planned to be used on moderately under-resourced languages, for which resources in the form of parallel corpora are scarce. The automatic word-alignment functionality therefore also incorporates information derived from non-parallel resources, in the form of pre-trained multilingual word embeddings from the MUSE library. | ['Rickard Domeij', 'Maria Skeppstedt', 'Gunnar Eriksson', 'Magnus Ahltorp'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['multilingual-word-embeddings'] | ['methodology'] | [ 2.50187159e-01 1.17580183e-01 -9.03557017e-02 -3.04922879e-01
-9.00469840e-01 -6.62790477e-01 4.11960751e-01 8.82932782e-01
-8.53558660e-01 5.98932266e-01 4.31299806e-01 -7.13214517e-01
-6.58022845e-03 -3.62085938e-01 -7.02302158e-02 -4.54364181e-01
2.65468620e-02 8.41594338e-01 3.81031245e-01 -4.49289888e-01
4.71477270e-01 2.80299842e-01 -9.11016524e-01 3.65857154e-01
5.82030475e-01 1.00079782e-01 6.23228490e-01 6.68218076e-01
-6.07687056e-01 -1.07627064e-02 -4.89215165e-01 -8.36830497e-01
1.91721559e-01 -5.01771629e-01 -1.02906072e+00 -3.22102338e-01
1.48318157e-01 -8.39159712e-02 2.79352158e-01 8.00324440e-01
4.87490147e-01 -4.47362006e-01 3.87407482e-01 -7.12245882e-01
-3.30738455e-01 5.71475804e-01 -3.75232369e-01 4.82643485e-01
6.75048590e-01 -1.79417312e-01 1.17563558e+00 -9.84385252e-01
1.12993610e+00 9.24487412e-01 5.52670360e-01 3.19118589e-01
-1.07995522e+00 -4.25298363e-01 -1.13087766e-01 9.17792767e-02
-9.86134529e-01 -3.32519054e-01 4.66168344e-01 -6.94906056e-01
1.48415136e+00 4.59865779e-01 5.80393314e-01 6.71328187e-01
3.36989045e-01 1.22593857e-01 7.03075588e-01 -1.21975672e+00
-3.94282117e-02 2.59085536e-01 1.67150885e-01 5.08115530e-01
2.00929970e-01 -3.76715153e-01 -5.17293751e-01 -3.23901206e-01
4.01931494e-01 -6.36762857e-01 -9.13004950e-02 -2.12556019e-01
-1.31806219e+00 5.72877526e-01 7.34676719e-02 9.10823345e-01
-4.05299157e-01 -4.54990327e-01 7.43478358e-01 2.43144870e-01
6.51218474e-01 5.02279699e-01 -4.46749687e-01 -3.21717650e-01
-7.70013154e-01 2.74671733e-01 6.52121305e-01 7.41445899e-01
5.60841501e-01 -4.58849579e-01 7.47041777e-02 9.86439705e-01
2.23676294e-01 3.37845385e-02 7.24513829e-01 -3.59537154e-01
5.66675246e-01 7.72336960e-01 1.59978539e-01 -5.58884323e-01
-3.87431413e-01 -2.80777097e-01 -3.05251386e-02 -3.47858132e-03
6.20505214e-01 -8.49032030e-02 -3.68963212e-01 1.37470579e+00
5.39070904e-01 -4.74930078e-01 5.62449917e-02 7.14872479e-01
7.46543527e-01 3.60081643e-01 3.85303162e-02 -4.26778346e-01
1.74031162e+00 -7.00478077e-01 -9.31419015e-01 -2.20139578e-01
1.06573582e+00 -1.43567848e+00 1.14622152e+00 2.09718734e-01
-1.10682988e+00 -3.54449421e-01 -1.16937625e+00 -1.42935097e-01
-7.19547451e-01 1.75410584e-02 4.96953763e-02 3.36348325e-01
-9.12018895e-01 5.55823565e-01 -8.94001007e-01 -8.12697768e-01
-2.81615287e-01 1.65146962e-01 -6.69507563e-01 1.98906884e-01
-1.26509833e+00 1.46648335e+00 4.05222058e-01 -8.54261220e-03
2.06750065e-01 -3.85672569e-01 -9.67864335e-01 -2.18479633e-01
1.65001422e-01 -4.33168650e-01 1.16051078e+00 -7.73595691e-01
-1.28563356e+00 1.38165820e+00 -1.62658721e-01 -1.61072150e-01
4.86063540e-01 -2.43874535e-01 -7.22070098e-01 8.93864036e-02
5.26915193e-01 4.39343929e-01 4.08712089e-01 -8.45363855e-01
-5.53842187e-01 -5.14625967e-01 -1.29598945e-01 1.82386801e-01
-3.26614380e-01 7.43640602e-01 -7.22269893e-01 -6.50473714e-01
-1.36443377e-01 -8.00119579e-01 -1.02980658e-01 -3.38579923e-01
-3.30581456e-01 -3.46989006e-01 6.73488915e-01 -1.29044497e+00
1.78580880e+00 -1.86382043e+00 5.50279440e-03 4.45304103e-02
-8.73725265e-02 4.02260512e-01 -1.48721531e-01 1.17331314e+00
-5.98788738e-01 6.09175377e-02 -1.38396502e-01 -2.12276965e-01
-5.79548907e-03 2.05731541e-01 1.17347501e-01 4.04301405e-01
2.48438716e-01 4.17938709e-01 -1.14401245e+00 -6.14362419e-01
4.39268872e-02 1.68232962e-01 -3.91382203e-02 2.08284840e-01
-6.99712941e-03 2.43533552e-01 -3.00861537e-01 4.10342887e-02
3.69786829e-01 3.08920324e-01 5.67076504e-01 -1.15479447e-01
-6.70927882e-01 8.34406018e-01 -1.03691053e+00 1.89369106e+00
-6.51993811e-01 5.37990034e-01 -2.00592145e-01 -6.18438840e-01
1.02257872e+00 5.14595151e-01 -5.76465130e-02 -2.99620420e-01
-5.03954403e-02 5.00021636e-01 3.13163251e-01 -9.29851234e-01
6.36925459e-01 1.07754581e-01 -4.75203656e-02 6.27748728e-01
6.05954826e-02 -5.42136021e-02 9.43710148e-01 2.23602399e-01
9.00507629e-01 4.85454023e-01 8.30393314e-01 -3.61496359e-01
8.32124531e-01 1.32355779e-01 3.31129372e-01 1.93211019e-01
2.70355105e-01 2.19399735e-01 7.36328244e-01 -5.64180911e-01
-1.62614965e+00 -6.40105844e-01 -5.45556188e-01 1.10404599e+00
-5.87805986e-01 -1.00793040e+00 -9.55628633e-01 -5.77632070e-01
-4.75473404e-01 9.47108448e-01 -5.34020245e-01 3.97553861e-01
-6.52518809e-01 -1.87357128e-01 1.35511696e-01 2.72775412e-01
-2.85922408e-01 -1.16102529e+00 -8.43778133e-01 3.01852942e-01
-4.01127301e-02 -6.78620696e-01 -5.36722004e-01 1.37033269e-01
-4.15049464e-01 -1.06843114e+00 -5.87278426e-01 -9.57167566e-01
7.58076549e-01 -1.97490275e-01 9.63533342e-01 2.64540426e-02
-2.45486453e-01 7.81423002e-02 -5.30568838e-01 -5.46841621e-01
-6.56430483e-01 2.32465938e-01 6.78543448e-02 -3.10097367e-01
6.97499216e-01 -4.14274782e-01 -1.48685113e-01 1.25129789e-01
-8.38131726e-01 1.71417952e-01 2.17122540e-01 7.41621792e-01
6.11241102e-01 -6.61946416e-01 5.13278008e-01 -7.62619734e-01
8.20595860e-01 -2.31337637e-01 -4.99959379e-01 3.36870551e-01
-2.50648797e-01 8.17446411e-02 4.69662309e-01 -1.06730282e-01
-4.12467718e-01 -5.02051376e-02 -6.80501819e-01 1.14632756e-01
-3.28600615e-01 6.09559774e-01 -2.16714218e-01 3.76476079e-01
5.26056111e-01 -1.42280400e-01 8.33256170e-02 -6.05597079e-01
5.11546612e-01 1.21619570e+00 6.78966761e-01 -2.53514290e-01
5.83239853e-01 -2.26838395e-01 -3.75369042e-01 -6.83131039e-01
-3.61538082e-01 -7.87024140e-01 -1.10263705e+00 -3.00587028e-01
9.41908240e-01 -4.72736239e-01 -1.27763137e-01 -3.66547965e-02
-1.38122475e+00 -8.13323185e-02 -1.91301197e-01 6.80024922e-01
-2.58961052e-01 3.44185799e-01 -3.74500245e-01 -4.98866349e-01
-4.85693246e-01 -1.16027033e+00 8.14955413e-01 3.20159346e-02
-1.17032516e+00 -9.87833977e-01 5.98700225e-01 1.13682874e-01
-6.46409988e-02 -1.00267269e-01 1.21283913e+00 -1.45445633e+00
4.24532354e-01 -7.07819462e-01 3.34235489e-01 2.52283096e-01
1.84125543e-01 1.40975356e-01 -4.96166557e-01 -6.63089678e-02
-1.40926659e-01 3.60828824e-02 2.31002539e-01 -2.20715433e-01
4.80303764e-01 -2.76493251e-01 -4.89742190e-01 -8.34052339e-02
1.06814003e+00 4.03084040e-01 5.63898504e-01 7.19000995e-01
3.07162642e-01 9.84913707e-01 4.10136402e-01 1.21871650e-01
-3.50407250e-02 1.02942252e+00 1.63538679e-02 5.14712185e-02
2.06478953e-01 2.82930396e-02 1.23941705e-01 1.19353712e+00
2.02382416e-01 -1.51285797e-01 -1.10418439e+00 7.56727934e-01
-1.58003092e+00 -7.74823308e-01 -8.31408799e-01 2.31170511e+00
9.89670157e-01 4.78937268e-01 3.33739966e-01 2.28091940e-01
7.41688251e-01 3.61051857e-02 1.07152216e-01 -9.58774149e-01
3.76158029e-01 3.73815626e-01 1.69779152e-01 8.91063929e-01
-6.72307968e-01 7.76574969e-01 5.36010933e+00 7.99519897e-01
-8.34696829e-01 3.01839501e-01 1.59212902e-01 6.54278547e-02
-3.96492034e-01 1.68369800e-01 -8.85020137e-01 4.49603498e-01
1.00322998e+00 -1.61081508e-01 9.58701000e-02 6.01535857e-01
5.43075800e-01 -2.96127707e-01 -1.20126998e+00 4.18173820e-01
1.36597484e-01 -1.36277175e+00 -1.20447911e-01 9.07335430e-02
1.01157904e-01 -1.35775320e-02 -4.53311861e-01 -1.88992873e-01
-7.16044083e-02 -6.16499066e-01 6.75505519e-01 2.12355703e-01
9.67246294e-01 -6.73721373e-01 1.06009543e+00 1.36602908e-01
-1.02312648e+00 5.66584170e-01 -1.28578424e-01 6.68475078e-03
5.95905662e-01 4.56442863e-01 -8.18453729e-01 8.37904751e-01
3.52292269e-01 1.62082151e-01 -4.53353226e-01 1.06807673e+00
-4.52302665e-01 3.84432197e-01 -2.32382134e-01 -1.04406208e-01
1.68515205e-01 -5.57519197e-01 7.35792816e-01 1.71221566e+00
2.50968963e-01 -2.35531449e-01 -3.27143185e-02 3.55089515e-01
3.53309721e-01 7.45118141e-01 -4.91225392e-01 -2.89283078e-02
4.63991195e-01 1.48084140e+00 -8.42239022e-01 -2.48360083e-01
-6.00130856e-01 9.30249810e-01 2.75859684e-01 -8.01374093e-02
-4.37893540e-01 -9.32084382e-01 3.72049898e-01 3.95513058e-01
2.95552075e-01 -2.39034563e-01 -3.83891352e-02 -3.55298638e-01
1.18875064e-01 -9.92095172e-01 3.47946227e-01 -1.01288235e+00
-1.03815985e+00 1.11226356e+00 1.85488850e-01 -1.09512794e+00
-5.32931387e-01 -6.36665106e-01 -7.87143707e-01 1.53946018e+00
-4.61691976e-01 -1.00661242e+00 2.46035695e-01 2.53483832e-01
5.41627645e-01 -1.59901962e-01 1.36193371e+00 2.08436728e-01
-6.60249531e-01 4.75456059e-01 3.48772407e-02 1.60608813e-01
8.17397714e-01 -1.22999597e+00 5.74236751e-01 6.89440072e-01
3.72687459e-01 8.65422547e-01 9.06551719e-01 -8.57792377e-01
-5.14357626e-01 -4.67753977e-01 2.03764963e+00 -2.69420236e-01
1.26235366e+00 -7.14938819e-01 -1.10074830e+00 4.99260902e-01
8.62300336e-01 -4.54751819e-01 1.18669426e+00 7.44950250e-02
-6.80043250e-02 1.57652572e-01 -7.41454422e-01 7.44008541e-01
5.07805943e-01 -5.60267746e-01 -1.11297607e+00 8.07473302e-01
5.48536122e-01 -4.65154648e-01 -7.97673762e-01 -5.77872321e-02
4.93001819e-01 -4.95082945e-01 4.60261792e-01 -7.81443238e-01
4.68235940e-01 -4.58110034e-01 1.78970605e-01 -1.38875413e+00
-1.49060443e-01 -8.43920469e-01 6.69205964e-01 1.47394168e+00
7.52448440e-01 -5.60946107e-01 1.36010483e-01 1.10690974e-01
-2.72911459e-01 -7.30853856e-01 -9.74530339e-01 -5.20281851e-01
-1.31517015e-02 -3.93877447e-01 4.82229531e-01 8.90266120e-01
8.44613671e-01 5.29097557e-01 1.08299039e-01 -1.86567083e-01
-8.08091536e-02 -4.51935172e-01 5.09066880e-01 -8.19309354e-01
-2.35053122e-01 -5.08030415e-01 -2.53233224e-01 -2.97717839e-01
1.69862516e-03 -1.10825038e+00 -5.91839366e-02 -1.62060738e+00
-9.40408260e-02 -2.44241413e-02 7.32359961e-02 5.23752332e-01
-3.36169392e-01 5.53526506e-02 1.29428536e-01 2.24127710e-01
3.24126482e-02 -1.12944782e-01 5.09025276e-01 2.79211223e-01
-1.61132261e-01 -3.29061896e-01 -3.77051622e-01 8.26079130e-01
8.44052672e-01 -8.35551739e-01 7.53563493e-02 -1.82163388e-01
3.62741798e-01 -2.38411918e-01 -1.15577795e-01 -6.46220386e-01
-2.68024541e-02 7.76916295e-02 9.77674499e-02 -5.21022320e-01
-4.59507927e-02 -6.97037101e-01 2.36789167e-01 5.98699570e-01
-4.80424345e-01 9.56803501e-01 4.65188473e-01 -1.17807724e-01
1.01850368e-01 -1.04344666e+00 4.00058240e-01 -2.55678505e-01
-2.82797992e-01 -3.69687617e-01 -6.54668570e-01 -2.00524941e-01
1.10664630e+00 -1.32603869e-01 -2.47206941e-01 -1.47705108e-01
-8.32711875e-01 1.16312586e-01 4.57348198e-01 3.03044438e-01
1.66426808e-01 -1.30248153e+00 -6.68777108e-01 2.21358657e-01
4.23984677e-01 -3.71092707e-01 1.15421981e-01 9.46944654e-01
-9.32252467e-01 5.05533516e-01 -2.60180265e-01 -1.25570625e-01
-1.61267519e+00 6.94391072e-01 -5.50343245e-02 -6.67625189e-01
-5.65481007e-01 7.27359712e-01 -2.70744562e-01 -4.05333877e-01
-1.25369668e-01 -4.48904969e-02 -4.50822264e-01 3.24848622e-01
6.28803372e-01 1.23837560e-01 4.15287226e-01 -8.51669133e-01
-6.48481071e-01 4.61105406e-01 -1.89245403e-01 -6.22690260e-01
1.44009018e+00 9.28545818e-02 -4.31043386e-01 7.25520611e-01
1.04874122e+00 6.03651702e-01 -5.26370347e-01 5.24569415e-02
3.45169693e-01 -2.51015514e-01 -3.80541533e-01 -8.01346242e-01
-2.00967282e-01 6.60263240e-01 3.88784438e-01 3.03360999e-01
6.49822772e-01 1.33677751e-01 4.02615756e-01 2.60524880e-02
2.01024935e-02 -1.08990240e+00 -2.81034231e-01 3.28382462e-01
1.21874809e+00 -4.41113055e-01 1.06928915e-01 -4.53491420e-01
-5.80452025e-01 1.41140890e+00 2.40454376e-01 1.33895323e-01
3.22900176e-01 3.48802626e-01 3.19559246e-01 -1.33730382e-01
-5.07731318e-01 -2.95163020e-02 2.25858465e-01 5.92241704e-01
9.07458663e-01 -1.76013261e-01 -1.35359156e+00 4.23166662e-01
-6.29203081e-01 -1.80669829e-01 4.90576327e-01 7.85339415e-01
-2.02001587e-01 -1.91375363e+00 -5.15063584e-01 2.68444330e-01
-5.70509791e-01 -5.56230009e-01 -6.55164838e-01 9.49156225e-01
4.00381178e-01 7.72470951e-01 5.38757801e-01 -1.41871393e-01
4.51004416e-01 4.56802756e-01 4.73371118e-01 -7.23841965e-01
-9.44810867e-01 2.83590436e-01 7.93234468e-01 -1.20573051e-01
-1.19197562e-01 -9.10968244e-01 -1.00608778e+00 -5.50710456e-03
-1.51115283e-01 3.08082819e-01 7.35836923e-01 1.02884114e+00
-6.73971837e-04 4.36252981e-01 -1.06833227e-01 -8.08306456e-01
-1.82715341e-01 -1.24787235e+00 -1.97063103e-01 1.84609726e-01
-2.22078398e-01 -2.95160592e-01 3.40047963e-02 9.13923457e-02] | [10.341002464294434, 10.031517028808594] |
2902ee6f-6a87-42e5-84b1-9213aed190b4 | bert-enhanced-relational-sentence-ordering | null | null | https://aclanthology.org/2020.emnlp-main.511 | https://aclanthology.org/2020.emnlp-main.511.pdf | BERT-enhanced Relational Sentence Ordering Network | In this paper, we introduce a novel BERT-enhanced Relational Sentence Ordering Network (referred to as BRSON) by leveraging BERT for capturing better dependency relationship among sentences to enhance the coherence modeling for the entire paragraph. In particular, we develop a new Relational Pointer Decoder (referred as RPD) by incorporating the relative ordering information into the pointer network with a Deep Relational Module (referred as DRM), which utilizes BERT to exploit the deep semantic connection and relative ordering between sentences.This enables us to strengthen both local and global dependencies among sentences. Extensive evaluations are conducted on six public datasets. The experimental results demonstrate the effectiveness and promise of our BRSON, showing a significant improvement over the state-of-the-art by a wide margin. | ['Zhongfei Zhang', 'Yingming Li', 'Baiyun Cui'] | null | null | null | null | emnlp-2020-11 | ['sentence-ordering'] | ['natural-language-processing'] | [-2.36924127e-01 2.68009812e-01 -3.24887186e-01 -7.73540676e-01
-5.51267624e-01 -2.42448643e-01 4.32283223e-01 1.67399079e-01
-1.62030503e-01 4.75120991e-01 9.01247919e-01 -2.80684501e-01
-2.70018667e-01 -8.19186151e-01 -7.55734324e-01 -2.48779088e-01
-2.53325328e-02 7.44329169e-02 6.04016840e-01 -5.70416212e-01
2.54484594e-01 1.27114788e-01 -7.08030760e-01 4.70013142e-01
9.06246245e-01 5.99149406e-01 4.19346243e-01 2.91557074e-01
-1.08669713e-01 1.22708225e+00 -6.16659224e-01 -5.51686347e-01
-1.65877730e-01 -2.97800630e-01 -8.55004132e-01 -1.19154617e-01
-8.73718485e-02 -6.39835060e-01 -8.76241207e-01 7.85783708e-01
2.85461128e-01 1.43860087e-01 -5.78480959e-02 -5.88990808e-01
-8.64451230e-01 1.53157258e+00 -9.46034610e-01 6.06356800e-01
5.34766018e-01 8.13076273e-03 1.65614128e+00 -7.44472027e-01
3.02579761e-01 1.48848176e+00 5.81694365e-01 1.45982280e-01
-1.05432212e+00 -5.60190558e-01 4.08539563e-01 3.83799374e-01
-1.52659702e+00 -4.50509936e-01 9.26170290e-01 1.09659180e-01
1.29376853e+00 1.61914721e-01 3.52817923e-01 8.27999175e-01
2.93169349e-01 9.20347691e-01 6.26924217e-01 -2.35237017e-01
-2.86234558e-01 -3.79203379e-01 8.60980332e-01 6.70924425e-01
1.35625586e-01 -3.09524953e-01 -6.29778922e-01 3.73994440e-01
5.28277636e-01 -1.07850969e-01 -2.57693201e-01 2.20989957e-01
-1.07413340e+00 5.94327807e-01 8.95813286e-01 4.37930495e-01
-2.68992841e-01 7.37129226e-02 6.46393955e-01 1.91330343e-01
5.84652781e-01 3.38219106e-01 -3.60661417e-01 -1.75100446e-01
-6.11557782e-01 -2.37279311e-02 5.97244442e-01 1.19048584e+00
5.41343689e-01 -5.31033516e-01 -6.58746958e-01 9.10034776e-01
7.35524178e-01 1.35904849e-01 2.17042074e-01 -8.32509816e-01
1.11183918e+00 9.02488470e-01 -4.89481598e-01 -1.36889005e+00
-6.98585391e-01 -8.60895038e-01 -1.12320185e+00 -1.03674746e+00
-3.59132886e-01 5.46719432e-02 -3.39507192e-01 1.85443401e+00
1.47345945e-01 2.33826041e-01 1.45401359e-01 6.81327522e-01
1.21440518e+00 8.31345081e-01 -3.29249710e-01 -2.68462300e-01
1.35246909e+00 -1.27986836e+00 -9.27109063e-01 -3.21440876e-01
7.97585845e-01 -4.98561531e-01 1.15136945e+00 7.71224201e-02
-1.03092480e+00 -7.13699043e-01 -1.28549254e+00 -5.74863255e-01
1.60577416e-01 -1.34856388e-01 6.50913775e-01 1.03812456e-01
-1.08166409e+00 5.38520277e-01 -1.14895332e+00 -1.40440047e-01
4.26063180e-01 4.02967274e-01 -1.84446573e-01 -1.45461529e-01
-1.68532276e+00 8.23195279e-01 5.88634431e-01 6.01702213e-01
-3.74572694e-01 -6.24647856e-01 -1.05613232e+00 3.09356928e-01
6.27019227e-01 -8.43489766e-01 1.17752767e+00 6.76527619e-02
-1.56821322e+00 4.71290886e-01 -4.55231935e-01 -5.55471241e-01
1.29300833e-01 -4.56136733e-01 -3.38775694e-01 2.67481923e-01
9.12536085e-02 5.66384614e-01 -5.43866046e-02 -1.02557743e+00
-2.43341714e-01 -1.46976322e-01 7.55973458e-01 4.13904101e-01
-5.36447108e-01 1.68043643e-01 -1.15296113e+00 -6.61140144e-01
1.83105350e-01 -5.93463778e-01 8.71269871e-03 -6.44998848e-01
-1.08380985e+00 -5.75937390e-01 5.75729430e-01 -7.65553415e-01
2.05285120e+00 -2.11871815e+00 4.58091289e-01 -1.00926235e-02
5.61617613e-01 3.22770216e-02 -2.24598303e-01 7.24286616e-01
-9.20480117e-03 2.49944627e-01 -2.32805878e-01 -6.84743226e-01
1.19785806e-02 5.41451037e-01 -3.44611794e-01 1.05526738e-01
3.02270859e-01 1.23976219e+00 -9.03421581e-01 -6.95236504e-01
-1.94181520e-02 2.02819273e-01 -5.60990036e-01 1.74157962e-01
-1.34164497e-01 3.02267224e-01 -6.28826141e-01 1.95849612e-01
7.56142557e-01 -5.82328498e-01 3.25924426e-01 -3.26205283e-01
1.84278369e-01 1.09761739e+00 -7.32211173e-01 1.91642451e+00
-4.27921116e-01 4.10944402e-01 -2.97011465e-01 -6.19265079e-01
1.02839756e+00 2.67189294e-02 2.42824733e-01 -8.04595768e-01
-4.91507128e-02 -2.70498365e-01 2.75929838e-01 -5.29922605e-01
8.06336343e-01 1.08115077e-02 -2.18375504e-01 4.24095392e-01
-1.43816084e-01 2.57165909e-01 4.42238599e-01 9.87827718e-01
1.38772202e+00 -1.03533879e-01 1.67376101e-01 -3.75391692e-01
7.84128845e-01 -4.97345328e-01 6.28555059e-01 4.56946522e-01
6.76976368e-02 3.76702517e-01 1.18944883e+00 1.21756695e-01
-7.19570637e-01 -8.54442358e-01 -1.82012338e-02 7.82706738e-01
4.80163634e-01 -9.81108487e-01 -4.90717053e-01 -7.34174550e-01
-1.99534088e-01 7.73739755e-01 -7.06591189e-01 -4.34333384e-01
-9.50749576e-01 -7.57392704e-01 4.21379358e-01 8.81799161e-01
8.81914377e-01 -7.44875908e-01 2.12618589e-01 7.69761652e-02
-4.64996487e-01 -1.47054231e+00 -6.58481300e-01 -4.19774652e-02
-8.32658052e-01 -7.90305018e-01 -1.04415990e-01 -5.80052257e-01
4.24470752e-01 5.62298179e-01 1.15129781e+00 1.88644692e-01
5.53139746e-01 -2.97892183e-01 -6.83887184e-01 2.08435848e-01
-2.45624512e-01 6.03238165e-01 -2.09622696e-01 -2.96153277e-01
1.64546505e-01 -5.05149901e-01 -6.47866964e-01 2.05171674e-01
-9.51721311e-01 3.58831555e-01 7.04524279e-01 7.70688713e-01
2.38215864e-01 3.89833540e-01 6.58382237e-01 -1.19947863e+00
8.80855441e-01 -5.43319583e-01 -2.64079362e-01 3.80236655e-01
-4.58722860e-01 2.52434522e-01 6.41041517e-01 2.73621649e-01
-1.14474916e+00 -7.65829325e-01 -3.62664670e-01 2.54639201e-02
5.77468395e-01 1.12996709e+00 -4.21462655e-01 7.01557994e-01
2.22075693e-02 -4.00811136e-02 -3.04696321e-01 -5.62995732e-01
5.14177144e-01 6.31443024e-01 6.33414030e-01 -5.01375437e-01
6.78492844e-01 2.42172718e-01 -3.68176308e-03 -2.97812998e-01
-1.52734160e+00 -5.18559098e-01 -7.98339844e-01 1.23735957e-01
7.44984210e-01 -9.53258574e-01 -7.45322466e-01 2.70019114e-01
-1.58439851e+00 -6.35498762e-02 -4.82034795e-02 2.39996672e-01
3.52388918e-02 6.66804612e-01 -1.12303746e+00 -3.12719673e-01
-4.42391604e-01 -1.21454620e+00 1.07626903e+00 2.10043594e-01
-1.07884012e-01 -1.10662031e+00 6.70679659e-03 5.19786298e-01
1.29462495e-01 -1.16139598e-01 9.83842850e-01 -5.82975447e-01
-6.51317596e-01 2.85353716e-02 -7.10702479e-01 3.94339442e-01
1.56950086e-01 1.03382371e-01 -5.52981973e-01 -1.04109921e-01
-1.27369702e-01 -6.58500493e-02 1.17957258e+00 -8.75265710e-03
1.10407233e+00 -2.40311682e-01 -3.73641521e-01 3.38793635e-01
1.11319113e+00 -1.29082948e-01 8.12335312e-01 2.64490128e-01
9.84709680e-01 5.75649619e-01 5.32763541e-01 3.60037327e-01
1.04068291e+00 7.68674433e-01 3.34934652e-01 9.68590155e-02
-2.61179239e-01 -3.92960429e-01 2.99916744e-01 1.86728978e+00
3.04741085e-01 -4.45674509e-01 -8.03767622e-01 2.22939998e-01
-1.99203050e+00 -7.60497749e-01 -3.76337677e-01 1.54356682e+00
1.18208945e+00 7.49536514e-01 -1.71837062e-01 1.22284114e-01
8.22089672e-01 5.67356229e-01 -3.74196440e-01 -2.97940522e-01
-2.14888737e-01 3.26247923e-02 6.93952292e-02 6.99416161e-01
-9.63648081e-01 9.66454089e-01 6.04785156e+00 7.94550896e-01
-5.75001478e-01 -7.79249966e-02 5.96394420e-01 1.57349497e-01
-5.10469854e-01 5.50087690e-02 -1.10361469e+00 5.10453582e-01
8.81412923e-01 -1.99999154e-01 2.25684285e-01 2.95571029e-01
4.83387977e-01 1.27573565e-01 -1.13743019e+00 4.77656752e-01
6.74951002e-02 -1.43872511e+00 -9.28053483e-02 -1.43305898e-01
3.93579692e-01 2.14757659e-02 -2.58563012e-01 4.68543142e-01
4.59954888e-01 -6.20302618e-01 4.34858143e-01 7.71434128e-01
4.34617668e-01 -8.39731812e-01 1.06833816e+00 3.16412449e-01
-1.37578428e+00 7.66340494e-02 -3.92176986e-01 8.77193734e-03
3.55319977e-01 8.50391090e-01 -4.55019057e-01 1.24918389e+00
5.16359210e-01 1.46546006e+00 -7.83534706e-01 5.70730031e-01
-5.87897539e-01 6.65415525e-01 -1.36044115e-01 -1.23903133e-01
3.35879356e-01 -1.60866544e-01 4.05379683e-01 1.32028055e+00
-1.13616310e-01 1.44618362e-01 1.02623902e-01 8.56468141e-01
-5.06414473e-01 -1.79148436e-01 -2.92905897e-01 1.36018917e-01
8.56870711e-01 1.36723542e+00 -4.28200066e-01 -3.19678754e-01
-5.48923731e-01 8.70727956e-01 6.24294937e-01 3.84585038e-02
-1.06258631e+00 -5.07950246e-01 2.21344724e-01 -2.27965757e-01
4.76133138e-01 -3.60629916e-01 -4.01066571e-01 -1.15154076e+00
2.72272527e-01 -5.40293992e-01 3.12210619e-01 -8.01216662e-01
-1.34281802e+00 7.38592625e-01 2.49069259e-01 -1.03451896e+00
1.12405181e-01 -2.47181118e-01 -6.29527688e-01 8.04907680e-01
-1.52000999e+00 -1.08938777e+00 -1.52729109e-01 1.62228972e-01
2.77688056e-01 1.87656194e-01 2.78333485e-01 3.55707169e-01
-1.20043194e+00 6.86207533e-01 -1.25484467e-01 3.17170590e-01
5.04229367e-01 -1.30620754e+00 5.69732487e-01 9.63222265e-01
-3.15546766e-02 1.24589992e+00 2.71791905e-01 -5.20136416e-01
-1.40626514e+00 -1.07685745e+00 1.27105331e+00 -3.09611142e-01
1.10235226e+00 -4.87927169e-01 -1.04621994e+00 8.08786452e-01
5.94430745e-01 -2.73383141e-01 6.42583728e-01 4.77257073e-01
-4.12897140e-01 -4.69698459e-01 -4.79878008e-01 5.43937027e-01
1.24236226e+00 -6.16623104e-01 -1.00969326e+00 3.16282719e-01
1.69870400e+00 -5.99301398e-01 -1.04621899e+00 5.89994192e-01
7.07243234e-02 -9.26367640e-01 8.46719086e-01 -3.25266182e-01
1.03602910e+00 -3.17156643e-01 -2.36912280e-01 -1.06328845e+00
-5.76667130e-01 -6.05649710e-01 -5.21266401e-01 1.79235518e+00
5.14808118e-01 -5.50455749e-01 4.56930846e-01 3.29816312e-01
-6.27400756e-01 -1.11705649e+00 -7.41371214e-01 -7.42474616e-01
-1.39084414e-01 -4.19610351e-01 7.57069290e-01 8.29582691e-01
3.03510964e-01 1.28325188e+00 -1.75530285e-01 2.18651757e-01
1.26095712e-01 1.69120163e-01 5.27300894e-01 -7.56532907e-01
-6.62583470e-01 -5.16370475e-01 -1.16329432e-01 -1.96385145e+00
3.51412952e-01 -1.04209769e+00 7.50447363e-02 -1.83617795e+00
7.13672996e-01 -3.69937062e-01 -5.37036538e-01 3.24846983e-01
-6.13557041e-01 -8.62200111e-02 2.37221956e-01 4.17343557e-01
-1.11642158e+00 9.71070766e-01 1.57423484e+00 -1.13505483e-01
-8.56704637e-02 -5.75849786e-02 -1.33419609e+00 4.36637878e-01
4.88188595e-01 -3.48124981e-01 -6.16756499e-01 -1.02693188e+00
5.82901478e-01 2.20286593e-01 3.68604772e-02 -5.84493458e-01
5.06163120e-01 3.93094122e-01 -2.00172588e-01 -1.15392494e+00
2.07973585e-01 -3.89725327e-01 -3.27114582e-01 2.31849954e-01
-7.57363260e-01 3.37317050e-01 1.27938524e-01 7.24855661e-01
-4.97381359e-01 -9.71389934e-02 3.38349640e-01 2.84020633e-01
-4.27663326e-01 1.39320999e-01 2.95521349e-01 9.22439769e-02
7.36151516e-01 4.18008298e-01 -9.60568488e-01 -3.14888030e-01
-3.41080785e-01 7.80470431e-01 -1.58709418e-02 4.46021080e-01
6.43005729e-01 -1.38907087e+00 -7.46497631e-01 -1.44144297e-01
1.09569043e-01 4.75516170e-01 3.22015256e-01 1.09371579e+00
-4.80266541e-01 8.27163517e-01 5.24700761e-01 -5.66189110e-01
-1.20822287e+00 5.13802171e-01 9.51772407e-02 -9.68791723e-01
-8.82879078e-01 1.02189732e+00 2.62769967e-01 -4.09924775e-01
9.01087821e-02 -6.90167665e-01 -3.74511719e-01 -2.37122625e-01
3.00276905e-01 1.57272935e-01 6.96032941e-02 -5.22805393e-01
-6.04745805e-01 3.00576091e-01 -4.90623921e-01 9.29637998e-02
1.33258021e+00 -4.77652043e-01 -8.84078681e-01 4.18814361e-01
1.38096869e+00 1.31118625e-01 -1.08024788e+00 -7.66825080e-01
2.77934223e-01 -2.17769653e-01 1.48376390e-01 -6.37295604e-01
-1.09189785e+00 8.23290706e-01 -4.36182201e-01 3.83171588e-01
1.18869829e+00 2.88326651e-01 1.04023314e+00 4.77310121e-01
1.22560170e-02 -6.51824534e-01 3.69556814e-01 9.14352715e-01
9.99010921e-01 -9.43269253e-01 3.87746841e-01 -7.95275211e-01
-7.84081578e-01 1.09372270e+00 6.17466092e-01 -9.01299417e-02
7.34897316e-01 3.97046685e-01 -3.66400361e-01 -2.07663745e-01
-1.10335147e+00 3.73396166e-02 3.90448123e-01 -1.73165709e-01
6.36717796e-01 -6.36363998e-02 -4.75433111e-01 8.88018072e-01
-1.52809218e-01 -2.23500058e-01 3.97336632e-01 6.70927048e-01
-3.90815675e-01 -1.31455159e+00 1.77283019e-01 3.85757446e-01
-1.41095877e-01 -6.29330158e-01 -4.15154725e-01 7.70824015e-01
-1.97862536e-01 1.16937971e+00 2.13538468e-01 -9.05680895e-01
4.04581249e-01 -5.10564327e-01 2.82019168e-01 -9.30240273e-01
-6.31867230e-01 1.61478847e-01 4.36551213e-01 -4.44766819e-01
-4.80166763e-01 -5.54303229e-01 -1.61405480e+00 -5.84329367e-01
-4.08296585e-01 -3.47086141e-04 5.11967763e-02 1.08897638e+00
6.11543596e-01 1.24544978e+00 9.51750994e-01 -1.15160914e-02
-5.06783366e-01 -1.38215888e+00 -3.54496986e-01 8.87603685e-02
1.31742433e-01 -3.69341969e-01 4.22065668e-02 -2.99044639e-01] | [10.931355476379395, 8.680809020996094] |
7e1b556b-cdcb-49d3-aba1-3731d0e5b641 | large-scale-domain-specific-pretraining-for | 2303.00915 | null | https://arxiv.org/abs/2303.00915v1 | https://arxiv.org/pdf/2303.00915v1.pdf | Large-Scale Domain-Specific Pretraining for Biomedical Vision-Language Processing | Contrastive pretraining on parallel image-text data has attained great success in vision-language processing (VLP), as exemplified by CLIP and related methods. However, prior explorations tend to focus on general domains in the web. Biomedical images and text are rather different, but publicly available datasets are small and skew toward chest X-ray, thus severely limiting progress. In this paper, we conducted by far the largest study on biomedical VLP, using 15 million figure-caption pairs extracted from biomedical research articles in PubMed Central. Our dataset (PMC-15M) is two orders of magnitude larger than existing biomedical image-text datasets such as MIMIC-CXR, and spans a diverse range of biomedical images. The standard CLIP method is suboptimal for the biomedical domain. We propose BiomedCLIP with domain-specific adaptations tailored to biomedical VLP. We conducted extensive experiments and ablation studies on standard biomedical imaging tasks from retrieval to classification to visual question-answering (VQA). BiomedCLIP established new state of the art in a wide range of standard datasets, substantially outperformed prior VLP approaches. Surprisingly, BiomedCLIP even outperformed radiology-specific state-of-the-art models such as BioViL on radiology-specific tasks such as RSNA pneumonia detection, thus highlighting the utility in large-scale pretraining across all biomedical image types. We will release our models at https://aka.ms/biomedclip to facilitate future research in biomedical VLP. | ['Hoifung Poon', 'Tristan Naumann', 'Matthew P. Lungren', 'Cliff Wong', 'Naveen Valluri', 'Mu Wei', 'Rajesh Rao', 'Sam Preston', 'Robert Tinn', 'Jaspreet Bagga', 'Naoto Usuyama', 'Yanbo Xu', 'Sheng Zhang'] | 2023-03-02 | null | null | null | null | ['pneumonia-detection'] | ['medical'] | [ 4.57112968e-01 -1.26574878e-02 -4.52057689e-01 -1.34568870e-01
-1.58490992e+00 -6.30644381e-01 4.93022949e-01 3.79611671e-01
-7.08071589e-01 7.17126846e-01 5.08416295e-01 -6.95084572e-01
9.87922624e-02 -1.96091175e-01 -8.25830042e-01 -4.76779550e-01
1.03786148e-01 6.02523744e-01 5.80491535e-02 2.06210703e-01
6.43453971e-02 4.77677107e-01 -9.22124386e-01 8.12084198e-01
4.29259658e-01 7.53789842e-01 5.74305356e-01 9.99899089e-01
-1.38221517e-01 7.74765193e-01 -3.61767530e-01 -2.33534053e-01
-8.05021003e-02 -3.94400924e-01 -9.87278879e-01 -9.26129967e-02
9.13254440e-01 -1.72648638e-01 -5.35749197e-01 7.37293780e-01
1.13405597e+00 -5.68282187e-01 1.00214946e+00 -9.84320700e-01
-1.05902135e+00 3.67679447e-01 -8.48687589e-01 7.71721244e-01
2.59601742e-01 5.09655416e-01 8.36353123e-01 -1.11210632e+00
1.04875314e+00 9.47487175e-01 8.49343479e-01 7.90462375e-01
-1.11380064e+00 -5.94216049e-01 -3.35654378e-01 2.57670701e-01
-1.31310010e+00 -4.55753535e-01 2.72874385e-01 -5.79326153e-01
1.14374948e+00 3.79989833e-01 4.48732376e-01 1.41593707e+00
3.93631548e-01 1.05360055e+00 1.24559283e+00 -3.82297486e-01
-1.46485716e-01 -2.07700469e-02 1.05725482e-01 7.86677659e-01
1.95096135e-01 -3.15965116e-01 -4.52600658e-01 -4.13752049e-01
6.22452855e-01 -1.91828206e-01 -6.62279665e-01 -1.41453385e-01
-1.78817034e+00 6.57473624e-01 3.44928890e-01 3.89537752e-01
-3.09031516e-01 1.36820003e-01 5.54145813e-01 3.41703594e-02
2.84579247e-01 4.21567708e-01 -2.89097339e-01 1.78490236e-01
-1.20372391e+00 6.75209612e-02 6.31180882e-01 9.03154910e-01
1.90647900e-01 -3.52016032e-01 -6.87392652e-01 1.01468563e+00
3.02960634e-01 7.79468417e-01 6.14749432e-01 -9.02233660e-01
5.01191914e-01 1.28848359e-01 -4.84304041e-01 -6.49036109e-01
-7.04811096e-01 -4.92978752e-01 -1.07387245e+00 -1.73771292e-01
4.59023714e-01 -1.04298994e-01 -1.50690556e+00 1.43781805e+00
-2.94486266e-02 -3.15608792e-02 3.72741818e-02 9.05719995e-01
1.58546793e+00 5.09367466e-01 5.08025050e-01 -1.03661381e-01
1.80014026e+00 -8.60657752e-01 -5.85010052e-01 -3.48352820e-01
7.14865983e-01 -9.05187011e-01 1.23305762e+00 2.56187409e-01
-1.21651721e+00 -2.21869588e-01 -8.78547311e-01 -3.75999868e-01
-3.20011586e-01 2.60740966e-01 5.70158660e-01 4.06185210e-01
-1.39462996e+00 -5.38184047e-02 -6.33568525e-01 -7.10264504e-01
9.72594023e-01 1.31687805e-01 -4.91641611e-01 -5.07603288e-01
-1.06830943e+00 8.98214936e-01 1.27368331e-01 -6.11834265e-02
-1.08964288e+00 -1.31633008e+00 -6.39862895e-01 -3.71041924e-01
3.11300099e-01 -1.24515295e+00 1.29488289e+00 -5.05194247e-01
-7.20088959e-01 1.61497307e+00 -5.88247105e-02 -5.93889475e-01
6.52466297e-01 1.82919756e-01 -4.20698851e-01 7.95566857e-01
4.02435899e-01 1.30624700e+00 7.13225901e-01 -1.10392499e+00
-2.34269693e-01 -2.59541571e-01 -3.87044817e-01 3.65589380e-01
-1.26015499e-01 6.09041043e-02 -1.02649093e+00 -6.64835095e-01
-3.86571765e-01 -7.64802277e-01 -2.18163192e-01 5.50751209e-01
-4.53782678e-01 -1.88103765e-02 4.59247649e-01 -7.60371745e-01
8.73619974e-01 -1.97057378e+00 -2.63336420e-01 -2.49822125e-01
5.64350724e-01 2.23211333e-01 -4.40677762e-01 1.99044704e-01
-3.00222367e-01 2.48641059e-01 -3.29958469e-01 -2.38633707e-01
-3.80593181e-01 8.62986743e-02 -7.31710196e-02 7.60260046e-01
1.71809390e-01 1.45914888e+00 -7.53413141e-01 -1.27167463e+00
7.45258108e-02 3.96126181e-01 -4.78689879e-01 1.53710237e-02
-1.70145527e-01 4.83251780e-01 -3.77071410e-01 9.94647443e-01
3.96763355e-01 -8.73876989e-01 -1.29226476e-01 -5.10609329e-01
1.57223493e-01 -1.98378339e-01 -1.90346718e-01 2.07930589e+00
-3.24219406e-01 7.50134885e-01 2.86831707e-01 -1.08628523e+00
9.92691889e-02 5.61939895e-01 8.03518176e-01 -8.27951193e-01
-7.04104379e-02 -1.01010213e-02 1.43587142e-02 -9.12182093e-01
-3.66035439e-02 -1.24513656e-01 2.67865360e-01 7.56915584e-02
9.89520624e-02 -2.79349536e-01 3.01788807e-01 5.57637751e-01
1.34413469e+00 -1.46660298e-01 2.62799144e-01 -1.90353781e-01
4.96477634e-01 5.46606660e-01 6.01366572e-02 1.25274265e+00
-4.75926101e-01 1.16667342e+00 4.50023085e-01 9.93138775e-02
-1.00284815e+00 -1.30911791e+00 -7.13759422e-01 8.28730941e-01
-3.73659939e-01 -3.76062840e-01 -5.15380621e-01 -7.87232161e-01
-5.97647093e-02 3.74245673e-01 -7.21714735e-01 4.01101857e-01
-5.28013349e-01 -1.10098386e+00 9.95350242e-01 4.38612491e-01
2.39573464e-01 -1.25400555e+00 -5.34545183e-01 1.30768895e-01
-3.49671155e-01 -1.60488212e+00 -5.92172384e-01 -1.37806283e-02
-7.51607656e-01 -1.17928743e+00 -1.69843435e+00 -9.93169785e-01
5.52094102e-01 1.38775438e-01 1.55786276e+00 -1.27268299e-01
-1.16547573e+00 1.00781798e+00 -8.58028904e-02 -7.53083408e-01
-5.27666926e-01 -2.03906018e-02 -4.94135141e-01 -6.04614556e-01
1.29330307e-01 -8.50898400e-03 -1.04291463e+00 -1.37284219e-01
-8.59891593e-01 3.61026615e-01 1.17398691e+00 1.05178344e+00
9.40245688e-01 -6.54162109e-01 4.93705332e-01 -1.10153627e+00
6.47413909e-01 -5.57558060e-01 2.00997368e-02 5.09592593e-01
-4.64108467e-01 -4.35826361e-01 2.63289690e-01 -2.98336565e-01
-8.58214438e-01 -2.35507682e-01 -3.70821387e-01 -4.40554678e-01
-3.89792353e-01 7.85672605e-01 4.59019333e-01 -3.54001559e-02
8.23901057e-01 3.81538868e-01 7.46275038e-02 -2.01041445e-01
3.81178200e-01 5.81507504e-01 8.49403501e-01 -4.06298459e-01
2.80185431e-01 5.81054032e-01 1.24172352e-01 -1.14435327e+00
-7.28333712e-01 -6.74038708e-01 -2.52868295e-01 -1.43172368e-01
1.20886064e+00 -1.07247484e+00 -4.58242983e-01 2.03626066e-01
-1.02282023e+00 -3.81210744e-01 -1.15280174e-01 3.62870097e-01
-4.58266646e-01 5.90317607e-01 -9.03948665e-01 -7.85811841e-02
-8.05330932e-01 -1.29118931e+00 1.34979773e+00 -1.50615558e-01
-2.45382234e-01 -8.93963814e-01 1.23593949e-01 6.78061724e-01
2.47865632e-01 1.74944639e-01 1.24057817e+00 -4.30178493e-01
-3.36515695e-01 8.91831666e-02 -7.44331062e-01 1.72103152e-01
-3.06387171e-02 -2.19654173e-01 -8.35252047e-01 -3.34049940e-01
-3.47803712e-01 -6.69038832e-01 1.15728581e+00 8.47293556e-01
1.62872458e+00 2.30261371e-01 -6.74672782e-01 6.77219450e-01
1.39316738e+00 -1.07884876e-01 5.22029340e-01 1.94531307e-01
7.92838752e-01 3.44117016e-01 3.99699509e-01 5.63094243e-02
3.48578364e-01 3.07230085e-01 2.39276901e-01 -7.73869157e-01
-7.49213636e-01 1.17580958e-01 -5.86710498e-02 6.59210622e-01
1.89861521e-01 -3.59088868e-01 -1.63383341e+00 8.61884117e-01
-1.42465544e+00 -4.73936468e-01 -1.14193717e-02 1.57372999e+00
1.15464854e+00 -8.73365104e-02 -1.78364947e-01 -5.48398435e-01
3.91795784e-01 2.13574901e-01 -5.08253634e-01 1.76425487e-01
-1.34486064e-01 3.42328191e-01 7.07753122e-01 1.62087560e-01
-1.22271121e+00 6.87349200e-01 6.74219656e+00 9.78490531e-01
-1.17140818e+00 4.51199412e-01 9.86587048e-01 -2.97228158e-01
-1.39462411e-01 -5.55558085e-01 -6.36676371e-01 1.66171819e-01
8.17227721e-01 5.98505232e-03 -1.20826356e-01 4.97446358e-01
4.87975813e-02 -2.05807999e-01 -1.28808749e+00 1.37921238e+00
3.38302255e-01 -1.91270840e+00 3.43645096e-01 -5.76820634e-02
4.47514087e-01 7.16176748e-01 4.39875782e-01 2.75544524e-01
-1.61188856e-01 -1.46733844e+00 1.81900144e-01 5.59820116e-01
1.53834999e+00 -1.27799377e-01 6.60199881e-01 -5.60046397e-02
-6.70489192e-01 4.23653841e-01 -3.28588694e-01 8.22794795e-01
8.22899714e-02 4.54960197e-01 -1.17926252e+00 4.88082021e-01
9.60435808e-01 8.48683298e-01 -8.93222332e-01 1.09722030e+00
2.60942668e-01 8.63196433e-01 -7.45118856e-02 1.02223277e-01
4.62314159e-01 4.24696326e-01 5.24652660e-01 1.90745807e+00
-2.31655017e-02 2.68180937e-01 7.03909248e-02 5.74553430e-01
-3.83892834e-01 3.95647317e-01 -6.16474092e-01 -2.42813721e-01
3.29998583e-02 1.37037480e+00 -6.91697001e-01 -5.64891219e-01
-8.53620172e-01 6.77564800e-01 1.93958074e-01 5.39384723e-01
-8.39280844e-01 -3.16244387e-03 1.14357144e-01 2.20521271e-01
1.42269105e-01 6.42817244e-02 -1.96154043e-01 -1.01281464e+00
-2.26792797e-01 -1.05078769e+00 6.91214800e-01 -1.08122957e+00
-1.65507507e+00 5.51970005e-01 1.02479860e-01 -1.09735954e+00
1.13533372e-02 -9.17728722e-01 -2.30126411e-01 8.36211622e-01
-1.79275322e+00 -1.28154027e+00 -3.13333809e-01 9.33674932e-01
6.28854811e-01 -2.07190350e-01 8.99048030e-01 4.67309922e-01
-2.48115867e-01 5.91991961e-01 5.72742857e-02 2.21055642e-01
1.32941675e+00 -1.22392392e+00 4.12865356e-02 3.63157153e-01
8.41107517e-02 4.71842736e-01 4.89560962e-01 -5.96926928e-01
-1.59489822e+00 -1.11540198e+00 4.38412398e-01 -6.26705647e-01
7.75500655e-01 -9.37160179e-02 -7.89862692e-01 6.13279223e-01
6.17596328e-01 2.78185248e-01 8.85708570e-01 -3.03676099e-01
-1.71998680e-01 2.10542858e-01 -1.02782178e+00 7.83188224e-01
9.44643736e-01 -6.45152867e-01 -6.74774647e-01 8.94165874e-01
5.31375289e-01 -5.91996551e-01 -1.11316025e+00 6.33751750e-01
3.99198890e-01 -4.32426840e-01 1.38458812e+00 -5.82953095e-01
7.10350752e-01 6.06252588e-02 9.23535042e-03 -8.60182941e-01
-3.88861775e-01 -1.29909799e-01 2.99721900e-02 7.58054554e-01
5.43757498e-01 -4.45527166e-01 9.38683033e-01 3.41112427e-02
-2.68145591e-01 -1.00574255e+00 -9.02675927e-01 -2.40459427e-01
5.64083517e-01 -5.16523898e-01 -2.80008256e-01 9.77708757e-01
-2.17674047e-01 4.11152005e-01 5.74223734e-02 -9.23404098e-02
7.81638443e-01 -7.34319910e-02 4.31860685e-01 -6.76139414e-01
-4.63404596e-01 -6.48211479e-01 -4.04500328e-02 -7.78814614e-01
6.77734762e-02 -1.47401822e+00 -1.68918833e-01 -1.99997580e+00
7.87377894e-01 -2.96374381e-01 -4.24972534e-01 5.80555797e-01
-2.72809327e-01 8.11122596e-01 9.09635574e-02 3.10256839e-01
-5.22322476e-01 -6.55751210e-03 1.68526602e+00 -6.92668021e-01
3.57927322e-01 -4.80695367e-01 -7.48887062e-01 6.03871644e-01
6.82441711e-01 -2.89997905e-01 -4.32266980e-01 -4.50424612e-01
-8.19779485e-02 3.66018176e-01 6.76889241e-01 -1.06789410e+00
2.29460105e-01 1.89398751e-01 8.79447520e-01 -9.16132569e-01
1.96414679e-01 -3.84879023e-01 -2.67755300e-01 4.37013030e-01
-4.58661109e-01 4.43657786e-02 5.69214582e-01 4.65981454e-01
-1.29933149e-01 2.57364716e-02 7.33418763e-01 -5.31376302e-01
-7.02633381e-01 5.94335496e-01 -5.83662152e-01 6.41796052e-01
7.40080059e-01 -9.52969268e-02 -5.57656467e-01 -2.04306826e-01
-8.88692796e-01 4.89848971e-01 1.40431643e-01 2.93892682e-01
8.51223290e-01 -8.37305367e-01 -1.27378094e+00 -3.04476231e-01
4.99984235e-01 -1.62024703e-02 5.77582717e-01 1.34863627e+00
-7.63790846e-01 8.88494432e-01 -1.15855642e-01 -1.11265004e+00
-1.43978727e+00 6.80905759e-01 2.41683707e-01 -5.70954978e-01
-9.50358272e-01 8.00128102e-01 7.27218568e-01 -2.22247139e-01
2.74421811e-01 -2.93370575e-01 -1.25020400e-01 -7.18881115e-02
5.08623660e-01 -2.10870907e-01 1.87436566e-01 -4.17602181e-01
-6.69251323e-01 5.65038443e-01 -5.46518087e-01 -1.59698606e-01
1.08818781e+00 5.15703112e-02 1.27497585e-02 1.97462484e-01
1.44901001e+00 -2.44801089e-01 -7.80039668e-01 -1.83892235e-01
-2.72129864e-01 1.27144739e-01 1.60714418e-01 -1.22798288e+00
-9.07054722e-01 1.00729918e+00 7.58032858e-01 -4.14022654e-01
1.08690035e+00 4.65347528e-01 5.81039548e-01 2.41961941e-01
1.14481233e-01 -7.97331929e-01 1.54279053e-01 2.62221962e-01
1.13749957e+00 -1.46786535e+00 4.51699525e-01 -2.88208067e-01
-8.47561955e-01 8.51561606e-01 4.62251276e-01 3.22992563e-01
6.24013603e-01 3.92768532e-01 2.52076209e-01 -4.38499838e-01
-7.02606797e-01 -7.88912848e-02 7.23641694e-01 7.68719673e-01
7.02402472e-01 -6.76334724e-02 -3.64208609e-01 2.00769916e-01
8.89454558e-02 2.05736399e-01 3.33725274e-01 9.33968425e-01
-3.56390770e-03 -7.81251669e-01 -3.60888541e-01 7.77391672e-01
-1.01907575e+00 -5.80902994e-01 -1.45567283e-01 1.04880679e+00
-2.55650561e-02 5.43164134e-01 -9.92318243e-02 4.10137504e-01
1.19462177e-01 -1.10432290e-01 8.15348983e-01 -6.75638318e-01
-5.10339022e-01 2.44438678e-01 1.57460228e-01 -6.44151092e-01
-4.46121126e-01 -6.14673913e-01 -1.26276422e+00 1.39623120e-01
2.78436244e-01 -3.13969463e-01 5.00396252e-01 6.71002567e-01
2.63703674e-01 7.77821958e-01 -2.59484351e-01 -5.48545182e-01
-1.60604522e-01 -7.89275169e-01 -2.61070132e-01 4.17395115e-01
4.06641066e-01 -3.03113014e-01 -5.25177596e-03 2.31827840e-01] | [14.98039436340332, -1.8184908628463745] |
40466585-e527-4163-9aa0-648ecd925c0b | virtual-bike-emulation-in-a-series-parallel | 2305.05569 | null | https://arxiv.org/abs/2305.05569v1 | https://arxiv.org/pdf/2305.05569v1.pdf | Virtual-bike emulation in a series-parallel human-powered electric bike | Combining the advantages of standard bicycles and electrified vehicles, electric bikes (e-Bikes) are promising vehicles to reduce emission and traffic. The current literature on e-Bikes ranges from works on the energy management to the vehicle control to properly govern the human-vehicle interaction. This last point is fundamental in chain-less series bikes, where the link between the human and the vehicle behavior is only given by a control law. In this work, we address this problem in a series-parallel bike. In particular, we provide an extension of the virtual-chain concept, born for series bikes, and then we improve it developing a virtual-bike framework. Experimental results are used to validate the effectiveness of the solutions, when the cyclist is actually riding the bike. | ['Sergio M. Savaresi', 'Giulio Panzani', 'Stefano Radrizzani'] | 2023-05-09 | null | null | null | null | ['energy-management'] | ['time-series'] | [-1.95153534e-01 1.25219837e-01 -5.90299964e-01 1.28626004e-01
3.89268786e-01 -3.49447280e-01 6.76943779e-01 -6.34788930e-01
-3.48024182e-02 9.07725453e-01 -5.46273589e-01 -5.49171746e-01
-7.14694381e-01 -1.09540725e+00 -6.07708752e-01 -8.92198384e-01
1.82728887e-01 3.32090527e-01 5.45947015e-01 -7.80002475e-01
6.01835782e-03 7.60648787e-01 -1.87265325e+00 -5.46515703e-01
9.58559513e-01 6.52651012e-01 2.18385771e-01 5.55436730e-01
1.42977670e-01 4.22631711e-01 5.46707818e-03 -2.13360339e-01
-2.51057029e-01 -6.93505049e-01 -3.30574244e-01 -2.13603720e-01
-4.55612272e-01 1.63783953e-01 -2.34044597e-01 5.21846056e-01
8.46677348e-02 4.34080094e-01 8.79558563e-01 -2.17416334e+00
1.16823487e-01 2.09837243e-01 -8.56037512e-02 1.12600639e-01
-2.12400004e-01 1.20666780e-01 3.71401638e-01 -4.95523363e-01
4.03275371e-01 9.59418476e-01 4.80958849e-01 5.82042754e-01
-1.05466640e+00 -8.02311540e-01 -1.21616639e-01 1.05482793e+00
-1.47121012e+00 -3.90705258e-01 9.52701271e-01 -3.36666316e-01
1.19738102e+00 4.95007426e-01 1.29550266e+00 4.96225566e-01
4.97355431e-01 4.59582418e-01 7.77984619e-01 -3.97100538e-01
9.42314640e-02 4.06956017e-01 5.26406765e-01 8.66764560e-02
5.74880958e-01 3.04973662e-01 1.32598411e-02 3.49991828e-01
-7.55456388e-02 -3.41847211e-01 4.33789939e-02 -7.84691930e-01
-4.89395618e-01 5.52300692e-01 -1.85562689e-02 6.82996690e-01
-5.02346873e-01 4.12386715e-01 4.73001748e-01 3.93375121e-02
-5.39792627e-02 -1.62610844e-01 1.55014411e-01 -2.83924222e-01
-7.47729421e-01 5.07048130e-01 9.19789374e-01 1.03586352e+00
7.06176639e-01 9.17608663e-02 1.57859232e-02 4.09751296e-01
2.26330891e-01 7.57109642e-01 -2.27095038e-01 -7.18875229e-01
3.77525799e-02 4.68755543e-01 2.80857801e-01 -7.29430079e-01
-4.88505512e-01 -1.21011928e-01 -7.37446427e-01 5.63369058e-02
8.45822692e-02 -2.25379661e-01 -2.38051817e-01 1.34177232e+00
4.19655710e-01 7.46079162e-02 -2.87135810e-01 5.26778817e-01
3.93509060e-01 8.90189826e-01 1.67084798e-01 -6.17658377e-01
1.01991189e+00 -8.56988251e-01 -1.31485438e+00 2.88398057e-01
4.99498576e-01 -1.31253093e-01 5.36098063e-01 3.77450675e-01
-1.25853360e+00 -5.65446079e-01 -1.29454923e+00 2.02817395e-01
-7.93012977e-01 -2.33641379e-02 -2.33737826e-01 1.23231459e+00
-1.08140171e+00 1.21481337e-01 -7.04907656e-01 -2.66815692e-01
-4.98275049e-02 3.57197583e-01 1.47100374e-01 3.71620029e-01
-1.51422346e+00 1.44797587e+00 1.59477115e-01 2.12089211e-01
-5.03927350e-01 -7.52738893e-01 -4.76292461e-01 1.18337594e-01
5.13201475e-01 -5.51938236e-01 1.15648723e+00 -3.10899854e-01
-1.82538140e+00 2.87864149e-01 -2.24836409e-01 -3.74519795e-01
7.30120480e-01 1.22264154e-01 -6.74937963e-01 8.42226893e-02
-1.95854515e-01 1.21043384e-01 4.00742084e-01 -1.30927730e+00
-5.99959612e-01 -1.04972228e-01 -4.19533283e-01 -1.19358696e-01
-1.92352623e-01 -2.64076412e-01 -1.79701850e-01 1.83054760e-01
-7.16002882e-01 -1.38014531e+00 1.58457190e-01 -4.76973891e-01
-4.38582510e-01 -1.07061625e+00 1.55699539e+00 -1.82214290e-01
1.45679581e+00 -1.80704772e+00 1.57753065e-01 6.61858201e-01
-3.53592247e-01 4.49774444e-01 4.37019408e-01 8.25214088e-01
2.20418274e-02 -7.29358569e-02 6.70662224e-02 -4.03135084e-02
1.41056731e-01 3.98269653e-01 1.04710180e-02 3.69301736e-01
-7.57925063e-02 1.09911466e+00 -6.92424297e-01 -3.48810196e-01
6.32062674e-01 5.54773450e-01 -2.91263849e-01 -2.21960679e-01
2.31372550e-01 3.18735689e-02 -4.48040307e-01 -8.38561580e-02
8.77775908e-01 4.72843140e-01 3.57807636e-01 -2.21981555e-01
-8.49748254e-01 -8.69296398e-03 -1.07152939e+00 7.99670160e-01
-7.48803794e-01 7.09900081e-01 4.64851618e-01 -1.44506943e+00
8.80787432e-01 2.95272470e-01 7.08726823e-01 -1.14107502e+00
1.98356494e-01 3.20136607e-01 -1.14710778e-02 -7.82169878e-01
7.10435033e-01 -9.03427452e-02 4.05767597e-02 2.77313411e-01
-2.64690667e-01 -4.24500793e-01 7.27269650e-01 1.42469583e-02
4.80788857e-01 1.50495753e-01 8.97900164e-02 -7.18849421e-01
8.66501331e-01 1.97726861e-01 2.40927130e-01 7.91295692e-02
-7.73254409e-02 -6.19752109e-01 4.49923903e-01 2.96446532e-01
-1.26345181e+00 -8.46415102e-01 -1.94442496e-01 5.38219452e-01
8.45594466e-01 -2.78488934e-01 -9.36996341e-01 -1.59087077e-01
1.18557913e-02 1.23694086e+00 -9.23377648e-02 -4.90709066e-01
-8.14787686e-01 -5.51811576e-01 2.01937363e-01 3.57675731e-01
5.12156606e-01 -4.50262159e-01 -8.68452013e-01 3.78773242e-01
-2.80377239e-01 -8.63728464e-01 -2.98770517e-01 6.18429109e-02
-3.35088968e-01 -1.18880022e+00 -2.33077586e-01 -6.11150742e-01
6.83860928e-02 4.93684143e-01 8.50686848e-01 2.59979963e-01
1.40605852e-01 4.43794191e-01 -1.76430374e-01 -6.59908652e-01
-5.94162703e-01 7.75742307e-02 8.25781524e-02 5.06651513e-02
5.41239321e-01 -5.37692845e-01 -6.03661537e-01 9.98519063e-01
-5.31195462e-01 2.46000051e-01 1.57685503e-02 4.68245983e-01
4.72289622e-01 5.76352298e-01 1.12431145e+00 -1.55542985e-01
4.12688285e-01 -6.45508111e-01 -6.25429094e-01 5.72475940e-02
-8.25041771e-01 -3.17468256e-01 5.35425067e-01 -1.62270635e-01
-1.19717050e+00 6.82150275e-02 -3.78429294e-01 1.11801378e-01
-2.85704229e-02 -1.74835160e-01 -8.66810083e-01 -7.81826973e-02
-4.47319485e-02 3.19058269e-01 6.90607727e-02 -2.57970870e-01
3.86902630e-01 7.25232244e-01 4.60269570e-01 -1.11658312e-01
6.65181696e-01 3.56761336e-01 5.72877645e-01 -1.12945414e+00
6.48599327e-01 -3.38205129e-01 -3.36237222e-01 -9.97165918e-01
9.12094772e-01 -2.99163729e-01 -1.52541852e+00 6.78379059e-01
-1.07668185e+00 -6.73785448e-01 -1.75123334e-01 5.56137919e-01
-8.29216957e-01 6.62630126e-02 -2.00022850e-02 -1.18598449e+00
-6.17982680e-03 -1.11506712e+00 5.55240750e-01 4.17492777e-01
-9.17985514e-02 -8.95350218e-01 1.99354768e-01 1.40266851e-01
4.77323502e-01 -1.69013832e-02 9.72653866e-01 2.36083902e-02
-7.01099217e-01 -1.64947491e-02 1.78115353e-01 1.45947859e-01
-1.91862866e-01 1.86567828e-01 -5.71361721e-01 -1.82375763e-04
-2.51453787e-01 3.92171949e-01 3.88820350e-01 4.68468070e-01
8.29372704e-01 -4.02229950e-02 -1.19957256e+00 9.07167643e-02
1.54475880e+00 8.08510900e-01 9.87153828e-01 1.65109411e-01
3.55155170e-01 7.43477046e-01 7.06566274e-01 -8.72482359e-03
7.44059265e-01 1.26385558e+00 5.15441358e-01 1.23144031e-01
-1.87493265e-01 -3.64021063e-01 1.20388821e-01 7.51986623e-01
-3.03751141e-01 -6.45924032e-01 -5.85857391e-01 8.51307094e-01
-1.67978299e+00 -1.10514176e+00 -1.09505069e+00 1.86016285e+00
3.91392738e-01 -5.10697924e-02 4.83483046e-01 7.55140245e-01
8.51667166e-01 -4.91546571e-01 -1.95785478e-01 -8.04095745e-01
2.31973469e-01 -5.97453490e-02 7.52398014e-01 5.23067236e-01
-4.63099241e-01 2.38808721e-01 6.32696199e+00 1.27898228e+00
-1.13394558e+00 3.59681427e-01 4.40240726e-02 1.26085430e-01
-5.18312931e-01 -1.24073327e-01 -9.51537073e-01 9.08459842e-01
1.36428463e+00 -6.56144202e-01 3.03733259e-01 4.84011561e-01
9.82187331e-01 -4.99873579e-01 -9.55241561e-01 6.99574351e-01
-2.18130216e-01 -9.96865034e-01 -5.25296450e-01 1.96749538e-01
5.08762181e-01 -3.36193472e-01 -4.42163497e-01 2.59776205e-01
-2.71886081e-01 -7.08610654e-01 7.89910316e-01 7.19922423e-01
5.06135523e-01 -1.26688075e+00 4.41781431e-01 7.74234474e-01
-1.59315634e+00 -7.59822130e-02 3.71095866e-01 -4.61594835e-02
9.16972935e-01 4.39210564e-01 -3.67875904e-01 7.91873276e-01
4.81364399e-01 4.10558909e-01 8.48569125e-02 1.07295573e+00
2.10065674e-02 4.84734267e-01 -3.95829260e-01 -7.57924139e-01
-2.69192383e-02 -7.25521386e-01 6.96805358e-01 1.01109898e+00
5.65897226e-01 -2.16993801e-02 -3.83477271e-01 8.76420081e-01
4.42421198e-01 -2.30595134e-02 -9.18047190e-01 4.70772535e-01
3.91019464e-01 1.21750808e+00 -7.16835976e-01 -3.82620901e-01
-3.01358998e-01 2.91868418e-01 -6.32826924e-01 3.78667921e-01
-1.21241033e+00 -9.05080795e-01 7.49184072e-01 6.64913356e-01
2.27749407e-01 -3.55465233e-01 -2.30277926e-01 -2.59725451e-01
-7.16679543e-02 1.26699090e-01 -2.53500402e-01 -8.12464416e-01
-5.91811478e-01 9.86865684e-02 8.68614793e-01 -1.13300562e+00
-5.19670308e-01 -4.68878210e-01 -6.70391381e-01 6.48957968e-01
-1.57497501e+00 -1.01166773e+00 -3.35590631e-01 4.54881340e-01
2.54645735e-01 2.57692218e-01 2.10667834e-01 9.96604681e-01
-5.17130077e-01 3.77641797e-01 4.43748057e-01 -6.71148896e-01
1.27779990e-02 -6.75529718e-01 1.97116137e-02 6.69904709e-01
-7.76178360e-01 -4.27693874e-03 9.82761264e-01 -5.28524995e-01
-1.76812553e+00 -7.05393195e-01 1.14151847e+00 -2.77620465e-01
5.14288604e-01 -3.46239954e-01 -6.44620121e-01 1.50997743e-01
4.71129596e-01 -4.84918594e-01 9.92652178e-02 -6.29264295e-01
8.59886050e-01 -4.53944385e-01 -1.02171707e+00 4.10325378e-01
1.19845092e+00 -2.79882133e-01 -6.49831817e-02 -8.09836537e-02
1.24470726e-01 1.85950235e-01 -5.16305685e-01 2.68970251e-01
6.25222802e-01 -7.41156399e-01 8.96640122e-01 8.73974338e-03
-2.62166381e-01 -4.09619153e-01 5.86856663e-01 -1.75204611e+00
-3.16009633e-02 -6.82284713e-01 -9.90490690e-02 1.17377245e+00
-2.61502881e-02 -6.53757513e-01 4.21654165e-01 3.66027087e-01
-4.64179486e-01 -5.15075326e-01 -1.29527724e+00 -1.14692020e+00
2.09395856e-01 -6.17380381e-01 9.10522461e-01 3.49526495e-01
4.93088692e-01 2.45948002e-01 -4.42911774e-01 -2.03748807e-01
5.91879487e-01 -2.49289677e-01 7.93390095e-01 -1.10932946e+00
2.43901893e-01 -7.39095271e-01 -3.04855496e-01 -9.80319321e-01
3.73334527e-01 -9.23147380e-01 2.12538153e-01 -1.69450867e+00
9.97660682e-03 -4.69404012e-01 2.70879060e-01 -1.69644281e-01
3.78423899e-01 2.89013684e-01 -1.88819971e-02 -3.69337380e-01
-2.64355779e-01 6.39793754e-01 1.16932738e+00 -1.28361836e-01
-4.39040959e-01 3.82180959e-01 1.53660879e-01 3.98396403e-01
8.27335060e-01 -2.90146261e-01 -8.28734815e-01 1.92279354e-01
2.77353287e-01 1.57727197e-01 5.95049202e-01 -8.44901264e-01
5.08368909e-01 -3.78095120e-01 -6.50607646e-01 -1.12435341e+00
2.59392768e-01 -1.36895514e+00 8.82627547e-01 7.59707630e-01
3.34224492e-01 -1.03992552e-01 9.53675136e-02 4.86260533e-01
-6.46925718e-02 -3.85921076e-02 8.67941678e-01 6.60865664e-01
-6.85312331e-01 -1.09088019e-01 -1.07142746e+00 -4.98513430e-01
1.85544932e+00 -5.57527661e-01 -3.12418252e-01 -1.40714943e-01
-6.08109832e-01 7.01469362e-01 1.24641083e-01 4.11316186e-01
1.67281121e-01 -1.49986362e+00 -2.54242301e-01 1.70271277e-01
-1.33622974e-01 -6.93186641e-01 7.06455410e-01 1.26835513e+00
-2.70116001e-01 8.93080115e-01 -5.13191462e-01 -5.57341099e-01
-1.29346669e+00 5.64534307e-01 7.78534889e-01 1.00784600e-01
-3.38255674e-01 -1.08665675e-01 -2.03315184e-01 3.63616832e-02
-2.22203940e-01 -3.18537891e-01 -4.69241619e-01 1.92220181e-01
1.64462075e-01 1.42686403e+00 4.13098246e-01 -9.17363405e-01
-6.18638456e-01 6.94941938e-01 1.04116344e+00 -7.80952871e-02
1.07456672e+00 -5.74854076e-01 -3.10122948e-02 2.77868271e-01
9.83556569e-01 -7.22339898e-02 -8.13921928e-01 4.76288527e-01
-2.61760831e-01 -1.04198664e-01 2.23913476e-01 -3.00814688e-01
-1.00233912e+00 7.24803507e-01 6.25614226e-01 7.69900918e-01
1.06214511e+00 -8.37137550e-02 9.24489081e-01 2.28724018e-01
4.66098487e-01 -1.61423123e+00 -3.37683827e-01 3.32669199e-01
5.15904486e-01 -4.44194525e-01 -3.61439586e-01 -5.14038622e-01
-1.93513438e-01 8.74765456e-01 5.67718744e-01 -3.50290239e-02
1.23143029e+00 5.84798455e-01 -6.55547619e-01 6.12334758e-02
-6.59206688e-01 -2.77531594e-01 -3.26504260e-02 5.55788398e-01
-6.60297722e-02 3.67887676e-01 -1.16560996e+00 7.73893952e-01
4.26804908e-02 6.68394506e-01 5.16249001e-01 6.68002903e-01
-4.27565694e-01 -1.25056851e+00 -4.10438061e-01 1.29542738e-01
5.66462837e-02 7.56160796e-01 5.90733327e-02 1.20114136e+00
5.45994580e-01 1.53752530e+00 2.85349190e-01 -3.08541209e-01
9.64353502e-01 2.48707905e-01 2.76177943e-01 2.67817378e-01
-2.22281232e-01 -3.25876653e-01 3.98683220e-01 -3.18819404e-01
-7.58711278e-01 -5.76457262e-01 -1.43940914e+00 -8.82048070e-01
-5.27714610e-01 6.56799853e-01 1.15334213e+00 9.36862290e-01
1.31949335e-01 6.93670928e-01 8.22817743e-01 -9.19436157e-01
7.71606788e-02 -7.39451647e-01 -8.12484562e-01 -4.18399498e-02
1.86942041e-01 -1.05304265e+00 -1.36842176e-01 -2.14017630e-01] | [5.595942974090576, 2.100679397583008] |
87238330-a8e5-440a-9216-b4de491b827f | distributional-reinforcement-learning-for-4 | 2110.13578 | null | https://arxiv.org/abs/2110.13578v1 | https://arxiv.org/pdf/2110.13578v1.pdf | Distributional Reinforcement Learning for Multi-Dimensional Reward Functions | A growing trend for value-based reinforcement learning (RL) algorithms is to capture more information than scalar value functions in the value network. One of the most well-known methods in this branch is distributional RL, which models return distribution instead of scalar value. In another line of work, hybrid reward architectures (HRA) in RL have studied to model source-specific value functions for each source of reward, which is also shown to be beneficial in performance. To fully inherit the benefits of distributional RL and hybrid reward architectures, we introduce Multi-Dimensional Distributional DQN (MD3QN), which extends distributional RL to model the joint return distribution from multiple reward sources. As a by-product of joint distribution modeling, MD3QN can capture not only the randomness in returns for each source of reward, but also the rich reward correlation between the randomness of different sources. We prove the convergence for the joint distributional Bellman operator and build our empirical algorithm by minimizing the Maximum Mean Discrepancy between joint return distribution and its Bellman target. In experiments, our method accurately models the joint return distribution in environments with richly correlated reward functions, and outperforms previous RL methods utilizing multi-dimensional reward functions in the control setting. | ['Tie-Yan Liu', 'Tao Qin', 'Wei Xiong', 'Li Zhao', 'Xiaoyu Chen', 'Pushi Zhang'] | 2021-10-26 | null | http://proceedings.neurips.cc/paper/2021/hash/0b9e57c46de934cee33b0e8d1839bfc2-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/0b9e57c46de934cee33b0e8d1839bfc2-Paper.pdf | neurips-2021-12 | ['distributional-reinforcement-learning'] | ['methodology'] | [-6.15680277e-01 -1.44339055e-01 -5.26651204e-01 -8.71506780e-02
-7.72368014e-01 -4.88979459e-01 4.31344122e-01 2.09038213e-01
-5.24607122e-01 1.06024015e+00 2.92884678e-01 5.85506074e-02
-6.47370815e-01 -9.14772451e-01 -5.30759931e-01 -9.17156577e-01
-5.80689490e-01 5.22203445e-01 -1.11431256e-01 -7.18340933e-01
2.22903773e-01 3.64730567e-01 -1.15105271e+00 -3.89512330e-01
7.03723967e-01 1.29350507e+00 3.09438616e-01 6.05287671e-01
-1.68887317e-01 1.28193820e+00 -9.63759542e-01 -1.33984923e-01
3.38924319e-01 -5.57909608e-01 -2.06033006e-01 -5.04124403e-01
-6.30335271e-01 -6.15839422e-01 -3.03326339e-01 9.42692637e-01
7.55693376e-01 2.39240736e-01 7.09198475e-01 -1.57559621e+00
-7.13153422e-01 9.92438674e-01 -7.39157736e-01 1.21075856e-02
-1.02387041e-01 3.77577126e-01 1.35180461e+00 -1.18512295e-01
1.29638493e-01 1.50719571e+00 2.15654746e-01 6.47208452e-01
-1.33682919e+00 -4.11459684e-01 1.20086096e-01 -1.45061940e-01
-7.84392834e-01 2.35983863e-01 7.30883956e-01 -2.29055688e-01
7.10896492e-01 -2.19279870e-01 6.98250532e-01 1.03678966e+00
4.83277291e-01 1.24947321e+00 1.27491760e+00 1.28189409e-02
6.21999085e-01 -1.98287982e-02 -2.24692792e-01 4.45332043e-02
1.10881917e-01 8.04958940e-01 -2.36446157e-01 -1.97643772e-01
1.09785426e+00 1.99339688e-01 8.66676718e-02 -7.16313839e-01
-9.05407786e-01 1.12478340e+00 5.56144357e-01 1.13096751e-01
-7.83773065e-01 8.80384505e-01 3.67186397e-01 5.86216509e-01
2.88796276e-01 4.83366370e-01 -3.99122328e-01 -6.64246619e-01
-5.06360292e-01 7.31988370e-01 6.94611609e-01 9.94878232e-01
5.54970920e-01 5.23692250e-01 -8.30910206e-01 6.86960697e-01
4.56437558e-01 9.48500156e-01 4.89710391e-01 -1.29026520e+00
6.00074410e-01 2.88686395e-01 5.51508963e-01 -5.17230392e-01
-4.17159200e-01 -8.69372785e-01 -6.41610682e-01 4.48616028e-01
5.60044348e-01 -5.18170178e-01 -2.15598166e-01 2.10782814e+00
-6.46275580e-02 -2.94270009e-01 3.92083228e-01 8.53613555e-01
3.36892083e-02 4.17738110e-01 -1.91979110e-01 -3.40157926e-01
6.34967387e-01 -6.33476079e-01 -5.57572842e-01 2.94601470e-01
6.43697381e-01 -2.16070950e-01 1.12263739e+00 3.16317707e-01
-1.19222176e+00 -2.12533250e-01 -7.34070837e-01 5.72003245e-01
1.92042425e-01 -2.80190974e-01 3.82149130e-01 5.22563040e-01
-9.18179333e-01 9.98137653e-01 -5.00730038e-01 2.80736953e-01
4.96993750e-01 2.50869572e-01 4.79786724e-01 -4.44367714e-02
-1.34736860e+00 7.92525411e-01 1.17613245e-02 -1.83052897e-01
-1.36262929e+00 -8.08169901e-01 -5.47214925e-01 2.52549261e-01
7.55774677e-01 -2.52569139e-01 1.54038155e+00 -8.60460222e-01
-1.97340715e+00 -1.42556176e-01 5.14442861e-01 -8.04908633e-01
5.95818281e-01 -1.55557171e-01 7.01660216e-02 -5.10542057e-02
8.86263978e-03 4.27190363e-01 9.21284676e-01 -1.07663286e+00
-4.94322598e-01 -2.92644501e-01 1.35176674e-01 2.28655770e-01
-4.58001867e-02 -4.05407339e-01 7.13338137e-01 -4.39551741e-01
-9.10802364e-01 -6.72560453e-01 -4.80214834e-01 -4.33880717e-01
-8.04238617e-02 -5.57452142e-01 2.84912407e-01 2.97164042e-02
1.06688404e+00 -2.04226947e+00 1.30792513e-01 2.96054840e-01
1.32432684e-01 8.60618874e-02 -5.52950561e-01 7.47930825e-01
3.78010720e-01 -3.72454375e-02 4.44381349e-02 -5.79695664e-02
6.85018957e-01 5.61802983e-01 -6.68306947e-01 3.24613035e-01
2.12953389e-01 1.12041962e+00 -1.30973291e+00 3.77738774e-02
-4.21740487e-03 2.32431099e-01 -5.79729795e-01 4.52806830e-01
-6.79921985e-01 3.82443547e-01 -6.82500899e-01 3.49891931e-01
5.54858744e-01 1.09665856e-01 -1.72785781e-02 5.34844458e-01
4.40170728e-02 3.64752375e-02 -1.12128544e+00 1.25387096e+00
-5.37808836e-01 -1.39527638e-02 1.01519510e-01 -9.75448132e-01
1.31276941e+00 6.95629865e-02 9.15481269e-01 -1.14777517e+00
3.46621484e-01 3.94200683e-01 2.61078060e-01 -2.05161199e-02
8.02485347e-01 -5.32489479e-01 -4.31532890e-01 7.97520280e-01
4.35392782e-02 -2.37709701e-01 1.67874575e-01 1.25860810e-01
1.04280114e+00 4.50327754e-01 7.24906996e-02 -3.01354736e-01
5.74019775e-02 -4.18578833e-01 6.10753715e-01 7.85269201e-01
-4.67970878e-01 2.32320372e-02 1.29368246e+00 6.17024414e-02
-8.63657594e-01 -1.35389388e+00 2.47168615e-01 1.14363003e+00
1.46915630e-01 -5.22140972e-02 -2.63048708e-01 -6.56584918e-01
6.22419119e-01 8.62747967e-01 -5.98202348e-01 -3.82232398e-01
-2.17622846e-01 -5.17942250e-01 5.49651265e-01 6.61821544e-01
2.43395507e-01 -1.14839220e+00 -8.97033811e-01 3.48536998e-01
1.09290510e-01 -3.94977570e-01 -6.08146250e-01 6.07656777e-01
-7.48672783e-01 -8.27563345e-01 -1.20918393e+00 -8.79454240e-03
1.40199913e-02 -2.41202023e-02 1.18240106e+00 -4.88667548e-01
5.54432869e-02 9.16916847e-01 -2.49955609e-01 -6.03145659e-01
-3.11825603e-01 -3.20242405e-01 1.22575566e-01 -1.05358637e-03
4.26385850e-02 -4.28221166e-01 -6.30642951e-01 3.04383337e-01
-9.58863318e-01 -8.50922346e-01 5.67287147e-01 8.57997358e-01
6.80624425e-01 3.18473540e-02 1.29945898e+00 -3.60044420e-01
1.43096364e+00 -9.28188860e-01 -8.22768450e-01 -3.73952501e-02
-7.09360421e-01 7.10720897e-01 1.00207722e+00 -5.89802504e-01
-8.71416628e-01 -4.69912410e-01 1.11864597e-01 -6.08371675e-01
3.71093959e-01 3.57327431e-01 1.83770090e-01 5.67016602e-01
4.81498033e-01 2.63948500e-01 4.78428394e-01 -1.42363369e-01
6.61913812e-01 3.06913197e-01 5.14409803e-02 -1.23733747e+00
4.91226196e-01 -6.37689233e-02 5.20590603e-01 -3.38568777e-01
-6.38853550e-01 -1.47352189e-01 -1.21508278e-02 -2.62991458e-01
3.79939109e-01 -6.44189894e-01 -1.27864611e+00 3.77370864e-01
-7.64177680e-01 -8.36395860e-01 -1.20767570e+00 6.88006997e-01
-1.38988245e+00 -1.09712526e-01 -4.57725406e-01 -1.44271827e+00
5.15814759e-02 -1.00213778e+00 6.07856572e-01 2.37594143e-01
4.70447421e-01 -1.10537362e+00 4.53848511e-01 -4.91147131e-01
8.72686267e-01 2.37591863e-01 9.28185761e-01 -5.80209315e-01
-3.58777016e-01 2.05716059e-01 -9.72521864e-03 7.61513233e-01
-8.13878998e-02 -5.13065398e-01 -3.48499537e-01 -3.09269965e-01
1.50544748e-01 -8.69984269e-01 6.48905516e-01 6.42167866e-01
7.17185676e-01 -2.36502349e-01 4.15525526e-01 6.76083118e-02
1.57203376e+00 2.81568378e-01 5.82576156e-01 1.82919934e-01
2.40517378e-01 5.75277567e-01 1.06436956e+00 1.22199690e+00
3.99022400e-01 4.23988640e-01 1.04888391e+00 3.18231612e-01
3.72910917e-01 -5.60887098e-01 8.43333304e-01 3.23741168e-01
-2.98547652e-02 -4.63064313e-02 -3.04448068e-01 4.33960080e-01
-2.07441378e+00 -1.23101878e+00 1.57869756e-01 2.43735361e+00
8.15402865e-01 -5.96825853e-02 8.83255303e-01 -2.15938792e-01
4.52358514e-01 7.85398707e-02 -1.16709876e+00 -7.39780962e-01
-2.42974632e-03 2.91341066e-01 7.89986134e-01 1.65499583e-01
-3.20313632e-01 5.61526835e-01 6.15086174e+00 1.05066204e+00
-8.68346870e-01 -2.34160110e-01 5.49230993e-01 -5.27310491e-01
-5.33837199e-01 -3.17440152e-01 -7.17637479e-01 5.05066276e-01
1.02287948e+00 -5.26701629e-01 7.69052088e-01 1.08385909e+00
4.21967059e-01 -1.16412573e-01 -1.13792825e+00 7.90428042e-01
-3.99849325e-01 -8.16245437e-01 -4.02290046e-01 5.03111243e-01
7.11313367e-01 1.65318280e-01 4.65263724e-01 9.23766732e-01
1.16229022e+00 -1.11022174e+00 7.90668786e-01 7.35358596e-01
4.72967058e-01 -1.34074557e+00 7.58952022e-01 5.52226663e-01
-1.10721290e+00 -6.30787194e-01 -6.13326430e-01 -1.66694179e-01
-7.00678602e-02 4.30879444e-01 -4.06981975e-01 4.72994924e-01
3.42001826e-01 8.04194152e-01 7.48486742e-02 8.96236241e-01
-3.01325709e-01 4.03324336e-01 -1.94190130e-01 -6.53283954e-01
5.57038844e-01 -3.81949604e-01 4.83350039e-01 6.38228953e-01
4.09362137e-01 -5.19035578e-01 2.28449494e-01 1.31407940e+00
-5.11024296e-02 4.98377047e-02 -7.11965442e-01 -1.88116983e-01
3.04393947e-01 1.03704095e+00 -9.60357934e-02 6.84805214e-02
2.68212929e-02 3.59573334e-01 4.10949767e-01 3.64787996e-01
-1.01983738e+00 -4.61534768e-01 9.14118290e-01 -6.16346933e-02
4.33374316e-01 -4.19127226e-01 1.22687183e-01 -6.71768427e-01
-4.66277078e-02 -5.92679024e-01 1.90365285e-01 -4.00194526e-01
-1.63657999e+00 1.98269486e-01 9.10919830e-02 -1.50868428e+00
-6.78235173e-01 -3.51030409e-01 -3.87731940e-01 9.82365370e-01
-1.95331812e+00 -5.48401594e-01 5.48180580e-01 9.01718020e-01
1.03101864e-01 -3.47264588e-01 4.09189224e-01 -7.09829479e-02
-2.29002818e-01 5.17003953e-01 7.12325454e-01 -1.18560895e-01
5.56906044e-01 -1.62512362e+00 -3.18712682e-01 7.53235817e-02
-2.00980633e-01 3.05860251e-01 5.12108445e-01 -3.89982671e-01
-1.69672120e+00 -1.09946048e+00 5.80292754e-02 -1.81752428e-01
1.04553950e+00 -5.28230183e-02 -4.77671266e-01 2.19668746e-01
8.02982375e-02 4.15050723e-02 3.51242572e-01 -6.58580586e-02
-3.06787342e-01 -3.39415878e-01 -1.31021488e+00 4.69535321e-01
5.82416892e-01 -1.13565795e-01 -2.95511216e-01 -2.07546175e-01
8.13996851e-01 -1.87482070e-02 -9.51523423e-01 -3.48696001e-02
3.24041098e-01 -1.06152272e+00 7.60613859e-01 -4.51821864e-01
6.18722081e-01 -3.11517939e-02 -4.35663581e-01 -1.99558008e+00
-2.98105758e-02 -9.06569183e-01 -2.59854674e-01 1.19077516e+00
-1.72791909e-03 -8.23309481e-01 4.19177562e-01 3.52502048e-01
2.02858634e-02 -1.03828514e+00 -1.05568027e+00 -1.51763463e+00
7.45224357e-01 -4.62060839e-01 7.27666318e-01 1.22861661e-01
2.84064144e-01 2.15953574e-01 -6.88517630e-01 -5.65423906e-01
9.15552497e-01 1.59189284e-01 6.24704480e-01 -9.21164870e-01
-7.31681705e-01 -8.84271145e-01 -5.98980710e-02 -1.27877629e+00
2.29941130e-01 -1.01434958e+00 2.12304428e-01 -1.43422306e+00
-1.76108591e-02 -7.17466652e-01 -6.69668317e-01 2.56728917e-01
2.00308621e-01 -3.80732983e-01 5.42501152e-01 7.89945498e-02
-8.89138758e-01 1.32455587e+00 1.71670306e+00 5.99301010e-02
-4.99872863e-01 4.05110925e-01 -8.88797402e-01 2.94808358e-01
1.02077973e+00 -5.80053985e-01 -8.65854561e-01 -4.66546118e-02
5.06552398e-01 7.06880927e-01 2.04451159e-01 -7.94349134e-01
-1.75879940e-01 -7.44005859e-01 -1.51011636e-02 -5.02240658e-01
1.84394836e-01 -7.85488605e-01 -5.48542619e-01 5.70399046e-01
-7.98976779e-01 9.08287428e-03 -1.23826385e-01 9.92013633e-01
-1.06282972e-01 -1.01992927e-01 8.70234251e-01 -8.83026421e-02
-3.66012566e-02 4.81023520e-01 -4.45768327e-01 6.28256142e-01
1.02169156e+00 3.09396148e-01 -4.06418294e-01 -7.51732469e-01
-3.47287357e-01 6.77078128e-01 -5.62539026e-02 2.35327587e-01
7.06547737e-01 -1.53930819e+00 -6.38932943e-01 -2.01817647e-01
-2.33043045e-01 -1.48992121e-01 7.90422931e-02 8.81955028e-01
3.05378824e-01 2.58780867e-01 -4.37754512e-01 -3.00167203e-01
-4.36018795e-01 6.34628415e-01 4.13400859e-01 -8.66512835e-01
-2.12917551e-01 3.10876369e-01 -1.15265986e-02 -4.17299747e-01
3.40307981e-01 -6.11374140e-01 -1.75321728e-01 6.39518946e-02
4.70999330e-01 7.09334552e-01 -5.57618320e-01 -1.77766755e-01
6.53181523e-02 3.04276735e-01 2.15427145e-01 -4.52064633e-01
1.48081160e+00 -8.56877938e-02 2.26620257e-01 7.73451149e-01
9.96333718e-01 -1.64919928e-01 -1.84037709e+00 -4.20027763e-01
-3.48108038e-02 -4.05180573e-01 -1.86673969e-01 -1.03527391e+00
-1.18778956e+00 1.05022728e+00 4.05493468e-01 3.99678767e-01
8.56531322e-01 -1.33719236e-01 7.41163492e-01 2.89504975e-01
7.69720912e-01 -1.45314217e+00 6.22743428e-01 8.34158003e-01
8.73343945e-01 -9.08604145e-01 -4.13917989e-01 6.42341554e-01
-1.20552742e+00 1.03713906e+00 3.74230504e-01 -7.36724496e-01
5.87128282e-01 4.46290553e-01 -2.64771610e-01 2.56487548e-01
-9.57817972e-01 -5.88006318e-01 -4.32824731e-01 9.96940434e-01
2.28698581e-01 3.95191193e-01 -3.11634898e-01 9.80963111e-01
7.87270889e-02 1.05059512e-01 7.53798604e-01 1.02966344e+00
-4.69115406e-01 -1.60565615e+00 -2.03675598e-01 4.58317548e-01
-4.39025998e-01 4.60312068e-02 3.43387946e-02 4.52126861e-01
-4.05037403e-01 1.03625691e+00 -4.94353175e-02 -2.78402060e-01
4.52989340e-01 -3.59743267e-01 5.91355979e-01 -3.29832077e-01
-8.19677234e-01 1.47147909e-01 -4.43066269e-01 -6.38575137e-01
-9.81871933e-02 -4.93759036e-01 -1.60573852e+00 -2.62518018e-01
1.70900315e-01 2.26240218e-01 6.77019715e-01 7.13506877e-01
2.79027849e-01 6.46617174e-01 1.19393444e+00 -6.12874031e-01
-1.68853569e+00 -7.55038381e-01 -1.39734244e+00 2.52282917e-01
4.93613482e-01 -8.36661518e-01 -4.10883665e-01 -9.37509954e-01] | [4.059291362762451, 2.4915127754211426] |
2064cde2-a9b3-4f7d-bc95-a580234d4353 | introduction-to-presentation-attack-detection | 2111.11794 | null | https://arxiv.org/abs/2111.11794v2 | https://arxiv.org/pdf/2111.11794v2.pdf | Introduction to Presentation Attack Detection in Face Biometrics and Recent Advances | The main scope of this chapter is to serve as an introduction to face presentation attack detection, including key resources and advances in the field in the last few years. The next pages present the different presentation attacks that a face recognition system can confront, in which an attacker presents to the sensor, mainly a camera, a Presentation Attack Instrument (PAI), that is generally a photograph, a video, or a mask, to try to impersonate a genuine user. First, we make an introduction of the current status of face recognition, its level of deployment, and its challenges. In addition, we present the vulnerabilities and the possible attacks that a face recognition system may be exposed to, showing that way the high importance of presentation attack detection methods. We review different types of presentation attack methods, from simpler to more complex ones, and in which cases they could be effective. Then, we summarize the most popular presentation attack detection methods to deal with these attacks. Finally, we introduce public datasets used by the research community for exploring vulnerabilities of face biometrics to presentation attacks and developing effective countermeasures against known PAIs. | ['Javier Galbally', 'Aythami Morales', 'Julian Fierrez', 'Javier Hernandez-Ortega'] | 2021-11-23 | null | null | null | null | ['face-presentation-attack-detection'] | ['computer-vision'] | [ 6.63740695e-01 -1.31455049e-01 1.28959775e-01 -1.34669825e-01
-3.13682407e-01 -1.01235533e+00 6.90765500e-01 -3.49167347e-01
4.69436459e-02 2.04024449e-01 -2.90347666e-01 -3.66563469e-01
-5.70935942e-02 -7.16655433e-01 -3.59540999e-01 -9.36858177e-01
-2.68488854e-01 -2.43561894e-01 -8.55619907e-02 -1.83581993e-01
4.69508797e-01 1.11391997e+00 -1.77308977e+00 3.91245335e-01
-1.95146546e-01 1.29929340e+00 -4.20293361e-01 8.27683330e-01
-3.95847037e-02 2.39586476e-02 -1.47799408e+00 -1.02503157e+00
3.23956907e-01 -2.67562181e-01 -4.45273399e-01 2.67401133e-02
8.36066425e-01 -6.42472744e-01 -2.73299962e-01 1.18292785e+00
7.09344745e-01 -5.40683389e-01 6.39318168e-01 -1.75071895e+00
-4.90462929e-01 3.34036589e-01 -5.83029330e-01 4.33823317e-01
1.02128661e+00 1.72282726e-01 -1.84580944e-02 -7.89564133e-01
3.78731519e-01 1.47965920e+00 3.83070678e-01 1.18931365e+00
-8.76819611e-01 -1.20083022e+00 -1.45248696e-01 1.41900927e-01
-1.37605286e+00 -7.65068114e-01 8.59337926e-01 -2.75455862e-01
6.29296601e-01 6.95323110e-01 3.08285654e-01 1.59605825e+00
2.79291958e-01 2.99818844e-01 1.20449257e+00 -5.89929223e-01
-6.40215259e-03 4.52689201e-01 4.57949698e-01 6.13408089e-01
5.52917242e-01 5.93056642e-02 -3.47167015e-01 -1.03978503e+00
3.15211535e-01 -2.08792314e-01 -2.85316944e-01 1.23977579e-01
-1.60711870e-01 5.81967473e-01 -1.38350517e-01 4.14438784e-01
-2.54507810e-01 -2.38570482e-01 3.90278995e-01 2.29447961e-01
-1.99572012e-01 -1.40060335e-01 1.33760720e-01 2.40626588e-01
-5.77099144e-01 9.61083546e-02 1.04421973e+00 3.94603044e-01
2.80120641e-01 1.56977445e-01 5.57435385e-04 5.15325844e-01
6.42841339e-01 1.06022632e+00 -4.17168178e-02 -2.73997158e-01
3.00006241e-01 4.21378668e-03 -2.73146302e-01 -1.30452228e+00
3.36293094e-02 3.62140894e-01 -4.02657747e-01 4.76762682e-01
2.85315216e-01 -3.03586900e-01 -6.02754474e-01 1.34110641e+00
1.41334474e-01 3.60296339e-01 1.14831448e-01 3.33701670e-01
1.04011929e+00 4.02195632e-01 1.36383250e-01 -3.90297353e-01
1.99690318e+00 -1.46444708e-01 -9.70037341e-01 2.11186484e-01
-9.20328423e-02 -1.15043664e+00 5.00885665e-01 4.77838546e-01
-1.05934727e+00 -2.31349945e-01 -1.26471484e+00 7.73017526e-01
-8.03024173e-01 -1.08226292e-01 3.05172235e-01 2.13195944e+00
-9.06122208e-01 2.38674790e-01 -3.46211642e-01 -5.69654167e-01
5.65456569e-01 7.50355124e-01 -5.66761792e-01 -4.54078503e-02
-1.13535321e+00 7.52383292e-01 -1.98145643e-01 1.94654718e-01
-7.15019882e-01 -5.07467747e-01 -7.68354833e-01 -7.16451779e-02
2.06884623e-01 -9.45010036e-02 8.78914297e-01 -6.72086716e-01
-1.47666144e+00 1.20982194e+00 -1.79328680e-01 -1.83100969e-01
1.55936077e-01 -3.02864630e-02 -1.03363776e+00 3.74475151e-01
-6.05126381e-01 2.20885146e-02 1.37939584e+00 -1.01765060e+00
-7.87771866e-02 -6.16351008e-01 7.01439977e-02 -5.01672268e-01
-7.12237537e-01 1.07724595e+00 -2.35381037e-01 -4.59304959e-01
-4.34587806e-01 -7.26571500e-01 3.88525903e-01 6.09785225e-03
-4.83929694e-01 -2.30149865e-01 1.70786083e+00 -4.64497328e-01
1.23242342e+00 -2.37721491e+00 -3.76312673e-01 5.66308796e-01
-1.56131968e-01 1.12465763e+00 -1.70903608e-01 9.20215905e-01
-3.05902690e-01 4.13098276e-01 -1.43029196e-02 -2.69856036e-01
-1.22810006e-01 8.74401852e-02 -6.79368436e-01 7.58582771e-01
2.50343025e-01 5.12554526e-01 -1.31441951e-01 2.39374116e-01
2.33753845e-01 9.68510091e-01 1.60697535e-01 2.49542043e-01
3.79342973e-01 1.42578617e-01 -4.28907096e-01 9.75940585e-01
1.40661037e+00 4.38525349e-01 2.00399294e-01 -2.16597959e-01
1.24901436e-01 7.73613676e-02 -1.31760347e+00 5.12325943e-01
5.51654696e-02 6.94371343e-01 4.37886447e-01 -6.06184542e-01
8.54443491e-01 6.44490659e-01 8.03633481e-02 -3.37314755e-01
3.99433315e-01 2.35670671e-01 -1.54070988e-01 -6.16636932e-01
-4.82959719e-03 3.98066729e-01 2.10702494e-01 7.44310915e-01
-7.45553374e-02 4.94112760e-01 -1.40418913e-02 -7.56311649e-03
1.08986068e+00 -5.81735909e-01 3.37522477e-01 -5.72987609e-02
1.39773023e+00 -9.57713246e-01 -9.68370289e-02 6.43643379e-01
-4.25597757e-01 2.58936081e-02 5.41422307e-01 -3.49743843e-01
-2.34086826e-01 -9.30009961e-01 -3.18230242e-01 6.00809395e-01
-3.58865485e-02 -7.73508251e-01 -1.16507113e+00 -1.07233882e+00
1.29070416e-01 -8.58603343e-02 -6.64868057e-01 -2.22649232e-01
-7.42670655e-01 -7.23537505e-01 1.21360755e+00 2.86800861e-01
7.10908532e-01 -1.27133942e+00 -7.16750979e-01 -3.59968722e-01
6.68203011e-02 -1.02394331e+00 -4.91051599e-02 -4.22641665e-01
-3.88612956e-01 -1.43999565e+00 -8.12042296e-01 -4.76961434e-01
7.48532474e-01 3.87930542e-01 6.05359077e-01 6.34468913e-01
-7.51841426e-01 9.67970192e-01 -3.67617421e-02 -7.05461442e-01
-5.39651930e-01 -5.63806832e-01 2.37984017e-01 3.63354862e-01
9.00794089e-01 -1.41742215e-01 -3.44691008e-01 3.17274779e-01
-1.06101537e+00 -1.19899559e+00 1.33925468e-01 2.99505442e-01
-2.77765274e-01 6.48600906e-02 7.46460781e-02 -8.79121363e-01
8.22856605e-01 -2.76117653e-01 -5.24968207e-01 4.35011864e-01
-9.71064940e-02 -5.72990954e-01 9.57099646e-02 -7.35059381e-01
-7.87892044e-01 6.31716028e-02 -4.80901331e-01 -2.55747527e-01
-5.24127722e-01 -4.54156026e-02 -7.74236560e-01 -1.02765179e+00
6.30894125e-01 3.17080230e-01 3.02577466e-01 -4.96544302e-01
-2.73688193e-02 8.97878051e-01 2.66686410e-01 -3.77262443e-01
1.11354673e+00 5.85787416e-01 1.96661308e-01 -1.45497930e+00
4.42270748e-02 -2.13665947e-01 -4.94806290e-01 -5.13931692e-01
4.37632650e-01 -1.50305852e-01 -1.18727779e+00 1.21254754e+00
-1.27685905e+00 4.35725242e-01 5.60239851e-01 -2.03018293e-01
-5.87720908e-02 7.99676061e-01 -7.85491288e-01 -1.22016990e+00
-3.86933327e-01 -1.27709174e+00 9.92904007e-01 5.29482663e-01
1.20975867e-01 -7.50137985e-01 -1.99578762e-01 3.28609198e-01
7.46349633e-01 4.82954711e-01 7.08323598e-01 -5.35440624e-01
-2.67337143e-01 -4.79593068e-01 -8.91923085e-02 4.28040087e-01
4.71765906e-01 4.79854733e-01 -1.49243140e+00 -6.22991085e-01
3.76708925e-01 -2.81591881e-02 5.11075377e-01 1.14817303e-02
1.09739399e+00 -5.94563901e-01 -6.45924985e-01 3.98558766e-01
1.25042748e+00 5.33241868e-01 1.19694388e+00 2.06731826e-01
1.42022103e-01 1.02147985e+00 1.73299789e-01 2.92881280e-01
-3.90443146e-01 8.54654014e-01 4.84756768e-01 9.04093497e-03
7.90114775e-02 1.97257921e-01 7.62736738e-01 -1.98590785e-01
-1.17465973e-01 -4.39889848e-01 -4.77671295e-01 -4.27810729e-01
-9.82813179e-01 -1.05305207e+00 -1.20296441e-01 2.36053967e+00
8.03012475e-02 -2.00271845e-01 5.14643013e-01 5.70987463e-01
1.03940201e+00 1.37848780e-01 -4.24028412e-02 -7.89796293e-01
-1.76850185e-01 4.47907895e-01 3.67556512e-01 2.21620902e-01
-1.30481517e+00 6.97556496e-01 7.37838364e+00 5.13372123e-01
-1.57293761e+00 -2.53964454e-01 3.59124213e-01 2.58749783e-01
3.22892874e-01 -4.53074455e-01 -1.25313926e+00 6.70861661e-01
1.17699027e+00 5.44630066e-02 1.80696279e-01 7.09095359e-01
-3.28268528e-01 8.29765107e-03 -8.21470916e-01 1.38159263e+00
9.34901536e-01 -1.01874828e+00 3.66945475e-01 6.24996960e-01
-6.16449527e-02 -7.90758073e-01 5.84147573e-01 -2.45787978e-01
-4.93204415e-01 -1.01204503e+00 1.34452581e-01 6.33419454e-02
5.25780022e-01 -8.84630620e-01 5.99132717e-01 -2.70101815e-01
-1.31590486e+00 -2.75304884e-01 -3.27387124e-01 2.28780821e-01
1.33374676e-01 -3.35667357e-02 -2.32463762e-01 4.48564857e-01
6.47164762e-01 2.64379997e-02 -7.26135433e-01 9.86541092e-01
-2.00731918e-01 5.47962010e-01 -4.98366088e-01 1.93280410e-02
-1.25282243e-01 1.23586543e-01 7.08083272e-01 1.27700388e+00
2.12428719e-01 2.84231484e-01 -1.62602633e-01 4.04050261e-01
2.27528006e-01 3.75387408e-02 -1.08705974e+00 -1.31195262e-01
4.27561671e-01 1.32225502e+00 -6.03283226e-01 -5.75963929e-02
-6.42190635e-01 1.08734477e+00 -2.65954137e-01 1.85254812e-01
-5.36247432e-01 -4.63271379e-01 8.93919826e-01 9.24049616e-02
1.36306599e-01 -4.60504629e-02 4.25646156e-01 -7.28465736e-01
2.37110004e-01 -1.37962532e+00 7.82067835e-01 -3.85703027e-01
-1.26294768e+00 8.59564781e-01 7.76989311e-02 -8.92785192e-01
-1.70812443e-01 -1.07419121e+00 -8.19386303e-01 1.03082502e+00
-1.20789611e+00 -1.17626786e+00 -8.66948217e-02 7.27070928e-01
1.85612306e-01 -4.93914753e-01 1.04906929e+00 4.91275996e-01
-7.41176724e-01 1.06395006e+00 -6.11419678e-01 3.44915390e-01
5.98967552e-01 -5.31554103e-01 3.98310035e-01 8.72960150e-01
1.10973097e-01 1.05171621e+00 4.68321264e-01 -5.93512118e-01
-1.72693884e+00 -2.78355300e-01 6.76090419e-01 -3.85612696e-01
2.98473507e-01 -7.11910546e-01 -9.14893687e-01 4.42698896e-01
3.74716431e-01 -2.67136991e-01 1.17543769e+00 -3.40619236e-01
-7.62512863e-01 1.86160393e-02 -1.75103247e+00 5.87168515e-01
5.32241821e-01 -6.71856940e-01 -1.54430941e-01 7.64605179e-02
-1.79282919e-01 -1.27180561e-01 -4.39430267e-01 3.29621345e-01
1.08153307e+00 -9.46796298e-01 1.11624348e+00 -4.90606517e-01
-4.05933291e-01 -1.34448558e-01 -2.57989615e-01 -4.81634438e-01
-2.89798882e-02 -1.01091874e+00 -3.44425946e-01 1.62588751e+00
-1.32773351e-02 -9.42392290e-01 6.72949016e-01 8.12058449e-01
7.53715992e-01 -2.73190111e-01 -7.85787880e-01 -8.88202727e-01
-1.75085783e-01 -1.69452876e-01 8.13562810e-01 7.78500140e-01
-3.86045948e-02 -2.43291304e-01 -3.95093232e-01 7.24289000e-01
9.95202124e-01 -6.38756096e-01 8.60551715e-01 -1.20371449e+00
3.77129437e-03 -5.40023804e-01 -1.00296974e+00 -5.91039360e-01
6.44410327e-02 -3.31361115e-01 -5.74549377e-01 -4.66383070e-01
1.65468883e-02 2.94087946e-01 -2.15444595e-01 3.72557789e-01
8.78086537e-02 7.44084060e-01 4.10328448e-01 2.65190322e-02
2.40888983e-01 -2.24635199e-01 3.60196650e-01 -1.38698071e-01
7.85205141e-02 3.92138273e-01 -5.45814097e-01 6.25288427e-01
5.87729394e-01 -3.26489508e-01 -2.75530756e-01 2.32090503e-01
-2.12648600e-01 -1.37726078e-02 3.59827518e-01 -9.24094081e-01
2.55357683e-01 1.97325721e-01 5.06580234e-01 -6.10385478e-01
4.96670306e-01 -1.06131971e+00 3.67373228e-02 7.56042898e-01
1.26320377e-01 2.66900808e-01 5.93010008e-01 1.89851135e-01
1.74862072e-01 -5.57963312e-01 8.60376477e-01 1.54217169e-01
-4.37388957e-01 2.60602981e-01 -8.14881623e-01 -7.61030614e-01
1.35161531e+00 -4.54341471e-01 -6.44137084e-01 -2.32880265e-01
-7.92497635e-01 -4.40536976e-01 2.42466420e-01 5.63061833e-01
8.09363008e-01 -1.01187432e+00 -5.49240470e-01 8.55613708e-01
1.08338259e-01 -1.33118153e+00 2.17245370e-01 3.30070347e-01
-3.80596578e-01 2.28870437e-01 -5.02229095e-01 -4.56358850e-01
-2.54967093e+00 7.34734416e-01 4.62703019e-01 3.60532224e-01
-4.10495341e-01 6.40344739e-01 1.02810316e-01 2.86265433e-01
6.04426563e-01 5.95897257e-01 -6.50531948e-01 2.80965287e-02
1.77063560e+00 3.69040161e-01 3.39433014e-01 -9.56016243e-01
-7.93857396e-01 7.57689595e-01 -2.89484710e-01 -6.11283369e-02
6.46662831e-01 -8.85728598e-02 -4.10919905e-01 -2.03878760e-01
1.21948338e+00 2.60259777e-01 -6.10837162e-01 5.11259176e-02
-6.98257089e-02 -8.74935746e-01 -5.51936686e-01 -6.86133444e-01
-1.31015575e+00 8.84163976e-01 9.86994445e-01 6.74711883e-01
1.15594208e+00 -1.08597673e-01 5.44519722e-01 3.72437656e-01
3.07588786e-01 -2.78800488e-01 1.03999749e-01 2.67004576e-02
9.97113645e-01 -7.53648818e-01 -1.99023589e-01 -1.06465447e+00
-1.17219195e-01 1.50783265e+00 4.45380837e-01 -8.52549225e-02
1.23116088e+00 6.16519332e-01 3.33976775e-01 -3.82467061e-01
-3.04162830e-01 1.92856461e-01 3.06714386e-01 1.30350518e+00
2.60716707e-01 -3.79221648e-01 -1.60293460e-01 1.85332090e-01
-1.11342564e-01 -4.31985438e-01 5.58959901e-01 1.05684876e+00
-7.76558891e-02 -1.74139917e+00 -1.04340398e+00 1.75879508e-01
-1.05507648e+00 2.76665986e-01 -1.13040340e+00 6.29094779e-01
-9.42583531e-02 1.27406526e+00 -9.66090783e-02 -4.19970036e-01
5.33452690e-01 1.45536542e-01 6.10761166e-01 -3.08633119e-01
-8.94229054e-01 -1.51782051e-01 -1.00172654e-01 -5.04362762e-01
-4.23327863e-01 -7.44062483e-01 -4.57548767e-01 -8.05987060e-01
-1.97755411e-01 -8.71344805e-02 9.97929037e-01 7.99561024e-01
1.39992192e-01 -1.43299997e-02 9.00305331e-01 -6.72702491e-01
-3.03691715e-01 -6.69489741e-01 -6.80388510e-01 1.89847156e-01
5.26178896e-01 -6.46400869e-01 -5.89038253e-01 -1.47708192e-01] | [13.080181121826172, 1.107970952987671] |
d115175c-b109-4cfd-a64e-4183284d21e6 | automatic-dialect-detection-in-arabic | 1509.06928 | null | http://arxiv.org/abs/1509.06928v2 | http://arxiv.org/pdf/1509.06928v2.pdf | Automatic Dialect Detection in Arabic Broadcast Speech | We investigate different approaches for dialect identification in Arabic
broadcast speech, using phonetic, lexical features obtained from a speech
recognition system, and acoustic features using the i-vector framework. We
studied both generative and discriminate classifiers, and we combined these
features using a multi-class Support Vector Machine (SVM). We validated our
results on an Arabic/English language identification task, with an accuracy of
100%. We used these features in a binary classifier to discriminate between
Modern Standard Arabic (MSA) and Dialectal Arabic, with an accuracy of 100%. We
further report results using the proposed method to discriminate between the
five most widely used dialects of Arabic: namely Egyptian, Gulf, Levantine,
North African, and MSA, with an accuracy of 52%. We discuss dialect
identification errors in the context of dialect code-switching between
Dialectal Arabic and MSA, and compare the error pattern between manually
labeled data, and the output from our classifier. We also release the train and
test data as standard corpus for dialect identification. | ['Ahmed Ali', 'Sree Harsha Yella', 'Sameer Khurana', 'Patrick Cardinal', 'Steve Renals', 'James Glass', 'Peter Bell', 'Najim Dehak'] | 2015-09-23 | null | null | null | null | ['spoken-language-identification'] | ['speech'] | [ 1.44336984e-01 -2.49989837e-01 3.56283814e-01 -6.99346721e-01
-9.93108630e-01 -1.07414913e+00 7.49854445e-01 9.17305350e-02
-4.30448920e-01 4.55228657e-01 -1.96106709e-03 -5.17655551e-01
-6.26621619e-02 -7.30046809e-01 -5.24837151e-02 -9.42120373e-01
-3.85832079e-02 9.31792080e-01 4.37626578e-02 -7.99123228e-01
4.65207309e-01 6.36209548e-01 -1.65652585e+00 5.23260295e-01
1.21262968e+00 7.86091626e-01 2.80513577e-02 6.71839297e-01
-3.82071622e-02 4.98809814e-01 -9.06414032e-01 -5.31107545e-01
-1.94121629e-01 -5.14618874e-01 -1.28466928e+00 5.46106212e-02
3.72680604e-01 -1.73940733e-02 3.55014727e-02 8.60313833e-01
4.75472480e-01 -2.06853405e-01 1.14965487e+00 -1.03622949e+00
-5.49239874e-01 9.82242346e-01 -1.99811235e-01 1.05446309e-01
6.09994650e-01 -2.65908808e-01 7.09516108e-01 -8.25522423e-01
3.24488580e-01 1.33081722e+00 7.05611169e-01 2.35561252e-01
-1.12022030e+00 -6.41244411e-01 -3.35736811e-01 3.22773844e-01
-1.64591086e+00 -7.32701421e-01 4.75197226e-01 -5.88471949e-01
8.15704465e-01 5.49020886e-01 3.94672185e-01 8.25147212e-01
-6.47877082e-02 6.16080880e-01 1.64846361e+00 -8.50189567e-01
6.15929924e-02 5.60669363e-01 3.13257456e-01 8.43262315e-01
-3.57666820e-01 -1.69937730e-01 -2.52877444e-01 -3.46413881e-01
3.00328851e-01 -8.21978569e-01 -2.81044811e-01 2.61198580e-01
-1.02094984e+00 1.20103431e+00 -1.28217295e-01 7.91374505e-01
-2.19218105e-01 -7.43633628e-01 2.46820420e-01 7.15153694e-01
3.81646544e-01 1.92434490e-01 -4.73271132e-01 -2.95222491e-01
-7.87016988e-01 -1.71351850e-01 1.00641906e+00 5.88880837e-01
7.39253938e-01 3.78660232e-01 4.76712942e-01 1.71541929e+00
4.32018161e-01 9.33451235e-01 9.23818350e-01 -3.27528983e-01
1.74698323e-01 2.95942068e-01 -1.76896244e-01 -7.69097209e-01
-3.80405933e-01 2.09913760e-01 -4.77133125e-01 3.00067961e-01
7.13067234e-01 -3.15958351e-01 -7.53727734e-01 1.32224524e+00
6.35626838e-02 -6.45231903e-01 3.46059382e-01 7.31523097e-01
6.82088315e-01 9.93325293e-01 -4.43777025e-01 -2.22087532e-01
1.25590372e+00 -7.52174973e-01 -5.82497835e-01 2.68020220e-02
7.04176426e-01 -1.27934754e+00 1.21580315e+00 6.24100745e-01
-8.26977491e-01 -4.60888296e-01 -1.04273379e+00 5.88762701e-01
-4.71575201e-01 4.42410499e-01 2.13998154e-01 1.40323246e+00
-1.06097555e+00 1.68934271e-01 -4.06311572e-01 -4.63478774e-01
-2.30135068e-01 5.22806227e-01 -6.25614285e-01 2.56250411e-01
-1.18834639e+00 1.17270672e+00 1.75138116e-01 -8.29014257e-02
-4.52952087e-01 4.02838252e-02 -1.03167903e+00 -2.09573299e-01
-5.13440251e-01 5.21469235e-01 9.87618625e-01 -1.39416909e+00
-1.76391435e+00 1.23175538e+00 -4.34062965e-02 -5.35317920e-02
3.65282446e-02 2.42743552e-01 -1.16257966e+00 -4.41487543e-02
-8.22656006e-02 1.40797853e-01 6.40391231e-01 -1.34209573e+00
-8.97109151e-01 -4.07779008e-01 -2.04214960e-01 9.63475928e-02
-1.79089144e-01 4.25993860e-01 2.32496798e-01 -5.99663556e-01
4.30702090e-01 -1.15286899e+00 4.17450160e-01 -8.69768381e-01
-2.19571814e-01 -1.67789981e-01 5.38366795e-01 -1.28511691e+00
1.10280275e+00 -2.06757474e+00 1.99515179e-01 7.02900708e-01
-4.19094086e-01 2.14801252e-01 -2.37280857e-02 5.34368813e-01
-3.85663770e-02 -3.08730155e-01 -5.93979001e-01 -7.18195457e-04
-4.76661511e-02 3.30990762e-01 -1.59640625e-01 5.30215979e-01
2.19861090e-01 1.47106633e-01 -5.37395597e-01 -2.51052439e-01
-8.19520000e-03 4.15355474e-01 -8.95004645e-02 8.39336663e-02
3.55058610e-01 7.33252838e-02 3.04592162e-01 1.03108644e+00
7.36208200e-01 6.78112626e-01 3.90760958e-01 -1.02896295e-01
-2.81284213e-01 2.97546685e-01 -1.19880486e+00 9.23592091e-01
-8.37866008e-01 8.11299145e-01 2.69691825e-01 -9.72808063e-01
1.30996358e+00 2.63272375e-01 -6.26372993e-02 -4.70896453e-01
3.08876157e-01 7.78435528e-01 4.57260281e-01 -3.10900748e-01
5.78048944e-01 3.98513786e-02 -2.12382749e-01 6.29905283e-01
3.79421562e-01 -2.90212065e-01 1.79332018e-01 -3.97583425e-01
3.44699293e-01 -2.55886078e-01 1.97508901e-01 -6.61628664e-01
9.54099417e-01 3.71902362e-02 -3.92487459e-02 5.95138907e-01
-1.27545998e-01 7.54669785e-01 3.95077676e-01 -2.53341734e-01
-5.22165418e-01 -1.19575131e+00 -6.12606525e-01 1.45165348e+00
-2.32364684e-01 -3.33436690e-02 -9.14103746e-01 -7.27996707e-01
-1.40346706e-01 7.83491552e-01 -3.20924133e-01 -2.66616640e-04
-9.00077522e-01 -1.34213388e+00 9.99153554e-01 4.38681198e-03
2.60975152e-01 -1.04594493e+00 1.70669004e-01 2.68354625e-01
-4.09189105e-01 -7.66862094e-01 -4.12743032e-01 2.99069256e-01
-2.51289636e-01 -1.13031280e+00 -7.24412739e-01 -1.48456907e+00
3.34058195e-01 -3.09760749e-01 8.15521121e-01 -9.91276726e-02
1.73375919e-01 2.99480140e-01 -6.14353776e-01 -2.00590312e-01
-1.36777484e+00 8.37643892e-02 2.42322743e-01 5.05851567e-01
3.71233523e-01 -1.23416819e-01 1.31752118e-01 7.03448236e-01
-6.28755450e-01 -6.17669821e-01 2.84667134e-01 7.55371690e-01
-9.90891159e-02 -2.28601694e-01 6.69429004e-01 -7.44127035e-01
5.25688887e-01 -4.72306103e-01 -4.86551940e-01 3.63434672e-01
-5.61743259e-01 7.47577986e-03 4.13908392e-01 -5.22848964e-01
-8.56084287e-01 2.51395404e-01 -1.03731823e+00 5.87956429e-01
-2.85286874e-01 4.76273477e-01 -3.14478248e-01 -3.39936316e-01
8.94464612e-01 6.53538585e-01 4.45424497e-01 -3.35900545e-01
2.14145318e-01 1.74087501e+00 5.38812160e-01 -4.12568510e-01
2.56961972e-01 5.09357452e-02 -6.26390755e-01 -1.30411267e+00
-1.41783226e-02 -5.12668014e-01 -7.97161400e-01 -2.82781839e-01
6.15803719e-01 -7.24023700e-01 -4.61066335e-01 1.31759787e+00
-1.02807665e+00 -1.58148348e-01 3.24764669e-01 4.94896173e-01
-5.36368847e-01 4.41375941e-01 -9.00143147e-01 -7.71449387e-01
-2.21568480e-01 -1.47866964e+00 8.53398681e-01 -1.06394067e-01
-4.25020754e-01 -1.11225712e+00 6.29381388e-02 1.60835475e-01
4.61469680e-01 -1.37573972e-01 1.30542874e+00 -9.91113245e-01
3.69091600e-01 6.91558281e-03 8.71095285e-02 4.21022832e-01
6.81137383e-01 3.03068370e-01 -1.25820458e+00 -2.58650512e-01
-9.63417534e-03 -3.33751827e-01 5.55653214e-01 3.85183655e-02
4.08856452e-01 -1.04355454e-01 1.20855086e-01 1.61447033e-01
1.04788637e+00 7.02919006e-01 3.45405012e-01 2.78700948e-01
5.03550768e-01 9.41822827e-01 5.87786674e-01 1.65439516e-01
4.60213721e-01 7.95977235e-01 5.44201098e-02 2.07139388e-01
1.22929879e-01 3.54974717e-01 7.32584834e-01 1.35289490e+00
3.00145783e-02 -8.47863406e-02 -1.31350541e+00 6.39642537e-01
-1.25448406e+00 -6.43296301e-01 -3.47969949e-01 2.28655362e+00
1.12411034e+00 -1.79091915e-01 6.03488624e-01 8.17348361e-01
8.04903686e-01 -9.57860872e-02 4.03483182e-01 -9.74491239e-01
-4.18777436e-01 4.22267705e-01 3.44965011e-01 1.04067934e+00
-1.28371644e+00 1.06057572e+00 6.60114574e+00 9.94669199e-01
-1.60936713e+00 1.25196613e-02 4.87825006e-01 5.18581450e-01
-6.32932037e-02 -2.41240472e-01 -6.67874157e-01 4.22278136e-01
1.36765492e+00 1.29259810e-01 4.65314478e-01 4.60770637e-01
-3.24221462e-01 -5.14782630e-02 -6.38885498e-01 1.06368029e+00
6.09875083e-01 -8.85223687e-01 1.79322399e-02 -2.16800421e-01
3.14934701e-01 4.81288182e-03 3.12960893e-02 2.11803645e-01
1.74923539e-01 -1.03044629e+00 8.77456367e-01 -1.22683577e-01
7.92685509e-01 -9.83671844e-01 9.06759024e-01 8.33557025e-02
-8.89690995e-01 -8.37297961e-02 -1.98806182e-01 1.54462025e-01
-2.83320192e-02 7.78336227e-02 -1.01595783e+00 4.08905327e-01
6.53932869e-01 2.42725730e-01 -5.94625652e-01 6.83195591e-01
1.62896037e-01 1.06348503e+00 -3.30358326e-01 -2.34962747e-01
2.68640012e-01 -4.99199927e-01 3.53137285e-01 1.49921429e+00
4.01044786e-01 -4.81627703e-01 -2.27374494e-01 3.10765840e-02
6.92973793e-01 6.64278567e-01 -3.10786843e-01 1.39008448e-01
3.97373199e-01 9.31702614e-01 -7.61008978e-01 -2.13118613e-01
-2.54266083e-01 1.15735149e+00 -6.99939579e-03 -3.72326374e-02
-6.68228149e-01 -8.10344040e-01 5.89780509e-01 -3.51935625e-01
1.05035439e-01 -7.35893324e-02 -4.51861769e-02 -8.31479251e-01
-2.14806646e-01 -1.29428792e+00 4.14804757e-01 -4.40580696e-01
-1.13850486e+00 1.23810327e+00 -2.71371268e-02 -1.16051745e+00
-6.70215666e-01 -9.84353185e-01 -2.02782184e-01 1.13842177e+00
-9.20191646e-01 -1.21603334e+00 5.75755499e-02 3.91797781e-01
4.14174557e-01 -9.47619319e-01 1.32288921e+00 5.46990871e-01
-1.87373221e-01 8.69914949e-01 5.67089438e-01 4.87219840e-01
8.57260883e-01 -1.40413296e+00 3.24757457e-01 2.75420755e-01
4.42704618e-01 1.07442155e-01 3.58697712e-01 -5.83819970e-02
-8.64092171e-01 -7.28532016e-01 1.26463175e+00 -3.48432928e-01
6.05503857e-01 -5.16487837e-01 -8.31309140e-01 4.44174796e-01
2.22155765e-01 -7.19058633e-01 9.97930288e-01 2.27544680e-01
-4.17824417e-01 -6.73983544e-02 -1.35655582e+00 3.25601190e-01
3.88079226e-01 -7.55324483e-01 -6.81365669e-01 3.10239553e-01
1.54002950e-01 -1.36670992e-01 -9.28841770e-01 -6.17739931e-02
8.71067286e-01 -9.55514610e-01 7.61185348e-01 -1.94624901e-01
-3.82764526e-02 -6.32201880e-02 -5.17420888e-01 -1.87868071e+00
-6.68792129e-02 -2.07661256e-01 8.16444397e-01 1.43036366e+00
7.34832764e-01 -1.01595235e+00 9.12705287e-02 -3.14802140e-01
-1.32098645e-01 6.07631588e-03 -9.71639514e-01 -7.33901858e-01
5.61019719e-01 -1.69490740e-01 5.98683894e-01 1.28939986e+00
5.77292964e-02 2.76769608e-01 -1.98672846e-01 3.10084075e-01
7.72683276e-03 2.06540853e-01 2.02881694e-01 -1.18466592e+00
-2.15642333e-01 -6.45485342e-01 -6.65326536e-01 -4.26482499e-01
4.54123080e-01 -1.11594212e+00 1.01617374e-01 -8.87837529e-01
-6.11269534e-01 -7.19389260e-01 1.24413177e-01 4.52808172e-01
1.13047741e-01 6.17948234e-01 6.00190088e-02 1.56157076e-01
3.63264769e-01 1.99165344e-01 4.12457496e-01 -4.47394907e-01
-3.21932495e-01 4.37776536e-01 -3.77977252e-01 5.84414482e-01
9.17704284e-01 -4.67123210e-01 -4.49083149e-02 -2.72638440e-01
-3.75521630e-01 2.91110594e-02 -2.94482023e-01 -9.93191540e-01
-3.21147680e-01 8.54915604e-02 4.52758595e-02 -1.82489738e-01
3.53965789e-01 -6.36363685e-01 -9.95498523e-02 6.96661294e-01
-8.63111392e-02 3.18721980e-01 2.79181302e-01 -3.41777176e-01
-5.74567020e-01 -5.62678218e-01 1.21788585e+00 3.36396784e-01
-7.38210380e-01 -3.35129708e-01 -1.16677701e+00 -3.09591174e-01
7.98766494e-01 -1.89439341e-01 -2.00679287e-01 -4.13973659e-01
-8.17728102e-01 -2.97078192e-01 2.81243891e-01 6.64948106e-01
3.16140085e-01 -1.11471915e+00 -1.18877089e+00 8.35597336e-01
3.43406528e-01 -1.02525425e+00 -1.84139967e-01 6.95867717e-01
-1.07298648e+00 1.48681909e-01 -6.12166941e-01 -7.46398091e-01
-1.74065518e+00 -5.80968596e-02 6.25964165e-01 4.49888885e-01
2.98905611e-01 6.79280579e-01 -4.25540686e-01 -1.27485025e+00
2.44827922e-02 3.71013163e-03 -7.69080698e-01 5.39892614e-01
3.50813538e-01 3.50132674e-01 5.14513969e-01 -1.43655145e+00
-5.91877103e-01 6.79124892e-01 3.26229595e-02 -5.63061595e-01
7.75540769e-01 -3.69380191e-02 -4.95118052e-01 7.10865855e-01
1.21266723e+00 7.76670277e-01 -2.75730789e-02 5.82721271e-02
1.49714649e-01 -1.82234839e-01 -2.60034382e-01 -9.35545743e-01
-5.91079831e-01 5.44781446e-01 1.09262633e+00 7.27179110e-01
9.57876623e-01 -6.32772818e-02 2.92867362e-01 2.90226638e-01
4.14161265e-01 -1.03892219e+00 -7.19702661e-01 9.76279616e-01
8.53474379e-01 -1.38748336e+00 -6.71762884e-01 -4.18731064e-01
-7.68597126e-01 1.39879453e+00 2.19784722e-01 8.66250843e-02
9.43287849e-01 1.58224255e-01 9.14678872e-01 3.01662236e-01
-1.06146827e-01 -3.06515396e-01 3.22400361e-01 9.64378119e-01
7.36428142e-01 2.11782724e-01 -5.87402046e-01 5.43974698e-01
-7.20595479e-01 -7.22804964e-01 7.15195715e-01 8.93550217e-01
-4.37750518e-01 -1.27455413e+00 -7.84312844e-01 5.44962108e-01
-3.55877221e-01 -1.93166703e-01 -5.50632536e-01 8.86482835e-01
2.23631854e-03 1.38761556e+00 4.00019020e-01 -7.69195497e-01
1.98315635e-01 4.39009815e-01 4.50591117e-01 -2.51722544e-01
-1.01586175e+00 -6.68038288e-03 4.59108353e-01 1.51500344e-01
-3.76267046e-01 -8.99122536e-01 -1.06396699e+00 -3.83348674e-01
-3.73824060e-01 5.01730919e-01 9.27359402e-01 1.00455022e+00
-3.53482515e-01 -1.79681942e-01 1.12308788e+00 -5.70109308e-01
-5.48873961e-01 -1.28647399e+00 -8.48480284e-01 3.02483261e-01
3.14531296e-01 -3.79299015e-01 -6.33839369e-01 5.45747504e-02] | [10.20938491821289, 10.689129829406738] |
7fd680a5-12f3-4964-89c9-14ed4b707315 | towards-robust-video-object-segmentation-with | 2207.00887 | null | https://arxiv.org/abs/2207.00887v1 | https://arxiv.org/pdf/2207.00887v1.pdf | Towards Robust Video Object Segmentation with Adaptive Object Calibration | In the booming video era, video segmentation attracts increasing research attention in the multimedia community. Semi-supervised video object segmentation (VOS) aims at segmenting objects in all target frames of a video, given annotated object masks of reference frames. Most existing methods build pixel-wise reference-target correlations and then perform pixel-wise tracking to obtain target masks. Due to neglecting object-level cues, pixel-level approaches make the tracking vulnerable to perturbations, and even indiscriminate among similar objects. Towards robust VOS, the key insight is to calibrate the representation and mask of each specific object to be expressive and discriminative. Accordingly, we propose a new deep network, which can adaptively construct object representations and calibrate object masks to achieve stronger robustness. First, we construct the object representations by applying an adaptive object proxy (AOP) aggregation method, where the proxies represent arbitrary-shaped segments at multi-levels for reference. Then, prototype masks are initially generated from the reference-target correlations based on AOP. Afterwards, such proto-masks are further calibrated through network modulation, conditioning on the object proxy representations. We consolidate this conditional mask calibration process in a progressive manner, where the object representations and proto-masks evolve to be discriminative iteratively. Extensive experiments are conducted on the standard VOS benchmarks, YouTube-VOS-18/19 and DAVIS-17. Our model achieves the state-of-the-art performance among existing published works, and also exhibits superior robustness against perturbations. Our project repo is at https://github.com/JerryX1110/Robust-Video-Object-Segmentation | ['Yan Lu', 'Xiang Ming', 'Jinglu Wang', 'Xiaohao Xu'] | 2022-07-02 | null | null | null | null | ['semi-supervised-video-object-segmentation', 'visual-object-tracking'] | ['computer-vision', 'computer-vision'] | [ 3.09663862e-01 -1.38715714e-01 -4.14691955e-01 -2.83194065e-01
-6.79818451e-01 -5.95029056e-01 3.90151113e-01 -2.23470718e-01
-2.66197443e-01 4.22927797e-01 -2.66451929e-02 1.02497332e-01
1.78584028e-02 -4.61060315e-01 -8.96185219e-01 -7.66902089e-01
3.46567817e-02 2.78576463e-01 9.03103113e-01 3.22744506e-03
9.30744410e-02 4.59240109e-01 -1.41387689e+00 2.47597605e-01
9.09156680e-01 1.16836381e+00 3.36582243e-01 6.38543069e-01
1.18426546e-01 5.55018604e-01 -5.82214773e-01 -1.61831796e-01
5.70832849e-01 -4.23029095e-01 -7.25611329e-01 5.29504478e-01
5.77820897e-01 -1.92072704e-01 -5.24677098e-01 1.26786351e+00
3.24781418e-01 3.14161986e-01 5.97703099e-01 -1.35679781e+00
-7.00826168e-01 7.37379551e-01 -8.60167384e-01 4.39237118e-01
1.62339732e-02 6.48264408e-01 8.36956799e-01 -7.50601470e-01
5.71650267e-01 1.27064800e+00 3.59081179e-01 7.01397717e-01
-1.39627409e+00 -7.50808835e-01 6.45673215e-01 1.46083117e-01
-1.54553401e+00 -5.00943184e-01 9.35613394e-01 -5.08052945e-01
1.98218539e-01 3.57501686e-01 5.93917251e-01 1.17005193e+00
-2.54109859e-01 1.19695425e+00 7.26521730e-01 5.98850995e-02
8.33133161e-02 6.14758581e-03 1.17210500e-01 4.92973834e-01
2.40095943e-01 -8.95668715e-02 -3.36171329e-01 2.41717130e-01
9.03221250e-01 9.24898908e-02 -5.89028299e-01 -4.23849702e-01
-1.30393755e+00 4.77347463e-01 4.95914400e-01 4.38367933e-01
-3.95476878e-01 3.04363221e-01 2.55279183e-01 -6.96765110e-02
2.84860224e-01 2.64041781e-01 -4.42049801e-01 1.63722679e-01
-1.28617537e+00 2.73390085e-01 2.88359493e-01 1.07795787e+00
6.82666421e-01 2.16645747e-01 -7.45006919e-01 7.94519961e-01
4.22384113e-01 2.02389836e-01 3.82855624e-01 -1.28629529e+00
5.20284295e-01 4.87372756e-01 1.52058020e-01 -1.07383227e+00
-7.47168958e-02 -5.75668633e-01 -7.73189545e-01 8.10391158e-02
5.30509353e-01 -1.36055559e-01 -1.30555987e+00 1.78640616e+00
5.44178426e-01 6.80458188e-01 -1.33414522e-01 1.19020796e+00
9.30521548e-01 6.76033318e-01 2.71560133e-01 -2.09907874e-01
1.10236275e+00 -1.24396443e+00 -5.12554944e-01 -2.02468649e-01
2.57796168e-01 -5.63700557e-01 1.06886995e+00 2.76159227e-01
-1.04077089e+00 -8.83320153e-01 -9.86937463e-01 2.70366430e-01
-6.29752129e-02 2.73428299e-02 1.30104512e-01 5.66348135e-01
-9.02468741e-01 5.75828612e-01 -8.29450130e-01 4.95707691e-02
9.07933772e-01 4.13553149e-01 -3.69413719e-02 1.32158548e-01
-8.98745954e-01 2.31704533e-01 6.62531912e-01 1.92800879e-01
-1.36488330e+00 -7.79964447e-01 -7.73978055e-01 -1.26133084e-01
7.17614710e-01 -5.26310802e-01 8.92887890e-01 -1.37324893e+00
-1.38022983e+00 7.09707916e-01 9.01067778e-02 -4.99287874e-01
7.95843601e-01 -1.32813841e-01 -3.85179818e-01 3.85060340e-01
1.27476349e-01 1.20881736e+00 1.24976921e+00 -1.74428368e+00
-7.27768540e-01 -2.81380694e-02 1.28181875e-01 5.12753278e-02
-2.89398253e-01 1.72660574e-01 -1.18136811e+00 -9.52197313e-01
1.75978482e-01 -8.79293740e-01 -4.03931499e-01 6.70263991e-02
-8.69819164e-01 -9.57462862e-02 1.01339889e+00 -5.81615210e-01
1.40124118e+00 -2.19111490e+00 2.58089364e-01 1.91235498e-01
3.28534484e-01 5.77125609e-01 -4.16974097e-01 -2.57179618e-01
-1.69197589e-01 2.90076673e-01 -4.04993206e-01 -3.57872933e-01
-2.11326003e-01 1.07279472e-01 -1.92340374e-01 4.60627794e-01
4.50161040e-01 1.01232147e+00 -8.66916239e-01 -7.99531221e-01
3.30465585e-01 4.19481277e-01 -6.46317601e-01 1.98451534e-01
-5.53160131e-01 7.26924121e-01 -5.51878810e-01 8.92549992e-01
6.64632380e-01 -2.63978928e-01 -9.29308534e-02 -3.28698039e-01
1.31049566e-02 -2.77011722e-01 -1.31381285e+00 1.41329503e+00
1.56202123e-01 5.14305234e-01 1.01129271e-01 -1.02476943e+00
7.63678372e-01 9.58100185e-02 8.18023860e-01 -3.08929354e-01
2.93851942e-01 -1.79741010e-02 9.67223123e-02 -5.48488617e-01
4.14477676e-01 4.34105963e-01 1.30735323e-01 -1.19499490e-01
-1.41736232e-02 9.80801508e-02 4.65242296e-01 2.18794361e-01
7.09540665e-01 3.27678800e-01 -1.74005687e-01 -2.04563111e-01
6.78356051e-01 -2.42946252e-01 8.53802919e-01 6.52602255e-01
-4.93263423e-01 1.12426829e+00 5.25734961e-01 -3.18226159e-01
-8.33558202e-01 -1.03503907e+00 -2.11215749e-01 8.63865674e-01
5.75452864e-01 -2.57644922e-01 -9.83477831e-01 -9.65076029e-01
-7.36840516e-02 6.40523016e-01 -6.31744266e-01 -9.63490158e-02
-7.31963634e-01 -5.85929990e-01 4.01487440e-01 5.93127787e-01
6.00611269e-01 -1.25744569e+00 -4.55876052e-01 2.56836683e-01
-1.59348354e-01 -1.24357915e+00 -8.36417735e-01 -1.29797727e-01
-7.63707638e-01 -9.99913394e-01 -9.88007128e-01 -8.61435235e-01
8.48078728e-01 2.96318024e-01 8.29897940e-01 1.94869444e-01
-8.27285051e-02 2.03176215e-01 -2.47846916e-01 8.62480104e-02
-3.63975376e-01 -2.00668555e-02 1.02517240e-01 6.13309205e-01
-5.14030755e-02 -5.46583533e-01 -9.06618118e-01 6.38062239e-01
-1.11766756e+00 3.13230008e-02 5.02667487e-01 6.13297760e-01
9.96490002e-01 -1.69186518e-01 4.59976375e-01 -7.32291400e-01
8.32871571e-02 -5.17457962e-01 -6.97747052e-01 1.30036980e-01
-1.42678902e-01 -3.15430671e-01 5.39391160e-01 -9.44692492e-01
-6.76607966e-01 2.76354820e-01 2.54939869e-02 -1.01684368e+00
-2.84315795e-01 1.01258561e-01 -5.61909318e-01 -1.41042009e-01
6.19632840e-01 1.81250960e-01 -3.07091147e-01 -2.60609388e-01
4.34509605e-01 4.07291889e-01 6.85297549e-01 -6.62026048e-01
9.99500275e-01 4.59961951e-01 -4.00040090e-01 -5.97505510e-01
-7.16539919e-01 -4.14853722e-01 -5.53558469e-01 -5.71301997e-01
1.05166733e+00 -7.74474561e-01 -4.37085420e-01 4.91393358e-01
-1.03368354e+00 -5.89454293e-01 -3.39255601e-01 1.94118410e-01
-4.29062724e-01 3.27351272e-01 -3.42056900e-01 -5.86620212e-01
-1.13842292e-02 -1.51005924e+00 1.02775478e+00 6.00513875e-01
-1.76467761e-01 -5.41427433e-01 -2.41367415e-01 3.57532829e-01
1.07525259e-01 4.32598442e-01 3.51387948e-01 -6.84874415e-01
-9.98073280e-01 -1.69967413e-02 -3.19850773e-01 4.96788293e-01
1.91933364e-01 2.92126536e-01 -7.52753794e-01 -3.03600937e-01
-2.40514979e-01 -7.44783357e-02 1.05529892e+00 6.04546010e-01
1.63826787e+00 -3.79682422e-01 -4.44025695e-01 9.93587375e-01
1.16801834e+00 3.28298360e-01 5.57161510e-01 2.54287809e-01
1.09920776e+00 4.06175613e-01 6.32265806e-01 2.25620925e-01
1.40928447e-01 7.95540810e-01 5.34125209e-01 -3.16878408e-02
-3.49494308e-01 -1.56399235e-01 2.63672382e-01 4.27872390e-01
-1.22191370e-01 -4.06429261e-01 -7.21008182e-01 7.33779550e-01
-1.79789329e+00 -8.61211360e-01 -1.88560262e-02 2.02317095e+00
8.44822526e-01 3.61611933e-01 3.92201275e-01 -4.26567830e-02
1.09449530e+00 5.02834737e-01 -7.95788169e-01 3.46865475e-01
-2.95053244e-01 -1.67297408e-01 4.89611179e-01 2.89403528e-01
-1.32771540e+00 1.19968283e+00 5.15089941e+00 9.81725097e-01
-1.14131999e+00 2.20706090e-01 1.07651484e+00 -2.35411540e-01
-1.32138401e-01 -1.47310421e-01 -8.23948562e-01 8.00390005e-01
4.42471474e-01 6.49727434e-02 3.04187387e-01 8.05187285e-01
3.07523847e-01 4.46432978e-02 -9.81249809e-01 8.78054678e-01
-1.40650049e-01 -1.47410429e+00 2.74591357e-01 -2.19201282e-01
1.17655337e+00 -6.93931133e-02 2.23868728e-01 1.09478496e-01
6.30962029e-02 -7.00595617e-01 1.11938214e+00 5.16480267e-01
5.88309050e-01 -5.45338511e-01 2.96038330e-01 5.96283078e-02
-1.36873305e+00 -2.00070038e-01 -1.85924143e-01 5.76250315e-01
2.89135307e-01 2.14367121e-01 -3.64360452e-01 3.03454131e-01
8.89294088e-01 7.10086107e-01 -5.23849189e-01 1.30598021e+00
-8.45056996e-02 8.09906900e-01 -2.31920645e-01 2.38080099e-01
3.58353227e-01 -2.05129638e-01 8.37520003e-01 1.17028439e+00
6.22589365e-02 1.03139520e-01 4.68517840e-01 1.02976918e+00
-2.29805127e-01 1.02751572e-02 -1.45232439e-01 7.44381398e-02
5.07069707e-01 1.27578330e+00 -1.21032774e+00 -5.14211118e-01
-1.66585088e-01 8.46196711e-01 7.07217678e-02 6.86277449e-01
-1.21314597e+00 -3.60109210e-02 8.09625149e-01 1.86179399e-01
7.57276177e-01 -2.04085439e-01 -3.15923214e-01 -9.75648165e-01
1.88783512e-01 -9.71414506e-01 3.29169482e-01 -5.61372578e-01
-1.02228236e+00 6.56266272e-01 2.17908069e-01 -1.40205026e+00
1.87858313e-01 -3.37297499e-01 -7.11337149e-01 4.58943307e-01
-1.24677753e+00 -1.03068542e+00 -4.50048208e-01 6.04160964e-01
9.47778761e-01 4.84705269e-02 -5.01762331e-02 4.28623319e-01
-9.67689753e-01 7.36817479e-01 -1.90215096e-01 3.72727871e-01
4.08949673e-01 -8.32271814e-01 3.75260055e-01 1.16997802e+00
3.41891497e-01 4.48076755e-01 6.32203400e-01 -8.43485415e-01
-1.02077937e+00 -1.58695424e+00 1.14007570e-01 -4.56762612e-01
6.16036117e-01 -3.18670869e-01 -1.14391446e+00 6.07192874e-01
5.06313220e-02 5.68222344e-01 -6.31713355e-03 -6.25624955e-01
-2.29423895e-01 -2.60501504e-01 -1.02529776e+00 8.50928366e-01
1.30392182e+00 -1.76306620e-01 -3.17630947e-01 4.51329678e-01
1.07537198e+00 -6.19598687e-01 -5.13560414e-01 6.05541706e-01
3.16677839e-01 -8.62498105e-01 1.07567835e+00 -5.88450253e-01
2.97944397e-01 -8.89745295e-01 1.52220754e-02 -8.73461485e-01
-3.06334227e-01 -9.65256572e-01 -3.06653142e-01 1.47190452e+00
3.15066904e-01 -3.38315398e-01 9.09348190e-01 5.43508947e-01
-3.79753977e-01 -8.96010518e-01 -8.33089888e-01 -9.35563028e-01
-2.88791895e-01 -6.19467437e-01 4.71049815e-01 8.68408501e-01
-6.32369697e-01 -2.31923491e-01 -3.38548720e-01 5.31533659e-01
5.94204128e-01 4.07089014e-04 7.64724314e-01 -7.93916821e-01
-4.20684993e-01 -7.74176657e-01 -5.06390989e-01 -1.37690091e+00
5.55439778e-02 -7.38747060e-01 2.65829861e-01 -1.18084347e+00
7.74167245e-03 -5.31946421e-01 -5.15668273e-01 4.48261291e-01
-5.49342632e-01 6.74822986e-01 4.76559132e-01 3.29955518e-01
-9.25405920e-01 4.67957675e-01 1.40336692e+00 -4.17423218e-01
-5.41918099e-01 1.61335260e-01 -6.29208028e-01 8.30908120e-01
7.08402455e-01 -5.99471033e-01 -3.56479317e-01 -3.43481123e-01
-4.90598708e-01 -1.52399793e-01 5.30626297e-01 -1.18372250e+00
1.33490399e-01 -3.11374426e-01 5.07545769e-01 -6.08211517e-01
2.72099346e-01 -6.43408418e-01 2.87636429e-01 4.39343393e-01
-3.40697736e-01 -2.47831032e-01 9.46131721e-02 5.88313282e-01
-1.46578059e-01 -2.12622449e-01 1.02345574e+00 2.32730005e-02
-9.00449634e-01 8.73716056e-01 -7.73079023e-02 3.84568810e-01
1.27513480e+00 -5.07645667e-01 -1.44879557e-02 -1.04542226e-01
-6.81197643e-01 4.58615869e-01 4.15048808e-01 6.24926209e-01
6.40099823e-01 -1.23391902e+00 -6.65727317e-01 5.69016002e-02
1.67122371e-02 3.76329422e-01 3.53578091e-01 8.71002376e-01
-4.00234073e-01 -1.15336604e-01 -1.45007461e-01 -1.00433123e+00
-1.12933826e+00 8.04469228e-01 3.52852583e-01 2.74613053e-01
-6.19043469e-01 1.10301757e+00 5.87964356e-01 2.19326928e-01
3.82504195e-01 -5.05810559e-01 -2.74763107e-01 -4.59499583e-02
4.26625699e-01 1.25243589e-01 -4.58232492e-01 -9.47258115e-01
-2.94429123e-01 7.69565225e-01 -9.43958834e-02 5.44867218e-02
1.13126421e+00 -1.42358363e-01 2.71449238e-01 2.50610262e-01
1.06445658e+00 -2.54359674e-02 -1.94872320e+00 -2.56102324e-01
1.21537028e-02 -6.02614462e-01 -1.99254975e-01 -3.95598054e-01
-1.64836371e+00 4.90786105e-01 5.27838707e-01 1.44503742e-01
1.12106574e+00 2.15065002e-01 7.63141870e-01 -7.07623586e-02
3.76686454e-02 -9.39682841e-01 4.01369393e-01 2.23689660e-01
8.44451249e-01 -1.10756218e+00 -1.79829404e-01 -5.05513787e-01
-6.98966086e-01 8.29379320e-01 9.59654391e-01 -2.38785014e-01
3.84325683e-01 1.32359132e-01 1.00126460e-01 6.24128915e-02
-2.97384173e-01 -2.68480390e-01 5.73122978e-01 6.24718487e-01
-4.05593216e-03 -1.05291925e-01 -2.63684429e-02 7.19804704e-01
1.77239448e-01 -2.34406158e-01 1.72769919e-01 4.96549696e-01
-3.53915155e-01 -8.47901464e-01 -5.05837023e-01 3.51952881e-01
-4.62014884e-01 -6.16255216e-02 -1.64635181e-01 7.08828092e-01
3.90245080e-01 7.39221871e-01 -3.04000191e-02 -4.98063952e-01
2.22261161e-01 -2.32027024e-01 1.59605712e-01 -4.41381544e-01
-5.13962746e-01 3.08379501e-01 -1.97940558e-01 -7.17445731e-01
-5.77835441e-01 -9.18748438e-01 -1.39426994e+00 1.85006291e-01
-3.65620971e-01 -4.34330888e-02 2.27629289e-01 8.44938278e-01
3.25784713e-01 7.59307504e-01 6.34775281e-01 -1.33941317e+00
-2.05762297e-01 -6.50093079e-01 -1.22942157e-01 5.73698699e-01
3.41381729e-01 -6.70164406e-01 -2.29421109e-01 3.64179671e-01] | [9.26196002960205, -0.08121488988399506] |
33e7e880-3dfd-497e-925c-69d76524ea6b | automated-mining-of-leaderboards-for | 2109.13089 | null | https://arxiv.org/abs/2109.13089v1 | https://arxiv.org/pdf/2109.13089v1.pdf | Automated Mining of Leaderboards for Empirical AI Research | With the rapid growth of research publications, empowering scientists to keep oversight over the scientific progress is of paramount importance. In this regard, the Leaderboards facet of information organization provides an overview on the state-of-the-art by aggregating empirical results from various studies addressing the same research challenge. Crowdsourcing efforts like PapersWithCode among others are devoted to the construction of Leaderboards predominantly for various subdomains in Artificial Intelligence. Leaderboards provide machine-readable scholarly knowledge that has proven to be directly useful for scientists to keep track of research progress. The construction of Leaderboards could be greatly expedited with automated text mining. This study presents a comprehensive approach for generating Leaderboards for knowledge-graph-based scholarly information organization. Specifically, we investigate the problem of automated Leaderboard construction using state-of-the-art transformer models, viz. Bert, SciBert, and XLNet. Our analysis reveals an optimal approach that significantly outperforms existing baselines for the task with evaluation scores above 90% in F1. This, in turn, offers new state-of-the-art results for Leaderboard extraction. As a result, a vast share of empirical AI research can be organized in the next-generation digital libraries as knowledge graphs. | ['Sören Auer', "Jennifer D'Souza", 'Salomon Kabongo'] | 2021-08-31 | null | null | null | null | ['scientific-results-extraction'] | ['natural-language-processing'] | [-3.51463646e-01 5.04136622e-01 -8.45569849e-01 1.58758894e-01
-8.91548097e-01 -9.97014940e-01 8.53948116e-01 6.66926086e-01
-2.25989912e-02 8.85024548e-01 2.83948630e-01 -4.81687039e-01
-4.44743007e-01 -1.04004455e+00 -8.54583263e-01 -1.08044684e-01
1.37262195e-01 6.93276167e-01 4.03481424e-02 -8.84592086e-02
8.28200758e-01 5.73019207e-01 -1.08844924e+00 -3.32455747e-02
1.04200232e+00 6.34202719e-01 1.83416739e-01 5.33804297e-01
-5.20751655e-01 1.46862066e+00 -8.16146433e-01 -8.95144224e-01
-2.42750589e-02 -7.96927363e-02 -1.17659283e+00 -1.94748580e-01
7.00188696e-01 2.32614011e-01 -4.37048078e-01 1.05909944e+00
1.56448647e-01 -3.16445641e-02 4.83346283e-01 -1.60895538e+00
-1.14304936e+00 9.42603648e-01 -3.51110399e-01 2.87662297e-01
5.56685150e-01 -1.45790085e-01 1.55980241e+00 -7.40629792e-01
1.41071403e+00 1.02957523e+00 5.19687712e-01 -1.03807807e-01
-8.56861830e-01 -5.96885324e-01 1.35191873e-01 5.78442216e-01
-1.34791696e+00 -3.59321654e-01 8.76458347e-01 -7.15084255e-01
7.50575244e-01 4.64321882e-01 7.22603440e-01 9.24110889e-01
5.65081835e-02 8.45973969e-01 1.16954982e+00 -4.49361086e-01
2.09135607e-01 3.86783332e-02 5.05917192e-01 1.14827025e+00
9.98880148e-01 -4.49050605e-01 -9.82999682e-01 -2.08694518e-01
7.62787938e-01 -4.15237516e-01 -1.53422132e-01 1.63349897e-01
-1.47980511e+00 5.92029572e-01 4.20933783e-01 3.94302011e-01
-3.59044403e-01 8.70154724e-02 1.18460707e-01 1.90399200e-01
5.19836426e-01 1.28473604e+00 5.41516244e-02 -4.45254952e-01
-1.16069067e+00 5.89564621e-01 1.40291214e+00 1.35853505e+00
6.38109505e-01 -1.10733181e-01 -4.72876370e-01 3.86386424e-01
1.35813221e-01 1.36424378e-01 3.51369604e-02 -1.13149261e+00
6.83875144e-01 1.21994698e+00 1.20118029e-01 -1.26594484e+00
-3.38333905e-01 -7.69115329e-01 -5.51740944e-01 -3.41723949e-01
3.50041807e-01 1.02600873e-01 -5.49350202e-01 9.32118654e-01
3.38052928e-01 1.05693124e-01 -2.04026788e-01 7.25566983e-01
1.38844180e+00 3.90379786e-01 -9.25233439e-02 2.03833893e-01
1.48066127e+00 -1.22645247e+00 -9.73531723e-01 -7.10534677e-02
4.73382831e-01 -7.65928626e-01 7.10102439e-01 4.51521456e-01
-9.12478149e-01 -3.12500298e-01 -9.38887894e-01 -4.79080439e-01
-6.96994364e-01 2.59149760e-01 8.48133206e-01 3.18567902e-01
-1.09215438e+00 7.49690711e-01 -4.87751454e-01 -1.36525258e-01
8.49808156e-01 -7.00291544e-02 -4.54097211e-01 -1.62204310e-01
-9.41484571e-01 1.05272031e+00 1.41213283e-01 -2.88097799e-01
-9.14876759e-01 -9.90008593e-01 -4.32412624e-01 -5.83372749e-02
6.87264681e-01 -7.28085697e-01 1.00889134e+00 -2.29535580e-01
-9.17805493e-01 1.04185486e+00 -1.12577766e-01 -3.63164604e-01
4.33567405e-01 -3.08608532e-01 -1.66057751e-01 1.87910840e-01
5.43091059e-01 2.10716188e-01 4.73771036e-01 -1.39221609e+00
-5.41067123e-01 -3.25422347e-01 4.66567948e-02 -6.99027255e-02
-6.13351941e-01 1.51818439e-01 -6.83109939e-01 -5.80751657e-01
-3.88007909e-01 -7.13018000e-01 -8.03337470e-02 -1.70050979e-01
-8.62564802e-01 -6.56143010e-01 5.57068586e-01 -6.49881482e-01
1.46573603e+00 -1.41980290e+00 3.21306914e-01 1.52059961e-02
1.02530217e+00 -2.05491614e-02 -1.69765707e-02 7.33630359e-01
4.45575118e-01 7.37467229e-01 2.71582931e-01 -2.35182628e-01
1.43702090e-01 -3.95130403e-02 -4.09957349e-01 4.02248472e-01
8.06768462e-02 1.31037927e+00 -1.35309947e+00 -6.91440284e-01
-2.38807216e-01 -9.63297039e-02 1.91398501e-01 2.84502357e-01
-3.11045855e-01 1.26077816e-01 -8.87401640e-01 1.08951700e+00
5.29173724e-02 -7.95679867e-01 1.61262557e-01 1.42357945e-01
-5.35532415e-01 3.92919123e-01 -6.44173920e-01 1.39741254e+00
7.13484585e-02 1.07631350e+00 1.80044964e-01 -8.54688942e-01
1.10873353e+00 1.82246208e-01 4.22044933e-01 -3.13326895e-01
-6.53927475e-02 1.34810790e-01 -1.44287452e-01 -4.33817059e-01
7.43298829e-01 5.12848318e-01 2.05772638e-01 5.22151291e-01
-9.23240855e-02 -4.02462095e-01 6.42436028e-01 7.14496732e-01
1.52547312e+00 8.71725380e-02 4.25106406e-01 -4.04643357e-01
1.31934226e-01 5.45015335e-01 3.43194276e-01 9.60122287e-01
-3.66357863e-02 6.90302029e-02 6.18071616e-01 -4.11845624e-01
-1.02511966e+00 -5.50826609e-01 3.34377475e-02 9.88235056e-01
-1.04918152e-01 -8.80478442e-01 -5.88004529e-01 -6.35274529e-01
4.61387128e-01 6.69077873e-01 -7.17851698e-01 3.05558771e-01
-3.38399351e-01 -3.66715014e-01 8.13382566e-01 3.93967867e-01
3.80022198e-01 -9.22630191e-01 -1.98768541e-01 2.42564306e-01
-1.85737982e-01 -1.50568235e+00 -1.47297624e-02 -1.80596992e-01
-6.92913294e-01 -1.48508155e+00 -6.75989509e-01 -7.27607727e-01
6.90574646e-01 5.36251247e-01 1.61205590e+00 2.46890619e-01
-2.87792474e-01 4.79129970e-01 -5.88491261e-01 -9.73305404e-01
-2.90126234e-01 4.16921437e-01 -2.76494354e-01 -5.11008978e-01
4.55853790e-01 -3.92381072e-01 -3.76129448e-01 4.94155660e-02
-5.31637967e-01 6.12117797e-02 5.69241107e-01 3.68151158e-01
6.00468874e-01 -9.53452811e-02 8.12064707e-01 -1.12445068e+00
1.11390710e+00 -7.61928141e-01 -6.57372892e-01 5.30316174e-01
-8.68097484e-01 -8.11662823e-02 7.12763011e-01 3.29279043e-02
-6.57951355e-01 -4.95494604e-01 3.85859370e-01 -2.64582038e-01
1.48430869e-01 8.78351510e-01 2.08892211e-01 -4.94803458e-01
7.35550046e-01 -2.74722744e-02 -4.46781158e-01 -4.79207426e-01
7.24984825e-01 8.36493433e-01 4.53280836e-01 -9.84787464e-01
8.88313830e-01 3.91548455e-01 4.40035671e-01 -8.99740040e-01
-1.34581006e+00 -7.13744283e-01 -3.93716395e-01 -5.41802585e-01
5.51225662e-01 -9.48871493e-01 -8.35885286e-01 -2.35557109e-02
-1.21189988e+00 -5.54037653e-02 -2.39996418e-01 1.25494763e-01
-1.65791452e-01 3.92874122e-01 -4.77234393e-01 -5.74708700e-01
-5.84263384e-01 -9.27082956e-01 1.06204987e+00 3.21738899e-01
-5.10648549e-01 -1.09501874e+00 3.03153187e-01 1.00996757e+00
1.70662552e-01 3.71587217e-01 9.61965799e-01 -1.01703501e+00
-1.02022326e+00 -3.61055523e-01 -4.93954509e-01 -1.09752305e-01
-1.16278365e-01 1.72621876e-01 -8.39751244e-01 1.51217014e-01
-5.41494370e-01 -2.64092058e-01 7.04544783e-01 1.31472304e-01
1.33211815e+00 -4.41986740e-01 -6.24117792e-01 2.80099362e-01
1.03355289e+00 -9.90062281e-02 4.79816347e-01 8.29094589e-01
1.10167074e+00 6.81070626e-01 3.93656075e-01 3.84287804e-01
1.08726430e+00 4.08128142e-01 3.57788175e-01 -3.70005332e-02
-2.60355115e-01 -4.25357491e-01 -6.13578521e-02 1.14397442e+00
-6.65778458e-01 -4.06773418e-01 -1.39542294e+00 8.64935160e-01
-1.93596780e+00 -9.18703556e-01 -6.76649630e-01 1.84440541e+00
1.04617822e+00 3.65454634e-03 3.65542918e-02 -1.14210770e-01
5.36480606e-01 1.53052554e-01 -2.69932330e-01 -1.72414798e-02
-2.14999661e-01 -4.26013917e-02 6.11764431e-01 1.91989601e-01
-8.75912666e-01 1.20069933e+00 6.56766081e+00 8.40938866e-01
-5.30694127e-01 -5.85069880e-02 4.41191375e-01 2.32906118e-01
-5.05956471e-01 3.28624904e-01 -1.11179912e+00 3.15929234e-01
7.94232130e-01 -7.66883671e-01 5.28359652e-01 8.36268544e-01
8.38436559e-02 2.42556185e-02 -1.19823742e+00 8.13985229e-01
8.32044333e-02 -2.00261831e+00 1.80837467e-01 1.64809763e-01
1.22318554e+00 2.54402421e-02 -3.12575340e-01 5.45870811e-02
9.71670270e-01 -1.15356100e+00 6.39701903e-01 7.99271047e-01
5.74102402e-01 -2.98193455e-01 5.63888907e-01 3.20354670e-01
-1.11279035e+00 3.24480645e-02 -4.59318608e-01 -2.22968519e-01
-1.10001691e-01 1.05110908e+00 -1.01222634e+00 9.46291566e-01
4.88499582e-01 1.07313871e+00 -7.44163215e-01 1.17255342e+00
-7.41556764e-01 9.66228008e-01 1.07347965e-01 -4.87104803e-01
1.68274865e-01 -1.84734523e-01 7.25849509e-01 1.34406555e+00
1.36011124e-01 -3.86005901e-02 4.07212019e-01 1.06293619e+00
-1.00050938e+00 2.18047306e-01 -8.65659952e-01 -8.30084085e-01
9.23405707e-01 1.57208300e+00 -7.14661121e-01 -4.94792938e-01
-3.39396149e-01 2.57675052e-01 9.03124154e-01 1.65616572e-01
-2.07056835e-01 -5.37003458e-01 2.12974146e-01 1.47501096e-01
-9.89393368e-02 -3.05115938e-01 -5.95257401e-01 -9.77508485e-01
1.90527346e-02 -8.27232599e-01 1.54868990e-01 -7.55476058e-01
-1.48369718e+00 2.83725590e-01 -1.68807447e-01 -7.33146071e-01
6.89186752e-02 -6.17672563e-01 -5.80097318e-01 5.90286672e-01
-1.78040159e+00 -1.31856680e+00 -6.19914889e-01 1.45114213e-01
2.81208932e-01 -3.80863190e-01 6.99993491e-01 -1.62999138e-01
-7.62809277e-01 2.03758061e-01 1.90830126e-01 3.30387175e-01
7.01702356e-01 -1.56317806e+00 6.91546500e-01 6.48559809e-01
4.68279213e-01 7.80538201e-01 3.69545877e-01 -1.19684398e+00
-2.03886700e+00 -1.15091312e+00 1.15480244e+00 -1.07283986e+00
1.31413329e+00 -1.72066674e-01 -8.16994905e-01 7.12089062e-01
2.23875016e-01 -2.50159413e-01 7.71671593e-01 4.67902601e-01
-3.29948992e-01 3.37877460e-02 -4.91749525e-01 5.65729976e-01
1.21323049e+00 -6.48698509e-01 -8.03421378e-01 8.16212177e-01
4.57937300e-01 -2.90066957e-01 -1.45546961e+00 -2.46973619e-01
4.19976324e-01 -3.06149065e-01 7.28409827e-01 -8.79070163e-01
9.77906346e-01 -9.75191072e-02 3.33808422e-01 -1.17236161e+00
-6.00731015e-01 -1.04251289e+00 -5.24499357e-01 1.24531400e+00
3.22528809e-01 -3.14337820e-01 8.95460308e-01 7.15328813e-01
-3.22164387e-01 -8.12996268e-01 -7.37438500e-01 -8.37365329e-01
9.08458531e-02 -1.16610527e-01 5.48833072e-01 1.29805756e+00
3.65241438e-01 6.41017497e-01 -5.97446822e-02 -2.21450359e-01
8.64671767e-01 4.48249549e-01 1.12951255e+00 -1.87649286e+00
2.85381466e-01 -7.71130025e-01 -4.21729326e-01 -7.76016653e-01
3.68478298e-01 -1.38090181e+00 -4.05028820e-01 -2.34351349e+00
5.59029818e-01 -3.43728840e-01 -1.12913184e-01 7.62446940e-01
-4.19065505e-01 -9.21042487e-02 2.25307152e-01 4.69221741e-01
-1.02659440e+00 2.42383540e-01 1.66686451e+00 -2.19793931e-01
-5.46888635e-03 -5.38063407e-01 -1.26673591e+00 6.17630005e-01
6.15004957e-01 -4.77169275e-01 -9.40272212e-02 -3.58631730e-01
8.60346496e-01 -1.07545905e-01 4.33885604e-01 -5.75245738e-01
8.55151057e-01 -2.31551096e-01 6.14562966e-02 -2.85922050e-01
-1.86944026e-02 -4.65935111e-01 5.23427874e-02 9.11303912e-04
-4.05579150e-01 1.79880306e-01 9.96376500e-02 6.02740526e-01
-2.27374747e-01 -3.41411114e-01 -1.30066037e-01 -3.22158307e-01
-4.89050120e-01 1.36523098e-01 -6.91834018e-02 4.10930037e-01
8.80590916e-01 1.00892093e-02 -1.21591723e+00 -3.23991865e-01
-5.21845073e-02 3.77238929e-01 3.39028001e-01 4.24670786e-01
3.83269757e-01 -1.04097176e+00 -1.16772580e+00 -5.45498908e-01
2.33495206e-01 -8.03627521e-02 -2.91799515e-01 6.85523272e-01
-5.55274963e-01 6.87745094e-01 -8.83769915e-02 1.46470308e-01
-9.83635485e-01 3.24424356e-01 -3.53596061e-01 -6.86064601e-01
-5.02769113e-01 8.07494283e-01 -3.61394823e-01 3.38453725e-02
9.70042497e-02 -2.30258238e-02 -3.85451853e-01 2.72409439e-01
4.75759387e-01 1.01258278e+00 6.83109090e-02 -3.16155046e-01
-2.50909030e-01 2.77100503e-01 -2.00824067e-01 3.41795027e-01
1.61357605e+00 4.54811424e-01 -6.18438125e-01 3.60547245e-01
7.08478272e-01 4.33398247e-01 -5.18703520e-01 3.35411802e-02
4.36352760e-01 -4.81776297e-01 3.88785094e-01 -1.04188609e+00
-7.60321617e-01 4.82660830e-01 -5.07301927e-01 4.84443665e-01
3.40510011e-01 1.91298187e-01 4.19999629e-01 8.75074387e-01
4.71610576e-01 -1.24308467e+00 -3.19137499e-02 3.47804010e-01
1.10382915e+00 -1.25892580e+00 7.26534128e-01 -6.16567016e-01
-4.84068990e-01 1.24596834e+00 2.95794845e-01 1.58614248e-01
4.87685084e-01 2.99858004e-01 -2.19336972e-01 -7.91306555e-01
-7.47519851e-01 -1.46030951e-02 8.58444750e-01 3.71487945e-01
6.13585830e-01 3.40505511e-01 -3.11637968e-01 6.99034572e-01
-4.65510070e-01 1.38858750e-01 5.99460244e-01 8.80432010e-01
-5.63034952e-01 -9.66442466e-01 -4.34124440e-01 9.17080820e-01
-6.50767326e-01 -1.92521885e-01 -1.26711857e+00 8.05853128e-01
-3.71680349e-01 1.20686185e+00 -4.93215531e-01 -2.31038570e-01
3.75074029e-01 -2.21434265e-01 2.28637993e-01 -6.53081656e-01
-8.77686262e-01 -4.20865625e-01 6.66427016e-01 -3.89818817e-01
-5.14031351e-01 -3.40855092e-01 -9.76388812e-01 -4.85356539e-01
-3.23301017e-01 3.77602309e-01 8.00039411e-01 7.32596278e-01
6.81492507e-01 6.87857926e-01 1.30645990e-01 -4.54009563e-01
-1.86245993e-01 -9.23006654e-01 -3.49586517e-01 1.74469143e-01
-2.34833688e-01 -6.36708140e-01 -9.32965353e-02 1.44165695e-01] | [9.6219482421875, 8.24588394165039] |
665a5995-1ce6-45d6-9926-816c1c226381 | towards-zero-resource-cross-lingual-entity | 1909.13180 | null | https://arxiv.org/abs/1909.13180v2 | https://arxiv.org/pdf/1909.13180v2.pdf | Towards Zero-resource Cross-lingual Entity Linking | Cross-lingual entity linking (XEL) grounds named entities in a source language to an English Knowledge Base (KB), such as Wikipedia. XEL is challenging for most languages because of limited availability of requisite resources. However, much previous work on XEL has been on simulated settings that actually use significant resources (e.g. source language Wikipedia, bilingual entity maps, multilingual embeddings) that are unavailable in truly low-resource languages. In this work, we first examine the effect of these resource assumptions and quantify how much the availability of these resource affects overall quality of existing XEL systems. Next, we propose three improvements to both entity candidate generation and disambiguation that make better use of the limited data we do have in resource-scarce scenarios. With experiments on four extremely low-resource languages, we show that our model results in gains of 6-23% in end-to-end linking accuracy. | ['Shuyan Zhou', 'Shruti Rijhwani', 'Graham Neubig'] | 2019-09-29 | towards-zero-resource-cross-lingual-entity-1 | https://aclanthology.org/D19-6127 | https://aclanthology.org/D19-6127.pdf | ws-2019-11 | ['cross-lingual-entity-linking'] | ['natural-language-processing'] | [-6.91704392e-01 1.44111648e-01 -5.46739638e-01 -9.19135064e-02
-1.20991194e+00 -8.77442956e-01 5.92493296e-01 3.33031178e-01
-9.51357126e-01 1.46456552e+00 4.75436687e-01 -3.43688756e-01
1.03018589e-01 -8.23115706e-01 -8.36055100e-01 3.80394220e-01
-1.09212846e-01 7.47540474e-01 4.13457751e-01 -4.72682029e-01
-1.87009498e-01 2.05812544e-01 -1.07521820e+00 -3.88975926e-02
1.17226183e+00 1.24857396e-01 1.01637781e-01 2.79766917e-01
-5.46203017e-01 4.36679542e-01 -3.86608809e-01 -7.00589597e-01
3.10552508e-01 -9.62052792e-02 -1.05327690e+00 -7.62462258e-01
4.26240534e-01 -8.58925879e-02 -1.68839678e-01 1.07673991e+00
7.89154351e-01 6.53376728e-02 3.24834347e-01 -1.28107941e+00
-6.87764049e-01 1.03760862e+00 -4.32027280e-01 2.72490472e-01
5.93069732e-01 -1.30076051e-01 1.00669885e+00 -1.21831489e+00
1.38067198e+00 9.36906636e-01 1.04666889e+00 3.15851539e-01
-1.03015530e+00 -7.02504933e-01 -1.32669613e-01 -3.45818698e-02
-1.61447644e+00 -7.94899464e-01 9.40562859e-02 -1.53446153e-01
1.45638025e+00 -1.29165694e-01 1.11681156e-01 7.10373521e-01
-2.17649609e-01 3.09459627e-01 1.04489875e+00 -8.19793403e-01
-2.99152076e-01 5.46864092e-01 1.97933868e-01 6.57435775e-01
9.59918201e-01 -1.81058913e-01 -7.16912091e-01 -2.93154538e-01
4.35994595e-01 -8.77049387e-01 -2.03058109e-01 -2.20831156e-01
-1.39884841e+00 4.86456156e-01 4.19565827e-01 3.27937335e-01
-2.73936838e-01 -4.53445986e-02 4.65168089e-01 2.61254340e-01
5.09794831e-01 8.85811627e-01 -7.52946436e-01 -2.25685894e-01
-9.21686351e-01 1.52623802e-01 1.22571564e+00 1.36351824e+00
9.26697910e-01 -2.52881914e-01 3.10609609e-01 9.11843777e-01
9.68280360e-02 7.35929787e-01 3.04955870e-01 -6.52104855e-01
1.05614889e+00 3.52823883e-01 6.04621351e-01 -7.15091884e-01
-5.25298119e-01 -1.57499015e-01 -1.65145338e-01 -1.89442441e-01
6.43053174e-01 -6.89354718e-01 -6.48213446e-01 2.11131668e+00
4.06342298e-01 6.55780509e-02 4.95629102e-01 6.13926113e-01
8.99436891e-01 3.98068607e-01 4.34533507e-01 2.48581506e-02
1.58009672e+00 -8.01635563e-01 -7.04554975e-01 -3.37502986e-01
1.05834770e+00 -9.46101964e-01 9.69374120e-01 -5.18289089e-01
-1.04967594e+00 -2.61606842e-01 -8.87728274e-01 -3.42351407e-01
-8.58816028e-01 2.61143357e-01 8.65465462e-01 5.67402661e-01
-1.19688559e+00 3.76047969e-01 -8.82917285e-01 -9.66012716e-01
-5.53177856e-02 3.63571733e-01 -9.86889660e-01 -1.35938793e-01
-1.84328485e+00 1.47868454e+00 7.28382230e-01 -1.70619980e-01
-1.52594605e-02 -9.23157573e-01 -8.68377030e-01 -1.50083631e-01
5.41955769e-01 -7.88203776e-01 1.04018068e+00 -5.51697731e-01
-6.86762750e-01 8.16210747e-01 -9.07397643e-02 -3.39665830e-01
4.58530694e-01 -4.25989807e-01 -7.32512116e-01 -1.39611989e-01
6.85464859e-01 8.45686495e-01 -2.57186323e-01 -1.03972208e+00
-6.46040678e-01 2.33643632e-02 2.59348124e-01 3.07396978e-01
-3.84919912e-01 3.06870520e-01 -9.10357118e-01 -4.52270299e-01
-4.30245757e-01 -1.06148243e+00 -1.13664716e-01 -3.11182022e-01
-4.12730694e-01 -1.81666780e-02 2.18977243e-01 -1.08101785e+00
1.31286585e+00 -1.64456093e+00 -2.74058104e-01 2.29595169e-01
-1.97315603e-01 3.21899831e-01 -1.80602536e-01 8.90375495e-01
-1.39197726e-02 5.90438604e-01 1.89269651e-02 5.79052940e-02
8.76566470e-02 8.59703049e-02 2.68363208e-02 3.74469459e-02
2.41891384e-01 9.41814303e-01 -1.05735707e+00 -8.83814335e-01
-2.99205363e-01 3.66257489e-01 -5.52962065e-01 -2.26923570e-01
-7.24706277e-02 -2.55039893e-02 -4.67417628e-01 6.82234347e-01
4.27782178e-01 -1.39054626e-01 6.36344850e-01 -4.03694123e-01
-3.38958770e-01 7.15117931e-01 -1.28639054e+00 1.87549233e+00
-8.02537739e-01 5.61846673e-01 -9.83949229e-02 -5.22483923e-02
4.13993716e-01 3.50337833e-01 2.23174155e-01 -6.06748044e-01
-3.13810915e-01 7.06705689e-01 8.37390404e-03 -3.02011102e-01
9.29459572e-01 -1.92095563e-02 -2.59460390e-01 5.14830589e-01
3.20552081e-01 4.24532324e-01 7.09375381e-01 5.29556513e-01
1.13966525e+00 2.67939448e-01 5.07580042e-01 -3.96822721e-01
1.24657370e-01 5.65392733e-01 7.29922235e-01 6.25161469e-01
-6.67038932e-02 -3.75276096e-02 2.27058187e-01 8.81759748e-02
-1.44427967e+00 -9.03203547e-01 -2.55292565e-01 8.89351428e-01
1.50934294e-01 -6.46617651e-01 -6.09333277e-01 -7.47236907e-01
1.97423592e-01 8.51395607e-01 -1.50520787e-01 1.50187656e-01
-6.95902109e-01 -9.62398946e-01 1.14911103e+00 6.61788106e-01
4.34451997e-01 -1.01192188e+00 -1.39034122e-01 3.51935267e-01
-3.88005108e-01 -1.43459058e+00 -3.18442196e-01 1.58636887e-02
-4.84591365e-01 -8.75710785e-01 -3.59339923e-01 -7.06006885e-01
5.18859327e-01 -7.08730593e-02 1.67365646e+00 -1.61210999e-01
-1.97169006e-01 2.99296081e-01 -2.93680400e-01 -1.54592767e-01
-3.20654571e-01 6.79762185e-01 4.45924520e-01 -8.84633899e-01
7.04095185e-01 -3.47564608e-01 -9.37209725e-02 1.92290545e-02
-6.26626015e-01 -5.27108610e-02 7.06069052e-01 7.24013209e-01
4.06975776e-01 -1.04567766e-01 9.71953869e-01 -1.48600924e+00
6.58078432e-01 -8.26141536e-01 -5.02224088e-01 6.91588402e-01
-8.68876636e-01 3.00558269e-01 4.24340308e-01 -1.05810910e-01
-1.34112716e+00 -2.89271951e-01 -2.67342124e-02 1.74777493e-01
1.38404831e-01 1.02579939e+00 -3.16984296e-01 -6.83403164e-02
6.50711715e-01 -4.94680226e-01 -5.51720142e-01 -7.08607554e-01
7.47096360e-01 5.37744939e-01 6.11926615e-01 -1.09420407e+00
8.30008626e-01 -6.71873689e-02 -3.87683451e-01 -4.52246845e-01
-6.58939779e-01 -3.54275674e-01 -8.65088046e-01 3.27673078e-01
6.45816684e-01 -1.49276698e+00 -1.46544009e-01 -7.08972514e-02
-9.16546643e-01 -2.28541866e-01 -2.03753665e-01 7.36193597e-01
-1.05389990e-01 2.28596717e-01 -6.78277433e-01 -2.77133584e-01
-3.59744638e-01 -7.72239327e-01 5.78863025e-01 2.31114715e-01
-2.81246811e-01 -1.21690476e+00 4.31138635e-01 1.14929065e-01
3.76923054e-01 6.77944422e-02 1.06775403e+00 -8.62177372e-01
-5.84508479e-01 -1.47330984e-01 -4.38651472e-01 -1.22486591e-01
1.81489751e-01 -1.08078003e-01 -5.18081844e-01 -3.99350286e-01
-1.02277803e+00 -4.48381960e-01 5.34703553e-01 -2.59659678e-01
7.85520300e-02 -7.74714425e-02 -7.17470884e-01 3.85560602e-01
1.96733725e+00 -1.73907474e-01 3.45182419e-01 7.15472519e-01
8.01928997e-01 5.52600741e-01 8.38851511e-01 3.32591198e-02
8.78355861e-01 8.39222491e-01 -4.04594392e-01 -2.11364314e-01
-5.29536784e-01 -5.99513829e-01 2.66058773e-01 1.19514370e+00
-3.23108882e-02 -3.28134656e-01 -1.48424089e+00 1.19957733e+00
-1.73517728e+00 -8.03513169e-01 1.19705223e-01 2.18986511e+00
1.34761858e+00 8.55827630e-02 -2.33735479e-02 -7.41103411e-01
8.63596082e-01 -2.64133483e-01 -4.66904879e-01 -5.86903021e-02
-3.64990383e-01 3.42161953e-01 1.14603341e+00 7.15714633e-01
-1.02936399e+00 1.41994870e+00 6.33928299e+00 6.31035209e-01
-7.35622227e-01 2.89715499e-01 -1.04214467e-01 -2.89364364e-02
-4.58712667e-01 3.59974980e-01 -1.26186597e+00 4.63513553e-01
1.12714541e+00 -6.78080618e-01 2.16917336e-01 7.35166967e-01
-2.52854586e-01 -2.20855977e-02 -1.04411530e+00 5.07416010e-01
-1.37172386e-01 -1.29389405e+00 -1.61546573e-01 -4.11509238e-02
7.90626705e-01 5.85692644e-01 -3.13762635e-01 5.26692688e-01
1.12187076e+00 -7.33357310e-01 5.16778767e-01 4.48152542e-01
1.28668952e+00 -8.56332898e-01 9.15130615e-01 1.53885499e-01
-1.20771718e+00 4.42422301e-01 -4.24035430e-01 4.16859865e-01
4.53846246e-01 6.26799047e-01 -1.02815974e+00 8.70402277e-01
5.37550151e-01 3.95359069e-01 -6.46844983e-01 9.05235410e-01
-3.64776701e-01 5.15709400e-01 -6.31298959e-01 1.34482339e-01
1.98628008e-01 1.83709919e-01 5.16416192e-01 1.59777522e+00
4.57913578e-01 -7.60605782e-02 3.54368597e-01 4.30670291e-01
-7.22898662e-01 4.20706719e-01 -7.64143586e-01 -1.97000355e-01
1.17856669e+00 1.27593875e+00 -4.76358831e-01 -3.52108598e-01
-7.86591649e-01 6.74211860e-01 9.32195008e-01 2.74559826e-01
-6.96884394e-01 -6.86886847e-01 7.10579455e-01 9.96117517e-02
1.23485811e-01 -2.85712123e-01 1.99821070e-01 -1.51824951e+00
2.09687501e-01 -7.83540130e-01 3.88396591e-01 -7.23290741e-01
-1.43692911e+00 6.38541579e-01 8.45831558e-02 -8.31731796e-01
-5.99666178e-01 -3.14633578e-01 -9.60620306e-03 1.22174287e+00
-1.77020919e+00 -1.44359994e+00 1.10408582e-01 2.20439136e-01
7.54042640e-02 -2.03908324e-01 9.43979442e-01 9.61926639e-01
-5.33385515e-01 9.87056375e-01 3.21685493e-01 4.91104454e-01
1.35418248e+00 -1.36515534e+00 8.04897785e-01 1.09396088e+00
3.70429873e-01 1.20910156e+00 6.06052995e-01 -1.06405115e+00
-1.21014535e+00 -1.11970639e+00 1.48357499e+00 -7.89796531e-01
9.42987204e-01 -2.45014295e-01 -8.78519952e-01 1.06920266e+00
2.03672379e-01 1.43609792e-01 7.60670245e-01 7.22223461e-01
-4.81844157e-01 1.78614318e-01 -1.06366336e+00 5.81913352e-01
1.14899087e+00 -6.55085623e-01 -6.02576733e-01 2.07744092e-01
5.86557865e-01 -5.41665554e-01 -1.33170116e+00 3.26664031e-01
4.89880472e-01 -2.05805227e-01 1.06552708e+00 -1.06141555e+00
1.68892443e-01 -6.23773277e-01 -2.83782780e-01 -1.51513493e+00
-2.69006044e-01 -1.59121916e-01 3.52136731e-01 1.86846054e+00
1.04635715e+00 -6.47839904e-01 3.95161092e-01 9.59956944e-01
4.22812365e-02 -2.93469667e-01 -7.78674960e-01 -8.85895908e-01
2.58159548e-01 -9.38767642e-02 6.65883482e-01 1.38577056e+00
1.26217067e-01 5.24247348e-01 -2.23823503e-01 4.00955975e-01
3.66529912e-01 5.74679859e-02 7.58862078e-01 -1.11531532e+00
-9.48039368e-02 1.95396990e-01 -1.57402962e-01 -5.33598602e-01
4.94392633e-01 -1.17326164e+00 -4.97545004e-02 -1.71876180e+00
3.15834373e-01 -1.28519905e+00 -2.57392287e-01 8.12425911e-01
-5.92311203e-01 2.14615628e-01 2.97331631e-01 3.09235096e-01
-6.79576814e-01 6.17766753e-02 3.51138055e-01 2.24872649e-01
-1.02903567e-01 -7.94921577e-01 -7.71621764e-01 6.47886276e-01
5.24261296e-01 -6.45339131e-01 -2.41551727e-01 -8.79009485e-01
6.75960958e-01 -6.47731349e-02 -5.21785170e-02 -8.27341259e-01
3.83974850e-01 -6.43953914e-03 1.97200999e-01 -2.43222922e-01
9.07580331e-02 -6.06214762e-01 3.93514961e-01 -5.81941567e-02
-2.39703804e-01 3.45694929e-01 4.73540843e-01 3.10584277e-01
-3.55782896e-01 -3.81189436e-01 3.80629629e-01 -2.19895229e-01
-1.14253378e+00 6.26752200e-03 1.02437973e-01 7.14179277e-01
8.12189043e-01 2.49871641e-01 -6.79298222e-01 1.37709291e-03
-5.33409953e-01 3.43397707e-01 9.13222373e-01 5.01271367e-01
-2.25412160e-01 -1.48296463e+00 -7.89878011e-01 -3.97206992e-01
3.34192932e-01 -4.25616741e-01 -2.08793268e-01 5.48671365e-01
-6.45883143e-01 7.10667312e-01 -4.43834931e-01 1.76324368e-01
-1.05535448e+00 4.66858268e-01 -1.35387052e-02 -6.66235149e-01
-3.97825122e-01 6.36177003e-01 -3.83802682e-01 -8.36013317e-01
-2.30072677e-01 2.54549563e-01 -5.73605224e-02 1.31428823e-01
1.14715688e-01 3.48440886e-01 1.66869476e-01 -8.54739845e-01
-7.57436097e-01 3.09130222e-01 -2.57365167e-01 -4.35896695e-01
1.24193561e+00 -2.29482070e-01 -5.27697317e-02 2.73774564e-01
8.90212774e-01 7.72951245e-01 -4.68075812e-01 -3.86246145e-01
5.44240355e-01 -2.48419225e-01 -1.18515566e-01 -1.01992047e+00
-7.45150805e-01 3.94969761e-01 2.58366108e-01 -3.73261720e-01
7.49153852e-01 1.17760729e-02 9.62570190e-01 6.17535293e-01
8.10544848e-01 -1.20966601e+00 -7.57118762e-01 5.10325193e-01
2.71508485e-01 -1.27452719e+00 2.50656247e-01 -5.59610128e-01
-5.96002698e-01 5.65874219e-01 7.58865595e-01 2.32692719e-01
5.99135101e-01 4.77707773e-01 2.14009449e-01 -1.38945475e-01
-8.55552793e-01 -5.88637054e-01 1.68974344e-02 5.65374434e-01
8.02928388e-01 2.75174584e-02 -6.29018307e-01 5.03830791e-01
-2.12103784e-01 6.76054955e-02 5.03327906e-01 9.18346524e-01
-1.19271897e-01 -1.42618740e+00 -4.87338528e-02 3.87691438e-01
-8.43237877e-01 -7.07425535e-01 -2.62611479e-01 1.41107440e+00
2.75769591e-01 5.84429264e-01 -7.47988075e-02 4.64129411e-02
4.67916042e-01 2.75521070e-01 4.01896417e-01 -8.05078685e-01
-5.51385343e-01 -4.30890590e-01 9.46884155e-01 -2.96911180e-01
-4.95945811e-01 -7.05527544e-01 -1.40130174e+00 -5.38036168e-01
-5.29697001e-01 3.75036478e-01 6.67943358e-01 6.80387676e-01
6.67985141e-01 1.76083729e-01 -1.14683487e-01 -7.49470741e-02
-1.44962162e-01 -7.78837502e-01 -3.99296969e-01 3.13725799e-01
-2.44004026e-01 -7.74223804e-01 1.12020440e-01 1.11805730e-01] | [9.565641403198242, 9.020813941955566] |
41f3133d-811b-45b5-ad56-50a2226b7fd8 | generativere-incorporating-a-novel-copy | null | null | https://aclanthology.org/2021.findings-emnlp.182 | https://aclanthology.org/2021.findings-emnlp.182.pdf | GenerativeRE: Incorporating a Novel Copy Mechanism and Pretrained Model for Joint Entity and Relation Extraction | Previous neural Seq2Seq models have shown the effectiveness for jointly extracting relation triplets. However, most of these models suffer from incompletion and disorder problems when they extract multi-token entities from input sentences. To tackle these problems, we propose a generative, multi-task learning framework, named GenerativeRE. We firstly propose a special entity labelling method on both input and output sequences. During the training stage, GenerativeRE fine-tunes the pre-trained generative model and learns the special entity labels simultaneously. During the inference stage, we propose a novel copy mechanism equipped with three mask strategies, to generate the most probable tokens by diminishing the scope of the model decoder. Experimental results show that our model achieves 4.6% and 0.9% F1 score improvements over the current state-of-the-art methods in the NYT24 and NYT29 benchmark datasets respectively. | ['Sophia Ananiadou', 'Jiarun Cao'] | null | null | null | null | findings-emnlp-2021-11 | ['joint-entity-and-relation-extraction'] | ['natural-language-processing'] | [ 3.90973955e-01 5.40128887e-01 -1.93792731e-02 -4.15506840e-01
-9.74290669e-01 -3.72229904e-01 3.90128076e-01 -2.17640460e-01
-5.46897650e-01 1.12039411e+00 1.84551850e-01 -1.99287519e-01
1.95440531e-01 -7.67044127e-01 -9.10135806e-01 -4.60304260e-01
2.78204530e-01 7.46634364e-01 1.48856565e-01 -1.13305077e-01
-2.77135540e-02 -2.40056679e-01 -1.24201906e+00 5.28473914e-01
1.30592632e+00 5.69059312e-01 3.84375453e-01 6.22794151e-01
-3.05200815e-01 7.39931405e-01 -8.14920068e-01 -7.47464418e-01
-1.22918323e-01 -5.90476632e-01 -8.12804639e-01 -1.38241932e-01
-3.78475599e-02 -2.52610028e-01 -1.99051753e-01 9.60003138e-01
7.57725835e-01 -1.65100947e-01 5.88290453e-01 -1.02955723e+00
-7.98120677e-01 1.30062664e+00 -5.68133891e-01 1.93244979e-01
2.43342519e-01 -5.20973885e-03 1.15492499e+00 -9.82113957e-01
7.89379895e-01 1.13025355e+00 4.33565229e-01 7.91399062e-01
-9.21202242e-01 -6.82815313e-01 1.53529540e-01 1.72159210e-01
-1.35934556e+00 -5.58770478e-01 4.31985438e-01 8.28512684e-02
1.52085900e+00 1.63620561e-01 2.95060068e-01 1.17195725e+00
1.76953316e-01 1.14698684e+00 7.43251443e-01 -4.39737678e-01
-6.64326847e-02 -1.83752730e-01 -8.33714083e-02 6.82953715e-01
9.76701528e-02 -2.94766843e-01 -5.95586777e-01 9.43061784e-02
4.73690301e-01 -4.16237652e-01 -1.01268195e-01 1.77165687e-01
-1.18412149e+00 4.73072827e-01 4.15427797e-03 3.95999670e-01
-4.83446360e-01 1.11569464e-01 4.54780668e-01 1.22913614e-01
5.91400087e-01 3.08387160e-01 -6.20226443e-01 -3.72995973e-01
-7.97792256e-01 1.21495478e-01 7.69956708e-01 1.43884230e+00
5.59410810e-01 -1.23133332e-01 -8.65688741e-01 9.13402438e-01
2.81523883e-01 2.88024962e-01 4.51537222e-01 -3.88155848e-01
1.09536290e+00 5.51388800e-01 -1.06726341e-01 -4.19955552e-01
-2.54994512e-01 -7.19436467e-01 -8.62518847e-01 -6.25325084e-01
1.39809385e-01 -6.51340008e-01 -1.07358384e+00 1.79820418e+00
3.71894032e-01 4.08982247e-01 4.82876837e-01 4.45033759e-01
9.83600616e-01 7.59063900e-01 2.69306630e-01 -1.70761436e-01
1.36168730e+00 -1.26857722e+00 -9.76559103e-01 -4.53322470e-01
7.95456529e-01 -6.79247558e-01 7.14784920e-01 1.50775775e-01
-1.26606059e+00 -4.87166137e-01 -8.69030833e-01 -2.22005427e-01
-1.87442288e-01 6.19160771e-01 4.33893442e-01 4.43911552e-01
-6.48085713e-01 4.57664281e-01 -9.03258741e-01 6.10417910e-02
4.16487247e-01 3.64766091e-01 -1.21889040e-01 -8.97990391e-02
-1.43318760e+00 8.69901717e-01 8.05692196e-01 4.90372926e-01
-7.91703224e-01 -6.13121808e-01 -9.54373002e-01 4.46850985e-01
6.05436683e-01 -9.29106653e-01 1.37769938e+00 -4.04784679e-01
-1.68852723e+00 5.46016872e-01 -5.01427472e-01 -3.41079533e-01
5.08303702e-01 -5.63066006e-01 -4.04561818e-01 -2.20103383e-01
1.85928062e-01 6.58284903e-01 3.98037374e-01 -7.65830040e-01
-7.29980230e-01 2.54411865e-02 -2.09700808e-01 2.18307480e-01
-8.87947157e-02 1.72970906e-01 -6.92786515e-01 -5.65292597e-01
-2.49542087e-01 -7.43125200e-01 -2.19744936e-01 -9.56727922e-01
-1.09929931e+00 -7.02909291e-01 5.15071988e-01 -6.52639210e-01
1.32304835e+00 -1.90546477e+00 2.62812644e-01 -3.21046263e-02
4.54985052e-02 7.38561213e-01 -3.31823111e-01 5.80054581e-01
1.14668429e-01 2.98324078e-01 -1.96587637e-01 -6.81007326e-01
1.51314139e-01 3.35696757e-01 -1.16540063e-02 -1.66497111e-01
7.48354614e-01 1.20806837e+00 -1.09065413e+00 -6.70146942e-01
-1.90465569e-01 4.21406120e-01 -4.85727668e-01 3.46370220e-01
-4.02411848e-01 4.01998758e-01 -6.59536719e-01 3.94011438e-01
6.53791964e-01 -3.08941931e-01 4.92774934e-01 5.47161996e-02
-2.09952220e-02 8.33929121e-01 -9.50768471e-01 1.94702280e+00
-3.68720800e-01 1.47808656e-01 -4.48291451e-01 -8.66380751e-01
7.99227774e-01 6.04155242e-01 -9.62283909e-02 -6.26129508e-01
2.21746609e-01 3.21393281e-01 1.77740023e-01 -7.92463481e-01
7.22134054e-01 2.69505708e-03 -2.50640631e-01 2.68321484e-01
3.33591729e-01 4.20282960e-01 4.73886579e-01 2.24579394e-01
1.12958527e+00 6.63788378e-01 1.89803824e-01 1.65927634e-01
4.63597119e-01 -2.53838837e-01 8.97905827e-01 5.84480703e-01
3.18061024e-01 6.06551349e-01 7.32108593e-01 6.78956956e-02
-9.20936942e-01 -7.20821440e-01 3.03657442e-01 9.08539176e-01
-1.74850777e-01 -6.41803205e-01 -9.04947877e-01 -9.06026244e-01
-2.58195549e-01 8.97924662e-01 -4.46735740e-01 -2.84796387e-01
-9.46372628e-01 -1.08379257e+00 9.51505363e-01 6.60577953e-01
6.64442778e-01 -1.19629824e+00 -2.78340012e-01 5.65504730e-01
-5.50924718e-01 -1.47553480e+00 -4.89350289e-01 2.65646696e-01
-6.19263649e-01 -6.99634731e-01 -5.94300032e-01 -9.05645251e-01
6.33247495e-01 -3.32008630e-01 1.24661160e+00 -6.05171733e-03
9.62065458e-02 -5.46664596e-01 -5.19880235e-01 -1.84370577e-01
-5.29761314e-01 7.90477276e-01 -4.30667400e-01 -3.07681184e-04
4.48274374e-01 -5.49120486e-01 -1.99598134e-01 -1.04052007e-01
-6.97551608e-01 4.13993448e-01 1.21060586e+00 1.05952299e+00
7.25593269e-01 -1.30971104e-01 1.05688918e+00 -1.49585986e+00
5.91768622e-01 -5.05292237e-01 -3.56080621e-01 6.64781988e-01
-6.37535632e-01 4.02069777e-01 7.82825589e-01 -2.65919775e-01
-1.44303226e+00 3.54445726e-02 -3.39842677e-01 -1.26411259e-01
-1.98012814e-01 6.62949741e-01 -5.20371854e-01 7.26207793e-01
1.32933348e-01 4.46768105e-01 -6.66045904e-01 -5.78387499e-01
3.47752094e-01 7.92677820e-01 5.74603081e-01 -5.93346179e-01
6.04268253e-01 -2.97455162e-01 -2.75473833e-01 -2.81300366e-01
-9.99896348e-01 -2.87646890e-01 -6.20625079e-01 2.30050027e-01
8.94828081e-01 -9.69677985e-01 -5.74062109e-01 6.38645113e-01
-1.48288119e+00 -2.60818899e-01 -4.39372659e-02 2.84159571e-01
-2.93366939e-01 2.71076739e-01 -7.94705510e-01 -6.79740071e-01
-7.40283310e-01 -9.24466550e-01 1.23523307e+00 3.96373928e-01
-6.09954000e-02 -7.78823793e-01 2.45463699e-02 3.73742372e-01
1.56690210e-01 1.05783544e-01 9.48248506e-01 -9.96912777e-01
-8.51497948e-01 7.59715810e-02 -2.08347425e-01 2.31876388e-01
5.10838255e-02 -1.45188466e-01 -9.47445810e-01 2.95741595e-02
-4.71686840e-01 -3.45150322e-01 8.93052697e-01 -1.31900206e-01
1.02293861e+00 -2.93053687e-01 -5.71735620e-01 4.76069272e-01
1.32866681e+00 2.22073600e-01 8.47596407e-01 5.91169968e-02
8.63086045e-01 3.38895053e-01 5.98667324e-01 2.19378233e-01
6.14643335e-01 5.04838824e-01 1.16110004e-01 1.52187124e-01
-2.70759881e-01 -6.03325248e-01 3.32083941e-01 1.12257671e+00
-7.01417681e-03 -7.60571897e-01 -7.59848237e-01 7.77603388e-01
-1.97202027e+00 -8.69340122e-01 -1.63696319e-01 1.67662907e+00
1.32904279e+00 3.53067875e-01 -1.96057498e-01 8.90228227e-02
7.52411246e-01 -5.74027747e-02 -4.64990377e-01 -3.29387337e-01
-1.03667393e-01 4.79437560e-01 3.73400986e-01 2.92293429e-01
-9.19070005e-01 1.36133802e+00 5.74936867e+00 1.09889507e+00
-7.74934590e-01 1.78428113e-01 5.08920312e-01 -2.95138992e-02
-4.55517977e-01 -2.14640833e-02 -1.37354815e+00 5.75344682e-01
1.16796052e+00 -1.51966047e-02 6.14036322e-02 4.24328268e-01
4.32284735e-02 7.91411009e-03 -1.13043153e+00 5.31133115e-01
7.14609697e-02 -1.01744211e+00 -2.31581107e-02 -1.49188906e-01
6.65949464e-01 7.98139274e-02 -2.51091272e-01 7.27922142e-01
3.94845337e-01 -1.01394701e+00 6.41276181e-01 5.16878843e-01
7.84896076e-01 -9.56497431e-01 9.48706448e-01 6.41619205e-01
-1.03423870e+00 3.49601269e-01 -2.32975796e-01 -1.08077116e-02
5.05779386e-01 8.56508076e-01 -1.12799418e+00 1.04946244e+00
3.24691951e-01 7.20534861e-01 -4.58869040e-01 9.76523638e-01
-8.65569890e-01 7.22384393e-01 -2.49434143e-01 -2.39631668e-01
2.07149461e-01 -1.91999488e-02 3.62332433e-01 1.38385117e+00
3.32822442e-01 2.80653145e-02 -6.00501662e-03 9.86947417e-01
-4.63623255e-01 -3.33062299e-02 -2.36844242e-01 -1.25809208e-01
6.58660829e-01 1.32143998e+00 -5.03645599e-01 -5.44131339e-01
-3.22027504e-01 1.17219758e+00 7.21019208e-01 2.90313810e-01
-9.43892658e-01 -8.93332899e-01 1.89084485e-01 -6.18297994e-01
7.44368374e-01 4.08837348e-02 -1.82933643e-01 -1.26775181e+00
3.06992799e-01 -7.45809913e-01 1.30508259e-01 -6.03832722e-01
-8.97403657e-01 7.80687213e-01 -9.77833197e-02 -7.40902603e-01
-6.25448704e-01 -1.51153222e-01 -5.35291433e-01 9.91355419e-01
-1.80688047e+00 -1.29561400e+00 1.41344607e-01 7.60014579e-02
5.35712361e-01 6.58211559e-02 7.99889088e-01 5.56319475e-01
-9.81695235e-01 9.64195371e-01 3.24093625e-02 4.31144327e-01
5.37533581e-01 -1.40479434e+00 9.41220582e-01 1.03182554e+00
7.96478540e-02 5.16452730e-01 4.64042664e-01 -7.71557510e-01
-1.15413249e+00 -1.32623327e+00 1.72510171e+00 -2.01967165e-01
2.78927773e-01 -5.57422042e-01 -8.52938354e-01 9.40554082e-01
4.96840537e-01 -2.58221090e-01 6.43647194e-01 2.65279710e-01
-1.40141323e-01 1.38444379e-01 -9.35641229e-01 4.69967067e-01
1.31605995e+00 -2.74457902e-01 -6.36634946e-01 2.01894835e-01
8.93514991e-01 -8.32390547e-01 -9.94757891e-01 3.78915638e-01
3.93574536e-01 -5.55225074e-01 6.58415973e-01 -5.62878489e-01
5.11453032e-01 -1.74365789e-01 1.67631269e-01 -1.48777521e+00
-2.61676490e-01 -7.61359215e-01 -3.64799798e-01 1.79641759e+00
9.99099791e-01 -5.00546098e-01 6.57997787e-01 1.94993734e-01
-5.44096708e-01 -9.81448531e-01 -9.02960181e-01 -7.70809770e-01
-6.58550188e-02 -1.45870045e-01 7.70755172e-01 5.33527732e-01
-6.74020424e-02 9.54383194e-01 -4.79647845e-01 8.62572119e-02
3.96648824e-01 1.10845171e-01 4.83013898e-01 -8.82385075e-01
-4.40513492e-01 -7.25677386e-02 1.87144190e-01 -1.28885674e+00
3.00592512e-01 -9.15385664e-01 3.05685937e-01 -1.69025004e+00
3.10885608e-01 -4.79890972e-01 -1.46956861e-01 8.55624497e-01
-7.35406518e-01 -1.33535480e-02 2.71718856e-03 -7.30359405e-02
-8.12576532e-01 7.50575423e-01 1.42629206e+00 7.24494979e-02
-8.30669999e-02 3.26486379e-02 -7.53421545e-01 1.77870601e-01
8.35668385e-01 -7.08278537e-01 -4.20934677e-01 -9.62044477e-01
4.28578138e-01 5.35688661e-02 -1.08822482e-02 -7.01680541e-01
2.47754782e-01 8.39410797e-02 2.31951132e-01 -8.59308362e-01
1.98945686e-01 -3.17636043e-01 4.79298495e-02 2.21273661e-01
-4.77986097e-01 5.37074320e-02 1.35246783e-01 4.59229290e-01
-2.15077534e-01 -3.45278978e-01 3.35153937e-01 -6.40403703e-02
-3.66725504e-01 1.96085453e-01 -2.75176048e-01 4.35637325e-01
7.85283506e-01 2.33490497e-01 -4.85617727e-01 1.01028211e-01
-5.00177205e-01 5.91645479e-01 -1.86633557e-01 4.31250006e-01
5.22246182e-01 -1.18414164e+00 -9.90760505e-01 1.83086589e-01
-4.32394780e-02 5.04147172e-01 1.78377688e-01 7.90216029e-01
-9.87645760e-02 6.53471172e-01 1.23175852e-01 -1.96077362e-01
-1.15184903e+00 3.60924929e-01 2.29378834e-01 -7.60976195e-01
-4.35712844e-01 1.03562880e+00 -1.65573761e-01 -4.09564912e-01
1.11941546e-02 -6.66853547e-01 -2.65354931e-01 -1.08466335e-01
2.78920054e-01 1.88548252e-01 2.29776144e-01 -4.49367076e-01
-2.61954874e-01 1.78298295e-01 -4.86659586e-01 5.91084510e-02
1.31956971e+00 -5.99507131e-02 -1.15307309e-02 2.15223819e-01
1.00889647e+00 -2.88117537e-03 -1.03115189e+00 -3.10488254e-01
2.56245136e-01 1.00882441e-01 -2.70305872e-01 -1.11015666e+00
-9.75476980e-01 7.77250469e-01 -8.59182477e-02 1.64471306e-02
1.08269978e+00 6.64465576e-02 1.34571505e+00 3.41984212e-01
1.09673560e-01 -9.93600667e-01 -3.36600780e-01 8.06893528e-01
2.56057858e-01 -9.02890146e-01 -3.86013091e-01 -8.19305241e-01
-5.32558680e-01 6.67149067e-01 8.69635701e-01 3.35026123e-02
1.68576524e-01 3.24133426e-01 -2.16106266e-01 -5.23357838e-03
-1.20722055e+00 -2.66560972e-01 1.70698076e-01 3.65891099e-01
6.30574703e-01 -6.68852963e-03 -5.41148663e-01 8.85010600e-01
-2.36812577e-01 1.53056666e-01 3.07120413e-01 7.89399385e-01
-3.10573846e-01 -1.61159122e+00 2.40310967e-01 3.03204507e-01
-8.16799223e-01 -4.97349232e-01 -2.37888873e-01 5.56136668e-01
3.41819495e-01 9.30746913e-01 -8.11565891e-02 -3.54199767e-01
4.79450285e-01 3.65789354e-01 5.09655237e-01 -8.91170681e-01
-7.76175201e-01 9.88369286e-02 5.86936116e-01 -1.18982010e-01
-4.37141627e-01 -7.36067295e-01 -1.34792030e+00 1.89630821e-01
-6.08666420e-01 4.25462335e-01 3.27168137e-01 1.20748234e+00
7.39455402e-01 1.04868531e+00 3.99283856e-01 -3.20458531e-01
-5.58962584e-01 -1.53493285e+00 -1.34302765e-01 6.36057854e-02
4.80031073e-02 -3.58648449e-01 1.45211831e-01 -3.73252593e-02] | [9.439807891845703, 8.811860084533691] |
f804afc7-ae90-4199-b048-4935bb08fb45 | disruption-precursor-onset-time-study-based | 2303.14965 | null | https://arxiv.org/abs/2303.14965v1 | https://arxiv.org/pdf/2303.14965v1.pdf | Disruption Precursor Onset Time Study Based on Semi-supervised Anomaly Detection | The full understanding of plasma disruption in tokamaks is currently lacking, and data-driven methods are extensively used for disruption prediction. However, most existing data-driven disruption predictors employ supervised learning techniques, which require labeled training data. The manual labeling of disruption precursors is a tedious and challenging task, as some precursors are difficult to accurately identify, limiting the potential of machine learning models. To address this issue, commonly used labeling methods assume that the precursor onset occurs at a fixed time before the disruption, which may not be consistent for different types of disruptions or even the same type of disruption, due to the different speeds at which plasma instabilities escalate. This leads to mislabeled samples and suboptimal performance of the supervised learning predictor. In this paper, we present a disruption prediction method based on anomaly detection that overcomes the drawbacks of unbalanced positive and negative data samples and inaccurately labeled disruption precursor samples. We demonstrate the effectiveness and reliability of anomaly detection predictors based on different algorithms on J-TEXT and EAST to evaluate the reliability of the precursor onset time inferred by the anomaly detection predictor. The precursor onset times inferred by these predictors reveal that the labeling methods have room for improvement as the onset times of different shots are not necessarily the same. Finally, we optimize precursor labeling using the onset times inferred by the anomaly detection predictor and test the optimized labels on supervised learning disruption predictors. The results on J-TEXT and EAST show that the models trained on the optimized labels outperform those trained on fixed onset time labels. | ['J-TEXT team', 'Yuan Pan', 'Yonghua Ding', 'Zhongyong Chen', 'Zhipeng Chen', 'Zhoujun Yang', 'Nengchao Wang', 'Yu Zhong', 'Bingjia Xiao', 'Bihao Guo', 'Chengshuo Shen', 'Dalong Chen', 'Ming Zhang', 'Wei Zheng', 'Xinkun Ai'] | 2023-03-27 | null | null | null | null | ['supervised-anomaly-detection', 'semi-supervised-anomaly-detection'] | ['computer-vision', 'computer-vision'] | [-3.26606423e-01 -2.32093528e-01 -9.42775160e-02 -4.29533392e-01
-6.81655169e-01 -5.07264197e-01 5.15939236e-01 7.10952461e-01
1.37433559e-01 5.11378586e-01 -1.06768692e-02 -3.18078846e-01
-2.34389350e-01 -4.98483360e-01 -2.01952875e-01 -9.02570486e-01
-3.00135940e-01 8.22299898e-01 3.75140399e-01 -3.20094496e-01
4.63151127e-01 9.03155625e-01 -1.30863094e+00 4.31529164e-01
6.58971429e-01 1.24576771e+00 -3.01215470e-01 4.95219857e-01
-3.24270129e-01 7.93874979e-01 -8.54628384e-01 1.96760476e-01
2.53567249e-01 -4.31485623e-01 -7.81195760e-01 -2.01924875e-01
-3.19090337e-01 -4.39630263e-02 -4.72087741e-01 5.63883305e-01
1.50223166e-01 1.93513751e-01 9.71577942e-01 -1.62191391e+00
3.08549162e-02 1.59109697e-01 -5.05088627e-01 7.68352568e-01
2.26323113e-01 4.01230790e-02 7.60110080e-01 -5.18151760e-01
3.92695695e-01 6.78957701e-01 6.42019272e-01 2.30709583e-01
-1.14632583e+00 -3.77748936e-01 7.61653036e-02 6.02144063e-01
-1.33464372e+00 -4.10868764e-01 9.62331831e-01 -8.23280394e-01
1.32790565e+00 2.37510636e-01 3.18785071e-01 7.62119293e-01
5.21771729e-01 3.93639922e-01 5.76813579e-01 -3.85498196e-01
4.82869685e-01 4.48756199e-03 2.66377270e-01 3.94960195e-01
-1.80504397e-02 5.11646926e-01 -6.62205577e-01 -5.10974765e-01
3.12279388e-02 -9.17458981e-02 -3.32800537e-01 -2.15235963e-01
-7.11967885e-01 8.53869915e-01 -1.39849663e-01 4.65063900e-01
-3.22583109e-01 -6.95125341e-01 9.72567916e-01 2.77981102e-01
6.75738096e-01 9.71639574e-01 -6.94484591e-01 -4.98106092e-01
-1.13497245e+00 2.18775913e-01 1.03739476e+00 7.58130610e-01
7.30418146e-01 2.87404060e-01 -2.66030908e-01 6.94677413e-01
-4.22964208e-02 9.67295021e-02 4.65102732e-01 -4.77354378e-01
4.93903756e-02 7.20247686e-01 2.28633553e-01 -9.35556769e-01
-9.13344204e-01 -8.99145752e-02 -5.92472970e-01 1.24736384e-01
5.33015430e-01 -2.36387193e-01 -9.33296442e-01 1.26726365e+00
1.77894384e-01 2.05267578e-01 2.01337501e-01 7.93974459e-01
3.86667639e-01 6.75919414e-01 -6.71067908e-02 -4.65216130e-01
7.16710150e-01 -8.47475052e-01 -8.83782566e-01 -1.35836422e-01
1.04546106e+00 -6.70324445e-01 7.52071559e-01 6.98565841e-01
-5.88252902e-01 -2.23935366e-01 -1.14647543e+00 5.15570462e-01
-3.93645793e-01 -2.21056506e-01 6.06473029e-01 4.02422875e-01
-4.39695030e-01 7.60967731e-01 -1.14083338e+00 -1.50267348e-01
-7.75825828e-02 9.19728279e-02 -2.22765028e-01 3.71176928e-01
-9.87892628e-01 1.01429451e+00 4.51637506e-01 -9.43888500e-02
-1.03580511e+00 -7.66472936e-01 -7.98033774e-01 -3.29507180e-02
3.07737350e-01 3.26796025e-01 1.48848367e+00 -4.09167767e-01
-1.19259262e+00 5.63951552e-01 -3.88229221e-01 -6.49770498e-01
2.54771769e-01 -8.51107463e-02 -1.00383425e+00 2.33119383e-01
9.33827311e-02 -2.21888036e-01 6.88091695e-01 -1.22276723e+00
-6.37346387e-01 3.36991996e-02 -4.63619620e-01 -2.69252867e-01
3.17571796e-02 -2.49694325e-02 -6.65471926e-02 -4.14957047e-01
4.24789011e-01 -6.43431723e-01 2.95367390e-02 -7.31213212e-01
-5.82570612e-01 -3.71327370e-01 1.22077286e+00 -6.38310611e-01
1.29047191e+00 -2.40581059e+00 -1.63328305e-01 3.82232308e-01
-8.28600749e-02 8.75623226e-02 3.24492335e-01 6.88220978e-01
-3.45542312e-01 4.18687910e-02 -3.58644754e-01 -2.48448655e-01
-1.75182268e-01 3.65264535e-01 -7.46150970e-01 4.37139750e-01
1.04695052e-01 2.84584403e-01 -8.95244062e-01 3.79005522e-02
3.70871484e-01 -2.68179476e-01 -2.54131585e-01 6.65014505e-01
-2.36793920e-01 7.51355290e-01 -5.92884794e-02 9.06062841e-01
4.01755333e-01 2.70262271e-01 -1.46556273e-01 -2.37206116e-01
-2.17389241e-01 3.26203465e-01 -7.57023275e-01 1.02998340e+00
-2.47541256e-02 5.35493135e-01 -1.82867587e-01 -1.47364378e+00
1.32711589e+00 4.53794569e-01 8.22321653e-01 -6.65905714e-01
-8.75246823e-02 1.93689808e-01 1.69010773e-01 -9.16068971e-01
5.30142248e-01 -6.42974198e-01 -1.83972210e-01 5.32681346e-01
-4.52247821e-02 -3.86763990e-01 2.01129392e-01 2.11821228e-01
1.42494166e+00 -2.99534202e-01 2.53015533e-02 -1.83003277e-01
3.43607366e-01 3.10799390e-01 1.15289342e+00 5.37149072e-01
-3.02859515e-01 6.91357374e-01 1.00900424e+00 -5.95599830e-01
-9.39121664e-01 -8.75805497e-01 -3.72984171e-01 5.96653342e-01
2.80519605e-01 -7.43064821e-01 -4.84800428e-01 -8.34728599e-01
6.84859827e-02 1.17985761e+00 -3.58187377e-01 -6.80619895e-01
-3.61551613e-01 -1.03833556e+00 7.54277885e-01 5.11430442e-01
3.42241302e-02 -1.04685056e+00 -2.52036601e-01 3.06468487e-01
-2.03746542e-01 -1.01544678e+00 -4.72914614e-02 5.75584590e-01
-6.18169546e-01 -1.37024093e+00 4.91165817e-01 -2.89571583e-01
5.34070253e-01 -5.74311391e-02 8.55886519e-01 2.96129733e-01
-2.71672249e-01 1.51587710e-01 -6.20130956e-01 -5.92243731e-01
-6.07996047e-01 -4.61916216e-02 7.03883708e-01 -6.52627274e-02
6.39167190e-01 -5.05873859e-01 -1.42721996e-01 5.92715740e-01
-7.27174699e-01 -3.77584070e-01 7.65010342e-02 8.75984669e-01
6.84309721e-01 5.53816259e-01 9.63431835e-01 -6.63183033e-01
3.42809707e-01 -9.07516003e-01 -4.34480995e-01 9.48021635e-02
-8.48952532e-01 -5.68187162e-02 9.63885367e-01 -3.16754460e-01
-1.18084276e+00 -3.24117750e-01 -1.49222121e-01 -7.00803518e-01
-4.80291188e-01 5.30212045e-01 -5.89827672e-02 1.88625798e-01
7.27998555e-01 -1.06278427e-01 -1.01748444e-01 -6.49448335e-01
-2.91836351e-01 7.35317528e-01 7.06068695e-01 -6.66006029e-01
4.84007180e-01 2.98427224e-01 -2.39896655e-01 -8.43055069e-01
-8.20893824e-01 -7.29250252e-01 -5.04524946e-01 -1.82587042e-01
4.20901984e-01 -5.52312672e-01 -3.47462773e-01 6.38615489e-01
-9.11481261e-01 -4.52906638e-01 -3.46614093e-01 4.56682116e-01
-4.05241847e-01 4.84067708e-01 -7.20300257e-01 -7.48716950e-01
-1.84897646e-01 -1.06974220e+00 7.27754653e-01 1.45427585e-01
-5.27995706e-01 -9.83929276e-01 5.38859703e-02 1.28871715e-02
2.27486491e-01 3.83560419e-01 1.47496188e+00 -1.22192276e+00
-1.67680785e-01 -4.76199567e-01 2.47174755e-01 2.22154379e-01
5.75977027e-01 2.76939124e-01 -9.06547606e-01 -3.11702102e-01
5.01386523e-01 -3.54814790e-02 5.07063985e-01 2.72353113e-01
1.31961668e+00 4.61065136e-02 -6.14622474e-01 5.65509498e-01
1.02437770e+00 7.59444535e-01 5.15892208e-01 4.49048370e-01
4.94824946e-01 6.70151651e-01 9.44533408e-01 6.07701361e-01
-2.93203760e-02 5.84348381e-01 2.54568517e-01 -5.08400761e-02
4.26195353e-01 -3.96529995e-02 2.82722235e-01 7.12638021e-01
3.57144326e-01 -1.94156528e-01 -1.28649020e+00 6.47881508e-01
-1.75818717e+00 -9.39574480e-01 -4.06066597e-01 2.24311495e+00
6.82555020e-01 4.05938566e-01 -2.99419817e-02 5.02675235e-01
4.69835103e-01 -1.36517555e-01 -6.97572410e-01 -9.09406066e-01
1.45159177e-02 -1.13002285e-01 1.42156318e-01 4.58080471e-01
-9.34378088e-01 6.84526682e-01 6.98327732e+00 6.22145414e-01
-1.29314029e+00 -2.49782633e-02 6.28883302e-01 -3.09247136e-01
7.84141868e-02 1.85614929e-01 -7.28738368e-01 7.10882127e-01
1.11202371e+00 -2.94363022e-01 2.07310244e-01 8.80906880e-01
5.36183059e-01 -4.54644471e-01 -1.51359999e+00 8.56167436e-01
2.53630131e-02 -1.24751389e+00 -4.35118973e-01 -1.05959065e-01
5.36395013e-01 1.39925936e-02 -6.83465660e-01 4.91575330e-01
-3.49200428e-01 -7.37634897e-01 5.66151500e-01 8.60076487e-01
3.56606960e-01 -9.09635603e-01 1.00624943e+00 4.45702851e-01
-7.83035100e-01 -4.66894582e-02 -1.72613353e-01 -2.04337642e-01
5.04150271e-01 1.12009168e+00 -1.18439567e+00 6.81904852e-01
7.75872290e-01 5.40856004e-01 -3.71780932e-01 1.05044413e+00
-1.35750577e-01 1.00916743e+00 -4.40047204e-01 5.51767230e-01
-1.43308952e-01 -2.30314419e-01 7.67417550e-01 1.06930244e+00
3.59631002e-01 5.64841218e-02 3.70394558e-01 6.25291944e-01
3.64478648e-01 -1.56967416e-01 -7.55594730e-01 -4.48093675e-02
6.01062119e-01 1.36751223e+00 -8.84288430e-01 -1.47653520e-01
-4.19741005e-01 5.29590905e-01 2.52863318e-01 2.94958264e-01
-8.16922069e-01 -2.60735154e-01 9.59600866e-01 2.41759747e-01
1.62901506e-02 -2.71366596e-01 -5.81169605e-01 -7.82552540e-01
-9.46320891e-02 -5.31876326e-01 5.79473257e-01 -4.44409609e-01
-1.65069079e+00 6.51077688e-01 2.42334127e-01 -1.35125124e+00
-5.33180833e-01 -4.66820002e-01 -1.21552730e+00 6.27184749e-01
-1.03937078e+00 -2.61442482e-01 -1.35116607e-01 8.46527740e-02
3.39516252e-01 -4.89715517e-01 8.68460953e-01 7.66739398e-02
-9.15348053e-01 4.61677760e-01 2.37382695e-01 -8.47220346e-02
8.47287416e-01 -1.39942336e+00 1.78589657e-01 8.48072410e-01
-6.48654625e-02 2.52655745e-01 9.11777735e-01 -9.71091449e-01
-1.00780892e+00 -9.25045013e-01 5.62016487e-01 -4.07690376e-01
6.14211202e-01 -1.91291466e-01 -1.68090785e+00 6.31605268e-01
-2.27465946e-02 3.68157625e-02 9.52089012e-01 3.24984461e-01
1.04887828e-01 -4.33840528e-02 -1.33694375e+00 1.64288521e-01
5.39144278e-01 -4.90836859e-01 -7.40309656e-01 4.63072985e-01
3.22789013e-01 -5.34201682e-01 -8.64395976e-01 6.66212678e-01
-2.64275610e-01 -1.05428016e+00 5.20910382e-01 -5.22579551e-01
8.19627866e-02 -4.82317388e-01 1.13599107e-01 -1.51513851e+00
-1.59358978e-01 -4.02072638e-01 -5.26751637e-01 1.45043540e+00
3.86485636e-01 -6.24072850e-01 6.64089084e-01 8.99865866e-01
-5.88883460e-01 -1.03903806e+00 -1.17315555e+00 -9.60925758e-01
-6.63430765e-02 -6.82533443e-01 7.43990242e-01 1.03781557e+00
6.81965888e-01 1.13574982e-01 -1.37109339e-01 4.61407989e-01
1.07746162e-01 2.23332942e-01 4.02853876e-01 -1.16650152e+00
-1.15377732e-01 -2.49410257e-01 -3.56305778e-01 -2.91340709e-01
1.68247774e-01 -8.66836846e-01 3.48153412e-01 -1.06269622e+00
-3.60161632e-01 -9.47116673e-01 -2.69713730e-01 7.41150439e-01
-5.50116226e-02 -1.33780390e-01 -4.45810437e-01 3.12617928e-01
-1.99971691e-01 6.11583471e-01 3.31923217e-01 -6.01257570e-02
-4.19660240e-01 -2.25973397e-01 -1.97883353e-01 8.27043176e-01
1.09529388e+00 -5.45261264e-01 -3.86908501e-01 6.34517446e-02
-8.01126882e-02 1.43196821e-01 6.71074465e-02 -1.11862147e+00
2.06203029e-01 -2.11023450e-01 3.73668790e-01 -6.83090091e-01
1.88949794e-01 -3.98825586e-01 4.28547598e-02 6.42020926e-02
2.19240651e-01 1.99243218e-01 6.28026307e-01 4.70797658e-01
-2.94600815e-01 -5.77253520e-01 5.23384154e-01 2.09671766e-01
-6.85664654e-01 2.54121631e-01 -8.54026139e-01 1.68459579e-01
1.30992067e+00 -6.36498407e-02 -3.62289473e-02 -1.96605980e-01
-6.61855757e-01 3.38618994e-01 5.06690919e-01 5.51181197e-01
4.66582030e-01 -1.00155187e+00 -3.02089810e-01 4.38735366e-01
2.41492167e-01 8.07893127e-02 2.72719681e-01 1.04257989e+00
-2.99279213e-01 1.51922792e-01 8.11378956e-02 -7.54842639e-01
-9.63271201e-01 7.31118739e-01 6.37903869e-01 -8.72799605e-02
-8.98545206e-01 5.26514649e-01 -1.01659685e-01 -2.67975658e-01
1.84181158e-03 -2.01921925e-01 -1.96432248e-01 1.29865319e-01
5.34049571e-01 4.86454546e-01 7.44316876e-01 -5.61054289e-01
-3.35212827e-01 8.01618621e-02 -4.47613686e-01 2.74387836e-01
1.05372190e+00 1.67011470e-01 -2.27411479e-01 6.42691791e-01
8.50017190e-01 -2.85433307e-02 -1.11044550e+00 8.04945827e-03
5.01312852e-01 -3.28935087e-01 4.79062647e-02 -8.74683082e-01
-9.85856831e-01 5.56846797e-01 3.91021073e-01 5.36363900e-01
1.04680061e+00 8.80313814e-02 8.72473300e-01 2.95383722e-01
4.24880385e-01 -1.21388412e+00 -1.88339889e-01 6.69648826e-01
7.64896691e-01 -1.09493589e+00 -3.26988131e-01 -3.46886218e-01
-6.67768717e-01 1.07635963e+00 9.89714265e-01 2.69939601e-01
5.53823471e-01 5.75970113e-01 2.42466852e-01 -4.55736428e-01
-7.83172607e-01 2.73119420e-01 -3.00681014e-02 3.71155888e-01
1.54194161e-01 1.73339218e-01 -3.27630967e-01 8.83288860e-01
-2.44315341e-01 -4.79896992e-01 6.20774031e-01 9.77007747e-01
-1.56363860e-01 -1.25764763e+00 -5.13787687e-01 1.03377390e+00
-3.10200632e-01 3.44459742e-01 -2.89140284e-01 4.29094672e-01
2.55145818e-01 1.36877096e+00 2.67530739e-01 -6.14967406e-01
4.64856893e-01 5.95377862e-01 -2.77096825e-03 -6.58158779e-01
-2.69693375e-01 -2.52721995e-01 2.06670195e-01 -4.45028394e-01
1.48586228e-01 -9.35022712e-01 -1.92039847e+00 -2.35908762e-01
-3.84611666e-01 8.39940548e-01 2.23586529e-01 1.42257023e+00
3.24574858e-01 6.59573734e-01 8.36409926e-01 -9.38623667e-01
-2.72158027e-01 -1.09831512e+00 -8.33815813e-01 6.40589654e-01
4.36692297e-01 -1.27885008e+00 -8.16784680e-01 -1.08829193e-01] | [7.397780418395996, 2.5938074588775635] |
e9b721b1-52da-4290-8108-37cfacbcbe9d | self-supervised-contrastive-learning-for-6 | null | null | https://ieeexplore.ieee.org/document/9763018 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9763018 | Self-Supervised Contrastive Learning for Singing Voices | This study introduces self-supervised contrastive learning to acquire feature representations of singing voices. To acquire robust representations in an unsupervised manner, regular self-supervised contrastive learning trains neural networks to make the feature representation of a sample close to those of its computationally transformed versions. Similarly, we employ two transformations—pitch shifting and time stretching—considering the nature of singing voices. Nevertheless, we use them reversely: we train networks to push away representations of the transformed versions. The networks then attempt to discriminate changes in vocal timbres introduced by pitch shifting without time stretching and those in singing expressions introduced by time stretching without pitch shifting. Consequently, the acquired representations become attentive to vocal timbre and singing expression. This was confirmed through a singer identification task, where we trained a classifier to learn the relationship between the feature representations to the corresponding singer labels of 500 singers. As a result, the employed transformations helped the classifier improve the classification accuracy by 9.12% (top-1 accuracy: 63.08%) compared with the case where the feature representations fed to the classifier were acquired without the transformations (top-1 accuracy: 53.96%). Furthermore, the proposed approach can be extended to acquire feature representations attentive to either vocal timbre or singing expression but not to the other by changing how the transformations are incorporated. We particularly explored the characteristics of such vocal timbre- or singing expression-oriented feature representations against song genre, singer gender, and vocal technique, and confirmed that they successfully capture different aspects of singing voices. | ['Masataka Goto', 'Kento Watanabe', 'Hiromu Yakura'] | 2022-04-26 | null | null | null | ieee-acm-transactions-on-audio-speech-and-7 | ['singer-identification', 'vocal-technique-classification'] | ['music', 'music'] | [ 4.22270656e-01 7.42234811e-02 1.09352954e-01 -2.79461652e-01
-3.81353915e-01 -9.54486847e-01 6.23894572e-01 -1.06497228e-01
-5.19142389e-01 5.70291877e-01 -5.60334735e-02 1.58744350e-01
-2.80541122e-01 -6.09081388e-01 -5.61336935e-01 -8.88156354e-01
-2.22099081e-01 1.34197339e-01 -1.00519516e-01 -4.23155665e-01
7.66337477e-03 7.53166080e-01 -2.37556624e+00 1.31268203e-01
6.71875775e-01 9.61458325e-01 -8.22833404e-02 7.99160719e-01
5.80993555e-02 2.96843469e-01 -9.99104381e-01 1.36138856e-01
2.10068926e-01 -7.12257385e-01 -5.70876181e-01 2.77727515e-01
4.75554168e-01 2.44794473e-01 1.44475684e-01 1.01246119e+00
3.63221705e-01 4.03643548e-01 8.42343807e-01 -8.37861478e-01
-3.86349022e-01 6.34296715e-01 -1.66499883e-01 1.43173963e-01
4.05567914e-01 7.48140141e-02 9.83582318e-01 -8.46946537e-01
4.09477681e-01 1.00401676e+00 6.37765884e-01 7.54444599e-01
-1.20665216e+00 -8.29949975e-01 -1.81893423e-01 1.28012046e-01
-1.07512534e+00 -5.93774796e-01 1.18367052e+00 -5.01942754e-01
4.81999159e-01 7.80806422e-01 8.21267128e-01 9.41543221e-01
-2.70671904e-01 2.37751290e-01 1.26080179e+00 -6.16746962e-01
-1.43667245e-02 3.71580869e-01 4.79425900e-02 2.58509547e-01
-3.85062814e-01 3.72871965e-01 -4.44862783e-01 7.32197240e-02
4.47204590e-01 -3.61072123e-01 -3.99174720e-01 3.67844589e-02
-1.01951146e+00 5.64374685e-01 4.36549038e-01 1.10176659e+00
-3.32326114e-01 -1.82644859e-01 4.10001457e-01 5.37449181e-01
2.93451339e-01 1.05628800e+00 -3.03283691e-01 -1.09161183e-01
-1.10661876e+00 8.70604143e-02 7.53450334e-01 4.51166153e-01
7.58934557e-01 6.68146491e-01 -1.90894157e-01 9.50215340e-01
-1.86356470e-01 5.15474856e-01 8.59602690e-01 -7.19936430e-01
-1.29156798e-01 2.24628031e-01 -1.99793547e-01 -1.04499626e+00
-5.09459674e-01 -8.34512889e-01 -7.07507908e-01 9.76879969e-02
5.37928641e-01 -5.28592356e-02 -6.83675289e-01 1.90293717e+00
1.54787257e-01 3.47726010e-02 2.13411689e-01 8.99207473e-01
8.50459158e-01 6.58715665e-01 -2.47546658e-01 -7.72631228e-01
1.17908323e+00 -7.82009959e-01 -1.08547485e+00 1.60062596e-01
1.54412881e-01 -1.00909519e+00 1.39706159e+00 4.10496026e-01
-1.11392748e+00 -1.20955801e+00 -1.17591751e+00 4.85692173e-01
-4.09289479e-01 1.86337203e-01 1.77617092e-02 7.48724818e-01
-8.03216994e-01 1.07828832e+00 -2.50353456e-01 -1.38749063e-01
-3.32560539e-02 4.11699742e-01 -3.55818778e-01 7.61536539e-01
-1.36319423e+00 6.27632499e-01 3.85521591e-01 3.03964704e-01
-8.19389105e-01 -5.77106714e-01 -5.29041588e-01 9.81112123e-02
9.72778574e-02 -2.25272968e-01 1.23931313e+00 -1.51936495e+00
-1.94606960e+00 8.14132571e-01 2.23013937e-01 -1.52560309e-01
4.03912842e-01 1.31379589e-01 -9.14395392e-01 1.51301876e-01
-2.24775001e-01 1.91442609e-01 1.52080023e+00 -1.23494744e+00
-4.14487630e-01 -1.14652254e-01 -2.96608031e-01 6.36539608e-02
-7.93026686e-01 2.09728777e-02 2.78972685e-01 -6.51364088e-01
1.05272815e-01 -8.39806020e-01 3.54651451e-01 -3.80508572e-01
-2.99260199e-01 -7.50229806e-02 8.51134658e-01 -4.48644310e-01
1.06695092e+00 -2.44421268e+00 1.95735186e-01 3.56417775e-01
3.15978676e-02 5.50453663e-01 -2.27402791e-01 3.48452598e-01
-4.98121470e-01 -5.47341555e-02 -2.78490037e-01 -8.32965523e-02
-9.79277641e-02 2.05293164e-01 -3.77464056e-01 3.37134898e-01
2.50086099e-01 4.58667517e-01 -9.39501643e-01 -2.04455614e-01
1.09966137e-01 5.35138607e-01 -4.87502903e-01 5.84119081e-01
1.11029208e-01 6.44130647e-01 2.21768711e-02 2.05644414e-01
4.24706429e-01 7.11794496e-01 1.96243331e-01 -4.61042821e-01
-3.15596968e-01 1.96172521e-01 -9.44354415e-01 1.07443929e+00
-7.40938783e-01 6.19525671e-01 1.84430704e-01 -1.01170695e+00
1.54052114e+00 6.01191461e-01 3.29332560e-01 -4.00622964e-01
2.66780108e-01 3.71282727e-01 4.81671631e-01 -4.70106870e-01
3.43008876e-01 -7.56847799e-01 3.41283120e-02 1.98284805e-01
3.56023759e-01 -6.00429118e-01 -4.54564095e-02 -7.36410320e-01
5.02995670e-01 3.27819446e-03 2.39218011e-01 -2.05878004e-01
8.60065639e-01 -4.77659881e-01 2.65569836e-01 5.42443931e-01
2.26854701e-02 4.17467415e-01 2.07137585e-01 -4.04445618e-01
-6.21505916e-01 -1.06352437e+00 -2.75981694e-01 1.34992683e+00
-2.88997591e-01 -2.10139573e-01 -8.51021469e-01 -5.41564822e-01
-2.12260917e-01 7.79509902e-01 -6.12275124e-01 -5.33025622e-01
-6.81151807e-01 -3.47109139e-01 8.18391323e-01 2.27974832e-01
2.62107819e-01 -1.48580742e+00 -4.89932865e-01 1.81921825e-01
1.85198933e-01 -7.10769951e-01 -4.24390823e-01 5.24139524e-01
-7.76259780e-01 -9.26641524e-01 -6.00167572e-01 -8.14156055e-01
3.66960734e-01 -2.11054668e-01 7.18621969e-01 -7.27740526e-02
-1.38731331e-01 3.21378499e-01 -3.21716487e-01 -3.98197532e-01
-8.92732561e-01 1.59018934e-02 4.39792395e-01 4.01895165e-01
-3.03127050e-01 -1.01223946e+00 -1.93650961e-01 4.63709652e-01
-9.88422871e-01 -4.51822668e-01 1.69705048e-01 9.20215786e-01
3.76720071e-01 1.59578681e-01 8.62176061e-01 -7.24456787e-01
9.07336473e-01 -3.63069624e-02 -2.06035629e-01 -1.29064038e-01
-3.15466940e-01 -1.95350647e-01 1.15448034e+00 -9.36422944e-01
-8.59319568e-01 2.17856020e-02 -4.05968308e-01 -5.09232700e-01
-5.11874497e-01 2.65382320e-01 -1.19878486e-01 5.46855014e-03
9.59110856e-01 3.01730514e-01 1.35893285e-01 -5.50156057e-01
3.80093366e-01 8.56273174e-01 8.11946154e-01 -4.82873440e-01
1.23941433e+00 4.48819213e-02 -2.17711732e-01 -1.19854593e+00
-7.12046146e-01 -2.37844840e-01 -7.18727708e-01 -5.95143855e-01
6.68695390e-01 -1.93937242e-01 -6.08831823e-01 5.65750778e-01
-9.40387428e-01 1.73806101e-01 -9.54959512e-01 6.68795049e-01
-7.60195553e-01 2.40582243e-01 -3.84893656e-01 -9.70447361e-01
-3.30391347e-01 -9.51400638e-01 5.33239603e-01 1.79012522e-01
-4.25463796e-01 -7.70495772e-01 9.46933031e-02 8.42835009e-02
5.80846369e-01 2.45088950e-01 1.10563910e+00 -9.95127559e-01
3.89495075e-01 -2.58253366e-01 4.35368448e-01 7.71310687e-01
7.18045950e-01 1.32305518e-01 -1.61426020e+00 -3.33240569e-01
4.37903166e-01 -1.46614164e-01 5.83685875e-01 4.29281108e-02
1.08529687e+00 -5.01162469e-01 3.60855252e-01 5.38811803e-01
6.84526682e-01 5.70525765e-01 3.57247144e-01 8.48284066e-02
4.83866066e-01 9.24990892e-01 5.81840932e-01 1.51164949e-01
-6.31798208e-01 9.34547424e-01 3.66921306e-01 -3.03969771e-01
-2.46860474e-01 -3.73804301e-01 3.58909339e-01 1.41860688e+00
-5.47284722e-01 9.32075381e-02 -4.41563547e-01 3.28147709e-01
-1.18160582e+00 -9.50612903e-01 9.24854074e-03 2.25578856e+00
8.81778002e-01 3.28347832e-02 3.93389672e-01 8.65594447e-01
9.39429462e-01 4.00038332e-01 -3.90515745e-01 -8.78809392e-01
-6.70114532e-02 7.08357275e-01 -2.45566338e-01 3.50450665e-01
-1.06288564e+00 7.88284004e-01 5.89077187e+00 8.56945813e-01
-1.57826781e+00 -8.00746232e-02 8.56979415e-02 -1.05404314e-02
-2.36201644e-01 -2.98401564e-01 -2.06000924e-01 1.61831647e-01
9.67548370e-01 -1.43310696e-01 7.60696769e-01 6.80494249e-01
2.08938330e-01 5.64912736e-01 -1.15758181e+00 8.59221637e-01
1.17767960e-01 -8.15135896e-01 1.42387018e-01 -2.13710696e-01
4.89179969e-01 -8.01655948e-01 2.59518653e-01 3.81156921e-01
-7.29849815e-01 -1.10470569e+00 7.24667728e-01 6.59252644e-01
8.60850215e-01 -8.00102174e-01 6.98633432e-01 2.65324444e-01
-1.13340652e+00 -8.94836634e-02 1.15923826e-02 -3.34162712e-01
-2.85468902e-02 2.28570819e-01 -1.12328672e+00 5.33694148e-01
4.78368193e-01 4.97240752e-01 -4.29953873e-01 6.42889678e-01
-2.28340387e-01 9.93372202e-01 -5.17054051e-02 -8.81462917e-02
4.12424421e-03 -1.79573566e-01 9.05180931e-01 1.12033319e+00
4.07730162e-01 -4.62578416e-01 -3.02109331e-01 9.32446599e-01
6.53914660e-02 5.31383097e-01 -5.75431168e-01 -3.19142252e-01
4.35715109e-01 1.27997553e+00 -3.94831359e-01 -1.34846196e-01
1.91692740e-01 7.01649964e-01 -7.36453161e-02 3.43281835e-01
-6.23566210e-01 -5.62070608e-01 5.70251226e-01 2.76251554e-01
2.53909141e-01 2.28857532e-01 3.10305089e-01 -6.23929620e-01
-1.71551615e-01 -1.06899059e+00 1.14141569e-01 -6.46008492e-01
-1.27262962e+00 1.15228045e+00 -8.24902207e-02 -1.43550634e+00
-7.02669919e-01 -4.36894357e-01 -8.48636448e-01 8.62295628e-01
-1.26881838e+00 -9.82334673e-01 -1.83642805e-01 4.79273617e-01
3.84925663e-01 -3.21110129e-01 1.24309933e+00 2.50972539e-01
-3.17537636e-01 7.45771468e-01 -7.20381513e-02 -1.05533898e-01
7.36384332e-01 -1.19872022e+00 -1.24806762e-01 3.31881970e-01
4.62481856e-01 5.77805400e-01 8.87200117e-01 4.38826717e-02
-1.00704145e+00 -8.27078760e-01 8.55775654e-01 2.58650631e-01
6.59042060e-01 -3.79290283e-01 -1.06618464e+00 1.36776239e-01
1.82397187e-01 1.82258859e-02 7.83327401e-01 7.29835108e-02
-3.53718877e-01 -4.41231608e-01 -1.02500165e+00 3.36416572e-01
8.56810629e-01 -8.51767957e-01 -1.07112038e+00 9.97685120e-02
5.24304152e-01 -6.41290471e-02 -1.00074077e+00 5.79426944e-01
7.69148588e-01 -1.13078785e+00 8.18797767e-01 -7.86270380e-01
1.39478117e-01 -4.40526187e-01 4.72082123e-02 -1.55859458e+00
-2.22623229e-01 -7.50194430e-01 2.34982848e-01 1.38448024e+00
2.92779326e-01 -7.22708166e-01 1.66219711e-01 -3.00664812e-01
-2.76615709e-01 -3.64177823e-01 -1.04364824e+00 -9.94040132e-01
1.34493291e-01 -2.65736431e-01 7.04998791e-01 1.18008745e+00
2.14483421e-02 3.35211545e-01 -4.44257468e-01 -8.27299878e-02
3.82134095e-02 3.21863979e-01 6.53240919e-01 -1.37215292e+00
-4.86322641e-01 -4.61653739e-01 -5.01917899e-01 -4.97433960e-01
3.24106246e-01 -1.00438321e+00 1.61937431e-01 -5.97298801e-01
-2.89538205e-01 -1.00935146e-01 -6.36381030e-01 3.52259457e-01
-6.15646690e-03 5.42219996e-01 4.88544852e-01 1.37695402e-01
2.07535267e-01 6.00405931e-01 1.44193399e+00 -1.70874566e-01
-5.65964878e-01 5.66067517e-01 -3.68562818e-01 9.77113187e-01
8.73354197e-01 -3.56162369e-01 -2.95078903e-01 2.81581491e-01
-1.78723827e-01 1.00770280e-01 1.96610406e-01 -1.19194043e+00
-6.37823939e-02 1.88203424e-01 1.89898163e-01 -3.03194106e-01
5.18405080e-01 -9.02205229e-01 3.59150976e-01 5.29267728e-01
-5.48200428e-01 -3.31923127e-01 3.80086660e-01 1.17094554e-01
-6.07642531e-01 -5.65285742e-01 8.75040054e-01 1.69628710e-01
-1.83333531e-01 -3.54888946e-01 -6.91010654e-01 -1.99181825e-01
7.18371153e-01 -4.38038051e-01 5.44212349e-02 -6.16170883e-01
-1.17858994e+00 -6.72830403e-01 1.60721987e-02 4.50520992e-01
3.47167790e-01 -1.23809564e+00 -5.79421401e-01 5.91981590e-01
4.97829318e-02 -4.65711534e-01 3.46773922e-01 8.13776731e-01
-3.00313067e-02 9.70927477e-02 -4.60762233e-01 -5.46346307e-01
-1.47394574e+00 5.38401127e-01 6.17527425e-01 1.16253540e-01
-1.44518569e-01 5.45545220e-01 2.20878050e-01 -7.20169604e-01
2.06570357e-01 -4.49543864e-01 -7.28333592e-01 5.97663105e-01
1.21847883e-01 3.86204749e-01 2.27658316e-01 -7.94079244e-01
-1.47982121e-01 9.08638835e-01 2.59600431e-01 1.21877797e-01
1.34926140e+00 2.46813849e-01 3.11062988e-02 9.02520776e-01
1.31465137e+00 5.60118020e-01 -7.65858054e-01 1.41838714e-01
-1.07786104e-01 -2.16910541e-01 -1.02592394e-01 -6.96383476e-01
-1.00187981e+00 9.06468987e-01 7.64970422e-01 6.05518878e-01
1.45960760e+00 -1.64864525e-01 3.01985502e-01 2.57504284e-01
1.73940472e-02 -9.24819231e-01 9.95012596e-02 4.01770115e-01
1.23065245e+00 -5.92126966e-01 -3.15266877e-01 -3.21029842e-01
-4.39387769e-01 1.45008445e+00 3.00470293e-01 -1.11211613e-01
4.49053317e-01 -3.78251560e-02 3.42183113e-01 -2.78254244e-02
-2.33927920e-01 -2.52677709e-01 7.66887605e-01 6.09053075e-01
5.24914742e-01 9.83038992e-02 -4.57872659e-01 7.63580501e-01
-9.21177030e-01 -4.14540857e-01 2.50526786e-01 3.20099175e-01
-3.36132497e-01 -1.06900752e+00 -6.04588449e-01 2.14456245e-01
-2.03936219e-01 9.70220342e-02 -6.67163908e-01 8.98913085e-01
4.22461957e-01 1.02802896e+00 1.63316771e-01 -7.54965425e-01
8.33446443e-01 3.90370816e-01 4.24610227e-01 -6.41680002e-01
-1.17954576e+00 3.53083372e-01 1.34733602e-01 -1.06747650e-01
-6.00301206e-01 -4.57014054e-01 -1.06465352e+00 2.31289014e-01
-4.89629894e-01 6.73934281e-01 6.73585057e-01 8.45376194e-01
-1.86313525e-01 9.01364803e-01 1.32600462e+00 -1.12150478e+00
-7.36953855e-01 -1.24648011e+00 -8.67027283e-01 5.50319552e-01
4.34190422e-01 -6.81781948e-01 -8.91813278e-01 9.31946039e-02] | [15.613308906555176, 5.732864856719971] |
bac66a7e-953f-4e27-ba80-f7de91cca8aa | implementation-of-tiny-machine-learning | 2207.12866 | null | https://arxiv.org/abs/2207.12866v1 | https://arxiv.org/pdf/2207.12866v1.pdf | Implementation Of Tiny Machine Learning Models On Arduino 33 BLE For Gesture And Speech Recognition | In this article gesture recognition and speech recognition applications are implemented on embedded systems with Tiny Machine Learning (TinyML). It features 3-axis accelerometer, 3-axis gyroscope and 3-axis magnetometer. The gesture recognition,provides an innovative approach nonverbal communication. It has wide applications in human-computer interaction and sign language. Here in the implementation of hand gesture recognition, TinyML model is trained and deployed from EdgeImpulse framework for hand gesture recognition and based on the hand movements, Arduino Nano 33 BLE device having 6-axis IMU can find out the direction of movement of hand. The Speech is a mode of communication. Speech recognition is a way by which the statements or commands of human speech is understood by the computer which reacts accordingly. The main aim of speech recognition is to achieve communication between man and machine. Here in the implementation of speech recognition, TinyML model is trained and deployed from EdgeImpulse framework for speech recognition and based on the keywords pronounced by human, Arduino Nano 33 BLE device having built-in microphone can make an RGB LED glow like red, green or blue based on keyword pronounced. The results of each application are obtained and listed in the results section and given the analysis upon the results. | ['Nishant Mohan', 'Viveka Simha PJ', 'Prem Chowdary Kakarla', 'Raghavendra Prasanna', 'Ramachandra A. C', 'Viswanatha V'] | 2022-07-23 | null | null | null | null | ['hand-gesture-recognition', 'hand-gesture-recognition-1', 'gesture-recognition'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-2.97181904e-01 -3.77855569e-01 -4.63066921e-02 -2.41627350e-01
1.89867571e-01 -4.68547314e-01 6.98743641e-01 -8.28392267e-01
-6.95400953e-01 3.83116305e-01 3.17358762e-01 -6.13877833e-01
1.09974548e-01 -4.23396468e-01 -2.22180024e-01 -7.62658834e-01
2.19546899e-01 1.71902597e-01 -5.98057322e-02 -1.30261183e-01
1.48762107e-01 6.73312843e-01 -1.48571968e+00 -6.60685897e-02
-3.84308159e-01 9.41066325e-01 1.34453773e-01 1.57697940e+00
-1.51463732e-01 6.29168451e-01 -7.79436648e-01 7.00875461e-01
1.39450744e-01 -2.50698268e-01 -5.58061063e-01 -1.63993180e-01
-4.81455237e-01 -7.49011576e-01 -2.25992680e-01 7.68552065e-01
9.74391043e-01 -3.91554683e-02 4.91937548e-01 -1.17212117e+00
-2.77027756e-01 3.15626144e-01 5.24462983e-02 8.13647062e-02
1.22993731e+00 3.53821546e-01 1.54897928e-01 -5.75760961e-01
5.00327647e-01 1.11986613e+00 3.00202101e-01 1.12621474e+00
-1.09349936e-01 -6.50105774e-01 -5.04343450e-01 2.87618846e-01
-1.11802185e+00 -2.38643706e-01 5.46942592e-01 -2.82784045e-01
1.65891254e+00 9.79375064e-01 8.74893844e-01 1.06813633e+00
4.17465270e-01 5.26095510e-01 1.20939505e+00 -7.73876607e-01
4.55161095e-01 5.11881597e-02 4.16493744e-01 6.01440907e-01
-3.61543953e-01 3.15807849e-01 -6.15446985e-01 5.14230765e-02
8.36152077e-01 4.15404499e-01 -2.15712041e-01 4.12517369e-01
-1.57303321e+00 4.75475900e-02 6.95005953e-02 6.72175348e-01
-2.81462729e-01 2.73939520e-01 5.46915531e-02 6.47365451e-01
-4.28023070e-01 -3.23171109e-01 -5.54273784e-01 -1.14922476e+00
-3.21195185e-01 -4.74471450e-01 1.25604045e+00 9.52523232e-01
2.13799085e-02 2.68910646e-01 4.84581977e-01 2.32353866e-01
1.12820280e+00 1.09299743e+00 1.24178517e+00 -2.32795894e-01
2.24201396e-01 7.97127843e-01 1.25049427e-01 -4.55116898e-01
-7.13971794e-01 6.86499476e-01 -5.37944615e-01 8.03699553e-01
2.79113442e-01 -6.89882278e-01 -9.39454436e-01 6.51763916e-01
7.87178755e-01 2.78355062e-01 3.29123467e-01 1.11164522e+00
1.46773350e+00 5.77653944e-01 -5.07050306e-02 -4.39786911e-02
1.66168904e+00 -5.65273881e-01 -8.28861713e-01 3.07978302e-01
7.33427286e-01 -9.78046536e-01 1.12201571e+00 4.32357907e-01
-2.57716447e-01 -3.46818089e-01 -1.21098900e+00 -2.51206420e-02
-7.08106875e-01 1.50051340e-01 6.10589504e-01 1.27250874e+00
-9.74666774e-01 2.04057381e-01 -1.45096028e+00 -9.95335460e-01
-4.47106838e-01 9.57526803e-01 -4.29011017e-01 7.60799527e-01
-6.36617959e-01 9.50928807e-01 1.07514009e-01 2.35717818e-01
-2.65439954e-02 2.85897970e-01 -5.28327286e-01 -4.20066506e-01
-3.94987732e-01 -4.11281377e-01 1.24301970e+00 -6.34820640e-01
-2.76747489e+00 8.46819401e-01 -6.07750826e-02 -1.15011871e-01
5.87699831e-01 1.49470251e-02 -9.32337761e-01 -4.97792521e-03
-6.74193680e-01 3.00280541e-01 8.52076530e-01 -3.54428589e-01
-6.53272569e-01 -7.59872913e-01 -3.39410454e-01 1.55953884e-01
-1.10662468e-01 4.97539192e-01 2.43039623e-01 -2.14447588e-01
3.49187851e-01 -8.49818707e-01 5.13465941e-01 -2.33688757e-01
-4.03041333e-01 -3.70923996e-01 1.72850966e+00 -5.44728816e-01
9.96951938e-01 -1.93984377e+00 -1.66143104e-01 2.23275244e-01
-7.65044391e-02 5.94241619e-01 5.52424073e-01 2.89209664e-01
1.45573840e-01 -1.53266743e-01 4.90739554e-01 1.34850457e-01
1.99769467e-01 4.94198918e-01 -1.81101143e-01 4.83165532e-01
-6.69081628e-01 8.03183854e-01 -5.80319166e-01 -7.49822780e-02
1.00808847e+00 7.90633380e-01 2.94486552e-01 3.75003219e-01
3.41782987e-01 9.62744117e-01 -7.15133190e-01 9.42036688e-01
2.42576331e-01 3.45988870e-01 -2.22309470e-01 1.94574326e-01
-3.92870128e-01 2.07692862e-01 -1.40372884e+00 1.24517047e+00
-5.50215304e-01 8.81745219e-01 1.72224060e-01 -5.30563474e-01
1.10383153e+00 9.02672589e-01 -1.41312080e-02 -4.81087238e-01
7.00012207e-01 5.40558457e-01 -1.02867216e-01 -1.35529816e+00
-3.53029191e-01 -8.26359987e-02 3.53722066e-01 8.97756755e-01
-2.75841624e-01 -2.85622180e-01 -6.10466897e-01 -4.85780001e-01
1.22784448e+00 2.56990910e-01 5.15021563e-01 4.63735998e-01
7.57742584e-01 -4.05087531e-01 -2.24366546e-01 2.98028618e-01
-3.21018964e-01 1.36576816e-01 -2.54947066e-01 -7.88940787e-01
-5.91137648e-01 -7.80513167e-01 9.98037010e-02 1.01586056e+00
-9.73667055e-02 -4.26395424e-02 -7.36065865e-01 -2.77115077e-01
-2.12198451e-01 2.75360048e-01 -7.43229762e-02 4.59382802e-01
-7.03799129e-01 -2.95721352e-01 7.59402692e-01 5.77745616e-01
7.61974752e-01 -1.60237753e+00 -1.48831725e+00 2.44817033e-01
5.81857860e-01 -1.06843448e+00 2.35382736e-01 2.20029727e-02
-9.95058954e-01 -1.00142622e+00 -6.76314771e-01 -9.15934503e-01
3.84242773e-01 -7.42921010e-02 2.57218301e-01 3.37568462e-01
-3.21988195e-01 1.05809009e+00 -7.44189024e-01 -7.04249382e-01
-1.97557330e-01 -3.86733830e-01 4.02910650e-01 -2.10231856e-01
1.17571819e+00 -6.98171079e-01 -5.90423226e-01 1.53956473e-01
-3.60914081e-01 -1.72418535e-01 2.16798663e-01 2.12889090e-01
-6.40585199e-02 -6.09465539e-01 -9.37817842e-02 1.78929597e-01
5.91549814e-01 -9.98753905e-02 -2.49477729e-01 1.33954778e-01
-2.77806193e-01 -1.43924162e-01 2.91957647e-01 -5.21713853e-01
-5.05169570e-01 2.71060526e-01 -4.41959262e-01 3.04123968e-01
-6.89553082e-01 -1.18948229e-01 -2.50866294e-01 -2.39929825e-01
4.34453994e-01 5.00612855e-01 1.37333777e-02 -5.75161517e-01
2.04333574e-01 2.12189317e+00 4.30795878e-01 1.06375612e-01
7.49904886e-02 3.88078779e-01 -2.92685300e-01 -1.41128671e+00
7.68042207e-01 -5.49169600e-01 -8.01417053e-01 -5.65886021e-01
1.09990883e+00 -4.27598268e-01 -1.47745597e+00 1.03350019e+00
-1.32327545e+00 -4.98877406e-01 2.35219836e-01 1.03379667e+00
-4.23635304e-01 1.24413818e-01 -5.43921471e-01 -1.34172547e+00
-8.00424278e-01 -1.00328493e+00 1.06142509e+00 6.34913504e-01
-6.48180187e-01 -8.48904788e-01 -9.73471254e-02 1.66486278e-01
2.81694442e-01 1.34875759e-01 2.42804006e-01 -6.61040425e-01
-2.22122416e-01 -7.79626310e-01 3.07017237e-01 5.25676757e-02
6.11227334e-01 1.91991374e-01 -7.64149666e-01 1.85658246e-01
3.05266142e-01 2.35431030e-01 9.78855938e-02 1.48495734e-01
5.41423678e-01 -3.55707645e-01 -4.80403036e-01 6.14829361e-01
1.10172641e+00 8.53434861e-01 6.26832485e-01 5.12923777e-01
8.52128208e-01 1.67152956e-01 1.26772642e-01 3.38350654e-01
3.90638143e-01 6.64564848e-01 1.12361878e-01 4.14447725e-01
-2.05563590e-01 1.56940088e-01 7.57742882e-01 1.04738653e+00
-6.14065170e-01 8.82103853e-03 -9.98467088e-01 -3.46709430e-01
-1.40713918e+00 -8.16023886e-01 -4.27351803e-01 2.17835927e+00
4.29830581e-01 -2.60905206e-01 3.58580977e-01 7.29985654e-01
5.87843597e-01 -6.82939664e-02 -3.70210111e-01 -9.27242577e-01
2.76751161e-01 5.07460058e-01 4.13031191e-01 8.88157845e-01
-7.63635278e-01 9.48040843e-01 5.27543449e+00 8.48549530e-02
-2.10312390e+00 3.30829442e-01 -3.34036440e-01 -8.29771906e-02
3.80951494e-01 -4.91708636e-01 -6.85300708e-01 7.87550151e-01
1.07186103e+00 3.71689916e-01 7.28806794e-01 7.19642937e-01
5.59456885e-01 -3.83573085e-01 -1.26218712e+00 1.56662917e+00
-1.50632545e-01 -8.27412128e-01 -3.25967968e-01 -1.18744388e-01
-3.09629329e-02 2.16977775e-01 -4.16491508e-01 -2.74675265e-02
-1.34884268e-01 -9.57334518e-01 5.57915270e-01 4.50425744e-01
9.22008276e-01 -2.11764321e-01 6.09681964e-01 5.89837074e-01
-1.27333915e+00 -1.76066265e-01 -7.03973249e-02 -1.00957870e+00
1.39316216e-01 -2.94935405e-01 -1.18957257e+00 -2.39186451e-01
6.02931738e-01 1.78029984e-01 7.67040476e-02 5.34973025e-01
-7.05034792e-01 6.67595804e-01 -9.01580095e-01 -1.26077509e+00
1.22965932e-01 -1.90432861e-01 7.85833716e-01 1.35157251e+00
6.39316916e-01 6.13950849e-01 -5.28046072e-01 2.00514674e-01
5.14760852e-01 2.25360870e-01 -7.65551984e-01 -1.00455262e-01
4.43737566e-01 1.00695777e+00 -8.64595294e-01 -4.87117201e-01
-1.24243259e-01 1.14656460e+00 -8.39233637e-01 3.14662606e-01
-3.11862588e-01 -7.14444041e-01 4.16557878e-01 -1.04043461e-01
-3.78042758e-02 -6.21404350e-01 -5.96507788e-01 -8.33288729e-01
2.26468518e-01 -6.54738486e-01 -1.70461964e-02 -1.05476582e+00
-2.52836496e-01 3.31906587e-01 -5.65905690e-01 -1.32251644e+00
-6.33092642e-01 -1.13063765e+00 -9.24698591e-01 7.59488940e-01
-5.86968243e-01 -1.25089014e+00 -4.54149842e-01 9.06988800e-01
6.37782991e-01 -4.89850849e-01 1.36003482e+00 -1.72693282e-02
-3.22449654e-01 1.98708773e-01 7.43112415e-02 5.23523331e-01
2.65314847e-01 -1.17071199e+00 2.53616869e-01 4.50343877e-01
2.13095352e-01 7.82463670e-01 5.48992336e-01 -6.00251675e-01
-2.04114199e+00 3.91131602e-02 1.00813401e+00 -6.89564288e-01
5.25367975e-01 -1.47363693e-01 -2.02183500e-01 8.51264417e-01
4.70688939e-02 1.06379770e-01 7.66613603e-01 -4.32358205e-01
2.18577846e-03 -1.66621134e-01 -1.39870203e+00 5.62165320e-01
7.63525546e-01 -6.57267869e-01 -7.93906987e-01 4.41485226e-01
1.50000870e-01 -8.34894896e-01 -5.34336686e-01 -3.73347908e-01
1.39713657e+00 -8.62372160e-01 7.19972014e-01 -6.18877351e-01
-3.05313379e-01 -3.12453985e-01 -2.58214921e-01 -8.88372421e-01
6.56895638e-01 -1.00902796e+00 -3.65596652e-01 8.69638264e-01
1.16800502e-01 -1.30568254e+00 6.98796034e-01 1.01151979e+00
4.04063731e-01 -2.78031230e-01 -1.40908539e+00 -5.72419465e-01
-5.94414890e-01 -8.28046978e-01 8.92797947e-01 4.38842058e-01
9.36869144e-01 2.42940173e-01 -1.28286704e-01 1.75792426e-01
1.61751628e-01 -5.87801754e-01 1.09907591e+00 -8.77020955e-01
4.77154367e-02 -3.76212686e-01 -1.27472067e+00 -1.38536286e+00
-3.82719040e-01 -3.87370199e-01 -3.51740569e-01 -1.39582479e+00
-7.10857093e-01 -8.07838142e-02 9.24445242e-02 4.05662596e-01
5.01864493e-01 2.74475180e-02 1.36761785e-01 2.67654121e-01
-1.99062452e-01 -2.67997295e-01 9.78328526e-01 -1.11908630e-01
-6.03861272e-01 4.37006921e-01 3.67890298e-01 7.96073735e-01
7.90341377e-01 -2.31730759e-01 2.19272580e-02 -3.43739122e-01
1.33895382e-01 2.73755997e-01 6.01794302e-01 -1.18975353e+00
8.10496628e-01 2.45543167e-01 5.31045139e-01 -4.08945471e-01
2.87802666e-01 -1.24678624e+00 3.99302430e-02 1.06314731e+00
3.37169349e-01 1.70724958e-01 -3.38337690e-01 9.02077481e-02
2.44536176e-01 -6.98090792e-02 3.60456705e-02 -4.82349396e-02
-1.01183307e+00 -3.19995582e-01 -8.65805686e-01 -1.04025865e+00
1.00015759e+00 -9.87409353e-01 -4.29071821e-02 -5.37220061e-01
-1.02182937e+00 -2.22567946e-01 9.54352692e-02 5.48531294e-01
7.98799157e-01 -1.29021764e+00 -8.37124959e-02 8.56084824e-01
-4.38118011e-01 -2.67794311e-01 -2.85379618e-01 8.69536936e-01
-1.14885843e+00 8.58543873e-01 -3.06778103e-01 -6.62891209e-01
-1.92901838e+00 -1.14394687e-01 5.22518277e-01 6.50023460e-01
-6.57548726e-01 8.96591842e-01 -1.13090968e+00 -6.34282947e-01
6.07006907e-01 -9.37761486e-01 -5.26365519e-01 -4.07038808e-01
1.09326935e+00 8.22138488e-01 4.63167205e-02 -8.93214762e-01
-8.19115639e-01 1.06244838e+00 6.88180745e-01 -7.63908744e-01
1.24887919e+00 -3.80597621e-01 -1.63143292e-01 7.83999801e-01
1.12997615e+00 -9.42688957e-02 -6.71488941e-01 4.80386615e-01
3.02173980e-02 -3.23091447e-01 -2.62159944e-01 -1.09946406e+00
-5.28305411e-01 9.55515683e-01 1.52347386e+00 1.99965760e-01
7.73261786e-01 -3.19923908e-01 9.99385595e-01 7.69147038e-01
7.73681700e-01 -1.33444536e+00 -3.36871833e-01 5.33589244e-01
7.77424395e-01 -1.23173285e+00 -2.71492869e-01 1.76690683e-01
-2.72685796e-01 1.62234378e+00 2.18765616e-01 6.52970001e-02
1.07753301e+00 8.20117414e-01 8.65409553e-01 -8.22264701e-02
-3.48369195e-03 5.77560961e-02 1.85614116e-02 1.09171331e+00
4.82526839e-01 6.97607040e-01 -6.77144587e-01 5.79360485e-01
-7.20409811e-01 5.54269731e-01 1.48216426e-01 1.29118633e+00
-8.13640773e-01 -9.57136631e-01 -9.40189421e-01 6.85998574e-02
-3.14780921e-01 2.59003520e-01 -6.02596998e-01 6.19185865e-01
1.89611852e-01 1.49612606e+00 2.04784125e-02 -9.72969830e-01
2.28185490e-01 2.13332713e-01 6.67251527e-01 -1.94211766e-01
-6.55050576e-01 -2.04743847e-01 -1.79810032e-01 -5.66205204e-01
-3.16669703e-01 -4.01452065e-01 -2.09845567e+00 -2.99819559e-01
-4.47916873e-02 -2.17160091e-01 1.61417997e+00 1.38265681e+00
1.24111958e-01 -3.26789469e-02 3.68966073e-01 -9.97613311e-01
-6.31983042e-01 -1.35816693e+00 -6.02422714e-01 -1.07773744e-01
6.23254359e-01 -2.92145789e-01 -3.77357244e-01 2.75046490e-02] | [6.517053604125977, -0.21946744620800018] |
b205444b-8691-400a-a482-28270aa42cc9 | probabilistic-deep-learning-for-instance | 2008.10678 | null | https://arxiv.org/abs/2008.10678v2 | https://arxiv.org/pdf/2008.10678v2.pdf | Probabilistic Deep Learning for Instance Segmentation | Probabilistic convolutional neural networks, which predict distributions of predictions instead of point estimates, led to recent advances in many areas of computer vision, from image reconstruction to semantic segmentation. Besides state of the art benchmark results, these networks made it possible to quantify local uncertainties in the predictions. These were used in active learning frameworks to target the labeling efforts of specialist annotators or to assess the quality of a prediction in a safety-critical environment. However, for instance segmentation problems these methods are not frequently used so far. We seek to close this gap by proposing a generic method to obtain model-inherent uncertainty estimates within proposal-free instance segmentation models. Furthermore, we analyze the quality of the uncertainty estimates with a metric adapted from semantic segmentation. We evaluate our method on the BBBC010 C.\ elegans dataset, where it yields competitive performance while also predicting uncertainty estimates that carry information about object-level inaccuracies like false splits and false merges. We perform a simulation to show the potential use of such uncertainty estimates in guided proofreading. | ['Dagmar Kainmueller', 'Lisa Mais', 'Josef Lorenz Rumberger'] | 2020-08-24 | null | null | null | null | ['probabilistic-deep-learning'] | ['computer-vision'] | [ 5.50849020e-01 7.72131920e-01 -1.89053953e-01 -8.67041707e-01
-1.18813443e+00 -6.41278267e-01 6.35637879e-01 5.76455653e-01
-7.39351451e-01 9.63275969e-01 -3.06297719e-01 -1.60898462e-01
-2.01902136e-01 -7.08924055e-01 -1.21256101e+00 -5.04065037e-01
-3.78730968e-02 8.82900417e-01 6.82095587e-01 3.15038115e-01
5.84083557e-01 3.92471820e-01 -1.50754702e+00 3.32620502e-01
9.49620366e-01 1.01625717e+00 2.04476997e-01 4.72114086e-01
-2.18489468e-01 3.39463532e-01 -7.22746193e-01 -6.66836441e-01
-5.07155284e-02 7.86652043e-03 -1.01354408e+00 -1.95111305e-01
4.14943427e-01 -2.77641296e-01 5.75652957e-01 1.26083994e+00
2.42256716e-01 -2.05136418e-01 7.51220584e-01 -1.35801136e+00
-3.52498367e-02 1.03840804e+00 -3.65730077e-01 6.35852218e-02
-7.92375803e-02 3.57936998e-03 9.19631720e-01 -5.14880598e-01
8.01765680e-01 1.15162671e+00 8.44740689e-01 4.59954828e-01
-1.31436825e+00 -4.20269012e-01 5.85287735e-02 3.61947447e-01
-1.10816514e+00 -4.40249383e-01 3.57196391e-01 -5.63451648e-01
6.68176234e-01 -1.07026845e-02 3.16674769e-01 1.15021002e+00
2.01635972e-01 9.26917315e-01 9.96320486e-01 -3.91106397e-01
7.31365383e-01 7.58596286e-02 1.66961089e-01 7.00414658e-01
4.12809163e-01 1.00314066e-01 -5.38810968e-01 -6.92460611e-02
4.50760216e-01 -4.48589593e-01 -8.61940458e-02 -6.39864862e-01
-1.14077044e+00 7.46914923e-01 3.95173609e-01 8.14902261e-02
8.97034034e-02 5.43801486e-01 2.85471857e-01 -3.27941865e-01
6.16906345e-01 4.61919308e-01 -5.75406909e-01 -1.63872451e-01
-1.29588068e+00 3.82583410e-01 7.00237036e-01 8.56893361e-01
6.45351529e-01 -3.44510287e-01 -7.06634820e-02 5.27555346e-01
5.92403293e-01 -1.64656639e-02 6.62942082e-02 -1.41443956e+00
2.60557830e-01 4.71189708e-01 2.77041376e-01 -7.27370918e-01
-5.99087715e-01 -3.07073265e-01 -4.07228529e-01 6.75084949e-01
9.57395315e-01 5.33651784e-02 -1.11808395e+00 1.74994433e+00
3.52105111e-01 1.78070918e-01 -2.52866000e-01 6.82553828e-01
4.79892284e-01 2.46033087e-01 1.29583731e-01 -4.56858985e-02
7.71724045e-01 -5.40265083e-01 -5.18059790e-01 9.42054484e-03
7.22887576e-01 -4.88658935e-01 6.03399277e-01 5.87560117e-01
-9.40880179e-01 -2.32506722e-01 -1.01049614e+00 2.70054229e-02
-5.36067009e-01 2.05798633e-02 5.40724039e-01 6.66759431e-01
-9.76659238e-01 1.11292338e+00 -1.10510874e+00 -2.77380764e-01
8.96470428e-01 3.79186749e-01 -1.83262020e-01 3.50314021e-01
-1.01396048e+00 1.03456914e+00 7.76848793e-01 9.93326455e-02
-9.88929689e-01 -8.40794384e-01 -7.73410916e-01 1.66842267e-01
5.31464994e-01 -2.87912637e-01 1.34144843e+00 -9.77779806e-01
-1.32824504e+00 9.88180339e-01 2.96126693e-01 -9.10786152e-01
9.07753706e-01 -1.35262743e-01 1.64841712e-01 -1.53028131e-01
1.61952153e-01 1.31626093e+00 4.56705302e-01 -1.29846275e+00
-6.90720558e-01 -5.39177001e-01 2.33450621e-01 -2.00784221e-01
3.62836629e-01 -1.58720315e-01 -2.33547434e-01 -3.88857067e-01
1.81772217e-01 -7.97300637e-01 -4.80764449e-01 7.28208363e-01
-7.05571294e-01 -1.42061144e-01 6.13806784e-01 -4.19825554e-01
4.21469092e-01 -1.65619218e+00 1.03880450e-01 1.36786520e-01
1.26102790e-01 1.76353022e-01 1.33738577e-01 -1.56573998e-03
1.02829032e-01 4.51642245e-01 -8.93040061e-01 -4.62990999e-01
1.53158352e-01 2.73554444e-01 -2.12183237e-01 5.24827182e-01
5.97727895e-01 7.62187243e-01 -7.95353591e-01 -7.18442142e-01
1.94464892e-01 3.03990811e-01 -4.30093557e-01 -6.62536323e-02
-7.30233669e-01 3.90022069e-01 -2.50818372e-01 6.18533432e-01
6.63680732e-01 -2.21674204e-01 1.03985080e-02 -1.37375973e-04
-8.16137940e-02 -5.14355525e-02 -9.98297155e-01 2.03333497e+00
-1.65601075e-02 5.26418686e-01 -4.88268733e-02 -1.31039548e+00
6.75670266e-01 -7.56228492e-02 3.43776286e-01 -3.37425560e-01
2.34943941e-01 2.19591528e-01 -5.85684506e-03 -1.07509762e-01
5.64789534e-01 1.04923934e-01 -5.46948612e-02 1.37601167e-01
4.03791785e-01 -4.10337567e-01 2.79992819e-01 2.43098691e-01
1.07462013e+00 8.54086339e-01 3.48829985e-01 -5.36292732e-01
2.39456251e-01 2.61038661e-01 7.63316989e-01 7.35531688e-01
-5.02362669e-01 9.13996637e-01 8.05198133e-01 -3.11927319e-01
-1.10902667e+00 -1.09408045e+00 -4.83805180e-01 6.40163779e-01
3.07862073e-01 -1.30155966e-01 -1.13778210e+00 -9.63444710e-01
-1.28784686e-01 1.03185534e+00 -7.27929413e-01 -5.50922602e-02
-1.97370440e-01 -8.09185863e-01 7.60941446e-01 5.39288044e-01
2.74955422e-01 -1.01928568e+00 -9.94273961e-01 2.35747591e-01
1.20399771e-02 -1.01979101e+00 3.40767801e-01 6.77483857e-01
-8.62430334e-01 -1.28473377e+00 -8.02817643e-01 -2.15654910e-01
4.99394268e-01 -6.81803882e-01 1.28537405e+00 -7.52485767e-02
-4.67010438e-01 2.37691164e-01 -3.63568872e-01 -8.27393115e-01
-5.04102111e-01 -4.48804311e-02 -2.75319610e-02 -5.56384504e-01
1.09539159e-01 -2.73138791e-01 -2.38419354e-01 3.91511887e-01
-1.17116964e+00 -5.56098707e-02 3.56688052e-01 7.93111682e-01
9.17777479e-01 -1.30950570e-01 5.69672525e-01 -1.15177202e+00
3.71961206e-01 -3.25160772e-01 -1.18777835e+00 5.29190958e-01
-7.41387606e-01 2.64855534e-01 3.33519042e-01 -1.55352876e-01
-1.08211458e+00 3.61347049e-01 -2.63910383e-01 -1.83383435e-01
-4.49336290e-01 3.60188097e-01 -1.35999963e-01 -2.59889215e-02
8.06828976e-01 -2.12640628e-01 -1.52857348e-01 -2.85970569e-01
3.50370914e-01 3.49148571e-01 5.45651495e-01 -6.61573112e-01
3.08667272e-01 3.65879238e-01 1.70837373e-01 -3.20138365e-01
-9.42973137e-01 -1.42223641e-01 -7.50752151e-01 -2.31807381e-01
7.87940919e-01 -5.01879632e-01 -6.79700434e-01 4.39651698e-01
-1.46243799e+00 -4.07934248e-01 -3.84343624e-01 2.21773520e-01
-1.09195626e+00 4.51337010e-01 -2.32643649e-01 -9.40913081e-01
-1.08640023e-01 -1.52326739e+00 1.35733235e+00 3.83054256e-01
-3.21084112e-01 -8.33664536e-01 1.23272590e-01 4.04531866e-01
1.71611935e-01 2.85232574e-01 9.17510927e-01 -8.60013068e-01
-7.01059878e-01 5.62402792e-03 -2.12640479e-01 1.74604416e-01
-4.61836874e-01 3.54637891e-01 -1.20754111e+00 2.46088460e-01
-4.55262125e-01 -5.40702879e-01 1.00605750e+00 7.31753230e-01
1.60235643e+00 1.57308131e-01 -5.63420951e-01 3.23302716e-01
1.32531130e+00 8.27259347e-02 7.84884095e-01 2.67544061e-01
1.94244981e-01 8.01467776e-01 7.28454053e-01 2.21063301e-01
4.54256944e-02 7.60873020e-01 9.86814380e-01 2.73475587e-01
8.73472169e-02 2.24762894e-02 -8.60794485e-02 -6.94469959e-02
8.44613910e-02 -5.96405149e-01 -1.23574507e+00 5.76888323e-01
-2.09120035e+00 -6.51227474e-01 -1.29522383e-01 2.34246659e+00
8.31305325e-01 6.28834128e-01 -1.76586404e-01 2.89923877e-01
8.01409245e-01 -2.90224940e-01 -7.94628978e-01 -2.60002434e-01
8.82362351e-02 1.06220640e-01 6.35556579e-01 3.84844661e-01
-1.14212239e+00 9.28008795e-01 6.42621183e+00 9.97941375e-01
-6.72936618e-01 -8.64301547e-02 1.08041239e+00 2.27582157e-01
-2.63480932e-01 2.01052696e-01 -8.75667989e-01 5.60757041e-01
8.73476624e-01 3.03121239e-01 -1.37212470e-01 9.83622253e-01
-5.94862625e-02 -6.79663360e-01 -1.38817692e+00 6.38213396e-01
-2.62597561e-01 -1.63693893e+00 -2.11809099e-01 -3.78920324e-02
4.89681840e-01 4.99827564e-02 -1.79184414e-02 1.54256195e-01
6.52691543e-01 -1.07672930e+00 9.81355071e-01 9.17110562e-01
4.85836506e-01 -7.62495756e-01 9.20222461e-01 5.79463124e-01
-4.30763215e-01 2.85752594e-01 -5.11055887e-01 2.02945784e-01
2.68382549e-01 9.71159697e-01 -1.17098451e+00 2.88342625e-01
7.60527313e-01 2.44941637e-01 -4.51724380e-01 1.40642834e+00
-1.93893373e-01 4.86035496e-01 -5.79769254e-01 -3.60041745e-02
6.74846917e-02 4.20831665e-02 5.39319754e-01 1.02750838e+00
1.63804725e-01 -3.38650554e-01 -1.93047762e-01 1.56859875e+00
-1.84894264e-01 -2.72334158e-01 -3.85013014e-01 -1.09747358e-01
2.79774070e-01 1.19031513e+00 -1.34049582e+00 -1.40017986e-01
3.80275846e-01 8.13074052e-01 4.78988290e-01 1.11537641e-02
-1.08171380e+00 -1.25596285e-01 3.51720482e-01 -5.03868535e-02
1.52142733e-01 -3.58741321e-02 -4.37357455e-01 -7.77208149e-01
-1.01941399e-01 -3.96476537e-01 1.95456576e-02 -8.61030936e-01
-1.00869882e+00 2.39388481e-01 2.56956607e-01 -8.89354527e-01
-3.80203873e-01 -8.37606013e-01 -3.88142347e-01 5.66413343e-01
-1.29519641e+00 -8.60102832e-01 1.21598676e-01 -1.46169871e-01
6.29811108e-01 1.08304925e-01 8.16787779e-01 -1.00910336e-01
-4.20048714e-01 4.59068239e-01 1.07006334e-01 -8.84223580e-02
5.04779100e-01 -1.51218283e+00 4.50339854e-01 8.62230957e-01
2.16054410e-01 4.43865806e-01 9.73893404e-01 -6.81963146e-01
-5.08701682e-01 -9.69391823e-01 4.97693479e-01 -5.24213374e-01
4.68351424e-01 -2.61666715e-01 -7.73564339e-01 4.38969433e-01
-2.87008703e-01 2.99667418e-01 3.57121289e-01 1.23813696e-01
-3.26046854e-01 2.58276135e-01 -1.60624087e+00 3.02604377e-01
8.91920209e-01 -8.91087875e-02 -4.01470155e-01 2.20972568e-01
6.60493791e-01 -4.76645321e-01 -6.63098037e-01 7.78971314e-01
6.92912102e-01 -1.30059159e+00 8.08316290e-01 -5.69099665e-01
6.53258681e-01 -2.26691201e-01 -1.94866702e-01 -1.17469168e+00
7.64206424e-02 5.05787097e-02 -1.16581498e-02 1.30397177e+00
8.13693285e-01 -4.35678624e-02 1.13147473e+00 6.65698886e-01
-4.13503319e-01 -6.49353862e-01 -1.26572835e+00 -5.75250268e-01
2.48906732e-01 -6.28151298e-01 2.88219184e-01 6.52110577e-01
-2.61098444e-01 -1.38781339e-01 -7.87186157e-03 1.26816437e-01
8.84487212e-01 -1.89535692e-01 3.96876723e-01 -1.46586001e+00
-2.14848846e-01 -4.85813916e-01 -8.08194995e-01 -5.95509171e-01
3.75823677e-01 -6.01305008e-01 6.76416039e-01 -1.55089653e+00
1.18552215e-01 -6.01653755e-01 -1.04045980e-01 5.60888708e-01
-2.29988322e-02 1.78694010e-01 -2.94833793e-03 -8.34896341e-02
-8.75751972e-01 3.63516033e-01 7.78691828e-01 -3.78270686e-01
2.70780981e-01 1.36754245e-01 -1.40876412e-01 1.14293039e+00
7.79402256e-01 -7.53336608e-01 -3.19190174e-01 -3.43979478e-01
5.77419817e-01 5.62213585e-02 5.68286300e-01 -1.28722727e+00
1.17920212e-01 -1.22679316e-03 4.94412750e-01 -7.43726850e-01
3.20219904e-01 -9.43457425e-01 1.71016395e-01 3.39609534e-01
-7.17977226e-01 -5.92768013e-01 1.08010061e-01 9.40487742e-01
4.18548323e-02 -7.31161177e-01 9.85199213e-01 -3.43420595e-01
-7.96719432e-01 6.46031946e-02 -3.49854439e-01 -1.01734437e-01
1.03311944e+00 -2.94900268e-01 -3.29468578e-01 -1.40512288e-01
-8.14917326e-01 4.70127091e-02 5.78160167e-01 1.53311610e-01
3.88073385e-01 -6.68380618e-01 -4.13034081e-01 -2.83250868e-01
2.88209379e-01 3.97074878e-01 5.27241901e-02 7.17152417e-01
-6.90075934e-01 3.06294739e-01 -2.90308803e-01 -1.08574021e+00
-1.09305751e+00 2.95325220e-01 3.69170189e-01 -2.35247493e-01
-1.59343660e-01 1.13353813e+00 -1.32561792e-02 -7.77721763e-01
4.72634465e-01 -5.86198211e-01 -2.18404531e-01 -2.11536866e-02
1.70985714e-01 3.37257057e-01 2.27350593e-01 -3.63973051e-01
-3.29867452e-01 2.09583059e-01 -7.67817795e-02 -1.25531629e-01
1.19496667e+00 -7.96843842e-02 6.74367100e-02 6.50538325e-01
6.74370110e-01 -4.72811639e-01 -1.43357742e+00 -2.61547770e-02
6.18388712e-01 -1.73026860e-01 9.89276096e-02 -1.21522522e+00
-8.77623677e-01 1.03098619e+00 9.64161575e-01 1.70039013e-02
6.39375210e-01 2.22037639e-02 2.00666741e-01 5.44284046e-01
7.48685658e-01 -1.18678546e+00 -2.84989059e-01 3.02670807e-01
5.35442591e-01 -1.64631522e+00 1.34135813e-01 -3.90715212e-01
-3.77240598e-01 1.24153972e+00 4.95344788e-01 1.95425063e-01
3.80086571e-01 5.50142407e-01 -7.02098161e-02 -4.09946404e-02
-4.52532798e-01 -7.56233335e-02 9.00236517e-02 7.73273349e-01
3.99346054e-01 -2.40265741e-04 -2.94261724e-01 4.92092311e-01
5.36419079e-02 3.11982423e-01 7.51704395e-01 8.30241680e-01
-7.12101638e-01 -1.09219408e+00 -2.41186753e-01 4.77861166e-01
-6.47235692e-01 5.50293401e-02 -5.07440388e-01 6.56324863e-01
2.84040958e-01 6.99218631e-01 2.13264748e-02 -1.00544102e-01
-1.93624139e-01 6.48590699e-02 4.97637480e-01 -5.94347179e-01
-2.31659561e-01 -2.39692509e-01 2.18987301e-01 -7.08823621e-01
-7.23140359e-01 -4.30809170e-01 -1.47396350e+00 2.07252055e-01
-7.35898674e-01 8.55186880e-02 1.10091877e+00 1.20061004e+00
3.85563970e-02 4.91380423e-01 -1.92588434e-01 -8.76752377e-01
-5.57718813e-01 -9.68320727e-01 -5.19606531e-01 1.03134830e-02
-1.20879665e-01 -8.95442188e-01 -2.10201174e-01 1.29059806e-01] | [7.868232727050781, 3.1213345527648926] |
c87699cd-9e3d-4703-909c-9f8aab0f762f | online-boosting-based-target-identification | null | null | https://www.mdpi.com/1424-8220/22/21/8422 | https://www.mdpi.com/1424-8220/22/21/8422/pdf | Online Boosting-Based Target Identification among Similar Appearance for Person-Following Robots | It is challenging for a mobile robot to follow a specific target person in a dynamic environment, comprising people wearing similar-colored clothes and having the same or similar height. This study describes a novel framework for a person identification model that identifies a target person by merging multiple features into a single joint feature online. The proposed framework exploits the deep learning output to extract four features for tracking the target person without prior knowledge making it generalizable and more robust. A modified intersection over union between the current frame and the last frame is proposed as a feature to distinguish people, in addition to color, height, and location. To improve the performance of target identification in a dynamic environment, an online boosting method was adapted by continuously updating the features in every frame. Through extensive real-life experiments, the effectiveness of the proposed method was demonstrated by showing experimental results that it outperformed the previous methods. | ['Redhwan Algabri'] | 2022-11-02 | null | null | null | sensors-2022-11 | ['person-identification'] | ['computer-vision'] | [-1.20856222e-02 -4.32833225e-01 2.30886832e-01 -3.79217684e-01
-2.42692977e-01 -3.56843591e-01 4.61254328e-01 1.34227172e-01
-8.82661462e-01 6.47905767e-01 -2.23971397e-01 4.44949836e-01
1.11670330e-01 -6.71387434e-01 -6.43395364e-01 -8.06772351e-01
-1.32639512e-01 3.23051721e-01 4.90460753e-01 9.42635760e-02
3.73906456e-02 4.05973345e-01 -1.57853401e+00 -3.51073712e-01
5.64704180e-01 9.49861884e-01 3.88972342e-01 4.82890755e-01
3.96136701e-01 1.45208895e-01 -7.74622679e-01 -3.70388746e-01
6.34872854e-01 -9.95176733e-02 -1.04127795e-01 4.87773478e-01
5.71016967e-01 -5.42667270e-01 -2.32599363e-01 9.38866913e-01
5.70586145e-01 3.58278126e-01 3.48459840e-01 -1.27377164e+00
-3.81110072e-01 7.64582008e-02 -6.94333851e-01 1.24089040e-01
6.62928760e-01 2.08909497e-01 3.59878570e-01 -5.60860991e-01
1.36251852e-01 1.43115938e+00 1.07614481e+00 4.50891137e-01
-7.35180795e-01 -7.89126396e-01 7.03135788e-01 4.48367447e-01
-1.57666516e+00 -1.35643274e-01 5.36217153e-01 -3.82654399e-01
3.88900161e-01 8.53752345e-02 1.11207378e+00 9.52248096e-01
1.71048462e-01 5.29515147e-01 1.16364789e+00 -4.36364204e-01
1.91911461e-03 2.48341963e-01 2.95903951e-01 1.03911960e+00
7.09296644e-01 2.68942446e-01 -3.09436828e-01 -1.80589512e-01
5.82749963e-01 4.19899166e-01 -1.19062431e-01 -4.74572390e-01
-1.17232800e+00 6.67825937e-01 5.63773394e-01 2.32259240e-02
-4.74392176e-01 -1.04985349e-01 1.76536158e-01 -1.56522244e-01
3.88029292e-02 -4.18402672e-01 -2.17872426e-01 1.01550154e-01
-3.65076005e-01 3.82997274e-01 5.60521126e-01 9.60889578e-01
8.21193397e-01 -1.05579287e-01 -3.42078686e-01 6.59761012e-01
4.64535952e-01 1.11420131e+00 4.63954419e-01 -5.88225543e-01
2.03634977e-01 6.88216507e-01 7.83565700e-01 -1.25592887e+00
-6.72022045e-01 -6.28755748e-01 -7.34225512e-01 2.99390405e-01
6.67692542e-01 -5.13586640e-01 -7.61569083e-01 1.71954954e+00
7.10023940e-01 3.70654821e-01 -6.28554821e-02 1.27191126e+00
4.63361144e-01 4.48644400e-01 1.46261632e-01 -2.07880080e-01
1.59954941e+00 -8.27596188e-01 -6.00899577e-01 -4.65095907e-01
-4.53824177e-02 -4.90735501e-01 5.00466704e-01 2.49774784e-01
-5.72178662e-01 -1.06079030e+00 -1.15030622e+00 5.43227553e-01
-3.13784778e-01 4.92399067e-01 4.48131293e-01 1.02243817e+00
-7.22004712e-01 3.07349920e-01 -6.31697893e-01 -7.98243701e-01
-8.76529291e-02 4.55241948e-01 -2.01436698e-01 -4.59153019e-02
-1.16996515e+00 8.40759635e-01 3.07445019e-01 3.87437165e-01
-6.76553667e-01 -8.26542228e-02 -8.17889869e-01 -2.47411042e-01
1.83442935e-01 -9.06493306e-01 1.07663631e+00 -1.18466425e+00
-1.68218756e+00 6.19195223e-01 -3.72528195e-01 -5.75088143e-01
8.72282565e-01 -5.90184033e-01 -4.42057550e-01 -9.59959403e-02
2.03458324e-01 3.56865227e-01 1.01068151e+00 -1.26594543e+00
-1.21806288e+00 -7.57779837e-01 6.56040460e-02 5.67056239e-01
-2.13982746e-01 -7.40539059e-02 -8.32762480e-01 -3.93052906e-01
-6.56290948e-02 -1.10815704e+00 -1.27561793e-01 -8.48162845e-02
-1.82586327e-01 -2.15189189e-01 7.46924281e-01 -7.86847591e-01
6.66089177e-01 -1.88163006e+00 -9.61919427e-02 4.03665841e-01
-1.81452945e-01 1.86966226e-01 2.12471575e-01 -6.94047734e-02
4.49867278e-01 -6.65509820e-01 1.72785237e-01 -2.89336979e-01
-1.69058163e-02 -1.97103396e-01 3.19154292e-01 8.41583073e-01
-1.04884207e-01 4.31687772e-01 -8.34633768e-01 -4.86506999e-01
4.05575782e-01 7.94486940e-01 1.19171165e-01 3.08359563e-01
4.44063008e-01 4.56537098e-01 -3.64649355e-01 5.92334390e-01
8.40509117e-01 1.31562471e-01 1.56218022e-01 -1.41863033e-01
-1.63118497e-01 -5.17298222e-01 -1.59509516e+00 1.04132545e+00
-2.53826678e-01 3.92334074e-01 1.52367294e-01 -5.64028800e-01
1.21940517e+00 7.92354941e-02 5.09268522e-01 -6.06437922e-01
3.25283557e-01 -2.11316511e-01 -3.61818150e-02 -2.81860232e-01
4.00745511e-01 1.36790738e-01 -2.20978349e-01 -2.29323413e-02
-4.06168669e-01 8.11418116e-01 1.72475681e-01 -3.39024395e-01
8.29804420e-01 1.78998873e-01 4.36807722e-01 -3.80635411e-02
9.33821619e-01 -1.56590581e-01 8.92198920e-01 1.07682478e+00
-7.65748680e-01 2.70715475e-01 -5.75319409e-01 -7.26599455e-01
-6.83225453e-01 -1.16485512e+00 1.82206854e-01 1.08196747e+00
8.51610363e-01 1.05488189e-01 -7.35596299e-01 -5.62207818e-01
1.95553273e-01 2.66087335e-02 -6.03098512e-01 -1.70725793e-01
-6.77486658e-01 -7.79283226e-01 4.22189265e-01 3.87427002e-01
1.01556563e+00 -6.70177579e-01 -1.00984800e+00 1.65197372e-01
-3.06172580e-01 -1.17093647e+00 -6.20568335e-01 -2.56522834e-01
-3.04363072e-01 -1.07090092e+00 -1.11898482e+00 -9.42404389e-01
7.42082834e-01 6.95950329e-01 4.34994072e-01 1.17770419e-01
-2.41308033e-01 5.83194256e-01 -2.46810108e-01 -6.20320320e-01
4.57678325e-02 -3.57826799e-01 4.93894488e-01 5.90471923e-01
3.59493196e-01 8.54005441e-02 -8.51815164e-01 4.97778505e-01
-6.42670086e-03 -6.35642931e-02 6.14817977e-01 5.03376484e-01
2.37397581e-01 3.71939272e-01 3.41269791e-01 -3.90195549e-01
4.03362811e-01 -1.88654155e-01 -6.02906585e-01 4.00065243e-01
-3.65699261e-01 -4.02140647e-01 4.89522129e-01 -7.06725717e-01
-1.23114491e+00 4.43208247e-01 3.08033139e-01 -1.60256159e-02
-2.91884780e-01 -1.87217191e-01 -3.77175808e-01 -1.76883563e-01
2.83721358e-01 2.90184706e-01 1.23675697e-01 -2.51519680e-01
1.02503620e-01 5.51190197e-01 7.96558678e-01 -2.64224559e-01
1.06057417e+00 5.89038074e-01 -9.12517756e-02 -7.11108208e-01
-6.34602487e-01 -6.55467331e-01 -8.76230001e-01 -7.15099871e-01
8.04342747e-01 -1.33837903e+00 -1.19070518e+00 9.93168473e-01
-1.04441631e+00 1.29812688e-01 3.92203540e-01 6.56545162e-01
-1.66964069e-01 6.86177909e-01 -4.13439304e-01 -1.16485846e+00
-4.50161189e-01 -7.54520595e-01 9.10148919e-01 8.15218985e-01
7.53826350e-02 -7.59217322e-01 -1.16179883e-01 2.46534526e-01
2.40916923e-01 3.77424151e-01 3.45332809e-02 -4.72967446e-01
-3.94184053e-01 -5.41681111e-01 -1.74181625e-01 -1.14942163e-01
3.26741248e-01 -2.31542379e-01 -6.23809874e-01 -7.22209930e-01
-1.72400609e-01 1.36638269e-01 6.09201193e-01 3.36958021e-01
4.57471758e-01 -9.94468704e-02 -6.33767307e-01 3.95131767e-01
1.38160563e+00 4.70433325e-01 1.23593777e-01 8.37543905e-01
7.67434478e-01 4.04926538e-01 9.45494652e-01 7.02751100e-01
6.92288399e-01 9.51750875e-01 1.79348841e-01 -9.31628346e-02
3.56892415e-04 1.20089233e-01 6.21474266e-01 2.62784272e-01
-2.66042411e-01 1.85147952e-02 -5.37796855e-01 3.75164926e-01
-2.08974886e+00 -1.10822487e+00 2.19971128e-02 2.40187192e+00
3.27874988e-01 2.95431823e-01 5.49241722e-01 -5.68140857e-02
1.41437066e+00 -2.80778557e-01 -6.26755595e-01 2.59137005e-01
-1.91148426e-02 -3.82338196e-01 7.78850138e-01 4.80092406e-01
-1.52932382e+00 6.50797009e-01 5.97884512e+00 3.16945523e-01
-9.88211751e-01 9.75386240e-03 1.90247402e-01 2.50053972e-01
6.42621994e-01 -3.35898221e-01 -1.30635262e+00 6.51044607e-01
5.19225240e-01 3.47567238e-02 3.12549740e-01 8.68890285e-01
2.88445860e-01 -2.23817423e-01 -6.48500860e-01 1.16646838e+00
3.23476791e-01 -5.20720840e-01 -2.32977137e-01 -2.76154131e-01
3.48405033e-01 -4.15140927e-01 9.46534351e-02 5.00050306e-01
3.68515968e-01 -4.30329740e-01 9.31895614e-01 7.24408448e-01
1.25407681e-01 -7.52000272e-01 9.95739281e-01 2.83064634e-01
-1.64727461e+00 -4.01138216e-01 -2.90225834e-01 -2.62266397e-01
1.36826426e-01 6.34836182e-02 -8.84167194e-01 6.70342147e-01
1.00125444e+00 5.14630079e-01 -8.77187252e-01 1.40521085e+00
1.46132410e-01 -2.35441723e-03 -3.57849211e-01 -1.41751841e-01
-5.27878143e-02 -1.21154785e-01 5.92223704e-01 1.29355443e+00
5.03268719e-01 -1.45376757e-01 8.28163087e-01 2.47193769e-01
5.54913044e-01 1.06594600e-02 -3.63005459e-01 8.22106838e-01
4.61584270e-01 1.40901566e+00 -7.32465088e-01 -3.40308756e-01
-5.20469546e-01 1.07467782e+00 2.00093850e-01 3.47142458e-01
-1.20672727e+00 -2.00640336e-01 3.23026538e-01 -5.75340092e-02
4.39974606e-01 -2.39931270e-01 2.59780232e-02 -8.44212532e-01
2.02140808e-01 -5.70974827e-01 5.84786117e-01 -4.02410299e-01
-1.25332010e+00 6.09095931e-01 7.75055448e-03 -1.42032552e+00
-1.19844504e-01 -5.79944432e-01 -4.28056926e-01 8.49687636e-01
-1.25359881e+00 -1.41584301e+00 -9.39862609e-01 8.38942289e-01
3.68929237e-01 -5.21834493e-01 5.42061269e-01 2.01241761e-01
-8.75992417e-01 6.48949981e-01 2.60249898e-03 2.91181952e-01
7.71576285e-01 -1.13153338e+00 1.03328153e-01 1.12108946e+00
-4.05554384e-01 6.51097476e-01 8.76712978e-01 -9.23479676e-01
-1.35595787e+00 -1.16265416e+00 2.46631876e-01 -2.17604175e-01
1.21523969e-01 -1.86606973e-01 -4.34188277e-01 5.85740626e-01
7.09800795e-02 -1.99232474e-02 4.13733274e-01 -2.21932605e-02
-4.43766415e-02 -5.09828746e-01 -1.26081872e+00 4.24073547e-01
8.05014849e-01 1.25969753e-01 -4.53570306e-01 2.46109292e-02
2.59612530e-01 -4.61972892e-01 -5.79967856e-01 4.36270326e-01
9.49531317e-01 -8.72423410e-01 1.05413711e+00 1.46959439e-01
-7.54190564e-01 -7.78626442e-01 -1.03173777e-02 -1.08745193e+00
-7.68282294e-01 -4.07895893e-01 -1.96245998e-01 1.17210007e+00
-2.36778691e-01 -6.26074374e-01 7.67879188e-01 4.67353731e-01
2.10453525e-01 -2.37559006e-01 -8.84823740e-01 -9.72891748e-01
-5.50212681e-01 1.60331592e-01 4.91279393e-01 2.88332254e-01
-3.76098752e-01 2.07440346e-01 -6.40168965e-01 8.48909616e-01
1.00879812e+00 1.19589530e-01 1.23351824e+00 -1.49154270e+00
-1.85850590e-01 -9.05452892e-02 -7.20368683e-01 -9.22715783e-01
5.33522256e-02 -3.57178003e-01 3.04124177e-01 -1.42093849e+00
4.63629603e-01 -2.37241238e-01 -4.32961494e-01 2.38282636e-01
-5.82157612e-01 2.15618610e-01 3.73475969e-01 1.28988951e-01
-7.50222325e-01 3.61500263e-01 7.58119047e-01 -4.37446028e-01
-2.03926533e-01 5.22231221e-01 -5.44846356e-01 7.48309493e-01
6.53779268e-01 -1.96994051e-01 -3.73755917e-02 -2.52677113e-01
-6.54994190e-01 -3.22893500e-01 5.99307477e-01 -1.67860019e+00
3.90092582e-01 -8.09204485e-03 1.08784091e+00 -6.48900986e-01
5.95679700e-01 -1.22832704e+00 2.95852423e-01 9.40106392e-01
1.72573224e-01 1.47073731e-01 1.95194915e-01 8.36414933e-01
2.11343154e-01 -5.13791703e-02 9.18020964e-01 -8.85736197e-02
-1.25560856e+00 2.55424023e-01 -1.43610090e-01 -6.50434852e-01
1.48783576e+00 -4.75469500e-01 -1.35325164e-01 -4.63605344e-01
-6.25813186e-01 2.33101621e-01 4.75516200e-01 6.61001801e-01
6.88984513e-01 -1.37406659e+00 -6.49455428e-01 1.49354026e-01
1.28572598e-01 -4.99524742e-01 3.52714181e-01 5.47429681e-01
-3.07967693e-01 2.15368137e-01 -5.70438504e-01 -7.29515016e-01
-1.73060656e+00 6.62714839e-01 5.17339051e-01 -1.18440188e-01
-6.88154638e-01 4.09044117e-01 1.22024268e-01 -3.50637227e-01
5.88752627e-01 1.01381615e-01 -4.13811892e-01 -2.45164767e-01
7.18833983e-01 4.50243860e-01 -3.21872592e-01 -1.12341881e+00
-5.52706242e-01 9.34826255e-01 -4.99332622e-02 -5.23349866e-02
8.21671546e-01 -6.59180105e-01 2.68518448e-01 4.23808903e-01
6.22694910e-01 -4.87288572e-02 -1.68539131e+00 -4.06051934e-01
-4.59057316e-02 -5.39465606e-01 -4.65781093e-01 -6.51437819e-01
-9.14422929e-01 1.37693524e-01 1.40371811e+00 -6.79189637e-02
9.83204305e-01 -5.32406628e-01 8.25484991e-01 4.45658505e-01
7.00840235e-01 -1.04995334e+00 5.97505532e-02 3.55857819e-01
4.95899171e-01 -1.42866838e+00 1.43854171e-01 -4.06089574e-01
-5.80073893e-01 1.03078175e+00 1.03570187e+00 -1.34206608e-01
4.59370583e-01 -1.44001886e-01 2.63289183e-01 2.10260510e-01
-5.47411069e-02 -6.06257975e-01 2.17040330e-01 1.05842340e+00
-1.19097307e-01 2.08926156e-01 -1.25771642e-01 6.53040707e-01
-1.28267527e-01 -1.76197499e-01 2.34736487e-01 9.85624731e-01
-7.10876703e-01 -7.64088929e-01 -1.01136541e+00 2.54561454e-02
-3.66851121e-01 6.43489182e-01 -2.18823671e-01 8.56177807e-01
6.00615799e-01 1.22619545e+00 9.99979228e-02 -4.17235732e-01
5.37095487e-01 -1.66094899e-01 5.08493960e-01 -1.39451683e-01
-5.04348278e-01 9.04592406e-03 -1.61520973e-01 -2.72737592e-01
-7.32980847e-01 -8.35588872e-01 -1.24934149e+00 -2.30331898e-01
-1.36802852e-01 6.58152485e-03 5.30806661e-01 9.78202403e-01
8.81684870e-02 2.80703843e-01 7.66356647e-01 -1.12419701e+00
-3.02903861e-01 -9.09003794e-01 -4.83017743e-01 6.04219675e-01
3.41822088e-01 -1.10851502e+00 7.21408725e-02 1.26875460e-01] | [6.82118034362793, -1.614709734916687] |
f8ecfee0-9470-47c7-8b9b-f2c7954fc60a | artificial-life-in-game-mods-for-intuitive | 2007.03787 | null | https://arxiv.org/abs/2007.03787v1 | https://arxiv.org/pdf/2007.03787v1.pdf | Artificial Life in Game Mods for Intuitive Evolution Education | The understanding and acceptance of evolution by natural selection has become a difficult issue in many parts of the world, particularly the United States of America. The use of games to improve intuition about evolution via natural selection is promising but can be challenging. We propose the use of modifications to commercial games using artificial life techniques to 'stealth teach' about evolution via natural selection, provide a proof-of-concept mod of the game Stardew Valley, and report on its initial reception. | ['Barbara Z. Johnson', 'Kevin Connors', 'Anya E. Vostinar'] | 2020-07-07 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [-2.37469316e-01 1.85323343e-01 3.61233205e-01 8.20659548e-02
3.05897892e-01 -7.60864437e-01 6.58142984e-01 -7.70541560e-03
-6.98427200e-01 1.04555249e+00 -9.72424224e-02 -8.10496688e-01
-3.33565660e-02 -9.09490049e-01 -4.79467303e-01 -3.76743466e-01
-3.25079471e-01 4.92401719e-01 5.54822385e-01 -1.24802268e+00
7.33512819e-01 1.70902118e-01 -1.76784098e+00 -3.28865349e-01
9.19173002e-01 -4.18631509e-02 -5.16125187e-02 1.07705951e+00
1.52365193e-01 5.93276203e-01 -1.03416538e+00 -6.54004216e-01
3.05834293e-01 -9.82732952e-01 -7.63390064e-01 -3.76723856e-01
-3.59682918e-01 3.80582325e-02 8.76499861e-02 9.65553939e-01
9.43682969e-01 1.28988206e-01 1.40968174e-01 -1.07345462e+00
-5.59899449e-01 5.48507929e-01 -4.09424305e-01 4.02497113e-01
5.38940370e-01 6.81554794e-01 9.42004025e-01 -9.55683738e-02
7.85719752e-01 1.22700918e+00 9.30918574e-01 7.24056363e-01
-1.05715787e+00 -6.96609020e-01 -1.68463379e-01 -2.24231184e-01
-1.52775395e+00 -1.63822547e-01 1.56294018e-01 -2.47774929e-01
1.04557860e+00 5.56068242e-01 1.66656113e+00 5.49525023e-01
6.99777722e-01 5.36530137e-01 1.11616719e+00 -6.27094924e-01
7.26389170e-01 1.25006348e-01 -6.12324417e-01 7.85533965e-01
5.20979404e-01 5.59803605e-01 -4.78999108e-01 -5.17153323e-01
9.72352982e-01 -7.22534478e-01 2.74168700e-02 -2.24740446e-01
-3.78235847e-01 1.27036178e+00 1.13482364e-01 3.17751735e-01
-2.92644233e-01 3.72115314e-01 1.92281827e-01 3.74977231e-01
4.13436919e-01 1.13153625e+00 -6.71587467e-01 -8.87436092e-01
-5.01198232e-01 7.23640263e-01 9.98168945e-01 3.55602533e-01
2.87950307e-01 9.95713100e-02 8.27746391e-01 6.39183998e-01
5.76270282e-01 -8.13100189e-02 5.53687513e-01 -1.07551515e+00
-5.79427719e-01 4.66125339e-01 4.87821698e-02 -7.79658079e-01
-2.85593927e-01 -4.14385408e-01 -7.60362744e-02 9.09779906e-01
1.83208942e-01 -6.85461521e-01 -7.47763097e-01 1.75330937e+00
4.51688170e-01 -5.49310148e-02 -2.49708183e-02 6.06763601e-01
4.53187853e-01 4.94897902e-01 1.50860131e-01 -1.73956364e-01
8.15483570e-01 -5.67762733e-01 -2.65505135e-01 -1.51009023e-01
5.23095906e-01 -4.84002411e-01 9.18144524e-01 6.48233414e-01
-1.18010211e+00 -2.80253470e-01 -1.21330273e+00 8.40121567e-01
-3.18535388e-01 -9.50354099e-01 1.30303085e+00 1.79021227e+00
-1.44843066e+00 6.11517310e-01 -9.15598035e-01 -9.21943426e-01
2.00025618e-01 6.14530444e-01 9.25919448e-04 4.08720493e-01
-1.36160660e+00 1.12669325e+00 6.46098793e-01 -2.93093115e-01
-5.70985258e-01 -6.39047384e-01 -5.62967777e-01 -2.94444382e-01
3.83951247e-01 -8.34079206e-01 8.81983638e-01 -1.10076869e+00
-1.93432403e+00 9.65473592e-01 5.46587944e-01 -3.41224432e-01
4.85364467e-01 8.06169882e-02 -2.79133558e-01 -4.29222345e-01
-1.67098284e-01 4.97853458e-01 2.75739402e-01 -1.02872133e+00
-7.56570220e-01 -1.07085325e-01 3.73246998e-01 3.73463660e-01
-3.64989415e-02 4.04012889e-01 5.42930216e-02 -5.33779860e-01
-5.16052186e-01 -1.06500936e+00 -9.12131369e-01 -1.15534820e-01
2.00728565e-01 -1.84614405e-01 1.89919576e-01 -3.50271642e-01
1.30634940e+00 -1.76755929e+00 1.58222154e-01 8.76887585e-04
2.61758387e-01 3.03419948e-01 1.88467912e-02 6.55534387e-01
9.54003818e-03 7.20015049e-01 3.12857539e-03 5.77469647e-01
-2.14281559e-01 3.08983058e-01 4.72415477e-01 3.64474148e-01
5.30824251e-02 6.71988130e-01 -1.11947477e+00 -2.36277908e-01
9.27629545e-02 6.19620197e-02 -9.36053097e-01 8.65119100e-02
-7.76293874e-02 1.58147782e-01 -2.26296097e-01 3.88126016e-01
2.38385931e-01 1.31988779e-01 1.18317164e-01 9.14289653e-01
-5.67873478e-01 1.45152181e-01 -9.91104364e-01 1.81126606e+00
5.39573692e-02 6.00555658e-01 -1.54319122e-01 -4.69028085e-01
6.30122721e-01 3.12571676e-04 2.85405695e-01 -4.44721192e-01
3.45041811e-01 1.88992724e-01 7.79838324e-01 -4.01842147e-01
4.50976700e-01 -7.99141407e-01 -2.17216849e-01 8.42356801e-01
-8.96569639e-02 -1.08078694e+00 1.51406214e-01 1.83376700e-01
1.25788391e+00 4.38524783e-01 6.54641867e-01 -7.52809048e-01
1.17488645e-01 4.33880687e-01 4.89328057e-01 9.25377131e-01
-4.49557126e-01 1.87472045e-01 1.52944446e-01 -5.38020670e-01
-1.02469003e+00 -9.23081219e-01 2.68082201e-01 1.10615623e+00
2.70597845e-01 -7.37020433e-01 -7.94928193e-01 -2.89005399e-01
-1.93585694e-01 1.29175997e+00 -8.74331295e-01 -6.67378247e-01
-1.73591420e-01 -1.12091911e+00 6.16776228e-01 -1.51218712e-01
3.57458800e-01 -9.94409561e-01 -1.26444614e+00 3.17314476e-01
3.36917281e-01 -1.16770379e-01 -6.24312125e-02 1.83555484e-01
-5.99184513e-01 -5.67323804e-01 -5.33569574e-01 -2.63613701e-01
7.61442855e-02 1.65425278e-02 1.32308674e+00 7.62961030e-01
-7.52734065e-01 3.89243573e-01 -5.52136660e-01 -9.18065667e-01
-9.53370094e-01 1.43273445e-02 2.24037647e-01 -9.40366149e-01
2.20990449e-01 -7.39533186e-01 -4.68671083e-01 3.32327962e-01
-8.54985893e-01 -1.39842942e-01 2.20210493e-01 8.35254133e-01
-3.53800207e-01 7.39826620e-01 2.87505507e-01 -7.89087415e-01
1.13668287e+00 -3.30253452e-01 -4.67901558e-01 4.86163609e-02
-6.21456265e-01 -5.25712632e-02 1.32565841e-01 -4.17848140e-01
-7.49338925e-01 -2.97980905e-01 -2.59378970e-01 5.30138433e-01
1.21762224e-01 6.15228117e-01 1.75951362e-01 -6.16224289e-01
1.05602622e+00 -9.09718350e-02 2.50444710e-01 8.70635808e-02
2.95091897e-01 4.24786210e-01 5.79467379e-02 -6.13546610e-01
7.78132439e-01 9.38229039e-02 -3.81409556e-01 -1.21718609e+00
1.12711988e-01 2.49050140e-01 -3.12037081e-01 -6.19695187e-01
5.41594028e-01 -5.63962162e-01 -7.64299989e-01 5.25442004e-01
-6.92500174e-01 -6.37329996e-01 -6.41542971e-01 2.03108460e-01
-3.33757997e-01 2.09003352e-02 -3.25967580e-01 -9.97169912e-01
2.79170461e-02 -6.12832546e-01 3.40102524e-01 1.02712202e+00
-8.48000109e-01 -1.20714235e+00 8.16010296e-01 1.46913556e-02
7.46493638e-01 1.41465962e-01 8.29559565e-01 -4.70146507e-01
-1.56115532e-01 -2.57551849e-01 4.40205693e-01 -1.18893065e-01
3.14462423e-01 6.73750520e-01 -3.07623029e-01 -2.44602993e-01
8.17592964e-02 -4.01045620e-01 2.47448593e-01 5.36748052e-01
1.86862826e-01 2.42241785e-01 -2.12007150e-01 8.65311995e-02
1.47396946e+00 7.10375071e-01 1.00901949e+00 9.11734223e-01
-1.59789875e-01 7.32742369e-01 6.47379339e-01 8.78234208e-01
2.94529080e-01 3.09894025e-01 3.03803563e-01 5.21406010e-02
4.10990447e-01 -7.66331553e-02 4.11321104e-01 8.08173597e-01
-4.70724523e-01 -3.65048230e-01 -1.06314337e+00 3.91193748e-01
-1.41657567e+00 -9.00788665e-01 2.06595641e-02 2.15450430e+00
8.18249345e-01 6.32971525e-01 6.69541836e-01 -2.20143020e-01
5.97328067e-01 -9.81245562e-02 -3.63536477e-01 -9.88344014e-01
-2.28670016e-01 6.22882009e-01 2.95734257e-01 6.73655629e-01
-3.78935248e-01 1.27161956e+00 8.74509239e+00 8.63798022e-01
-9.29962695e-01 -1.62799910e-01 7.10916400e-01 7.60505199e-02
-6.50706947e-01 4.70021874e-01 -2.75407463e-01 3.51285785e-01
9.32414591e-01 -8.63078654e-01 4.67105657e-01 3.27906340e-01
3.55162978e-01 -6.32543087e-01 -3.33273262e-01 2.10366204e-01
-3.68749276e-02 -1.15290308e+00 -1.69339657e-01 8.11384767e-02
5.47224700e-01 -2.75639653e-01 1.05707545e-03 3.01191956e-01
1.26651013e+00 -1.29672253e+00 6.49625421e-01 4.42410856e-02
1.79199353e-01 -1.13263798e+00 4.91412610e-01 2.87730753e-01
-5.01950800e-01 1.59302019e-02 -4.16012436e-01 -1.00283134e+00
1.21492110e-01 -1.75662823e-02 -7.28832841e-01 8.43085423e-02
8.71659577e-01 -8.28928575e-02 -7.22294688e-01 1.37481725e+00
-2.75223434e-01 7.84955025e-01 -5.56045651e-01 -9.33135986e-01
1.08266734e-01 -2.42236927e-01 5.84377229e-01 6.58023000e-01
2.96607703e-01 5.56213617e-01 -2.74791062e-01 8.11361253e-01
7.25554883e-01 2.38762841e-01 -5.88379800e-01 -5.03369689e-01
1.89104691e-01 4.98417407e-01 -9.91500854e-01 2.30147131e-02
6.70274496e-02 1.04028857e+00 -1.92770064e-01 -1.34031594e-01
-1.00140595e+00 -4.70332652e-01 9.23930466e-01 -2.19650373e-01
1.05254292e-01 -1.63836241e-01 -4.53796625e-01 -8.43704283e-01
-8.34478974e-01 -1.36028028e+00 2.29510829e-01 -7.67844737e-01
-9.15751338e-01 4.18147385e-01 -1.39004484e-01 -6.70060873e-01
-2.99780101e-01 -3.33815038e-01 -8.65300834e-01 7.10767925e-01
-2.16748938e-01 -6.93366289e-01 2.57983893e-01 -1.19327269e-04
5.66992223e-01 -4.18805599e-01 7.91896939e-01 -3.90011698e-01
-4.19321597e-01 4.96638238e-01 -4.37926948e-02 -7.01937854e-01
3.80529672e-01 -1.10693467e+00 1.02278745e+00 5.89991570e-01
-9.65654403e-02 8.62985969e-01 1.57363701e+00 -1.01858020e+00
-1.34728432e+00 6.29307777e-02 2.28972524e-01 -4.43403423e-01
7.53991961e-01 -3.19187820e-01 -3.43603194e-01 2.61574924e-01
6.08698487e-01 -9.15432513e-01 1.01514196e+00 1.86239406e-01
3.00164759e-01 6.29957676e-01 -1.23536217e+00 1.52386725e+00
1.00398314e+00 -3.57833281e-02 -4.09908146e-01 -4.23764922e-02
8.91494274e-01 -1.27850309e-01 -5.64589918e-01 9.96743441e-02
9.16907191e-01 -1.09646940e+00 9.11995649e-01 -8.14389586e-01
3.65601301e-01 -1.81530476e-01 1.53529167e-01 -1.72873771e+00
-5.62752485e-01 -1.44233203e+00 7.42028534e-01 1.05360425e+00
7.10658371e-01 -8.52196038e-01 1.16577876e+00 9.71686423e-01
2.83469558e-01 -5.76252997e-01 -7.23603725e-01 -5.40302396e-01
6.17335320e-01 -3.06879193e-01 3.31074357e-01 1.06221449e+00
4.14186865e-01 2.88650393e-01 -4.18621272e-01 -4.67618182e-02
3.09920073e-01 -5.16322792e-01 1.10031629e+00 -1.24196589e+00
-9.81179595e-01 -8.27138066e-01 -9.66829419e-01 -4.02434856e-01
-4.81753170e-01 -1.24774911e-01 1.20573588e-01 -1.13602042e+00
1.32117599e-01 -8.91404897e-02 8.62149671e-02 -7.36853182e-02
-4.67010647e-01 1.87623963e-01 1.05158933e-01 -3.08807731e-01
-4.05214459e-01 5.06330490e-01 1.08087468e+00 2.49190852e-01
-4.33127612e-01 -6.83442131e-02 -1.31775308e+00 6.92566693e-01
8.08579743e-01 -5.98164320e-01 -4.13004220e-01 1.27020702e-01
7.42123187e-01 -1.21190265e-01 -3.54951680e-01 -8.87876034e-01
5.55226877e-02 -7.24295855e-01 8.27484131e-02 2.20665053e-01
1.11453921e-01 -2.65679926e-01 7.01760232e-01 1.20343113e+00
-1.21039502e-01 4.37494487e-01 8.43562484e-01 2.47336239e-01
2.58311987e-01 -4.44865823e-01 7.47853220e-01 -5.07389724e-01
-8.15199256e-01 -3.64307523e-01 -1.40382206e+00 1.33130342e-01
1.31843197e+00 -4.99140799e-01 -8.59537497e-02 -8.38686943e-01
-3.76326740e-01 3.75815243e-01 1.15745056e+00 2.03096256e-01
4.90664691e-01 -7.86372066e-01 -8.35085213e-01 -2.30863076e-02
-9.03059393e-02 -6.97160602e-01 7.80308247e-02 3.74118924e-01
-1.40459859e+00 -2.59269923e-01 -8.27902079e-01 -2.69683339e-02
-1.52903223e+00 3.38143289e-01 5.56478381e-01 7.34091923e-02
-2.50483841e-01 1.51154399e+00 -4.60693762e-02 -5.09491444e-01
-3.74159008e-01 5.83360255e-01 -6.27506673e-02 -4.16182607e-01
4.53241944e-01 6.38414174e-02 -6.30591810e-01 -2.21429408e-01
-4.93476033e-01 2.29817808e-01 -3.00463419e-02 -8.23479176e-01
1.41811478e+00 -1.88860670e-01 -4.31899548e-01 5.82691908e-01
2.61222333e-01 1.06072702e-01 -6.80098176e-01 6.20785296e-01
-1.40364304e-01 -7.20973134e-01 -2.95695007e-01 -1.04777420e+00
-6.01015091e-01 5.55020452e-01 6.97208345e-01 3.91095668e-01
9.64298546e-01 -3.26722383e-01 1.95527777e-01 1.08990021e-01
7.99686313e-01 -1.03591883e+00 -1.16057634e-01 4.17368650e-01
5.65542221e-01 -7.50943303e-01 4.39884633e-01 -2.62964785e-01
-5.97882986e-01 9.25174296e-01 6.39840364e-01 -3.69104236e-01
7.79815674e-01 2.49082983e-01 1.54542685e-01 -2.39646181e-01
-1.00682950e+00 -1.08665921e-01 -5.23932219e-01 8.78879964e-01
7.24765420e-01 -4.12715133e-03 -1.07518601e+00 2.46768236e-01
-5.74041724e-01 9.73089486e-02 1.03689301e+00 1.47248316e+00
-6.47038639e-01 -1.38428903e+00 -3.71891171e-01 2.11302489e-01
-4.59937572e-01 -1.81442320e-01 -1.14374030e+00 1.01645577e+00
2.98930198e-01 8.45274925e-01 -3.28449607e-01 -8.67861390e-01
2.39352025e-02 -2.06869498e-01 7.02055216e-01 -6.22791350e-01
-1.14028919e+00 -8.99969891e-04 3.98366481e-01 -9.23875645e-02
-1.82203069e-01 -8.60692024e-01 -9.14746523e-01 -9.14278150e-01
-7.41431653e-01 5.97340465e-01 4.83716041e-01 8.73962402e-01
7.98574239e-02 3.14407021e-01 2.86590427e-01 -7.58696079e-01
-1.98574826e-01 -5.16595006e-01 -8.35901022e-01 3.24329212e-02
-5.59142649e-01 -6.03951693e-01 -3.57804060e-01 -3.36755961e-01] | [5.491997718811035, 4.015431880950928] |
0ae75577-32ce-4e1a-9386-8644af94b1cf | peek-into-the-future-camera-based-occupant | 2212.11950 | null | https://arxiv.org/abs/2212.11950v1 | https://arxiv.org/pdf/2212.11950v1.pdf | Peek into the Future Camera-based Occupant Sensing in Configurable Cabins for Autonomous Vehicles | The development of fully autonomous vehicles (AVs) can potentially eliminate drivers and introduce unprecedented seating design. However, highly flexible seat configurations may lead to occupants' unconventional poses and actions. Understanding occupant behaviors and prioritize safety features become eye-catching topics in the AV research frontier. Visual sensors have the advantages of cost-efficiency and high-fidelity imaging and become more widely applied for in-car sensing purposes. Occlusion is one big concern for this type of system in crowded car cabins. It is important but largely unknown about how a visual-sensing framework will look like to support 2-D and 3-D human pose tracking towards highly configurable seats. As one of the first studies to touch this topic, we peek into the future camera-based sensing framework via a simulation experiment. Constructed representative car-cabin, seat layouts, and occupant sizes, camera coverage from different angles and positions is simulated and calculated. The comprehensive coverage data are synthesized through an optimization process to determine the camera layout and overall occupant coverage. The results show the needs and design of a different number of cameras to fully or partially cover all the occupants with changeable configurations of up to six seats. | ['Saeed Barbat', 'Srinivasan Sundararajan', 'Jialiang Le', 'Lingxi Li', 'Renran Tian', 'Avinash Prabu'] | 2022-12-22 | null | null | null | null | ['pose-tracking'] | ['computer-vision'] | [-1.00994088e-01 3.56552415e-02 -7.83872530e-02 -3.89218479e-01
-3.17042410e-01 -6.54745877e-01 1.44692779e-01 -4.33674514e-01
-2.42675871e-01 5.36907434e-01 1.07011020e-01 -4.61939067e-01
2.55080331e-02 -4.81551081e-01 -4.08748746e-01 -5.44421017e-01
4.72647190e-01 2.59932905e-01 5.98018527e-01 -4.13698912e-01
-7.81276003e-02 9.17255223e-01 -2.24794722e+00 7.91518390e-03
5.05348504e-01 9.01243389e-01 5.28086662e-01 4.79596436e-01
3.83714855e-01 3.07566524e-01 -3.81801516e-01 -1.44543543e-01
4.16926324e-01 2.27936029e-01 9.67737064e-02 3.45627785e-01
3.22567105e-01 -3.76747310e-01 -2.10006833e-01 4.46249455e-01
4.68729913e-01 2.32176453e-01 3.44581932e-01 -1.40231740e+00
-2.88684815e-01 -8.35186690e-02 -4.34942275e-01 1.10520571e-01
4.58029598e-01 6.02674365e-01 4.48631763e-01 -7.38096058e-01
5.89796752e-02 1.07131875e+00 5.86854637e-01 5.91934741e-01
-7.82904088e-01 -6.00208521e-01 2.80558378e-01 2.16926500e-01
-1.54460835e+00 -6.89090014e-01 9.33695436e-01 -4.65000987e-01
6.87477410e-01 8.82651508e-01 9.36877549e-01 1.04401362e+00
2.24978626e-01 4.08337474e-01 1.19463551e+00 -1.65033475e-01
2.94320911e-01 8.66825163e-01 1.16685465e-01 6.34722888e-01
8.58263612e-01 2.65211999e-01 -3.04704577e-01 2.02142209e-01
2.86438406e-01 4.77026463e-01 -4.64918055e-02 -5.51922977e-01
-9.12374973e-01 6.66241348e-01 1.70582727e-01 3.29607353e-02
-1.08680047e-01 7.59897307e-02 2.31956299e-02 -2.99200445e-01
-1.27073795e-01 3.82531822e-01 -2.06414640e-01 -2.16322392e-01
-6.33894801e-01 3.35187733e-01 4.24200773e-01 1.26683068e+00
5.87364137e-01 9.71192122e-02 -1.16754197e-01 6.21131599e-01
4.76231694e-01 9.39677954e-01 -1.31424069e-01 -1.22791696e+00
2.47885853e-01 7.90582538e-01 3.75382781e-01 -8.57567489e-01
-5.72123647e-01 -2.09823892e-01 -2.56560206e-01 3.82578254e-01
6.15604632e-02 -4.20360506e-01 -8.26676428e-01 1.18884671e+00
5.72266579e-01 -6.39832437e-01 -4.81238037e-01 1.20022285e+00
7.20031202e-01 5.18222630e-01 -1.43510029e-01 -1.74140230e-01
1.72880888e+00 -7.20359981e-01 -7.80263245e-01 -6.67932630e-01
1.53414190e-01 -5.10400593e-01 1.22600245e+00 1.54418886e-01
-9.12305892e-01 -7.14048803e-01 -1.31661165e+00 1.18396528e-01
-4.66274083e-01 2.33358130e-01 4.29508418e-01 1.33894086e+00
-8.47611845e-01 -4.08726484e-01 -8.60908508e-01 -3.96438301e-01
9.38932598e-03 2.31527448e-01 -9.16528851e-02 -5.73520996e-02
-1.02934504e+00 1.06399131e+00 -4.28488821e-01 3.04124504e-01
-9.66679156e-01 -6.03013039e-01 -7.95865238e-01 -1.57205567e-01
6.37296200e-01 -7.04145968e-01 1.10919917e+00 -3.16460729e-01
-1.48141897e+00 5.93968391e-01 -2.32849687e-01 -3.20966020e-02
4.18915719e-01 -1.52814001e-01 -7.53754497e-01 -1.05837084e-01
2.18251683e-02 6.09657109e-01 6.57807767e-01 -1.63695920e+00
-5.55188119e-01 -4.99607325e-01 1.53205171e-01 2.14829132e-01
-1.98090583e-01 -3.63076702e-02 -2.10042790e-01 1.45141125e-01
-1.67586356e-01 -1.31548774e+00 -2.61015803e-01 2.55924370e-02
-3.99352372e-01 1.41847491e-01 1.17361665e+00 -2.59291112e-01
1.16762435e+00 -2.01538515e+00 -3.74098450e-01 7.98882768e-02
2.36794129e-01 1.93103671e-01 5.93496561e-01 3.93470705e-01
3.72497529e-01 -3.26189756e-01 -6.40714094e-02 -4.59475219e-01
6.87503070e-02 1.96158826e-01 2.94244111e-01 6.02707148e-01
-2.16558680e-01 5.68200588e-01 -4.06843722e-01 -5.63805580e-01
8.33407640e-01 5.86250186e-01 -6.42062485e-01 2.25167140e-01
2.99135745e-01 3.92277598e-01 -4.37233448e-01 1.00276470e+00
7.36226261e-01 2.69509077e-01 -1.73476160e-01 -2.51153886e-01
-6.35155499e-01 -1.69550121e-01 -1.32139349e+00 7.92771995e-01
-4.56691504e-01 7.19154179e-01 6.12125874e-01 -4.95416433e-01
1.05779421e+00 1.68262161e-02 5.08307099e-01 -8.43345225e-01
3.10338378e-01 -7.07740150e-03 -3.56444955e-01 -9.28939939e-01
9.26750898e-01 -1.54682130e-01 -4.84551817e-01 -1.92975223e-01
-7.64792204e-01 -6.66497871e-02 -3.02032024e-01 -3.52335691e-01
8.06632698e-01 -2.96110660e-01 -6.72168753e-05 -4.66152072e-01
6.41301274e-01 2.28029147e-01 4.00336623e-01 3.96242887e-01
-6.42589986e-01 5.16490042e-01 -1.60096988e-01 -4.09035653e-01
-1.04369545e+00 -1.09493625e+00 -2.49660507e-01 1.04352450e+00
6.73795819e-01 6.68326244e-02 -1.02819705e+00 -5.88449724e-02
1.36825323e-01 9.71936643e-01 -3.14803600e-01 -3.06012005e-01
-5.89434922e-01 -2.55944788e-01 7.44555369e-02 6.08851850e-01
4.45811391e-01 -4.44045007e-01 -1.48761046e+00 -2.42015019e-01
4.05129008e-02 -1.20268476e+00 -5.60411155e-01 -1.30938128e-01
-1.69550315e-01 -8.55995238e-01 -5.57916403e-01 -4.96646494e-01
6.57282412e-01 8.94639492e-01 6.70307755e-01 -2.12362736e-01
-3.51249218e-01 7.95105219e-01 -9.72819850e-02 -8.35312366e-01
-2.48928815e-01 7.34191313e-02 3.04376870e-01 1.39369428e-01
3.99814874e-01 -1.37333348e-01 -1.24049270e+00 1.04834259e+00
-3.30896586e-01 2.09164515e-01 4.26840007e-01 1.43351033e-01
4.65810150e-01 -3.20055038e-02 3.39537412e-01 -5.81576049e-01
4.29625213e-01 -4.18407440e-01 -7.73585916e-01 9.72064510e-02
-5.37479639e-01 -3.89941990e-01 5.11237144e-01 -2.54229575e-01
-1.20919633e+00 2.87780374e-01 -4.72732149e-02 -3.48407954e-01
-4.65029716e-01 -3.42735112e-01 -5.49698591e-01 2.64341921e-01
5.38313925e-01 -1.76082611e-01 2.27695420e-01 -1.78346246e-01
2.85540849e-01 9.45491612e-01 4.30188715e-01 -2.56927222e-01
7.05750525e-01 8.17272902e-01 -2.41486039e-02 -1.22173584e+00
-3.56533706e-01 -6.52437270e-01 -4.26518410e-01 -9.15600955e-01
1.10320342e+00 -1.00962734e+00 -1.27939403e+00 9.03158411e-02
-6.70393527e-01 -9.78104249e-02 -2.10133746e-01 4.02758360e-01
-1.42780796e-01 -5.03295586e-02 2.19940096e-01 -1.30132926e+00
1.89267062e-02 -1.64780974e+00 1.39858711e+00 3.74058574e-01
-4.44065988e-01 -4.81184930e-01 -1.19952202e-01 1.19352031e+00
5.82534909e-01 5.44783100e-02 3.00669193e-01 1.46982849e-01
-8.01581144e-01 -4.61902320e-01 2.46867493e-01 5.72825707e-02
1.04478352e-01 -8.39278772e-02 -1.24432898e+00 -4.63770270e-01
-5.52427880e-02 3.35552275e-01 4.28203732e-01 4.78659779e-01
8.72229218e-01 -1.80569261e-01 -8.69488060e-01 3.02695930e-01
1.19532299e+00 4.79249358e-01 4.15081561e-01 1.16772905e-01
8.03148806e-01 8.64092529e-01 8.70380342e-01 4.70385730e-01
6.15607738e-01 9.74661648e-01 6.60812080e-01 -2.30168432e-01
-9.33979675e-02 -4.27454948e-01 5.52773297e-01 7.52174675e-01
-8.85618106e-02 -2.48966649e-01 -8.57621074e-01 5.16193509e-01
-1.35238576e+00 -9.28802371e-01 -2.26642013e-01 2.25765085e+00
1.70584291e-01 1.95414618e-01 5.22169292e-01 4.53526303e-02
7.27611184e-01 6.97843507e-02 -5.96091449e-01 -6.32759213e-01
4.59631473e-01 -5.00641048e-01 1.15762937e+00 5.94019532e-01
-6.86897397e-01 2.93684751e-01 6.51475286e+00 2.46848017e-01
-8.55045617e-01 2.48067483e-01 5.54026604e-01 -5.10036588e-01
-6.23774767e-01 1.91095620e-02 -1.43117404e+00 5.51845133e-01
9.95916486e-01 3.43881100e-01 2.99667150e-01 1.12057924e+00
5.55317521e-01 -4.52771127e-01 -8.66613388e-01 9.95757222e-01
1.84616998e-01 -1.09135175e+00 -5.31656981e-01 3.52107078e-01
5.42772532e-01 -1.32477790e-01 4.94853295e-02 2.69755006e-01
-1.26627624e-01 -7.50611246e-01 1.25801051e+00 5.00790298e-01
8.95565808e-01 -7.43285179e-01 4.85658199e-01 4.23558354e-01
-1.49785256e+00 -5.24131477e-01 -2.72033393e-01 -2.05609664e-01
5.34309447e-01 1.41828552e-01 -7.17687964e-01 -1.99125066e-01
9.57957983e-01 -2.17289150e-01 -5.97297847e-01 7.10588396e-01
4.73443896e-01 2.81030238e-01 -4.24110800e-01 -6.94231331e-01
-5.33423573e-02 -5.60337640e-02 6.58277571e-01 8.33349824e-01
3.23538065e-01 8.23408961e-02 1.09692752e-01 9.78883505e-01
5.20669341e-01 -5.53463638e-01 -1.04175878e+00 5.77082098e-01
8.87645185e-01 1.34961498e+00 -6.12714827e-01 1.51758194e-01
-6.66096091e-01 2.17088655e-01 -3.17384034e-01 3.60714167e-01
-1.30060971e+00 -1.98846981e-02 1.11218834e+00 1.19830203e+00
1.63035437e-01 -5.36029935e-01 -3.71547610e-01 -3.47007751e-01
8.09423253e-02 -3.55868280e-01 -1.39746860e-01 -7.12051988e-01
-6.98286891e-01 2.93710440e-01 5.43263555e-01 -1.15388012e+00
2.41598174e-01 -7.58707643e-01 -3.41066450e-01 4.46215391e-01
-1.03035927e+00 -1.11059546e+00 -6.77344382e-01 4.75609243e-01
6.85121834e-01 -9.39006731e-02 1.82203516e-01 5.93371987e-01
-7.98321187e-01 8.69145572e-01 -2.42335200e-02 -5.05732179e-01
2.38618597e-01 -6.04835749e-01 7.72259235e-02 7.03908801e-01
-5.68039656e-01 4.16085780e-01 1.07542455e+00 -4.95719105e-01
-2.10921669e+00 -9.21015441e-01 3.26038390e-01 -8.90520096e-01
-6.70958534e-02 -8.56412649e-01 -4.15912777e-01 5.21206796e-01
1.66540846e-01 -1.57377824e-01 2.85069883e-01 -2.42834181e-01
5.41449070e-01 -6.04368031e-01 -1.17841387e+00 9.80150521e-01
1.12137246e+00 -3.49469304e-01 -2.35021457e-01 -1.83605552e-02
7.97623694e-01 -4.17056113e-01 -3.79082114e-01 3.99159998e-01
7.74442255e-01 -1.09798610e+00 1.03680611e+00 2.27957398e-01
-3.59642923e-01 -5.36460936e-01 -3.72954875e-01 -7.73506463e-01
-4.07642901e-01 -6.43826902e-01 2.47494563e-01 1.15844607e+00
4.82860416e-01 -8.93814921e-01 8.93803120e-01 1.12475049e+00
-8.12111497e-01 -7.40008473e-01 -1.17452884e+00 -5.65369725e-01
-3.95654887e-01 -5.12860954e-01 7.54859924e-01 3.01251769e-01
-2.30029654e-02 3.22130680e-01 -2.94442445e-01 4.24357653e-01
3.24759483e-01 -1.57022059e-01 9.94682968e-01 -1.05194712e+00
8.23298022e-02 -2.48907804e-01 -2.98523903e-01 -8.90831292e-01
-3.26256603e-01 -1.47010103e-01 2.83322930e-01 -1.52489090e+00
6.67975396e-02 -3.60706389e-01 3.56911242e-01 2.19103880e-02
1.02489777e-01 -2.12499812e-01 -4.08733264e-03 -1.14055283e-01
-4.22143072e-01 3.81078899e-01 1.28545702e+00 -1.13196582e-01
-7.93167874e-02 3.41946542e-01 -6.78446174e-01 6.57299817e-01
7.35997617e-01 -1.74121391e-02 -8.49319637e-01 -1.64010912e-01
2.24168643e-01 8.90626013e-02 4.81783748e-01 -1.37924254e+00
4.10952181e-01 -5.28325140e-01 1.34150580e-01 -8.31934988e-01
8.11812937e-01 -1.49839616e+00 6.98286355e-01 4.09131140e-01
9.70311686e-02 3.67788434e-01 4.18028086e-01 5.24266660e-01
2.24112630e-01 8.43784809e-02 8.47011626e-01 -1.72976032e-01
-7.68136561e-01 5.05244061e-02 -8.42722178e-01 -3.72882634e-01
1.71075833e+00 -1.12352777e+00 -2.98597395e-01 -2.66385674e-01
-2.16147915e-01 2.82800525e-01 7.24794447e-01 7.02440977e-01
6.36606097e-01 -1.23886669e+00 -1.57910556e-01 6.78073406e-01
2.60975331e-01 1.77972633e-02 5.88314474e-01 6.43886447e-01
-5.39542377e-01 8.58123362e-01 -1.62755072e-01 -8.69121075e-01
-1.37598121e+00 7.54657328e-01 6.27763033e-01 6.15426898e-01
-5.50604641e-01 4.24138844e-01 2.69713372e-01 -2.38309070e-01
-6.00210056e-02 -4.24380124e-01 -2.93539502e-02 -2.38227144e-01
3.03987622e-01 9.36151743e-01 2.64369193e-02 -1.02832425e+00
-6.31284654e-01 6.63983405e-01 4.23725158e-01 5.45482598e-02
7.77213752e-01 -7.75913656e-01 6.65171504e-01 4.31844532e-01
8.17818344e-01 4.16757494e-01 -1.72693872e+00 4.33967412e-01
-4.45781857e-01 -5.28378606e-01 4.91092615e-02 -4.89185244e-01
-9.50624168e-01 6.57876015e-01 1.02597904e+00 2.08492741e-01
8.26115727e-01 4.10639822e-01 7.76220381e-01 -4.56743650e-02
6.80204690e-01 -1.37139189e+00 -1.00592360e-01 -1.67382941e-01
7.37540245e-01 -1.23804736e+00 -2.97022448e-03 -4.52972203e-01
-8.42644751e-01 5.31098306e-01 9.73375916e-01 4.13241863e-01
6.40629530e-01 4.67967719e-01 -6.53916225e-02 -6.48981094e-01
-5.73639274e-01 -1.69464350e-01 1.40972659e-01 7.45291352e-01
6.88681751e-02 6.17246091e-01 9.74679515e-02 5.78529894e-01
-3.29439223e-01 -4.75433201e-01 2.66410857e-01 9.79038954e-01
-8.70478630e-01 -5.41447401e-01 -8.96410286e-01 4.04582888e-01
8.99002552e-02 6.36482656e-01 -1.10330775e-01 8.97981167e-01
3.98640096e-01 1.31247759e+00 3.18358056e-02 -6.42645478e-01
9.57438707e-01 -6.65677488e-02 1.00094013e-01 -4.46671695e-01
-2.91164964e-01 -2.66151100e-01 3.49469066e-01 -4.24330831e-01
-8.50888118e-02 -8.28244269e-01 -9.56512392e-01 -6.89763427e-01
-3.98417413e-01 -2.80259401e-01 7.12424934e-01 6.77811682e-01
5.03406763e-01 6.54609144e-01 7.14751840e-01 -8.71914804e-01
-2.06095919e-01 -5.20656466e-01 -5.22494912e-01 2.00101398e-02
4.32224184e-01 -1.10399175e+00 -2.58925110e-01 -2.65049309e-01] | [7.859551429748535, -1.043502926826477] |
81e49ff7-4035-4aa0-86cd-e0ff8600488e | learning-a-universal-human-prior-for | 2304.04602 | null | https://arxiv.org/abs/2304.04602v1 | https://arxiv.org/pdf/2304.04602v1.pdf | Learning a Universal Human Prior for Dexterous Manipulation from Human Preference | Generating human-like behavior on robots is a great challenge especially in dexterous manipulation tasks with robotic hands. Even in simulation with no sample constraints, scripting controllers is intractable due to high degrees of freedom, and manual reward engineering can also be hard and lead to non-realistic motions. Leveraging the recent progress on Reinforcement Learning from Human Feedback (RLHF), we propose a framework to learn a universal human prior using direct human preference feedback over videos, for efficiently tuning the RL policy on 20 dual-hand robot manipulation tasks in simulation, without a single human demonstration. One task-agnostic reward model is trained through iteratively generating diverse polices and collecting human preference over the trajectories; it is then applied for regularizing the behavior of polices in the fine-tuning stage. Our method empirically demonstrates more human-like behaviors on robot hands in diverse tasks including even unseen tasks, indicating its generalization capability. | ['Chi Jin', 'Hao Dong', 'Shixiang Shane Gu', 'Allen Z. Ren', 'Yuanpei Chen', 'Zihan Ding'] | 2023-04-10 | null | null | null | null | ['robot-manipulation'] | ['robots'] | [ 1.72001421e-02 2.16884881e-01 -1.40008211e-01 -3.08302715e-02
-3.57491851e-01 -6.29469573e-01 3.53118479e-01 -7.14920402e-01
-5.51886737e-01 1.20960903e+00 1.09819151e-01 -2.00693265e-01
-2.54052788e-01 -9.35636386e-02 -7.41008699e-01 -6.25834107e-01
-4.78899986e-01 8.37552845e-01 2.11707175e-01 -4.58318442e-01
1.63602069e-01 6.02423012e-01 -1.39125025e+00 -2.08340004e-01
8.42765272e-01 5.99567175e-01 7.50494361e-01 9.10082817e-01
7.78284788e-01 1.08577085e+00 -5.12625337e-01 1.97893545e-01
5.89038014e-01 -3.38075042e-01 -7.38667667e-01 3.71006310e-01
-1.57481745e-01 -9.26761627e-01 -7.05641150e-01 8.21735799e-01
5.40895045e-01 5.78835249e-01 7.78710604e-01 -1.60223317e+00
-6.40897453e-01 6.18948162e-01 -3.77188265e-01 -4.25080091e-01
5.75929701e-01 1.16267073e+00 6.75439596e-01 -3.22830170e-01
7.78939188e-01 1.65064895e+00 1.92968130e-01 9.95024085e-01
-9.81307805e-01 -6.06242657e-01 7.04939067e-02 8.42692778e-02
-9.54406321e-01 2.25114107e-01 3.19892794e-01 -6.49505079e-01
8.27133119e-01 -2.58077919e-01 8.60114574e-01 1.79835176e+00
2.90608615e-01 9.83183980e-01 9.74632382e-01 -1.44710382e-02
2.84111440e-01 -2.48265117e-01 -4.87703770e-01 7.34535336e-01
1.36904776e-01 5.69140136e-01 -2.14040093e-02 -1.29331544e-01
1.44300044e+00 2.11421371e-01 -3.04032415e-01 -8.31153989e-01
-1.63513350e+00 6.31531119e-01 4.61858958e-01 -1.67126775e-01
-7.96498716e-01 5.41531086e-01 3.35669935e-01 5.46733677e-01
-6.87539160e-01 1.04707789e+00 -6.31056845e-01 -4.41108614e-01
-1.16718933e-01 9.34344888e-01 9.87485647e-01 1.56318891e+00
4.64539886e-01 8.65880474e-02 -6.16050839e-01 4.54068899e-01
-5.49191050e-02 6.88217103e-01 2.86574960e-01 -1.69715357e+00
4.63327557e-01 3.16339970e-01 9.60631669e-01 -4.98139739e-01
-4.71980542e-01 1.36298031e-01 -4.16888267e-01 7.92542100e-01
5.29761314e-01 -6.06344938e-01 -1.01448822e+00 1.71517694e+00
2.51380622e-01 -3.45936656e-01 5.09299561e-02 1.40408766e+00
-2.46266965e-02 5.07991850e-01 2.29692101e-01 9.83351842e-02
1.02346897e+00 -1.22794962e+00 -4.92963791e-01 -2.16730267e-01
3.37255597e-01 -2.56324112e-01 1.47817516e+00 7.02810526e-01
-8.40909123e-01 -4.53893661e-01 -8.69737923e-01 2.26862371e-01
2.06433833e-01 3.03031415e-01 6.17734075e-01 -8.24164823e-02
-5.82335830e-01 1.03142798e+00 -9.33832288e-01 -4.36154723e-01
2.15362832e-01 5.95163405e-01 -3.18620741e-01 -1.76794514e-01
-9.36698139e-01 1.24772155e+00 4.68920648e-01 -8.84486549e-03
-1.81302691e+00 -2.39492595e-01 -8.11970294e-01 -1.14140004e-01
9.64755774e-01 -7.64713943e-01 1.65595210e+00 -5.75415969e-01
-2.05459785e+00 3.40422958e-01 5.39245963e-01 -2.64483005e-01
9.19246614e-01 -5.39861381e-01 2.40562350e-01 1.17639862e-01
1.66296169e-01 1.12275445e+00 1.26025081e+00 -1.37372434e+00
-5.42069793e-01 1.45127028e-01 3.62525642e-01 3.88602763e-01
4.54971101e-03 -2.40047842e-01 -2.38714188e-01 -5.50033987e-01
-5.86150765e-01 -1.58837235e+00 -6.77314937e-01 7.37714618e-02
-2.47110248e-01 -3.57742816e-01 7.84243286e-01 -4.09262836e-01
5.32823741e-01 -1.96956527e+00 8.17390561e-01 7.61103183e-02
-9.90682617e-02 1.21010013e-01 -4.81216431e-01 3.96896720e-01
3.73844504e-01 -4.28334594e-01 -1.26047954e-01 2.79433995e-01
4.87118125e-01 5.12599885e-01 -3.47805113e-01 2.65052974e-01
3.81118268e-01 8.75200748e-01 -1.60563684e+00 -3.19790572e-01
1.52197018e-01 9.48674306e-02 -7.49748349e-01 7.30994999e-01
-6.07324958e-01 1.06125116e+00 -7.68490374e-01 5.71192622e-01
-3.04793101e-02 -2.37787098e-01 3.52431267e-01 3.43191087e-01
1.40217632e-01 -2.07015902e-01 -9.64361846e-01 1.58018696e+00
-4.36942607e-01 1.84416652e-01 2.68551797e-01 -5.07420480e-01
7.24301636e-01 3.18197787e-01 3.91815484e-01 -2.78041273e-01
3.28400075e-01 2.27474004e-01 3.15450281e-01 -1.03708696e+00
4.71548200e-01 9.43335071e-02 -3.54671299e-01 4.73984748e-01
2.45180115e-01 -8.13414097e-01 3.30532104e-01 -1.06770188e-01
1.41795290e+00 8.13387632e-01 1.64407745e-01 -1.19414153e-02
6.87738284e-02 3.69180560e-01 4.33603853e-01 9.10193205e-01
-4.29192990e-01 3.99901986e-01 5.37556350e-01 -2.05598950e-01
-1.46301126e+00 -9.47834253e-01 6.60869658e-01 1.36500835e+00
1.37254059e-01 1.14318609e-01 -6.67396486e-01 -6.46489382e-01
2.80485064e-01 7.40214765e-01 -5.63419998e-01 -3.04430425e-01
-8.74610007e-01 -1.64725304e-01 2.34797463e-01 6.83469176e-01
8.95363763e-02 -1.77294016e+00 -1.27775037e+00 2.35349849e-01
4.90745865e-02 -1.12671709e+00 -6.93098664e-01 3.10307890e-01
-5.62090933e-01 -1.11216021e+00 -1.21810901e+00 -8.06989491e-01
8.27970743e-01 8.32723826e-02 7.68375516e-01 -1.77237868e-01
-5.05335808e-01 7.59673238e-01 -3.25312197e-01 -1.93648547e-01
-5.52479506e-01 -1.64039060e-01 4.34161097e-01 -7.11991429e-01
-1.38173088e-01 -3.92484128e-01 -5.96118569e-01 4.35954332e-01
-5.26099265e-01 -2.32686758e-01 1.01460230e+00 1.30639899e+00
1.61154583e-01 -2.27741092e-01 4.61557269e-01 -2.00241879e-01
1.00219524e+00 -4.41775560e-01 -7.94582248e-01 -1.26454458e-02
-1.39082640e-01 2.68739820e-01 8.29183400e-01 -1.05531824e+00
-9.43818867e-01 3.90933961e-01 4.44274396e-01 -8.66723776e-01
-2.16524899e-01 -3.90328988e-02 8.32711160e-02 1.05399862e-01
8.12510073e-01 -1.22325674e-01 2.63087451e-01 1.45085726e-03
4.54433292e-01 4.18677747e-01 7.03991532e-01 -1.19039071e+00
7.45928466e-01 4.88859639e-02 -8.59759897e-02 -6.51061296e-01
-3.13287854e-01 -1.35453165e-01 -6.40065968e-01 -2.89135784e-01
8.64772260e-01 -7.42452204e-01 -1.36219883e+00 2.57688910e-01
-1.16398633e+00 -1.25085402e+00 -2.89822131e-01 6.79423511e-01
-1.44586647e+00 2.08348826e-01 -7.72310019e-01 -8.31344366e-01
1.09565169e-01 -1.51644123e+00 1.11838055e+00 1.73664644e-01
-6.27556562e-01 -3.32845092e-01 -2.45424937e-02 -3.23232599e-02
1.93794131e-01 1.55098110e-01 7.46814132e-01 -1.43341541e-01
-6.96384251e-01 -5.18483669e-02 6.11622073e-02 2.26596713e-01
-1.51473610e-02 -1.78530321e-01 -3.62776726e-01 -6.63689613e-01
-2.33753756e-01 -1.13769412e+00 1.47836596e-01 1.29492357e-01
1.10284114e+00 -3.83708239e-01 -2.87318289e-01 9.20965075e-02
8.10312033e-01 3.13657016e-01 4.06277001e-01 2.67173141e-01
6.45987689e-01 6.35600567e-01 1.38540268e+00 7.75358796e-01
1.18860029e-01 5.37926376e-01 6.47069335e-01 3.87820810e-01
1.77065834e-01 -4.59101498e-01 7.43571520e-01 1.50800541e-01
-5.71865201e-01 -1.20453291e-01 -6.95803165e-01 4.77263242e-01
-2.15951777e+00 -9.82512951e-01 3.25254738e-01 1.94787502e+00
9.16330695e-01 -6.44308627e-02 4.90025550e-01 -3.37929845e-01
6.38861239e-01 -4.65950370e-01 -1.12159395e+00 -2.57367771e-02
4.89921540e-01 2.59325188e-03 6.03708148e-01 2.43350714e-01
-8.67522538e-01 1.13109934e+00 6.35622883e+00 4.87852097e-01
-8.98509562e-01 -2.17484504e-01 -7.90793151e-02 -1.87587053e-01
2.52905071e-01 -1.56143919e-01 -4.57102835e-01 3.67000014e-01
4.01484966e-01 6.17070980e-02 9.78288233e-01 1.29245627e+00
4.34949398e-01 -1.91076130e-01 -1.34438956e+00 9.12722886e-01
-4.01292145e-01 -5.87888956e-01 -2.43886635e-01 2.60822140e-02
7.49973357e-01 -1.20044164e-02 -1.00735370e-02 7.99108744e-01
1.22172260e+00 -1.15093064e+00 9.00436878e-01 3.23097676e-01
7.63192415e-01 -6.68757737e-01 4.14525926e-01 7.83166349e-01
-5.98943949e-01 -6.44212842e-01 -4.61844027e-01 -1.76774412e-01
4.15109128e-01 -3.42428595e-01 -1.18115783e+00 -1.52786877e-02
6.25940979e-01 4.83434677e-01 1.02732368e-01 7.74426937e-01
-6.74181521e-01 5.57200648e-02 -1.62925661e-01 -4.93536741e-01
5.36184549e-01 -1.55314445e-01 7.13590324e-01 9.49169755e-01
3.08983624e-01 2.46068642e-01 7.81200826e-01 9.49265540e-01
2.87328243e-01 -4.65479970e-01 -9.28126335e-01 -3.08913887e-01
4.43762988e-01 1.19067287e+00 -4.51408207e-01 -1.97523996e-01
1.43147156e-01 1.25182104e+00 5.58392465e-01 6.26606345e-01
-1.11711574e+00 -4.20647711e-01 6.19766951e-01 -2.09338129e-01
3.82083118e-01 -7.65842736e-01 3.52224201e-01 -8.96615922e-01
-9.88732651e-02 -1.24824834e+00 -7.61336833e-02 -8.76250803e-01
-1.32354128e+00 3.29406172e-01 8.00706819e-02 -1.52378523e+00
-8.26119900e-01 -1.05081904e+00 -2.40962088e-01 5.73233545e-01
-9.57112432e-01 -6.76207781e-01 -3.66183549e-01 7.26749003e-01
5.76007009e-01 -2.58790702e-01 6.51286602e-01 -1.20481215e-01
-2.37494946e-01 2.61151671e-01 -2.45158017e-01 -1.05784588e-01
9.47196662e-01 -1.21891642e+00 1.20449044e-01 1.28374889e-01
-8.40890229e-01 7.22903371e-01 9.76695120e-01 -6.90377951e-01
-1.82823944e+00 -8.22301567e-01 -1.84136286e-01 -5.35254300e-01
1.01742256e+00 -2.20892906e-01 -5.29104352e-01 8.54219735e-01
2.39427537e-01 -2.43626982e-01 -8.14080909e-02 -4.63797539e-01
9.61264446e-02 4.69700456e-01 -1.06723762e+00 1.06676912e+00
1.23537433e+00 -8.15377235e-02 -7.21384108e-01 5.22053957e-01
8.42584252e-01 -6.62708938e-01 -8.15136611e-01 3.90163600e-01
6.49345696e-01 -3.44996780e-01 8.73715460e-01 -9.26479220e-01
5.06027877e-01 -5.71089208e-01 5.19230515e-02 -1.59071171e+00
-5.22866011e-01 -1.15375447e+00 -9.07947198e-02 4.19872314e-01
8.77313539e-02 -2.82475263e-01 5.77634215e-01 5.09769917e-01
-1.22966468e-01 -4.75806236e-01 -4.77469534e-01 -1.23465312e+00
-2.60168482e-02 2.99185272e-02 3.63507152e-01 6.09021962e-01
3.88227075e-01 3.64690334e-01 -9.78095472e-01 5.17844334e-02
4.88752991e-01 -2.09884599e-01 1.36017966e+00 -8.41954231e-01
-7.95512259e-01 -2.70359635e-01 5.31892516e-02 -1.34446776e+00
4.24274385e-01 -5.40839434e-01 1.00922489e+00 -1.17848134e+00
2.14880541e-01 -4.78923529e-01 2.93323010e-01 5.21986365e-01
-1.24428540e-01 -4.20441926e-01 5.63945889e-01 2.22251356e-01
-7.72024453e-01 8.31386149e-01 2.15465689e+00 -1.29067525e-01
-2.65575856e-01 4.51504253e-02 -1.30018190e-01 6.58961713e-01
7.06360936e-01 -2.80301303e-01 -6.39851987e-01 -3.63428980e-01
-9.42414030e-02 5.18442094e-01 5.22692621e-01 -1.09288788e+00
5.71311675e-02 -7.26783872e-01 1.68146491e-01 -1.63568974e-01
3.21766466e-01 -9.35818255e-01 2.88267402e-05 9.71970081e-01
-4.24287677e-01 1.10972427e-01 -6.79595247e-02 8.46366704e-01
2.35027671e-01 -2.56403834e-01 7.30028331e-01 -5.36072373e-01
-8.82106006e-01 3.72903943e-01 -6.37677431e-01 2.32120547e-02
1.25963151e+00 1.96830571e-01 -1.03460148e-01 -5.54319680e-01
-8.98663521e-01 8.45914602e-01 6.35344505e-01 5.44607043e-01
4.54167247e-01 -1.16470945e+00 -4.04920846e-01 -1.30557597e-01
3.96536179e-02 -2.23635081e-02 1.26200030e-02 5.63937902e-01
-4.34090972e-01 2.22371489e-01 -7.68622339e-01 -5.88555813e-01
-7.35733449e-01 8.12768877e-01 5.19505590e-02 -2.97625870e-01
-7.04403281e-01 4.52324688e-01 1.74292982e-01 -7.65612721e-01
5.33622861e-01 -4.34260339e-01 4.78239320e-02 -6.22038126e-01
1.85686603e-01 4.97365057e-01 -7.29347587e-01 -6.88478500e-02
3.00719719e-02 3.11985999e-01 8.01691711e-02 -3.56288522e-01
1.24523187e+00 2.44981587e-01 1.65184602e-01 1.45407706e-01
5.13332427e-01 -5.38592637e-01 -2.27831030e+00 3.11152369e-01
-2.80152168e-02 -3.28607231e-01 -8.48419666e-01 -8.14757586e-01
-4.59740728e-01 8.39086413e-01 1.80630594e-01 -3.72993708e-01
5.12862384e-01 -2.03410819e-01 6.83055580e-01 1.27464545e+00
1.16718650e+00 -1.39848650e+00 8.01774621e-01 1.00261343e+00
1.32243276e+00 -1.36916709e+00 -2.67811865e-01 -1.75698102e-02
-1.23958051e+00 1.09483409e+00 1.00936425e+00 -5.82124650e-01
1.67582601e-01 3.03563416e-01 -1.32941678e-01 3.16400826e-02
-7.44117200e-01 -5.90104014e-02 -1.63191557e-01 9.11972225e-01
-1.06648408e-01 2.24015564e-01 -4.22301367e-02 4.33597952e-01
-7.72183612e-02 4.14495856e-01 6.67870879e-01 1.21783841e+00
-5.36714554e-01 -8.59366119e-01 -3.82785738e-01 6.94244802e-02
1.45339683e-01 5.25703967e-01 -1.92926526e-01 1.02546418e+00
-2.36819431e-01 6.30133450e-01 -3.20407778e-01 -3.74794960e-01
5.42848587e-01 -1.92694366e-01 1.02270794e+00 -8.32553089e-01
-3.61307263e-01 -2.11327314e-01 -5.70817888e-02 -9.78981972e-01
3.72186340e-02 -6.73666477e-01 -1.54968047e+00 -1.18210427e-01
8.28240663e-02 -2.09146902e-01 2.08941355e-01 7.20167339e-01
2.84872174e-01 6.04506195e-01 4.22120243e-01 -1.76115644e+00
-1.52276480e+00 -1.09327519e+00 -5.35287976e-01 6.27135396e-01
3.79545689e-01 -1.25253892e+00 -1.92029551e-01 1.48418322e-02] | [4.586612701416016, 0.8007894158363342] |
23bca87b-16c3-4980-a9a7-317865623a61 | misinformation-as-information-pollution | 2306.12466 | null | https://arxiv.org/abs/2306.12466v1 | https://arxiv.org/pdf/2306.12466v1.pdf | Misinformation as Information Pollution | Social media feed algorithms are designed to optimize online social engagements for the purpose of maximizing advertising profits, and therefore have an incentive to promote controversial posts including misinformation. By thinking about misinformation as information pollution, we can draw parallels with environmental policy for countering pollution such as carbon taxes. Similar to pollution, a Pigouvian tax on misinformation provides economic incentives for social media companies to control the spread of misinformation more effectively to avoid or reduce their misinformation tax, while preserving some degree of freedom in platforms' response. In this paper, we highlight a bird's eye view of a Pigouvian misinformation tax and discuss the key questions and next steps for implementing such a taxing scheme. | ['Rada Mihalcea', 'Ashkan Kazemi'] | 2023-06-21 | null | null | null | null | ['misinformation'] | ['miscellaneous'] | [ 1.44128025e-01 7.69446254e-01 -9.83774304e-01 -1.71230972e-01
-4.19462204e-01 -6.13955140e-01 7.40873218e-01 4.81800377e-01
-6.37174368e-01 6.59027636e-01 8.13919306e-01 -8.43610704e-01
-5.87453917e-02 -1.15598822e+00 -6.40586257e-01 -1.68152779e-01
2.82034785e-01 1.85173422e-01 4.71489727e-02 -4.37594950e-01
4.19544399e-01 -5.03516421e-02 -1.18996060e+00 5.78265414e-02
6.76993072e-01 6.43719196e-01 6.39343634e-02 2.08455682e-01
-6.87110305e-01 7.86332667e-01 -4.89883304e-01 -9.58679199e-01
4.52379465e-01 -1.94939002e-01 -4.24073756e-01 -1.06773332e-01
3.66159260e-01 -2.77423471e-01 -2.05155075e-01 1.57574809e+00
-6.41367510e-02 -1.89934596e-01 2.31426850e-01 -8.84159803e-01
-6.49574280e-01 1.09337652e+00 -7.16007948e-01 1.56652093e-01
5.08947335e-02 1.70268938e-01 1.54257143e+00 -4.28151749e-02
7.73922086e-01 1.36020172e+00 1.70889363e-01 2.02823356e-01
-1.18181431e+00 -8.90506089e-01 5.10312736e-01 -1.79074436e-01
-6.19959652e-01 -2.45764896e-01 4.31318849e-01 -5.51375628e-01
3.21575046e-01 1.05721402e+00 8.85077238e-01 8.64893079e-01
3.86631250e-01 4.74799275e-01 1.35893488e+00 -7.11610317e-02
2.61979848e-01 5.47400117e-01 1.53576374e-01 1.88131347e-01
9.31589663e-01 3.56767297e-01 -4.20113832e-01 -4.17349905e-01
1.46886393e-01 5.94468340e-02 2.05652878e-01 2.77319908e-01
-8.22288096e-01 1.32012784e+00 5.87146223e-01 3.40811461e-01
-4.51151788e-01 3.22481960e-01 1.22782653e-02 5.32637417e-01
9.97820854e-01 8.84269536e-01 -2.95980692e-01 1.66891038e-01
-2.22327322e-01 4.79521036e-01 9.02945876e-01 4.31565553e-01
9.92345154e-01 -1.88475624e-01 -1.59149498e-01 4.28782254e-01
7.79783309e-01 7.59371877e-01 -3.62724550e-02 -1.13068080e+00
3.70755970e-01 5.17925203e-01 4.22450781e-01 -1.33283758e+00
-1.05286345e-01 -6.93382621e-01 -2.77425200e-01 2.33321086e-01
4.63774264e-01 -3.94274652e-01 -2.00628281e-01 1.32253861e+00
2.78389722e-01 -4.94341195e-01 -4.99252200e-01 9.97295737e-01
3.06929983e-02 7.17577994e-01 4.83755589e-01 -3.61984313e-01
1.53240192e+00 -2.88318038e-01 -9.82565939e-01 -6.08509660e-01
4.35819626e-01 -8.14631045e-01 8.76212776e-01 2.10475013e-01
-8.36490273e-01 4.04019624e-01 -5.76437771e-01 2.59535283e-01
-5.20165026e-01 -8.53008747e-01 8.72085631e-01 7.29437351e-01
-4.54502195e-01 5.69296896e-01 -1.68454424e-01 -6.23139143e-02
7.64990032e-01 5.39750867e-02 5.52682281e-01 2.39462405e-01
-1.29709601e+00 9.39390421e-01 -1.46163106e-01 -4.98069197e-01
-6.47502601e-01 -8.35843682e-01 -2.44997635e-01 2.21118480e-01
8.24057102e-01 -4.55449641e-01 1.36720407e+00 -1.26652730e+00
-1.08296001e+00 7.89773226e-01 2.47286990e-01 -6.99443996e-01
5.59695601e-01 -7.23451898e-02 -6.71145320e-01 -2.60360986e-01
4.20822829e-01 4.04973835e-01 8.72997940e-01 -8.36401820e-01
-9.42502558e-01 -5.57090700e-01 3.36731046e-01 8.50852951e-02
-5.87258518e-01 4.25215632e-01 3.91845375e-01 -3.71380091e-01
-2.62135178e-01 -8.87404442e-01 -8.02347004e-01 -1.23934887e-01
-4.36282784e-01 -2.06796259e-01 8.58120203e-01 -3.40637684e-01
1.19508517e+00 -1.82875133e+00 -6.58403575e-01 2.80905992e-01
6.63302302e-01 1.31255552e-01 1.34080559e-01 2.76511908e-01
4.54791158e-01 8.25951338e-01 2.32365236e-01 2.86352873e-01
2.40951329e-01 -1.64728239e-02 -2.00721458e-01 5.50501347e-01
-4.16013561e-02 7.57252634e-01 -1.04570925e+00 7.09667206e-02
-1.94498301e-01 -1.79656342e-01 -7.94220686e-01 -4.52440202e-01
-9.76563573e-01 3.07254195e-01 -7.93158829e-01 4.73043919e-01
5.04617512e-01 -4.35668319e-01 5.30982494e-01 2.72930503e-01
-8.01184952e-01 9.14826512e-01 -4.80883896e-01 6.51522756e-01
-4.08978164e-01 7.98723042e-01 5.94309330e-01 -3.64170700e-01
6.00908458e-01 -2.33514830e-01 7.37312973e-01 -9.27278101e-01
5.27005136e-01 1.33768067e-01 7.12361187e-03 -3.41894954e-01
7.14292109e-01 -2.65092522e-01 -1.84969142e-01 9.30548072e-01
-7.65975833e-01 7.62712732e-02 -9.36241969e-02 5.10729492e-01
8.37933481e-01 -5.12158334e-01 1.19792195e-02 -6.56871140e-01
-6.83245510e-02 4.36277390e-01 1.27402067e-01 7.32569098e-01
-2.06518605e-01 -5.36088586e-01 4.23882991e-01 -1.98081091e-01
-9.87811089e-01 -5.18269062e-01 1.92806423e-01 1.35234165e+00
1.13977760e-01 -4.99778181e-01 -5.14804900e-01 -7.59139240e-01
4.82238829e-01 1.15606940e+00 -2.16668636e-01 2.26463214e-01
-1.84238374e-01 -8.41231346e-01 -1.10639378e-01 -4.19980735e-01
5.05860031e-01 -2.03282326e-01 -1.75404012e-01 1.51448727e-01
-2.12251082e-01 -7.52719820e-01 -8.45324636e-01 -1.65057972e-01
-4.34623718e-01 -9.02696908e-01 -2.04786852e-01 9.86247584e-02
3.78780603e-01 7.57682860e-01 8.87199223e-01 -1.33698985e-01
1.51080295e-01 3.41688573e-01 4.12326446e-03 -1.36817598e+00
-6.92378700e-01 2.60425508e-02 -2.63872027e-01 1.50187775e-01
7.39262104e-01 -5.47034144e-01 -7.71793544e-01 2.89194614e-01
-9.25549328e-01 -2.38506138e-01 2.83487737e-01 9.70889777e-02
1.28515303e-01 2.83279754e-02 7.10491598e-01 -1.24841344e+00
1.02859390e+00 -1.05828381e+00 -1.25110793e+00 -4.02276307e-01
-1.15093827e+00 -1.61653176e-01 5.43969534e-02 -2.97523618e-01
-1.17708719e+00 -3.60208303e-01 5.19165583e-02 3.13115627e-01
5.41370869e-01 5.90384066e-01 2.32225910e-01 -3.16518277e-01
1.02716315e+00 -5.42593062e-01 4.22813892e-01 -5.69152057e-01
7.54716098e-01 7.04824448e-01 -3.38642687e-01 -8.87738466e-02
6.50872290e-01 8.85228038e-01 -2.50531554e-01 -9.37878013e-01
-1.07742226e+00 -5.97210646e-01 4.10135418e-01 -5.11367321e-01
8.74012589e-01 -8.18839192e-01 -1.25744402e+00 -9.32378843e-02
-1.00261688e+00 1.57042131e-01 -5.07455528e-01 5.33803582e-01
-3.36440932e-03 1.35558292e-01 -2.24677876e-01 -1.03967071e+00
1.77900475e-02 -4.70938861e-01 2.27897435e-01 1.11699700e-01
-2.74523914e-01 -9.20927584e-01 -3.40347551e-02 8.06672633e-01
1.18189800e+00 4.83499840e-02 2.89485067e-01 -7.77264595e-01
-1.16098344e+00 -2.50340790e-01 -2.49664381e-01 1.39803112e-01
1.27323577e-02 -3.45906377e-01 -8.07517350e-01 1.28807127e-01
2.42874175e-01 9.62374434e-02 7.33629167e-01 5.11197507e-01
6.69573069e-01 -1.24666750e+00 -4.45775211e-01 -1.49318933e-01
1.33782554e+00 -3.83632779e-02 5.00719368e-01 4.60307479e-01
3.12056422e-01 8.13928545e-01 5.91252387e-01 7.02028453e-01
4.56775635e-01 3.35831374e-01 8.24961662e-01 1.33765519e-01
1.28813863e-01 -8.13494265e-01 4.24926698e-01 4.52866137e-01
2.17801809e-01 -2.41449997e-02 -3.21856230e-01 3.90086979e-01
-1.51265228e+00 -1.00795674e+00 -4.80575562e-01 2.00738430e+00
5.29674649e-01 3.55265439e-01 5.97778082e-01 -3.64118814e-01
9.52222109e-01 6.27580047e-01 -2.42680416e-01 -5.59912205e-01
2.10907776e-02 -3.96576345e-01 1.74519038e+00 8.84831488e-01
-6.06304407e-01 9.24048185e-01 6.86656713e+00 7.00026214e-01
-1.00022328e+00 5.75861752e-01 7.21449852e-01 -2.18629599e-01
-1.26750398e+00 4.21212554e-01 -1.01625097e+00 6.46208882e-01
1.12721765e+00 -8.23397219e-01 7.07570016e-01 6.37420833e-01
8.70792091e-01 -4.51541245e-01 -3.02108794e-01 1.75845653e-01
-5.79869688e-01 -1.83811200e+00 -1.51696086e-01 5.99221706e-01
8.63878131e-01 4.33550984e-01 1.97507769e-01 -5.00885397e-02
1.21919370e+00 -4.94186193e-01 7.56897271e-01 3.43523324e-01
2.41174638e-01 -3.79120797e-01 -1.65725425e-02 5.91767788e-01
-3.92674834e-01 -3.86705667e-01 -2.72446990e-01 -6.72827959e-01
4.14713442e-01 1.16079843e+00 -9.12840426e-01 -1.07114322e-01
2.18385100e-01 4.40306634e-01 -1.81662217e-01 7.21195936e-01
-2.12810457e-01 6.86019242e-01 -1.97561413e-01 -7.60159492e-01
7.10413992e-01 -5.81173301e-01 9.94732797e-01 6.50321305e-01
1.70693453e-02 5.13087027e-02 -1.38165593e-01 1.09070730e+00
-2.52865523e-01 1.97808012e-01 -7.74709523e-01 -8.64312112e-01
6.23028576e-01 9.10672724e-01 -1.79051176e-01 -9.77933556e-02
-5.88523090e-01 -1.14310989e-02 -1.90882400e-01 1.25488788e-01
-6.16674840e-01 2.04674944e-01 9.06911910e-01 9.81089711e-01
-2.40488619e-01 -2.21410114e-02 -6.07474089e-01 -8.45841348e-01
-5.05417526e-01 -8.43564034e-01 3.35335404e-01 -3.49894494e-01
-1.14248347e+00 -3.71425122e-01 -2.66919196e-01 -6.83251560e-01
3.74991894e-01 -2.86409467e-01 -5.22437036e-01 6.55374587e-01
-1.86026418e+00 -6.50124490e-01 3.68371516e-01 -5.67576177e-02
2.50152022e-01 1.74434558e-01 1.14618711e-01 3.10667515e-01
5.63725829e-02 -1.17269993e-01 1.35906875e-01 -5.93410015e-01
5.00280380e-01 -7.42931902e-01 3.61475557e-01 2.68904865e-01
-7.48335347e-02 4.88507509e-01 9.97480094e-01 -9.65683639e-01
-1.61744654e+00 -1.35343099e+00 1.13305783e+00 -4.55741316e-01
1.38927555e+00 -4.05227356e-02 -3.05055290e-01 7.82182157e-01
3.99133652e-01 -8.27672303e-01 5.88583052e-01 4.11617130e-01
-2.25225732e-01 -1.95061043e-01 -1.30990827e+00 8.22375357e-01
9.46291387e-01 -2.32864171e-01 -3.36972997e-02 1.12846327e+00
1.14071238e+00 2.72473067e-01 -4.76048112e-01 -5.42443097e-01
6.15784347e-01 -6.37644947e-01 7.11494625e-01 -6.85512304e-01
4.03624505e-01 1.18011422e-01 -2.73789138e-01 -1.41459906e+00
-4.72770333e-01 -1.06050193e+00 4.70023841e-01 6.87289298e-01
6.60596013e-01 -1.13746572e+00 8.91440332e-01 6.61058664e-01
9.36242640e-02 -1.53993949e-01 -6.32418931e-01 -6.53798759e-01
6.22013025e-03 -4.71169084e-01 6.62769496e-01 1.09014595e+00
1.07264660e-01 3.12744558e-01 -4.63966578e-01 -2.29391898e-03
9.24776196e-01 -4.35882300e-01 6.13559723e-01 -1.16093767e+00
-1.52720273e-01 -5.73721290e-01 2.37271383e-01 -8.86520684e-01
1.22496835e-03 -1.01104796e+00 -3.43250394e-01 -1.08176148e+00
1.77486297e-02 -4.91360575e-01 1.96288805e-02 -8.55837092e-02
3.22239786e-01 3.58841978e-02 3.48851383e-01 2.89316297e-01
-5.13638616e-01 -1.00707531e-01 1.54370320e+00 -5.23668706e-01
-2.91017681e-01 2.31908634e-01 -1.94256532e+00 6.96648300e-01
8.57627332e-01 -8.17475021e-01 -1.76957980e-01 -3.40798855e-01
1.18151486e+00 -1.09227598e-01 3.24470580e-01 1.29858674e-02
3.28333020e-01 -9.89035547e-01 -5.40436804e-01 -3.39926898e-01
7.10188150e-02 -8.66974175e-01 4.69912827e-01 6.81552708e-01
-7.86124587e-01 -3.89096588e-01 -2.40749970e-01 1.05586600e+00
1.04440421e-01 -1.79307580e-01 9.08371925e-01 -5.77096343e-01
-5.37111759e-02 1.13184884e-01 -8.87756109e-01 -7.86105767e-02
8.97929251e-01 2.92443722e-01 -9.35526013e-01 -8.13036263e-01
-6.89466953e-01 4.92800385e-01 4.20730650e-01 5.04995525e-01
-9.63824540e-02 -1.08487248e+00 -7.66038835e-01 -2.39590585e-01
-1.14593163e-01 -7.42970407e-01 -6.89059421e-02 7.74768293e-01
-3.60586047e-02 4.85144794e-01 1.71392485e-01 -4.96400446e-02
-8.87371004e-01 4.30697918e-01 2.69375056e-01 -1.93472311e-01
1.57848354e-02 5.76156378e-01 6.82965517e-02 1.55930305e-02
1.63444746e-02 -1.48680983e-02 -9.22368243e-02 3.91306728e-01
7.86623597e-01 7.43204474e-01 -2.01972172e-01 -1.75462738e-01
-1.43483236e-01 -2.38332286e-01 -7.55412653e-02 -2.02741712e-01
1.18766379e+00 -4.35749948e-01 -3.31821740e-01 1.58708528e-01
9.61756527e-01 6.85248137e-01 -7.19055355e-01 -2.36602202e-01
2.61077046e-01 -1.18357813e+00 6.24762297e-01 -1.12324691e+00
-9.71201003e-01 2.45036051e-01 2.38893494e-01 1.08438218e+00
3.29038233e-01 1.67883143e-01 7.09364295e-01 7.48451129e-02
1.92236781e-01 -1.34563673e+00 -6.09175041e-02 -1.25355795e-02
7.51643658e-01 -1.00556231e+00 1.80043578e-01 -7.58074880e-01
-6.47392631e-01 3.19332838e-01 7.57798105e-02 -2.60699373e-02
9.52405214e-01 2.13954095e-02 -2.78906316e-01 -6.57428801e-01
-8.45158875e-01 -3.92817706e-01 -1.97569504e-01 3.89655709e-01
2.63389051e-01 6.28805339e-01 -1.28014839e+00 5.50438464e-01
5.47588952e-02 -1.26640394e-01 7.39642143e-01 5.86065054e-01
-1.30301154e+00 -1.07367885e+00 -5.58763623e-01 1.02718627e+00
-9.11353707e-01 -2.31270306e-02 -8.69959414e-01 4.73962307e-01
6.00624271e-03 1.14051235e+00 -1.63252503e-02 -1.75807297e-01
2.31511220e-01 -3.41944367e-01 -9.57966894e-02 -5.01828611e-01
-6.60645604e-01 2.53489584e-01 8.78822923e-01 -4.04243529e-01
-3.86854380e-01 -7.47696877e-01 -5.17657697e-01 -8.55342686e-01
-5.56824863e-01 2.91737467e-01 1.14728415e+00 6.71108544e-01
5.66893637e-01 7.29824975e-02 1.03912067e+00 -9.81854796e-02
-9.73342359e-01 -5.46614230e-01 -7.49739468e-01 1.28697261e-01
1.26855567e-01 -3.79228145e-01 -6.65070534e-01 -4.84939933e-01] | [9.066611289978027, 5.8225579261779785] |
f54ed931-aebb-4991-95c9-f9383457c82c | stock-index-prediction-using-cointegration | 2109.15045 | null | https://arxiv.org/abs/2109.15045v1 | https://arxiv.org/pdf/2109.15045v1.pdf | Stock Index Prediction using Cointegration test and Quantile Loss | Recent researches on stock prediction using deep learning methods has been actively studied. This is the task to predict the movement of stock prices in the future based on historical trends. The approach to predicting the movement based solely on the pattern of the historical movement of it on charts, not on fundamental values, is called the Technical Analysis, which can be divided into univariate and multivariate methods in the regression task. According to the latter approach, it is important to select different factors well as inputs to enhance the performance of the model. Moreover, its performance can depend on which loss is used to train the model. However, most studies tend to focus on building the structures of models, not on how to select informative factors as inputs to train them. In this paper, we propose a method that can get better performance in terms of returns when selecting informative factors using the cointegration test and learning the model using quantile loss. We compare the two RNN variants with quantile loss with only five factors obtained through the cointegration test among the entire 15 stock index factors collected in the experiment. The Cumulative return and Sharpe ratio were used to evaluate the performance of trained models. Our experimental results show that our proposed method outperforms the other conventional approaches. | ['Minjung Kang', 'Heejoon Lee', 'Jaeyoung Cheong'] | 2021-09-29 | null | null | null | null | ['stock-prediction'] | ['time-series'] | [-6.34338021e-01 -6.25337720e-01 -1.89030007e-01 -2.98058897e-01
-5.29640377e-01 -5.18466115e-01 4.90391374e-01 -5.72196953e-02
-5.90001285e-01 8.17168832e-01 1.69108197e-01 -5.69404721e-01
-2.21555963e-01 -1.08998883e+00 -7.73131073e-01 -8.02756250e-01
-2.24710748e-01 3.54048342e-01 2.59820700e-01 -4.17832762e-01
4.31878179e-01 5.29904127e-01 -1.45779669e+00 -4.48994488e-02
5.73193669e-01 1.39596498e+00 -7.54474252e-02 3.33194882e-01
-2.98638135e-01 1.07694411e+00 -6.86939776e-01 -6.34596050e-01
5.74115872e-01 -4.03700799e-01 -1.28137782e-01 -3.98140341e-01
-8.87179300e-02 -4.75167245e-01 -1.36761814e-01 1.06308389e+00
3.94709855e-01 2.86896192e-02 7.38905072e-01 -1.05995297e+00
-4.58424240e-01 1.05529249e+00 -5.59845686e-01 5.12952447e-01
-2.39123195e-01 -1.08661920e-01 1.43432319e+00 -8.60388100e-01
1.25613049e-01 9.85393345e-01 6.00018680e-01 4.17864211e-02
-1.02206290e+00 -9.20564830e-01 1.73308685e-01 3.27275246e-01
-9.06027913e-01 5.23177460e-02 9.64315474e-01 -7.49097407e-01
4.08377528e-01 6.70324415e-02 6.23019636e-01 8.09817851e-01
4.95457798e-01 8.28389466e-01 9.33871031e-01 -2.28049383e-01
2.61165410e-01 2.62025148e-01 2.86143035e-01 4.85945977e-02
2.82420695e-01 3.22816342e-01 -3.12990487e-01 -8.05094466e-03
6.77305162e-01 1.74930394e-01 -2.73252875e-01 -8.88854638e-03
-9.12209749e-01 1.07837629e+00 5.15878499e-01 4.86817211e-01
-7.06199050e-01 5.83647937e-02 3.80960405e-01 5.57007849e-01
6.29038990e-01 4.14106935e-01 -8.09840858e-01 1.04981270e-02
-1.25917566e+00 4.27555263e-01 7.76753962e-01 2.89678395e-01
4.63780642e-01 4.14478749e-01 -1.76690280e-01 3.91041815e-01
1.74629614e-01 4.71355319e-01 8.26300919e-01 -2.53084898e-01
5.68073630e-01 4.15299356e-01 2.43483886e-01 -1.05798459e+00
-4.30007994e-01 -9.14274991e-01 -7.74913847e-01 4.78732884e-01
6.04052663e-01 -5.33205926e-01 -7.14839578e-01 1.57821238e+00
-1.79082379e-01 2.03861639e-01 -7.74181113e-02 6.90130591e-01
2.11624682e-01 1.03200352e+00 -1.30759060e-01 -2.74822623e-01
8.08990836e-01 -5.90716243e-01 -8.20308387e-01 3.29030603e-01
4.13135082e-01 -7.69437134e-01 6.91474795e-01 5.52830577e-01
-9.11138594e-01 -6.25176311e-01 -1.11351180e+00 4.68181193e-01
-5.22960901e-01 2.46549025e-01 3.03157389e-01 3.61296117e-01
-7.52635241e-01 1.00614154e+00 -7.78694749e-01 6.05337441e-01
1.35013208e-01 2.21261919e-01 1.60134107e-01 6.06064558e-01
-1.60871458e+00 9.47882593e-01 5.32402277e-01 6.13918342e-02
-5.15038252e-01 -8.54279101e-01 -3.02116275e-01 4.73911136e-01
7.53758699e-02 -2.60350376e-01 1.07337213e+00 -1.08168566e+00
-1.30957842e+00 7.86490440e-02 5.19124329e-01 -1.01910162e+00
9.91424143e-01 -3.67358238e-01 -3.86297405e-01 -2.78361112e-01
-1.22427031e-01 -4.12106700e-02 7.39616394e-01 -7.93385923e-01
-1.04415941e+00 -3.23438525e-01 -1.08064450e-01 -2.22714573e-01
-2.25709394e-01 6.58138022e-02 -7.70890787e-02 -1.14785612e+00
1.98097695e-02 -6.43216372e-01 -2.04863802e-01 -4.18095440e-01
-2.04659864e-01 -3.32362264e-01 3.90646487e-01 -1.05953038e+00
1.36218107e+00 -1.94715655e+00 -2.07702756e-01 4.00834054e-01
-2.30991423e-01 1.56362161e-01 2.86223829e-01 3.28477800e-01
-2.03305423e-01 5.68374023e-02 -1.48173831e-02 -3.81597388e-03
2.68817157e-01 -3.85641128e-01 -9.00358617e-01 3.53940785e-01
9.96081829e-02 6.47562444e-01 -2.63916612e-01 -4.46078368e-02
-4.83573116e-02 4.50186193e-01 -2.12866038e-01 3.43965888e-01
-2.90066391e-01 1.23771258e-01 -4.07751381e-01 2.98705310e-01
7.55026996e-01 -5.74842803e-02 -2.31468081e-01 6.39178157e-02
-2.82448560e-01 3.71310949e-01 -1.17334783e+00 7.02099681e-01
-4.19576079e-01 5.10647535e-01 -4.90629047e-01 -1.13714623e+00
1.30919921e+00 3.63954246e-01 4.38727170e-01 -6.98182642e-01
2.08497301e-01 5.95260739e-01 1.25521466e-01 -7.93845952e-02
3.17941427e-01 -4.95826036e-01 1.05250746e-01 4.32305366e-01
-2.42279872e-01 5.11225998e-01 3.44463766e-01 -3.55202645e-01
4.51840401e-01 1.69320509e-01 3.39644812e-02 -1.70861587e-01
5.60919404e-01 -2.17039689e-01 8.09703231e-01 4.49812651e-01
2.11020321e-01 2.80831724e-01 8.96879911e-01 -7.75917053e-01
-9.05359328e-01 -1.03118467e+00 -2.80224919e-01 9.04913306e-01
-2.38603681e-01 2.55365491e-01 -3.28615934e-01 -6.41923010e-01
1.84643209e-01 1.08092201e+00 -6.54400349e-01 5.73327430e-02
-4.46786851e-01 -1.16303575e+00 1.85065955e-01 6.57347023e-01
4.64233369e-01 -1.23534167e+00 -5.31915307e-01 3.81333321e-01
2.94002622e-01 -5.14286458e-01 -1.72221705e-01 3.21101516e-01
-8.75084639e-01 -9.97974396e-01 -1.09863937e+00 -4.82239127e-01
2.13814035e-01 -2.33727753e-01 1.13469422e+00 -2.89063364e-01
4.21829283e-01 -2.57830262e-01 -3.12808394e-01 -7.03958094e-01
-1.27610594e-01 2.61361450e-01 -1.71420842e-01 3.58923048e-01
3.57965082e-01 -7.22141981e-01 -5.99606156e-01 3.16764146e-01
-8.97231877e-01 -4.05332983e-01 7.85373807e-01 7.20677078e-01
4.44226056e-01 2.98191637e-01 1.00715756e+00 -7.45765030e-01
6.11731768e-01 -6.08512998e-01 -1.19423926e+00 4.14770305e-01
-8.28417718e-01 3.72517377e-01 8.71278346e-01 -3.43118072e-01
-8.70455325e-01 -2.61965126e-01 -1.28973007e-01 -3.62546623e-01
4.80911821e-01 8.00331771e-01 -1.00491695e-01 5.22880435e-01
2.05916867e-01 4.73277330e-01 -2.02211261e-01 -7.14400649e-01
-6.87451959e-02 2.63687819e-01 2.96484113e-01 -1.27339363e-01
7.99605846e-01 5.35472594e-02 5.87007925e-02 -3.99402171e-01
-7.77397513e-01 -2.01442510e-01 -5.28741658e-01 -1.90623671e-01
6.92264438e-01 -6.95819974e-01 -6.58849418e-01 6.13174021e-01
-8.88460279e-01 3.78629453e-02 -2.19980076e-01 9.23253357e-01
-2.82503992e-01 -1.75159454e-01 -6.35526538e-01 -1.13850331e+00
-5.06068707e-01 -1.04591310e+00 3.68013442e-01 1.31939679e-01
2.35181138e-01 -1.16393542e+00 3.40994567e-01 -2.20070839e-01
5.94555020e-01 3.10752124e-01 9.66212153e-01 -1.21287549e+00
-5.76785922e-01 -5.99334121e-01 -9.63140875e-02 7.03498244e-01
-1.51355593e-02 7.70440847e-02 -7.85873711e-01 -1.02860063e-01
4.05895531e-01 -3.52976061e-02 1.19934201e+00 6.64145887e-01
8.02313030e-01 -3.65542442e-01 2.25991189e-01 5.66658020e-01
1.53313029e+00 6.98066413e-01 7.45885909e-01 7.85048604e-01
3.60516638e-01 6.39943600e-01 5.29426932e-01 4.30019736e-01
-4.93294224e-02 2.53475457e-01 2.66305059e-01 2.12132335e-01
5.88400185e-01 -4.03379083e-01 5.62559307e-01 9.35127914e-01
-1.10149629e-01 -8.60486925e-02 -6.86992049e-01 2.75882423e-01
-1.78003812e+00 -1.19585299e+00 1.38578489e-01 2.30199075e+00
5.00308812e-01 6.14240766e-01 2.50844747e-01 3.06516230e-01
7.85857558e-01 3.36263627e-01 -6.27362609e-01 -7.07314387e-02
-3.26361507e-01 3.13969627e-02 6.99355364e-01 2.34617591e-01
-1.22960913e+00 4.19005036e-01 5.36423779e+00 9.56367850e-01
-1.56248820e+00 -3.57710063e-01 1.08951557e+00 8.91830996e-02
-3.77098083e-01 -2.37566754e-01 -1.11408830e+00 8.19889188e-01
9.09472287e-01 -2.80219853e-01 2.76408461e-03 8.99482310e-01
3.69960666e-01 2.63714522e-01 -7.60643005e-01 6.84454858e-01
-3.26094836e-01 -1.29652357e+00 1.82267070e-01 -4.45697717e-02
6.06854081e-01 -1.66238964e-01 3.52727175e-01 5.16415477e-01
2.93031428e-02 -9.02478456e-01 8.49662185e-01 1.04864883e+00
-1.94855947e-02 -1.24957407e+00 1.29543006e+00 4.39702928e-01
-9.93239224e-01 -9.84678939e-02 -3.92310172e-01 -4.24549952e-02
1.68954153e-02 6.80726349e-01 -5.26357293e-01 5.42085111e-01
4.26957577e-01 6.50059223e-01 -4.13964599e-01 1.22441745e+00
-2.00144872e-01 8.26933086e-01 -2.05146581e-01 -4.15560633e-01
2.93465078e-01 -5.19386470e-01 2.12750986e-01 7.83925891e-01
6.83436930e-01 -3.42756033e-01 -5.72342239e-03 8.25585783e-01
-6.46360219e-02 5.01816332e-01 -3.04368347e-01 1.45566329e-01
6.51045814e-02 7.19516456e-01 -6.18481815e-01 -2.23336175e-01
-5.26682079e-01 1.54231772e-01 -3.00494190e-02 1.70829773e-01
-7.17894912e-01 -5.35232067e-01 3.30283254e-01 3.57060552e-01
7.48412967e-01 3.11502982e-02 -3.11825305e-01 -1.13366044e+00
3.09362471e-01 -6.49482429e-01 4.99902844e-01 -3.95989090e-01
-1.56318569e+00 8.58852625e-01 -7.67996982e-02 -1.40869832e+00
-4.42854732e-01 -7.88229287e-01 -9.44089293e-01 1.10714841e+00
-1.69636631e+00 -5.76808274e-01 4.07717794e-01 3.98793250e-01
3.56860369e-01 -5.79170346e-01 1.61931902e-01 2.18035251e-01
-5.76018453e-01 4.45610046e-01 6.31026626e-01 4.49385315e-01
4.07522917e-01 -1.36276615e+00 3.10415328e-01 9.43898737e-01
1.65725887e-01 2.44051188e-01 6.97679877e-01 -6.47737086e-01
-5.71643949e-01 -8.76115680e-01 9.32809472e-01 1.27568254e-02
1.01743269e+00 5.13844974e-02 -1.09771025e+00 6.98353648e-01
5.44052683e-02 -1.88319340e-01 4.30892199e-01 -3.60932127e-02
-1.89384922e-01 -5.02192736e-01 -7.38106489e-01 2.79103398e-01
-2.17253398e-02 -6.84004277e-02 -6.92634225e-01 -2.11647153e-02
6.76696062e-01 2.44953454e-01 -6.96546912e-01 4.02098775e-01
6.70719147e-01 -1.24073327e+00 6.47008002e-01 -7.12198138e-01
5.11789203e-01 -1.18709795e-01 -3.39232013e-02 -1.39943373e+00
-2.65497476e-01 -3.65063876e-01 7.31170848e-02 1.27726126e+00
8.98723185e-01 -8.92336547e-01 8.42458248e-01 4.48796898e-01
3.62384081e-01 -9.04902935e-01 -7.45561004e-01 -7.47913480e-01
4.60679799e-01 -3.35876465e-01 1.01325655e+00 8.10497999e-01
-5.94361842e-01 3.14737111e-01 -4.95544016e-01 7.53887817e-02
4.36527491e-01 5.21771014e-01 5.59597254e-01 -1.33032715e+00
-4.15802181e-01 -7.85048962e-01 -4.01984930e-01 -7.58957386e-01
8.25684443e-02 -5.84499776e-01 -2.50811517e-01 -1.09142649e+00
-2.98138738e-01 -2.69849122e-01 -9.04533684e-01 -8.85151792e-03
-2.29113460e-01 -3.59391868e-01 3.32581758e-01 3.91779095e-01
-7.05993697e-02 8.82725656e-01 9.73428905e-01 -6.40261918e-02
-3.14746797e-01 8.63489211e-01 -4.14000064e-01 7.01990545e-01
1.07382989e+00 -4.21040446e-01 -2.32846811e-01 4.12282087e-02
5.35306931e-01 3.67933691e-01 -4.96316813e-02 -9.32674825e-01
1.68212518e-01 -4.88223471e-02 6.54379547e-01 -9.86503482e-01
5.57247270e-03 -8.14042747e-01 4.77033854e-02 6.96963549e-01
-4.97635990e-01 6.70407176e-01 -5.11336178e-02 4.67471242e-01
-6.74559712e-01 -4.15864855e-01 6.37669444e-01 -2.00128481e-01
-3.11849594e-01 3.63530606e-01 -1.85563356e-01 1.89589914e-02
8.41494501e-01 2.91562766e-01 2.23768335e-02 -6.29426777e-01
-5.70312500e-01 2.38619655e-01 -1.17991388e-01 2.01320916e-01
4.26596671e-01 -1.45824015e+00 -9.81115460e-01 1.20234147e-01
-4.01280999e-01 -3.42939734e-01 5.55917528e-03 6.41850233e-01
-7.72483587e-01 5.84004462e-01 -2.82219350e-01 -7.09297433e-02
-5.42763233e-01 7.01148510e-01 5.05866826e-01 -8.21005285e-01
-3.00061494e-01 5.15504301e-01 2.67847061e-01 2.39962842e-02
4.43440646e-01 -4.27380115e-01 -7.11480618e-01 7.24573195e-01
6.99057817e-01 5.92834830e-01 2.17629418e-01 -3.36340874e-01
5.91398478e-02 5.05091727e-01 -9.42836553e-02 -2.08670899e-01
1.74491727e+00 2.93141812e-01 -9.49439928e-02 9.25851941e-01
1.31599581e+00 7.37157017e-02 -1.32446957e+00 -3.19712967e-01
4.79700029e-01 -2.10429549e-01 8.44363123e-02 -6.21332943e-01
-1.56539166e+00 9.42376435e-01 6.02725625e-01 6.10667884e-01
1.05196357e+00 -5.29040873e-01 9.05980289e-01 2.50368446e-01
1.45143941e-01 -1.03063881e+00 -2.89648980e-01 3.43427300e-01
9.49304521e-01 -1.10552812e+00 -1.44088447e-01 2.57121593e-01
-6.56089544e-01 1.44598484e+00 1.98398679e-01 -6.24430299e-01
1.34755051e+00 1.56266510e-01 3.11493278e-01 2.32347939e-02
-7.48925865e-01 5.93130989e-03 4.28621858e-01 -9.06820446e-02
3.29608560e-01 6.32912144e-02 -3.57839406e-01 9.80211794e-01
-6.01299286e-01 -1.76599830e-01 3.49502981e-01 4.83486831e-01
-4.49444503e-01 -8.21607292e-01 -4.25888509e-01 6.13709629e-01
-1.02741933e+00 -4.65924526e-03 -1.24242306e-01 9.20547843e-01
-1.97483733e-01 5.10521472e-01 3.39832574e-01 -3.57572228e-01
4.50874865e-01 2.89522737e-01 -2.72286296e-01 -4.95529398e-02
-7.40466356e-01 1.47903532e-01 -2.47573465e-01 -1.00680687e-01
-2.32541040e-01 -8.22430134e-01 -8.46222281e-01 -2.76744843e-01
-3.44098181e-01 4.18453723e-01 5.53484559e-01 8.80018294e-01
-1.52229249e-01 4.93895948e-01 1.19044054e+00 -5.59573770e-01
-1.15696120e+00 -1.01662993e+00 -9.81591523e-01 2.15927437e-01
5.38205743e-01 -6.18336976e-01 -7.64868617e-01 -2.10626423e-01] | [4.439486026763916, 4.224885940551758] |
a7d74a7d-a2c0-438e-ad83-850b09ff2297 | few-shot-class-incremental-learning-via-class | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhao_Few-Shot_Class-Incremental_Learning_via_Class-Aware_Bilateral_Distillation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhao_Few-Shot_Class-Incremental_Learning_via_Class-Aware_Bilateral_Distillation_CVPR_2023_paper.pdf | Few-Shot Class-Incremental Learning via Class-Aware Bilateral Distillation | Few-Shot Class-Incremental Learning (FSCIL) aims to continually learn novel classes based on only few training samples, which poses a more challenging task than the well-studied Class-Incremental Learning (CIL) due to data scarcity. While knowledge distillation, a prevailing technique in CIL, can alleviate the catastrophic forgetting of older classes by regularizing outputs between current and previous model, it fails to consider the overfitting risk of novel classes in FSCIL. To adapt the powerful distillation technique for FSCIL, we propose a novel distillation structure, by taking the unique challenge of overfitting into account. Concretely, we draw knowledge from two complementary teachers. One is the model trained on abundant data from base classes that carries rich general knowledge, which can be leveraged for easing the overfitting of current novel classes. The other is the updated model from last incremental session that contains the adapted knowledge of previous novel classes, which is used for alleviating their forgetting. To combine the guidances, an adaptive strategy conditioned on the class-wise semantic similarities is introduced. Besides, for better preserving base class knowledge when accommodating novel concepts, we adopt a two-branch network with an attention-based aggregation module to dynamically merge predictions from two complementary branches. Extensive experiments on 3 popular FSCIL datasets: mini-ImageNet, CIFAR100 and CUB200 validate the effectiveness of our method by surpassing existing works by a significant margin. | ['Xiangzhong Fang', 'Yi Niu', 'Dashan Guo', 'Zhanzhan Cheng', 'Yunlu Xu', 'Jing Lu', 'Linglan Zhao'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['class-incremental-learning', 'few-shot-class-incremental-learning', 'incremental-learning', 'general-knowledge', 'novel-concepts'] | ['computer-vision', 'methodology', 'methodology', 'miscellaneous', 'reasoning'] | [ 3.16978455e-01 2.65790790e-01 -2.48004213e-01 -1.69461682e-01
-1.72748163e-01 -1.29644334e-01 3.42816651e-01 3.51013154e-01
-5.15949488e-01 8.60481977e-01 1.65025353e-01 3.66035402e-02
-2.64028698e-01 -9.27779973e-01 -8.11957836e-01 -8.52455020e-01
2.75572211e-01 1.28737643e-01 7.45230675e-01 -2.89267927e-01
5.73517978e-02 1.03128016e-01 -1.80436575e+00 2.90439427e-01
1.50254881e+00 1.15920925e+00 3.57131958e-01 -7.68366754e-02
-2.72564650e-01 7.87920415e-01 -3.22542042e-01 -4.37217176e-01
5.74453734e-02 -3.90435427e-01 -4.97297496e-01 -2.47256234e-02
4.12683249e-01 -2.82876432e-01 -3.18599641e-01 8.96899581e-01
5.20152092e-01 4.17195976e-01 2.43574768e-01 -1.06613171e+00
-8.88399184e-01 9.14820254e-01 -4.36550170e-01 2.96861023e-01
-8.86806250e-02 1.83649614e-01 6.69673741e-01 -1.19723868e+00
3.95924181e-01 1.11814034e+00 7.34265149e-01 6.71436429e-01
-9.56228614e-01 -9.14861441e-01 9.64058757e-01 5.56353331e-01
-1.36984408e+00 -3.27660710e-01 8.75531554e-01 -9.59580988e-02
6.46597087e-01 -1.55414883e-02 8.13639104e-01 1.06531644e+00
-1.24643713e-01 1.24193823e+00 7.72813261e-01 -4.02473301e-01
3.36076409e-01 4.11681384e-01 4.17895764e-01 5.49608231e-01
3.37515742e-01 -1.61913946e-01 -5.72164297e-01 9.66035500e-02
2.65866369e-01 6.41422570e-01 -4.12738979e-01 -5.30898869e-01
-8.35384429e-01 6.03942752e-01 6.68313026e-01 4.21824813e-01
-2.47095630e-01 -4.90275621e-01 3.74808788e-01 2.85992086e-01
5.34253359e-01 2.40562394e-01 -6.71788216e-01 1.92111418e-01
-7.13652313e-01 1.73890516e-02 1.82151943e-01 9.74471509e-01
9.80898261e-01 1.81272298e-01 -4.69252557e-01 9.43627596e-01
4.34905291e-02 2.28752106e-01 1.06549966e+00 -3.47311974e-01
3.84954840e-01 1.16378140e+00 -3.82433593e-01 -8.49900246e-01
-9.23174322e-02 -1.13271844e+00 -8.74352515e-01 -6.20565303e-02
-1.32356569e-01 4.81264293e-02 -9.92543817e-01 1.89476418e+00
5.23679316e-01 5.89438200e-01 2.49359533e-02 3.13624442e-01
7.60799468e-01 5.14981270e-01 2.00967953e-01 -6.82156801e-01
8.62888396e-01 -1.13464701e+00 -4.79570419e-01 -2.51976788e-01
4.69962955e-01 -1.62572220e-01 1.03875601e+00 4.23523277e-01
-7.37491071e-01 -8.82599473e-01 -1.28273952e+00 7.29737207e-02
-5.15738428e-01 -2.42372796e-01 5.80044150e-01 3.06190848e-01
-6.32380664e-01 7.41687059e-01 -5.15801549e-01 -1.05814546e-01
9.38911855e-01 2.57200263e-02 -7.39399791e-02 -3.64024132e-01
-1.24850416e+00 6.58478677e-01 9.48039889e-01 -9.68917683e-02
-7.67643034e-01 -1.15525448e+00 -6.79452062e-01 3.85876238e-01
7.78223276e-01 -4.79970783e-01 1.09861016e+00 -1.09508252e+00
-1.36407459e+00 2.28388608e-01 3.11723724e-02 -5.46120942e-01
5.13459086e-01 -2.84942001e-01 -4.07541692e-01 -6.82201609e-02
8.70468095e-03 6.11260772e-01 1.09967363e+00 -1.07439792e+00
-9.69615936e-01 -4.32238996e-01 -2.50258204e-02 4.58329678e-01
-1.03216410e+00 -8.07444870e-01 -3.94583046e-01 -8.21992934e-01
1.34903952e-01 -6.21817231e-01 -1.63900107e-01 -7.28290156e-02
-3.02476697e-02 -5.13917625e-01 9.86136854e-01 -3.27908635e-01
1.77744484e+00 -2.21862006e+00 1.12320572e-01 -1.90379754e-01
2.72801697e-01 8.80909801e-01 -2.85859078e-01 4.19400111e-02
-7.98200518e-02 -2.27019191e-01 -3.39524448e-01 -2.69292265e-01
-3.09774995e-01 2.80057877e-01 -5.76948404e-01 -1.84927613e-01
2.53988057e-01 9.36919451e-01 -1.24369240e+00 -2.89512813e-01
8.19297433e-02 2.94844449e-01 -6.13607824e-01 -7.34495558e-03
-3.79906952e-01 2.36163855e-01 -2.96505332e-01 5.39452672e-01
7.42212892e-01 -6.45074472e-02 -9.92272571e-02 -1.50285363e-01
4.11593094e-02 -7.59471729e-02 -1.06170404e+00 1.81835818e+00
-4.51299369e-01 -6.42811321e-03 -4.47405696e-01 -1.20717800e+00
8.21056962e-01 1.51757434e-01 2.50168651e-01 -8.67064178e-01
-1.24413200e-01 1.40118748e-01 -7.20248595e-02 -2.73712575e-01
1.65769950e-01 -2.46676698e-01 1.41508073e-01 2.29822606e-01
4.69768345e-01 3.29083949e-01 2.31571570e-01 4.21824515e-01
7.90042043e-01 1.61628142e-01 2.52829820e-01 -3.98294926e-02
6.28727913e-01 -3.96884024e-01 1.10399020e+00 8.87269080e-01
-3.10931206e-01 2.61421263e-01 3.32273692e-02 -6.11116827e-01
-4.13343996e-01 -1.20274699e+00 -1.15652688e-01 1.30839348e+00
1.84338212e-01 -3.48156542e-01 -4.48007137e-01 -1.13217700e+00
1.84617192e-01 9.86793756e-01 -6.79786861e-01 -1.02519846e+00
-3.63610744e-01 -9.64617670e-01 -4.83908951e-02 7.20762134e-01
6.07521772e-01 -9.53207254e-01 -4.98140663e-01 5.41122496e-01
4.23058793e-02 -6.17381513e-01 -3.93050730e-01 2.14655489e-01
-1.22004104e+00 -1.09157693e+00 -8.50516319e-01 -7.31828392e-01
7.66822815e-01 6.38356090e-01 9.17521477e-01 1.67037696e-01
-2.49380767e-01 3.34004223e-01 -5.51988482e-01 -7.07229972e-01
-2.84897047e-03 2.40186602e-01 3.11938345e-01 3.65292668e-01
4.98436242e-01 -8.85957956e-01 -5.37735343e-01 6.54540658e-02
-9.73057687e-01 2.49099195e-01 7.66457081e-01 1.00884438e+00
6.46475017e-01 4.67286885e-01 1.12021399e+00 -1.14053094e+00
3.15174431e-01 -7.22351015e-01 -7.60089904e-02 4.73109066e-01
-1.02548826e+00 -3.68515030e-02 9.44397986e-01 -8.64771307e-01
-1.50763190e+00 -1.01306342e-01 8.99825692e-02 -6.18053615e-01
1.25170186e-01 5.04113972e-01 -3.19268465e-01 8.22470263e-02
5.19614398e-01 5.55778623e-01 -2.11675137e-01 -6.67907715e-01
5.81343949e-01 4.04605597e-01 6.40537858e-01 -5.03374398e-01
7.89373159e-01 4.09284562e-01 -6.34207666e-01 -5.33724368e-01
-1.45934868e+00 -2.49920994e-01 -6.41141951e-01 -1.95208155e-02
3.10569614e-01 -1.01529372e+00 -3.98067683e-01 7.38981009e-01
-7.18255162e-01 -1.41747102e-01 -7.49143541e-01 3.17465693e-01
-6.84428737e-02 2.81566024e-01 -3.35218906e-01 -6.07188284e-01
-4.21047091e-01 -6.62604988e-01 4.98621643e-01 7.94796944e-01
3.20470154e-01 -8.39739680e-01 4.49366309e-02 1.36485681e-01
6.00733757e-01 -7.56540224e-02 1.09334373e+00 -8.79663706e-01
-4.46687788e-01 -1.72693700e-01 -1.19060285e-01 5.84923625e-01
3.06569129e-01 -3.43409002e-01 -1.13486648e+00 -6.33209467e-01
7.35550374e-02 -5.48347473e-01 1.51376927e+00 -8.15595910e-02
1.41309524e+00 -3.93207341e-01 -3.51126790e-01 3.78585070e-01
1.29586256e+00 3.51625562e-01 3.58263552e-01 1.44231915e-01
7.55494595e-01 2.64252335e-01 5.11032462e-01 4.45964485e-01
5.66702425e-01 3.67384017e-01 3.70483309e-01 1.44651845e-01
-3.66054714e-01 -6.57041430e-01 2.43630141e-01 1.16394782e+00
7.22957179e-02 1.20036215e-01 -5.43299079e-01 6.31347597e-01
-1.99202955e+00 -9.28052008e-01 6.80357456e-01 2.43444061e+00
1.15588558e+00 4.06616986e-01 -2.17808396e-01 2.58781195e-01
5.14715910e-01 1.03056002e-02 -1.05586517e+00 1.45702019e-01
-9.71644595e-02 2.50504524e-01 -1.14483222e-01 1.42478362e-01
-1.01032782e+00 9.20101285e-01 4.48288584e+00 1.26996505e+00
-1.02761555e+00 2.87634552e-01 7.25843370e-01 -3.07058990e-01
-3.14246476e-01 1.24102691e-03 -1.20675588e+00 6.99102700e-01
5.74587643e-01 -3.11246574e-01 1.97914988e-01 9.95829821e-01
-2.75607318e-01 -8.18625242e-02 -9.32887256e-01 7.22608685e-01
3.46445203e-01 -1.24881732e+00 4.15626019e-01 -6.06607795e-01
1.02187860e+00 -1.11652754e-01 1.72300294e-01 1.13847506e+00
1.93818435e-01 -3.48429084e-01 5.77375889e-01 8.13531101e-01
4.93326306e-01 -7.69638956e-01 6.21230423e-01 6.09928727e-01
-1.11677241e+00 -6.35970175e-01 -5.53239882e-01 -1.64660066e-02
-1.53882802e-01 8.32696617e-01 -5.11417150e-01 6.36674166e-01
8.35847199e-01 1.16294944e+00 -9.94107842e-01 1.15856349e+00
-2.90120929e-01 7.04298913e-01 -1.45574436e-01 3.21562588e-01
9.21061486e-02 5.27934991e-02 3.62673700e-01 7.89398789e-01
2.78344274e-01 2.84391016e-01 2.96537846e-01 5.87875903e-01
-1.80729225e-01 1.17426030e-01 -4.95595366e-01 1.52238265e-01
6.71195149e-01 1.01157606e+00 -4.52977836e-01 -6.39883518e-01
-4.89221871e-01 9.55580771e-01 6.06257796e-01 3.26324254e-01
-5.73723614e-01 -5.64060867e-01 2.59627938e-01 6.86589302e-03
5.08847117e-01 2.38953307e-01 -8.84453952e-02 -1.43748927e+00
1.94105849e-01 -5.99400401e-01 7.04057992e-01 -5.34519672e-01
-1.31572354e+00 4.80190456e-01 -4.60922159e-02 -1.21450734e+00
1.12342939e-01 -1.83756232e-01 -9.62866127e-01 4.19255376e-01
-1.81633031e+00 -1.15910399e+00 -3.97711366e-01 5.55132747e-01
7.98407316e-01 -4.17903006e-01 5.17315149e-01 4.83720750e-01
-8.78874421e-01 9.45004046e-01 1.37051046e-01 -2.38764480e-01
6.24002397e-01 -9.12997782e-01 3.41331288e-02 8.41814756e-01
1.51786264e-02 5.99013448e-01 2.00540632e-01 -7.56208241e-01
-1.03505576e+00 -1.51921892e+00 7.24418998e-01 -2.69884646e-01
4.73667920e-01 -2.64825970e-01 -1.46254933e+00 5.59845090e-01
-1.94628075e-01 9.31480080e-02 5.28531194e-01 5.04740104e-02
-6.24940217e-01 -6.54859960e-01 -7.90085375e-01 5.52885354e-01
1.20377409e+00 -2.17067674e-01 -9.00816441e-01 2.02202648e-01
1.22429729e+00 -3.28541622e-02 -5.12230933e-01 6.91536725e-01
3.57270032e-01 -8.99990499e-01 9.84456062e-01 -6.32941425e-01
1.48049504e-01 -2.76438892e-01 1.69197172e-01 -1.50748181e+00
-5.03351331e-01 -3.11983138e-01 -5.03939152e-01 1.46021831e+00
2.15481967e-01 -6.91869438e-01 6.57530904e-01 2.45982751e-01
-3.51653695e-01 -1.03259027e+00 -9.36005473e-01 -1.00631988e+00
1.20475695e-01 -1.66420519e-01 6.12976968e-01 1.05548179e+00
-1.39507964e-01 5.11103332e-01 -3.61763626e-01 -2.07410723e-01
5.25505900e-01 1.49108633e-01 4.50006753e-01 -1.49025464e+00
-2.60638446e-01 -4.16742980e-01 -2.00179130e-01 -8.94635260e-01
4.13520522e-02 -1.07557535e+00 -6.73787668e-02 -1.21016347e+00
4.39640462e-01 -7.01871455e-01 -9.61182058e-01 1.01223755e+00
-7.77741790e-01 4.76337373e-02 3.79016340e-01 4.78746556e-02
-1.04922104e+00 1.19284189e+00 1.16072297e+00 -2.43564725e-01
-5.36427438e-01 1.37842000e-01 -1.13202870e+00 8.33304763e-01
5.97261667e-01 -4.76671904e-01 -9.53593671e-01 -3.57754648e-01
1.23259604e-01 -4.93009329e-01 1.31178930e-01 -1.26889300e+00
4.94409770e-01 1.25835333e-02 4.34800684e-01 -5.77051282e-01
1.62758306e-02 -8.24933469e-01 -2.31012613e-01 5.84897041e-01
-2.77947098e-01 -4.49605823e-01 3.09625119e-01 1.05413902e+00
-1.39242604e-01 -2.52222061e-01 9.55772400e-01 -2.12835401e-01
-9.73354697e-01 6.68938994e-01 9.35336947e-02 1.81614265e-01
1.07053232e+00 -1.74722046e-01 -3.36745054e-01 1.16616420e-01
-7.63718903e-01 5.59534788e-01 1.52098536e-01 8.08137596e-01
7.50650287e-01 -1.49491143e+00 -6.18616939e-01 5.69257259e-01
3.79653305e-01 2.93552011e-01 7.13462293e-01 7.86998391e-01
2.88566232e-01 7.84410238e-02 -1.40700281e-01 -3.28157723e-01
-7.49905109e-01 1.05708885e+00 4.22385558e-02 -3.30323130e-01
-6.38015687e-01 1.07785869e+00 4.74595487e-01 -3.57453197e-01
4.54831183e-01 -1.93568885e-01 -3.28291923e-01 3.60844731e-01
9.34163690e-01 3.43864113e-01 2.78208345e-01 -7.15299100e-02
-1.25376657e-01 2.62237281e-01 -8.21218848e-01 5.89789450e-01
1.43470848e+00 -3.48893076e-01 1.00182705e-01 7.33705401e-01
7.95121491e-01 -2.79008716e-01 -1.52794266e+00 -8.97121668e-01
-6.32545203e-02 -1.41466230e-01 -1.03123160e-02 -9.24166381e-01
-1.12101233e+00 9.35802460e-01 7.56664336e-01 -2.35040367e-01
1.38163579e+00 -2.85882205e-01 1.03521705e+00 4.83543158e-01
4.65855002e-01 -1.26002491e+00 3.50629210e-01 5.92292249e-01
6.42521501e-01 -1.20270455e+00 -3.67433093e-02 -1.47172853e-01
-4.73394781e-01 8.95704806e-01 1.00121737e+00 1.72331795e-01
7.55816698e-01 -3.45817268e-01 -3.19622189e-01 2.83935517e-01
-9.40787554e-01 -1.21283747e-01 2.22630680e-01 4.64150816e-01
-9.46014524e-02 -2.14200333e-01 -3.17027867e-01 1.24023926e+00
3.07391822e-01 1.84842542e-01 2.58296072e-01 9.73408163e-01
-8.62728477e-01 -8.81810129e-01 1.43417329e-01 6.32747710e-01
6.31968379e-02 -3.26552004e-01 -4.09250408e-02 3.76388818e-01
7.76190221e-01 6.30979836e-01 -2.91338190e-02 -4.64566439e-01
4.00639117e-01 3.42310578e-01 2.15579078e-01 -8.51385772e-01
-4.06160086e-01 -2.87608951e-01 -5.81846833e-01 -4.23825204e-01
-2.41365105e-01 -3.95199567e-01 -9.94195044e-01 1.33847341e-01
-4.99584794e-01 1.82042465e-01 -1.51787726e-02 1.08550525e+00
6.20459795e-01 6.74547434e-01 7.27695823e-01 -4.89623368e-01
-6.87073648e-01 -8.68269801e-01 -4.78508592e-01 4.44386065e-01
2.09195256e-01 -9.27092671e-01 -3.49934727e-01 -9.35800597e-02] | [9.81306266784668, 3.3963847160339355] |
22d7f4b0-f19c-471c-a202-a4b98a452a6b | vptr-efficient-transformers-for-video | 2203.15836 | null | https://arxiv.org/abs/2203.15836v1 | https://arxiv.org/pdf/2203.15836v1.pdf | VPTR: Efficient Transformers for Video Prediction | In this paper, we propose a new Transformer block for video future frames prediction based on an efficient local spatial-temporal separation attention mechanism. Based on this new Transformer block, a fully autoregressive video future frames prediction Transformer is proposed. In addition, a non-autoregressive video prediction Transformer is also proposed to increase the inference speed and reduce the accumulated inference errors of its autoregressive counterpart. In order to avoid the prediction of very similar future frames, a contrastive feature loss is applied to maximize the mutual information between predicted and ground-truth future frame features. This work is the first that makes a formal comparison of the two types of attention-based video future frames prediction models over different scenarios. The proposed models reach a performance competitive with more complex state-of-the-art models. The source code is available at \emph{https://github.com/XiYe20/VPTR}. | ['Guillaume-Alexandre Bilodeau', 'Xi Ye'] | 2022-03-29 | null | null | null | null | ['video-prediction'] | ['computer-vision'] | [ 9.26415846e-02 2.18456641e-01 -1.68559954e-01 -4.65775728e-01
-5.29616177e-01 1.01729311e-01 5.65728486e-01 -3.19467336e-01
-2.07024977e-01 5.77403009e-01 2.81553447e-01 -1.42458498e-01
1.33987382e-01 -4.67599481e-01 -7.35024869e-01 -6.07047200e-01
8.38230848e-02 2.73893718e-02 4.66219425e-01 -1.84543803e-02
2.01282233e-01 2.02386707e-01 -1.66103423e+00 5.00721514e-01
7.04144955e-01 1.15806854e+00 7.00377941e-01 7.43072093e-01
4.42217030e-02 1.27375829e+00 -7.40715414e-02 -5.48951805e-01
2.70675987e-01 -4.95243728e-01 -6.08140826e-01 4.25290763e-02
2.59759873e-01 -6.52236700e-01 -8.16609561e-01 9.22278643e-01
3.09196472e-01 4.41665977e-01 3.76891911e-01 -1.34363174e+00
-6.22042239e-01 4.92107749e-01 -4.75046456e-01 6.77855074e-01
4.10055995e-01 1.83083385e-01 1.06595385e+00 -1.00182617e+00
4.27378744e-01 1.10608029e+00 4.44885254e-01 6.89698935e-01
-6.97895825e-01 -8.56643617e-01 5.76387584e-01 9.24175322e-01
-1.25505733e+00 -6.84300303e-01 8.76498997e-01 -2.66898692e-01
1.14936352e+00 1.62172213e-01 7.40509868e-01 9.76265013e-01
4.56000626e-01 1.09214604e+00 4.04607028e-01 -3.41388166e-01
-1.21714477e-03 -1.22645862e-01 -4.70285192e-02 6.68014467e-01
-4.78145361e-01 1.89465180e-01 -3.94810677e-01 2.20814198e-01
4.98015463e-01 3.55520636e-01 -4.98551726e-01 -1.13810733e-01
-1.03602314e+00 5.87141275e-01 2.26248458e-01 3.41321975e-01
-6.55745268e-01 2.07368791e-01 2.97513574e-01 7.25657586e-03
8.13757300e-01 1.59166083e-02 -4.11629170e-01 -4.05666471e-01
-1.02710140e+00 3.78334634e-02 4.34855640e-01 1.08212686e+00
6.66743040e-01 2.63005704e-01 -4.21017855e-01 4.99844998e-01
6.05332732e-01 3.11670393e-01 5.18262088e-01 -1.30704224e+00
5.10869563e-01 1.93033427e-01 4.54562902e-02 -1.14313805e+00
-1.98323317e-02 -3.18476677e-01 -6.45824015e-01 -1.46629229e-01
3.16409133e-02 -1.68361932e-01 -6.47794008e-01 1.46569157e+00
2.41053671e-01 1.01970685e+00 -1.65454447e-01 9.25044060e-01
5.73255837e-01 1.04887402e+00 2.14763969e-01 -4.82273668e-01
9.23329592e-01 -1.27622426e+00 -7.91472912e-01 -8.70505944e-02
2.46461436e-01 -7.33918548e-01 6.65836990e-01 2.89195329e-01
-1.27298892e+00 -8.91428769e-01 -8.19904387e-01 -2.43388548e-01
1.18582264e-01 2.20613196e-01 4.15206552e-01 2.01614857e-01
-1.09172595e+00 5.36745012e-01 -1.12229133e+00 -2.77537137e-01
4.10026461e-01 2.03579366e-01 -9.74114761e-02 -2.23329291e-02
-1.16451073e+00 6.57610118e-01 2.60776013e-01 2.61402160e-01
-8.40507507e-01 -6.96619332e-01 -8.18445086e-01 3.54082942e-01
3.51846933e-01 -5.31954169e-01 1.26104069e+00 -1.27317464e+00
-1.56477833e+00 4.47239548e-01 -6.86362982e-01 -8.03971708e-01
4.22278583e-01 -4.85374689e-01 -3.44181240e-01 2.89988101e-01
-1.02499900e-02 7.03672290e-01 1.02176905e+00 -8.18445861e-01
-8.51144612e-01 -1.82756692e-01 2.16850773e-01 3.12949121e-01
-4.45772648e-01 5.78736886e-02 -8.22937667e-01 -8.48478556e-01
-2.14554742e-01 -9.09853220e-01 -1.30163774e-01 -6.61611110e-02
3.46455798e-02 -1.98927402e-01 1.00410795e+00 -9.00159061e-01
1.49360120e+00 -2.27264190e+00 2.06722632e-01 -2.67780721e-01
2.78554469e-01 3.02501708e-01 -1.19345859e-01 -1.12795243e-02
-2.16806844e-01 -5.40539734e-02 1.36037603e-01 -5.22496700e-01
-1.82181835e-01 7.05639049e-02 -5.22153556e-01 3.01304519e-01
1.94419414e-01 7.22401440e-01 -8.68521273e-01 -4.26102132e-01
6.01466835e-01 7.10784316e-01 -6.02772355e-01 3.63330394e-01
-3.10804158e-01 4.18358684e-01 -4.84694034e-01 2.44785473e-01
6.61316097e-01 -3.12814474e-01 1.68842703e-01 -2.50716805e-01
-2.24815547e-01 1.43326789e-01 -6.56256795e-01 1.49947464e+00
-3.53916615e-01 7.44457364e-01 -3.98858517e-01 -1.14052677e+00
7.52278328e-01 5.87056935e-01 7.29341567e-01 -5.56052446e-01
2.56120771e-01 -2.41086572e-01 -1.86050907e-01 -5.35118461e-01
6.28936231e-01 5.87516874e-02 3.84427309e-01 4.59154556e-03
1.58457577e-01 4.38139439e-01 2.53441215e-01 1.52703643e-01
8.50719333e-01 6.22503459e-01 3.31507564e-01 -6.36584237e-02
9.13034558e-01 -3.52093518e-01 1.10868609e+00 4.72740948e-01
-4.90436584e-01 7.20399439e-01 3.09084177e-01 -6.51407838e-01
-1.06472027e+00 -9.22917604e-01 2.73668051e-01 1.03693521e+00
2.66786963e-01 -6.03565395e-01 -5.49124837e-01 -6.58398509e-01
-3.80919576e-01 8.92374337e-01 -5.40458739e-01 -1.73289388e-01
-5.71056008e-01 -3.29267979e-01 5.88681251e-02 6.77831352e-01
5.87557733e-01 -9.26592946e-01 -6.77610576e-01 1.96209550e-01
-6.21361494e-01 -1.24463010e+00 -5.20954728e-01 -4.42323387e-01
-8.60245883e-01 -9.03552949e-01 -9.32587802e-01 -5.38478792e-01
2.91010857e-01 3.56796354e-01 9.35368121e-01 1.45543516e-01
2.46286675e-01 4.37710792e-01 -6.23172939e-01 -2.77443230e-01
-2.03247190e-01 -1.55115768e-01 -1.50887119e-02 2.76679724e-01
6.63137436e-01 -4.83182281e-01 -7.30877280e-01 1.37746096e-01
-5.38438261e-01 4.43518430e-01 2.50896484e-01 9.00174201e-01
6.89391077e-01 1.05216196e-02 2.48545900e-01 -3.68855059e-01
-5.04318252e-02 -7.11251199e-01 -6.60120666e-01 3.17802221e-01
-3.96171004e-01 -6.31787926e-02 8.08682919e-01 -3.61283541e-01
-1.50320709e+00 4.82609347e-02 -2.90897131e-01 -1.00510764e+00
-5.01070134e-02 2.67094165e-01 5.80124818e-02 3.85549814e-01
-1.02745434e-02 3.89726907e-01 -2.05196157e-01 -4.62538689e-01
4.36641648e-02 3.94801885e-01 3.11766684e-01 -4.15546112e-02
2.86615103e-01 4.62717950e-01 -2.36410618e-01 -7.18047023e-01
-8.23576570e-01 -3.84661794e-01 -5.04507780e-01 -4.72956717e-01
1.06521726e+00 -1.05191481e+00 -5.42166948e-01 4.17246103e-01
-1.33858609e+00 -3.98089409e-01 -1.67599767e-01 8.10008943e-01
-8.68211091e-01 6.28410220e-01 -8.32598448e-01 -9.49528217e-01
-3.00521344e-01 -1.24211287e+00 9.24513459e-01 2.05470502e-01
-7.21563026e-02 -8.84754896e-01 -1.86756849e-01 3.45306188e-01
1.87893033e-01 -1.05960362e-01 3.54586691e-01 -4.33116466e-01
-8.20499778e-01 5.54010160e-02 -1.72725171e-01 3.68614942e-01
-4.88418750e-02 2.52851218e-01 -8.67759764e-01 -1.83772177e-01
1.76701263e-01 2.49567345e-01 9.94759023e-01 7.13015139e-01
1.21695769e+00 -2.79106408e-01 -3.41932863e-01 6.63709462e-01
1.16365325e+00 6.58520579e-01 1.02879798e+00 2.17129365e-01
6.29543066e-01 3.24606925e-01 9.12203372e-01 8.02850842e-01
6.52016580e-01 7.66183972e-01 3.84869933e-01 2.39660427e-01
-1.16026051e-01 -1.61115587e-01 5.58570385e-01 7.26379395e-01
-3.61489683e-01 -6.54398501e-01 -7.99119055e-01 5.20628273e-01
-2.10244274e+00 -1.43757617e+00 -1.13835394e-01 2.03591013e+00
1.93300515e-01 -9.91016347e-03 -1.10712752e-01 -4.27840278e-02
8.63587797e-01 3.76441300e-01 -3.79400939e-01 -2.23792329e-01
1.37787864e-01 -9.41225290e-02 8.18137527e-02 6.14791989e-01
-1.39788258e+00 1.03790164e+00 5.67346668e+00 8.73714328e-01
-1.08638966e+00 2.33757958e-01 8.53994131e-01 -1.20682038e-01
-1.22046627e-01 5.66631183e-03 -8.19035769e-01 8.43833208e-01
1.07368743e+00 -2.24987894e-01 3.14585894e-01 7.76465118e-01
5.20471811e-01 1.90181527e-02 -7.43896484e-01 1.02457023e+00
2.10919842e-01 -1.37531376e+00 1.64562538e-01 -2.60370433e-01
4.70998198e-01 1.69864029e-01 6.70565292e-02 2.79904038e-01
-4.66953292e-02 -4.14638549e-01 8.86058509e-01 1.03193021e+00
4.10855114e-01 -7.16208100e-01 7.64333904e-01 1.64733678e-01
-1.51170075e+00 -5.19388974e-01 -2.46474937e-01 -1.88394159e-01
5.37465751e-01 4.11533326e-01 -3.03292364e-01 5.94747663e-01
9.99318302e-01 1.23972094e+00 -4.86978918e-01 9.54297066e-01
-2.12309957e-01 6.62704766e-01 -4.05524075e-02 2.71310717e-01
6.03721887e-02 -1.19132832e-01 7.72786856e-01 1.05606794e+00
6.88940227e-01 3.03363830e-01 9.99948010e-04 5.34652710e-01
2.20137298e-01 -7.40287006e-02 -4.49834198e-01 2.66758978e-01
2.99728602e-01 8.66067708e-01 -4.53601271e-01 -6.47263467e-01
-8.20707738e-01 1.19044018e+00 2.09583372e-01 4.63872880e-01
-1.37458026e+00 3.95995192e-02 6.41971290e-01 1.21916778e-01
6.31526649e-01 -6.58714920e-02 9.96257961e-02 -1.47689033e+00
1.91090003e-01 -5.64778149e-01 4.70388025e-01 -1.05718541e+00
-1.01045632e+00 7.62666821e-01 -1.33909537e-02 -1.53723073e+00
-5.92679620e-01 -2.95076400e-01 -5.68065822e-01 5.42070150e-01
-1.51072180e+00 -1.08113086e+00 -2.26905331e-01 6.17811263e-01
9.99225438e-01 -2.89075166e-01 4.75392789e-01 4.37800258e-01
-5.70798099e-01 4.99132544e-01 1.06795859e-02 -4.53780442e-02
3.97788167e-01 -8.18168819e-01 3.84082764e-01 1.19513583e+00
-6.33431077e-02 2.55808115e-01 7.20712960e-01 -6.58201754e-01
-1.08943248e+00 -1.17425406e+00 1.02710307e+00 -1.85823455e-01
5.57271004e-01 -4.72328477e-02 -8.81773770e-01 9.72973049e-01
2.00027257e-01 2.89804414e-02 4.24227059e-01 -3.91878843e-01
-4.64916006e-02 -2.57963270e-01 -8.03234935e-01 6.47670567e-01
9.04522061e-01 -2.85007238e-01 -4.04646873e-01 1.59442917e-01
8.27305436e-01 -1.92984775e-01 -8.45019042e-01 5.62069774e-01
7.23098755e-01 -1.20273519e+00 9.53134418e-01 -2.97331035e-01
5.16509295e-01 -2.94219196e-01 -1.75026655e-01 -6.37508392e-01
-7.10709631e-01 -5.78757823e-01 -6.91485882e-01 1.07499814e+00
3.36836427e-01 -4.04544383e-01 8.90587449e-01 7.28286684e-01
-2.05843315e-01 -8.24750423e-01 -9.38795686e-01 -6.99728549e-01
-2.43595287e-01 -6.07276857e-01 2.98769295e-01 5.82815766e-01
-4.82512861e-02 1.73709050e-01 -7.62171924e-01 1.75319135e-01
3.48952353e-01 8.23970325e-03 5.03521979e-01 -7.41880536e-01
-4.99481112e-01 -1.86440378e-01 -6.50915504e-01 -1.49225724e+00
3.96905720e-01 -5.64739823e-01 -7.35597592e-03 -1.32933700e+00
2.94756025e-01 3.86044234e-02 -4.62710768e-01 2.76946336e-01
-2.28106588e-01 1.57494068e-01 6.90105259e-01 1.14218637e-01
-8.37847412e-01 9.58212912e-01 9.94291067e-01 4.34095822e-02
-1.56143442e-01 2.21918538e-01 -2.84207702e-01 8.65178168e-01
1.02390397e+00 -3.05436820e-01 -6.00631297e-01 -6.23716056e-01
-2.58022904e-01 3.67097348e-01 4.78878379e-01 -1.17062843e+00
3.35686415e-01 -1.42125160e-01 3.42295200e-01 -7.27041841e-01
7.16272414e-01 -8.37920666e-01 3.34743470e-01 3.84269029e-01
-2.76023030e-01 2.07520276e-01 7.82487020e-02 7.44355798e-01
-3.24981928e-01 -9.06746369e-03 6.59262300e-01 -1.00605555e-01
-1.07000911e+00 5.59046686e-01 -6.11787677e-01 -3.40745389e-01
1.25528848e+00 -2.80925214e-01 -1.04074664e-01 -7.00918734e-01
-8.07596028e-01 2.05339581e-01 2.60686636e-01 6.21875525e-01
9.58158255e-01 -1.23505771e+00 -6.81555212e-01 5.79945557e-02
-1.48319393e-01 -5.46504259e-01 8.60422909e-01 1.05390310e+00
-2.74403721e-01 4.45435226e-01 -1.66190371e-01 -5.10898411e-01
-1.42657459e+00 9.16260123e-01 2.77852982e-01 -3.41418982e-01
-7.15748131e-01 7.74392903e-01 4.77139860e-01 1.46224394e-01
5.78016602e-03 -1.08843304e-01 -3.95175040e-01 -2.40784436e-01
8.20047200e-01 4.90644068e-01 -4.04809564e-01 -1.02808630e+00
-3.05627346e-01 3.93782735e-01 -6.98167980e-02 2.65192259e-02
1.27261961e+00 -6.31573856e-01 1.50455385e-01 3.16161662e-01
1.09326708e+00 -3.62254143e-01 -1.76988363e+00 -1.19274437e-01
-1.31151229e-01 -7.12257624e-01 1.66576058e-01 -2.42190540e-01
-1.38077736e+00 9.96075928e-01 6.45009279e-01 -4.49568182e-02
1.59772825e+00 -1.30563527e-01 8.91487718e-01 2.17336863e-02
2.31359541e-01 -8.75431836e-01 -1.25522390e-01 7.20905125e-01
9.10219610e-01 -1.17881989e+00 -4.81282733e-02 -4.08816099e-01
-7.04687655e-01 1.09641457e+00 7.79116154e-01 -1.60155714e-01
8.18771005e-01 1.12952814e-01 -3.47779185e-01 2.34208018e-01
-1.10058737e+00 -2.07330436e-01 4.06149983e-01 4.65245903e-01
5.38964689e-01 -2.78360516e-01 -2.96889395e-01 4.69047785e-01
2.22162724e-01 4.06751722e-01 3.77056152e-01 4.81570184e-01
-1.76848784e-01 -8.16659749e-01 -8.52943435e-02 3.47962707e-01
-6.93032384e-01 -3.08848977e-01 2.30720833e-01 4.91961360e-01
6.45440593e-02 1.07550824e+00 3.91015083e-01 -6.30133748e-01
-7.56427795e-02 -2.30308138e-02 3.50398988e-01 -2.29986906e-01
-1.26185253e-01 2.76504993e-01 -4.94856574e-02 -8.42304349e-01
-6.97846293e-01 -6.92666829e-01 -1.03632700e+00 -3.22921127e-01
-2.25781187e-01 -4.23025154e-02 2.87627224e-02 8.69970024e-01
5.99303067e-01 5.84509671e-01 6.11575603e-01 -1.08015907e+00
-1.95120741e-02 -8.02087784e-01 -1.93933964e-01 2.61992455e-01
2.49039635e-01 -6.30403340e-01 -2.35184908e-01 4.24683958e-01] | [8.787407875061035, 0.2459997534751892] |
f003a3ff-390f-49d4-bd72-40a37f1b5d0a | keyphrase-generation-a-text-summarization-1 | null | null | https://aclanthology.org/N19-1070 | https://aclanthology.org/N19-1070.pdf | Keyphrase Generation: A Text Summarization Struggle | Authors{'} keyphrases assigned to scientific articles are essential for recognizing content and topic aspects. Most of the proposed supervised and unsupervised methods for keyphrase generation are unable to produce terms that are valuable but do not appear in the text. In this paper, we explore the possibility of considering the keyphrase string as an abstractive summary of the title and the abstract. First, we collect, process and release a large dataset of scientific paper metadata that contains 2.2 million records. Then we experiment with popular text summarization neural architectures. Despite using advanced deep learning models, large quantities of data and many days of computation, our systematic evaluation on four test datasets reveals that the explored text summarization methods could not produce better keyphrases than the simpler unsupervised methods, or the existing supervised ones. | ['Ond{\\v{r}}ej Bojar', 'Erion {\\c{C}}ano'] | 2019-06-01 | null | null | null | naacl-2019-6 | ['keyphrase-generation'] | ['natural-language-processing'] | [ 2.68751174e-01 3.62022340e-01 -1.96112171e-01 1.70770347e-01
-8.64916205e-01 -7.41216540e-01 1.08493364e+00 9.78551090e-01
-4.21296597e-01 1.19145751e+00 8.73137534e-01 -2.93285608e-01
-3.04752409e-01 -7.62936056e-01 -8.52926016e-01 -5.38738966e-01
-9.66483578e-02 3.35962385e-01 -2.07414269e-01 1.56289443e-01
1.05455506e+00 4.36344355e-01 -1.64872921e+00 3.71598870e-01
1.09803343e+00 7.33473003e-01 1.58519328e-01 8.80573750e-01
-8.41142833e-01 8.09040546e-01 -1.24594378e+00 -2.37203836e-01
-1.29127786e-01 -4.95850980e-01 -9.00659144e-01 -1.56560004e-01
1.01994824e+00 -3.23194772e-01 -3.85505170e-01 1.05283105e+00
5.33491194e-01 7.04277605e-02 9.10757124e-01 -9.50701058e-01
-9.33510602e-01 1.39985478e+00 -5.95514655e-01 4.57367867e-01
4.17263687e-01 -3.43874872e-01 1.22170091e+00 -5.53640783e-01
9.73561525e-01 9.30179060e-01 5.46344042e-01 2.04732329e-01
-8.66060019e-01 -3.75854939e-01 -1.31215706e-01 -8.15179870e-02
-8.49303663e-01 -4.48934168e-01 9.17668343e-01 -2.91604131e-01
9.61393416e-01 5.70202947e-01 5.55746794e-01 1.25173104e+00
5.55541337e-01 8.62586856e-01 8.46937895e-01 -3.85309517e-01
3.80002528e-01 -8.51100832e-02 6.88432932e-01 3.95117879e-01
1.03777254e+00 -5.79514742e-01 -7.00929999e-01 -3.96754593e-01
1.13919444e-01 -4.68679294e-02 -1.23238765e-01 3.41191351e-01
-1.60644531e+00 6.38053000e-01 -1.10714898e-01 3.59203279e-01
-8.34117055e-01 8.41112956e-02 6.08900309e-01 3.14479589e-01
5.80857992e-01 1.08856571e+00 -3.03132623e-01 -4.31816280e-02
-1.51607943e+00 6.98863328e-01 1.21574998e+00 1.04880488e+00
5.05046248e-01 1.46290988e-01 -7.29887128e-01 4.44065750e-01
-2.01896623e-01 3.37420017e-01 6.68642640e-01 -8.75184119e-01
5.17406285e-01 8.04153919e-01 1.60451755e-01 -1.17438185e+00
-3.82165551e-01 -5.78565478e-01 -1.09417009e+00 -4.25734729e-01
-4.29528356e-02 -3.82579088e-01 -9.75653112e-01 9.77507293e-01
-2.79704392e-01 -4.11134511e-01 3.04845601e-01 1.11951157e-01
1.68956792e+00 1.09245408e+00 2.63429480e-03 -6.83302581e-01
1.35012925e+00 -8.19739223e-01 -1.14525926e+00 2.24372953e-01
2.89266944e-01 -8.68559957e-01 5.08304715e-01 7.44088888e-01
-1.31767333e+00 -4.87525523e-01 -1.07534552e+00 -2.98224270e-01
-7.84705460e-01 4.80606824e-01 6.57350302e-01 1.78299665e-01
-1.17848933e+00 8.98119330e-01 -2.81157851e-01 -5.39186239e-01
4.88847554e-01 -8.86854976e-02 -2.15642571e-01 4.79885399e-01
-9.35932934e-01 5.53835273e-01 8.45090210e-01 -2.50158817e-01
-4.90016192e-01 -8.86643291e-01 -4.40973431e-01 4.98778880e-01
2.40983933e-01 -7.39477038e-01 1.24789596e+00 -3.98963302e-01
-1.32016325e+00 7.77907491e-01 1.49080306e-02 -7.65012383e-01
3.25219840e-01 -6.03593171e-01 -1.49369016e-01 4.60822642e-01
-1.11618228e-02 6.17972136e-01 7.10094571e-01 -1.10538697e+00
-8.52089047e-01 -1.41574845e-01 -2.85642296e-01 -1.21714540e-01
-6.76680863e-01 -2.28277482e-02 -1.11212060e-01 -1.14434242e+00
-7.94924702e-03 -3.70132864e-01 -3.33822742e-02 -7.55524039e-01
-9.16935146e-01 -6.78168595e-01 5.13125122e-01 -1.01565373e+00
1.56012845e+00 -1.43842435e+00 2.98906386e-01 -1.46048546e-01
3.84106576e-01 4.38739248e-02 1.21637993e-01 1.00802231e+00
1.51904896e-01 5.09811938e-01 -1.18874609e-01 -1.55909389e-01
7.19006360e-02 -1.65868521e-01 -1.00387430e+00 7.90138319e-02
-1.70620650e-01 8.84820342e-01 -8.24593842e-01 -6.78013325e-01
-4.04407442e-01 -1.43695906e-01 -2.25746378e-01 2.79608041e-01
-5.74005902e-01 -7.80557282e-03 -6.81309044e-01 6.81705832e-01
5.51517487e-01 -2.05059364e-01 -1.61092892e-01 -3.19671988e-01
-3.95264596e-01 5.87531209e-01 -8.04985702e-01 1.70230854e+00
1.02584794e-01 1.09970510e+00 -4.28110182e-01 -1.05604494e+00
8.98652315e-01 2.85721183e-01 7.35769153e-01 -4.00252461e-01
2.15133116e-01 2.62445807e-01 -3.94965500e-01 -5.18894970e-01
1.23873019e+00 5.49319029e-01 -2.16031060e-01 7.25232244e-01
2.53638029e-01 -1.68986127e-01 7.23209858e-01 6.65982366e-01
1.04253876e+00 -1.31686460e-02 3.48321766e-01 -5.10714948e-01
4.14295346e-01 1.48014143e-01 7.04546273e-02 1.34058738e+00
4.76785570e-01 6.99033141e-01 8.52244675e-01 -4.27962512e-01
-1.22601426e+00 -7.17507005e-01 2.37734672e-02 8.28391075e-01
-3.36262196e-01 -6.84902072e-01 -8.78639281e-01 -4.98412937e-01
9.68689919e-02 9.74691510e-01 -4.67841923e-01 -9.34375972e-02
-5.12094259e-01 -8.08298051e-01 8.63867640e-01 1.68987647e-01
3.10837626e-01 -1.26590753e+00 -7.19298601e-01 2.87411302e-01
-1.76871002e-01 -7.34998286e-01 -2.14774653e-01 1.83809966e-01
-9.53900158e-01 -6.09164417e-01 -1.16862309e+00 -6.67440772e-01
6.29898190e-01 7.97469765e-02 1.22360015e+00 -2.47789845e-01
-1.49708271e-01 3.68132591e-01 -5.48099339e-01 -7.53000617e-01
-6.56542182e-01 6.71050251e-01 1.31309964e-03 -3.60563546e-01
2.53500670e-01 -7.39260435e-01 -3.31308335e-01 -1.00216925e+00
-1.13021159e+00 1.05468661e-01 9.57969248e-01 6.74416006e-01
2.54731804e-01 -5.90250976e-02 8.74194145e-01 -1.03132856e+00
1.31219232e+00 -4.85743165e-01 -3.32577199e-01 3.24203700e-01
-9.29860353e-01 5.14785409e-01 7.17744470e-01 -4.08459455e-01
-9.64903533e-01 -5.87274075e-01 1.88846752e-01 1.79953128e-02
-2.33690232e-01 8.96646917e-01 2.26212710e-01 6.93117142e-01
6.25470161e-01 8.12821329e-01 -3.28131020e-01 -1.02456725e+00
5.17371714e-01 8.64322841e-01 6.44052088e-01 -6.49038374e-01
5.08942366e-01 1.96522653e-01 -1.30493373e-01 -1.02142549e+00
-8.78822923e-01 -4.14781988e-01 -5.31721175e-01 7.34552369e-02
6.10607564e-01 -7.29241371e-01 -5.48566997e-01 4.48962927e-01
-1.72143590e+00 5.63739836e-01 -4.72054958e-01 2.91952670e-01
-2.31809467e-01 7.47839570e-01 -5.67106247e-01 -3.97727370e-01
-1.24517417e+00 -5.64987719e-01 9.82324660e-01 6.37249529e-01
-6.23564184e-01 -4.56423551e-01 1.39758363e-01 1.30342051e-01
3.63809973e-01 2.90782869e-01 1.02304184e+00 -1.30067933e+00
-2.73812354e-01 -3.83260489e-01 -1.43071011e-01 1.44094929e-01
3.40211183e-01 4.82850939e-01 -7.57758021e-01 -6.92725778e-02
-1.48787394e-01 -1.92768663e-01 1.46307707e+00 4.65353191e-01
1.53670585e+00 -1.07465482e+00 -2.07990035e-01 4.15781766e-01
1.03024352e+00 2.42353722e-01 5.23815870e-01 5.26504338e-01
8.26066792e-01 7.58942187e-01 1.20543346e-01 5.41536331e-01
3.05725634e-01 -1.07834719e-01 -1.01619117e-01 2.33567834e-01
5.92847914e-02 -2.56783128e-01 2.15718895e-01 1.25109100e+00
-2.03141257e-01 -6.28070354e-01 -8.56057644e-01 7.40805149e-01
-1.72180068e+00 -1.30816710e+00 -2.77470946e-01 1.75259709e+00
1.31358385e+00 2.69829035e-01 -1.33236393e-01 -1.69823468e-02
3.28199178e-01 5.81076145e-01 -2.24047080e-01 -6.52851284e-01
-3.55671167e-01 2.06478029e-01 6.17718160e-01 1.34562880e-01
-1.13994753e+00 8.61500263e-01 6.69114399e+00 7.49524593e-01
-9.47142303e-01 -3.69741201e-01 5.52681386e-01 -3.94048318e-02
-5.87135315e-01 -1.13273367e-01 -8.95673633e-01 4.71388757e-01
1.05219424e+00 -7.53608584e-01 -8.60857815e-02 8.09596241e-01
1.83437705e-01 -3.06716859e-01 -1.14399040e+00 7.97377527e-01
3.72984439e-01 -2.00884724e+00 8.88230205e-01 -7.74715245e-02
1.10274470e+00 -2.75698274e-01 -2.16498137e-01 4.58733290e-02
3.07286382e-01 -9.13252771e-01 6.33126080e-01 1.10451376e+00
2.61429757e-01 -6.48979843e-01 7.91536808e-01 2.35694066e-01
-2.83830523e-01 1.29768625e-01 -5.75033784e-01 -8.20557252e-02
-3.54403645e-01 8.56833994e-01 -4.76791203e-01 7.78659821e-01
7.21359611e-01 8.54214251e-01 -8.76294613e-01 9.83711541e-01
1.83348916e-02 7.94808745e-01 -3.68619263e-02 -5.54404140e-01
3.04980010e-01 -1.69989228e-01 7.71322668e-01 1.64311683e+00
4.18274373e-01 -2.17440967e-02 -1.01315752e-01 9.20162082e-01
-6.04000509e-01 2.84667820e-01 -7.03823388e-01 -8.25773478e-01
3.68209243e-01 1.33941519e+00 -1.02875030e+00 -9.51279342e-01
-4.37386287e-03 7.85385132e-01 4.29110928e-03 4.70618159e-01
-2.78395504e-01 -8.44039619e-01 2.55218726e-02 -1.64371103e-01
1.56797349e-01 -1.75165936e-01 -6.30174994e-01 -1.21958470e+00
1.05551325e-01 -1.00775993e+00 3.22752863e-01 -8.73746753e-01
-1.13177419e+00 5.16659141e-01 1.77223250e-01 -9.92619514e-01
-3.36656392e-01 -2.89525121e-01 -8.85080695e-01 5.43567240e-01
-1.01994383e+00 -9.61132109e-01 -1.47413880e-01 -2.41918728e-01
8.23411524e-01 -6.69730306e-01 7.09418356e-01 -2.12770477e-01
-4.35537189e-01 2.30970412e-01 6.52542710e-01 4.98799309e-02
6.64765239e-01 -1.48591638e+00 4.39923406e-01 8.02482307e-01
1.78800479e-01 1.03405786e+00 1.21537423e+00 -9.54097927e-01
-1.71859503e+00 -6.84938610e-01 1.17136049e+00 -3.74249011e-01
7.29012907e-01 -1.43145949e-01 -1.00724566e+00 3.45059186e-01
1.22214091e+00 -8.44456136e-01 4.99172837e-01 -1.37745321e-01
-6.67274594e-02 -1.49546325e-01 -6.47286594e-01 7.38103509e-01
6.86340153e-01 -9.60973725e-02 -1.43953001e+00 4.03053522e-01
1.12542892e+00 -8.39282274e-02 -8.05988252e-01 1.63477942e-01
4.40717876e-01 -3.58397424e-01 8.58696282e-01 -8.49583447e-01
1.12753677e+00 1.29726887e-01 3.17699492e-01 -1.31861997e+00
-1.92304328e-01 -8.23473334e-01 -7.74718702e-01 1.67022812e+00
2.97583908e-01 -3.14060390e-01 5.38558483e-01 3.64168696e-02
-1.35029152e-01 -5.29469669e-01 -5.79637468e-01 -2.54428655e-01
4.05307055e-01 2.04426467e-01 4.65885639e-01 9.80364203e-01
1.02057539e-01 5.20559430e-01 -3.11697394e-01 -4.70055640e-01
7.86583722e-01 4.18556541e-01 7.40841269e-01 -1.60721588e+00
3.22998881e-01 -1.00477624e+00 8.19976255e-02 -6.31732583e-01
5.18212244e-02 -8.50011766e-01 -1.68006420e-01 -2.21893930e+00
4.19369191e-01 2.85347134e-01 -3.39654028e-01 3.29247057e-01
-3.75574142e-01 -5.00362873e-01 -3.36246789e-01 4.53033715e-01
-5.82319260e-01 5.50411463e-01 9.47534442e-01 -7.12647557e-01
-2.28231549e-01 -2.33745456e-01 -1.19255590e+00 6.16147637e-01
8.26099992e-01 -5.51043272e-01 -1.24208964e-01 -1.71803579e-01
5.02731860e-01 -2.13509381e-01 1.56539306e-01 -8.25702250e-01
6.27306044e-01 -3.80289555e-01 5.37985682e-01 -1.40191519e+00
-4.71844047e-01 -1.59758925e-01 -2.01368913e-01 2.97615618e-01
-9.92071629e-01 2.55866170e-01 2.87550598e-01 3.83796185e-01
-2.00947151e-01 -4.30948347e-01 1.38106793e-01 -4.29169625e-01
-3.55783165e-01 -1.05746649e-01 -8.38281989e-01 4.56418172e-02
3.19903344e-01 8.23509991e-02 -5.53918302e-01 -1.84335515e-01
-7.82452598e-02 1.17331408e-01 1.46522492e-01 5.76939762e-01
5.84040880e-01 -9.60026920e-01 -1.21530950e+00 -2.86831915e-01
-4.87390794e-02 9.33235288e-02 -6.63087070e-02 4.64550048e-01
-8.94732654e-01 9.05762553e-01 -3.33333731e-01 2.01869663e-02
-8.69482458e-01 7.23337054e-01 -3.78688872e-01 -4.66957718e-01
-9.31952655e-01 3.66262197e-01 -2.39803925e-01 -8.30254108e-02
5.77618837e-01 -7.49903798e-01 -7.85550237e-01 8.46687198e-01
6.53465509e-01 5.77542722e-01 9.16217268e-02 -1.86241314e-01
1.20997854e-01 1.95894435e-01 -4.75052476e-01 -3.20579903e-03
1.69293392e+00 5.35434410e-02 -6.36281431e-01 6.54958308e-01
1.12687564e+00 3.20990413e-01 -3.96179527e-01 -1.82971984e-01
5.86009681e-01 2.93785691e-01 2.70646453e-01 -6.58854127e-01
-6.31242871e-01 6.78966463e-01 1.61301419e-01 6.52431548e-01
8.93302679e-01 -1.07306905e-01 8.19880843e-01 1.07182372e+00
-2.51374245e-01 -1.38914180e+00 -3.83755304e-02 5.45305610e-01
1.09705102e+00 -9.84950244e-01 6.66295290e-01 3.25836271e-01
-2.55714476e-01 1.36675930e+00 2.57636726e-01 -7.84988627e-02
1.70809790e-01 4.78942208e-02 -1.70366272e-01 -5.44702888e-01
-7.86645055e-01 2.04313397e-01 5.91194630e-01 2.88238320e-02
6.41664505e-01 -1.56824872e-01 -1.00075710e+00 8.32151949e-01
-6.02170348e-01 -1.69403329e-02 1.06246316e+00 9.09498453e-01
-7.10622489e-01 -4.85809833e-01 -4.64427888e-01 1.01833701e+00
-1.05221546e+00 -2.05376193e-01 -8.86746347e-01 3.60820651e-01
-3.54782641e-01 6.65299952e-01 8.56042653e-02 -4.88095991e-02
7.77783170e-02 1.96875021e-01 8.54702294e-02 -5.30347347e-01
-6.94272339e-01 6.00235201e-02 2.25187428e-02 2.04908222e-01
-4.26375806e-01 -6.13561928e-01 -1.09094524e+00 -3.74677688e-01
9.26198363e-02 5.67464113e-01 6.86308682e-01 8.00266862e-01
4.74548876e-01 9.25298989e-01 4.94206190e-01 -8.67689550e-01
-6.41618729e-01 -1.48034024e+00 -2.65249968e-01 2.72920400e-01
4.74970937e-01 -5.20812348e-02 -2.84102529e-01 5.25761187e-01] | [12.485389709472656, 9.380666732788086] |
87003252-bcad-4ff8-9dab-0ce2c975557e | generalization-bounds-for-magnitude-based | 2305.18789 | null | https://arxiv.org/abs/2305.18789v2 | https://arxiv.org/pdf/2305.18789v2.pdf | Generalization Bounds for Magnitude-Based Pruning via Sparse Matrix Sketching | In this paper, we derive a novel bound on the generalization error of Magnitude-Based pruning of overparameterized neural networks. Our work builds on the bounds in Arora et al. [2018] where the error depends on one, the approximation induced by pruning, and two, the number of parameters in the pruned model, and improves upon standard norm-based generalization bounds. The pruned estimates obtained using our new Magnitude-Based compression algorithm are close to the unpruned functions with high probability, which improves the first criteria. Using Sparse Matrix Sketching, the space of the pruned matrices can be efficiently represented in the space of dense matrices of much smaller dimensions, thereby lowering the second criterion. This leads to stronger generalization bound than many state-of-the-art methods, thereby breaking new ground in the algorithm development for pruning and bounding generalization error of overparameterized models. Beyond this, we extend our results to obtain generalization bound for Iterative Pruning [Frankle and Carbin, 2018]. We empirically verify the success of this new method on ReLU-activated Feed Forward Networks on the MNIST and CIFAR10 datasets. | ['Xiaoming Huo', 'Prasanjit Dubey', 'Etash Kumar Guha'] | 2023-05-30 | null | null | null | null | ['generalization-bounds'] | ['methodology'] | [ 2.79676855e-01 3.41951519e-01 -2.37284943e-01 -2.63146251e-01
-3.33188593e-01 -4.89564955e-01 6.32481873e-02 1.74642280e-01
-7.52603292e-01 8.14457893e-01 -1.16753325e-01 -4.99865770e-01
-6.51308239e-01 -7.67019451e-01 -1.02884042e+00 -7.59409547e-01
-3.26519340e-01 4.30709720e-01 2.93890357e-01 2.26357318e-02
1.47963122e-01 6.54848039e-01 -1.51297104e+00 2.85778761e-01
5.19254565e-01 1.24611115e+00 -1.38214216e-01 6.04491413e-01
3.04697037e-01 3.87232721e-01 -4.32557762e-01 -5.94075143e-01
5.77509999e-01 -5.95063530e-02 -8.22029591e-01 -3.21065128e-01
8.57420802e-01 -3.56050372e-01 -5.71277738e-01 1.22270381e+00
4.21035029e-02 3.71078432e-01 5.20465910e-01 -1.03458881e+00
-5.92534065e-01 1.51152396e+00 -4.74974930e-01 4.07118291e-01
-4.54344839e-01 -4.44553226e-01 1.13666904e+00 -9.24443483e-01
3.45730811e-01 1.11670518e+00 9.98828888e-01 5.45504510e-01
-1.42161369e+00 -9.55203652e-01 5.11400640e-01 2.43223175e-01
-1.48060405e+00 -3.90921205e-01 3.80025744e-01 -3.01860869e-01
1.12072277e+00 2.56117642e-01 8.24245036e-01 9.51189935e-01
-4.88682002e-01 6.36304677e-01 4.42067474e-01 -6.11427426e-01
4.02884960e-01 2.25017563e-01 7.15916872e-01 6.70540631e-01
1.04529381e+00 1.60369009e-01 -4.03637886e-01 -4.78122383e-01
7.64050663e-01 1.18950300e-01 -5.38830936e-01 -5.36631644e-01
-7.52672493e-01 1.04486716e+00 3.66920382e-01 2.32340798e-01
-1.16834193e-01 3.61745536e-01 3.56930107e-01 3.62956792e-01
3.90964270e-01 5.79369724e-01 -7.48907208e-01 -1.71697542e-01
-1.12086499e+00 3.49262178e-01 9.85049546e-01 9.57746208e-01
6.36502743e-01 1.27909124e-01 5.31879365e-02 1.01256454e+00
4.50285673e-02 2.56606013e-01 4.18197721e-01 -1.02089429e+00
6.14357948e-01 3.72216254e-01 -1.52491435e-01 -7.92892933e-01
-2.93676376e-01 -8.00245702e-01 -9.11256969e-01 -1.11554876e-01
6.40412092e-01 -2.59096593e-01 -7.73875594e-01 2.17340684e+00
6.06238469e-02 1.06370069e-01 -5.12231365e-02 5.03622353e-01
2.34473675e-01 5.55526137e-01 -1.63447365e-01 -2.41063073e-01
1.00633693e+00 -9.23349619e-01 -3.84911418e-01 -1.33681968e-01
9.23267365e-01 -4.17996310e-02 9.74937022e-01 8.85714710e-01
-1.11833477e+00 -1.63383365e-01 -1.40089369e+00 1.08472437e-01
-2.19108030e-01 2.70943522e-01 1.07060468e+00 9.40170109e-01
-1.07941484e+00 1.06126487e+00 -9.11743045e-01 -1.28204778e-01
8.55555415e-01 5.15709102e-01 -2.57226020e-01 -5.96591569e-02
-1.05170846e+00 7.91828990e-01 7.74809003e-01 6.20961525e-02
-5.36831856e-01 -1.11461699e+00 -4.27876323e-01 4.70475495e-01
3.35969299e-01 -4.95819598e-01 1.19174719e+00 -4.51315939e-01
-1.24824107e+00 3.69488001e-01 1.07557818e-01 -1.28637910e+00
3.49480420e-01 -3.47920626e-01 4.46233042e-02 3.53471100e-01
-7.20068276e-01 6.00383461e-01 6.76627278e-01 -8.44931066e-01
-6.56540573e-01 -4.54170942e-01 1.64230704e-01 -1.09902568e-01
-7.95029044e-01 -4.01449353e-01 -4.36684377e-02 -6.28756046e-01
2.60260820e-01 -8.71404052e-01 -2.28471830e-01 -4.32487354e-02
-2.25711003e-01 -2.42123365e-01 3.41628194e-01 -2.64831901e-01
1.45545411e+00 -2.14423060e+00 1.99261725e-01 4.55807894e-01
5.63152373e-01 3.63863498e-01 -1.68195009e-01 2.54489720e-01
-3.20398510e-01 3.65727931e-01 -2.92098016e-01 -3.25433165e-01
1.45555690e-01 2.85778224e-01 -8.37096632e-01 5.24486125e-01
-4.39904071e-02 4.76469696e-01 -5.37942946e-01 -7.77014196e-02
-3.61445963e-01 3.98151308e-01 -8.40239942e-01 -3.37262988e-01
-1.41884694e-02 -4.87217933e-01 -1.76161304e-02 1.43024713e-01
7.19455183e-01 -1.92860916e-01 -3.35367434e-02 -2.45084122e-01
3.23796183e-01 3.77044559e-01 -1.34562325e+00 1.31427658e+00
-3.77922922e-01 6.22044623e-01 -1.62034258e-02 -1.21289086e+00
6.95494652e-01 9.45673659e-02 2.61206537e-01 1.29013611e-02
1.12016521e-01 2.70562857e-01 1.80347890e-01 5.86179420e-02
2.09722504e-01 -8.03235397e-02 4.44982737e-01 4.55745876e-01
2.92093396e-01 1.72865316e-01 6.34418547e-01 2.53296286e-01
1.17951262e+00 -2.77763098e-01 1.22682542e-01 -3.99409413e-01
1.31072089e-01 -3.61436576e-01 3.95800978e-01 1.08125436e+00
1.00051798e-01 1.60130158e-01 7.69616485e-01 -3.52461010e-01
-1.11155546e+00 -1.07899785e+00 -6.87730610e-01 1.24479258e+00
-5.00722706e-01 -6.79629922e-01 -8.91174734e-01 -5.38735151e-01
2.61681020e-01 7.57694662e-01 -1.04234362e+00 -4.69025195e-01
-5.16093254e-01 -8.51994872e-01 9.11103487e-01 8.13955486e-01
1.13372810e-01 -4.39437509e-01 -5.90613484e-01 -1.11878933e-02
2.10999161e-01 -1.11715353e+00 -1.29851475e-01 7.20449567e-01
-1.46220863e+00 -1.05291843e+00 -5.61880291e-01 -5.50122261e-01
6.47683561e-01 4.56512906e-02 7.67254710e-01 1.33751720e-01
-7.02702329e-02 -3.55099328e-02 -1.47907704e-01 -5.66402376e-01
-2.64988929e-01 3.88722420e-01 4.19106513e-01 -3.54454219e-01
3.49809319e-01 -1.06548965e+00 -3.32389534e-01 9.25667807e-02
-8.50344062e-01 -4.49847504e-02 6.60406291e-01 1.03531563e+00
6.74033582e-01 1.75561056e-01 5.27623415e-01 -9.21666443e-01
5.85390389e-01 -3.42512786e-01 -8.83194149e-01 1.50864869e-01
-8.11352491e-01 4.29223508e-01 9.19909894e-01 -1.12785041e+00
-3.68252754e-01 4.21846472e-02 -1.28464699e-02 -7.32870877e-01
2.47733146e-01 4.04118866e-01 2.13871896e-01 -2.51575768e-01
9.30366278e-01 1.10337079e-01 -3.19381803e-01 -6.41215801e-01
4.32963461e-01 2.82370299e-01 4.36093748e-01 -7.44667113e-01
7.21910655e-01 3.00971717e-01 1.61262557e-01 -6.51899517e-01
-1.15031087e+00 -5.44409193e-02 -4.37417686e-01 3.91165197e-01
8.44252929e-02 -6.51935697e-01 -7.01919615e-01 -7.71662395e-04
-9.82588589e-01 -3.87084424e-01 -7.07257569e-01 6.87033117e-01
-4.86587226e-01 3.11519384e-01 -8.64681304e-01 -1.01557016e+00
-2.95941710e-01 -7.72887945e-01 4.52511549e-01 -1.50010630e-01
-1.67931527e-01 -6.47929907e-01 2.57539703e-03 -1.10708058e-01
3.36330026e-01 -1.97376758e-01 1.30328655e+00 -1.13566029e+00
-2.48599052e-01 -2.65357256e-01 -3.07981074e-01 6.05147958e-01
-5.51795006e-01 -1.40554160e-01 -9.07146871e-01 -3.70890915e-01
3.25376868e-01 -3.07991713e-01 1.47788835e+00 5.00284314e-01
1.50925565e+00 -8.20351362e-01 -3.21122140e-01 1.06603158e+00
1.29796088e+00 -2.06181780e-01 3.90516996e-01 2.44293675e-01
4.93172199e-01 3.12942028e-01 1.58795267e-01 5.24160206e-01
-2.27983877e-01 2.43875265e-01 3.32240820e-01 3.71553570e-01
1.56802148e-01 -2.35288382e-01 1.40155654e-03 7.41034985e-01
-1.13099270e-01 5.05920686e-03 -6.42402530e-01 4.86562371e-01
-1.82749772e+00 -9.23940301e-01 1.85796306e-01 2.39408302e+00
1.18589568e+00 3.87987018e-01 1.80246949e-01 6.08075440e-01
5.47807753e-01 -1.76513478e-01 -7.15202212e-01 -5.68245232e-01
-1.44310370e-01 5.23316205e-01 8.39409649e-01 5.74883640e-01
-9.30883110e-01 7.07956970e-01 6.85247469e+00 1.11602092e+00
-7.05706656e-01 1.44443661e-01 6.42674804e-01 -8.11655700e-01
-1.08128145e-01 -1.05669305e-01 -1.29131663e+00 1.19768798e-01
1.35173500e+00 -2.03527048e-01 9.04291570e-01 1.27806306e+00
-4.54842925e-01 1.14961535e-01 -1.56617701e+00 1.10739732e+00
2.28929259e-02 -1.39867508e+00 4.06117178e-02 1.94861934e-01
7.17376590e-01 2.52578974e-01 1.82825372e-01 5.82343280e-01
2.67658293e-01 -1.20700812e+00 7.70794094e-01 8.56921300e-02
6.14529669e-01 -9.03446615e-01 4.82612014e-01 3.69029850e-01
-8.23533595e-01 -5.39541006e-01 -1.08217633e+00 -2.39268374e-02
-2.50071079e-01 9.12068963e-01 -5.92525005e-01 -1.27462819e-01
7.12562442e-01 2.56495386e-01 -5.65394819e-01 1.11595285e+00
-7.23245963e-02 1.03465140e+00 -1.02779090e+00 -3.16691279e-01
1.38767615e-01 -1.08952858e-01 4.54222798e-01 1.17446887e+00
2.11496517e-01 6.16553016e-02 -3.45298111e-01 9.95587230e-01
-4.89615649e-01 5.44622121e-03 -3.99327070e-01 -1.64296880e-01
7.61504054e-01 1.03459442e+00 -5.30744433e-01 -2.48529643e-01
1.50274718e-03 2.94572175e-01 9.64658558e-01 4.77473199e-01
-8.83920372e-01 -7.00755119e-01 4.71884727e-01 7.81189874e-02
8.94706011e-01 -1.82473630e-01 -5.43267012e-01 -9.49839413e-01
2.86035061e-01 -8.66465569e-01 3.91876101e-01 -1.12362288e-01
-1.09872866e+00 6.51590586e-01 3.93204451e-01 -6.93120241e-01
-8.78560394e-02 -9.01363552e-01 -1.27862826e-01 6.60181880e-01
-9.71650183e-01 -4.88975435e-01 2.58850485e-01 1.68957338e-01
1.30025625e-01 -1.63589448e-01 8.40246975e-01 2.79830724e-01
-7.80617297e-01 1.11277926e+00 3.50932300e-01 -8.80755559e-02
2.30129421e-01 -9.09532964e-01 2.23778620e-01 9.31106925e-01
3.74675661e-01 1.02157962e+00 6.29302979e-01 -3.36590558e-01
-1.20372069e+00 -8.82759392e-01 4.67587501e-01 -1.71555161e-01
8.39455783e-01 -5.39121449e-01 -1.11876202e+00 8.40896249e-01
-4.02349353e-01 -1.84208546e-02 6.72666609e-01 5.81460416e-01
-8.39402258e-01 -2.73965269e-01 -1.09022570e+00 6.93644524e-01
1.28654087e+00 -1.47508100e-01 -5.71695030e-01 4.20773804e-01
9.04428959e-01 -3.00462723e-01 -9.21642900e-01 5.08696496e-01
6.56089365e-01 -8.57580602e-01 9.43388820e-01 -1.04881895e+00
2.46773720e-01 3.61389279e-01 -4.44680601e-01 -1.13803434e+00
-3.84503573e-01 -6.38951182e-01 -1.06392264e+00 8.01986873e-01
7.53789306e-01 -7.83699095e-01 9.31857347e-01 5.55055559e-01
-6.95126876e-02 -1.48719347e+00 -1.11301053e+00 -1.15262556e+00
5.10695159e-01 -6.17739737e-01 7.27387011e-01 5.46235144e-01
2.74243087e-01 2.49763653e-01 5.66047132e-02 2.49859076e-02
5.57634354e-01 -2.70118117e-01 2.81382710e-01 -1.38286817e+00
-7.96250761e-01 -1.01566768e+00 -3.42429429e-01 -1.31295979e+00
4.38365862e-02 -9.49269593e-01 -2.65019298e-01 -9.56892133e-01
2.65225828e-01 -5.58883905e-01 -4.19628322e-01 6.96723521e-01
2.18532979e-01 1.52256817e-01 1.25401527e-01 1.74219102e-01
-3.53186220e-01 4.97914910e-01 6.85129762e-01 9.96726528e-02
-9.01695266e-02 -1.65704377e-02 -9.14330542e-01 8.37830663e-01
9.15454924e-01 -7.55614758e-01 -5.88339269e-01 -4.22865421e-01
5.44840276e-01 -4.34200734e-01 1.83584124e-01 -1.23623502e+00
1.74564958e-01 1.00597851e-01 3.39037597e-01 -6.41200483e-01
4.48560059e-01 -7.39039421e-01 -2.80214369e-01 5.83856285e-01
-8.53382826e-01 2.97993096e-03 5.26510119e-01 7.63226926e-01
3.34796488e-01 -6.45779014e-01 1.00162697e+00 2.69819677e-01
1.74778834e-01 4.54303086e-01 -8.27937573e-02 3.09733987e-01
6.54369414e-01 -3.40377688e-01 -4.69659358e-01 -1.51136965e-01
-6.83755338e-01 -1.00026235e-01 3.89741883e-02 -1.14449888e-01
5.97828329e-01 -1.20398998e+00 -5.77974260e-01 2.67997414e-01
-2.52333373e-01 8.98844078e-02 3.26195583e-02 1.07539284e+00
-3.53069097e-01 5.52361310e-01 5.73588796e-02 -4.00886029e-01
-1.02817380e+00 6.47258759e-01 3.07572395e-01 -3.47729832e-01
-7.18454838e-01 1.31101906e+00 7.74616525e-02 -5.72161097e-03
8.43186796e-01 -8.17992926e-01 1.66682735e-01 -3.06700349e-01
7.95974851e-01 6.13294184e-01 3.39057505e-01 1.69768259e-01
-2.14231461e-01 8.98238644e-02 -4.60729331e-01 -1.47335693e-01
1.53623796e+00 2.79905140e-01 2.22881045e-02 4.42129195e-01
1.40458703e+00 -4.10497561e-03 -1.14700067e+00 -2.74213672e-01
-1.63724303e-01 -2.19710842e-01 1.09070510e-01 -6.41982436e-01
-1.20489824e+00 9.26433623e-01 4.55107242e-01 3.60355794e-01
1.09206152e+00 -1.26282096e-01 6.71720445e-01 1.29472661e+00
1.96657822e-01 -1.04309106e+00 -3.70472819e-01 7.18834400e-01
9.07006025e-01 -5.84310234e-01 2.60054827e-01 -3.97197753e-01
-1.59511283e-01 1.21384394e+00 4.44790304e-01 -3.90564024e-01
7.88844228e-01 5.12560189e-01 -6.61797643e-01 1.85722977e-01
-1.10119128e+00 3.38162243e-01 2.64307678e-01 5.66052079e-01
1.21598907e-01 -1.77179024e-01 -2.86499023e-01 1.11705613e+00
-7.20678687e-01 -9.90538150e-02 3.19176048e-01 5.02557814e-01
-7.57970393e-01 -6.48691654e-01 -1.98347911e-01 8.30555260e-01
-4.58217800e-01 -5.56493580e-01 -2.89188325e-01 8.29431891e-01
-1.71772674e-01 6.28784299e-01 5.73269650e-02 -4.86507207e-01
8.10116827e-02 2.18773484e-01 9.41073239e-01 -4.67610866e-01
-3.42686743e-01 -4.53384489e-01 1.50654867e-01 -3.68105680e-01
2.91071981e-01 -3.83931518e-01 -1.07301760e+00 -3.20303380e-01
-6.93037510e-01 2.47398823e-01 6.10828340e-01 8.54349971e-01
4.00574535e-01 1.75020859e-01 2.49580413e-01 -6.67115211e-01
-1.44896352e+00 -1.16854894e+00 -6.09920859e-01 1.54267222e-01
1.24684431e-01 -7.04432487e-01 -8.00054789e-01 -3.64116848e-01] | [8.396909713745117, 3.372965097427368] |
8f317bd5-02cc-457b-82f3-c24ef20388af | towards-overcoming-false-positives-in-visual | 2012.12510 | null | https://arxiv.org/abs/2012.12510v2 | https://arxiv.org/pdf/2012.12510v2.pdf | Towards Overcoming False Positives in Visual Relationship Detection | In this paper, we investigate the cause of the high false positive rate in Visual Relationship Detection (VRD). We observe that during training, the relationship proposal distribution is highly imbalanced: most of the negative relationship proposals are easy to identify, e.g., the inaccurate object detection, which leads to the under-fitting of low-frequency difficult proposals. This paper presents Spatially-Aware Balanced negative pRoposal sAmpling (SABRA), a robust VRD framework that alleviates the influence of false positives. To effectively optimize the model under imbalanced distribution, SABRA adopts Balanced Negative Proposal Sampling (BNPS) strategy for mini-batch sampling. BNPS divides proposals into 5 well defined sub-classes and generates a balanced training distribution according to the inverse frequency. BNPS gives an easier optimization landscape and significantly reduces the number of false positives. To further resolve the low-frequency challenging false positive proposals with high spatial ambiguity, we improve the spatial modeling ability of SABRA on two aspects: a simple and efficient multi-head heterogeneous graph attention network (MH-GAT) that models the global spatial interactions of objects, and a spatial mask decoder that learns the local spatial configuration. SABRA outperforms SOTA methods by a large margin on two human-object interaction (HOI) datasets and one general VRD dataset. | ['Xianglong Liu', 'Hongsheng Li', 'Zhoujun Li', 'Shuai Yi', 'Haiyu Zhao', 'Mingyuan Zhang', 'Jiashu Tao', 'Yizhuo Zhou', 'Chongzhi Zhang', 'Xiao Ma', 'Daisheng Jin'] | 2020-12-23 | null | null | null | null | ['visual-relationship-detection'] | ['computer-vision'] | [-1.47705674e-01 1.03436805e-01 -2.69436598e-01 -2.52020746e-01
-5.05631745e-01 -2.99651265e-01 3.37937921e-01 -5.61460368e-02
-2.91486770e-01 5.34584701e-01 3.14118788e-02 -2.82857060e-01
-7.36147165e-02 -7.51389980e-01 -7.36553967e-01 -6.63196385e-01
1.08835876e-01 7.93855965e-01 8.62994432e-01 -1.75260007e-01
1.79085001e-01 5.97810805e-01 -1.58065856e+00 2.72108078e-01
9.90822136e-01 7.99744248e-01 7.12826073e-01 4.38235849e-01
-1.44295380e-01 6.25454903e-01 -9.68828321e-01 -1.49951607e-01
1.33746222e-01 -2.19541758e-01 -4.48554337e-01 -2.49685809e-01
4.73637760e-01 -4.14717853e-01 -3.28856438e-01 1.03221917e+00
9.63515580e-01 1.61827892e-01 9.10562932e-01 -1.48827279e+00
-8.13794911e-01 3.92910242e-01 -1.37427330e+00 5.42143524e-01
5.06474823e-02 2.99878806e-01 9.46355700e-01 -1.16576767e+00
6.34079218e-01 1.62316847e+00 3.43036294e-01 2.86161155e-01
-1.15338445e+00 -7.95781136e-01 4.63033706e-01 3.66924584e-01
-1.98705018e+00 -1.06292721e-02 8.71159971e-01 -4.94236320e-01
1.00827503e+00 3.28062654e-01 8.06577146e-01 1.03650451e+00
4.28740345e-02 9.31209445e-01 5.33026099e-01 -2.03565717e-01
2.23675266e-01 1.23349473e-01 1.93915516e-01 4.55563098e-01
7.37090766e-01 9.72521231e-02 -6.39380813e-01 -6.10223599e-02
1.14192224e+00 -1.59563407e-01 -2.58951515e-01 -7.38458514e-01
-9.36764598e-01 5.67631423e-01 7.66098678e-01 2.37706482e-01
-1.92937106e-01 -8.52618292e-02 4.53817956e-02 -2.41738603e-01
2.39899740e-01 3.16770047e-01 -2.16328770e-01 3.27298313e-01
-4.79061157e-01 3.87980461e-01 1.66405544e-01 1.11986518e+00
5.85627019e-01 -7.15248194e-03 -6.94785118e-01 1.14729989e+00
5.12219369e-01 5.49349666e-01 2.93877214e-01 -4.57094222e-01
6.55472636e-01 8.35870504e-01 1.37045622e-01 -1.54586732e+00
-6.01397395e-01 -1.08337736e+00 -9.26689148e-01 -4.12742887e-03
2.41663024e-01 4.23120201e-01 -9.75652039e-01 1.86759388e+00
5.53398252e-01 -2.68221140e-01 -4.46879715e-01 1.45969450e+00
1.19380450e+00 5.86864650e-01 1.46326527e-01 -3.63288261e-02
1.26537800e+00 -1.04742670e+00 -8.09146464e-01 -5.85020006e-01
5.34657419e-01 -4.63241637e-01 1.19941282e+00 1.05015002e-01
-8.92926872e-01 -6.47413015e-01 -1.06672275e+00 -1.62153602e-01
-2.83646852e-01 4.33435977e-01 6.72532022e-01 2.81713098e-01
-1.02434635e+00 -4.94746417e-02 -2.57527292e-01 -5.15053831e-02
7.83612013e-01 4.37376440e-01 -1.54673025e-01 -1.81396827e-01
-9.87537086e-01 7.70449638e-01 3.99874598e-01 2.50385910e-01
-7.22329140e-01 -5.70618689e-01 -8.48693609e-01 7.31045678e-02
6.57429576e-01 -6.25179172e-01 8.27996910e-01 -7.32481182e-01
-8.74863982e-01 8.03309798e-01 -2.74105787e-01 -1.02469162e-03
4.38069284e-01 6.84012324e-02 -1.24695718e-01 -1.38991416e-01
3.69474322e-01 1.04227364e+00 7.48360217e-01 -1.57407284e+00
-2.82968730e-01 -5.97812235e-01 -1.89069912e-01 6.07330203e-01
-1.32700682e-01 -3.10934275e-01 -9.35644507e-01 -7.74159074e-01
4.68138903e-01 -6.94868445e-01 -8.35844651e-02 5.02726361e-02
-7.09343076e-01 -4.38984483e-01 5.46255350e-01 -3.38037014e-01
1.43384755e+00 -2.23786211e+00 2.47795880e-01 2.35591382e-01
6.28524959e-01 2.94599593e-01 -3.99355561e-01 -1.36991143e-01
-1.54960493e-03 1.21371716e-01 -2.07586102e-02 -2.66986698e-01
-1.41487956e-01 1.57236710e-01 -2.06873223e-01 2.69179791e-01
3.20171714e-01 1.05291879e+00 -1.00136924e+00 -7.01940119e-01
1.67340323e-01 3.81561607e-01 -5.17025232e-01 3.23301554e-01
-6.63630292e-02 2.54570037e-01 -3.33080769e-01 8.03977966e-01
1.23229384e+00 -6.08401418e-01 1.27695024e-01 -5.65247655e-01
1.43524662e-01 1.83153600e-01 -1.36633074e+00 1.15669394e+00
7.40373060e-02 4.48997080e-01 -1.10970765e-01 -5.87641299e-01
9.57228005e-01 -2.95799732e-01 7.78965950e-02 -1.10175633e+00
1.37380838e-01 2.83291321e-02 1.79561406e-01 -3.70674819e-01
6.02771997e-01 3.73635590e-01 1.76119581e-01 1.63141623e-01
-7.55483359e-02 1.94724232e-01 -7.18182772e-02 4.18432802e-01
8.42367828e-01 -1.20680071e-02 5.12443602e-01 -2.05545932e-01
1.17088690e-01 -2.54229665e-01 5.82825422e-01 9.84005690e-01
-4.11823034e-01 8.92819881e-01 5.98271072e-01 -3.62297356e-01
-5.95894217e-01 -1.22648156e+00 -2.64325142e-02 9.85301793e-01
8.43267322e-01 -2.30908513e-01 -2.66434997e-01 -7.51140893e-01
7.24274516e-02 6.50415480e-01 -6.56498015e-01 -4.91783053e-01
-4.92502689e-01 -9.92672324e-01 7.41384923e-02 6.16442144e-01
3.86142552e-01 -1.10283339e+00 -3.75112891e-01 -1.23362891e-01
-2.48747796e-01 -9.56112504e-01 -5.65059662e-01 1.02013700e-01
-5.05245209e-01 -1.03902638e+00 -7.60004520e-01 -7.50944734e-01
9.26545918e-01 8.88264537e-01 1.35298359e+00 1.64927363e-01
-3.19167495e-01 -1.57579109e-01 -2.00807050e-01 -4.50382829e-01
1.06559590e-01 1.56717733e-01 -8.71505663e-02 -1.70493647e-01
3.89275372e-01 -2.35096350e-01 -6.60746455e-01 6.59744799e-01
-4.81244177e-01 3.19285959e-01 8.93351138e-01 7.96797156e-01
8.13211501e-01 -5.82428724e-02 5.64997494e-01 -5.65771341e-01
3.57421815e-01 -4.93044227e-01 -5.85702479e-01 1.60176456e-01
-2.58256286e-01 -7.75069967e-02 2.44708136e-01 -7.40167439e-01
-8.93008828e-01 -1.62253052e-01 1.08354382e-01 -6.07065082e-01
1.27128154e-01 -2.95367874e-02 -3.45591456e-01 -2.21920028e-01
8.04524481e-01 8.75698477e-02 -2.96187341e-01 -2.96104968e-01
1.13833174e-01 3.56815666e-01 1.52690992e-01 -2.23676085e-01
6.88161075e-01 1.57956541e-01 -6.32128343e-02 -8.59888315e-01
-7.35044301e-01 -5.41210949e-01 -2.10664198e-01 -3.08323473e-01
7.56233990e-01 -9.44431901e-01 -6.53460622e-01 3.92649978e-01
-1.33486140e+00 -5.09704411e-01 -1.06080443e-01 2.94592679e-01
-2.03679558e-02 2.02550739e-01 -2.71246284e-01 -1.07082725e+00
-7.05390796e-02 -1.15264869e+00 1.42800963e+00 2.94804841e-01
-2.20207766e-01 -4.64091629e-01 -2.72114456e-01 2.01415837e-01
2.89823413e-01 -1.95871979e-01 9.50103641e-01 -4.10660118e-01
-1.11730719e+00 1.52967781e-01 -8.14454496e-01 -1.72272906e-01
-1.92905590e-01 -1.19602732e-01 -9.58899558e-01 -1.72596097e-01
-3.85078758e-01 -1.43704310e-01 8.26099992e-01 7.03196943e-01
1.39990354e+00 -3.37209739e-02 -8.38961959e-01 4.46267575e-01
1.19801188e+00 1.89629510e-01 5.81080079e-01 5.60704954e-02
1.11494350e+00 5.59085488e-01 8.55222821e-01 2.61470944e-01
6.22321308e-01 1.20953190e+00 8.14647973e-01 -4.60516244e-01
-3.47659081e-01 -2.92540133e-01 -1.74169749e-01 3.27406526e-01
-9.10990871e-03 -6.68760002e-01 -7.79396892e-01 6.30619943e-01
-1.90126717e+00 -6.99840248e-01 -3.01534504e-01 2.23435831e+00
6.30746305e-01 2.55309284e-01 2.57563710e-01 -1.07366666e-01
9.05157566e-01 2.84788668e-01 -4.74106401e-01 2.52257735e-01
-3.58612061e-01 -3.36476654e-01 4.82748389e-01 5.54037511e-01
-8.80447567e-01 8.27720761e-01 5.66601276e+00 1.20285857e+00
-8.91179264e-01 8.76352414e-02 8.29918623e-01 -2.01626062e-01
-4.30115074e-01 -4.16108459e-01 -1.04830289e+00 5.68132758e-01
4.44520861e-02 5.62018454e-01 3.14753026e-01 9.57154036e-01
9.72588137e-02 -2.95709550e-01 -7.04591334e-01 1.36701655e+00
2.33468235e-01 -1.24757707e+00 3.04378211e-01 1.26511946e-01
3.88526052e-01 -6.92981482e-02 2.38571316e-01 3.53895783e-01
1.14809409e-01 -8.76575291e-01 1.00339413e+00 2.38671646e-01
6.71953201e-01 -5.71755230e-01 7.80625343e-01 4.29162085e-01
-1.27268922e+00 -1.36761278e-01 -6.09720051e-01 2.12203860e-01
8.17517042e-02 6.98485672e-01 -9.46438491e-01 2.85642177e-01
9.52369034e-01 4.43025768e-01 -9.61724877e-01 1.16482484e+00
-1.80622026e-01 1.96343020e-01 -4.46759850e-01 -2.83083290e-01
-2.36972556e-01 -6.92395344e-02 8.70573640e-01 8.02893817e-01
1.15198836e-01 1.00799635e-01 -1.16102761e-02 1.18666041e+00
1.03936754e-01 -1.22627616e-02 -6.90717161e-01 4.10685360e-01
8.14660549e-01 1.07942140e+00 -8.17078233e-01 -1.66675359e-01
-8.30549449e-02 7.69234598e-01 7.99914420e-01 5.81089854e-01
-1.06358707e+00 6.58293366e-02 2.18629256e-01 3.09660941e-01
4.41769212e-01 5.42242406e-03 -5.54763258e-01 -9.24867332e-01
2.91888624e-01 -5.68190515e-01 3.97866905e-01 -1.00425363e+00
-1.19102371e+00 5.96647382e-01 1.54008254e-01 -1.17543936e+00
1.50636643e-01 -3.97534788e-01 -3.08644950e-01 7.61193991e-01
-1.38043559e+00 -9.97155070e-01 -5.32581389e-01 8.19678679e-02
4.75195438e-01 1.32128328e-01 2.72696048e-01 5.44616222e-01
-8.59191775e-01 9.15885210e-01 -6.68486059e-01 -1.73300803e-01
6.22063160e-01 -1.06219542e+00 3.37172240e-01 6.33732200e-01
-2.18189377e-02 5.94946802e-01 4.74807441e-01 -8.66508484e-01
-1.03135455e+00 -1.22332704e+00 7.98697948e-01 -4.52845573e-01
-7.43075386e-02 -6.02352321e-01 -1.08270097e+00 3.55250865e-01
-3.98321062e-01 2.48045832e-01 3.54595929e-01 -1.75302569e-02
-2.49990791e-01 -1.72654927e-01 -1.07142556e+00 8.67820859e-01
1.61522055e+00 -1.76031157e-01 -1.63942263e-01 3.86412203e-01
7.71771610e-01 -6.03929043e-01 -1.32812575e-01 6.04475021e-01
2.97239363e-01 -1.18680131e+00 1.26623070e+00 1.61256958e-02
-1.50041366e-02 -6.70731783e-01 2.72475015e-02 -9.20550883e-01
-7.02746093e-01 -1.89995602e-01 -4.12162960e-01 1.02338588e+00
4.13915277e-01 -6.23161554e-01 7.03578532e-01 2.05457747e-01
-2.32802704e-01 -1.02738822e+00 -9.37649906e-01 -6.35193586e-01
-6.32897496e-01 -2.72823185e-01 5.94590068e-01 7.96322882e-01
-5.80722570e-01 6.88817203e-01 -3.06598127e-01 3.04087460e-01
5.39170265e-01 -8.42802972e-02 7.66761363e-01 -1.05046153e+00
-5.28926313e-01 -5.83988130e-01 -4.32011038e-01 -1.47704470e+00
-3.64402860e-01 -6.25740111e-01 7.66507238e-02 -1.54079080e+00
3.50002587e-01 -6.49057508e-01 -1.94668137e-02 3.65046501e-01
-4.37264442e-01 5.08194149e-01 -5.67169785e-02 1.90953180e-01
-9.36586559e-01 7.22815990e-01 1.56077576e+00 -2.07868949e-01
-4.12243009e-01 -5.19796349e-02 -7.00071871e-01 5.47773123e-01
4.76056457e-01 -4.64537084e-01 -6.35297120e-01 -4.80972052e-01
4.12207484e-01 -1.66211531e-01 6.35132611e-01 -7.20768392e-01
2.48430967e-02 -1.75936148e-01 6.92962468e-01 -1.28320396e+00
4.28961605e-01 -7.43446887e-01 5.85407466e-02 2.91000485e-01
-8.73972476e-02 -3.25346030e-02 2.62547225e-01 6.76020026e-01
1.01912677e-01 5.95334172e-02 6.65987730e-01 8.68047848e-02
-6.64200783e-01 2.60700345e-01 -1.13054693e-01 1.47966787e-01
1.07062757e+00 -4.68166023e-01 -6.25834882e-01 -2.24544808e-01
-2.64807671e-01 4.45243627e-01 4.37390149e-01 4.49360102e-01
8.06742013e-01 -1.53680027e+00 -3.18539441e-01 4.83950645e-01
2.74769217e-01 6.33351564e-01 6.60321712e-01 9.08737361e-01
-2.50584990e-01 1.70620501e-01 1.50826126e-01 -9.91119981e-01
-1.10403669e+00 5.28553426e-01 4.51890022e-01 -2.04738945e-01
-4.26414013e-01 1.23479331e+00 8.53898346e-01 -3.75301749e-01
3.90413314e-01 -1.59872383e-01 -5.01628339e-01 -1.51907252e-02
4.19030339e-01 3.42973769e-01 -8.22567716e-02 -7.09288418e-01
-6.96628928e-01 5.87884665e-01 -1.52172565e-01 2.85809398e-01
1.02006578e+00 -7.36000612e-02 -2.28653499e-03 2.45001629e-01
9.33790922e-01 -3.28357257e-02 -1.25736964e+00 -2.97777895e-02
-4.98253644e-01 -7.94722974e-01 4.07991558e-02 -7.48081028e-01
-1.04151833e+00 7.20144272e-01 6.77866459e-01 1.55893981e-01
8.45961213e-01 3.31591189e-01 3.08621556e-01 6.35092556e-02
2.51831114e-01 -9.45908785e-01 4.72882181e-01 3.34887683e-01
1.09940672e+00 -1.32029641e+00 3.02113183e-02 -1.01465476e+00
-7.98855901e-01 6.17083490e-01 1.29942632e+00 1.75011039e-01
3.58322382e-01 7.94833452e-02 -2.34956011e-01 -5.05088270e-01
-3.92197579e-01 -6.74587190e-02 5.77899873e-01 8.59178841e-01
-3.51786078e-03 1.15012024e-02 -1.74943596e-01 8.32010746e-01
7.86587074e-02 -4.07573372e-01 6.88853413e-02 5.75948298e-01
-2.85395145e-01 -4.21799004e-01 -3.37411165e-01 5.77111483e-01
1.38684645e-01 -8.39327425e-02 -4.36726034e-01 8.60082567e-01
2.35194340e-01 5.75664461e-01 3.21004242e-01 -4.58623588e-01
3.77484024e-01 -4.73986477e-01 4.97132033e-01 -6.01245105e-01
-1.15003444e-01 7.51027539e-02 -1.32930120e-02 -5.44090092e-01
-6.59206361e-02 -1.99840680e-01 -1.18672967e+00 1.30989142e-02
-7.37227857e-01 -2.08267838e-01 1.63114473e-01 6.90728545e-01
6.85005665e-01 8.85280192e-01 3.38977575e-01 -1.00628614e+00
-3.74038100e-01 -9.10375357e-01 -4.37868744e-01 2.61195689e-01
2.59197086e-01 -1.29988670e+00 -4.23070520e-01 -7.88798332e-01] | [9.1362886428833, 0.625058650970459] |
e771ea64-aa29-4083-ba70-e35b61793bfe | academic-writing-with-gpt-3-5-reflections-on | 2304.11079 | null | https://arxiv.org/abs/2304.11079v1 | https://arxiv.org/pdf/2304.11079v1.pdf | Academic Writing with GPT-3.5: Reflections on Practices, Efficacy and Transparency | The debate around the use of GPT 3.5 has been a popular topic among academics since the release of ChatGPT. Whilst some have argued for the advantages of GPT 3.5 in enhancing academic writing, others have raised concerns such as plagiarism, the spread of false information, and ecological issues. The need for finding ways to use GPT 3.5 models transparently has been voiced, and suggestions have been made on social media as to how to use GPT 3.5 models in a smart way. Nevertheless, to date, there is a lack of literature which clearly outlines how to use GPT 3.5 models in academic writing, how effective they are, and how to use them transparently. To address this, I conducted a personal experience experiment with GPT 3.5, specifically by using OpenAI text davinci 003 model, for writing this article. I identified five ways of using GPT 3.5: Chunk Stylist, Bullet to Paragraph, Talk Textualizer, Research Buddy, and Polisher. I reflected on their efficacy, and commented on their potential impact on writing ethics. Additionally, I provided a comprehensive document which shows the prompts I used, results I got from GPT 3.5, the final edits and visually compares those by showing the differences in percentages. | ["Oğuz 'Oz' Buruk"] | 2023-02-12 | null | null | null | null | ['ethics'] | ['miscellaneous'] | [-3.59280020e-01 5.60352445e-01 -2.33921006e-01 2.51401991e-01
-3.36032152e-01 -8.43372345e-01 6.49490535e-01 3.94863069e-01
-2.24639252e-01 6.89850867e-01 9.69612598e-01 -8.40129852e-01
-3.47586900e-01 -3.60196143e-01 -3.00022632e-01 -3.85724232e-02
6.78538442e-01 1.46820650e-01 1.17962426e-02 -2.22376227e-01
1.46486223e+00 2.29217820e-02 -9.04810905e-01 1.11294292e-01
1.21090758e+00 -3.01697552e-01 6.78560287e-02 4.51516867e-01
-7.82879770e-01 9.82175648e-01 -1.26948488e+00 -8.11508477e-01
-1.22199260e-01 -3.96887779e-01 -1.11004829e+00 -1.41552150e-01
2.11789384e-01 -3.15724820e-01 2.81259939e-02 8.75981867e-01
6.19873941e-01 -2.33089905e-02 4.10368055e-01 -9.96105969e-01
-1.16604221e+00 9.82231438e-01 -3.78919572e-01 5.35323501e-01
7.02749610e-01 2.06541583e-01 6.66204929e-01 -5.82782388e-01
8.36321890e-01 1.44527280e+00 7.31010258e-01 2.59808838e-01
-9.57210600e-01 -9.77981627e-01 -3.59596461e-01 -7.37335756e-02
-1.04018211e+00 -2.95967788e-01 2.29013160e-01 -8.21195304e-01
9.52968776e-01 4.01853442e-01 1.01080537e+00 1.14187944e+00
5.50710559e-01 1.31509587e-01 1.56241584e+00 -6.72533870e-01
-9.58102271e-02 6.92690074e-01 4.07520354e-01 2.26026133e-01
4.16531235e-01 -5.57768881e-01 -7.51227796e-01 -3.09853345e-01
4.98456985e-01 -3.56725782e-01 -2.60852307e-01 6.06457412e-01
-9.77484882e-01 9.08226311e-01 -1.03700131e-01 6.92202032e-01
8.77738222e-02 -1.91322938e-01 7.18025327e-01 3.84654820e-01
8.75785768e-01 7.97630966e-01 1.19407654e-01 -1.09135580e+00
-9.17631149e-01 1.85429305e-01 1.19381809e+00 7.86727607e-01
-4.47516777e-02 -2.45056316e-01 -2.46751428e-01 8.21410120e-01
5.14885187e-01 1.75881103e-01 5.63074708e-01 -9.89961624e-01
5.57335258e-01 7.15586722e-01 6.55487552e-02 -1.24042010e+00
1.86406448e-02 -2.30555922e-01 -6.66419342e-02 5.12631498e-02
5.32080531e-01 -4.11611497e-01 -2.05226675e-01 1.10297608e+00
-1.77895665e-01 -4.21713859e-01 -6.35945871e-02 5.24804831e-01
1.14043033e+00 6.46961570e-01 1.14042237e-01 -3.85420434e-02
1.11676431e+00 -6.06225491e-01 -1.12890017e+00 -2.23473489e-01
1.09661746e+00 -1.64212763e+00 1.24123144e+00 2.85719454e-01
-1.20196974e+00 -5.41171208e-02 -9.41112697e-01 -2.98389435e-01
-2.98204839e-01 -8.05097297e-02 1.58935547e-01 1.22059727e+00
-8.97760868e-01 8.20597053e-01 -5.75568140e-01 -9.78461027e-01
3.86263043e-01 -3.13723981e-01 -2.44194061e-01 1.90183267e-01
-8.05207849e-01 1.27707803e+00 2.22656727e-02 1.00300841e-01
3.34459215e-01 -7.18870282e-01 -3.71081591e-01 -1.86499998e-01
2.11795881e-01 -4.01833773e-01 1.19467270e+00 -6.03080273e-01
-1.60857129e+00 6.48383021e-01 -7.92666078e-02 -8.52221102e-02
5.56678772e-01 -5.49385846e-01 -1.69492498e-01 2.01807152e-02
5.25590241e-01 2.39868745e-01 3.06753218e-01 -7.55719543e-01
-9.21680927e-02 -1.06524728e-01 -5.55588445e-03 1.99581623e-01
-6.30936921e-01 7.90204167e-01 1.51140630e-01 -8.37153673e-01
-9.59570631e-02 -7.61881232e-01 3.61548543e-01 -2.40800202e-01
-3.92337859e-01 -4.79977369e-01 1.01395464e+00 -1.16831970e+00
1.63100469e+00 -2.15580201e+00 -4.66451347e-01 2.78663747e-02
3.89252931e-01 3.02755356e-01 3.87185216e-01 1.37773478e+00
4.23993170e-01 1.17452455e+00 3.51216942e-01 -2.66680289e-02
8.87561440e-02 8.70950520e-02 -1.43262327e-01 3.43379349e-01
-2.17807204e-01 6.37650549e-01 -9.70718861e-01 -4.96391147e-01
-1.36174768e-01 5.91864467e-01 -9.61909145e-02 -3.53485703e-01
2.82021642e-01 2.51246452e-01 -3.68885994e-01 3.76373738e-01
4.93040591e-01 -3.38116646e-01 -1.58547219e-02 4.25956666e-01
-9.27490413e-01 1.00517285e+00 -8.63958597e-01 8.67024541e-01
-7.33316243e-02 1.28738904e+00 -1.70963407e-01 -2.46714145e-01
1.13422060e+00 6.23769522e-01 -1.19296178e-01 -3.86257142e-01
6.82553723e-02 3.06414604e-01 1.96171924e-01 -7.93028414e-01
7.31605768e-01 -2.73370128e-02 1.87401280e-01 1.10936773e+00
-2.96282858e-01 -4.07609820e-01 4.40467075e-02 5.46324968e-01
8.76890004e-01 2.47114867e-01 2.54730374e-01 -3.94606680e-01
-2.24994674e-01 5.43703474e-02 -4.86819446e-02 8.48620355e-01
-1.25855491e-01 4.47625935e-01 7.40972161e-01 3.73340882e-02
-9.34621692e-01 -1.55148059e-01 2.58876234e-02 5.40139735e-01
-2.45781243e-01 -1.06344974e+00 -7.81414211e-01 -2.66774744e-01
-3.67956340e-01 1.33709407e+00 -3.19299608e-01 2.30162051e-02
-2.00183034e-01 -8.83081630e-02 6.69939697e-01 -8.20169374e-02
6.65514588e-01 -9.40879226e-01 -9.71457303e-01 2.01967284e-01
-2.79311389e-01 -7.14370310e-01 -5.61637521e-01 -2.16087610e-01
-9.02362525e-01 -8.18862855e-01 -6.44305348e-01 -4.36069787e-01
3.35678220e-01 5.74788511e-01 6.77418530e-01 4.90942776e-01
1.61736235e-01 5.02265096e-01 -6.80894017e-01 -5.94782948e-01
-7.58312881e-01 3.68654020e-02 -5.82823992e-01 -8.70868802e-01
4.21228021e-01 -5.14935017e-01 -1.42510712e-01 -5.37538491e-02
-7.48734117e-01 2.89288729e-01 3.48198801e-01 5.44237614e-01
-5.96141219e-01 -1.18095785e-01 4.45742458e-01 -1.20215404e+00
1.36815810e+00 -4.52900350e-01 2.70626217e-01 -1.80795193e-02
-8.69228840e-01 -4.09453273e-01 2.06341490e-01 -1.86505213e-01
-1.03266335e+00 -1.02719045e+00 -2.08555888e-02 2.60832459e-01
1.27481699e-01 6.66275918e-01 3.48776937e-01 -4.07491684e-01
7.74393559e-01 -3.79509449e-01 4.64873224e-01 -2.35966772e-01
1.77142471e-01 9.37769771e-01 -1.29644910e-03 -5.13458312e-01
6.55245721e-01 -2.06483528e-01 -6.85215712e-01 -1.07818806e+00
-4.61681932e-01 -3.46253723e-01 -2.03683361e-01 -4.17117029e-01
4.69972193e-01 -6.84910119e-01 -7.37954676e-01 1.56763241e-01
-1.12248957e+00 -4.77443129e-01 -1.20884813e-01 5.74983954e-01
1.33817689e-02 6.52433991e-01 -8.72329533e-01 -7.84269750e-01
-4.98714864e-01 -8.18523943e-01 1.52790457e-01 6.05360091e-01
-1.35696387e+00 -1.31942546e+00 -5.08014411e-02 9.27517593e-01
6.48797631e-01 1.81832135e-01 9.07125115e-01 -6.88909352e-01
-1.54491529e-01 -9.67756510e-02 -2.67361730e-01 8.68771747e-02
3.19539547e-01 5.06762385e-01 -5.90804636e-01 5.44481650e-02
-1.18674049e-02 -2.53805608e-01 -1.12980574e-01 -2.73595117e-02
5.89916050e-01 -1.28291285e+00 -2.06898272e-01 -1.71491817e-01
1.22651112e+00 1.24916643e-01 4.78608072e-01 9.85286117e-01
4.76377189e-01 7.26806939e-01 4.06580299e-01 4.34233069e-01
3.42824817e-01 3.84265512e-01 -2.35483004e-03 5.92888296e-01
-9.53632966e-02 -3.79605174e-01 5.46329916e-01 1.23850298e+00
-1.60674453e-02 -1.38285324e-01 -9.83913243e-01 4.73165184e-01
-1.43658984e+00 -9.30681586e-01 -7.24705696e-01 1.91001999e+00
7.38854766e-01 5.96072733e-01 1.95564464e-01 3.47585082e-02
4.77602422e-01 1.44488394e-01 2.24459678e-01 -1.03845668e+00
-4.38265651e-02 2.18212828e-01 3.24903220e-01 6.89449489e-01
-9.59122032e-02 7.16015279e-01 6.60313749e+00 4.77017105e-01
-1.10890162e+00 1.58674657e-01 2.52261549e-01 1.48395509e-01
-5.28867185e-01 6.07419848e-01 -6.75602555e-01 7.78184235e-01
8.40747535e-01 -7.55339026e-01 -8.54277462e-02 4.36497808e-01
7.61429310e-01 -4.82702971e-01 -5.45786619e-01 5.36620736e-01
1.21047504e-01 -1.43276823e+00 -4.13272604e-02 2.72310972e-01
4.49688137e-01 -3.34438562e-01 -1.08715475e-01 1.63747534e-01
2.90220767e-01 -9.74032521e-01 8.58330607e-01 3.01025718e-01
1.75017491e-02 -2.14702576e-01 7.17700601e-01 3.61688793e-01
-2.36122027e-01 2.89144069e-01 -1.23637460e-01 -7.82293141e-01
1.57544330e-01 3.26939583e-01 -1.23409474e+00 3.30063581e-01
6.11041486e-01 7.36117303e-01 -7.56182551e-01 1.28173709e+00
-3.62798244e-01 1.14215326e+00 -2.16549620e-01 -4.46061522e-01
1.27028823e-01 -3.94582272e-01 8.65015328e-01 1.39100921e+00
5.27337313e-01 6.53073266e-02 -5.32671690e-01 7.95915008e-01
1.86987683e-01 2.58433223e-01 -7.66878307e-01 -6.23697937e-01
9.45459306e-01 1.10653079e+00 -8.76272857e-01 -1.19475961e-01
-3.34339559e-01 7.42067575e-01 4.43182588e-02 1.36796191e-01
-3.49818885e-01 -4.25963163e-01 3.13326530e-02 6.24243081e-01
-4.28040922e-02 -3.28817636e-01 -8.19063306e-01 -6.13209486e-01
-3.83487977e-02 -1.21619916e+00 -1.08545437e-01 -1.12814581e+00
-1.07341588e+00 1.16428636e-01 2.19508320e-01 -8.21400285e-01
1.68937221e-01 -2.37066045e-01 -9.37423766e-01 1.13103318e+00
-5.64974248e-01 -5.87242126e-01 -1.10670649e-01 -1.08299226e-01
4.99176353e-01 4.12483186e-01 5.84682167e-01 9.36710835e-02
-4.26280349e-01 4.28401262e-01 4.73187454e-02 -4.82619405e-02
1.19938028e+00 -1.00041246e+00 3.25345606e-01 6.01968408e-01
-2.49103948e-01 1.20677102e+00 1.05906999e+00 -9.41772103e-01
-8.54297042e-01 -2.35159576e-01 1.51812422e+00 -5.83996773e-01
9.39061821e-01 1.86900884e-01 -1.06550312e+00 8.01266074e-01
1.02639937e+00 -1.01236606e+00 9.47488010e-01 1.12149259e-02
3.90394824e-04 6.62571967e-01 -1.05007923e+00 9.96083975e-01
6.53671443e-01 -3.67069334e-01 -1.07686996e+00 5.56842089e-01
4.92246538e-01 -5.46308577e-01 -1.10833943e+00 -5.11802971e-01
5.65521896e-01 -1.07383907e+00 4.39503670e-01 -8.31343532e-02
6.92428648e-01 1.25981629e-01 5.40822625e-01 -1.36306417e+00
-2.01900274e-01 -1.13969123e+00 4.84302521e-01 1.78256190e+00
3.73379529e-01 -1.06080627e+00 4.91163760e-01 1.07960558e+00
-4.16832089e-01 -3.37869138e-01 -8.82404327e-01 -4.75909382e-01
3.74466538e-01 -1.75047696e-01 7.10903779e-02 1.29815722e+00
7.92533219e-01 3.19916278e-01 -2.29912415e-01 -4.25390512e-01
1.65002197e-01 -5.25721014e-01 9.13312018e-01 -1.21326888e+00
-4.87346165e-02 -8.25375915e-01 2.70042638e-03 -7.64184713e-01
-5.19659400e-01 -5.63535869e-01 -6.62550628e-01 -2.03891087e+00
3.49571377e-01 -1.17803611e-01 9.02177930e-01 3.83863956e-01
-9.21613947e-02 -3.75964791e-02 6.63815558e-01 7.28144109e-01
3.09505258e-02 5.47299199e-02 1.49779677e+00 1.98784426e-01
-5.40799260e-01 -3.91264737e-01 -1.38601255e+00 4.21284586e-01
1.22425985e+00 -5.72491109e-01 -2.23938122e-01 -2.66354363e-02
7.69477308e-01 -2.05312923e-01 4.05184507e-01 -7.23834693e-01
1.74335852e-01 -1.80839598e-01 1.40424386e-01 -3.33837599e-01
1.28825739e-01 -2.89159834e-01 3.23710501e-01 3.33776295e-01
-3.90853435e-01 3.09321821e-01 5.05985320e-01 -1.29990041e-01
-1.86341517e-02 -9.23598528e-01 1.50638506e-01 -2.51110077e-01
8.54807794e-02 -6.34126127e-01 -9.32940006e-01 2.46596009e-01
7.02860594e-01 -8.74474406e-01 -8.14822912e-01 -8.90587509e-01
-2.95703679e-01 7.58692250e-02 6.96710765e-01 4.79409903e-01
2.13731661e-01 -9.11840320e-01 -5.08727610e-01 -2.86536157e-01
-1.45064890e-01 -2.48773307e-01 2.86093690e-02 1.15915382e+00
-7.25448310e-01 4.30716574e-01 -3.99256378e-01 -7.71491155e-02
-1.48707783e+00 -7.53442943e-03 -8.15858468e-02 5.94879836e-02
-1.09592283e+00 5.51387966e-01 -4.48493630e-01 -2.00470731e-01
-6.40535653e-02 -1.20196849e-01 -4.88708675e-01 2.01785058e-01
4.00720686e-01 8.62687647e-01 -9.05062258e-02 -5.90114832e-01
9.83597804e-03 3.31111133e-01 -2.85844117e-01 -5.03814518e-01
1.07271481e+00 -4.99197006e-01 -4.58014786e-01 9.27553177e-01
8.72104704e-01 6.10509634e-01 -5.07863462e-01 2.26262152e-01
2.31140107e-01 -8.69261205e-01 -4.43887562e-01 -8.86795223e-01
-3.41840424e-02 5.94335616e-01 -1.25044271e-01 7.76272833e-01
2.62545377e-01 -2.13535205e-01 3.56081873e-01 1.08902775e-01
-1.61132812e-01 -1.28776062e+00 2.03304023e-01 6.36809468e-01
1.11012495e+00 -7.08203256e-01 3.90897483e-01 -3.69007081e-01
-7.20330358e-01 1.52800047e+00 3.83100390e-01 2.70253897e-01
5.01637936e-01 -2.53478866e-02 2.07431868e-01 -4.36601460e-01
-5.30869007e-01 6.64064288e-01 -2.55590409e-01 5.42106807e-01
1.14895725e+00 -8.14372003e-02 -1.40712488e+00 2.54950434e-01
-9.19737756e-01 3.06372702e-01 1.35305572e+00 1.11853623e+00
-3.58461976e-01 -1.27985251e+00 -8.85417461e-01 4.34041739e-01
-8.84399533e-01 -3.06547340e-02 -9.11136687e-01 9.92824495e-01
-3.66396695e-01 1.25895703e+00 -4.56105888e-01 -1.94431350e-01
-1.11717328e-01 2.49951780e-01 1.97048813e-01 -6.48396671e-01
-1.19682467e+00 -7.72405416e-02 6.38478458e-01 2.04136908e-01
-3.33836049e-01 -8.69945407e-01 -5.98192692e-01 -1.13822150e+00
-4.63776201e-01 4.02335703e-01 5.26127458e-01 1.02068770e+00
4.92656797e-01 1.87870070e-01 -1.29354475e-02 -3.48529279e-01
-4.31985199e-01 -1.47461832e+00 -1.95512205e-01 -1.14681974e-01
-6.66128173e-02 -2.84038961e-01 -4.01905745e-01 -2.90492177e-01] | [10.209488868713379, 7.400301456451416] |
ad9976fc-58dd-4d86-8398-a1242a7b8c9f | towards-the-semantic-weak-generalization | 2204.11280 | null | https://arxiv.org/abs/2204.11280v3 | https://arxiv.org/pdf/2204.11280v3.pdf | Deconstructed Generation-Based Zero-Shot Model | Recent research on Generalized Zero-Shot Learning (GZSL) has focused primarily on generation-based methods. However, current literature has overlooked the fundamental principles of these methods and has made limited progress in a complex manner. In this paper, we aim to deconstruct the generator-classifier framework and provide guidance for its improvement and extension. We begin by breaking down the generator-learned unseen class distribution into class-level and instance-level distributions. Through our analysis of the role of these two types of distributions in solving the GZSL problem, we generalize the focus of the generation-based approach, emphasizing the importance of (i) attribute generalization in generator learning and (ii) independent classifier learning with partially biased data. We present a simple method based on this analysis that outperforms SotAs on four public GZSL datasets, demonstrating the validity of our deconstruction. Furthermore, our proposed method remains effective even without a generative model, representing a step towards simplifying the generator-classifier structure. Our code is available at \url{https://github.com/cdb342/DGZ}. | ['Philip H. S. Torr', 'Haofeng Zhang', 'Yuming Shen', 'Dubing Chen'] | 2022-04-24 | null | null | null | null | ['generalized-zero-shot-learning', 'generalized-zero-shot-learning'] | ['computer-vision', 'methodology'] | [ 3.60532343e-01 2.96629697e-01 -1.23628944e-01 -3.61669600e-01
-9.82465088e-01 -5.58145940e-01 7.98765838e-01 9.96146873e-02
1.02095686e-01 8.68440628e-01 5.75838871e-02 -3.35475028e-01
-4.07884836e-01 -1.05019724e+00 -4.20628607e-01 -8.88058782e-01
1.19239822e-01 5.01879096e-01 5.55041321e-02 -2.14514211e-01
1.10823989e-01 2.02982262e-01 -1.93770337e+00 1.85539082e-01
8.81075442e-01 7.91022480e-01 -4.54247296e-02 5.03103137e-01
3.22574712e-02 7.44028568e-01 -6.61736190e-01 -5.49032688e-01
2.53464460e-01 -7.44882047e-01 -7.66243815e-01 7.46210143e-02
1.79236859e-01 -5.85948005e-02 -1.10319816e-01 1.08642673e+00
7.71862149e-01 3.07457387e-01 8.20027471e-01 -1.68207002e+00
-7.92014241e-01 7.79238105e-01 -1.32828414e-01 -1.39046218e-02
1.16750404e-01 5.01582436e-02 1.15219569e+00 -9.23930764e-01
5.36372960e-01 1.00890911e+00 6.54314876e-01 8.37322354e-01
-1.16206098e+00 -6.55385137e-01 1.81993227e-02 3.61845791e-01
-1.58834004e+00 -4.70346808e-01 8.50417674e-01 -5.52960813e-01
8.15366387e-01 1.73872888e-01 4.37485576e-01 1.47292888e+00
-5.52433319e-02 9.82774079e-01 1.20343530e+00 -7.73251534e-01
5.63662946e-01 2.49454737e-01 3.70358944e-01 4.13582444e-01
5.77296853e-01 2.98105687e-01 -4.27406788e-01 -1.71498701e-01
4.14633602e-01 -1.82339653e-01 -1.39283717e-01 -5.77112436e-01
-7.32259333e-01 1.10073805e+00 1.08421966e-01 2.90059537e-01
-3.52683440e-02 -8.79830047e-02 1.96058095e-01 5.86584657e-02
7.38501310e-01 2.74548173e-01 -9.64821503e-02 -1.18849903e-01
-8.64188969e-01 4.30485606e-01 9.90233183e-01 1.30358505e+00
7.80186534e-01 1.48170874e-01 -3.71808499e-01 9.70579028e-01
6.61482066e-02 2.48162687e-01 5.60694754e-01 -5.45339823e-01
6.58510253e-02 5.24658084e-01 -1.60031006e-01 -5.96607745e-01
-4.42524403e-02 -7.17301548e-01 -6.92761064e-01 2.26329383e-03
4.02554840e-01 -1.70095980e-01 -7.93920100e-01 1.90569842e+00
1.51747599e-01 3.18228096e-01 1.84655845e-01 4.89712059e-01
7.31088221e-01 2.06626385e-01 1.94876760e-01 1.89208128e-02
1.05234969e+00 -7.63759315e-01 -5.35805285e-01 -2.83932239e-02
6.77962005e-01 -4.07944053e-01 1.10749781e+00 3.00333232e-01
-8.68086100e-01 -4.01111960e-01 -1.10365450e+00 1.56917865e-03
-5.83023250e-01 1.44378364e-01 7.67680526e-01 1.00041187e+00
-8.87236476e-01 6.43654048e-01 -5.94795406e-01 -5.28064966e-01
6.05292916e-01 7.60062188e-02 -6.13370501e-02 -1.72789350e-01
-1.38681579e+00 6.06707692e-01 4.69267488e-01 -2.04572469e-01
-8.14169884e-01 -5.96245766e-01 -7.78948545e-01 2.02639833e-01
5.64262748e-01 -7.21862733e-01 1.42531061e+00 -5.51227212e-01
-1.33670437e+00 7.16935754e-01 -4.59000506e-02 -4.77712750e-01
5.12262046e-01 -1.79283731e-02 -2.70890266e-01 -8.76296312e-02
1.17893770e-01 3.00817370e-01 6.93318903e-01 -1.30210471e+00
-6.78981423e-01 -3.26166362e-01 1.02303691e-01 5.71737848e-02
-3.34952146e-01 -3.81968290e-01 -1.02231227e-01 -7.52196193e-01
-1.59992173e-01 -8.16869795e-01 -9.12270173e-02 -4.73165423e-01
-3.94772947e-01 -4.94860381e-01 4.94923979e-01 5.19069785e-04
1.44175529e+00 -2.02668571e+00 -1.20691977e-01 -3.08217555e-02
2.08754092e-01 3.05188835e-01 -6.85737804e-02 8.79409254e-01
-2.91368008e-01 1.31208673e-01 -4.47982848e-01 -4.19952542e-01
1.26883700e-01 1.39549017e-01 -5.72696507e-01 2.87635714e-01
2.33688921e-01 9.09633040e-01 -1.33955455e+00 -1.81329265e-01
1.62587538e-01 3.73231858e-01 -5.37239254e-01 1.05991915e-01
-6.98361322e-02 1.66284740e-01 -2.96544671e-01 8.73086989e-01
5.11331618e-01 -1.68407485e-01 2.25949049e-01 5.92252426e-03
4.70884480e-02 2.07050502e-01 -1.12081289e+00 1.57845867e+00
-2.56029397e-01 2.71684885e-01 -4.27370250e-01 -1.14851081e+00
1.03507483e+00 2.00432956e-01 4.32237566e-01 -2.29846597e-01
2.18948841e-01 2.32788563e-01 4.45347838e-02 -2.54265159e-01
3.15470040e-01 -4.69996125e-01 -2.28514180e-01 6.20151341e-01
6.11499429e-01 -1.07271440e-01 3.75876337e-01 1.82726547e-01
9.51643825e-01 3.10919344e-01 6.69428110e-01 -3.25907499e-01
2.93659240e-01 -1.57066241e-01 4.62758213e-01 1.22775078e+00
-1.85069263e-01 6.62236035e-01 5.52280843e-01 -1.29392162e-01
-8.50656211e-01 -1.22151232e+00 -2.03454897e-01 1.07536662e+00
-1.00541815e-01 -6.46965802e-01 -8.85965645e-01 -6.73411608e-01
-9.92010161e-03 1.29799402e+00 -6.87993228e-01 -5.53265870e-01
-8.41304734e-02 -1.29877985e+00 6.82301939e-01 5.22638202e-01
3.08095455e-01 -8.09749722e-01 -5.77526152e-01 6.71444088e-02
-1.41363204e-01 -7.63715029e-01 1.03499994e-01 2.89530575e-01
-7.21472085e-01 -1.17837536e+00 -4.30348039e-01 -4.35915291e-01
5.20726800e-01 1.94753170e-01 1.07501602e+00 -2.05339491e-01
-4.47023094e-01 4.88991499e-01 -6.83901608e-01 -7.59729743e-01
-5.28348029e-01 4.62418199e-01 -3.45017724e-02 1.19110510e-01
7.40822136e-01 -7.65638113e-01 -3.06319445e-01 1.27544686e-01
-9.25183594e-01 6.85635507e-02 5.72569311e-01 9.07495975e-01
2.73067892e-01 5.13137430e-02 9.37087953e-01 -1.31436658e+00
9.99399960e-01 -8.86831403e-01 -3.31861228e-01 3.13604206e-01
-1.02538955e+00 -2.48266589e-02 3.76188427e-01 -2.65838832e-01
-1.09558535e+00 -3.34853530e-02 -2.38458127e-01 -3.87401491e-01
-2.97154963e-01 2.98708200e-01 -3.42410326e-01 1.78146288e-01
8.33046675e-01 3.90940696e-01 2.69208616e-03 -5.11195183e-01
5.14529705e-01 7.95454860e-01 4.21771944e-01 -7.43171334e-01
7.52467334e-01 4.02469426e-01 -2.35444084e-02 -7.17939734e-01
-9.89745557e-01 -3.30302477e-01 -7.26118386e-01 -1.21326864e-01
3.58946204e-01 -7.67669797e-01 -3.60346735e-01 5.64999640e-01
-6.35147810e-01 -3.53484541e-01 -8.24920833e-01 2.55787373e-01
-8.31145048e-01 3.27245206e-01 -3.97339731e-01 -9.94971216e-01
-2.59861290e-01 -8.49570096e-01 1.00543070e+00 4.32981923e-02
-1.82261363e-01 -1.05625856e+00 2.62423813e-01 1.22639492e-01
4.27835554e-01 1.76257536e-01 9.82870340e-01 -1.15485978e+00
-4.90432978e-01 -5.50568640e-01 6.53314665e-02 4.52856988e-01
2.80673146e-01 -1.74897656e-01 -1.29374743e+00 -2.91218966e-01
1.84733957e-01 -3.69618922e-01 9.54364836e-01 2.05961555e-01
1.21976769e+00 -1.66891776e-02 -2.11584166e-01 6.98646426e-01
1.59253442e+00 9.51099992e-02 6.13355756e-01 6.04588091e-02
5.01495898e-01 6.00591421e-01 5.07392108e-01 6.11914456e-01
2.89839089e-01 4.71579820e-01 2.21337706e-01 1.82236955e-01
-3.57148230e-01 -6.34606838e-01 8.31892118e-02 5.51431000e-01
-2.60633349e-01 -5.12988865e-01 -8.64449143e-01 5.17030239e-01
-1.89861226e+00 -1.13380897e+00 6.40924498e-02 2.38416433e+00
6.33134663e-01 -7.54875243e-02 1.53848633e-01 3.27720582e-01
7.56591797e-01 4.15612869e-02 -3.88199180e-01 -8.22692961e-02
-1.67142108e-01 4.34765846e-01 2.41986424e-01 3.88426453e-01
-1.04993868e+00 9.55761075e-01 6.65141869e+00 1.20022929e+00
-7.77130246e-01 2.07824543e-01 4.87553865e-01 -1.10666320e-01
-4.38231647e-01 1.84925750e-01 -1.05066013e+00 4.47073668e-01
7.82460213e-01 -7.23872602e-01 3.98471773e-01 1.11001623e+00
-1.18610233e-01 -1.96441915e-02 -1.20479405e+00 8.97516847e-01
2.87343889e-01 -1.20210505e+00 2.87380874e-01 1.50512561e-01
6.41335189e-01 -1.73773021e-01 3.80326398e-02 7.13360786e-01
3.98209929e-01 -9.35359597e-01 7.18989849e-01 3.01703662e-01
7.23743975e-01 -6.66062057e-01 6.23037219e-01 4.22316223e-01
-1.00537634e+00 -3.05288613e-01 -3.04425240e-01 -2.78309941e-01
-1.00260556e-01 5.61780393e-01 -9.78700876e-01 8.81507397e-01
3.22708130e-01 6.44058943e-01 -6.99123323e-01 9.10689056e-01
-4.26340789e-01 8.10724676e-01 3.55599895e-02 2.43268870e-02
-4.74784598e-02 -2.21271023e-01 5.87043345e-01 1.20771396e+00
4.35378343e-01 4.41269912e-02 -1.86709519e-02 1.07560396e+00
1.24113441e-01 -2.51528900e-02 -7.84735858e-01 -9.29499120e-02
6.21768117e-01 1.24512362e+00 -8.48581076e-01 -3.81673604e-01
-3.47964853e-01 6.78860426e-01 4.15460259e-01 3.92382681e-01
-7.99900532e-01 -4.13778156e-01 5.34883618e-01 1.53671518e-01
2.60242254e-01 3.00921276e-02 -2.06094310e-01 -1.33298171e+00
-1.47107750e-01 -8.11683178e-01 4.70673084e-01 -4.35149014e-01
-1.60634649e+00 5.83573639e-01 4.85841006e-01 -1.37533665e+00
-6.23532295e-01 -5.35384595e-01 -5.48725963e-01 9.08618271e-01
-1.21907616e+00 -1.21385789e+00 -3.06607097e-01 5.06898105e-01
6.25033200e-01 -2.81729966e-01 9.89212036e-01 2.51132518e-01
-5.41554809e-01 8.05302143e-01 1.24135166e-01 -1.09176770e-01
5.27037323e-01 -1.22834551e+00 3.79799604e-01 8.71033192e-01
7.00971782e-02 7.55668581e-01 5.93553603e-01 -6.06557250e-01
-1.13339591e+00 -1.15092063e+00 8.33550513e-01 -6.94512665e-01
6.56041682e-01 -6.98341012e-01 -7.41896868e-01 6.06005311e-01
-1.52522638e-01 -1.31570116e-01 1.01739407e+00 1.39880762e-01
-2.62748122e-01 4.05479968e-02 -1.19093299e+00 4.83213335e-01
1.26734686e+00 -4.21085387e-01 -5.73304892e-01 2.08545387e-01
3.40885758e-01 -1.46781415e-01 -4.65513200e-01 4.34451371e-01
4.11093563e-01 -1.16442037e+00 8.39031219e-01 -6.83635294e-01
3.08027029e-01 -1.79008823e-02 -3.11778545e-01 -1.37713957e+00
-4.28861916e-01 -4.88669872e-01 -1.78192317e-01 1.30923140e+00
2.12122455e-01 -8.54816616e-01 6.33935511e-01 4.21511799e-01
-2.77799726e-01 -9.67082143e-01 -7.94399023e-01 -1.00315726e+00
2.93668568e-01 -6.30566716e-01 7.04161108e-01 8.84169638e-01
-1.47664808e-02 3.60824019e-01 -3.94916654e-01 -1.87989980e-01
8.26789975e-01 1.44736990e-01 7.15555191e-01 -1.33543468e+00
-3.90085936e-01 -3.74632716e-01 -2.99440950e-01 -6.23628080e-01
1.28833622e-01 -1.28603387e+00 -1.09922431e-01 -1.33484912e+00
3.88221473e-01 -4.53163773e-01 -3.04867983e-01 5.46396196e-01
-2.61638284e-01 2.27046818e-01 1.38057292e-01 2.39995003e-01
-3.56369197e-01 5.44980228e-01 8.37448299e-01 2.78714210e-01
3.05721872e-02 1.52916193e-01 -1.06951654e+00 5.57378352e-01
1.05226624e+00 -4.02163744e-01 -9.42541420e-01 9.99687007e-04
-3.69326584e-02 -4.71883088e-01 5.15249252e-01 -1.03899872e+00
4.51524295e-02 -7.29639828e-03 1.51186839e-01 -2.87368566e-01
2.03416124e-01 -3.56299728e-01 1.51074603e-01 3.39981616e-01
-3.68674368e-01 -3.63889843e-01 -5.43877594e-02 6.96983993e-01
-4.32694657e-03 -3.83611441e-01 6.26021206e-01 -3.18541378e-01
-6.52687371e-01 3.06303591e-01 -2.81564146e-01 2.91131228e-01
1.03879833e+00 -2.38288209e-01 -2.67370224e-01 -2.99323052e-01
-7.89570630e-01 -1.43511057e-01 4.82698470e-01 4.73753542e-01
4.57582712e-01 -1.21570587e+00 -6.80039883e-01 4.91073459e-01
4.26905692e-01 -2.35939503e-01 3.02208453e-01 5.83839715e-01
2.07240991e-02 3.29640150e-01 -1.33388340e-01 -3.95135820e-01
-8.88501823e-01 5.96121609e-01 2.47307643e-01 -2.49748483e-01
-5.29756606e-01 7.21110761e-01 2.03779653e-01 -3.05157900e-01
1.51830390e-01 -1.55836204e-02 -6.94536120e-02 2.36512974e-01
3.93895596e-01 5.30743420e-01 2.01726884e-01 -4.57885176e-01
-1.97799116e-01 1.77597746e-01 1.45303542e-02 -3.64899710e-02
1.21270752e+00 2.87763192e-03 1.20457955e-01 6.64004266e-01
8.84562731e-01 -1.85469300e-01 -1.02078176e+00 -3.03210337e-02
-9.51995626e-02 -4.69283640e-01 -1.39021009e-01 -7.03695476e-01
-6.56409919e-01 8.99052143e-01 3.06799471e-01 2.65590191e-01
1.19493091e+00 1.09581783e-01 3.72270852e-01 3.97086367e-02
7.44901657e-01 -9.19516325e-01 -1.50296062e-01 4.24944103e-01
6.22647166e-01 -9.93153453e-01 -1.77975386e-01 -5.73557258e-01
-4.39714134e-01 9.65224028e-01 4.59488720e-01 -8.90133306e-02
6.26445472e-01 1.25954002e-01 -2.45067343e-01 -1.20027736e-01
-8.17036688e-01 -5.27382791e-01 1.75020158e-01 7.72859812e-01
4.79862243e-01 1.93900064e-01 -3.78450096e-01 9.27933633e-01
-3.53794605e-01 2.67957062e-01 5.86917579e-01 1.03707922e+00
-2.23520100e-01 -1.39398587e+00 -4.66972589e-02 3.96385044e-01
-1.61189109e-01 -2.04675555e-01 -3.72018009e-01 8.97966206e-01
2.52245843e-01 9.15594697e-01 -5.36985509e-02 -4.48206753e-01
2.34940886e-01 4.31355476e-01 6.02319539e-01 -1.00533998e+00
-1.93223536e-01 -1.14025399e-01 -1.73633844e-02 -2.94878036e-01
-2.38748595e-01 -6.20614231e-01 -6.17997766e-01 -1.12764142e-01
-2.15673894e-01 1.48559153e-01 3.71404469e-01 6.97189510e-01
4.77324009e-01 3.41569871e-01 5.09213030e-01 -7.46454835e-01
-9.57402945e-01 -9.49459255e-01 -8.34020078e-01 4.12176460e-01
-4.49891165e-02 -9.58159029e-01 -6.11017168e-01 -3.01993024e-02] | [9.862812042236328, 2.9193367958068848] |
140066fe-97b1-4c03-b0a7-16cd788c8816 | animepose-multi-person-3d-pose-estimation-and | 2002.02792 | null | https://arxiv.org/abs/2002.02792v1 | https://arxiv.org/pdf/2002.02792v1.pdf | AnimePose: Multi-person 3D pose estimation and animation | 3D animation of humans in action is quite challenging as it involves using a huge setup with several motion trackers all over the person's body to track the movements of every limb. This is time-consuming and may cause the person discomfort in wearing exoskeleton body suits with motion sensors. In this work, we present a trivial yet effective solution to generate 3D animation of multiple persons from a 2D video using deep learning. Although significant improvement has been achieved recently in 3D human pose estimation, most of the prior works work well in case of single person pose estimation and multi-person pose estimation is still a challenging problem. In this work, we firstly propose a supervised multi-person 3D pose estimation and animation framework namely AnimePose for a given input RGB video sequence. The pipeline of the proposed system consists of various modules: i) Person detection and segmentation, ii) Depth Map estimation, iii) Lifting 2D to 3D information for person localization iv) Person trajectory prediction and human pose tracking. Our proposed system produces comparable results on previous state-of-the-art 3D multi-person pose estimation methods on publicly available datasets MuCo-3DHP and MuPoTS-3D datasets and it also outperforms previous state-of-the-art human pose tracking methods by a significant margin of 11.7% performance gain on MOTA score on Posetrack 2018 dataset. | ['Laxman Kumarapu', 'Prerana Mukherjee'] | 2020-02-06 | null | null | null | null | ['3d-multi-person-pose-estimation'] | ['computer-vision'] | [-1.80905327e-01 -6.17977679e-02 2.06046835e-01 8.68652761e-02
-5.53781450e-01 -3.73171896e-01 3.23807150e-01 -3.97900939e-01
-7.26791859e-01 6.90254569e-01 3.92336190e-01 4.14740086e-01
3.75064939e-01 -3.84797096e-01 -5.42574227e-01 -4.84132886e-01
-6.14930950e-02 1.10357451e+00 3.04460675e-01 -2.81368822e-01
-4.25366938e-01 5.33776581e-01 -1.35259044e+00 -3.52484316e-01
3.05873215e-01 5.00613689e-01 -3.05410355e-01 9.52221334e-01
6.80800080e-01 1.04602717e-01 -7.20850646e-01 -3.81491959e-01
5.85978806e-01 -4.90540832e-01 -6.64236426e-01 4.69664454e-01
9.14288640e-01 -6.46109819e-01 -5.66804349e-01 5.89730859e-01
1.24391127e+00 2.55930930e-01 5.41665792e-01 -1.53258145e+00
2.42843151e-01 6.35964498e-02 -8.56211185e-01 -7.97096360e-03
1.26376092e+00 5.03310680e-01 8.66568163e-02 -4.04923439e-01
8.75905216e-01 1.53959429e+00 1.17918408e+00 1.09287024e+00
-7.09472239e-01 -5.33403993e-01 -9.15201753e-03 -2.77932244e-03
-1.36108613e+00 -3.88842672e-02 5.62773049e-01 -5.99421799e-01
9.47431624e-01 1.22758672e-01 1.43616641e+00 1.53489840e+00
1.43960312e-01 9.33221638e-01 1.02406371e+00 -2.78465837e-01
-2.37674594e-01 -1.00495145e-01 2.31796876e-02 9.92449939e-01
4.94616389e-01 1.34289980e-01 -5.02747118e-01 1.35947213e-01
8.97716820e-01 -8.67249966e-02 -2.25313962e-01 -3.08713615e-01
-1.25003970e+00 3.44574690e-01 3.33215207e-01 -1.71524584e-01
-5.90930581e-01 3.35515678e-01 6.08255923e-01 -8.81118625e-02
1.54450938e-01 -2.23108426e-01 -3.82720351e-01 -2.93278366e-01
-8.78241837e-01 9.77800846e-01 6.54415011e-01 1.08556867e+00
-1.49371609e-01 -1.07272163e-01 -3.51227671e-01 2.03877121e-01
4.43656266e-01 6.84249640e-01 2.17990890e-01 -6.47131562e-01
5.31775355e-01 7.07076848e-01 4.54116404e-01 -6.10501587e-01
-9.60784137e-01 -1.57813191e-01 -6.54653788e-01 3.04666847e-01
9.84400690e-01 -6.81852341e-01 -1.01992440e+00 1.57399559e+00
1.04399788e+00 1.42784957e-02 -2.50138938e-01 1.44357860e+00
1.00306487e+00 3.81491691e-01 3.23417246e-01 1.30742505e-01
1.81821907e+00 -1.05887532e+00 -7.40988612e-01 -4.45206881e-01
3.55708420e-01 -4.94779617e-01 6.64991021e-01 2.61438251e-01
-1.26308882e+00 -7.55172968e-01 -9.17491615e-01 -7.36751482e-02
-1.84275150e-01 5.00636935e-01 5.66790879e-01 1.10389686e+00
-6.33773029e-01 4.79923457e-01 -9.71389294e-01 -9.47766840e-01
4.21646774e-01 7.20613956e-01 -5.71895003e-01 2.37431735e-01
-9.69274461e-01 1.10730767e+00 3.09442610e-01 3.28279674e-01
-8.15175772e-01 -3.89499366e-01 -9.27360237e-01 -5.53055227e-01
3.83298844e-01 -1.50807488e+00 9.91097629e-01 -2.16136739e-01
-1.59149694e+00 1.18869662e+00 1.21745303e-01 -3.62019122e-01
1.45226157e+00 -9.40897644e-01 -2.19143510e-01 6.60974458e-02
-4.08576950e-02 8.19991112e-01 7.05024660e-01 -8.03542733e-01
-5.32100379e-01 -8.43227625e-01 -2.68209845e-01 5.92900932e-01
1.04281260e-02 2.64508963e-01 -9.18709517e-01 -5.55072963e-01
-3.50758135e-02 -1.49445307e+00 -2.37067774e-01 4.47508782e-01
-7.12154150e-01 -3.11877161e-01 7.81895161e-01 -1.03826451e+00
7.87761927e-01 -1.41145504e+00 5.60100257e-01 -1.69691741e-01
6.14256188e-02 3.46233100e-01 4.81909454e-01 4.26421203e-02
2.76597053e-01 -5.31727791e-01 8.38801339e-02 -7.86079943e-01
1.68255895e-01 -2.93452889e-01 4.98113573e-01 1.08925235e+00
-3.78373384e-01 7.83612609e-01 -5.95716119e-01 -7.82995045e-01
5.67951918e-01 8.56087625e-01 -3.63083601e-01 3.05840313e-01
1.52434558e-01 8.83046031e-01 -1.95511341e-01 9.46572423e-01
5.59247494e-01 3.30343135e-02 2.54539363e-02 -3.00802737e-01
9.05499980e-02 -3.05329919e-01 -1.69643390e+00 2.11967063e+00
1.72091097e-01 3.09884757e-01 -6.21793456e-02 -4.19800580e-01
4.53918248e-01 6.53875172e-01 8.51018131e-01 1.28025383e-01
6.25281394e-01 -7.05347285e-02 -2.48432711e-01 -6.75695240e-01
3.37079585e-01 -5.48922494e-02 -4.25983697e-01 2.01822996e-01
8.57309401e-02 2.79070199e-01 4.18080129e-02 -2.04468295e-01
8.56088817e-01 1.02091670e+00 3.20698053e-01 7.28824437e-02
4.92154837e-01 1.13090165e-01 4.70082164e-01 3.42904180e-01
-9.07173693e-01 5.62193215e-01 -1.60685256e-01 -6.67657018e-01
-1.05450070e+00 -1.12663424e+00 3.43314886e-01 6.21749103e-01
2.26979300e-01 -3.90782237e-01 -1.22319865e+00 -6.27127051e-01
1.98600784e-01 -1.16355374e-01 -5.38342953e-01 1.30729109e-01
-1.03808975e+00 -7.82989264e-01 8.98277760e-01 8.81276667e-01
9.26914155e-01 -9.64013219e-01 -1.11843967e+00 5.70235327e-02
-3.74768972e-01 -1.57997930e+00 -6.28834188e-01 -4.84800428e-01
-6.22506142e-01 -1.12511683e+00 -1.29635406e+00 -7.72868156e-01
5.14570177e-01 -1.96500659e-01 8.63356888e-01 -3.41503203e-01
-7.50951171e-01 5.25509953e-01 -6.00357316e-02 -5.68505943e-01
1.39410526e-01 -5.31685576e-02 7.68008590e-01 -3.85244310e-01
4.86714035e-01 -1.70228943e-01 -9.47076201e-01 4.04248208e-01
1.08680278e-01 -4.99503920e-03 3.51088017e-01 2.24850520e-01
4.57364261e-01 -1.31600842e-01 1.03346873e-02 -4.07451630e-01
3.14423114e-01 8.86148587e-02 -3.40453058e-01 7.85590187e-02
1.03745773e-01 -5.59159398e-01 5.51544540e-02 -5.93961656e-01
-8.79856229e-01 8.41983676e-01 -3.69281113e-01 -3.47360075e-01
-5.81354082e-01 -5.09072423e-01 -3.37758362e-01 -4.76029441e-02
7.35148191e-01 -3.86548042e-02 5.57939112e-02 -4.34903651e-01
1.49713010e-01 5.22565901e-01 8.56957912e-01 -3.13485652e-01
1.01603913e+00 6.59239709e-01 2.63216883e-01 -7.11615443e-01
-6.12836480e-01 -5.94416499e-01 -1.28897643e+00 -8.89936805e-01
1.43118572e+00 -1.26131094e+00 -1.34788299e+00 1.07668829e+00
-1.10049486e+00 -1.95530221e-01 1.77377127e-02 6.08719587e-01
-6.23767853e-01 3.92211795e-01 -6.68173194e-01 -1.00398314e+00
-7.21489310e-01 -1.00079596e+00 1.46136427e+00 4.06852514e-01
-7.15180278e-01 -7.66772568e-01 7.66223446e-02 6.91407382e-01
-1.13265656e-01 1.21717250e+00 -2.37217575e-01 -6.60441294e-02
-2.18728751e-01 -7.75432289e-01 4.01720107e-01 -3.64294112e-01
-1.14719756e-01 -6.15074337e-01 -7.21131384e-01 -6.56270862e-01
-4.19807911e-01 -3.65585119e-01 4.89976734e-01 7.88101137e-01
5.01745999e-01 -6.14620820e-02 -7.56724119e-01 6.21992826e-01
1.04454803e+00 -4.41890955e-01 4.05927807e-01 4.52426732e-01
1.17716813e+00 6.63194239e-01 7.10565388e-01 5.51602602e-01
5.16359568e-01 1.16032934e+00 1.76285028e-01 -1.42764347e-02
-4.99614507e-01 -2.18238831e-01 6.06659830e-01 7.44921565e-02
-7.71509349e-01 -9.33868662e-02 -7.73792565e-01 2.93207377e-01
-1.76820433e+00 -1.02658772e+00 -4.33406055e-01 2.09595633e+00
4.85320389e-01 1.95474327e-01 1.17777395e+00 2.73387611e-01
9.23848450e-01 -2.50276476e-01 -5.79344332e-01 3.15508485e-01
1.18635543e-01 -1.01537742e-01 6.11729860e-01 2.71703154e-01
-1.53028977e+00 9.57929671e-01 5.56806278e+00 6.65007010e-02
-6.29184425e-01 1.98925272e-01 -1.20513141e-01 -5.59317470e-01
7.93058753e-01 -5.17059445e-01 -1.32093287e+00 3.88736039e-01
5.82863033e-01 3.71214688e-01 -1.79346561e-01 7.30005622e-01
2.43251085e-01 -1.77483976e-01 -1.13276315e+00 1.41510272e+00
2.24768013e-01 -6.96061492e-01 -2.97371030e-01 1.43993184e-01
4.97230142e-01 -4.53631699e-01 -3.14823240e-01 2.06164852e-01
-9.32329968e-02 -9.05505598e-01 8.46689105e-01 5.77940822e-01
5.89538455e-01 -7.60099888e-01 6.75652921e-01 5.07210195e-01
-1.40563333e+00 1.28094047e-01 -4.91307378e-02 -4.84242700e-02
8.63279343e-01 9.86841768e-02 -6.26227856e-01 3.39084655e-01
1.00847173e+00 4.78786200e-01 -4.68682170e-01 1.34479547e+00
-2.60667026e-01 1.05325140e-01 -5.36411166e-01 4.46845032e-03
-7.90154487e-02 1.01292908e-01 9.16484654e-01 1.30641067e+00
3.18208754e-01 1.52067989e-01 5.98708749e-01 3.16785038e-01
1.64481521e-01 -1.49592906e-01 -4.41307604e-01 5.32698810e-01
4.33640592e-02 1.16273808e+00 -7.59988129e-01 -3.89245749e-01
-6.08872399e-02 1.57594705e+00 -1.07292160e-01 -1.47580713e-01
-1.16276360e+00 -8.68984833e-02 6.72059000e-01 5.08543491e-01
5.83333485e-02 -4.33263779e-01 8.47549736e-02 -1.04260623e+00
1.90266728e-01 -6.91391349e-01 7.34426558e-01 -6.00300074e-01
-8.37807715e-01 4.53752756e-01 2.06131816e-01 -1.50362575e+00
-5.30662656e-01 -6.77130342e-01 -2.49860853e-01 8.43439519e-01
-7.98432231e-01 -1.66526854e+00 -8.39537323e-01 8.93036127e-01
6.92892611e-01 -1.40015155e-01 8.56116235e-01 4.18637186e-01
-7.43680239e-01 7.21318722e-01 -7.57591724e-01 3.85196596e-01
8.27849030e-01 -1.34223831e+00 6.50419950e-01 8.83591950e-01
-2.40692109e-01 3.84362191e-01 1.05178249e+00 -1.09803450e+00
-1.68547273e+00 -9.58251715e-01 6.37273788e-01 -1.02661395e+00
-7.91923478e-02 -4.50067997e-01 -1.60492539e-01 9.63712335e-01
1.92102753e-02 5.93985692e-02 5.75935662e-01 -1.95792481e-01
2.28090256e-01 3.04461420e-01 -1.57706726e+00 6.85373068e-01
1.56722116e+00 8.38290825e-02 -5.13302922e-01 6.80246234e-01
1.68137059e-01 -1.19746029e+00 -9.60333169e-01 2.26224557e-01
1.03616786e+00 -6.38015866e-01 1.37215579e+00 -4.44309115e-01
-1.05043471e-01 -5.72014332e-01 2.10150838e-01 -8.79592478e-01
-1.48051366e-01 -8.99656594e-01 -6.03142142e-01 7.53065407e-01
-2.83441275e-01 8.62071216e-02 1.37307966e+00 6.58677578e-01
8.98896754e-02 -5.27297616e-01 -9.41911101e-01 -7.49922872e-01
-2.94963747e-01 -1.76201239e-01 1.38936624e-01 3.81780922e-01
-2.39680156e-01 2.12787479e-01 -9.49439764e-01 5.16871274e-01
1.31574619e+00 -2.19348654e-01 1.35828900e+00 -1.25414717e+00
-2.26927966e-01 -4.87678312e-02 -7.98878729e-01 -9.08365130e-01
-1.17716260e-01 -3.99565130e-01 -1.02684237e-01 -1.82203650e+00
2.04312265e-01 2.88960487e-01 5.29983759e-01 3.97781014e-01
-2.55154520e-01 5.15018523e-01 3.07397246e-01 6.65592402e-02
-6.18295848e-01 2.10235909e-01 1.29948044e+00 4.23804633e-02
-3.40487927e-01 3.71339351e-01 -4.96129170e-02 9.54720497e-01
3.72631580e-01 -2.80573964e-01 -2.06238218e-02 -1.03545338e-01
-3.67841750e-01 1.08696535e-01 9.79114115e-01 -1.67968273e+00
3.91210198e-01 3.68097275e-01 1.18615377e+00 -1.01922643e+00
8.67606044e-01 -7.48325169e-01 5.13541639e-01 1.07808602e+00
2.85134792e-01 1.69137612e-01 3.18084598e-01 3.77150416e-01
5.08312106e-01 2.51184881e-01 8.60285878e-01 -5.84587395e-01
-8.54692638e-01 5.80458999e-01 -1.50276899e-01 -6.28700554e-02
1.39564085e+00 -6.15218341e-01 3.75031270e-02 -2.00714037e-01
-1.05513239e+00 2.39043280e-01 3.73518884e-01 5.89342177e-01
4.88329858e-01 -1.44134736e+00 -8.30628395e-01 -1.30395576e-01
-1.49573579e-01 -1.51746785e-02 3.69480252e-01 8.23543489e-01
-7.53173947e-01 3.69646400e-01 -6.78274214e-01 -7.48695195e-01
-1.86952865e+00 1.14276432e-01 5.62804639e-01 -1.78833276e-01
-1.12181175e+00 8.13860774e-01 -6.22169375e-01 -4.92242157e-01
4.22604710e-01 1.36968726e-02 -1.92536697e-01 3.39413364e-03
4.83632356e-01 8.53690028e-01 -2.94723123e-01 -1.26566982e+00
-7.51139104e-01 1.13972294e+00 2.39349708e-01 -2.28752062e-01
1.19225109e+00 -1.67764366e-01 4.60309535e-01 -7.11974408e-03
7.87837386e-01 -3.03637147e-01 -1.56726396e+00 1.40472025e-01
-4.00194824e-01 -4.20495212e-01 -3.27240765e-01 -8.50728810e-01
-8.66610467e-01 5.60415506e-01 1.15385616e+00 -4.57144648e-01
8.41856122e-01 1.24645062e-01 1.16125405e+00 1.03719942e-01
6.33460343e-01 -1.33973968e+00 2.32336774e-01 1.21700190e-01
7.26759076e-01 -1.22939479e+00 4.38879550e-01 -5.34273624e-01
-7.52535343e-01 7.96854198e-01 9.73682046e-01 -2.22130224e-01
4.14269656e-01 1.74900502e-01 8.01531877e-03 -3.25856984e-01
1.83527306e-01 -3.54833961e-01 5.90589881e-01 8.03136408e-01
3.28734636e-01 1.41964585e-01 -1.80873409e-01 4.68820453e-01
-4.73800182e-01 1.14626780e-01 1.80298895e-01 1.02321565e+00
-1.40083879e-01 -8.78610253e-01 -8.30217302e-01 -9.25170928e-02
-5.55830061e-01 7.77726889e-01 -3.78162354e-01 1.25727713e+00
3.02276105e-01 6.80567026e-01 -2.80929446e-01 -3.14487398e-01
8.44360352e-01 1.21851124e-01 1.09378040e+00 -4.01922047e-01
-7.96090245e-01 1.09544389e-01 2.09961146e-01 -6.77190065e-01
-7.01365232e-01 -1.22838175e+00 -1.31641471e+00 -5.23390830e-01
-4.08009328e-02 -5.55254102e-01 6.97576821e-01 9.80042577e-01
-8.19280371e-02 5.70922613e-01 -1.74796671e-01 -1.54665017e+00
-4.13686007e-01 -1.09398603e+00 -4.50255394e-01 6.43623412e-01
1.47158384e-01 -9.49355364e-01 1.66232407e-01 1.74037114e-01] | [7.031610488891602, -0.8457633852958679] |
083b0279-f262-46b8-abb0-8cb228e7589b | minimum-divergence-vs-maximum-margin-an | null | null | https://openreview.net/forum?id=H1xD9sR5Fm | https://openreview.net/pdf?id=H1xD9sR5Fm | Minimum Divergence vs. Maximum Margin: an Empirical Comparison on Seq2Seq Models | Sequence to sequence (seq2seq) models have become a popular framework for neural sequence prediction. While traditional seq2seq models are trained by Maximum Likelihood Estimation (MLE), much recent work has made various attempts to optimize evaluation scores directly to solve the mismatch between training and evaluation, since model predictions are usually evaluated by a task specific evaluation metric like BLEU or ROUGE scores instead of perplexity. This paper for the first time puts this existing work into two categories, a) minimum divergence, and b) maximum margin. We introduce a new training criterion based on the analysis of existing work, and empirically compare models in the two categories. Our experimental results show that our new training criterion can usually work better than existing methods, on both the tasks of machine translation and sentence summarization. | ['Huan Zhang', 'Hai Zhao'] | 2019-05-01 | null | null | null | iclr-2019-5 | ['abstractive-sentence-summarization'] | ['natural-language-processing'] | [ 6.21213615e-01 1.74406990e-01 -6.21563137e-01 -6.52280331e-01
-1.16833854e+00 -5.99850953e-01 6.28527641e-01 9.34436619e-02
-6.17435932e-01 1.21522963e+00 3.86243939e-01 -4.30165082e-01
2.95077622e-01 -3.14887017e-01 -5.06957829e-01 -4.19750988e-01
3.56037199e-01 5.19884646e-01 2.57668346e-01 -2.46978730e-01
5.34675360e-01 1.10891357e-01 -1.18816698e+00 4.36153531e-01
1.23891819e+00 5.46011508e-01 3.42312157e-01 8.66909742e-01
-2.94771761e-01 7.71042466e-01 -7.91077197e-01 -6.76154673e-01
-1.60579421e-02 -9.62691009e-01 -1.03142965e+00 -4.19695169e-01
3.67052972e-01 -1.08302467e-01 4.90392707e-02 1.17912233e+00
7.77797997e-01 1.30958527e-01 6.45704746e-01 -9.01542723e-01
-4.84763384e-01 6.58918321e-01 -2.49899372e-01 2.71084487e-01
4.15384978e-01 5.88588193e-02 1.17838466e+00 -7.21440971e-01
6.82971358e-01 1.13676834e+00 7.76650369e-01 8.57703567e-01
-1.13136256e+00 -3.11850280e-01 -1.14300158e-02 3.21513921e-01
-9.83683586e-01 -4.71848011e-01 3.00866455e-01 -2.90526003e-01
1.31932771e+00 3.77810895e-01 3.28989089e-01 1.17003214e+00
3.86380821e-01 1.01870787e+00 9.37854946e-01 -4.65425849e-01
2.49904022e-01 1.51655406e-01 1.75592616e-01 3.91381860e-01
-9.65273902e-02 -1.26267403e-01 -4.54572111e-01 -1.65692449e-01
4.83478993e-01 -5.18638313e-01 -3.87734085e-01 7.32007325e-02
-1.39112091e+00 8.74827504e-01 -1.19461134e-01 2.50787735e-01
-3.47134352e-01 -6.79035187e-02 6.51753843e-01 2.95465469e-01
6.50867343e-01 7.11698234e-01 -7.01579392e-01 -6.43135607e-01
-1.11017191e+00 9.87190828e-02 1.08209491e+00 7.76351869e-01
4.68233585e-01 2.43163574e-02 -5.26838005e-01 1.16327929e+00
1.68673918e-01 2.09828228e-01 8.75270844e-01 -8.07555556e-01
6.54822469e-01 1.55986011e-01 1.28378287e-01 -5.26859760e-01
-7.07888156e-02 -1.81202292e-01 -7.34760463e-01 -3.22157502e-01
3.09936047e-01 -4.88337159e-01 -6.60579443e-01 1.87704313e+00
-5.94550408e-02 1.26790151e-01 1.50514588e-01 7.88734674e-01
7.05738246e-01 1.06382632e+00 5.33308536e-02 -7.51129508e-01
7.92271495e-01 -1.40386534e+00 -9.44281161e-01 -1.44801557e-01
1.04096580e+00 -8.89584184e-01 9.76942062e-01 3.11635703e-01
-1.21859741e+00 -4.04751271e-01 -9.26086724e-01 9.64843780e-02
4.86870483e-02 8.24994370e-02 4.16333646e-01 7.62641311e-01
-1.22625887e+00 8.76842916e-01 -7.63901412e-01 -4.72621709e-01
-8.85728151e-02 3.60158652e-01 -5.74064553e-02 2.39068493e-01
-1.36658025e+00 1.25191450e+00 5.32510400e-01 -8.29626545e-02
-4.86892462e-01 -3.48534286e-01 -6.33336067e-01 7.16143567e-03
4.25153151e-02 -7.12045133e-01 1.59074616e+00 -1.42607403e+00
-2.00780725e+00 6.99376285e-01 -3.95027667e-01 -5.54796398e-01
3.36673409e-01 -3.60549778e-01 -2.06671745e-01 -1.05240010e-01
-2.44960785e-01 7.33413875e-01 2.81744868e-01 -8.26259732e-01
-4.28143799e-01 -3.26540507e-02 -1.63632318e-01 4.91398484e-01
-3.55724156e-01 4.94426847e-01 -3.19556892e-01 -6.91515625e-01
-4.92136985e-01 -8.44706833e-01 -3.40070665e-01 -5.60457289e-01
-3.95314127e-01 -5.51012039e-01 9.95509550e-02 -8.92515600e-01
1.58053648e+00 -1.53833818e+00 3.32902849e-01 -3.26619059e-01
-3.31376910e-01 8.62340868e-01 -4.63146031e-01 6.41038239e-01
-5.90896718e-02 1.87693194e-01 -3.98387551e-01 -3.46319348e-01
-1.42786577e-01 1.47598878e-01 -2.24762321e-01 -7.08111376e-02
1.93545252e-01 8.84702325e-01 -1.07178020e+00 -5.87905288e-01
-1.84798077e-01 2.30038464e-01 -4.95105714e-01 2.45299295e-01
-3.47973883e-01 2.74913847e-01 -2.78408915e-01 2.21314192e-01
3.69452208e-01 -9.78300422e-02 4.97315437e-01 2.72375703e-01
-1.16583109e-01 7.52127707e-01 -6.52821660e-01 1.74889433e+00
-1.92661598e-01 7.41518855e-01 -2.91062504e-01 -1.08753455e+00
9.79719758e-01 4.86042380e-01 2.26089478e-01 -3.77478927e-01
-6.45250678e-02 2.42971912e-01 1.40512690e-01 -5.72325408e-01
7.23376751e-01 -1.17999509e-01 1.34071633e-01 3.28421295e-01
1.50510475e-01 5.61551042e-02 3.91527891e-01 -1.43039776e-02
1.09888947e+00 4.67585504e-01 6.14084482e-01 -2.08512440e-01
4.41079944e-01 1.57086611e-01 8.48292828e-01 7.57565439e-01
-3.03040981e-01 7.61840045e-01 6.08628929e-01 -2.07560718e-01
-1.28218365e+00 -7.73692191e-01 1.44115239e-01 1.10744333e+00
-1.83374584e-01 -6.69951856e-01 -1.12243760e+00 -9.71929014e-01
-4.45056170e-01 9.68977928e-01 -2.66599983e-01 -1.70125198e-02
-8.09796095e-01 -1.01107335e+00 8.02872062e-01 5.86220860e-01
1.10205360e-01 -1.14452016e+00 -2.25286022e-01 3.56620401e-01
-4.74355459e-01 -7.54068136e-01 -6.40341699e-01 7.27040023e-02
-1.26469600e+00 -5.58667004e-01 -8.15144420e-01 -8.75231981e-01
2.52858698e-01 7.63708577e-02 1.22935319e+00 4.77398857e-02
2.25232989e-01 -5.02625145e-02 -5.66639066e-01 -4.11420733e-01
-7.77114868e-01 4.04950261e-01 1.11013897e-01 -3.34075063e-01
5.64255893e-01 -4.29036111e-01 -4.24210966e-01 2.90394127e-01
-8.56028676e-01 2.32393727e-01 9.52868521e-01 1.00227189e+00
4.67282683e-01 -7.63208687e-01 1.00246334e+00 -9.38358068e-01
9.42222238e-01 -3.82177234e-01 -2.38084212e-01 6.61688745e-01
-8.52342784e-01 2.45910794e-01 6.80531859e-01 -4.18557107e-01
-1.04402435e+00 2.55950764e-02 -5.64092040e-01 -1.14973530e-01
-1.82932213e-01 8.49742234e-01 -3.71367522e-02 3.17232370e-01
5.90076208e-01 4.26979959e-01 4.24391180e-02 -5.46519339e-01
8.40059891e-02 9.19855833e-01 2.16095865e-01 -2.10828558e-01
2.25010127e-01 -3.43453169e-01 -3.66839588e-01 -7.92262137e-01
-9.18100893e-01 -4.56257373e-01 -8.34320188e-01 -9.00768936e-02
8.65779698e-01 -6.68358147e-01 -3.38234037e-01 2.71144450e-01
-1.38305867e+00 -4.24806356e-01 6.74098209e-02 7.27286041e-01
-6.62003338e-01 8.37602496e-01 -7.92577326e-01 -6.95261598e-01
-6.09487414e-01 -1.04498672e+00 9.57630634e-01 1.22040272e-01
-4.91022229e-01 -1.20825636e+00 5.58728218e-01 4.02692914e-01
3.89575183e-01 1.54263489e-02 9.34162855e-01 -9.30035114e-01
-2.23002896e-01 -8.37513283e-02 5.03632687e-02 6.75662398e-01
-1.53550953e-01 -7.95225892e-03 -7.21001387e-01 -2.00121865e-01
2.52316426e-02 -4.02510554e-01 9.31152701e-01 3.48986864e-01
8.92463744e-01 -4.01760608e-01 -1.80246070e-01 2.39154041e-01
1.39299190e+00 2.41697699e-01 8.86096239e-01 3.70817989e-01
4.58328426e-01 4.82174277e-01 8.07488561e-01 2.18409806e-01
2.15462744e-01 8.36556852e-01 7.40187988e-02 2.79022783e-01
-1.40383672e-02 -3.89831185e-01 8.66175115e-01 1.44357145e+00
-1.30472600e-01 -6.04670048e-01 -9.46948886e-01 4.17211652e-01
-2.22366309e+00 -1.06962168e+00 -7.95498639e-02 2.23954630e+00
1.21341908e+00 1.00914083e-01 1.61495954e-01 -3.03589821e-01
8.24406981e-01 7.83096775e-02 -3.42465639e-01 -8.94008577e-01
-2.01693654e-01 -6.00448018e-03 3.73636276e-01 5.64034104e-01
-7.95608997e-01 9.83109176e-01 7.39183903e+00 1.07299125e+00
-1.10762000e+00 -2.59862747e-02 5.50226748e-01 -2.39587232e-01
-3.13397408e-01 3.33013758e-02 -9.85856533e-01 6.32289052e-01
1.39835882e+00 -8.08098316e-02 1.48428902e-01 6.93487346e-01
2.92983025e-01 -8.42974149e-03 -1.25480032e+00 7.52859890e-01
2.67743260e-01 -1.13814080e+00 2.34701127e-01 -1.60356928e-02
9.12927568e-01 2.03622833e-01 -1.64327830e-01 4.71123248e-01
3.02383840e-01 -1.06224310e+00 3.77027929e-01 5.99043131e-01
5.55476427e-01 -5.82760632e-01 9.91532385e-01 7.29137540e-01
-5.61408281e-01 1.81201264e-01 -5.30477703e-01 -1.29909843e-01
5.07363737e-01 4.80160177e-01 -1.03423727e+00 5.76497316e-01
1.63018197e-01 7.28527069e-01 -2.97742993e-01 1.22396088e+00
-3.26673537e-01 1.09686506e+00 3.85877937e-02 -5.58391452e-01
4.09384310e-01 -3.49035323e-01 7.35247433e-01 1.58035481e+00
2.82769412e-01 -1.59273893e-01 8.55753496e-02 4.01244372e-01
-2.94715986e-02 4.82773006e-01 -3.88102591e-01 -6.19370714e-02
2.60282427e-01 8.49336326e-01 -3.60903680e-01 -4.65800494e-01
-4.41241533e-01 1.12399924e+00 5.28555632e-01 1.88203231e-01
-7.98525989e-01 -5.03391027e-01 5.65331340e-01 -2.63907969e-01
3.37101907e-01 -2.72912383e-01 -2.77794540e-01 -1.29143393e+00
5.23445830e-02 -9.36766624e-01 1.88248202e-01 -6.31277323e-01
-1.09501028e+00 6.67658627e-01 -5.59404679e-02 -1.22778845e+00
-6.02912068e-01 -5.82310975e-01 -5.95585525e-01 9.04704690e-01
-1.44856286e+00 -6.09576225e-01 4.03202891e-01 -2.17570260e-01
1.01437116e+00 -9.04150233e-02 9.92252111e-01 2.12996483e-01
-8.32228005e-01 7.86175549e-01 7.55326569e-01 -5.40997647e-02
9.99778032e-01 -1.34817755e+00 5.64692736e-01 6.56284392e-01
2.14389730e-02 5.40993333e-01 7.98951149e-01 -6.71903551e-01
-9.56557930e-01 -9.00621235e-01 1.55747783e+00 -5.50750077e-01
3.17784697e-01 -1.59344420e-01 -8.26678097e-01 5.91412067e-01
4.42909569e-01 -8.31490099e-01 9.90143955e-01 2.35371932e-01
-1.21837355e-01 2.02878892e-01 -6.42475605e-01 5.71202636e-01
8.88486981e-01 -2.76361912e-01 -7.29598880e-01 4.24781293e-01
6.76419556e-01 -1.69715837e-01 -1.01399028e+00 6.63183391e-01
5.88699996e-01 -8.63346457e-01 5.24631798e-01 -9.89565074e-01
7.51968503e-01 -3.49461809e-02 -8.39258134e-02 -1.63826263e+00
-2.53391296e-01 -7.32979536e-01 2.31884513e-02 1.25736833e+00
8.48662734e-01 -2.26397112e-01 7.94509232e-01 2.72276968e-01
-4.08295244e-01 -1.07918727e+00 -6.79324925e-01 -9.33551848e-01
2.94748127e-01 -1.67904988e-01 3.86955202e-01 9.68274534e-01
2.97140688e-01 8.77625287e-01 -6.48194730e-01 -4.21681672e-01
1.86324745e-01 -7.98818655e-03 5.68687379e-01 -1.05858409e+00
-5.27574718e-01 -6.58701897e-01 -2.83738554e-01 -1.52854764e+00
2.95951843e-01 -9.71181810e-01 4.02507752e-01 -1.68260014e+00
6.00676358e-01 1.37879431e-01 -2.56570280e-01 2.96461880e-01
-5.96007049e-01 7.11167678e-02 3.45887579e-02 3.02175820e-01
-8.35924625e-01 6.42509580e-01 1.11171865e+00 1.97029963e-01
-1.38673365e-01 8.62416327e-02 -4.89975542e-01 5.32375634e-01
1.02700579e+00 -4.75079358e-01 -2.21643150e-01 -7.02547371e-01
4.55688566e-01 2.17736006e-01 -2.90517569e-01 -6.90998256e-01
1.02631800e-01 -2.83730775e-01 4.28254865e-02 -5.70210457e-01
1.94902077e-01 -1.27565652e-01 -1.74705341e-01 4.33925122e-01
-8.05602014e-01 1.49133593e-01 -5.71962669e-02 5.10105968e-01
-3.96530449e-01 -6.68147922e-01 7.13249743e-01 -2.09385812e-01
-6.86189592e-01 -1.18184581e-01 -5.81674814e-01 2.10928377e-02
7.32495844e-01 -2.81845152e-01 -2.17673734e-01 -5.12249351e-01
-5.44432580e-01 1.74151078e-01 4.20120150e-01 3.87968868e-01
5.06528914e-01 -1.14676738e+00 -1.10566151e+00 -2.75983363e-01
-1.04159564e-01 -5.18835604e-01 -5.74774891e-02 9.59780872e-01
-4.61589366e-01 7.93632627e-01 -2.01290250e-01 -5.13459742e-01
-1.51555860e+00 3.53703678e-01 1.52293801e-01 -6.70491278e-01
-1.40257962e-02 8.61786246e-01 9.20483395e-02 -7.47633994e-01
3.23743820e-01 1.36958927e-01 -4.66329247e-01 -1.12105176e-01
5.67694366e-01 4.09050971e-01 1.12371005e-01 -6.22683764e-01
-5.93991801e-02 2.40580395e-01 -2.36504570e-01 -1.58375964e-01
1.17240620e+00 -1.25481740e-01 -1.19385704e-01 6.04941905e-01
1.14825165e+00 -3.10462743e-01 -9.69791174e-01 -2.59578973e-01
3.07707757e-01 -2.03772828e-01 -2.71330088e-01 -9.13520396e-01
-3.26741844e-01 1.04943156e+00 2.53893256e-01 1.57094285e-01
9.77310181e-01 -3.29290986e-01 9.78699327e-01 6.11755848e-01
7.52046928e-02 -1.22411299e+00 7.13845901e-03 1.06975555e+00
5.70072651e-01 -1.12731695e+00 -2.35904291e-01 -2.57962018e-01
-8.89226675e-01 1.08878076e+00 6.26144230e-01 9.08373520e-02
1.40197381e-01 5.90126216e-02 1.03155293e-01 3.70658964e-01
-1.19612253e+00 -1.04539534e-02 3.82550687e-01 4.42707479e-01
9.97303486e-01 -5.80705814e-02 -9.83587980e-01 4.46336925e-01
-2.09347337e-01 2.41863672e-02 4.04971480e-01 8.69428098e-01
-8.50633681e-01 -1.50548041e+00 -6.46668002e-02 5.80248833e-01
-7.75000572e-01 -4.52983081e-01 -8.10327113e-01 3.37372780e-01
-3.02146852e-01 8.42399955e-01 -4.72418219e-02 -5.17997980e-01
1.66930422e-01 4.68952119e-01 5.33182800e-01 -7.95654476e-01
-8.19575548e-01 -7.59035200e-02 4.82596874e-01 -2.43927196e-01
-5.21989405e-01 -8.35020661e-01 -7.49013662e-01 -1.11042686e-01
-6.12521648e-01 4.42934126e-01 7.17187703e-01 9.74331498e-01
2.92115092e-01 1.49746656e-01 6.47910476e-01 -5.88665485e-01
-1.05412805e+00 -1.36684072e+00 -2.42137820e-01 2.57698119e-01
6.56011254e-02 7.02302381e-02 -9.06573981e-02 1.51156172e-01] | [11.9368896484375, 9.34886646270752] |
c617b679-b1ef-4223-b61f-37e6c0a99829 | when-evolutionary-computation-meets-privacy | 2304.01205 | null | https://arxiv.org/abs/2304.01205v1 | https://arxiv.org/pdf/2304.01205v1.pdf | When Evolutionary Computation Meets Privacy | Recently, evolutionary computation (EC) has been promoted by machine learning, distributed computing, and big data technologies, resulting in new research directions of EC like distributed EC and surrogate-assisted EC. These advances have significantly improved the performance and the application scope of EC, but also trigger privacy leakages, such as the leakage of optimal results and surrogate model. Accordingly, evolutionary computation combined with privacy protection is becoming an emerging topic. However, privacy concerns in evolutionary computation lack a systematic exploration, especially for the object, motivation, position, and method of privacy protection. To this end, in this paper, we discuss three typical optimization paradigms (i.e., \textit{centralized optimization, distributed optimization, and data-driven optimization}) to characterize optimization modes of evolutionary computation and propose BOOM to sort out privacy concerns in evolutionary computation. Specifically, the centralized optimization paradigm allows clients to outsource optimization problems to the centralized server and obtain optimization solutions from the server. While the distributed optimization paradigm exploits the storage and computational power of distributed devices to solve optimization problems. Also, the data-driven optimization paradigm utilizes data collected in history to tackle optimization problems lacking explicit objective functions. Particularly, this paper adopts BOOM to characterize the object and motivation of privacy protection in three typical optimization paradigms and discusses potential privacy-preserving technologies balancing optimization performance and privacy guarantees in three typical optimization paradigms. Furthermore, this paper attempts to foresee some new research directions of privacy-preserving evolutionary computation. | ['Jun Zhang', 'Qingqi Pei', 'Ximeng Liu', 'Xiaoguo Li', 'Wei-neng Chen', 'Bowen Zhao'] | 2023-03-22 | null | null | null | null | ['distributed-optimization'] | ['methodology'] | [-2.66940475e-01 -4.10194069e-01 -3.06634102e-02 -4.50882524e-01
-4.11922544e-01 -8.77085328e-01 2.44786724e-01 1.06118195e-01
-4.85323846e-01 8.34040463e-01 5.06746322e-02 -2.87449472e-02
-4.72401768e-01 -1.06208885e+00 -4.30032969e-01 -1.12503147e+00
9.69667956e-02 1.27488196e-01 -4.17136341e-01 1.94596145e-02
1.56198636e-01 6.37355089e-01 -1.50875175e+00 6.87821954e-02
1.13720143e+00 1.39898682e+00 -3.59633774e-01 2.82595336e-01
-7.66461939e-02 9.85506102e-02 -6.68212950e-01 -9.08350646e-01
6.98552251e-01 -2.48264328e-01 -3.87296796e-01 -3.23717952e-01
-2.71538526e-01 -2.16390908e-01 -8.29301998e-02 1.37664366e+00
8.50823343e-01 2.11861387e-01 1.65122095e-02 -1.76211476e+00
-7.99238741e-01 3.18319470e-01 -3.18533957e-01 -3.45867157e-01
5.76790385e-02 2.76817441e-01 8.03586602e-01 -6.57765031e-01
7.43387282e-01 6.75121665e-01 4.89066124e-01 5.54470479e-01
-8.69571984e-01 -4.88861442e-01 -2.66347621e-02 1.56910822e-01
-1.60347056e+00 -2.86037177e-01 6.09430850e-01 -1.56645074e-01
6.85241997e-01 1.18514681e+00 8.93933296e-01 4.80950415e-01
1.93135634e-01 7.82346487e-01 1.17021048e+00 2.11405084e-02
7.86794245e-01 6.07634127e-01 -2.81137433e-02 3.10647398e-01
4.83307093e-01 6.28716707e-01 -6.54164433e-01 -7.52468705e-01
-2.24027876e-03 2.68923134e-01 -4.19542015e-01 -6.19180024e-01
-6.22076929e-01 5.95251679e-01 2.89710253e-01 4.78338636e-02
-4.32533920e-01 -1.66548759e-01 6.49461448e-01 4.73268449e-01
4.54389453e-01 4.04774040e-01 -6.27778649e-01 -1.42585441e-01
-6.78655982e-01 4.78786081e-01 1.00146055e+00 1.19855642e+00
6.96535766e-01 -2.94498764e-02 -2.21716225e-01 3.54334325e-01
1.07298046e-01 1.56492174e-01 3.67314965e-01 -8.06542516e-01
3.89511287e-01 8.08681309e-01 6.55228421e-02 -1.37270761e+00
-8.57092347e-03 -3.86651367e-01 -1.00099182e+00 2.34912217e-01
5.71540557e-02 -5.20218134e-01 -7.28655159e-02 1.56259215e+00
7.92117059e-01 -5.42787075e-01 2.66795188e-01 1.25279164e+00
4.35086995e-01 6.58174098e-01 -1.00234859e-01 -4.45797563e-01
1.25627017e+00 -9.25628364e-01 -8.12416852e-01 4.76208746e-01
5.41493773e-01 -3.52317154e-01 6.32928908e-01 3.08092654e-01
-1.07412839e+00 8.34216848e-02 -8.44351530e-01 -1.07220605e-01
-9.02837217e-01 -1.68385189e-02 8.88817132e-01 1.16555977e+00
-9.89663363e-01 5.37645459e-01 -7.32108235e-01 -2.06292704e-01
6.64845705e-01 5.70964456e-01 -2.28242576e-01 1.48291916e-01
-1.00561893e+00 4.75519836e-01 4.33606714e-01 2.21770316e-01
-5.22020638e-01 -8.40200782e-01 -3.21843565e-01 3.34759474e-01
4.65237260e-01 -1.12335575e+00 4.54664230e-01 -7.85272002e-01
-1.65133131e+00 7.13866889e-01 6.19836710e-02 -3.34646374e-01
7.88598657e-01 -1.33546725e-01 -4.94879156e-01 -3.24502259e-01
-4.41965401e-01 -1.08359389e-01 3.76282722e-01 -9.58763540e-01
-7.37129807e-01 -9.66589987e-01 -4.22264248e-01 1.97891265e-01
-8.15544784e-01 2.84378946e-01 -2.43095681e-01 -5.59052050e-01
-3.24168950e-01 -7.51661003e-01 -3.74443144e-01 5.26983857e-01
-4.50973511e-01 2.30882749e-01 1.00442791e+00 -5.97629130e-01
1.45123029e+00 -2.22625542e+00 2.37161532e-01 5.17015576e-01
2.93268770e-01 1.15104496e-01 3.09371799e-01 4.53749776e-01
3.98099244e-01 4.82661426e-01 -3.86421502e-01 -3.70491982e-01
3.86695832e-01 8.53138566e-02 -3.35713327e-01 7.47804046e-01
-3.01261812e-01 1.01055229e+00 -4.33543652e-01 -2.76452154e-01
-9.90006402e-02 3.18878323e-01 -7.35191941e-01 3.85278493e-01
-2.29543552e-01 5.39692223e-01 -8.43990684e-01 1.00248873e+00
1.04973197e+00 -1.44229503e-02 2.19704524e-01 -3.61000821e-02
-5.70989490e-01 -4.64047909e-01 -1.06678283e+00 1.64657986e+00
-1.60562977e-01 5.06687313e-02 7.41400540e-01 -7.93520033e-01
1.00654912e+00 3.59739959e-01 8.25375497e-01 -2.91853994e-01
3.54377151e-01 4.42047030e-01 -2.47296914e-01 -3.54231507e-01
6.68747842e-01 3.62444043e-01 -7.96131119e-02 6.33099079e-01
-5.05056441e-01 -1.13311820e-01 -3.06150287e-01 -2.74958044e-01
8.20960224e-01 -1.27076998e-01 2.34391913e-01 -3.95122260e-01
5.67474484e-01 5.87616749e-02 1.03358459e+00 4.02939558e-01
-4.52696264e-01 2.84405470e-01 1.75896809e-01 -6.04390860e-01
-8.02011549e-01 -8.07705700e-01 -4.48833443e-02 9.33321059e-01
4.33176577e-01 -6.67517483e-01 -6.38683379e-01 -5.36067963e-01
3.32345754e-01 5.25808990e-01 -1.72776639e-01 -4.22542989e-01
-3.60360801e-01 -1.23636687e+00 4.62572455e-01 -6.34472445e-02
8.65398526e-01 -6.20090902e-01 -7.39978433e-01 -1.06098153e-01
1.45587310e-01 -4.22441483e-01 -6.09572232e-01 -1.31694466e-01
-8.48580301e-01 -5.85587502e-01 -3.98777992e-01 -2.35633403e-01
7.48606920e-01 -3.04875791e-01 7.15139806e-01 3.67893539e-02
-5.38140714e-01 2.57967591e-01 2.03529578e-02 -6.28042340e-01
1.86529025e-01 -1.35124490e-01 5.49174994e-02 4.05968040e-01
4.03627872e-01 -7.39293039e-01 -7.69565105e-01 3.71368140e-01
-8.11838269e-01 -4.34402198e-01 2.42527127e-01 6.70339227e-01
9.15382504e-01 2.90218860e-01 1.60612047e-01 -7.50461280e-01
9.25095201e-01 -6.49885416e-01 -1.06625807e+00 7.61492193e-01
-1.27288425e+00 -8.21565613e-02 8.94029498e-01 -1.13128096e-01
-1.16795576e+00 6.94090873e-02 2.87869602e-01 -4.98124868e-01
1.26782641e-01 4.04697120e-01 -7.05555201e-01 -3.63884002e-01
2.22997010e-01 5.11476398e-01 6.48592934e-02 -5.99081695e-01
3.18700552e-01 7.44896531e-01 2.30129585e-01 -8.41552317e-01
4.77680206e-01 6.19240105e-01 1.89466536e-01 -3.32066178e-01
1.11894019e-01 -4.58506532e-02 4.21549119e-02 1.00172974e-01
5.66168368e-01 -5.88511944e-01 -1.24788427e+00 5.36262095e-01
-1.13573241e+00 6.40073538e-01 -6.95685208e-01 3.04934651e-01
-3.81792635e-01 1.39938951e-01 -3.26346338e-01 -1.12730420e+00
-1.03592598e+00 -1.22113109e+00 7.03058898e-01 5.79625666e-01
8.18169937e-02 -7.88979232e-01 2.17509065e-02 2.13254437e-01
9.44335759e-01 5.42943537e-01 6.80248320e-01 -7.85158098e-01
-1.05473638e+00 -2.96813279e-01 -2.80436501e-03 4.05678265e-02
-1.65032834e-01 2.63965148e-02 -8.72690201e-01 -5.24330378e-01
5.01549184e-01 4.28923182e-02 1.61924005e-01 1.28004536e-01
1.75180089e+00 -8.23852956e-01 -3.75211686e-01 1.48777163e+00
1.49377942e+00 3.57910931e-01 4.36643273e-01 1.43253759e-01
2.39686817e-01 8.31162453e-01 5.00720441e-01 9.00279522e-01
2.05447644e-01 4.92925882e-01 4.14631099e-01 1.90411583e-01
7.05917478e-01 -2.06923485e-01 1.60457879e-01 7.39767015e-01
-1.83793157e-01 -4.12292242e-01 -4.82399523e-01 3.32916349e-01
-1.92808843e+00 -6.60097420e-01 1.46822274e-01 2.35259151e+00
6.79506600e-01 -8.30631077e-01 3.69417593e-02 -5.08835137e-01
8.30830872e-01 1.16793722e-01 -9.53103483e-01 -7.75351346e-01
-4.49493110e-01 5.33070900e-02 6.63398266e-01 -9.71772447e-02
-7.27034926e-01 3.96113098e-01 5.49075460e+00 8.36464882e-01
-1.11022401e+00 3.31055045e-01 7.35653520e-01 -5.30810714e-01
-7.74799526e-01 2.48400986e-01 -5.90553284e-01 7.91358829e-01
6.99432671e-01 -9.44743812e-01 7.48045743e-01 1.15170729e+00
-5.96204475e-02 2.80546874e-01 -9.70777631e-01 1.18672335e+00
-3.53685945e-01 -1.47842550e+00 -1.43423066e-01 4.87461627e-01
8.21962833e-01 -1.61737397e-01 3.51959258e-01 -1.27795950e-01
1.34575889e-01 -8.62717986e-01 4.53090459e-01 6.14648938e-01
6.07236683e-01 -9.92756486e-01 6.00464046e-01 4.35169399e-01
-9.60327446e-01 -3.01692754e-01 -4.12685871e-01 6.10980801e-02
1.59106568e-01 6.55039370e-01 7.45165274e-02 1.10824561e+00
1.05827916e+00 2.99738646e-01 -1.87519833e-01 1.03233838e+00
6.89059049e-02 -1.03253936e-02 -5.19694209e-01 -4.45803374e-01
-2.41719827e-01 -1.05074644e+00 8.75323594e-01 5.08672655e-01
5.50420940e-01 3.43529344e-01 -6.87559647e-03 1.46118069e+00
-2.69849807e-01 3.71981978e-01 -5.33441067e-01 -4.43876117e-01
7.50931263e-01 1.22396207e+00 -7.36965016e-02 1.04495600e-01
-9.13029611e-02 9.46454585e-01 1.46012962e-01 2.85918087e-01
-5.63939631e-01 -6.19298577e-01 1.22406650e+00 -3.82940233e-01
-4.21717502e-02 3.94353941e-02 -7.85462320e-01 -1.35641754e+00
3.93814653e-01 -8.99814010e-01 8.10718656e-01 -1.57875523e-01
-1.36900997e+00 5.60850620e-01 -2.93434441e-01 -9.06480253e-01
1.10440515e-01 -2.96972543e-01 -8.54743779e-01 1.14873898e+00
-1.10716403e+00 -8.50946248e-01 -2.06840485e-01 8.38456988e-01
-1.96500778e-01 -4.00195330e-01 9.25954223e-01 2.91434556e-01
-8.02869081e-01 9.42413449e-01 8.09068382e-01 -2.68835306e-01
3.30844492e-01 -6.96455419e-01 -3.99139002e-02 8.89879048e-01
-1.69413939e-01 9.17819500e-01 3.99506688e-01 -5.72433591e-01
-2.18470025e+00 -8.76696765e-01 8.45028043e-01 -2.97773808e-01
2.67059594e-01 -4.84152675e-01 -5.99482536e-01 2.72815853e-01
-4.19837385e-02 2.96642184e-01 8.65076125e-01 -1.46135986e-01
5.98905794e-02 -5.87814808e-01 -1.95968664e+00 7.37821817e-01
8.88550639e-01 -5.50070226e-01 3.05475891e-02 4.18148637e-01
8.52880716e-01 -3.58022422e-01 -1.23373449e+00 1.30492941e-01
5.66383660e-01 -9.03865993e-01 9.45997536e-01 -7.76001692e-01
1.05011351e-01 -2.29247630e-01 -3.66173893e-01 -9.64776695e-01
1.70866266e-01 -1.33588672e+00 -3.55951011e-01 1.38176334e+00
1.40312806e-01 -1.25467265e+00 1.12033892e+00 1.71262634e+00
2.12003499e-01 -9.24105465e-01 -1.07484639e+00 -7.98678696e-01
6.85059726e-02 -4.08851169e-02 1.51060939e+00 1.05454147e+00
-1.15693815e-03 -6.63192391e-01 -4.47022885e-01 2.59914935e-01
7.05588877e-01 7.92574525e-01 8.61813307e-01 -7.12670445e-01
-3.24639946e-01 -3.38630021e-01 -3.07434350e-01 -5.43458581e-01
-7.36003518e-02 -1.00622487e+00 -5.28671622e-01 -7.24187016e-01
-6.86059892e-02 -5.30121565e-01 -3.58097613e-01 2.58552015e-01
1.21014699e-01 -3.44695836e-01 2.35203907e-01 2.65929103e-01
-2.64168531e-01 9.83619809e-01 1.01675844e+00 1.83622576e-02
-3.57251227e-01 1.96617126e-01 -7.41155028e-01 2.22528756e-01
6.61640406e-01 -6.04840577e-01 -2.45573044e-01 -5.71870685e-01
2.12701440e-01 1.18280083e-01 3.43600303e-01 -4.09673214e-01
8.86534214e-01 -6.24019384e-01 1.54266179e-01 -3.53724539e-01
1.85240582e-01 -1.41519582e+00 9.21413600e-01 5.95618486e-01
-1.01570502e-01 3.04882705e-01 -7.15862438e-02 6.01643920e-01
-3.29285294e-01 1.74397230e-01 5.41991353e-01 -6.40597045e-02
-4.47478265e-01 8.50440264e-01 1.01871490e-01 -6.27833158e-02
1.50999367e+00 -1.94887057e-01 -4.38155830e-01 -9.80507117e-03
-6.04262233e-01 6.26819134e-01 8.16164255e-01 1.00417860e-01
6.14870548e-01 -1.17905951e+00 -6.17955923e-01 4.65316683e-01
-1.36924848e-01 4.66325507e-03 2.42944643e-01 9.48276937e-01
-4.83502775e-01 2.10848615e-01 -2.04626963e-01 -2.38343507e-01
-1.18935406e+00 7.68625259e-01 2.99467593e-01 -5.20518236e-02
-3.72183442e-01 8.74818027e-01 -4.94061112e-02 -7.63674855e-01
2.03420937e-01 2.38945842e-01 4.70525593e-01 -2.86271840e-01
3.88416439e-01 8.44811261e-01 -1.75053701e-02 -1.80114284e-01
-6.15619361e-01 3.33481133e-01 1.46199390e-01 7.32134208e-02
1.39016891e+00 -2.91596979e-01 -8.40626121e-01 -1.85294509e-01
1.49927235e+00 1.90798581e-01 -7.26634204e-01 -3.18046696e-02
-4.05193828e-02 -1.08401608e+00 -3.20614837e-02 -1.10588539e+00
-1.52759290e+00 5.36315203e-01 6.15280211e-01 -1.41710311e-01
1.57938933e+00 -5.75060427e-01 7.50101030e-01 2.28403077e-01
7.06693590e-01 -1.20070672e+00 -9.01902020e-01 -1.77526399e-02
8.21258068e-01 -9.27971900e-01 3.13716888e-01 -2.93169945e-01
-5.06708682e-01 9.45117831e-01 5.91120720e-01 2.43049204e-01
8.29230249e-01 4.43762839e-01 -3.03158402e-01 -2.63242364e-01
-7.10963488e-01 5.93088806e-01 1.44913241e-01 4.66870785e-01
-2.30821207e-01 1.77575573e-01 -7.14999259e-01 1.25569284e+00
-3.08207095e-01 3.99118774e-02 -1.47109345e-01 1.07370603e+00
2.53494769e-01 -1.33868897e+00 -4.47359085e-01 3.58875692e-01
-5.10228872e-01 3.58169638e-02 -8.10049593e-01 4.49722230e-01
3.20752144e-01 7.34381080e-01 -8.25339630e-02 -3.42987210e-01
2.09686622e-01 -8.51920843e-02 -1.32165596e-01 3.32515836e-02
-1.39804971e+00 -4.41224277e-01 -1.12846635e-01 -8.54250610e-01
1.45356402e-01 -6.07963681e-01 -9.83797133e-01 -7.57661045e-01
-4.61051345e-01 6.25198305e-01 9.53486443e-01 4.33602661e-01
1.16854560e+00 -9.75524858e-02 8.67182672e-01 -7.92310536e-02
-1.08156013e+00 4.11512479e-02 -8.37164462e-01 1.42644465e-01
-7.24318847e-02 7.33928457e-02 -4.89789218e-01 -4.38438594e-01] | [5.890294551849365, 6.517660140991211] |
fcbff802-89fe-41dd-b35c-efcb1654d1d7 | auto-absa-automatic-detection-of-aspects-in | 2202.00484 | null | https://arxiv.org/abs/2202.00484v2 | https://arxiv.org/pdf/2202.00484v2.pdf | Auto-ABSA: Automatic Detection of Aspects in Aspect-Based Sentiment Analysis | After transformer is proposed, lots of pre-trained language models have been come up with and sentiment analysis (SA) task has been improved. In this paper, we proposed a method that uses an auxiliary sentence about aspects that the sentence contains to help sentiment prediction. The first is aspect detection, which uses a multi-aspects detection model to predict all aspects that the sentence has. Combining the predicted aspects and the original sentence as Sentiment Analysis (SA) model's input. The second is to do out-of-domain aspect-based sentiment analysis(ABSA), train sentiment classification model with one kind of dataset and validate it with another kind of dataset. Finally, we created two baselines, they use no aspect and all aspects as sentiment classification model's input, respectively. Compare two baselines performance to our method, found that our method really makes sense. | ['Yijie Tong', 'Bolun Sun', 'Teng Wang'] | 2022-01-05 | null | null | null | null | ['aspect-based-sentiment-analysis'] | ['natural-language-processing'] | [ 1.08641475e-01 3.45435977e-01 -1.74127564e-01 -8.76790226e-01
-5.45849621e-01 -6.76176846e-01 7.86067069e-01 2.60808766e-01
-4.17701811e-01 5.13497770e-01 5.25537133e-01 -2.61784822e-01
6.59017265e-01 -1.01010501e+00 -4.30400461e-01 -3.45413983e-01
5.55038452e-01 4.61021155e-01 3.38115633e-01 -8.35109711e-01
6.15420222e-01 -2.30196953e-01 -1.31198108e+00 9.79889870e-01
3.67350101e-01 1.10894001e+00 -3.46011557e-02 7.07103133e-01
-9.42951739e-01 1.03423262e+00 -8.66968513e-01 -7.72712886e-01
6.75991774e-02 -3.26248944e-01 -9.92004454e-01 1.30756006e-01
-1.67803988e-01 2.96661824e-01 5.97366393e-01 8.35004032e-01
2.66826987e-01 -1.39216930e-02 7.03964651e-01 -1.30807137e+00
-6.63890600e-01 6.03919387e-01 -7.40172207e-01 -6.28221855e-02
4.84701902e-01 -6.22059554e-02 1.06968963e+00 -7.42798626e-01
5.67384660e-01 1.22941542e+00 8.51225674e-01 5.30781329e-01
-5.78828216e-01 -2.99519569e-01 5.29755831e-01 -1.53136477e-01
-4.33252126e-01 -9.51177776e-02 9.52050447e-01 -2.87973583e-01
1.46444559e+00 2.47372538e-01 1.03967631e+00 9.51216459e-01
4.06437278e-01 1.19387197e+00 1.49242473e+00 -4.09136772e-01
2.62141407e-01 1.04808033e+00 8.51788342e-01 1.21364981e-01
1.35512814e-01 -4.88654524e-01 -5.46624303e-01 -7.76366815e-02
-4.10756201e-01 -1.63787022e-01 2.86237121e-01 2.53622383e-01
-8.97407532e-01 8.04611623e-01 -8.07783082e-02 4.29714233e-01
-4.44362819e-01 -5.75459778e-01 8.41884971e-01 5.92786491e-01
9.63924766e-01 5.59083998e-01 -1.40853095e+00 -2.16165394e-01
-7.57769525e-01 1.61535174e-01 1.11477017e+00 8.63863766e-01
8.00696611e-01 -1.38228357e-01 -2.33561218e-01 8.48236322e-01
4.10351574e-01 6.86077297e-01 8.76509130e-01 -7.68207945e-03
5.82075119e-01 1.40897381e+00 -1.32571116e-01 -8.11135828e-01
-4.71669644e-01 -3.52491111e-01 -4.76240426e-01 1.37115479e-01
-2.48493150e-01 -4.28477585e-01 -1.02722275e+00 1.11910892e+00
3.91981304e-01 -3.89203340e-01 5.49405158e-01 5.99594355e-01
1.07992613e+00 7.57942379e-01 1.93733707e-01 -7.08672777e-02
1.70144379e+00 -1.45227158e+00 -6.73804522e-01 -6.65048122e-01
8.57431948e-01 -1.31587648e+00 1.43208158e+00 5.52822530e-01
-7.94650674e-01 -5.94257414e-01 -1.02933300e+00 8.76053870e-02
-1.14493954e+00 1.83006421e-01 8.58646572e-01 8.17504942e-01
-1.07761586e+00 1.71451226e-01 -5.44490397e-01 -4.24593717e-01
2.82725304e-01 2.61100262e-01 -4.53017592e-01 3.49083215e-01
-9.25568998e-01 8.11633050e-01 3.24860454e-01 -4.19974715e-01
-3.86356324e-01 -4.54499811e-01 -9.96503770e-01 -8.19534734e-02
1.40163705e-01 -7.09507167e-01 1.37537622e+00 -1.77831733e+00
-1.35973883e+00 9.25670385e-01 -6.87617600e-01 -5.30684218e-02
-1.80252224e-01 -4.25041109e-01 -6.39600754e-01 -2.14474067e-01
6.31546557e-01 3.11046392e-01 7.20536172e-01 -1.30808258e+00
-7.75315285e-01 -7.03086197e-01 4.73093420e-01 2.12743238e-01
-4.37259167e-01 3.70659173e-01 -4.22547132e-01 -6.00713551e-01
-1.43373638e-01 -7.74694026e-01 -5.21940887e-01 -8.28113019e-01
-7.16922104e-01 -3.43209773e-01 9.85591531e-01 -4.87881303e-01
1.24780381e+00 -1.96421576e+00 -5.39124429e-01 1.91112906e-01
-4.32278275e-01 3.23341936e-01 -4.81778175e-01 4.57330436e-01
-2.99207658e-01 2.31901497e-01 -2.77938396e-01 -7.14955568e-01
-2.18522206e-01 -5.84739894e-02 -4.69585866e-01 -4.46004778e-01
5.76167047e-01 7.81328917e-01 -5.22330165e-01 -4.75637943e-01
-2.01450124e-01 3.30714315e-01 -6.25666559e-01 3.55650783e-01
-5.84904432e-01 2.47364774e-01 -7.61547267e-01 7.38210857e-01
8.61280084e-01 2.94146314e-02 -1.35046542e-01 -2.42530003e-01
4.07523039e-05 6.80823386e-01 -9.25099730e-01 1.37457132e+00
-7.55765915e-01 4.02392030e-01 -4.75709140e-01 -8.82114589e-01
1.16846263e+00 2.80425936e-01 2.15295300e-01 -5.63225269e-01
3.70391935e-01 -1.35271221e-01 -2.42251828e-01 -7.89524853e-01
7.13050783e-01 -6.38743162e-01 -2.58463264e-01 7.81174600e-01
1.34858415e-01 -4.40318704e-01 4.52322483e-01 4.08039004e-01
8.84595931e-01 3.29526484e-01 4.64973807e-01 -2.29035035e-01
1.20236742e+00 6.76173210e-01 3.31342161e-01 2.62061089e-01
5.08727469e-02 6.49798214e-01 9.71270204e-01 -6.54368758e-01
-9.17387187e-01 -3.57540399e-01 1.57445014e-01 1.12293470e+00
-8.78795087e-02 -9.19630289e-01 -6.41680777e-01 -1.31359565e+00
-4.49716866e-01 9.18602526e-01 -1.02550936e+00 4.09326106e-02
-1.13104008e-01 -9.04868543e-01 2.37239916e-02 5.78049421e-01
5.83345413e-01 -1.27082896e+00 -1.71180725e-01 3.84574272e-02
-6.66673528e-03 -1.03103054e+00 -1.11091264e-01 3.11267734e-01
-6.82330966e-01 -1.01764584e+00 -2.15430304e-01 -8.07826042e-01
6.09878957e-01 1.83773935e-01 1.42970741e+00 -8.47886130e-02
4.32542443e-01 1.30873904e-01 -8.74957085e-01 -1.17554712e+00
-3.54484260e-01 2.02572167e-01 -3.16050023e-01 -1.10317804e-01
1.17987359e+00 -3.28505963e-01 -4.67009336e-01 3.73974219e-02
-1.05152547e+00 -1.88192725e-03 8.39623451e-01 3.40614796e-01
6.44584715e-01 9.63037927e-03 5.62103033e-01 -1.72319686e+00
9.34562266e-01 -5.63070297e-01 2.94155697e-03 1.28452644e-01
-7.56034911e-01 -2.05153902e-03 9.19062197e-01 -4.70317230e-02
-1.31990099e+00 -1.58711687e-01 -8.15280557e-01 3.97594541e-01
-3.53037119e-01 8.53011131e-01 -2.99503297e-01 6.93673134e-01
3.47455949e-01 3.54142755e-01 -1.65733974e-02 -2.53973603e-01
7.18045458e-02 1.11797678e+00 -2.70980448e-01 -9.60624516e-02
4.16320622e-01 4.70493674e-01 -1.91851869e-01 -5.75876951e-01
-1.77132034e+00 -8.30943167e-01 -6.67575896e-01 3.49949785e-02
8.02926481e-01 -1.02074993e+00 -3.04915398e-01 4.27519292e-01
-1.21713185e+00 9.27983001e-02 -3.18170279e-01 2.82765120e-01
-3.14603031e-01 2.08983704e-01 -3.14299852e-01 -9.72086191e-01
-8.79483461e-01 -1.01152265e+00 1.26669753e+00 3.23918819e-01
-4.27293092e-01 -1.06789029e+00 4.96476263e-01 5.44745982e-01
5.62909663e-01 -5.07693775e-02 6.55160189e-01 -1.30655181e+00
1.17072947e-01 -5.08802056e-01 9.14051980e-02 8.76244485e-01
2.94964254e-01 -3.29079404e-02 -1.16939771e+00 1.48325577e-01
4.39922988e-01 -3.38864446e-01 7.56072402e-01 1.84889466e-01
8.21249723e-01 -3.05857986e-01 -1.01741336e-01 2.33389378e-01
1.45731199e+00 2.83810765e-01 7.17688322e-01 1.04075730e+00
4.45483893e-01 7.23720014e-01 1.22233796e+00 1.33589968e-01
7.57599771e-01 7.12914988e-02 -2.46587452e-02 -4.50098887e-02
7.99039155e-02 -9.81844962e-02 8.83719683e-01 1.30664265e+00
1.10716060e-01 -1.18373677e-01 -5.95257521e-01 5.70538104e-01
-1.62283635e+00 -7.27722585e-01 -4.31887716e-01 1.45377851e+00
7.89075911e-01 7.13763177e-01 2.21010819e-01 2.22810075e-01
9.48472470e-02 4.27072257e-01 -3.78215402e-01 -1.26773095e+00
-4.49676007e-01 2.99161315e-01 -7.65248090e-02 3.08841825e-01
-1.33152246e+00 1.14058602e+00 5.93190432e+00 6.00723445e-01
-1.18117142e+00 2.16262043e-01 9.30289805e-01 2.68880099e-01
-7.04056501e-01 2.24009693e-01 -9.51384127e-01 3.46121520e-01
9.62109327e-01 8.45933631e-02 -4.22966093e-01 1.34140730e+00
1.80034712e-01 -4.19507504e-01 -6.73931658e-01 4.41656083e-01
6.39942408e-01 -9.29948807e-01 5.31808257e-01 -6.01508319e-01
7.26264596e-01 -1.19000956e-01 -1.21884659e-01 8.16319048e-01
9.68166590e-02 -7.02586591e-01 3.76698375e-01 4.68450278e-01
2.54941016e-01 -7.34155595e-01 1.55494106e+00 3.52638364e-01
-1.11307633e+00 3.21879566e-01 -4.77413952e-01 -3.51510942e-01
6.67993948e-02 9.05663252e-01 -7.20933676e-01 5.87309659e-01
8.60482633e-01 9.67119217e-01 -8.39668810e-01 4.42026913e-01
-5.57359159e-01 8.37600350e-01 5.08422852e-02 -4.68250722e-01
4.09132510e-01 -3.02654624e-01 5.16567767e-01 1.24648488e+00
1.61520997e-03 -9.74142998e-02 -7.25575536e-02 3.57784301e-01
2.26443902e-01 6.44946337e-01 -6.56585753e-01 -8.11308175e-02
-3.29106629e-01 1.72457170e+00 -7.44852602e-01 -8.22825670e-01
-9.03278768e-01 7.40542948e-01 4.47065197e-02 6.16429076e-02
-3.23596835e-01 -5.29046774e-01 4.55250382e-01 -2.92594600e-02
2.17046008e-01 4.50753689e-01 -9.58962023e-01 -1.47236776e+00
2.63977408e-01 -1.04964364e+00 4.70843762e-01 -1.04331219e+00
-1.36123478e+00 1.18608928e+00 -5.16761363e-01 -1.30496264e+00
-1.37857094e-01 -8.93526793e-01 -1.25767124e+00 7.91889191e-01
-1.87009954e+00 -1.63807714e+00 -5.90145588e-04 5.34302354e-01
9.72971499e-01 -5.15400290e-01 8.22828710e-01 2.12610736e-02
-2.19695047e-01 4.02565509e-01 -6.39680326e-01 2.12065443e-01
8.08808923e-01 -1.49671161e+00 5.34126878e-01 7.48049200e-01
-1.13817647e-01 5.82835734e-01 9.17072833e-01 -7.54892766e-01
-1.28716385e+00 -1.08554256e+00 1.42403877e+00 -9.77342963e-01
9.47646916e-01 -1.07996985e-01 -6.42091691e-01 8.74088645e-01
5.86730540e-01 -6.12010002e-01 1.21222675e+00 4.57568347e-01
-3.87032479e-01 -2.19683260e-01 -1.04453599e+00 2.49172509e-01
3.15266997e-01 -3.34541053e-01 -1.16977441e+00 6.91913068e-02
9.48733926e-01 1.86868608e-01 -4.91082788e-01 6.02162957e-01
5.12075663e-01 -1.02053308e+00 5.69971621e-01 -9.41884935e-01
1.10837245e+00 -4.00337785e-01 -1.56325042e-01 -1.60597837e+00
2.18023553e-01 1.66396722e-01 4.10492778e-01 1.66368067e+00
1.14981687e+00 -6.07557118e-01 8.25390995e-01 5.65541029e-01
-9.90064442e-02 -1.08867466e+00 -8.57569203e-02 -2.91072100e-01
5.05238324e-02 -7.42111921e-01 8.44414473e-01 8.68262708e-01
3.26164514e-01 1.21910977e+00 3.93388718e-02 -4.40669060e-02
-5.63414805e-02 5.28578520e-01 1.04490471e+00 -8.08865309e-01
-2.71181554e-01 -2.67754555e-01 -3.18230391e-01 -6.78008080e-01
1.41499102e-01 -7.50695825e-01 -3.27619761e-01 -1.88162851e+00
5.04492223e-01 -8.37729499e-02 -2.89909393e-01 6.59550190e-01
-5.99640667e-01 2.23040402e-01 6.95114657e-02 -2.08402202e-01
-8.26977134e-01 5.81276417e-01 1.24727309e+00 -3.68303746e-01
-2.12464571e-01 5.81946194e-01 -1.38201702e+00 1.03390920e+00
1.08049357e+00 -5.03481805e-01 -4.28422064e-01 -8.55508447e-02
8.98205698e-01 -3.90854567e-01 -5.37946343e-01 -7.32500076e-01
-3.45825404e-02 -2.10774187e-02 4.40147847e-01 -1.03555810e+00
2.44491354e-01 -8.54734242e-01 -6.04020119e-01 2.87925988e-01
-3.80752742e-01 4.00313288e-01 2.82972783e-01 5.74432611e-02
-9.07212496e-01 -4.55961496e-01 4.09588039e-01 -2.30011284e-01
-7.24856138e-01 1.16005853e-01 -3.81798625e-01 6.97705075e-02
9.12756979e-01 -7.74298236e-02 -2.58607358e-01 -3.84833813e-01
-7.18533099e-01 2.42577299e-01 2.65632898e-01 5.49322844e-01
4.48852450e-01 -9.22571719e-01 -4.26181316e-01 1.58003733e-01
4.21961904e-01 1.89439040e-02 2.08832398e-02 7.31632948e-01
-2.91573375e-01 2.54954219e-01 -3.47794965e-02 -2.61153966e-01
-1.35273349e+00 6.95433557e-01 -1.91713914e-01 -7.37961352e-01
-6.50425479e-02 7.32823730e-01 2.81596836e-02 -1.12864065e+00
-4.13668364e-01 -1.28767952e-01 -1.14846849e+00 3.18914860e-01
7.62631059e-01 -4.01536584e-01 1.19932108e-01 -7.62871802e-01
-3.52520585e-01 7.71149397e-01 -2.35224590e-01 -1.10505775e-01
1.69599962e+00 -2.42998451e-01 -5.90616167e-01 7.64570236e-01
1.31288826e+00 3.55357707e-01 -3.90614033e-01 4.17828597e-02
1.75776660e-01 -2.91026473e-01 -2.20444292e-01 -1.19231486e+00
-1.08177233e+00 8.17517638e-01 1.76112741e-01 5.45409203e-01
1.33396530e+00 1.53345361e-01 9.54996288e-01 5.23461044e-01
3.89869921e-02 -1.23431861e+00 1.08744510e-01 9.86832380e-01
6.11916065e-01 -1.73492587e+00 -6.73464220e-03 -4.50687408e-01
-1.44723499e+00 9.66001511e-01 8.75118971e-01 -3.12373340e-01
1.20839167e+00 4.54464257e-01 7.36215532e-01 -4.59412962e-01
-1.12770677e+00 -3.50782573e-01 2.76832014e-01 5.80165923e-01
7.47285485e-01 1.73325483e-02 -5.86823225e-01 1.43864024e+00
-9.31162417e-01 -7.71421194e-02 5.89747727e-01 9.72944260e-01
-4.45814520e-01 -1.32036316e+00 -9.24770609e-02 8.63941371e-01
-9.25897956e-01 -3.79676312e-01 -5.68281710e-01 5.03087521e-01
5.97917587e-02 1.26337576e+00 -1.72744170e-01 -5.22404790e-01
6.98032141e-01 2.45135322e-01 -5.09783447e-01 -9.34176981e-01
-1.15847766e+00 -5.63403741e-02 4.35269237e-01 -4.78914142e-01
-7.80500054e-01 -5.42974412e-01 -9.69815671e-01 1.03260040e-01
-2.14153871e-01 5.01611173e-01 1.11134887e+00 1.29262781e+00
3.60374719e-01 3.55798811e-01 1.01227117e+00 -1.80725813e-01
8.05573985e-02 -1.27703905e+00 -4.75383937e-01 6.78629041e-01
2.66730666e-01 -4.84687574e-02 -5.20157456e-01 1.35589048e-01] | [11.402005195617676, 6.700481414794922] |
91b81324-e99e-46d0-8fc2-8f25c20e0d5d | covid-19-and-misinformation-a-large-scale | null | null | https://aclanthology.org/2021.acl-srw.13 | https://aclanthology.org/2021.acl-srw.13.pdf | COVID-19 and Misinformation: A Large-Scale Lexical Analysis on Twitter | Social media is often used by individuals and organisations as a platform to spread misinformation. With the recent coronavirus pandemic we have seen a surge of misinformation on Twitter, posing a danger to public health. In this paper, we compile a large COVID-19 Twitter misinformation corpus and perform an analysis to discover patterns with respect to vocabulary usage. Among others, our analysis reveals that the variety of topics and vocabulary usage are considerably more limited and negative in tweets related to misinformation than in randomly extracted tweets. In addition to our qualitative analysis, our experimental results show that a simple linear model based only on lexical features is effective in identifying misinformation-related tweets (with accuracy over 80{\%}), providing evidence to the fact that the vocabulary used in misinformation largely differs from generic tweets. | ['David Rogers', 'Alun Preece', 'Jose Camacho-Collados', 'Dimosthenis Antypas'] | 2021-08-01 | null | null | null | acl-2021-5 | ['lexical-analysis'] | ['natural-language-processing'] | [ 4.67609614e-02 1.18432581e-01 -5.81042886e-01 -3.29356007e-02
-4.85204518e-01 -7.47824728e-01 1.19454467e+00 9.88876343e-01
-6.78086877e-01 7.17885196e-01 9.19119298e-01 -5.27139723e-01
2.27678478e-01 -8.79485130e-01 -2.85189509e-01 -3.96881014e-01
-1.28558651e-01 3.69160622e-01 1.26056090e-01 -6.88730538e-01
4.93858904e-01 1.76185697e-01 -1.25750470e+00 1.99764490e-01
4.49414194e-01 5.27680933e-01 -1.26416653e-01 3.46456379e-01
-8.62189114e-01 8.92696679e-01 -9.29436862e-01 -3.28074455e-01
-3.02067995e-01 -1.26031369e-01 -8.09047341e-01 -7.86783844e-02
6.90916926e-02 -1.02920227e-01 -2.24130958e-01 1.08398175e+00
3.41707230e-01 -3.75534236e-01 6.95308328e-01 -8.12835813e-01
1.27209783e-01 8.06362987e-01 -4.28012282e-01 8.16621244e-01
6.30132318e-01 -1.05289869e-01 6.49177790e-01 -5.18370867e-01
1.11960864e+00 1.27427244e+00 7.71212339e-01 -1.21115848e-01
-9.00575399e-01 -8.14094245e-01 -1.10561706e-01 -3.97572875e-01
-1.53571391e+00 -3.58560413e-01 5.41647673e-01 -9.58810389e-01
9.24680769e-01 5.11131525e-01 5.98975599e-01 1.22032177e+00
4.44346547e-01 1.76008463e-01 1.17869544e+00 -1.04549117e-01
-6.07184917e-02 6.46525621e-01 2.28422448e-01 2.51821458e-01
6.65544093e-01 -1.43989056e-01 -3.47683519e-01 -8.40937555e-01
-7.48452470e-02 3.06767941e-01 -1.56279236e-01 5.93813241e-01
-1.19472730e+00 1.46570396e+00 2.12247878e-01 9.00188327e-01
-5.50183415e-01 -3.70513439e-01 6.40127063e-01 4.11969304e-01
1.18741882e+00 5.44738591e-01 -2.87387490e-01 -3.54552835e-01
-7.54902244e-01 3.43757421e-01 1.04906821e+00 5.62920153e-01
8.26812923e-01 -3.35276335e-01 2.09505409e-01 8.62855315e-01
2.23151922e-01 8.49602222e-01 5.94006121e-01 -1.31364986e-01
6.34753048e-01 7.64717996e-01 1.51302576e-01 -2.04591894e+00
-9.00457442e-01 -4.80865330e-01 -6.16732180e-01 -6.95345819e-01
1.94416270e-01 -4.30985898e-01 -4.39988285e-01 1.42046738e+00
2.82194346e-01 -1.27022624e-01 -3.91324401e-01 2.97111839e-01
1.00159907e+00 6.34952486e-01 4.89003718e-01 -5.85355639e-01
1.42363107e+00 2.09229723e-01 -1.05314195e+00 -2.98865616e-01
9.37626719e-01 -9.68850136e-01 6.01002395e-01 -3.06653287e-02
-5.39604187e-01 1.13609269e-01 -6.82340086e-01 3.85094732e-01
-1.04662597e+00 -1.04916310e+00 2.23140627e-01 9.22900856e-01
-7.55998909e-01 2.71707863e-01 -2.27310851e-01 -8.17134202e-01
4.47655886e-01 -2.61895120e-01 -5.93682863e-02 6.83082864e-02
-1.41218626e+00 9.14456248e-01 6.21918082e-01 -7.42999673e-01
-2.76895761e-01 -5.92698097e-01 -6.75979137e-01 -6.80353284e-01
3.36402625e-01 -3.11319917e-01 8.36035371e-01 -7.72307158e-01
-6.57278299e-01 1.05716538e+00 -2.24081814e-01 -4.27977681e-01
4.18378562e-01 1.67609006e-01 -9.85800982e-01 2.39989936e-01
4.94870454e-01 2.70312637e-01 8.56489599e-01 -9.48481143e-01
-7.33560681e-01 -2.98955083e-01 -8.99849534e-02 -2.44837046e-01
-4.41124439e-01 6.46786392e-01 1.13584474e-01 -9.03315485e-01
-2.66920030e-01 -7.21839607e-01 -1.33603394e-01 -9.48864937e-01
-3.73868853e-01 -2.69246161e-01 7.15748012e-01 -6.42245531e-01
2.01484203e+00 -2.01267934e+00 -5.17145157e-01 4.50992942e-01
7.36058235e-01 8.81962180e-02 5.16684175e-01 1.09986115e+00
2.86186218e-01 9.13634896e-01 1.77415349e-02 7.71544650e-02
-2.17136472e-01 1.26199111e-01 -5.86372912e-01 8.62322807e-01
-1.28098503e-01 8.03977430e-01 -1.14247072e+00 -5.13841927e-01
3.02160196e-02 4.68769312e-01 -3.00671935e-01 -1.61707953e-01
-1.46851137e-01 2.36296326e-01 -6.92867756e-01 3.65072906e-01
5.80579042e-01 -6.02821529e-01 1.65148973e-01 1.20717481e-01
-5.23439467e-01 7.12907493e-01 -3.67091656e-01 6.12242639e-01
-2.25704849e-01 9.16722476e-01 -1.58061415e-01 -3.06374162e-01
5.84038258e-01 3.15393925e-01 4.44153190e-01 -5.35655320e-01
6.98426783e-01 -2.76305582e-02 -4.22001451e-01 -6.00729167e-01
7.11997449e-01 -2.62837857e-01 -5.10963738e-01 6.85149014e-01
-4.60556626e-01 7.49569386e-02 -1.05624693e-02 2.93192744e-01
7.78542280e-01 -1.10604799e+00 8.16268265e-01 -6.02817655e-01
3.65771592e-01 5.40518582e-01 -5.60734347e-02 7.08908498e-01
1.22699114e-02 1.29436865e-01 2.61689574e-01 -4.24300134e-01
-7.33613789e-01 -4.44301188e-01 -4.83766377e-01 1.01794994e+00
3.27893198e-02 -7.98920035e-01 -6.73569739e-01 -4.71355170e-01
6.05785772e-02 8.30834091e-01 -7.57555962e-01 3.54528308e-01
-4.55440879e-01 -1.09622049e+00 5.39110243e-01 -3.17887038e-01
3.30178887e-01 -8.05230558e-01 -5.92241466e-01 3.98897022e-01
-5.46452761e-01 -1.09106040e+00 -3.97387713e-01 -3.21130067e-01
-5.26175618e-01 -1.08179605e+00 -4.09663975e-01 -3.67232382e-01
6.08025610e-01 6.10884905e-01 1.12833965e+00 3.21257740e-01
-3.75281908e-02 2.82249719e-01 -4.86614555e-01 -8.88379991e-01
-7.66160429e-01 2.52300173e-01 -2.69920193e-02 -1.30177423e-01
6.36680424e-01 -1.62322491e-01 -4.51544493e-01 2.03044340e-01
-1.32554197e+00 -4.17516649e-01 1.54416505e-02 2.25193024e-01
-1.44575566e-01 1.96429074e-01 5.75134397e-01 -1.37592459e+00
8.10736895e-01 -1.49881482e+00 -1.75240695e-01 -6.56022787e-01
-6.49485052e-01 -4.22186047e-01 1.08469054e-01 -2.68704057e-01
-5.39207637e-01 -7.87776768e-01 -3.11132163e-01 3.74377251e-01
-3.29637557e-01 9.72283542e-01 7.04799294e-01 1.30734649e-02
8.74792874e-01 3.77514996e-02 2.37363309e-01 -4.60763246e-01
1.68908611e-02 1.11202180e+00 -3.25753927e-01 5.13088524e-01
7.81019330e-01 8.88145447e-01 -5.35108447e-01 -1.78536057e+00
-8.44219506e-01 -1.01059389e+00 -5.60960844e-02 -3.17297459e-01
1.00404823e+00 -8.14718544e-01 -7.69840121e-01 4.09146249e-01
-9.10445511e-01 6.23786747e-02 2.43099302e-01 2.66534448e-01
6.75832331e-02 2.90674925e-01 -5.08048892e-01 -7.24910796e-01
-3.51087824e-02 -6.73333883e-01 5.36150634e-01 -2.99391121e-01
-1.03925979e+00 -1.62871253e+00 3.70718896e-01 3.27545911e-01
1.03267813e+00 6.26033962e-01 6.52932048e-01 -1.35284221e+00
8.26664343e-02 -3.20553958e-01 -3.32920015e-01 -5.40722013e-01
5.68994522e-01 -4.09585476e-01 -8.46669853e-01 -3.28485340e-01
1.67251796e-01 -1.11179329e-01 6.30726576e-01 1.94193676e-01
3.95260692e-01 -1.01349080e+00 -6.98280632e-01 -1.28311701e-02
1.43337488e+00 -3.32328156e-02 1.94366857e-01 8.16666603e-01
3.99107963e-01 8.36056769e-01 2.40121379e-01 6.91520810e-01
4.93703693e-01 5.61217725e-01 2.48615608e-01 1.25113696e-01
3.28579545e-01 -4.56325024e-01 3.17100376e-01 9.56484556e-01
3.57565731e-01 -3.28032792e-01 -1.26050234e+00 5.43392479e-01
-1.17592299e+00 -1.11931598e+00 -3.04122537e-01 1.99967957e+00
7.61607230e-01 2.02937797e-01 5.13415992e-01 -8.27243999e-02
6.41560435e-01 7.17782736e-01 2.80823022e-01 -3.68661225e-01
-1.58369005e-01 -2.84479052e-01 1.04087591e+00 8.43157530e-01
-1.11913693e+00 6.01808250e-01 7.71452665e+00 7.73612678e-01
-1.33797920e+00 2.60606766e-01 4.04869348e-01 3.06040674e-01
-6.40039325e-01 -5.04862785e-01 -9.52353895e-01 8.02438140e-01
1.28335214e+00 -4.48601037e-01 1.02885105e-01 3.53690535e-01
4.54573572e-01 -1.50983319e-01 -6.00523949e-02 6.83920205e-01
3.32781583e-01 -1.61890471e+00 2.07351178e-01 3.58891129e-01
5.50542951e-01 5.61595380e-01 -3.17083411e-02 7.29911553e-04
1.39572442e-01 -9.44548845e-01 5.32203138e-01 2.28757769e-01
5.25148630e-01 -6.80350959e-01 9.97504056e-01 7.42147267e-01
-6.82100713e-01 3.31986725e-01 -3.26805934e-02 -2.11941883e-01
4.16699290e-01 7.95394599e-01 -1.53691375e+00 4.36922349e-02
4.90649521e-01 8.77940118e-01 -6.47268534e-01 6.50112987e-01
2.85908014e-01 9.34663355e-01 -2.66054541e-01 -4.50842232e-01
5.21031439e-01 1.19836502e-01 9.54017341e-01 1.96689641e+00
-3.75951603e-02 -1.00812621e-01 1.09614119e-01 3.75671715e-01
9.82112512e-02 5.05710542e-01 -1.21064425e+00 -6.96853280e-01
3.91240090e-01 8.30356777e-01 -9.19722021e-01 -4.68566835e-01
-3.96528721e-01 3.05298030e-01 -7.49025494e-02 1.58394784e-01
-4.71486717e-01 -2.58849859e-01 5.15283048e-01 6.59172893e-01
4.29946519e-02 -2.07233459e-01 -1.14957206e-01 -8.76137793e-01
-5.79236567e-01 -8.21874857e-01 3.61379355e-01 1.30147208e-02
-1.36370254e+00 6.83445334e-01 2.46561885e-01 -1.01423633e+00
-3.93747747e-01 -1.53164670e-01 -2.01842755e-01 5.88078499e-01
-1.44678307e+00 -6.06150746e-01 -7.50884861e-02 4.45837677e-01
2.72169769e-01 6.43026829e-02 7.25910544e-01 3.59704882e-01
-1.05600558e-01 2.69160390e-01 -1.41186165e-02 1.53294116e-01
5.11271179e-01 -7.08426118e-01 3.66303504e-01 4.58127223e-02
-2.70358950e-01 7.39205122e-01 1.29958010e+00 -1.07458282e+00
-9.82570946e-01 -1.29552484e+00 1.76638436e+00 -7.11546183e-01
1.25769305e+00 -2.19399795e-01 -8.46458793e-01 6.68795824e-01
5.39833188e-01 -8.65264893e-01 1.06269515e+00 -4.68716472e-02
-2.95594603e-01 5.95887601e-01 -1.31133282e+00 5.32550514e-01
7.30641127e-01 -5.54467022e-01 -7.54746675e-01 6.36704206e-01
8.55014920e-01 2.04461459e-02 -8.09405625e-01 -6.29927069e-02
2.52638668e-01 -7.69130051e-01 7.57349491e-01 -6.41309977e-01
6.77690059e-02 6.76957965e-02 -1.67788342e-01 -1.18311763e+00
-2.06378147e-01 -8.29535484e-01 5.20025790e-01 1.12168026e+00
5.56053996e-01 -1.21483040e+00 3.47702384e-01 1.42721787e-01
3.87934446e-01 -1.01083755e-01 -7.92698026e-01 -4.15533006e-01
1.69755816e-01 -7.43169427e-01 3.89153868e-01 1.41332567e+00
3.72830927e-01 2.83040702e-01 -2.69131780e-01 -2.31604546e-01
3.81634593e-01 -3.49924326e-01 5.79088748e-01 -1.26959801e+00
3.32554251e-01 -6.07981265e-01 -3.39536667e-01 -6.38136446e-01
-1.06265746e-01 -7.45784819e-01 -3.39675337e-01 -1.09895718e+00
2.98835844e-01 -4.74782676e-01 2.50832975e-01 -1.64351523e-01
1.19774587e-01 5.61191082e-01 -3.08777820e-02 4.63293403e-01
-4.82698172e-01 -2.81911105e-01 8.54099393e-01 -2.78825343e-01
-2.97719121e-01 -4.11509909e-02 -9.18081343e-01 9.81536686e-01
8.11628401e-01 -7.65002012e-01 6.60654828e-02 2.17775464e-01
1.16690373e+00 -3.46413761e-01 -2.17566788e-02 -3.47051352e-01
-5.14853187e-02 -1.68746084e-01 -1.59940481e-01 -7.15064108e-01
-3.70425470e-02 -7.60378420e-01 2.90248662e-01 6.87292039e-01
-2.91699052e-01 3.02513808e-01 3.24496806e-01 5.60154438e-01
-2.70855278e-01 9.13506523e-02 4.87540841e-01 -1.10011660e-01
-2.88931251e-01 6.78343922e-02 -1.34494364e+00 6.14691973e-01
7.37159014e-01 7.19810352e-02 -5.80095649e-01 -7.32326627e-01
-4.93956685e-01 -8.39122012e-03 5.93129039e-01 4.67424124e-01
1.91291258e-01 -1.13512874e+00 -8.50723267e-01 -3.90691049e-02
4.73411083e-01 -6.32850468e-01 3.41108888e-02 1.11769891e+00
-5.83795130e-01 8.75794590e-01 1.79341525e-01 -4.62971538e-01
-1.05159652e+00 6.76423728e-01 -1.22213297e-01 -8.18768442e-02
-5.42144418e-01 3.26429099e-01 -1.29473269e-01 -2.55396575e-01
3.23645025e-02 -1.96714848e-01 -6.26989126e-01 9.14242625e-01
1.08680868e+00 7.00483561e-01 -6.45093471e-02 -1.37374103e+00
-4.44040447e-01 3.06413114e-01 -5.10823615e-02 7.94909336e-03
1.07665682e+00 -7.08471060e-01 -3.65425706e-01 9.42899764e-01
1.72202659e+00 5.45239031e-01 5.69240339e-02 -2.99656570e-01
3.42753708e-01 -5.72633684e-01 6.80422932e-02 -4.85631108e-01
-4.49851632e-01 3.12540829e-01 9.96720865e-02 1.19921374e+00
4.79555458e-01 3.46914649e-01 8.63985360e-01 4.26058359e-02
3.28448534e-01 -8.20932269e-01 -3.31010193e-01 6.58623934e-01
8.23261559e-01 -1.27824390e+00 -5.44312820e-02 -4.05297339e-01
-6.55269861e-01 8.74022961e-01 -4.66858655e-01 1.34431556e-01
1.45254612e+00 1.51030660e-01 4.17412370e-01 -6.98656738e-01
-3.36193323e-01 -2.07608834e-01 1.28792211e-01 3.29303414e-01
5.50600708e-01 2.23288372e-01 -8.86970639e-01 1.20449528e-01
-4.16999638e-01 -1.99137479e-01 6.46473765e-01 9.19238687e-01
-8.67375731e-01 -7.33398199e-01 -4.59950477e-01 6.06288612e-01
-1.27776361e+00 -4.55976464e-02 -6.83476269e-01 8.35413277e-01
3.18537978e-03 1.43310785e+00 3.15468252e-01 -5.65037727e-01
-1.22619398e-01 1.41075514e-02 -4.25416887e-01 -5.99863648e-01
-9.81202304e-01 4.73088212e-02 6.32638097e-01 -5.19954145e-01
-6.76921487e-01 -7.69624591e-01 -9.76648390e-01 -7.82552183e-01
-1.97657987e-01 2.47982606e-01 6.47604048e-01 1.16818190e+00
3.62941563e-01 6.25907332e-02 7.07409859e-01 -3.92531246e-01
-2.42155753e-02 -1.05620956e+00 -4.10869926e-01 6.21750355e-01
8.25801373e-01 -4.43164736e-01 -7.42393315e-01 -2.25299090e-01] | [8.460530281066895, 9.820080757141113] |
84810460-02d9-412f-88b7-e0427d97d52e | artificial-intelligence-distinguishes-covid | null | null | https://pubs.rsna.org/doi/10.1148/radiol.2020200905 | https://pubs.rsna.org/doi/pdf/10.1148/radiol.2020200905 | Artificial Intelligence Distinguishes COVID-19 from Community Acquired Pneumonia on Chest CT | Background
Coronavirus disease has widely spread all over the world since the beginning of 2020. It is desirable to develop automatic and accurate detection of COVID-19 using chest CT.
Purpose
To develop a fully automatic framework to detect COVID-19 using chest CT and evaluate its performances.
Materials and Methods
In this retrospective and multi-center study, a deep learning model, COVID-19 detection neural network (COVNet), was developed to extract visual features from volumetric chest CT exams for the detection of COVID-19. Community acquired pneumonia (CAP) and other non-pneumonia CT exams were included to test the robustness of the model. The datasets were collected from 6 hospitals between August 2016 and February 2020. Diagnostic performance was assessed by the area under the receiver operating characteristic curve (AUC), sensitivity and specificity.
Results
The collected dataset consisted of 4356 chest CT exams from 3,322 patients. The average age is 49±15 years and there were slightly more male patients than female (1838 vs 1484; p-value=0.29). The per-exam sensitivity and specificity for detecting COVID-19 in the independent test set was 114 of 127 (90% [95% CI: 83%, 94%]) and 294 of 307 (96% [95% CI: 93%, 98%]), respectively, with an AUC of 0.96 (p-value<0.001). The per-exam sensitivity and specificity for detecting CAP in the independent test set was 87% (152 of 175) and 92% (239 of 259), respectively, with an AUC of 0.95 (95% CI: 0.93, 0.97).
Conclusions
A deep learning model can accurately detect COVID-19 and differentiate it from community acquired pneumonia and other lung diseases. | ['Shiqin Zhang', 'Daliang Liu', 'Kunlin Cao', 'Yi Lu', 'Qi Song', 'Lixin Qin', 'Jun Xia', 'Bin Kong', 'Qizhong Xu', 'Guisheng Wang', 'Youbing Yin', 'Lin Li', 'Junjie Bai', 'Zhenghan Fang', 'Zeguo Xu', 'Xisheng Fang', 'Xin Wang', 'Juan Xia'] | 2020-02-19 | null | null | null | radiology-2020-2 | ['covid-19-image-segmentation'] | ['computer-vision'] | [-8.48772451e-02 -3.82039666e-01 -1.14027917e-01 1.09120579e-02
-6.79344833e-01 -6.73056006e-01 1.61764815e-01 4.82963592e-01
-7.33435094e-01 7.06217170e-01 -1.64642045e-03 -6.99766934e-01
-1.40728250e-01 -7.76121199e-01 -5.49536586e-01 -7.35080540e-01
-3.05241674e-01 8.10423434e-01 2.51691520e-01 7.98367739e-01
-4.40812320e-01 6.24817491e-01 -8.82828414e-01 4.73688871e-01
7.54226387e-01 8.34709167e-01 5.05435884e-01 1.30971026e+00
3.47564578e-01 7.40242183e-01 -5.41961849e-01 -6.18604310e-02
2.16889963e-01 -2.98373371e-01 -3.71422499e-01 -2.24446908e-01
1.73340812e-01 -9.31862354e-01 -1.79072931e-01 3.77882361e-01
7.22166479e-01 -4.28856462e-01 1.39044321e+00 -9.75991428e-01
-2.04466924e-01 -1.93418220e-01 -5.69127202e-01 6.04022503e-01
-3.12869608e-01 3.93708736e-01 5.78444302e-01 -7.80817747e-01
5.83635151e-01 5.86073816e-01 1.05456197e+00 4.02368754e-01
-5.99986970e-01 -7.28643835e-01 -4.92603868e-01 -1.64865792e-01
-1.28331769e+00 5.57338536e-01 -1.53554291e-01 -1.11251485e+00
7.98006892e-01 2.58306205e-01 1.07825935e+00 9.03472900e-01
6.38212621e-01 3.55695754e-01 9.24598932e-01 -6.31095991e-02
-1.28141746e-01 1.71085313e-01 -1.17475856e-02 8.63348544e-01
1.00946271e+00 4.27274853e-02 6.08877957e-01 -6.43321931e-01
1.11592448e+00 6.21228874e-01 -2.22424522e-01 -3.50141883e-01
-1.24928141e+00 1.29042399e+00 4.49732840e-01 2.77782381e-01
-5.56911349e-01 -3.23181182e-01 7.09695876e-01 -2.34550819e-01
-1.57645941e-01 6.96678236e-02 -5.69841683e-01 2.83224583e-01
-6.16324008e-01 1.01661794e-01 4.65337545e-01 5.23261845e-01
-9.73230302e-02 -8.34322199e-02 -5.27326763e-01 7.97766030e-01
3.59939903e-01 1.22534621e+00 6.05096221e-01 -6.81865215e-01
3.53437990e-01 4.43749070e-01 1.33569866e-01 -5.56869030e-01
-9.22688901e-01 -4.85685319e-01 -1.33766925e+00 -3.10715377e-01
3.19886476e-01 -7.36590028e-01 -1.02885222e+00 1.40488887e+00
8.19665194e-03 -1.99939221e-01 -3.01778186e-02 8.36631298e-01
9.42236006e-01 5.83400548e-01 3.70510250e-01 -4.69497025e-01
1.87346244e+00 -6.27406657e-01 -1.97735295e-01 1.22956084e-02
8.44456434e-01 -3.73329967e-01 7.29926109e-01 7.62701109e-02
-6.49690688e-01 -3.01219523e-01 -9.27865684e-01 6.52139962e-01
-1.62854604e-02 4.15749043e-01 -9.96330660e-03 7.41680741e-01
-8.33831966e-01 -3.20272148e-02 -9.90695119e-01 -6.22839093e-01
8.09235096e-01 2.03792974e-01 -1.70514092e-01 -2.52044916e-01
-1.07287097e+00 7.85654247e-01 2.98177242e-01 -3.24290484e-01
-1.01681125e+00 -7.10969806e-01 -3.39146763e-01 3.37364376e-02
3.83009808e-03 -1.45190728e+00 1.15861332e+00 -3.97682846e-01
-2.25003466e-01 1.13324976e+00 -1.88676771e-02 -3.17549914e-01
8.00750434e-01 -1.62628278e-01 -2.70278901e-01 5.68293452e-01
2.75482506e-01 3.92227173e-01 4.78544563e-01 -1.13977456e+00
-8.50000441e-01 -4.53229785e-01 -4.91353214e-01 -5.49714528e-02
-1.25843599e-01 3.00704360e-01 -1.93016559e-01 -6.08996034e-01
-4.77041453e-01 -1.19318807e+00 -2.56306380e-01 -2.55958140e-02
-1.25565737e-01 -3.01996827e-01 9.22052562e-01 -1.01985443e+00
8.43337357e-01 -1.95492101e+00 -9.53967571e-01 1.75963789e-01
1.00297058e+00 8.72963130e-01 2.69047678e-01 1.72536269e-01
-1.72449276e-01 3.77940118e-01 -5.83137751e-01 2.82771111e-01
-5.23234785e-01 -9.19084176e-02 1.94443867e-01 6.21387303e-01
2.73342907e-01 1.13968360e+00 -7.32483327e-01 -7.60930419e-01
4.44027960e-01 8.16402197e-01 -3.49443138e-01 5.71025670e-01
3.05982888e-01 4.69147176e-01 -6.07404947e-01 3.92039806e-01
6.72049165e-01 -8.70628119e-01 2.43381515e-01 1.97518751e-01
9.71470475e-02 -3.19371521e-01 -6.54480159e-01 6.79495931e-01
-2.77920872e-01 8.89583826e-01 1.00951763e-02 -6.23664737e-01
7.24511683e-01 6.73111141e-01 7.79262304e-01 -4.06461269e-01
4.13617045e-01 2.86122113e-01 2.50895560e-01 -8.94653499e-01
-3.92206162e-01 -4.74746734e-01 2.53910929e-01 5.63952684e-01
-5.54839909e-01 -6.58471212e-02 -1.77111644e-02 9.75831077e-02
1.18984389e+00 -6.69893146e-01 5.16990364e-01 -1.22933730e-01
4.44820166e-01 2.25276291e-01 3.34779710e-01 1.08748817e+00
-5.24041176e-01 1.02496195e+00 4.56383169e-01 -5.31179368e-01
-9.42368209e-01 -1.55756545e+00 -5.59317470e-01 4.27728564e-01
-5.64254522e-01 1.44989923e-01 -8.28824282e-01 -8.88433039e-01
-4.13804017e-02 3.55091184e-01 -4.26042110e-01 4.07961383e-02
-8.08721006e-01 -9.22695100e-01 6.93051040e-01 1.06218576e+00
4.91715610e-01 -1.15920579e+00 -1.25065362e+00 4.69520949e-02
-3.28075707e-01 -9.11994100e-01 -2.62050748e-01 8.70718434e-02
-9.37014818e-01 -1.47557414e+00 -1.33695126e+00 -9.95622694e-01
4.67543811e-01 1.54995188e-01 1.10568821e+00 3.60247195e-01
-7.95398653e-01 4.03815746e-01 -3.69245857e-02 -6.71305120e-01
-4.81161803e-01 -1.92952588e-01 1.17406867e-01 -6.13029718e-01
4.69663292e-01 -9.37829912e-03 -1.04447412e+00 1.95158407e-01
-8.37732375e-01 -3.65406275e-01 9.46657538e-01 7.44676411e-01
6.33195877e-01 -3.03041995e-01 5.92366755e-01 -6.43857777e-01
4.17370617e-01 -6.55853689e-01 -3.62884402e-01 9.48735885e-03
-6.44621909e-01 -6.49294615e-01 4.83932793e-01 -2.22588524e-01
-6.03608131e-01 -8.16147625e-02 7.12872371e-02 -6.23992324e-01
-2.71051198e-01 2.78361589e-01 6.48857296e-01 7.61083782e-01
5.33110976e-01 6.61230981e-02 1.41598150e-01 -3.25276196e-01
-3.91793340e-01 9.72488225e-01 4.67319965e-01 -8.95927474e-02
6.33649647e-01 4.10731971e-01 5.20522706e-02 -9.03620481e-01
-6.19208872e-01 -9.17273998e-01 -6.36245668e-01 -1.25998065e-01
1.52031112e+00 -1.15432644e+00 -7.85587370e-01 4.95880336e-01
-1.05339575e+00 -1.70464963e-01 -3.45934667e-02 1.09320390e+00
-3.10622245e-01 4.23137814e-01 -7.55244493e-01 -5.48652053e-01
-1.06037176e+00 -1.16047871e+00 8.16223443e-01 1.14141777e-01
-5.13576210e-01 -9.71070886e-01 4.56432432e-01 4.69750792e-01
4.63769943e-01 6.10920131e-01 1.27333009e+00 -1.12863612e+00
-2.26694688e-01 -3.94242644e-01 -1.03069437e+00 5.54782748e-01
4.18381542e-01 2.01092333e-01 -7.92572260e-01 -5.37565649e-01
1.37583300e-01 -6.11035414e-02 7.18809962e-01 1.01903236e+00
1.06434369e+00 -1.09246761e-01 -7.35860109e-01 4.62847888e-01
1.62376547e+00 8.93413961e-01 4.22398984e-01 1.80654958e-01
7.73036778e-01 2.22474426e-01 2.32627079e-01 5.34543574e-01
1.41456276e-01 3.41278054e-02 4.44678217e-01 -2.51819938e-01
4.55803052e-02 1.94472268e-01 -1.01535469e-01 4.97216582e-01
-3.44494849e-01 -4.55953449e-01 -1.47752964e+00 7.71107316e-01
-1.14985514e+00 -7.05818653e-01 -7.54138887e-01 1.96554208e+00
2.96349794e-01 2.87731853e-03 3.39651436e-01 -1.81723416e-01
1.06902611e+00 -1.65376812e-01 -5.34517109e-01 -5.38847387e-01
1.41176209e-01 2.46720612e-01 4.76538271e-01 -1.89907476e-01
-1.30089533e+00 -1.43415257e-01 5.78188133e+00 2.03281730e-01
-1.12875175e+00 8.52054134e-02 7.92335033e-01 5.87531403e-02
3.91117662e-01 -7.47473836e-01 -3.06190848e-01 4.48892951e-01
8.97826850e-01 2.62714289e-02 -2.48469248e-01 8.57526243e-01
2.12276384e-01 -1.37679270e-02 -7.45768130e-01 7.37481534e-01
1.66396767e-01 -1.32160723e+00 -1.46356210e-01 6.41158149e-02
8.62451851e-01 6.03420317e-01 2.86192354e-02 3.59345138e-01
8.65027234e-02 -1.14017618e+00 5.66724464e-02 2.31083423e-01
1.45594847e+00 -6.12237513e-01 1.43988347e+00 3.13122004e-01
-1.18990159e+00 1.00342348e-01 -2.95525417e-02 2.49316871e-01
1.27838343e-01 5.05674779e-01 -1.64055490e+00 1.16799429e-01
1.15925097e+00 2.40144521e-01 -4.06429946e-01 1.14211762e+00
-4.81356271e-02 8.16748083e-01 -3.78929466e-01 -2.87617832e-01
3.15520376e-01 2.03406483e-01 4.40004557e-01 1.60988212e+00
3.05089593e-01 1.47642538e-01 -9.33036730e-02 7.15549231e-01
-1.15515245e-02 2.78980792e-01 -8.15555871e-01 4.19102274e-02
2.50100493e-01 1.08745837e+00 -8.39043796e-01 -5.62664211e-01
-6.36400342e-01 2.53968716e-01 -2.57465720e-01 1.65051311e-01
-1.14898121e+00 -3.95705968e-01 2.37347096e-01 4.60643977e-01
5.28367817e-01 1.99125960e-01 -4.28825855e-01 -6.28518879e-01
-1.91033825e-01 -6.99429929e-01 8.80513608e-01 -7.91687191e-01
-1.35643470e+00 4.55363244e-01 7.39433542e-02 -1.23642337e+00
-2.15407550e-01 -8.10040712e-01 -7.25076616e-01 8.89279485e-01
-9.86658454e-01 -6.95746779e-01 -5.60999393e-01 2.94061691e-01
2.40126029e-01 -1.60713360e-01 8.87078822e-01 -2.79002003e-02
-3.46390992e-01 4.39784437e-01 3.95394474e-01 6.12835765e-01
5.44217706e-01 -9.20458019e-01 -4.62233787e-03 3.90783727e-01
-7.88661540e-01 7.45133936e-01 -1.46970689e-01 -8.27778280e-01
-5.77672839e-01 -1.55032670e+00 9.43984509e-01 -5.47721267e-01
2.57122070e-01 2.16204062e-01 -7.75475681e-01 5.96586585e-01
4.90844250e-03 -8.55727419e-02 8.33992302e-01 -7.10970521e-01
-2.87016392e-01 3.97017539e-01 -1.43973267e+00 2.37290442e-01
4.83858079e-01 -2.86432475e-01 -7.25903928e-01 3.64682376e-01
7.00996637e-01 -2.65881717e-01 -1.03631961e+00 1.00664783e+00
9.04247344e-01 -9.84318852e-01 1.23003733e+00 -6.38837457e-01
4.29407477e-01 -3.09241954e-02 -7.11014792e-02 -5.60189247e-01
-5.74644148e-01 3.08994770e-01 2.66453028e-01 3.98634821e-01
3.07710409e-01 -8.79461586e-01 7.04437673e-01 4.74197306e-02
-3.72891054e-02 -1.11931884e+00 -7.87651300e-01 -5.46989143e-01
3.57920378e-01 -2.11179525e-01 4.18886065e-01 9.05698895e-01
-7.95131743e-01 4.03633304e-02 1.90874830e-01 1.73156664e-01
2.10906148e-01 -5.96596226e-02 3.16018254e-01 -1.35614502e+00
-5.95447049e-02 -4.35123205e-01 5.39676398e-02 -2.41320759e-01
-5.85073054e-01 -7.24688888e-01 -3.91708851e-01 -2.34732771e+00
1.06147850e+00 -3.47877741e-01 -5.99084139e-01 3.62314641e-01
-2.08622143e-01 9.99578908e-02 2.84827240e-02 4.00042027e-01
-3.62162385e-03 -8.56331959e-02 1.36441565e+00 -1.88931242e-01
-3.11533418e-02 3.42894703e-01 -2.56390661e-01 7.25873172e-01
1.12802887e+00 -7.04404116e-01 -2.41928175e-01 -9.80618447e-02
-1.31819397e-01 2.64323711e-01 5.79408944e-01 -9.18648899e-01
-5.67717314e-01 -1.44560486e-01 8.47629786e-01 -1.12173378e+00
-9.83195528e-02 -6.14670098e-01 1.47517443e-01 1.55205393e+00
2.30004042e-01 4.26343501e-01 3.35882664e-01 4.53475058e-01
2.19464764e-01 -7.02122748e-02 8.76486421e-01 -1.36820972e-01
-1.00003362e-01 2.49846131e-01 -1.16009200e+00 5.22738993e-01
1.13962233e+00 -1.73736468e-01 -5.60083508e-01 -1.63952529e-01
-3.93509209e-01 -1.54664665e-02 2.93825507e-01 1.98490918e-01
5.98600566e-01 -9.34587657e-01 -9.05866921e-01 1.74705610e-01
3.24212581e-01 2.76452541e-01 3.48197937e-01 1.24697530e+00
-1.24506831e+00 7.43309319e-01 -1.38404354e-01 -1.09448075e+00
-1.41752660e+00 5.89644432e-01 5.40354729e-01 -5.96611202e-01
-6.80299044e-01 6.99793696e-01 7.25807369e-01 -3.52029115e-01
4.52056900e-02 -5.15043795e-01 -1.99318886e-01 1.33824712e-02
3.53178710e-01 5.67429304e-01 4.90085259e-02 -4.74986702e-01
-7.99452603e-01 7.15519607e-01 -1.73283201e-02 4.70759302e-01
1.24545658e+00 3.56064141e-01 -1.09348625e-01 1.31770670e-01
1.54493487e+00 -1.69664919e-01 -6.30957007e-01 2.12039143e-01
-5.08834302e-01 -3.11393123e-02 -4.22837108e-01 -9.66904998e-01
-8.84801090e-01 9.21613634e-01 1.29549909e+00 1.25484571e-01
8.75429332e-01 3.22353840e-01 8.99438739e-01 1.76338419e-01
-1.35880873e-01 -5.47536790e-01 8.65354836e-02 4.41973597e-01
5.82934618e-01 -1.17133403e+00 -1.48470625e-01 -2.44094282e-02
-7.90252686e-01 8.44211340e-01 4.61308330e-01 -4.46307063e-02
6.87808096e-01 1.15377143e-01 2.83225775e-01 -5.84065199e-01
-4.67286199e-01 3.28862250e-01 9.56231803e-02 8.23185384e-01
4.66299057e-01 5.85110664e-01 -3.38074088e-01 7.95467556e-01
1.77784413e-01 -1.65734868e-02 3.04370016e-01 9.63011861e-01
-5.19157350e-01 -2.60284096e-01 -5.82809269e-01 1.25857282e+00
-6.96989119e-01 1.30369931e-01 -2.23516300e-01 1.19430661e+00
3.35754991e-01 8.23756576e-01 2.62665927e-01 -8.70397836e-02
1.56714767e-01 1.57359004e-01 4.82328907e-02 -5.40597081e-01
-4.45904702e-01 6.30743429e-02 7.70256147e-02 9.16891694e-02
-3.13816488e-01 -7.41854668e-01 -1.63840723e+00 -2.62202360e-02
-1.64430991e-01 -7.87276626e-02 3.79384905e-01 7.25331545e-01
6.83345497e-02 7.28921056e-01 4.28439885e-01 -5.83513156e-02
-4.31380361e-01 -8.69959414e-01 -4.96367097e-01 1.28688931e-01
5.89493513e-01 -6.20963037e-01 -5.31036198e-01 3.20279971e-02] | [15.503130912780762, -1.7930952310562134] |
9080c0f3-fdcf-44a7-8e7b-3282cd2301f6 | cdnet-contrastive-disentangled-network-for | 2206.08524 | null | https://arxiv.org/abs/2206.08524v1 | https://arxiv.org/pdf/2206.08524v1.pdf | CDNet: Contrastive Disentangled Network for Fine-Grained Image Categorization of Ocular B-Scan Ultrasound | Precise and rapid categorization of images in the B-scan ultrasound modality is vital for diagnosing ocular diseases. Nevertheless, distinguishing various diseases in ultrasound still challenges experienced ophthalmologists. Thus a novel contrastive disentangled network (CDNet) is developed in this work, aiming to tackle the fine-grained image categorization (FGIC) challenges of ocular abnormalities in ultrasound images, including intraocular tumor (IOT), retinal detachment (RD), posterior scleral staphyloma (PSS), and vitreous hemorrhage (VH). Three essential components of CDNet are the weakly-supervised lesion localization module (WSLL), contrastive multi-zoom (CMZ) strategy, and hyperspherical contrastive disentangled loss (HCD-Loss), respectively. These components facilitate feature disentanglement for fine-grained recognition in both the input and output aspects. The proposed CDNet is validated on our ZJU Ocular Ultrasound Dataset (ZJUOUSD), consisting of 5213 samples. Furthermore, the generalization ability of CDNet is validated on two public and widely-used chest X-ray FGIC benchmarks. Quantitative and qualitative results demonstrate the efficacy of our proposed CDNet, which achieves state-of-the-art performance in the FGIC task. Code is available at: https://github.com/ZeroOneGame/CDNet-for-OUS-FGIC . | ['Yaqi Wang', 'Qun Jin', 'Juan Ye', 'Ruiquan Ge', 'Gangyong Jia', 'Yijie Wang', 'Yunxiang Li', 'Ruilong Dan'] | 2022-06-17 | null | null | null | null | ['image-categorization'] | ['computer-vision'] | [ 1.14661641e-01 -2.04299510e-01 -1.36050209e-01 -1.22339413e-01
-8.12333584e-01 -3.38108987e-01 4.42956924e-01 -1.99137270e-01
2.05009580e-02 4.50512916e-01 1.30514011e-01 -2.21264064e-01
-7.11643279e-01 -2.99813658e-01 -1.60145551e-01 -1.06382656e+00
9.06903967e-02 2.81766474e-01 -4.98420447e-02 2.66119808e-01
7.84789398e-02 8.01349998e-01 -1.73659778e+00 3.84958237e-01
1.10282373e+00 1.32226121e+00 8.24913965e-04 8.77077281e-01
3.18151772e-01 8.89152408e-01 -1.65539548e-01 -3.47877264e-01
2.39023983e-01 -2.65668541e-01 -7.69494474e-01 1.97279230e-01
7.41965890e-01 -2.69949257e-01 -3.39106917e-01 9.97533917e-01
9.62586522e-01 -1.91121042e-01 1.01511037e+00 -1.03148305e+00
-3.40450436e-01 -2.53023542e-02 -8.07286203e-01 5.64144433e-01
-1.47587359e-01 4.49787885e-01 8.00221860e-01 -1.02867270e+00
3.73324603e-01 9.39595520e-01 3.69159490e-01 4.70245779e-01
-9.20799553e-01 -9.19143081e-01 -3.47753376e-01 2.54068553e-01
-1.31209016e+00 -3.46976817e-01 2.72092253e-01 -9.33874249e-01
4.33521509e-01 6.77661300e-01 6.81164026e-01 8.24129105e-01
4.17862952e-01 5.87516308e-01 1.55329645e+00 -1.97915688e-01
-2.39935815e-01 -6.46937191e-02 3.18807781e-01 1.13829339e+00
4.42076772e-01 2.58246928e-01 -4.09922272e-01 -2.28078857e-01
7.77942717e-01 -2.21627373e-02 -6.34548306e-01 -3.85320902e-01
-1.07160127e+00 5.59568942e-01 3.71136695e-01 -2.17024572e-02
-3.00870955e-01 -1.77043587e-01 3.12607974e-01 1.51793554e-01
5.66302478e-01 4.92107868e-01 -1.33287936e-01 -6.19806815e-03
-5.10281146e-01 -1.80484131e-01 3.71336877e-01 4.96737599e-01
2.79693007e-01 -2.95756161e-01 -4.34629261e-01 8.78609300e-01
2.85516322e-01 4.70314056e-01 5.10047436e-01 -5.32047153e-01
1.57688394e-01 6.67002916e-01 -2.52670079e-01 -8.11981916e-01
-5.55814922e-01 -8.99820268e-01 -1.26724768e+00 2.43877932e-01
2.65489966e-01 -1.45622358e-01 -1.06365895e+00 1.16101432e+00
4.80204821e-01 5.18212855e-01 -2.11748749e-01 1.24328661e+00
1.45230210e+00 1.48286503e-02 -9.56253707e-02 -9.28459838e-02
1.58036482e+00 -7.77706742e-01 -3.24912280e-01 1.61608845e-01
4.07131940e-01 -6.69725418e-01 1.04724610e+00 5.54083526e-01
-9.29719627e-01 -3.81223977e-01 -7.92416453e-01 -9.03553814e-02
1.06475353e-01 5.97038150e-01 5.51062584e-01 4.73207414e-01
-7.76498675e-01 2.61341244e-01 -8.11512887e-01 -1.61996976e-01
8.07127655e-01 5.13290226e-01 -5.41205704e-01 -3.14514518e-01
-8.13753009e-01 6.05762661e-01 -2.51290083e-01 2.71898448e-01
-6.53081357e-01 -1.05541158e+00 -5.80116391e-01 -5.45539260e-02
2.63737261e-01 -1.13827443e+00 7.98153102e-01 -4.00191009e-01
-1.20885468e+00 1.15800965e+00 5.93256466e-02 -1.95677385e-01
4.82540548e-01 -5.59418043e-03 -4.39320803e-01 5.95277369e-01
-7.57013261e-02 3.38567823e-01 1.12764049e+00 -1.09293354e+00
-6.09377742e-01 -5.86525261e-01 -1.23618387e-01 1.55892283e-01
-1.34095564e-01 5.39075658e-02 -3.72453958e-01 -6.80605233e-01
8.41717348e-02 -1.02401364e+00 2.67790675e-01 2.54803628e-01
-7.37452626e-01 -3.32707435e-01 4.79003757e-01 -6.52906120e-01
1.08320475e+00 -2.37937403e+00 2.46614784e-01 6.70324191e-02
1.14522576e+00 4.75047469e-01 -1.91134140e-01 -3.48488279e-02
-3.43724221e-01 1.65491834e-01 2.27460966e-01 -4.47342396e-01
-2.45669857e-01 -2.73365945e-01 8.54977742e-02 8.55584264e-01
4.12756532e-01 9.01440322e-01 -6.37206376e-01 -5.29439926e-01
2.81248778e-01 3.78561050e-01 -3.50078702e-01 1.70953169e-01
1.46931902e-01 7.46083677e-01 -3.21448058e-01 9.04125035e-01
6.76859796e-01 -8.61273468e-01 -1.35265023e-01 -7.52862394e-01
-3.86845693e-02 -1.71693534e-01 -8.05474997e-01 1.30468309e+00
-4.13128555e-01 6.72470570e-01 -5.22344783e-02 -4.66418594e-01
4.96862322e-01 1.76682323e-01 4.58533883e-01 -6.21189475e-01
4.57508415e-01 2.14317560e-01 3.77456874e-01 -8.53384614e-01
-1.95029497e-01 -9.53627452e-02 4.32195634e-01 2.13982075e-01
9.95139778e-02 7.14383125e-02 6.64010718e-02 3.27741802e-02
9.63730216e-01 -3.47634017e-01 4.29347187e-01 -2.14653715e-01
6.09223664e-01 -2.77505428e-01 2.87015736e-01 6.12202168e-01
-3.29749554e-01 9.07702744e-01 8.40765953e-01 -3.04017574e-01
-4.80109036e-01 -9.16426480e-01 -5.64134181e-01 4.74726826e-01
2.85980105e-01 -1.04289554e-01 -4.15590078e-01 -7.65149713e-01
1.81056485e-01 6.36945814e-02 -7.96267331e-01 -3.31125073e-02
-1.60776049e-01 -7.65110612e-01 4.72844452e-01 1.44998923e-01
4.67353910e-01 -5.05413830e-01 -3.41355294e-01 -4.81081337e-01
-1.27266064e-01 -1.03218937e+00 -4.48716819e-01 -4.32287395e-01
-5.43837547e-01 -1.68100023e+00 -8.14100087e-01 -7.10550725e-01
7.23928928e-01 3.97527665e-01 9.44993496e-01 8.22412595e-02
-9.78826344e-01 4.38003600e-01 -2.08378598e-01 -4.36643124e-01
-2.06560537e-01 -2.08588690e-01 -2.23797862e-03 4.63900775e-01
2.22913340e-01 -2.78241605e-01 -1.23659050e+00 5.23422301e-01
-6.08956754e-01 2.20422074e-01 9.50636566e-01 1.07938719e+00
7.34183788e-01 -1.48240209e-01 2.80031532e-01 -8.93171132e-01
5.99586964e-01 -3.82257551e-01 -5.48603594e-01 3.21898282e-01
-6.11724615e-01 -5.08221507e-01 1.48863479e-01 -2.96070278e-01
-7.46152580e-01 -3.82175475e-01 6.66731223e-02 -8.76016557e-01
-2.74727732e-01 4.46136266e-01 3.18732113e-01 -5.99993467e-01
7.32990623e-01 1.72722161e-01 4.42380428e-01 -3.63262802e-01
6.33319616e-02 1.08594215e+00 3.63421708e-01 -1.73655674e-01
5.08502722e-01 5.00977695e-01 3.60255033e-01 -9.15917397e-01
-9.74887311e-01 -7.25992084e-01 -2.54028678e-01 -2.77880698e-01
7.19681382e-01 -1.00611460e+00 -1.01488686e+00 7.09897518e-01
-6.84710145e-01 9.05796662e-02 -3.07607651e-01 7.68825293e-01
-2.87756801e-01 3.79856110e-01 -5.64806342e-01 -4.07818079e-01
-5.24780571e-01 -1.43748212e+00 1.02075231e+00 3.70466799e-01
9.34876576e-02 -6.70853376e-01 -1.05011664e-01 8.68377805e-01
2.91411668e-01 3.41607869e-01 1.27583015e+00 -5.14981687e-01
-6.87220395e-01 -1.19304165e-01 -6.89305723e-01 7.78509259e-01
1.90334141e-01 1.35875791e-01 -9.33467507e-01 -4.23219949e-01
-8.76236707e-02 -5.42580545e-01 9.75648463e-01 6.45649791e-01
1.44950902e+00 -1.53334647e-01 -8.19362476e-02 1.00484693e+00
1.19915521e+00 -1.16359726e-01 5.13604343e-01 -4.76546809e-02
6.65711641e-01 5.31326115e-01 6.09284520e-01 5.02900362e-01
3.21501702e-01 4.21068966e-01 5.65340161e-01 -4.59155351e-01
-7.55615771e-01 3.20402622e-01 -4.01891023e-01 8.99252534e-01
-3.34388465e-01 -3.44498307e-01 -1.05276895e+00 4.19552147e-01
-1.40489280e+00 -4.28350627e-01 -1.93528384e-01 2.02899861e+00
6.59690082e-01 -3.19430202e-01 -1.44729540e-01 -1.61539897e-01
5.60607076e-01 -5.95963784e-02 -7.62756646e-01 1.74948722e-01
-7.17660487e-02 3.37483853e-01 2.23858207e-01 1.44459933e-01
-1.34430039e+00 3.61194432e-01 4.38041925e+00 1.07153642e+00
-1.43731344e+00 2.73716360e-01 7.17177629e-01 -4.02477771e-01
3.00386436e-02 -5.11669815e-01 -5.72754979e-01 5.66228032e-01
1.77334964e-01 1.33207813e-01 2.27705792e-01 2.25820050e-01
7.42969215e-02 -4.28471789e-02 -8.51160884e-01 1.39907777e+00
1.68795541e-01 -1.53081691e+00 9.01842564e-02 -1.69026689e-03
5.07233083e-01 2.89286256e-01 6.60971224e-01 -9.60314423e-02
-3.91351849e-01 -1.06897664e+00 -1.20370403e-01 8.07720423e-01
1.66005433e+00 -4.32752877e-01 9.11695540e-01 -7.20794722e-02
-7.06558406e-01 -8.38336647e-02 1.96863953e-02 4.56166059e-01
-4.32865649e-01 5.93706071e-01 -8.71507943e-01 5.54500341e-01
7.52179682e-01 8.92993689e-01 -6.74680471e-01 1.44757986e+00
-6.87283278e-03 6.87950373e-01 2.00344428e-01 1.04598098e-01
8.73749927e-02 -2.39150733e-01 9.47147727e-01 8.82277727e-01
1.15424633e-01 4.67321903e-01 -3.42034429e-01 5.46516895e-01
-2.20633242e-02 2.48944804e-01 -1.26035243e-01 -6.37540966e-02
1.76177099e-01 1.45812666e+00 -4.61473107e-01 6.05742540e-03
-2.92944252e-01 4.87162679e-01 -1.01260200e-01 4.35640752e-01
-5.52008867e-01 -3.92809242e-01 8.59838068e-01 1.98932514e-01
4.15614732e-02 3.11982512e-01 -1.27452299e-01 -1.28776014e+00
-8.27134848e-02 -1.06849360e+00 4.39688295e-01 -8.19940627e-01
-1.54987907e+00 7.14932084e-01 -1.99932426e-01 -1.72176290e+00
1.67079657e-01 -6.26251101e-01 -2.13947445e-01 9.99767780e-01
-1.82924676e+00 -1.30834687e+00 -8.97702754e-01 5.05961657e-01
2.82766193e-01 -5.46506345e-01 6.71109855e-01 4.77347761e-01
-7.76848555e-01 6.98503971e-01 1.14501074e-01 3.10488883e-02
7.48350561e-01 -1.14811158e+00 -1.30675569e-01 5.93101203e-01
-2.71485776e-01 5.73785245e-01 3.04818332e-01 -3.80095899e-01
-1.18677926e+00 -1.22588897e+00 3.17019016e-01 -1.90591708e-01
7.27178812e-01 -8.61103088e-02 -7.28439689e-01 1.75019637e-01
-5.96301332e-02 6.10162318e-01 9.93142247e-01 -1.94380194e-01
-3.90089840e-01 -1.16738670e-01 -1.14144480e+00 4.83561724e-01
8.56023550e-01 -5.02375424e-01 -1.02490678e-01 8.57567310e-01
4.86810029e-01 -8.61203313e-01 -1.18493629e+00 8.65939081e-01
7.89136410e-01 -1.06869841e+00 9.88563061e-01 -9.02835667e-01
5.75396478e-01 -1.71412364e-01 9.99145657e-02 -1.10972309e+00
-4.24782783e-01 -5.96132517e-01 -2.78564364e-01 6.55047476e-01
9.05941799e-02 -1.14796627e+00 6.21668100e-01 7.78773800e-02
-2.11852700e-01 -1.47549093e+00 -9.42690730e-01 -3.84133995e-01
-1.96788281e-01 -1.60848618e-01 3.88065010e-01 7.60632515e-01
-4.70366687e-01 2.20415473e-01 -1.37804702e-01 4.96982366e-01
7.58594334e-01 3.46889228e-01 4.10451114e-01 -1.38355827e+00
-4.18492228e-01 -5.37237048e-01 -8.43476772e-01 -6.51721597e-01
-1.63926929e-01 -9.39442515e-01 -5.05202472e-01 -1.45314729e+00
4.58347499e-01 -5.88421404e-01 -4.12601918e-01 3.01066548e-01
-1.93836436e-01 5.63623369e-01 1.30073518e-01 4.96913373e-01
-4.67603832e-01 3.93123567e-01 1.77095604e+00 -2.90093958e-01
5.56551479e-02 4.17231023e-01 -7.66588748e-01 7.03162551e-01
7.01515317e-01 -3.23059767e-01 -3.62445563e-01 -1.68819577e-01
1.97815169e-02 2.33159989e-01 7.93135345e-01 -7.66554117e-01
1.46926492e-01 1.51620448e-01 1.86977208e-01 -5.33909380e-01
3.70659858e-01 -3.73604059e-01 -9.73279327e-02 5.35104573e-01
-1.50836900e-01 -4.57030505e-01 1.04479969e-01 5.14155746e-01
-4.13110673e-01 8.98192748e-02 9.71472323e-01 1.60351589e-01
-4.40697342e-01 5.54250240e-01 -5.35002351e-02 2.21057281e-01
1.05969775e+00 -9.73739475e-02 -9.89844143e-01 7.28634745e-02
-8.94300401e-01 2.43389323e-01 4.48580310e-02 2.51209527e-01
9.66688514e-01 -6.77894235e-01 -1.01576471e+00 5.24714351e-01
5.65414906e-01 1.23092361e-01 8.18273902e-01 1.68081141e+00
-7.73023665e-01 4.85739052e-01 -4.06888872e-03 -8.23650718e-01
-1.73744869e+00 -1.30095899e-01 7.34996319e-01 -1.96506921e-02
-5.71210206e-01 1.10450709e+00 5.14211535e-01 -2.07143784e-01
2.84071624e-01 -5.15704751e-01 -4.37646300e-01 1.67667001e-01
3.39213312e-01 5.60716808e-01 3.65607232e-01 -4.95073438e-01
-2.79616266e-01 9.13637817e-01 -3.98688614e-01 6.11111760e-01
1.09943759e+00 1.59085393e-02 -4.26018864e-01 6.06948771e-02
1.31224167e+00 -4.17264067e-02 -8.69605601e-01 -4.59472984e-01
-7.48368979e-01 -6.05887830e-01 3.61143738e-01 -1.13292587e+00
-1.06476176e+00 9.73256707e-01 1.20275438e+00 -2.67169587e-02
1.36075759e+00 1.36105567e-01 4.89451438e-01 2.89431028e-03
1.13586903e-01 -1.51922405e-01 2.65113443e-01 -5.18207322e-04
1.11705005e+00 -1.30076849e+00 1.63669676e-01 -8.31690907e-01
-7.52939343e-01 9.34608817e-01 5.31800926e-01 1.66644350e-01
8.04395914e-01 7.06549734e-03 1.29399508e-01 -5.61494350e-01
-6.08149529e-01 -3.61915141e-01 9.63233888e-01 3.50548565e-01
9.83191431e-02 2.15225011e-01 -2.43384227e-01 5.47447801e-01
1.76683843e-01 4.00753170e-02 4.52257544e-01 4.16588008e-01
-1.01386674e-03 -4.49243218e-01 6.42115474e-02 1.21899343e+00
-5.42047679e-01 -2.19745979e-01 -1.91132650e-01 8.14664602e-01
3.16259503e-01 6.85381413e-01 9.34742913e-02 -5.35795748e-01
2.42229551e-01 -4.37940121e-01 4.33442771e-01 -7.20975876e-01
-4.79258507e-01 1.77123815e-01 -7.11144805e-02 -6.81402206e-01
-3.84978861e-01 -5.85682392e-01 -7.19023407e-01 1.88122377e-01
-5.75396895e-01 -1.50227115e-01 4.99258101e-01 6.30420625e-01
6.58528268e-01 6.48195148e-01 8.17654252e-01 -3.61691356e-01
-5.80289185e-01 -9.03908789e-01 -8.17762911e-01 9.15238634e-02
8.25830996e-01 -8.37322831e-01 -5.82921565e-01 -3.27040195e-01] | [15.762152671813965, -3.926283359527588] |
0ec2d3a2-9007-4adb-8ef3-c9dcda8176f1 | aggression-identification-in-english-hindi | null | null | https://aclanthology.org/2020.trac-1.12 | https://aclanthology.org/2020.trac-1.12.pdf | Aggression Identification in English, Hindi and Bangla Text using BERT, RoBERTa and SVM | This paper presents the results of the classifiers we developed for the shared tasks in aggression identification and misogynistic aggression identification. These two shared tasks were held as part of the second workshop on Trolling, Aggression and Cyberbullying (TRAC). Both the subtasks were held for English, Hindi and Bangla language. In our study, we used English BERT (En-BERT), RoBERTa, DistilRoBERTa, and SVM based classifiers for English language. For Hindi and Bangla language, multilingual BERT (M-BERT), XLM-RoBERTa and SVM classifiers were used. Our best performing models are EN-BERT for English Subtask A (Weighted F1 score of 0.73, Rank 5/16), SVM for English Subtask B (Weighted F1 score of 0.87, Rank 2/15), SVM for Hindi Subtask A (Weighted F1 score of 0.79, Rank 2/10), XLMRoBERTa for Hindi Subtask B (Weighted F1 score of 0.87, Rank 2/10), SVM for Bangla Subtask A (Weighted F1 score of 0.81, Rank 2/10), and SVM for Bangla Subtask B (Weighted F1 score of 0.93, Rank 4/8). It is seen that the superior performance of the SVM classifier was achieved mainly because of its better prediction of the majority class. BERT based classifiers were found to predict the minority classes better. | ['Kuntal Dey', 'Kaushik Das', 'Ferdous Barbhuiya', 'Arup Baruah'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['misogynistic-aggression-identification', 'aggression-identification'] | ['natural-language-processing', 'natural-language-processing'] | [-7.99898982e-01 -1.64653897e-01 7.58360252e-02 -2.82179087e-01
-6.10604107e-01 -2.99621671e-01 6.19480848e-01 2.66568244e-01
-9.67619956e-01 1.04395020e+00 1.97510064e-01 -3.50632340e-01
-5.95194876e-01 -5.00404775e-01 5.34271263e-02 -5.66015244e-01
-4.89902981e-02 6.55510604e-01 5.14230371e-01 -9.45677161e-01
6.12304091e-01 6.48257792e-01 -1.15094733e+00 5.62209964e-01
4.37592149e-01 4.60645437e-01 -2.90884286e-01 1.19817579e+00
1.63791731e-01 1.11755455e+00 -6.57901466e-01 -1.00629520e+00
-7.67225474e-02 -2.71775901e-01 -1.47045934e+00 -7.98456609e-01
4.46798950e-01 -6.24209702e-01 -5.33215165e-01 7.48502254e-01
4.83129442e-01 3.42492968e-01 1.01712596e+00 -1.43946469e+00
7.51924142e-02 4.86137390e-01 -5.66818714e-01 7.03419983e-01
6.04554832e-01 8.29840973e-02 1.02212417e+00 -5.37034988e-01
5.39090455e-01 1.24320352e+00 8.62989783e-01 7.93680847e-01
-8.74295771e-01 -1.00479758e+00 -3.95348012e-01 4.56139207e-01
-1.26340818e+00 -2.59301484e-01 5.98396182e-01 -7.06343532e-01
1.20088828e+00 4.91827160e-01 6.68098569e-01 1.43317032e+00
1.48688242e-01 8.01097393e-01 1.49140131e+00 -1.27463236e-01
1.56461203e-03 5.25589943e-01 8.53884757e-01 4.55923021e-01
-6.46386743e-02 3.47633362e-01 -4.84938264e-01 -4.19336647e-01
4.74301845e-01 -7.36357629e-01 5.61929822e-01 5.24690211e-01
-1.78661272e-01 1.27902567e+00 -8.86304229e-02 5.69753289e-01
-2.05612089e-02 -2.38017246e-01 1.00398850e+00 6.37777328e-01
6.65781274e-02 3.35934609e-01 -3.26096892e-01 -9.35493052e-01
-7.50632107e-01 7.74533868e-01 6.27411842e-01 5.00333250e-01
6.39558807e-02 1.99847430e-01 1.23667335e-02 1.66166329e+00
2.31275439e-01 2.89697438e-01 6.10661387e-01 -6.17444575e-01
5.14137268e-01 2.39038080e-01 -2.03692243e-01 -8.42477679e-01
-9.11796570e-01 1.16304107e-01 -2.42827922e-01 4.61437613e-01
6.98761165e-01 -1.41054451e-01 -4.54275459e-01 1.47637916e+00
2.93931663e-01 -4.29173976e-01 1.35998994e-01 6.87182188e-01
1.22774994e+00 3.92200679e-01 3.69617254e-01 3.85984145e-02
9.93156254e-01 -1.03388453e+00 -2.12036505e-01 1.45387232e-01
1.00720918e+00 -7.24483967e-01 7.40231633e-01 7.18299389e-01
-1.06199098e+00 -4.31525797e-01 -7.27611423e-01 2.28783578e-01
-4.29256231e-01 -2.55670786e-01 3.83686662e-01 1.23345971e+00
-7.19744146e-01 3.24398369e-01 -4.48782027e-01 -7.40557969e-01
1.69316217e-01 6.15894794e-01 -9.32520568e-01 3.59937280e-01
-1.09775841e+00 1.32720387e+00 2.53240883e-01 -4.43789601e-01
-7.65295744e-01 -3.71297210e-01 -7.00039208e-01 -3.99069011e-01
-7.95075968e-02 2.95254737e-01 9.99084473e-01 -5.12971699e-01
-1.38324833e+00 1.26924121e+00 3.52289826e-01 -1.66854039e-01
7.27265954e-01 -6.31488115e-02 -6.27024531e-01 4.63083684e-02
2.59756267e-01 7.70987943e-02 3.34397733e-01 -9.31869626e-01
-9.62773561e-01 -7.09217191e-01 3.71197075e-01 3.92652489e-02
-1.44628078e-01 1.00662673e+00 6.19931757e-01 -4.66878742e-01
-4.20324862e-01 -6.45818174e-01 1.67942494e-01 -7.13300228e-01
-3.96211058e-01 -4.83996958e-01 7.86491096e-01 -1.16659677e+00
1.57265377e+00 -2.07113981e+00 9.74032283e-02 3.03700775e-01
3.06164950e-01 7.61872530e-01 2.25865245e-01 8.56040180e-01
-3.90333772e-01 2.81300638e-02 5.27373999e-02 -6.34600446e-02
-6.70026317e-02 4.06135201e-01 2.33814925e-01 5.09618700e-01
-2.69208819e-01 3.67403030e-01 -6.85650647e-01 -5.09929657e-01
3.09117943e-01 -9.94375646e-02 -5.72591841e-01 1.31918654e-01
9.46109653e-01 3.45403761e-01 -2.40201578e-01 3.81485224e-01
4.92989749e-01 9.57436800e-01 -4.65164095e-01 3.90647858e-01
-2.86252260e-01 1.55162856e-01 -1.20634818e+00 7.54601598e-01
-6.17349207e-01 7.34334946e-01 4.04414415e-01 -1.09029746e+00
1.18389702e+00 3.52095842e-01 3.74532968e-01 -5.40970862e-01
4.63809758e-01 5.33271611e-01 5.71445942e-01 -8.29113781e-01
4.04474884e-01 -5.26640475e-01 -2.22810671e-01 2.72383034e-01
1.10230908e-01 -2.40198821e-01 3.29834431e-01 2.51107812e-01
1.06682646e+00 -3.79333973e-01 3.27246100e-01 -2.04148665e-01
1.04117465e+00 -9.47564989e-02 1.24929748e-01 7.48679459e-01
-7.96505094e-01 4.42647457e-01 7.98051894e-01 -4.93800431e-01
-9.41657066e-01 -8.86941373e-01 -4.39690530e-01 1.31542456e+00
-2.93821514e-01 -4.99030471e-01 -7.66624331e-01 -1.01278114e+00
-1.03578400e-02 1.12037706e+00 -5.40812850e-01 -4.27621543e-01
-7.34904110e-01 -4.97671455e-01 1.26835740e+00 5.52405119e-01
3.99212778e-01 -1.00095069e+00 -4.37723219e-01 1.01919942e-01
1.08930819e-01 -9.18683946e-01 -9.63788554e-02 8.55930522e-02
-3.56593847e-01 -9.50650334e-01 -4.55262125e-01 -4.68180835e-01
-8.32161754e-02 -2.76530594e-01 3.65971088e-01 3.49458128e-01
-4.61008459e-01 1.43287197e-01 -7.13679314e-01 -2.56044984e-01
-5.34452796e-01 -1.01019226e-01 3.41153473e-01 -2.46995196e-01
4.80810612e-01 -3.15262139e-01 1.13116011e-01 3.70220512e-01
-2.66378343e-01 -1.60746098e-01 -9.51602608e-02 1.21962595e+00
-5.62604368e-01 -1.71736225e-01 5.20590663e-01 -7.40654886e-01
8.53274047e-01 -5.82470894e-01 -7.21843168e-02 -3.56938154e-01
-1.86531544e-01 -7.63162494e-01 6.26526177e-01 -5.17081738e-01
-6.94071114e-01 -5.52211642e-01 -1.03109550e+00 2.92119700e-02
-5.53962767e-01 2.81808555e-01 2.53429174e-01 -1.24893300e-01
8.65284741e-01 -8.94563459e-03 -7.69233629e-02 -6.78037584e-01
-3.75814646e-01 1.40218437e+00 2.34014899e-01 -6.86146677e-01
6.48475528e-01 -1.86734885e-01 5.23146838e-02 -1.44188178e+00
-4.46621418e-01 -6.25583768e-01 -8.90678525e-01 -5.70170403e-01
1.04941046e+00 -4.58098471e-01 -8.27061296e-01 8.96145046e-01
-6.96714044e-01 -3.12125772e-01 3.29697132e-02 7.93257296e-01
-5.35675824e-01 1.76045626e-01 -6.29456818e-01 -1.29289305e+00
-2.43074864e-01 -1.30736673e+00 6.28269613e-01 3.03338282e-02
-1.02195573e+00 -1.02560329e+00 2.99061835e-01 1.07349718e+00
1.62692145e-01 9.75693986e-02 1.18786979e+00 -1.57341695e+00
1.10417330e+00 -4.06692266e-01 -2.61763930e-01 6.12317383e-01
-7.63022155e-02 2.10271195e-01 -7.26552248e-01 -1.77834481e-01
-3.26503694e-01 -5.67917049e-01 3.10012996e-01 -2.15006992e-01
6.03090286e-01 -1.89972430e-01 1.27108499e-01 2.15036124e-01
6.10406458e-01 7.51591742e-01 4.27082896e-01 5.60876191e-01
6.07847631e-01 8.77593875e-01 6.01071000e-01 4.42707121e-01
4.22241539e-01 1.27171052e+00 3.04634608e-02 3.23404551e-01
-1.50198683e-01 -8.54670182e-02 4.51649189e-01 5.27713478e-01
-4.66234535e-01 4.82957989e-01 -1.36307704e+00 6.52004898e-01
-1.57213962e+00 -1.15571570e+00 -9.90628362e-01 2.02172780e+00
5.64801633e-01 1.98527366e-01 1.39186954e+00 7.56177604e-01
5.22915661e-01 -1.86584920e-01 1.10486373e-01 -1.75707424e+00
2.78389513e-01 4.86678392e-01 6.65856674e-02 8.89481008e-01
-9.75056767e-01 1.24610746e+00 5.64067411e+00 1.25725937e+00
-9.55297351e-01 2.74158388e-01 2.45532945e-01 -3.64145607e-01
4.65497226e-01 -2.00871050e-01 -8.45675528e-01 8.12130451e-01
1.25537634e+00 -1.99818552e-01 2.80049801e-01 7.84379303e-01
-4.58444282e-02 -3.98558021e-01 -7.56291807e-01 1.03151953e+00
7.99186702e-04 -5.66605747e-01 -2.59332865e-01 2.01289296e-01
3.09329093e-01 -2.27672577e-01 -1.15540899e-01 7.45964766e-01
4.47231919e-01 -1.35884261e+00 7.74421871e-01 1.25755593e-02
3.85111034e-01 -1.02997077e+00 1.20965326e+00 5.96200824e-01
-6.99142158e-01 -4.27090824e-01 -2.67568439e-01 -4.79697943e-01
-8.18339661e-02 -2.65681267e-01 -5.86704731e-01 5.20849288e-01
1.05192757e+00 2.83606470e-01 -3.09508771e-01 9.65196073e-01
-2.86423624e-01 8.93582642e-01 -3.55429322e-01 -5.94462276e-01
6.55385256e-01 -1.32778004e-01 9.31765914e-01 1.50642633e+00
-2.61908859e-01 2.27533266e-01 -3.52584161e-02 -4.87863757e-02
8.39537919e-01 6.72970593e-01 -3.89833301e-01 2.28205115e-01
3.04834787e-02 1.07840168e+00 -1.32027015e-01 5.42179076e-03
-2.55309880e-01 6.31438315e-01 4.76095408e-01 -1.96288645e-01
-9.82604027e-01 -6.75630271e-01 7.40997910e-01 2.87255108e-01
-2.33626708e-01 4.87998948e-02 -5.71258724e-01 -6.08260989e-01
-4.27245498e-01 -1.00400901e+00 8.40612531e-01 -4.78524476e-01
-1.32939041e+00 5.88113606e-01 5.89891553e-01 -8.08894455e-01
-1.92283854e-01 -9.62670565e-01 -8.42465699e-01 9.81089056e-01
-1.95965990e-01 -1.28682280e+00 1.15414843e-01 7.36509740e-01
6.33794069e-01 -9.14342999e-01 5.71290553e-01 4.38941866e-01
-8.53398919e-01 9.88734245e-01 1.98030043e-02 3.51976722e-01
4.69789296e-01 -1.04870415e+00 -3.14900458e-01 7.07196072e-02
-1.10684089e-01 5.39483488e-01 7.02299654e-01 -5.17659187e-01
-5.51102459e-01 -4.46413636e-01 1.16951537e+00 -5.92534423e-01
1.13373137e+00 -1.87228441e-01 -6.59518898e-01 4.05788660e-01
-1.94352373e-01 -7.15141594e-01 9.48523521e-01 4.45485234e-01
-4.49012756e-01 -2.31400393e-02 -1.83202970e+00 5.13432324e-01
8.20711195e-01 -4.86600548e-01 -7.76540101e-01 4.87119198e-01
-3.11719149e-01 -3.57688755e-01 -1.12587583e+00 9.02505999e-05
9.78152275e-01 -1.26469564e+00 6.63788915e-01 -1.27487266e+00
7.94256449e-01 4.91244435e-01 -2.51831114e-01 -1.03883886e+00
-3.38254690e-01 -1.50541037e-01 5.23341835e-01 1.00824308e+00
5.19179702e-01 -8.87384772e-01 6.40910804e-01 6.66260719e-01
-7.10230246e-02 -8.30586553e-01 -1.72992730e+00 -9.34200764e-01
9.29889023e-01 -5.76171339e-01 -1.08597474e-02 8.99219394e-01
5.83192527e-01 2.12688208e-01 -5.58588445e-01 -3.00399959e-01
1.45849898e-01 -6.73872948e-01 1.08297706e+00 -1.11512208e+00
-2.30399162e-01 -7.80331850e-01 -1.18648767e+00 -1.13044702e-01
1.25456572e-01 -8.45118523e-01 -7.65452862e-01 -1.22763228e+00
2.70658016e-01 -2.52379626e-01 4.79246974e-02 6.14017010e-01
8.78352299e-02 4.32245493e-01 2.60359108e-01 -1.82349160e-01
1.51795074e-01 1.68550938e-01 5.68068326e-01 3.67231667e-02
-2.87775528e-02 2.29164153e-01 -5.90183437e-01 9.15918767e-01
9.08857465e-01 -6.34100080e-01 -1.73497468e-01 1.06545866e-01
-2.80751795e-01 8.28267336e-02 3.13636154e-01 -9.95877564e-01
-8.46548826e-02 -5.27800024e-01 1.59244806e-01 -2.09400252e-01
8.29086959e-01 -4.19759125e-01 -5.05721569e-02 7.17134178e-01
-2.47344315e-01 -1.21474899e-01 2.24411473e-01 -4.49707329e-01
-1.91867679e-01 -9.13430214e-01 1.19521093e+00 9.85445827e-02
-3.54879588e-01 -1.12634152e-02 -1.03314602e+00 2.22735628e-01
1.53624570e+00 -7.40352154e-01 -2.61505634e-01 -6.03495657e-01
-1.12323666e+00 2.85811885e-03 -8.75326321e-02 4.54908967e-01
4.77455586e-01 -8.19379270e-01 -9.70227540e-01 2.60861009e-01
3.30032520e-02 -9.76764083e-01 3.08954775e-01 1.17992198e+00
-8.02797258e-01 2.97028065e-01 -8.03018272e-01 -5.12272865e-02
-2.13006711e+00 -2.00975705e-02 1.56618953e-01 -1.78279459e-01
-2.71923155e-01 1.18585265e+00 -1.81738004e-01 -6.70282483e-01
-1.10774554e-01 7.33009219e-01 -6.72303438e-01 1.37390688e-01
5.19574404e-01 1.23297977e+00 -2.28351839e-02 -1.77549565e+00
-5.48058093e-01 6.13831460e-01 -6.79479837e-01 -4.67062980e-01
1.44172716e+00 2.99806118e-01 -2.40681142e-01 4.45931017e-01
1.28254461e+00 3.31491798e-01 9.12974775e-02 4.77690071e-01
-8.28999188e-03 -7.02105105e-01 -1.77760541e-01 -9.76557493e-01
-6.54314935e-01 9.03437197e-01 2.76191175e-01 2.60845810e-01
6.37808084e-01 -8.97539333e-02 5.11637688e-01 2.12592874e-02
5.95661581e-01 -1.40744627e+00 -1.59736872e-01 1.02274942e+00
1.14447474e+00 -8.88300955e-01 -9.30068344e-02 -2.97573417e-01
-1.17697453e+00 1.29112887e+00 1.19255424e+00 -2.22173840e-01
6.07294619e-01 7.20577613e-02 1.33319229e-01 -1.16644688e-01
-3.92534107e-01 2.41788588e-02 1.02789514e-01 8.37110937e-01
4.29427028e-01 2.27538064e-01 -1.11560261e+00 1.10183907e+00
-1.07671881e+00 -6.86597764e-01 4.55575645e-01 6.64636075e-01
-3.29649180e-01 -1.11749291e+00 -4.20135260e-01 7.18818247e-01
-5.53223848e-01 3.37465644e-01 -8.99844646e-01 1.27130759e+00
1.35163352e-01 1.32019842e+00 -2.04137117e-01 -8.81815195e-01
4.98810709e-01 1.33102834e-01 3.55521947e-01 -4.09176320e-01
-1.41120696e+00 -3.72598559e-01 9.04716671e-01 -4.08824980e-01
4.67855006e-01 -9.72995937e-01 -1.09582698e+00 -7.35006213e-01
-2.52360523e-01 4.67507005e-01 6.18519425e-01 9.61524367e-01
-6.54966474e-01 -2.88820751e-02 6.19372725e-01 -2.40517005e-01
-6.00249588e-01 -1.23736751e+00 -1.03217638e+00 4.60502774e-01
-2.03506816e-02 -1.09407723e+00 -4.34754729e-01 -8.20071220e-01] | [8.80667781829834, 10.769123077392578] |
b9189938-2e9d-43a0-9817-5b15fcd2daea | poseur-direct-human-pose-regression-with | 2201.07412 | null | https://arxiv.org/abs/2201.07412v2 | https://arxiv.org/pdf/2201.07412v2.pdf | Poseur: Direct Human Pose Regression with Transformers | We propose a direct, regression-based approach to 2D human pose estimation from single images. We formulate the problem as a sequence prediction task, which we solve using a Transformer network. This network directly learns a regression mapping from images to the keypoint coordinates, without resorting to intermediate representations such as heatmaps. This approach avoids much of the complexity associated with heatmap-based approaches. To overcome the feature misalignment issues of previous regression-based methods, we propose an attention mechanism that adaptively attends to the features that are most relevant to the target keypoints, considerably improving the accuracy. Importantly, our framework is end-to-end differentiable, and naturally learns to exploit the dependencies between keypoints. Experiments on MS-COCO and MPII, two predominant pose-estimation datasets, demonstrate that our method significantly improves upon the state-of-the-art in regression-based pose estimation. More notably, ours is the first regression-based approach to perform favorably compared to the best heatmap-based pose estimation methods. | ['Anton Van Den Hengel', 'Zhibin Wang', 'Xinlong Wang', 'Zhi Tian', 'Chunhua Shen', 'Yongtao Ge', 'Weian Mao'] | 2022-01-19 | null | null | null | null | ['2d-human-pose-estimation'] | ['computer-vision'] | [ 3.72917540e-02 8.14889849e-04 -1.97461098e-01 -2.63767660e-01
-1.03231442e+00 -3.37110847e-01 5.43878078e-01 -1.08416229e-01
-5.77530026e-01 5.89967728e-01 3.22238535e-01 2.15028226e-01
4.42569777e-02 -2.84221679e-01 -9.82164979e-01 -3.26850146e-01
-2.62950957e-01 8.29471231e-01 1.16953170e-02 -3.65195602e-01
3.23287010e-01 3.94667566e-01 -1.36268222e+00 -1.49871945e-01
4.34774637e-01 1.05439436e+00 8.21901932e-02 8.14177573e-01
3.81932408e-01 6.97632968e-01 -4.63790745e-01 -2.61416435e-01
2.18909770e-01 -2.23231018e-01 -8.38457286e-01 -1.16593555e-01
9.58829284e-01 -4.41468239e-01 -6.62456214e-01 5.31347871e-01
5.43263137e-01 2.41847470e-01 6.87766254e-01 -1.32736337e+00
-4.61483359e-01 -5.49627580e-02 -8.40745747e-01 1.08010665e-01
6.59282446e-01 2.82385517e-02 1.02369368e+00 -1.11231852e+00
7.61478424e-01 1.38363743e+00 9.58578408e-01 4.86404628e-01
-1.40919471e+00 -4.73106861e-01 2.53707081e-01 2.43636295e-01
-1.45328295e+00 -3.58267188e-01 7.11944997e-01 -5.49503803e-01
9.25266087e-01 5.15227020e-02 9.11902308e-01 1.16783941e+00
1.69289544e-01 1.11121118e+00 8.76218259e-01 -3.58440369e-01
-1.32437199e-01 -3.06427389e-01 -3.15714657e-01 9.76116717e-01
-1.72580883e-01 1.35212213e-01 -9.17010188e-01 4.60938178e-02
1.08458376e+00 -1.20938718e-02 -4.03587580e-01 -1.00915956e+00
-1.47367823e+00 7.24706054e-01 7.63637066e-01 -1.38485938e-01
-3.55386674e-01 6.72570109e-01 2.22826988e-01 5.65158688e-02
5.31488776e-01 5.95691800e-01 -6.30298257e-01 -3.62724781e-01
-8.66853952e-01 6.59251153e-01 4.20764446e-01 9.12340939e-01
6.72088861e-01 -4.29718912e-01 -4.06468995e-02 5.95395565e-01
3.65986317e-01 2.65170455e-01 2.08461061e-01 -1.02141511e+00
7.36987591e-01 4.04698581e-01 2.39136755e-01 -1.21403611e+00
-5.93138635e-01 -3.53742540e-01 -3.70848119e-01 2.57331878e-01
4.95658875e-01 -5.50567126e-03 -9.22989011e-01 1.79167366e+00
3.53065431e-01 3.73584814e-02 -4.48371917e-01 1.05666375e+00
5.45400739e-01 4.21431184e-01 -3.10858320e-02 1.30622804e-01
1.13757241e+00 -1.26723278e+00 -3.13000947e-01 -4.34111744e-01
4.14265990e-01 -4.86751646e-01 1.11922419e+00 2.97223151e-01
-1.06477571e+00 -5.87340593e-01 -1.06039214e+00 -1.77872524e-01
-7.93625712e-02 2.41557404e-01 6.26783848e-01 2.73054808e-01
-1.08415806e+00 8.21357965e-01 -1.09035897e+00 -4.69879448e-01
3.43094856e-01 4.49816167e-01 -6.31788790e-01 2.04551443e-01
-8.74229670e-01 1.25108826e+00 2.11946681e-01 2.82668531e-01
-8.12101364e-01 -8.80636156e-01 -1.05014884e+00 -3.05317193e-01
5.51678479e-01 -7.90376067e-01 1.46061587e+00 -6.97935641e-01
-1.56083524e+00 7.99290717e-01 -2.85877705e-01 -3.28075200e-01
8.43152881e-01 -8.20675313e-01 2.48403013e-01 3.10730338e-01
1.36316001e-01 1.20341289e+00 8.80231380e-01 -1.18078470e+00
-4.93764222e-01 -3.80280703e-01 1.94142852e-02 5.82216978e-01
-1.13827884e-01 -1.41608432e-01 -8.35844457e-01 -5.87896407e-01
5.62020577e-02 -1.36545861e+00 -3.79333168e-01 4.02322739e-01
-5.59280336e-01 -1.97430089e-01 6.00301683e-01 -7.06020117e-01
9.51766551e-01 -1.66088033e+00 7.25456178e-01 1.24505192e-01
4.18015510e-01 -8.61473233e-02 -9.79603380e-02 3.94525617e-01
-1.12389803e-01 -2.32753187e-01 -1.13470361e-01 -5.72100222e-01
2.18011048e-02 -1.33553788e-01 -7.59136081e-02 7.04412222e-01
3.43173891e-01 1.25513446e+00 -8.58568788e-01 -4.85234559e-01
4.51470464e-01 6.19524360e-01 -6.57168567e-01 3.76851946e-01
-2.17220649e-01 7.90080428e-01 -3.10139745e-01 5.41820526e-01
1.96155101e-01 -3.05670112e-01 1.59316793e-01 -4.66616720e-01
1.02346212e-01 2.31953666e-01 -7.87752271e-01 2.30292749e+00
-5.65006256e-01 8.78109634e-01 -3.08654636e-01 -8.89781117e-01
7.78541327e-01 5.00386544e-02 7.50236034e-01 -3.72358173e-01
6.46407902e-03 -9.30826813e-02 -3.25054109e-01 -3.24631929e-01
6.21976972e-01 1.50210395e-01 -3.27507585e-01 1.80406585e-01
2.01615587e-01 -1.34237140e-01 -7.33079761e-02 8.89586061e-02
8.04544151e-01 9.54408348e-01 3.97888631e-01 -4.40132283e-02
3.07448953e-01 7.10472539e-02 3.68757248e-01 4.25816804e-01
-2.59567201e-01 8.47443342e-01 4.24581200e-01 -6.92476511e-01
-1.21201181e+00 -1.15115643e+00 2.18701228e-01 1.27123237e+00
2.01977640e-01 -7.48724639e-01 -8.02127004e-01 -7.58103788e-01
2.03272164e-01 7.99289793e-02 -9.68479097e-01 1.63861457e-02
-8.43963504e-01 -2.55838037e-01 4.01337624e-01 9.01861131e-01
3.40697885e-01 -8.11545253e-01 -8.01264822e-01 1.45665854e-01
-4.94677335e-01 -9.57635701e-01 -7.55287707e-01 4.97699045e-02
-7.22079098e-01 -1.11142254e+00 -1.00563407e+00 -6.83151603e-01
7.32439578e-01 -2.73100343e-02 1.30344999e+00 -1.03892520e-01
-6.31818175e-01 5.60992420e-01 -1.64145991e-01 -1.12668768e-01
1.32255778e-01 3.71027172e-01 4.29682843e-02 -3.41509908e-01
2.76817679e-01 -3.54469627e-01 -8.64406645e-01 4.02520925e-01
-1.19722195e-01 8.00280124e-02 6.25550926e-01 9.36027586e-01
6.30101800e-01 -6.62617207e-01 2.51043409e-01 -4.76353556e-01
4.31137025e-01 -9.08691809e-02 -3.86806846e-01 1.69475898e-01
-3.05535346e-01 3.36724997e-01 3.21367562e-01 -3.45596373e-01
-6.71827734e-01 5.92657387e-01 8.70486051e-02 -5.66560090e-01
-1.04709171e-01 3.41635823e-01 1.81047887e-01 -3.12551051e-01
6.15785480e-01 1.53679758e-01 2.25174889e-01 -5.18997788e-01
4.13042337e-01 2.43115604e-01 8.33221138e-01 -5.65879643e-01
9.65646029e-01 2.83159792e-01 1.80832148e-01 -6.16568625e-01
-8.12367201e-01 -4.14573848e-01 -1.09073865e+00 -3.28836203e-01
1.02649045e+00 -1.00886452e+00 -1.09057128e+00 3.82242441e-01
-1.04217482e+00 -3.48627925e-01 7.96858072e-02 4.44735080e-01
-1.20719957e+00 3.24139416e-01 -6.21947229e-01 -5.58610260e-01
-1.86527163e-01 -1.17399955e+00 1.46226394e+00 -5.91545291e-02
-6.90324664e-01 -7.47815490e-01 3.26738477e-01 3.02395016e-01
1.99025378e-01 6.44904375e-01 4.97387052e-01 -1.83825269e-01
-6.05980217e-01 -3.56665373e-01 -8.94915089e-02 -2.48008579e-01
-1.26466200e-01 -1.57217309e-01 -6.69604063e-01 -6.12953246e-01
-5.52351832e-01 -7.89039910e-01 8.04704905e-01 4.35851663e-01
1.19795167e+00 -6.15324713e-02 -5.01120448e-01 6.96220219e-01
1.05683827e+00 -2.88934559e-01 4.14787322e-01 5.57787299e-01
9.71872151e-01 5.80616355e-01 8.81461263e-01 3.42833132e-01
7.41591632e-01 1.27529490e+00 4.40713882e-01 -2.50832140e-01
7.09483074e-03 -7.51014948e-01 2.09563464e-01 4.30324137e-01
-3.29523593e-01 2.65507340e-01 -8.61070395e-01 1.93392962e-01
-2.04651594e+00 -8.28119099e-01 3.55779558e-01 2.20551157e+00
6.86094940e-01 3.63903455e-02 5.18520355e-01 -8.96960348e-02
5.30649781e-01 2.86371201e-01 -6.02621615e-01 -5.39882295e-02
4.99179989e-01 1.65379673e-01 4.68825847e-01 3.94643992e-01
-1.52469754e+00 1.06029332e+00 7.30840731e+00 4.01449412e-01
-1.00191069e+00 -2.22515941e-01 4.66450691e-01 -4.37647462e-01
1.74397528e-01 -3.07087690e-01 -6.20306671e-01 1.63746879e-01
5.07646084e-01 1.39225870e-01 5.51637053e-01 9.07332480e-01
5.07482812e-02 3.07228900e-02 -1.58469570e+00 1.26224899e+00
3.18461627e-01 -1.10120332e+00 -2.74715960e-01 1.16625585e-01
5.08914173e-01 -7.20731020e-02 1.75026014e-01 2.29134262e-01
1.79262891e-01 -1.29844308e+00 9.52248514e-01 5.80796897e-01
8.77781451e-01 -9.27166402e-01 3.69424820e-01 1.67387322e-01
-1.43065023e+00 4.40983288e-02 -1.89063653e-01 -5.29440157e-02
2.30858132e-01 -6.24269340e-03 -8.78916025e-01 2.66241699e-01
8.48537564e-01 9.08908308e-01 -6.43295050e-01 1.06784403e+00
-4.92445022e-01 1.33151282e-02 -2.01860040e-01 -1.20547742e-01
1.83089688e-01 2.02254772e-01 4.34450984e-01 1.07340443e+00
1.56795010e-01 -2.30114862e-01 5.03797770e-01 6.36477113e-01
-6.11096732e-02 -2.97695473e-02 -6.14589572e-01 2.09903866e-01
4.39380884e-01 1.09545219e+00 -5.22371829e-01 -1.39011204e-01
-2.18976200e-01 1.21853185e+00 8.50256622e-01 2.52381295e-01
-9.13284957e-01 -4.34388012e-01 7.64218867e-01 1.49346426e-01
4.89928067e-01 -6.07614696e-01 -4.63815220e-02 -1.10249758e+00
2.13597044e-01 -9.21189725e-01 1.82034165e-01 -8.33111584e-01
-9.59177017e-01 5.34737825e-01 1.95608065e-01 -1.44095230e+00
-7.37947166e-01 -6.25616729e-01 -2.70641565e-01 7.09933400e-01
-1.20168698e+00 -1.30884778e+00 -4.03735906e-01 5.29111147e-01
5.62904596e-01 9.50021818e-02 7.97954559e-01 1.80898160e-02
-2.75959402e-01 7.89751232e-01 -2.16495141e-01 2.30736613e-01
8.72705817e-01 -1.38200772e+00 7.97222018e-01 5.59981406e-01
2.44118616e-01 5.97608328e-01 7.40389287e-01 -4.20367777e-01
-1.51906109e+00 -6.94260120e-01 6.84026897e-01 -8.34360421e-01
4.37244177e-01 -6.06217921e-01 -4.58805531e-01 9.22371626e-01
-1.03753097e-01 1.57660529e-01 3.84066194e-01 4.31073576e-01
-4.40459758e-01 1.61820967e-02 -7.58582115e-01 7.08077967e-01
1.29929245e+00 -4.93026555e-01 -6.38324022e-01 3.10607255e-01
5.89727163e-01 -8.20630968e-01 -8.72010291e-01 3.36608917e-01
9.19083357e-01 -7.00489998e-01 1.36812282e+00 -6.44728661e-01
6.46877646e-01 -2.49591842e-01 4.85879183e-02 -1.52953303e+00
-5.35031736e-01 -5.86973190e-01 -3.26810688e-01 5.64273953e-01
4.50439394e-01 -2.03812510e-01 1.19727659e+00 5.29028296e-01
1.04935229e-01 -9.07248557e-01 -9.26711738e-01 -5.88650823e-01
1.73185486e-02 -1.38898671e-01 3.30446661e-01 6.09642565e-01
1.97312176e-01 2.52847075e-01 -8.04678977e-01 -1.52085394e-01
7.83702850e-01 -8.39189533e-03 1.29365706e+00 -1.11852157e+00
-4.40604419e-01 -4.28964376e-01 -7.73871660e-01 -1.51372361e+00
1.36894807e-01 -5.36910057e-01 5.14835417e-01 -1.30925763e+00
1.73331082e-01 -2.37699896e-01 -7.13515952e-02 5.38938999e-01
-3.18723440e-01 6.24996960e-01 4.11275089e-01 2.74595827e-01
-8.02762926e-01 6.45468116e-01 1.30377257e+00 -1.06651470e-01
-2.60053158e-01 -1.18782029e-01 -5.42736351e-01 7.23062396e-01
5.45051098e-01 -2.84919739e-01 -2.97579318e-01 -4.12776411e-01
1.30920246e-01 2.14231521e-01 5.08338153e-01 -1.08692408e+00
3.15519363e-01 -7.97854587e-02 8.40135455e-01 -7.77624309e-01
7.44264126e-01 -5.34349263e-01 -5.17647378e-02 4.90040898e-01
-6.63181245e-01 3.92721087e-01 1.03223221e-02 6.24307513e-01
-1.59228276e-02 2.31204167e-01 5.94444513e-01 -1.65614679e-01
-8.28172147e-01 5.07356167e-01 -1.13796629e-01 1.56015873e-01
1.06519485e+00 -2.78627872e-01 3.96687686e-02 -5.76922774e-01
-8.49195480e-01 2.29377508e-01 5.24212420e-01 6.35804951e-01
7.46024549e-01 -1.53428495e+00 -7.03112781e-01 5.55056632e-02
3.48925471e-01 1.26104280e-02 -2.91660298e-02 8.89885247e-01
-5.96427977e-01 5.27454555e-01 -2.53566533e-01 -8.45733523e-01
-1.24777508e+00 5.50585151e-01 4.10995007e-01 -3.48447680e-01
-8.41264546e-01 9.31231558e-01 1.29852608e-01 -6.31768346e-01
3.88842076e-01 -1.20687298e-01 -7.08125718e-03 -2.21939877e-01
4.11615163e-01 4.24949467e-01 -9.92147159e-03 -8.15053046e-01
-5.98430336e-01 9.46015537e-01 -1.82884648e-01 -3.07484657e-01
1.28064227e+00 -8.76294747e-02 2.22204700e-01 3.15369368e-01
1.41136396e+00 -3.13921064e-01 -1.84888101e+00 -1.54345080e-01
-8.12852383e-02 -7.25895166e-01 -1.29587129e-01 -8.57988596e-01
-9.37610209e-01 8.69321704e-01 3.81791174e-01 -4.41490799e-01
7.89562345e-01 -1.88964400e-02 7.31672943e-01 4.90975291e-01
4.95262325e-01 -1.14489341e+00 4.81516987e-01 3.87002081e-01
1.09506178e+00 -1.25127339e+00 1.48786306e-01 -2.86035150e-01
-6.22880042e-01 1.17288542e+00 9.16275501e-01 -4.51896489e-01
3.65727574e-01 1.10883951e-01 9.18691680e-02 -1.23016775e-01
-7.37701118e-01 -1.11861117e-01 6.49118781e-01 7.29803681e-01
5.84205925e-01 -6.24651387e-02 5.77198602e-02 -1.10949725e-02
-4.62174118e-01 -8.20762850e-03 3.70033197e-02 1.00043809e+00
-3.96686584e-01 -8.81860137e-01 -4.64397848e-01 2.17741221e-01
-1.73375815e-01 2.97788922e-02 -5.24182141e-01 1.00269043e+00
-3.09756607e-01 5.89723408e-01 2.61759907e-01 -6.58743322e-01
4.44956183e-01 -7.63850361e-02 8.79746795e-01 -3.76644701e-01
-3.98666769e-01 1.01938946e-02 -2.15108115e-02 -1.08338916e+00
-3.13437432e-01 -6.01662636e-01 -1.02611578e+00 -3.14908862e-01
-1.05094999e-01 -1.62988424e-01 5.58797777e-01 8.46428752e-01
3.30460548e-01 3.80189866e-01 4.89215463e-01 -1.62110484e+00
-5.82472801e-01 -7.73533523e-01 -1.02165386e-01 4.73057419e-01
4.44560558e-01 -1.10673606e+00 1.64004430e-01 -2.57842839e-01] | [6.98205041885376, -0.9448651671409607] |
c1fd8626-e7fe-48f3-aee4-60550aebc2eb | multi-step-joint-modality-attention-network | 2001.06206 | null | https://arxiv.org/abs/2001.06206v1 | https://arxiv.org/pdf/2001.06206v1.pdf | Multi-step Joint-Modality Attention Network for Scene-Aware Dialogue System | Understanding dynamic scenes and dialogue contexts in order to converse with users has been challenging for multimodal dialogue systems. The 8-th Dialog System Technology Challenge (DSTC8) proposed an Audio Visual Scene-Aware Dialog (AVSD) task, which contains multiple modalities including audio, vision, and language, to evaluate how dialogue systems understand different modalities and response to users. In this paper, we proposed a multi-step joint-modality attention network (JMAN) based on recurrent neural network (RNN) to reason on videos. Our model performs a multi-step attention mechanism and jointly considers both visual and textual representations in each reasoning process to better integrate information from the two different modalities. Compared to the baseline released by AVSD organizers, our model achieves a relative 12.1% and 22.4% improvement over the baseline on ROUGE-L score and CIDEr score. | ['Lun-Wei Ku', 'Chao-Chun Hsu', 'Kuan-Yen Lin', 'Yun-Wei Chu'] | 2020-01-17 | null | null | null | null | ['scene-aware-dialogue'] | ['computer-vision'] | [-2.10566580e-01 7.84765184e-03 1.17354147e-01 -4.14514363e-01
-8.72601449e-01 -7.35632241e-01 1.04104745e+00 -3.69028330e-01
-4.40667838e-01 4.58352983e-01 1.15048075e+00 -5.50416932e-02
5.55056930e-01 -1.05490245e-01 -1.24457352e-01 -1.71973035e-01
4.93178219e-01 6.70754611e-01 -3.81600996e-03 -6.45810723e-01
1.14650384e-01 3.47755283e-01 -1.09736717e+00 1.31719148e+00
5.57616591e-01 8.48315954e-01 2.77088314e-01 1.37132585e+00
-3.46199989e-01 1.59058869e+00 -5.96271455e-01 -3.88475984e-01
-4.74959135e-01 -6.70449436e-01 -1.42548490e+00 2.98178852e-01
3.30224127e-01 -9.24014509e-01 -8.69140029e-01 5.54891169e-01
6.72141075e-01 5.54908037e-01 8.29868257e-01 -1.45151579e+00
-8.66665244e-01 4.94163483e-01 -4.32683915e-01 -8.43455922e-03
1.02270091e+00 5.38304031e-01 1.00879228e+00 -1.18269300e+00
7.63348520e-01 2.10739303e+00 1.97274074e-01 9.31781530e-01
-8.00096512e-01 -1.54532000e-01 1.66428357e-01 5.47059894e-01
-7.89705098e-01 -6.93351269e-01 3.83431584e-01 -3.01457852e-01
1.32011485e+00 3.45969379e-01 2.97288239e-01 1.60966265e+00
-2.74943143e-01 1.47999561e+00 6.98990226e-01 -4.02553737e-01
-2.85621673e-01 2.67541379e-01 -2.64453073e-03 6.08745039e-01
-1.00513017e+00 -6.54945731e-01 -6.72135234e-01 6.66130707e-02
6.20360017e-01 -7.68444315e-02 -2.91081786e-01 -1.60740644e-01
-1.38798535e+00 9.48111475e-01 5.21702528e-01 3.26501429e-01
-3.54061007e-01 2.14268621e-02 8.13308954e-01 1.85973674e-01
2.29574829e-01 4.06205267e-01 -1.39387101e-01 -4.71746713e-01
-1.92226082e-01 1.38471685e-02 9.84116316e-01 7.87677824e-01
-5.55116832e-02 1.44486250e-02 -9.09517348e-01 1.27293050e+00
6.28968239e-01 5.72326124e-01 1.63087368e-01 -1.43262255e+00
8.18007231e-01 6.81520164e-01 2.44487464e-01 -7.97455192e-01
-5.77672958e-01 6.56600356e-01 -8.36399555e-01 -1.32348880e-01
4.25958872e-01 -4.48474228e-01 -6.08640671e-01 1.62055266e+00
2.27372184e-01 -3.16765070e-01 5.19800007e-01 1.17437935e+00
1.76267612e+00 1.05134463e+00 3.59813482e-01 -9.14962515e-02
1.63293040e+00 -1.58772206e+00 -1.18158007e+00 -1.03271432e-01
5.03277063e-01 -8.84760082e-01 1.33889818e+00 1.55374035e-01
-1.34821844e+00 -6.97386742e-01 -5.93102932e-01 -6.06295228e-01
-3.89955282e-01 3.88053864e-01 1.27279907e-01 1.49401665e-01
-1.30911303e+00 -3.54941458e-01 -3.01778287e-01 -8.23287010e-01
-1.89667335e-03 -3.26732509e-02 -4.34093595e-01 -1.28154112e-02
-1.35658884e+00 1.06747055e+00 -6.03941120e-02 2.30176032e-01
-1.14219880e+00 -2.16444075e-01 -9.60430443e-01 2.93649640e-02
4.41810429e-01 -7.00285554e-01 1.75965619e+00 -7.06760168e-01
-1.91840732e+00 8.32621038e-01 -3.11419249e-01 -4.14472550e-01
4.28547174e-01 -4.68827397e-01 -4.02392775e-01 7.73835421e-01
-1.33514851e-01 1.17241776e+00 5.51469982e-01 -1.30131280e+00
-4.69303787e-01 -1.00397386e-01 5.09572923e-01 9.37525392e-01
-1.35274202e-01 3.93881768e-01 -9.98758554e-01 -2.28853226e-01
-3.82311523e-01 -8.54792655e-01 1.10331148e-01 -1.28802940e-01
-4.82078075e-01 -5.33946514e-01 1.24686861e+00 -9.05864418e-01
7.67858624e-01 -2.08061504e+00 6.57169521e-01 -5.26537478e-01
3.57563049e-01 2.58571416e-01 -4.74334329e-01 7.86314547e-01
3.09188247e-01 -1.66512355e-01 3.03405941e-01 -8.12389314e-01
8.64922255e-02 7.69608479e-04 -4.03176516e-01 8.38975981e-02
2.04389215e-01 9.28828895e-01 -6.84145749e-01 -5.53200960e-01
5.82053900e-01 7.49208927e-01 -3.43122214e-01 6.39121413e-01
-5.01237273e-01 8.15510690e-01 -1.50614202e-01 5.43598413e-01
3.00321609e-01 -4.87445623e-01 3.60771388e-01 -4.89825726e-01
1.59350216e-01 1.12875603e-01 -6.32411122e-01 1.86102259e+00
-7.84402907e-01 1.14513528e+00 5.21599174e-01 -3.91293615e-01
4.51206684e-01 9.51059997e-01 2.29727030e-01 -9.50650752e-01
2.59743482e-01 -5.37394822e-01 -3.86906862e-01 -9.49764490e-01
9.26672578e-01 2.70799011e-01 -5.30016780e-01 5.40851951e-01
4.08989668e-01 2.64832228e-02 1.89683046e-02 1.01539195e+00
7.28928328e-01 -1.24357633e-01 5.51192127e-02 4.00310487e-01
9.57592070e-01 -8.21328759e-02 1.76419187e-02 8.19422662e-01
-4.45922971e-01 6.06871963e-01 8.63591909e-01 -1.68581843e-01
-7.48575687e-01 -8.66443753e-01 5.79103291e-01 1.86101639e+00
2.95441270e-01 -3.54431421e-01 -5.54822624e-01 -7.70890594e-01
-2.57958382e-01 7.86891282e-01 -5.59160590e-01 -3.06454189e-02
-2.96770096e-01 -1.06884904e-01 8.12895179e-01 5.88888824e-01
8.86831880e-01 -1.40053546e+00 -4.00064528e-01 -9.92814377e-02
-8.51320148e-01 -1.64545262e+00 -6.31271124e-01 -6.24032974e-01
-7.06726983e-02 -9.41031098e-01 -9.58765626e-01 -6.40957892e-01
1.30639926e-01 5.07308722e-01 1.02791548e+00 -2.32977062e-01
-8.05185959e-02 1.15821242e+00 -6.08130395e-01 -3.46377231e-02
-4.89149541e-01 -2.06322923e-01 -8.32661167e-02 1.89084277e-01
1.00103043e-01 3.20470482e-01 -4.46291298e-01 4.45317954e-01
-4.74061400e-01 3.74872267e-01 2.21565247e-01 9.40384448e-01
-1.29442617e-01 -6.48436844e-01 4.58099157e-01 -5.85062802e-01
1.03984535e+00 -5.01315773e-01 -4.64513935e-02 7.45274484e-01
1.98783398e-01 -3.20863098e-01 3.16984802e-01 -2.77161390e-01
-1.80343664e+00 6.00174163e-03 -6.69011250e-02 -4.08761024e-01
-3.88029128e-01 2.63549268e-01 -2.68578440e-01 2.79925883e-01
3.44439685e-01 -2.89074965e-02 6.22731633e-02 -3.02154362e-01
8.48989189e-01 1.11545312e+00 8.57766330e-01 -3.36806476e-01
3.73896807e-02 3.35077733e-01 -6.74530208e-01 -1.03269255e+00
-5.95155895e-01 -7.28095055e-01 -4.75595027e-01 -7.86328852e-01
1.54469717e+00 -1.28996837e+00 -1.46346879e+00 4.89290208e-01
-1.53965843e+00 -5.98491728e-01 3.38948637e-01 3.41014296e-01
-5.45321643e-01 6.18826687e-01 -1.05573487e+00 -1.24849129e+00
-3.93084526e-01 -1.34013546e+00 1.24843299e+00 4.33841914e-01
-3.42027426e-01 -9.26600099e-01 -4.57377769e-02 1.12909281e+00
4.08753633e-01 -1.57894194e-01 4.41439539e-01 -8.02235126e-01
-2.91536480e-01 1.64233461e-01 -7.92662382e-01 -9.43382159e-02
-8.93631279e-02 -5.43525442e-02 -1.16966116e+00 -1.94795355e-01
-1.38898730e-01 -1.12541544e+00 9.20697331e-01 2.24158272e-01
9.02745962e-01 -3.43765795e-01 5.59057742e-02 -1.55806407e-01
6.33915603e-01 4.20080304e-01 5.71274281e-01 -9.89057124e-02
9.82249081e-01 8.17626238e-01 6.39413476e-01 6.08336687e-01
9.07182157e-01 9.21173453e-01 4.92690355e-01 -3.16106915e-01
-2.23963797e-01 -3.21059749e-02 6.08298242e-01 6.26796961e-01
2.25274891e-01 -8.24701548e-01 -8.93242300e-01 4.47768569e-01
-2.09562159e+00 -1.00691187e+00 -9.99375433e-02 1.44033766e+00
5.70521176e-01 -3.72990310e-01 5.70011854e-01 -4.61426049e-01
6.80079103e-01 3.76905024e-01 -4.04919624e-01 -7.15414822e-01
-3.06479216e-01 -7.29265511e-01 -2.36571237e-01 7.09555387e-01
-1.17050517e+00 1.13374138e+00 6.03409004e+00 4.63665396e-01
-8.69245231e-01 1.68391049e-01 5.37235022e-01 -8.05524886e-02
4.16919515e-02 -3.36973011e-01 -5.82522810e-01 4.56314683e-02
1.09693468e+00 1.92187533e-01 5.56690216e-01 4.78487939e-01
2.04525486e-01 -2.26654604e-01 -8.33159626e-01 1.03151000e+00
4.91224617e-01 -1.24524546e+00 1.55495033e-01 -2.89986849e-01
4.42202479e-01 -3.10420394e-02 1.97502881e-01 7.08686769e-01
4.77687865e-01 -1.18481529e+00 2.92458653e-01 7.20271230e-01
7.17125714e-01 -4.71533149e-01 8.15152705e-01 8.52584541e-02
-1.28942561e+00 -3.96549553e-02 1.90740943e-01 3.09745312e-01
5.84934413e-01 -5.21799386e-01 -1.32604361e+00 3.11868012e-01
7.13525474e-01 7.70608783e-01 -4.47717607e-01 5.19195914e-01
3.64992097e-02 1.35380283e-01 1.75150335e-01 -1.25225544e-01
3.54704827e-01 2.08893120e-01 6.60624027e-01 1.54855847e+00
-2.39127561e-01 1.96077093e-01 1.85411066e-01 4.29323435e-01
-4.15227056e-01 -7.81287253e-02 -6.72358453e-01 -3.86511296e-01
3.72158080e-01 1.32983828e+00 -4.00195718e-01 -5.71354926e-01
-6.61445916e-01 1.34008980e+00 2.34288290e-01 6.28251433e-01
-8.40801716e-01 -2.11710721e-01 4.43139851e-01 -8.43097448e-01
1.42886400e-01 -2.08307460e-01 5.35127595e-02 -1.22667277e+00
-4.97804224e-01 -1.04942715e+00 8.72656465e-01 -1.42183113e+00
-1.32878804e+00 8.32312107e-01 -6.24440350e-02 -1.07518840e+00
-5.35738289e-01 -6.09479666e-01 -3.92803848e-01 6.92633808e-01
-1.02658677e+00 -1.43610239e+00 -4.89838451e-01 9.40972805e-01
1.39690387e+00 -4.57418680e-01 1.00153184e+00 4.60580587e-02
-6.63382471e-01 5.52936673e-01 -2.30443612e-01 3.54802668e-01
1.18460274e+00 -1.31010640e+00 -1.74806230e-02 3.19987983e-01
-5.18280081e-02 2.46075302e-01 6.45563841e-01 -4.30744737e-01
-1.56561840e+00 -7.60940194e-01 8.22004259e-01 -6.34195209e-01
5.62458694e-01 -2.51286596e-01 -8.04267287e-01 6.66076124e-01
1.23206007e+00 -6.36225045e-01 6.20164633e-01 1.38293937e-01
-5.10184348e-01 2.88971603e-01 -1.09440506e+00 8.81806612e-01
5.74375629e-01 -1.12564862e+00 -7.06099391e-01 5.36760986e-01
8.43883395e-01 -6.56564772e-01 -9.77643549e-01 2.43549749e-01
5.72367907e-01 -8.11091363e-01 1.21259642e+00 -7.36307919e-01
6.91828907e-01 9.89700481e-03 -3.22620571e-01 -9.12979841e-01
2.11513340e-01 -6.19932413e-01 -2.19887942e-01 1.28899622e+00
3.01480412e-01 5.34907617e-02 1.42889172e-01 8.45407367e-01
-6.31403998e-02 -1.50904551e-01 -7.48900056e-01 4.79127839e-02
-4.16783333e-01 -2.47289747e-01 7.78504834e-02 8.51997614e-01
5.04935980e-01 1.22536850e+00 -1.08083510e+00 -4.04121317e-02
-4.04343568e-02 -9.35695544e-02 1.03824484e+00 -7.76919067e-01
-8.84308144e-02 -5.74200690e-01 1.87975422e-01 -1.38288677e+00
3.57088715e-01 -4.06141460e-01 1.02801740e-01 -1.99478185e+00
3.57916266e-01 5.38560152e-01 1.13181852e-01 3.56053740e-01
-3.11604515e-02 1.79664493e-01 6.40060186e-01 1.70167521e-01
-1.55606186e+00 9.93197381e-01 1.30543327e+00 -4.32029068e-01
-4.24831510e-01 -3.18372816e-01 -5.71050048e-01 6.20099425e-01
4.20660347e-01 3.34538996e-01 -3.47463608e-01 -5.45803368e-01
-1.23854667e-01 8.66677463e-01 4.50036854e-01 -5.88719130e-01
5.91948867e-01 -3.53705063e-02 4.19204354e-01 -1.10734546e+00
1.16272628e+00 -6.14944100e-01 -4.07814384e-01 2.33546615e-01
-1.00928354e+00 7.52752647e-02 4.31224853e-01 6.77971363e-01
-4.01495248e-01 2.75877327e-01 4.72210974e-01 -2.01285690e-01
-1.10503912e+00 -2.51903713e-01 -9.85315204e-01 -3.59902228e-03
6.47214472e-01 2.33833835e-01 -8.19067478e-01 -1.41934812e+00
-1.00200522e+00 9.44076717e-01 -4.28645089e-02 8.01080525e-01
9.09873724e-01 -1.37842333e+00 -8.21613431e-01 -4.14875388e-01
3.27961177e-01 -5.02718210e-01 9.83720064e-01 8.65971923e-01
-4.14186031e-01 8.16769540e-01 -3.16457629e-01 -6.79582417e-01
-1.65331042e+00 2.47809485e-01 3.30682963e-01 -2.48987824e-01
-5.00520051e-01 9.63101506e-01 3.06802005e-01 -7.51896203e-01
7.19754875e-01 2.35108465e-01 -9.58735704e-01 4.28039849e-01
7.56092310e-01 2.74809599e-01 -4.33655292e-01 -1.05371714e+00
-2.67363220e-01 2.09697291e-01 -2.75719404e-01 -6.85640275e-01
9.01527643e-01 -7.48243630e-01 1.18702009e-01 6.85037315e-01
1.12301552e+00 -3.47074479e-01 -1.16687644e+00 -4.98204440e-01
-3.51929277e-01 -1.42330945e-01 -2.23719314e-01 -1.21669245e+00
-8.13383102e-01 1.11035001e+00 5.42364001e-01 4.05967534e-01
1.00227869e+00 2.32968450e-01 7.22873330e-01 7.92478383e-01
-2.91314512e-01 -1.23629797e+00 8.05815041e-01 9.91157949e-01
1.57851589e+00 -1.63498127e+00 -3.94220680e-01 -1.21192716e-01
-1.89070451e+00 1.22277939e+00 1.03862679e+00 4.50471342e-01
1.14648275e-01 -3.16914111e-01 3.92384768e-01 -2.15476647e-01
-1.26125133e+00 -2.79204518e-01 5.26268482e-01 4.17348057e-01
4.69362766e-01 -2.05225617e-01 3.16038221e-01 5.03562510e-01
4.43279147e-01 -3.14666361e-01 4.75736737e-01 6.06488943e-01
-1.60223752e-01 -7.09690034e-01 -3.22523326e-01 -4.76919524e-02
-9.32000801e-02 1.49164751e-01 -1.11666977e+00 7.91226089e-01
-7.19574571e-01 1.53764963e+00 1.22565240e-01 -5.53277016e-01
4.60012197e-01 1.98890507e-01 1.07816011e-01 -4.40186441e-01
-7.55165040e-01 2.69659609e-01 7.01916039e-01 -6.64591312e-01
-6.40099108e-01 -4.74433869e-01 -1.19854105e+00 -2.22247303e-01
1.58179834e-01 -6.05041049e-02 3.90926331e-01 1.09189844e+00
4.14059252e-01 7.28523672e-01 7.00329185e-01 -8.74912381e-01
-5.55484146e-02 -1.24892795e+00 -1.07252091e-01 3.69894147e-01
4.49910611e-01 -5.39339304e-01 -1.40848607e-01 -1.39280543e-01] | [10.887042045593262, 1.3133264780044556] |
13463e49-5d30-40c9-8d27-a7501716b347 | context-aware-image-denoising-with-auto | 2101.05833 | null | https://arxiv.org/abs/2101.05833v1 | https://arxiv.org/pdf/2101.05833v1.pdf | Context-Aware Image Denoising with Auto-Threshold Canny Edge Detection to Suppress Adversarial Perturbation | This paper presents a novel context-aware image denoising algorithm that combines an adaptive image smoothing technique and color reduction techniques to remove perturbation from adversarial images. Adaptive image smoothing is achieved using auto-threshold canny edge detection to produce an accurate edge map used to produce a blurred image that preserves more edge features. The proposed algorithm then uses color reduction techniques to reconstruct the image using only a few representative colors. Through this technique, the algorithm can reduce the effects of adversarial perturbations on images. We also discuss experimental data on classification accuracy. Our results showed that the proposed approach reduces adversarial perturbation in adversarial attacks and increases the robustness of the deep convolutional neural network models. | ['Wu-chi Feng', 'Yeganeh Jalalpour', 'Li-Yun Wang'] | 2021-01-14 | null | null | null | null | ['image-smoothing'] | ['computer-vision'] | [ 4.59607840e-01 -4.28733081e-01 6.96045816e-01 -1.29496187e-01
-3.04551065e-01 -8.61469269e-01 3.70443612e-01 -1.82419077e-01
-7.71220803e-01 6.45937324e-01 -2.21156534e-02 -2.03699559e-01
1.39472649e-01 -8.53191614e-01 -8.23642969e-01 -8.73112798e-01
1.35997370e-01 -6.39203370e-01 4.00205404e-01 -4.42130357e-01
4.97619748e-01 9.73345160e-01 -1.32014644e+00 3.03583592e-01
7.37700760e-01 8.14217150e-01 -9.65683386e-02 1.32915986e+00
8.47070143e-02 8.69640648e-01 -9.56421375e-01 -3.62282991e-01
8.86487365e-01 -2.95246005e-01 -3.07106853e-01 -5.44137508e-02
7.21087992e-01 -5.94352543e-01 -4.92197484e-01 1.62230563e+00
5.27105868e-01 3.43226135e-01 5.19141018e-01 -1.06110668e+00
-1.13419008e+00 1.40266463e-01 -6.94507301e-01 6.74158573e-01
-1.35128051e-01 2.76220739e-01 -1.21663764e-01 -3.89861315e-01
4.94537801e-01 1.30886304e+00 7.35960543e-01 6.56484246e-01
-1.05403769e+00 -8.72640967e-01 -1.70800045e-01 5.40914893e-01
-1.13877106e+00 -2.56735563e-01 1.14737272e+00 9.87798199e-02
6.40390754e-01 6.32057309e-01 2.61844009e-01 5.21256268e-01
7.10298061e-01 8.72962922e-02 1.48327148e+00 -7.43385136e-01
1.42628029e-01 6.04994521e-02 1.69830937e-02 7.64080107e-01
2.99826443e-01 4.13690567e-01 -9.22651812e-02 -5.74846789e-02
6.35614634e-01 1.73948154e-01 -4.15602177e-01 2.95810759e-01
-4.60025579e-01 6.10207498e-01 6.97192252e-01 4.00692940e-01
-5.64614594e-01 3.43779027e-01 4.52443898e-01 5.54307818e-01
2.83210695e-01 -1.27778221e-02 -1.01494901e-01 3.08956563e-01
-6.36374295e-01 -4.30106297e-02 5.39559364e-01 6.43036783e-01
6.42646790e-01 6.84973001e-01 1.01291813e-01 5.76760232e-01
1.15766071e-01 5.79379976e-01 4.33301955e-01 -1.08956480e+00
5.91387115e-02 2.75511503e-01 5.99752106e-02 -1.31510139e+00
1.36743829e-01 -6.97269440e-02 -8.07578027e-01 1.42616951e+00
2.79564738e-01 -3.48381609e-01 -1.44591808e+00 1.37638068e+00
1.93931520e-01 4.79536831e-01 3.82757574e-01 7.77682066e-01
5.67986727e-01 7.91278481e-01 2.42194116e-01 -8.89004581e-03
1.09852839e+00 -6.53803110e-01 -9.84355271e-01 -2.17273071e-01
-1.69867188e-01 -1.16188300e+00 7.04369247e-01 6.03213668e-01
-1.04838312e+00 -7.43832529e-01 -1.43777037e+00 1.55952737e-01
-7.63027966e-01 -3.40703994e-01 3.90014976e-01 1.20185256e+00
-1.07059765e+00 7.52640367e-01 -5.69235444e-01 1.55858360e-02
5.40844858e-01 5.10744929e-01 -7.22173691e-01 -1.31513089e-01
-1.21074867e+00 1.02124727e+00 5.62364519e-01 1.10867053e-01
-6.73868418e-01 -5.04012823e-01 -7.13258684e-01 -1.14052646e-01
-2.07522646e-01 -3.27113420e-01 6.07347012e-01 -1.63543594e+00
-1.42503846e+00 7.27568328e-01 4.10704501e-02 -7.81755328e-01
6.30745173e-01 -2.11852938e-01 -6.66499436e-01 5.72368979e-01
-7.34466434e-01 3.80842328e-01 1.42597187e+00 -1.51578033e+00
-5.22736371e-01 -3.13455462e-01 -2.91146904e-01 1.77795768e-01
-2.71085054e-01 7.87896290e-02 -2.99727321e-01 -1.13191783e+00
-3.28362174e-02 -7.34357834e-01 -2.21594021e-01 9.31136236e-02
7.49616139e-03 6.11549854e-01 1.56891906e+00 -1.02781498e+00
8.25121045e-01 -2.36134839e+00 -3.92924279e-01 3.92003387e-01
2.59612203e-02 7.11480916e-01 -1.67604193e-01 2.56551147e-01
-4.85762537e-01 1.28744662e-01 -3.64197195e-01 -1.80430375e-02
-3.70528549e-01 3.00321788e-01 -2.79238462e-01 7.52856672e-01
-3.29478830e-02 4.65538651e-01 -4.40152436e-01 -3.26123327e-01
5.39377153e-01 9.16099727e-01 -2.15064377e-01 7.96668828e-02
2.70147383e-01 1.94372341e-01 -2.99330242e-02 3.52437466e-01
1.41614246e+00 9.20912981e-01 -5.17260849e-01 -4.51039076e-01
1.80293709e-01 -7.87346959e-01 -1.32762098e+00 1.02270353e+00
-3.54936719e-01 9.74862933e-01 3.36047083e-01 -8.89766455e-01
1.09278357e+00 1.70228511e-01 -1.12187266e-02 -3.79859149e-01
5.96481144e-01 -1.01619087e-01 1.14618734e-01 -5.38570642e-01
3.10559034e-01 -1.03559256e-01 3.81727248e-01 1.65791899e-01
-3.43783885e-01 3.73449400e-02 -2.08659619e-01 2.30101615e-01
9.97041345e-01 -1.15730166e-02 6.69960007e-02 -2.17681274e-01
8.64620566e-01 1.15946002e-01 5.99189341e-01 8.32087934e-01
-6.77400887e-01 5.59555709e-01 -1.01019226e-01 -6.54140115e-01
-1.11890233e+00 -1.03734589e+00 1.00749217e-01 7.87550092e-01
3.10352623e-01 4.92215604e-01 -1.24862158e+00 -5.99635065e-01
1.34005491e-02 7.72580385e-01 -8.41840863e-01 -5.77396333e-01
-6.79966092e-01 -6.27697051e-01 6.34918451e-01 3.08172464e-01
1.14754164e+00 -1.14004505e+00 -3.09352487e-01 -6.53026700e-02
3.40887070e-01 -7.41020679e-01 -5.36548674e-01 1.55134164e-02
-9.82562542e-01 -1.00073493e+00 -3.48917693e-01 -1.03860772e+00
1.04633796e+00 3.19096863e-01 5.33849835e-01 4.40059990e-01
-5.53050637e-01 3.08934391e-01 -4.49181020e-01 -6.78059280e-01
-8.64111602e-01 -8.67016852e-01 -1.03644915e-01 1.10486224e-01
5.83068013e-01 -6.25578284e-01 -8.18095386e-01 -2.05301344e-01
-1.29760635e+00 -6.41525090e-01 6.57689095e-01 6.22852147e-01
4.28248852e-01 8.02802920e-01 3.38598013e-01 -1.03163397e+00
8.25761080e-01 2.11619982e-03 -6.27578199e-01 5.64277358e-02
-4.78093266e-01 1.31722867e-01 1.10895002e+00 -8.18733394e-01
-1.38590753e+00 1.44435227e-01 -5.90035841e-02 -6.47181153e-01
-5.50128877e-01 -3.09432186e-02 -2.28845775e-02 -8.96066904e-01
1.00982034e+00 4.66717571e-01 1.64895564e-01 -3.90191436e-01
5.37244141e-01 7.57393360e-01 1.09958172e+00 1.32809458e-02
1.39289701e+00 6.26485050e-01 1.52531981e-01 -8.00799251e-01
1.98872741e-02 -8.05509239e-02 -4.23886955e-01 -4.25558954e-01
7.94896722e-01 -6.07101738e-01 -5.33856332e-01 1.06807482e+00
-9.92490888e-01 3.46339191e-03 2.36691639e-01 2.28693575e-01
-1.29659697e-02 6.71309531e-01 -8.12457442e-01 -8.35827529e-01
-6.92596793e-01 -9.87300992e-01 2.52325326e-01 5.55229962e-01
3.99149597e-01 -1.12775898e+00 -2.78260678e-01 9.56984982e-02
6.07153296e-01 9.29069698e-01 5.73418796e-01 -5.04639387e-01
-4.89401430e-01 -7.03987062e-01 -2.18537543e-02 8.59156489e-01
3.70535970e-01 1.87217936e-01 -1.00030136e+00 -3.20750803e-01
3.15953881e-01 2.22503603e-01 9.81735468e-01 3.16696256e-01
1.05225134e+00 -5.20980835e-01 1.03784628e-01 8.56827438e-01
2.04367638e+00 5.30531585e-01 1.24233735e+00 6.96153522e-01
6.32698119e-01 8.86934027e-02 3.64113092e-01 -8.77961814e-02
-5.81696212e-01 -2.82721128e-02 6.86010182e-01 -4.15899545e-01
-3.40752423e-01 2.74178058e-01 2.85203308e-01 1.43009216e-01
-1.31763414e-01 -7.36901760e-02 -3.97660971e-01 3.74471843e-01
-1.29936123e+00 -1.22383904e+00 -2.13714600e-01 1.88399148e+00
8.50706339e-01 2.17198327e-01 -3.65177721e-01 4.14775580e-01
1.14719033e+00 3.59055661e-02 -4.63317126e-01 -1.10038579e+00
-7.09597170e-02 5.03512800e-01 1.23681998e+00 8.58265400e-01
-1.24292743e+00 1.05063736e+00 6.59738827e+00 6.57183170e-01
-1.26959002e+00 -3.84462029e-02 6.45918906e-01 2.48709336e-01
1.73152894e-01 -3.92385721e-01 -2.31661826e-01 4.54516619e-01
5.93670428e-01 -1.78467408e-01 5.65931439e-01 8.94240558e-01
3.13046187e-01 -1.72815040e-01 -2.39734903e-01 8.12384784e-01
1.92585737e-01 -1.14664638e+00 2.50294060e-01 -2.63586849e-01
7.53209352e-01 -4.08636659e-01 3.57276082e-01 -3.09257269e-01
6.68530703e-01 -8.09261382e-01 4.73683685e-01 7.90352404e-01
4.63969797e-01 -1.32272029e+00 8.51531327e-01 7.86369443e-02
-8.69705081e-01 -2.27277458e-01 -6.44448340e-01 -3.09963431e-02
-2.71608293e-01 1.48953632e-01 -6.21687055e-01 1.44183055e-01
8.39220405e-01 4.98802699e-02 -7.00360775e-01 9.43139911e-01
-1.94808334e-01 5.01036465e-01 -8.15047473e-02 1.78215176e-01
1.57146022e-01 -3.26167375e-01 7.08683670e-01 1.32893932e+00
1.04414366e-01 3.69466841e-01 -3.85714889e-01 4.16583776e-01
-1.96421102e-01 2.98978165e-02 -7.09513068e-01 3.01497340e-01
6.05691135e-01 1.23892117e+00 -9.55238998e-01 -3.05478841e-01
-3.04464728e-01 1.49356878e+00 -9.78232697e-02 5.70440650e-01
-7.63711274e-01 -9.21496212e-01 7.07846045e-01 -2.37943053e-01
3.14815968e-01 -1.92491502e-01 -4.57892329e-01 -3.97063255e-01
-2.23137155e-01 -9.41496253e-01 4.20711458e-01 -7.78841674e-01
-9.45921063e-01 7.28574991e-01 -2.04498902e-01 -1.03178990e+00
2.78333873e-01 -6.15605712e-01 -1.09789145e+00 1.17926025e+00
-1.24521744e+00 -9.71658349e-01 -3.93761575e-01 9.41368759e-01
4.97187048e-01 -4.82432634e-01 6.23629153e-01 -9.43360105e-03
-4.02336448e-01 5.95941544e-01 3.33473116e-01 3.59067500e-01
8.29854190e-01 -1.31282473e+00 4.03354883e-01 1.84467936e+00
-2.31820643e-01 7.10458279e-01 1.15366948e+00 -7.84194827e-01
-1.19910765e+00 -1.21135998e+00 7.25292340e-02 -1.07235787e-02
2.32283682e-01 9.23285410e-02 -1.09058940e+00 6.08041763e-01
7.60471344e-01 1.82164103e-01 4.55416799e-01 -9.19861317e-01
-3.61590207e-01 -3.88559103e-01 -1.92123318e+00 7.32055128e-01
2.46560380e-01 -2.68165827e-01 -7.24077106e-01 6.58564363e-03
5.57220221e-01 -4.06455547e-01 -6.37221932e-01 -2.27208678e-02
2.68808007e-01 -9.84232426e-01 1.11072540e+00 -3.21691036e-01
4.38047759e-02 -6.36360466e-01 6.10958263e-02 -1.29831600e+00
-4.12694603e-01 -8.28654528e-01 2.61386544e-01 1.12927401e+00
4.88476828e-02 -6.72397852e-01 6.20499015e-01 6.84417248e-01
1.44169573e-02 7.01122880e-02 -4.93900299e-01 -5.45726180e-01
5.13321953e-03 -5.78780398e-02 2.32572034e-01 8.19425762e-01
-6.74168348e-01 -6.28920913e-01 -3.90390515e-01 7.56070316e-01
9.20482039e-01 -6.21836841e-01 4.56192851e-01 -5.97818673e-01
3.84531617e-02 -1.27859265e-01 -8.90536010e-01 9.13970917e-02
-6.07639775e-02 -3.26164454e-01 1.24679692e-01 -1.18599522e+00
-2.25393549e-01 2.17421323e-01 -5.66437125e-01 4.07660753e-01
-6.92901552e-01 1.02105808e+00 2.01545238e-01 -2.14532509e-01
4.98868376e-02 -2.40120769e-01 1.11892974e+00 -2.09628657e-01
-5.21848984e-02 -1.91943258e-01 -7.62198746e-01 7.84346759e-01
1.19855785e+00 -7.12275505e-01 -4.38992292e-01 -4.32310551e-01
-3.79759908e-01 -5.79958320e-01 4.83002335e-01 -1.29199338e+00
1.39427602e-01 -8.22966546e-02 9.02989686e-01 -2.04713315e-01
-3.68343666e-02 -1.32226157e+00 2.81922787e-01 9.14553881e-01
-1.62478700e-01 1.20024078e-01 6.39065444e-01 6.71515763e-01
-1.02794074e-01 -5.21040142e-01 1.49986935e+00 -2.30757594e-01
-1.11382008e+00 -1.96804549e-03 -5.45352697e-01 -6.53355658e-01
1.38319600e+00 -5.29985845e-01 -3.14040869e-01 -4.30246681e-01
-8.10347557e-01 -4.78264093e-01 6.16855085e-01 2.70232588e-01
8.28354657e-01 -1.23303652e+00 -7.57170260e-01 2.74546742e-01
-5.41132987e-01 -5.51930130e-01 3.01279545e-01 1.65345073e-01
-1.49011207e+00 -4.03460562e-01 -7.68069327e-01 1.36504490e-02
-2.04580879e+00 9.72206593e-01 5.38294077e-01 3.51948678e-01
-6.99434102e-01 9.99979258e-01 -4.41735119e-01 3.55147302e-01
3.18401605e-01 9.92597416e-02 -4.48830515e-01 -6.02132380e-01
9.42723691e-01 4.91693228e-01 -1.86547358e-02 -7.51752794e-01
-2.15332359e-01 6.24381125e-01 -2.85366684e-01 -1.46412030e-01
1.25558722e+00 -1.96640402e-01 -2.95844793e-01 -3.15492868e-01
1.36759651e+00 3.98487419e-01 -1.55173266e+00 1.00517049e-01
-3.72892320e-01 -8.21168661e-01 5.37829459e-01 -9.22382057e-01
-1.47234404e+00 5.22779882e-01 1.33798492e+00 2.63218731e-01
1.90909529e+00 -9.34317231e-01 8.62327576e-01 2.80876070e-01
-3.79315346e-01 -1.10861671e+00 -1.58837855e-01 4.67261262e-02
8.74089718e-01 -1.05886173e+00 7.33387470e-02 -3.37372780e-01
-5.45025766e-01 1.35870993e+00 4.41859007e-01 -8.63390446e-01
6.19633257e-01 6.62919521e-01 6.70183361e-01 3.86733025e-01
-2.39808142e-01 1.54953357e-02 1.60776898e-01 1.13779938e+00
-1.01575673e-01 -3.79625589e-01 -4.53669876e-01 7.38897622e-02
-5.92937320e-02 -1.34856731e-01 7.23401546e-01 1.09951723e+00
-8.54706526e-01 -9.99278009e-01 -1.16946971e+00 8.18924382e-02
-1.08548152e+00 -1.64727747e-01 -5.04448473e-01 7.11559355e-01
2.86595017e-01 1.04263794e+00 -1.90982282e-01 -5.09972274e-01
2.21002355e-01 -2.21175347e-02 4.88541871e-01 8.06245506e-02
-7.83512235e-01 -2.93816328e-02 -2.32074484e-01 -3.66751909e-01
-3.06729555e-01 -2.83235580e-01 -1.15215707e+00 -5.84805071e-01
-1.35324597e-01 4.68538795e-03 9.68492746e-01 7.53912091e-01
1.04490325e-01 7.21619129e-01 7.82072842e-01 -7.84294069e-01
-2.53242254e-01 -8.31987739e-01 -3.82625580e-01 8.68378639e-01
6.36973560e-01 -1.26587391e-01 -6.48158908e-01 7.69564033e-01] | [5.51540470123291, 7.953231334686279] |
166720ad-5f63-4977-a5b1-93b08c6f55fb | a-two-step-approach-for-handling-zero | 2302.09887 | null | https://arxiv.org/abs/2302.09887v1 | https://arxiv.org/pdf/2302.09887v1.pdf | A Two-step Approach for Handling Zero-Cardinality in Relation Extraction | Relation tuple extraction from text is an important task for building knowledge bases. Recently, joint entity and relation extraction models have achieved very high F1 scores in this task. However, the experimental settings used by these models are restrictive and the datasets used in the experiments are not realistic. They do not include sentences with zero tuples (zero-cardinality). In this paper, we evaluate the state-of-the-art joint entity and relation extraction models in a more realistic setting. We include sentences that do not contain any tuples in our experiments. Our experiments show that there is significant drop ($\sim 10-15\%$ in one dataset and $\sim 6-14\%$ in another dataset) in their F1 score in this setting. We also propose a two-step modeling using a simple BERT-based classifier that leads to improvement in the overall performance of these models in this realistic experimental setup. | ['Indrajit Bhattacharya', 'Samiran Pal', 'Tapas Nayak', 'Pratik Saini'] | 2023-02-20 | null | null | null | null | ['joint-entity-and-relation-extraction'] | ['natural-language-processing'] | [-1.53942287e-01 4.80173469e-01 -3.60716134e-01 -5.47535896e-01
-8.83971810e-01 -2.75759697e-01 6.73512757e-01 6.47598147e-01
-5.28423369e-01 1.08989918e+00 -8.92323703e-02 -3.68683636e-01
-1.69249669e-01 -1.16099298e+00 -8.53704691e-01 -1.45018071e-01
-2.87922800e-01 7.82013357e-01 6.17970407e-01 -3.24420542e-01
-1.86638862e-01 2.47106671e-01 -1.39671624e+00 4.68993485e-01
7.65111625e-01 8.53042424e-01 -3.32732826e-01 6.80417359e-01
-2.56940216e-01 7.92684555e-01 -7.30671227e-01 -9.55756426e-01
1.55383885e-01 -1.39148876e-01 -1.03153312e+00 -2.59242535e-01
2.17736974e-01 -1.13185100e-01 -1.47184566e-01 8.58595669e-01
1.68353364e-01 4.19975184e-02 6.55526340e-01 -1.17821836e+00
-1.85601026e-01 1.03901947e+00 -4.48237509e-01 1.13659285e-01
4.66474831e-01 -4.08718616e-01 1.30002546e+00 -7.32893407e-01
5.76233745e-01 1.11202657e+00 6.00342274e-01 1.58096895e-01
-1.16088378e+00 -6.29037201e-01 1.01008557e-01 9.85748768e-02
-1.64687181e+00 -5.45321643e-01 2.64269024e-01 -1.82680115e-01
1.54413116e+00 5.22214770e-01 1.57362044e-01 5.92803299e-01
-3.75751182e-02 7.42114961e-01 8.23495209e-01 -7.69211173e-01
8.06689784e-02 5.67370176e-01 5.14720857e-01 4.31585938e-01
7.58232057e-01 -2.55797356e-01 -6.51810586e-01 -2.40647614e-01
4.60960060e-01 -2.76385844e-01 -4.64840867e-02 -1.11576051e-01
-8.41302037e-01 8.18668664e-01 1.82982415e-01 3.99766654e-01
-1.97787434e-01 -1.99034989e-01 1.23207889e-01 3.10417145e-01
5.89476764e-01 4.35785443e-01 -6.89041138e-01 -1.57166719e-01
-7.59004354e-01 5.34179270e-01 1.38080788e+00 1.36044407e+00
5.57754278e-01 -6.12824440e-01 -1.87715217e-01 8.67785573e-01
6.57999963e-02 9.46931541e-02 -7.76144043e-02 -6.94859028e-01
1.05057335e+00 7.10704088e-01 4.10287499e-01 -8.20303142e-01
-3.91418964e-01 -1.99392706e-01 -5.11410117e-01 -3.11544687e-01
4.73038286e-01 -1.79231495e-01 -8.81423116e-01 1.53230691e+00
3.85467619e-01 1.31237537e-01 3.82836699e-01 4.08144176e-01
9.95904803e-01 4.21563298e-01 2.99642891e-01 -4.15862411e-01
1.62008202e+00 -8.39393854e-01 -9.35984552e-01 -2.10876450e-01
9.14743245e-01 -8.63571644e-01 7.87284672e-01 1.03536665e-01
-9.93857741e-01 -3.12670261e-01 -1.25322437e+00 4.29819562e-02
-5.74318588e-01 1.58468962e-01 9.61562872e-01 7.97122777e-01
-5.63032150e-01 5.88198543e-01 -9.34412062e-01 -5.11964798e-01
2.92199433e-01 5.97863913e-01 -4.92442757e-01 -1.31199202e-02
-1.32526422e+00 1.05307734e+00 6.79405570e-01 -2.86548585e-01
-4.07521911e-02 -4.87817079e-01 -9.24933076e-01 1.39979601e-01
8.60317111e-01 -5.44477344e-01 1.28877378e+00 1.01381883e-01
-1.12234688e+00 5.86169839e-01 -5.00252843e-01 -6.67029858e-01
2.81409919e-01 -5.41285336e-01 -6.25514150e-01 -2.53644586e-01
8.69626105e-02 3.41965109e-01 3.85613460e-03 -1.16194570e+00
-9.68368530e-01 -1.84825122e-01 4.02920842e-01 1.62891950e-02
-9.42262411e-02 2.74184406e-01 -7.49738812e-01 -4.44106281e-01
-4.79766109e-04 -8.46819222e-01 -1.71723619e-01 -6.80040836e-01
-6.95396006e-01 -4.26352143e-01 5.02440393e-01 -4.04546112e-01
1.60000741e+00 -1.70151794e+00 -6.22493029e-01 1.99191228e-01
8.12004209e-02 5.13220906e-01 2.32962683e-01 6.77977979e-01
-5.63297682e-02 5.16787350e-01 -1.72984570e-01 -6.17211521e-01
-1.49536654e-01 3.45451176e-01 -2.00419739e-01 -1.69289544e-01
4.82452482e-01 7.84647942e-01 -5.77517033e-01 -6.66064560e-01
-3.56909558e-02 3.20596755e-01 -2.35335767e-01 2.35453025e-01
-2.05542043e-01 -2.27301419e-01 -5.15694678e-01 5.43214023e-01
6.23859763e-01 -2.30474129e-01 4.87591803e-01 -2.34361649e-01
-9.62761417e-03 6.89480007e-01 -1.43631935e+00 1.17324889e+00
-2.37861753e-01 3.19267631e-01 -2.35186219e-01 -9.37637687e-01
7.50411630e-01 4.40936267e-01 4.28806931e-01 -3.88898134e-01
8.91737789e-02 8.34511444e-02 1.90206975e-01 -4.51086223e-01
6.37482345e-01 -1.21850252e-01 -1.05028212e-01 2.96227068e-01
3.68144587e-02 -5.56244068e-02 7.04577684e-01 2.31848404e-01
1.33660591e+00 -2.27019355e-01 5.42038679e-01 -7.59727731e-02
3.50621551e-01 4.62716781e-02 8.67312193e-01 8.14971566e-01
1.93229780e-01 5.82604945e-01 7.29568064e-01 -1.71378806e-01
-9.01494503e-01 -6.86297238e-01 -3.31527084e-01 6.56449497e-01
9.59659517e-02 -8.65973949e-01 -4.93865669e-01 -8.45303118e-01
1.77574813e-01 8.51651013e-01 -4.98706192e-01 1.28302515e-01
-5.93715847e-01 -1.31231415e+00 6.64222777e-01 7.62530923e-01
4.85554606e-01 -7.93150723e-01 -2.09332332e-01 3.47696930e-01
-2.82674551e-01 -1.71460974e+00 1.38265193e-01 4.44156438e-01
-6.62394583e-01 -1.12554860e+00 -5.59433289e-02 -5.97242057e-01
4.11560267e-01 -5.43793403e-02 1.42975247e+00 -1.80854261e-01
-1.89057998e-02 -2.43803680e-01 -6.12070739e-01 -9.24002945e-01
-3.07724625e-01 3.69742841e-01 1.60106726e-03 -2.41333976e-01
9.31695282e-01 -2.58423805e-01 -2.51910537e-01 1.82951614e-01
-7.73144543e-01 -1.79828852e-01 6.81907177e-01 5.53228855e-01
5.03329873e-01 3.51416856e-01 5.32638431e-01 -1.64140725e+00
7.73595512e-01 -4.96510237e-01 -4.10268068e-01 4.97239321e-01
-9.24296975e-01 1.35524869e-01 3.74540269e-01 -6.72634542e-02
-8.93668115e-01 -9.09333825e-02 -1.55257761e-01 1.15954123e-01
-3.10206473e-01 7.98213780e-01 -3.29704404e-01 3.19213033e-01
5.46508312e-01 -4.78144109e-01 -3.55156839e-01 -6.77293181e-01
1.12184286e-01 8.55075777e-01 3.75726759e-01 -4.33314055e-01
6.33967698e-01 2.02484727e-01 4.91571985e-02 -5.41647434e-01
-1.05057645e+00 -6.81368470e-01 -5.88372409e-01 6.16574705e-01
3.90298903e-01 -1.08441591e+00 -7.76135683e-01 1.65094852e-01
-1.02750361e+00 -1.33596301e-01 -2.03961119e-01 6.57175004e-01
-1.22253098e-01 3.64283681e-01 -6.44975960e-01 -1.02334952e+00
-4.17593747e-01 -9.88869727e-01 8.26630652e-01 2.79749870e-01
-4.69838917e-01 -6.74165249e-01 -2.50487532e-02 3.81312698e-01
9.53109562e-02 2.81754851e-01 9.70045328e-01 -1.10241437e+00
-5.22774398e-01 -4.44792628e-01 -3.34080100e-01 8.84224027e-02
2.58854002e-01 5.47121093e-02 -9.01304483e-01 5.80971129e-02
-4.09452796e-01 -1.99268773e-01 9.28245604e-01 -2.01549269e-02
6.49048448e-01 -1.56275183e-01 -8.16894054e-01 -1.47982151e-03
1.26870823e+00 3.36996317e-01 6.88977301e-01 2.18630925e-01
4.90999132e-01 7.37379849e-01 9.64856565e-01 3.07101995e-01
6.33026004e-01 9.19957161e-01 -7.01545388e-04 4.26200638e-03
2.39978563e-02 -1.55768007e-01 1.03068508e-01 4.77422029e-01
-1.28640205e-01 -4.92236912e-01 -9.17207778e-01 8.03017378e-01
-2.11919570e+00 -7.51445949e-01 -4.41001356e-01 2.27913666e+00
1.22117186e+00 5.58402121e-01 1.69644311e-01 4.30619597e-01
4.84966040e-01 -1.25971228e-01 -9.18536931e-02 -4.69362438e-01
-1.15461044e-01 6.55713618e-01 5.95316350e-01 4.72221047e-01
-1.45461261e+00 1.02987921e+00 6.46282816e+00 6.01687729e-01
-7.31757045e-01 -1.62777990e-01 4.25800949e-01 4.85580489e-02
-3.52210999e-02 2.18620583e-01 -1.36455703e+00 3.26842308e-01
1.30276847e+00 -3.00861951e-02 -2.17190340e-01 6.19156301e-01
-1.54402286e-01 -5.62069952e-01 -1.37128258e+00 7.93728232e-01
-1.94815114e-01 -1.30817223e+00 -2.43893370e-01 7.42963478e-02
5.07925153e-01 -1.58248857e-01 -3.59546840e-01 5.13677478e-01
5.64531744e-01 -1.12250018e+00 3.41727912e-01 3.24084252e-01
7.06410408e-01 -9.62764025e-01 1.27997303e+00 4.45027500e-01
-1.12391806e+00 2.62968481e-01 -3.27085435e-01 -4.74730097e-02
2.75433838e-01 1.08002114e+00 -1.21878064e+00 9.08376038e-01
7.86966801e-01 2.62052208e-01 -5.65819919e-01 8.97005737e-01
-1.22639164e-01 6.84003711e-01 -6.92482471e-01 -1.07396044e-01
-2.42476642e-01 1.00555606e-01 2.51215547e-01 1.43127739e+00
2.78786663e-02 3.02370697e-01 9.73071828e-02 5.32479346e-01
-2.69541770e-01 3.14430267e-01 -6.30154669e-01 -4.72573675e-02
6.33724809e-01 1.00058496e+00 -6.00880921e-01 -3.60375941e-01
-6.56537473e-01 3.81702453e-01 5.24185359e-01 2.21369669e-01
-6.57037497e-01 -7.55283177e-01 6.73802793e-01 1.54123053e-01
6.17036641e-01 -4.30987366e-02 -4.55693245e-01 -1.18998289e+00
3.13843071e-01 -5.56531608e-01 4.33212847e-01 -1.06303006e-01
-1.25078833e+00 6.22284472e-01 4.04685229e-01 -8.25250208e-01
-5.36830962e-01 -5.06570458e-01 -2.90774941e-01 8.63369703e-01
-1.36866057e+00 -1.05528903e+00 -1.39583245e-01 2.25714862e-01
2.03874148e-02 -4.37305719e-02 1.07360232e+00 5.02250314e-01
-7.27627754e-01 1.03317022e+00 -1.43623233e-01 5.92897832e-01
8.23213398e-01 -1.36991799e+00 5.00862002e-01 8.18451941e-01
3.93054664e-01 9.22728121e-01 8.93033743e-01 -7.31567144e-01
-9.53447998e-01 -1.09514761e+00 1.47633505e+00 -7.06329525e-01
4.27323580e-01 -5.12388647e-01 -9.69659805e-01 8.57701838e-01
8.02184865e-02 -4.25968086e-03 9.61142421e-01 8.41902852e-01
-3.19340229e-01 -1.54705480e-01 -1.27712154e+00 4.33972746e-01
9.03371036e-01 -3.78121585e-01 -4.67639565e-01 2.36544251e-01
7.12276518e-01 -3.08880031e-01 -1.35001075e+00 1.03159952e+00
4.99802619e-01 -8.91539276e-01 7.59286523e-01 -7.42040336e-01
2.27186084e-01 -2.74616271e-01 -2.15443581e-01 -8.21594656e-01
6.70972392e-02 -3.61558199e-01 -5.88975251e-01 1.69958234e+00
1.08513844e+00 -6.28360927e-01 1.05081904e+00 1.13145828e+00
2.67917335e-01 -9.58816051e-01 -6.27979815e-01 -8.44799101e-01
-2.80723511e-03 -4.28100884e-01 7.36252844e-01 7.67556787e-01
2.20686674e-01 7.12076962e-01 -5.10995723e-02 1.45200804e-01
3.67618799e-01 2.63840884e-01 8.45122039e-01 -1.23840976e+00
-3.86456102e-01 4.05074190e-03 -3.19568396e-01 -8.39978933e-01
-7.56731490e-03 -5.80806673e-01 -1.77908540e-01 -1.53086519e+00
3.00414592e-01 -8.89949024e-01 -1.90266922e-01 6.31668508e-01
-4.32684630e-01 9.40644220e-02 7.92531148e-02 1.29660651e-01
-4.44797516e-01 1.86096504e-01 5.77811778e-01 1.60237029e-03
-2.73375958e-01 2.48772472e-01 -8.74114990e-01 5.16079545e-01
7.58133292e-01 -6.21664405e-01 -3.98227066e-01 -1.78410903e-01
1.61957011e-01 -6.05362393e-02 -1.94739386e-01 -7.80380189e-01
1.17275998e-01 -1.07263394e-01 1.77864686e-01 -5.72616041e-01
5.47206640e-01 -6.51771963e-01 1.96009487e-01 7.19321743e-02
-1.23952121e-01 -1.55590534e-01 1.44609988e-01 4.82165664e-01
-5.20146251e-01 -3.35697770e-01 3.25515777e-01 -1.06795989e-02
-3.88183594e-01 -2.44449154e-02 1.95289165e-01 3.41782868e-02
1.18807447e+00 -5.76742738e-02 -5.18361092e-01 -3.42314035e-01
-7.46029377e-01 2.18781784e-01 2.08462670e-01 3.29401642e-01
2.53193557e-01 -1.13683200e+00 -8.56204629e-01 -5.06198928e-02
3.44671428e-01 3.94635201e-01 -3.30081642e-01 7.99478173e-01
-4.21283841e-01 8.71272981e-01 2.95322180e-01 -1.89683899e-01
-1.46764648e+00 4.37891483e-01 -4.40227240e-02 -9.81260717e-01
-1.34972319e-01 9.73166704e-01 -1.79393977e-01 -3.92059863e-01
1.76983133e-01 -4.97730285e-01 -4.14411724e-01 1.41443789e-01
3.71622056e-01 2.62022197e-01 5.68389475e-01 -4.39526111e-01
-5.22189021e-01 8.11544433e-03 -5.66405118e-01 -2.15327680e-01
1.27914810e+00 1.08239740e-01 -3.99204157e-02 4.87844557e-01
1.05763721e+00 2.50139713e-01 -3.92019331e-01 -3.66231322e-01
4.64548588e-01 -2.62494355e-01 -3.96594971e-01 -7.45825171e-01
-7.67535329e-01 4.35775250e-01 1.01274632e-01 3.59795451e-01
7.90561378e-01 2.50584871e-01 8.58376801e-01 4.77674991e-01
6.74328804e-01 -9.38089669e-01 -5.92904627e-01 6.49323642e-01
5.34042001e-01 -1.45874715e+00 4.11105752e-01 -1.23125494e+00
-5.63385069e-01 6.67011619e-01 7.04790533e-01 1.08653009e-01
9.52579379e-01 6.18138552e-01 -9.15097818e-03 -1.14576072e-01
-1.07121181e+00 -5.69985688e-01 1.87341183e-01 3.79515052e-01
7.87706137e-01 2.58799672e-01 -6.76354468e-01 9.06405270e-01
-4.17722106e-01 -9.68837962e-02 4.24880803e-01 1.18135214e+00
-1.21412791e-01 -1.71087158e+00 -1.21479698e-01 7.74224222e-01
-8.66820633e-01 -1.05081014e-01 -5.35140812e-01 9.86168325e-01
7.40780085e-02 1.38123345e+00 -8.67470577e-02 -5.95665634e-01
6.38120532e-01 2.68007755e-01 4.30571228e-01 -8.71484220e-01
-6.56259656e-01 -1.06451467e-01 8.28296721e-01 -3.44260812e-01
-4.61423397e-01 -7.70969689e-01 -1.48828876e+00 -3.61944705e-01
-7.13482857e-01 4.61385429e-01 4.24118221e-01 1.05679154e+00
4.93837982e-01 3.83817613e-01 1.79706678e-01 1.12133525e-01
-2.95693427e-01 -1.34644842e+00 -5.84361851e-01 3.28954220e-01
-3.58572304e-02 -8.10532093e-01 1.03709191e-01 8.36123601e-02] | [9.331939697265625, 8.680254936218262] |
e5f6d893-5743-484d-b231-50a112cc24b4 | blindspotnet-seeing-where-we-cannot-see | 2207.03870 | null | https://arxiv.org/abs/2207.03870v1 | https://arxiv.org/pdf/2207.03870v1.pdf | BlindSpotNet: Seeing Where We Cannot See | We introduce 2D blind spot estimation as a critical visual task for road scene understanding. By automatically detecting road regions that are occluded from the vehicle's vantage point, we can proactively alert a manual driver or a self-driving system to potential causes of accidents (e.g., draw attention to a road region from which a child may spring out). Detecting blind spots in full 3D would be challenging, as 3D reasoning on the fly even if the car is equipped with LiDAR would be prohibitively expensive and error prone. We instead propose to learn to estimate blind spots in 2D, just from a monocular camera. We achieve this in two steps. We first introduce an automatic method for generating ``ground-truth'' blind spot training data for arbitrary driving videos by leveraging monocular depth estimation, semantic segmentation, and SLAM. The key idea is to reason in 3D but from 2D images by defining blind spots as those road regions that are currently invisible but become visible in the near future. We construct a large-scale dataset with this automatic offline blind spot estimation, which we refer to as Road Blind Spot (RBS) dataset. Next, we introduce BlindSpotNet (BSN), a simple network that fully leverages this dataset for fully automatic estimation of frame-wise blind spot probability maps for arbitrary driving videos. Extensive experimental results demonstrate the validity of our RBS Dataset and the effectiveness of our BSN. | ['Ko Nishino', 'Shohei Nobuhara', 'Shinya Ishizaki', 'Kotaro Hasegawa', 'Taichi Fukuda'] | 2022-07-08 | null | null | null | null | ['road-scene-understanding'] | ['computer-vision'] | [ 1.05284497e-01 2.76206315e-01 -1.66881382e-01 -4.71395850e-01
-6.64493024e-01 -7.04668522e-01 4.53137994e-01 -5.43881416e-01
-3.63991261e-01 4.52845275e-01 1.36863470e-01 -7.56594658e-01
1.96146935e-01 -7.29413986e-01 -1.00198889e+00 -2.41933361e-01
4.35235471e-01 4.75928605e-01 6.89950049e-01 1.64831564e-01
2.08744451e-01 6.45477057e-01 -1.96751773e+00 -3.78999501e-01
9.44337785e-01 6.73423588e-01 3.40886146e-01 7.75377095e-01
-2.79852282e-02 7.67517030e-01 -3.24764967e-01 -3.73382926e-01
4.80367392e-01 -1.24812163e-01 -5.10000408e-01 2.59580135e-01
8.45861018e-01 -8.72897923e-01 -6.43787801e-01 1.00197101e+00
3.49222720e-01 -1.61224037e-01 4.56231236e-01 -1.52037466e+00
-1.92635864e-01 -7.02943206e-02 -6.65216267e-01 1.02140546e-01
5.99797249e-01 6.23860538e-01 4.66430664e-01 -8.58678043e-01
3.39547455e-01 1.15782428e+00 5.14684021e-01 6.06462002e-01
-7.43454874e-01 -7.79430807e-01 4.63005692e-01 2.32843831e-01
-1.45145214e+00 -7.56249964e-01 6.53325677e-01 -5.80371678e-01
6.61732435e-01 8.06594118e-02 6.62876010e-01 6.93874180e-01
-2.84678876e-01 9.11554158e-01 1.12833703e+00 -1.57612711e-01
2.74250537e-01 -1.06555723e-01 -1.79133460e-01 8.28473032e-01
3.11742693e-01 5.00347674e-01 -6.75682425e-01 3.90302598e-01
7.11038947e-01 -2.96883495e-03 -2.40093261e-01 -6.81045949e-01
-1.05334139e+00 7.53261685e-01 7.63380408e-01 -3.42664093e-01
-2.71879554e-01 2.20934331e-01 -3.71579528e-01 8.27351287e-02
3.25802326e-01 -7.85892457e-02 -2.75346696e-01 1.02129564e-01
-9.25631642e-01 1.88582569e-01 3.51711333e-01 9.43039954e-01
1.34051383e+00 -2.37383559e-01 6.34558685e-03 4.24495339e-01
3.61109763e-01 6.95156515e-01 -2.85682052e-01 -1.15562403e+00
3.40377748e-01 6.89015329e-01 3.69670123e-01 -5.30302644e-01
-2.59370297e-01 -1.60140246e-01 -3.04557025e-01 6.42916024e-01
4.20595884e-01 -2.23615363e-01 -1.55237186e+00 1.54506612e+00
5.41514814e-01 4.75671619e-01 -1.83195949e-01 1.26099789e+00
7.92733967e-01 2.03019723e-01 -1.68843091e-01 1.78125873e-01
1.04767668e+00 -7.60805845e-01 -2.91974723e-01 -9.99749243e-01
4.10833389e-01 -5.06805182e-01 7.44701028e-01 2.33479291e-01
-8.73149157e-01 -1.78943887e-01 -7.67710209e-01 -1.21495053e-01
-3.76078278e-01 2.27738813e-01 6.83088183e-01 7.22813606e-01
-1.39218688e+00 -2.42999673e-01 -6.44927442e-01 -5.32252371e-01
7.99448967e-01 1.90037027e-01 -3.62012714e-01 -6.15716994e-01
-9.30891633e-01 9.09668148e-01 -1.49449157e-02 6.49621263e-02
-1.41684103e+00 -7.29788840e-01 -1.16962492e+00 -3.18905383e-01
4.79635388e-01 -7.03584909e-01 1.34494662e+00 -5.39699733e-01
-9.33223009e-01 1.23386109e+00 -9.22954261e-01 -4.97670293e-01
5.96168339e-01 -2.48065740e-01 -2.88858980e-01 3.18862587e-01
5.80427170e-01 1.04759896e+00 1.11993146e+00 -1.53210878e+00
-9.59501743e-01 -4.72456634e-01 2.90551573e-01 2.05336243e-01
3.17062587e-01 1.25110328e-01 -8.69451344e-01 -1.10400997e-01
4.24056023e-01 -8.11342061e-01 -1.65592834e-01 3.84252131e-01
-7.30108857e-01 2.78040394e-02 7.71093011e-01 -7.27359653e-01
7.38316476e-01 -1.97827435e+00 -4.13983703e-01 1.27672821e-01
5.89388669e-01 3.12669963e-01 1.71982348e-01 -8.52791891e-02
2.29113907e-01 -2.58235782e-01 -5.22132397e-01 -3.44603837e-01
-1.01039425e-01 2.30016708e-01 -4.78430212e-01 6.89428270e-01
2.90489227e-01 9.58431244e-01 -1.00617075e+00 -5.76140046e-01
7.59571671e-01 4.60110962e-01 -4.75488722e-01 3.61048073e-01
-1.63123265e-01 6.16730750e-01 -2.54285336e-01 9.22653198e-01
1.08876431e+00 1.07255369e-01 -5.00869155e-01 2.70774364e-01
-4.95786607e-01 4.23303157e-01 -8.92777801e-01 1.60542643e+00
-5.62608123e-01 8.20587039e-01 2.60476023e-01 -7.00902760e-01
6.10737443e-01 -8.19461569e-02 -9.04792454e-03 -8.22964370e-01
-4.11708280e-02 6.62131875e-04 -6.14344001e-01 -3.80363792e-01
2.14939043e-01 2.02404335e-01 6.30234405e-02 4.61900502e-01
-2.38801137e-01 -1.65169179e-01 -1.55275494e-01 4.21377182e-01
1.30210090e+00 -6.60597458e-02 -2.26911858e-01 3.50966215e-01
4.29241478e-01 1.45936996e-01 5.38616061e-01 8.52926552e-01
-4.69457418e-01 8.01913261e-01 3.18660080e-01 -3.72458935e-01
-5.87933123e-01 -1.57535839e+00 -2.50459742e-02 6.70838356e-01
7.25664020e-01 1.52798384e-01 -7.09418833e-01 -9.31891978e-01
3.16020727e-01 7.26987183e-01 -4.34437156e-01 -1.05142025e-02
-3.08719635e-01 -2.03136802e-01 1.86926708e-01 4.69541460e-01
6.43283308e-01 -6.31240308e-01 -7.63132453e-01 -3.34810823e-01
-1.86203629e-01 -1.31057119e+00 -3.40692312e-01 -2.14162603e-01
-1.76316485e-01 -1.23632026e+00 -7.36664414e-01 -7.24166274e-01
9.99649227e-01 1.09407616e+00 1.09674859e+00 1.48559332e-01
-5.24527431e-01 4.48193431e-01 -6.47545094e-03 -5.27964711e-01
-2.03388542e-01 -4.32001501e-01 7.77888745e-02 1.02544978e-01
6.09954774e-01 -3.73950660e-01 -1.07114065e+00 5.57403386e-01
-2.96658903e-01 1.49946466e-01 7.51217127e-01 1.54447481e-01
4.79082972e-01 6.49299771e-02 3.23589861e-01 -3.75242442e-01
-1.20888889e-01 -4.66081470e-01 -1.09537840e+00 -1.73637215e-02
-2.32325584e-01 -1.73449412e-01 -1.33774832e-01 1.01993456e-01
-8.75827253e-01 4.51206416e-01 -1.59489568e-02 -7.07075715e-01
-5.55681646e-01 -3.63834292e-01 -3.86489093e-01 -3.39911938e-01
4.73960608e-01 2.61379123e-01 -9.00476500e-02 -2.54477799e-01
5.88911653e-01 8.13811123e-01 9.21032488e-01 -5.85122742e-02
1.23417091e+00 9.51692522e-01 -1.32023856e-01 -6.01246715e-01
-1.01880491e+00 -8.13051403e-01 -5.51661134e-01 -6.53240502e-01
1.05532956e+00 -1.36192203e+00 -8.89726400e-01 5.77045560e-01
-1.16451430e+00 -2.41251871e-01 9.85644013e-02 3.36916953e-01
-4.25113201e-01 1.47826061e-01 1.58700362e-01 -1.09024048e+00
1.45204559e-01 -1.08312106e+00 1.44311035e+00 2.18494728e-01
1.02276467e-01 -5.23913860e-01 -2.51767844e-01 7.03737497e-01
1.08551100e-01 1.42912511e-02 4.12190527e-01 4.57456298e-02
-1.19378364e+00 -1.81523189e-01 -7.58749902e-01 1.99968159e-01
-2.35122982e-02 -4.65016067e-01 -1.49666667e+00 -4.70613167e-02
-5.44206560e-01 -1.11558244e-01 1.11692214e+00 4.35955673e-01
9.39482391e-01 -9.63360593e-02 -6.83428764e-01 6.39993012e-01
1.22168398e+00 -1.69190541e-01 3.44323695e-01 9.88238603e-02
9.44355726e-01 9.10224199e-01 6.52856886e-01 -1.21334884e-02
9.69906151e-01 5.30320883e-01 8.96428287e-01 -2.48636395e-01
-5.87218761e-01 -8.58453572e-01 2.54866123e-01 5.94244758e-03
3.83798003e-01 1.10131474e-02 -1.19282925e+00 1.20350230e+00
-1.57446182e+00 -7.01189578e-01 -1.51368290e-01 2.46041703e+00
3.27411145e-01 1.30628020e-01 2.29113862e-01 9.11358446e-02
9.58961666e-01 -1.94795102e-01 -9.13332164e-01 1.35319918e-01
-1.41548112e-01 -2.30416730e-01 1.02358902e+00 8.97578239e-01
-1.07279849e+00 1.31841087e+00 5.46311092e+00 1.48399040e-01
-8.35909426e-01 2.27224559e-01 4.15696174e-01 -1.22242235e-01
-4.37750727e-01 4.62739646e-01 -7.69418180e-01 4.69102710e-01
4.79906559e-01 2.95281589e-01 5.66530943e-01 5.66626549e-01
2.62068123e-01 -6.16584182e-01 -8.81900370e-01 1.12473619e+00
1.46125451e-01 -1.17548990e+00 -2.68770576e-01 1.12466156e-01
6.77442491e-01 5.51233649e-01 6.39521182e-02 -2.32946090e-02
8.02659988e-01 -9.83713746e-01 8.38632226e-01 3.90442878e-01
1.16065300e+00 -6.15665555e-01 1.50234923e-01 5.22406220e-01
-1.07235122e+00 -1.63164467e-01 -2.06133947e-01 -7.36564398e-02
3.84142131e-01 6.40851796e-01 -1.24816155e+00 8.66379887e-02
7.77676880e-01 7.23119140e-01 -7.05055475e-01 1.42016637e+00
-7.70761669e-01 3.26485246e-01 -3.57790202e-01 4.38470066e-01
1.05791897e-01 2.30238307e-02 7.55536199e-01 5.78454852e-01
3.26764137e-01 -1.64587572e-01 -1.63196877e-01 1.12690878e+00
2.50786319e-02 -6.63866401e-01 -1.00982165e+00 3.61949384e-01
6.76685929e-01 9.18090999e-01 -4.99640018e-01 2.28280704e-02
-6.29115224e-01 9.52435493e-01 5.44863455e-02 6.51965082e-01
-7.27662444e-01 -3.52172375e-01 1.12030256e+00 3.65800679e-01
3.45545441e-01 -1.00987874e-01 -3.79815161e-01 -9.75990713e-01
1.92094877e-01 -1.93955094e-01 -2.58374754e-02 -1.26179028e+00
-7.69210517e-01 3.00323784e-01 -3.07571411e-01 -1.29904974e+00
-4.44700755e-02 -4.89991635e-01 -6.79779351e-01 9.92438555e-01
-2.30453086e+00 -1.17846036e+00 -7.71123052e-01 8.78393531e-01
4.63692546e-01 -3.06653194e-02 1.98526487e-01 1.05052985e-01
-4.22730863e-01 5.11862457e-01 -4.98813838e-01 1.19853109e-01
5.47269583e-01 -9.90762353e-01 9.84112382e-01 1.51561296e+00
1.03946859e-02 1.47696212e-01 5.49505711e-01 -7.80459464e-01
-1.41586018e+00 -1.37746441e+00 1.13232362e+00 -9.96204674e-01
5.44003069e-01 -5.42036295e-01 -5.79435647e-01 6.62005663e-01
-4.03488368e-01 1.09462347e-02 -1.16035692e-01 -2.90349960e-01
-3.30792397e-01 -1.73005491e-01 -1.19097388e+00 6.93461359e-01
1.42616546e+00 -8.43768597e-01 -4.98941779e-01 4.65053767e-01
8.54948401e-01 -3.30740750e-01 1.05512492e-01 5.02337873e-01
1.62378833e-01 -1.25572705e+00 1.21104574e+00 -6.45265579e-02
-1.24468960e-01 -6.06486320e-01 -2.17501782e-02 -1.07060981e+00
4.01360601e-01 -8.42063069e-01 -2.27666683e-02 9.52818513e-01
2.93020099e-01 -7.17025101e-01 9.30093825e-01 5.98731399e-01
-2.54078597e-01 -1.78339198e-01 -1.07371688e+00 -5.74982107e-01
-2.63375998e-01 -1.10149097e+00 9.34285879e-01 3.91276687e-01
-4.83202130e-01 -5.57379797e-02 -2.57352918e-01 9.67562795e-01
1.05401838e+00 -9.45411250e-02 1.06976616e+00 -1.41160786e+00
3.48323524e-01 -1.70130834e-01 -6.87644362e-01 -1.35092664e+00
3.54469955e-01 -6.61302447e-01 5.20449281e-01 -2.00443077e+00
-1.01031318e-01 -5.39585948e-01 2.39145495e-02 4.26185101e-01
-2.26765558e-01 3.30122769e-01 -2.42029145e-01 2.54673302e-01
-6.19634986e-01 1.37671560e-01 1.11901522e+00 -1.82842836e-02
-1.02783389e-01 3.28009218e-01 -6.30122185e-01 9.08857167e-01
5.32199144e-01 -1.94492564e-01 -5.00705421e-01 -8.88767421e-01
4.03994262e-01 2.29382947e-01 9.83714521e-01 -1.09430790e+00
6.15380585e-01 -2.78384946e-02 1.72031820e-01 -8.27751815e-01
4.89870906e-01 -7.73800492e-01 -4.23284948e-01 9.72376540e-02
2.30527729e-01 -4.10592377e-01 1.36224374e-01 7.77326941e-01
1.77151151e-02 2.00708196e-01 6.33333504e-01 7.70319849e-02
-1.23512244e+00 6.71164751e-01 -2.52142221e-01 8.78952891e-02
1.22721076e+00 -4.49181855e-01 -5.44135988e-01 -5.74923635e-01
-3.12717289e-01 6.28700435e-01 8.83650482e-01 7.09341109e-01
1.05725431e+00 -9.23561931e-01 -6.28192604e-01 8.58518541e-01
5.99211037e-01 4.05712754e-01 3.76298547e-01 7.07681715e-01
-5.65491378e-01 6.38965249e-01 1.87588930e-01 -8.04885089e-01
-9.35776591e-01 5.62511623e-01 3.54491860e-01 8.09975624e-01
-8.86257052e-01 1.09018910e+00 7.05712318e-01 -4.34882760e-01
4.40894991e-01 -1.67731702e-01 -8.75154957e-02 -4.11091506e-01
5.97606957e-01 2.06506655e-01 3.08529548e-02 -1.02856445e+00
-6.26757443e-01 8.10728908e-01 1.04035467e-01 -2.01000735e-01
7.95338333e-01 -6.09352589e-01 -1.94169842e-02 -1.92179214e-02
7.88484216e-01 -4.33502952e-03 -1.72257376e+00 -3.34936470e-01
-4.11314011e-01 -8.92047882e-01 4.37266797e-01 -8.76879334e-01
-1.02889585e+00 9.34132576e-01 5.68245113e-01 -2.89401174e-01
1.06684995e+00 6.57742083e-01 9.07635868e-01 3.95519892e-03
7.03315735e-01 -8.21366549e-01 -3.57705593e-01 1.95832893e-01
4.59404409e-01 -1.72626126e+00 -4.28579599e-01 -5.25108516e-01
-6.37187958e-01 4.60921675e-01 7.43806362e-01 2.48749912e-01
5.16003430e-01 6.01300560e-02 3.97811472e-01 -3.07627708e-01
-3.78486276e-01 -7.76963472e-01 2.24518776e-01 1.12348747e+00
-6.35056496e-01 2.13455021e-01 7.22087920e-01 -1.30328012e-03
-3.58753920e-01 -1.02490187e-01 6.06615126e-01 7.56787360e-01
-7.43842483e-01 -5.23065388e-01 -5.35145998e-01 3.77642155e-01
1.78330243e-01 1.75425112e-02 -6.94588363e-01 4.39601064e-01
3.28687161e-01 1.44873774e+00 3.19717050e-01 -2.54559338e-01
1.42634347e-01 -8.24416727e-02 3.60787779e-01 -5.73971629e-01
5.35154581e-01 -5.55797398e-01 -1.00323938e-01 -8.49216580e-01
-1.62232086e-01 -6.23075426e-01 -1.06166112e+00 -1.79568499e-01
6.02420866e-02 -2.32719630e-01 9.68213558e-01 1.05800533e+00
3.15294147e-01 9.38578621e-02 8.26824307e-01 -8.04536223e-01
2.62280554e-01 -4.04340416e-01 -3.30685794e-01 -2.27087557e-01
1.12401187e+00 -8.83778572e-01 -6.26545429e-01 -2.68383443e-01] | [8.092247009277344, -2.334517002105713] |
08ccaf1f-8c89-48d0-a2b7-e16c7060f30b | robustl2s-speaker-specific-lip-to-speech | 2307.01233 | null | https://arxiv.org/abs/2307.01233v1 | https://arxiv.org/pdf/2307.01233v1.pdf | RobustL2S: Speaker-Specific Lip-to-Speech Synthesis exploiting Self-Supervised Representations | Significant progress has been made in speaker dependent Lip-to-Speech synthesis, which aims to generate speech from silent videos of talking faces. Current state-of-the-art approaches primarily employ non-autoregressive sequence-to-sequence architectures to directly predict mel-spectrograms or audio waveforms from lip representations. We hypothesize that the direct mel-prediction hampers training/model efficiency due to the entanglement of speech content with ambient information and speaker characteristics. To this end, we propose RobustL2S, a modularized framework for Lip-to-Speech synthesis. First, a non-autoregressive sequence-to-sequence model maps self-supervised visual features to a representation of disentangled speech content. A vocoder then converts the speech features into raw waveforms. Extensive evaluations confirm the effectiveness of our setup, achieving state-of-the-art performance on the unconstrained Lip2Wav dataset and the constrained GRID and TCD-TIMIT datasets. Speech samples from RobustL2S can be found at https://neha-sherin.github.io/RobustL2S/ | ['Vineet Gandhi', 'Vishal Tambrahalli', 'Neil Shah', 'Neha Sahipjohn'] | 2023-07-03 | null | null | null | null | ['speaker-specific-lip-to-speech-synthesis', 'lip-to-speech-synthesis', 'speech-synthesis'] | ['computer-vision', 'computer-vision', 'speech'] | [ 2.07120880e-01 1.16948970e-01 -3.66045535e-01 -3.66264999e-01
-1.43372941e+00 -4.05675501e-01 7.60106444e-01 -8.29050958e-01
3.61213446e-01 5.73421836e-01 7.95112967e-01 -1.56920433e-01
5.15733778e-01 -8.64120051e-02 -6.45388484e-01 -7.84053683e-01
2.11258948e-01 1.32609993e-01 -2.24838629e-01 1.15262233e-01
-5.83503135e-02 3.32779467e-01 -1.89290500e+00 7.37358809e-01
3.11399400e-01 1.11749768e+00 1.86891273e-01 1.13701046e+00
-6.44212887e-02 7.76136220e-01 -2.77063668e-01 -3.82663310e-01
1.69797122e-01 -7.22626030e-01 -3.56225193e-01 -7.06885979e-02
5.17826796e-01 -3.65769625e-01 -5.54268479e-01 8.22035909e-01
9.29762959e-01 -1.20123662e-01 9.55331743e-01 -1.53755963e+00
-6.12255752e-01 3.52814794e-01 -2.26463273e-01 -3.48195940e-01
3.98255318e-01 6.31741285e-01 8.66295695e-01 -1.47805774e+00
5.00466228e-01 1.53653514e+00 4.95043427e-01 1.00394654e+00
-1.48273909e+00 -9.28068638e-01 -3.54138464e-01 2.98214853e-01
-1.46076453e+00 -1.85917914e+00 1.01683164e+00 -4.70829606e-01
7.00171769e-01 2.39468560e-01 3.18242580e-01 1.60396183e+00
-2.71520495e-01 1.01413369e+00 1.12479997e+00 -3.71980906e-01
3.76181267e-02 2.56780505e-01 -6.38393581e-01 5.60125649e-01
-4.71210629e-01 3.63224655e-01 -1.27941918e+00 6.39467984e-02
3.85702252e-01 -7.63285458e-01 -3.43204200e-01 -3.88927847e-01
-8.69373500e-01 7.52498090e-01 -9.81474891e-02 -8.39107037e-02
-2.83818632e-01 1.73596218e-01 4.38860685e-01 -3.32278050e-02
5.99953473e-01 -5.04194126e-02 -2.46657208e-01 -2.22406179e-01
-1.25290799e+00 1.63492009e-01 5.29682159e-01 9.32492495e-01
4.10618007e-01 6.44482553e-01 -3.15044582e-01 9.76416945e-01
6.56543374e-01 9.27349389e-01 4.95456070e-01 -1.08305049e+00
4.25921082e-01 -2.74711132e-01 -2.13015661e-01 -4.29147720e-01
-3.34323198e-02 1.87213674e-01 -6.24238074e-01 4.33747321e-01
3.11481118e-01 -1.61545977e-01 -9.26815450e-01 1.85012722e+00
2.26177663e-01 4.78058755e-01 2.93612182e-01 8.64444017e-01
1.18395019e+00 7.66468942e-01 7.58764446e-02 -3.80993307e-01
1.18124986e+00 -8.46639991e-01 -9.33670104e-01 -1.02536790e-01
2.14731395e-02 -1.00602734e+00 1.08428919e+00 1.01704977e-01
-1.44559479e+00 -7.24637151e-01 -7.45865643e-01 -2.07479656e-01
-1.52012214e-01 5.22790551e-01 -1.74280594e-03 7.25410223e-01
-1.23727787e+00 1.87415704e-01 -5.75922668e-01 -3.69851701e-02
5.54515362e-01 4.66553569e-02 -4.90334779e-01 1.40174821e-01
-8.93986940e-01 5.24856269e-01 -1.23477258e-01 -7.97874108e-02
-1.16113436e+00 -9.11391199e-01 -1.08279765e+00 -4.64358255e-02
8.62825066e-02 -4.06278700e-01 1.40167820e+00 -8.15891266e-01
-2.19820881e+00 9.74742234e-01 -8.39850485e-01 -4.27471489e-01
4.76089686e-01 1.26416013e-01 -4.76863980e-01 4.09370512e-01
-1.37483671e-01 1.03005016e+00 1.46052039e+00 -1.47147417e+00
-1.84301153e-01 -4.29384708e-02 -6.63489580e-01 2.22320795e-01
-2.55904458e-02 2.31512725e-01 -2.91266680e-01 -7.49900579e-01
-2.92125434e-01 -9.26902652e-01 4.84110445e-01 3.18526894e-01
-5.32705069e-01 -2.23637864e-01 7.94153690e-01 -8.91723216e-01
6.39260650e-01 -2.28735423e+00 4.64401841e-02 -4.02492434e-01
-1.09391259e-02 4.61087048e-01 -6.38828993e-01 4.16425616e-01
-2.98210442e-01 1.41108230e-01 9.97120515e-02 -1.10575259e+00
1.92497209e-01 -3.18143994e-01 -6.82375073e-01 5.60702085e-01
4.88837421e-01 1.03137100e+00 -5.07921755e-01 -7.33562469e-01
4.39478874e-01 9.53480959e-01 -5.07458508e-01 6.03463769e-01
-3.07741284e-01 5.91603100e-01 1.68564141e-01 7.79567719e-01
6.28310800e-01 2.01597050e-01 3.67437974e-02 -3.73366654e-01
-8.93571675e-02 6.84882581e-01 -5.93807995e-01 1.74020720e+00
-7.08721101e-01 1.18459666e+00 5.01773417e-01 -7.06414640e-01
8.70456278e-01 7.88689852e-01 4.36546504e-01 -5.18466532e-01
7.04537556e-02 1.30618989e-01 -3.66428494e-01 -7.16619551e-01
1.63384512e-01 -2.62650967e-01 3.60303819e-01 1.15501598e-01
4.31476653e-01 -5.07467985e-01 -2.61413306e-01 -9.25804302e-02
5.49131215e-01 3.21074307e-01 1.37538612e-01 3.47893238e-02
6.33365154e-01 -7.19335139e-01 2.70555943e-01 1.77013159e-01
-2.68704653e-01 9.03052986e-01 3.72263253e-01 3.94091994e-01
-9.30844843e-01 -1.54092085e+00 -1.36342689e-01 1.26265991e+00
-3.59886467e-01 -5.20591855e-01 -8.89077544e-01 -3.76579940e-01
-2.01812983e-01 9.42684650e-01 -4.19317424e-01 -6.50157854e-02
-2.97010452e-01 9.80441347e-02 9.33667183e-01 2.62487590e-01
6.89898431e-02 -1.34112430e+00 -4.26960848e-02 -8.24261382e-02
-3.82041216e-01 -1.42680550e+00 -7.29300022e-01 -2.53211290e-01
-2.10349634e-01 -6.24031603e-01 -9.43932712e-01 -6.98980570e-01
2.51302928e-01 1.25149548e-01 7.73074269e-01 -5.03238916e-01
-3.61765772e-01 4.16867882e-01 -4.12058420e-02 -5.00597000e-01
-9.11879003e-01 -2.55524457e-01 4.14530575e-01 4.63525176e-01
6.04515448e-02 -6.22765362e-01 -5.87736845e-01 3.09211761e-01
-3.69359016e-01 1.98875085e-01 4.48752314e-01 7.05958247e-01
4.78838235e-01 -5.48508525e-01 9.19309855e-01 -2.99886744e-02
5.13362825e-01 -3.30125540e-01 -5.06077826e-01 -7.88819268e-02
-2.29524702e-01 2.62657199e-02 5.89153230e-01 -6.50554478e-01
-1.18712723e+00 1.77115694e-01 -3.68767768e-01 -1.09851336e+00
-4.05086190e-01 -3.31470296e-02 -5.36472023e-01 1.86608821e-01
5.48652232e-01 5.34152448e-01 4.10362929e-01 -4.78770912e-01
6.45506740e-01 1.22620440e+00 8.46658945e-01 -3.73550415e-01
6.31853819e-01 3.72786105e-01 5.21146134e-03 -1.31659269e+00
-3.63330811e-01 -2.95133501e-01 -3.07122946e-01 -3.32064062e-01
8.99957299e-01 -1.15704846e+00 -9.91395652e-01 7.64302731e-01
-1.21675563e+00 -7.60386884e-01 -3.29429865e-01 4.76844281e-01
-1.12901247e+00 9.87233892e-02 -4.06828403e-01 -1.27505720e+00
-3.21766794e-01 -1.25288570e+00 1.35427976e+00 7.82083794e-02
-2.92863160e-01 -5.91892421e-01 1.09752402e-01 7.26072609e-01
4.87496942e-01 -1.71568945e-01 5.12614191e-01 -5.21713555e-01
-4.53867853e-01 1.03377379e-01 -1.08352184e-01 5.73417842e-01
1.84354588e-01 8.42071399e-02 -1.69585562e+00 -1.18969768e-01
-2.42447495e-01 -7.10016429e-01 7.09008396e-01 5.48765063e-01
1.08593488e+00 -6.72832012e-01 -1.03390113e-01 6.87321782e-01
8.27044189e-01 -1.33011073e-01 6.38368607e-01 -6.86560631e-01
5.39555371e-01 8.58880997e-01 2.37110987e-01 4.17396575e-01
3.38166684e-01 1.01732183e+00 1.72776401e-01 1.47638109e-03
-1.16612768e+00 -8.75544727e-01 7.83314764e-01 8.00579071e-01
3.90343845e-01 -3.20113868e-01 -7.28723228e-01 7.34317303e-01
-1.39845490e+00 -1.17795086e+00 1.97121888e-01 1.98037648e+00
1.06435955e+00 -3.16314638e-01 2.62778223e-01 8.05126131e-02
7.72590101e-01 4.37331080e-01 -6.39924526e-01 -1.47653103e-01
-8.67739841e-02 2.49943748e-01 1.05359867e-01 8.11704218e-01
-8.78576994e-01 1.14126408e+00 5.74760675e+00 1.19298613e+00
-1.56239343e+00 2.15833068e-01 5.53448677e-01 -5.20979285e-01
-2.27508813e-01 -4.41787928e-01 -7.93082297e-01 5.72609007e-01
1.30411589e+00 -2.63707846e-01 7.83853650e-01 7.10295856e-01
6.49383545e-01 2.41051927e-01 -1.17469323e+00 1.38125086e+00
3.91903222e-01 -1.41627038e+00 -1.97929546e-01 3.72287668e-02
4.50238854e-01 2.62748778e-01 6.02467477e-01 1.17438801e-01
4.20156792e-02 -1.28183448e+00 1.11102366e+00 5.67901909e-01
1.71229959e+00 -4.08088148e-01 -1.08248882e-01 2.03199357e-01
-1.38594294e+00 1.40961289e-01 7.12338313e-02 5.13828516e-01
3.21109116e-01 -8.08404293e-04 -1.42553735e+00 4.51770797e-02
3.92511517e-01 5.64597785e-01 -7.33738020e-02 6.57212436e-01
-3.09330642e-01 8.86489093e-01 -1.48274451e-01 1.36422843e-01
-3.38230669e-01 2.55856961e-01 8.47960293e-01 1.38137889e+00
3.41361552e-01 -1.21200196e-01 -3.00722986e-01 1.04036796e+00
-2.44827867e-01 1.05925292e-01 -8.55337977e-01 -2.78317660e-01
5.65944135e-01 1.16636503e+00 -5.42905852e-02 -1.26001686e-01
-1.70851886e-01 6.59565270e-01 -1.00057036e-01 5.98424613e-01
-7.01886415e-01 -1.99906364e-01 1.11332583e+00 3.18182409e-01
3.78983378e-01 -1.87388852e-01 -9.87496972e-02 -1.03221297e+00
-1.24591015e-01 -1.05714691e+00 -3.06987911e-01 -1.27087641e+00
-1.13266551e+00 5.58622181e-01 -1.59803659e-01 -1.13888431e+00
-8.28998268e-01 -5.16058803e-01 -5.44270277e-01 1.14658344e+00
-1.60435724e+00 -1.54921901e+00 -1.30900621e-01 7.17561781e-01
1.13403833e+00 -5.65078855e-01 1.00352299e+00 -6.89374357e-02
-2.53816634e-01 9.68307316e-01 -9.49868038e-02 7.47259781e-02
9.35326755e-01 -6.76053047e-01 5.72962761e-01 6.86118007e-01
1.87938735e-01 1.44832805e-01 8.58645558e-01 -4.08740908e-01
-1.48085856e+00 -9.53357935e-01 7.86508799e-01 -2.90301025e-01
5.70973754e-01 -8.68978262e-01 -7.05193102e-01 3.17985922e-01
4.74576056e-01 3.89711767e-01 7.53717899e-01 -4.93761361e-01
-5.93986154e-01 -1.94650069e-01 -1.10327184e+00 7.89973140e-01
9.95132506e-01 -1.10700238e+00 -2.33322516e-01 1.07720509e-01
6.38045132e-01 -1.46530449e-01 -4.86852646e-01 2.97335654e-01
7.91181087e-01 -9.71379101e-01 1.00555992e+00 -4.75318938e-01
3.84418935e-01 -1.56935558e-01 -4.60952073e-01 -1.30513322e+00
2.48874590e-01 -1.27329373e+00 -2.45243326e-01 1.55399013e+00
4.30583388e-01 -5.09954214e-01 7.70948887e-01 3.29680115e-01
-6.16301298e-02 -5.64670265e-01 -1.22264838e+00 -6.60259128e-01
1.42309666e-01 -6.46992743e-01 5.07041276e-01 3.86756659e-01
1.52656510e-01 3.62003297e-01 -5.90926588e-01 1.08146526e-01
7.37371564e-01 -1.55910611e-01 1.03785241e+00 -8.90305340e-01
-1.18180439e-01 -4.34252590e-01 -9.93269160e-02 -1.08414912e+00
9.10988569e-01 -8.80790770e-01 5.69417179e-01 -1.23765135e+00
-2.04702109e-01 -1.28246412e-01 2.07896203e-01 4.98252422e-01
1.84656873e-01 3.17555189e-01 5.05877137e-01 6.41735643e-02
-1.47034213e-01 9.38961267e-01 1.01197922e+00 -1.45866290e-01
-1.97313845e-01 1.25843570e-01 -3.56950015e-01 5.97994745e-01
8.41567636e-01 -3.17810267e-01 -5.77850878e-01 5.64078763e-02
-3.83992523e-01 6.55460656e-01 4.87742364e-01 -8.60805511e-01
1.94774494e-01 -1.71084255e-01 1.47407159e-01 -4.71435010e-01
1.31257474e+00 -4.72093821e-01 -9.32712927e-02 -7.20137358e-02
-6.33727074e-01 -6.22291386e-01 3.68270725e-01 2.78860271e-01
-6.60286620e-02 7.06030503e-02 1.00239539e+00 2.64953464e-01
-1.42269731e-01 2.71197855e-01 -5.01907408e-01 1.86015680e-01
6.36687219e-01 -2.93365773e-02 -3.26207280e-01 -9.80990291e-01
-6.76107526e-01 -3.02705169e-01 2.76464432e-01 6.54057205e-01
8.30908179e-01 -1.36853814e+00 -1.04855287e+00 5.67375124e-01
1.49811618e-02 -3.42584372e-01 3.51108134e-01 7.24615335e-01
1.15839846e-01 7.02382565e-01 -1.25197247e-01 -6.47341192e-01
-1.69603789e+00 4.70146269e-01 3.17826599e-01 4.59486961e-01
-3.36411476e-01 8.05618227e-01 4.35293704e-01 -1.38454795e-01
3.03112924e-01 5.38995266e-02 6.23215921e-02 1.10970967e-01
5.24792552e-01 3.20263028e-01 -2.37325773e-01 -1.24385130e+00
-3.60856295e-01 5.79799473e-01 4.39073771e-01 -7.15343893e-01
1.09154427e+00 -2.67846882e-01 4.45247859e-01 5.25889874e-01
1.64204621e+00 4.38713789e-01 -1.61712337e+00 2.54460443e-02
-6.30412459e-01 -3.23308229e-01 6.62131310e-02 -6.92571998e-01
-9.20378506e-01 1.31874061e+00 4.64941710e-01 -1.08467825e-02
1.00753605e+00 1.99876398e-01 5.91168344e-01 -6.00190237e-02
-8.93155485e-03 -6.01087868e-01 2.44901001e-01 2.36861810e-01
1.36608267e+00 -1.15260398e+00 -5.24551332e-01 -4.13880676e-01
-9.83626306e-01 9.97353911e-01 1.60473198e-01 3.26314330e-01
6.20037496e-01 5.58318198e-01 3.46905649e-01 4.33819264e-01
-9.80055153e-01 -4.04370874e-01 5.27291656e-01 1.02015734e+00
3.41159552e-01 9.80545208e-02 5.44157386e-01 4.96398449e-01
-5.29488742e-01 1.88337371e-01 2.05296427e-01 7.23166838e-02
-8.65417123e-02 -8.27041447e-01 -3.49426657e-01 -5.72820455e-02
-3.66498470e-01 -2.90823132e-01 -2.85078257e-01 2.23947853e-01
-2.23235220e-01 1.32512617e+00 -2.56762784e-02 -2.74360865e-01
-2.86147427e-02 5.84597826e-01 3.74350309e-01 -3.18334967e-01
1.38581172e-01 4.53873366e-01 1.68859720e-01 -6.47734821e-01
-4.40909676e-02 -7.88071334e-01 -1.11162210e+00 -1.27444550e-01
-1.99286360e-02 -1.17312014e-01 1.15339088e+00 4.85466033e-01
7.14595556e-01 2.33206451e-01 7.94521391e-01 -1.41866088e+00
-6.05520844e-01 -1.07265449e+00 -2.96871006e-01 1.26284823e-01
8.88559818e-01 -4.86211449e-01 -7.19087720e-01 4.37946737e-01] | [14.361948013305664, 5.0094709396362305] |
14adc557-9f70-4e05-b096-e163d43f69cd | a-simple-baseline-for-audio-visual-scene-1 | null | null | http://openaccess.thecvf.com/content_CVPR_2019/html/Schwartz_A_Simple_Baseline_for_Audio-Visual_Scene-Aware_Dialog_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Schwartz_A_Simple_Baseline_for_Audio-Visual_Scene-Aware_Dialog_CVPR_2019_paper.pdf | A Simple Baseline for Audio-Visual Scene-Aware Dialog | The recently proposed audio-visual scene-aware dialog task paves the way to a more data-driven way of learning virtual assistants, smart speakers and car navigation systems. However, very little is known to date about how to effectively extract meaningful information from a plethora of sensors that pound the computational engine of those devices. Therefore, in this paper, we provide and carefully analyze a simple baseline for audio-visual scene-aware dialog which is trained end-to-end. Our method differentiates in a data-driven manner useful signals from distracting ones using an attention mechanism. We evaluate the proposed approach on the recently introduced and challenging audio-visual scene-aware dataset, and demonstrate the key features that permit to outperform the current state-of-the-art by more than 20% on CIDEr.
| [' Tamir Hazan', ' Alexander G. Schwing', 'Idan Schwartz'] | 2019-06-01 | null | null | null | cvpr-2019-6 | ['scene-aware-dialogue'] | ['computer-vision'] | [ 9.51396972e-02 7.63815418e-02 1.80651113e-01 -5.88408232e-01
-1.07941854e+00 -6.71756268e-01 1.01431215e+00 -1.04449429e-02
-5.19669652e-01 5.08053243e-01 5.68458319e-01 -1.60764784e-01
1.83514680e-03 -2.32292518e-01 -3.57071310e-01 -4.56393331e-01
2.43220866e-01 5.79329133e-01 4.44174379e-01 -6.55078769e-01
1.47729427e-01 4.44029927e-01 -1.96602488e+00 3.62670332e-01
3.96077424e-01 1.07701111e+00 4.50965524e-01 9.30591941e-01
-3.07997882e-01 9.72187281e-01 -7.25148439e-01 -1.69837594e-01
-1.10981055e-01 -2.27051169e-01 -8.78396690e-01 1.39968008e-01
7.31439412e-01 -3.94388229e-01 -2.98569083e-01 7.38609731e-01
7.16324389e-01 3.08594942e-01 4.50978041e-01 -1.23664343e+00
-4.01694477e-01 4.00442749e-01 -1.23624280e-01 3.04714441e-01
5.03409982e-01 3.98937076e-01 1.20681930e+00 -1.12946939e+00
8.40420067e-01 1.44367445e+00 1.34435177e-01 6.67008519e-01
-1.23995852e+00 -4.52351689e-01 3.97523046e-01 3.73516470e-01
-1.08139682e+00 -8.45501781e-01 1.21254981e+00 -4.70091283e-01
8.76616895e-01 4.68423098e-01 3.26680601e-01 1.61805010e+00
-3.49614054e-01 1.06628251e+00 9.82317030e-01 -1.36344403e-01
4.25129801e-01 6.08801186e-01 1.50807068e-01 3.50945294e-01
-4.20984358e-01 3.89469080e-02 -9.95469868e-01 1.94386691e-02
1.11129597e-01 -3.19968998e-01 -1.99875519e-01 -7.44530678e-01
-9.60203052e-01 8.90087068e-01 2.78725147e-01 3.98568779e-01
-3.24181795e-01 6.70701964e-03 5.19019961e-01 2.88424969e-01
3.33454251e-01 4.12184566e-01 -1.88741609e-01 -4.30309772e-01
-7.78525054e-01 1.91817433e-01 5.79466760e-01 8.83794129e-01
4.21904355e-01 2.79916763e-01 -4.60467637e-01 8.57239485e-01
4.68904078e-01 5.55345118e-01 1.78444862e-01 -9.97108757e-01
3.91329587e-01 2.96959758e-01 2.91607112e-01 -5.69338202e-01
-4.92329001e-01 -1.35084823e-01 -5.38956821e-01 3.95142615e-01
6.09019220e-01 -4.35517281e-02 -5.43771148e-01 1.52677095e+00
4.21036690e-01 -2.19831206e-02 -1.25106111e-01 1.24165368e+00
1.09272254e+00 5.30172110e-01 1.49500087e-01 -4.89685731e-03
1.46425152e+00 -9.18043673e-01 -9.79603529e-01 -3.07112962e-01
9.43950787e-02 -6.09860897e-01 1.69870663e+00 7.33185112e-01
-9.27671671e-01 -7.24208236e-01 -1.21470606e+00 -2.10946724e-01
-3.83678406e-01 -2.71842867e-01 5.28454661e-01 7.21998990e-01
-1.14663291e+00 1.74121305e-01 -4.85127151e-01 -4.15480494e-01
2.60469943e-01 3.02234367e-02 -2.89492995e-01 1.36660069e-01
-1.02855468e+00 8.38478208e-01 -7.86657038e-04 -2.28953645e-01
-1.27975237e+00 -6.03037596e-01 -7.23576009e-01 5.05943075e-02
7.67410100e-01 -4.97878015e-01 1.62136543e+00 -6.80528879e-01
-1.76140893e+00 8.51761520e-01 -2.05916613e-01 -5.15400171e-01
5.21079659e-01 -4.33411658e-01 -3.91242236e-01 4.10898894e-01
-1.49762467e-01 6.41793013e-01 9.95318353e-01 -1.62311530e+00
-6.03049815e-01 -3.88651848e-01 1.15040280e-01 1.04283720e-01
-3.86765271e-01 5.70236966e-02 -5.03118515e-01 -3.81931096e-01
-4.83332366e-01 -7.06290007e-01 -1.39427871e-01 1.59923673e-01
-4.69222248e-01 -1.74862489e-01 1.03974271e+00 -3.21390003e-01
7.69446373e-01 -2.41111732e+00 1.65605560e-01 -1.22636691e-01
4.14274842e-01 2.56415039e-01 -9.93727744e-02 6.14551067e-01
1.72098652e-01 -2.96780765e-01 -5.73813543e-03 -8.58392358e-01
3.38910073e-01 1.02324180e-01 -5.49120903e-01 3.04605603e-01
2.15222314e-01 6.92197144e-01 -8.23372185e-01 -3.87780309e-01
8.41554940e-01 7.11800158e-01 -5.68353295e-01 4.06561643e-01
-2.81218767e-01 6.62432253e-01 -3.48800093e-01 5.52777767e-01
5.37116945e-01 7.42640421e-02 -3.93945444e-03 2.40651648e-02
-3.39974791e-01 6.04213953e-01 -1.04010785e+00 1.99552906e+00
-6.23587251e-01 1.05421472e+00 5.12936890e-01 -8.46954346e-01
9.28349555e-01 3.80481064e-01 3.47003967e-01 -9.23232377e-01
3.44314277e-01 -1.98701531e-01 -3.49992931e-01 -5.86711884e-01
7.58540511e-01 -9.16904211e-02 -1.82839885e-01 -3.58742326e-02
2.84601152e-01 -1.70008168e-01 -1.50658563e-01 3.53151649e-01
9.73043919e-01 -1.73858941e-01 1.78174689e-01 -1.93596631e-01
7.46474683e-01 4.72708084e-02 -1.97929908e-02 8.25070262e-01
-5.34180582e-01 5.69685876e-01 1.89036146e-01 -5.35738990e-02
-7.24259973e-01 -1.29309881e+00 -4.94856648e-02 1.59255672e+00
1.58481926e-01 -4.13804591e-01 -5.36393642e-01 -6.05512559e-01
-2.30814535e-02 9.64842200e-01 -3.37867647e-01 -2.52706259e-02
-1.46433488e-01 1.64000958e-01 4.25410271e-01 3.96663398e-01
3.31719428e-01 -1.10370672e+00 -7.52212882e-01 2.03357533e-01
-1.61379814e-01 -1.37248945e+00 -2.29425505e-01 1.72495127e-01
-4.88330215e-01 -7.07702339e-01 -7.03286469e-01 -3.76845956e-01
-1.39934465e-01 5.68374991e-01 1.13974881e+00 -5.68573534e-01
-2.39829630e-01 7.98868716e-01 -2.70817190e-01 -6.15808249e-01
-4.40564245e-01 -1.41210660e-01 -6.05547801e-02 2.36165971e-01
4.14029866e-01 -5.88106275e-01 -3.68749529e-01 1.95205927e-01
-4.23087001e-01 -1.61984652e-01 3.74719083e-01 4.79678899e-01
3.39300126e-01 -4.72901583e-01 5.65811276e-01 -7.06061661e-01
7.74365485e-01 -2.39667445e-01 -5.99162042e-01 -1.83524236e-01
-2.45411977e-01 2.07597330e-01 6.63102925e-01 -3.74909878e-01
-1.20455825e+00 4.09192264e-01 -4.04571116e-01 -5.39103866e-01
-6.83248818e-01 1.06556676e-01 -5.33218861e-01 1.82191372e-01
7.41433799e-01 6.48515671e-02 -2.99964279e-01 -6.07965946e-01
9.51069951e-01 1.08292496e+00 1.13199854e+00 -2.66079158e-01
7.34936833e-01 7.71211445e-01 -2.07949832e-01 -1.29755855e+00
-4.64861959e-01 -9.82063711e-01 -3.84640992e-01 -4.81309265e-01
9.46801126e-01 -9.63385761e-01 -1.08537245e+00 1.82592005e-01
-1.14532614e+00 -1.65111080e-01 -5.16218483e-01 2.49443606e-01
-7.07696915e-01 2.21499309e-01 -1.95432991e-01 -1.39791143e+00
-2.26575702e-01 -1.16490400e+00 1.49533260e+00 -1.28144398e-01
-3.67496729e-01 -5.06571889e-01 2.56782830e-01 6.12086236e-01
4.81860340e-01 -8.21324624e-03 3.93552095e-01 -7.55227387e-01
-7.75986373e-01 -4.29662457e-03 -1.69880956e-01 7.97845870e-02
-1.72818035e-01 -3.66526991e-01 -1.96368277e+00 -3.47709879e-02
9.03403610e-02 -4.42068428e-01 9.16410863e-01 1.60104290e-01
9.25400436e-01 -8.09125751e-02 -8.32733065e-02 2.33675808e-01
1.13115966e+00 3.01030815e-01 4.49882448e-01 5.32795601e-02
5.79108775e-01 7.81224489e-01 6.68451130e-01 6.70468390e-01
3.89119238e-01 1.12130833e+00 8.39460313e-01 -9.67176855e-02
-1.82558417e-01 -2.34859884e-01 2.75558233e-01 4.15670961e-01
2.67626584e-01 -3.44061792e-01 -6.64776027e-01 8.60671163e-01
-1.74248755e+00 -9.78057146e-01 -1.38584254e-02 2.03301191e+00
5.13534188e-01 2.60113835e-01 4.79974031e-01 3.59191507e-01
3.14518332e-01 5.59199452e-01 -4.70342278e-01 -5.11617064e-01
2.89405994e-02 1.12112887e-01 7.97674954e-02 6.90130591e-01
-1.22466481e+00 9.44540679e-01 6.02598476e+00 5.12097418e-01
-1.13446391e+00 2.24768505e-01 1.04809977e-01 -1.59968406e-01
-2.28493690e-01 -3.53491545e-01 -5.26150465e-01 7.74071291e-02
1.10293508e+00 6.53780252e-02 3.87280166e-01 1.30416846e+00
2.97609627e-01 -7.69836605e-02 -1.32677293e+00 1.38326931e+00
1.73970088e-01 -1.14390934e+00 -1.98226228e-01 -6.82656020e-02
7.20647052e-02 1.56179607e-01 2.96670884e-01 3.32703829e-01
3.18166941e-01 -9.37075913e-01 8.47340882e-01 3.44179481e-01
5.29018879e-01 -6.44218326e-01 3.32983792e-01 2.37687260e-01
-1.27851713e+00 -6.93326220e-02 -2.34758612e-02 5.20000840e-03
3.40224057e-01 2.39126444e-01 -1.09354162e+00 3.63945365e-01
6.99662387e-01 4.54542547e-01 -4.47070032e-01 8.27554762e-01
-9.90606099e-02 5.20779252e-01 -9.06905681e-02 -2.82536447e-01
2.86584258e-01 2.31704801e-01 8.90658677e-01 1.35059774e+00
-1.75900497e-02 -2.11574763e-01 7.74152344e-03 7.51546025e-01
5.90926073e-02 1.33583903e-01 -9.81906235e-01 4.13134135e-02
3.99316400e-01 1.20145822e+00 -4.44813430e-01 -1.88427418e-01
-5.39523482e-01 9.80871022e-01 1.46149009e-01 2.23012313e-01
-6.62396789e-01 -3.28740180e-01 9.94277835e-01 -6.07073121e-02
4.87611622e-01 -2.75049835e-01 -1.98463678e-01 -8.67048860e-01
-5.67844398e-02 -8.76924098e-01 1.95315644e-01 -9.62608218e-01
-1.16845310e+00 6.96602464e-01 -6.00208975e-02 -1.07189679e+00
-5.35014331e-01 -6.54316604e-01 -3.52578789e-01 5.71635604e-01
-1.42695093e+00 -8.25913608e-01 -5.15261650e-01 8.85086715e-01
9.69137669e-01 -3.07673037e-01 9.02218282e-01 2.53371477e-01
5.30520231e-02 4.80298221e-01 1.76971536e-02 -4.31422144e-02
7.76728213e-01 -1.27738893e+00 4.90453362e-01 6.06719196e-01
5.84199727e-01 2.56047249e-01 1.19951963e+00 -2.13063266e-02
-1.64957201e+00 -6.99720860e-01 7.57006347e-01 -8.20339322e-01
5.48315644e-01 -9.46355462e-01 -8.20704699e-01 4.48531121e-01
5.35699844e-01 -2.68408746e-01 4.15311217e-01 2.83057928e-01
-5.44313252e-01 -2.51662612e-01 -1.06706631e+00 6.65263295e-01
1.02303600e+00 -9.14673209e-01 -7.18661010e-01 -4.28170292e-03
8.13705325e-01 -2.14167416e-01 -2.78216273e-01 3.42413373e-02
4.55069065e-01 -1.29378760e+00 1.41462970e+00 -4.52776462e-01
1.10891640e-01 -1.29915252e-01 -4.37533259e-01 -1.26879513e+00
4.79518399e-02 -8.64206672e-01 -8.68185014e-02 1.36124682e+00
1.84925154e-01 -3.67068380e-01 6.96495652e-01 3.40589672e-01
-4.30461109e-01 -1.94898754e-01 -1.22593510e+00 -6.50733829e-01
-2.76482940e-01 -1.01311624e+00 3.33502114e-01 7.29289472e-01
2.12272972e-01 7.95535743e-01 -5.93987644e-01 5.14376536e-02
4.22031850e-01 -5.04083112e-02 1.11341715e+00 -1.52716732e+00
-3.86847049e-01 -4.40125078e-01 -4.25244778e-01 -1.33256149e+00
5.30828759e-02 -5.63103259e-01 1.82756618e-01 -1.33364499e+00
-7.69000426e-02 7.34902918e-02 -1.46289930e-01 6.26051426e-02
1.03812128e-01 4.34744470e-02 6.04685903e-01 -3.27558130e-01
-9.11493123e-01 7.11277008e-01 9.53461230e-01 -4.95566726e-01
-3.31957728e-01 -1.12522289e-01 -8.95807266e-01 6.12416148e-01
5.49477458e-01 -9.63433683e-02 -5.92765749e-01 -7.33318105e-02
-1.69956133e-01 1.22617699e-01 5.83286107e-01 -1.02449226e+00
2.37296060e-01 5.39713539e-02 -1.11929923e-01 -8.47940922e-01
1.12403607e+00 -1.06843793e+00 -3.09998572e-01 9.32531059e-02
-5.74705303e-01 -4.36079264e-01 4.01558667e-01 6.51257336e-01
-2.48580754e-01 1.63044661e-01 6.33161068e-01 1.45978406e-01
-1.00212550e+00 -1.46442965e-01 -7.21938491e-01 2.03769132e-01
7.33874977e-01 2.23255865e-02 -3.96947771e-01 -9.89977658e-01
-6.22010529e-01 1.28053352e-01 -8.71374905e-02 8.73499811e-01
6.16817653e-01 -1.14897847e+00 -4.01587099e-01 1.20740891e-01
4.58683044e-01 -2.38659441e-01 4.45902646e-01 5.23500144e-01
2.58168634e-02 8.53146434e-01 -2.30282217e-01 -8.82301390e-01
-1.59611189e+00 8.36922646e-01 6.98897243e-02 1.29197374e-01
-5.94712615e-01 7.93417096e-01 2.89416373e-01 -3.26619685e-01
7.98527837e-01 -4.65335727e-01 -3.34547043e-01 2.16426298e-01
7.39680886e-01 2.53738046e-01 8.31943825e-02 -8.07356477e-01
-5.32216966e-01 2.29738727e-01 2.02988729e-01 -6.03271663e-01
1.07363629e+00 -2.98614085e-01 7.99521863e-01 8.90068293e-01
9.87603366e-01 9.75395888e-02 -1.43814087e+00 -2.72407591e-01
-1.59711633e-02 -3.44035864e-01 2.35434189e-01 -7.86922455e-01
-6.76411271e-01 1.29516029e+00 8.52140307e-01 4.18029130e-01
1.00714993e+00 3.08534980e-01 5.06536782e-01 5.90784550e-01
3.28780264e-01 -1.00636768e+00 3.06029886e-01 4.82219905e-01
1.09637260e+00 -1.42969525e+00 -4.85456347e-01 -3.95921081e-01
-1.04878676e+00 6.85673833e-01 3.42619359e-01 1.60335913e-01
4.96616572e-01 3.89150262e-01 3.50779831e-01 -3.03469896e-01
-8.21182907e-01 -6.22347355e-01 3.25302154e-01 1.14542794e+00
1.57648742e-01 -1.83938444e-01 4.00978029e-01 6.26438856e-01
-3.13944578e-01 -2.42372960e-01 1.85705140e-01 6.93392158e-01
-6.91176593e-01 -7.96751976e-01 -4.03796792e-01 -2.16954872e-01
-2.21102029e-01 8.96411464e-02 -8.74922752e-01 9.15295303e-01
-3.08686465e-01 1.53588629e+00 -6.30427152e-04 -4.82325703e-01
7.36260831e-01 3.40937793e-01 1.35831892e-01 -3.36578935e-01
-5.98035693e-01 1.98357508e-01 4.02046144e-01 -8.51807177e-01
-5.12680054e-01 -5.62892318e-01 -1.09237456e+00 -8.97810832e-02
4.33534905e-02 -1.91114485e-01 8.16975176e-01 9.15359199e-01
3.27233464e-01 7.76878238e-01 6.24388933e-01 -1.22580016e+00
-4.96438235e-01 -8.64618301e-01 -4.05958146e-01 5.39857686e-01
6.67121470e-01 -7.60072529e-01 -2.53932774e-01 -1.63799495e-01] | [14.556314468383789, 5.050055980682373] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.