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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
8c394854-8713-4b2d-8632-c5b61117f8fa | sub-label-dependencies-for-neural | null | null | https://aclanthology.org/W18-3904 | https://aclanthology.org/W18-3904.pdf | Sub-label dependencies for Neural Morphological Tagging -- The Joint Submission of University of Colorado and University of Helsinki for VarDial 2018 | This paper presents the submission of the UH{\&}CU team (Joint University of Colorado and University of Helsinki team) for the VarDial 2018 shared task on morphosyntactic tagging of Croatian, Slovenian and Serbian tweets. Our system is a bidirectional LSTM tagger which emits tags as character sequences using an LSTM generator in order to be able to handle unknown tags and combinations of several tags for one token which occur in the shared task data sets. To the best of our knowledge, using an LSTM generator is a novel approach. The system delivers sizable improvements of more than 6{\%}-points over a baseline trigram tagger. Overall, the performance of our system is quite even for all three languages. | ['Senka Drobac', 'Miikka Silfverberg'] | 2018-08-01 | null | null | null | coling-2018-8 | ['morphological-tagging'] | ['natural-language-processing'] | [-1.82575490e-02 2.29612365e-01 -7.68639743e-02 -3.74233454e-01
-1.09011388e+00 -7.78758943e-01 8.38977575e-01 3.30567509e-01
-8.35693538e-01 9.82099414e-01 2.16081634e-01 -7.51663387e-01
5.47193289e-01 -7.98847139e-01 -6.56713963e-01 -3.84813935e-01
-1.11813307e-01 1.06878996e+00 2.81452626e-01 -4.21628922e-01
-3.20213228e-01 1.19065240e-01 -9.39057350e-01 4.79300678e-01
4.79768097e-01 5.96516907e-01 7.38055035e-02 6.32284582e-01
-3.32869887e-01 7.70949364e-01 -4.33796376e-01 -7.04299748e-01
3.69243473e-02 -1.31166577e-01 -1.25627339e+00 -7.19507813e-01
5.66115081e-01 5.35831340e-02 -2.35067472e-01 9.33363616e-01
5.77150941e-01 2.39174813e-01 6.05781198e-01 -7.03880727e-01
-4.10738111e-01 1.54038930e+00 -9.45334435e-02 3.45416903e-01
1.80084974e-01 -3.70514005e-01 1.30269337e+00 -8.94289076e-01
6.00862563e-01 8.87695730e-01 7.68064618e-01 5.97741783e-01
-1.06343317e+00 -9.40658867e-01 -8.59687626e-02 8.33493546e-02
-1.53231931e+00 -3.39307547e-01 2.74454445e-01 -6.09254420e-01
1.31780100e+00 2.16089040e-01 1.91843182e-01 1.41231763e+00
-7.53587559e-02 7.57487893e-01 1.10121763e+00 -6.03963077e-01
-7.63788298e-02 3.24092835e-01 4.59602028e-02 5.98506570e-01
-1.21445343e-01 -1.42802626e-01 -3.94045293e-01 -3.05162042e-01
2.57258385e-01 -5.12883067e-01 2.74634928e-01 5.03870428e-01
-1.10266137e+00 9.19612348e-01 9.65862796e-02 7.01939642e-01
1.64298154e-02 3.56809050e-01 6.46995962e-01 1.57721400e-01
8.52022231e-01 4.25931215e-01 -6.22730970e-01 -6.05735362e-01
-1.00076950e+00 7.51518458e-02 8.53786767e-01 1.19030821e+00
7.65060425e-01 2.61707127e-01 -3.31869721e-01 9.62863624e-01
9.33953822e-02 5.71107030e-01 7.45325565e-01 -3.27632725e-01
6.81659877e-01 -4.29514796e-02 1.54549852e-01 -6.04205489e-01
-4.90359664e-01 -2.80353218e-01 -5.32987237e-01 -5.39866149e-01
4.56861943e-01 -8.02134633e-01 -8.90140414e-01 1.80666983e+00
8.38320181e-02 5.21036386e-01 -1.37706697e-01 4.04713631e-01
8.62910151e-01 8.23982298e-01 6.05834305e-01 1.27050757e-01
1.27465880e+00 -7.49257803e-01 -7.16868520e-01 -5.19371450e-01
1.02132833e+00 -1.01888883e+00 8.31985474e-01 -7.07743019e-02
-1.12224686e+00 -5.61815441e-01 -7.44206727e-01 -2.86432467e-02
-6.57379568e-01 -2.57760361e-02 4.92401332e-01 9.42575336e-01
-1.15504098e+00 4.48065251e-01 -8.21976900e-01 -4.51791704e-01
-8.80849138e-02 3.16015273e-01 -1.78707749e-01 2.81221002e-01
-1.82291448e+00 8.48005176e-01 6.23343885e-01 -1.08291313e-01
-4.64472771e-01 -6.56996608e-01 -1.02371824e+00 -1.93209514e-01
-2.61512697e-01 -1.80977434e-01 1.41680825e+00 -6.91681504e-01
-1.30267596e+00 1.27505887e+00 -2.57544935e-01 -8.25911045e-01
5.84260404e-01 -1.89574704e-01 -6.75672412e-01 -4.41277564e-01
9.52638462e-02 4.81451184e-01 1.86704308e-01 -8.36844087e-01
-1.03814662e+00 1.17601223e-01 -4.77991879e-01 -9.47713032e-02
-3.89729142e-01 3.69627625e-01 -1.06793828e-02 -7.38516390e-01
-2.73850888e-01 -1.07679915e+00 4.00525667e-02 -1.25244224e+00
-4.47111964e-01 -4.17504191e-01 3.69337410e-01 -8.92094493e-01
1.19999003e+00 -1.88217962e+00 -4.92633998e-01 2.42193431e-01
-2.42030174e-01 4.73931283e-01 -4.03078422e-02 6.91118062e-01
7.70457089e-02 4.03750092e-01 -5.08593246e-02 -7.17855394e-01
2.02140272e-01 3.73803645e-01 -6.94057107e-01 5.10695994e-01
-9.29573327e-02 1.00855994e+00 -1.13072717e+00 -3.63779038e-01
-1.12161383e-01 4.08392549e-01 -2.16932699e-01 9.04479325e-02
-1.75207425e-02 2.53148288e-01 -9.76021588e-03 1.57026261e-01
5.03580451e-01 2.75625199e-01 4.18728292e-01 3.70093971e-01
-3.61855417e-01 9.79314923e-01 -7.82076240e-01 1.35593402e+00
-6.77220523e-01 7.11204648e-01 -3.86603385e-01 -5.17767310e-01
1.17340028e+00 6.84758306e-01 3.93830240e-01 -6.16345465e-01
1.79899186e-01 6.10798419e-01 -5.30220382e-02 -1.65635735e-01
9.39889491e-01 -5.16061664e-01 -5.02686918e-01 4.31689858e-01
1.97241306e-01 1.36151329e-01 9.27163959e-02 3.42053585e-02
9.80198085e-01 1.21031143e-02 1.24924414e-01 -3.42786312e-01
3.23855102e-01 -2.69063413e-01 5.61322153e-01 1.00289917e+00
1.65473507e-03 6.28476143e-01 1.85710922e-01 -3.65101367e-01
-1.18543398e+00 -9.32915926e-01 -1.98529959e-01 1.52826321e+00
-5.03676891e-01 -4.63798761e-01 -5.74532151e-01 -7.30309725e-01
-2.28294760e-01 9.57274437e-01 -6.57859743e-01 2.65105888e-02
-1.08262622e+00 -6.45814836e-01 1.40195274e+00 5.29402316e-01
2.71572977e-01 -1.42733359e+00 1.00052319e-01 6.01666272e-01
-4.34629738e-01 -1.55273354e+00 -5.11479497e-01 4.28506851e-01
-2.50350863e-01 -3.60965699e-01 -3.07588369e-01 -1.14031553e+00
3.83952819e-02 -3.84659022e-01 1.20048881e+00 -2.19932020e-01
1.43020242e-01 -1.55713663e-01 -5.99196494e-01 -4.41754133e-01
-6.26614630e-01 6.81445479e-01 -7.10588023e-02 -6.74971938e-02
6.32914722e-01 -3.79768819e-01 4.85083088e-04 -3.74576598e-02
-5.80759943e-01 -2.09402055e-01 1.47822529e-01 7.30070293e-01
2.98835367e-01 -4.87854779e-01 2.91763902e-01 -1.62563896e+00
4.41600829e-01 -8.35580051e-01 -3.87564361e-01 -1.52475193e-01
-3.86300415e-01 -1.30692005e-01 7.63992012e-01 -1.76993161e-01
-1.08191895e+00 -1.39212281e-01 -6.48854673e-01 1.98544160e-01
-3.75509113e-01 4.83842015e-01 -6.19987696e-02 4.85769659e-01
4.51595247e-01 1.88879266e-01 -6.71311498e-01 -7.19708443e-01
4.97665405e-01 9.09015357e-01 6.29163504e-01 -6.45314634e-01
6.72874808e-01 2.69387007e-01 -3.55661094e-01 -6.13307297e-01
-1.03290653e+00 -7.48516500e-01 -7.16332376e-01 3.29236612e-02
9.24746096e-01 -9.17201221e-01 -3.59925687e-01 6.52075171e-01
-1.29629278e+00 -7.25247622e-01 -3.21736515e-01 3.91353816e-01
-3.45167786e-01 3.94499786e-02 -1.14839685e+00 -5.24103940e-01
-4.24424917e-01 -7.09053814e-01 7.73036242e-01 -5.92958853e-02
-5.75173855e-01 -1.57410765e+00 4.73573923e-01 5.14218509e-02
4.24139917e-01 1.03572998e-02 8.85812163e-01 -1.41509926e+00
2.12925196e-01 -2.85257161e-01 2.63755098e-02 9.37414616e-02
-4.09969911e-02 -1.88779801e-01 -1.25728357e+00 -2.75853842e-01
-4.82699245e-01 -2.67478973e-01 8.70698571e-01 1.78821236e-02
7.28566766e-01 -7.11295426e-01 -3.19300413e-01 4.87140536e-01
1.10344160e+00 9.68743190e-02 6.85229480e-01 4.35839862e-01
7.62920797e-01 1.79510757e-01 3.17127466e-01 5.31275928e-01
8.20741296e-01 7.68319845e-01 1.22965090e-01 -1.37336515e-02
-1.86173081e-01 -6.73624516e-01 5.09721994e-01 1.19958031e+00
1.29128262e-01 -5.45659721e-01 -1.29567778e+00 9.70810533e-01
-1.68113840e+00 -1.03314078e+00 -5.86446762e-01 2.29196048e+00
1.16058004e+00 1.47279948e-01 5.92660643e-02 -1.81151062e-01
9.29405093e-01 2.43649140e-01 5.36578417e-01 -1.04925942e+00
-3.35754119e-02 8.81581128e-01 9.43419456e-01 9.34397876e-01
-1.38514662e+00 1.62909925e+00 6.23463297e+00 1.04526246e+00
-1.29703033e+00 5.26692986e-01 3.75763446e-01 2.66132295e-01
-5.43980956e-01 9.02261809e-02 -1.24809980e+00 6.97952151e-01
1.55026281e+00 -3.97778787e-02 1.38699159e-01 4.76295561e-01
-2.42668558e-02 2.25961015e-01 -7.44154096e-01 5.78315794e-01
3.83166187e-02 -1.08057153e+00 -2.05742270e-01 1.59003347e-01
8.96802366e-01 7.52432227e-01 2.25267351e-01 5.68956017e-01
1.03056288e+00 -1.12107956e+00 9.62481141e-01 3.08800966e-01
7.72586167e-01 -8.78544807e-01 1.09524381e+00 3.81509662e-01
-9.34844971e-01 5.37453890e-01 -3.19115251e-01 -6.74543530e-02
3.64200324e-01 7.14981675e-01 -1.61254144e+00 3.39933664e-01
4.84365046e-01 4.98038024e-01 -5.23294568e-01 9.69531417e-01
-4.32854623e-01 1.53362954e+00 -3.77425998e-01 -1.15269803e-01
7.36926913e-01 8.76551569e-02 5.27491212e-01 2.06846309e+00
4.19483155e-01 -3.39523852e-01 2.84204990e-01 3.45421493e-01
-3.98469388e-01 3.42499226e-01 -5.32924056e-01 4.39480990e-02
7.60786474e-01 1.39752460e+00 -5.45499265e-01 -3.89318645e-01
-1.30692482e-01 8.82427752e-01 6.20683908e-01 1.59279972e-01
-8.04479182e-01 -4.28485662e-01 4.44446981e-01 1.78985551e-01
3.37316871e-01 -4.48228002e-01 -3.55361700e-01 -8.49035382e-01
-4.25339133e-01 -4.09789592e-01 6.98895693e-01 -2.31652007e-01
-1.28870392e+00 8.87642264e-01 -2.32624218e-01 -8.55623960e-01
-7.81218469e-01 -3.32302749e-01 -4.18827802e-01 1.02490950e+00
-1.43230343e+00 -1.41348886e+00 3.23412716e-01 1.83091551e-01
1.91297047e-02 -2.20041841e-01 1.07337964e+00 6.07361615e-01
-1.15469828e-01 8.49431634e-01 2.87279844e-01 6.38843715e-01
1.00992846e+00 -1.49931955e+00 1.06375730e+00 7.62800157e-01
4.11980510e-01 3.60725462e-01 8.66814077e-01 -6.50815964e-01
-4.19397593e-01 -1.46774232e+00 2.14848828e+00 -4.55936313e-01
1.03569889e+00 -7.84553766e-01 -7.76841581e-01 1.29102457e+00
4.54307705e-01 1.53545300e-02 7.77810931e-01 3.20744127e-01
-4.08071250e-01 1.76498204e-01 -7.62432575e-01 3.30483347e-01
7.34423459e-01 -5.79946697e-01 -3.58938038e-01 3.49313915e-01
5.86360395e-01 -6.04948640e-01 -7.77452767e-01 1.64251104e-01
2.76936173e-01 -6.71492398e-01 5.83154738e-01 -7.13090360e-01
1.70910284e-02 8.39094445e-02 -1.86215028e-01 -1.36296105e+00
-3.61930788e-01 -7.86805630e-01 3.59474659e-01 1.84756315e+00
8.54841292e-01 -5.93361855e-01 5.71582556e-01 2.53791571e-01
-3.25635642e-01 -1.04353301e-01 -1.32944286e+00 -8.50483477e-01
6.19707167e-01 -8.54620278e-01 4.81134653e-01 1.22492063e+00
1.92407057e-01 2.66408533e-01 -6.40093803e-01 -2.00286970e-01
1.35160536e-01 -3.82399112e-01 3.44089687e-01 -1.07200885e+00
-1.66630983e-01 -2.66923904e-01 -3.47185016e-01 -6.88249528e-01
8.16409826e-01 -1.40210617e+00 2.84707636e-01 -1.35545003e+00
-1.64961264e-01 -9.50506985e-01 -3.07814777e-01 1.01903260e+00
6.60663098e-02 7.74560571e-01 1.47462666e-01 -1.11402161e-02
-4.47877139e-01 1.99389219e-01 4.57195699e-01 -1.44313604e-01
-4.78786929e-03 7.17469975e-02 -3.64708364e-01 4.68264580e-01
9.10484374e-01 -8.08737099e-01 2.73206532e-01 -7.10999548e-01
6.59902215e-01 -3.00949752e-01 5.85292093e-02 -8.36113632e-01
9.27028507e-02 5.02585955e-02 -7.02114701e-02 -3.12225252e-01
2.75458604e-01 -3.57157826e-01 1.28911063e-01 2.42357925e-01
-4.40755457e-01 3.05578440e-01 3.73127311e-01 2.10121814e-02
-2.18727574e-01 -3.04536998e-01 7.98729837e-01 -2.72950113e-01
-5.00290871e-01 2.92659163e-01 -7.66517818e-01 5.84587514e-01
6.68303013e-01 2.36586750e-01 -3.32637995e-01 -3.72563690e-01
-6.77421272e-01 1.54050170e-02 5.18447496e-02 5.52848697e-01
6.66014710e-03 -1.40478098e+00 -9.53021228e-01 -3.83283235e-02
-6.37965724e-02 -3.78372788e-01 1.31515209e-02 5.87968767e-01
-4.68964338e-01 5.71856558e-01 2.47388110e-02 -1.48958843e-02
-9.34162974e-01 -2.73092296e-02 2.61621267e-01 -6.31404638e-01
-4.10602987e-01 1.10436511e+00 -3.75120997e-01 -8.20001006e-01
-1.71734005e-01 2.12252419e-02 -2.62743562e-01 5.11914313e-01
2.87164390e-01 3.20668846e-01 6.55917108e-01 -1.02827215e+00
-4.50531393e-01 -1.91140324e-02 -1.30944252e-01 -5.46685994e-01
1.26318383e+00 -5.05226366e-02 -2.36352950e-01 7.78868735e-01
1.32397377e+00 4.74927872e-01 -5.75057566e-01 -3.55158091e-01
2.27295458e-01 1.22426853e-01 -1.45546317e-01 -7.21528351e-01
-6.94931984e-01 9.39618945e-01 2.23841280e-01 1.60797417e-01
1.33084163e-01 1.29057780e-01 1.26977038e+00 8.21461752e-02
2.04666719e-01 -1.21680403e+00 -6.65225506e-01 1.32009327e+00
1.91814691e-01 -7.36874402e-01 -7.70439267e-01 6.19885996e-02
-7.25868344e-01 9.04588938e-01 3.99066597e-01 -1.25211611e-01
5.65061212e-01 4.90808487e-01 3.77935499e-01 2.38242954e-01
-6.27723455e-01 -4.40999597e-01 2.42810488e-01 6.46512449e-01
8.74778807e-01 3.67741704e-01 -5.86624622e-01 5.52130580e-01
-8.25021029e-01 -5.08573711e-01 7.06873000e-01 6.57751143e-01
-3.76018792e-01 -1.39835310e+00 -1.84469093e-02 2.37980545e-01
-9.54894841e-01 -7.12442756e-01 -2.33095646e-01 7.92586029e-01
2.24877641e-01 8.26912761e-01 4.37255383e-01 -5.08196890e-01
6.28516600e-02 4.67867762e-01 2.02488706e-01 -9.45067406e-01
-1.02780330e+00 5.81765324e-02 5.57457209e-01 -1.59010608e-02
-1.09068513e-01 -8.66297424e-01 -1.48430383e+00 -2.84669220e-01
7.79578686e-02 5.46011865e-01 5.38784385e-01 1.16797066e+00
-4.57311273e-02 1.48626611e-01 3.70169908e-01 -3.89676392e-01
-3.51878554e-01 -1.29612911e+00 -8.15106213e-01 5.09408116e-01
7.59831816e-02 -2.21253723e-01 -9.87742320e-02 4.41744290e-02] | [9.97833251953125, 9.789222717285156] |
0ba73b45-266d-4987-8989-e74f0dd2cd21 | wssod-a-new-pipeline-for-weakly-and-semi | 2105.11293 | null | https://arxiv.org/abs/2105.11293v1 | https://arxiv.org/pdf/2105.11293v1.pdf | WSSOD: A New Pipeline for Weakly- and Semi-Supervised Object Detection | The performance of object detection, to a great extent, depends on the availability of large annotated datasets. To alleviate the annotation cost, the research community has explored a number of ways to exploit unlabeled or weakly labeled data. However, such efforts have met with limited success so far. In this work, we revisit the problem with a pragmatic standpoint, trying to explore a new balance between detection performance and annotation cost by jointly exploiting fully and weakly annotated data. Specifically, we propose a weakly- and semi-supervised object detection framework (WSSOD), which involves a two-stage learning procedure. An agent detector is first trained on a joint dataset and then used to predict pseudo bounding boxes on weakly-annotated images. The underlying assumptions in the current as well as common semi-supervised pipelines are also carefully examined under a unified EM formulation. On top of this framework, weakly-supervised loss (WSL), label attention and random pseudo-label sampling (RPS) strategies are introduced to relax these assumptions, bringing additional improvement on the efficacy of the detection pipeline. The proposed framework demonstrates remarkable performance on PASCAL-VOC and MSCOCO benchmark, achieving a high performance comparable to those obtained in fully-supervised settings, with only one third of the annotations. | ['Wayne Zhang', 'Dahua Lin', 'Kai Chen', 'Xinjiang Wang', 'Yuhang Cao', 'Shijie Fang'] | 2021-05-21 | null | null | null | null | ['semi-supervised-object-detection'] | ['computer-vision'] | [ 3.67131323e-01 2.93003201e-01 -2.37717882e-01 -5.19725442e-01
-1.01810575e+00 -3.66313815e-01 7.43725300e-01 8.71475339e-02
-8.25812221e-01 6.43785179e-01 -1.73804343e-01 8.00480768e-02
3.95615399e-01 -3.09507757e-01 -6.45747125e-01 -8.98062825e-01
3.67952466e-01 3.51014048e-01 6.38636887e-01 2.77781278e-01
3.03765777e-02 2.47583300e-01 -1.58030057e+00 1.43269837e-01
7.61587203e-01 8.82663786e-01 5.01615703e-01 5.50826728e-01
3.51445563e-02 9.52396214e-01 -3.86157572e-01 -6.45483792e-01
2.83756554e-01 -2.33793080e-01 -8.28930199e-01 4.81284380e-01
3.17901790e-01 -2.99490064e-01 2.59936415e-02 1.16945863e+00
4.60226208e-01 -1.82347283e-01 6.19496167e-01 -1.06723440e+00
-2.29586691e-01 4.94938970e-01 -9.25658882e-01 2.42650546e-02
-9.16304160e-03 2.80124873e-01 1.11583698e+00 -1.16024685e+00
4.18265700e-01 9.71286595e-01 6.65829122e-01 4.70140547e-01
-1.24357021e+00 -3.96900833e-01 3.04915339e-01 -3.11919004e-02
-1.52945554e+00 -3.28371555e-01 6.77925706e-01 -4.82452512e-01
6.40079379e-01 -6.13427255e-03 3.36716354e-01 9.58386421e-01
-2.59438157e-01 1.22076511e+00 1.19009256e+00 -7.48179615e-01
1.82578951e-01 7.58938015e-01 1.62471563e-01 7.84289718e-01
1.90794766e-01 8.79017562e-02 -2.45706111e-01 -9.46686640e-02
4.12958950e-01 -2.10317284e-01 -7.28517100e-02 -7.62507498e-01
-1.01466763e+00 7.29653478e-01 2.99092919e-01 2.68605262e-01
-2.93891847e-01 -1.17880985e-01 4.87700641e-01 -2.16267049e-01
7.24207103e-01 1.73597008e-01 -4.40034777e-01 1.91301256e-01
-1.05524600e+00 -4.86332923e-03 5.62336683e-01 9.51109290e-01
7.51170933e-01 -1.33090898e-01 -3.31837416e-01 9.04820979e-01
5.08408427e-01 2.79243916e-01 1.69479951e-01 -5.80751061e-01
4.18750495e-01 6.64576292e-01 2.72532761e-01 -7.03208923e-01
-4.20140773e-01 -7.57921457e-01 -4.51556504e-01 1.31442249e-01
4.54180837e-01 -2.42480393e-02 -9.53993976e-01 1.69301546e+00
6.58738256e-01 1.33938417e-01 -5.19521013e-02 9.57426131e-01
4.63679373e-01 3.15814197e-01 5.38630128e-01 -1.44809738e-01
1.34760892e+00 -1.26838768e+00 -6.57465398e-01 -3.32040608e-01
7.46597230e-01 -8.23840618e-01 9.97126639e-01 2.32709229e-01
-9.99756932e-01 -7.16906071e-01 -1.07293618e+00 -6.41784221e-02
-3.06807578e-01 7.30401695e-01 4.61101592e-01 8.24327648e-01
-8.67232203e-01 1.61814228e-01 -1.03786528e+00 -4.89290714e-01
5.75261652e-01 2.95868099e-01 -2.62708902e-01 -9.81153771e-02
-7.04637170e-01 1.00303876e+00 7.17853189e-01 3.55457902e-01
-1.02476978e+00 -4.04800713e-01 -7.91138470e-01 -1.54569924e-01
7.51835287e-01 -4.18772787e-01 1.25366974e+00 -9.74978983e-01
-1.34672403e+00 1.20977378e+00 5.77179454e-02 -6.10214293e-01
8.15050423e-01 -3.95522863e-01 -9.90312546e-02 9.69085619e-02
2.05456972e-01 9.49129999e-01 7.38327503e-01 -1.35539997e+00
-1.03451610e+00 -4.17333484e-01 8.73404145e-02 1.98278129e-01
-4.13183957e-01 2.61804312e-01 -8.15269232e-01 -5.41007638e-01
-2.12386772e-01 -1.09633338e+00 -3.47747535e-01 5.13888858e-02
-5.28501689e-01 -4.17317301e-01 8.75468373e-01 -4.23661977e-01
9.66122091e-01 -2.14066529e+00 3.27837542e-02 -2.65859306e-01
7.33789355e-02 6.16936147e-01 -2.88013332e-02 1.40803412e-01
1.36557907e-01 -3.11774611e-01 -5.04772842e-01 -9.53388393e-01
-6.67424127e-02 2.22175866e-01 -8.70291367e-02 6.60098612e-01
6.15528464e-01 8.16863894e-01 -8.96365285e-01 -7.68031359e-01
3.52107674e-01 3.67328823e-01 -4.05362159e-01 3.43472004e-01
-3.34082454e-01 4.77377117e-01 -4.58555847e-01 8.27022016e-01
6.78715408e-01 -3.17323357e-01 1.60272017e-01 -1.85771286e-01
-2.53123462e-01 -6.18671626e-02 -1.29348099e+00 1.59820533e+00
-2.49451756e-01 3.83894891e-01 2.77422100e-01 -1.13680398e+00
6.42654538e-01 2.16202959e-01 3.46455157e-01 -3.78247410e-01
2.20831975e-01 2.29497820e-01 -1.70491993e-01 -5.09634733e-01
4.42636520e-01 -9.10294726e-02 5.51961772e-02 2.61786193e-01
3.08496088e-01 8.27428028e-02 3.39899659e-01 8.90930519e-02
8.22749674e-01 6.20230198e-01 4.87766027e-01 -1.48920298e-01
7.25432873e-01 1.66903958e-01 6.29976690e-01 9.06443894e-01
-5.34147501e-01 6.43676400e-01 2.47363508e-01 -2.11350948e-01
-9.70610023e-01 -7.58470297e-01 -3.44599962e-01 1.29454219e+00
2.42109567e-01 -2.35021919e-01 -1.00248218e+00 -1.01402462e+00
-2.01784253e-01 5.67547977e-01 -5.42895257e-01 -4.01070081e-02
-4.68213648e-01 -1.10199261e+00 5.91907799e-01 6.82710350e-01
6.00498259e-01 -9.65968251e-01 -6.53558254e-01 1.36751950e-01
4.73201163e-02 -1.46048844e+00 -2.14641064e-01 5.72127819e-01
-6.47590101e-01 -1.03983951e+00 -7.21834898e-01 -7.83443570e-01
7.87471950e-01 3.55032504e-01 9.56788778e-01 -3.76026854e-02
-4.91556019e-01 2.95522392e-01 -4.58738178e-01 -4.72785592e-01
-3.15341800e-01 9.93165076e-02 7.54702166e-02 2.56121814e-01
5.55859983e-01 4.65976670e-02 -4.82990205e-01 4.63462532e-01
-8.78304899e-01 3.82413939e-02 9.71991420e-01 9.09642994e-01
6.89491510e-01 -2.18498394e-01 5.22491336e-01 -1.02455473e+00
-9.72118527e-02 -2.30382919e-01 -7.96812713e-01 2.00663239e-01
-6.19105458e-01 -6.82095140e-02 3.53380680e-01 -3.61472934e-01
-1.43617928e+00 6.95254862e-01 -1.63137481e-01 -3.59709591e-01
-3.61286938e-01 5.53951338e-02 -3.62532854e-01 -1.29488349e-01
4.76306289e-01 1.15701571e-01 -2.54013002e-01 -6.02853954e-01
4.70570087e-01 8.62390041e-01 5.59014320e-01 -4.45402890e-01
8.71502042e-01 7.62023330e-01 -3.86067212e-01 -7.54974365e-01
-1.38208497e+00 -8.58942568e-01 -1.03322589e+00 -1.45722181e-01
1.00275445e+00 -1.05782211e+00 -1.98349550e-01 4.39469665e-01
-1.02461791e+00 -2.87301749e-01 -3.10942918e-01 6.04968965e-01
-3.73923421e-01 4.97002184e-01 -5.14445961e-01 -1.07936656e+00
-9.85205472e-02 -1.20447886e+00 1.28898776e+00 1.11424431e-01
1.59366429e-02 -8.11356127e-01 -3.37602533e-02 6.95886135e-01
1.44688830e-01 1.35086522e-01 3.24304610e-01 -8.56622100e-01
-6.22951329e-01 -5.05660415e-01 -5.42753577e-01 7.15698719e-01
-1.66895539e-01 -1.94763362e-01 -1.30989945e+00 -3.74362320e-01
2.02300437e-02 -6.81663275e-01 1.02621627e+00 2.40579396e-01
9.21273947e-01 2.55567044e-01 -5.42117417e-01 2.70121306e-01
1.30571055e+00 -1.91354081e-01 2.36215636e-01 3.91610056e-01
7.38305211e-01 7.97274709e-01 9.65277851e-01 3.73255074e-01
3.54808152e-01 7.34084427e-01 5.36098540e-01 -3.13656300e-01
-2.15325043e-01 -1.77395135e-01 2.41138592e-01 4.04607207e-01
1.97820924e-02 -1.56537950e-01 -7.94841230e-01 5.61924458e-01
-1.94441855e+00 -6.24057353e-01 -1.85454711e-01 2.16919184e+00
7.45471776e-01 4.93927121e-01 1.87763885e-01 8.21871087e-02
8.25711071e-01 1.49515912e-01 -5.54967940e-01 1.61539719e-01
-1.72529295e-01 -3.04628611e-01 5.68993926e-01 2.51443475e-01
-1.53929877e+00 9.79266703e-01 6.12234735e+00 8.34713757e-01
-9.17114973e-01 3.50496322e-01 6.34223819e-01 -3.71336658e-03
3.93527985e-01 1.00515030e-01 -1.16737676e+00 2.89635509e-01
6.28895700e-01 4.57199603e-01 -2.43637502e-01 1.22019899e+00
2.12168232e-01 -1.95994407e-01 -1.08814788e+00 7.46161163e-01
2.85071582e-01 -8.56408238e-01 -2.76388884e-01 -1.81844607e-02
7.92113483e-01 2.53012925e-01 -1.19396048e-02 5.50453722e-01
1.62845701e-01 -5.20992517e-01 9.12697077e-01 1.07019559e-01
5.53516686e-01 -5.22275507e-01 9.12103832e-01 6.26871645e-01
-1.20041287e+00 -1.63718387e-01 -3.62403572e-01 -2.99906284e-02
1.88283861e-01 6.38969481e-01 -8.14755380e-01 4.45473492e-01
6.49796903e-01 5.16192555e-01 -7.69769788e-01 1.16161573e+00
-4.36104447e-01 6.29073262e-01 -2.59658545e-01 5.93949109e-02
3.48366141e-01 3.01764421e-02 3.53686631e-01 1.46140814e+00
-1.72680199e-01 -1.12853013e-01 5.73030353e-01 9.02182758e-01
-1.00459374e-01 6.17315806e-02 -1.63922578e-01 2.08663002e-01
2.10310847e-01 1.66593337e+00 -9.63172615e-01 -3.14168304e-01
-7.22459435e-01 9.42982852e-01 6.22974992e-01 2.12807059e-01
-9.78803873e-01 -7.73263630e-03 2.60377880e-02 1.47467107e-02
3.91043484e-01 -2.38901135e-02 -1.47088423e-01 -1.08683479e+00
1.84606522e-01 -5.37680864e-01 4.52804029e-01 -4.55066293e-01
-1.13442409e+00 4.79354352e-01 5.84387854e-02 -1.20630729e+00
6.38175104e-03 -6.04534209e-01 -4.41411823e-01 5.54747224e-01
-1.87820745e+00 -1.50607610e+00 -3.17857057e-01 1.38549134e-01
7.87886262e-01 5.85382245e-02 5.53892076e-01 5.37602603e-01
-9.86224294e-01 4.94550705e-01 1.78626031e-02 1.29897937e-01
7.73313344e-01 -1.38348544e+00 4.82502319e-02 1.02646041e+00
3.00111800e-01 3.90122294e-01 6.82183743e-01 -3.82530749e-01
-1.04159856e+00 -1.32265866e+00 6.22507155e-01 -5.16074657e-01
4.96202856e-01 -4.92871106e-01 -9.59504962e-01 6.24324203e-01
1.09560303e-01 4.34253752e-01 5.72234511e-01 -1.02410324e-01
-2.23892644e-01 1.47706315e-01 -1.06669307e+00 3.53308082e-01
8.37474525e-01 -4.43499178e-01 -5.87789237e-01 5.46034694e-01
5.52122474e-01 -3.02331954e-01 -4.35973585e-01 6.10068560e-01
2.72385180e-01 -9.02162790e-01 8.35629404e-01 -2.09020197e-01
4.61772308e-02 -5.90098441e-01 -8.54002759e-02 -8.01287770e-01
-8.21704268e-02 -3.31881881e-01 -1.70392022e-01 1.45145571e+00
3.24780345e-01 -2.88398713e-01 1.05317879e+00 5.36711931e-01
-1.76680759e-01 -7.86917686e-01 -6.94750369e-01 -7.30128825e-01
-2.69030452e-01 -4.63147819e-01 -5.74167632e-02 6.97606087e-01
-2.66778052e-01 3.40146095e-01 -4.45499510e-01 3.29650342e-01
8.39827716e-01 8.26362800e-03 8.51076603e-01 -1.09257674e+00
-4.55154896e-01 -2.55554855e-01 -2.63346732e-01 -1.17830884e+00
1.39977217e-01 -7.29820013e-01 4.21664655e-01 -1.18289983e+00
6.87607825e-01 -4.42881018e-01 -3.89394939e-01 4.47766662e-01
-5.10913074e-01 6.93436861e-01 8.94713253e-02 4.72329468e-01
-1.12119472e+00 5.68957150e-01 7.64322698e-01 -6.93987459e-02
-4.84477766e-02 4.49076062e-04 -4.32036459e-01 1.14380944e+00
5.11783540e-01 -5.79093099e-01 -1.15872517e-01 -4.42262501e-01
-2.44905859e-01 -3.16248149e-01 4.89584804e-01 -9.63673055e-01
3.56353641e-01 2.23246232e-01 2.66267985e-01 -7.90474355e-01
3.22886407e-01 -8.00592959e-01 -3.25680196e-01 4.19484824e-01
-4.35138881e-01 -4.92109597e-01 6.85735745e-03 9.78067338e-01
-3.70318025e-01 -3.97751123e-01 1.18644059e+00 -1.18010296e-02
-7.78652906e-01 1.09736152e-01 -2.74303705e-02 -1.01493724e-01
1.44775653e+00 -2.62408376e-01 5.14449999e-02 2.65938967e-01
-6.65807962e-01 3.96297812e-01 3.84311736e-01 3.77325922e-01
2.12167934e-01 -1.00124168e+00 -7.53089666e-01 -2.27594562e-02
4.82379913e-01 1.16510540e-01 1.29349455e-01 9.77638483e-01
-2.20959067e-01 4.72754210e-01 1.91758111e-01 -8.91179442e-01
-1.22751558e+00 6.92473888e-01 1.54262692e-01 -5.05354464e-01
-4.94222969e-01 9.59858477e-01 3.36010188e-01 -4.16307002e-01
5.78652024e-01 3.07294596e-02 -2.31326908e-01 1.46651581e-01
5.62475145e-01 3.19998562e-01 4.02659290e-02 -8.12370658e-01
-3.46321046e-01 4.40278322e-01 -3.77248645e-01 1.42854765e-01
1.17979848e+00 -3.65201861e-01 1.82215154e-01 2.90275276e-01
1.00934613e+00 -1.56794623e-01 -1.58077979e+00 -5.52808285e-01
4.37129617e-01 -3.74546409e-01 1.75806656e-01 -6.28739238e-01
-8.68369699e-01 8.15030694e-01 6.83171451e-01 9.86865759e-02
9.22427595e-01 1.97173983e-01 5.40839195e-01 2.55058706e-01
4.05716777e-01 -1.22742963e+00 2.99018204e-01 1.44980654e-01
4.12685454e-01 -1.68220961e+00 -7.00171068e-02 -5.34936786e-01
-6.77211404e-01 6.99820757e-01 7.21808195e-01 1.62026826e-02
2.88758665e-01 3.25989097e-01 -6.47812057e-03 1.03384210e-02
-4.64229286e-01 -6.22733057e-01 2.55834699e-01 4.67953265e-01
4.94434893e-01 -2.52121985e-01 -3.15540820e-01 4.69829619e-01
5.89570761e-01 -2.63195243e-02 1.88076928e-01 9.82556462e-01
-6.19269729e-01 -1.06078541e+00 -4.13384706e-01 2.71044403e-01
-6.63139403e-01 4.35820408e-02 -3.67585570e-01 8.48329723e-01
3.48256677e-01 1.02112222e+00 -2.27062896e-01 5.05780391e-02
2.77244538e-01 3.26205380e-02 2.13821530e-01 -8.54498267e-01
-4.17488545e-01 3.07365149e-01 -9.49405483e-04 -4.37692851e-01
-8.77856970e-01 -6.46226525e-01 -1.01848328e+00 4.86148834e-01
-8.94832969e-01 3.67549877e-03 6.51958227e-01 1.01331103e+00
1.19506739e-01 4.78716999e-01 4.24942136e-01 -1.05382979e+00
-7.80184567e-01 -1.05220759e+00 -5.51851451e-01 4.55663711e-01
8.41052309e-02 -7.80181170e-01 -3.67685258e-01 1.86914995e-01] | [9.237126350402832, 1.2222390174865723] |
5e62f93f-25a9-4cec-bd29-108e3eb5dbc9 | how-to-evaluate-the-next-system-automatic | 1912.04664 | null | https://arxiv.org/abs/1912.04664v1 | https://arxiv.org/pdf/1912.04664v1.pdf | How to Evaluate the Next System: Automatic Dialogue Evaluation from the Perspective of Continual Learning | Automatic dialogue evaluation plays a crucial role in open-domain dialogue research. Previous works train neural networks with limited annotation for conducting automatic dialogue evaluation, which would naturally affect the evaluation fairness as dialogue systems close to the scope of training corpus would have more preference than the other ones. In this paper, we study alleviating this problem from the perspective of continual learning: given an existing neural dialogue evaluator and the next system to be evaluated, we fine-tune the learned neural evaluator by selectively forgetting/updating its parameters, to jointly fit dialogue systems have been and will be evaluated. Our motivation is to seek for a lifelong and low-cost automatic evaluation for dialogue systems, rather than to reconstruct the evaluator over and over again. Experimental results show that our continual evaluator achieves comparable performance with reconstructing new evaluators, while requires significantly lower resources. | ['dianhai yu', 'Zhongheng He', 'Xiangyang Zhou', 'Lu Li'] | 2019-12-10 | null | null | null | null | ['dialogue-evaluation'] | ['natural-language-processing'] | [ 1.27446085e-01 5.75007796e-01 1.04317367e-01 -6.35895252e-01
-6.20406568e-01 -7.53198922e-01 5.95564306e-01 9.34540331e-02
-8.55364621e-01 1.24686849e+00 4.36118513e-01 -3.40231836e-01
2.20238611e-01 -5.75193942e-01 -9.10970196e-02 -4.22808826e-01
8.14079586e-03 1.09710872e+00 9.69861075e-02 -6.39158368e-01
8.93467143e-02 -3.31039284e-03 -1.11200440e+00 2.93390006e-01
1.05413747e+00 4.47176456e-01 4.19102088e-02 1.07798755e+00
-1.14144636e-02 1.15858841e+00 -1.04117656e+00 -8.99306715e-01
-5.45454472e-02 -6.78705990e-01 -1.58765209e+00 -1.13344956e-02
1.75233051e-01 -9.41649199e-01 -1.16692483e-01 9.01421070e-01
7.22197890e-01 4.71537828e-01 4.97775614e-01 -8.64857376e-01
-3.43953192e-01 1.17048955e+00 1.22958213e-01 -1.57123283e-01
3.83619398e-01 3.96869868e-01 1.16089499e+00 -5.63382626e-01
4.82722074e-01 1.42419887e+00 5.51933765e-01 1.07762110e+00
-1.34142339e+00 -9.30765867e-02 6.74151927e-02 -7.02004554e-03
-8.68919134e-01 -8.43733966e-01 6.71719253e-01 -1.70508683e-01
9.39318955e-01 2.59274900e-01 4.77837354e-01 1.16546690e+00
-2.23301828e-01 1.03398085e+00 8.21249843e-01 -5.94784379e-01
1.44097790e-01 6.56541049e-01 3.18194628e-01 7.89187908e-01
-3.57223958e-01 -2.03069195e-01 -2.71017611e-01 -4.28106815e-01
4.57920879e-01 -8.03530633e-01 -2.73449063e-01 -2.48173922e-01
-8.43185782e-01 1.05197024e+00 -3.22542489e-02 4.55413729e-01
-4.02166218e-01 -2.05632105e-01 8.98765326e-01 9.44544673e-01
4.02546614e-01 1.17818868e+00 -7.06872344e-01 -6.15505517e-01
-5.98489046e-01 1.17915653e-01 1.58545601e+00 4.77322489e-01
7.30511427e-01 1.10530306e-03 -5.41493118e-01 1.40235245e+00
-7.37743303e-02 -1.46480545e-01 6.32487297e-01 -1.41703844e+00
3.42131853e-01 6.94818914e-01 7.38555640e-02 -4.25370514e-01
-3.74420077e-01 2.86488235e-02 -7.18208432e-01 2.14093804e-01
5.59595525e-01 -8.44674766e-01 -2.68471658e-01 1.97160208e+00
3.06729347e-01 -5.65876484e-01 4.12348360e-01 7.78836429e-01
6.94561362e-01 6.07352853e-01 -1.20954867e-02 -5.08063734e-01
9.31373119e-01 -1.32177329e+00 -7.67314792e-01 4.58253659e-02
8.74652624e-01 -5.58282018e-01 1.39129031e+00 4.99675035e-01
-1.29004037e+00 -5.81961751e-01 -9.13575768e-01 2.52545383e-02
6.54683933e-02 1.36059642e-01 5.69095790e-01 6.72655046e-01
-1.26937687e+00 6.73101723e-01 -4.87326562e-01 -2.79534996e-01
-1.64207324e-01 3.61497849e-01 7.07363244e-03 5.05435884e-01
-1.64607823e+00 1.43186903e+00 6.77556396e-01 8.57926533e-02
-9.68989074e-01 -1.32422462e-01 -6.03848279e-01 1.58006638e-01
5.59538662e-01 -5.44454575e-01 2.03708601e+00 -1.12397754e+00
-2.07171869e+00 6.07677698e-01 2.72450209e-01 -5.95047712e-01
7.10564137e-01 -2.41803333e-01 7.34348549e-03 -1.16598226e-01
-1.91951662e-01 7.38735259e-01 2.77805358e-01 -1.29430199e+00
-6.29214048e-01 -1.80406690e-01 7.73974955e-01 7.97866285e-01
-7.78308213e-01 -3.37718368e-01 -4.04042095e-01 7.39429444e-02
-4.39418733e-01 -8.10184836e-01 -1.72653049e-01 -5.03951252e-01
-4.94466066e-01 -6.38822258e-01 4.70500648e-01 -6.27285242e-01
1.57126808e+00 -1.53877854e+00 2.89912581e-01 -1.15827799e-01
2.67723054e-01 6.46190524e-01 -1.86839819e-01 4.45070803e-01
2.66319662e-01 -2.96431947e-02 -7.41630197e-02 -5.30235171e-01
1.75515220e-01 2.22987086e-01 2.85847690e-02 -5.09566814e-02
5.08508570e-02 6.42149746e-01 -8.63859534e-01 -6.50133133e-01
-4.61248346e-02 -1.96992867e-02 -5.97453058e-01 7.04674125e-01
-5.42285383e-01 3.99702668e-01 -3.69929194e-01 1.95095479e-01
1.30314812e-01 -7.34055564e-02 4.70803618e-01 -9.70675424e-02
-6.60021007e-02 4.92009968e-01 -9.76390541e-01 1.53791714e+00
-8.29586327e-01 6.34785235e-01 3.52775961e-01 -7.35416591e-01
9.92893577e-01 5.52630842e-01 2.37704992e-01 -7.79920220e-01
1.45780489e-01 -8.79209489e-02 4.22661752e-01 -6.23805821e-01
1.02264738e+00 -1.12784296e-01 -2.11770892e-01 1.01796210e+00
4.11806822e-01 4.70169634e-03 5.55177510e-01 2.31524631e-01
9.71083641e-01 5.37488051e-02 2.92950124e-01 -9.30619612e-03
7.89295495e-01 6.80597872e-02 3.45122844e-01 8.84582102e-01
-4.97872084e-01 3.13678309e-02 6.82083011e-01 -4.52263892e-01
-1.15604722e+00 -5.10600448e-01 -7.05724163e-03 1.74834645e+00
-2.57710785e-01 -1.69823006e-01 -1.22174251e+00 -8.89926136e-01
-4.31980491e-01 1.01002371e+00 -4.02901441e-01 -3.06720883e-01
-8.13206196e-01 -6.01315320e-01 1.06438923e+00 1.13595463e-01
7.06566155e-01 -1.38919592e+00 -5.26171446e-01 6.34686708e-01
-4.01095271e-01 -4.35078830e-01 -4.43736434e-01 4.08151418e-01
-5.73003471e-01 -8.32351685e-01 -7.04464674e-01 -7.58567929e-01
3.27122569e-01 -3.37665260e-01 1.49541998e+00 2.29726732e-01
2.38135532e-01 2.30828509e-01 -1.85651332e-01 1.88893173e-02
-1.22348368e+00 6.11851573e-01 5.12842983e-02 -1.78077355e-01
2.70654619e-01 -3.20626915e-01 -2.52963483e-01 2.76495844e-01
-6.14665329e-01 4.85775545e-02 3.37814212e-01 1.35080254e+00
-2.27862075e-01 -1.54401129e-02 9.64728653e-01 -1.39616096e+00
1.72410977e+00 -1.16410553e-01 -3.33498448e-01 7.05272496e-01
-9.36150610e-01 3.24608535e-01 7.59621263e-01 -4.59104002e-01
-1.65227830e+00 -1.37239024e-01 -3.82662535e-01 4.89405394e-02
-1.09076492e-01 5.00968099e-01 -1.47141591e-01 2.40005195e-01
1.03605258e+00 -7.28959143e-02 3.51338759e-02 -2.68308640e-01
3.82071435e-01 8.42927933e-01 3.35844934e-01 -9.00966287e-01
1.65841386e-01 -4.05678183e-01 -7.91501105e-01 -6.55243695e-01
-7.82012939e-01 -2.86830872e-01 -6.38829172e-01 -4.52159107e-01
6.16422355e-01 -5.22041082e-01 -9.61180091e-01 3.55021805e-01
-1.48075581e+00 -7.23424256e-01 -3.70670110e-01 2.04078019e-01
-5.83339691e-01 3.09147030e-01 -9.33008313e-01 -1.09413362e+00
-5.73086023e-01 -1.28372717e+00 5.04225731e-01 2.64146656e-01
-7.75552809e-01 -1.42382014e+00 7.01764584e-01 4.30296898e-01
5.12450337e-01 -3.30142945e-01 9.69536424e-01 -1.09140360e+00
-2.51041185e-02 -3.32934223e-02 -3.85299884e-02 7.77732253e-01
-4.26165238e-02 -1.58844426e-01 -1.20867562e+00 -3.33452404e-01
1.12330526e-01 -1.18050599e+00 5.08230746e-01 -4.57636192e-02
7.58887947e-01 -7.24112153e-01 1.38817906e-01 -2.97058597e-02
8.33263457e-01 3.62380147e-01 4.38605219e-01 2.27878481e-01
4.63632852e-01 1.00095594e+00 5.32897949e-01 5.28048038e-01
3.51804733e-01 5.40703893e-01 -7.18153790e-02 -8.95583183e-02
4.59477678e-02 -1.04345329e-01 4.42834377e-01 1.10651064e+00
4.37570810e-02 -5.08252919e-01 -7.74750352e-01 3.58751953e-01
-1.87245083e+00 -7.16551542e-01 3.85949969e-01 2.17813087e+00
1.53694403e+00 1.85806260e-01 1.80022925e-01 -2.55934358e-01
7.30333805e-01 1.23490728e-01 -7.39494443e-01 -8.67392600e-01
3.70903835e-02 6.81912005e-02 -1.18302330e-01 1.06695437e+00
-8.55745137e-01 9.69627857e-01 6.08871222e+00 6.14977002e-01
-8.23598623e-01 1.47793248e-01 9.15219665e-01 1.43474057e-01
-2.41057649e-01 6.23204485e-02 -6.57001913e-01 1.03459448e-01
1.35146129e+00 -2.01858997e-01 4.54018980e-01 7.30445385e-01
2.13816576e-02 -2.00391918e-01 -1.36459064e+00 3.13300341e-01
-5.77227818e-03 -8.23470294e-01 6.77813739e-02 -1.85463011e-01
5.90234697e-01 -2.72766441e-01 -3.97038221e-01 9.57804680e-01
7.98352659e-01 -7.86943018e-01 2.20809504e-01 5.23262739e-01
4.95378941e-01 -8.00344408e-01 8.63749504e-01 6.46576345e-01
-4.29978371e-01 1.75812930e-01 -3.51186663e-01 -3.49109918e-02
1.86085686e-01 -4.61981073e-03 -1.61029613e+00 -2.69330442e-02
9.98939648e-02 -1.76796675e-01 -4.80824769e-01 6.89321697e-01
-1.48227826e-01 5.73111236e-01 1.31224664e-02 -5.94264448e-01
3.40717733e-01 -1.80112213e-01 4.90844756e-01 1.18158686e+00
-1.67311549e-01 -1.00749269e-01 4.01454121e-01 5.90533018e-01
-2.41366759e-01 2.14609176e-01 -5.51302671e-01 -1.11230761e-01
4.13803786e-01 1.25010216e+00 -1.78137168e-01 -5.21109104e-01
3.28818299e-02 1.11981213e+00 7.31151104e-01 1.54775634e-01
-6.01103246e-01 -5.08211970e-01 1.77269608e-01 -2.73064524e-01
-2.89229751e-01 1.00784697e-01 -4.77682725e-02 -1.10256171e+00
-2.08462387e-01 -1.22882426e+00 3.75534743e-01 -3.78635705e-01
-1.15384078e+00 1.05009222e+00 -2.36427158e-01 -7.89549649e-01
-9.01918948e-01 -3.03120196e-01 -5.08151472e-01 8.62242460e-01
-1.04017270e+00 -7.47558415e-01 -4.42858115e-02 4.83586252e-01
9.81860876e-01 -4.14522737e-01 1.38898873e+00 2.30967626e-01
-5.28212965e-01 8.24189305e-01 3.02576311e-02 2.20330030e-01
9.24236894e-01 -1.57083130e+00 6.57152459e-02 1.68965951e-01
-1.37581840e-01 6.96793795e-01 8.14784169e-01 -4.25727844e-01
-1.03995252e+00 -6.38840914e-01 1.00833237e+00 -3.51650000e-01
6.73627317e-01 -1.05877139e-01 -9.99394178e-01 2.58064568e-01
8.39119375e-01 -9.43235099e-01 7.17293441e-01 7.48971403e-01
1.27419084e-01 1.05315909e-01 -1.02220881e+00 8.70953441e-01
5.63606739e-01 -6.32458270e-01 -7.62642980e-01 3.67983401e-01
6.41877890e-01 -4.86059129e-01 -9.17965412e-01 3.10457647e-01
5.08540809e-01 -9.67944443e-01 6.34695172e-01 -7.29727209e-01
3.59501958e-01 1.42560855e-01 2.25424632e-01 -1.70939350e+00
-2.46963445e-02 -8.80869031e-01 -1.28431633e-01 1.49041629e+00
8.40975106e-01 -3.17823172e-01 8.53452325e-01 9.46125031e-01
-1.85555980e-01 -8.26430142e-01 -6.80873573e-01 -3.45086187e-01
3.11788231e-01 1.94062278e-01 3.60318094e-01 9.40460026e-01
3.47615063e-01 1.29277611e+00 -6.19857252e-01 -4.21314210e-01
2.36311212e-01 -7.84981176e-02 9.41394210e-01 -1.19557273e+00
-6.99095547e-01 -5.94782293e-01 4.21083033e-01 -1.29025996e+00
3.28469574e-01 -4.71985996e-01 5.28801024e-01 -1.17085922e+00
1.30740821e-01 -5.44586062e-01 8.57425332e-02 3.76926541e-01
-3.17430824e-01 -1.21327765e-01 -1.07429460e-01 1.86520383e-01
-1.08145595e+00 8.22333634e-01 1.45155120e+00 -2.63922423e-01
-5.95790684e-01 3.56026769e-01 -7.40088344e-01 6.57768667e-01
7.75960505e-01 -1.26697972e-01 -7.34542251e-01 -3.73488158e-01
6.27389327e-02 5.96030533e-01 -1.93640873e-01 -6.80980682e-01
5.51153600e-01 1.15114920e-01 3.40366475e-02 -2.87040591e-01
4.26317632e-01 -3.15680325e-01 -2.71223843e-01 2.46417493e-01
-1.08243203e+00 1.59133688e-01 -8.02519470e-02 6.82852417e-02
-3.45965177e-01 -1.03419137e+00 7.56797373e-01 -4.68493402e-01
-5.72184563e-01 -3.27534229e-02 -7.05930531e-01 2.13747561e-01
6.82177663e-01 -5.47474548e-02 -3.31172317e-01 -7.60575473e-01
-9.49582160e-01 5.92080712e-01 1.74787238e-01 2.86865514e-02
3.35649073e-01 -9.41132307e-01 -7.84117281e-01 -2.12633446e-01
1.81243625e-02 -7.60951862e-02 1.11540645e-01 3.16125363e-01
-3.41884106e-01 4.27701682e-01 -7.41173699e-02 -3.32082391e-01
-1.32162905e+00 2.46958863e-02 5.36860764e-01 -8.84046972e-01
-1.70053810e-01 9.38030124e-01 -5.60588837e-02 -1.21412778e+00
8.20233703e-01 2.25834787e-01 -5.75218141e-01 4.58602935e-01
2.66806304e-01 3.68975163e-01 1.51605755e-01 -2.55660266e-01
4.32415068e-01 -4.98360723e-01 -5.81596792e-01 -5.03147662e-01
1.14790487e+00 -3.05199325e-01 -1.31641597e-01 7.83746183e-01
1.11802649e+00 -2.81653434e-01 -1.28099275e+00 -5.00444531e-01
4.16937694e-02 5.15332073e-02 -1.19268604e-01 -1.23675191e+00
-5.52484155e-01 7.45558918e-01 3.06697428e-01 6.70597970e-01
7.42172480e-01 -2.73135066e-01 6.89938366e-01 1.28908074e+00
3.74440014e-01 -1.56104946e+00 3.86452377e-01 1.16785622e+00
9.07733500e-01 -1.30151534e+00 -1.32615000e-01 3.60498995e-01
-1.02270055e+00 1.26574922e+00 1.28936756e+00 1.51188388e-01
1.00435533e-01 4.66133393e-02 2.74729103e-01 -2.92099297e-01
-1.19241989e+00 5.73775172e-02 -5.46115562e-02 3.49575728e-01
7.60370731e-01 8.27464536e-02 -2.86107838e-01 4.17223513e-01
-3.34747314e-01 -2.80714452e-01 6.20759606e-01 7.28298008e-01
-5.27858377e-01 -1.39061666e+00 6.26154169e-02 3.52631807e-01
-1.29207537e-01 -1.37451321e-01 -8.68460953e-01 5.44938684e-01
-1.58087865e-01 9.09036338e-01 -1.47334501e-01 -3.86530012e-01
4.82469887e-01 4.34924394e-01 3.44539583e-01 -6.54584885e-01
-1.11562777e+00 -6.22072369e-02 1.00468993e+00 -1.41207710e-01
-3.26965868e-01 -5.56199610e-01 -9.72966850e-01 -3.08912784e-01
-7.72283614e-01 7.72965074e-01 3.76363575e-01 8.82250786e-01
-8.68483931e-02 6.06351435e-01 9.41217244e-01 -5.12346268e-01
-1.35275114e+00 -1.53231275e+00 -2.43228674e-01 1.09939478e-01
2.26180553e-01 -4.20654774e-01 -1.82195112e-01 -9.92608443e-02] | [12.90770149230957, 8.001869201660156] |
4eddf75d-56dd-431e-908f-6bbf14671a0a | causal-discovery-from-conditionally-1 | 2110.06257 | null | https://arxiv.org/abs/2110.06257v1 | https://arxiv.org/pdf/2110.06257v1.pdf | Causal discovery from conditionally stationary time-series | Causal discovery, i.e., inferring underlying cause-effect relationships from observations of a scene or system, is an inherent mechanism in human cognition, but has been shown to be highly challenging to automate. The majority of approaches in the literature aiming for this task consider constrained scenarios with fully observed variables or data from stationary time-series. In this work we aim for causal discovery in a more general class of scenarios, scenes with non-stationary behavior over time. For our purposes we here regard a scene as a composition objects interacting with each other over time. Non-stationarity is modeled as stationarity conditioned on an underlying variable, a state, which can be of varying dimension, more or less hidden given observations of the scene, and also depend more or less directly on these observations. We propose a probabilistic deep learning approach called State-Dependent Causal Inference (SDCI) for causal discovery in such conditionally stationary time-series data. Results in two different synthetic scenarios show that this method is able to recover the underlying causal dependencies with high accuracy even in cases with hidden states. | ['Hedvig Kjellstrom', 'Ruibo Tu', 'Carles Balsells Rodas'] | 2021-10-12 | causal-discovery-from-conditionally | https://openreview.net/forum?id=q9zIvzRaU94 | https://openreview.net/pdf?id=q9zIvzRaU94 | null | ['probabilistic-deep-learning'] | ['computer-vision'] | [ 4.82684761e-01 -7.45157152e-02 3.48813832e-02 -4.11812097e-01
-3.72322500e-01 -5.34687817e-01 1.23581302e+00 2.54701972e-01
6.69829100e-02 9.38605070e-01 3.60381305e-01 -3.25070381e-01
-5.56434095e-01 -8.35624218e-01 -9.91803050e-01 -8.47472847e-01
-5.60402691e-01 6.35285258e-01 1.57260969e-01 2.63443828e-01
1.23575419e-01 5.54618955e-01 -1.63984299e+00 5.87267578e-02
4.43432301e-01 3.86022091e-01 3.20092380e-01 8.68987620e-01
1.13695092e-01 9.65237558e-01 -4.99895245e-01 4.76920567e-02
-2.35115618e-01 -3.51998776e-01 -8.33801568e-01 -4.60907398e-03
2.08278671e-01 -1.27893165e-01 -3.51383448e-01 7.42729902e-01
-8.24360773e-02 -3.69835347e-02 8.00101161e-01 -1.41797125e+00
-3.93999398e-01 6.55504465e-01 -2.75554180e-01 2.53501445e-01
3.36451322e-01 1.06141545e-01 9.97322500e-01 -2.75340855e-01
3.74601811e-01 1.72814775e+00 1.86605126e-01 2.15500146e-02
-1.77751255e+00 -5.18861234e-01 4.24829602e-01 5.13992310e-01
-1.13722825e+00 -2.74948925e-01 9.02386367e-01 -9.22900021e-01
6.02702439e-01 1.76897630e-01 3.72280568e-01 1.69740963e+00
5.86231947e-01 5.60701847e-01 1.20386302e+00 -2.00335756e-01
6.33916676e-01 -4.32365417e-01 1.89394578e-01 2.68564939e-01
2.31230959e-01 2.56810308e-01 -6.08304918e-01 -2.18112707e-01
7.55407631e-01 3.08906641e-02 -9.64172632e-02 -4.23784882e-01
-1.52862954e+00 8.13948929e-01 2.44961306e-01 2.32554615e-01
-7.27604210e-01 5.86332858e-01 8.21411312e-02 2.08778173e-01
3.11345875e-01 2.30978802e-01 -6.50888681e-01 1.31676733e-01
-6.83841407e-01 5.15025735e-01 8.06854606e-01 6.25774682e-01
4.83197927e-01 -8.32761824e-02 -3.25579524e-01 6.06032126e-02
2.81852841e-01 7.38196492e-01 -1.58208728e-01 -6.20017648e-01
8.12126398e-02 1.87795326e-01 6.04057550e-01 -7.92400420e-01
-5.22936881e-01 -3.33836466e-01 -9.43253100e-01 1.75219178e-01
5.50249517e-01 -1.49318516e-01 -1.05888665e+00 2.25715494e+00
3.84603977e-01 7.95684934e-01 2.08910815e-02 8.45784605e-01
5.54552555e-01 8.29941452e-01 3.24096948e-01 -5.73262453e-01
1.19435191e+00 -9.46682598e-03 -9.72661734e-01 9.84717975e-04
-2.92901725e-01 -6.62476897e-01 7.07054913e-01 4.01979178e-01
-6.27709806e-01 -4.76798326e-01 -6.63922787e-01 2.78375834e-01
-2.24288061e-01 -1.50882840e-01 8.13359797e-01 1.54228106e-01
-7.89787173e-01 3.42898518e-01 -1.32330024e+00 -3.35445732e-01
9.80728045e-02 3.10136288e-01 -2.37100601e-01 -6.65892661e-02
-1.39744425e+00 6.60453022e-01 4.44251984e-01 -1.44654457e-02
-1.69776261e+00 -7.01107144e-01 -4.95469123e-01 2.11723447e-01
6.72102511e-01 -7.91032255e-01 1.22035289e+00 -6.63646042e-01
-1.03156233e+00 2.71732479e-01 -5.35048962e-01 -5.24522483e-01
4.84493494e-01 -3.81455719e-01 -6.20705426e-01 -1.57489866e-01
2.84379929e-01 1.31418020e-01 1.06604290e+00 -1.46578646e+00
-3.97133946e-01 -3.93571973e-01 2.24251598e-01 -2.83565640e-01
3.51495594e-01 -4.89364341e-02 -5.58513254e-02 -3.11945498e-01
7.75401592e-02 -1.04602671e+00 -2.12677777e-01 -4.18663621e-01
-9.13616776e-01 -2.94253170e-01 9.79979932e-01 -5.53881288e-01
8.87124062e-01 -1.77917588e+00 3.90332848e-01 -1.26429752e-01
1.71633705e-01 -1.42846599e-01 8.94126948e-03 6.99135363e-01
-5.63746929e-01 -1.25111791e-03 -4.44163233e-01 -3.45632464e-01
-1.44801602e-01 5.21458030e-01 -6.49469614e-01 7.10566461e-01
6.47692204e-01 5.96896052e-01 -1.23159468e+00 -2.02870294e-01
4.71861929e-01 5.68727136e-01 -1.47838682e-01 3.63656044e-01
-6.26236081e-01 1.17029619e+00 -5.11829853e-01 1.43294111e-01
4.09990013e-01 -3.32479149e-01 2.30669111e-01 1.01286933e-01
-4.89407182e-01 2.40492001e-01 -1.42061257e+00 1.47850215e+00
-5.49748421e-01 7.14489162e-01 -3.06636214e-01 -1.16186798e+00
5.58597267e-01 7.94183552e-01 4.97747719e-01 -3.31556141e-01
-4.56879660e-02 -2.45525882e-01 2.28955656e-01 -8.31442475e-01
-1.92643739e-02 -3.00793499e-01 -1.60525009e-01 3.81271213e-01
8.90091136e-02 1.96841925e-01 1.46145135e-01 1.98447064e-01
1.51718140e+00 2.43241061e-02 3.53428423e-01 -2.27970213e-01
3.18609655e-01 2.18944382e-02 6.39576495e-01 9.12944615e-01
1.77938581e-01 4.22485769e-01 8.29947531e-01 -4.19938743e-01
-1.14127731e+00 -1.36786473e+00 -3.72517437e-01 5.53816199e-01
1.47723377e-01 -5.46376295e-02 -2.29251623e-01 -2.67021090e-01
1.33821066e-03 1.28482199e+00 -9.64804113e-01 -7.44561628e-02
-5.30658484e-01 -9.05336082e-01 -5.61704934e-02 1.51914269e-01
2.06842333e-01 -1.12555766e+00 -8.11662316e-01 4.44368482e-01
-3.61442178e-01 -1.15575027e+00 3.64869207e-01 2.40169898e-01
-5.85834801e-01 -1.17168665e+00 -2.88783818e-01 6.31045178e-02
3.07955742e-01 1.84244029e-02 1.21558714e+00 -5.87054670e-01
-4.46097404e-01 4.16310042e-01 -1.28043488e-01 -5.61865509e-01
-5.28024673e-01 -4.14043456e-01 1.81723878e-01 5.10461509e-01
4.81254645e-02 -8.10319245e-01 -3.72594804e-01 9.92845818e-02
-1.11140394e+00 1.02759548e-01 4.98229742e-01 6.08659744e-01
4.48195279e-01 7.35345840e-01 3.11942726e-01 -7.40220845e-01
2.11832553e-01 -9.51012731e-01 -7.53925502e-01 1.93896770e-01
-1.23541184e-01 4.22244430e-01 5.69638610e-01 -5.62641442e-01
-1.38039041e+00 1.94384888e-01 3.65341753e-01 -5.29286683e-01
-7.44365573e-01 6.35849655e-01 -4.46942240e-01 9.32666838e-01
5.34926832e-01 3.68017741e-02 -4.73379016e-01 -3.85088950e-01
3.40903193e-01 -5.91170825e-02 5.68758368e-01 -6.26281202e-01
7.97344387e-01 8.92289877e-01 6.61728203e-01 -5.87235749e-01
-9.72434580e-01 -3.57084215e-01 -1.02148449e+00 -3.37685585e-01
1.09060860e+00 -9.46573496e-01 -7.61073709e-01 2.31399238e-01
-1.48651266e+00 -4.67430145e-01 -5.00475839e-02 7.56182969e-01
-4.98885632e-01 -3.06218117e-01 -2.48086736e-01 -1.22675920e+00
4.65479344e-01 -8.32202852e-01 1.27087653e+00 -7.14819506e-02
-3.80124897e-01 -1.36911976e+00 3.73682976e-01 -9.12095532e-02
-3.71593772e-03 6.87075555e-01 1.02117014e+00 -3.12426925e-01
-8.06336939e-01 -3.49808484e-01 -1.53473139e-01 -2.01234445e-01
3.94256741e-01 2.88176090e-01 -1.08509421e+00 -2.12405194e-02
5.05721476e-03 2.50069588e-01 7.48171866e-01 7.53735960e-01
1.03233981e+00 -3.47891420e-01 -5.50217569e-01 -6.95576593e-02
1.40742087e+00 2.48210803e-01 4.33026493e-01 -7.10303709e-02
8.84932697e-01 8.99554193e-01 5.77962279e-01 5.56967020e-01
2.39083648e-01 6.82563543e-01 9.78805006e-01 -5.08495308e-02
3.98840047e-02 -2.48603895e-01 1.35748833e-01 1.44360498e-01
-1.51472434e-01 -6.48647606e-01 -1.10918140e+00 9.10987377e-01
-2.12681127e+00 -1.36772227e+00 -1.08729219e+00 2.24884534e+00
5.41768670e-01 -4.92504612e-02 -1.20053530e-01 5.41938916e-02
9.29488897e-01 -1.45158628e-02 -8.88655782e-01 2.04611570e-01
-2.03997046e-01 2.65732091e-02 4.71628934e-01 6.14568233e-01
-1.23040271e+00 6.78304851e-01 6.18952370e+00 1.31434157e-01
-1.16949975e+00 1.97640479e-01 3.47133040e-01 -1.53429613e-01
-3.58434856e-01 3.49721819e-01 -4.44948137e-01 4.86860305e-01
1.17333913e+00 -1.01472273e-01 2.32010096e-01 3.15873653e-01
8.40043068e-01 -1.99816257e-01 -1.43595254e+00 5.66516042e-01
-5.12696505e-01 -1.07946026e+00 -8.09428766e-02 1.97074115e-01
9.15962160e-01 -1.15281746e-01 6.43762723e-02 -8.53728205e-02
9.49277043e-01 -1.05155718e+00 8.99667323e-01 9.82627869e-01
4.67759162e-01 -4.69149321e-01 4.11796421e-01 5.19960165e-01
-8.25804830e-01 -1.35260716e-01 1.10964134e-01 -4.78518695e-01
5.17409921e-01 1.18851161e+00 -8.27710092e-01 4.61656630e-01
5.91547668e-01 8.49343061e-01 -1.83174476e-01 8.95878911e-01
-7.53219485e-01 1.14521718e+00 -2.17083395e-01 1.70836136e-01
3.91237848e-02 9.98359844e-02 9.33094561e-01 1.03562510e+00
2.75709093e-01 1.27159029e-01 4.80671376e-02 1.26898301e+00
5.12040377e-01 -6.38674259e-01 -6.14484727e-01 1.53757155e-01
2.95402944e-01 9.41533089e-01 -7.48822391e-01 -4.60322052e-01
-2.25883469e-01 7.24846661e-01 9.04597342e-02 5.66363931e-01
-1.13464820e+00 5.08251846e-01 6.70069158e-01 6.82722107e-02
2.16156349e-01 -6.23534799e-01 -1.19633585e-01 -1.26089704e+00
-1.39696956e-01 -2.68252879e-01 3.26335102e-01 -9.44172561e-01
-1.38189995e+00 1.84630379e-01 5.29840171e-01 -9.97244120e-01
-4.00874555e-01 -2.66557723e-01 -7.39542842e-01 1.05769861e+00
-1.36874163e+00 -1.18470705e+00 -2.18473837e-01 8.78077865e-01
5.27446151e-01 3.15777868e-01 7.80939460e-01 -2.89697386e-02
-7.07286596e-01 -4.77405667e-01 2.39695366e-02 -2.96857476e-01
4.78038162e-01 -1.52889884e+00 4.45204467e-01 1.11415398e+00
2.18934044e-01 6.76604450e-01 1.49132729e+00 -9.85074997e-01
-1.40434694e+00 -1.31090486e+00 7.96760023e-01 -6.47304296e-01
1.11656559e+00 -6.29801214e-01 -1.10389900e+00 7.61101663e-01
1.84854105e-01 -1.21426240e-01 4.00516391e-01 4.37389344e-01
-3.17026407e-01 1.44771054e-01 -7.50615060e-01 5.74439108e-01
9.69455540e-01 -4.15286660e-01 -5.49218178e-01 5.63085914e-01
8.51667583e-01 9.40670744e-02 -6.29392028e-01 4.14001614e-01
2.28705078e-01 -1.01262236e+00 9.12812173e-01 -6.96593761e-01
7.04419076e-01 -5.95368266e-01 -7.96418786e-02 -1.38160944e+00
-6.38884664e-01 -4.45266336e-01 -2.56427765e-01 1.20613909e+00
1.69420078e-01 -4.05664206e-01 2.51072407e-01 5.28977811e-01
1.10677361e-01 2.13624313e-01 -1.04598284e+00 -8.14728081e-01
-7.88142905e-02 -7.87181377e-01 7.21864760e-01 1.17666638e+00
-5.22754014e-01 7.78335810e-01 -4.46005821e-01 1.03580058e+00
9.38053250e-01 3.33756626e-01 7.58597553e-01 -1.65892351e+00
-3.26252848e-01 -7.93300495e-02 -2.55805790e-01 -5.11583030e-01
4.28928792e-01 -2.32341424e-01 2.61267334e-01 -1.54648101e+00
2.47517928e-01 -4.35836017e-01 -1.65585160e-01 3.43801528e-01
-2.93310404e-01 -3.61823350e-01 -2.13480607e-01 4.38243330e-01
-2.61573106e-01 5.54136992e-01 1.00818813e+00 -1.50414929e-01
-5.26774712e-02 1.99353650e-01 -7.37566873e-02 6.93571270e-01
5.87490737e-01 -5.01654029e-01 -6.23927057e-01 -2.39550456e-01
2.64502972e-01 6.99921608e-01 1.14929008e+00 -8.75986576e-01
5.22760786e-02 -6.43989682e-01 2.00651288e-01 -7.07830846e-01
4.10895318e-01 -9.73192632e-01 7.08028793e-01 4.20995742e-01
-4.23372030e-01 -1.42283872e-01 1.12585977e-01 1.04646051e+00
-2.12907910e-01 1.74170643e-01 3.21795762e-01 -3.72239214e-04
-7.33344257e-01 3.04108292e-01 -6.18609846e-01 -5.29147506e-01
1.11699629e+00 4.59773034e-01 -9.12300870e-02 -5.40905774e-01
-1.16060781e+00 4.56610210e-02 -4.72109839e-02 5.25695026e-01
2.96263576e-01 -1.12230611e+00 -8.89028788e-01 -2.02143505e-01
1.83252662e-01 3.29944305e-03 4.42811549e-01 8.11840951e-01
1.96941033e-01 7.10518539e-01 8.72959718e-02 -9.42923367e-01
-1.15651917e+00 9.90359247e-01 9.52764004e-02 -2.14011267e-01
-4.43484455e-01 5.38564503e-01 6.60781562e-01 -6.00229725e-02
-2.22326294e-01 -4.57787395e-01 -3.41388732e-01 7.12690204e-02
3.99546981e-01 1.69311658e-01 -2.49042854e-01 -6.77226722e-01
-1.49214879e-01 1.16101466e-01 3.78572643e-01 -3.66277814e-01
1.33736932e+00 -1.79548293e-01 -1.95565954e-01 1.40981889e+00
8.82159829e-01 -3.10448170e-01 -1.60392547e+00 -2.30650350e-01
-1.14085034e-01 -3.95374179e-01 1.69225082e-01 -8.76380980e-01
-7.03713119e-01 8.72876227e-01 5.80553532e-01 8.03894043e-01
8.93669903e-01 4.55146044e-01 9.40096825e-02 -1.72462925e-01
5.14505625e-01 -3.11453849e-01 -1.64773434e-01 2.64065653e-01
1.13432288e+00 -1.24966800e+00 -1.30115822e-01 -4.88748580e-01
-7.07199797e-02 9.59277868e-01 1.36990383e-01 -2.42469758e-01
8.07754040e-01 7.21766278e-02 -4.12426621e-01 -4.09419537e-01
-1.15217745e+00 -4.22445714e-01 2.43840277e-01 4.39406365e-01
3.83183330e-01 5.63000679e-01 1.14528850e-01 -2.96981484e-02
-1.07585348e-01 -1.24718621e-01 5.58582067e-01 5.85959852e-01
3.05523779e-02 -8.06296229e-01 -8.33528697e-01 2.21677080e-01
-2.17976317e-01 -6.56806529e-02 -2.66806066e-01 7.69571126e-01
2.13548258e-01 1.28273332e+00 3.88397843e-01 9.26637053e-02
2.75794566e-01 -1.46130309e-01 4.37271535e-01 -5.81564188e-01
1.34385601e-01 1.73483714e-01 1.99463014e-02 -6.07910156e-01
-7.21733391e-01 -1.42201161e+00 -1.09458017e+00 -2.44179785e-01
-4.64744121e-02 -1.77449867e-01 7.31365025e-01 1.33674383e+00
9.61616412e-02 1.11211038e+00 4.93265837e-01 -7.01095223e-01
4.78068367e-02 -9.19718623e-01 -5.86271703e-01 5.27882695e-01
9.12320077e-01 -1.08053863e+00 -4.30299222e-01 6.58528209e-01] | [7.844395160675049, 5.057373046875] |
771b5c9b-473b-4300-8f8b-e461fa7a718a | uncovering-and-categorizing-social-biases-in | 2305.16253 | null | https://arxiv.org/abs/2305.16253v2 | https://arxiv.org/pdf/2305.16253v2.pdf | Uncovering and Categorizing Social Biases in Text-to-SQL | Content Warning: This work contains examples that potentially implicate stereotypes, associations, and other harms that could be offensive to individuals in certain social groups.} Large pre-trained language models are acknowledged to carry social biases towards different demographics, which can further amplify existing stereotypes in our society and cause even more harm. Text-to-SQL is an important task, models of which are mainly adopted by administrative industries, where unfair decisions may lead to catastrophic consequences. However, existing Text-to-SQL models are trained on clean, neutral datasets, such as Spider and WikiSQL. This, to some extent, cover up social bias in models under ideal conditions, which nevertheless may emerge in real application scenarios. In this work, we aim to uncover and categorize social biases in Text-to-SQL models. We summarize the categories of social biases that may occur in structured data for Text-to-SQL models. We build test benchmarks and reveal that models with similar task accuracy can contain social biases at very different rates. We show how to take advantage of our methodology to uncover and assess social biases in the downstream Text-to-SQL task. We will release our code and data. | ['Jian-Guang Lou', 'Elliott Ash', 'Xiaokang Chen', 'Zhe Su', 'Yan Gao', 'Yan Liu'] | 2023-05-25 | null | null | null | null | ['text-to-sql'] | ['computer-code'] | [-1.97361261e-02 5.07210732e-01 -2.53215164e-01 -6.96775436e-01
-2.00155675e-01 -5.48217118e-01 8.23402524e-01 7.42986917e-01
-3.81210089e-01 6.98386014e-01 7.68935919e-01 -4.79692400e-01
-3.56333554e-02 -9.32226062e-01 -6.67678416e-01 -2.18215317e-01
-1.23654090e-01 3.48655730e-01 -2.15028211e-01 -8.01241636e-01
6.22981846e-01 1.77469760e-01 -1.35631084e+00 5.61010957e-01
1.04599583e+00 3.89710337e-01 -7.28980064e-01 3.89507264e-01
-1.76486865e-01 1.13180447e+00 -8.30568373e-01 -1.42121470e+00
2.41824791e-01 3.26237530e-01 -7.42668390e-01 -5.29465556e-01
9.12056804e-01 -2.33412132e-01 -6.31787404e-02 1.45232224e+00
8.67306411e-01 -4.28087413e-01 7.22232878e-01 -1.39700174e+00
-8.85094941e-01 1.11084116e+00 -5.87851584e-01 1.66458324e-01
5.51326215e-01 2.71500051e-01 9.95879710e-01 -5.89393795e-01
6.04745984e-01 1.92960250e+00 8.86428475e-01 4.88537580e-01
-1.44192314e+00 -1.01595676e+00 2.39615351e-01 -3.81031871e-01
-8.97924840e-01 -5.58738410e-01 5.21472514e-01 -6.77181065e-01
7.29359448e-01 8.29250157e-01 3.90608788e-01 1.90235209e+00
4.43322837e-01 5.81905782e-01 1.39135396e+00 1.23267420e-01
-1.31534250e-03 4.99792665e-01 4.80976015e-01 3.96053672e-01
8.64537537e-01 2.32181475e-01 -7.70392239e-01 -9.56994832e-01
-2.57316858e-01 5.63531928e-03 1.73041105e-01 -1.79809257e-01
-1.03256750e+00 1.15453649e+00 4.43262041e-01 -6.03548363e-02
-1.24775663e-01 -1.63761191e-02 9.17677820e-01 4.19728160e-01
8.96953821e-01 7.85238087e-01 -1.99968591e-01 -2.29006428e-02
-5.42365968e-01 7.35253990e-01 1.00718987e+00 8.32525730e-01
4.38773930e-01 -2.54874229e-01 -3.95165265e-01 9.55195367e-01
6.42651692e-02 7.81275272e-01 3.47529888e-01 -9.09037411e-01
7.05777884e-01 6.77732766e-01 4.31864373e-02 -1.60846579e+00
-3.77561390e-01 -6.17686033e-01 -8.75843227e-01 -7.93955475e-02
4.42433476e-01 -2.80255973e-01 -1.66403353e-01 1.76771402e+00
7.89846852e-02 -5.14852524e-01 -1.69050306e-01 4.16005135e-01
6.98009253e-01 -2.78676860e-02 6.08500361e-01 2.85325199e-01
1.22153521e+00 -1.60500288e-01 -6.81806207e-01 -6.95667803e-01
1.22543406e+00 -8.04979622e-01 1.38104117e+00 1.29113644e-01
-1.15504086e+00 7.80221447e-02 -7.09034741e-01 7.76685923e-02
-8.16390157e-01 -4.63822424e-01 6.02995455e-01 1.19907963e+00
-6.85554385e-01 8.15422595e-01 7.35593364e-02 -6.77014828e-01
8.36522996e-01 2.93556694e-02 -1.01868466e-01 2.34923661e-01
-1.68936169e+00 1.01919150e+00 -1.42291524e-02 -4.13620085e-01
-4.66770053e-01 -1.01913095e+00 -6.23639584e-01 -8.09925422e-02
5.29457211e-01 -5.78753352e-01 8.35505843e-01 -8.18060875e-01
-4.90850002e-01 1.40708625e+00 1.02861643e-01 -4.93100673e-01
9.97971416e-01 -4.03083771e-01 -4.97293860e-01 -5.43707430e-01
4.08651084e-01 3.13772321e-01 6.81024134e-01 -1.26854861e+00
-1.68295816e-01 -5.97881079e-01 2.95868129e-01 -7.15852976e-02
-8.31120968e-01 5.77960908e-01 4.69623327e-01 -8.35162878e-01
-6.96741700e-01 -9.61339295e-01 -3.89076740e-01 -2.63805509e-01
-9.92742300e-01 -1.48162917e-05 2.63999015e-01 -4.92632508e-01
1.56652224e+00 -1.91693962e+00 -3.77248585e-01 2.91500449e-01
6.25397623e-01 2.95169741e-01 -4.73615676e-02 6.03872776e-01
-1.76296860e-01 1.01415277e+00 3.00113857e-02 -1.69004709e-01
1.50610998e-01 -1.57614127e-01 -6.73313916e-01 4.28977221e-01
6.68986738e-02 7.75377035e-01 -1.01245689e+00 -5.85343838e-01
-2.36603826e-01 -6.55822307e-02 -9.74783003e-01 -2.78322753e-02
1.03048280e-01 -2.18261063e-01 -2.96427727e-01 6.18052661e-01
7.23662853e-01 4.35630381e-02 7.25433901e-02 8.94420147e-02
-6.91607371e-02 5.99924088e-01 -5.88466763e-01 6.64785445e-01
-1.02712132e-01 5.37038803e-01 5.60827181e-02 -6.97383463e-01
8.84362817e-01 -4.43042219e-01 1.26298696e-01 -6.64532542e-01
1.31864220e-01 1.19297005e-01 1.73422918e-01 -7.26134777e-01
7.20943213e-01 -2.35704556e-01 -4.28251028e-01 5.12639105e-01
-6.01593018e-01 -5.16562387e-02 -7.40560330e-03 6.83976948e-01
1.03267801e+00 -7.44728267e-01 2.93467194e-01 -9.09637809e-01
3.35903227e-01 -1.84955612e-01 3.40884298e-01 1.10584044e+00
-4.17026728e-01 1.47983506e-01 1.09891689e+00 -4.97371167e-01
-9.22888458e-01 -9.67415929e-01 -1.67301849e-01 1.50645590e+00
-1.12799987e-01 -6.42635763e-01 -5.14330328e-01 -1.16937733e+00
6.53415680e-01 1.19142127e+00 -8.02504778e-01 -5.11565447e-01
-3.81951816e-02 -1.08408761e+00 9.50933397e-01 -5.05002029e-02
2.39864528e-01 -7.42020369e-01 -3.70685220e-01 -1.89621687e-01
-1.05866171e-01 -8.36525619e-01 -3.26493829e-01 -9.46776420e-02
-6.78298771e-01 -1.18312979e+00 -3.93975288e-01 -6.57515377e-02
5.28111160e-01 5.81905991e-02 1.64355767e+00 3.57059270e-01
-1.21570297e-01 1.68717206e-01 -1.55313769e-02 -1.22213125e+00
-7.16035783e-01 2.56313562e-01 2.12780550e-01 -1.57497302e-01
8.22998047e-01 -3.77314180e-01 -4.56712335e-01 4.92541134e-01
-9.30067122e-01 -1.07937798e-01 2.42290035e-01 3.39699626e-01
-5.46185851e-01 -4.62081790e-01 8.03318501e-01 -1.73732340e+00
1.32604074e+00 -9.69785690e-01 -1.50245903e-02 9.56910774e-02
-7.94372082e-01 -1.35197148e-01 5.12742758e-01 -3.69281858e-01
-8.33014607e-01 -7.29449272e-01 3.66465673e-02 2.92133212e-01
-4.97557260e-02 4.59349066e-01 -5.99892810e-02 9.06292349e-02
1.36368263e+00 -3.29287559e-01 1.67839751e-01 -2.59190023e-01
1.68157801e-01 9.77235258e-01 -1.48476899e-01 -6.70257866e-01
1.03786361e+00 6.65921569e-01 -2.20828086e-01 -8.22910368e-01
-1.08109713e+00 -9.52162817e-02 3.02392375e-02 -2.42329314e-01
4.34711069e-01 -8.63258243e-01 -9.04462516e-01 3.99137050e-01
-9.12400961e-01 -4.03667808e-01 -4.45047468e-02 -9.39903855e-02
2.44567692e-02 4.89508182e-01 -5.29502809e-01 -8.20918858e-01
-2.14997530e-01 -6.85339391e-01 7.82780409e-01 -1.80210471e-01
-8.79108310e-01 -1.11669123e+00 -1.38350904e-01 5.04688263e-01
7.70292342e-01 3.13651443e-01 1.18314397e+00 -1.18115020e+00
7.64790252e-02 -2.40195915e-01 -3.60804856e-01 3.01080197e-01
-1.15722172e-01 1.69290975e-01 -9.50847745e-01 -2.91170478e-01
-1.56533167e-01 -6.11831784e-01 7.21684098e-01 -1.67886205e-02
1.41969621e+00 -7.75061071e-01 -3.59721243e-01 1.38430327e-01
9.70518649e-01 -3.76379937e-01 6.59023821e-01 3.20977986e-01
6.27201438e-01 1.47585678e+00 7.24274039e-01 6.18947744e-01
4.71979707e-01 2.78085113e-01 6.48157060e-01 -1.33291647e-01
4.34035569e-01 -3.74300987e-01 6.15608633e-01 4.73911107e-01
2.27061808e-01 -1.73578724e-01 -1.19417346e+00 4.95267272e-01
-1.52684462e+00 -1.14076555e+00 -4.36717093e-01 2.25355005e+00
8.78207386e-01 4.49451774e-01 1.69699252e-01 -1.79306060e-01
8.53261769e-01 6.28450751e-01 -4.93468851e-01 -8.94325852e-01
-3.09266150e-01 -4.06520873e-01 7.73203850e-01 4.79933172e-01
-9.35361087e-01 8.03924799e-01 6.90730572e+00 5.55703878e-01
-1.02397645e+00 -1.56453043e-01 1.10156715e+00 -5.18752038e-01
-8.70818853e-01 -1.67159155e-01 -7.95972466e-01 6.96156204e-01
8.99309278e-01 -6.27156615e-01 1.76192164e-01 9.71262753e-01
2.79443324e-01 2.09786296e-01 -1.27179062e+00 6.91365063e-01
1.99617594e-01 -1.06498659e+00 3.33780944e-01 4.31977183e-01
6.94498658e-01 1.46424860e-01 3.27799290e-01 5.61457753e-01
8.14294398e-01 -1.19867039e+00 6.66109264e-01 3.69171828e-01
4.50614840e-01 -6.64370179e-01 6.56855285e-01 4.81252402e-01
-5.10510243e-02 -2.73053944e-01 -6.42098188e-01 -3.40038270e-01
-1.38179719e-01 1.17960930e+00 -6.15353048e-01 6.30566627e-02
9.99099612e-01 8.17981720e-01 -1.21938944e+00 4.40876782e-01
2.61250347e-01 5.31692743e-01 2.91212890e-02 -4.85480487e-01
-5.27174622e-02 3.58842835e-02 7.46292889e-01 1.35620272e+00
1.12161465e-01 -3.75434339e-01 -1.86370984e-01 9.14286852e-01
-2.37124309e-01 3.24271470e-01 -1.31135786e+00 -2.06994697e-01
3.41235191e-01 1.04949749e+00 3.58234085e-02 -4.70731437e-01
-3.48980486e-01 3.17594618e-01 4.03824478e-01 2.47974619e-01
-7.47246742e-01 -2.93718755e-01 9.92750406e-01 7.74162710e-01
-8.13799560e-01 2.57128894e-01 -5.41547298e-01 -1.28064883e+00
-1.30673811e-01 -1.54354823e+00 4.43414897e-01 -4.40539688e-01
-1.96019590e+00 1.00957729e-01 -1.59926858e-05 -7.27858186e-01
-2.54042279e-02 -5.29944003e-01 -5.75393379e-01 7.15883672e-01
-1.36143291e+00 -7.33412385e-01 -1.92554861e-01 3.14484507e-01
1.54633686e-01 -4.30007488e-01 5.83600521e-01 2.94241130e-01
-5.41452646e-01 7.74601519e-01 -2.58426785e-01 1.99836046e-01
1.48640621e+00 -1.28103447e+00 6.49312317e-01 4.37945604e-01
-5.96759379e-01 1.19913185e+00 1.02926707e+00 -1.03466773e+00
-8.67462158e-01 -1.30875790e+00 1.36228454e+00 -1.08472478e+00
9.77690279e-01 -6.27923310e-01 -7.87234783e-01 8.34657729e-01
1.73261851e-01 -4.08157855e-01 1.16345918e+00 6.04402542e-01
-7.37284958e-01 -6.80070445e-02 -1.16780770e+00 1.29012620e+00
1.47067678e+00 -6.77873671e-01 -5.75944841e-01 6.22372270e-01
5.86037099e-01 2.61935331e-02 -6.68929875e-01 2.51008630e-01
6.23892426e-01 -1.62727511e+00 1.13497686e+00 -1.35424471e+00
1.00913286e+00 5.18503308e-01 -1.57402962e-01 -1.47356403e+00
-2.59269685e-01 -8.02304089e-01 3.29340309e-01 1.45854712e+00
5.40551126e-01 -1.18949640e+00 6.45133674e-01 1.12952888e+00
3.20415616e-01 -5.15764832e-01 -4.31217492e-01 -6.66487455e-01
6.55248284e-01 -3.93640965e-01 8.66655588e-01 1.48847592e+00
1.07888438e-01 1.93569720e-01 -3.64741951e-01 -8.03459138e-02
1.00741708e+00 -3.68906021e-01 1.02719736e+00 -1.48063850e+00
1.89262986e-01 -6.08539402e-01 -9.44218412e-02 -3.96873951e-01
4.18760568e-01 -1.00857544e+00 -5.78337252e-01 -8.63949120e-01
4.81008619e-01 -6.21332526e-01 6.05635196e-02 1.35676280e-01
-4.55936998e-01 1.54396936e-01 3.07212055e-01 -6.34177253e-02
-4.37004089e-01 2.96698838e-01 8.85553539e-01 -3.49715382e-01
1.50987193e-01 -1.06558412e-01 -1.68156326e+00 1.03978682e+00
9.75071192e-01 -7.45833278e-01 -8.44699740e-02 -3.26850235e-01
1.23831367e+00 -6.69802308e-01 6.71244621e-01 -5.52785337e-01
-2.18031153e-01 -4.40980226e-01 1.16212510e-01 -5.82444668e-02
-2.61904597e-02 -6.43765032e-01 -2.27784112e-01 8.13757539e-01
-1.03333306e+00 1.38796747e-01 -9.75717604e-02 4.01167125e-01
7.54941553e-02 -1.52174130e-01 5.79606712e-01 -2.11076468e-01
3.28897983e-01 4.75347787e-02 -4.43265259e-01 6.17865801e-01
8.51071179e-01 2.89852262e-01 -1.17206955e+00 -7.72812843e-01
-3.11009437e-01 4.37523693e-01 4.84186441e-01 7.29962230e-01
1.57443210e-01 -1.34163010e+00 -1.07655501e+00 -1.41031876e-01
5.62915742e-01 -6.30993009e-01 3.03732660e-02 8.41340125e-01
-3.32071006e-01 1.92959070e-01 -1.20438978e-01 -2.24839985e-01
-1.46106410e+00 6.11319840e-01 1.68847531e-01 -2.75834471e-01
1.20916635e-01 5.53822100e-01 4.26676989e-01 -8.33210766e-01
1.71428427e-01 -1.46776363e-01 -2.12652072e-01 4.78524417e-01
4.52707529e-01 7.61609256e-01 -1.92251548e-01 -4.92933303e-01
-4.34155077e-01 1.61499858e-01 -3.14874947e-01 3.45302194e-01
1.03523076e+00 -1.21085286e-01 -4.45397496e-01 3.80435497e-01
1.11103547e+00 6.99658930e-01 -3.50993335e-01 -2.36315783e-02
1.98453918e-01 -7.73008883e-01 -4.63066131e-01 -8.97214174e-01
-7.39069700e-01 9.66967702e-01 1.45994648e-01 7.15094268e-01
3.60761672e-01 -2.00055614e-01 4.54643399e-01 4.00597781e-01
1.50731474e-01 -1.39744985e+00 1.19700998e-01 3.85700911e-01
1.22016931e+00 -1.43490863e+00 2.21517801e-01 -6.54915392e-01
-8.24232459e-01 5.81004202e-01 6.60774529e-01 -1.18640758e-01
7.72815168e-01 1.84881613e-01 1.91543773e-01 -3.45324039e-01
-8.42458248e-01 2.26912603e-01 -8.84020776e-02 5.92763066e-01
7.33810723e-01 2.42583737e-01 -6.52336717e-01 6.08676255e-01
-5.77157736e-01 -4.53514338e-01 8.13401639e-01 5.74110270e-01
-1.36850476e-01 -1.01111019e+00 -4.98558640e-01 9.49208856e-01
-9.59470391e-01 -3.26567590e-01 -1.14954841e+00 5.92447460e-01
-1.53202973e-02 9.71589148e-01 -9.60133895e-02 -4.92704868e-01
4.80718523e-01 1.65784791e-01 -2.28548691e-01 -5.56837738e-01
-1.40311766e+00 -5.47598779e-01 8.31935287e-01 -7.88287401e-01
1.11906715e-02 -7.46750176e-01 -5.45998752e-01 -1.06856835e+00
1.14699833e-01 -2.11147621e-01 2.93308109e-01 4.10358012e-01
5.91070056e-01 1.69785172e-01 5.11514366e-01 -3.37878734e-01
-1.03961694e+00 -9.20510590e-01 -4.24914986e-01 1.10890853e+00
1.51120439e-01 -3.64796966e-01 -7.32132077e-01 -4.81423914e-01] | [9.111984252929688, 10.214293479919434] |
415d2ebd-6451-4e15-8eac-6765b86d95a0 | autobots-lt-edi-eacl2021-one-world-one-family | null | null | https://aclanthology.org/2021.ltedi-1.21 | https://aclanthology.org/2021.ltedi-1.21.pdf | Autobots@LT-EDI-EACL2021: One World, One Family: Hope Speech Detection with BERT Transformer Model | The rapid rise of online social networks like YouTube, Facebook, Twitter allows people to express their views more widely online. However, at the same time, it can lead to an increase in conflict and hatred among consumers in the form of freedom of speech. Therefore, it is essential to take a positive strengthening method to research on encouraging, positive, helping, and supportive social media content. In this paper, we describe a Transformer-based BERT model for Hope speech detection for equality, diversity, and inclusion, submitted for LT-EDI-2021 Task 2. Our model achieves a weighted averaged f1-score of 0.93 on the test set. | ['Radhika Mamidi', 'Sunil Gundapu'] | null | null | null | null | eacl-ltedi-2021-4 | ['hope-speech-detection'] | ['natural-language-processing'] | [-5.85325897e-01 4.46673125e-01 -7.20601022e-01 -1.95825681e-01
-5.35598755e-01 -3.59565467e-01 7.53869355e-01 2.28184074e-01
-2.40500420e-01 5.67322731e-01 1.02724671e+00 -2.63620287e-01
1.56264186e-01 -3.97059411e-01 1.30583212e-01 -1.97464079e-01
2.97584742e-01 -1.87977210e-01 -3.10301390e-02 -5.01502454e-01
1.23578236e-01 1.13961063e-01 -9.74640787e-01 3.51237655e-01
1.03271616e+00 5.63180745e-01 -3.16723764e-01 1.58810958e-01
-1.64049253e-01 1.21607900e+00 -7.20621407e-01 -9.96165156e-01
-2.87589431e-01 -3.91666174e-01 -6.24651790e-01 -1.78835377e-01
1.48470595e-01 -5.57298779e-01 -4.29291964e-01 1.15324366e+00
7.17683434e-01 1.31501886e-03 3.39090556e-01 -1.39357328e+00
-8.92657638e-01 1.19803178e+00 -4.22288775e-01 1.55968487e-01
6.48387969e-01 7.23060742e-02 1.09477699e+00 -8.14110518e-01
1.03943729e+00 1.37060654e+00 4.02761817e-01 5.00200987e-01
-9.17861581e-01 -8.81975710e-01 -1.77147314e-01 4.52449434e-02
-1.07103992e+00 -7.70325959e-01 7.97253728e-01 -2.85840362e-01
8.30359399e-01 5.63719094e-01 1.25655603e+00 1.57944763e+00
-3.92331021e-05 9.78199601e-01 1.05901515e+00 -8.40084255e-02
6.65571094e-02 4.83039588e-01 -1.65555447e-01 3.08456510e-01
-3.46157551e-02 -2.20878229e-01 -8.08678508e-01 -9.04867351e-02
-7.45995343e-02 -1.13745011e-01 -3.14605832e-01 3.27364028e-01
-9.16307807e-01 1.16469657e+00 4.84634131e-01 6.10823929e-01
-1.29176334e-01 -6.35373056e-01 5.49156189e-01 4.19791371e-01
1.11226201e+00 3.19936067e-01 3.20294499e-01 -7.67549098e-01
-7.57363439e-01 1.88552454e-01 8.55246186e-01 2.80706853e-01
-1.05582811e-01 7.23796431e-03 -3.27387035e-01 1.10362077e+00
5.47985792e-01 4.57068771e-01 4.83635217e-01 -8.47712398e-01
3.03366363e-01 5.80258727e-01 -3.08318198e-01 -1.42449045e+00
-3.73924196e-01 -7.83164978e-01 -8.91500056e-01 -2.59175718e-01
5.31181134e-02 -2.05202878e-01 -1.91781461e-01 1.68519485e+00
2.34594554e-01 -3.58883709e-01 -8.64739716e-02 6.46562338e-01
1.06189072e+00 7.18126595e-01 1.24535471e-01 -4.90957081e-01
9.16747808e-01 -7.68603444e-01 -1.37100995e+00 -6.62233606e-02
1.02548718e+00 -1.04992318e+00 1.02757406e+00 2.79465765e-01
-1.00489557e+00 1.40672415e-01 -7.81992555e-01 -2.12506533e-01
-2.36530140e-01 -3.64595950e-01 3.32126647e-01 1.03859413e+00
-9.59549308e-01 4.98295754e-01 -1.91045180e-01 -5.68696797e-01
7.27675498e-01 -3.16911310e-01 -3.39323670e-01 4.94665578e-02
-1.31249368e+00 9.72376406e-01 -7.53535777e-02 -2.02780411e-01
-2.08216459e-01 -5.02532840e-01 -5.91625750e-01 -1.33329749e-01
9.06730518e-02 -1.16165698e-01 9.52252865e-01 -9.84413207e-01
-1.27572978e+00 1.05457783e+00 1.17738247e-01 -2.80572116e-01
6.06521904e-01 -3.25305462e-01 -7.71659851e-01 -1.80244297e-01
5.08100353e-02 6.97337210e-01 5.69402039e-01 -6.18198037e-01
-2.36607790e-01 -3.80408704e-01 3.35879698e-02 4.62285466e-02
-1.01048732e+00 7.05830097e-01 4.34829086e-01 -4.69908983e-01
-2.02955559e-01 -6.92444384e-01 2.01432288e-01 -1.78465441e-01
-7.35495687e-01 -5.72908103e-01 1.14498854e+00 -1.00447083e+00
1.66516161e+00 -2.65100741e+00 -2.25130618e-02 -5.84953949e-02
7.26895452e-01 4.37463880e-01 1.81531936e-01 5.57902455e-01
1.49507165e-01 9.25832212e-01 3.22815001e-01 -3.46123308e-01
7.40790591e-02 -4.59874496e-02 -1.80148542e-01 4.28123921e-01
5.18398993e-02 7.16027498e-01 -1.00622654e+00 -4.06056613e-01
-1.14796162e-01 6.59303546e-01 -7.12885559e-01 -4.52210531e-02
3.48094672e-01 1.42824501e-01 -1.64374620e-01 4.86023068e-01
4.66245413e-01 -1.76100165e-01 1.43594757e-01 2.97929406e-01
-5.66509783e-01 8.82491767e-01 -4.33709294e-01 9.29737329e-01
-1.49165029e-02 1.14303255e+00 1.34312943e-01 -5.22657275e-01
9.18174207e-01 2.74909079e-01 3.37090552e-01 -1.02804613e+00
3.44400585e-01 1.49541214e-01 2.10404947e-01 -4.25993502e-01
6.27928436e-01 -7.79524595e-02 -9.77264997e-03 3.15091670e-01
-3.80703837e-01 -9.48948413e-02 -4.55488153e-02 7.69680023e-01
7.41464317e-01 -5.42085350e-01 4.09356862e-01 -2.04309046e-01
1.02555379e-02 -5.27058363e-01 4.14086729e-01 1.18337929e-01
-6.16379499e-01 1.13622710e-01 8.89853299e-01 3.95642705e-02
-9.06637490e-01 -5.75125277e-01 -1.19606189e-01 1.01251543e+00
-3.31269622e-01 -8.68342280e-01 -6.30813479e-01 -4.52252239e-01
-5.74332848e-02 1.10192013e+00 -9.17949826e-02 -4.32769209e-01
-1.77388579e-01 -2.82608867e-01 6.41818106e-01 -2.58788556e-01
7.64315188e-01 -8.79522502e-01 -1.21179648e-01 1.23824375e-02
-7.90344238e-01 -1.05890822e+00 -4.20081019e-01 -4.46420968e-01
-4.62223053e-01 -5.55555522e-01 -7.58742511e-01 -4.80357409e-01
2.91359276e-02 5.30783057e-01 7.90103197e-01 2.21691832e-01
2.24988595e-01 1.50059946e-02 -5.64571619e-01 -3.84756714e-01
-5.24958253e-01 1.54047966e-01 -3.23656872e-02 2.57785115e-02
1.45776540e-01 -4.50314164e-01 -1.87884495e-01 2.03254282e-01
-5.71016788e-01 1.85789093e-01 -2.30864048e-01 4.41666126e-01
-2.54857808e-01 -3.27409744e-01 8.96387935e-01 -5.15640259e-01
1.05693781e+00 -8.15479815e-01 1.01133630e-01 -1.96539402e-01
-5.91240227e-01 -6.69428349e-01 2.84563214e-01 -3.65716517e-01
-8.03548515e-01 -6.50815606e-01 -4.32512343e-01 5.41956723e-02
2.83932596e-01 6.02702975e-01 7.70233804e-03 1.83581203e-01
6.79620683e-01 -1.94430515e-01 2.36314237e-01 -3.73400003e-01
3.56858969e-01 1.29942918e+00 -7.47718811e-02 2.71179587e-01
5.54732263e-01 7.28200153e-02 -5.40201962e-01 -1.38066208e+00
-8.80303681e-01 -3.13268393e-01 1.63755223e-01 -6.79447055e-01
7.99011528e-01 -1.04829121e+00 -7.34270811e-01 5.21455169e-01
-1.08291304e+00 -1.77745059e-01 -2.01704979e-01 4.59615141e-01
1.46183193e-01 3.88400853e-01 -5.64200580e-01 -1.15405905e+00
-3.37041944e-01 -5.86242676e-01 2.51670659e-01 2.51015663e-01
-8.13146412e-01 -8.54333401e-01 4.86931466e-02 7.74203718e-01
8.06826651e-01 3.25234711e-01 6.20961487e-01 -9.19364572e-01
-3.48008201e-02 -1.00983940e-01 -1.09581962e-01 4.82778996e-01
1.12174295e-01 3.61260593e-01 -8.57500434e-01 -2.21816618e-02
-6.05347827e-02 -7.64839888e-01 4.76050138e-01 1.40910789e-01
8.06364596e-01 -9.41054821e-01 -8.77330229e-02 -3.98749672e-02
5.79057217e-01 1.32966757e-01 8.86679530e-01 2.15249181e-01
2.60184973e-01 6.63271904e-01 4.16442245e-01 9.00883138e-01
4.45160151e-01 6.18298769e-01 4.18852359e-01 6.89859614e-02
-5.33744022e-02 -5.74932635e-01 7.18014657e-01 1.17029011e+00
7.35722408e-02 -4.52011734e-01 -7.94871807e-01 3.23635370e-01
-1.49878764e+00 -1.49866509e+00 -3.30724657e-01 1.98591018e+00
7.61061966e-01 2.63701767e-01 6.48557305e-01 3.61523479e-01
6.92619920e-01 5.91703832e-01 1.57016218e-02 -7.13630915e-01
-3.98952723e-01 -3.80931139e-01 -6.12721704e-02 5.91428518e-01
-5.93364537e-01 7.36336410e-01 6.07812405e+00 7.89369464e-01
-1.28448689e+00 3.73592973e-01 1.16292703e+00 -5.16634703e-01
-7.57112741e-01 -2.39643574e-01 -5.73273957e-01 6.15954041e-01
1.03662813e+00 -7.16539741e-01 4.79402542e-01 7.44847059e-01
5.38646281e-01 3.58769181e-03 -4.43472981e-01 1.00992858e+00
1.86178565e-01 -1.13818455e+00 -5.14868736e-01 2.53527403e-01
5.46731532e-01 7.62736276e-02 4.22992587e-01 3.81114513e-01
-2.39183716e-02 -9.11127269e-01 7.39956379e-01 3.68157140e-04
6.67752028e-01 -6.75648212e-01 4.65121090e-01 6.77257240e-01
-3.63190711e-01 -6.74642995e-02 -6.67929426e-02 -4.34953243e-01
3.20193559e-01 7.85192788e-01 -6.89345479e-01 -1.44475615e-02
4.91723746e-01 9.14320350e-01 -4.29709822e-01 9.37962472e-01
-3.44645977e-01 9.79068458e-01 -2.63089955e-01 -5.98119259e-01
1.50154591e-01 3.81327495e-02 6.52224600e-01 1.07692850e+00
2.93081939e-01 4.67305146e-02 -2.01126382e-01 5.47448933e-01
-5.04972637e-01 4.69948977e-01 -9.88698125e-01 -7.20869839e-01
5.29708326e-01 1.24924314e+00 -2.98585534e-01 -2.03808174e-02
-3.75938624e-01 5.60474634e-01 3.26288879e-01 -2.33017020e-02
-7.64074445e-01 -2.86723915e-02 3.83642405e-01 5.77480674e-01
-5.57287931e-01 -1.65644780e-01 -1.03898324e-01 -1.06769323e+00
-2.21738636e-01 -1.00952113e+00 1.68687608e-02 -6.16346955e-01
-1.21733963e+00 1.60259664e-01 -2.66852438e-01 -7.08661139e-01
-2.53730565e-01 2.29636654e-02 -6.23592436e-01 4.82265741e-01
-1.09783435e+00 -6.13857448e-01 4.19604592e-02 2.82533824e-01
4.03626710e-01 -1.78256303e-01 5.44589996e-01 4.73891407e-01
-5.91224432e-01 6.52913392e-01 3.91910896e-02 9.51506644e-02
8.44102979e-01 -5.41263878e-01 1.85178354e-01 3.74244630e-01
-4.88895364e-02 3.02339911e-01 7.34654129e-01 -6.13237500e-01
-1.04544413e+00 -5.78460097e-01 1.73184907e+00 -7.88305327e-02
9.07128632e-01 -3.30509990e-01 -6.43181264e-01 3.71016085e-01
6.14285529e-01 -5.96028090e-01 1.01771939e+00 5.05357563e-01
-4.12050903e-01 1.87392766e-03 -1.41484141e+00 7.97432482e-01
1.21562517e+00 -7.30172932e-01 -3.26526552e-01 7.42874920e-01
8.32798481e-01 2.22079139e-02 -8.01851690e-01 -2.24332035e-01
6.13964856e-01 -1.08537364e+00 6.46885633e-01 -5.99946856e-01
8.64152908e-01 6.29738092e-01 -8.74548256e-02 -1.39488375e+00
-4.29971248e-01 -1.11108077e+00 2.17621610e-01 1.64359748e+00
5.27619243e-01 -7.27899730e-01 4.48931307e-01 8.76677871e-01
-1.26760244e-01 -5.28410256e-01 -1.27726758e+00 -5.02050817e-01
6.50618225e-02 -2.60478377e-01 1.24040633e-01 1.47564447e+00
8.14621329e-01 7.44561851e-01 -6.91843331e-01 -5.83540201e-01
2.03577265e-01 -7.08690882e-01 6.39923751e-01 -1.26474047e+00
1.21475048e-01 -8.33596349e-01 -1.63551122e-01 -5.80122292e-01
4.21233028e-02 -9.40155923e-01 -7.16346145e-01 -1.43731022e+00
3.65962744e-01 -1.51307732e-01 1.77457884e-01 2.89849967e-01
3.67388189e-01 3.77668478e-02 5.64933836e-01 8.61325413e-02
-8.03868949e-01 6.51000798e-01 1.31634772e+00 -1.52051076e-01
-2.75545925e-01 -2.02153578e-01 -1.14047539e+00 4.33177650e-01
1.16861832e+00 -4.21412081e-01 -2.23690167e-01 1.06212176e-01
8.78168225e-01 1.73181087e-01 1.89397171e-01 -5.80559850e-01
-6.08357452e-02 -2.09413663e-01 -1.99752897e-01 -3.43307227e-01
4.37905669e-01 -4.08615321e-01 5.76077662e-02 6.49865687e-01
-5.93773305e-01 -1.10084601e-02 -1.24593355e-01 8.69756658e-03
-1.60633504e-01 -4.61805612e-02 8.02444816e-01 3.59566540e-01
2.66302109e-01 -1.49845168e-01 -6.07191205e-01 4.55037504e-01
6.50830328e-01 6.81830943e-02 -8.44429433e-01 -1.11106634e+00
-4.26070899e-01 7.71366879e-02 2.68495798e-01 7.94734538e-01
6.55976951e-01 -1.44760609e+00 -1.03869104e+00 -1.05840042e-01
-3.81295085e-02 -7.66885638e-01 3.37707698e-01 1.07476437e+00
4.69963066e-02 2.11207613e-01 -1.67572156e-01 -1.55412748e-01
-1.51425815e+00 1.44270346e-01 4.30220775e-02 -4.16070223e-02
-4.59934592e-01 7.95237422e-01 -3.69277894e-01 -2.18294650e-01
1.90492138e-01 2.32171014e-01 -4.27381277e-01 5.53914189e-01
8.11048388e-01 8.65044117e-01 -3.55953664e-01 -1.10734928e+00
-3.24444324e-01 -2.56539166e-01 -5.62337227e-02 -1.45576835e-01
1.20870531e+00 7.82084372e-03 -3.76638532e-01 4.45368111e-01
1.41695106e+00 4.30631191e-01 -2.67100453e-01 1.70749381e-01
-9.06245410e-03 -8.69849563e-01 1.86792776e-01 -8.44007909e-01
-8.29489350e-01 7.03675210e-01 3.70456964e-01 9.62540448e-01
5.64828575e-01 1.78141564e-01 7.40333259e-01 4.05946746e-02
-2.17635393e-01 -1.36977124e+00 2.97036588e-01 3.76568735e-01
1.17205048e+00 -1.19019186e+00 -1.92904323e-01 -4.78512466e-01
-8.35488439e-01 7.82357991e-01 2.52785236e-01 5.41059911e-01
7.24282563e-01 -2.03942787e-02 -1.10042304e-01 -3.61368922e-03
-8.50946367e-01 2.35538363e-01 -4.32673804e-02 5.16665399e-01
1.00456119e+00 5.00020146e-01 -8.63566935e-01 5.29718578e-01
-3.63444835e-01 -6.85556307e-02 6.81183815e-01 2.82581717e-01
-8.31249714e-01 -7.41805613e-01 -2.24564597e-02 3.24224681e-01
-8.76502752e-01 -1.04563730e-02 -8.62537086e-01 5.68033814e-01
-2.09978878e-01 1.35784030e+00 1.60241574e-02 -5.98284125e-01
-3.46447900e-02 1.57355428e-01 -2.02793673e-01 -1.93197280e-01
-8.08829129e-01 7.48840626e-03 8.79469156e-01 -4.28329736e-01
-3.76637965e-01 -8.17735136e-01 -8.81063759e-01 -1.04795897e+00
-2.85541445e-01 6.21219724e-03 7.88235009e-01 7.05237508e-01
5.18887162e-01 5.64442351e-02 8.43892992e-01 -8.44256729e-02
-5.55926859e-01 -1.05093789e+00 -4.00291324e-01 2.58282423e-01
6.23656027e-02 -2.86212653e-01 -6.62279427e-01 -7.48714387e-01] | [8.905678749084473, 10.622347831726074] |
5abe884a-a357-4e60-80ba-fb2cca7f9f42 | mapping-the-buried-cable-by-ground | 2201.11253 | null | https://arxiv.org/abs/2201.11253v1 | https://arxiv.org/pdf/2201.11253v1.pdf | Mapping the Buried Cable by Ground Penetrating Radar and Gaussian-Process Regression | With the rapid expansion of urban areas and the increasingly use of electricity, the need for locating buried cables is becoming urgent. In this paper, a noval method to locate underground cables based on Ground Penetrating Radar (GPR) and Gaussian-process regression is proposed. Firstly, the coordinate system of the detected area is conducted, and the input and output of locating buried cables are determined. The GPR is moved along the established parallel detection lines, and the hyperbolic signatures generated by buried cables are identified and fitted, thus the positions and depths of some points on the cable could be derived. On the basis of the established coordinate system and the derived points on the cable, the clustering method and cable fitting algorithm based on Gaussian-process regression are proposed to find the most likely locations of the underground cables. Furthermore, the confidence intervals of the cable's locations are also obtained. Both the position and depth noises are taken into account in our method, ensuring the robustness and feasibility in different environments and equipments. Experiments on real-world datasets are conducted, and the obtained results demonstrate the effectiveness of the proposed method. | ['Huanhuan Chen', 'Shengfei Lyu', 'Qiuju Chen', 'Xiren Zhou'] | 2022-01-25 | null | null | null | null | ['gpr', 'gpr'] | ['computer-vision', 'miscellaneous'] | [-1.89337507e-01 -4.35014814e-01 5.98661423e-01 -1.30926207e-01
-7.48156726e-01 -4.16889250e-01 -1.96077645e-01 6.69285804e-02
-2.51478255e-01 4.78625953e-01 -1.60975978e-01 -1.14134699e-01
-6.08812034e-01 -8.43605578e-01 -2.37644091e-01 -1.22853208e+00
-9.13645253e-02 2.76558101e-01 3.12392443e-01 3.37869555e-01
4.06887442e-01 5.23128271e-01 -9.24440205e-01 -2.83099234e-01
8.65607500e-01 1.23979831e+00 6.44097447e-01 -5.21797733e-03
2.06612870e-01 5.21752634e-04 -6.48580432e-01 1.00521995e-02
-1.63906664e-01 1.27403587e-02 9.68865491e-03 1.32638410e-01
-9.89814699e-01 -2.46989131e-01 -4.96412814e-01 1.23342824e+00
7.31724620e-01 5.88917099e-02 7.15729177e-01 -8.10711205e-01
-8.56601074e-02 5.35077751e-01 -1.14185059e+00 -3.09029538e-02
2.29878664e-01 -4.14912283e-01 3.38905245e-01 -9.29004431e-01
-3.62740345e-02 1.02551234e+00 1.05499029e+00 -2.29652122e-01
-7.53918052e-01 -7.34465182e-01 -8.90601128e-02 1.99226364e-01
-1.99883306e+00 9.96664688e-02 9.50888872e-01 -3.90179187e-01
2.36920044e-01 -1.54245645e-01 7.26695478e-01 7.45397508e-01
2.40216240e-01 4.67804104e-01 7.56467998e-01 -5.54754555e-01
2.63133109e-01 2.31693447e-01 8.99311230e-02 2.95684814e-01
5.32292068e-01 -1.57177746e-01 2.65297830e-01 -4.93341774e-01
7.57104933e-01 2.22745091e-01 -5.84381878e-01 -7.83799440e-02
-8.36581767e-01 6.73596799e-01 9.43617001e-02 4.24290299e-01
-4.99806881e-01 -2.64007866e-01 4.43610340e-01 -5.95723152e-01
3.51034850e-01 -1.95028976e-01 -1.55128583e-01 -1.12049721e-01
-9.22782779e-01 -1.53599694e-01 6.45404100e-01 1.19137466e+00
3.66579413e-01 -1.11230120e-01 2.91438282e-01 6.67973638e-01
6.36441886e-01 6.71287715e-01 2.51957383e-02 -3.42095971e-01
5.97800195e-01 6.11828314e-03 6.64125308e-02 -1.56039548e+00
-4.16147083e-01 -3.47964674e-01 -8.73816311e-01 -5.05010664e-01
6.50442624e-03 -8.57402146e-01 -3.33704919e-01 8.05857241e-01
1.47837311e-01 4.72142011e-01 3.94768417e-02 7.17355371e-01
3.33432376e-01 1.02527356e+00 -2.51731705e-02 -5.14576137e-01
1.25415146e+00 -7.21151829e-02 -8.70543480e-01 8.42556357e-02
3.63951892e-01 -6.41180098e-01 2.72697359e-01 7.59395361e-01
-5.47174990e-01 -3.56491566e-01 -9.85293269e-01 8.28701615e-01
4.36827451e-01 5.96937478e-01 5.27146518e-01 4.84105945e-01
-5.16283274e-01 4.62481320e-01 -9.22092915e-01 -1.88154578e-01
3.02771121e-01 -2.06650987e-01 1.45270869e-01 -1.95920959e-01
-1.10951328e+00 5.20451546e-01 4.22834516e-01 1.09014928e+00
-7.34636605e-01 -2.30497271e-01 -4.46802676e-01 7.48921558e-02
4.94107977e-02 6.01815917e-02 9.41403091e-01 -1.89867362e-01
-7.99389660e-01 -3.26158702e-02 6.21140562e-02 9.67612341e-02
4.82102603e-01 -2.32691187e-02 -7.74635017e-01 4.81522709e-01
2.80584902e-01 -5.23041904e-01 4.01067972e-01 -1.36409318e+00
-8.09802592e-01 -5.39559245e-01 -7.29532421e-01 -1.31938174e-01
-1.90218061e-01 1.45505086e-01 -3.48618776e-01 -7.09091648e-02
9.00877178e-01 -5.55927217e-01 -3.94641042e-01 -2.94682026e-01
-8.21144760e-01 -8.52876808e-03 8.63681376e-01 -1.24554610e+00
1.19668078e+00 -2.38010311e+00 -1.53144583e-01 1.06057978e+00
-9.92509425e-02 -2.50760287e-01 4.82549161e-01 5.04942954e-01
-5.73605709e-02 -1.68545201e-01 -4.59062755e-01 9.26627517e-02
-4.52200711e-01 -1.31304845e-01 -1.58418268e-01 8.00123870e-01
-2.93221921e-01 -1.71436146e-02 -7.00388610e-01 -5.29770195e-01
-6.40485957e-02 4.41072702e-01 6.78087845e-02 5.49586043e-02
3.23409587e-01 5.98770201e-01 -1.04731739e+00 6.95649207e-01
1.15955985e+00 3.73819113e-01 9.94692296e-02 -3.62490565e-01
-3.37238282e-01 -3.68505895e-01 -1.56961262e+00 1.28189921e+00
-2.26460800e-01 2.87242949e-01 6.72240108e-02 -9.28541720e-01
1.64979756e+00 5.51120698e-01 6.68554187e-01 -4.17277217e-01
4.27724391e-01 3.42957169e-01 -2.18683511e-01 -1.05452299e+00
1.16083100e-01 -1.29840150e-01 -1.17371045e-01 5.04454449e-02
-4.58096951e-01 -2.38744780e-01 -1.69528663e-01 6.97966991e-03
8.02204609e-01 1.77161396e-01 -1.95548132e-01 -3.00030261e-01
2.01214641e-01 -2.48541776e-02 8.62701535e-01 -1.92755484e-03
3.43507409e-01 4.59333062e-01 3.73019457e-01 -1.73512585e-02
-8.30740392e-01 -9.45987880e-01 -5.33646107e-01 -1.59547418e-01
5.86548388e-01 5.92730455e-02 -6.54835224e-01 -2.75751632e-02
-3.47399898e-02 8.50422144e-01 -5.26418686e-02 -9.15977582e-02
-2.09344670e-01 -9.38558698e-01 5.94457150e-01 7.02648342e-01
5.81348956e-01 -7.53381908e-01 -2.76245922e-01 5.41549623e-01
-4.76194769e-01 -9.96975005e-01 3.39024991e-01 -2.92246584e-02
-9.91783202e-01 -1.01311791e+00 -7.90889382e-01 -8.66034031e-01
9.57835615e-01 2.71340281e-01 2.48024091e-02 -4.62030955e-02
-3.09785098e-01 2.66872913e-01 -4.32144463e-01 -4.00674522e-01
1.27075270e-01 -4.65980500e-01 7.95051903e-02 1.01552263e-01
2.35466629e-01 -7.80615687e-01 -4.73826230e-01 4.93700325e-01
-4.19510722e-01 -5.24681151e-01 4.88936543e-01 4.67003882e-01
5.06476700e-01 1.24318671e+00 5.91744423e-01 -2.53127456e-01
8.04612815e-01 -8.15610349e-01 -8.70783031e-01 4.66150381e-02
-2.36222118e-01 -5.85700810e-01 1.60948142e-01 -2.81381100e-01
-1.21505499e+00 -2.26957336e-01 -2.37081766e-01 -2.44033441e-01
-1.91656157e-01 8.91895533e-01 -8.33587945e-01 2.69044131e-01
3.25672299e-01 1.58591092e-01 -1.45556077e-01 -6.52056456e-01
-1.42553732e-01 9.63439524e-01 4.20396626e-01 -7.68805861e-01
7.26873338e-01 4.27821368e-01 -9.28906202e-02 -9.50763226e-01
-3.12661439e-01 -5.44570506e-01 -6.31555259e-01 -3.14297736e-01
7.85207212e-01 -8.42629850e-01 -6.79009140e-01 6.39973342e-01
-1.58464193e+00 4.75507945e-01 3.00454646e-01 9.92561519e-01
-1.98551387e-01 9.19685721e-01 -5.85854828e-01 -1.52881265e+00
-1.91261098e-01 -9.06626701e-01 5.55074275e-01 3.44510674e-01
8.00588876e-02 -8.79803419e-01 -2.75127679e-01 -2.73815304e-01
-9.96772125e-02 2.70402193e-01 9.23904479e-01 -3.32170516e-01
-3.62643689e-01 -6.52237713e-01 -2.63597548e-01 3.19883376e-01
-9.80885774e-02 3.25023174e-01 -7.62255609e-01 -1.34135723e-01
5.00896513e-01 6.37862384e-01 2.89119571e-01 7.97646523e-01
1.20319963e+00 8.72669443e-02 -9.19275343e-01 5.52061260e-01
1.69165647e+00 9.58383799e-01 1.10323632e+00 4.17127848e-01
4.03363168e-01 6.81025982e-01 8.75540495e-01 8.41939807e-01
1.44985527e-01 5.79895489e-02 4.64646786e-01 -1.49879694e-01
1.01318252e+00 -2.63420671e-01 -1.48093939e-01 8.44037771e-01
-3.64783734e-01 -1.84447870e-01 -7.27208316e-01 5.83187699e-01
-1.67912805e+00 -9.83503401e-01 -8.04635465e-01 2.19095302e+00
2.14513853e-01 1.95253059e-01 -1.93547338e-01 6.20981276e-01
1.24782312e+00 -5.49266219e-01 -2.53933996e-01 1.79576084e-01
-2.88427155e-03 -2.51636297e-01 5.86707413e-01 -1.02354005e-01
-6.87392414e-01 -6.92101195e-02 5.49554253e+00 8.29126120e-01
-8.64616632e-01 -1.04554132e-01 4.19207692e-01 4.86362755e-01
-2.38922581e-01 2.46028155e-01 -8.61554861e-01 8.33824337e-01
1.71670750e-01 -1.20361522e-02 -4.52023977e-03 9.49489594e-01
6.40929699e-01 -3.75134706e-01 -3.62198561e-01 1.14636993e+00
-1.95596173e-01 -5.77464223e-01 -5.53507268e-01 4.23197150e-02
3.42083156e-01 -4.38235074e-01 -2.35239014e-01 -1.20437421e-01
-1.12768054e-01 -3.73048455e-01 8.74285161e-01 1.01337218e+00
7.27605164e-01 -1.10958707e+00 1.09808707e+00 5.36257625e-01
-1.20904958e+00 -3.30240250e-01 -7.20904887e-01 1.30942062e-01
7.36640096e-01 1.33014214e+00 -6.99354231e-01 1.02280712e+00
7.92258024e-01 5.58263123e-01 1.04950316e-01 1.60679555e+00
-6.81257546e-01 8.23114097e-01 -5.85304201e-01 -8.11392292e-02
-2.46776175e-02 -7.96429515e-01 4.70380545e-01 9.60150421e-01
1.28506815e+00 2.72729188e-01 3.03882584e-02 8.42613101e-01
6.23776793e-01 1.54149607e-01 -4.36997712e-01 3.13017249e-01
1.00074148e+00 1.35360622e+00 -7.29844570e-01 1.30215704e-01
-2.06502706e-01 4.01549608e-01 -4.72571820e-01 5.28256118e-01
-1.15156174e+00 -8.07813883e-01 -4.13964577e-02 5.20002484e-01
4.04861927e-01 -3.59817415e-01 -3.11975539e-01 -6.49736345e-01
2.80840814e-01 -1.50557607e-01 2.38441572e-01 -1.13197589e+00
-1.33127892e+00 4.98537779e-01 7.26310238e-02 -1.35405695e+00
2.77266234e-01 -1.42487019e-01 -1.19521594e+00 1.02156782e+00
-1.11654162e+00 -8.02521527e-01 -4.86245006e-01 3.66309553e-01
-8.36019125e-03 5.35221174e-02 7.55849600e-01 3.82276922e-01
-6.81035399e-01 2.82671712e-02 6.97064340e-01 3.07683468e-01
1.86209694e-01 -3.67463887e-01 -1.84869498e-01 8.01106036e-01
-5.02759159e-01 5.62708139e-01 6.72397017e-01 -9.45867538e-01
-9.22116220e-01 -7.57472932e-01 3.27543795e-01 3.31072301e-01
7.48883963e-01 -3.05240452e-01 -7.89467633e-01 5.14914632e-01
-2.92781919e-01 -4.56009179e-01 3.96356791e-01 4.72423173e-02
2.57078350e-01 -2.50570595e-01 -1.26426482e+00 2.09014475e-01
5.56155503e-01 -5.60531542e-02 -4.28138167e-01 3.32625985e-01
4.75306034e-01 -2.11984590e-01 -9.29732680e-01 4.25986022e-01
2.95715511e-01 -4.25593406e-01 6.49142146e-01 2.44495794e-01
4.97802012e-02 -3.87020260e-01 -1.07198246e-01 -1.31836164e+00
-4.07810360e-01 -1.44867599e-01 6.74291015e-01 1.61930132e+00
4.19727296e-01 -7.88648605e-01 4.09836173e-01 3.94280583e-01
-2.42395014e-01 -7.41157591e-01 -1.04347718e+00 -4.71038193e-01
-6.08496785e-01 -4.52723682e-01 4.42669153e-01 6.96897089e-01
2.30858847e-01 5.26457019e-02 -8.59375075e-02 8.84624422e-01
8.64961147e-01 -2.38732561e-01 2.49227464e-01 -1.49067235e+00
-1.97754279e-01 2.77850628e-01 -5.88672638e-01 -1.12641311e+00
-2.19360590e-01 -3.58798206e-01 4.11504775e-01 -1.95425224e+00
1.06622241e-01 -8.98288667e-01 -3.68049592e-02 -7.03448355e-02
1.61438942e-01 -4.92275745e-01 -6.00769758e-01 4.21667427e-01
6.57314807e-02 6.15814745e-01 8.73618782e-01 9.76748671e-03
4.25655134e-02 6.58401549e-01 -4.82189953e-01 1.07106924e+00
6.44890666e-01 -5.63098192e-01 -1.48628607e-01 -5.03030121e-01
3.42627279e-02 5.50764561e-01 2.79891729e-01 -1.13619435e+00
4.79017913e-01 3.69686067e-01 5.44164181e-01 -1.07492769e+00
3.93634617e-01 -1.24886239e+00 2.03058913e-01 3.57947022e-01
4.61563706e-01 -8.14015642e-02 1.99487343e-01 8.63896966e-01
-3.09027154e-02 -6.35676026e-01 4.30388659e-01 2.31016114e-01
-4.17297095e-01 2.53681481e-01 -6.17387533e-01 -5.54299116e-01
1.29009426e+00 -4.02616143e-01 1.63331017e-01 -3.71200770e-01
-7.87336648e-01 3.20497751e-01 -9.06342193e-02 -2.30347946e-01
9.35720026e-01 -1.13794279e+00 -5.79955101e-01 1.07268013e-01
-1.61994204e-01 3.89928341e-01 6.39158666e-01 1.03549719e+00
-4.01604503e-01 4.98592388e-03 1.40651599e-01 -7.35535681e-01
-1.05670536e+00 3.65828842e-01 2.82196403e-01 3.48509401e-01
-7.60743201e-01 6.07639074e-01 -2.68944860e-01 5.35330176e-02
1.82593182e-01 -2.89379537e-01 -3.70921165e-01 1.21509716e-01
3.77033770e-01 5.65167904e-01 -5.41418791e-02 -3.40548009e-01
-1.30987018e-01 8.75748098e-01 4.21347350e-01 -1.44223675e-01
1.60561895e+00 -2.96712965e-01 -4.12365049e-01 3.31024975e-01
8.14746380e-01 1.39174685e-01 -9.70156431e-01 -1.03357211e-01
8.19288846e-03 -4.86731440e-01 -1.47501662e-01 -3.27977806e-01
-9.56906676e-01 9.30732965e-01 3.09642434e-01 1.71553388e-01
1.04924893e+00 -2.68539071e-01 8.60280812e-01 7.44794914e-03
1.00535429e+00 -1.02250302e+00 -4.99609143e-01 3.92061062e-02
3.87708485e-01 -2.34120741e-01 8.05791616e-02 -7.35121012e-01
-5.52735090e-01 1.33612645e+00 2.68318981e-01 -1.08177602e-01
8.88194442e-01 6.42772317e-01 -2.46390954e-01 -1.17170498e-01
1.11394450e-01 3.65799397e-01 -6.14756286e-01 8.81039321e-01
-7.94858560e-02 -3.07581797e-02 -3.70655596e-01 1.39496648e+00
1.54565409e-01 -1.71737134e-01 5.61155736e-01 8.48555923e-01
-8.42682600e-01 -4.84414876e-01 -9.69893932e-01 2.73718357e-01
-2.63686985e-01 6.15369491e-02 4.90580767e-01 5.38953543e-01
6.39882311e-02 1.27075255e+00 5.57326525e-02 -3.55949581e-01
6.51821554e-01 -3.13815683e-01 2.06989437e-01 -2.94788510e-01
4.18416709e-01 4.83190000e-01 5.95708303e-02 1.68790296e-01
4.36498970e-02 -7.86521852e-01 -1.81054103e+00 -2.39173528e-02
-1.27532923e+00 9.26317334e-01 1.10067689e+00 8.33587527e-01
-2.38566548e-02 4.57549781e-01 1.09123707e+00 -5.98945439e-01
-7.74597764e-01 -1.08041966e+00 -1.44776428e+00 -2.12602258e-01
-6.52511775e-01 -8.68286490e-01 -7.83879399e-01 -1.95283517e-01] | [6.770392894744873, 1.468284249305725] |
8f751ca4-7252-4f94-bd5f-3dd3f215feba | rotational-projection-statistics-for-3d-local | 1304.3192 | null | http://arxiv.org/abs/1304.3192v1 | http://arxiv.org/pdf/1304.3192v1.pdf | Rotational Projection Statistics for 3D Local Surface Description and Object Recognition | Recognizing 3D objects in the presence of noise, varying mesh resolution,
occlusion and clutter is a very challenging task. This paper presents a novel
method named Rotational Projection Statistics (RoPS). It has three major
modules: Local Reference Frame (LRF) definition, RoPS feature description and
3D object recognition. We propose a novel technique to define the LRF by
calculating the scatter matrix of all points lying on the local surface. RoPS
feature descriptors are obtained by rotationally projecting the neighboring
points of a feature point onto 2D planes and calculating a set of statistics
(including low-order central moments and entropy) of the distribution of these
projected points. Using the proposed LRF and RoPS descriptor, we present a
hierarchical 3D object recognition algorithm. The performance of the proposed
LRF, RoPS descriptor and object recognition algorithm was rigorously tested on
a number of popular and publicly available datasets. Our proposed techniques
exhibited superior performance compared to existing techniques. We also showed
that our method is robust with respect to noise and varying mesh resolution.
Our RoPS based algorithm achieved recognition rates of 100%, 98.9%, 95.4% and
96.0% respectively when tested on the Bologna, UWA, Queen's and Ca' Foscari
Venezia Datasets. | ['Mohammed Bennamoun', 'Ferdous Sohel', 'Min Lu', 'Jianwei Wan', 'Yulan Guo'] | 2013-04-11 | null | null | null | null | ['3d-object-recognition'] | ['computer-vision'] | [ 7.61908293e-02 -7.34123051e-01 8.40030685e-02 -3.26100439e-01
-7.01526761e-01 -5.08726120e-01 7.93801367e-01 1.81375191e-01
-1.36719629e-01 3.80293012e-01 6.98865801e-02 1.04268342e-01
-4.77740675e-01 -7.56429672e-01 -4.80180293e-01 -7.97136605e-01
-2.64350325e-01 5.24020731e-01 4.33050722e-01 1.50951073e-01
7.67283320e-01 1.48640597e+00 -1.86054814e+00 7.92666450e-02
2.40522578e-01 1.20420361e+00 -8.64368901e-02 8.36187959e-01
8.63405019e-02 2.15967856e-02 -5.76076388e-01 2.25227818e-01
5.37457228e-01 2.45891258e-01 -6.13290250e-01 3.08048338e-01
4.96506333e-01 4.00903486e-02 3.32716294e-02 7.49181151e-01
4.43999022e-01 4.20752853e-01 1.20741522e+00 -8.56704831e-01
-4.47190374e-01 -5.12838364e-01 -7.85577476e-01 2.82362431e-01
7.43586421e-01 -3.88222814e-01 6.58146858e-01 -1.41290426e+00
8.78223956e-01 1.39422095e+00 6.52612209e-01 1.27120823e-01
-1.21721995e+00 -4.19989139e-01 -3.57753128e-01 1.92274023e-02
-1.82835829e+00 -3.28372389e-01 7.53485322e-01 -7.33007193e-01
1.02273464e+00 7.04347551e-01 3.27284366e-01 4.92804855e-01
3.31524134e-01 1.76017985e-01 1.33728623e+00 -6.67485952e-01
1.66370839e-01 7.08370656e-02 4.20548409e-01 5.41740358e-01
4.65790898e-01 -1.24197066e-01 -4.20625955e-01 -5.88759959e-01
8.71993899e-01 -3.81268635e-02 9.98117402e-02 -7.20681012e-01
-1.02823436e+00 4.76138622e-01 -6.91524968e-02 2.70606875e-01
-4.46658134e-01 -4.23476011e-01 2.21493483e-01 -1.20165825e-01
3.45309407e-01 -1.36221692e-01 -2.76474178e-01 -5.40832579e-02
-5.70088625e-01 2.30512425e-01 8.69285226e-01 8.60969007e-01
4.53007460e-01 -3.20421040e-01 1.57329038e-01 9.72026765e-01
5.97579479e-01 7.91396141e-01 2.31350616e-01 -6.27807856e-01
3.56677204e-01 7.13077843e-01 3.54816347e-01 -1.62898481e+00
-4.31185037e-01 -5.12682982e-02 -7.19811022e-01 3.60885322e-01
1.96908534e-01 4.60378915e-01 -7.98939407e-01 1.10711479e+00
7.27962255e-01 -7.11795762e-02 6.40590712e-02 8.20653677e-01
8.41806173e-01 5.42595029e-01 -3.51210892e-01 -2.55153090e-01
1.20129609e+00 -2.41636708e-01 -3.36962461e-01 6.52263045e-01
1.80341333e-01 -1.20083523e+00 5.44633448e-01 3.70838940e-01
-8.88705313e-01 -6.09816670e-01 -9.26382661e-01 3.58424097e-01
-3.40113103e-01 3.58105212e-01 2.95133412e-01 7.18868434e-01
-7.04907715e-01 5.38489282e-01 -8.36631954e-01 -5.78880250e-01
2.68691063e-01 6.46609426e-01 -7.60909081e-01 1.09417066e-01
-3.75667423e-01 8.43370020e-01 -1.45511040e-02 7.72473812e-02
-3.90409321e-01 -1.78905755e-01 -6.26485467e-01 -2.85957634e-01
-1.06745720e-01 -2.74556071e-01 6.08600676e-01 -5.27140349e-02
-1.54630542e+00 1.12377536e+00 -4.69391704e-01 1.25951141e-01
3.52271348e-01 -5.04041426e-02 -4.33911979e-01 1.28191486e-01
-4.54831384e-02 -1.54395467e-02 5.46944857e-01 -1.14274049e+00
-3.69046271e-01 -9.54643190e-01 -5.83046257e-01 2.29160398e-01
1.86058655e-01 1.22813597e-01 -1.52222827e-01 -3.79629701e-01
8.15189540e-01 -8.90483975e-01 7.60204419e-02 -5.33098429e-02
-2.76055992e-01 -5.61602294e-01 1.09998000e+00 -2.88579226e-01
7.30389476e-01 -2.25708365e+00 -1.10085728e-02 8.65221679e-01
-4.45478670e-02 4.01888847e-01 5.98086976e-02 5.23612916e-01
-3.63550931e-02 -2.06232920e-01 1.39427364e-01 1.32797167e-01
-1.77202359e-01 1.09031470e-02 -2.72310656e-02 1.02085567e+00
1.06984913e-01 2.73635715e-01 -3.95729750e-01 -4.22075242e-01
5.08290291e-01 8.27441990e-01 -2.82567918e-01 -3.22603993e-02
5.13651133e-01 2.72408336e-01 -5.80342770e-01 9.00109231e-01
9.73301053e-01 2.48252809e-01 -2.71123260e-01 -2.77037591e-01
-2.71829993e-01 -3.42543088e-02 -1.94781649e+00 1.01212311e+00
4.39047404e-02 3.95306081e-01 -2.58855850e-01 -9.15600836e-01
1.84176743e+00 1.59791410e-01 6.79801285e-01 -4.67357010e-01
2.20156431e-01 3.86939853e-01 -4.02858824e-01 -4.83037353e-01
3.59269142e-01 -4.72308956e-02 1.36136591e-01 1.59304082e-01
-1.53265774e-01 9.65873972e-02 -4.89504300e-02 -3.81903529e-01
9.84482408e-01 -8.27870369e-02 5.48365355e-01 -5.61104596e-01
1.18771410e+00 -1.25429377e-01 2.80022711e-01 4.88971144e-01
-3.12035382e-01 6.60439849e-01 2.58304745e-01 -6.24880612e-01
-9.69993651e-01 -1.26325083e+00 -8.14731181e-01 4.56289023e-01
1.94114625e-01 -2.58304298e-01 -3.40973139e-01 -2.83523411e-01
2.75476396e-01 1.77301005e-01 -5.87227821e-01 1.13672458e-01
-5.21361053e-01 -6.17939591e-01 1.10835038e-01 2.55144358e-01
5.46785235e-01 -5.57788491e-01 -6.63559318e-01 -7.95849338e-02
4.36160356e-01 -9.55973625e-01 -5.80910407e-02 -2.10356712e-01
-1.08326685e+00 -1.14508903e+00 -6.15302503e-01 -7.86669672e-01
8.12203348e-01 4.14221317e-01 8.12327743e-01 -4.34863359e-01
-5.66293478e-01 5.78038335e-01 -2.22499639e-01 -2.38944143e-01
2.76629254e-03 -2.75463760e-01 3.92179489e-01 9.19505358e-02
6.11989915e-01 -4.48118716e-01 -4.79717582e-01 7.06058145e-01
-4.18345988e-01 -4.81999516e-01 4.97560352e-01 5.95411777e-01
1.03902876e+00 2.12585945e-02 1.56395108e-01 -5.79655528e-01
5.12496233e-01 -2.63946593e-01 -9.08642232e-01 -6.01904616e-02
-1.46366581e-01 -1.31438911e-01 1.54372305e-01 -2.73308218e-01
-8.25724185e-01 3.19480896e-01 5.43835163e-02 -3.93957853e-01
-6.07078016e-01 8.39112848e-02 1.25507396e-02 -4.23666388e-01
7.04024196e-01 3.30979168e-01 -4.79084328e-02 -8.71611714e-01
-1.25515863e-01 1.12303424e+00 4.61590976e-01 -6.63753688e-01
9.26458895e-01 7.87231565e-01 7.19776630e-01 -1.32048023e+00
-2.85520494e-01 -1.02132928e+00 -1.05406129e+00 -2.49305084e-01
5.54847598e-01 -4.53466743e-01 -1.03459907e+00 4.15711492e-01
-1.00795603e+00 6.97409332e-01 -2.01826721e-01 7.00758576e-01
-5.11610985e-01 4.90491688e-01 -2.30669752e-02 -1.32746267e+00
-3.63751888e-01 -1.00352979e+00 1.20377445e+00 1.34741947e-01
-2.16328993e-01 -6.21512234e-01 3.09667170e-01 1.85374066e-01
3.52780730e-01 7.71852434e-01 7.35621572e-01 -6.88368917e-01
-2.29601771e-01 -8.31110835e-01 -3.40954900e-01 2.59430379e-01
3.31111431e-01 3.22782665e-01 -7.32403934e-01 -9.01756287e-02
1.39857829e-01 1.70898333e-01 3.70845258e-01 2.19439507e-01
8.01618040e-01 -7.06869811e-02 -3.26115519e-01 4.61196721e-01
1.59696698e+00 4.66504276e-01 4.82598394e-01 3.17020684e-01
4.46896493e-01 4.30785209e-01 6.87392414e-01 7.41212010e-01
3.24785449e-02 8.52620602e-01 5.68483956e-02 3.81004810e-01
8.82186054e-04 1.45479709e-01 2.78257430e-01 8.02526891e-01
-5.19775867e-01 2.64930487e-01 -1.03402925e+00 3.09235156e-01
-1.49681187e+00 -9.11740303e-01 -4.65600282e-01 2.58218312e+00
2.27463067e-01 7.50120506e-02 2.31822297e-01 5.05633414e-01
8.78232539e-01 -1.42367095e-01 -2.47961029e-01 -6.55068278e-01
-9.48160067e-02 3.15025717e-01 4.38484848e-01 4.17090505e-01
-1.24416697e+00 4.49655086e-01 6.07952070e+00 6.20477259e-01
-9.57584918e-01 -3.52166146e-01 2.11974293e-01 1.93894416e-01
3.07110280e-01 -2.81534851e-01 -1.24104345e+00 1.42840847e-01
7.71588206e-01 -2.04323921e-02 -7.29437843e-02 8.50805759e-01
1.71955720e-01 -3.26308310e-01 -7.74937868e-01 1.22758532e+00
3.35265756e-01 -1.10188067e+00 2.72299379e-01 1.55125275e-01
6.40856087e-01 -2.11463012e-02 5.50191961e-02 -2.49440253e-01
-3.25201392e-01 -9.83829200e-01 3.89461249e-01 7.30574965e-01
6.05399251e-01 -7.96710670e-01 7.50668347e-01 1.12809770e-01
-1.36552000e+00 2.82979250e-01 -5.15157878e-01 4.93671857e-02
-3.40905011e-01 5.39203703e-01 -8.14682364e-01 5.04268229e-01
5.71925223e-01 5.26929557e-01 -5.27748942e-01 1.23519492e+00
4.42456186e-01 1.48664802e-01 -8.01633179e-01 -2.56732970e-01
1.49466053e-01 -3.86745691e-01 8.16392243e-01 1.27656150e+00
3.46943438e-01 3.04409534e-01 -9.26524103e-02 4.75823849e-01
3.41001779e-01 6.38442636e-01 -7.77944803e-01 2.57715374e-01
2.85203665e-01 9.72384870e-01 -8.99390876e-01 -8.82978663e-02
-3.33598763e-01 4.02524710e-01 -1.02092542e-01 5.25364503e-02
-2.91490883e-01 -6.49005711e-01 5.62436640e-01 2.34245792e-01
3.82143974e-01 -5.80215931e-01 -2.43519321e-01 -9.61791158e-01
2.37094447e-01 -4.17465806e-01 4.43220109e-01 -4.92120534e-01
-1.19919348e+00 5.81233442e-01 3.02531421e-01 -1.36786366e+00
1.40788659e-01 -1.01922417e+00 -4.09115732e-01 9.75587130e-01
-9.77888525e-01 -7.50910938e-01 -4.31562781e-01 6.80535495e-01
2.83510447e-01 -4.35795635e-01 9.46838498e-01 1.42113447e-01
-3.08551997e-01 2.79930770e-01 5.59897244e-01 -2.72477251e-02
3.05351377e-01 -8.52604151e-01 6.91822395e-02 4.42221075e-01
1.47114620e-01 6.37593865e-01 5.69768906e-01 -5.62593162e-01
-1.67397368e+00 -7.06315815e-01 8.97699118e-01 -5.34015357e-01
2.98945546e-01 -3.03056717e-01 -7.12817967e-01 2.79175580e-01
-6.59971714e-01 2.39294380e-01 8.15056264e-01 1.14299199e-02
-3.60505819e-01 -1.11675419e-01 -1.50730932e+00 2.22296551e-01
8.57841671e-01 -3.28603268e-01 -7.42028415e-01 5.40697932e-01
-2.28658274e-01 -5.15726864e-01 -1.24484980e+00 5.40928960e-01
9.60463703e-01 -1.03340828e+00 1.16042936e+00 -3.44923049e-01
-2.48502687e-01 -4.98708814e-01 -7.46192515e-01 -4.79425192e-01
-3.73867303e-01 -3.64580750e-01 1.81431130e-01 1.11042237e+00
2.58845612e-02 -7.17296124e-01 7.82459080e-01 5.23403049e-01
1.60611391e-01 -8.14661682e-01 -1.22727621e+00 -1.01428950e+00
-3.25056076e-01 -4.31493431e-01 4.03547555e-01 5.61277747e-01
-2.37520918e-01 5.06688543e-02 8.68514255e-02 4.24177438e-01
8.23545575e-01 2.79547602e-01 9.49324369e-01 -1.54404414e+00
1.83071867e-01 -1.09069265e-01 -1.30844843e+00 -5.03889799e-01
-1.90350682e-01 -7.91437328e-01 -3.67825121e-01 -1.29975688e+00
2.13375092e-01 -4.94336307e-01 -2.80830622e-01 -5.12345731e-02
5.18716037e-01 3.79889786e-01 4.47515249e-02 5.27496338e-01
-2.92364568e-01 2.92487234e-01 1.09811890e+00 2.23347962e-01
-2.82272726e-01 2.29872808e-01 1.66876800e-02 7.13892758e-01
7.47090161e-01 -4.33089226e-01 1.49359748e-01 -2.93772575e-02
-2.53769785e-01 -2.31101632e-01 3.02269280e-01 -1.16529894e+00
7.86186531e-02 -1.99155122e-01 7.23688543e-01 -1.19706476e+00
5.39331436e-01 -8.91362548e-01 4.20516312e-01 3.64366174e-01
8.49865079e-02 7.63245299e-02 9.83498320e-02 4.85419452e-01
-7.73976743e-02 -2.34200269e-01 9.89472508e-01 6.64175451e-02
-4.93405044e-01 -6.70816144e-03 -2.30526358e-01 -5.13322175e-01
1.40514839e+00 -7.34855950e-01 -3.98669988e-02 1.55260712e-01
-8.18568647e-01 -3.67259532e-01 4.23942178e-01 2.36085474e-01
9.10897672e-01 -1.48626685e+00 -5.56316257e-01 7.29500234e-01
1.47066578e-01 -6.42907992e-02 1.14648342e-02 6.81307793e-01
-7.96142578e-01 7.53664196e-01 -4.61349607e-01 -1.01930439e+00
-1.66692793e+00 2.75767654e-01 1.68288941e-03 1.32652491e-01
-5.97640932e-01 5.66058695e-01 -2.42969126e-01 -3.04998010e-01
2.30611145e-01 -3.08761537e-01 -3.94178152e-01 5.74190216e-03
2.43030265e-01 7.87846386e-01 2.45802894e-01 -1.33015049e+00
-8.07835221e-01 1.54932725e+00 -3.59282410e-03 -7.33680576e-02
1.40032232e+00 1.67975891e-02 -2.93456353e-02 4.65197563e-01
1.50002086e+00 4.55625989e-02 -6.19444847e-01 -1.42621279e-01
9.80964303e-02 -9.14974213e-01 -1.99065939e-01 -3.35917234e-01
-6.14859641e-01 6.08074188e-01 1.03563619e+00 9.92299542e-02
8.68979216e-01 8.05345476e-02 3.28322470e-01 4.84592408e-01
4.04099524e-01 -9.10932720e-01 -2.66313016e-01 6.35411322e-01
1.06034207e+00 -9.22326684e-01 2.97680527e-01 -6.23436689e-01
-2.84425974e-01 1.35276926e+00 1.74707308e-01 -6.60516620e-01
9.60231066e-01 -3.14342268e-02 -1.71472624e-01 -3.23375791e-01
-5.34072042e-01 -2.39046011e-02 5.98729730e-01 6.59863770e-01
3.32875073e-01 5.19908071e-02 -5.68305910e-01 1.01647995e-01
-1.25834659e-01 -6.52372688e-02 2.43588567e-01 1.19411230e+00
-5.84128618e-01 -7.61785626e-01 -8.84249985e-01 4.61221546e-01
-5.14179945e-01 7.07661450e-01 -3.58686000e-01 9.25188005e-01
2.67442651e-02 5.48044026e-01 1.78557768e-01 -3.78739774e-01
6.46167994e-01 6.38980940e-02 5.72300434e-01 -4.52162951e-01
-6.53551146e-02 2.06660658e-01 -5.31842932e-02 -5.01897991e-01
-5.82974136e-01 -1.08185542e+00 -1.03300107e+00 -1.35300845e-01
-3.51992011e-01 2.15571240e-01 1.17393327e+00 6.01682603e-01
4.29927468e-01 -2.63251394e-01 8.22920263e-01 -1.03362274e+00
-5.68880796e-01 -6.71862543e-01 -8.08156610e-01 5.28119266e-01
8.03386420e-02 -1.10819781e+00 -3.45959365e-01 -5.49426675e-03] | [8.171692848205566, -2.3587357997894287] |
8451526a-a6d6-412a-8c14-ab7e24152518 | deep-multi-facial-patches-aggregation-network | 2002.09298 | null | https://arxiv.org/abs/2002.09298v1 | https://arxiv.org/pdf/2002.09298v1.pdf | Deep Multi-Facial Patches Aggregation Network For Facial Expression Recognition | In this paper, we propose an approach for Facial Expressions Recognition (FER) based on a deep multi-facial patches aggregation network. Deep features are learned from facial patches using deep sub-networks and aggregated within one deep architecture for expression classification . Several problems may affect the performance of deep-learning based FER approaches, in particular, the small size of existing FER datasets which might not be sufficient to train large deep learning networks. Moreover, it is extremely time-consuming to collect and annotate a large number of facial images. To account for this, we propose two data augmentation techniques for facial expression generation to expand FER labeled training datasets. We evaluate the proposed framework on three FER datasets. Results show that the proposed approach achieves state-of-art FER deep learning approaches performance when the model is trained and tested on images from the same dataset. Moreover, the proposed data augmentation techniques improve the expression recognition rate, and thus can be a solution for training deep learning FER models using small datasets. The accuracy degrades significantly when testing for dataset bias. | ['Ahmed Rachid Hazourli', 'Alice Othmani', 'Amine Djeghri', 'Hanan Salam'] | 2020-02-20 | null | null | null | null | ['facial-expression-generation'] | ['computer-vision'] | [ 3.36557329e-01 1.27773985e-01 7.65363798e-02 -6.85970008e-01
-5.87252498e-01 6.35062307e-02 2.68108994e-01 -2.62764752e-01
-3.93813342e-01 9.49582756e-01 -3.03097934e-01 2.61620939e-01
2.15472594e-01 -8.15611243e-01 -7.70309448e-01 -9.03026640e-01
-5.76305948e-02 1.81653291e-01 -3.03723872e-01 -3.49356025e-01
-2.95283973e-01 8.89366865e-01 -2.03181291e+00 5.52100599e-01
5.07855117e-01 1.82252443e+00 -3.28190207e-01 3.27944964e-01
-3.11273020e-02 1.35885799e+00 -8.67322087e-01 -6.78566992e-01
8.32072571e-02 -4.81124938e-01 -6.74228668e-01 2.24483237e-01
5.90956986e-01 -5.60123980e-01 1.04098432e-02 9.07422125e-01
5.43966353e-01 1.92524288e-02 5.26533782e-01 -1.52292562e+00
-4.06976908e-01 2.13959306e-01 -7.13758469e-01 -5.42182922e-02
1.87768191e-01 -2.50155568e-01 5.91853917e-01 -1.03138494e+00
4.65325832e-01 1.22083354e+00 8.06768656e-01 7.68519104e-01
-9.54215467e-01 -9.68895853e-01 -1.25553906e-01 1.48414755e-02
-1.46536195e+00 -7.42102861e-01 8.16094935e-01 -3.29096437e-01
7.57069170e-01 1.66925564e-01 6.85159802e-01 1.18667018e+00
-2.47300535e-01 8.91839147e-01 1.31552064e+00 -4.64786202e-01
2.15322562e-02 1.82640657e-01 -1.08427033e-01 7.64489055e-01
-2.12934185e-02 -1.55792892e-01 -5.26830137e-01 -2.58078992e-01
7.41426468e-01 -1.01938672e-01 -7.80075556e-03 2.23598555e-01
-4.02718186e-01 9.01407659e-01 3.71665686e-01 5.10002673e-01
-6.48883104e-01 1.81162447e-01 7.61692584e-01 5.44382989e-01
9.28138494e-01 2.02465847e-01 -3.94441903e-01 -1.47031739e-01
-1.03309357e+00 3.01574022e-01 4.41126049e-01 5.60010076e-01
1.18004775e+00 3.67103696e-01 -1.47857532e-01 1.22462189e+00
7.02261627e-02 3.51345330e-01 3.75832886e-01 -7.61317194e-01
-4.92056645e-03 6.60597146e-01 -1.64748639e-01 -1.13799965e+00
-2.12417275e-01 3.24902619e-04 -1.01379120e+00 3.71777833e-01
2.87953407e-01 -4.49336201e-01 -7.52449334e-01 1.80135727e+00
1.70811966e-01 1.15350217e-01 9.22656208e-02 7.82016695e-01
9.85297561e-01 4.92618918e-01 5.45874715e-01 -3.03926677e-01
1.11711144e+00 -9.51129913e-01 -8.09787333e-01 1.47934750e-01
8.45220804e-01 -6.45948529e-01 7.70307958e-01 5.62632382e-01
-9.69356537e-01 -7.71333992e-01 -7.49502182e-01 1.85496911e-01
-3.42269242e-01 7.85386443e-01 9.62362885e-01 6.54767573e-01
-1.14109600e+00 3.61852646e-01 -4.42702800e-01 -1.98332652e-01
9.77903128e-01 8.43920827e-01 -7.35596001e-01 3.40490118e-02
-1.12936091e+00 7.13358760e-01 1.92015424e-01 3.88497561e-01
-9.40109611e-01 -3.38206619e-01 -7.55801558e-01 7.52338618e-02
-2.42410883e-01 -1.37810394e-01 1.12161303e+00 -2.21043038e+00
-1.50910270e+00 1.10287297e+00 -4.31993641e-02 -3.32287490e-01
4.02896732e-01 -2.41206497e-01 -4.66787338e-01 2.17436090e-01
-2.79424995e-01 7.70035028e-01 9.34049487e-01 -1.11606991e+00
-3.20923328e-01 -5.09402931e-01 9.57358256e-02 -2.33983427e-01
-6.40806615e-01 3.79089892e-01 5.55706723e-03 -4.38849449e-01
-5.73293746e-01 -8.01350832e-01 -1.00614935e-01 1.16061360e-01
4.57423329e-02 -3.70653331e-01 9.34053302e-01 -5.61892271e-01
1.04618371e+00 -2.23762560e+00 -1.96834564e-01 2.33391657e-01
8.53187777e-03 5.36436498e-01 -3.18390578e-01 -1.25702232e-01
-3.79301399e-01 3.43587659e-02 2.17986315e-01 -4.06335503e-01
-2.46486038e-01 1.92064032e-01 4.29595308e-03 3.31693858e-01
6.55750275e-01 6.79878950e-01 -4.22302693e-01 -4.98088688e-01
3.28798709e-03 7.49945700e-01 -4.72833812e-01 4.77062017e-01
-1.16505079e-01 4.75320399e-01 -4.52096820e-01 8.51609170e-01
1.02437425e+00 -2.43425407e-02 -1.08764078e-02 -4.14541900e-01
3.69554847e-01 -3.74177098e-01 -7.60662377e-01 1.42565739e+00
-6.77414000e-01 6.86355889e-01 9.86425020e-03 -1.43392944e+00
1.34303367e+00 5.25159240e-01 7.72350729e-01 -7.10388541e-01
6.89376950e-01 3.37593973e-01 -1.93655044e-01 -5.77969193e-01
2.97704905e-01 -4.16679084e-01 6.20781332e-02 2.48426110e-01
6.42788649e-01 3.43136430e-01 2.66937852e-01 -4.32542205e-01
9.18681085e-01 1.48135975e-01 2.01820340e-02 7.12145120e-02
8.49209726e-01 -2.40752667e-01 4.98765886e-01 1.69246882e-01
-4.11933847e-02 1.08085312e-01 6.88019097e-01 -9.37519193e-01
-9.70742881e-01 -3.65481049e-01 -3.05945724e-01 1.44286025e+00
-4.47243571e-01 -1.23189934e-01 -9.63210940e-01 -8.20389509e-01
-1.92077637e-01 6.45783395e-02 -9.99179125e-01 -1.60709321e-01
-2.49119043e-01 -9.77766156e-01 1.01069558e+00 5.04432619e-01
7.30731368e-01 -1.24335217e+00 -5.57840705e-01 4.46748100e-02
-1.37376875e-01 -1.24266219e+00 3.67369145e-01 -4.07159403e-02
-6.02499247e-01 -1.01555800e+00 -8.87674928e-01 -7.74503708e-01
8.71486008e-01 -2.83116460e-01 1.00464272e+00 3.86111110e-01
-3.37631822e-01 1.39666975e-01 -5.28307319e-01 -5.99551976e-01
-4.74435300e-01 -2.01292202e-01 -6.24730363e-02 5.89804649e-01
7.46626735e-01 -4.19927150e-01 -4.49224681e-01 1.45421937e-01
-9.01051283e-01 -1.73897013e-01 5.81363976e-01 9.89144921e-01
4.79094714e-01 -1.94147214e-01 7.75541306e-01 -6.75374150e-01
6.20109439e-01 -4.50510472e-01 -3.67604613e-01 3.46636474e-01
-1.00619167e-01 -2.01887131e-01 5.43019414e-01 -5.10356903e-01
-1.20349264e+00 1.60682902e-01 -4.46878880e-01 -6.97140038e-01
-2.98532754e-01 4.23553765e-01 -5.48022538e-02 -5.83089948e-01
6.73223913e-01 -1.85732439e-01 3.41013402e-01 -2.34637707e-01
-1.93056837e-02 9.00081933e-01 -1.34418225e-02 -6.25410080e-01
3.12645850e-03 1.83131248e-01 2.37972170e-01 -7.26315260e-01
-7.26088524e-01 6.58111200e-02 -7.00335383e-01 -5.27400851e-01
7.22941041e-01 -1.28216004e+00 -6.95153892e-01 6.89190209e-01
-1.19090188e+00 -2.19161019e-01 -2.13618800e-02 1.70348257e-01
-5.14516950e-01 4.05834690e-02 -6.11959040e-01 -9.59513783e-01
-5.60262024e-01 -1.16987312e+00 1.38520610e+00 2.30395496e-01
-2.63413727e-01 -8.73730302e-01 4.78239693e-02 3.29178780e-01
4.98176575e-01 8.38347971e-01 4.85729635e-01 -5.50795674e-01
-4.78345975e-02 -3.08371603e-01 -3.70204300e-01 8.74586463e-01
8.77000839e-02 2.66902000e-01 -1.51319039e+00 -8.43200982e-02
-3.26615542e-01 -1.37810075e+00 6.10116661e-01 -1.09856538e-02
1.61669779e+00 -3.30573231e-01 2.25693849e-03 5.36252737e-01
1.39859748e+00 1.73926517e-01 8.05524826e-01 2.22545117e-01
3.82581234e-01 6.96927011e-01 7.61914313e-01 7.21671939e-01
-1.07445136e-01 5.78613698e-01 2.75193155e-01 -6.66586041e-01
7.05380887e-02 1.60040408e-01 2.64896542e-01 2.06310302e-01
-6.03207946e-01 -3.81564498e-02 -6.41410112e-01 3.18380237e-01
-1.62691736e+00 -9.60096598e-01 1.15054019e-01 1.74095595e+00
8.16516876e-01 -5.56872070e-01 1.40734419e-01 1.33981735e-01
3.13123405e-01 -1.70332231e-02 -2.79224187e-01 -8.19727540e-01
-3.26598853e-01 8.72834086e-01 -8.58799741e-02 1.17372327e-01
-1.20968866e+00 1.05401707e+00 6.01239729e+00 7.35795319e-01
-1.77054894e+00 2.69913375e-01 1.24435055e+00 -1.11967854e-01
1.51380107e-01 -7.41389334e-01 -4.98232991e-01 1.70782566e-01
1.08466923e+00 1.25836909e-01 1.27263054e-01 1.23953652e+00
8.23429823e-02 2.84862388e-02 -7.93302953e-01 1.37494636e+00
3.20624620e-01 -9.93792593e-01 2.17781290e-01 -7.72459507e-02
6.49965167e-01 -1.27638489e-01 3.48578170e-02 4.48920548e-01
-2.65424419e-02 -1.37482154e+00 2.30140507e-01 4.20908183e-01
1.16986513e+00 -9.32851136e-01 1.12132597e+00 -1.68286338e-01
-7.47269869e-01 -6.78481683e-02 -6.38049126e-01 -8.59809741e-02
-1.44697472e-01 4.05189484e-01 -5.83044112e-01 3.37405652e-01
9.19429183e-01 5.46953321e-01 -5.57937562e-01 5.09583592e-01
6.66544735e-02 3.82719636e-01 -2.92119682e-01 -4.91517410e-02
1.91550344e-01 -9.92813241e-03 -3.94781381e-01 1.15347564e+00
5.81951559e-01 -1.17141649e-01 -6.14473633e-02 6.44817114e-01
-3.49196881e-01 3.69607329e-01 -6.67772293e-01 -2.84396470e-01
-1.56752735e-01 1.65347075e+00 -2.21098393e-01 -4.11673397e-01
-5.53610861e-01 1.01139653e+00 5.17744124e-01 3.43497604e-01
-9.65041459e-01 -1.81934722e-02 7.72998512e-01 5.42960577e-02
5.61421737e-02 3.96479815e-01 1.55411795e-01 -1.00009668e+00
-1.04560688e-01 -1.10908782e+00 2.06886560e-01 -8.44667435e-01
-1.11391556e+00 1.17904389e+00 -2.48420313e-01 -1.07509494e+00
-3.30543488e-01 -8.47622633e-01 -3.04426700e-01 6.50478244e-01
-1.55238605e+00 -1.52679253e+00 -8.02946329e-01 9.77892220e-01
1.18713371e-01 -5.57259321e-01 1.34227753e+00 7.39404798e-01
-5.58384776e-01 9.54819322e-01 -3.08261901e-01 4.10099924e-01
7.17540145e-01 -6.06401742e-01 -3.92342269e-01 4.85365570e-01
5.06349355e-02 2.16538429e-01 2.87300169e-01 -6.94730058e-02
-1.03967166e+00 -1.19835222e+00 7.34773040e-01 5.02523631e-02
3.32665443e-01 -2.75818497e-01 -7.70368278e-01 6.82607353e-01
1.94597960e-01 5.13361216e-01 1.23929799e+00 1.68137953e-01
-2.81038404e-01 -4.28021699e-01 -1.38260674e+00 9.32687223e-02
6.83875918e-01 -4.22620475e-01 3.93125750e-02 3.82923454e-01
-6.40692608e-03 -1.75304353e-01 -1.16461802e+00 6.12887442e-01
7.94791818e-01 -1.05443561e+00 5.05408525e-01 -7.96181381e-01
7.50602305e-01 6.21640161e-02 -2.46754467e-01 -1.10330820e+00
-5.79846352e-02 -3.22857559e-01 1.34545431e-01 1.17892325e+00
3.27683240e-01 -2.70369262e-01 9.00672495e-01 7.01567769e-01
3.10473889e-01 -1.03058159e+00 -9.75967288e-01 -3.31941932e-01
-1.04993805e-01 -2.01506987e-01 8.04351926e-01 1.00801086e+00
-2.48472542e-01 2.65988469e-01 -5.53558469e-01 -3.06596220e-01
2.18817875e-01 -8.75883847e-02 1.06923640e+00 -1.18441510e+00
1.58198159e-02 -2.64234036e-01 -7.99806952e-01 -3.74760598e-01
6.59642875e-01 -5.54625809e-01 -2.30370551e-01 -9.57221448e-01
1.79376230e-01 -6.33269668e-01 -4.46592838e-01 1.00019383e+00
8.46977383e-02 7.19855368e-01 5.87407090e-02 -9.79927331e-02
-5.84823132e-01 8.13820958e-01 1.10374987e+00 -1.17100276e-01
1.07696936e-01 -1.32048473e-01 -5.98835588e-01 6.87066972e-01
7.49116659e-01 -1.24602430e-01 -3.57951283e-01 -4.34807032e-01
8.89495164e-02 -1.41132146e-01 3.42006892e-01 -1.00660825e+00
-2.66654432e-01 6.97878525e-02 8.63765597e-01 2.40246970e-02
4.56067234e-01 -9.83000398e-01 1.35267392e-01 5.13876602e-02
-1.52778894e-01 -8.69049132e-02 5.66349626e-01 -8.01654980e-02
-7.68924296e-01 6.02513105e-02 1.07932127e+00 -9.06962827e-02
-7.74636507e-01 5.35347879e-01 -2.84067094e-01 -5.02559662e-01
1.17346358e+00 -1.93510681e-01 -2.87642665e-02 -4.18873280e-01
-8.17433596e-01 -3.22322428e-01 2.13965118e-01 4.53723103e-01
5.96613228e-01 -1.61043060e+00 -5.77533066e-01 3.92232269e-01
3.57460171e-01 -2.08970174e-01 3.74610543e-01 8.49800229e-01
-5.51979959e-01 4.51527089e-02 -8.44477594e-01 -4.49072629e-01
-1.72526562e+00 2.60980070e-01 6.68780982e-01 -2.50900358e-01
7.08107054e-02 1.04371428e+00 1.70165166e-01 -2.72885054e-01
1.11677423e-01 -2.00528111e-02 -4.73269165e-01 3.20163518e-01
8.23834360e-01 -9.66417491e-02 1.16223522e-01 -1.06873584e+00
-2.17135772e-01 5.94024241e-01 -2.78043784e-02 2.87690043e-01
1.47157311e+00 3.27140063e-01 -4.48873162e-01 4.27347161e-02
1.46521580e+00 -4.06082779e-01 -9.56295729e-01 -1.72216445e-02
-3.34403306e-01 -4.02233750e-01 1.32115260e-01 -6.89315498e-01
-1.55669963e+00 1.10279131e+00 8.65292907e-01 -2.92579800e-01
1.72574723e+00 -2.06639469e-01 5.82034945e-01 3.60634029e-01
3.66747350e-01 -1.11903870e+00 2.68098086e-01 2.24859342e-01
9.60676908e-01 -1.31600881e+00 -2.29862392e-01 -1.73821151e-01
-6.90859795e-01 1.46755898e+00 9.88479495e-01 -5.47203459e-02
7.32285261e-01 3.31385583e-01 2.77475536e-01 -3.14336985e-01
-6.27461135e-01 -1.36536762e-01 6.55490831e-02 4.25700694e-01
9.41274583e-01 -1.02860622e-01 -2.37947479e-01 6.62852645e-01
-8.17318819e-03 5.94200134e-01 9.19783264e-02 6.84111536e-01
-1.82272971e-01 -1.18915939e+00 -1.73791081e-01 4.74212706e-01
-9.44416046e-01 1.74639091e-01 -5.85602701e-01 8.62726331e-01
5.15864074e-01 6.64679945e-01 1.16048977e-01 -5.40037572e-01
3.49293053e-01 2.23361507e-01 7.25969434e-01 -2.31013998e-01
-5.88473201e-01 3.03030014e-03 3.16211343e-01 -5.74224591e-01
-1.11401796e+00 -4.18009281e-01 -8.58988702e-01 -2.67231584e-01
-2.23333523e-01 5.62487207e-02 4.13257778e-01 8.90536606e-01
4.37371999e-01 2.87496567e-01 7.91428328e-01 -7.45709181e-01
-2.07666218e-01 -1.26982439e+00 -7.34015107e-01 8.29298437e-01
1.66414231e-01 -8.72708678e-01 -2.35539019e-01 2.25402579e-01] | [13.596585273742676, 1.7560434341430664] |
bb5caddc-06c6-4322-8f1b-5a1ea78324bf | examining-risks-of-racial-biases-in-nlp-tools | 2305.19409 | null | https://arxiv.org/abs/2305.19409v1 | https://arxiv.org/pdf/2305.19409v1.pdf | Examining risks of racial biases in NLP tools for child protective services | Although much literature has established the presence of demographic bias in natural language processing (NLP) models, most work relies on curated bias metrics that may not be reflective of real-world applications. At the same time, practitioners are increasingly using algorithmic tools in high-stakes settings, with particular recent interest in NLP. In this work, we focus on one such setting: child protective services (CPS). CPS workers often write copious free-form text notes about families they are working with, and CPS agencies are actively seeking to deploy NLP models to leverage these data. Given well-established racial bias in this setting, we investigate possible ways deployed NLP is liable to increase racial disparities. We specifically examine word statistics within notes and algorithmic fairness in risk prediction, coreference resolution, and named entity recognition (NER). We document consistent algorithmic unfairness in NER models, possible algorithmic unfairness in coreference resolution models, and little evidence of exacerbated racial bias in risk prediction. While there is existing pronounced criticism of risk prediction, our results expose previously undocumented risks of racial bias in realistic information extraction systems, highlighting potential concerns in deploying them, even though they may appear more benign. Our work serves as a rare realistic examination of NLP algorithmic fairness in a potential deployed setting and a timely investigation of a specific risk associated with deploying NLP in CPS settings. | ['Yulia Tsvetkov', 'David Steier', 'Emily Putnam-Hornstein', 'Alexandra Chouldechova', 'Nupoor Gandhi', 'Amanda Coston', 'Anjalie Field'] | 2023-05-30 | null | null | null | null | ['named-entity-recognition-ner', 'coreference-resolution'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.76296103e-01 5.03797174e-01 -6.73489213e-01 -7.02195227e-01
-9.21072006e-01 -7.52169371e-01 5.81919014e-01 7.56854653e-01
-8.60015452e-01 7.95657694e-01 1.10064912e+00 -8.11705172e-01
-3.96411389e-01 -8.05262268e-01 -5.37273586e-01 -3.50599885e-01
2.49726698e-01 4.63544369e-01 -6.90159023e-01 1.49132952e-01
5.36889613e-01 7.43035376e-01 -8.64488006e-01 1.05979405e-01
1.29471171e+00 2.82476217e-01 -5.92508912e-01 4.12640393e-01
-2.16834754e-01 1.09306324e+00 -6.15370095e-01 -1.09850097e+00
4.21351165e-01 1.19283445e-01 -9.32002723e-01 -7.47463822e-01
9.23096359e-01 -4.30273712e-01 -2.40167066e-01 9.81038094e-01
8.39738011e-01 1.27576873e-01 7.31920421e-01 -1.40914285e+00
-7.19134450e-01 1.11501253e+00 -7.69887388e-01 5.06218433e-01
3.56967300e-01 4.45020646e-01 1.20776367e+00 -2.33988419e-01
6.64334059e-01 1.30453408e+00 9.33307648e-01 5.11758149e-01
-1.45053709e+00 -1.30205774e+00 -9.89385545e-02 -4.89860803e-01
-1.06262255e+00 -8.90516222e-01 1.92342862e-01 -7.38648117e-01
9.92671788e-01 4.34371024e-01 3.52420062e-01 1.18376398e+00
6.94885403e-02 2.34885663e-01 1.13488328e+00 -1.54026002e-01
1.13178141e-01 -1.43239766e-01 5.99842191e-01 4.00275260e-01
9.64939535e-01 3.00439328e-01 -6.47069812e-01 -1.05020058e+00
1.69605717e-01 -1.17304698e-01 -3.47663723e-02 5.51766790e-02
-8.16064537e-01 8.14459682e-01 1.68422624e-01 9.30204168e-02
-3.71753186e-01 -6.75647035e-02 4.26307231e-01 -1.49948016e-01
7.93336034e-01 7.26423264e-01 -3.41870129e-01 -5.55784822e-01
-1.19435978e+00 6.07367218e-01 9.11373913e-01 7.71913290e-01
3.53552759e-01 -3.03819358e-01 -4.26348418e-01 8.64045680e-01
-7.22100511e-02 7.30714381e-01 -6.45613596e-02 -1.15020907e+00
1.09615219e+00 5.27364731e-01 2.26639062e-01 -1.28132510e+00
-5.98038256e-01 -4.03891243e-02 -6.32467151e-01 1.10986978e-01
8.01754415e-01 -5.04474401e-01 -3.64169240e-01 2.00729704e+00
1.40703738e-01 -1.26929656e-01 -9.77355242e-02 6.45410478e-01
2.73065865e-01 3.49315286e-01 9.39968467e-01 -3.84359062e-01
1.26789951e+00 -1.26411512e-01 -5.20301640e-01 -2.97346681e-01
7.30402768e-01 -7.04571366e-01 7.55532146e-01 -6.10838085e-02
-1.07839382e+00 3.20617616e-01 -3.45346868e-01 -1.91695899e-01
-1.50126949e-01 -5.46167254e-01 8.69879305e-01 1.31610191e+00
-6.52899921e-01 7.15193212e-01 -3.89604747e-01 -8.20879877e-01
9.93914425e-01 7.24259764e-02 -4.62952673e-01 -1.02620378e-01
-9.90830898e-01 9.29553747e-01 -1.50540397e-01 -7.69545808e-02
-1.87320054e-01 -1.50725925e+00 -7.85241067e-01 2.51513839e-01
3.32225055e-01 -5.67642450e-01 9.39900517e-01 -7.87825644e-01
-3.42913419e-01 1.00013292e+00 -2.93041587e-01 -3.08978885e-01
5.44000983e-01 -2.28360876e-01 -5.75533092e-01 -7.44423866e-02
4.87749577e-01 4.84840602e-01 1.91280514e-01 -8.52969110e-01
-5.83634675e-01 -5.60274243e-01 4.95318063e-02 1.34379804e-01
-3.35571826e-01 8.87742162e-01 8.22031736e-01 -7.20782638e-01
-3.56502146e-01 -5.50803900e-01 -2.66151577e-01 -1.62077472e-01
-3.32487345e-01 -1.57018721e-01 3.05967834e-02 -6.95926666e-01
1.51197815e+00 -1.84311628e+00 -8.73004317e-01 1.48419008e-01
2.15329826e-01 3.20968747e-01 -1.07041769e-01 3.51904273e-01
-2.33122885e-01 9.91166115e-01 -1.78641215e-01 1.17324642e-03
9.04192030e-02 -2.80538291e-01 -4.97671455e-01 5.98669171e-01
2.16080561e-01 5.75918615e-01 -1.01617253e+00 -6.90800428e-01
-1.03990138e-01 1.26713127e-01 -8.92730951e-01 -1.32193297e-01
4.03006792e-01 -1.41162157e-01 -1.01195220e-02 8.49137366e-01
9.62554038e-01 3.29849094e-01 2.23392114e-01 2.91108713e-02
-6.83844805e-01 7.20149457e-01 -6.75457597e-01 1.04965973e+00
-8.02384596e-03 5.78096390e-01 2.40211219e-01 -6.16023839e-01
4.91819441e-01 -3.22174728e-02 4.00122523e-01 -4.23998982e-01
6.46137521e-02 -9.48794093e-03 2.43891209e-01 -4.97079402e-01
9.79409754e-01 -4.43255812e-01 -8.04543495e-02 9.09238100e-01
-3.18524629e-01 3.77701558e-02 -1.78127378e-01 3.14142466e-01
1.41145122e+00 -4.91390139e-01 3.43456894e-01 -6.48538589e-01
-3.19765270e-01 1.87085479e-01 1.02772748e+00 9.13964570e-01
-7.07569957e-01 6.20593846e-01 7.71748185e-01 -2.33671531e-01
-1.02085507e+00 -1.04415739e+00 -4.07260090e-01 1.05618048e+00
-4.09429908e-01 -4.02615935e-01 -5.30626237e-01 -8.05715382e-01
2.10971907e-01 1.16883600e+00 -5.33705652e-01 -9.32305008e-02
-2.83579469e-01 -1.00686455e+00 1.17980671e+00 4.04672831e-01
4.52596769e-02 -7.39049375e-01 -7.74141133e-01 -1.55918911e-01
-1.03833005e-01 -9.16089892e-01 -7.21251786e-01 -2.97424167e-01
-6.88582957e-01 -1.29214108e+00 -5.55839181e-01 -1.66834593e-01
5.18137097e-01 9.47408304e-02 1.08977747e+00 -1.05377756e-01
-3.04581970e-01 5.35666823e-01 -7.49138743e-02 -8.23852658e-01
-4.19166178e-01 2.33835261e-02 1.86768800e-01 -4.40880746e-01
9.91447508e-01 -4.21294838e-01 -5.09884655e-01 -2.27901816e-01
-6.76030576e-01 -3.37657601e-01 2.03299522e-01 4.20606613e-01
-2.97720969e-01 -4.09215033e-01 5.34810364e-01 -1.36454093e+00
8.22008908e-01 -8.74538124e-01 -2.46393129e-01 4.00341749e-01
-1.02207088e+00 -3.22607249e-01 1.44708917e-01 -2.64870703e-01
-1.40079033e+00 -8.40982020e-01 2.10415989e-01 1.89305604e-01
-5.89229405e-01 2.67727286e-01 -2.90754866e-02 2.31675327e-01
8.98223221e-01 -9.33246374e-01 -1.72581911e-01 -2.60761678e-01
-6.28992617e-02 8.28766048e-01 1.31104842e-01 -1.05790365e+00
7.08315969e-01 3.50202411e-01 -3.62469405e-01 -6.16327584e-01
-8.76941621e-01 -1.67327955e-01 -1.11039631e-01 9.79846269e-02
8.21242690e-01 -7.53759563e-01 -1.10896933e+00 7.67502636e-02
-9.93028045e-01 -3.93039346e-01 -2.49341056e-01 5.09105563e-01
-5.47535233e-02 2.60978252e-01 -5.83282351e-01 -1.28205419e+00
-5.36070883e-01 -6.68004453e-01 5.81986129e-01 3.87634844e-01
-9.73938525e-01 -7.50019133e-01 2.13417098e-01 8.49370599e-01
3.83638918e-01 2.58506060e-01 1.34258723e+00 -1.06664562e+00
-9.53442156e-02 -9.78900865e-02 -5.76541185e-01 -1.02248229e-01
8.70521739e-02 3.94177794e-01 -1.02495837e+00 6.86615855e-02
-4.29732442e-01 -1.65971085e-01 3.41339618e-01 3.23221862e-01
8.07780266e-01 -6.36904061e-01 -2.45200783e-01 3.14170390e-01
1.25833654e+00 2.16726437e-01 2.89968550e-01 1.08756788e-01
4.66462851e-01 1.20253026e+00 5.61173856e-01 6.15582764e-01
5.25487900e-01 3.18613648e-01 -1.73958838e-01 1.21696718e-01
3.36818188e-01 -7.84901083e-01 2.52580792e-01 1.57833786e-03
6.34929314e-02 -1.06569588e-01 -1.61080492e+00 6.61574364e-01
-1.50997925e+00 -1.41988182e+00 4.69798632e-02 2.23994160e+00
7.10481465e-01 9.17424783e-02 1.89260542e-01 -2.40962446e-01
1.10783589e+00 3.97892416e-01 -4.10275310e-01 -1.00439346e+00
-1.45614758e-01 3.15154672e-01 7.75473177e-01 3.01741242e-01
-6.53331816e-01 5.92893422e-01 7.21029472e+00 4.41541463e-01
-6.88365757e-01 2.95406971e-02 1.08843839e+00 -5.92286050e-01
-7.61097670e-01 2.48607561e-01 -4.92849857e-01 5.51100791e-01
9.72102284e-01 -6.66812301e-01 2.97121704e-01 5.04540682e-01
4.51603502e-01 -2.57574737e-01 -1.18049550e+00 7.03953743e-01
-2.85028368e-02 -1.23269093e+00 -4.77657504e-02 2.70394355e-01
6.19564056e-01 -4.19752449e-02 1.25286937e-01 2.59459108e-01
7.75905609e-01 -1.22817910e+00 7.49707282e-01 3.57453406e-01
8.91279757e-01 -9.34473574e-01 5.36862314e-01 2.29606688e-01
-3.55906874e-01 -3.58961552e-01 -3.51288050e-01 -4.37528908e-01
2.62989014e-01 8.70206475e-01 -7.36454904e-01 1.32554635e-01
7.61195779e-01 1.52318448e-01 -1.87857151e-01 7.98283815e-01
1.54895708e-01 8.88753831e-01 -2.89111733e-01 1.54818371e-01
1.11360639e-01 -2.68354407e-03 4.81348902e-01 1.31926644e+00
1.91559017e-01 7.41052926e-01 -5.47297180e-01 8.63371670e-01
-4.45605457e-01 3.55515420e-01 -1.11266184e+00 -5.33653140e-01
9.03450906e-01 1.23925567e+00 -3.41849208e-01 1.10277776e-02
-4.46589321e-01 7.65971988e-02 6.19707644e-01 2.69749492e-01
-5.31958520e-01 -1.06620803e-01 1.23796809e+00 3.14330578e-01
-7.05357373e-01 2.21877739e-01 -8.77272248e-01 -1.05204403e+00
-2.39510939e-01 -1.03510237e+00 6.90932035e-01 -6.38898373e-01
-1.54076827e+00 -1.74159274e-01 1.27843186e-01 -3.78000289e-01
1.71048403e-01 -1.25967979e-01 -6.84649229e-01 1.02203190e+00
-1.12556922e+00 -6.98794782e-01 1.99091583e-01 5.99965900e-02
3.48495841e-02 -1.07339509e-01 7.03067780e-01 2.64742792e-01
-7.25624979e-01 1.11131895e+00 -3.99598479e-01 2.37489849e-01
1.21121502e+00 -8.81623805e-01 4.57591265e-01 1.05694318e+00
-1.74632400e-01 1.18454623e+00 6.37706101e-01 -1.02789831e+00
-8.16498578e-01 -8.44216883e-01 1.45364475e+00 -8.35025609e-01
6.14811361e-01 -2.10643262e-01 -6.12424791e-01 7.92354465e-01
6.30868748e-02 -5.12699842e-01 1.22271466e+00 7.55290926e-01
-7.27121890e-01 9.70534142e-03 -1.78479958e+00 8.27651083e-01
1.51194382e+00 -7.01791763e-01 -7.83677340e-01 7.12135434e-02
6.70799553e-01 9.07319691e-03 -8.20346773e-01 3.96752268e-01
6.29232824e-01 -8.73281896e-01 6.07486129e-01 -1.19111383e+00
7.55227923e-01 4.63234663e-01 -1.00566804e-01 -1.08466613e+00
-4.02218819e-01 -7.41096854e-01 7.09338009e-01 2.05214000e+00
6.94085121e-01 -7.90648103e-01 8.25883269e-01 2.21140766e+00
2.30823815e-01 -3.12115163e-01 -7.62355089e-01 -2.18124717e-01
6.08038127e-01 -7.19461083e-01 8.93500447e-01 1.70784891e+00
4.01541024e-01 -5.31739853e-02 -8.66111591e-02 4.25146669e-01
7.45195210e-01 -2.37872049e-01 4.75579441e-01 -1.04217744e+00
2.20828354e-01 -4.96209592e-01 -1.15928665e-01 1.10298790e-01
6.63639545e-01 -9.61499572e-01 -3.04039717e-01 -1.21221364e+00
6.97812140e-01 -9.37150240e-01 -4.95917350e-02 5.69198847e-01
-3.42339277e-01 -3.96909922e-01 4.69719857e-01 -1.64363563e-01
-9.92304385e-02 4.05232124e-02 4.32297975e-01 -1.13287792e-01
-5.85866757e-02 -1.12119727e-01 -1.54520512e+00 6.99951470e-01
7.49792516e-01 -7.79633522e-01 -1.82127014e-01 -6.07571840e-01
5.65756202e-01 -1.37296230e-01 3.49401474e-01 -6.09956861e-01
3.80195767e-01 -7.12923944e-01 1.32415146e-01 -5.98040223e-02
-3.50034446e-01 -7.81798780e-01 6.93896115e-02 1.92957297e-01
-8.66633177e-01 1.52857870e-01 2.24137157e-01 1.75849527e-01
3.82810712e-01 -4.26286966e-01 4.40429837e-01 6.68550357e-02
1.40689820e-01 3.34333032e-01 -5.01462817e-01 6.74149632e-01
7.07519829e-01 3.35544571e-02 -9.91647065e-01 -1.57563657e-01
-6.49814084e-02 3.05361748e-01 8.63427460e-01 3.25506866e-01
-1.57316625e-02 -9.72460568e-01 -6.98878944e-01 -1.91995636e-01
1.13705046e-01 -3.16462725e-01 2.92897463e-01 7.65873790e-01
-4.43733275e-01 1.62864462e-01 -4.74080406e-02 2.54003912e-01
-1.21984804e+00 6.50084496e-01 7.89725557e-02 6.18331581e-02
-2.75620818e-01 7.13975787e-01 1.96666315e-01 -4.82960582e-01
1.71550766e-01 -1.90086827e-01 8.42929706e-02 3.56984466e-01
6.69537842e-01 9.40295517e-01 -2.84821570e-01 -3.99214774e-01
-5.29162586e-01 -1.79071091e-02 -4.91117202e-02 -2.78793305e-01
1.30675220e+00 -3.25374678e-02 -3.13055575e-01 2.30942860e-01
8.64511132e-01 6.59524918e-01 -7.65855014e-01 1.98807195e-01
2.54710287e-01 -7.11913288e-01 -4.48326260e-01 -8.79989028e-01
-7.78459191e-01 7.14596689e-01 1.39750481e-01 1.04981493e-02
8.46633732e-01 -3.15251827e-01 4.49392557e-01 1.58953160e-01
4.88112479e-01 -1.16344416e+00 -8.43454540e-01 1.02607012e-01
4.93240476e-01 -1.08920717e+00 1.80360511e-01 -3.03903937e-01
-6.28857970e-01 4.77148622e-01 6.09047234e-01 3.10687751e-01
5.92631996e-01 5.15971780e-01 3.61455157e-02 -9.32413489e-02
-8.14786136e-01 2.91137785e-01 -3.92527848e-01 7.16496110e-01
7.77013600e-01 3.40673149e-01 -8.15008938e-01 9.52697337e-01
-5.43896794e-01 -1.03149816e-01 7.71740854e-01 6.37436330e-01
-1.24521457e-01 -8.46693933e-01 -4.65102166e-01 9.66162622e-01
-1.09803939e+00 -5.68893850e-01 -5.93525350e-01 5.49399853e-01
1.50279030e-01 1.20746374e+00 3.01655799e-01 -9.25532952e-02
4.18271810e-01 3.04692328e-01 1.12125829e-01 -6.38628542e-01
-1.17299664e+00 -7.66505897e-01 7.68720686e-01 -6.52207494e-01
-3.89477193e-01 -1.19215775e+00 -8.25001895e-01 -1.09602773e+00
1.05562005e-02 1.71349868e-01 3.37489575e-01 7.89061010e-01
4.33379978e-01 -3.09765279e-01 1.77985772e-01 -1.60246998e-01
-3.99373114e-01 -6.59934700e-01 -5.43603122e-01 5.59556723e-01
1.22877866e-01 -2.89094895e-01 -3.89267713e-01 -5.50810814e-01] | [8.878366470336914, 5.675219535827637] |
f28c67d3-1967-42b8-bbb2-fe592cfc5251 | evadedroid-a-practical-evasion-attack-on | 2110.03301 | null | https://arxiv.org/abs/2110.03301v3 | https://arxiv.org/pdf/2110.03301v3.pdf | EvadeDroid: A Practical Evasion Attack on Machine Learning for Black-box Android Malware Detection | Over the last decade, researchers have extensively explored the vulnerabilities of Android malware detectors to adversarial examples through the development of evasion attacks; however, the practicality of these attacks in real-world scenarios remains arguable. The majority of studies have assumed attackers know the details of the target classifiers used for malware detection, while in reality, malicious actors have limited access to the target classifiers. This paper introduces EvadeDroid, a practical decision-based adversarial attack designed to effectively evade black-box Android malware detectors in real-world scenarios. In addition to generating real-world adversarial malware, the proposed evasion attack can also preserve the functionality of the original malware applications (apps). EvadeDroid constructs a collection of functionality-preserving transformations derived from benign donors that share opcode-level similarity with malware apps by leveraging an n-gram-based approach. These transformations are then used to morph malware instances into benign ones via an iterative and incremental manipulation strategy. The proposed manipulation technique is a novel, query-efficient optimization algorithm that can find and inject optimal sequences of transformations into malware apps. Our empirical evaluation demonstrates the efficacy of EvadeDroid under soft- and hard-label attacks. Furthermore, EvadeDroid exhibits the capability to generate real-world adversarial examples that can effectively evade a wide range of black-box ML-based malware detectors with minimal query requirements. Finally, we show that the proposed problem-space adversarial attack is able to preserve its stealthiness against five popular commercial antiviruses, thus demonstrating its feasibility in the real world. | ['Veelasha Moonsamy', 'Hamid Bostani'] | 2021-10-07 | null | null | null | null | ['android-malware-detection'] | ['miscellaneous'] | [ 6.41496003e-01 -1.14205293e-01 -4.22907054e-01 1.59377053e-01
-7.58356750e-01 -1.39449096e+00 7.67545819e-01 -2.08140790e-01
-6.33470993e-03 5.51389456e-01 -5.37524402e-01 -8.22820663e-01
1.35121390e-01 -9.72182870e-01 -9.27035987e-01 -5.54100215e-01
-3.33167106e-01 2.05920801e-01 3.69613498e-01 -2.43237704e-01
3.50473583e-01 5.94614565e-01 -1.49735737e+00 2.54969329e-01
9.60793018e-01 7.35518813e-01 -2.56579757e-01 9.09320116e-01
9.79113802e-02 4.26742792e-01 -9.92547870e-01 -8.95343482e-01
7.50303864e-01 -2.18837723e-01 -4.47043955e-01 -4.96176392e-01
2.66061872e-01 -4.60949063e-01 -2.62067690e-02 1.46527994e+00
2.98309684e-01 -5.27405262e-01 4.59219038e-01 -1.65018630e+00
-5.66753924e-01 2.83398598e-01 -4.21330661e-01 -1.26709044e-02
5.63353717e-01 6.39513373e-01 7.00301707e-01 -3.97428662e-01
5.10931075e-01 1.33536220e+00 5.14199197e-01 6.62776649e-01
-1.07949173e+00 -1.04059565e+00 -3.17304701e-01 -8.81537795e-02
-1.15257025e+00 -3.17013592e-01 6.49543107e-01 -4.54839528e-01
6.72551751e-01 7.18669236e-01 5.75350344e-01 1.52275538e+00
6.61853850e-01 2.20968604e-01 1.39138901e+00 1.73694920e-02
3.47013712e-01 5.43384194e-01 3.04901246e-02 8.20062995e-01
7.85295904e-01 5.11342704e-01 -6.15162216e-02 -1.08238554e+00
5.54713607e-02 2.70778745e-01 -2.23866865e-01 -3.31979215e-01
-5.60381353e-01 1.07866549e+00 2.10744902e-01 5.68180382e-02
-2.70539951e-02 2.33279052e-03 5.95245123e-01 1.23973407e-01
1.78299710e-01 7.26027310e-01 -3.62750173e-01 -1.55788913e-01
-4.74839747e-01 1.21978350e-01 1.04117763e+00 6.35276675e-01
7.34300256e-01 1.80966496e-01 3.21429446e-02 8.41562748e-02
3.72251928e-01 1.13590646e+00 6.07934356e-01 -6.59368157e-01
1.62520573e-01 7.52923191e-01 -2.24422872e-01 -1.31036091e+00
3.53043020e-01 -5.95861450e-02 -1.84999034e-01 4.61235702e-01
2.42787182e-01 -1.46316856e-01 -6.25663817e-01 1.68168306e+00
5.86948991e-01 2.82854348e-01 1.14793926e-01 2.38395765e-01
2.17825770e-01 6.30696535e-01 1.84146520e-02 -2.77713984e-01
1.57840526e+00 -4.55300838e-01 -4.65760142e-01 -2.80016392e-01
5.21231055e-01 -6.00774586e-01 1.37151921e+00 1.71087265e-01
-5.97671390e-01 -1.07338786e-01 -1.24477053e+00 5.71703374e-01
-7.50591278e-01 -2.51562417e-01 4.49424058e-01 1.59167504e+00
-6.95945919e-01 8.40809271e-02 -8.98237765e-01 -9.39469859e-02
7.81815827e-01 7.21602082e-01 -2.40100637e-01 2.06683785e-01
-9.05565977e-01 5.28028190e-01 2.34118804e-01 -4.50921714e-01
-1.59949338e+00 -6.24849796e-01 -7.07846165e-01 -8.30595940e-02
9.27077889e-01 -2.31845766e-01 1.04448998e+00 -1.11700702e+00
-1.54034460e+00 6.84313297e-01 6.60366863e-02 -6.44681633e-01
4.17113960e-01 -7.30282888e-02 -4.86877054e-01 3.02124326e-03
-8.25352669e-02 2.07607538e-01 1.48774076e+00 -1.29667044e+00
-3.37198943e-01 -4.31867123e-01 4.09405947e-01 -3.49940836e-01
-7.14924335e-01 1.27135247e-01 1.12663731e-02 -6.95923984e-01
-8.60931516e-01 -1.43994403e+00 4.28042710e-02 -2.28827730e-01
-7.18810081e-01 3.38945448e-01 1.62233162e+00 -5.83538651e-01
1.07125866e+00 -2.17090917e+00 -8.50023627e-02 4.79552895e-01
2.85331517e-01 1.01090229e+00 -2.48847887e-01 3.79790395e-01
2.42369249e-01 7.12722421e-01 -5.41668355e-01 2.59367526e-01
-1.55557871e-01 7.42568523e-02 -8.18720818e-01 4.43895012e-01
1.07269816e-01 1.01081896e+00 -8.10096860e-01 -2.69081891e-01
-1.40571311e-01 5.66591442e-01 -6.69673502e-01 2.18954250e-01
-6.87030494e-01 2.95590788e-01 -5.82533121e-01 7.99245059e-01
6.31421983e-01 1.50338501e-01 4.35682058e-01 1.84256643e-01
4.25010264e-01 4.10079136e-02 -5.73912740e-01 6.10990882e-01
-4.80223715e-01 4.07574922e-01 9.56308022e-02 -2.72295117e-01
6.23858035e-01 -1.93396643e-01 2.23416850e-01 -3.02391827e-01
3.74774933e-01 4.14081484e-01 2.66430467e-01 -2.89337993e-01
3.02684426e-01 4.03809875e-01 -1.39356419e-01 7.63604641e-01
-5.10502517e-01 -2.41585895e-02 -1.41609296e-01 4.33430336e-02
1.54936147e+00 -1.84236035e-01 5.49179673e-01 1.25805782e-02
1.00277925e+00 4.73760217e-02 2.96652108e-01 6.75265074e-01
-2.11555913e-01 -4.22408521e-01 6.28149152e-01 -3.86091024e-02
-6.67101741e-01 -1.12295616e+00 2.87904114e-01 1.01187396e+00
1.86844811e-01 -5.02870440e-01 -1.25061500e+00 -1.51816595e+00
1.70089235e-03 5.63282430e-01 -6.11961365e-01 -5.55879474e-01
-8.74975324e-01 -5.10382652e-01 1.20407224e+00 -1.67157695e-01
7.12425232e-01 -8.72413635e-01 -7.05526650e-01 -1.67950138e-01
2.90813088e-01 -1.06169951e+00 -5.59163511e-01 -4.09621954e-01
-5.11775732e-01 -1.71591139e+00 -5.26638106e-02 -5.00044942e-01
7.57627189e-01 3.38681161e-01 5.06668508e-01 3.96148801e-01
-5.20378113e-01 4.15308982e-01 -2.78336465e-01 -4.55760807e-01
-1.27366018e+00 1.42019689e-01 1.91226322e-02 1.67454273e-01
3.05667222e-01 -2.71658540e-01 -1.92944258e-01 4.54127878e-01
-1.16996145e+00 -8.28178465e-01 5.68747759e-01 6.62360489e-01
4.76751149e-01 2.63576508e-01 4.43400502e-01 -1.15349007e+00
9.49491024e-01 -4.99503732e-01 -1.12513781e+00 2.78821945e-01
-5.53188205e-01 -7.06017017e-02 1.14050758e+00 -1.21223438e+00
-7.84701467e-01 1.42518148e-01 -2.15304449e-01 -3.17699462e-01
1.54040486e-01 -5.92307076e-02 -7.01676726e-01 -7.09783196e-01
9.09922242e-01 3.42559904e-01 3.32909137e-01 2.88003068e-02
5.49032152e-01 7.62953401e-01 4.46839631e-01 -5.34497082e-01
1.49891400e+00 5.72926939e-01 8.28687400e-02 -7.97896624e-01
-1.60277143e-01 2.23990172e-01 1.05336681e-01 3.91220823e-02
8.58093381e-01 -3.28010499e-01 -1.08793485e+00 6.72564864e-01
-9.89416420e-01 -2.31089577e-01 -1.06155232e-01 -2.37974256e-01
-2.07733825e-01 6.22134686e-01 -3.42767596e-01 -7.03240395e-01
-5.16043782e-01 -1.76557374e+00 1.01780105e+00 -7.94464722e-02
-7.83101935e-03 -8.30314040e-01 1.69843361e-01 3.16812485e-01
4.93796319e-01 5.14787138e-01 9.82733011e-01 -1.22975254e+00
-8.31185460e-01 -5.85574389e-01 9.94632468e-02 3.89929742e-01
4.85306323e-01 2.63487786e-01 -8.51822019e-01 -5.79170465e-01
4.25556004e-01 1.60014592e-02 1.54146463e-01 -2.49353036e-01
1.08558011e+00 -1.13593042e+00 -5.75021863e-01 5.43002784e-01
1.22663045e+00 5.54237783e-01 5.57481825e-01 9.44696292e-02
7.71165311e-01 4.22146112e-01 5.99040329e-01 9.37673971e-02
-1.11010373e-01 8.24353456e-01 8.61388206e-01 2.39294410e-01
2.33245805e-01 -3.71989191e-01 1.04064143e+00 2.09948376e-01
3.93905461e-01 -1.98627815e-01 -5.73554635e-01 -3.87026258e-02
-1.13716567e+00 -9.26080465e-01 1.66578427e-01 2.44624615e+00
7.28723049e-01 3.31242830e-01 4.24183428e-01 5.56536093e-02
7.67564654e-01 1.42884061e-01 -6.94014132e-01 -8.16854715e-01
1.76460132e-01 4.46019471e-01 8.51196289e-01 4.96256709e-01
-1.00681973e+00 1.09431422e+00 5.61728096e+00 1.04154551e+00
-1.36528659e+00 6.04656041e-01 3.51159900e-01 3.25062633e-01
-2.64440447e-01 1.79175451e-01 -7.68796384e-01 6.71260059e-01
1.20727158e+00 -3.21407080e-01 7.83188283e-01 1.18203235e+00
-3.37469988e-02 3.02705288e-01 -9.16497469e-01 5.13332844e-01
2.60921568e-01 -1.24088264e+00 3.93932201e-02 6.47077918e-01
6.37870789e-01 -3.72948915e-01 5.44573724e-01 2.87277430e-01
4.90497828e-01 -1.03757167e+00 3.14327627e-01 -1.23462871e-01
9.57337558e-01 -9.13111746e-01 5.89470625e-01 2.94140011e-01
-1.09206915e+00 -6.31249428e-01 9.07201096e-02 2.87432879e-01
-1.25110611e-01 4.68240231e-02 -1.20837069e+00 1.46790251e-01
3.27854425e-01 1.51238993e-01 -1.04019225e+00 3.79842907e-01
-3.84864032e-01 8.86826456e-01 -7.46205226e-02 -2.14185894e-01
7.10337907e-02 2.25622114e-02 8.22836578e-01 8.56950700e-01
2.76438832e-01 -3.64813268e-01 2.02498078e-01 8.02000821e-01
-1.35665968e-01 1.38230681e-01 -1.32942033e+00 -5.56669354e-01
6.83058977e-01 1.22717690e+00 -5.26957870e-01 -2.59635985e-01
-2.16010734e-02 7.46920228e-01 -6.80649951e-02 1.31707555e-02
-1.25943363e+00 -4.58564460e-01 1.01772285e+00 2.63402045e-01
1.89613700e-01 -1.24706723e-01 -8.23063329e-02 -7.36815095e-01
-4.93411347e-02 -1.94687831e+00 2.04918310e-01 -1.26817450e-01
-1.06502306e+00 8.25907052e-01 1.58860251e-01 -1.06134152e+00
-3.65327984e-01 -6.46632850e-01 -7.37342775e-01 3.98036242e-01
-8.20601046e-01 -1.17610776e+00 -8.97156969e-02 6.78102851e-01
3.41184080e-01 -6.32610559e-01 7.40681708e-01 7.05651939e-02
-5.14945924e-01 9.28147197e-01 -7.99192861e-02 -5.36726788e-03
3.69228035e-01 -7.13673949e-01 5.67902923e-01 1.14354002e+00
-1.71627924e-02 1.01802659e+00 6.37580395e-01 -1.31320477e+00
-1.78771722e+00 -1.34077072e+00 5.02069406e-02 -5.56134224e-01
8.23296249e-01 -7.06845105e-01 -8.26404631e-01 7.13575006e-01
2.89304033e-02 -1.04134284e-01 5.57348192e-01 -8.25313151e-01
-8.01909745e-01 -2.02451035e-01 -1.75082600e+00 8.55543554e-01
8.86760831e-01 -7.82748401e-01 -2.05962174e-02 5.36189318e-01
1.00425720e+00 -9.96166319e-02 -5.00750303e-01 3.98314416e-01
5.89736879e-01 -7.65435696e-01 1.17809248e+00 -5.89722753e-01
-1.74280629e-02 -3.50083262e-01 -2.01071709e-01 -6.90653920e-01
6.18138254e-01 -1.19155049e+00 -5.37835479e-01 1.06648278e+00
9.01945084e-02 -1.22833884e+00 5.81392944e-01 4.02178569e-03
3.29047233e-01 -7.11694300e-01 -8.48893404e-01 -1.12565970e+00
-2.74020284e-02 -3.49860460e-01 1.00747287e+00 6.65893078e-01
-4.96584147e-01 -1.16166666e-01 -3.36225122e-01 3.17379981e-01
5.44502139e-01 -4.85032767e-01 1.25615132e+00 -8.71609211e-01
-6.15398586e-01 -3.97758991e-01 -5.21416247e-01 -3.99885029e-01
5.67682385e-01 -8.83931875e-01 -2.01466292e-01 -3.61845523e-01
1.04722768e-01 -3.70681286e-01 3.84136230e-01 5.05308151e-01
-7.39985704e-02 5.55412054e-01 2.68448919e-01 2.72639513e-01
2.79080179e-02 6.38457462e-02 8.43162775e-01 -3.46586168e-01
-5.10576904e-01 2.25325108e-01 -6.53612375e-01 7.36451924e-01
8.70479941e-01 -8.53268802e-01 -8.01072717e-01 3.83527845e-01
1.69532686e-01 -4.55890417e-01 4.79940951e-01 -7.15239346e-01
-3.23199570e-01 -2.30076760e-01 -4.02565658e-01 -3.38071026e-02
2.94809073e-01 -9.91255701e-01 4.70643133e-01 1.09028006e+00
-4.86690700e-02 3.81358147e-01 2.33119234e-01 8.00557494e-01
3.28057021e-01 -1.73466116e-01 8.88749719e-01 6.96369484e-02
-2.52964348e-01 3.08850259e-01 -5.20484686e-01 3.30717824e-02
1.74684572e+00 -2.67369092e-01 -8.19363415e-01 -1.23611726e-01
-3.10332552e-02 -4.58236158e-01 9.28446412e-01 4.22899067e-01
4.79363710e-01 -9.05078113e-01 -3.41147780e-01 3.78591180e-01
-1.09713010e-01 -8.49095464e-01 -3.23428631e-01 6.37870491e-01
-7.03688443e-01 1.00400023e-01 -3.82554352e-01 -5.33925474e-01
-1.92998517e+00 1.12438548e+00 2.82214046e-01 -5.26221097e-01
-1.69588745e-01 2.65627712e-01 8.95373896e-02 -4.75133866e-01
9.76554751e-02 1.32368892e-01 9.98653919e-02 -5.19581378e-01
5.09221971e-01 4.48473305e-01 -1.07311413e-01 -8.80146682e-01
-5.92015862e-01 4.32275414e-01 9.71864723e-03 5.88796735e-02
5.80835879e-01 4.91086572e-01 -4.24845010e-01 -4.00712013e-01
1.26523483e+00 8.70582879e-01 -7.10536897e-01 3.94096613e-01
-6.37351051e-02 -7.93926716e-01 -6.66919291e-01 -6.63473725e-01
-9.74057972e-01 6.27150953e-01 9.23758984e-01 5.55307329e-01
1.08290148e+00 -3.04085195e-01 1.05361247e+00 5.68078995e-01
4.30499792e-01 -3.13988835e-01 3.88928413e-01 2.13655233e-02
5.44220746e-01 -1.01923406e+00 -1.10373639e-01 -8.33695948e-01
-3.29544306e-01 6.93619192e-01 6.33556128e-01 -5.00653237e-02
4.73978937e-01 4.12782401e-01 -1.73018053e-01 -1.14759907e-01
-2.50401378e-01 3.75628382e-01 2.16524705e-01 1.08123386e+00
-4.74872082e-01 1.69681013e-01 -3.61906826e-01 3.67093474e-01
-2.73281097e-01 -4.97931391e-01 6.46219194e-01 9.75170195e-01
-2.31317163e-01 -1.59294009e+00 -6.72235668e-01 3.43301833e-01
-7.46434033e-01 1.05188660e-01 -1.04112101e+00 9.65807378e-01
2.44651869e-01 1.10398149e+00 -6.04779780e-01 -8.40965509e-01
-9.04774889e-02 -8.52486044e-02 1.16794840e-01 -6.50478184e-01
-1.11806548e+00 -2.84524113e-01 2.94003021e-02 -5.13784647e-01
5.66086136e-02 -3.03690225e-01 -9.64828610e-01 -4.77455169e-01
-3.57160509e-01 -2.89930739e-02 8.15837801e-01 7.33121455e-01
7.09260941e-01 1.77584752e-01 9.76034224e-01 -5.45216441e-01
-9.46644723e-01 -3.92432004e-01 2.23563641e-01 2.88508266e-01
2.09502101e-01 -6.75223887e-01 -4.68500316e-01 -8.57535228e-02] | [14.40007209777832, 9.66585636138916] |
122fe8db-f2cb-422a-96bd-160fa7e83a51 | deepskeleton-skeleton-map-for-3d-human-pose | 1711.10796 | null | http://arxiv.org/abs/1711.10796v1 | http://arxiv.org/pdf/1711.10796v1.pdf | DeepSkeleton: Skeleton Map for 3D Human Pose Regression | Despite recent success on 2D human pose estimation, 3D human pose estimation
still remains an open problem. A key challenge is the ill-posed depth ambiguity
nature. This paper presents a novel intermediate feature representation named
skeleton map for regression. It distills structural context from irrelavant
properties of RGB image e.g. illumination and texture. It is simple, clean and
can be easily generated via deconvolution network. For the first time, we show
that training regression network from skeleton map alone is capable of meeting
the performance of state-of-theart 3D human pose estimation works. We further
exploit the power of multiple 3D hypothesis generation to obtain reasonbale 3D
pose in consistent with 2D pose detection. The effectiveness of our approach is
validated on challenging in-the-wild dataset MPII and indoor dataset Human3.6M. | ['xiangyang xue', 'Wei zhang', 'Qingfu Wan'] | 2017-11-29 | null | null | null | null | ['2d-human-pose-estimation'] | ['computer-vision'] | [ 2.75859952e-01 3.08309853e-01 1.26565829e-01 -3.61035913e-01
-7.12131262e-01 -3.25612605e-01 5.11397600e-01 -2.63851702e-01
-6.71107292e-01 7.34056175e-01 7.30741099e-02 1.49383768e-01
-6.08503036e-02 -2.65917480e-01 -8.15045357e-01 -3.74339521e-01
-4.44674119e-02 8.35819185e-01 1.50164887e-01 -4.08176512e-01
1.51155274e-02 6.34967387e-01 -1.62441552e+00 -8.89502242e-02
2.62903571e-01 9.76768851e-01 8.43860302e-03 7.15688884e-01
4.15530533e-01 4.16194886e-01 -5.82861543e-01 -1.09647088e-01
5.36870062e-01 -3.46809357e-01 -6.36666298e-01 2.56979465e-01
6.67852819e-01 -5.45087397e-01 -4.04103220e-01 4.29833442e-01
9.05984223e-01 1.66110471e-02 6.30962610e-01 -1.26169753e+00
7.93838277e-02 -1.89472392e-01 -4.64455485e-01 -6.87170997e-02
7.99898744e-01 3.29376720e-02 6.99507236e-01 -9.82178986e-01
8.74777317e-01 1.33235133e+00 9.76142347e-01 6.52211308e-01
-1.23026133e+00 -3.55349660e-01 1.35302944e-02 -1.49166748e-01
-1.25854909e+00 -3.20051759e-01 6.68137789e-01 -5.04874408e-01
9.46515441e-01 3.39446455e-01 7.44000196e-01 1.47055352e+00
8.99679959e-02 8.27740312e-01 1.54300392e+00 -4.70026493e-01
-9.67524871e-02 -1.39157653e-01 -2.37385929e-01 9.42633331e-01
2.92053759e-01 2.65002280e-01 -9.69838858e-01 4.68823463e-02
1.00618994e+00 -2.42476597e-01 -2.63474077e-01 -7.85989642e-01
-1.28764331e+00 4.13494140e-01 5.47811925e-01 -2.62322903e-01
-2.54655540e-01 4.35448319e-01 8.63185376e-02 -4.29883823e-02
3.53276849e-01 3.22944313e-01 -7.18686163e-01 -2.83202201e-01
-8.05816412e-01 6.70789003e-01 5.79887986e-01 8.28857064e-01
4.88158971e-01 -2.05191806e-01 1.80988163e-01 4.40107316e-01
4.46895808e-01 7.53140032e-01 1.32453471e-01 -1.00679779e+00
4.96065259e-01 3.85795414e-01 2.55567759e-01 -9.22283947e-01
-1.02383900e+00 -4.83931214e-01 -5.52652776e-01 4.86722201e-01
7.03631938e-01 -5.86280935e-02 -1.02305806e+00 1.61401904e+00
6.05845988e-01 -2.71106124e-01 -2.63555408e-01 1.24998057e+00
8.98495615e-01 1.38797075e-01 -2.56898791e-01 2.32319146e-01
1.22400355e+00 -6.24589682e-01 -3.69363993e-01 -4.85146552e-01
6.59781545e-02 -4.01330590e-01 5.68668306e-01 5.52987576e-01
-1.11144972e+00 -6.46884024e-01 -1.12300372e+00 -1.94114134e-01
-1.82314783e-01 1.49299949e-01 6.42587900e-01 6.95186973e-01
-8.05208087e-01 6.84764087e-01 -1.02739501e+00 -7.38240063e-01
2.76434928e-01 5.89012206e-01 -7.88650095e-01 1.50083780e-01
-9.88724113e-01 1.14563298e+00 3.47964048e-01 3.73807997e-01
-7.47620881e-01 -4.86454219e-01 -9.86852705e-01 -6.15059853e-01
5.43048620e-01 -1.05275738e+00 1.07381558e+00 -2.46866897e-01
-1.55908966e+00 1.28991008e+00 -8.83275047e-02 -4.06231821e-01
1.03323066e+00 -8.19833517e-01 1.06669672e-01 3.59517306e-01
8.47163722e-02 9.85228360e-01 9.20795023e-01 -1.47678232e+00
-2.94299334e-01 -8.09251487e-01 -1.42363235e-01 3.64655048e-01
2.45943919e-01 -2.12348208e-01 -5.42918682e-01 -5.19068241e-01
7.36893296e-01 -1.14651752e+00 -3.47996473e-01 2.64415383e-01
-5.41815400e-01 3.63065064e-01 2.13929951e-01 -9.10001338e-01
6.55015111e-01 -1.64807642e+00 6.12950504e-01 3.58974904e-01
2.87739724e-01 -9.47308466e-02 1.71964929e-01 2.57101506e-01
-3.72396223e-02 -2.06560895e-01 -2.08339021e-01 -6.28061593e-01
1.68730929e-01 2.50974268e-01 9.04089585e-03 8.71978164e-01
4.45311159e-01 9.97193813e-01 -6.49129510e-01 -4.89577591e-01
3.87408406e-01 8.06319833e-01 -5.75215995e-01 3.68865252e-01
3.96258794e-02 1.02547252e+00 -2.81648993e-01 8.15073311e-01
5.26050210e-01 -1.09679796e-01 -5.96427135e-02 -3.45216721e-01
2.28621289e-02 1.77237242e-01 -1.24568617e+00 2.18830991e+00
-9.78928953e-02 4.34691489e-01 3.24305543e-03 -6.12620175e-01
7.19571114e-01 1.55739546e-01 5.41011989e-01 -5.22453666e-01
4.03743446e-01 3.37003618e-01 -2.92410672e-01 -3.70884836e-01
4.49373126e-01 -1.61475658e-01 -2.58440018e-01 1.48007810e-01
2.47074232e-01 -3.03770870e-01 -2.99198151e-01 -1.73266590e-01
1.03014517e+00 1.00949967e+00 4.38102633e-01 -4.81617413e-02
2.79772073e-01 -2.41964664e-02 2.80239642e-01 6.15310133e-01
-3.66109699e-01 1.09602046e+00 3.30652475e-01 -4.45098072e-01
-1.03755450e+00 -1.40003753e+00 -1.74446002e-01 8.41565311e-01
1.12953097e-01 -4.22622204e-01 -6.79267466e-01 -6.28090143e-01
2.04937115e-01 1.11397738e-02 -8.18792999e-01 1.36564642e-01
-8.45590830e-01 -7.34069407e-01 6.52966857e-01 5.85598230e-01
4.28336084e-01 -6.94463968e-01 -1.22689605e+00 2.80597210e-02
-3.38822186e-01 -1.39322436e+00 1.75733194e-01 4.51946706e-01
-8.45741868e-01 -1.15372753e+00 -8.51489246e-01 -3.23345274e-01
5.66122472e-01 -1.97936650e-02 1.32144666e+00 -2.69330144e-01
-6.76708758e-01 6.37597442e-01 -2.56998837e-01 -4.16248500e-01
-2.26024762e-02 2.03003902e-02 3.63939762e-01 -3.87924641e-01
5.75466938e-02 -5.94807982e-01 -6.37180269e-01 4.63195533e-01
-2.31163427e-01 4.62541617e-02 5.19449651e-01 7.12353051e-01
8.00233364e-01 -1.37892172e-01 4.07429151e-02 -5.24964094e-01
8.64211321e-02 2.85232402e-02 -3.17424804e-01 -1.41149804e-01
-1.71597302e-01 1.91309109e-01 -9.91877094e-02 -1.67921081e-01
-8.86402726e-01 5.86799562e-01 -2.49832019e-01 -2.58186489e-01
-5.28088391e-01 1.12430900e-02 -1.67343840e-01 -1.52908847e-01
8.09513688e-01 -1.56686783e-01 2.15125099e-01 -6.21484816e-01
2.35245377e-01 3.06627274e-01 8.40425014e-01 -7.07277060e-01
1.03604710e+00 8.32460284e-01 5.73264897e-01 -8.43467891e-01
-8.78738582e-01 -4.98745590e-01 -1.24369049e+00 -4.71547872e-01
9.94238019e-01 -1.07717061e+00 -8.55815291e-01 5.29620588e-01
-1.21269715e+00 -1.97159082e-01 -2.70979293e-02 3.35503727e-01
-8.32974076e-01 3.79016668e-01 -4.36163247e-01 -1.07309699e+00
-1.03960447e-01 -8.63663971e-01 1.58476901e+00 -2.18452871e-01
-6.07652724e-01 -6.14900529e-01 2.67225374e-02 6.39926374e-01
-7.63946027e-02 1.04388905e+00 2.02448085e-01 -1.68998390e-01
-5.29722512e-01 -4.94561523e-01 -5.05386554e-02 1.74092548e-03
-2.20636547e-01 -4.54603583e-01 -1.21721029e+00 -2.24767148e-01
-1.05484806e-01 -5.83230317e-01 8.94441009e-01 2.92695880e-01
7.52255619e-01 2.28141963e-01 -1.83584675e-01 6.58445895e-01
1.07259154e+00 -5.91652453e-01 5.56745112e-01 5.43011010e-01
9.65627015e-01 9.37752545e-01 7.21448779e-01 6.25253081e-01
4.66399521e-01 9.20508087e-01 6.21412635e-01 -1.49985552e-01
-2.86830544e-01 -3.62580210e-01 2.20205441e-01 3.49799991e-01
-5.79980850e-01 1.77143648e-01 -1.04821169e+00 2.42776256e-02
-1.77734089e+00 -5.88910460e-01 -1.86252087e-01 2.10403895e+00
4.72764760e-01 4.47093397e-01 3.86928827e-01 4.75776017e-01
2.39964068e-01 -1.22515047e-02 -3.18408996e-01 7.25765750e-02
-1.17119789e-01 4.35842216e-01 6.46531641e-01 4.48326826e-01
-1.24976063e+00 6.32645249e-01 6.20779610e+00 2.67088681e-01
-6.37680650e-01 -9.17398836e-03 1.24744304e-01 -2.43272930e-01
1.21352084e-01 -2.07853183e-01 -8.90237033e-01 -2.41648369e-02
4.30820018e-01 6.98365092e-01 1.54819846e-01 7.10921288e-01
1.56710390e-02 -3.16884637e-01 -1.27638638e+00 1.31803644e+00
3.12026620e-01 -6.37912810e-01 -3.83586794e-01 1.51287556e-01
4.02871877e-01 -1.60356041e-03 5.84289432e-02 1.30991742e-01
-2.33561501e-01 -1.21871364e+00 1.04864383e+00 6.02664411e-01
8.75832677e-01 -6.42947078e-01 6.14691794e-01 5.80514967e-01
-1.06453907e+00 1.86233968e-01 2.35346630e-02 -2.63398468e-01
2.60810018e-01 4.12248731e-01 -9.02452528e-01 5.56475103e-01
7.88919508e-01 5.52689016e-01 -8.45425963e-01 8.87638271e-01
-4.58527803e-01 -8.03702176e-02 -5.54919422e-01 2.53278285e-01
-9.84180346e-02 1.78402662e-01 5.95747292e-01 1.15275693e+00
1.68629065e-01 -3.61647792e-02 2.86597401e-01 6.14149213e-01
2.22021952e-01 -1.91146448e-01 -5.84695935e-01 2.79007584e-01
-8.17332864e-02 9.97051656e-01 -8.65475953e-01 1.65226564e-01
5.70904613e-02 1.39979446e+00 3.10274035e-01 1.83104366e-01
-8.38928580e-01 2.76749637e-02 4.59038764e-01 2.93517172e-01
2.94379652e-01 -5.82499683e-01 -2.78334528e-01 -1.15635109e+00
3.44775707e-01 -7.32336283e-01 1.41153440e-01 -8.05767655e-01
-1.03398681e+00 4.71119314e-01 1.75028577e-01 -1.14321494e+00
-5.01384318e-01 -1.19324386e+00 1.45515025e-01 6.16447449e-01
-1.28849435e+00 -1.31140089e+00 -5.81027031e-01 4.65586632e-01
2.24894851e-01 1.89216420e-01 1.02721369e+00 5.49378619e-02
-3.58092606e-01 3.65422636e-01 -3.94499213e-01 5.47406189e-02
6.66232884e-01 -1.42740095e+00 5.64719737e-01 6.68430567e-01
8.76350701e-02 5.16949832e-01 9.85253632e-01 -6.63628042e-01
-1.64037764e+00 -5.33502817e-01 7.64211595e-01 -1.25891066e+00
1.99946254e-01 -7.47042000e-01 -2.89688408e-01 4.87421602e-01
-4.68978167e-01 7.34780822e-03 4.48153019e-01 9.49681103e-02
-4.59909707e-01 1.26617894e-01 -1.17317283e+00 4.02788788e-01
1.36353767e+00 -4.87265289e-01 -7.13805974e-01 2.11068690e-01
3.65865827e-01 -9.79041696e-01 -7.55448818e-01 9.01467621e-01
9.22829032e-01 -1.27280140e+00 1.44513547e+00 -3.35114270e-01
2.50145257e-01 -1.78810492e-01 -4.82739210e-01 -8.00460100e-01
1.38624400e-01 -6.22371495e-01 -4.49637562e-01 6.74318075e-01
5.67036234e-02 -1.35258913e-01 1.20485103e+00 5.35989702e-01
1.77788556e-01 -6.14555061e-01 -1.08286893e+00 -8.04936945e-01
-1.30508378e-01 -6.66187465e-01 1.21936351e-01 4.32885677e-01
-3.54757935e-01 3.27613652e-01 -6.74229085e-01 2.04581201e-01
1.07499647e+00 -1.57683611e-01 1.24542117e+00 -1.56001604e+00
-5.82239211e-01 -3.06790601e-02 -7.73914337e-01 -1.22341549e+00
9.50361490e-02 -4.49493587e-01 2.92810142e-01 -1.40064418e+00
-6.62434846e-02 -1.25832736e-01 5.22745661e-02 3.46205056e-01
5.31215556e-02 6.60613239e-01 2.61007965e-01 7.46713504e-02
-5.96561670e-01 4.13889915e-01 1.12761247e+00 2.98514664e-01
-7.51550961e-03 -6.64244639e-03 -2.06963435e-01 9.20870304e-01
7.21489370e-01 -3.49161237e-01 -1.60747483e-01 -2.65164435e-01
3.46975535e-01 9.55851376e-02 9.00217116e-01 -1.36464357e+00
-1.10921502e-01 2.67558217e-01 8.91855121e-01 -9.69125330e-01
9.46446061e-01 -8.77637088e-01 1.95916116e-01 6.22299135e-01
-2.27125851e-03 -3.59410159e-02 1.50343299e-01 5.98425508e-01
2.18676552e-01 1.80589378e-01 4.88770992e-01 -4.73912090e-01
-8.11223328e-01 1.42545640e-01 1.78041399e-01 1.76974788e-01
7.13229775e-01 -5.76819301e-01 1.74154878e-01 -1.88674018e-01
-8.33315849e-01 -3.04311868e-02 5.46500385e-01 3.90917093e-01
7.78895378e-01 -1.26735532e+00 -7.61250436e-01 3.32207561e-01
1.85311437e-01 2.06577405e-01 2.22216677e-02 8.98055494e-01
-7.22778738e-01 4.34848160e-01 -3.36719960e-01 -9.47382033e-01
-1.28093529e+00 2.32434466e-01 3.65376085e-01 -6.47042096e-02
-6.55731797e-01 9.62081313e-01 -5.12939505e-02 -7.00803638e-01
4.47013587e-01 -1.75163805e-01 1.65382057e-01 -7.68469349e-02
2.53489256e-01 5.34430921e-01 1.25832185e-01 -1.01586139e+00
-6.83280766e-01 9.22633708e-01 4.06884670e-01 -4.08534378e-01
1.39986753e+00 -2.16645017e-01 4.51042354e-02 5.34277856e-01
1.18055725e+00 -8.57262090e-02 -1.54807866e+00 1.07778005e-01
-1.14866741e-01 -5.99794984e-01 -3.07991415e-01 -8.54687154e-01
-5.86221039e-01 9.75115597e-01 7.61465847e-01 -4.19822246e-01
8.63278508e-01 1.48719877e-01 6.05636120e-01 5.95830798e-01
7.34421372e-01 -1.13896680e+00 3.40341717e-01 4.05077696e-01
9.60801423e-01 -1.49701083e+00 4.04581457e-01 -5.75003326e-01
-2.86612988e-01 1.04591012e+00 6.29147708e-01 -2.35129416e-01
4.32165533e-01 2.18972534e-01 8.07354972e-02 -3.47629160e-01
-1.40499994e-01 -4.76359427e-01 5.67529917e-01 8.91259313e-01
3.97733569e-01 -4.31578420e-02 3.00050452e-02 4.33064580e-01
-4.85337943e-01 -1.77835360e-01 7.69965649e-02 1.18152940e+00
-3.77508193e-01 -1.01889491e+00 -7.88417757e-01 -1.77107573e-01
-4.24319923e-01 2.43738666e-01 -5.78835607e-01 1.11103094e+00
2.67912656e-01 6.81852579e-01 -2.71875471e-01 -4.68439847e-01
6.36928141e-01 6.90598562e-02 1.18202329e+00 -4.13684309e-01
-4.03299123e-01 1.48344874e-01 2.24292859e-01 -8.77675056e-01
-6.93431795e-01 -6.89114213e-01 -1.15224886e+00 -1.07284822e-02
-1.87347993e-01 -4.76874679e-01 9.57960069e-01 9.46937382e-01
7.88491517e-02 3.55935246e-01 9.26761478e-02 -1.60883796e+00
-4.82697994e-01 -6.33897364e-01 -4.86844301e-01 5.05034447e-01
5.31025529e-01 -1.07711601e+00 -2.67656803e-01 -8.76763687e-02] | [6.956085205078125, -0.9579131603240967] |
7600da50-4b33-4635-b367-a5763a49cd96 | what-is-your-metric-telling-you-evaluating | 2205.11454 | null | https://arxiv.org/abs/2205.11454v1 | https://arxiv.org/pdf/2205.11454v1.pdf | What is Your Metric Telling You? Evaluating Classifier Calibration under Context-Specific Definitions of Reliability | Classifier calibration has received recent attention from the machine learning community due both to its practical utility in facilitating decision making, as well as the observation that modern neural network classifiers are poorly calibrated. Much of this focus has been towards the goal of learning classifiers such that their output with largest magnitude (the "predicted class") is calibrated. However, this narrow interpretation of classifier outputs does not adequately capture the variety of practical use cases in which classifiers can aid in decision making. In this work, we argue that more expressive metrics must be developed that accurately measure calibration error for the specific context in which a classifier will be deployed. To this end, we derive a number of different metrics using a generalization of Expected Calibration Error (ECE) that measure calibration error under different definitions of reliability. We then provide an extensive empirical evaluation of commonly used neural network architectures and calibration techniques with respect to these metrics. We find that: 1) definitions of ECE that focus solely on the predicted class fail to accurately measure calibration error under a selection of practically useful definitions of reliability and 2) many common calibration techniques fail to improve calibration performance uniformly across ECE metrics derived from these diverse definitions of reliability. | ['Eric Heim', 'Jacob Oaks', 'John Kirchenbauer'] | 2022-05-23 | null | null | null | null | ['classifier-calibration', 'classifier-calibration'] | ['computer-vision', 'miscellaneous'] | [ 2.47496977e-01 -6.46270290e-02 -1.93387225e-01 -9.96157587e-01
-6.93144321e-01 -5.91432095e-01 4.70675498e-01 4.10491914e-01
-4.00790900e-01 6.75747991e-01 -1.00675762e-01 -5.99071562e-01
-4.15463507e-01 -7.01303959e-01 -4.23288345e-01 -4.88197893e-01
2.55129248e-01 2.87040681e-01 -1.65603608e-01 -1.20951757e-01
3.07371378e-01 6.05100214e-01 -1.45649123e+00 6.22345917e-02
8.60386133e-01 1.35516882e+00 -3.03177029e-01 4.82772112e-01
1.49341598e-01 7.25336134e-01 -9.08232450e-01 -6.58922315e-01
1.40844718e-01 -3.61526161e-01 -7.43992984e-01 -3.54993701e-01
5.34960151e-01 6.03047237e-02 6.88334852e-02 9.66234684e-01
3.01226079e-01 -2.01361865e-01 9.14595068e-01 -1.48693013e+00
-5.96231520e-01 8.08010936e-01 -6.30111694e-02 2.86818773e-01
8.32392573e-02 -2.34202191e-01 1.15279019e+00 -6.72634184e-01
6.30255118e-02 9.66505706e-01 1.15875244e+00 5.03570795e-01
-1.19814157e+00 -8.34814608e-01 2.17167333e-01 -1.04005933e-01
-1.33610487e+00 -2.83064753e-01 6.40448511e-01 -5.31120896e-01
6.78359032e-01 3.01400423e-01 2.10621983e-01 1.05543602e+00
1.55671403e-01 3.96673560e-01 1.07967103e+00 -6.39663219e-01
4.49925423e-01 5.52318096e-01 6.96104646e-01 2.28121683e-01
7.63757169e-01 2.82637358e-01 -2.84129351e-01 -1.34358048e-01
4.40818161e-01 -3.01872611e-01 -2.41699517e-01 -2.61240780e-01
-8.27244639e-01 1.08105028e+00 3.76594394e-01 4.37697351e-01
-8.83096308e-02 1.59259319e-01 3.22248280e-01 4.52524185e-01
3.17846686e-01 7.87839353e-01 -5.62573910e-01 -1.12372115e-01
-7.21834660e-01 1.15695417e-01 9.73489821e-01 9.59938705e-01
6.22733116e-01 1.64636120e-01 1.37300178e-01 1.00682843e+00
-3.79459560e-02 3.22190851e-01 4.26401794e-01 -1.00513208e+00
3.66121650e-01 5.42326987e-01 1.06931351e-01 -8.99816751e-01
-4.69344407e-01 -7.84977615e-01 -8.20526779e-01 2.11183563e-01
4.97461528e-01 -1.65531084e-01 -4.49147791e-01 1.97086573e+00
-1.12567350e-01 -3.11924189e-01 3.00025672e-01 6.11088037e-01
5.51141560e-01 6.57393858e-02 2.08416581e-01 -3.82040553e-02
1.02537751e+00 -4.56996799e-01 -2.78264016e-01 -5.24591327e-01
7.02933609e-01 -6.13778889e-01 1.04276550e+00 3.70690644e-01
-7.24234581e-01 -6.71445012e-01 -1.60664976e+00 2.39650026e-01
-5.05164444e-01 -9.20042675e-03 7.28660822e-01 1.23247302e+00
-7.81624258e-01 6.76921010e-01 -4.95930254e-01 -2.41769150e-01
2.72627056e-01 2.66244084e-01 -2.06747711e-01 -2.26543713e-02
-1.10181499e+00 1.36573029e+00 4.93846744e-01 2.30262671e-02
-3.95388305e-01 -5.83631516e-01 -5.69509387e-01 7.39780888e-02
7.74057135e-02 -6.14425182e-01 1.59965432e+00 -1.45927072e+00
-9.60994184e-01 6.24094844e-01 2.46454760e-01 -5.51306844e-01
5.34213722e-01 -2.94146419e-01 -5.45786738e-01 -3.30385238e-01
-7.87199438e-02 4.15132046e-01 4.16431040e-01 -1.32237494e+00
-9.40551341e-01 -2.01003253e-01 1.79250225e-01 2.53104158e-02
-5.42688251e-01 -8.04802924e-02 1.83979850e-02 -6.32168710e-01
6.19946979e-02 -8.81923318e-01 1.38751045e-01 -1.63270116e-01
-7.44181871e-02 -2.56158352e-01 5.96382082e-01 -1.52786255e-01
1.42299974e+00 -2.04790473e+00 -1.95590422e-01 4.05013770e-01
2.04300582e-01 1.92524418e-01 4.11943020e-03 2.47542784e-01
-3.59244078e-01 1.61608636e-01 -4.72188979e-01 -5.97418211e-02
-4.91007380e-02 3.49602550e-01 -5.08274913e-01 3.30505431e-01
3.54070038e-01 6.45606935e-01 -7.32940018e-01 -2.28284016e-01
7.44917169e-02 5.13734281e-01 -3.01352799e-01 2.41357580e-01
2.33185887e-01 3.30308676e-02 -2.54478306e-01 6.09620154e-01
3.52497131e-01 -3.02081883e-01 1.92122519e-01 -1.59627587e-01
1.79309502e-01 1.00830242e-01 -9.57796812e-01 1.02032340e+00
-5.37261546e-01 7.06970811e-01 -2.89937109e-01 -1.20924413e+00
1.13000941e+00 9.79223847e-02 3.52187723e-01 -3.66169512e-01
3.52266222e-01 4.01699841e-01 3.00915420e-01 -7.89341778e-02
2.68452495e-01 -1.88765526e-01 -2.05827817e-01 4.49201465e-01
-1.95653103e-02 -3.52025665e-02 -5.19604608e-02 -8.78003687e-02
1.04396677e+00 -1.48031294e-01 5.19437492e-01 -3.78699869e-01
4.09205467e-01 2.54609510e-02 6.35590434e-01 9.25099492e-01
-2.75497496e-01 5.09297252e-01 4.49405819e-01 -3.46593857e-01
-1.26719272e+00 -8.81925523e-01 -5.67766964e-01 9.69750583e-01
-2.74027944e-01 -4.81104702e-02 -8.54725957e-01 -5.89319408e-01
3.06900024e-01 8.63253951e-01 -7.90166318e-01 -3.69253516e-01
-4.12423998e-01 -9.07103896e-01 8.29467893e-01 1.07280350e+00
4.06489909e-01 -6.56279206e-01 -9.91095364e-01 8.82505551e-02
3.07394955e-02 -1.02267194e+00 -8.00458714e-02 5.64460635e-01
-1.05807936e+00 -1.45124543e+00 -4.46125716e-01 -5.44798791e-01
5.66480517e-01 1.75663918e-01 1.45258033e+00 1.88558996e-01
1.98935121e-01 5.76251864e-01 -4.13342476e-01 -6.75345898e-01
-6.75783217e-01 2.66730785e-01 1.57447249e-01 -3.39069247e-01
8.55793417e-01 -4.76823509e-01 -2.41025209e-01 5.67475080e-01
-7.55529761e-01 -2.30464444e-01 8.09732139e-01 9.84375954e-01
1.31292716e-01 -1.21485233e-01 1.00242805e+00 -1.18171954e+00
1.10175812e+00 -6.04434669e-01 -5.56592345e-01 5.91759801e-01
-1.36422467e+00 1.25368044e-01 5.66836298e-01 -5.86755931e-01
-7.47413933e-01 -1.95691615e-01 -5.27483039e-02 -2.54093260e-01
-9.73045751e-02 7.03711510e-01 5.07176993e-03 -1.00094356e-01
1.09881794e+00 -2.07495481e-01 -9.89331827e-02 -2.13071987e-01
1.44725703e-02 9.18945551e-01 7.25549936e-01 -8.48495424e-01
5.58256507e-01 -1.62592888e-01 7.67199993e-02 -3.68623435e-01
-1.04448450e+00 -2.84199089e-01 -6.22766495e-01 -7.70019665e-02
3.23029071e-01 -6.79032505e-01 -5.69914639e-01 2.54051805e-01
-8.53963971e-01 -1.71042413e-01 -1.38862148e-01 4.02898967e-01
-4.94776756e-01 2.09673062e-01 -2.54173368e-01 -9.58331645e-01
-3.88177872e-01 -1.12527382e+00 5.39034843e-01 1.32263511e-01
-5.57397068e-01 -1.08271694e+00 -1.15629926e-01 -3.18818749e-03
5.55279136e-01 3.94925058e-01 1.03952384e+00 -8.48142207e-01
4.13376018e-02 -5.69499195e-01 -2.49581441e-01 9.05709267e-01
8.17072093e-02 1.16532922e-01 -1.23476064e+00 -2.37875178e-01
6.97713718e-02 -2.58486837e-01 6.00734651e-01 3.81961673e-01
1.13328779e+00 -1.11681640e-01 -1.91854104e-01 4.30726290e-01
1.46703541e+00 3.91866088e-01 3.47742319e-01 3.78497541e-01
2.62921661e-01 7.90519834e-01 4.49468046e-01 3.04038096e-02
1.24274150e-01 5.24228692e-01 1.36860758e-01 2.39923343e-01
2.63446003e-01 -9.17052105e-02 -8.00563302e-03 7.36318648e-01
1.98905319e-02 -1.63481101e-01 -1.20192564e+00 2.81742483e-01
-1.82403648e+00 -6.58730388e-01 5.07529378e-02 2.52321267e+00
7.52628982e-01 5.11295557e-01 1.72990575e-01 6.02868736e-01
7.78411984e-01 -3.60837728e-01 -6.07289314e-01 -4.65924472e-01
-1.76458254e-01 1.69401482e-01 5.87066412e-01 3.87028873e-01
-9.01859164e-01 3.30062032e-01 7.70025682e+00 3.91813368e-01
-9.99483168e-01 3.00512183e-02 1.00161409e+00 2.85718232e-01
-1.00804739e-01 -6.48855465e-03 -8.08973074e-01 3.61578375e-01
1.27431726e+00 -1.57929778e-01 6.03065975e-02 1.21511257e+00
-3.62340897e-01 -1.75188910e-02 -1.70264220e+00 1.03757751e+00
-3.33878174e-02 -9.55501139e-01 -2.50955731e-01 -1.15838446e-01
5.97872257e-01 -2.93206275e-01 2.67945290e-01 3.59483361e-01
4.21376705e-01 -1.24541461e+00 8.24262917e-01 4.17961389e-01
8.93335879e-01 -8.53691876e-01 9.21974480e-01 2.27534413e-01
-7.80715346e-01 -5.35989463e-01 -4.74229723e-01 -1.32916003e-01
-3.75072330e-01 7.45838940e-01 -7.49819577e-01 2.09264815e-01
6.18272781e-01 3.48783344e-01 -6.84284866e-01 9.29524183e-01
-1.16726488e-01 7.00441003e-01 -2.11890012e-01 -7.96879753e-02
1.32926688e-01 4.69981283e-02 -2.66831126e-02 1.35613334e+00
2.95909733e-01 2.06856313e-03 -3.43495250e-01 7.62236774e-01
7.21159065e-03 -2.15523541e-01 -4.66062307e-01 9.66816023e-02
1.02528763e+00 9.57292020e-01 -4.43390727e-01 -2.67674416e-01
-5.14637828e-01 2.39019915e-01 4.70071614e-01 2.30928078e-01
-7.23440230e-01 -4.11876291e-01 6.21849179e-01 -1.20292850e-01
-1.77164301e-01 -1.39577404e-01 -7.75619090e-01 -9.34320927e-01
1.59196824e-01 -8.51820529e-01 5.24593353e-01 -6.65340424e-01
-1.60150278e+00 6.71005547e-01 3.10318261e-01 -1.20729005e+00
-4.28241611e-01 -8.20109248e-01 -4.64929730e-01 9.34440315e-01
-1.15523839e+00 -8.39817166e-01 -6.79505885e-01 1.21722087e-01
1.62259340e-01 -2.58258581e-01 1.00693345e+00 2.97652990e-01
-5.89062512e-01 1.06792402e+00 7.94591531e-02 2.57210910e-01
8.11009288e-01 -1.24093914e+00 3.75017136e-01 7.83974886e-01
-1.71965957e-02 6.84452713e-01 8.79352987e-01 -3.52853954e-01
-9.17110085e-01 -9.51054394e-01 7.70381987e-01 -9.34369266e-01
3.96492124e-01 -7.48116523e-03 -8.82574320e-01 8.21093380e-01
-2.69873589e-01 -1.72639802e-01 8.47031891e-01 4.65112597e-01
-8.91412318e-01 -5.15590012e-01 -1.13891125e+00 2.42027849e-01
8.24388921e-01 -4.68627781e-01 -6.43392503e-01 1.35642337e-02
3.87989968e-01 -2.19499826e-01 -1.24017799e+00 5.67186117e-01
7.74340510e-01 -1.08601606e+00 8.45560193e-01 -6.08422995e-01
4.19145107e-01 5.23392297e-02 -5.23627698e-01 -1.23240912e+00
-3.15428466e-01 -4.59294394e-02 -6.70126006e-02 1.19773257e+00
5.60856819e-01 -9.78005588e-01 6.18580699e-01 1.16058886e+00
-1.73349544e-01 -9.91383553e-01 -8.20570886e-01 -8.28705370e-01
5.25756717e-01 -6.31380200e-01 7.77363896e-01 1.09435558e+00
-1.30863920e-01 3.99295300e-01 -1.30797047e-02 -8.00990984e-02
5.13524830e-01 -2.23480880e-01 6.59828067e-01 -1.63486528e+00
-1.63836703e-01 -8.32473576e-01 -5.73629141e-01 -6.63014889e-01
7.48352322e-04 -4.67932791e-01 -7.56260827e-02 -9.87501144e-01
2.02545166e-01 -9.89503980e-01 -5.57738125e-01 4.92241293e-01
-1.72151312e-01 9.71765071e-02 -8.01620334e-02 4.50532377e-01
2.52361502e-02 -1.95473246e-02 4.13801730e-01 -3.11694369e-02
9.17112455e-02 1.25659868e-01 -1.21958959e+00 6.35061324e-01
8.46267223e-01 -4.21722978e-01 -5.28843582e-01 -4.20028567e-01
4.05154645e-01 -1.24394007e-01 2.65351474e-01 -1.40176868e+00
1.61058694e-01 -2.18005046e-01 5.86551428e-01 -1.39566854e-01
1.27964646e-01 -9.49616492e-01 3.37082624e-01 3.21269333e-01
-6.93912089e-01 4.64746952e-01 3.53136659e-02 5.41052461e-01
-2.88254142e-01 -5.21955907e-01 9.65536296e-01 1.25928700e-01
-5.62471032e-01 -4.52353880e-02 5.06859086e-02 3.60025056e-02
9.66358721e-01 -4.46421981e-01 -4.31174159e-01 -3.40612471e-01
-2.82116711e-01 -5.15728593e-02 5.25265455e-01 4.54557955e-01
3.52438837e-01 -1.32446396e+00 -6.60718620e-01 -7.20131546e-02
4.47186261e-01 -2.11482689e-01 -3.72474968e-01 5.99303961e-01
-3.33055019e-01 6.34144485e-01 -1.96748018e-01 -5.90713739e-01
-9.75210071e-01 4.78200972e-01 6.12239599e-01 6.03793375e-02
-2.66964227e-01 7.44294524e-01 -5.57420850e-02 -2.82040983e-01
4.02172416e-01 -4.13239092e-01 -7.44485408e-02 -7.69220069e-02
3.81096452e-01 3.27540755e-01 4.28700805e-01 -5.94694614e-01
-2.55354524e-01 3.31316620e-01 2.77125519e-02 4.35886309e-02
1.04972768e+00 1.58824727e-01 2.66445994e-01 6.06630862e-01
9.88350093e-01 -4.68943149e-01 -1.07214153e+00 -1.70798630e-01
2.90017754e-01 -3.45757306e-01 -1.91868190e-02 -1.08186865e+00
-9.41701353e-01 9.00548339e-01 6.65053368e-01 3.46723855e-01
1.26554036e+00 -4.27862823e-01 2.73068696e-01 6.45635664e-01
4.28360224e-01 -1.22254062e+00 -1.63678363e-01 3.62255722e-01
7.09352911e-01 -1.10394883e+00 5.36475256e-02 -1.96981937e-01
-6.71777785e-01 1.34782684e+00 7.05411911e-01 -4.58532805e-03
6.69522166e-01 1.48362994e-01 2.08133027e-01 1.11413106e-01
-5.86726785e-01 2.34012559e-01 2.61662841e-01 7.25087047e-01
7.11167157e-01 1.57243192e-01 -4.06884789e-01 7.14493096e-01
-5.01249135e-01 -6.04197569e-02 4.32230860e-01 8.88096511e-01
-5.51553369e-01 -1.11314380e+00 -3.45796525e-01 7.04726458e-01
-3.36201906e-01 1.57234207e-01 -4.41957325e-01 8.82173657e-01
4.25490318e-03 1.17087770e+00 8.19508657e-02 -6.67260170e-01
3.50203812e-01 1.74701273e-01 3.44763517e-01 -3.49965662e-01
-6.55823469e-01 -4.76442367e-01 2.09091321e-01 -2.08743513e-01
-5.99000335e-01 -5.54087758e-01 -6.88099980e-01 -5.30374467e-01
-5.37351251e-01 3.23281884e-01 7.93817759e-01 9.77521002e-01
7.44417012e-02 2.98677653e-01 4.98115867e-01 -3.27286750e-01
-1.20550776e+00 -1.08648944e+00 -4.27686661e-01 4.54189211e-01
1.70053393e-01 -8.81788909e-01 -3.24974388e-01 -1.68135002e-01] | [8.589374542236328, 4.3698577880859375] |
2ce230cd-a85a-4cc3-870b-81d7154026a6 | stargan-v2-diverse-image-synthesis-for | 1912.01865 | null | https://arxiv.org/abs/1912.01865v2 | https://arxiv.org/pdf/1912.01865v2.pdf | StarGAN v2: Diverse Image Synthesis for Multiple Domains | A good image-to-image translation model should learn a mapping between different visual domains while satisfying the following properties: 1) diversity of generated images and 2) scalability over multiple domains. Existing methods address either of the issues, having limited diversity or multiple models for all domains. We propose StarGAN v2, a single framework that tackles both and shows significantly improved results over the baselines. Experiments on CelebA-HQ and a new animal faces dataset (AFHQ) validate our superiority in terms of visual quality, diversity, and scalability. To better assess image-to-image translation models, we release AFHQ, high-quality animal faces with large inter- and intra-domain differences. The code, pretrained models, and dataset can be found at https://github.com/clovaai/stargan-v2. | ['Jung-Woo Ha', 'Jaejun Yoo', 'Yunjey Choi', 'Youngjung Uh'] | 2019-12-04 | stargan-v2-diverse-image-synthesis-for-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Choi_StarGAN_v2_Diverse_Image_Synthesis_for_Multiple_Domains_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Choi_StarGAN_v2_Diverse_Image_Synthesis_for_Multiple_Domains_CVPR_2020_paper.pdf | cvpr-2020-6 | ['fundus-to-angiography-generation', 'multimodal-unsupervised-image-to-image'] | ['computer-vision', 'computer-vision'] | [ 7.39113381e-03 -2.55295068e-01 -2.60004133e-01 -5.34082770e-01
-7.61210561e-01 -7.29566395e-01 7.00714707e-01 -6.19199216e-01
-1.90157015e-02 5.88301480e-01 9.43608880e-02 -1.91667959e-01
3.62306833e-01 -6.26716554e-01 -7.59844363e-01 -3.95696104e-01
5.17231487e-02 4.39083546e-01 1.86345994e-01 -1.84962526e-01
-5.12501076e-02 3.37794214e-01 -1.35302556e+00 4.63088959e-01
7.49880373e-01 1.00042343e+00 1.56774476e-01 5.57611883e-01
1.81399733e-01 7.04992115e-01 -3.34374726e-01 -9.17470694e-01
5.70618629e-01 -8.07586849e-01 -7.15225756e-01 8.26566964e-02
9.48688805e-01 -5.82050502e-01 -3.29210639e-01 1.05530024e+00
6.58235788e-01 -3.94948065e-01 5.41952550e-01 -1.86113250e+00
-1.26685011e+00 1.85598075e-01 -6.87641263e-01 1.33490292e-02
2.77184397e-01 5.56542456e-01 8.13161254e-01 -9.79485631e-01
9.44676101e-01 1.46536696e+00 5.00679731e-01 8.15321565e-01
-1.38929474e+00 -1.00322425e+00 -2.51009315e-01 6.46168441e-02
-1.47083902e+00 -9.34666395e-01 4.19097364e-01 -5.03614783e-01
6.45002723e-01 2.25006789e-02 4.05484170e-01 1.40898645e+00
1.11272231e-01 5.56280792e-01 1.30797398e+00 -8.66562873e-02
1.32111907e-02 2.48255089e-01 -5.64702213e-01 6.47848964e-01
2.45121226e-01 3.09684783e-01 -5.93072414e-01 -1.26692533e-01
8.99262846e-01 -4.36540663e-01 -2.25234374e-01 -5.36261618e-01
-1.20465136e+00 7.96841860e-01 4.75711584e-01 2.42160931e-02
-1.45005494e-01 2.42614880e-01 3.57663006e-01 5.73797762e-01
3.90681237e-01 3.54021430e-01 -3.81332606e-01 1.93893775e-01
-8.18419278e-01 4.31016922e-01 4.96856689e-01 1.28049290e+00
7.92757988e-01 2.56547123e-01 -2.43241861e-01 8.23520482e-01
2.14227736e-01 9.98562813e-01 1.40366822e-01 -1.31821287e+00
4.33697432e-01 1.47392318e-01 1.75029740e-01 -1.13243949e+00
2.12338697e-02 -2.29645252e-01 -7.95317531e-01 2.97489047e-01
3.38738680e-01 -1.26529902e-01 -1.06447828e+00 2.14340973e+00
3.19263399e-01 1.74783915e-02 1.22576118e-01 9.28868234e-01
1.28012443e+00 6.80284917e-01 1.58215940e-01 7.93865100e-02
1.54357946e+00 -1.07060742e+00 -5.87518573e-01 -4.94930416e-01
2.54515499e-01 -1.00062156e+00 1.25288045e+00 -4.64217141e-02
-1.33148146e+00 -6.90561533e-01 -9.10757005e-01 -1.72844172e-01
-1.56412497e-01 3.21874171e-01 2.94086069e-01 5.27411282e-01
-1.56337106e+00 1.52477309e-01 -3.69311839e-01 -5.71006536e-01
7.23016739e-01 2.77298748e-01 -6.35620415e-01 -2.50987709e-01
-1.02102566e+00 7.22823858e-01 1.80325583e-01 -5.51262975e-01
-1.17770362e+00 -7.58750796e-01 -8.57975662e-01 -1.39068365e-01
7.60393515e-02 -8.69376421e-01 1.37643218e+00 -1.40276241e+00
-1.23951232e+00 1.30411720e+00 -4.44546081e-02 -2.15976849e-01
7.75469899e-01 1.26549482e-01 -5.95932484e-01 3.04393649e-01
4.20556456e-01 1.37648642e+00 9.15194452e-01 -1.42402411e+00
-4.34317023e-01 -2.33389646e-01 -1.72094535e-02 5.37018366e-02
-2.20313147e-01 4.33762372e-02 -6.17794275e-01 -6.26988649e-01
-4.59327608e-01 -1.10575271e+00 2.29850382e-01 2.48259827e-01
-1.27324894e-01 7.90696144e-02 7.68384159e-01 -7.98830569e-01
7.26719618e-01 -2.37094283e+00 2.45680800e-03 -2.14315191e-01
1.20475203e-01 3.75148654e-01 -9.51761484e-01 3.01491618e-01
-1.89846620e-01 1.76728204e-01 -2.30270952e-01 -3.90721053e-01
-1.65475816e-01 -3.52839194e-02 -8.88430104e-02 4.13789928e-01
4.68559563e-01 1.23156369e+00 -7.92642713e-01 -5.64670026e-01
-1.48293778e-01 4.98517513e-01 -5.26690960e-01 3.08320850e-01
-1.88150853e-01 5.62921524e-01 -6.76035732e-02 7.15905130e-01
1.11764789e+00 -5.16920507e-01 3.08291376e-01 -3.01670611e-01
2.77829349e-01 -6.99277371e-02 -8.47099423e-01 1.63493931e+00
-2.16712654e-01 7.05051422e-01 1.48902178e-01 -3.43563527e-01
8.50917816e-01 3.09149593e-01 3.56478304e-01 -1.20143998e+00
-5.62417991e-02 3.64942133e-01 -1.07451238e-01 -3.68598104e-01
3.18748742e-01 -3.69450226e-02 1.38449639e-01 2.94812888e-01
3.35062087e-01 -1.59178197e-01 2.84730107e-01 3.21363896e-01
8.28254223e-01 1.48905158e-01 2.51471788e-01 -3.24317604e-01
2.47288257e-01 6.34293929e-02 6.38244212e-01 4.42468971e-01
-3.76159698e-01 9.73830462e-01 4.92688119e-01 -3.93682390e-01
-1.45289600e+00 -1.14233923e+00 7.42034987e-02 1.02954996e+00
2.19065204e-01 -3.35391313e-01 -8.32937360e-01 -6.72922254e-01
5.44240735e-02 5.24888575e-01 -6.93246603e-01 2.61926409e-02
-3.84960860e-01 -3.28243613e-01 6.97096050e-01 2.61726916e-01
8.33242953e-01 -1.07492280e+00 -4.75099653e-01 -2.32105717e-01
-5.31971395e-01 -1.40456247e+00 -7.83157945e-01 -5.50912201e-01
-5.47454596e-01 -1.10981572e+00 -8.50410700e-01 -9.21320379e-01
5.88663042e-01 6.46070123e-01 1.56291449e+00 1.13163516e-01
-1.20378986e-01 3.03312063e-01 -2.85233468e-01 -2.08554029e-01
-6.65581942e-01 -8.65996932e-04 -8.00186470e-02 -2.07032651e-01
1.56080157e-01 -4.29991096e-01 -6.93483293e-01 5.38026035e-01
-9.89992261e-01 3.05271208e-01 6.15777850e-01 7.34593153e-01
4.82090324e-01 -3.12605143e-01 5.02595186e-01 -7.32327640e-01
4.27329719e-01 -5.16194284e-01 -6.92990482e-01 3.45398009e-01
-5.54937899e-01 -1.91190049e-01 3.74136358e-01 -2.75961727e-01
-9.88643467e-01 1.05754241e-01 -1.78415496e-02 -4.41698700e-01
-1.80005819e-01 1.15059204e-01 -3.69704753e-01 -1.41080424e-01
8.15951109e-01 2.41217092e-01 2.45061904e-01 -1.70729920e-01
5.51477134e-01 6.04133189e-01 5.71436763e-01 -5.30533612e-01
9.06065047e-01 4.72499311e-01 -3.22848976e-01 -5.46416819e-01
-3.95847917e-01 -2.97699142e-02 -3.84236306e-01 -6.25891238e-02
7.89236903e-01 -1.45167518e+00 -3.10077995e-01 5.50141335e-01
-1.16861629e+00 -6.14258170e-01 -7.07238391e-02 1.39854580e-01
-6.70165062e-01 1.97236508e-01 -5.15533090e-01 -2.98783123e-01
-4.55480605e-01 -1.13909161e+00 1.16156054e+00 1.77511618e-01
-1.13940954e-01 -7.58739710e-01 4.82226573e-02 4.99141395e-01
5.93805969e-01 3.14050198e-01 6.60770237e-01 -2.09507331e-01
-5.99687815e-01 2.85639793e-01 -5.76957941e-01 4.12311137e-01
1.23046070e-01 9.29564685e-02 -7.56243110e-01 -7.02393174e-01
-3.20037454e-01 -6.98404253e-01 7.38858402e-01 2.81513751e-01
7.46778727e-01 -5.64580858e-01 -1.37131646e-01 8.90388727e-01
1.53027952e+00 7.49046132e-02 9.33649361e-01 2.76250929e-01
5.17818451e-01 5.05248010e-01 5.69894493e-01 3.55110288e-01
6.44062400e-01 8.64386439e-01 4.10103500e-01 -3.84836763e-01
-6.01790786e-01 -3.95243317e-01 6.06771648e-01 5.22980511e-01
6.22386597e-02 -4.23529834e-01 -8.21450353e-01 8.25486898e-01
-1.70645547e+00 -1.07699037e+00 4.34422642e-02 1.76140654e+00
7.64982283e-01 -3.21731299e-01 5.64270973e-01 -5.08541107e-01
6.34953380e-01 2.13372216e-01 -5.75281918e-01 -3.30115259e-01
-4.24301893e-01 -8.04441944e-02 4.85316247e-01 3.46073359e-01
-8.85229826e-01 1.18199301e+00 6.55548048e+00 7.75658846e-01
-1.33848619e+00 5.07321954e-01 9.28364456e-01 -1.26931965e-01
-3.65448415e-01 -9.30063725e-02 -6.55370474e-01 5.33836782e-01
7.94892550e-01 -1.60107985e-01 6.09001219e-01 7.69009292e-01
-4.79580741e-03 2.68606573e-01 -9.37060714e-01 1.10925055e+00
2.34545767e-01 -1.42373669e+00 2.39340231e-01 2.06532285e-01
1.04466152e+00 3.29632342e-01 4.49828237e-01 9.96099114e-02
5.18891573e-01 -1.09450495e+00 1.04935563e+00 5.36498763e-02
1.41002357e+00 -5.24680078e-01 4.11534935e-01 -1.12479664e-01
-1.04040265e+00 1.35740057e-01 -4.38147575e-01 2.91806370e-01
-9.02680531e-02 1.47505105e-01 -6.08856916e-01 5.14661252e-01
9.11873758e-01 8.29431355e-01 -9.29282844e-01 7.60998905e-01
-2.76739031e-01 3.73640388e-01 -9.12791193e-02 5.49136817e-01
1.06923908e-01 -1.04719818e-01 2.45326981e-01 1.17460454e+00
6.49395168e-01 -1.45503250e-03 3.58726569e-02 1.06255376e+00
-3.06068659e-01 2.03052275e-02 -8.53732049e-01 -6.74597844e-02
7.10850835e-01 1.18577945e+00 -5.64891994e-01 -3.54008883e-01
-5.73868752e-01 1.16773367e+00 1.65918365e-01 4.34445858e-01
-1.09660947e+00 -5.82845621e-02 1.01341522e+00 3.43534023e-01
4.64721024e-01 -8.38749781e-02 -2.25234464e-01 -1.30236745e+00
-1.90651659e-02 -1.55788016e+00 2.91604131e-01 -8.47598612e-01
-1.23663294e+00 9.07135785e-01 8.44124239e-03 -1.33544028e+00
-1.40816420e-01 -4.72267687e-01 -1.53199553e-01 6.30548000e-01
-1.53128088e+00 -1.74765384e+00 -5.93363881e-01 7.84889936e-01
4.24602568e-01 -4.02386487e-01 7.03155577e-01 3.79179120e-01
-2.22150534e-01 9.70409989e-01 -2.87615675e-02 1.47634521e-01
1.10995257e+00 -7.62839377e-01 8.90709996e-01 8.94220054e-01
5.45048565e-02 1.70500979e-01 5.53471744e-01 -5.64990878e-01
-1.23535204e+00 -1.36428797e+00 8.44764590e-01 -6.47405922e-01
2.16415673e-01 -4.07304764e-01 -7.58462787e-01 7.52571523e-01
7.20274329e-01 1.38636351e-01 6.17098451e-01 -2.81349659e-01
-7.36975431e-01 -2.08637431e-01 -1.31140387e+00 6.56137705e-01
1.26126170e+00 -3.88820708e-01 8.90112966e-02 2.96970874e-01
6.32091582e-01 -3.98691595e-01 -8.36419761e-01 3.19393009e-01
6.64698482e-01 -1.27130115e+00 1.08361411e+00 -3.73137474e-01
7.89701462e-01 -4.09756690e-01 -3.71214718e-01 -1.28073418e+00
-6.25189841e-01 -5.22059917e-01 1.30155787e-01 1.37252772e+00
2.87144303e-01 -5.99873424e-01 5.11268139e-01 2.19232902e-01
3.51483494e-01 -3.92088532e-01 -7.17424154e-01 -1.03843629e+00
1.52988434e-01 3.46743464e-02 1.12802887e+00 1.07244194e+00
-4.75825489e-01 6.51977479e-01 -8.12007725e-01 4.04227488e-02
7.50037074e-01 1.38216704e-01 1.18434215e+00 -8.13219845e-01
-2.20992863e-01 -3.45017523e-01 -4.35946077e-01 -8.01431537e-01
1.29543409e-01 -8.55193198e-01 -1.37042210e-01 -1.40473270e+00
4.72881258e-01 -4.17023569e-01 5.97641468e-02 8.02942395e-01
-3.85860801e-02 8.88086200e-01 5.68721831e-01 4.81151611e-01
-6.58275247e-01 4.42023933e-01 1.42351007e+00 -1.41962513e-01
1.55288279e-01 -5.29306233e-01 -9.62434053e-01 2.23008484e-01
9.49817777e-01 -4.19629157e-01 -5.60549736e-01 -9.89148200e-01
-1.53512415e-03 8.04490745e-02 5.22602081e-01 -1.00432777e+00
-1.30848169e-01 -3.48634869e-01 4.89372432e-01 -1.17223032e-01
3.73770058e-01 -6.08398795e-01 6.57013655e-01 4.64941591e-01
-2.09324583e-01 3.27438295e-01 3.19008708e-01 2.47861117e-01
-2.88498789e-01 3.29397619e-01 1.28853285e+00 -1.36858150e-01
-8.93048644e-01 5.47104061e-01 1.69579297e-01 2.89575547e-01
1.11179745e+00 -1.74708009e-01 -6.65421844e-01 -7.06887841e-01
-2.29226202e-01 1.88291237e-01 1.14343488e+00 7.28863597e-01
6.51435196e-01 -1.73168731e+00 -1.24800348e+00 3.72989595e-01
4.55153316e-01 -4.41630870e-01 1.81108117e-01 6.62977636e-01
-7.61830688e-01 2.54372358e-01 -8.10739994e-01 -6.26717031e-01
-1.56458580e+00 5.81735909e-01 2.82925010e-01 -5.06451577e-02
-2.00639620e-01 8.32076609e-01 4.98610795e-01 -4.95019019e-01
-1.55043796e-01 2.58296728e-01 1.51843712e-01 -1.34629339e-01
4.88433748e-01 1.54880762e-01 -1.40317068e-01 -1.20112062e+00
-5.30489087e-01 5.78429639e-01 7.01359212e-02 -4.67175506e-02
1.24791229e+00 -3.24533045e-01 -6.90130219e-02 -2.25976124e-01
1.27264082e+00 -2.49044523e-01 -1.41709042e+00 -2.90727466e-01
-4.16907430e-01 -9.08417106e-01 -3.32325995e-01 -9.44048703e-01
-1.50434566e+00 7.27793753e-01 9.42025840e-01 -1.34932593e-01
1.38869154e+00 3.85520756e-02 8.04023266e-01 -1.61380291e-01
2.96729267e-01 -6.62251592e-01 3.80275697e-01 3.17236245e-01
1.14564180e+00 -1.51063073e+00 1.41652243e-03 -3.22177202e-01
-1.16758454e+00 6.50564849e-01 8.79184842e-01 7.70768598e-02
3.76418412e-01 2.28478268e-01 4.97066557e-01 -3.61384936e-02
-9.05398369e-01 -1.39105588e-01 1.60509259e-01 8.93758297e-01
5.38225472e-01 1.01027541e-01 -5.89959212e-02 5.16493879e-02
-2.02073663e-01 2.70569533e-01 3.08602244e-01 5.58914661e-01
4.64945659e-03 -1.22837329e+00 -2.62812197e-01 1.58835389e-02
-4.42983061e-01 6.44806481e-04 -5.70795834e-01 7.69033551e-01
2.37582535e-01 1.02517104e+00 6.91262865e-03 -4.45859820e-01
3.25221241e-01 -3.01709384e-01 6.14300549e-01 -3.45827341e-01
-4.41754103e-01 -4.59583849e-03 6.90413862e-02 -7.76368380e-01
-4.57066029e-01 -5.42099774e-01 -7.10195422e-01 -8.23499084e-01
2.78438106e-02 -1.76562816e-01 5.01385570e-01 3.02287370e-01
9.91074979e-01 8.75024647e-02 6.29864752e-01 -7.20148861e-01
-3.94500375e-01 -8.22708070e-01 -4.18430001e-01 7.55363405e-01
3.30503315e-01 -3.76968592e-01 -7.82626793e-02 3.83019596e-01] | [11.68975830078125, -0.3038899600505829] |
8502efce-9415-4e47-b416-437cb8cd825d | optimal-energy-system-scheduling-using-a | 2305.05484 | null | https://arxiv.org/abs/2305.05484v1 | https://arxiv.org/pdf/2305.05484v1.pdf | Optimal Energy System Scheduling Using A Constraint-Aware Reinforcement Learning Algorithm | The massive integration of renewable-based distributed energy resources (DERs) inherently increases the energy system's complexity, especially when it comes to defining its operational schedule. Deep reinforcement learning (DRL) algorithms arise as a promising solution due to their data-driven and model-free features. However, current DRL algorithms fail to enforce rigorous operational constraints (e.g., power balance, ramping up or down constraints) limiting their implementation in real systems. To overcome this, in this paper, a DRL algorithm (namely MIP-DQN) is proposed, capable of \textit{strictly} enforcing all operational constraints in the action space, ensuring the feasibility of the defined schedule in real-time operation. This is done by leveraging recent optimization advances for deep neural networks (DNNs) that allow their representation as a MIP formulation, enabling further consideration of any action space constraints. Comprehensive numerical simulations show that the proposed algorithm outperforms existing state-of-the-art DRL algorithms, obtaining a lower error when compared with the optimal global solution (upper boundary) obtained after solving a mathematical programming formulation with perfect forecast information; while strictly enforcing all operational constraints (even in unseen test days). | ['Peter Palensky', 'Edgar Mauricio Salazar Duque', 'Pedro P. Vergara', 'Hou Shengren'] | 2023-05-09 | null | null | null | null | ['energy-management'] | ['time-series'] | [-2.50367045e-01 -3.27283174e-01 -5.30458212e-01 -3.03789433e-02
-3.41904193e-01 -6.91242874e-01 2.18352437e-01 2.14376986e-01
-1.86385959e-01 1.22366941e+00 -2.97822356e-01 -4.66086805e-01
-8.35813463e-01 -1.02590799e+00 -3.76775712e-01 -1.00404751e+00
-3.88672054e-01 5.20350754e-01 -6.29763842e-01 -2.74651825e-01
1.60466969e-01 6.63023353e-01 -1.20979691e+00 -5.52404284e-01
1.27570546e+00 1.26262128e+00 2.42055386e-01 -2.98914369e-02
2.08519235e-01 5.48573792e-01 -6.24297500e-01 1.50090531e-01
4.17650551e-01 -2.43441433e-01 -3.65697473e-01 -5.55364564e-02
-4.56611633e-01 -7.29422867e-01 -1.79779023e-01 1.23152471e+00
6.73082411e-01 5.89876354e-01 2.80826718e-01 -1.59742129e+00
-7.94616938e-01 4.26270306e-01 -4.38109040e-01 2.26705775e-01
1.73454002e-01 1.35326266e-01 1.10993588e+00 -6.40248120e-01
2.44795635e-01 5.89179754e-01 2.64745355e-01 2.85326540e-01
-1.24449408e+00 -4.11371499e-01 3.62582535e-01 5.90657055e-01
-1.52100205e+00 4.21792157e-02 1.04220486e+00 -2.17332661e-01
1.52017832e+00 3.70490968e-01 1.02650833e+00 8.68187249e-01
3.09456170e-01 7.20278263e-01 1.04919016e+00 -2.49319732e-01
8.44300687e-01 -3.85966033e-01 -1.24239668e-01 5.52817620e-02
5.28083324e-01 4.15343225e-01 -3.58835496e-02 2.57122546e-01
4.30889785e-01 -1.16957612e-01 -3.52306515e-01 -6.15278423e-01
-7.31154799e-01 8.17071199e-01 5.46001792e-01 4.85365957e-01
-6.83635652e-01 -6.25888854e-02 4.15988743e-01 6.92541897e-02
2.77500302e-01 4.03730810e-01 -3.90798390e-01 -4.88175564e-02
-1.00939190e+00 1.23637132e-01 6.28598034e-01 7.88961172e-01
4.46066499e-01 9.49410498e-01 -5.47105968e-02 4.68704104e-01
2.25658119e-01 4.23015386e-01 5.05256653e-01 -6.68043971e-01
4.74281251e-01 7.23041415e-01 4.96161103e-01 -9.40247953e-01
-7.86570728e-01 -7.91616440e-01 -1.23418748e+00 3.27735156e-01
-1.31160483e-01 -5.32866836e-01 -3.89548898e-01 1.63548887e+00
2.11875692e-01 -7.06725493e-02 1.40338421e-01 1.05055201e+00
-2.51747831e-03 9.84365582e-01 -1.33948298e-02 -8.73015642e-01
1.00260377e+00 -5.33448875e-01 -1.07901740e+00 5.89023419e-02
2.17576504e-01 -1.55459777e-01 4.62744772e-01 6.74588263e-01
-8.12195301e-01 -4.13434029e-01 -1.35194957e+00 4.91987437e-01
-6.41810834e-01 3.18325877e-01 6.32788181e-01 6.83997631e-01
-1.15211189e+00 4.26271141e-01 -6.94580019e-01 -3.73794213e-02
1.23660259e-01 5.97651660e-01 1.66495815e-01 2.14966089e-01
-1.42896497e+00 1.22755420e+00 8.96137416e-01 9.00902212e-01
-9.67065394e-01 -7.33583689e-01 -7.88198769e-01 4.11555499e-01
6.60545647e-01 -4.82883126e-01 1.13880217e+00 -8.42545807e-01
-1.78484070e+00 -1.85405985e-02 5.27868032e-01 -7.04043627e-01
5.33279777e-01 -1.06909260e-01 -7.23384559e-01 5.66773824e-02
-4.31249678e-01 8.30520838e-02 6.49310589e-01 -1.20868146e+00
-5.94781399e-01 -6.80992156e-02 4.87584472e-01 2.23311856e-01
-4.55222428e-01 -3.82463336e-01 2.46546030e-01 -5.94221950e-01
-4.16282475e-01 -5.13505340e-01 -4.21127200e-01 -5.74866533e-01
-4.08634305e-01 -4.55741882e-01 8.18209291e-01 -8.32509160e-01
1.59007323e+00 -1.54230881e+00 2.67779648e-01 2.71945328e-01
-2.70231217e-01 4.90463406e-01 1.30553901e-01 8.17780912e-01
-1.95061773e-01 -4.11401168e-02 -2.47058064e-01 -2.04193890e-01
5.94724000e-01 5.44407547e-01 -9.24735591e-02 4.45851147e-01
2.64880598e-01 8.12192321e-01 -9.90264654e-01 5.14832549e-02
7.43594587e-01 3.41126025e-01 -2.52810836e-01 1.40975490e-01
-5.08245528e-01 3.11379373e-01 -5.51261902e-01 5.32985806e-01
7.01633751e-01 1.35462239e-01 6.22291505e-01 -2.91768461e-01
-3.51584107e-01 -2.84950256e-01 -1.49005616e+00 1.34113467e+00
-8.74048114e-01 3.95404041e-01 3.09739977e-01 -1.70834243e+00
9.62967336e-01 3.18104863e-01 1.15002942e+00 -1.22955847e+00
1.05002090e-01 2.23590076e-01 -5.60764931e-02 -3.29196393e-01
2.11091161e-01 -5.48739396e-02 -4.91458699e-02 2.88150012e-01
9.48760379e-03 9.50895399e-02 6.47315025e-01 -3.18915546e-01
5.60726881e-01 1.77045807e-01 4.96385664e-01 -6.43893719e-01
8.12730670e-01 -2.23899201e-01 9.42385614e-01 3.17152202e-01
-1.32339209e-01 -2.09306985e-01 4.66244102e-01 -3.87637496e-01
-9.10660684e-01 -9.47673142e-01 -1.15729295e-01 4.21281457e-01
1.29990071e-01 1.18839480e-01 -4.53391165e-01 -5.80873251e-01
9.34652761e-02 1.26093042e+00 -2.78027177e-01 -1.71982244e-01
-5.91355741e-01 -1.08471930e+00 -2.42711052e-01 4.65945601e-01
6.58431828e-01 -7.71202981e-01 -1.01220441e+00 6.54730260e-01
1.41795129e-01 -9.70624924e-01 -1.38678864e-01 5.41701138e-01
-4.82160896e-01 -1.08688331e+00 -6.71270609e-01 -4.13464934e-01
7.62674451e-01 -2.83532560e-01 1.08549821e+00 -1.21679991e-01
-1.19753286e-01 3.42200190e-01 -1.24782003e-01 1.67430320e-03
-1.29610091e-01 -8.23530275e-03 3.07792515e-01 -4.89750057e-02
3.22126746e-02 -6.08008802e-01 -6.04969919e-01 2.59195358e-01
-9.14073646e-01 -9.33775231e-02 4.98882413e-01 9.77248549e-01
5.97489655e-01 8.94346654e-01 1.27205646e+00 -1.29587755e-01
8.94289434e-01 -5.20218790e-01 -1.36469305e+00 5.67013621e-01
-1.00474727e+00 -4.43572849e-02 1.20764351e+00 7.68538332e-03
-1.01003730e+00 -1.69217736e-01 -8.03851932e-02 -4.86757964e-01
9.30524021e-02 5.90704560e-01 -5.23751438e-01 1.82716530e-02
-1.82542041e-01 4.53215271e-01 -3.03150713e-01 -2.21655548e-01
1.08674936e-01 4.04702783e-01 2.76746243e-01 -6.84702218e-01
9.35813069e-01 -1.16925940e-01 2.92645723e-01 -2.49266371e-01
-7.09772587e-01 -1.61328256e-01 -5.33406138e-01 -5.42467117e-01
6.18742049e-01 -6.13636971e-01 -1.29766595e+00 3.54841292e-01
-7.79489756e-01 -1.26397952e-01 -2.57809550e-01 3.90417635e-01
-6.99528813e-01 2.36944795e-01 -1.64410964e-01 -1.11854661e+00
-4.89951849e-01 -1.18535328e+00 2.89880276e-01 4.40942526e-01
3.03146183e-01 -1.24096239e+00 -1.93200074e-02 5.65879010e-02
7.06357777e-01 8.17609489e-01 9.93581653e-01 -3.66576135e-01
-4.41930890e-01 -3.11804935e-02 9.20959041e-02 7.86645174e-01
3.30774337e-01 2.23722355e-03 -3.84001642e-01 -8.31846654e-01
8.56245160e-02 -5.55704683e-02 5.26893958e-02 4.45738226e-01
1.15202677e+00 -6.32219613e-01 1.56284217e-02 2.52297699e-01
2.25860023e+00 8.06854129e-01 2.30810404e-01 4.91395086e-01
1.79756582e-01 2.06941783e-01 6.48072660e-01 9.25658941e-01
3.95247310e-01 6.78990304e-01 1.11543357e+00 -9.39300358e-02
4.49341089e-01 4.54151332e-02 2.99365461e-01 6.90643847e-01
4.02690955e-02 -6.26688838e-01 -7.45819807e-01 4.77619022e-01
-2.00274110e+00 -8.21687162e-01 1.97431862e-01 2.24456716e+00
5.05266249e-01 2.95455605e-01 -7.73442611e-02 3.21794212e-01
6.83254540e-01 3.01250786e-01 -1.00145733e+00 -9.44371343e-01
-2.12545842e-01 1.83363512e-01 4.82533038e-01 2.89867371e-01
-9.20071363e-01 1.77993402e-01 5.61846924e+00 8.54693294e-01
-1.22890496e+00 -1.24786578e-01 6.37215078e-01 -9.75276753e-02
-2.92049617e-01 -4.49653834e-01 -4.85771984e-01 6.92184925e-01
1.05238366e+00 -6.84070170e-01 9.29613173e-01 7.73891747e-01
9.77112293e-01 -3.35128486e-01 -1.06715643e+00 8.06893468e-01
-2.30647147e-01 -1.27850115e+00 -1.20694108e-01 1.14335313e-01
1.10913205e+00 -1.83784902e-01 -1.53125912e-01 5.92301846e-01
2.14011759e-01 -9.39446986e-01 7.68086255e-01 4.86041814e-01
4.12177533e-01 -1.39299810e+00 9.47468877e-01 3.94287407e-01
-1.22092557e+00 -6.19820476e-01 -2.87333932e-02 -1.94773912e-01
3.31956178e-01 8.51038992e-01 -4.20113683e-01 1.42924607e+00
4.55768287e-01 6.36406779e-01 1.07050966e-02 7.93513834e-01
-3.69854242e-01 3.23343992e-01 -4.76911485e-01 -2.50462890e-01
4.65744495e-01 -5.55667758e-01 3.26947898e-01 6.89933836e-01
3.05755049e-01 -9.51150134e-02 5.05984426e-01 9.93430257e-01
7.00309798e-02 -2.02473059e-01 -2.14729056e-01 -5.90061694e-02
5.82079887e-01 1.26372004e+00 -4.75208938e-01 3.38807516e-02
-3.47133905e-01 4.82428133e-01 7.28628505e-03 6.33910179e-01
-1.08593357e+00 -2.50760108e-01 7.04350531e-01 -5.12418747e-01
2.69973248e-01 -5.86321831e-01 -2.88446039e-01 -8.41252089e-01
2.52116621e-01 -5.69220841e-01 3.91791463e-01 -5.76483130e-01
-1.38855052e+00 1.97067410e-01 7.44415307e-03 -1.41877699e+00
-5.59878469e-01 -6.87703788e-01 -6.95855916e-01 7.78221786e-01
-2.05492806e+00 -8.52620959e-01 3.00142430e-02 5.19106328e-01
5.19891560e-01 -1.06725626e-01 6.78538382e-01 6.18389606e-01
-1.28903866e+00 2.49563649e-01 6.77458465e-01 -2.26755902e-01
-2.37511396e-01 -1.42802191e+00 -3.97369355e-01 1.18537581e+00
-2.47566223e-01 1.12077624e-01 7.53937900e-01 -2.01636836e-01
-1.97924006e+00 -8.11467111e-01 2.90749639e-01 5.19464076e-01
9.78036106e-01 -1.75003663e-01 -4.60382044e-01 3.49135607e-01
6.48755193e-01 -4.28882912e-02 3.81597966e-01 -4.03893560e-01
4.26268965e-01 -5.43559313e-01 -1.28168464e+00 3.09708148e-01
4.72978175e-01 -1.42903700e-01 -3.65891933e-01 2.23539874e-01
2.66724885e-01 -2.76844800e-01 -1.15185750e+00 6.63031280e-01
6.28239065e-02 -6.87287629e-01 8.67991745e-01 -3.92440528e-01
3.48671228e-02 -5.29593587e-01 -1.53966114e-01 -1.73955905e+00
-9.42882672e-02 -6.66840196e-01 -7.66478479e-01 1.33456397e+00
2.80104894e-02 -7.21585631e-01 3.97217661e-01 4.77400005e-01
-3.91044766e-01 -1.11619210e+00 -1.63255870e+00 -1.17142618e+00
3.02748289e-02 -2.74799228e-01 8.42095912e-01 1.05458081e+00
3.10463030e-02 -2.63101757e-01 -2.95012325e-01 5.97340524e-01
6.16145194e-01 4.26571637e-01 -4.96468432e-02 -7.11612165e-01
-4.65575717e-02 -6.93382800e-01 -1.03944317e-01 -5.57433367e-01
3.37294370e-01 -5.60421824e-01 -1.41537711e-01 -1.91335857e+00
-4.43837851e-01 -4.69018131e-01 -8.06103706e-01 5.47018349e-01
3.21090013e-01 -2.77460188e-01 3.41315180e-01 -3.04496944e-01
-4.09772933e-01 1.12674534e+00 1.02638435e+00 -4.55499172e-01
1.40213016e-02 1.02275070e-02 -1.82862043e-01 3.24257702e-01
1.28726125e+00 1.44115701e-01 -6.42512500e-01 -4.64952052e-01
3.36881846e-01 3.59600216e-01 1.40084550e-01 -1.16937232e+00
1.20848864e-01 -5.16559005e-01 2.00688705e-01 -7.66287327e-01
-6.97949855e-03 -1.35791385e+00 4.83861536e-01 7.91048288e-01
1.25265136e-01 4.05480653e-01 2.32585534e-01 3.38142097e-01
-3.09047610e-01 -3.94878268e-01 5.41444004e-01 2.73617953e-01
-1.14362264e+00 2.45140076e-01 -4.44319904e-01 -3.10977459e-01
1.53483963e+00 -5.30374721e-02 -3.12914908e-01 -6.06662780e-02
-7.24396348e-01 9.47998762e-01 -2.08317302e-02 5.59457779e-01
2.51831472e-01 -1.34978902e+00 -3.76622140e-01 -5.51029295e-02
-4.36526150e-01 -1.17090724e-01 4.95807618e-01 7.04729378e-01
-3.52353483e-01 8.37370992e-01 -3.36926788e-01 -3.71040255e-01
-2.18657449e-01 7.71890879e-01 8.06121588e-01 -7.36674726e-01
-3.63666177e-01 -5.88793568e-02 -5.70304811e-01 -2.70562470e-01
1.77063242e-01 -4.64399874e-01 -8.99713188e-02 4.96522248e-01
5.08177765e-02 6.57494128e-01 5.35212576e-01 -1.02415055e-01
-4.29141581e-01 2.73388475e-01 5.37372053e-01 4.12119657e-01
1.48685586e+00 -2.32566714e-01 4.99207601e-02 7.44313300e-02
9.10058856e-01 -4.07739192e-01 -1.30105257e+00 2.75609285e-01
1.62599176e-01 -3.26093026e-02 4.89058226e-01 -1.24701297e+00
-1.47796941e+00 4.95103627e-01 6.59209132e-01 6.53724670e-01
1.56946969e+00 -9.54487860e-01 6.92637026e-01 3.34570795e-01
8.32170427e-01 -1.61191237e+00 -3.18096489e-01 3.56764257e-01
9.25569355e-01 -9.12867844e-01 1.10342331e-01 2.37661704e-01
-8.73057023e-02 1.36929071e+00 6.17850184e-01 -1.65759027e-01
5.96752882e-01 3.25669676e-01 -5.20503163e-01 3.35501730e-01
-6.39587402e-01 -5.10197075e-04 -4.68915924e-02 4.91273642e-01
-1.86743587e-01 3.13198924e-01 -7.99976945e-01 5.69711924e-01
2.12111875e-01 1.01209350e-01 5.12412131e-01 9.26451921e-01
-1.46238089e-01 -8.45490098e-01 -2.39291176e-01 2.39913613e-01
-1.11785457e-01 2.47017711e-01 4.83573109e-01 1.03952098e+00
2.70114750e-01 9.57888901e-01 4.32737619e-02 2.22036064e-01
4.45206791e-01 -2.88708150e-01 2.05538258e-01 -7.93284774e-02
-5.31237245e-01 -1.55444324e-01 3.93884405e-02 -4.57165390e-01
-5.07463634e-01 -5.62690139e-01 -1.46944773e+00 -1.77825436e-01
-2.95781791e-01 3.12909931e-01 7.83758461e-01 1.16550946e+00
1.19720384e-01 9.69899952e-01 1.16780579e+00 -8.98402214e-01
-1.07309973e+00 -5.53431332e-01 -7.00881720e-01 -2.36259505e-01
2.53568381e-01 -7.73292959e-01 -3.63336742e-01 -4.24527824e-01] | [5.636808395385742, 2.5613152980804443] |
fec37987-8382-4e06-a4c7-109d5e9567b5 | learning-to-estimate-6dof-pose-from-limited | 2306.07598 | null | https://arxiv.org/abs/2306.07598v1 | https://arxiv.org/pdf/2306.07598v1.pdf | Learning to Estimate 6DoF Pose from Limited Data: A Few-Shot, Generalizable Approach using RGB Images | The accurate estimation of six degrees-of-freedom (6DoF) object poses is essential for many applications in robotics and augmented reality. However, existing methods for 6DoF pose estimation often depend on CAD templates or dense support views, restricting their usefulness in realworld situations. In this study, we present a new cascade framework named Cas6D for few-shot 6DoF pose estimation that is generalizable and uses only RGB images. To address the false positives of target object detection in the extreme few-shot setting, our framework utilizes a selfsupervised pre-trained ViT to learn robust feature representations. Then, we initialize the nearest top-K pose candidates based on similarity score and refine the initial poses using feature pyramids to formulate and update the cascade warped feature volume, which encodes context at increasingly finer scales. By discretizing the pose search range using multiple pose bins and progressively narrowing the pose search range in each stage using predictions from the previous stage, Cas6D can overcome the large gap between pose candidates and ground truth poses, which is a common failure mode in sparse-view scenarios. Experimental results on the LINEMOD and GenMOP datasets demonstrate that Cas6D outperforms state-of-the-art methods by 9.2% and 3.8% accuracy (Proj-5) under the 32-shot setting compared to OnePose++ and Gen6D. | ['Zhangyang Wang', 'Chenxin Li', 'Peihao Wang', 'Brandon Y. Feng', 'Zhiwen Fan', 'Panwang Pan'] | 2023-06-13 | null | null | null | null | ['pose-estimation'] | ['computer-vision'] | [-1.24257617e-01 -7.70023763e-02 -2.16581702e-01 -2.14608237e-01
-8.76739323e-01 -3.83759290e-01 3.28114122e-01 -3.07610005e-01
-1.87151089e-01 3.97042155e-01 1.50813699e-01 4.49273765e-01
-2.22129092e-01 -5.35728395e-01 -7.81825185e-01 -5.35341740e-01
1.53685361e-01 8.57929528e-01 5.85799038e-01 -3.05596411e-01
1.85097367e-01 7.30923533e-01 -1.58533609e+00 -1.16035461e-01
5.73611259e-01 1.13936043e+00 3.72266501e-01 3.64201725e-01
4.01431829e-01 4.20480967e-01 -4.05665100e-01 -1.48533344e-01
6.56491220e-01 -9.04428884e-02 -1.11991100e-01 4.32468176e-01
7.77269006e-01 -7.39859462e-01 -5.43716908e-01 9.64785516e-01
5.43599427e-01 2.63450861e-01 5.11213064e-01 -1.12902546e+00
-1.71159878e-01 -3.88485454e-02 -8.14430773e-01 4.47267592e-02
7.24638104e-01 1.35968655e-01 7.60049522e-01 -1.53566015e+00
8.06309164e-01 1.23375010e+00 8.09731483e-01 4.24732238e-01
-1.00144815e+00 -6.02360308e-01 2.75426179e-01 4.38503250e-02
-1.59810698e+00 -3.18989336e-01 7.63768911e-01 -3.54129732e-01
1.04150867e+00 -6.56563696e-03 9.22202229e-01 9.07299161e-01
3.22217643e-01 6.71441257e-01 7.25134611e-01 -8.12194869e-02
2.71209002e-01 -3.53791744e-01 -2.15020180e-01 8.17442536e-01
3.79410982e-01 1.56584397e-01 -9.95886087e-01 -9.85754803e-02
1.21497643e+00 4.25975829e-01 -2.04451457e-01 -1.14686179e+00
-1.54479635e+00 7.81879783e-01 7.61043131e-01 -3.14061463e-01
-4.39665109e-01 2.38893509e-01 -7.62349274e-03 -7.63281658e-02
3.25191677e-01 4.53091234e-01 -5.14319777e-01 4.11297269e-02
-6.33741319e-01 3.74790400e-01 4.90956038e-01 1.28679836e+00
8.61353219e-01 -1.33781001e-01 2.98429430e-02 5.65008640e-01
5.29729962e-01 6.82309687e-01 2.09850088e-01 -1.05151677e+00
4.42212641e-01 6.82498336e-01 5.59227824e-01 -9.28639054e-01
-4.29616392e-01 -6.14775956e-01 -3.65426481e-01 1.06009841e-01
3.02038759e-01 1.57605380e-01 -1.29984033e+00 1.16044652e+00
9.43151951e-01 2.36131012e-01 -1.49257272e-01 1.28213096e+00
7.87596762e-01 4.37734008e-01 -4.69748944e-01 -2.05851555e-01
1.10170507e+00 -9.57673073e-01 -4.54681933e-01 -6.45983517e-01
3.66308361e-01 -8.09559524e-01 9.29775357e-01 3.64427954e-01
-6.49804831e-01 -5.31601250e-01 -1.16799128e+00 -9.41621810e-02
-5.63477315e-02 2.70559639e-01 7.41216600e-01 2.91631013e-01
-6.47111058e-01 2.45804891e-01 -9.98509765e-01 -3.32957983e-01
3.20212364e-01 4.27247763e-01 -5.70124805e-01 -4.57815975e-01
-7.11945534e-01 1.03366566e+00 4.12351996e-01 2.08306849e-01
-1.04649913e+00 -7.38249958e-01 -1.10709679e+00 -2.67174572e-01
8.41806948e-01 -7.33821571e-01 1.20899808e+00 -2.94057429e-02
-1.53473401e+00 7.20791399e-01 -6.59061084e-03 -1.88531503e-01
6.34484529e-01 -7.13545382e-01 1.61786169e-01 1.85373113e-01
2.76717246e-01 6.76109016e-01 9.52014625e-01 -1.24633622e+00
-6.46844804e-01 -8.16091299e-01 1.81559727e-01 7.69818842e-01
1.84439331e-01 -4.35517699e-01 -8.71216774e-01 -4.36443657e-01
1.11860311e+00 -1.15620673e+00 -5.07113636e-01 4.19540286e-01
-6.98795170e-02 -7.72936121e-02 8.41326535e-01 -3.38985384e-01
5.52456498e-01 -2.13659048e+00 4.38917637e-01 6.34689480e-02
2.50677347e-01 -1.65229663e-01 4.65538651e-02 2.79057741e-01
3.28689694e-01 -6.46088004e-01 2.05018446e-01 -4.93688941e-01
-1.97276458e-01 2.51859248e-01 -1.77750647e-01 8.76005888e-01
1.22612379e-01 8.65728080e-01 -1.04466498e+00 -3.65852654e-01
6.81747317e-01 5.45257330e-01 -7.09493399e-01 1.94842160e-01
-1.69213042e-01 4.46986854e-01 -6.12345099e-01 9.99809086e-01
6.69036508e-01 -2.42228225e-01 -2.53661007e-01 -5.69433272e-01
5.26256301e-02 -1.24898605e-01 -1.47445667e+00 2.36647558e+00
-2.70540088e-01 9.01549757e-02 -2.84170419e-01 -3.91612500e-01
1.03770077e+00 -1.00411348e-01 7.19057560e-01 -2.13484094e-01
2.28705898e-01 2.35181332e-01 -2.96918184e-01 -2.01902449e-01
4.66741771e-01 -1.14879645e-01 -3.83366913e-01 -1.01360388e-01
4.09358561e-01 -7.60876834e-01 -1.39838949e-01 1.05296344e-01
9.86740828e-01 5.40761352e-01 5.59792876e-01 2.49753192e-01
1.82264030e-01 7.07723945e-02 6.43982708e-01 5.07494748e-01
-3.84849787e-01 1.02374351e+00 -8.88526887e-02 -5.00774622e-01
-8.59583855e-01 -1.33459377e+00 -5.47478087e-02 6.70077384e-01
6.56460702e-01 -4.21859562e-01 -2.82133102e-01 -6.97539032e-01
2.93619722e-01 3.85917962e-01 -5.29408395e-01 -3.25193256e-01
-4.78202403e-01 -4.67630744e-01 -2.74591267e-01 6.50277078e-01
4.51711506e-01 -4.66680557e-01 -9.67661142e-01 1.92870185e-01
-1.54706091e-01 -1.21681964e+00 -3.71208370e-01 2.66717315e-01
-8.52867484e-01 -1.05754697e+00 -8.19623351e-01 -5.47909975e-01
7.17115641e-01 6.94113851e-01 7.08293200e-01 -3.71373445e-01
-2.95524329e-01 5.81112027e-01 -5.80705404e-01 -3.28919858e-01
1.98639557e-01 -1.56972542e-01 3.60585392e-01 -9.46860090e-02
2.56982058e-01 -3.78704578e-01 -8.14237595e-01 5.98960996e-01
-3.17935139e-01 -6.09058626e-02 6.90797329e-01 7.81803131e-01
1.00532877e+00 -4.90295410e-01 2.32595727e-01 -3.43311578e-01
-2.56047398e-01 -2.48067379e-01 -7.33860731e-01 -5.75251579e-02
-3.05802464e-01 -1.51745617e-01 3.92555036e-02 -5.04728258e-01
-9.28377330e-01 7.57226706e-01 1.85818955e-01 -1.15611184e+00
4.27187979e-02 4.05558914e-01 -1.72011927e-01 -4.34818268e-01
7.67051816e-01 4.98859026e-02 -2.87167318e-02 -3.81211400e-01
3.27655435e-01 2.89450914e-01 6.18677735e-01 -1.84670627e-01
1.11767161e+00 6.05119586e-01 1.70868002e-02 -5.55837929e-01
-1.23281097e+00 -8.33239019e-01 -8.26448381e-01 -5.46912253e-01
7.00396240e-01 -1.39082634e+00 -3.89930576e-01 2.87665159e-01
-9.89496768e-01 9.34997797e-02 -3.74367028e-01 8.14995766e-01
-8.15684199e-01 1.74143746e-01 -3.03598523e-01 -7.22777247e-01
-7.72599354e-02 -1.06446636e+00 1.74044633e+00 2.38372579e-01
-7.67427981e-02 -8.41737464e-02 -1.09033898e-01 5.12789786e-01
-1.93209961e-01 4.60704535e-01 8.70618150e-02 -2.39319667e-01
-8.93985868e-01 -5.12988329e-01 7.84154907e-02 -1.43788382e-01
-1.37644717e-02 -3.90670776e-01 -8.13490868e-01 -4.96920377e-01
1.22183301e-01 -4.73927230e-01 6.10204697e-01 3.40698510e-01
6.69876337e-01 3.00375611e-01 -4.14606333e-01 7.80648828e-01
1.42075384e+00 1.27542736e-02 4.86997932e-01 2.48914376e-01
8.45649481e-01 2.25102097e-01 1.41632116e+00 7.42185235e-01
3.76857996e-01 8.73307586e-01 9.60826516e-01 3.32652837e-01
-8.78576115e-02 -5.57463288e-01 2.72789925e-01 6.41753316e-01
-1.12937197e-01 1.46634936e-01 -9.14535046e-01 3.48647803e-01
-2.00230455e+00 -5.42207003e-01 1.75492793e-01 2.26958656e+00
4.27077591e-01 2.58738935e-01 -6.69275224e-02 -8.30402784e-03
5.75153291e-01 1.73413098e-01 -8.71017337e-01 6.06651843e-01
2.40161449e-01 -2.22282812e-01 5.22053659e-01 3.66603583e-01
-9.88818645e-01 1.19178247e+00 5.46031284e+00 5.76738954e-01
-9.51642454e-01 4.63339686e-02 4.09403853e-02 -3.43233377e-01
2.45169640e-01 -3.00066061e-02 -1.17415333e+00 1.57753490e-02
2.82380819e-01 2.09353223e-01 2.15159297e-01 1.11609435e+00
-1.29579574e-01 -4.38160866e-01 -1.18545842e+00 1.44733083e+00
3.86779308e-01 -1.14514768e+00 -2.26924598e-01 2.81299260e-02
8.39542568e-01 2.69830614e-01 -1.55811206e-01 3.42273057e-01
4.59152907e-02 -5.72892606e-01 9.89203095e-01 4.29705143e-01
8.01836252e-01 -6.88215137e-01 6.52632296e-01 5.56002438e-01
-1.25931752e+00 -2.22760737e-01 -5.79633534e-01 -8.70252177e-02
2.17345923e-01 5.36920965e-01 -1.09697211e+00 6.01805210e-01
9.32094574e-01 8.63691688e-01 -5.30322552e-01 1.25558221e+00
-2.76128739e-01 -8.26628059e-02 -6.81971848e-01 6.24367334e-02
4.02535535e-02 1.36165485e-01 7.22665846e-01 3.23819637e-01
5.79959393e-01 3.35942626e-01 4.57847953e-01 3.80107731e-01
2.68924117e-01 -3.22178423e-01 -5.40312231e-01 3.40353549e-01
6.05083525e-01 1.13692188e+00 -9.84843493e-01 -1.20679915e-01
-3.30440223e-01 1.09810138e+00 2.76147276e-01 1.44568413e-01
-8.81245434e-01 -6.66809008e-02 5.36858797e-01 2.35034868e-01
6.12553716e-01 -5.84229648e-01 -4.49743420e-02 -1.48444617e+00
2.58220404e-01 -6.41324103e-01 2.82427788e-01 -9.49351668e-01
-1.09342396e+00 3.04472953e-01 2.36687601e-01 -1.81915832e+00
-3.40712935e-01 -5.68656862e-01 -6.23749010e-03 4.18168753e-01
-1.14892876e+00 -1.40070009e+00 -7.00838566e-01 5.72121859e-01
9.60778713e-01 1.34854808e-01 7.65778899e-01 -3.28840613e-02
-1.26535267e-01 3.02283823e-01 -2.37874508e-01 -1.19231537e-01
7.10566282e-01 -8.86126399e-01 2.39219666e-01 6.71380401e-01
1.80541754e-01 3.46909553e-01 7.52240479e-01 -8.61674547e-01
-1.98316622e+00 -1.01803875e+00 4.12872970e-01 -6.90863431e-01
3.72968912e-01 -5.67403257e-01 -4.62275386e-01 7.37485766e-01
-6.34622931e-01 6.96654975e-01 1.94528341e-01 -1.35195464e-01
-2.98703462e-01 -1.19685039e-01 -1.25599480e+00 4.63140637e-01
1.58694232e+00 -3.64734024e-01 -7.20687270e-01 3.61416698e-01
7.73963511e-01 -1.17182803e+00 -9.34161305e-01 7.27119863e-01
6.57001555e-01 -8.62402976e-01 1.16501927e+00 -9.55191925e-02
3.95015441e-02 -8.01746070e-01 -5.16496658e-01 -1.15267026e+00
-3.71194631e-01 -5.05170107e-01 -6.05538189e-01 7.13152289e-01
-1.43278971e-01 -3.69902253e-01 1.06551719e+00 2.60204047e-01
-3.40849459e-01 -8.23459685e-01 -1.19887638e+00 -6.94424748e-01
-7.01772153e-01 -3.28546911e-01 3.99996877e-01 5.04468083e-01
-3.20229173e-01 4.11117196e-01 -2.86207050e-01 6.10849738e-01
8.28260958e-01 3.56202155e-01 1.20678449e+00 -1.14511871e+00
-2.55923331e-01 1.82231024e-01 -9.87024605e-01 -1.42682254e+00
-2.43734121e-01 -4.14474100e-01 4.25003588e-01 -1.52321959e+00
8.21665227e-02 -1.83559075e-01 -5.05756438e-02 3.24800432e-01
-1.98638290e-01 5.40008426e-01 2.87958980e-01 2.83812135e-01
-7.24824786e-01 8.36099148e-01 1.33351064e+00 -7.23093841e-03
-1.99810341e-01 1.30176274e-02 -2.38042697e-01 8.59782696e-01
2.17118114e-01 -3.72268975e-01 -4.23188329e-01 -2.84574121e-01
2.19626352e-01 1.62663043e-01 5.81690013e-01 -1.25741959e+00
1.96130812e-01 -2.80751288e-01 8.26211214e-01 -1.27709401e+00
9.47328925e-01 -1.03723502e+00 1.91564083e-01 4.63852972e-01
2.83936799e-01 -1.06237851e-01 -5.69869466e-02 9.69720483e-01
-7.88026974e-02 -4.19789329e-02 6.29683554e-01 -2.51870096e-01
-1.17914617e+00 6.33159697e-01 1.32885560e-01 -2.32843608e-01
1.38430357e+00 -6.24163985e-01 1.04377598e-01 -3.16868573e-01
-7.68518984e-01 3.74896497e-01 6.95299506e-01 5.46421707e-01
1.05627263e+00 -1.41631413e+00 -3.90917480e-01 2.57482350e-01
5.47020316e-01 7.83310115e-01 2.78663248e-01 8.39668334e-01
-5.28374851e-01 5.96298799e-02 -1.15237147e-01 -1.19891250e+00
-9.31877792e-01 4.84142482e-01 3.03797662e-01 3.62382457e-02
-7.08666027e-01 1.09172940e+00 3.18632543e-01 -5.03035486e-01
3.93655807e-01 -3.51525337e-01 -4.51187231e-03 4.82196882e-02
3.54053438e-01 2.89907247e-01 1.04705669e-01 -8.28713298e-01
-5.22556603e-01 9.67965245e-01 -9.67003778e-02 1.99559350e-02
1.53994572e+00 -2.47574076e-01 4.54349816e-01 4.68785852e-01
9.46286440e-01 -1.71338037e-01 -1.91413069e+00 -2.22857356e-01
-2.98251331e-01 -9.21003997e-01 8.62728953e-02 -4.98686016e-01
-8.43590438e-01 5.53934455e-01 7.33350635e-01 -5.71335733e-01
8.63387764e-01 3.71609539e-01 5.66739023e-01 5.46285510e-01
1.04900813e+00 -8.73953581e-01 5.83331466e-01 6.48675263e-01
1.26447403e+00 -1.31034601e+00 3.20061177e-01 -8.33688498e-01
-4.64598805e-01 1.06480432e+00 1.09239185e+00 -4.90953922e-01
5.01855254e-01 1.17744759e-01 -6.79425523e-02 -4.05323625e-01
-2.99712598e-01 -2.28210539e-02 4.24926549e-01 6.19961441e-01
-2.17313722e-01 -1.84291288e-01 8.61885920e-02 4.25576121e-01
-2.15935081e-01 -6.37705624e-02 2.24071115e-01 1.31589568e+00
-5.04421711e-01 -5.03421962e-01 -6.19647443e-01 4.02579159e-01
8.04465860e-02 3.55645835e-01 -2.87802547e-01 8.07286143e-01
2.01011851e-01 5.40273666e-01 1.84537936e-02 -6.35558665e-01
6.18617296e-01 -2.74195343e-01 9.75063324e-01 -9.68654633e-01
-2.04418153e-01 4.73468125e-01 -9.16303843e-02 -9.52391028e-01
-4.77418989e-01 -8.61695468e-01 -1.31397033e+00 3.14373255e-01
-8.54695082e-01 -4.37003881e-01 5.22846103e-01 9.26428437e-01
3.48378778e-01 2.58722067e-01 4.29484576e-01 -1.72292733e+00
-8.95228624e-01 -8.50346029e-01 -4.12715346e-01 7.48618320e-02
4.20057416e-01 -1.28587174e+00 -2.74527788e-01 -1.93154931e-01] | [7.407684326171875, -2.551175117492676] |
3dfb0940-e8d8-4c32-99eb-ae03c040f8b8 | geometric-structure-preserving-warp-for | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Du_Geometric_Structure_Preserving_Warp_for_Natural_Image_Stitching_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Du_Geometric_Structure_Preserving_Warp_for_Natural_Image_Stitching_CVPR_2022_paper.pdf | Geometric Structure Preserving Warp for Natural Image Stitching | Preserving geometric structures in the scene plays a vital role in image stitching. However, most of the existing methods ignore the large-scale layouts reflected by straight lines or curves, decreasing overall stitching quality. To address this issue, this work presents a structure-preserving stitching approach that produces images with natural visual effects and less distortion. Our method first employs deep learning-based edge detection to extract various types of large-scale edges. Then, the extracted edges are sampled to construct multiple groups of triangles to represent geometric structures. Meanwhile, a GEometric Structure preserving (GES) energy term is introduced to make these triangles undergo similarity transformation. Further, an optimized GES energy term is presented to reasonably determine the weights of the sampling points on the geometric structure, and the term is added into the Global Similarity Prior (GSP) stitching model called GES-GSP to achieve a smooth transition between local alignment and geometric structure preservation. The effectiveness of GES-GSP is validated through comprehensive experiments on a stitching dataset. The experimental results show that the proposed method outperforms several state-of-the-art methods in geometric structure preservation and obtains more natural stitching results. The code and dataset are available at https://github.com/flowerDuo/GES-GSP-Stitching. | ['Jiaxin Wang', 'Xinchao Wang', 'Shaoli Huang', 'Jiguang Cui', 'Jifeng Ning', 'Peng Du'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['image-stitching', 'edge-detection'] | ['computer-vision', 'computer-vision'] | [ 3.04269165e-01 -5.07407844e-01 -2.16931343e-01 -2.81048939e-02
-2.04360083e-01 -5.09621322e-01 3.81434888e-01 -6.64171129e-02
7.23932162e-02 2.72957832e-01 2.62937814e-01 7.55111203e-02
-6.83828490e-04 -7.40175605e-01 -6.19523585e-01 -9.87905264e-01
2.25881875e-01 -1.16692930e-01 4.21193480e-01 -3.08574259e-01
6.76361322e-01 3.51757348e-01 -1.19626451e+00 -1.22519225e-01
1.37519276e+00 9.28497195e-01 5.63903749e-02 2.56411612e-01
7.26843998e-02 1.54679924e-01 -4.51470226e-01 -2.93447167e-01
4.07424033e-01 -3.81600946e-01 -1.35337919e-01 3.69519919e-01
6.79150462e-01 -5.25055528e-01 -4.77750421e-01 1.33103716e+00
4.01792467e-01 1.34029537e-01 3.99628699e-01 -1.24768782e+00
-8.51011693e-01 8.92474949e-02 -1.34831738e+00 -1.41186416e-01
9.43000540e-02 2.12780043e-01 7.51510978e-01 -6.39660776e-01
4.48346853e-01 1.31384456e+00 6.69988215e-01 1.08707352e-02
-1.12302148e+00 -9.85992432e-01 -2.25010693e-01 1.26177566e-02
-1.36879969e+00 -2.42249608e-01 1.44930065e+00 -9.47797969e-02
3.11125606e-01 2.38781184e-01 5.99007249e-01 6.63594067e-01
6.03281498e-01 7.61294663e-01 8.00544143e-01 -3.58493805e-01
-2.00649664e-01 -4.03275609e-01 -4.34855511e-03 9.27490234e-01
6.09591126e-01 5.13849318e-01 -3.61354917e-01 -5.38950413e-02
1.27990615e+00 2.25284904e-01 -4.84974593e-01 -4.89909261e-01
-1.18546283e+00 5.69323063e-01 8.15971971e-01 4.60092247e-01
-8.23275466e-03 3.23139429e-01 5.94480395e-01 1.31585856e-03
4.42152470e-01 1.55112565e-01 1.46656722e-01 2.41696209e-01
-8.95793378e-01 6.21129572e-02 1.77222818e-01 9.55545008e-01
8.37129474e-01 1.94297373e-01 -1.19905464e-01 8.92914295e-01
1.87584251e-01 4.58339185e-01 2.39294022e-01 -9.57238436e-01
3.81231427e-01 6.92304313e-01 -5.49103245e-02 -1.81441939e+00
1.28902465e-01 -5.11678636e-01 -1.11091232e+00 3.09163481e-02
5.80523815e-03 2.02155765e-02 -7.67493188e-01 1.15255177e+00
4.41300780e-01 5.64220667e-01 -3.11040461e-01 9.27394092e-01
6.86466515e-01 7.67108262e-01 -3.80552560e-01 -6.11035191e-02
1.34012079e+00 -1.17304456e+00 -9.32838023e-01 -2.20372155e-01
4.03065950e-01 -1.12211645e+00 9.03216064e-01 2.28852689e-01
-8.64937544e-01 -7.44527757e-01 -1.28071046e+00 -2.90464371e-01
4.51065414e-02 2.43431062e-01 4.91230249e-01 4.69522387e-01
-8.34544122e-01 4.24175680e-01 -6.62880301e-01 7.32439533e-02
4.68121409e-01 1.89021483e-01 -7.39703923e-02 -2.72043198e-01
-8.70643735e-01 4.65927213e-01 5.88696957e-01 2.33184278e-01
-3.92243207e-01 -6.49659634e-01 -1.14046323e+00 1.75693154e-01
4.36841875e-01 -5.76988816e-01 8.31765175e-01 -9.62257862e-01
-1.31406999e+00 7.07051098e-01 -1.05311461e-01 3.96546572e-02
3.62915516e-01 -1.13384172e-01 -4.04428631e-01 2.64883667e-01
-2.15857681e-02 5.73233366e-01 1.23695755e+00 -1.63071477e+00
-5.28225601e-01 -1.94600731e-01 -4.24248219e-01 2.39517152e-01
-4.24229473e-01 -5.17031029e-02 -6.52896583e-01 -1.15928853e+00
1.84548214e-01 -7.34508097e-01 -3.12012006e-02 5.92069440e-02
-5.00623882e-01 1.17518902e-01 1.49307668e+00 -8.76753569e-01
1.59726644e+00 -2.32291031e+00 6.49272278e-02 2.27702364e-01
2.58531302e-01 4.27657664e-01 -2.89949149e-01 5.37349284e-01
-3.83511223e-02 -8.86013433e-02 -5.06076694e-01 -4.13019359e-01
-3.54911774e-01 -6.57772273e-02 -2.65636593e-01 4.71048445e-01
-9.20874067e-03 1.06757987e+00 -1.00534320e+00 -5.59338868e-01
5.79125345e-01 5.86360812e-01 -3.97461176e-01 -5.55779226e-02
1.72646828e-02 3.62674296e-01 -5.40921926e-01 7.77532458e-01
1.32073116e+00 8.14738777e-03 -1.22410186e-01 -5.58310390e-01
-3.92490089e-01 -3.22432667e-01 -8.62055302e-01 1.64739954e+00
-1.53431267e-01 5.66921473e-01 7.23243207e-02 -6.49239779e-01
1.26843989e+00 -2.79940367e-02 5.33727467e-01 -5.35873115e-01
4.88130420e-01 4.73581910e-01 -2.39939049e-01 -3.61842632e-01
6.62790358e-01 1.39204443e-01 3.34767997e-02 1.47762075e-01
-4.75399464e-01 -4.87519830e-01 -3.37699689e-02 1.00483261e-01
5.03564239e-01 -7.01630069e-03 2.26344075e-02 -4.07361358e-01
6.55490398e-01 -8.84770453e-02 8.00440729e-01 -4.67022173e-02
-1.41957536e-01 8.17937434e-01 2.22141638e-01 -4.92294699e-01
-1.20952809e+00 -7.42873609e-01 4.55654860e-02 3.87747943e-01
1.00900340e+00 -3.56698006e-01 -9.19986904e-01 -4.01672602e-01
-1.58556238e-01 6.37929738e-01 -5.20964444e-01 -5.17285585e-01
-7.72886395e-01 -3.18602324e-01 3.41682583e-01 3.22671026e-01
1.23861217e+00 -9.04690564e-01 -4.91435766e-01 -2.46121157e-02
-3.07092190e-01 -8.59800696e-01 -1.48783934e+00 -6.52159393e-01
-1.00753939e+00 -9.22343194e-01 -7.90726781e-01 -1.25861418e+00
9.12095726e-01 9.91067111e-01 4.90086913e-01 4.75768596e-01
-2.98455358e-01 -5.40252179e-02 -2.90087104e-01 5.08452095e-02
-9.65210497e-02 -2.35856161e-01 -4.11242187e-01 1.97694123e-01
-1.29999414e-01 -6.66874766e-01 -9.37363386e-01 5.66282332e-01
-1.28137887e+00 3.42315286e-01 6.39825523e-01 9.44802940e-01
4.50168580e-01 6.21602654e-01 -1.21149294e-01 -3.76331061e-01
5.27814209e-01 1.06421143e-01 -7.55742490e-01 2.11203143e-01
-4.10816938e-01 -8.37545842e-03 4.98689890e-01 -5.18176198e-01
-1.12870288e+00 -1.58031985e-01 1.49220645e-01 -7.87540734e-01
2.74218172e-01 5.20141721e-01 -2.02781618e-01 -4.59551930e-01
8.28956366e-02 6.44225895e-01 1.18616998e-01 -3.51201534e-01
1.56367108e-01 2.87231475e-01 6.71398103e-01 -4.63977158e-01
1.24580133e+00 7.06056118e-01 5.35753556e-02 -8.63709867e-01
-4.56838578e-01 -3.46809298e-01 -4.76853997e-01 -2.91373163e-01
5.98770559e-01 -4.63968217e-01 -4.65736657e-01 1.00088918e+00
-1.15991521e+00 -1.77695081e-01 6.43366501e-02 1.64881513e-01
-3.28979939e-01 1.11118329e+00 -7.06544757e-01 -4.83692974e-01
-7.13499486e-01 -1.16575980e+00 1.15816844e+00 5.98051786e-01
3.33680272e-01 -1.02043962e+00 1.83478549e-01 3.51083875e-01
3.33161861e-01 6.64690375e-01 7.64183819e-01 3.15841407e-01
-7.56536961e-01 -2.30508536e-01 -5.18680036e-01 1.83464870e-01
5.38912416e-01 3.63401353e-01 -5.43908358e-01 -5.95948637e-01
-2.19179504e-02 1.07310921e-01 9.68304753e-01 6.32301450e-01
1.03616607e+00 -3.26844066e-01 -5.40859640e-01 1.00669527e+00
1.60608625e+00 3.31732601e-01 9.78832185e-01 3.31539243e-01
1.08408928e+00 4.25937146e-01 8.76057446e-01 2.29495078e-01
6.15101643e-02 6.74152195e-01 4.94290054e-01 -5.55071771e-01
-1.99021026e-01 -5.84056199e-01 2.78322458e-01 6.69196963e-01
-3.53014953e-02 -2.79597759e-01 -5.17360568e-01 3.68857861e-01
-1.79357660e+00 -9.03030336e-01 -2.94545025e-01 2.32667208e+00
5.45214951e-01 8.25365782e-02 -2.46244818e-01 1.68760225e-01
1.25764394e+00 4.82415110e-01 -4.74029601e-01 -1.65659964e-01
-2.28269890e-01 -2.27543935e-02 6.57171130e-01 5.88095486e-01
-1.04878628e+00 1.14853477e+00 4.72337866e+00 1.47662604e+00
-1.30264258e+00 -2.64355570e-01 5.61773181e-01 4.69353497e-01
-4.83911395e-01 2.54004657e-01 -3.99420917e-01 6.99715972e-01
-9.66224894e-02 -1.81723401e-01 3.77929866e-01 3.49408478e-01
4.38839316e-01 3.06771193e-02 -2.34871134e-01 1.12982690e+00
4.63566147e-02 -1.28235948e+00 9.03035998e-02 1.78748950e-01
1.01766193e+00 -4.96612042e-01 2.93589681e-01 -2.95880049e-01
2.70552151e-02 -6.54323876e-01 6.50316596e-01 2.76880652e-01
1.02346754e+00 -1.09233832e+00 5.66509426e-01 -5.61135076e-02
-1.76526594e+00 -2.49489043e-02 -3.02836984e-01 5.40714383e-01
3.26426327e-01 7.56121039e-01 -2.18609527e-01 1.03634536e+00
4.04710501e-01 1.02865970e+00 -4.29243714e-01 1.27803361e+00
-3.08782041e-01 3.91558141e-01 -1.84249058e-01 3.59982759e-01
4.42210644e-01 -7.39115715e-01 7.97286034e-01 9.54011679e-01
5.03288150e-01 1.74576104e-01 1.57693639e-01 8.05377126e-01
-3.19243111e-02 4.60067615e-02 -5.18224061e-01 6.47542179e-02
7.50938594e-01 1.31923139e+00 -9.61846292e-01 -2.29521498e-01
-1.82060301e-01 1.17347407e+00 -6.48904890e-02 2.46681854e-01
-8.11495721e-01 -7.16319501e-01 4.23700780e-01 1.74960122e-01
3.06405723e-01 -5.00774205e-01 -3.65907878e-01 -1.09968019e+00
-1.05399825e-02 -8.45165253e-01 2.47816131e-01 -8.37322533e-01
-1.10697591e+00 2.26687551e-01 -9.16749611e-02 -1.59717178e+00
4.15142238e-01 -1.97394699e-01 -1.11200571e+00 5.48826456e-01
-1.39604068e+00 -1.48642087e+00 -6.17638111e-01 3.42075348e-01
5.35722613e-01 2.92764992e-01 6.56665564e-02 1.63353816e-01
-8.02589357e-01 5.77918112e-01 2.08930269e-01 1.85020059e-01
9.14838850e-01 -6.81705713e-01 5.58165729e-01 1.13914609e+00
-3.65268916e-01 7.21544027e-01 6.05431259e-01 -1.05405056e+00
-1.22180557e+00 -1.05020523e+00 5.51897705e-01 2.19929814e-01
4.08423156e-01 -1.01845972e-01 -1.09227300e+00 3.31953675e-01
5.49480855e-01 -1.67566299e-01 9.56605449e-02 -6.87636614e-01
-4.11080271e-01 -1.02975316e-01 -1.19828832e+00 7.78288960e-01
1.12484229e+00 -1.17610887e-01 -2.90942550e-01 -1.19753383e-01
8.77891183e-01 -4.30404425e-01 -5.95216990e-01 5.32458007e-01
6.53910041e-01 -1.04167092e+00 1.01679659e+00 3.15307498e-01
5.73951542e-01 -7.31481910e-01 3.52784991e-01 -1.29210007e+00
-4.94564712e-01 -9.21368837e-01 6.07441887e-02 1.34974432e+00
-3.11990589e-01 -8.92828763e-01 5.98620832e-01 1.29727989e-01
-3.70270789e-01 -5.75077832e-01 -5.13177156e-01 -8.42898607e-01
-2.19643682e-01 4.01833385e-01 6.52754366e-01 1.30314624e+00
-2.41941646e-01 8.63791034e-02 -4.57599461e-01 1.34427980e-01
9.72787082e-01 3.38476419e-01 6.89673364e-01 -9.56029356e-01
9.58969668e-02 -7.08468795e-01 -2.90115356e-01 -1.14707446e+00
-1.04232728e-01 -5.47541559e-01 -1.04818083e-01 -1.35354650e+00
1.74830943e-01 -4.90817934e-01 3.91278937e-02 2.05626652e-01
-4.62219596e-01 3.36028099e-01 5.98633836e-04 3.23099732e-01
-5.89540824e-02 1.00683010e+00 1.83397090e+00 -1.85540035e-01
-3.83432209e-01 -2.98289746e-01 -5.71470618e-01 5.55868030e-01
1.00475764e+00 -1.78864986e-01 -3.62275243e-01 -5.60748816e-01
-3.32411826e-01 -2.25366261e-02 2.88092524e-01 -9.41021681e-01
3.22089165e-01 -2.13652864e-01 3.37393612e-01 -8.97191107e-01
1.68701589e-01 -8.95902753e-01 2.19008997e-01 9.44270074e-01
2.82380767e-02 -7.80641884e-02 3.63684237e-01 4.82282519e-01
-2.58212000e-01 -1.65194839e-01 1.00584400e+00 3.82720560e-01
-4.81313795e-01 5.02029061e-01 1.95862472e-01 -2.33575210e-01
9.64731455e-01 -6.12471700e-01 -3.42054427e-01 -2.36455798e-01
2.75689334e-01 2.24357173e-01 9.80537653e-01 3.70473981e-01
9.96322572e-01 -1.56089509e+00 -6.38248920e-01 3.63778085e-01
3.13205197e-02 1.57409295e-01 5.61099946e-01 7.25535393e-01
-1.02154434e+00 1.79753661e-01 -4.87847924e-01 -5.03413022e-01
-1.55324900e+00 7.53166318e-01 1.13642499e-01 -1.95982590e-01
-6.96304500e-01 7.48527169e-01 4.44157094e-01 -7.33961398e-03
-2.66131833e-02 -3.54325712e-01 8.41680542e-02 -3.23426485e-01
4.37210113e-01 4.78509665e-01 -2.81919569e-01 -7.52604067e-01
-9.07897726e-02 1.21880746e+00 -2.53193706e-01 -5.72562516e-02
1.07725644e+00 -2.39504576e-01 -2.33926758e-01 -2.98547298e-01
1.25787914e+00 3.33295733e-01 -1.58882952e+00 -2.30278492e-01
-3.60156149e-01 -1.07749307e+00 2.59326607e-01 -2.63537824e-01
-1.54527485e+00 7.90418386e-01 4.56539273e-01 -1.34879231e-01
1.38134003e+00 -4.47575986e-01 1.48658693e+00 -1.76129922e-01
3.29060078e-01 -7.43323445e-01 4.94291693e-01 2.04746455e-01
9.91829634e-01 -8.93874407e-01 1.68629393e-01 -8.64568830e-01
-5.13576984e-01 1.12822807e+00 6.21416926e-01 -5.23428023e-01
2.07789570e-01 2.19256327e-01 -1.70424104e-01 -1.95869580e-01
-7.88635463e-02 7.53580257e-02 4.13950741e-01 4.75421071e-01
1.19691066e-01 -1.28451467e-01 -4.54749316e-01 -8.38933289e-02
-3.08029726e-02 -2.52992690e-01 2.58268565e-01 9.61489081e-01
-5.75834036e-01 -1.17843080e+00 -6.43940151e-01 6.05434701e-02
1.79552380e-02 -1.33670568e-01 -5.13370335e-01 6.15979791e-01
1.57603562e-01 9.19384181e-01 -3.92064452e-02 -6.41648650e-01
1.22565240e-01 -7.04117596e-01 5.61029673e-01 -1.83794782e-01
-4.67698544e-01 5.70018113e-01 -3.11548203e-01 -3.51404727e-01
-1.84770226e-01 -5.06637037e-01 -1.18388903e+00 -5.94659984e-01
-5.66286683e-01 -6.11038767e-02 3.25383097e-01 4.24833953e-01
3.57590556e-01 2.77255684e-01 1.02941942e+00 -1.06846094e+00
3.99757810e-02 -5.59819162e-01 -4.62777972e-01 3.87216300e-01
4.02382702e-01 -6.32355273e-01 -4.29608375e-01 1.79956049e-01] | [9.346170425415039, -2.360717535018921] |
62a9b15b-77eb-40ac-8fcd-c83b506a20e5 | weighted-training-for-cross-task-learning | 2105.14095 | null | https://arxiv.org/abs/2105.14095v2 | https://arxiv.org/pdf/2105.14095v2.pdf | Weighted Training for Cross-Task Learning | In this paper, we introduce Target-Aware Weighted Training (TAWT), a weighted training algorithm for cross-task learning based on minimizing a representation-based task distance between the source and target tasks. We show that TAWT is easy to implement, is computationally efficient, requires little hyperparameter tuning, and enjoys non-asymptotic learning-theoretic guarantees. The effectiveness of TAWT is corroborated through extensive experiments with BERT on four sequence tagging tasks in natural language processing (NLP), including part-of-speech (PoS) tagging, chunking, predicate detection, and named entity recognition (NER). As a byproduct, the proposed representation-based task distance allows one to reason in a theoretically principled way about several critical aspects of cross-task learning, such as the choice of the source data and the impact of fine-tuning. | ['Weijie J. Su', 'Dan Roth', 'Hangfeng He', 'Koby Crammer', 'Shuxiao Chen'] | 2021-05-28 | weighted-training-for-cross-task-learning-1 | https://openreview.net/forum?id=ltM1RMZntpu | https://openreview.net/pdf?id=ltM1RMZntpu | iclr-2022-4 | ['predicate-detection'] | ['natural-language-processing'] | [ 4.22329366e-01 2.07023159e-01 -3.86678845e-01 -4.07228380e-01
-1.37771451e+00 -6.94262445e-01 5.49118638e-01 3.81158739e-01
-9.48844314e-01 6.65932655e-01 4.09199923e-01 -4.87751514e-01
-4.86735970e-01 -2.88298875e-01 -5.21352470e-01 -6.75080597e-01
-3.05981845e-01 3.32472056e-01 2.88072318e-01 3.80612314e-02
2.51224577e-01 2.42089003e-01 -1.01292098e+00 2.60695875e-01
8.32598567e-01 1.01794279e+00 3.32094789e-01 5.70647120e-01
-2.70627081e-01 6.48264349e-01 -4.06745821e-01 -6.62201643e-01
6.32287264e-02 -5.16171269e-02 -1.03632843e+00 -3.25877577e-01
8.10629204e-02 7.76233912e-01 2.62965709e-01 8.84431541e-01
6.86770678e-01 4.37257886e-01 6.93569064e-01 -9.18121576e-01
-4.39294994e-01 7.16596663e-01 -4.30037379e-01 3.01644772e-01
-1.97955081e-03 -3.80710773e-02 1.39582098e+00 -1.07136178e+00
4.08023775e-01 1.04946196e+00 7.87212312e-01 5.91391087e-01
-1.22867358e+00 -5.48071444e-01 6.33102432e-02 -6.08355999e-02
-1.15622211e+00 -5.51348269e-01 4.03535843e-01 -5.11976659e-01
9.68546510e-01 6.45653754e-02 -7.24729476e-03 1.08334470e+00
3.62252355e-01 9.88208055e-01 8.49548340e-01 -7.58144319e-01
3.95330012e-01 -3.97248305e-02 2.87422836e-01 6.73340559e-01
2.19502196e-01 7.86260813e-02 -7.28505015e-01 -4.99727666e-01
3.29335868e-01 -5.09501815e-01 -2.77077645e-01 -7.58414209e-01
-1.41409206e+00 8.12327683e-01 -8.73266067e-03 6.73373818e-01
-2.86358416e-01 1.32646654e-02 9.30204153e-01 3.27464372e-01
8.14595997e-01 5.80481052e-01 -1.20884717e+00 -2.60451555e-01
-7.42614210e-01 -1.18314512e-01 7.15494692e-01 6.66265249e-01
3.78432512e-01 1.03430942e-01 -4.45945472e-01 1.13235223e+00
1.64812252e-01 4.56284940e-01 9.19412076e-01 -6.33899868e-01
8.45448196e-01 1.72091946e-02 2.20304802e-01 -3.38104427e-01
-6.18315399e-01 -3.05951029e-01 -7.20660329e-01 -1.92658901e-01
5.74444413e-01 -5.84275723e-01 -4.83962804e-01 2.13116741e+00
3.13689888e-01 6.95995167e-02 3.71827453e-01 5.69740295e-01
3.72389942e-01 3.72100204e-01 7.34400690e-01 -4.12094146e-01
1.58340716e+00 -7.09593713e-01 -5.00995994e-01 -4.45616037e-01
1.20321226e+00 -7.22399235e-01 1.12683475e+00 6.21550120e-02
-1.02465427e+00 -5.77492654e-01 -6.10084772e-01 5.75124100e-02
-3.03455234e-01 1.69161290e-01 5.54201007e-01 7.09721684e-01
-6.35473073e-01 6.32853687e-01 -6.31724775e-01 -3.38978350e-01
2.03090027e-01 2.09137052e-01 -5.11750340e-01 5.99790476e-02
-1.32171929e+00 1.10356534e+00 8.87300551e-01 -3.08477849e-01
-4.51508313e-01 -1.11583853e+00 -9.13074136e-01 3.34464818e-01
2.77170300e-01 -6.57576025e-01 1.43231344e+00 -9.16922569e-01
-1.40769100e+00 8.99518311e-01 -2.10246816e-01 -6.29777670e-01
2.40120500e-01 -4.60673958e-01 -4.11117464e-01 -1.58020481e-01
8.68006423e-02 3.72368753e-01 8.01380157e-01 -8.14214468e-01
-8.44617307e-01 -3.32186311e-01 -4.69207883e-01 2.18729556e-01
-3.71317863e-01 3.50889921e-01 -1.31309107e-01 -8.26395094e-01
-4.15371656e-01 -6.13861680e-01 -1.72789410e-01 -3.51440132e-01
-1.18435413e-01 -6.83836639e-01 2.98800468e-01 -5.36739469e-01
1.15825450e+00 -2.45055771e+00 8.67141634e-02 -1.04594432e-01
-1.21903345e-01 6.75538599e-01 -4.61149096e-01 5.02723932e-01
-2.85474777e-01 5.75274639e-02 -3.54522467e-01 -3.95162404e-01
7.21707195e-02 1.46804051e-02 -3.54002237e-01 3.99162680e-01
2.17309982e-01 7.89135814e-01 -1.12032282e+00 -6.98940814e-01
3.05737648e-02 2.92547762e-01 -9.91333723e-02 2.59168357e-01
-1.80109233e-01 7.68943280e-02 -5.27873993e-01 1.85483396e-01
4.84007806e-01 -1.85403340e-02 4.61144686e-01 -1.47580996e-01
-2.00853407e-01 6.38494074e-01 -8.94310415e-01 1.51400805e+00
-8.17519546e-01 3.85373950e-01 3.88319977e-02 -1.19114697e+00
7.44428575e-01 4.94855285e-01 4.85936522e-01 -6.06064081e-01
9.32546854e-02 1.40200406e-01 -4.33712341e-02 -4.11650985e-01
3.11815768e-01 -3.84685725e-01 -3.95083576e-01 6.33568048e-01
5.71932316e-01 3.31167042e-01 1.65457040e-01 -5.45064658e-02
1.12612832e+00 1.13471016e-01 1.06230998e+00 -5.88128507e-01
4.04417962e-01 -8.88504386e-02 6.85193658e-01 6.74735188e-01
-2.56120920e-01 1.54930562e-01 4.11085933e-01 -1.45305172e-01
-8.10687900e-01 -1.03379738e+00 -1.33285880e-01 1.84930134e+00
-3.86667162e-01 -1.76125497e-01 -6.21996999e-01 -1.05823350e+00
-1.64593291e-02 8.01760793e-01 -6.14380479e-01 -3.28361183e-01
-7.76755273e-01 -7.24719226e-01 7.72162974e-01 5.00101686e-01
1.69976577e-01 -1.16556215e+00 -5.96040428e-01 1.76622018e-01
-1.85669959e-01 -1.18126178e+00 -8.91001821e-01 7.33184755e-01
-1.00960374e+00 -1.17566681e+00 -8.61719370e-01 -9.94231641e-01
2.36366183e-01 2.91754037e-01 1.16082168e+00 -5.24584532e-01
1.95590988e-01 4.23739314e-01 -5.26166379e-01 -3.50096881e-01
-2.79940933e-01 3.88555139e-01 1.32261351e-01 7.00361328e-03
3.54995847e-01 -2.54776359e-01 -2.11814158e-02 1.37712643e-01
-5.86682200e-01 -3.36300105e-01 8.79031956e-01 1.03580165e+00
4.61371154e-01 -1.44946193e-02 7.12133884e-01 -1.23793423e+00
7.07711458e-01 -4.94323611e-01 -5.32964826e-01 8.08978677e-01
-5.04250169e-01 5.26591122e-01 5.25631368e-01 -5.30050576e-01
-1.43921793e+00 1.03562154e-01 5.11720367e-02 -1.63236767e-01
-4.04719375e-02 4.64703202e-01 -3.01919609e-01 3.87492478e-02
6.26861751e-01 2.07135946e-01 -3.18559468e-01 -5.72605550e-01
5.00241697e-01 3.99746746e-01 3.45047712e-01 -7.39568830e-01
6.49196029e-01 -4.93914671e-02 -2.03388780e-01 -7.78409243e-01
-1.44835162e+00 -8.24411929e-01 -7.17191815e-01 2.34952897e-01
8.76356244e-01 -8.75963926e-01 -7.08562255e-01 2.52764188e-02
-1.16467571e+00 -6.96788907e-01 -3.76294911e-01 6.14682376e-01
-6.74486995e-01 3.13651830e-01 -3.35838795e-01 -9.05515611e-01
-5.37489235e-01 -4.28195238e-01 9.07045484e-01 -2.91344702e-01
-2.72068143e-01 -1.48965025e+00 2.10337862e-01 3.01314950e-01
7.83148706e-02 -2.55311906e-01 1.15040028e+00 -1.19466674e+00
2.30940938e-01 1.30627245e-01 -3.23953718e-01 3.45242143e-01
6.87997639e-02 -3.89554769e-01 -8.48984897e-01 -1.73292652e-01
1.07511513e-01 -3.01794052e-01 9.57254052e-01 3.54052514e-01
9.31225717e-01 -3.78501594e-01 -3.08637530e-01 2.41260946e-01
1.34417474e+00 -5.03443554e-02 1.88844696e-01 1.98323071e-01
5.71177959e-01 8.81283283e-01 9.71622169e-01 4.75042015e-01
2.14780450e-01 6.35850608e-01 -1.51155069e-01 2.36917645e-01
-7.19150379e-02 -2.20762596e-01 4.87705946e-01 7.68288136e-01
2.56298304e-01 -3.01198483e-01 -1.06884778e+00 7.19264925e-01
-1.87634909e+00 -9.82526898e-01 1.43564954e-01 2.40087867e+00
1.01256490e+00 8.36648121e-02 1.08535603e-01 -3.17886285e-02
8.36164176e-01 2.47845337e-01 -2.92904049e-01 -4.24418062e-01
-3.92396227e-02 5.56287706e-01 7.37529933e-01 4.46829259e-01
-1.35858536e+00 1.07990408e+00 6.41984797e+00 1.20230281e+00
-8.25998127e-01 6.08058095e-01 2.22764015e-01 1.44221872e-01
-6.73271269e-02 -2.35741854e-01 -7.01963246e-01 3.87030095e-01
1.11072552e+00 -3.51464689e-01 -1.40106246e-01 8.40969563e-01
1.54589698e-01 6.24363832e-02 -1.11217821e+00 5.55302024e-01
-8.28952715e-02 -1.06136286e+00 -2.82253116e-01 -1.60478950e-01
4.24929768e-01 3.34869623e-01 -4.72018495e-02 5.96208870e-01
6.11848891e-01 -4.74122077e-01 6.64975762e-01 8.46458003e-02
6.82298362e-01 -6.90505743e-01 7.11403370e-01 4.79739457e-01
-1.13671017e+00 -2.91422814e-01 -4.12686199e-01 3.66469979e-01
2.27514803e-01 9.09783065e-01 -7.31695831e-01 5.91436863e-01
4.48173314e-01 2.12254688e-01 -1.42092198e-01 1.04510415e+00
-2.32141420e-01 8.65834594e-01 8.38446841e-02 -3.12882178e-02
2.52447456e-01 2.33060852e-01 3.28672230e-01 1.74365485e+00
2.55226582e-01 2.17242050e-04 1.81094974e-01 5.37000932e-02
-2.35355869e-01 3.42584610e-01 -4.44565922e-01 1.51060913e-02
6.44689560e-01 1.08670056e+00 -4.54029053e-01 -3.39618325e-02
-2.65694171e-01 7.86294043e-01 7.61357129e-01 2.53494889e-01
-6.82411194e-01 -4.73197103e-01 8.47708464e-01 -2.94484526e-01
7.89872169e-01 -3.64056498e-01 -3.85398567e-01 -1.11703205e+00
-9.58960205e-02 -5.16499281e-01 7.65792608e-01 -2.22617194e-01
-1.45325506e+00 3.20543766e-01 2.30408274e-02 -1.04292822e+00
-5.46602011e-02 -6.95147932e-01 -5.66429377e-01 6.92963779e-01
-1.58581364e+00 -8.01637888e-01 5.85004985e-01 3.82891983e-01
5.06729543e-01 -1.76436216e-01 8.81070018e-01 1.45411506e-01
-4.62276846e-01 7.67485976e-01 3.01476419e-01 2.54911780e-01
8.80942345e-01 -1.17432642e+00 3.20394337e-01 7.53600538e-01
3.67330819e-01 4.30778235e-01 5.70510447e-01 -4.68402654e-01
-9.75257099e-01 -1.35896790e+00 1.58984709e+00 -3.49516332e-01
8.94798696e-01 -3.29182655e-01 -8.48574400e-01 6.04674518e-01
5.56482328e-03 5.52819967e-02 1.17276192e+00 6.36720061e-01
-8.20370257e-01 -1.09820284e-01 -1.04823732e+00 3.02496016e-01
9.34508383e-01 -5.11044025e-01 -9.25298333e-01 5.58298469e-01
8.93882334e-01 -9.96581316e-02 -9.50347006e-01 1.73550516e-01
4.45966780e-01 -6.54610634e-01 9.61302817e-01 -8.67340267e-01
-1.69460401e-01 3.42270553e-01 -2.51289010e-01 -1.55106044e+00
-5.04180908e-01 -6.26791179e-01 -5.03241159e-02 1.36732650e+00
6.71346426e-01 -6.41867995e-01 4.45600301e-01 3.70860010e-01
-2.50659138e-01 -5.36923110e-01 -1.16715646e+00 -1.21394956e+00
3.43020916e-01 -6.68225408e-01 2.02960983e-01 9.14412379e-01
2.50893325e-01 6.62623763e-01 -4.40947503e-01 1.05285123e-01
6.06731951e-01 -1.64199308e-01 9.93318185e-02 -1.36679530e+00
-2.55998224e-01 -4.62775528e-01 8.23091622e-03 -8.26762497e-01
6.47050083e-01 -7.62715161e-01 3.06224436e-01 -1.07188833e+00
-9.59071144e-03 -6.89256012e-01 -6.36883438e-01 5.87696850e-01
-2.56937116e-01 -1.95900559e-01 3.40421170e-01 1.76103398e-01
-1.00866914e+00 4.47006255e-01 9.13927794e-01 1.78909361e-01
-1.94794200e-02 2.69255221e-01 -6.77919328e-01 6.66109443e-01
1.06869686e+00 -8.70667219e-01 -2.32278854e-01 -5.16287386e-01
1.60289988e-01 8.70587900e-02 -9.32186991e-02 -6.18348777e-01
8.07447433e-02 -2.64295280e-01 -3.30236368e-02 -1.06094196e-01
1.81560203e-01 -5.22647321e-01 -5.19358814e-01 5.22656143e-01
-8.38472426e-01 -2.15364620e-01 1.07214106e-02 6.87352061e-01
-3.58069167e-02 -4.65302140e-01 9.27750587e-01 2.70343628e-02
-6.07388377e-01 2.09842965e-01 -4.40618217e-01 7.14775145e-01
9.62306023e-01 1.62432164e-01 -3.06446943e-02 8.66813362e-02
-6.21106148e-01 1.77269757e-01 -2.21520513e-01 3.94460201e-01
1.56033918e-01 -1.15362561e+00 -7.01476872e-01 -9.36852843e-02
3.39423478e-01 -4.54613388e-01 2.75724590e-01 8.47917140e-01
1.76613092e-01 9.02112424e-01 1.19319543e-01 -1.72244206e-01
-1.32910299e+00 6.58539593e-01 1.21980896e-02 -8.72782111e-01
-2.28699267e-01 8.40248764e-01 3.01853716e-01 -6.39807284e-01
3.16260278e-01 -2.51532882e-01 -2.21917436e-01 1.91579103e-01
4.98229951e-01 3.54945004e-01 1.89223930e-01 -4.88233805e-01
-4.85610068e-01 3.61992091e-01 -2.86306031e-02 -1.58662364e-01
1.01550674e+00 -6.20273091e-02 2.51767546e-01 5.87889373e-01
1.20862210e+00 -9.79582295e-02 -1.11081254e+00 -7.67821670e-01
6.68409288e-01 -1.93403915e-01 -1.09826900e-01 -7.28140712e-01
-6.32422030e-01 1.01543534e+00 3.14557850e-01 1.16985530e-01
8.11326027e-01 2.40080550e-01 8.58610094e-01 6.39227748e-01
4.25271809e-01 -1.26436663e+00 7.01649487e-02 8.71991634e-01
4.91153657e-01 -1.11775768e+00 -2.96888411e-01 -3.00084561e-01
-9.31566477e-01 8.89948428e-01 3.22636634e-01 8.84236097e-02
7.55359650e-01 1.62419453e-01 2.65795793e-02 6.63195401e-02
-1.01557052e+00 -4.97132719e-01 3.85553956e-01 1.02071333e+00
7.53722429e-01 2.09652185e-01 -3.50062847e-01 5.72741032e-01
4.01804149e-02 -2.70718604e-01 -1.04728676e-01 7.16955543e-01
-6.41554832e-01 -1.31882739e+00 -7.38956556e-02 2.08550557e-01
-7.57436752e-01 -2.79195040e-01 -4.04805467e-02 6.29015446e-01
1.33594468e-01 8.34239423e-01 -2.59722173e-02 3.63059305e-02
2.69561291e-01 4.72461760e-01 4.85574543e-01 -7.93069124e-01
-7.82978594e-01 -3.04839790e-01 3.89261425e-01 -4.01699305e-01
-5.40822268e-01 -7.93584585e-01 -1.06615174e+00 3.65692303e-02
-3.63451600e-01 4.87896591e-01 5.21052361e-01 1.27013290e+00
5.88463902e-01 3.03866267e-01 5.08407533e-01 -5.10471702e-01
-9.36926961e-01 -9.77159977e-01 -4.45044279e-01 2.68311888e-01
1.68839663e-01 -6.89360976e-01 -2.08335415e-01 -1.01904795e-01] | [10.572081565856934, 8.797649383544922] |
80f93d27-2c5a-474f-8c09-9820e3dbc127 | exploiting-geometric-constraints-on-dense | 1909.13258 | null | https://arxiv.org/abs/1909.13258v2 | https://arxiv.org/pdf/1909.13258v2.pdf | EpO-Net: Exploiting Geometric Constraints on Dense Trajectories for Motion Saliency | The existing approaches for salient motion segmentation are unable to explicitly learn geometric cues and often give false detections on prominent static objects. We exploit multiview geometric constraints to avoid such shortcomings. To handle the nonrigid background like a sea, we also propose a robust fusion mechanism between motion and appearance-based features. We find dense trajectories, covering every pixel in the video, and propose trajectory-based epipolar distances to distinguish between background and foreground regions. Trajectory epipolar distances are data-independent and can be readily computed given a few features' correspondences between the images. We show that by combining epipolar distances with optical flow, a powerful motion network can be learned. Enabling the network to leverage both of these features, we propose a simple mechanism, we call input-dropout. Comparing the motion-only networks, we outperform the previous state of the art on DAVIS-2016 dataset by 5.2% in the mean IoU score. By robustly fusing our motion network with an appearance network using the input-dropout mechanism, we also outperform the previous methods on DAVIS-2016, 2017 and Segtrackv2 dataset. | ['Mohsen Ali', 'Richard Hartley', 'Muhammad Faisal', 'Ijaz Akhter'] | 2019-09-29 | epo-net-exploiting-geometric-constraints-on | https://arxiv.org/abs/1909.13258 | https://arxiv.org/pdf/1909.13258 | wacv-2020-3 | ['unsupervised-video-object-segmentation'] | ['computer-vision'] | [-1.41362362e-02 -1.68536395e-01 -2.97549844e-01 -1.91488490e-01
-6.20762527e-01 -8.55158925e-01 6.51351511e-01 -2.32522249e-01
-5.23668885e-01 4.92002517e-01 1.17485709e-01 -1.31180137e-01
2.43305698e-01 -4.30201620e-01 -9.56678391e-01 -7.18859673e-01
-4.74425554e-02 7.60397390e-02 8.18356335e-01 6.68672621e-02
5.97013049e-02 4.51304555e-01 -1.36376464e+00 7.79149234e-02
7.24302173e-01 9.42886531e-01 1.35533825e-01 8.21651280e-01
2.10095972e-01 8.56583059e-01 -9.20222029e-02 -1.00725479e-01
5.92212379e-01 -2.51137137e-01 -8.31719100e-01 2.54244983e-01
1.20467114e+00 -7.91675985e-01 -8.06962430e-01 9.84911025e-01
1.83778778e-01 4.95618254e-01 5.84549904e-01 -1.25879228e+00
-6.19343042e-01 2.27744982e-01 -9.04163241e-01 5.42627335e-01
2.90690213e-01 4.33363646e-01 9.16974247e-01 -1.09250200e+00
1.08048844e+00 1.34346914e+00 7.14656949e-01 6.69784009e-01
-1.23963988e+00 -2.62518495e-01 5.42337716e-01 1.67309612e-01
-1.09675944e+00 -4.99914318e-01 7.62135208e-01 -5.02854526e-01
6.87550187e-01 9.90117639e-02 6.37898982e-01 1.23787713e+00
4.46068347e-02 1.20694971e+00 5.48484921e-01 -9.09824744e-02
2.04986762e-02 -2.82860309e-01 1.32363975e-01 9.86493528e-01
2.75657743e-01 1.71912890e-02 -3.87705475e-01 1.16919845e-01
1.02732730e+00 2.94593990e-01 -4.38711852e-01 -7.37669766e-01
-1.38288736e+00 5.80150485e-01 5.90065122e-01 -5.82662299e-02
-8.11276119e-03 5.77525198e-01 2.37774834e-01 -2.61288155e-02
4.38905656e-01 7.98457637e-02 -3.89017552e-01 7.08479658e-02
-1.01376307e+00 2.52988815e-01 5.00664055e-01 1.07045007e+00
9.77645218e-01 2.29147956e-01 -1.23901926e-01 3.34590167e-01
3.95976722e-01 5.96712589e-01 5.33377603e-02 -1.62103808e+00
4.60366011e-01 4.71288115e-01 2.62123197e-01 -1.14158618e+00
-2.94230074e-01 -1.02090441e-01 -7.60263383e-01 1.63945556e-01
6.94794714e-01 -2.44471088e-01 -1.23518765e+00 1.67803550e+00
7.59808242e-01 6.50900126e-01 -8.17255452e-02 1.10429001e+00
9.27965283e-01 5.37621319e-01 -2.51330435e-01 8.61704499e-02
9.52326834e-01 -1.38052094e+00 -5.35634398e-01 -3.56665075e-01
6.91231370e-01 -4.07716334e-01 8.01134884e-01 9.21139214e-03
-1.21074390e+00 -4.32329357e-01 -1.01651132e+00 -3.59074056e-01
-6.21429235e-02 4.31173965e-02 6.75161600e-01 3.35807860e-01
-1.26938212e+00 8.90290380e-01 -1.26457930e+00 -2.06051096e-01
5.26321769e-01 2.69429177e-01 -4.75452662e-01 -1.72849581e-01
-6.96356058e-01 5.55247962e-01 3.04631852e-02 2.31731281e-01
-1.17004967e+00 -7.92532682e-01 -1.14160562e+00 -2.92244554e-01
3.69407743e-01 -1.04517579e+00 1.05459988e+00 -1.03093827e+00
-1.25276983e+00 7.09407926e-01 -3.82397622e-01 -4.20315206e-01
8.74192119e-01 -3.83866966e-01 6.03184290e-02 6.85276926e-01
9.91592929e-02 1.01145208e+00 7.87334383e-01 -1.32029355e+00
-8.30549657e-01 -4.68847342e-02 1.92082569e-01 1.73976019e-01
-4.80839461e-02 -1.06610291e-01 -7.82214165e-01 -5.69517851e-01
1.20271765e-01 -1.03193378e+00 -3.81685466e-01 5.38951278e-01
-4.20242608e-01 9.76731032e-02 1.25145233e+00 -6.17562234e-01
8.34279895e-01 -2.09255862e+00 2.58195639e-01 -2.63593972e-01
3.59825611e-01 2.70964652e-01 -1.86373278e-01 -3.17801893e-01
3.73668641e-01 1.22113876e-01 -3.62902671e-01 -7.15804398e-01
-2.74049938e-01 3.30647469e-01 -2.17310071e-01 9.27483439e-01
5.02910137e-01 1.04015982e+00 -1.16315758e+00 -4.92755651e-01
4.20719713e-01 5.27250767e-01 -7.82033741e-01 -3.51130543e-03
-3.30159336e-01 6.10349655e-01 -4.91882175e-01 7.29670227e-01
7.64554322e-01 -4.41209495e-01 -2.21734881e-01 -9.24873203e-02
-8.29160959e-02 6.15189746e-02 -1.19130838e+00 1.99169350e+00
3.85667756e-02 9.89196360e-01 2.94824928e-01 -6.96817994e-01
4.59850758e-01 1.02635778e-01 7.05098331e-01 -2.46494472e-01
1.22004010e-01 7.01947734e-02 -2.50859231e-01 -4.55962360e-01
5.05790770e-01 4.23126787e-01 3.08911920e-01 9.77380872e-02
1.39309987e-01 2.05757484e-01 1.84206605e-01 3.72536749e-01
1.20496094e+00 6.71081662e-01 -2.59459257e-01 -2.01076433e-01
4.42310691e-01 -8.65363181e-02 1.09165323e+00 6.19665384e-01
-6.03646219e-01 1.17526972e+00 2.62919903e-01 -6.98065817e-01
-9.71747398e-01 -1.17179453e+00 1.31130457e-01 6.99217021e-01
5.87956905e-01 -2.01169878e-01 -4.95256752e-01 -9.81442869e-01
3.84585559e-02 1.33498743e-01 -6.37605786e-01 1.02103896e-01
-8.47291410e-01 -4.81820494e-01 3.51672888e-01 8.05305660e-01
5.44105232e-01 -6.06331050e-01 -6.29725993e-01 2.23660320e-01
-3.76977295e-01 -1.61024916e+00 -8.56352746e-01 -1.40769752e-02
-9.77699101e-01 -9.44553435e-01 -8.33550215e-01 -7.08827257e-01
6.09321058e-01 7.97885120e-01 1.03914821e+00 1.11087039e-01
-3.12168598e-01 4.62832719e-01 -2.87568480e-01 4.32363562e-02
-4.59851883e-02 -1.19014934e-01 7.26067424e-02 1.38987392e-01
2.14454923e-02 -5.79344809e-01 -1.12060130e+00 3.66320908e-01
-7.52849758e-01 3.32825780e-02 3.63089234e-01 6.30534351e-01
5.01322567e-01 -7.17463136e-01 1.56928539e-01 -4.63234633e-01
-5.08360684e-01 -6.20243788e-01 -5.31195164e-01 4.56575900e-02
-7.89448023e-02 6.18345961e-02 3.53287011e-01 -5.49176395e-01
-7.56008744e-01 5.21643579e-01 2.56359875e-01 -1.14535403e+00
-1.64384335e-01 -1.66081056e-01 -9.27286595e-02 -3.79719019e-01
4.07889038e-01 6.84974790e-02 -1.76794961e-01 -2.99267590e-01
6.29179418e-01 -1.26796132e-02 8.89703929e-01 -4.60409790e-01
9.20669258e-01 1.21959341e+00 9.03368145e-02 -8.17583144e-01
-8.45210671e-01 -7.78499961e-01 -9.51125443e-01 -2.20675305e-01
1.28972495e+00 -1.19360900e+00 -6.06621087e-01 3.00399184e-01
-1.32757771e+00 -4.56752479e-01 -2.41446402e-02 6.21689260e-01
-6.16250932e-01 6.99541628e-01 -7.59057462e-01 -6.58386648e-01
-2.02096611e-01 -1.25586236e+00 1.12651455e+00 2.61389434e-01
-5.90385161e-02 -1.11665118e+00 -1.37667283e-01 1.65985510e-01
1.59798130e-01 7.01564312e-01 2.92431653e-01 -2.63800651e-01
-1.11529946e+00 1.27849251e-01 -2.74293393e-01 2.59371638e-01
6.20975420e-02 4.37343359e-01 -1.03929532e+00 -3.52309167e-01
-2.52203852e-01 -1.84908301e-01 1.42888987e+00 6.70360327e-01
8.39576244e-01 -2.26876557e-01 -4.79448676e-01 1.27123892e+00
1.23111510e+00 -1.44212067e-01 4.34425056e-01 3.99220794e-01
1.34830022e+00 4.90245461e-01 4.95056421e-01 1.87440515e-01
5.26853025e-01 5.43097854e-01 6.63253963e-01 -3.12737316e-01
-3.50825071e-01 -9.24417302e-02 6.99950635e-01 6.74248636e-01
-2.94664204e-01 -5.38475476e-02 -6.81894422e-01 8.88744235e-01
-2.14015460e+00 -1.05916667e+00 -3.34382772e-01 1.95698118e+00
4.86154079e-01 5.50176911e-02 1.98746249e-01 -4.10089821e-01
5.46513975e-01 5.32369196e-01 -6.87435031e-01 4.77384925e-02
-3.27062160e-01 -4.49176967e-01 8.07399333e-01 7.19446719e-01
-1.60679638e+00 1.04177427e+00 6.23783207e+00 3.30807388e-01
-1.01570535e+00 5.15047945e-02 6.83918238e-01 -3.10639083e-01
-3.01865369e-01 1.48155183e-01 -9.31410134e-01 3.02848756e-01
6.16176903e-01 2.22459152e-01 7.41429105e-02 7.21296310e-01
3.71140599e-01 -6.12116493e-02 -1.20903003e+00 8.55413795e-01
1.42692432e-01 -1.66498125e+00 1.65683001e-01 -1.93178114e-02
1.09159791e+00 3.79301429e-01 1.22686185e-01 -9.81762335e-02
4.75617141e-01 -7.62008607e-01 8.64453554e-01 5.13516963e-01
4.03115273e-01 -5.71183324e-01 3.11431348e-01 1.09909944e-01
-1.49547207e+00 1.10820569e-01 -3.70190114e-01 7.89196789e-02
3.88633728e-01 3.69464517e-01 -4.05119658e-01 7.98735440e-01
8.23363066e-01 1.35367501e+00 -4.51026767e-01 1.21636939e+00
-3.21400912e-05 3.91205907e-01 -4.16172504e-01 3.54271650e-01
6.07581615e-01 -3.87441397e-01 8.46909642e-01 1.20709395e+00
2.22564355e-01 -1.35579079e-01 3.52204382e-01 9.92500842e-01
-8.19325075e-02 -2.93136328e-01 -8.08720469e-01 3.53346735e-01
1.32700071e-01 1.30004668e+00 -9.02891040e-01 -4.34555203e-01
-7.41051733e-01 1.26602113e+00 2.27434948e-01 6.42669857e-01
-8.51382792e-01 1.31668178e-02 9.58002567e-01 -1.58332363e-02
6.40258729e-01 -5.70630372e-01 8.00025910e-02 -1.60801780e+00
1.96032673e-01 -5.89701831e-01 1.83382258e-01 -4.92523134e-01
-1.03788662e+00 1.98530957e-01 -6.50324114e-03 -1.30668199e+00
-8.92282426e-02 -6.95730150e-01 -8.13255668e-01 4.85514104e-01
-1.60576928e+00 -1.24314761e+00 -2.86358625e-01 6.06350899e-01
6.81687117e-01 2.97424138e-01 1.97559699e-01 3.08464199e-01
-7.30461776e-01 3.13900769e-01 1.42440572e-01 5.01296163e-01
8.09139609e-01 -1.36014986e+00 6.61329567e-01 1.29610801e+00
2.55995244e-01 5.67539573e-01 4.93794471e-01 -6.33050680e-01
-1.65974236e+00 -1.46680093e+00 4.02202308e-01 -7.70448864e-01
6.43120170e-01 -3.73109519e-01 -1.08286619e+00 8.90943289e-01
1.33306712e-01 7.81638086e-01 1.76096931e-01 -4.63284403e-01
-3.39007825e-01 1.28839463e-01 -8.15033555e-01 7.32174218e-01
1.46335280e+00 -2.26932123e-01 -4.23620701e-01 1.17745742e-01
1.05777109e+00 -5.17378986e-01 -5.53717256e-01 4.64456201e-01
4.50519204e-01 -8.00886571e-01 1.28728700e+00 -7.11567998e-01
4.19112831e-01 -6.58829689e-01 -2.33565316e-01 -9.88310575e-01
-2.48986736e-01 -1.02169812e+00 -4.36899215e-01 1.02567863e+00
3.10618848e-01 -1.94367036e-01 9.29454684e-01 5.61419189e-01
-3.16351563e-01 -6.63109958e-01 -8.67272854e-01 -8.38978887e-01
1.61164612e-01 -3.09908360e-01 5.52267320e-02 1.01944387e+00
-4.92682457e-01 1.75406598e-02 -5.46872616e-01 3.76141250e-01
7.46226907e-01 5.26092499e-02 8.24178219e-01 -1.08348978e+00
-3.03584069e-01 -4.59769636e-01 -5.27824819e-01 -1.54331636e+00
1.72498494e-01 -8.91687155e-01 1.41907141e-01 -1.48314857e+00
2.23665461e-01 -4.64094424e-04 -1.13257356e-01 3.46875608e-01
-3.43184829e-01 2.76106447e-01 2.80278027e-01 2.28222534e-01
-9.26256597e-01 6.11407816e-01 1.55668819e+00 -1.98099852e-01
-2.83905357e-01 -2.23485723e-01 -2.33864427e-01 1.07808733e+00
4.40348476e-01 -3.39116991e-01 -2.81442821e-01 -6.94499016e-01
-1.93618294e-02 -7.71524534e-02 6.51404202e-01 -9.01797295e-01
4.05879378e-01 -2.50674844e-01 5.63077986e-01 -7.62608051e-01
3.75885874e-01 -4.97842312e-01 -1.80390596e-01 3.84962618e-01
6.70329994e-03 2.26587892e-01 1.94266200e-01 1.06187069e+00
-4.31583598e-02 2.11662054e-01 6.49054945e-01 -7.16842860e-02
-1.06489933e+00 7.06545472e-01 -3.94096047e-01 2.81712145e-01
1.04009843e+00 -2.89732873e-01 -4.60201621e-01 -3.25710595e-01
-7.00710177e-01 4.67514038e-01 7.38689721e-01 4.55803514e-01
7.21044838e-01 -1.27403891e+00 -5.45421779e-01 2.44482029e-02
-1.72851205e-01 4.62036073e-01 1.95416644e-01 1.26165175e+00
-7.37386942e-01 3.14434245e-02 -1.06906101e-01 -1.07278419e+00
-1.11631441e+00 5.38854659e-01 5.08895755e-01 2.07083642e-01
-1.00101936e+00 7.76168287e-01 5.24524570e-01 -1.28756300e-01
3.90219092e-01 -6.46703005e-01 -9.70508344e-03 -6.28325269e-02
5.69456100e-01 3.45831543e-01 -3.66615027e-01 -7.25257874e-01
-5.17205179e-01 8.31317306e-01 -1.68183423e-03 -1.88420892e-01
1.25590384e+00 -3.34865659e-01 1.61301002e-01 4.58352417e-01
1.47865820e+00 -1.14865504e-01 -2.18672585e+00 -2.56023109e-01
-5.77413104e-02 -7.17841923e-01 -3.55421081e-02 -3.62082683e-02
-1.37787139e+00 1.07116258e+00 4.15344387e-01 -1.57264933e-01
7.13583350e-01 -9.28527862e-02 7.31489182e-01 3.53724480e-01
8.45718384e-02 -8.95640731e-01 3.29836071e-01 6.88148022e-01
5.10494769e-01 -1.50259972e+00 -9.25512314e-02 -5.16581059e-01
-6.12391174e-01 1.13527179e+00 6.86413109e-01 -3.94723773e-01
5.09942055e-01 2.26767957e-01 7.80442730e-02 9.01312456e-02
-7.90654302e-01 -4.00834203e-01 5.64302027e-01 6.19006276e-01
2.11635694e-01 -3.54281515e-01 1.40989274e-01 6.25630021e-02
3.60400707e-01 -3.40552241e-01 7.29765117e-01 9.34316814e-01
-5.33578575e-01 -6.84729099e-01 -1.08936377e-01 1.45134717e-01
-5.91865242e-01 7.35161677e-02 -3.81388694e-01 8.60173583e-01
-1.19772553e-01 8.04245651e-01 2.88984984e-01 -3.11145753e-01
6.68597966e-02 -2.29971960e-01 3.90382916e-01 -4.08913881e-01
-2.25215107e-01 2.22186968e-01 1.00100432e-02 -1.04046953e+00
-7.60295749e-01 -9.26260531e-01 -1.36145592e+00 -3.52371961e-01
-1.85175121e-01 -4.76343066e-01 1.47297516e-01 8.81444216e-01
4.81553614e-01 4.02470350e-01 5.89077055e-01 -1.39987385e+00
-2.69310802e-01 -4.14723307e-01 -1.91902667e-01 5.50369382e-01
9.02348876e-01 -7.50888228e-01 -7.13144422e-01 2.75400728e-01] | [8.989665985107422, -0.5231369733810425] |
4e24c826-d670-4f04-8975-8df52679ead7 | efficient-and-direct-inference-of-heart-rate | 2303.13637 | null | https://arxiv.org/abs/2303.13637v1 | https://arxiv.org/pdf/2303.13637v1.pdf | Efficient and Direct Inference of Heart Rate Variability using Both Signal Processing and Machine Learning | Heart Rate Variability (HRV) measures the variation of the time between consecutive heartbeats and is a major indicator of physical and mental health. Recent research has demonstrated that photoplethysmography (PPG) sensors can be used to infer HRV. However, many prior studies had high errors because they only employed signal processing or machine learning (ML), or because they indirectly inferred HRV, or because there lacks large training datasets. Many prior studies may also require large ML models. The low accuracy and large model sizes limit their applications to small embedded devices and potential future use in healthcare. To address the above issues, we first collected a large dataset of PPG signals and HRV ground truth. With this dataset, we developed HRV models that combine signal processing and ML to directly infer HRV. Evaluation results show that our method had errors between 3.5% to 25.7% and outperformed signal-processing-only and ML-only methods. We also explored different ML models, which showed that Decision Trees and Multi-level Perceptrons have 13.0% and 9.1% errors on average with models at most hundreds of KB and inference time less than 1ms. Hence, they are more suitable for small embedded devices and potentially enable the future use of PPG-based HRV monitoring in healthcare. | ['Wei Wang', 'Houbing Song', 'Dakai Zhu', 'Mimi Xie', 'Jingye Xu', 'Yuntong Zhang'] | 2023-03-23 | null | null | null | null | ['photoplethysmography-ppg', 'heart-rate-variability'] | ['medical', 'medical'] | [ 2.75267094e-01 -1.11909909e-02 -1.83745399e-01 -4.55872893e-01
-3.35262865e-01 -1.83515862e-01 -1.15647145e-01 1.32921278e-01
-2.43411943e-01 9.92256165e-01 1.08755872e-01 -3.59619766e-01
3.09114426e-01 -1.03880882e+00 -1.22692764e-01 -4.63484317e-01
-1.02954343e-01 -1.18764304e-01 -3.34293962e-01 1.30753025e-01
-3.94722596e-02 3.30452472e-01 -1.36084926e+00 7.07869977e-02
8.02477539e-01 1.22231400e+00 -3.70539606e-01 9.64660525e-01
2.62228847e-01 4.37495798e-01 -8.79067302e-01 1.96453989e-01
1.03075221e-01 -5.72066605e-01 -1.67092636e-01 -5.82756698e-01
1.80833384e-01 -6.40567482e-01 -2.97382474e-01 4.04584497e-01
1.13525605e+00 -1.37467861e-01 2.46418640e-02 -1.18822038e+00
-4.04250532e-01 3.44813347e-01 -2.93923765e-01 4.25975740e-01
7.00939119e-01 4.40125585e-01 3.77900362e-01 -3.63829643e-01
-2.86743902e-02 7.38441467e-01 1.31728148e+00 5.82149327e-01
-1.51165044e+00 -8.55438352e-01 -7.57172704e-01 9.71927717e-02
-1.47052956e+00 -8.16812932e-01 5.53865969e-01 -3.67054999e-01
1.44552398e+00 6.00457847e-01 1.13249183e+00 9.26167071e-01
6.06997192e-01 -2.03581184e-01 1.51741612e+00 -1.80408701e-01
1.77144453e-01 7.87154585e-02 3.04398775e-01 5.65114379e-01
6.31216228e-01 3.22711200e-01 -4.78311747e-01 -4.96073276e-01
8.98480594e-01 1.48333624e-01 -4.83954698e-01 7.96457350e-01
-1.33708882e+00 5.65705657e-01 1.32947072e-01 1.55495137e-01
-5.72272956e-01 1.36122257e-01 4.81061578e-01 2.03009963e-01
3.25637341e-01 3.19437742e-01 -5.03861547e-01 -4.94429618e-01
-9.62016463e-01 -1.51329130e-01 1.18614399e+00 5.46819150e-01
3.55218410e-01 1.92034289e-01 -2.78048873e-01 6.59965694e-01
2.65255481e-01 7.94330060e-01 5.85655868e-01 -1.07693100e+00
3.48827869e-01 4.88634467e-01 -1.28277361e-01 -1.30532217e+00
-8.78399312e-01 -7.13910609e-02 -1.43060696e+00 -2.70759046e-01
2.61709541e-01 -3.77237648e-01 -6.67249203e-01 1.49737871e+00
1.91114664e-01 6.43168509e-01 -7.04201236e-02 8.48182321e-01
1.28317070e+00 5.74649930e-01 6.00924119e-02 -8.55867684e-01
1.61498082e+00 -2.61274993e-01 -1.11923110e+00 -2.45646983e-01
2.65613496e-01 -3.52599442e-01 9.90324795e-01 2.16754988e-01
-9.65939224e-01 -8.14490020e-01 -1.13963544e+00 -8.18081014e-03
3.64707075e-02 -1.48961276e-01 4.03623849e-01 1.23253787e+00
-1.04553843e+00 9.73978341e-01 -1.06860137e+00 -3.53943288e-01
2.38325015e-01 2.62341857e-01 1.67821795e-01 3.31936568e-01
-1.50428474e+00 8.87719393e-01 5.01142442e-02 3.59303534e-01
4.56947647e-02 -8.35000038e-01 -8.46593380e-01 2.92825326e-02
-3.30408245e-01 -9.68608201e-01 7.89520860e-01 -1.04663052e-01
-1.79977214e+00 7.12551653e-01 -3.31924558e-01 -5.99427640e-01
2.47500345e-01 -1.18076950e-01 -8.14147830e-01 2.67452866e-01
-4.14900750e-01 8.41878727e-02 7.19847977e-01 -3.93566221e-01
1.09147005e-01 -4.60077792e-01 -5.85004866e-01 -1.92760810e-01
-3.26143205e-02 -2.54696995e-01 6.36288226e-02 -1.61120012e-01
2.65400469e-01 -8.68064404e-01 3.65140364e-02 -2.94062555e-01
-1.57662585e-01 3.12398285e-01 2.63122141e-01 -1.18109822e+00
1.69865060e+00 -1.89136410e+00 -6.53617382e-01 1.25331342e-01
4.50456440e-01 3.14369708e-01 4.18093204e-01 2.40287274e-01
1.37855664e-01 3.46584290e-01 -8.90838653e-02 -1.00447379e-01
-2.66393304e-01 2.66197294e-01 -2.36486897e-01 6.14197850e-01
-2.85771519e-01 1.02788532e+00 -5.19498765e-01 -5.35393536e-01
6.10459805e-01 8.94252896e-01 -2.51865417e-01 1.80622116e-01
4.60924208e-01 6.17689371e-01 9.23222154e-02 7.91076958e-01
4.97901648e-01 -2.39305332e-01 2.60526061e-01 -4.54819828e-01
1.19232789e-01 4.56078768e-01 -8.41885269e-01 1.12268984e+00
-3.58609796e-01 6.88850462e-01 -3.23222369e-01 -6.67079270e-01
1.19483542e+00 6.40603900e-01 6.66714668e-01 -7.76074231e-01
1.91539839e-01 -1.23826526e-02 9.05411392e-02 -8.71959448e-01
-9.05271154e-03 -4.85441744e-01 -2.34216917e-02 3.16773534e-01
-4.15054262e-01 -8.14363062e-02 -2.61783183e-01 -6.04651988e-01
1.25852215e+00 -1.60348177e-01 9.23636973e-01 1.87610239e-02
3.69957477e-01 -4.62799042e-01 1.01717651e+00 7.33417928e-01
-7.96299577e-01 5.44234037e-01 2.42650077e-01 -6.22610331e-01
-5.62451243e-01 -8.30645800e-01 -4.93917733e-01 4.76759613e-01
-2.56862223e-01 -5.53887427e-01 -4.54982102e-01 1.30034521e-01
2.41160899e-01 7.09627986e-01 -2.80147374e-01 -1.19318597e-01
-6.25833750e-01 -1.04930925e+00 1.07324195e+00 6.10919774e-01
6.93108022e-01 -1.20305443e+00 -1.04589820e+00 3.80672097e-01
-6.19346857e-01 -1.31322134e+00 -1.05438046e-01 -2.75809705e-01
-1.45485294e+00 -9.44375038e-01 -2.79820204e-01 -1.53543398e-01
2.28695005e-01 -1.81021720e-01 1.15609586e+00 8.43310356e-02
-6.17405176e-01 3.04315597e-01 1.08980387e-01 -5.71958482e-01
-2.81610161e-01 -2.84171045e-01 3.85933429e-01 -3.35170031e-01
7.61248469e-01 -9.81515765e-01 -1.05517101e+00 -4.17693965e-02
-1.05276972e-01 -1.59756422e-01 4.26706761e-01 5.49635649e-01
6.93385780e-01 -1.19310088e-01 8.43618274e-01 -5.48764765e-01
7.69647479e-01 -2.71464467e-01 -4.44464236e-01 -1.72564283e-01
-1.15483224e+00 -4.13084477e-01 5.63123703e-01 -3.11580688e-01
-4.98552859e-01 -2.79263437e-01 -1.58864304e-01 -3.10847729e-01
-2.33028859e-01 5.05831778e-01 2.11533412e-01 2.12376118e-01
8.17586124e-01 2.20659077e-01 3.99374872e-01 -1.41131058e-01
-2.22501025e-01 9.25213277e-01 4.56676483e-01 -1.71509400e-01
1.75085634e-01 1.26164213e-01 1.62464216e-01 -1.27051008e+00
-2.20016211e-01 -2.25853726e-01 -3.64370078e-01 -2.04043493e-01
8.33490968e-01 -1.05055392e+00 -1.39903855e+00 4.18978184e-01
-8.81266057e-01 -3.34585339e-01 -6.36597052e-02 7.59814739e-01
-4.51278687e-01 4.62140560e-01 -1.05454290e+00 -1.16608727e+00
-1.24430215e+00 -4.20608878e-01 7.16299415e-01 3.58794481e-01
-8.37892056e-01 -7.59424984e-01 1.44188717e-01 6.58540249e-01
6.81907654e-01 7.52820969e-01 6.34816885e-01 -1.79768443e-01
8.14174581e-03 -3.25145721e-01 -1.10887416e-01 2.23932594e-01
3.61586988e-01 -1.61867857e-01 -1.29661775e+00 -3.24970126e-01
4.71090376e-01 -3.93090621e-02 3.39885712e-01 6.84815705e-01
1.25938916e+00 -4.83481914e-01 -2.44166031e-01 6.07614636e-01
1.38095343e+00 2.64518499e-01 1.06427097e+00 -1.31103918e-01
4.96936858e-01 1.97934493e-01 2.05491513e-01 6.62838757e-01
4.71179098e-01 3.77043933e-01 -2.21956700e-01 -7.89367557e-02
1.42917514e-01 5.05458340e-02 2.90788680e-01 1.28421640e+00
-6.23945832e-01 3.57530594e-01 -8.33620191e-01 1.63935021e-01
-1.45903528e+00 -1.09478974e+00 -3.57812583e-01 2.58995128e+00
1.18327451e+00 1.30758226e-01 2.06153959e-01 4.45272595e-01
6.21097684e-01 9.44411568e-03 -7.47511268e-01 -6.79152489e-01
4.81970608e-01 5.47340393e-01 4.47117627e-01 1.22125953e-01
-7.29771733e-01 2.44475938e-02 7.18825293e+00 -3.40490699e-01
-1.24562573e+00 -2.29617655e-02 5.55183768e-01 -3.02088469e-01
2.98930526e-01 -4.94265139e-01 -6.33114517e-01 8.27405691e-01
1.73834825e+00 -2.04007804e-01 5.15668392e-01 5.12764931e-01
6.96178496e-01 -1.64575323e-01 -9.97562766e-01 1.54985416e+00
-1.62814274e-01 -8.60321522e-01 -6.82829201e-01 1.83562711e-01
1.75732240e-01 -2.63667703e-02 -4.88656849e-01 4.56696421e-01
-6.03651881e-01 -1.22553718e+00 -2.59135723e-01 8.08269501e-01
1.15120375e+00 -4.63170171e-01 9.95008588e-01 2.46402115e-01
-1.24275744e+00 1.35484129e-01 -3.70074064e-01 -6.30392790e-01
7.26851672e-02 1.18125987e+00 -9.35485601e-01 1.65586218e-01
6.97847545e-01 5.06986797e-01 -1.49617583e-01 7.78156698e-01
-7.77157992e-02 9.67784286e-01 -5.85155487e-01 -1.67278975e-01
-8.55585098e-01 -1.61928773e-01 2.22640261e-01 1.09024429e+00
5.26877284e-01 6.02957010e-01 1.12635354e-02 8.34265828e-01
2.03920364e-01 -1.72045454e-01 -5.88490725e-01 -1.55281827e-01
7.97759116e-01 1.14365017e+00 -4.68211770e-01 -4.15753245e-01
-6.49313331e-01 5.99864483e-01 -3.05484504e-01 2.45156676e-01
-1.06140625e+00 -6.27924263e-01 8.74166846e-01 3.08194041e-01
-2.06425890e-01 -2.14683667e-01 -9.83696401e-01 -1.09941339e+00
6.37407079e-02 -7.91501582e-01 3.82801175e-01 -5.16383767e-01
-1.14358222e+00 3.38637024e-01 -3.81007761e-01 -1.17272699e+00
-3.20769072e-01 -1.11138359e-01 -5.47055900e-01 1.13081431e+00
-1.34004581e+00 -3.20245266e-01 -7.34614491e-01 4.20973867e-01
-2.89645530e-02 4.66463685e-01 1.40067959e+00 3.89904171e-01
-8.92940402e-01 4.80014712e-01 -4.26472485e-01 1.36274174e-01
6.53550267e-01 -1.02917600e+00 3.73040706e-01 5.47503710e-01
-5.38364768e-01 8.51086974e-01 6.09561265e-01 -6.80304945e-01
-1.74944878e+00 -9.23737943e-01 1.10906327e+00 -4.63605165e-01
8.36709365e-02 -5.68693168e-02 -9.41731513e-01 4.81632411e-01
-9.56484750e-02 2.56026745e-01 1.29357719e+00 1.23217002e-01
1.51576534e-01 -5.86379349e-01 -1.57281232e+00 2.62823015e-01
7.81828344e-01 -6.25689745e-01 -7.85042405e-01 -3.01321924e-01
4.84202713e-01 -5.79267442e-01 -1.62760389e+00 6.78926051e-01
1.06517971e+00 -9.55281138e-01 1.04593039e+00 1.80376675e-02
1.99149981e-01 -1.21293068e-01 1.23992629e-01 -1.00352407e+00
-2.83773631e-01 -5.71902990e-01 -8.17647517e-01 1.03811598e+00
1.08241878e-01 -1.24128771e+00 4.15414453e-01 1.29946136e+00
2.72523761e-01 -4.54065353e-01 -7.34311461e-01 -5.24072945e-01
-5.65813065e-01 -4.45630848e-01 6.05669498e-01 1.01142728e+00
5.63755989e-01 4.05655742e-01 -4.38120455e-01 1.01768970e-01
6.85623705e-01 3.35591018e-01 5.23841321e-01 -1.38810861e+00
-3.08170170e-01 -1.45372273e-02 -4.96136814e-01 -3.80920172e-01
-4.21490401e-01 -6.53228164e-01 7.97344446e-02 -1.56966472e+00
-5.68985082e-02 -1.90572158e-01 -5.53821802e-01 6.99545264e-01
-3.15997809e-01 5.60273707e-01 -1.47985563e-01 1.41165212e-01
2.71698803e-01 1.00946084e-01 9.62236166e-01 2.20709920e-01
-9.35053229e-01 1.05843939e-01 -6.49623454e-01 5.10647416e-01
1.28775525e+00 -2.39530608e-01 -3.66566867e-01 3.33716661e-01
9.74650830e-02 6.86372101e-01 3.87320817e-01 -1.15967405e+00
-1.64148845e-02 -5.66268228e-02 9.93104160e-01 -4.23988879e-01
4.21501517e-01 -5.42815328e-01 4.85921383e-01 9.29408252e-01
8.89542997e-02 -2.58316338e-01 4.17567104e-01 2.78883040e-01
1.56974822e-01 3.23504120e-01 5.84775746e-01 -1.65422440e-01
-2.51956165e-01 7.64629841e-02 -4.88952816e-01 -6.11593090e-02
7.96053290e-01 -4.94907349e-01 -3.57941508e-01 -3.14326972e-01
-7.61895180e-01 -9.21063125e-02 5.85607141e-02 4.46146645e-04
8.68615687e-01 -1.15827000e+00 -5.78948438e-01 5.75932026e-01
-4.99100015e-02 -2.92735398e-01 1.88883945e-01 1.22452188e+00
-5.83580375e-01 6.06638551e-01 -3.70924324e-01 -7.73476958e-01
-1.47601223e+00 4.53902870e-01 2.96362996e-01 1.32418856e-01
-1.01943588e+00 4.74194847e-02 -6.80438280e-01 -3.13097239e-01
7.31062964e-02 -7.97252715e-01 -2.04874799e-01 -1.98963732e-02
8.29210520e-01 9.22763646e-01 1.07545294e-01 -2.20585763e-01
-6.05526388e-01 5.07623613e-01 7.27599502e-01 1.60986021e-01
1.01420867e+00 -4.20380503e-01 -2.45836258e-01 7.73546696e-01
1.08941436e+00 -2.31522039e-01 -5.97640395e-01 6.72641844e-02
-3.34480107e-01 -3.13403219e-01 6.11788742e-02 -6.77804947e-01
-8.30688655e-01 7.76489735e-01 7.97277510e-01 4.26565051e-01
1.54488850e+00 -6.57672107e-01 1.10202754e+00 2.82503486e-01
5.08504570e-01 -9.64708984e-01 -4.96374786e-01 7.75222033e-02
5.06670713e-01 -9.83377576e-01 3.52710158e-01 -3.97570759e-01
-3.78767043e-01 1.24085164e+00 4.05236721e-01 2.21721068e-01
7.69640267e-01 3.32088321e-01 1.82875410e-01 -1.66708082e-02
-7.56530344e-01 9.51693878e-02 1.53306112e-01 5.38563550e-01
8.40481877e-01 4.94296640e-01 -7.36647546e-01 6.84959888e-01
-4.84597147e-01 8.58095229e-01 4.17692751e-01 8.06212306e-01
-2.82655030e-01 -5.92344463e-01 -4.67060357e-01 1.07113791e+00
-6.10414684e-01 -1.61643162e-01 3.47816460e-02 3.54676753e-01
-5.80929071e-02 1.38669479e+00 8.79789740e-02 -5.18605530e-01
3.24650884e-01 5.69644690e-01 4.67703909e-01 -3.93562764e-01
-5.96194804e-01 -4.84001227e-02 2.98928678e-01 -8.73922765e-01
-3.91196609e-01 -3.89859557e-01 -1.27412510e+00 -7.99909651e-01
1.36728510e-02 -2.87055790e-01 5.51014245e-01 6.71557784e-01
7.46647835e-01 7.12454975e-01 4.36407238e-01 -4.16117758e-01
-4.15893465e-01 -1.18572736e+00 -5.84450901e-01 -3.75252701e-02
4.44271296e-01 1.77337378e-02 -3.81086320e-01 1.25280872e-01] | [13.900629043579102, 3.0652058124542236] |
f8b11423-64b4-4797-b1d3-b70dfa951ce4 | learning-continuous-grasping-function-with-a | 2207.05053 | null | https://arxiv.org/abs/2207.05053v3 | https://arxiv.org/pdf/2207.05053v3.pdf | Learning Continuous Grasping Function with a Dexterous Hand from Human Demonstrations | We propose to learn to generate grasping motion for manipulation with a dexterous hand using implicit functions. With continuous time inputs, the model can generate a continuous and smooth grasping plan. We name the proposed model Continuous Grasping Function (CGF). CGF is learned via generative modeling with a Conditional Variational Autoencoder using 3D human demonstrations. We will first convert the large-scale human-object interaction trajectories to robot demonstrations via motion retargeting, and then use these demonstrations to train CGF. During inference, we perform sampling with CGF to generate different grasping plans in the simulator and select the successful ones to transfer to the real robot. By training on diverse human data, our CGF allows generalization to manipulate multiple objects. Compared to previous planning algorithms, CGF is more efficient and achieves significant improvement on success rate when transferred to grasping with the real Allegro Hand. Our project page is available at https://jianglongye.com/cgf . | ['Xiaolong Wang', 'Yuzhe Qin', 'Binghao Huang', 'Jiashun Wang', 'Jianglong Ye'] | 2022-07-11 | null | null | null | null | ['motion-retargeting'] | ['computer-vision'] | [-1.27961233e-01 3.82135838e-01 -2.64494307e-03 -2.42817387e-01
-5.07890224e-01 -5.46307921e-01 5.54192781e-01 -5.91459990e-01
-1.37096882e-01 8.94106627e-01 4.80450876e-03 -6.43512905e-02
-1.85576439e-01 -8.57779682e-01 -1.16836345e+00 -6.28193259e-01
-3.07517320e-01 1.11057949e+00 -9.00692344e-02 -8.32692608e-02
2.26915315e-01 7.48933971e-01 -1.59940827e+00 2.75467962e-01
8.74938965e-01 3.48281682e-01 9.87813354e-01 8.69516015e-01
2.44901538e-01 6.71346128e-01 -5.73774695e-01 3.03061247e-01
4.97802645e-01 -1.66783273e-01 -7.42476046e-01 8.49482939e-02
-2.80779392e-01 -1.08888459e+00 -4.36965168e-01 8.24171603e-01
3.80582303e-01 3.35628510e-01 9.70735371e-01 -1.74459302e+00
-6.47276640e-01 1.08159757e+00 1.54445264e-02 -7.75916040e-01
4.55839694e-01 6.43096626e-01 5.94028771e-01 -8.23569059e-01
1.15165401e+00 1.74368906e+00 1.16621152e-01 9.99082208e-01
-1.02241313e+00 -6.24899805e-01 -1.44298732e-01 -7.69856945e-02
-8.51572871e-01 2.38242149e-01 4.04604882e-01 -6.12099051e-01
9.57263470e-01 -2.46948659e-01 7.89630711e-01 1.62567139e+00
4.87824619e-01 1.30434048e+00 7.46360242e-01 -2.04767436e-01
3.83245170e-01 -5.00723064e-01 -4.00734216e-01 7.76377261e-01
-5.17487526e-04 6.01186037e-01 -7.55227506e-02 -1.32822022e-01
1.51209462e+00 3.33100736e-01 -3.08134407e-01 -6.96369469e-01
-1.68761051e+00 8.53521705e-01 5.89324594e-01 -3.25812548e-02
-9.12223339e-01 6.18709326e-01 1.52766835e-02 4.17648554e-01
-4.07224566e-01 7.42358685e-01 -3.97008032e-01 -2.29609579e-01
-2.42354438e-01 1.08671010e+00 1.24681675e+00 1.63748407e+00
5.21260500e-01 3.49804386e-02 -4.03528750e-01 4.29789841e-01
3.36643666e-01 7.40352452e-01 2.22393423e-01 -1.78739512e+00
2.88322687e-01 2.16147602e-01 6.92871630e-01 -3.53533030e-01
-2.10931480e-01 2.42291152e-01 -2.93741584e-01 9.97009277e-01
1.77202642e-01 -3.92304540e-01 -1.25405574e+00 1.51736510e+00
3.05003017e-01 -3.34055990e-01 1.49035543e-01 1.04407835e+00
3.41260761e-01 7.16161489e-01 1.78389251e-01 2.90701836e-01
6.63056612e-01 -1.06737626e+00 -5.42963505e-01 1.25098795e-01
2.67917216e-01 -3.51886779e-01 1.15747035e+00 6.84735060e-01
-1.07852066e+00 -5.55552542e-01 -9.13678765e-01 3.95155698e-02
-8.86275060e-03 3.36726338e-01 7.71168530e-01 -2.88867921e-01
-8.02005947e-01 1.30221546e+00 -1.50168610e+00 -2.58890688e-01
3.65789264e-01 3.84315044e-01 -2.99634635e-01 -2.05519468e-01
-7.11517155e-01 1.05588686e+00 7.63550341e-01 -6.58241734e-02
-1.87496305e+00 -3.84472668e-01 -7.79368758e-01 -9.17521194e-02
2.89799005e-01 -8.66043270e-01 1.68972743e+00 -2.14167997e-01
-2.15730906e+00 2.42025018e-01 2.81560659e-01 -1.59556180e-01
8.88070822e-01 -7.03957677e-01 5.21969795e-01 1.76946819e-01
2.42013454e-01 1.29734910e+00 9.86832440e-01 -1.69195998e+00
-3.52751464e-01 4.15619127e-02 1.34953707e-01 1.41074091e-01
4.41422433e-01 -5.36116004e-01 -7.47827217e-02 -3.93952191e-01
3.83667052e-02 -1.34338331e+00 -2.81874627e-01 2.07887560e-01
-5.18670321e-01 -5.95852196e-01 8.84840548e-01 -6.28611505e-01
1.38345897e-01 -1.78461552e+00 1.10587168e+00 -2.48492341e-02
-4.57994230e-02 -1.30554378e-01 -4.05858696e-01 9.21671689e-01
4.09824908e-01 -4.15648162e-01 -3.47273320e-01 -2.03586861e-01
4.96567428e-01 5.86219549e-01 -5.00488698e-01 1.42992571e-01
2.41890371e-01 1.25639343e+00 -1.29163420e+00 -1.55525014e-01
4.15215611e-01 4.56637025e-01 -7.72660494e-01 7.65771866e-01
-9.02747512e-01 9.25435185e-01 -8.68597925e-01 8.71754885e-01
3.63676667e-01 -3.15802693e-02 1.95188373e-01 5.87535761e-02
-5.25598712e-02 -2.25919425e-01 -8.12777996e-01 1.99139476e+00
-4.48595613e-01 1.47324711e-01 2.72048116e-01 -6.44120574e-01
9.84959245e-01 2.03731611e-01 3.99685830e-01 2.36324579e-01
3.04982096e-01 3.92199099e-01 -3.84084396e-02 -9.00481105e-01
3.15578371e-01 1.09091289e-01 -2.24448871e-02 4.02205765e-01
5.27525127e-01 -8.46733928e-01 3.17188315e-02 -9.22455173e-03
1.00828564e+00 1.27551830e+00 -1.32349744e-01 -8.37820247e-02
-2.69366443e-01 2.53069699e-01 1.08967893e-01 8.78790975e-01
2.10613925e-02 5.08630455e-01 2.50506639e-01 -1.25485823e-01
-1.26187408e+00 -1.62866175e+00 3.20509791e-01 7.61370718e-01
2.23394126e-01 7.56409839e-02 -6.38464332e-01 -4.09299195e-01
4.66869354e-01 1.18776143e+00 -5.82666039e-01 2.15584822e-02
-8.96930516e-01 -1.84366982e-02 1.23589925e-01 7.55450428e-01
3.66633534e-01 -1.91135991e+00 -1.27076375e+00 4.03298199e-01
-3.56812254e-02 -6.45008028e-01 -5.36023043e-02 1.55185252e-01
-8.74229789e-01 -1.14893138e+00 -1.11384809e+00 -1.01301587e+00
6.44091487e-01 -1.49675071e-01 6.33121848e-01 -2.85338700e-01
-5.92414975e-01 7.08500087e-01 -5.11785805e-01 -4.91294146e-01
-8.27797413e-01 -1.36048064e-01 7.16793090e-02 -8.79117906e-01
-1.90758139e-01 -6.95159316e-01 -5.68718255e-01 7.73370340e-02
-6.66240692e-01 1.35155872e-01 8.93265903e-01 1.14695537e+00
5.24548173e-01 -2.65412569e-01 4.44360048e-01 -2.11630672e-01
8.16266656e-01 -4.44141775e-01 -5.74031770e-01 9.93297086e-04
-1.49251178e-01 2.15887085e-01 3.56386423e-01 -8.69007409e-01
-1.06212223e+00 2.59761393e-01 1.84535570e-02 -1.03055668e+00
-3.92803311e-01 1.18929423e-01 8.69562626e-02 2.81815588e-01
5.04177928e-01 1.00789778e-01 3.52379143e-01 -3.64199609e-01
6.60103142e-01 3.52982134e-01 6.20432615e-01 -1.05101168e+00
6.36094987e-01 1.71670452e-01 -1.25586882e-01 -5.07343113e-01
1.86858639e-01 3.32181156e-01 -7.78196573e-01 -2.41076678e-01
1.00082314e+00 -5.18799841e-01 -1.33319640e+00 4.47744399e-01
-1.33586943e+00 -1.19327760e+00 -3.76449943e-01 8.23100448e-01
-1.41456318e+00 2.35734358e-02 -9.56667781e-01 -9.30118382e-01
-2.58469284e-01 -1.48551059e+00 1.55586338e+00 1.04153134e-01
-3.33169907e-01 -4.25238490e-01 -1.26267269e-01 -4.01109427e-01
1.50053799e-01 7.21465051e-01 8.13029647e-01 -1.31947443e-01
-9.78607774e-01 -1.36778682e-01 2.40636826e-01 1.73818365e-01
1.10959187e-01 -1.29396608e-02 -2.60446191e-01 -6.55285656e-01
-2.43636131e-01 -7.05595195e-01 5.59962213e-01 5.31575918e-01
1.18699253e+00 -2.92711645e-01 -7.65435815e-01 3.23133379e-01
1.08398175e+00 6.07112825e-01 5.72450459e-01 1.88323021e-01
5.01404643e-01 5.97335100e-01 1.05838215e+00 5.01140356e-01
4.16028500e-02 3.03195059e-01 7.47831047e-01 6.79896712e-01
4.71728146e-02 -7.17428684e-01 4.96544749e-01 3.38921189e-01
-5.53265393e-01 -1.91648871e-01 -6.67944193e-01 4.64331776e-01
-1.92069399e+00 -8.91136944e-01 3.83442968e-01 1.81524801e+00
5.71534216e-01 -1.90784946e-01 6.29386082e-02 -3.34756076e-01
5.34693241e-01 -4.56661403e-01 -8.98386836e-01 -6.82769530e-03
4.95083153e-01 3.20428997e-01 1.88456580e-01 6.64760292e-01
-7.61221766e-01 1.12489772e+00 5.99524593e+00 3.37293804e-01
-9.31930542e-01 -2.09683910e-01 -2.66408831e-01 -8.56261030e-02
-1.83033556e-01 -6.24224544e-04 -4.43240970e-01 2.68118858e-01
2.54528522e-01 -3.68513912e-02 8.35593700e-01 1.25614369e+00
7.09981099e-03 8.53160920e-04 -1.49398983e+00 5.58932304e-01
-3.90822768e-01 -9.26838696e-01 4.72411551e-02 -1.89890474e-01
4.99018312e-01 6.19741566e-02 -2.11701646e-01 8.53789389e-01
9.89561975e-01 -1.23028743e+00 7.77115703e-01 5.75539649e-01
5.34725428e-01 -4.56694514e-01 2.73828179e-01 7.45796680e-01
-6.43298209e-01 -3.20140809e-01 -4.45343018e-01 2.51454115e-02
6.02672040e-01 -2.12869853e-01 -1.20019245e+00 4.20388073e-01
6.17354453e-01 5.06100655e-01 4.37621385e-01 6.68430328e-01
-6.14233494e-01 7.09399283e-02 -1.87137455e-01 -6.20380580e-01
1.71931148e-01 -2.94030011e-01 8.75908732e-01 6.98980391e-01
6.69762492e-01 1.88484237e-01 5.98445356e-01 1.43573177e+00
3.33768010e-01 -5.54604292e-01 -7.62177587e-01 -3.01501304e-01
5.69853902e-01 8.58201802e-01 -4.00711358e-01 -3.96026105e-01
4.10253108e-01 1.30841398e+00 3.66882175e-01 5.50709188e-01
-8.08707118e-01 -6.25807881e-01 3.47338825e-01 -2.78057873e-01
3.58914554e-01 -7.42264569e-01 2.80964613e-01 -1.01380444e+00
-9.53026116e-02 -7.12997615e-01 -3.00284982e-01 -1.21891177e+00
-1.13526750e+00 6.06278479e-01 5.54628611e-01 -1.18709731e+00
-7.41171598e-01 -8.17401052e-01 -3.99322599e-01 8.39114368e-01
-8.32020342e-01 -1.16203487e+00 -6.50789022e-01 3.64326388e-01
8.78593445e-01 -1.81793675e-01 1.00205505e+00 -3.55545640e-01
3.48277658e-01 -1.25547708e-03 -2.10471481e-01 -8.07227492e-02
4.43397015e-01 -1.30053103e+00 2.79431790e-01 3.95424291e-02
-5.06642938e-01 7.44530499e-01 8.38403463e-01 -1.02586293e+00
-1.68721318e+00 -9.50659871e-01 -1.22919105e-01 -4.46819395e-01
4.57850158e-01 -1.55187309e-01 -7.26272583e-01 1.20946348e+00
2.78743356e-01 -5.62548220e-01 -3.36024404e-01 -4.96111900e-01
1.17825679e-01 6.99545503e-01 -1.33732045e+00 8.75059903e-01
1.33755863e+00 8.45269784e-02 -9.09307241e-01 5.67835331e-01
8.43174219e-01 -9.26231384e-01 -1.07129490e+00 6.16701305e-01
9.10248935e-01 -4.01799172e-01 8.76458406e-01 -7.65685022e-01
9.00882781e-01 -2.05168054e-01 -2.67624110e-01 -1.70802832e+00
-4.48729545e-01 -6.43979967e-01 -3.40454966e-01 3.97439182e-01
2.55455114e-02 -6.47574782e-01 6.56621397e-01 2.68125683e-01
-4.11924273e-01 -7.41063535e-01 -5.45031607e-01 -9.76521015e-01
6.08376980e-01 9.64362323e-02 6.07753992e-01 4.43565965e-01
2.21221119e-01 -2.46283859e-01 -2.89791197e-01 2.96705484e-01
8.16434085e-01 5.63003540e-01 1.08311641e+00 -9.38195348e-01
-7.36755550e-01 -2.19376266e-01 5.20779341e-02 -1.34548736e+00
4.61688459e-01 -1.10309935e+00 7.39860535e-01 -1.76239288e+00
-5.38729541e-02 -4.68123049e-01 5.36662877e-01 5.88575006e-01
4.86961715e-02 -6.07915401e-01 5.39819479e-01 3.32724988e-01
9.31569114e-02 9.80214357e-01 2.11474442e+00 -1.87334821e-01
-4.47981030e-01 9.65393521e-03 1.73854753e-01 4.05398756e-01
9.59182560e-01 -1.57458976e-01 -3.54794919e-01 -5.39407194e-01
-5.48371971e-01 7.32518375e-01 6.62951052e-01 -9.18286800e-01
-1.07642949e-01 -5.16154766e-01 5.37640929e-01 -6.11503899e-01
5.13010621e-01 -5.66427112e-01 3.41619611e-01 1.15877807e+00
-4.44270700e-01 -1.04004823e-01 5.21590002e-02 6.86922431e-01
1.62923902e-01 -2.95997977e-01 4.19821620e-01 -5.15327930e-01
-6.28510952e-01 4.28359061e-01 -5.26347637e-01 -5.32886982e-01
1.20681238e+00 1.95765689e-01 -1.15247751e-02 -4.94709015e-01
-1.26029313e+00 5.51255584e-01 5.18982232e-01 7.22534657e-01
8.39344919e-01 -1.27763069e+00 -7.25260317e-01 1.02978058e-01
-2.18148440e-01 3.69887650e-01 1.21373042e-01 1.60204932e-01
-7.85961926e-01 6.85314834e-02 -6.51082635e-01 -7.92354763e-01
-6.38922453e-01 5.87252140e-01 1.20741613e-01 1.63722232e-01
-1.04726100e+00 7.25505829e-01 -2.03729384e-02 -9.62622702e-01
2.67276943e-01 -5.32162428e-01 2.32458666e-01 -7.73316324e-01
6.78458959e-02 5.15695333e-01 -7.17185915e-01 1.08797304e-01
1.20698012e-01 2.30259508e-01 2.46355310e-01 -4.59828526e-01
1.49580264e+00 5.11110663e-01 1.17106391e-02 2.86354274e-01
9.13096607e-01 -4.72685665e-01 -1.97904038e+00 4.93844748e-01
-4.10264462e-01 -3.49902689e-01 -7.18154728e-01 -9.00163949e-01
-6.56560957e-01 7.53819227e-01 3.25468600e-01 -4.03327674e-01
3.82652968e-01 1.47235423e-01 5.88565588e-01 9.65721190e-01
1.16824019e+00 -7.83447742e-01 4.74669427e-01 6.78648531e-01
1.78267622e+00 -1.09990847e+00 -3.49015117e-01 -3.22707534e-01
-6.97991610e-01 1.34222257e+00 8.05318832e-01 -8.89575183e-01
5.03028154e-01 4.97348785e-01 -1.13842271e-01 -2.80771572e-02
-5.95314205e-01 2.64783889e-01 -2.57253826e-01 9.44540083e-01
-1.15824692e-01 4.74561095e-01 -1.12088725e-01 3.40871036e-01
-4.10748988e-01 5.98456025e-01 5.12604892e-01 1.39944279e+00
-4.81124848e-01 -1.10577452e+00 -1.90647393e-01 3.41210485e-01
3.14516366e-01 5.07231772e-01 -6.74302597e-03 1.05652988e+00
-2.05847517e-01 5.64970553e-01 -4.12725843e-02 -3.37503970e-01
3.06414217e-01 1.23254679e-01 1.21327853e+00 -6.80743575e-01
-2.16596782e-01 -1.52363375e-01 -2.62820363e-01 -9.43347037e-01
-1.01197831e-01 -7.57197142e-01 -1.70455837e+00 -2.37904880e-02
-2.76943874e-02 -2.68500187e-02 7.74693429e-01 5.86400688e-01
4.07375246e-01 7.21974969e-01 2.62355357e-01 -1.90386510e+00
-1.30960119e+00 -1.30815327e+00 -4.27906692e-01 4.36824530e-01
2.57960767e-01 -1.27143109e+00 -1.44719347e-01 5.83993178e-03] | [4.747663497924805, 0.5648707747459412] |
76872cc2-56b6-45df-a528-f4ea1ea87ded | finstreder-simple-and-fast-spoken-language | 2206.14589 | null | https://arxiv.org/abs/2206.14589v1 | https://arxiv.org/pdf/2206.14589v1.pdf | Finstreder: Simple and fast Spoken Language Understanding with Finite State Transducers using modern Speech-to-Text models | In Spoken Language Understanding (SLU) the task is to extract important information from audio commands, like the intent of what a user wants the system to do and special entities like locations or numbers. This paper presents a simple method for embedding intents and entities into Finite State Transducers, and, in combination with a pretrained general-purpose Speech-to-Text model, allows building SLU-models without any additional training. Building those models is very fast and only takes a few seconds. It is also completely language independent. With a comparison on different benchmarks it is shown that this method can outperform multiple other, more resource demanding SLU approaches. | ['Wolfgang Reif', 'Alexander Poeppel', 'Daniel Bermuth'] | 2022-06-29 | null | null | null | null | ['slot-filling'] | ['natural-language-processing'] | [ 2.75385499e-01 3.89502317e-01 -6.95472807e-02 -7.07462132e-01
-1.05450046e+00 -7.32850611e-01 7.50944555e-01 4.40816820e-01
-5.76219141e-01 7.64466822e-01 5.88102818e-01 -5.57872057e-01
4.66157436e-01 -5.08883297e-01 -6.39085352e-01 -1.56972840e-01
-3.25374693e-01 6.04095101e-01 4.71528530e-01 -4.85640317e-01
9.58183631e-02 2.43219554e-01 -1.42016351e+00 1.06036440e-01
3.18456411e-01 9.24933374e-01 2.41324529e-01 1.24899137e+00
-5.94206929e-01 1.17044640e+00 -8.90595853e-01 3.94663036e-01
-2.98773825e-01 -4.03635800e-01 -1.18208420e+00 -8.10177624e-02
2.30525434e-01 -5.02219737e-01 -4.12872672e-01 5.90805650e-01
5.58033049e-01 3.55065137e-01 5.75903594e-01 -1.06549442e+00
-1.38430715e-01 1.01825535e+00 4.25012529e-01 3.55083227e-01
1.07095480e+00 -2.84376949e-01 1.21479404e+00 -6.55377209e-01
5.32807887e-01 1.26091707e+00 4.32668358e-01 6.93587899e-01
-1.06727123e+00 -2.13315889e-01 3.23480576e-01 -2.38115534e-01
-1.35814655e+00 -8.61074328e-01 3.14938664e-01 -2.25094154e-01
1.62721384e+00 4.93436337e-01 -1.00317456e-01 8.38965952e-01
-4.84359972e-02 9.95098412e-01 7.48983622e-01 -6.79493666e-01
4.71462131e-01 5.31145811e-01 7.44969070e-01 6.52451396e-01
-3.44620943e-01 -4.18426126e-01 -5.28859735e-01 -3.29379886e-01
5.55965066e-01 -2.16618806e-01 -4.40772533e-01 1.45095035e-01
-1.23888052e+00 6.83442533e-01 -1.84103623e-02 7.22569942e-01
-1.91391364e-01 2.49221742e-01 6.77861929e-01 5.81820369e-01
3.61639023e-01 1.64009795e-01 -6.73710108e-01 -7.66560733e-01
-7.34376371e-01 -2.33670324e-01 1.55035436e+00 1.22131228e+00
7.03194141e-01 4.01517689e-01 2.45445132e-01 6.21157110e-01
2.41807625e-01 5.14523983e-01 6.58716202e-01 -4.83942866e-01
4.49784786e-01 3.08279186e-01 1.64746553e-01 -3.49006236e-01
-4.59409744e-01 1.31349236e-01 -4.04396355e-01 -1.66145504e-01
4.50483486e-02 -7.76412189e-01 -1.07977998e+00 1.68519926e+00
3.00523099e-02 4.16743666e-01 4.83276606e-01 5.78074813e-01
1.09817457e+00 1.42054665e+00 8.86451080e-02 -1.82784617e-01
1.41833889e+00 -5.58124125e-01 -1.08726466e+00 -6.47840381e-01
9.47091818e-01 -5.48104763e-01 8.43891859e-01 2.13530660e-01
-9.46851075e-01 -4.39385056e-01 -8.81648839e-01 -1.41553268e-01
-7.14123964e-01 2.51428664e-01 7.14472771e-01 5.25604308e-01
-1.42041016e+00 2.14214593e-01 -7.82693207e-01 -6.09311044e-01
-3.81893367e-01 5.75910985e-01 -5.32665491e-01 4.12952751e-01
-1.56450713e+00 9.40746605e-01 6.64052606e-01 -1.52148157e-01
-7.79592156e-01 -2.74091512e-01 -1.62204480e+00 3.11057866e-01
5.62511981e-01 -2.22402498e-01 1.81887186e+00 -5.13324916e-01
-1.87483180e+00 4.33167785e-01 -6.07382119e-01 -7.62790382e-01
-2.35295415e-01 -2.82997578e-01 -5.08827031e-01 2.95652356e-02
-2.57127166e-01 5.53874075e-01 6.35188818e-01 -9.10080791e-01
-8.44359159e-01 -1.97510675e-01 1.58561528e-01 2.41831452e-01
-1.51083976e-01 6.27355933e-01 -3.09205472e-01 -3.41963589e-01
2.38883551e-02 -6.08332694e-01 -1.35663211e-01 -7.07201481e-01
-2.55084008e-01 -5.31860590e-01 1.01587129e+00 -9.00109589e-01
1.73912597e+00 -1.98128927e+00 3.04848820e-01 8.29142481e-02
-1.58376381e-01 1.81500539e-01 -9.86536145e-02 9.47094262e-01
-1.17798723e-01 5.23856170e-02 -1.07242204e-01 -5.19346654e-01
2.00348839e-01 6.52299583e-01 -6.35702848e-01 -1.14579510e-03
2.44221061e-01 7.39563584e-01 -7.94257045e-01 -3.01712632e-01
4.38868374e-01 3.17276001e-01 -2.42455795e-01 4.12271112e-01
-2.97653615e-01 4.90741655e-02 -3.38714063e-01 3.57260793e-01
3.43095176e-02 1.48284823e-01 2.13776737e-01 2.62830913e-01
-3.06164980e-01 1.21881497e+00 -1.43583357e+00 1.59138215e+00
-1.03391862e+00 8.91351104e-01 2.95593679e-01 -1.11393785e+00
8.16615880e-01 9.63409960e-01 8.62679407e-02 1.46758156e-02
2.53390104e-01 3.53830516e-01 -2.79077947e-01 -3.22391599e-01
3.52282345e-01 -1.59533054e-01 -5.49347281e-01 8.79110515e-01
6.67686164e-01 -4.24122691e-01 1.30535707e-01 4.59518701e-01
1.17004502e+00 -3.23308289e-01 7.81279802e-01 -1.01014480e-01
7.04172134e-01 -2.76857674e-01 5.15500642e-03 6.40320122e-01
4.86623459e-02 2.78997272e-01 3.54054868e-01 -3.07545185e-01
-5.96507788e-01 -6.58474624e-01 1.49958953e-01 1.48855221e+00
-3.11113268e-01 -7.46271908e-01 -8.70946884e-01 -5.52401364e-01
-3.59343857e-01 1.23103130e+00 -2.81684339e-01 4.19385619e-02
-5.79893768e-01 -6.27706945e-02 7.79098094e-01 7.42116332e-01
-3.81968766e-02 -1.18140566e+00 -3.53187650e-01 4.60242778e-01
-1.43161252e-01 -1.50037098e+00 -6.96690202e-01 7.00252414e-01
-6.87419415e-01 -4.19442743e-01 -5.59717119e-01 -1.15592885e+00
4.91936266e-01 -1.48127198e-01 1.01967430e+00 -3.37852001e-01
-5.07844472e-03 6.40527070e-01 -5.46454370e-01 -8.39825988e-01
-7.67183959e-01 1.13470539e-01 3.03356260e-01 -1.20329395e-01
5.86501360e-01 -3.62241268e-01 3.76271129e-01 -1.13067791e-01
-1.02126002e+00 -3.11372817e-01 2.95515150e-01 6.02703869e-01
1.14853151e-01 -1.25930026e-01 3.99090886e-01 -8.15580964e-01
9.18161511e-01 -2.14551851e-01 -2.90241331e-01 3.05675328e-01
-4.00372632e-02 3.51743668e-01 7.69729674e-01 -2.84929693e-01
-7.81670153e-01 2.94742256e-01 -6.23992026e-01 1.39185950e-01
-8.70996416e-01 6.32511854e-01 -2.29672626e-01 -1.21430522e-02
4.67077136e-01 7.86910236e-01 -8.43855590e-02 -5.86143374e-01
3.24096739e-01 1.34403479e+00 4.04045224e-01 -2.48908818e-01
3.78308803e-01 -2.63645891e-02 -7.16985941e-01 -1.68994045e+00
-6.02280438e-01 -1.09800124e+00 -8.27206969e-01 1.39393404e-01
5.65162122e-01 -7.75018573e-01 -4.12663758e-01 4.15680379e-01
-1.51078844e+00 -6.30292714e-01 -3.69887471e-01 6.84691548e-01
-6.08127534e-01 5.10482192e-01 -8.23952973e-01 -1.14866352e+00
-4.06439990e-01 -9.06123340e-01 1.27821076e+00 1.29498631e-01
-5.96634865e-01 -1.30921245e+00 3.76983285e-02 -8.45566839e-02
2.63449460e-01 -4.29270744e-01 5.40491343e-01 -1.41835797e+00
-1.49703547e-01 -5.33095360e-01 1.36958793e-01 6.30763292e-01
4.00375247e-01 -1.13511622e-01 -9.50778425e-01 -7.83194155e-02
-1.46965487e-02 -3.64934504e-01 4.48210806e-01 1.32206962e-01
5.78985453e-01 -6.83202982e-01 -1.93505034e-01 5.12894504e-02
1.09383321e+00 4.52567637e-01 3.56020182e-01 -3.34015250e-01
4.02392417e-01 4.67336178e-01 4.14347231e-01 4.77964938e-01
5.72321534e-01 6.08378649e-01 2.38135681e-02 4.59488966e-02
2.11566046e-01 -2.16439471e-01 8.58532369e-01 1.20659792e+00
7.55313113e-02 -3.61556143e-01 -9.76936579e-01 7.46234238e-01
-1.72160316e+00 -7.77132869e-01 1.34026185e-01 2.01112080e+00
9.75324690e-01 6.07918166e-02 -7.22261071e-02 1.78928897e-01
2.46249884e-01 3.77431810e-01 -7.52606466e-02 -9.54199851e-01
1.82327017e-01 4.41285223e-01 3.99493158e-01 1.16354668e+00
-1.17224538e+00 1.33370781e+00 7.59465885e+00 6.40019000e-01
-1.22613668e+00 1.22094586e-01 1.63453788e-01 4.00612861e-01
-2.39932328e-01 -1.70818478e-01 -1.03422749e+00 1.14908673e-01
1.67977393e+00 -2.36771449e-01 3.89424860e-01 7.46914208e-01
2.68836528e-01 -2.18474761e-01 -1.39357162e+00 8.76082301e-01
3.20813835e-01 -1.09038854e+00 4.17010263e-02 -4.09934014e-01
1.50515109e-01 1.85120612e-01 -5.98879635e-01 4.45769668e-01
4.93903518e-01 -9.25068498e-01 4.77903724e-01 1.30465269e-01
6.78251565e-01 -6.01885736e-01 9.09505010e-01 8.15184295e-01
-1.39637733e+00 1.10493295e-01 -2.58367240e-01 -3.15986991e-01
5.74184537e-01 -1.07311076e-02 -1.11337137e+00 4.52191472e-01
4.26578611e-01 3.66244614e-01 1.17111877e-02 4.26956028e-01
-4.70653236e-01 1.02223396e+00 -8.84270251e-01 -5.44705272e-01
7.66996801e-01 8.14904943e-02 6.76663458e-01 1.72522092e+00
3.43245476e-01 4.83393341e-01 1.98760852e-01 3.12513918e-01
2.22211912e-01 4.03555691e-01 -1.05943263e+00 -3.63860428e-01
3.21296036e-01 8.78816187e-01 -3.88406992e-01 -7.08480418e-01
-5.57831824e-01 1.24536765e+00 -5.83470203e-02 2.86332816e-01
-3.90363753e-01 -8.17634881e-01 6.98446512e-01 -8.15461576e-02
2.52243489e-01 -5.16210794e-01 1.93925664e-01 -1.20103395e+00
-1.75061733e-01 -8.02509904e-01 3.55983287e-01 -8.54858875e-01
-5.26555419e-01 8.77494693e-01 5.07153198e-02 -8.51462424e-01
-1.05631030e+00 -6.51971877e-01 -8.18700671e-01 9.98329341e-01
-1.14028144e+00 -8.09597909e-01 2.76778370e-01 4.81121272e-01
1.00089765e+00 -1.53973088e-01 1.33780563e+00 1.33453250e-01
-3.79482180e-01 2.73383170e-01 -2.51284450e-01 4.26153272e-01
5.10438442e-01 -1.60193086e+00 6.99353874e-01 8.14927638e-01
6.97468281e-01 8.80757570e-01 1.08194208e+00 -2.57856458e-01
-1.25706184e+00 -6.48528159e-01 1.52802074e+00 -5.75819492e-01
9.33969498e-01 -6.62148416e-01 -8.59678805e-01 1.20482862e+00
5.57821453e-01 -2.79748708e-01 9.65563416e-01 1.14389956e-01
-2.46143695e-02 1.78479463e-01 -6.68636858e-01 2.86973476e-01
4.65171874e-01 -8.96286011e-01 -1.02942169e+00 4.14738029e-01
1.08558130e+00 -4.78426695e-01 -7.80304313e-01 -4.91386652e-02
1.40636861e-01 -4.84671563e-01 7.50030994e-01 -1.09325325e+00
-2.54426032e-01 -1.55926660e-01 -1.60416290e-01 -1.48401201e+00
1.78450927e-01 -1.02552569e+00 -3.26443076e-01 1.07223022e+00
6.56813085e-01 -9.28815484e-01 3.23482543e-01 6.89367950e-01
-2.69440532e-01 -3.22185814e-01 -9.53355491e-01 -5.62937856e-01
-4.00600761e-01 -8.84121418e-01 2.89646447e-01 5.48593879e-01
6.98284149e-01 9.16946292e-01 -4.46723610e-01 4.67898458e-01
1.02655150e-01 -7.64439479e-02 6.92612648e-01 -1.11346877e+00
-1.03929184e-01 -1.54597178e-01 -6.26624227e-01 -1.65906250e+00
3.02512258e-01 -4.70618397e-01 4.42663580e-01 -1.69084156e+00
-5.07151961e-01 -8.02465677e-02 -1.95214078e-01 8.91539752e-01
9.32354182e-02 -3.01032990e-01 -6.18639365e-02 -3.08346063e-01
-6.78074718e-01 3.17279309e-01 4.62677032e-01 -1.04668356e-01
-4.29674149e-01 4.49841678e-01 -3.73074263e-01 7.84387827e-01
8.26737165e-01 -2.68800884e-01 -4.64535147e-01 -1.19516253e-01
-8.78909826e-02 3.84711981e-01 -1.20398782e-01 -1.02789867e+00
4.89551812e-01 1.06388107e-01 -5.62352240e-01 -6.28153622e-01
4.96894151e-01 -1.00794113e+00 -5.30793548e-01 3.47473204e-01
-4.06047851e-01 -1.82003021e-01 4.94746774e-01 2.86624402e-01
-8.68892074e-01 -7.75035739e-01 1.72542885e-01 -1.16960697e-01
-1.26521122e+00 -2.82394774e-02 -1.08624542e+00 -1.13127202e-01
9.18660164e-01 -4.94739749e-02 4.31222349e-01 -1.17572320e+00
-1.05217946e+00 2.61421919e-01 -2.39925519e-01 4.75272357e-01
5.83860099e-01 -8.65001202e-01 -4.33293283e-01 3.00705731e-01
4.81045991e-02 -5.85497878e-02 1.22798570e-01 5.76628387e-01
-5.14451504e-01 1.30644190e+00 2.43301600e-01 -3.47433180e-01
-1.60453784e+00 1.82208583e-01 2.41776451e-01 -1.42945200e-01
-3.39531988e-01 1.03435397e+00 -1.21459395e-01 -8.75730813e-01
6.78231120e-01 -9.66631234e-01 -4.63710248e-01 1.85571983e-01
9.96233046e-01 -7.66723752e-02 9.90716368e-02 -7.64687121e-01
-4.78744030e-01 2.18639866e-01 1.22766785e-01 -4.17435110e-01
1.18356538e+00 -2.87746459e-01 -4.90229875e-02 1.12297583e+00
1.11788607e+00 1.01829022e-01 -6.94087148e-01 -4.78149295e-01
1.13439120e-01 4.28423762e-01 1.31659299e-01 -4.94046986e-01
-4.55678970e-01 1.24877596e+00 4.72975038e-02 7.27940559e-01
8.31923366e-01 2.10813329e-01 9.67827797e-01 1.04270530e+00
6.44699812e-01 -1.11414015e+00 -3.90894771e-01 1.03066969e+00
8.32020044e-01 -1.29393613e+00 -5.21633983e-01 -3.61972123e-01
-7.47745752e-01 1.25719142e+00 2.22225368e-01 2.55433023e-01
8.20797861e-01 7.50521719e-01 4.49371815e-01 5.56793287e-02
-1.06900847e+00 -4.39175785e-01 2.35249996e-01 5.32435775e-01
6.33483827e-01 7.40910098e-02 -5.96816540e-02 5.79589963e-01
-2.17398316e-01 -5.02352007e-02 5.99824607e-01 1.11717236e+00
-8.80782366e-01 -1.16371465e+00 -2.49641418e-01 3.70375454e-01
-4.96923089e-01 -2.69270957e-01 -6.03057206e-01 6.91788614e-01
-4.32870120e-01 9.52547371e-01 1.15552217e-01 -4.03585941e-01
3.64354551e-01 6.43944502e-01 2.64157597e-02 -1.27601337e+00
-5.71623564e-01 1.04351647e-01 5.32279789e-01 -5.88868737e-01
-1.28507495e-01 -5.97243547e-01 -1.76394928e+00 1.01700440e-01
-3.54022950e-01 5.05298197e-01 7.34406769e-01 1.05932248e+00
1.90263182e-01 5.13750732e-01 5.41623175e-01 -7.39402235e-01
-5.57655275e-01 -1.27639985e+00 -5.65649986e-01 -1.50860816e-01
5.24414301e-01 4.36661057e-02 -5.31545758e-01 1.40060723e-01] | [14.051546096801758, 6.940502643585205] |
b86e611c-bcaa-40db-a65f-dc933b179a3e | using-neural-network-for-identifying | 1806.07713 | null | http://arxiv.org/abs/1806.07713v1 | http://arxiv.org/pdf/1806.07713v1.pdf | Using Neural Network for Identifying Clickbaits in Online News Media | Online news media sometimes use misleading headlines to lure users to open
the news article. These catchy headlines that attract users but disappointed
them at the end, are called Clickbaits. Because of the importance of automatic
clickbait detection in online medias, lots of machine learning methods were
proposed and employed to find the clickbait headlines. In this research, a
model using deep learning methods is proposed to find the clickbaits in
Clickbait Challenge 2017's dataset. The proposed model gained the first rank in
the Clickbait Challenge 2017 in terms of Mean Squared Error. Also, data
analytics and visualization techniques are employed to explore and discover the
provided dataset to get more insight from the data. | ['Hui Jiang', 'Amin Omidvar', 'Aijun An'] | 2018-06-20 | null | null | null | null | ['clickbait-detection'] | ['natural-language-processing'] | [-6.38222456e-01 -3.74392033e-01 -4.64619517e-01 -3.19732964e-01
-8.61340344e-01 -3.04291427e-01 7.30885029e-01 5.95411479e-01
-2.73890406e-01 8.09610784e-01 3.09242457e-01 -3.86904120e-01
-2.36320034e-01 -5.96763611e-01 -7.39321113e-01 -1.52230188e-01
-1.65243432e-01 1.03354767e-01 4.27597344e-01 -1.57179788e-01
8.50479007e-01 -1.72682613e-01 -1.36610293e+00 5.00069976e-01
6.55037522e-01 1.55592990e+00 1.25476360e-01 4.06931758e-01
-5.13251662e-01 1.02415836e+00 -7.27089942e-01 -2.91889310e-01
1.19132958e-01 -1.27073601e-01 -4.77779180e-01 -2.66846746e-01
8.31432521e-01 -5.66829205e-01 -6.48661315e-01 9.94438589e-01
1.72271773e-01 -1.57645747e-01 1.35236531e-01 -1.21150815e+00
-8.68981421e-01 4.67778653e-01 -7.72265732e-01 7.50796735e-01
2.84757137e-01 -1.65403515e-01 1.49251091e+00 -9.14546251e-01
4.45945650e-01 8.78540158e-01 3.76026779e-01 -9.12544206e-02
-8.59694004e-01 -7.28216827e-01 6.45503476e-02 4.64143485e-01
-1.06640685e+00 1.17991373e-01 7.76321590e-01 -9.21912611e-01
3.27541530e-01 7.93905735e-01 8.08997035e-01 1.19421124e+00
2.35880047e-01 1.50404096e+00 1.14171159e+00 1.15532786e-01
4.18101586e-02 3.30906212e-01 7.25190818e-01 3.19546640e-01
3.59696984e-01 -7.23277926e-02 -6.41497135e-01 -1.13845415e-01
8.07730332e-02 5.14411926e-01 4.87454869e-02 1.86170831e-01
-7.81076074e-01 1.33611262e+00 9.43470538e-01 2.73358345e-01
-3.70975047e-01 7.48746842e-02 5.86549938e-01 2.20601544e-01
9.07766938e-01 6.89224720e-01 -4.02596623e-01 -3.01301062e-01
-1.09459186e+00 6.48937047e-01 7.77680397e-01 5.76514602e-01
7.27841377e-01 -3.64175200e-01 -4.08326425e-02 7.33517706e-01
2.73676187e-01 1.68520153e-01 6.82190239e-01 -1.60606593e-01
5.82922757e-01 8.41592133e-01 2.89953530e-01 -1.60397685e+00
-4.15349305e-01 -9.83758450e-01 -4.77054507e-01 -1.75324365e-01
1.80246532e-01 -4.60179895e-03 -7.31042624e-01 7.20221102e-01
3.80872376e-02 -3.03591520e-01 -7.87886560e-01 1.02099264e+00
8.70743215e-01 8.03170562e-01 -1.21227056e-01 3.40697616e-02
1.26111877e+00 -9.33729351e-01 -1.02182293e+00 -2.64958084e-01
5.99896967e-01 -8.79565358e-01 1.22755659e+00 3.72460335e-01
-7.97738075e-01 -3.34606320e-01 -1.16097808e+00 -2.29630500e-01
-9.10884559e-01 1.11610308e-01 5.53794682e-01 4.24684444e-03
-1.61550805e-01 5.34028530e-01 -4.83920783e-01 -5.73824681e-02
9.09458339e-01 -3.35346669e-01 4.65250283e-01 1.90588474e-01
-1.25237930e+00 6.27873003e-01 4.28448021e-01 -7.99535438e-02
-8.10593367e-01 -9.43386674e-01 -1.32700354e-02 1.73707187e-01
6.55743182e-01 -1.10171065e-02 1.34688139e+00 -7.92507172e-01
-7.44243860e-01 5.14574826e-01 2.93369591e-01 -7.23946214e-01
9.79108989e-01 -8.49617660e-01 -4.77078319e-01 -1.96767986e-01
2.33525947e-01 -1.06537975e-01 8.31792355e-01 -4.91403937e-01
-9.99168575e-01 -3.00714314e-01 -1.00270815e-01 -4.60920602e-01
-4.12494272e-01 1.74863655e-02 -1.14352174e-01 -5.00854194e-01
1.05023846e-01 -4.26453769e-01 2.49441400e-01 -3.00225645e-01
-9.43235040e-01 -7.10609078e-01 1.03693449e+00 -1.00548697e+00
1.94189703e+00 -2.06663394e+00 -5.82339525e-01 1.07607894e-01
1.01288712e+00 2.86611617e-01 4.45107430e-01 7.36800015e-01
2.33663037e-01 5.00816524e-01 8.01472247e-01 3.78870040e-01
1.47707477e-01 -4.67063397e-01 -6.19212747e-01 3.52824867e-01
-3.38551700e-01 8.57854605e-01 -6.73422694e-01 -1.31047100e-01
-1.83092430e-01 -1.03798702e-01 -2.69879103e-01 1.33548304e-01
-6.06757998e-01 5.92337884e-02 -1.06755304e+00 4.03747499e-01
7.31434286e-01 -8.94668043e-01 -6.16732955e-01 -3.09412420e-01
-6.97065532e-01 7.89003372e-01 -3.91731828e-01 7.34443247e-01
-1.06857039e-01 1.24674487e+00 -3.12099159e-01 -5.86312473e-01
1.00957382e+00 -1.35197878e-01 3.39344084e-01 -1.23445737e+00
2.63498962e-01 4.38181341e-01 -7.52808973e-02 -8.43469143e-01
5.54501176e-01 3.70306671e-01 1.35244399e-01 4.94520962e-01
-6.79767430e-01 6.77709162e-01 2.13454396e-01 4.42288160e-01
8.75742316e-01 -3.77238154e-01 1.64252460e-01 1.54503258e-02
1.67283654e-01 2.75742859e-01 6.36197180e-02 8.08870494e-01
-1.66835055e-01 2.18200013e-01 5.62395275e-01 -1.04518557e+00
-1.19324362e+00 -4.77971256e-01 -4.58270460e-02 1.49177337e+00
-1.27112895e-01 -5.35990834e-01 -4.25860047e-01 -7.61171937e-01
2.03216359e-01 6.38162613e-01 -6.77215159e-01 1.62638411e-01
-2.25370362e-01 -3.99288625e-01 1.76705763e-01 2.23475218e-01
1.03232038e+00 -5.88416398e-01 -5.71664214e-01 2.46826321e-01
-2.02913642e-01 -6.20678067e-01 -4.93025810e-01 -1.99446827e-01
-4.65848863e-01 -1.31921268e+00 -1.01924789e+00 -5.29861629e-01
4.84342009e-01 2.12272227e-01 7.02760637e-01 -1.17657962e-03
-4.14410561e-01 -3.61830831e-01 -4.40110862e-01 -5.80414593e-01
-6.79177493e-02 4.67037976e-01 -3.92563522e-01 1.23608641e-01
6.53774798e-01 1.78801769e-03 -9.18975949e-01 4.15568769e-01
-8.51295412e-01 -9.74095836e-02 3.76934886e-01 6.50075316e-01
-3.50359902e-02 -8.72230381e-02 7.12191284e-01 -7.53189623e-01
1.14841485e+00 -8.66364658e-01 -9.73159611e-01 -3.99473533e-02
-6.76170170e-01 -1.52335420e-01 2.39457875e-01 -3.62743944e-01
-5.64002812e-01 -7.37866163e-01 6.42118528e-02 4.39926088e-02
2.04585597e-01 1.15749002e+00 5.80331862e-01 1.77296937e-01
7.72222340e-01 4.21082675e-01 -2.21132636e-01 -1.13295710e+00
-5.93715869e-02 8.75182390e-01 -3.37142587e-01 4.40760523e-01
5.93149543e-01 1.40544921e-01 -5.00655830e-01 -6.44704044e-01
-1.62080932e+00 -6.34810805e-01 1.14785880e-01 -3.87014091e-01
6.82341695e-01 -6.43653870e-01 -1.15585947e+00 4.41750139e-01
-1.05033839e+00 5.33066511e-01 3.60250413e-01 4.84613568e-01
7.93822780e-02 -4.05227430e-02 -4.75033730e-01 -7.11915970e-01
-1.00808971e-01 -5.80685258e-01 3.63115996e-01 2.77002156e-01
-2.50694364e-01 -8.15371871e-01 1.29230306e-01 5.81351101e-01
1.09738219e+00 2.69840136e-02 6.87669873e-01 -1.60867143e+00
-8.81242096e-01 -9.65822995e-01 -6.54658258e-01 4.00990188e-01
1.17000669e-01 -1.63257673e-01 -6.88018262e-01 -4.86567914e-02
-1.56990409e-01 -3.91492784e-01 8.21833968e-01 4.42589462e-01
1.42661190e+00 -1.11318743e+00 -4.40670222e-01 1.08843312e-01
1.11085856e+00 1.23652071e-01 3.54562283e-01 1.06324434e+00
5.04847348e-01 4.20897365e-01 6.17480814e-01 6.58136725e-01
1.64978594e-01 4.78112757e-01 7.20119596e-01 1.19656898e-01
2.74902970e-01 -8.00138414e-01 -9.11351144e-02 7.81382978e-01
3.74191910e-01 5.59783243e-02 -7.83695042e-01 3.30177814e-01
-1.70312715e+00 -7.87969768e-01 -3.24366033e-01 1.95420873e+00
5.19300938e-01 4.82160091e-01 3.28192323e-01 4.26917709e-02
7.05505371e-01 5.52851081e-01 -6.17977440e-01 -8.57395753e-02
2.03185409e-01 -4.40499902e-01 7.27857232e-01 1.33930817e-01
-1.14959943e+00 4.95408952e-01 5.70573044e+00 1.00874782e+00
-1.35742033e+00 5.64772375e-02 8.60039592e-01 -1.08157039e-01
1.23364637e-02 -3.22047412e-01 -9.85958695e-01 1.03178906e+00
5.81580937e-01 -2.85621047e-01 8.46153200e-02 1.26607966e+00
7.23554254e-01 -2.25087032e-01 -7.00010836e-01 9.24418092e-01
-4.89040278e-02 -1.95022047e+00 -2.08495885e-01 3.68477583e-01
5.17069519e-01 2.95142919e-01 3.83805782e-01 3.21757436e-01
6.05290979e-02 -6.18193269e-01 3.87209743e-01 6.52923226e-01
-6.82027787e-02 -3.65144908e-01 7.48171687e-01 4.69957769e-01
-4.32396114e-01 -3.72755885e-01 -2.12957919e-01 -3.12969923e-01
6.85163587e-03 8.61233830e-01 -1.22381532e+00 -2.19902560e-01
6.65429890e-01 9.41071153e-01 -1.06837595e+00 1.76644528e+00
1.00324847e-01 9.34320807e-01 -7.58176968e-02 -1.00849569e+00
7.27476358e-01 7.53023475e-02 6.01699650e-01 7.78772116e-01
3.03144544e-01 -7.93591797e-01 7.58511573e-02 9.18720782e-01
-1.82688951e-01 2.71909118e-01 -3.19074571e-01 -7.84513652e-01
2.06906751e-01 1.10596740e+00 -3.81558895e-01 -3.74162525e-01
-3.24174255e-01 4.65014368e-01 1.10308826e-01 3.62084508e-01
-9.53149974e-01 -6.86656833e-01 1.28467023e-01 8.77934456e-01
3.07662547e-01 -3.19230855e-01 -1.34181827e-01 -9.39693034e-01
8.31021070e-02 -4.96007055e-01 2.12377355e-01 -8.80794227e-01
-1.54833519e+00 4.35203969e-01 -1.95612207e-01 -1.09863877e+00
3.00545841e-01 -4.37601715e-01 -6.45537734e-01 4.82822418e-01
-1.42560160e+00 -7.08428502e-01 -5.05304694e-01 1.94697499e-01
8.19627941e-01 -2.34597445e-01 -6.31848425e-02 6.84990227e-01
-5.01347244e-01 3.85743320e-01 6.13374233e-01 3.37167025e-01
3.89176458e-01 -9.94762540e-01 3.78123283e-01 1.59724310e-01
-2.89215557e-02 7.06513464e-01 1.04145873e+00 -8.22434723e-01
-1.12835729e+00 -7.85508037e-01 1.08261490e+00 -5.44143245e-02
1.54753888e+00 -1.54380649e-01 -9.06526089e-01 6.76484108e-01
2.70632267e-01 -3.31004143e-01 6.09261990e-01 3.96421254e-01
-2.49936074e-01 -4.27813053e-01 -5.95227420e-01 3.46563995e-01
7.89808258e-02 -2.22826988e-01 -6.89922333e-01 8.80543709e-01
7.29507983e-01 1.14717461e-01 -3.13197017e-01 -3.52093637e-01
6.25890553e-01 -7.64020205e-01 5.77954054e-01 -1.05370224e+00
7.46111274e-01 2.67705441e-01 -3.35641541e-02 -1.36467123e+00
-1.55380040e-01 -4.52996224e-01 3.40787843e-02 1.14878917e+00
5.72788656e-01 -7.10089207e-01 8.66682410e-01 5.15653253e-01
3.41829471e-02 -5.67600429e-01 -8.72970700e-01 -5.58699012e-01
-2.92747140e-01 -1.16477087e-01 4.38494861e-01 6.79016113e-01
2.48190295e-02 5.55953562e-01 -7.88506150e-01 -4.93360668e-01
4.42792624e-01 7.05753639e-02 7.95193732e-01 -1.21921563e+00
7.06031546e-02 -6.24160945e-01 -2.56185740e-01 -1.61163521e+00
-7.58395016e-01 -7.30415046e-01 -3.02136332e-01 -1.35977650e+00
1.31157547e-01 -1.93788216e-01 -4.58057821e-01 -2.70283930e-02
3.12500931e-02 1.62854195e-02 6.76706880e-02 5.54305434e-01
-1.05806911e+00 3.94480020e-01 1.21840298e+00 -4.08513367e-01
-2.91656554e-02 5.49506187e-01 -8.55475664e-01 7.56666183e-01
7.56157696e-01 -5.40732384e-01 2.70549595e-01 -1.97801456e-01
1.13247609e+00 -3.38607490e-01 4.80232596e-01 -7.04108536e-01
2.61671156e-01 -4.58547398e-02 2.53795892e-01 -1.19204903e+00
1.00063927e-01 -6.47082508e-01 -2.94157624e-01 4.86789554e-01
-1.26408505e+00 5.37866466e-02 -2.48670578e-01 7.83117890e-01
-5.67196727e-01 -1.81382641e-01 4.91894603e-01 -1.61753237e-01
-3.68199855e-01 2.67290682e-01 -5.48971236e-01 1.62593395e-01
8.02961648e-01 1.70085385e-01 -7.81524479e-01 -6.51086271e-01
-6.89427555e-01 2.92872667e-01 -5.63238382e-01 5.74615777e-01
5.60055196e-01 -1.30378520e+00 -7.60551274e-01 -1.67318974e-02
2.56095231e-01 -5.74079216e-01 2.28098795e-01 1.00904834e+00
-6.72423482e-01 8.31921458e-01 -3.80901620e-02 -3.67731929e-01
-1.03129816e+00 4.16986376e-01 -2.72399504e-02 -1.83564484e-01
-5.17425954e-01 8.72852206e-01 5.58736213e-02 2.41262227e-01
4.69795287e-01 -1.97853193e-01 -2.29573563e-01 6.17930830e-01
9.19506788e-01 8.82195532e-01 -7.00064050e-03 -6.81073368e-02
-9.64757875e-02 -1.63159788e-01 -7.96150804e-01 3.12175572e-01
1.39734232e+00 -3.85548860e-01 -5.39732464e-02 6.13282681e-01
1.82624173e+00 -6.76107258e-02 -9.63994801e-01 -3.61670583e-01
5.95897317e-01 -1.09630561e+00 2.94859856e-01 -1.10470164e+00
-9.93786395e-01 1.04933214e+00 4.53842580e-01 1.26177406e+00
3.10034931e-01 1.54737219e-01 1.25484645e+00 4.72073376e-01
-2.51566648e-01 -1.42133975e+00 4.57760513e-01 4.06157196e-01
1.16279781e+00 -1.61833286e+00 2.51414210e-01 2.60976464e-01
-4.85657990e-01 1.22607875e+00 4.10491467e-01 -2.80392200e-01
8.52719665e-01 -3.23986709e-01 -1.12862393e-01 -5.46809375e-01
-6.27611399e-01 2.52197951e-01 3.73538077e-01 -7.36491904e-02
3.09942752e-01 -3.35180163e-02 -8.04955304e-01 6.31088436e-01
-1.15521856e-01 1.04016729e-01 3.63867790e-01 6.37592494e-01
-1.16698909e+00 -3.21113557e-01 -9.92627814e-02 1.05798686e+00
-9.34985816e-01 -1.40146106e-01 -3.74206871e-01 7.23758280e-01
-5.32001615e-01 6.87134326e-01 1.53608341e-03 -3.86604398e-01
-6.45891428e-02 -2.56268531e-02 -5.92818379e-01 -6.33265683e-03
-3.91482502e-01 3.19887370e-01 1.59085914e-01 -4.61130619e-01
1.48228496e-01 -1.73333943e-01 -4.46557343e-01 -2.38673061e-01
-6.12568080e-01 5.67600846e-01 1.30681920e+00 8.85884464e-01
4.86040175e-01 5.37880182e-01 8.99597585e-01 -1.41998038e-01
-8.57953370e-01 -1.24734056e+00 -5.75312436e-01 3.64590973e-01
6.64157569e-01 -5.28500438e-01 -7.13463962e-01 -5.88198543e-01] | [7.751511096954346, 9.784372329711914] |
937e318d-b180-4653-a417-a319de82fb75 | a-baseline-framework-for-part-level-action | 2110.03368 | null | https://arxiv.org/abs/2110.03368v2 | https://arxiv.org/pdf/2110.03368v2.pdf | A Baseline Framework for Part-level Action Parsing and Action Recognition | This technical report introduces our 2nd place solution to Kinetics-TPS Track on Part-level Action Parsing in ICCV DeeperAction Workshop 2021. Our entry is mainly based on YOLOF for instance and part detection, HRNet for human pose estimation, and CSN for video-level action recognition and frame-level part state parsing. We describe technical details for the Kinetics-TPS dataset, together with some experimental results. In the competition, we achieved 61.37% mAP on the test set of Kinetics-TPS. | ['Tao Mei', 'Wu Liu', 'Kun Liu', 'Xinchen Liu', 'Xiaodong Chen'] | 2021-10-07 | null | null | null | null | ['action-parsing'] | ['natural-language-processing'] | [ 1.29449368e-01 5.01026034e-01 -3.15005660e-01 -3.35261226e-01
-1.23837554e+00 -5.44795573e-01 1.44560084e-01 -5.14227152e-01
-5.91864586e-01 5.18362045e-01 5.31619966e-01 2.86917865e-01
7.57282913e-01 1.34054702e-02 -9.59290862e-01 -4.23977435e-01
-2.46787235e-01 6.09620869e-01 9.17989433e-01 -1.94182813e-01
-2.06068590e-01 4.33764875e-01 -1.07997906e+00 8.96237969e-01
-4.47916538e-01 1.31790507e+00 -4.62047532e-02 1.39197958e+00
1.82182059e-01 1.02010250e+00 -7.96528339e-01 -7.73493469e-01
5.40679038e-01 -4.86153930e-01 -1.03249490e+00 3.95292550e-01
1.15701425e+00 -4.69690055e-01 -8.55627775e-01 6.18335903e-01
7.58384645e-01 -1.47370556e-02 1.12143137e-01 -1.43237078e+00
5.33565223e-01 6.02578878e-01 -3.47264409e-01 7.43721485e-01
7.77003646e-01 1.73488438e-01 5.76781631e-01 -7.86605120e-01
1.18822384e+00 1.42092752e+00 8.18714559e-01 9.49300766e-01
-4.92400378e-01 -3.48081976e-01 3.19779366e-01 3.35906565e-01
-1.08967280e+00 -7.75616884e-01 6.60524219e-02 -3.62483561e-02
1.64588916e+00 9.88951996e-02 1.03648663e+00 1.34390342e+00
4.65874135e-01 1.59182227e+00 7.58920193e-01 -5.68087921e-02
1.51403919e-01 -5.13992906e-01 6.01493828e-02 9.61029053e-01
-1.25810042e-01 -2.31879428e-02 -9.12976682e-01 3.14544469e-01
8.75944316e-01 -5.06294608e-01 -2.91178022e-02 -4.81588572e-01
-1.07814121e+00 2.34240770e-01 2.32960731e-01 -7.69350817e-03
-2.46367469e-01 6.97575808e-01 1.07478309e+00 -1.25601277e-01
2.01910138e-01 -1.17398523e-01 -9.98760819e-01 -7.37966120e-01
-8.05630684e-01 1.87061384e-01 5.96854687e-01 1.44704914e+00
-3.62742037e-01 3.13254371e-02 -5.36519110e-01 3.59071702e-01
3.20157051e-01 4.46289212e-01 1.86030731e-01 -1.41940808e+00
9.27189052e-01 5.42212874e-02 1.25200730e-02 -2.12030318e-02
-6.22847915e-01 1.43310362e-02 -2.08189130e-01 6.19320013e-02
6.12489581e-01 -2.74887472e-01 -1.43508852e+00 1.21507728e+00
3.80655795e-01 2.08759591e-01 -3.75324227e-02 8.90383005e-01
1.29059422e+00 5.09961307e-01 6.07456744e-01 -9.79195908e-02
1.69031072e+00 -1.49005556e+00 -7.79851675e-01 -5.06969869e-01
8.00494492e-01 -4.98199522e-01 2.70795971e-01 4.68454778e-01
-1.51029098e+00 -7.58489966e-01 -9.23224747e-01 -1.31564915e-01
4.78889095e-03 6.14604294e-01 7.93763518e-01 9.11484361e-01
-1.13825047e+00 4.72324342e-01 -1.41994822e+00 -5.89039087e-01
6.77132308e-01 1.41535029e-01 -8.06606412e-01 -1.27681807e-01
-9.02623832e-01 8.91396940e-01 5.45470059e-01 2.91982114e-01
-9.70322132e-01 -2.26634398e-01 -1.06383145e+00 -1.63728058e-01
4.65018362e-01 -3.34042281e-01 1.76882887e+00 -4.73129988e-01
-1.57653928e+00 1.07505679e+00 2.21075751e-02 -8.83994401e-01
9.47579026e-01 -6.53846383e-01 -4.13074881e-01 5.51363111e-01
-1.42293964e-02 1.16015649e+00 4.63513881e-01 -3.98388505e-01
-9.11601543e-01 -3.21254700e-01 -2.37483960e-02 6.12853050e-01
6.13981307e-01 3.99808466e-01 -1.23830664e+00 -3.17305118e-01
3.44960719e-01 -1.02639973e+00 -3.98607939e-01 -3.71893100e-03
-3.49029124e-01 -1.93862736e-01 6.97540879e-01 -9.29087102e-01
6.77291453e-01 -2.06910372e+00 6.84134141e-02 -5.08007884e-01
-6.42663846e-03 2.99940884e-01 -1.58966094e-01 1.73512071e-01
-2.33237401e-01 -3.35811943e-01 2.26334453e-01 -8.30009878e-01
-1.65813655e-01 2.05487847e-01 9.72851738e-02 7.94692159e-01
-4.26176563e-02 1.30141854e+00 -4.14488822e-01 -8.53479028e-01
4.91944939e-01 1.89763144e-01 -3.17389727e-01 -1.08871564e-01
-7.55770504e-02 2.79412925e-01 -4.47408646e-01 1.02361107e+00
6.71497941e-01 2.51895487e-01 9.83005390e-02 -4.15815473e-01
1.37068368e-02 2.41728365e-01 -1.17372262e+00 2.17285991e+00
2.14270756e-01 8.12932312e-01 2.18531221e-01 -6.00226283e-01
6.07474186e-02 5.52345157e-01 7.88004220e-01 -7.22941041e-01
2.86396593e-01 -2.11363360e-01 -1.63513333e-01 -1.27644077e-01
4.34742749e-01 1.44868389e-01 -3.21182221e-01 -3.21623266e-01
4.83832955e-01 7.09472084e-03 5.28698564e-01 3.30953449e-01
1.57799518e+00 8.83556843e-01 3.17597210e-01 9.38162953e-02
3.56379032e-01 2.85281450e-01 5.84729612e-01 5.65566838e-01
-1.15986764e+00 1.03005624e+00 5.43958604e-01 -5.96973717e-01
-6.53051913e-01 -1.07986891e+00 1.23329228e-02 9.07667696e-01
8.59192461e-02 -8.34278703e-01 -1.06347156e+00 -1.27664983e+00
-5.71496069e-01 3.06522071e-01 -6.87672675e-01 1.95540860e-01
-9.85878944e-01 -1.99590698e-01 9.54487324e-01 1.37452674e+00
8.80312622e-01 -1.49201751e+00 -8.55318844e-01 1.43383294e-01
-3.28820050e-01 -1.99931324e+00 -4.34571654e-01 4.11992282e-01
-1.05426919e+00 -1.36695731e+00 -8.41228306e-01 -6.09645605e-01
1.66407436e-01 -4.00393568e-02 1.45990562e+00 -4.05268103e-01
-6.03763819e-01 5.33668160e-01 -5.30757487e-01 -2.23575562e-01
-1.19991124e-01 -1.81036726e-01 -9.16277841e-02 -6.86530352e-01
1.21650100e-01 2.74063677e-01 -6.52573228e-01 5.89344561e-01
-1.78769335e-01 -1.99675977e-01 8.81405711e-01 9.58617702e-02
7.97839105e-01 -3.48342746e-01 -4.25483227e-01 -6.56954408e-01
-4.94939506e-01 5.44831574e-01 -4.09801483e-01 3.04368347e-01
3.55266780e-01 -3.30881685e-01 -3.04015696e-01 8.80139396e-02
-1.20678616e+00 1.08400023e+00 -4.31297004e-01 -5.84382832e-01
-3.55674058e-01 -3.93206835e-01 -4.18063432e-01 -1.52818605e-01
4.98111039e-01 2.60432869e-01 -4.43094879e-01 -3.25834900e-01
3.65718007e-01 8.46160427e-02 9.45648909e-01 -1.76216424e-01
2.16903925e-01 4.29908216e-01 -3.71507276e-03 -7.32548058e-01
-8.43310058e-01 -8.22280526e-01 -9.96219218e-01 -3.38666826e-01
1.72922254e+00 -1.49428892e+00 -6.11984372e-01 7.17037737e-01
-1.25204480e+00 -5.57578802e-01 -4.54826504e-01 6.76797807e-01
-1.02276194e+00 3.19575965e-01 -1.02206230e+00 -5.08380234e-01
-1.83881238e-01 -1.14559340e+00 1.70960045e+00 1.38669983e-01
-4.27975617e-02 -4.74248141e-01 1.97586983e-01 8.37575674e-01
-2.35918343e-01 1.74193010e-01 -3.87861520e-01 -5.49643815e-01
-5.43981791e-01 -7.34415233e-01 3.81741449e-02 2.66354084e-01
-6.47322297e-01 -1.36136770e-01 -7.86833405e-01 -1.40587136e-01
-3.58791173e-01 -6.59429491e-01 8.03881705e-01 9.84962106e-01
1.13254678e+00 4.52819526e-01 -4.90698993e-01 5.63347340e-01
8.07337463e-01 2.37155825e-01 1.20476651e+00 1.56763732e-01
6.79585516e-01 1.69781014e-01 1.11383522e+00 3.59131634e-01
1.14738017e-01 1.06667280e+00 4.18087393e-01 1.19183667e-01
-5.90977550e-01 -2.48057574e-01 7.28749871e-01 1.18369736e-01
-6.69922531e-01 -3.73108625e-01 -6.06346071e-01 4.50956039e-02
-1.64859927e+00 -1.16003728e+00 -5.47852993e-01 1.71319187e+00
1.14477061e-01 6.53470933e-01 6.60706580e-01 -2.15005994e-01
6.82098448e-01 2.24567071e-01 -9.81186926e-02 -2.59476185e-01
-3.05923074e-02 1.85363799e-01 1.04475284e+00 2.17482194e-01
-1.76724041e+00 1.44339848e+00 7.45440054e+00 8.26655328e-01
-2.88571745e-01 3.99478793e-01 5.43742537e-01 -2.59441853e-01
9.04262304e-01 -2.84375638e-01 -1.33036089e+00 -5.84114753e-02
9.72410977e-01 6.34213328e-01 -4.51034874e-01 1.31099010e+00
-6.92356527e-02 -5.15701175e-01 -1.17696369e+00 1.13725281e+00
9.97372642e-02 -1.32868290e+00 -6.38353646e-01 1.22851200e-01
3.11597943e-01 3.06707233e-01 -6.18100584e-01 9.10232544e-01
-8.72973800e-02 -8.50223005e-01 1.01967037e+00 7.17959553e-02
6.43835664e-01 -6.32629633e-01 9.14477468e-01 8.03876966e-02
-1.73806262e+00 2.94333160e-01 -3.84427369e-01 2.37563744e-01
6.97302401e-01 -2.47312173e-01 -6.60389423e-01 4.14038152e-01
9.60118234e-01 8.36432993e-01 -8.45454872e-01 1.22579443e+00
-4.53502178e-01 4.61608529e-01 -3.22137773e-01 2.76009858e-01
8.87333453e-02 5.40384948e-01 4.23599392e-01 1.43754530e+00
-1.10286571e-01 6.91514552e-01 5.39761424e-01 -4.24122214e-01
6.59457073e-02 -4.79273766e-01 -2.30306432e-01 -1.51120769e-02
-2.44149178e-01 1.28444946e+00 -1.35980284e+00 -7.01045215e-01
-2.74607092e-01 1.72386026e+00 -1.08147286e-01 -2.34134927e-01
-1.39837420e+00 2.44942650e-01 6.62708044e-01 4.26836200e-02
6.24225199e-01 -2.41061509e-01 1.43843144e-02 -1.14284062e+00
-2.46938365e-03 -7.02742815e-01 6.84941351e-01 -9.92216587e-01
-4.73072201e-01 7.54815817e-01 3.59711558e-01 -1.18472528e+00
-1.64060652e-01 -1.04144049e+00 -2.49089822e-01 1.60810918e-01
-4.30169821e-01 -1.17230761e+00 -2.63897270e-01 6.12221956e-01
1.23813963e+00 1.48705309e-02 6.54435635e-01 2.71863341e-01
-5.83878160e-01 6.91067576e-01 -7.06463397e-01 5.85097909e-01
3.97725731e-01 -1.25243831e+00 1.13481224e+00 1.03549695e+00
4.72922713e-01 -1.58993125e-01 7.16211915e-01 -9.57493842e-01
-1.35565674e+00 -7.19318688e-01 6.16629243e-01 -1.13111019e+00
2.71399289e-01 -5.33596098e-01 1.52769294e-02 1.10502028e+00
1.15685001e-01 6.79469168e-01 5.66621944e-02 -3.34770769e-01
-5.40502034e-02 4.33247596e-01 -1.23946142e+00 2.68356800e-01
1.63567400e+00 -5.65501451e-02 -4.10998344e-01 8.51353586e-01
5.58389068e-01 -1.42491055e+00 -6.64381623e-01 4.22273934e-01
6.66519284e-01 -1.09183609e+00 1.09543717e+00 -8.40922773e-01
3.16805720e-01 -1.97261989e-01 -4.04120088e-01 -5.10115802e-01
-2.55683154e-01 -3.74395996e-01 -2.86352038e-01 6.54638767e-01
1.37749746e-01 4.40284252e-01 1.60216641e+00 6.04805350e-01
-5.66500306e-01 -4.14392859e-01 -1.46262085e+00 -8.71435404e-01
-5.18429816e-01 -8.24704885e-01 -6.35619044e-01 -9.79407802e-02
-3.23549330e-01 7.02769533e-02 -3.35057527e-01 1.18928105e-01
5.61719954e-01 -5.33932686e-01 8.20464611e-01 -4.09902513e-01
-5.46531916e-01 -1.03230059e-01 -1.26731861e+00 -1.10692251e+00
-1.65361777e-01 -1.60439253e-01 4.63675320e-01 -1.79404461e+00
9.34837162e-02 5.39713442e-01 5.51596060e-02 6.87495053e-01
1.66604847e-01 7.27429330e-01 6.04141831e-01 -1.98314294e-01
-1.20573711e+00 2.53308564e-01 1.16115880e+00 1.05109632e-01
1.14555679e-01 4.99450326e-01 1.50588140e-01 7.92700350e-01
3.49219918e-01 -5.14806211e-01 -2.43137732e-01 6.20483160e-02
-4.03638810e-01 5.34193873e-01 3.66208762e-01 -1.58597016e+00
-1.80625424e-01 9.94597077e-02 8.84082317e-01 -1.17307746e+00
8.60058963e-01 -6.19164348e-01 -5.71285337e-02 1.05285299e+00
-2.65000053e-02 1.84781656e-01 5.03004253e-01 5.09864330e-01
-4.86863479e-02 1.48160741e-01 8.02539170e-01 -4.52266365e-01
-1.60884881e+00 3.61362427e-01 -4.99881566e-01 3.18749338e-01
1.31496656e+00 -5.26674807e-01 -2.78886050e-01 -3.27246904e-01
-1.47601819e+00 2.41155863e-01 3.91198359e-02 5.51317811e-01
5.69419265e-01 -1.09025216e+00 -6.87597096e-01 -2.55496114e-01
1.43900573e-01 -3.08747798e-01 5.84362686e-01 1.15928698e+00
-9.67476308e-01 7.85554051e-01 -5.70679069e-01 -5.58416367e-01
-1.82645655e+00 1.71944514e-01 4.58522826e-01 -3.16892236e-01
-9.34840262e-01 1.58548939e+00 -6.31198734e-02 -4.18113545e-02
4.84279454e-01 -3.66249651e-01 -1.26500875e-02 -3.25830989e-02
5.70757747e-01 3.78314793e-01 2.65458345e-01 -8.31736743e-01
-1.23626149e+00 3.87347013e-01 1.01786338e-01 -5.38453102e-01
8.88423622e-01 1.57904044e-01 6.57854140e-01 -6.80399463e-02
1.03035164e+00 -2.29094490e-01 -1.69575644e+00 5.40843368e-01
-2.40528598e-01 -2.81216174e-01 -6.21371754e-02 -1.11411929e+00
-1.30312765e+00 7.12055385e-01 1.00154006e+00 -4.45956677e-01
7.64176250e-01 5.60876310e-01 7.33864605e-01 4.09643739e-01
6.39277518e-01 -1.37226379e+00 1.24686748e-01 5.84126353e-01
7.84814417e-01 -1.10110044e+00 3.04605097e-01 -7.77744412e-01
-1.08303404e+00 1.14328086e+00 1.10671592e+00 -2.06638545e-01
2.71360666e-01 5.06629109e-01 1.98602051e-01 -3.22579652e-01
-5.95857978e-01 -3.19573343e-01 3.61875504e-01 7.95347571e-01
5.26561320e-01 5.68302115e-03 -2.25267768e-01 3.21347237e-01
-5.73716573e-02 -1.37573972e-01 3.58605087e-01 1.27195442e+00
-3.65199119e-01 -9.39164221e-01 -1.74571559e-01 -9.81100351e-02
-5.60835123e-01 1.59544364e-01 -7.09163427e-01 1.33486283e+00
1.14361040e-01 5.64621568e-01 -5.58866300e-02 -3.53058010e-01
7.60266483e-01 2.43444279e-01 1.08498228e+00 -7.40274429e-01
-6.38801634e-01 3.05388153e-01 8.15421641e-01 -1.74054885e+00
-3.95553172e-01 -1.22438836e+00 -1.56501114e+00 2.69943643e-02
1.22018225e-01 -2.63890743e-01 8.23876202e-01 1.11493337e+00
-2.59564370e-02 6.75453842e-01 -1.44636527e-01 -1.31724966e+00
-1.40608132e-01 -9.35683727e-01 -5.05505204e-01 -1.88867837e-01
-2.08872274e-01 -5.79193890e-01 7.75178596e-02 3.16413492e-02] | [8.069623947143555, 0.3946479856967926] |
ebce1245-bfbe-4677-9322-84aa0b18f005 | deep-analysis-of-visual-product-reviews | 2207.09499 | null | https://arxiv.org/abs/2207.09499v1 | https://arxiv.org/pdf/2207.09499v1.pdf | Deep Analysis of Visual Product Reviews | With the proliferation of the e-commerce industry, analyzing customer feedback is becoming indispensable to a service provider. In recent days, it can be noticed that customers upload the purchased product images with their review scores. In this paper, we undertake the task of analyzing such visual reviews, which is very new of its kind. In the past, the researchers worked on analyzing language feedback, but here we do not take any assistance from linguistic reviews that may be absent, since a recent trend can be observed where customers prefer to quickly upload the visual feedback instead of typing language feedback. We propose a hierarchical architecture, where the higher-level model engages in product categorization, and the lower-level model pays attention to predicting the review score from a customer-provided product image. We generated a database by procuring real visual product reviews, which was quite challenging. Our architecture obtained some promising results by performing extensive experiments on the employed database. The proposed hierarchical architecture attained a 57.48% performance improvement over the single-level best comparable architecture. | ['Muhammad Saqib', 'Soumi Chattopadhyay', 'Chandranath Adak'] | 2022-07-19 | null | null | null | null | ['product-categorization'] | ['miscellaneous'] | [ 1.51281431e-01 2.59531066e-02 -3.41563970e-02 -5.96660852e-01
-5.76907814e-01 -4.91935849e-01 5.92633545e-01 7.25821316e-01
-4.42845762e-01 4.36681211e-01 2.70056054e-02 -4.94971156e-01
2.21081659e-01 -6.09806895e-01 -3.93103272e-01 -5.54643631e-01
1.01554886e-01 2.19111964e-01 4.13467810e-02 -3.39500904e-01
7.04434693e-01 3.47133055e-02 -1.57542431e+00 6.95054770e-01
7.40739167e-01 1.17063010e+00 4.36523587e-01 6.56759024e-01
-3.99880141e-01 8.70481491e-01 -7.15442717e-01 -8.30278158e-01
1.79139793e-01 -3.21799904e-01 -5.84809542e-01 6.17272496e-01
2.94369131e-01 -3.08837071e-02 4.23985481e-01 1.15164793e+00
2.19882399e-01 -1.91473797e-01 7.05755591e-01 -1.04597950e+00
-1.37278235e+00 3.91616613e-01 -9.61507261e-01 1.07128881e-01
4.97975022e-01 -2.00670421e-01 1.35087371e+00 -1.11897326e+00
5.93460441e-01 9.59932804e-01 1.00395456e-01 1.65559173e-01
-9.10577893e-01 -2.23625183e-01 6.58680439e-01 4.14997965e-01
-1.17382443e+00 -9.92492437e-02 9.85311270e-01 -7.19311893e-01
6.99392498e-01 2.25068226e-01 6.60609901e-01 6.33842349e-01
2.81278074e-01 9.36562598e-01 1.26227129e+00 -7.44591892e-01
1.68502316e-01 1.03841996e+00 2.13135675e-01 4.71960545e-01
2.95144916e-02 -3.72514397e-01 -3.66044194e-01 2.25346729e-01
1.09422348e-01 1.99268535e-01 -1.15602873e-01 -5.33610761e-01
-6.30245447e-01 8.29012156e-01 4.84720320e-01 5.40652096e-01
-4.06208783e-01 -4.53215390e-01 4.84339565e-01 6.59786105e-01
7.45761395e-01 -8.28062836e-03 -3.10215771e-01 -1.71914175e-02
-7.70027578e-01 -1.03487447e-01 6.60993218e-01 9.15914655e-01
6.19066179e-01 -2.87647009e-01 3.78721982e-01 1.00534081e+00
5.27926743e-01 1.72513828e-01 6.31079197e-01 -2.93859452e-01
4.79130566e-01 8.92849505e-01 2.39975810e-01 -1.45582354e+00
-1.47467256e-01 -5.39142609e-01 -8.65771115e-01 4.24799442e-01
3.33609253e-01 1.77167714e-01 -7.20351100e-01 1.12219787e+00
1.00189738e-01 -6.48642361e-01 5.54017350e-02 1.03197265e+00
7.82583237e-01 6.32702231e-01 -1.70537606e-02 -5.06952107e-01
1.40603793e+00 -1.11576438e+00 -8.02639782e-01 -1.16139702e-01
6.16372526e-01 -1.23132730e+00 1.48277521e+00 1.07492387e+00
-9.98273194e-01 -9.33795571e-01 -1.27707016e+00 1.18889950e-01
-6.45551741e-01 4.63100880e-01 6.61705852e-01 7.80676544e-01
-1.16558075e+00 2.40947857e-01 -2.86782622e-01 -6.80177212e-01
1.26378193e-01 1.04595236e-01 -4.40182120e-01 -5.88577725e-02
-9.29089189e-01 8.02760124e-01 -2.22968295e-01 5.24317205e-01
-4.52379853e-01 -1.57763109e-01 -6.93958342e-01 -1.42591983e-01
4.34746057e-01 -2.11727336e-01 1.26560736e+00 -1.57780492e+00
-1.31625569e+00 9.23048377e-01 -2.20899776e-01 -2.29925022e-01
5.10450304e-01 -1.62175700e-01 -7.20871747e-01 -7.86605999e-02
-1.56779841e-01 3.84125918e-01 9.43945110e-01 -1.47659409e+00
-8.80874038e-01 -5.20679653e-01 3.46772909e-01 6.86657727e-02
-6.95964336e-01 -4.91527393e-02 -4.10993665e-01 -3.09838921e-01
-4.97011095e-02 -7.01382279e-01 -9.98824537e-02 -2.36194119e-01
-7.95228407e-02 -1.92996114e-01 5.86588562e-01 -4.82691199e-01
1.35970533e+00 -2.24765801e+00 -1.53965592e-01 1.17080294e-01
2.01665416e-01 2.61270612e-01 -5.24142496e-02 6.07325435e-01
-3.72647308e-02 2.33320609e-01 -1.38909854e-02 -5.15814960e-01
1.33255087e-02 -1.57217562e-01 -3.57659549e-01 2.12083608e-01
6.66410476e-02 6.57242477e-01 -7.67732322e-01 -5.93107462e-01
-2.78877304e-03 4.01652873e-01 -5.05952656e-01 1.92521930e-01
-1.52074434e-02 4.22647931e-02 -2.96160907e-01 8.71547878e-01
6.97447360e-01 -3.53139967e-01 2.29056641e-01 -1.97925791e-01
-3.08610320e-01 4.45085727e-02 -9.01120186e-01 1.44461787e+00
-6.09933138e-01 7.12297142e-01 5.69889955e-02 -9.78648722e-01
1.30255318e+00 1.84979677e-01 3.30258936e-01 -1.01515114e+00
2.51155227e-01 1.23088911e-01 1.14939056e-01 -5.79897106e-01
9.71048534e-01 -1.13288313e-01 -4.41977382e-02 5.11719704e-01
-4.15295511e-01 6.51571527e-02 4.11335737e-01 3.49160641e-01
7.77709723e-01 5.46780275e-03 5.21070182e-01 -6.33907542e-02
1.00059354e+00 1.79346919e-01 -1.58568680e-01 4.35751021e-01
-3.36019158e-01 5.82879782e-01 4.87622082e-01 -6.41250730e-01
-9.45642233e-01 -8.29862773e-01 1.37505442e-01 1.20268404e+00
8.70471597e-02 -5.16807020e-01 -5.53870082e-01 -7.18600929e-01
-1.17598943e-01 2.55919904e-01 -6.06549919e-01 6.42754734e-02
-1.14014089e-01 -5.80219328e-01 -3.39624614e-01 2.20007315e-01
3.09365511e-01 -1.22939265e+00 -4.12224293e-01 1.57136947e-01
2.66094953e-01 -7.12648749e-01 -4.89616007e-01 7.97189325e-02
-7.58070350e-01 -7.76735008e-01 -8.96134436e-01 -1.06139469e+00
8.62051249e-01 4.07816559e-01 1.22252035e+00 2.21231490e-01
-2.31645063e-01 3.53210300e-01 -7.24033356e-01 -4.92593646e-01
-4.15535420e-01 3.48774731e-01 -2.15701699e-01 4.57633585e-01
7.14928091e-01 -4.13831085e-01 -6.68567061e-01 3.10160637e-01
-9.36880291e-01 -1.75189301e-02 9.08309042e-01 6.99561954e-01
2.95528829e-01 6.05879501e-02 8.35128009e-01 -1.26815999e+00
8.88915598e-01 -2.57937163e-01 -3.96017909e-01 3.22284758e-01
-8.21291327e-01 -1.05435021e-01 7.97413945e-01 -2.54401773e-01
-1.08615160e+00 1.20090358e-01 -1.79913312e-01 2.31665537e-01
-2.09786788e-01 7.84852505e-01 7.93393049e-03 1.87203109e-01
4.89649713e-01 1.51591733e-01 -1.42433241e-01 -5.87215185e-01
3.33732128e-01 1.28076887e+00 2.64819503e-01 -1.73346743e-01
4.00966734e-01 2.74834186e-01 -5.37450135e-01 -8.31062376e-01
-7.64677048e-01 -7.06509829e-01 -7.46641040e-01 -5.69690704e-01
5.84750235e-01 -6.31677389e-01 -9.39460278e-01 1.67635664e-01
-1.11564088e+00 1.76015824e-01 -1.95251200e-02 1.55006111e-01
-2.01025531e-01 4.30044502e-01 -5.02726138e-01 -1.11316848e+00
-1.62103757e-01 -1.14285529e+00 8.43605459e-01 6.04134500e-02
-2.28112459e-01 -8.48602891e-01 4.95263329e-03 5.04445136e-01
4.28510517e-01 -2.18565688e-01 9.39171910e-01 -4.61476624e-01
-4.46919024e-01 -4.32079196e-01 -3.79782557e-01 6.36556804e-01
3.33935052e-01 1.24958284e-01 -1.02330840e+00 -2.23659024e-01
1.18684717e-01 -3.53671998e-01 6.92274451e-01 8.49334523e-02
7.65552819e-01 -2.54619867e-02 1.04804702e-01 -1.14456624e-01
1.44766057e+00 5.15863895e-01 6.83958352e-01 3.29609364e-01
4.78945971e-01 9.67831552e-01 1.03562057e+00 4.67647046e-01
4.52460378e-01 5.79155028e-01 5.61485648e-01 -2.48406887e-01
2.78456628e-01 -1.67775854e-01 4.31018233e-01 1.38246691e+00
-6.31861538e-02 -3.33505005e-01 -6.44072831e-01 6.02971435e-01
-1.79330516e+00 -6.70790255e-01 -1.25792533e-01 2.06784678e+00
5.52992702e-01 4.90258753e-01 1.91476136e-01 2.93775141e-01
5.96494555e-01 2.14027226e-01 -4.37379509e-01 -8.79725575e-01
1.66862994e-01 -1.91008821e-01 1.15473941e-01 6.48191214e-01
-9.14797246e-01 6.08189404e-01 5.72763824e+00 4.77929324e-01
-1.18799448e+00 1.21968701e-01 8.94020617e-01 4.04194929e-02
-3.67661387e-01 -2.36682191e-01 -7.90424407e-01 3.57444167e-01
7.98874676e-01 -9.19675156e-02 -3.07679158e-02 1.05909228e+00
3.28401059e-01 -3.27051491e-01 -9.26911294e-01 1.05127037e+00
4.35962468e-01 -7.33230889e-01 1.27785891e-01 9.68639627e-02
4.97258335e-01 -5.81638813e-01 3.64509761e-01 2.70453095e-01
-1.99357625e-02 -8.59362125e-01 7.70073831e-01 4.23157722e-01
4.44017947e-01 -8.82519603e-01 9.73641694e-01 3.92796427e-01
-1.20416605e+00 -1.26453653e-01 -5.40019929e-01 -3.78392607e-01
2.54210830e-01 6.13808095e-01 -1.03151155e+00 6.48515940e-01
6.19466662e-01 1.00008953e+00 -9.09143627e-01 8.82476389e-01
4.80530746e-02 1.98864460e-01 3.19689155e-01 -4.71150070e-01
1.01885125e-01 -5.64904153e-01 -1.20024875e-01 1.13986540e+00
4.88772422e-01 -2.18072191e-01 8.02645534e-02 4.30632621e-01
6.53360710e-02 8.18183839e-01 -6.55039072e-01 -3.29942197e-01
-2.99107224e-01 1.66846442e+00 -9.03948843e-01 -3.47143292e-01
-7.66700387e-01 9.82974112e-01 3.01849633e-01 1.07597098e-01
-6.27547860e-01 -3.64228547e-01 3.30490977e-01 2.68713504e-01
3.11352730e-01 -8.63072500e-02 -3.11317921e-01 -8.65781665e-01
3.90066206e-01 -7.79666722e-01 2.27834091e-01 -1.04563463e+00
-1.25174844e+00 1.06203854e+00 -5.56220412e-01 -1.51005936e+00
-2.64731497e-01 -7.67675400e-01 -1.01841874e-02 8.35402906e-01
-1.53139460e+00 -8.85836780e-01 -1.90987661e-01 2.79239655e-01
8.29223752e-01 -1.62814364e-01 5.52425325e-01 7.22884417e-01
-1.76589936e-01 4.56245929e-01 -1.35788843e-01 -2.33838245e-01
9.77765977e-01 -1.29772997e+00 -2.39499044e-02 6.07282043e-01
3.87274265e-01 6.71005607e-01 9.46370244e-01 -4.67180520e-01
-1.16781187e+00 -5.62024593e-01 1.33548808e+00 -4.00076002e-01
9.02142107e-01 -4.70984966e-01 -7.59315729e-01 2.49188468e-01
6.11205339e-01 -3.61509800e-01 6.67576253e-01 1.63032100e-01
-2.73144692e-01 -4.90710825e-01 -8.46892774e-01 5.33303082e-01
7.68800139e-01 -6.00963950e-01 -5.07571518e-01 1.68625206e-01
5.56031704e-01 1.09302521e-01 -5.70056140e-01 -3.93168591e-02
6.46526456e-01 -1.12038934e+00 5.96637189e-01 -3.35689843e-01
6.78640544e-01 -5.58139443e-01 1.93065982e-02 -1.31823599e+00
-7.57733658e-02 -1.18275292e-01 2.83409983e-01 1.45864809e+00
7.12327659e-01 -3.92512321e-01 8.57873678e-01 3.85139853e-01
-1.46581560e-01 -7.30680645e-01 -2.45818779e-01 -3.26283216e-01
-3.47839564e-01 -4.76817966e-01 2.89419800e-01 4.65360492e-01
3.04851472e-01 5.37784278e-01 -5.49295306e-01 -5.92055060e-02
1.07636116e-01 4.29862648e-01 7.45876014e-01 -1.24034452e+00
-4.66456175e-01 -3.31442446e-01 -3.98710370e-01 -1.36018360e+00
-3.29195231e-01 -7.19454944e-01 1.23024397e-01 -1.56625807e+00
3.18923801e-01 -8.49656537e-02 -4.81981605e-01 8.93638581e-02
1.05078686e-02 6.58708036e-01 2.04008251e-01 1.41926691e-01
-8.96812201e-01 9.07363892e-02 1.37761426e+00 -3.96740496e-01
-1.88574046e-01 7.71683902e-02 -1.19344175e+00 6.09511077e-01
5.66390097e-01 -2.71170080e-01 -6.30123377e-01 -2.45374903e-01
7.74489582e-01 -1.35368675e-01 -6.78386837e-02 -6.19457245e-01
3.70396167e-01 1.75631747e-01 2.70801902e-01 -7.57303357e-01
2.72482634e-01 -1.12507069e+00 5.67170642e-02 2.37587526e-01
-3.68429005e-01 5.47802329e-01 -2.17519850e-01 5.15395224e-01
-7.09525883e-01 -2.37442195e-01 5.00849426e-01 -1.88133210e-01
-7.49207258e-01 -9.93784741e-02 -3.99492323e-01 -5.81857324e-01
8.65215421e-01 -2.67885149e-01 -3.25304836e-01 -5.95523059e-01
-8.53745043e-01 -7.17171431e-02 4.88140196e-01 5.67911625e-01
6.66076064e-01 -1.20133245e+00 -3.73231888e-01 2.42261022e-01
4.99153733e-01 -5.34783423e-01 3.88053417e-01 8.43026221e-01
-4.28486496e-01 5.70018470e-01 -1.04312040e-01 -6.38294578e-01
-1.57423449e+00 9.30368304e-01 -2.47370124e-01 -3.60228628e-01
-2.48227239e-01 6.25458360e-01 2.51013368e-01 -9.53896195e-02
2.75679827e-01 -3.31766158e-01 -7.67837167e-01 6.74988091e-01
8.13607514e-01 -2.31349431e-02 4.54145819e-01 -7.68740594e-01
-1.76610962e-01 7.20444500e-01 -5.17073095e-01 -1.38399750e-01
1.35678840e+00 -5.50162673e-01 4.75946907e-03 8.73467982e-01
1.24989223e+00 1.60968035e-01 -9.35377479e-01 -1.09754205e-01
2.47894913e-01 -5.95945120e-01 -1.87343940e-01 -8.27968776e-01
-1.14072955e+00 1.06805944e+00 6.51246965e-01 1.03278148e+00
1.38864875e+00 -2.37967409e-02 5.44258058e-01 2.62770206e-01
6.12384617e-01 -1.11592615e+00 3.64427753e-02 2.40927368e-01
8.47622216e-01 -1.67598236e+00 8.69847313e-02 -4.30217028e-01
-9.48438883e-01 8.98924887e-01 3.74238640e-01 -8.99506956e-02
5.99230886e-01 3.40417819e-03 4.84271049e-01 -2.08287582e-01
-1.01541114e+00 -2.23815113e-01 3.23645443e-01 3.80509496e-01
8.60202849e-01 -2.87839733e-02 -6.43998742e-01 6.15432620e-01
-1.41752556e-01 2.78877914e-02 6.60819173e-01 8.50019455e-01
-5.92751622e-01 -1.47798502e+00 -2.66593128e-01 3.02485794e-01
-5.24010897e-01 -1.36896610e-01 -5.55922329e-01 8.69236946e-01
-7.69381747e-02 1.16383862e+00 -6.62230998e-02 -3.98417056e-01
6.15096390e-01 7.67140910e-02 1.97929472e-01 -6.51639819e-01
-6.89493954e-01 7.56174773e-02 -1.06549147e-03 -3.30653965e-01
-6.97934031e-01 -6.37780786e-01 -7.57646203e-01 -2.22768500e-01
-1.10975198e-01 5.15745223e-01 9.41192031e-01 6.56524062e-01
1.14329383e-01 4.13455099e-01 8.01734686e-01 -7.33209252e-01
-2.38575801e-01 -9.53867495e-01 -8.00442636e-01 5.50639153e-01
4.00394648e-01 -4.39096481e-01 -5.53360879e-01 4.41613436e-01] | [11.207559585571289, 6.625596523284912] |
6887c18f-2531-4d10-a15c-82e64d47057d | small-data-no-problem-exploring-the-viability | null | null | https://aclanthology.org/2021.mrl-1.11 | https://aclanthology.org/2021.mrl-1.11.pdf | Small Data? No Problem! Exploring the Viability of Pretrained Multilingual Language Models for Low-resourced Languages | Pretrained multilingual language models have been shown to work well on many languages for a variety of downstream NLP tasks. However, these models are known to require a lot of training data. This consequently leaves out a huge percentage of the world’s languages as they are under-resourced. Furthermore, a major motivation behind these models is that lower-resource languages benefit from joint training with higher-resource languages. In this work, we challenge this assumption and present the first attempt at training a multilingual language model on only low-resource languages. We show that it is possible to train competitive multilingual language models on less than 1 GB of text. Our model, named AfriBERTa, covers 11 African languages, including the first language model for 4 of these languages. Evaluations on named entity recognition and text classification spanning 10 languages show that our model outperforms mBERT and XLM-Rin several languages and is very competitive overall. Results suggest that our “small data” approach based on similar languages may sometimes work better than joint training on large datasets with high-resource languages. Code, data and models are released at https://github.com/keleog/afriberta. | ['Jimmy Lin', 'Yuxin Zhu', 'Kelechi Ogueji'] | null | null | null | null | emnlp-mrl-2021-11 | ['pretrained-multilingual-language-models'] | ['natural-language-processing'] | [-6.85164094e-01 -5.76098412e-02 -5.90571702e-01 -4.13081735e-01
-1.25777161e+00 -8.23069692e-01 6.50800347e-01 1.58275187e-01
-8.94757926e-01 1.05588078e+00 3.10544699e-01 -8.72607231e-01
3.89865965e-01 -6.03391051e-01 -6.89758241e-01 -1.71305716e-01
2.24862844e-01 8.58659744e-01 -3.20001990e-02 -3.13476652e-01
-2.14758366e-01 3.77833396e-01 -6.60434663e-01 3.79128754e-01
1.08465028e+00 2.76243016e-02 3.92496109e-01 4.58530605e-01
-7.04263985e-01 7.54198849e-01 -3.93138111e-01 -6.32212818e-01
2.08591700e-01 -2.07505897e-01 -1.13162839e+00 -4.57075000e-01
4.53892261e-01 -3.96747189e-03 -4.31299582e-02 7.95076072e-01
5.20398378e-01 -2.54945189e-01 3.66568029e-01 -9.44417000e-01
-5.17099738e-01 1.20830131e+00 -4.38443929e-01 1.28267661e-01
2.67332584e-01 -1.15936667e-01 1.14176214e+00 -1.08077896e+00
1.07504058e+00 1.24202061e+00 8.77341509e-01 2.73388267e-01
-1.11714458e+00 -9.66717005e-01 1.43167615e-01 -4.38076615e-01
-1.69795465e+00 -7.56328702e-01 2.26192147e-01 -4.26872581e-01
1.56347442e+00 -4.45595710e-03 1.73829228e-01 8.70577872e-01
1.49752036e-01 8.22427452e-01 1.27422607e+00 -7.33315051e-01
-4.28404987e-01 6.54178441e-01 -6.53340891e-02 6.50074422e-01
4.19307798e-01 -1.96738765e-01 -4.48396772e-01 -3.35276961e-01
3.64451945e-01 -3.87161940e-01 -1.78753600e-01 2.54222564e-02
-1.38216853e+00 1.06126010e+00 1.70694396e-01 9.24169540e-01
-1.04735941e-01 -3.00323725e-01 4.46123511e-01 4.85248178e-01
9.63267148e-01 4.60018754e-01 -1.07438791e+00 -7.41451606e-02
-1.03909230e+00 -5.55117093e-02 1.13263798e+00 1.03017282e+00
8.56281161e-01 9.18272510e-02 5.29265940e-01 1.26167560e+00
3.04849893e-01 7.16016710e-01 4.63341296e-01 -4.18125957e-01
9.96903002e-01 5.24802804e-01 -9.91098583e-02 -5.33086777e-01
-5.63179612e-01 -2.48776376e-01 -6.73270106e-01 -2.32460499e-01
8.23283195e-01 -4.47727174e-01 -8.60113263e-01 1.66900718e+00
1.17506087e-01 -4.67538267e-01 4.22959685e-01 5.04890144e-01
6.19728684e-01 9.09226418e-01 3.22792768e-01 3.82330492e-02
1.30006814e+00 -9.50882912e-01 -4.46693748e-01 -4.85640794e-01
1.25059569e+00 -1.28500319e+00 9.33737755e-01 6.33764043e-02
-9.69971478e-01 -3.28091800e-01 -5.93393028e-01 -2.66640365e-01
-8.62102389e-01 3.85100573e-01 9.74122584e-01 8.32089126e-01
-1.22096324e+00 7.62276649e-02 -8.50681961e-01 -9.70298409e-01
9.77399871e-02 2.95719117e-01 -7.35910714e-01 -2.82709122e-01
-1.34135473e+00 1.06788206e+00 6.82767153e-01 -3.72241102e-02
-5.47256172e-01 -5.49845934e-01 -8.43372405e-01 -4.01832551e-01
8.63158107e-02 -2.95642406e-01 9.96497333e-01 -8.45405698e-01
-1.03642070e+00 1.22545993e+00 -2.66268432e-01 -2.77788728e-01
5.38953483e-01 -2.64752775e-01 -7.62007594e-01 -2.77966648e-01
4.45403963e-01 7.27482617e-01 -1.06700351e-02 -9.22150314e-01
-8.51873398e-01 -1.95617184e-01 -1.83428496e-01 1.64135396e-01
-3.61962795e-01 7.42420256e-01 -5.49208343e-01 -6.10870421e-01
-1.58405259e-01 -9.68783498e-01 -2.92846411e-01 -6.87652230e-01
-3.35244209e-01 -2.14256197e-01 4.31990027e-01 -1.20651782e+00
1.17926180e+00 -1.85736370e+00 -2.80667484e-01 7.62964934e-02
-2.08878189e-01 2.85198927e-01 -2.72096723e-01 7.48978198e-01
1.30029600e-02 6.88366711e-01 -6.63622171e-02 -1.52627259e-01
-1.50147796e-01 2.48169556e-01 -2.74629861e-01 4.92725134e-01
3.23331267e-01 7.08071291e-01 -7.47830868e-01 -4.44579571e-01
-1.05728291e-01 6.53078735e-01 -5.36619425e-01 -2.11333875e-02
-9.45677608e-02 5.71645796e-01 -2.88164228e-01 8.04410338e-01
5.95918596e-01 -3.22301686e-02 5.99624336e-01 2.60006309e-01
-6.12270355e-01 8.95047665e-01 -9.74928141e-01 1.74878395e+00
-7.13413894e-01 5.65850556e-01 1.88743711e-01 -5.86220503e-01
7.34333575e-01 4.75944012e-01 2.07302213e-01 -5.13273180e-01
-2.23026872e-01 9.02126908e-01 2.14077979e-01 -1.34350464e-01
5.55949569e-01 -2.06466928e-01 -3.80095512e-01 5.21508932e-01
2.53312439e-01 -9.42493528e-02 3.49463761e-01 2.97720402e-01
6.77785218e-01 2.14740798e-01 5.87792575e-01 -4.55543488e-01
5.11774421e-01 3.71006817e-01 7.45702744e-01 5.23985863e-01
-2.24364579e-01 1.22105628e-01 2.42252424e-01 -4.03947949e-01
-1.28957868e+00 -7.43524671e-01 -4.49719757e-01 1.45996118e+00
-6.89027488e-01 -4.06091839e-01 -4.00117904e-01 -6.74293220e-01
-4.72773500e-02 8.05783927e-01 -5.65121556e-03 6.09756291e-01
-8.26599240e-01 -1.02633274e+00 1.14947391e+00 2.47959167e-01
3.02553743e-01 -1.05869961e+00 2.44285673e-01 2.70498395e-01
-2.33740687e-01 -1.22476065e+00 -4.24333602e-01 3.04744899e-01
-6.72824979e-01 -7.66474903e-01 -9.42607164e-01 -1.08952487e+00
4.72298622e-01 -1.57386214e-01 1.40886056e+00 -2.16759324e-01
8.97760168e-02 1.33548826e-01 -1.88875914e-01 -4.26542073e-01
-8.32735419e-01 7.36856818e-01 1.64358109e-01 -3.98892760e-01
6.59492731e-01 -1.57898203e-01 2.35302746e-02 2.75223330e-02
-5.96446574e-01 7.48922974e-02 6.74063683e-01 6.44730449e-01
4.59454775e-01 -1.85621753e-01 6.34205103e-01 -1.35243940e+00
3.99879992e-01 -7.02031136e-01 -6.28406882e-01 5.20920813e-01
-5.78864038e-01 4.43067215e-02 6.21052682e-01 -2.24397838e-01
-1.18179929e+00 -3.14088091e-02 -2.90882170e-01 1.48039550e-01
-3.62988442e-01 8.16427886e-01 -1.77456379e-01 1.72195345e-01
4.84842777e-01 -1.48251886e-02 -5.38607121e-01 -8.38052630e-01
3.91266465e-01 9.98608649e-01 -1.67334657e-02 -6.39914751e-01
5.91196060e-01 2.64402241e-01 -6.98867738e-01 -1.03405571e+00
-6.93524301e-01 -7.79685438e-01 -8.31803381e-01 2.71281809e-01
8.71564627e-01 -1.44500530e+00 2.99103446e-02 4.87730622e-01
-1.16785049e+00 -6.12820268e-01 1.72718078e-01 8.55678856e-01
7.63419047e-02 7.07335249e-02 -1.01748598e+00 -7.33734787e-01
-3.60973150e-01 -9.78742838e-01 8.39705825e-01 -1.27020776e-01
-2.67839223e-01 -1.53044832e+00 3.92789304e-01 3.61649424e-01
3.87515455e-01 -2.34772608e-01 8.97626102e-01 -1.23183763e+00
-2.55497813e-01 2.91473158e-02 -1.56331390e-01 2.77932316e-01
2.40801409e-01 -5.79026677e-02 -7.29327202e-01 -6.41020715e-01
-3.62377197e-01 -6.32553756e-01 6.83139741e-01 1.18162736e-01
2.67497420e-01 -1.72723755e-01 -3.81440461e-01 5.23665726e-01
1.58012915e+00 -7.51187839e-03 1.52932584e-01 5.23215532e-01
8.90577793e-01 6.36896610e-01 3.36327136e-01 2.23628283e-02
6.86993301e-01 3.23492438e-01 -5.11935115e-01 -5.10961413e-01
-6.60519227e-02 -4.02972043e-01 6.68136656e-01 1.45530760e+00
1.68273300e-01 -2.26920784e-01 -1.70348370e+00 8.59002650e-01
-1.37597430e+00 -6.67329669e-01 -3.06383461e-01 2.13257408e+00
1.20328355e+00 -1.23303883e-01 1.05775505e-01 -6.27616286e-01
7.01512277e-01 6.85539842e-02 -1.74145386e-01 -5.41570246e-01
-6.13719106e-01 3.46074015e-01 8.36497486e-01 7.95700908e-01
-1.13604641e+00 1.59516490e+00 6.51618481e+00 6.75505340e-01
-1.21884942e+00 4.25793648e-01 7.24361598e-01 1.81780353e-01
-3.60003948e-01 4.04563278e-01 -1.41800165e+00 2.49563947e-01
1.43137515e+00 -3.04560572e-01 2.90521741e-01 6.70207858e-01
2.02541858e-01 -1.31979573e-03 -8.13003480e-01 5.20965695e-01
1.51217664e-02 -1.04651821e+00 1.44084603e-01 2.01750457e-01
8.74076724e-01 9.86550868e-01 -2.65594929e-01 5.67594051e-01
9.02642727e-01 -1.00642467e+00 5.89011669e-01 1.38065964e-01
9.90065575e-01 -7.42900848e-01 7.74746656e-01 5.18306673e-01
-1.18894005e+00 4.18555379e-01 -6.02171361e-01 1.50259942e-01
2.38827705e-01 6.02496684e-01 -1.06000459e+00 7.00699985e-01
6.82884037e-01 5.87506115e-01 -6.78410232e-01 7.25080371e-01
-1.32601812e-01 9.36208248e-01 -5.91964662e-01 2.38268077e-01
4.84205782e-01 -1.71913147e-01 2.58292854e-01 1.78167200e+00
3.93625200e-01 -2.57677376e-01 5.30499518e-01 3.44675004e-01
-3.83407027e-01 9.37931120e-01 -9.32814121e-01 -4.99348700e-01
3.05380940e-01 1.11084104e+00 -5.96524715e-01 -3.90423447e-01
-9.23716426e-01 8.46926570e-01 7.54879534e-01 3.26358408e-01
-4.98588234e-01 -4.04406786e-01 3.87365609e-01 -5.46648633e-03
1.11869844e-02 -3.84774029e-01 4.47766436e-03 -1.59435582e+00
-2.54319340e-01 -1.14892733e+00 7.51900673e-01 -4.65393871e-01
-1.39504790e+00 7.27433980e-01 -2.77446449e-01 -8.26912105e-01
-3.84811699e-01 -8.77668738e-01 1.72114968e-02 1.36852384e+00
-1.75395930e+00 -1.57627535e+00 4.77712750e-01 6.50999427e-01
4.69126850e-01 -4.84784603e-01 9.97824311e-01 6.08192265e-01
-5.75826705e-01 6.09110296e-01 3.23445350e-01 6.87192082e-01
1.18029010e+00 -1.07681191e+00 5.53961217e-01 8.70339751e-01
5.54097712e-01 9.28002357e-01 2.81878501e-01 -8.03101838e-01
-1.16903734e+00 -1.07623017e+00 1.84434319e+00 -6.25604391e-01
1.11596644e+00 -5.76365590e-01 -9.97694552e-01 1.25764430e+00
6.61781549e-01 -3.31825137e-01 1.07857394e+00 5.72378099e-01
-6.15952730e-01 1.69789374e-01 -8.43693912e-01 4.27554667e-01
7.41390705e-01 -7.88984478e-01 -5.94498456e-01 6.80710018e-01
5.27969837e-01 -2.15330750e-01 -1.08197868e+00 9.32660773e-02
3.66985351e-01 -6.03003800e-01 5.21425545e-01 -7.99016237e-01
3.96409482e-02 -1.09852977e-01 -1.57327861e-01 -1.35540092e+00
6.25512376e-03 -5.27720690e-01 6.65586054e-01 1.60929692e+00
1.13048446e+00 -1.15036941e+00 3.71980131e-01 3.84038717e-01
1.04269668e-01 -1.77763298e-01 -9.16554689e-01 -9.93117630e-01
8.95440817e-01 -5.60559452e-01 3.10117066e-01 1.71282864e+00
-6.98735416e-02 6.46555364e-01 -2.08892733e-01 1.91900507e-01
3.33202899e-01 -5.89131191e-02 7.66094863e-01 -1.02972734e+00
-1.87178373e-01 -2.77545780e-01 1.67616103e-02 -7.09490418e-01
5.52061975e-01 -1.47590411e+00 -1.60786107e-01 -1.54142594e+00
3.08796525e-01 -1.00315797e+00 -1.83460459e-01 9.29693341e-01
-5.55522367e-02 1.09292574e-01 1.34604231e-01 4.52702850e-01
-1.92873061e-01 -7.92420469e-03 6.19626641e-01 5.59700206e-02
-2.45085910e-01 -2.79855788e-01 -6.48493648e-01 7.28884220e-01
9.64497745e-01 -5.87656558e-01 1.11825995e-01 -8.66990745e-01
3.61736804e-01 -1.79870382e-01 -3.11424911e-01 -6.49602830e-01
1.60807550e-01 -2.36209199e-01 3.50241363e-01 -5.10150909e-01
-1.50190413e-01 -5.46367645e-01 1.31605595e-01 4.51745063e-01
-1.13978334e-01 2.49252647e-01 5.00213921e-01 -6.28375262e-02
-3.84594053e-01 -1.09276019e-01 6.82928979e-01 -3.42278957e-01
-7.56839037e-01 1.25600636e-01 -3.95427585e-01 4.13106292e-01
6.31824076e-01 3.58113468e-01 -2.94658154e-01 -1.77680030e-01
-5.11385083e-01 1.87338084e-01 5.75719118e-01 5.10946274e-01
-1.12979248e-01 -1.03248966e+00 -1.15447319e+00 1.32530183e-01
1.18704930e-01 -3.88453960e-01 -1.73354760e-01 7.42980599e-01
-8.31430972e-01 1.08316779e+00 6.74449503e-02 -3.81565273e-01
-9.59737062e-01 3.87819529e-01 2.55185366e-01 -7.69391894e-01
-2.85467416e-01 7.68711984e-01 1.39006793e-01 -1.35324454e+00
-2.42427558e-01 -1.40553758e-01 -4.82031070e-02 1.45787939e-01
1.22595899e-01 -6.38280576e-03 1.05522528e-01 -1.16307998e+00
-6.05836928e-01 3.30933750e-01 -3.26191247e-01 -5.01448750e-01
1.36593270e+00 -6.95885643e-02 -3.09339672e-01 6.74398065e-01
1.22208679e+00 8.85664761e-01 -4.91962403e-01 -2.34508798e-01
4.82517093e-01 -1.09338485e-01 -5.71032763e-02 -9.44410563e-01
-7.92788625e-01 7.71410406e-01 1.79046705e-01 -1.66637719e-01
8.57185721e-01 1.72152773e-01 6.96076691e-01 5.27294815e-01
6.49448156e-01 -1.20637321e+00 -7.37750351e-01 9.01272833e-01
6.08280122e-01 -1.45975733e+00 4.10673320e-02 -2.11718693e-01
-5.74116230e-01 7.71691680e-01 4.29237306e-01 2.90486991e-01
7.54039884e-01 4.96742964e-01 6.16114795e-01 4.80449982e-02
-7.31043994e-01 -2.61523306e-01 5.16390130e-02 2.85077572e-01
1.28463590e+00 3.15963238e-01 -5.91963530e-01 3.21196586e-01
-4.45728064e-01 -2.11018160e-01 2.74974227e-01 9.76603985e-01
-1.41443670e-01 -1.60040617e+00 -4.21647310e-01 1.69202849e-01
-1.15651238e+00 -7.38152206e-01 -4.20234591e-01 1.29785073e+00
5.61324246e-02 8.37658703e-01 -8.06048047e-03 1.17601715e-01
1.31724387e-01 6.05738282e-01 5.90618216e-02 -8.58899951e-01
-7.85203815e-01 3.78993630e-01 5.14229894e-01 -2.62686938e-01
-4.78433967e-01 -8.06045115e-01 -1.15107596e+00 -4.89079893e-01
-2.26451114e-01 3.84682387e-01 7.68753469e-01 6.79553449e-01
2.80234873e-01 -1.25679433e-01 3.17895085e-01 -3.12135994e-01
-1.43131092e-01 -1.05188298e+00 -4.97846335e-01 6.69292510e-02
-1.22619800e-01 -2.12529246e-02 -4.03831065e-01 1.04271486e-01] | [10.652469635009766, 9.897492408752441] |
7566eff4-fbef-443b-ad40-b4615b2976da | fact-federated-adversarial-cross-training | 2306.00607 | null | https://arxiv.org/abs/2306.00607v1 | https://arxiv.org/pdf/2306.00607v1.pdf | FACT: Federated Adversarial Cross Training | Federated Learning (FL) facilitates distributed model development to aggregate multiple confidential data sources. The information transfer among clients can be compromised by distributional differences, i.e., by non-i.i.d. data. A particularly challenging scenario is the federated model adaptation to a target client without access to annotated data. We propose Federated Adversarial Cross Training (FACT), which uses the implicit domain differences between source clients to identify domain shifts in the target domain. In each round of FL, FACT cross initializes a pair of source clients to generate domain specialized representations which are then used as a direct adversary to learn a domain invariant data representation. We empirically show that FACT outperforms state-of-the-art federated, non-federated and source-free domain adaptation models on three popular multi-source-single-target benchmarks, and state-of-the-art Unsupervised Domain Adaptation (UDA) models on single-source-single-target experiments. We further study FACT's behavior with respect to communication restrictions and the number of participating clients. | ['Michael Altenbuchinger', 'Andreas Schäfer', 'Jonas Lippl', 'Stefan Schrod'] | 2023-06-01 | null | null | null | null | ['source-free-domain-adaptation', 'unsupervised-domain-adaptation'] | ['computer-vision', 'methodology'] | [ 6.66956753e-02 -1.36012316e-01 -4.63988125e-01 -4.95012522e-01
-1.28454006e+00 -1.34223127e+00 8.89299035e-01 4.00100425e-02
-3.86505604e-01 9.47245002e-01 -9.30513963e-02 -4.19718117e-01
1.77466720e-01 -7.90791571e-01 -9.57672417e-01 -6.82936132e-01
-1.53544173e-01 1.08061028e+00 1.86530337e-01 -1.41914397e-01
-3.41303200e-01 5.89464366e-01 -7.69142449e-01 6.23818517e-01
6.50477469e-01 9.50908422e-01 -5.89232564e-01 5.19931972e-01
-3.35143507e-01 6.46189690e-01 -9.17639375e-01 -9.73325253e-01
6.19034350e-01 -3.02264124e-01 -8.94792557e-01 -3.41445982e-01
3.28181267e-01 -5.03283143e-01 -4.14857149e-01 1.04244089e+00
5.65199196e-01 -5.41489273e-02 8.07678223e-01 -1.77355230e+00
-8.44643950e-01 6.86547220e-01 -3.78744990e-01 1.55404717e-01
1.51451007e-01 2.51501232e-01 4.60485190e-01 -5.53160787e-01
7.69056976e-01 1.18126333e+00 4.20514733e-01 9.21776593e-01
-1.31634653e+00 -1.19704425e+00 1.06703021e-01 -7.93233216e-02
-9.76930857e-01 -5.97193241e-01 8.39695632e-01 -2.13899091e-01
5.08643210e-01 -5.67790121e-03 -4.84526157e-01 1.74575508e+00
-1.84387520e-01 5.96988976e-01 1.09071660e+00 -2.89036125e-01
7.31495976e-01 5.73144138e-01 -8.99101198e-02 5.38021699e-02
3.05164725e-01 3.86660725e-01 -5.40464818e-01 -1.02465892e+00
3.31175596e-01 5.55345640e-02 -7.79156461e-02 -7.71380126e-01
-8.79101157e-01 1.10979426e+00 2.07575321e-01 -4.01542000e-02
-3.70832622e-01 -3.20942670e-01 8.73490036e-01 8.38458121e-01
4.45906669e-01 1.81207955e-01 -1.03255570e+00 8.38259235e-02
-4.71130759e-01 4.24474031e-01 1.13390088e+00 1.11288857e+00
7.02420771e-01 -4.68736701e-02 8.43960196e-02 7.36876667e-01
6.83098659e-03 7.49587238e-01 5.98069191e-01 -8.74572515e-01
9.87842023e-01 3.52470130e-01 1.82025909e-01 -3.05251926e-01
3.10976923e-01 -4.13825810e-02 -7.54793108e-01 4.24747229e-01
6.90370560e-01 -6.10451400e-01 -6.77410185e-01 1.99235451e+00
6.40803337e-01 2.72547722e-01 6.56257629e-01 7.72429883e-01
3.23276490e-01 3.13717425e-01 3.50758880e-01 8.97930861e-02
1.11871564e+00 -7.27141917e-01 -2.19245121e-01 -1.23335056e-01
6.35687232e-01 -4.67914790e-01 7.23853290e-01 2.80749679e-01
-8.72481525e-01 -5.43783046e-02 -6.94331288e-01 3.68848890e-01
-7.95605302e-01 -5.34321547e-01 4.06942546e-01 8.52536142e-01
-5.44406950e-01 3.70943695e-01 -6.93639159e-01 -2.75197059e-01
1.10941970e+00 4.77246761e-01 -8.62903297e-01 -3.12010348e-01
-1.45634592e+00 5.62919736e-01 2.03283742e-01 -1.13495243e+00
-1.30868840e+00 -1.05426645e+00 -7.11228549e-01 -1.49181351e-01
1.98030546e-01 -5.97145617e-01 1.59468508e+00 -1.33157682e+00
-1.35920024e+00 1.13255644e+00 9.70335305e-02 -8.98747802e-01
7.50370383e-01 -8.03207979e-03 -9.83602285e-01 1.94771156e-01
2.97077205e-02 1.44349799e-01 9.08266425e-01 -1.48384333e+00
-7.43791699e-01 -7.73506582e-01 -1.14189796e-01 -1.37078270e-01
-4.17172343e-01 3.86843383e-01 -8.09922144e-02 -6.51275337e-01
-5.19578338e-01 -7.79058456e-01 5.99496402e-02 2.49785736e-01
-4.50968772e-01 -1.26448900e-01 1.26384056e+00 -4.73637491e-01
6.09717607e-01 -2.21546650e+00 -2.90193737e-01 3.58937800e-01
3.61089185e-02 5.53456843e-01 -3.49416971e-01 4.04454678e-01
-2.61979222e-01 1.13994390e-01 -2.48320490e-01 -6.60206318e-01
1.58147603e-01 4.03260261e-01 -6.64360523e-01 6.12901747e-01
-5.29594161e-03 4.90949214e-01 -7.25386202e-01 -4.34028357e-01
-2.82769531e-01 3.24284196e-01 -4.87428606e-01 6.97471857e-01
-5.28212428e-01 5.17300546e-01 -6.55053675e-01 6.77922964e-01
1.11418676e+00 -2.39576846e-01 3.11067551e-01 2.83975065e-01
5.14314830e-01 3.06713939e-01 -1.13033450e+00 1.75569963e+00
-4.52896506e-01 2.90077478e-02 5.51099181e-01 -8.93044829e-01
8.68687391e-01 5.26070416e-01 5.19522965e-01 -5.94133317e-01
7.55861774e-02 2.10157663e-01 -5.45565903e-01 4.20351103e-02
2.16685086e-02 -1.16428092e-01 -4.17898118e-01 1.01177418e+00
3.08072120e-01 5.56622744e-01 -4.99262422e-01 3.94330442e-01
1.32291067e+00 -2.72324443e-01 3.20871413e-01 6.10717461e-02
4.91899461e-01 -5.37347868e-02 5.78192890e-01 8.25904608e-01
-6.15276575e-01 3.57246041e-01 4.30622280e-01 -2.81648993e-01
-9.50737894e-01 -1.52140415e+00 2.59353966e-01 1.65884495e+00
1.47860855e-01 2.88092107e-01 -6.88290238e-01 -1.61017001e+00
7.18038857e-01 8.21769178e-01 -6.99684680e-01 -4.04151440e-01
-4.48645413e-01 -1.27975389e-01 1.17157400e+00 3.12033772e-01
4.84193921e-01 -8.80876958e-01 1.00160781e-02 2.94371933e-01
5.57044744e-02 -1.02761221e+00 -6.27403200e-01 4.53207850e-01
-4.82478529e-01 -1.16755140e+00 -6.41735375e-01 -6.24291956e-01
5.25264204e-01 9.28303599e-02 1.18264389e+00 -6.81116104e-01
2.92156637e-01 4.96468186e-01 -2.13561475e-01 -4.32757527e-01
-1.03821623e+00 2.97988921e-01 2.77552575e-01 1.82516128e-01
7.39833176e-01 -7.23104239e-01 -4.40403998e-01 5.95669091e-01
-1.08461928e+00 -7.57486045e-01 2.04676986e-01 8.60556364e-01
4.46542978e-01 -1.31671429e-01 9.08310771e-01 -1.51349902e+00
6.84023499e-01 -1.21803689e+00 -5.77927828e-01 4.53177929e-01
-5.03713787e-01 1.68721437e-01 1.25412226e+00 -9.94386852e-01
-1.38074183e+00 -4.59442474e-02 3.55328739e-01 -7.45601416e-01
-6.85164869e-01 -2.25441605e-01 -6.84752166e-01 -2.03964800e-01
1.25803030e+00 1.65322095e-01 2.40488797e-01 -7.40404904e-01
6.29899383e-01 1.04621005e+00 8.52134347e-01 -1.05305052e+00
1.10944140e+00 5.68358600e-01 -4.49411839e-01 5.30217774e-02
-3.01963121e-01 -4.61659014e-01 -2.58121073e-01 5.15837550e-01
3.46513599e-01 -9.90384817e-01 -5.24665713e-01 7.07854867e-01
-1.19373620e+00 -5.31935036e-01 -4.86307979e-01 1.47597075e-01
-5.99220753e-01 1.69039801e-01 -4.14569914e-01 -4.19972420e-01
-6.05043352e-01 -7.23454595e-01 6.55112982e-01 9.28514302e-02
-7.51071572e-02 -1.26728261e+00 4.47296411e-01 3.20176512e-01
6.88782275e-01 3.03833932e-01 9.07183170e-01 -1.79860198e+00
-1.13514401e-01 -3.04228365e-01 6.57286048e-02 5.03192246e-01
2.63016313e-01 -4.85118598e-01 -1.08184683e+00 -5.51319003e-01
-7.71290511e-02 -7.62140930e-01 2.60387748e-01 -3.00786287e-01
1.14317870e+00 -9.00591493e-01 -4.82537776e-01 8.13668847e-01
1.36428726e+00 6.59728497e-02 2.81464994e-01 3.85641783e-01
2.65590221e-01 2.06317917e-01 4.16476876e-01 7.43355691e-01
2.39827871e-01 5.24227977e-01 4.41761762e-01 6.16568998e-02
9.67346579e-02 -4.80843276e-01 3.27963144e-01 -2.69663960e-01
7.10093796e-01 -3.81981730e-01 -7.71681070e-01 7.00272560e-01
-1.49378622e+00 -8.99041414e-01 5.58645785e-01 2.30463958e+00
1.26197982e+00 -1.19264036e-01 4.98943090e-01 -3.38275135e-01
8.59341204e-01 -1.11531600e-01 -1.14757884e+00 -5.36986053e-01
-2.19493195e-01 4.28263962e-01 1.00577056e+00 8.33964422e-02
-1.16646504e+00 7.52764344e-01 5.19652367e+00 6.66613042e-01
-1.17963123e+00 6.26587570e-01 6.47545218e-01 -1.76706478e-01
-3.01699668e-01 -3.12536329e-01 -6.88833535e-01 7.96038151e-01
1.21554339e+00 -6.65873528e-01 7.87571549e-01 1.46965742e+00
-6.19129717e-01 6.50685430e-01 -1.28374541e+00 6.79407656e-01
-3.38008851e-01 -1.32880664e+00 2.57306147e-05 1.90776438e-01
9.07619953e-01 5.70148289e-01 3.11529785e-01 5.30901849e-01
1.30717325e+00 -7.47190237e-01 3.21033508e-01 -1.16686579e-02
1.28084517e+00 -8.83817911e-01 4.92384434e-01 5.23364842e-01
-6.87813580e-01 -2.78423846e-01 -2.13148713e-01 7.14759350e-01
-2.78795660e-01 1.00722998e-01 -9.07378316e-01 6.39092267e-01
5.95727324e-01 2.95792092e-02 -1.98350295e-01 5.82298219e-01
1.97171032e-01 8.14435184e-01 -5.91122389e-01 5.06035984e-01
-5.19059636e-02 4.20306176e-01 4.80691612e-01 1.18468904e+00
1.03475407e-01 -6.88033700e-02 9.56753865e-02 7.84446955e-01
-8.25613201e-01 -1.03726491e-01 -9.20976400e-01 1.33038566e-01
1.06204247e+00 8.35608780e-01 3.19500357e-01 -3.02565753e-01
-5.52098334e-01 1.25477338e+00 4.33533967e-01 6.71101451e-01
-6.43990874e-01 -2.11114183e-01 1.53571284e+00 6.48814067e-02
4.16828483e-01 2.03413218e-01 -1.86385036e-01 -1.07801676e+00
-1.47027105e-01 -1.42061126e+00 1.18832409e+00 -9.70586017e-02
-2.21561861e+00 4.93617684e-01 -1.42164677e-01 -1.16460884e+00
-5.30645609e-01 -2.26550341e-01 -7.37882018e-01 1.25328565e+00
-1.74311209e+00 -1.45476627e+00 -3.15017737e-02 1.64944530e+00
4.72579291e-03 -8.52103770e-01 1.37958717e+00 2.87342131e-01
-2.89974242e-01 1.52117538e+00 7.16461599e-01 6.39890254e-01
1.28594947e+00 -1.06033826e+00 7.06702650e-01 6.15691304e-01
-1.43211424e-01 5.99324524e-01 3.24920177e-01 -5.70967257e-01
-1.34230614e+00 -1.62826645e+00 4.67743278e-01 -5.74082494e-01
5.99949241e-01 -4.68689531e-01 -1.23410225e+00 9.56613600e-01
8.46081041e-03 8.70282888e-01 1.05120659e+00 -1.96960747e-01
-1.35251081e+00 -3.85377258e-01 -2.22111678e+00 1.07477993e-01
6.93671823e-01 -7.97349870e-01 -3.67928565e-01 3.37714225e-01
9.71648872e-01 -2.67237723e-01 -1.07137513e+00 -1.30729273e-01
2.61195749e-01 -6.74246132e-01 1.03688776e+00 -1.40925527e+00
-4.48820218e-02 1.32730454e-01 -3.16772074e-01 -1.37507677e+00
-8.49535093e-02 -9.75925982e-01 -4.75143582e-01 1.64425433e+00
1.76942497e-01 -1.24034739e+00 1.07046652e+00 8.32434416e-01
5.14514208e-01 -7.96277225e-02 -1.32346642e+00 -1.16528416e+00
8.51590931e-01 -1.21500157e-01 1.27831364e+00 1.33442879e+00
-4.68603112e-02 -1.96184903e-01 -1.06180251e-01 5.10077953e-01
1.08444834e+00 -5.05320635e-03 1.19358361e+00 -1.16466939e+00
-4.02830571e-01 -6.97303712e-02 -9.78331789e-02 -6.57631993e-01
6.53252363e-01 -7.87246585e-01 -4.00010765e-01 -6.15753889e-01
-1.99097455e-01 -8.29528928e-01 -6.24118924e-01 7.36745179e-01
1.43978521e-01 -2.81753659e-01 3.31614912e-02 3.34639877e-01
-5.72746933e-01 3.21979314e-01 6.14142120e-01 -2.62868434e-01
-8.39925855e-02 4.37542558e-01 -1.11071408e+00 1.23470828e-01
1.16007590e+00 -8.57148588e-01 -4.38497424e-01 -3.12196642e-01
-6.43455505e-01 2.45917290e-01 3.91226590e-01 -7.82727182e-01
2.91921258e-01 -5.05107522e-01 2.99926758e-01 -5.47342040e-02
-6.04723720e-03 -1.22303951e+00 6.59108311e-02 1.04298353e-01
-4.96181816e-01 -3.04199845e-01 2.12385982e-01 6.72281384e-01
-2.55727377e-02 1.61601558e-01 1.14742792e+00 -1.80616584e-02
-4.74930137e-01 6.55214667e-01 2.03650028e-01 5.26018262e-01
1.28643882e+00 1.08709671e-01 -8.32323134e-01 -4.04283285e-01
-4.38832611e-01 2.29161218e-01 8.61667097e-01 4.34118181e-01
2.32851237e-01 -1.33234763e+00 -8.18126798e-01 5.64040065e-01
4.75251287e-01 6.96873888e-02 1.23290032e-01 -4.75261286e-02
9.53633487e-02 1.30380571e-01 -2.64054179e-01 -2.32713625e-01
-1.31917500e+00 9.49700296e-01 5.25345564e-01 -4.65810627e-01
-2.97823489e-01 8.32041144e-01 1.40349641e-01 -9.38595116e-01
4.58374232e-01 3.29094738e-01 4.58048642e-01 -3.61961991e-01
7.02687681e-01 3.46672684e-01 1.15024880e-01 -4.81052727e-01
-6.36653066e-01 -2.63029605e-01 -4.27592993e-01 -1.81929171e-01
1.04956234e+00 1.48767129e-01 3.02189827e-01 -1.99561447e-01
1.65201426e+00 -2.17769798e-02 -1.50730741e+00 -1.03713953e+00
-2.65072197e-01 -5.84097326e-01 -4.73893493e-01 -1.36352587e+00
-1.16431820e+00 6.75051510e-01 7.49603391e-01 -5.98168224e-02
1.05879748e+00 7.96748176e-02 8.99432182e-01 2.82883465e-01
7.35039830e-01 -8.36344838e-01 -2.61166822e-02 2.70272642e-01
5.91270685e-01 -1.42209733e+00 -3.41532975e-01 3.84930857e-02
-9.19037640e-01 7.56766915e-01 7.64326394e-01 1.34769678e-01
6.65226877e-01 4.43279713e-01 4.15637523e-01 3.74671072e-01
-8.75255227e-01 4.37010467e-01 -2.70360947e-01 1.46232009e+00
-3.31608355e-01 1.59042776e-01 5.99354863e-01 1.13952255e+00
1.51922420e-01 6.14033528e-02 1.75360627e-02 1.01302338e+00
5.75415716e-02 -1.55351663e+00 -6.42524362e-01 5.92256635e-02
-8.50043178e-01 2.89555758e-01 -5.49247026e-01 5.74917078e-01
5.34764826e-02 9.36581612e-01 -6.79203346e-02 -2.08767742e-01
2.92404145e-01 4.45958406e-01 -5.02394401e-02 -4.58136171e-01
-9.11446691e-01 -6.55343771e-01 -1.50956914e-01 -5.69983006e-01
1.86359823e-01 -5.89979529e-01 -1.26107788e+00 -7.34061360e-01
1.16668403e-01 3.25234294e-01 7.57773697e-01 5.34030259e-01
8.26224625e-01 -1.26890168e-01 1.13216770e+00 -2.00028092e-01
-1.49949920e+00 -6.94733441e-01 -6.32661164e-01 9.73066866e-01
5.01598775e-01 -3.65901828e-01 -5.30827403e-01 1.01465292e-01] | [10.355576515197754, 3.169701099395752] |
ae2e96f3-3a72-42f1-a1ba-9d25c6aeaf5d | learning-dynamic-preference-structure | 2111.11886 | null | https://arxiv.org/abs/2111.11886v1 | https://arxiv.org/pdf/2111.11886v1.pdf | Learning Dynamic Preference Structure Embedding From Temporal Networks | The dynamics of temporal networks lie in the continuous interactions between nodes, which exhibit the dynamic node preferences with time elapsing. The challenges of mining temporal networks are thus two-fold: the dynamic structure of networks and the dynamic node preferences. In this paper, we investigate the dynamic graph sampling problem, aiming to capture the preference structure of nodes dynamically in cooperation with GNNs. Our proposed Dynamic Preference Structure (DPS) framework consists of two stages: structure sampling and graph fusion. In the first stage, two parameterized samplers are designed to learn the preference structure adaptively with network reconstruction tasks. In the second stage, an additional attention layer is designed to fuse two sampled temporal subgraphs of a node, generating temporal node embeddings for downstream tasks. Experimental results on many real-life temporal networks show that our DPS outperforms several state-of-the-art methods substantially owing to learning an adaptive preference structure. The code will be released soon at https://github.com/doujiang-zheng/Dynamic-Preference-Structure. | ['Hao Xu', 'Chun Chen', 'Xinyu Wang', 'Xingen Wang', 'Mingli Song', 'Chengchao Shen', 'Yu Wang', 'Zunlei Feng', 'Tongya Zheng'] | 2021-11-23 | null | null | null | null | ['graph-sampling'] | ['graphs'] | [-3.66853811e-02 3.11666071e-01 -4.09075946e-01 -2.66911149e-01
2.37082206e-02 -4.58620489e-01 7.08902299e-01 -9.25724953e-02
-1.26945019e-01 6.41576469e-01 2.84212261e-01 -3.04657280e-01
-7.54151344e-01 -7.90286005e-01 -3.47773761e-01 -9.72106040e-01
-5.61202645e-01 6.93485796e-01 4.11705494e-01 -1.49642065e-01
-1.39870301e-01 2.80053496e-01 -9.50874090e-01 -1.44842044e-01
7.85785854e-01 8.06048214e-01 4.72098976e-01 6.19470537e-01
9.52583924e-03 6.24206066e-01 2.40407243e-01 -4.75363433e-01
2.44610712e-01 -1.90584913e-01 -8.33012462e-01 1.93676054e-01
-4.62138653e-01 7.24021951e-03 -8.65130723e-01 1.02526903e+00
4.87460881e-01 2.86193341e-01 5.85225046e-01 -1.49991143e+00
-5.00582516e-01 1.37312353e+00 -5.26943743e-01 5.13750732e-01
-9.01096221e-03 3.97296041e-01 1.44245172e+00 -6.08774841e-01
5.98996520e-01 1.40596390e+00 3.90605390e-01 6.22769594e-01
-1.34885669e+00 -7.68209398e-01 9.56728816e-01 2.95763612e-01
-1.30740857e+00 -3.36306393e-01 1.17341268e+00 -3.15627098e-01
4.77912515e-01 -3.83224413e-02 1.17425263e+00 1.53068316e+00
-9.33450237e-02 7.63211727e-01 5.51059961e-01 2.75627196e-01
1.55710891e-01 -2.05049306e-01 1.12989686e-01 7.08663642e-01
4.93121892e-02 1.11864910e-01 -5.39925933e-01 -3.77174616e-01
8.48194003e-01 3.78386706e-01 -2.88326174e-01 -4.97743189e-01
-8.89361858e-01 7.06677914e-01 5.60430586e-01 3.02704245e-01
-4.22291875e-01 3.12142313e-01 2.53498793e-01 5.18549383e-01
6.56099737e-01 -1.01632908e-01 -6.53219402e-01 -8.73628780e-02
-6.39118195e-01 4.64473478e-02 7.39255011e-01 1.14140272e+00
6.79402947e-01 -2.49464959e-01 -2.06634045e-01 6.95151210e-01
6.07954502e-01 -9.10747424e-02 3.04783463e-01 -7.89065838e-01
5.05492747e-01 6.42932057e-01 -1.78010464e-01 -9.77060914e-01
-4.65983868e-01 -7.35643983e-01 -1.29522645e+00 -6.26581609e-01
3.60368341e-01 -4.67565298e-01 -6.23376727e-01 2.07805204e+00
5.80736518e-01 5.85564554e-01 -3.13529938e-01 6.02199733e-01
6.49931431e-01 5.62497497e-01 6.80546835e-02 -4.77894723e-01
1.03885818e+00 -9.60782588e-01 -5.14251471e-01 -1.51584953e-01
2.52596587e-01 -4.58936468e-02 9.46170986e-01 5.44008501e-02
-9.88284171e-01 -1.89965054e-01 -6.66789114e-01 2.49792501e-01
3.56640108e-02 -1.86309487e-01 6.45544112e-01 1.79604605e-01
-1.28165507e+00 9.73516643e-01 -8.81567657e-01 -6.46309972e-01
5.34816206e-01 5.02667904e-01 3.37842926e-02 1.20334640e-01
-1.33556402e+00 2.11296719e-03 3.58680099e-01 3.12529296e-01
-1.08529544e+00 -9.34499443e-01 -6.55690432e-01 3.04012060e-01
6.97809041e-01 -8.17494690e-01 1.22354376e+00 -6.78946197e-01
-1.41286290e+00 4.92893815e-01 -2.46793613e-01 -3.54472578e-01
7.96575248e-01 2.66481370e-01 -5.67966461e-01 -1.93943959e-02
1.94401406e-02 2.56975859e-01 8.43778491e-01 -1.03441656e+00
-5.05400181e-01 -2.78505445e-01 1.41267836e-01 2.05261737e-01
-6.92787468e-01 -3.68018776e-01 -9.63461101e-01 -6.58892453e-01
-1.41663877e-02 -1.04098165e+00 -7.26593614e-01 2.22517736e-03
-4.45583552e-01 -5.33050537e-01 7.90425360e-01 -5.95431868e-03
1.61259687e+00 -1.92360175e+00 5.32587051e-01 5.61995029e-01
5.93950748e-01 -1.83468699e-01 -3.33258241e-01 6.68856680e-01
4.80467454e-02 2.44742900e-01 -9.77601856e-02 -4.92098778e-01
-9.62429643e-02 3.13859612e-01 -9.56865847e-02 3.05948496e-01
1.72771383e-02 9.22312796e-01 -1.44614327e+00 -5.01448154e-01
-1.92442611e-01 3.16474527e-01 -3.86062473e-01 1.33793145e-01
-2.79146254e-01 5.34931779e-01 -8.99385452e-01 5.12942255e-01
6.01099908e-01 -7.37632155e-01 9.37277555e-01 -7.44474083e-02
1.94589660e-01 3.01623434e-01 -1.07275987e+00 1.51959753e+00
-9.63966921e-02 1.10613368e-01 4.04284447e-01 -8.29803348e-01
6.63950264e-01 3.87227654e-01 7.24767327e-01 -2.51753300e-01
2.92363465e-01 2.51144171e-02 3.08478743e-01 -3.57448339e-01
8.09553936e-02 3.08965594e-01 7.54441097e-02 6.83834493e-01
1.72662377e-01 1.96454078e-01 4.66373920e-01 6.29895568e-01
1.35514069e+00 -4.07332063e-01 9.13545117e-02 -6.63301587e-01
5.20008326e-01 -5.50908804e-01 1.08922875e+00 5.84011793e-01
-5.27821958e-01 1.51748985e-01 1.03240705e+00 -4.68389273e-01
-7.32827187e-01 -1.00749993e+00 3.70827079e-01 9.54358280e-01
2.30764315e-01 -6.48370385e-01 -4.31457311e-01 -9.50356066e-01
2.09877267e-02 1.63501978e-01 -9.37494934e-01 -3.32493275e-01
-4.13181692e-01 -7.68649161e-01 5.59004880e-02 3.85897338e-01
2.66581327e-01 -1.09206915e+00 1.72700416e-02 2.50944406e-01
-6.21006042e-02 -8.06236327e-01 -1.15958846e+00 8.01020041e-02
-1.13426602e+00 -1.23661566e+00 -5.03766775e-01 -7.04985142e-01
7.77500272e-01 3.17148209e-01 1.14934075e+00 2.07688138e-01
5.99789657e-02 4.67845678e-01 -3.40973020e-01 2.45621931e-02
1.82540387e-01 6.67051435e-01 2.00124189e-01 3.63492996e-01
1.04204133e-01 -1.45173264e+00 -8.52557659e-01 4.23084885e-01
-8.02094281e-01 -7.89393559e-02 6.74230754e-01 8.53876889e-01
5.09841621e-01 3.70360970e-01 3.36228818e-01 -1.35349309e+00
8.18545401e-01 -1.10049903e+00 -5.03527761e-01 2.77648002e-01
-8.25113654e-01 2.33763605e-01 6.25556409e-01 -7.85915911e-01
-9.28614557e-01 1.56230867e-01 1.58313766e-01 -6.27936065e-01
4.03328508e-01 6.88609362e-01 -2.68628716e-01 2.57639199e-01
2.17739224e-01 6.79081678e-02 6.44133389e-02 -3.47825557e-01
2.90299118e-01 2.29198039e-02 4.80716955e-03 -7.33121514e-01
1.10947442e+00 5.00195742e-01 1.52452616e-02 -6.14375830e-01
-8.00318718e-01 -4.61777180e-01 -5.76197624e-01 -5.04782379e-01
1.97267100e-01 -7.01866090e-01 -6.48745418e-01 5.36498010e-01
-8.09761286e-01 -7.10165203e-01 -2.04917803e-01 1.42752513e-01
-3.94300491e-01 2.79185951e-01 -8.34202647e-01 -8.38787854e-01
-2.96661377e-01 -8.07616889e-01 5.35551429e-01 3.61108631e-01
-8.50487649e-02 -1.31284106e+00 7.85907879e-02 -8.88738781e-02
3.19691956e-01 -4.13025990e-02 6.13263011e-01 -4.63929564e-01
-7.20148146e-01 2.47474983e-01 -7.74974450e-02 -4.65576231e-01
3.80244464e-01 2.74558663e-01 -4.99305546e-01 -4.75062519e-01
-3.54108661e-01 -1.64395973e-01 9.79533911e-01 5.30632794e-01
1.36213231e+00 -5.04548013e-01 -6.95365071e-01 7.40150571e-01
1.08049357e+00 -2.67792158e-02 8.42727795e-02 -2.86928833e-01
5.71375489e-01 7.39419162e-01 5.84282637e-01 7.82757342e-01
5.22714496e-01 2.85298556e-01 7.27951348e-01 4.65807945e-01
2.97751576e-01 -7.11322784e-01 2.27081180e-01 8.26585591e-01
-7.56650195e-02 -5.62436640e-01 -7.66351104e-01 8.28906596e-01
-2.36373687e+00 -1.05396938e+00 7.97459036e-02 1.72026229e+00
7.86904275e-01 3.00684273e-01 5.53855002e-01 -1.08352862e-01
8.46909881e-01 5.88306546e-01 -9.69047487e-01 2.82745421e-01
9.13953185e-02 -5.10318935e-01 4.22838986e-01 5.19113898e-01
-8.92163515e-01 9.44536209e-01 4.90524673e+00 8.73374820e-01
-7.47643054e-01 4.14792858e-02 6.74081206e-01 -1.69683233e-01
-7.08302855e-01 1.44452721e-01 -6.27475262e-01 6.03580475e-01
8.26398075e-01 -5.48917055e-01 7.54721403e-01 5.88788450e-01
4.66650039e-01 6.10205472e-01 -7.83452749e-01 7.47642815e-01
-6.38729453e-01 -1.28649044e+00 -1.21053359e-04 1.50056079e-01
7.00842559e-01 2.59716630e-01 7.70351961e-02 4.27996889e-02
9.59259808e-01 -7.18453169e-01 5.71191669e-01 6.55824780e-01
6.33019030e-01 -7.19555795e-01 2.91241825e-01 5.00354588e-01
-1.92132080e+00 -4.30768788e-01 -3.03057551e-01 1.49939701e-01
3.00091654e-01 9.11632955e-01 -5.51061034e-01 6.34533465e-01
8.96335244e-01 1.47790754e+00 -3.37231040e-01 8.36664557e-01
-2.33238831e-01 6.36875749e-01 -3.70894223e-01 -3.00960749e-01
2.49899358e-01 -4.90662336e-01 8.54145288e-01 8.59362423e-01
3.24956536e-01 1.84622169e-01 8.95992219e-02 9.60900426e-01
-2.62328148e-01 -3.05953830e-01 -6.67219937e-01 -4.35056895e-01
9.99224186e-01 1.59700608e+00 -7.20462441e-01 9.63701159e-02
-1.56175181e-01 8.26644421e-01 7.20918655e-01 5.02011478e-01
-8.30186605e-01 -7.26780668e-02 9.16106582e-01 3.73820871e-01
4.56545323e-01 -2.49791816e-01 3.22346538e-01 -1.00701416e+00
-6.82481676e-02 -3.68998021e-01 8.70899677e-01 -2.13428989e-01
-1.69943738e+00 7.06435502e-01 -5.33317141e-02 -1.12754178e+00
1.20953776e-01 -2.99801141e-01 -1.04104590e+00 5.71892262e-01
-1.06153655e+00 -1.17225552e+00 -2.42085248e-01 7.77810633e-01
4.90617037e-01 6.35804934e-03 2.43072867e-01 2.07140490e-01
-7.76837409e-01 5.68261087e-01 -8.38839337e-02 1.19864412e-01
2.21231773e-01 -1.20681131e+00 2.57232815e-01 8.29891324e-01
-4.16776612e-02 6.66642129e-01 6.19421363e-01 -7.28901625e-01
-1.31580222e+00 -1.46949863e+00 7.10572004e-01 4.57094386e-02
1.16044497e+00 -5.25130510e-01 -5.75176716e-01 7.64623702e-01
9.45712700e-02 3.26358497e-01 5.86025894e-01 2.04050422e-01
-2.26910844e-01 -4.33098704e-01 -9.37303603e-01 8.41293454e-01
1.81456029e+00 -3.57731313e-01 1.81726784e-01 3.08933288e-01
1.02624130e+00 -1.77698601e-02 -7.73235321e-01 3.74763548e-01
4.39690828e-01 -7.01297283e-01 8.52722764e-01 -6.25779450e-01
1.00261167e-01 -2.23237500e-01 3.59953523e-01 -1.45351899e+00
-7.64685869e-01 -1.31713748e+00 -6.01283848e-01 1.50481176e+00
6.24732852e-01 -7.91392386e-01 1.06720877e+00 3.75633806e-01
2.10757434e-01 -1.02029753e+00 -8.50028932e-01 -5.80304265e-01
-2.21730530e-01 -1.91188693e-01 7.31912315e-01 8.17523479e-01
1.94442924e-02 6.45205796e-01 -4.76219505e-01 1.71356708e-01
7.45962977e-01 1.35432154e-01 3.92672896e-01 -1.52165067e+00
-5.43051124e-01 -7.10327446e-01 1.99324548e-01 -1.34638214e+00
2.47762561e-01 -8.69564891e-01 -1.77368239e-01 -1.28213596e+00
2.60229856e-01 -5.04054844e-01 -5.51791131e-01 1.08561188e-01
-2.28127703e-01 -5.81270933e-01 -2.02365279e-01 1.34368137e-01
-9.12049830e-01 1.08198118e+00 1.26900065e+00 -2.80432910e-01
-3.93836021e-01 4.95103329e-01 -9.84132826e-01 4.79265362e-01
1.16078889e+00 -6.00945830e-01 -1.15714657e+00 -2.48961031e-01
2.69377500e-01 2.53493100e-01 1.17272794e-01 -5.30532837e-01
5.66527069e-01 -3.76557767e-01 -2.20720559e-01 -7.09472358e-01
2.36793801e-01 -9.61872637e-01 4.71304715e-01 6.59051120e-01
-3.18165779e-01 2.04714015e-01 -3.47582102e-01 1.35991418e+00
2.12439656e-01 1.95779309e-01 6.42203987e-01 -2.16474101e-01
-4.63440478e-01 1.38403749e+00 -3.71899366e-01 2.33702093e-01
9.40987706e-01 8.00700635e-02 -1.39095992e-01 -5.57947874e-01
-9.42939818e-01 1.18895113e+00 2.19596609e-01 4.94498223e-01
6.43849790e-01 -1.36404204e+00 -5.21215916e-01 -3.73799503e-02
-5.46396859e-02 3.64154041e-01 4.58364338e-01 1.03151870e+00
7.26466104e-02 -2.09473632e-02 1.36573568e-01 -2.98765391e-01
-1.14849854e+00 6.66355491e-01 3.97293389e-01 -7.93830037e-01
-4.61683959e-01 1.12321830e+00 -7.76030272e-02 -5.79058409e-01
3.94113630e-01 -3.32251750e-02 -4.64490086e-01 2.20350027e-01
1.26860484e-01 3.87949437e-01 -6.27796113e-01 -1.87891573e-01
-3.72359902e-01 2.02246547e-01 -1.66640386e-01 6.16814848e-03
1.62696242e+00 -4.07745332e-01 -3.27510595e-01 5.86901426e-01
9.41041946e-01 -2.35464156e-01 -1.49660230e+00 -7.33670950e-01
1.52556434e-01 -3.27700704e-01 -3.05782109e-01 -8.54229107e-02
-1.60772526e+00 4.68662143e-01 -1.76950637e-02 5.47986746e-01
1.24725008e+00 2.95201629e-01 6.29776478e-01 1.66424364e-01
3.28345627e-01 -9.52140570e-01 1.25780806e-01 7.07375884e-01
7.24827945e-01 -8.49942029e-01 -2.59225518e-01 -6.31599188e-01
-5.52761853e-01 9.05144453e-01 8.84014010e-01 -8.29291716e-02
1.58464277e+00 2.20162291e-02 -5.25048196e-01 -4.71458048e-01
-1.47986197e+00 -1.92039311e-01 6.65570274e-02 4.44009185e-01
2.24172786e-01 7.42709637e-02 -3.25692177e-01 9.79714751e-01
7.37626553e-02 -2.97288567e-01 2.02230647e-01 4.83448565e-01
1.56009814e-03 -1.43093121e+00 5.21873474e-01 7.35656261e-01
-1.37705848e-01 -7.49130547e-02 -7.46913195e-01 4.79762942e-01
-1.53746814e-01 9.38670218e-01 -6.31583333e-02 -7.66741574e-01
1.40688971e-01 -3.22730809e-01 1.03717014e-01 -6.06390834e-01
-3.82543325e-01 -2.52726395e-03 7.74278492e-02 -7.68901944e-01
-3.92568499e-01 -1.06609106e+00 -9.24627066e-01 -6.49137974e-01
-8.86007845e-02 4.89658386e-01 -3.21627349e-01 4.18640286e-01
5.82350075e-01 5.94507873e-01 8.62583280e-01 -8.39270055e-01
-5.52440524e-01 -9.76276875e-01 -7.96273172e-01 2.34489352e-01
4.20089155e-01 -7.16879845e-01 -5.67004025e-01 -4.36408967e-01] | [7.286764144897461, 5.990217208862305] |
52ad832f-eb6b-4d62-aa13-a88438a8bed9 | vampnet-music-generation-via-masked-acoustic | 2307.04686 | null | https://arxiv.org/abs/2307.04686v1 | https://arxiv.org/pdf/2307.04686v1.pdf | VampNet: Music Generation via Masked Acoustic Token Modeling | We introduce VampNet, a masked acoustic token modeling approach to music synthesis, compression, inpainting, and variation. We use a variable masking schedule during training which allows us to sample coherent music from the model by applying a variety of masking approaches (called prompts) during inference. VampNet is non-autoregressive, leveraging a bidirectional transformer architecture that attends to all tokens in a forward pass. With just 36 sampling passes, VampNet can generate coherent high-fidelity musical waveforms. We show that by prompting VampNet in various ways, we can apply it to tasks like music compression, inpainting, outpainting, continuation, and looping with variation (vamping). Appropriately prompted, VampNet is capable of maintaining style, genre, instrumentation, and other high-level aspects of the music. This flexible prompting capability makes VampNet a powerful music co-creation tool. Code and audio samples are available online. | ['Bryan Pardo', 'Rithesh Kumar', 'Prem Seetharaman', 'Hugo Flores Garcia'] | 2023-07-10 | null | null | null | null | ['music-compression', 'music-generation', 'music-generation'] | ['audio', 'audio', 'music'] | [ 2.15405107e-01 -1.79195121e-01 -3.44350599e-02 1.21637680e-01
-1.17945921e+00 -7.96639383e-01 4.72286135e-01 -4.12482798e-01
6.96833879e-02 3.90764862e-01 6.51835859e-01 -1.73256606e-01
-2.39113607e-02 -4.69549209e-01 -7.65547156e-01 -5.31489909e-01
-2.06062853e-01 2.41732731e-01 -2.10704476e-01 -2.04649493e-01
2.93527097e-02 2.01045066e-01 -1.49524295e+00 7.26756513e-01
2.70301700e-01 6.15270972e-01 3.01363468e-01 1.31044996e+00
-4.06107977e-02 1.06413209e+00 -9.96712327e-01 -2.98450887e-01
2.04057276e-01 -7.81695187e-01 -3.95712048e-01 -2.69504368e-01
2.76013672e-01 -3.92582893e-01 -1.32848904e-01 6.68523431e-01
7.20723391e-01 2.53499091e-01 4.90221262e-01 -1.15392685e+00
-2.97636896e-01 1.48627412e+00 -2.79006988e-01 -1.20102599e-01
5.03716886e-01 2.33858079e-01 1.29546356e+00 -1.09640539e+00
4.24815416e-01 1.15752757e+00 1.05092347e+00 4.25713629e-01
-1.67288041e+00 -1.01414073e+00 -3.65604609e-01 2.88498010e-02
-1.34636128e+00 -9.63634253e-01 1.05659604e+00 -2.91563213e-01
1.00371301e+00 6.84373617e-01 9.23062861e-01 1.30675793e+00
-6.16429597e-02 8.35532784e-01 5.88204980e-01 -6.33666754e-01
1.69404909e-01 -4.09232974e-01 -5.50291181e-01 2.77674645e-01
-8.22861314e-01 5.39114416e-01 -1.29088223e+00 -4.72400457e-01
1.00279164e+00 -5.57063580e-01 -1.93944365e-01 3.40517640e-01
-1.29984224e+00 4.08269823e-01 -5.72634265e-02 1.37735069e-01
-4.95887190e-01 8.54739249e-01 5.56209862e-01 7.04021692e-01
1.18217871e-01 9.23117459e-01 -4.33156013e-01 -6.40734315e-01
-1.51240742e+00 5.76954544e-01 6.50666535e-01 9.28318501e-01
3.77566010e-01 9.82278943e-01 -4.87184763e-01 1.07673824e+00
5.46547631e-03 4.25496787e-01 6.16156340e-01 -1.35415018e+00
1.84816718e-01 -5.94561815e-01 -9.73281115e-02 -7.19330907e-01
-9.24229473e-02 -6.17316604e-01 -7.43386328e-01 3.01108956e-01
-1.01137571e-01 -3.37257802e-01 -7.82677114e-01 1.97811413e+00
-7.12433606e-02 8.28356326e-01 -4.92855385e-02 4.40770298e-01
7.68211663e-01 8.89227152e-01 -1.97814941e-01 -6.17819168e-02
1.21564996e+00 -9.44968700e-01 -8.82790864e-01 9.99801513e-03
1.16920397e-01 -1.33022845e+00 1.26158452e+00 9.34323549e-01
-1.63778329e+00 -8.66062939e-01 -1.14160800e+00 -2.80034631e-01
3.88555944e-01 -2.21786299e-03 5.16619802e-01 4.51099485e-01
-1.19773877e+00 1.08048785e+00 -8.52580488e-01 5.66838920e-01
3.93274203e-02 1.83730006e-01 -6.91703409e-02 3.77948344e-01
-9.59502637e-01 3.83998722e-01 1.17754355e-01 -1.19835667e-01
-1.14789283e+00 -1.19626212e+00 -8.30749512e-01 2.58104056e-01
-8.01782086e-02 -7.47409403e-01 1.93552208e+00 -8.98804188e-01
-1.98655140e+00 2.05447122e-01 -2.60955662e-01 -6.63035035e-01
1.69118494e-01 -4.11845118e-01 -6.75964117e-01 4.77047302e-02
-3.94726992e-02 6.51973844e-01 1.41457176e+00 -8.99231493e-01
-2.63541669e-01 4.00534481e-01 -5.16449153e-01 1.54994309e-01
1.20699108e-01 2.28076264e-01 -3.25943172e-01 -1.34071636e+00
4.11110111e-02 -7.25425661e-01 -6.33995906e-02 -3.25417757e-01
-6.39722288e-01 3.40183228e-01 8.21637630e-01 -7.42066681e-01
1.25036955e+00 -2.51038766e+00 1.98001862e-01 2.55002707e-01
-1.54755354e-01 5.25155477e-02 -4.48648959e-01 6.77561164e-01
-2.87457466e-01 8.36004168e-02 -1.53397143e-01 -7.12541819e-01
2.40351230e-01 1.15933165e-01 -9.48423922e-01 -3.95526104e-02
3.33294660e-01 9.46292102e-01 -7.23598480e-01 -2.58377105e-01
2.43945450e-01 7.03245819e-01 -1.06266379e+00 3.60016555e-01
-5.84931195e-01 6.46438956e-01 2.77179331e-01 5.72625101e-01
1.61891043e-01 3.98664922e-01 1.06453896e-03 -1.77607343e-01
-2.17044041e-01 8.48806202e-01 -1.34758663e+00 2.07745242e+00
-8.15174341e-01 8.92071545e-01 2.78214335e-01 -3.57238859e-01
9.40952480e-01 7.90139973e-01 2.98122823e-01 -2.96489924e-01
1.60508212e-02 2.41639450e-01 -8.35333019e-02 -1.46912739e-01
8.36718917e-01 -3.24853688e-01 9.50514451e-02 5.79305232e-01
3.11723381e-01 -7.92930424e-01 -5.31147458e-02 4.01878357e-02
1.13333464e+00 2.15592548e-01 3.00014783e-02 1.98664218e-01
-1.38451740e-01 -3.80198479e-01 2.61648029e-01 8.62564266e-01
4.69537318e-01 7.99079776e-01 3.18015069e-01 1.86327606e-01
-1.28366256e+00 -1.35733199e+00 1.72293603e-01 1.29828000e+00
-7.59006679e-01 -9.05153155e-01 -5.59413135e-01 4.94801849e-01
-1.93311527e-01 9.32444334e-01 -2.00495526e-01 -2.23144189e-01
-7.44346440e-01 -7.41329640e-02 1.12913394e+00 4.77128804e-01
1.07708678e-01 -1.46190941e+00 -2.74928093e-01 6.39687598e-01
-2.37217695e-01 -5.90765655e-01 -9.80937004e-01 5.23333013e-01
-7.56674469e-01 -5.18903553e-01 -4.86061335e-01 -8.35396647e-01
-2.85665959e-01 -1.45072356e-01 1.48188019e+00 -3.53763737e-02
-8.04196969e-02 9.50416028e-02 -2.60184467e-01 -6.12384558e-01
-8.52179766e-01 8.00635666e-02 3.11884806e-02 -1.90584943e-01
-3.02229792e-01 -1.37749159e+00 -1.63442492e-01 -1.12389490e-01
-9.06102598e-01 2.88465023e-01 4.36082929e-01 8.94984245e-01
7.91871369e-01 -2.74011232e-02 5.33887267e-01 -6.34474576e-01
8.55444491e-01 -1.59302562e-01 -3.15053493e-01 -4.16660160e-01
-1.49739474e-01 -6.83208555e-02 8.54795456e-01 -8.70488167e-01
-6.68530464e-01 -1.03269398e-01 -6.14394844e-01 -8.47833633e-01
2.83136457e-01 5.32720447e-01 1.89992618e-02 2.94386119e-01
8.79401028e-01 2.66851306e-01 -7.94270560e-02 -6.13315165e-01
6.73933387e-01 4.61796910e-01 1.17819643e+00 -7.85340130e-01
9.30787504e-01 -1.93987396e-02 -2.61330307e-01 -8.04933488e-01
-5.02059579e-01 1.05042793e-02 -1.09830007e-01 -8.52150172e-02
3.19535047e-01 -1.01734996e+00 -8.85196269e-01 2.86199421e-01
-9.66852069e-01 -7.15276003e-01 -9.74832356e-01 4.23877776e-01
-8.93921912e-01 -2.47833785e-02 -9.84857917e-01 -8.21839750e-01
-5.98906755e-01 -8.52445722e-01 9.60552216e-01 -2.91648179e-01
-8.26201916e-01 -5.20862699e-01 2.49690831e-01 -3.70683670e-01
6.44374847e-01 1.11172900e-01 9.14875805e-01 -9.74438339e-02
-5.22237480e-01 7.80939832e-02 5.67476153e-01 4.46091205e-01
6.03245245e-03 2.49346897e-01 -1.30241740e+00 -7.26819783e-02
1.07878260e-01 -3.04628879e-01 7.71923900e-01 6.53859496e-01
1.24606705e+00 -6.20069504e-01 3.19861293e-01 1.06308234e+00
7.93525279e-01 1.70492932e-01 7.84068644e-01 -3.84267978e-02
4.72581714e-01 -4.19062749e-03 2.31539935e-01 6.42131567e-01
-2.48116463e-01 7.82619774e-01 2.10828349e-01 -1.86088383e-01
-5.67723632e-01 -6.35624468e-01 6.02046013e-01 1.30007184e+00
1.11766785e-01 1.03250436e-01 -2.74619937e-01 4.85767990e-01
-1.35344660e+00 -1.55096376e+00 1.12688623e-01 1.93800211e+00
1.43427014e+00 6.43853396e-02 1.84217602e-01 7.29333401e-01
4.30393070e-01 3.13551456e-01 -4.83506888e-01 -8.11461627e-01
-7.14963526e-02 1.04547155e+00 1.23505138e-01 7.15095222e-01
-7.52303362e-01 8.89808536e-01 7.38055038e+00 1.02727509e+00
-1.06203949e+00 9.58531275e-02 7.01851249e-02 -4.61282939e-01
-7.38880873e-01 -6.86670048e-03 -3.11953545e-01 2.38999486e-01
1.15952063e+00 -1.22472852e-01 1.07781994e+00 5.47227740e-01
4.16701108e-01 5.41680872e-01 -1.27059233e+00 9.73143935e-01
-1.60396174e-01 -1.72160423e+00 1.65491290e-02 -4.77207959e-01
5.65392375e-01 -8.66793543e-02 4.73876387e-01 4.76931632e-01
3.99879873e-01 -1.33654904e+00 1.29716206e+00 4.20152754e-01
1.30757344e+00 -9.81103361e-01 -1.26604423e-01 1.12927228e-01
-1.35034299e+00 -9.07476805e-03 7.13928565e-02 -1.45361498e-01
3.47380430e-01 5.14640987e-01 -1.16674709e+00 3.03440243e-01
3.45779419e-01 6.85927451e-01 1.10253595e-01 1.01652622e+00
-3.59366953e-01 1.18116343e+00 -2.80352980e-01 5.38728058e-01
-1.53480649e-01 3.75961363e-02 6.84155583e-01 1.36881554e+00
4.43477988e-01 -1.07257001e-01 2.75352765e-02 1.07079923e+00
-2.28267401e-01 -8.84601250e-02 -2.49604672e-01 -1.68184519e-01
9.24517155e-01 7.92340696e-01 -7.37385973e-02 -2.78483391e-01
1.77612573e-01 8.94949913e-01 -3.23541313e-01 3.62895221e-01
-7.77703822e-01 -5.17816603e-01 8.33872199e-01 -1.07562207e-01
4.04780716e-01 -4.64156181e-01 -3.58558834e-01 -8.43373179e-01
-2.78015912e-01 -1.34068418e+00 -3.12783197e-02 -1.21054292e+00
-9.66468155e-01 6.64562881e-01 -2.33251452e-01 -1.26010406e+00
-8.48037183e-01 3.29347327e-02 -8.81322265e-01 1.10639000e+00
-1.05088401e+00 -9.38191950e-01 1.81527212e-01 6.50558710e-01
7.72134006e-01 -2.24414602e-01 1.23846889e+00 3.60088855e-01
-7.36757666e-02 5.88272691e-01 -1.53364092e-01 -4.35595587e-02
7.18229413e-01 -1.15302241e+00 9.38432693e-01 7.17768371e-01
8.50321114e-01 6.06199086e-01 1.04505324e+00 -3.86628002e-01
-1.27195871e+00 -1.00148904e+00 5.63421190e-01 -1.17797285e-01
5.87122679e-01 -4.49708730e-01 -7.65620470e-01 9.12131786e-01
3.45424324e-01 -5.45992613e-01 8.02722275e-01 2.55773067e-01
-6.74786508e-01 -1.13644719e-01 -6.36913598e-01 9.27579999e-01
7.52938867e-01 -1.03171277e+00 -5.42697847e-01 1.96157411e-01
1.07502019e+00 -8.61506224e-01 -6.39248371e-01 1.34645021e-02
7.70134270e-01 -9.20280278e-01 1.22452283e+00 -3.60213429e-01
6.46436751e-01 -3.17183882e-01 -2.25210354e-01 -1.43223190e+00
-3.35167587e-01 -1.55616570e+00 -3.16661417e-01 1.47402096e+00
4.40446109e-01 -2.14308619e-01 5.38312197e-01 -9.14915651e-02
-5.61721802e-01 -2.42291778e-01 -6.72230482e-01 -7.26382136e-01
-1.41033098e-01 -1.20463574e+00 8.98286164e-01 7.83951283e-01
-4.44960855e-02 1.26760691e-01 -9.06496763e-01 9.63979736e-02
3.92329067e-01 4.31080572e-02 9.72405374e-01 -7.70521164e-01
-1.28399086e+00 -4.42549288e-01 2.46954590e-01 -1.04550433e+00
-2.59804409e-02 -9.97786641e-01 2.21525609e-01 -8.95615160e-01
-3.74264657e-01 -4.68870461e-01 -1.83201224e-01 6.51516020e-01
2.40718365e-01 4.82501119e-01 5.58281004e-01 2.66407579e-01
2.89522111e-01 5.05437970e-01 1.00603294e+00 -1.61192507e-01
-5.91010094e-01 2.14424551e-01 -6.75408006e-01 5.88241458e-01
7.86099553e-01 -5.77267885e-01 -4.06694382e-01 -4.78194654e-01
2.27818400e-01 5.82524836e-01 3.33167523e-01 -1.31432426e+00
2.17915207e-01 1.04059190e-01 4.84842896e-01 -7.28980601e-01
1.03196144e+00 -3.12166691e-01 8.72594953e-01 1.88895792e-01
-7.77394593e-01 2.66973615e-01 4.39595014e-01 1.01544224e-01
-3.42535734e-01 -1.37569413e-01 6.25065565e-01 -8.02856907e-02
-9.67655480e-02 -1.16604874e-02 -4.92014676e-01 1.66342840e-01
1.02111392e-01 7.94737861e-02 1.73242733e-01 -7.21123457e-01
-8.56403649e-01 -3.66583586e-01 1.73165664e-01 1.57895803e-01
6.00085318e-01 -1.46826100e+00 -7.90084124e-01 5.01726806e-01
-4.26139057e-01 -1.48981094e-01 2.25691870e-01 4.31011409e-01
-2.83003241e-01 -9.67550501e-02 3.85933332e-02 -4.86451894e-01
-1.21839786e+00 2.47757778e-01 2.64890313e-01 6.76187426e-02
-7.79590249e-01 1.08488870e+00 -2.18762979e-01 -1.01993866e-01
3.76833498e-01 -3.27104568e-01 2.78504699e-01 -1.87689349e-01
7.82057464e-01 1.12381384e-01 3.97853181e-02 -1.40689105e-01
8.40215907e-02 3.50903779e-01 2.59750098e-01 -8.96345615e-01
1.22144639e+00 1.79372087e-01 -9.82640162e-02 8.50332737e-01
8.60721469e-01 7.68178642e-01 -1.11798465e+00 2.61097625e-02
-2.85852641e-01 -3.57745767e-01 9.68960524e-02 -7.73314059e-01
-8.71470571e-01 7.16610253e-01 -4.61225733e-02 1.42805189e-01
1.26648021e+00 -2.83910304e-01 1.02782500e+00 1.76595449e-01
1.05858363e-01 -8.59439015e-01 2.31171384e-01 6.73321903e-01
1.19414115e+00 -8.56899545e-02 -2.60263622e-01 -6.47680759e-02
-6.42631233e-01 9.89708841e-01 -6.77299723e-02 -4.17139709e-01
6.51846886e-01 1.00416005e+00 3.40343148e-01 2.13988245e-01
-1.07137370e+00 1.92511752e-01 1.12839192e-01 5.26702464e-01
5.72102666e-01 6.41964450e-02 3.19981873e-01 5.94830513e-01
-1.27220881e+00 -1.50964111e-01 6.53555393e-01 5.16141355e-01
-3.37206781e-01 -1.37243474e+00 -7.18670428e-01 2.23701939e-01
-5.45297801e-01 -7.08066523e-01 -1.48059368e-01 3.59939128e-01
-1.05152302e-03 1.11752331e+00 1.70527428e-01 -6.03959382e-01
1.81482553e-01 2.49736980e-01 4.84863192e-01 -6.55563593e-01
-1.03631783e+00 6.82185352e-01 2.83587575e-01 -6.05019748e-01
-9.74454451e-03 -5.66387653e-01 -1.13843811e+00 -2.77416170e-01
-2.30681181e-01 2.89170533e-01 7.21527994e-01 4.26551402e-01
3.81032884e-01 1.14265251e+00 7.45694101e-01 -1.24708867e+00
-4.85981584e-01 -9.99247670e-01 -7.83488810e-01 -4.18209136e-02
6.38706088e-01 8.75261892e-03 -2.40609691e-01 3.00388724e-01] | [15.626253128051758, 5.8331298828125] |
06c53016-2a45-404b-8a04-807be5a042c9 | machine-learning-for-large-scale-optimization | 2301.03377 | null | https://arxiv.org/abs/2301.03377v1 | https://arxiv.org/pdf/2301.03377v1.pdf | Machine Learning for Large-Scale Optimization in 6G Wireless Networks | The sixth generation (6G) wireless systems are envisioned to enable the paradigm shift from "connected things" to "connected intelligence", featured by ultra high density, large-scale, dynamic heterogeneity, diversified functional requirements and machine learning capabilities, which leads to a growing need for highly efficient intelligent algorithms. The classic optimization-based algorithms usually require highly precise mathematical model of data links and suffer from poor performance with high computational cost in realistic 6G applications. Based on domain knowledge (e.g., optimization models and theoretical tools), machine learning (ML) stands out as a promising and viable methodology for many complex large-scale optimization problems in 6G, due to its superior performance, generalizability, computational efficiency and robustness. In this paper, we systematically review the most representative "learning to optimize" techniques in diverse domains of 6G wireless networks by identifying the inherent feature of the underlying optimization problem and investigating the specifically designed ML frameworks from the perspective of optimization. In particular, we will cover algorithm unrolling, learning to branch-and-bound, graph neural network for structured optimization, deep reinforcement learning for stochastic optimization, end-to-end learning for semantic optimization, as well as federated learning for distributed optimization, for solving challenging large-scale optimization problems arising from various important wireless applications. Through the in-depth discussion, we shed light on the excellent performance of ML-based optimization algorithms with respect to the classical methods, and provide insightful guidance to develop advanced ML techniques in 6G networks. | ['Wei zhang', 'Jun Zhang', 'Lin Bai', 'Liqun Fu', 'Yong Zhou', 'Zixin Wang', 'Yuanming Shi', 'Lixiang Lian', 'Yandong Shi'] | 2023-01-03 | null | null | null | null | ['distributed-optimization'] | ['methodology'] | [-1.59352064e-01 4.92196754e-02 -7.27486968e-01 -2.62702644e-01
-3.70929122e-01 -3.67160648e-01 -1.69388726e-01 3.68354246e-02
1.45461783e-01 9.19793963e-01 -3.35159898e-02 -5.99238932e-01
-9.51233804e-01 -7.83896923e-01 -3.21833134e-01 -7.42251873e-01
-7.87845314e-01 8.18903565e-01 -5.22126555e-01 -2.68526584e-01
-1.42820654e-02 5.32004416e-01 -1.08178008e+00 -3.65324438e-01
1.06966615e+00 1.59889662e+00 3.21532190e-01 5.88414371e-01
-7.83527046e-02 5.68089247e-01 -5.49563110e-01 -5.43690503e-01
3.72603506e-01 -1.56094939e-01 -7.10694730e-01 2.73812175e-01
-1.23936713e-01 2.85151042e-02 -4.98101473e-01 9.14063096e-01
8.61366987e-01 -1.88970007e-02 1.40952080e-01 -1.48819530e+00
-3.46856564e-01 8.87038171e-01 -1.57049835e-01 1.00248747e-01
1.85306206e-01 9.87822115e-02 1.30421722e+00 -4.15798038e-01
3.22439909e-01 1.03207064e+00 4.24042732e-01 3.54029834e-01
-9.09549057e-01 -4.07371402e-01 5.44828534e-01 4.59044814e-01
-1.11859596e+00 -6.40287638e-01 7.91069925e-01 -5.93047142e-02
9.33921099e-01 3.86129707e-01 9.85925257e-01 6.42863274e-01
1.74491659e-01 5.83413184e-01 5.32488465e-01 -2.77421147e-01
4.83707577e-01 3.05194799e-02 -3.53249401e-01 9.58127081e-01
3.61905307e-01 3.00338447e-01 -2.74228901e-01 -1.89992607e-01
2.79866606e-01 3.51206362e-02 -8.64272937e-02 -6.09142661e-01
-1.07662487e+00 6.86808050e-01 6.58617020e-01 -1.16243875e-02
-4.89780724e-01 4.39838350e-01 1.44301847e-01 6.36985064e-01
1.98571816e-01 4.60681379e-01 -7.62859225e-01 -1.47991478e-01
-8.46177340e-01 3.91469263e-02 1.23594129e+00 1.23838019e+00
8.67695808e-01 5.42351544e-01 -3.04368526e-01 6.02433443e-01
6.60877943e-01 5.50048232e-01 6.20670663e-03 -8.38198960e-01
7.79652357e-01 5.31741381e-01 -2.54436135e-01 -1.06094837e+00
-8.06753457e-01 -1.34561753e+00 -1.36128366e+00 -1.58829719e-01
7.31766149e-02 -5.71469247e-01 -4.97009128e-01 1.62316775e+00
2.76252925e-01 1.42650485e-01 -8.30200016e-02 7.52249837e-01
6.74573600e-01 5.44172764e-01 -1.11834265e-01 -5.89940250e-01
8.48548591e-01 -1.08572125e+00 -5.51904142e-01 -1.77557573e-01
6.95091128e-01 -4.07363176e-01 6.24435008e-01 1.63571402e-01
-9.06561196e-01 -1.80933565e-01 -1.08636177e+00 6.17558658e-01
-2.17703238e-01 -3.61911431e-02 9.80714798e-01 1.13162792e+00
-1.02926171e+00 5.33806264e-01 -4.91378397e-01 -5.27837992e-01
8.76872003e-01 1.02206099e+00 2.35560432e-01 -7.97883794e-02
-9.91400778e-01 5.19091249e-01 4.16032165e-01 1.27478316e-01
-9.46975708e-01 -8.25255930e-01 -3.35145712e-01 4.18271244e-01
8.29875231e-01 -1.02479172e+00 8.65851462e-01 -6.66205764e-01
-1.89213884e+00 4.55076277e-01 1.36297673e-01 -6.66958272e-01
3.08625251e-01 1.14206515e-01 -7.53173769e-01 -2.36940291e-02
-3.30234617e-01 -2.00756431e-01 9.87744629e-01 -8.87499928e-01
-5.90391755e-01 -4.92936730e-01 2.40282431e-01 6.13037869e-03
-5.88025570e-01 -1.18401587e-01 -3.04011405e-01 -6.39991045e-01
1.26306802e-01 -8.50388288e-01 -8.65520895e-01 -3.13540161e-01
-4.59185064e-01 7.97140121e-04 8.54933023e-01 -3.37583274e-01
1.59004116e+00 -1.77449489e+00 3.22616667e-01 6.66894019e-01
8.36079955e-01 1.80827901e-01 -1.76089421e-01 4.12086874e-01
2.94343919e-01 3.20013374e-01 1.34627730e-01 7.84746334e-02
2.15459779e-01 1.23522967e-01 3.38372797e-01 3.25998485e-01
-2.83876061e-01 1.29243684e+00 -1.02441001e+00 -2.50022680e-01
3.46756369e-01 -1.45592570e-01 -7.72532463e-01 3.28875870e-01
-4.78572637e-01 5.26721835e-01 -9.80535328e-01 1.17010474e+00
4.37247545e-01 -7.69511580e-01 4.22328413e-01 -5.78699291e-01
2.35195801e-01 -1.30743653e-01 -1.03534842e+00 1.50888288e+00
-8.71614516e-01 1.33378610e-01 4.92573440e-01 -1.76035988e+00
8.59092414e-01 4.10926729e-01 1.02618122e+00 -7.23864853e-01
9.45673063e-02 4.42487210e-01 2.55209029e-01 -5.92550635e-01
-3.65819991e-01 9.80377421e-02 -1.41510189e-01 3.51420105e-01
2.23817125e-01 -3.67086940e-02 2.44213700e-01 9.01511535e-02
1.15377569e+00 -5.74525058e-01 4.90119129e-01 -2.62395442e-01
5.91107726e-01 -3.92791301e-01 6.16979420e-01 1.00067747e+00
-4.17937875e-01 -7.10328147e-02 4.08671081e-01 -4.46550339e-01
-6.67244077e-01 -9.32131171e-01 3.29959512e-01 1.04480433e+00
2.49475703e-01 -4.55030799e-01 -3.89634281e-01 -7.33746707e-01
8.77917260e-02 2.77010769e-01 7.49953911e-02 -3.41069877e-01
-4.02091175e-01 -1.23727441e+00 1.58005476e-01 -1.16661593e-01
7.30945468e-01 -5.63835919e-01 8.10965374e-02 7.00382352e-01
2.53803045e-01 -1.39798737e+00 -1.73628911e-01 3.64503175e-01
-1.08235157e+00 -7.44520783e-01 -4.04139489e-01 -6.99046016e-01
2.35431716e-01 1.37416139e-01 1.46983480e+00 1.26277074e-01
-2.27384493e-01 4.00323272e-01 -2.53203690e-01 -1.92233667e-01
-2.11583123e-01 6.00899935e-01 2.04537854e-01 2.56937593e-01
-2.42595822e-01 -1.16908276e+00 -5.49367487e-01 4.69853431e-01
-1.78091824e-01 -2.73104608e-01 1.07721853e+00 7.26908207e-01
4.99906212e-01 3.73104900e-01 9.57904935e-01 -8.31433356e-01
5.49039364e-01 -7.24383116e-01 -8.58375251e-01 6.04739666e-01
-1.00929236e+00 1.18305080e-01 1.01722860e+00 -6.23408239e-03
-5.53309798e-01 -3.06594580e-01 -2.76930835e-02 -2.44669586e-01
3.27930957e-01 6.84336901e-01 -8.24234545e-01 -7.28637516e-01
4.63935554e-01 4.99799429e-03 -8.46501887e-02 -3.40676576e-01
5.26340365e-01 5.26430190e-01 -9.31915343e-02 -8.43418479e-01
1.20841455e+00 3.19458842e-01 3.57107103e-01 -7.59558499e-01
-9.70516264e-01 -2.54058987e-01 -1.17841996e-01 -3.29879969e-01
3.95802945e-01 -6.33281648e-01 -9.13548648e-01 8.00636858e-02
-7.53705561e-01 -3.23357165e-01 -3.02257836e-01 5.04900277e-01
-6.31095469e-01 9.08029452e-02 -1.96413904e-01 -6.12779200e-01
-6.05942547e-01 -1.10746825e+00 4.55410659e-01 3.48478138e-01
3.30659360e-01 -1.30857289e+00 -4.20859307e-01 6.42813027e-01
1.01786017e+00 1.13359652e-01 1.25254738e+00 -4.92458791e-01
-1.08425391e+00 -1.51790828e-01 -2.53552765e-01 1.16079278e-01
1.34283662e-01 -3.47527295e-01 -3.82691324e-01 -4.32665825e-01
-1.95146084e-01 -2.42036078e-02 3.78365099e-01 6.42930388e-01
1.58306110e+00 -7.76160896e-01 -4.20524389e-01 1.29911017e+00
1.65797102e+00 1.62827015e-01 2.95679480e-01 6.90119639e-02
7.47437418e-01 2.56312788e-01 3.65407169e-01 9.16356444e-01
3.26959312e-01 7.34482765e-01 1.05403578e+00 -2.53492668e-02
1.24170100e-02 -2.59040557e-02 -1.82059146e-02 8.74332547e-01
-3.27686667e-01 -6.72809243e-01 -5.82082152e-01 -3.07199173e-03
-1.97884178e+00 -9.10094619e-01 3.10205191e-01 2.14440966e+00
9.88938287e-02 1.58778310e-01 3.39790136e-01 5.95972389e-02
7.36253619e-01 3.86913091e-01 -8.00922692e-01 -2.18573779e-01
-1.82889163e-01 2.04314858e-01 8.65047395e-01 1.83734477e-01
-9.99667287e-01 1.02865160e+00 5.95632696e+00 1.19520915e+00
-1.09018850e+00 2.42639422e-01 9.00798857e-01 -1.68904379e-01
-3.32908452e-01 6.02101274e-02 -5.00820756e-01 4.31294531e-01
1.03322124e+00 -3.05505574e-01 1.09744406e+00 8.25087011e-01
7.28925705e-01 5.51033795e-01 -9.26390827e-01 1.44890523e+00
-5.10704696e-01 -1.76564324e+00 8.89343172e-02 2.86761880e-01
1.02837026e+00 3.55767608e-01 6.79282919e-02 2.43984222e-01
1.63092524e-01 -1.23864090e+00 2.89737880e-01 1.97095796e-01
6.86589897e-01 -6.70986414e-01 6.30557656e-01 5.22329211e-02
-1.28957272e+00 -6.32948756e-01 -1.38156220e-01 -4.54222783e-02
4.17772412e-01 1.14542317e+00 -4.72541749e-01 8.15603197e-01
5.36280632e-01 9.18937504e-01 -1.92419812e-01 1.40386868e+00
1.19443893e-01 6.88230872e-01 -3.89894783e-01 -6.23227835e-01
1.39656350e-01 -4.45161015e-01 8.78523707e-01 7.30571508e-01
5.29823005e-01 -1.79332539e-01 5.76546907e-01 6.67028844e-01
-3.64996761e-01 2.29599789e-01 -2.11839557e-01 -4.56689358e-01
6.03918910e-01 1.40421605e+00 -4.45411533e-01 -9.80894044e-02
-6.17845416e-01 4.64393020e-01 9.58988741e-02 7.14223683e-01
-7.74311006e-01 -3.43571454e-01 6.63672924e-01 1.29661635e-01
1.34698421e-01 -3.23068798e-01 -5.46312392e-01 -1.11787045e+00
9.96406078e-02 -1.02552497e+00 4.85876948e-01 5.04594892e-02
-1.27710843e+00 4.05976474e-01 -4.92134601e-01 -1.15195119e+00
6.71398733e-03 -5.89500546e-01 -5.64886630e-01 3.46361130e-01
-1.63456309e+00 -9.84555542e-01 -4.18537587e-01 7.51410484e-01
3.48353297e-01 -8.54275107e-01 7.27443933e-01 7.38494992e-01
-7.15495944e-01 8.07722867e-01 3.49836081e-01 -1.99129283e-01
7.28776827e-02 -1.04866993e+00 -1.03287399e-01 4.50528473e-01
6.64961487e-02 1.24871165e-01 3.79139781e-01 -2.39275277e-01
-2.18542600e+00 -1.04493892e+00 4.02905524e-01 2.42420167e-01
9.20775771e-01 -2.42320895e-01 -2.16921028e-02 2.29343712e-01
-4.15176451e-01 4.34587538e-01 5.79702616e-01 4.35268641e-01
1.79589137e-01 -9.87142563e-01 -1.10173535e+00 6.51354015e-01
1.42875493e+00 -9.95000154e-02 4.41318959e-01 8.82920086e-01
6.56078756e-01 -1.86828256e-01 -1.08447516e+00 5.10958672e-01
4.89125252e-01 -7.66892016e-01 1.23463321e+00 -7.32996225e-01
-1.71252072e-01 -6.49350509e-02 -4.89550203e-01 -1.08777940e+00
-3.66388083e-01 -1.51070333e+00 -1.00111914e+00 8.03987861e-01
6.09466732e-01 -8.16743016e-01 1.23653126e+00 6.93634152e-02
-2.44938821e-01 -1.29199374e+00 -1.06054890e+00 -9.19472754e-01
-4.70392406e-01 -4.49360847e-01 8.90773237e-01 4.96846080e-01
-1.51545838e-01 4.15382147e-01 -6.61825180e-01 2.41089657e-01
8.73748779e-01 2.08580509e-01 7.51765430e-01 -1.29041147e+00
-7.93945253e-01 -7.36615300e-01 -8.18445802e-01 -1.40671504e+00
-4.88201641e-02 -9.84799266e-01 -5.65327525e-01 -1.52332997e+00
-1.70983076e-01 -9.78111744e-01 -4.66477722e-01 -8.85393098e-02
1.69723541e-01 -1.18134759e-01 8.21311846e-02 3.48399207e-02
-1.01053762e+00 6.69994354e-01 1.31398964e+00 -4.43099886e-01
-3.52741748e-01 1.00237012e+00 -9.72817838e-01 4.85858083e-01
9.49106395e-01 -1.25367403e-01 -5.56289911e-01 -4.19432193e-01
4.98545259e-01 3.56484622e-01 5.96544482e-02 -1.15211427e+00
2.32662469e-01 -4.84256297e-01 -5.19840010e-02 -5.47930226e-02
7.87869617e-02 -1.06648529e+00 -2.72368602e-02 6.25338852e-01
2.96688396e-02 -1.71412155e-01 -4.09920305e-01 7.26233125e-01
-6.64360151e-02 -3.64295170e-02 7.33780086e-01 -8.09512511e-02
-8.91363680e-01 1.28706861e+00 -8.36140215e-02 3.74624759e-01
9.49271560e-01 -2.37414613e-01 6.33911043e-02 -7.49878466e-01
-1.05452168e+00 5.29654741e-01 -1.75051928e-01 1.57502502e-01
5.09348333e-01 -1.20655406e+00 -7.79579520e-01 -3.75633575e-02
-6.67928010e-02 -4.47000980e-01 1.43774331e-01 9.63895202e-01
-4.98553574e-01 3.98376405e-01 1.52307212e-01 -5.41591227e-01
-7.16657639e-01 5.73363066e-01 7.45730996e-01 -6.47476912e-01
-2.37613797e-01 7.69591987e-01 -3.71086329e-01 -4.48755473e-01
4.87955958e-01 -5.46930842e-02 1.48812771e-01 -3.00743788e-01
-1.63020417e-01 7.89294958e-01 2.44192049e-01 8.69873271e-04
-4.82992977e-01 5.61930299e-01 2.71501601e-01 3.64623368e-01
1.44879842e+00 -4.12763327e-01 -1.41895249e-01 -1.21126421e-01
1.27205002e+00 -8.19000378e-02 -7.34775603e-01 -2.89699376e-01
2.09949747e-01 -2.78471351e-01 2.35265017e-01 -8.10248315e-01
-1.62950563e+00 5.34990728e-01 7.31563807e-01 5.35348594e-01
1.37413347e+00 1.66927040e-01 8.62010896e-01 7.48918056e-01
8.77337396e-01 -1.23272789e+00 1.43996686e-01 4.70697969e-01
4.25666511e-01 -1.39409924e+00 2.11589366e-01 -6.20272875e-01
1.73500732e-01 1.23331463e+00 4.01743442e-01 1.41670302e-01
1.22380912e+00 5.17187417e-02 -2.90316314e-01 -2.02468738e-01
-6.25658989e-01 -2.92645752e-01 2.05962911e-01 8.58933389e-01
7.67694414e-02 2.35120803e-01 -1.26523092e-01 5.08730769e-01
-2.53563434e-01 -4.16307271e-01 -5.46754487e-02 3.81532043e-01
-4.68476772e-01 -1.28321159e+00 1.85190402e-02 1.04123402e+00
-3.96611452e-01 1.84365101e-02 9.31862742e-02 5.10417342e-01
1.74341038e-01 1.23977518e+00 -6.31549478e-01 -4.08931434e-01
9.09543037e-02 -7.62023687e-01 4.83992040e-01 -4.56702590e-01
-3.58450383e-01 -1.21116340e-01 2.97616929e-01 -7.88664579e-01
-3.26632589e-01 -2.14757949e-01 -8.50042045e-01 -4.05377746e-01
-4.02144790e-01 2.23083600e-01 7.36843228e-01 1.33406639e+00
6.22627974e-01 6.08369410e-01 1.12283647e+00 -5.24242818e-01
-7.83501446e-01 -3.52786958e-01 -6.37932539e-01 -2.41755843e-01
2.86028177e-01 -3.21884811e-01 -3.49986762e-01 -5.92884719e-01] | [5.974393844604492, 1.6959322690963745] |
89e8a48d-237d-4cf2-808c-99026b1c6e37 | precognition-in-task-oriented-dialogue | 2203.03244 | null | https://arxiv.org/abs/2203.03244v1 | https://arxiv.org/pdf/2203.03244v1.pdf | Precognition in Task-oriented Dialogue Understanding: Posterior Regularization by Future Context | Task-oriented dialogue systems have become overwhelmingly popular in recent researches. Dialogue understanding is widely used to comprehend users' intent, emotion and dialogue state in task-oriented dialogue systems. Most previous works on such discriminative tasks only models current query or historical conversations. Even if in some work the entire dialogue flow was modeled, it is not suitable for the real-world task-oriented conversations as the future contexts are not visible in such cases. In this paper, we propose to jointly model historical and future information through the posterior regularization method. More specifically, by modeling the current utterance and past contexts as prior, and the entire dialogue flow as posterior, we optimize the KL distance between these distributions to regularize our model during training. And only historical information is used for inference. Extensive experiments on two dialogue datasets validate the effectiveness of our proposed method, achieving superior results compared with all baseline models. | ['Yongliang Wang', 'Bingzhu Du', 'Chao Liu', 'Yuchi Zhang', 'Nan Su'] | 2022-03-07 | null | null | null | null | ['dialogue-understanding'] | ['natural-language-processing'] | [-1.02297321e-01 1.06884770e-01 -3.34614068e-01 -7.87955880e-01
-4.44725037e-01 -4.87040818e-01 9.76653218e-01 5.45337237e-02
-4.83739853e-01 9.86815155e-01 6.22877777e-01 -2.34897330e-01
4.68732089e-01 -4.37702537e-01 9.02778469e-03 -4.92721885e-01
1.27309814e-01 3.19146127e-01 1.04657836e-01 -5.10345519e-01
3.10802102e-01 -2.10284386e-02 -1.06999969e+00 1.97853237e-01
1.01946008e+00 6.96170747e-01 2.79241294e-01 6.81817055e-01
-4.71342891e-01 9.27863836e-01 -7.95568526e-01 -4.52804476e-01
-2.19555855e-01 -5.74319363e-01 -1.08738840e+00 3.34250242e-01
-3.41738611e-01 -4.01588559e-01 -4.23133165e-01 8.84302258e-01
4.17821109e-01 5.70063591e-01 4.97842103e-01 -9.22282815e-01
-1.83537662e-01 4.61145163e-01 -4.66672421e-01 7.23043084e-02
7.02851295e-01 1.28481716e-01 1.06909299e+00 -6.38156414e-01
3.16637784e-01 1.54634643e+00 2.56395519e-01 8.32381606e-01
-1.04152775e+00 -1.83363423e-01 7.54046321e-01 2.04683438e-01
-8.86677325e-01 -5.22917271e-01 1.13336027e+00 -2.73155093e-01
7.47734725e-01 2.95026124e-01 6.61579072e-01 1.29238963e+00
1.32519200e-01 1.32152104e+00 9.41305816e-01 -3.05735618e-01
3.84494513e-02 5.28715193e-01 4.26712215e-01 3.58177483e-01
-7.57485688e-01 -5.45657516e-01 -4.43997264e-01 -3.88215810e-01
3.96641403e-01 -3.31566744e-02 -4.59888190e-01 1.80481344e-01
-9.99445736e-01 9.53427434e-01 -1.88870817e-01 3.33210289e-01
-4.82247502e-01 -3.62306386e-01 6.74685657e-01 2.90495098e-01
8.36847901e-01 1.73707083e-01 -4.27259266e-01 -6.82534099e-01
-5.04012823e-01 2.37720892e-01 1.29290724e+00 7.65635252e-01
6.36307955e-01 -3.00698727e-01 -4.24984545e-01 1.32376111e+00
3.44273239e-01 6.86865896e-02 4.35666472e-01 -8.12426150e-01
5.88389575e-01 5.05756378e-01 4.47847337e-01 -8.69496346e-01
-4.02133375e-01 -4.18963470e-02 -8.30264747e-01 -6.75769806e-01
5.29393435e-01 -7.81867504e-01 -3.82845074e-01 1.75802565e+00
3.88645232e-01 -2.86234953e-02 2.82962948e-01 8.77454638e-01
5.92777550e-01 1.10203636e+00 1.75703004e-01 -6.83463633e-01
1.28434050e+00 -1.10614014e+00 -1.19322431e+00 -4.22028899e-01
6.04668379e-01 -9.32123780e-01 1.10088718e+00 4.40438390e-01
-8.55529130e-01 -5.50520718e-01 -7.44559348e-01 -3.68791074e-02
6.25854880e-02 2.13547245e-01 7.62733221e-01 4.96495247e-01
-6.63094640e-01 2.46266142e-01 -6.66349947e-01 -3.58630270e-01
-2.58688569e-01 1.10131331e-01 1.22507088e-01 2.74911582e-01
-1.61355364e+00 8.79702687e-01 4.16165143e-01 4.70726907e-01
-5.35375178e-01 -1.09334134e-01 -7.48705029e-01 -6.56116307e-02
4.35363442e-01 -2.96290576e-01 1.76418519e+00 -6.51763439e-01
-2.22361636e+00 5.31307042e-01 -5.26461124e-01 -2.80688554e-01
6.14551127e-01 -4.56324220e-01 -2.50848204e-01 -2.51267970e-01
-3.89417470e-01 3.60095829e-01 6.55873895e-01 -1.12673390e+00
-7.63032019e-01 -1.82000905e-01 5.47333598e-01 7.18934357e-01
-5.38834929e-01 -2.24708751e-01 -8.46700490e-01 -2.26164818e-01
-6.32381886e-02 -9.63458836e-01 -3.96022081e-01 -2.83647299e-01
-3.88714492e-01 -8.08854401e-01 1.03601241e+00 -7.67497838e-01
1.49713182e+00 -1.92963159e+00 1.91396579e-01 -2.30559051e-01
-1.24514066e-01 2.67220825e-01 1.47615373e-01 7.29959667e-01
4.80251968e-01 -1.74673006e-01 -7.88352937e-02 -6.22702599e-01
1.55508220e-01 5.06876349e-01 -5.18598497e-01 3.42873335e-01
-2.88681537e-02 5.39704144e-01 -8.41604888e-01 -5.93519568e-01
3.03347081e-01 3.23084742e-01 -3.84991020e-01 8.41243207e-01
-6.19642019e-01 8.65902543e-01 -9.04243529e-01 8.30570832e-02
4.84662622e-01 -7.32379556e-02 4.80049908e-01 -4.46641818e-02
-5.23226745e-02 5.10031402e-01 -7.16890216e-01 1.94821489e+00
-1.00071990e+00 5.45876980e-01 3.57689321e-01 -9.79193926e-01
1.05066514e+00 6.34087861e-01 2.91170686e-01 -5.82851112e-01
2.28166468e-02 -2.82439917e-01 3.42457113e-03 -8.44224870e-01
7.00290382e-01 -2.84976065e-01 -2.74627119e-01 4.65868980e-01
-1.12001836e-01 -1.87214479e-01 2.29692176e-01 2.60930359e-01
4.38759059e-01 7.43140578e-02 3.10609311e-01 9.68698561e-02
1.06987345e+00 -3.29398960e-01 7.74660945e-01 5.39156795e-01
-2.86228061e-01 3.20335865e-01 9.76950347e-01 -1.32822990e-01
-5.95342875e-01 -5.78602791e-01 -2.01824848e-02 1.40157878e+00
2.78728545e-01 -4.99151558e-01 -6.39151275e-01 -9.76924598e-01
-4.41092640e-01 1.05044806e+00 -4.05255198e-01 -1.09235466e-01
-6.14743948e-01 -8.99965167e-01 2.14073181e-01 8.16322789e-02
6.38966918e-01 -9.41828132e-01 -2.49174327e-01 4.29454178e-01
-6.87303185e-01 -9.78141904e-01 -6.73014998e-01 -3.46737146e-01
-6.86701775e-01 -7.10712075e-01 -8.83609891e-01 -4.78478193e-01
2.10251883e-01 1.03824576e-02 9.61177886e-01 -7.84811005e-02
1.41129643e-01 3.47413719e-01 -3.05465907e-01 -2.69821614e-01
-5.45101285e-01 9.58818048e-02 -5.14757931e-02 3.18824053e-01
2.66639829e-01 -4.16802734e-01 -6.69548750e-01 2.70650685e-01
-5.32139003e-01 1.77559018e-01 1.87788174e-01 1.14275217e+00
-1.09603465e-01 -1.64379224e-01 8.42022955e-01 -1.20569789e+00
1.23060536e+00 -6.54699087e-01 -2.05630362e-01 4.96929526e-01
-4.18148756e-01 -2.34127063e-02 8.19144368e-01 -4.62872326e-01
-1.85372496e+00 -2.38604426e-01 -5.83929539e-01 1.06369190e-01
-2.40041912e-01 6.38755381e-01 -3.12475324e-01 6.34076953e-01
1.38558537e-01 2.53356248e-01 -4.94667813e-02 -5.03407538e-01
2.71473378e-01 1.13900328e+00 1.38293684e-01 -8.50501776e-01
1.00897409e-01 1.01584069e-01 -6.32955909e-01 -1.07322884e+00
-9.79667604e-01 -7.30790317e-01 -5.89749455e-01 -2.47546911e-01
7.69553304e-01 -6.90126836e-01 -1.05284107e+00 4.94620293e-01
-1.56512570e+00 -1.52530521e-01 2.34029084e-01 5.84794402e-01
-4.70130295e-01 7.65744507e-01 -5.94003022e-01 -1.54502690e+00
-2.97930747e-01 -1.22395217e+00 7.92039573e-01 5.27376831e-01
-3.83542717e-01 -1.47594750e+00 1.58099487e-01 3.34305227e-01
2.51400769e-01 -1.09744772e-01 8.43048930e-01 -9.55247998e-01
-3.18444148e-02 -1.77173197e-01 -8.84348080e-02 3.20075899e-01
3.50448400e-01 -1.35478422e-01 -1.09579420e+00 -1.02777511e-01
4.57017004e-01 -5.69970727e-01 6.09646261e-01 2.60184221e-02
1.08600068e+00 -5.35219312e-01 -2.64537781e-01 -1.85403936e-02
7.72047341e-01 5.29171884e-01 3.97051543e-01 -2.26180449e-01
4.81106818e-01 1.10980594e+00 1.04639482e+00 8.69741619e-01
6.20785058e-01 7.40685165e-01 1.22583276e-02 1.82184234e-01
5.62867403e-01 -2.50052214e-01 4.06864285e-01 1.26751816e+00
1.98025137e-01 -6.15078211e-01 -6.62499964e-01 5.51689208e-01
-2.24997330e+00 -8.54319334e-01 2.57717562e-03 2.06636477e+00
1.22642505e+00 1.95587769e-01 -3.81226949e-02 -3.99823099e-01
6.86212659e-01 7.44573414e-01 -6.51871681e-01 -4.62155223e-01
1.82268649e-01 -2.42480949e-01 -3.18199426e-01 7.51737952e-01
-1.09712470e+00 8.61361682e-01 5.72221279e+00 8.24545324e-01
-1.01772141e+00 6.50773123e-02 8.05188715e-01 1.81534544e-01
-2.89343119e-01 2.64593810e-01 -7.48815954e-01 4.44136471e-01
8.08282673e-01 -3.68157148e-01 1.37236252e-01 7.19599783e-01
5.30677080e-01 -4.41992551e-01 -1.05405605e+00 8.98672462e-01
-9.06049013e-02 -7.61801958e-01 -2.51574844e-01 -3.19303781e-01
3.49791974e-01 -5.97234249e-01 -1.61099225e-01 7.45548546e-01
2.19645917e-01 -7.53840029e-01 1.42948002e-01 7.60116339e-01
3.57629746e-01 -7.80072451e-01 8.43634605e-01 9.10923481e-01
-8.41448247e-01 5.09179719e-02 -2.45449379e-01 -2.14711979e-01
5.85704327e-01 4.15428966e-01 -1.15340543e+00 5.14500856e-01
1.47083119e-01 7.71977782e-01 -9.12730917e-02 4.70866084e-01
-2.07466975e-01 8.43685746e-01 -1.78822994e-01 -4.07847822e-01
3.44516069e-01 -5.37994027e-01 5.46141624e-01 1.32797146e+00
-1.14046253e-01 2.56032676e-01 5.90148866e-01 4.57624555e-01
1.99729856e-02 3.45259845e-01 -5.04204035e-01 -1.62294656e-01
3.33673716e-01 1.25880623e+00 -2.57390082e-01 -3.02925706e-01
-5.40597260e-01 1.25490332e+00 2.93028504e-01 4.77965534e-01
-7.21988022e-01 -4.10083443e-01 6.22278750e-01 -4.64746833e-01
-1.99605629e-01 -3.39746445e-01 -1.12907803e-02 -1.39252782e+00
1.66055951e-02 -7.73430705e-01 4.07842487e-01 -2.17964724e-01
-1.38432217e+00 5.25219977e-01 9.73629858e-03 -9.18025970e-01
-7.28637516e-01 -2.30597019e-01 -9.11473811e-01 1.06179857e+00
-1.33001637e+00 -9.33446109e-01 1.21739728e-03 4.53028470e-01
1.14501214e+00 2.94928942e-02 9.82257366e-01 -2.92162951e-02
-8.11373055e-01 4.26049083e-01 1.08939677e-01 3.46473232e-02
1.05504537e+00 -1.28991723e+00 1.56933829e-01 3.70874256e-01
-1.63888916e-01 8.05592120e-01 9.24512029e-01 -4.98448968e-01
-1.23482847e+00 -6.03675365e-01 8.37588966e-01 -9.45750177e-02
6.67404354e-01 -4.85875309e-01 -1.13018990e+00 4.12163258e-01
6.26464248e-01 -6.51514649e-01 7.84481883e-01 5.83159983e-01
1.37121350e-01 -5.49932569e-02 -8.31757426e-01 7.50975132e-01
6.67836607e-01 -5.76186597e-01 -7.34235227e-01 5.16572058e-01
6.89137101e-01 -5.01148224e-01 -9.30560589e-01 2.47035399e-01
5.39673269e-01 -9.34590161e-01 4.94287491e-01 -4.36341912e-01
7.24493787e-02 1.31807655e-01 1.46286964e-01 -1.41777802e+00
3.48176867e-01 -1.02697396e+00 -1.74180403e-01 1.43136728e+00
3.49311650e-01 -6.21800244e-01 6.61986470e-01 9.64164674e-01
3.33569720e-02 -8.02264929e-01 -6.79440737e-01 -9.87941474e-02
-8.33823457e-02 -2.32788041e-01 3.19257736e-01 9.18867290e-01
4.53024358e-01 8.58388364e-01 -9.76466000e-01 -1.80310205e-01
1.06951892e-01 3.26496363e-01 9.64233935e-01 -1.04766858e+00
-4.69647616e-01 -2.02377960e-01 2.23453984e-01 -1.94548702e+00
5.34515679e-01 -2.11170882e-01 3.86725068e-01 -1.42031622e+00
6.53897002e-02 -3.24564338e-01 1.98756710e-01 6.96452036e-02
-5.64005136e-01 -5.96136749e-01 -4.74350154e-02 1.89912915e-01
-6.51607335e-01 1.19454646e+00 1.34529448e+00 -1.11030430e-01
-5.37334263e-01 5.99553347e-01 -3.71239632e-01 8.42671752e-01
8.32768202e-01 1.04986597e-02 -8.54227901e-01 -1.61545277e-01
-2.44548663e-01 7.45382667e-01 -1.45387486e-01 -4.48146760e-01
3.78820539e-01 -5.71999252e-01 -2.29684282e-02 -6.45402193e-01
7.82535851e-01 -5.81790686e-01 -4.28024709e-01 1.96834859e-02
-6.49571836e-01 -3.69820863e-01 -2.93410830e-02 8.83409202e-01
-5.24303138e-01 -3.42706978e-01 5.48883259e-01 -1.08320706e-01
-6.23773813e-01 2.86051720e-01 -5.67369819e-01 6.83655739e-02
7.86404252e-01 1.86304867e-01 -2.14825720e-02 -1.07665110e+00
-8.09648395e-01 7.86281824e-01 1.48817897e-01 5.67589700e-01
4.76149201e-01 -8.24320674e-01 -6.43509805e-01 -1.43185928e-01
-5.21793030e-04 4.43815328e-02 6.03864193e-01 7.01351762e-01
-1.22859841e-02 4.76192534e-01 2.43334204e-01 -5.49794197e-01
-1.21834981e+00 1.95779681e-01 2.01956332e-01 -5.64460993e-01
-3.53775650e-01 7.08273768e-01 4.50740278e-01 -5.63420713e-01
6.18074477e-01 1.33353397e-01 -6.96612954e-01 2.47721538e-01
4.65470225e-01 -7.90552050e-03 -4.47712421e-01 -4.39714104e-01
2.56441962e-02 2.10804176e-02 -5.33021688e-01 -4.78133976e-01
1.04568291e+00 -6.79056108e-01 -1.33858606e-01 1.11369252e+00
1.23558295e+00 1.53250359e-02 -1.36553812e+00 -4.81595248e-01
2.15227485e-01 -3.87121141e-01 -2.74619311e-01 -4.82476294e-01
-6.10053658e-01 1.25935912e+00 9.51421484e-02 5.00083089e-01
8.98634851e-01 -3.28448921e-01 9.34371889e-01 6.59147739e-01
2.75000721e-01 -1.23035741e+00 1.00177228e-01 9.95514929e-01
8.64242136e-01 -1.36227143e+00 -1.53466731e-01 -3.63501608e-01
-1.22006750e+00 1.19938481e+00 8.66723239e-01 2.93587804e-01
6.40769005e-01 -1.30598634e-01 1.92984641e-01 1.75377235e-01
-1.17490351e+00 1.47139067e-02 1.54768795e-01 1.90457046e-01
8.96720946e-01 -2.18824863e-01 -7.31450677e-01 6.35626018e-01
-1.41875129e-02 -2.47041136e-01 4.20796752e-01 9.64435697e-01
-2.82144159e-01 -1.42059636e+00 7.03617930e-03 3.25491011e-01
-4.66212362e-01 1.66201398e-01 -5.14391780e-01 4.32311833e-01
-4.39814806e-01 1.35910857e+00 -5.31272516e-02 -2.47405723e-01
3.45124215e-01 4.63637829e-01 7.24591017e-02 -7.55236328e-01
-5.74497402e-01 2.92110831e-01 6.35190725e-01 -1.62399665e-01
-4.54909414e-01 -4.85496014e-01 -1.20990157e+00 -1.04417793e-01
-5.28339148e-01 7.20108271e-01 5.40090203e-01 1.04508924e+00
2.56584268e-02 2.40654081e-01 1.17998707e+00 -5.89971125e-01
-7.59279072e-01 -1.43528056e+00 -5.13783216e-01 2.92833120e-01
3.14187974e-01 -4.74564523e-01 -2.54260480e-01 -1.19528167e-01] | [12.695860862731934, 7.747383117675781] |
832ab623-616e-4712-bf80-93a82ebab24c | a-two-step-approach-to-sentence-compression | null | null | https://aclanthology.org/P12-2033 | https://aclanthology.org/P12-2033.pdf | A Two-step Approach to Sentence Compression of Spoken Utterances | null | ['Yang Liu', 'Xian Qian', 'Dong Wang'] | 2012-07-01 | null | null | null | acl-2012-7 | ['meeting-summarization'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.34088134765625, 3.6750025749206543] |
67f606c4-652c-4887-91e8-666bbe2ccf50 | taen-temporal-aware-embedding-network-for-few | 2004.10141 | null | https://arxiv.org/abs/2004.10141v2 | https://arxiv.org/pdf/2004.10141v2.pdf | TAEN: Temporal Aware Embedding Network for Few-Shot Action Recognition | Classification of new class entities requires collecting and annotating hundreds or thousands of samples that is often prohibitively costly. Few-shot learning suggests learning to classify new classes using just a few examples. Only a small number of studies address the challenge of few-shot learning on spatio-temporal patterns such as videos. In this paper, we present the Temporal Aware Embedding Network (TAEN) for few-shot action recognition, that learns to represent actions, in a metric space as a trajectory, conveying both short term semantics and longer term connectivity between action parts. We demonstrate the effectiveness of TAEN on two few shot tasks, video classification and temporal action detection and evaluate our method on the Kinetics-400 and on ActivityNet 1.2 few-shot benchmarks. With training of just a few fully connected layers we reach comparable results to prior art on both few shot video classification and temporal detection tasks, while reaching state-of-the-art in certain scenarios. | ['Udi Barzelay', 'Rami Ben-Ari', 'Mor Shpigel', 'Ophir Azulai', 'Daniel Rotman'] | 2020-04-21 | null | null | null | null | ['few-shot-action-recognition'] | ['computer-vision'] | [ 1.70797259e-01 -8.79036412e-02 -6.42387509e-01 -4.11483198e-01
-5.76957643e-01 -4.13288653e-01 9.88989949e-01 1.91784933e-01
-6.27995372e-01 6.60762489e-01 5.90654731e-01 3.36609602e-01
-1.81793690e-01 -6.82172716e-01 -7.32128620e-01 -3.61267716e-01
-7.63846815e-01 1.82187468e-01 7.23939359e-01 1.05258584e-01
-2.02579852e-02 2.64265180e-01 -1.59217703e+00 6.11717522e-01
-7.38201812e-02 1.19541311e+00 -2.75802851e-01 8.28454196e-01
6.96070194e-02 1.35812044e+00 -1.69955745e-01 -1.50177225e-01
2.09754080e-01 -5.41535854e-01 -8.89446378e-01 5.92214540e-02
7.66569138e-01 -6.49745464e-01 -9.06335831e-01 5.37996769e-01
2.79146552e-01 7.19165564e-01 7.98345327e-01 -1.48979354e+00
-5.39961636e-01 7.20175624e-01 -3.10595214e-01 1.02435648e+00
4.87059087e-01 6.14928663e-01 1.08579302e+00 -8.82297635e-01
1.12146699e+00 8.71706128e-01 9.44134474e-01 7.46145546e-01
-1.00379658e+00 -3.54390144e-01 3.30123484e-01 5.28442621e-01
-1.16828465e+00 -5.90441942e-01 4.41806585e-01 -7.48183429e-01
1.61042237e+00 -2.66468078e-01 9.95021462e-01 1.56366348e+00
1.28262818e-01 1.05192924e+00 4.29307222e-01 1.35992154e-01
5.15583038e-01 -4.90223795e-01 3.69525820e-01 7.96658218e-01
2.54116878e-02 1.01644859e-01 -8.98385584e-01 1.39652058e-01
4.47268873e-01 4.84440058e-01 2.50835833e-03 -5.84260881e-01
-1.34035480e+00 6.58023179e-01 3.61976355e-01 5.85865676e-01
-4.86916542e-01 9.71333385e-01 9.38270688e-01 2.93275297e-01
7.53382206e-01 3.71720314e-01 -4.40396518e-01 -8.88109922e-01
-9.70931292e-01 1.65072411e-01 6.37290120e-01 8.46696794e-01
5.87711513e-01 1.83892757e-01 -5.93290865e-01 4.25821960e-01
-3.99120390e-01 -3.81068476e-02 5.97335696e-01 -9.36349809e-01
4.64330643e-01 4.19420004e-01 2.19358336e-02 -6.65145636e-01
-3.80258799e-01 3.14010903e-02 -3.63963068e-01 -2.88046747e-02
4.02836591e-01 -2.13895693e-01 -8.92094135e-01 1.74955952e+00
1.58923000e-01 9.27392840e-01 -2.52687603e-01 7.86551654e-01
7.04572976e-01 5.82735896e-01 3.93055737e-01 -1.39960021e-01
9.86035347e-01 -1.05076754e+00 -5.79979837e-01 -3.10769707e-01
1.05427051e+00 1.94757178e-01 1.01985717e+00 1.54090347e-02
-9.40378010e-01 -6.00542188e-01 -9.23051417e-01 -5.71677200e-02
-8.25773537e-01 -1.13455556e-01 8.42720568e-01 1.68464661e-01
-8.68312836e-01 1.16880727e+00 -1.31172347e+00 -8.61843109e-01
1.15278828e+00 1.13710277e-01 -5.83578944e-01 -1.79566532e-01
-1.16743004e+00 7.98927128e-01 3.92577499e-01 -4.38553482e-01
-1.38256836e+00 -1.04286027e+00 -1.24947429e+00 1.25931770e-01
5.67887008e-01 -5.36671221e-01 1.27970612e+00 -6.65471077e-01
-1.16951644e+00 7.29290009e-01 9.55894142e-02 -9.41948473e-01
5.32168508e-01 -2.87777245e-01 -5.47138333e-01 4.73028183e-01
1.85139716e-01 7.94040561e-01 7.21542060e-01 -1.78699270e-01
-8.77923131e-01 -1.90391362e-01 3.44043791e-01 5.83958551e-02
-6.53054178e-01 -4.59438823e-02 -1.78563535e-01 -6.03478253e-01
-4.87972528e-01 -5.14823794e-01 -6.58268901e-03 5.09069085e-01
1.76506937e-01 -5.50416291e-01 1.01063180e+00 -3.03116918e-01
1.15419364e+00 -2.21750784e+00 2.64475107e-01 -5.58585286e-01
3.40944648e-01 9.16539505e-02 -3.00904065e-01 4.06470180e-01
-1.27332836e-01 -5.50533831e-02 -5.21318093e-02 -2.45544657e-01
1.79773062e-01 1.36220172e-01 -1.44766748e-01 7.58518696e-01
3.83863747e-01 1.36819983e+00 -1.38069093e+00 -2.31593475e-01
3.96295816e-01 3.16918552e-01 -4.05023456e-01 -5.04942201e-02
-2.95190811e-01 -3.15847956e-02 -3.83969009e-01 8.61653626e-01
-1.96795329e-01 -3.89712721e-01 -1.14044033e-01 -1.56638384e-01
-1.58728510e-02 1.90438345e-01 -9.05678570e-01 2.13710809e+00
-3.16407681e-01 1.09733713e+00 -6.90729499e-01 -1.17215836e+00
2.66111672e-01 4.29391414e-01 9.74110126e-01 -6.61239922e-01
5.33116609e-02 -1.98967785e-01 -1.66325361e-01 -7.90146410e-01
1.61282301e-01 -2.20256373e-01 -2.73100108e-01 6.71397567e-01
8.38482797e-01 4.47059780e-01 6.22692227e-01 4.52006221e-01
1.86543155e+00 3.42544705e-01 5.56052327e-01 8.93127397e-02
-1.31713912e-01 7.41457790e-02 5.28969824e-01 7.88904190e-01
-6.57232285e-01 4.01874155e-01 5.67718685e-01 -8.93535256e-01
-9.71331418e-01 -1.06531143e+00 3.46096575e-01 1.55259395e+00
-4.59955931e-02 -7.67694473e-01 -3.77706110e-01 -1.04877162e+00
2.34501794e-01 5.45503259e-01 -1.11140490e+00 -4.49288219e-01
-6.78619802e-01 -3.22568506e-01 6.21598005e-01 1.16365230e+00
4.05247897e-01 -1.19472837e+00 -1.01812470e+00 3.87481421e-01
2.60491908e-01 -1.30155659e+00 -4.72101986e-01 1.44319683e-01
-8.78207088e-01 -1.17013729e+00 -8.34323525e-01 -5.18903196e-01
9.26606655e-02 2.02829942e-01 1.09870589e+00 -3.64310235e-01
-7.72728145e-01 8.10648680e-01 -5.30144155e-01 -1.74210101e-01
2.18120411e-01 7.37738460e-02 3.18150669e-01 1.80978831e-02
8.17186892e-01 -8.52899909e-01 -6.63614869e-01 4.93466854e-02
-5.98337412e-01 -3.01826090e-01 2.58321583e-01 6.67941391e-01
3.49365145e-01 -2.61089444e-01 6.11781716e-01 -7.55633593e-01
1.84504032e-01 -7.84518838e-01 -2.01372966e-01 1.33538753e-01
-1.20764405e-01 -2.57156640e-01 5.40485859e-01 -8.42751443e-01
-8.03606093e-01 1.83192477e-01 4.17251825e-01 -8.22114050e-01
-3.55924457e-01 4.11848485e-01 3.03771019e-01 1.31833389e-01
9.33225632e-01 2.03121960e-01 -4.62613881e-01 -4.30475622e-01
7.95431554e-01 1.93874493e-01 4.00828153e-01 -1.62849247e-01
4.97596622e-01 8.60383689e-01 -1.45398587e-01 -8.89915049e-01
-1.12633109e+00 -8.69438350e-01 -1.12725151e+00 -5.47400832e-01
1.22197902e+00 -9.51988280e-01 -4.50078428e-01 3.45596224e-01
-7.91911244e-01 -7.54070878e-01 -9.28558767e-01 7.00310647e-01
-1.05337679e+00 1.14212573e-01 -7.89991140e-01 -4.24526423e-01
4.83333506e-02 -4.54824179e-01 1.10188317e+00 -7.87512362e-02
-6.07450366e-01 -1.05956697e+00 4.42189336e-01 -1.24511542e-02
2.54128665e-01 4.91295606e-01 5.23002744e-01 -6.00962400e-01
-5.48889399e-01 -4.23700124e-01 4.44791727e-02 -2.62822360e-02
1.94234416e-01 -2.87693471e-01 -9.85342622e-01 -3.35732579e-01
-3.49901199e-01 -7.78504193e-01 1.47206831e+00 3.73466134e-01
1.17784166e+00 -1.72096550e-01 -6.33470535e-01 6.43746197e-01
1.29068911e+00 8.46330300e-02 4.86042053e-01 -1.89226251e-02
7.77703643e-01 1.96790949e-01 7.76369393e-01 6.59554005e-01
4.87100706e-02 5.57170153e-01 2.10342571e-01 3.86112481e-01
-3.43231678e-01 -2.01839581e-01 4.52638179e-01 3.81242037e-01
-3.54867995e-01 -8.55537355e-02 -9.40964162e-01 8.98838043e-01
-2.11613655e+00 -1.81088626e+00 5.85188866e-01 1.73795426e+00
6.30056500e-01 3.13518316e-01 3.97351116e-01 -2.07851633e-01
4.48985368e-01 8.47498894e-01 -8.53390455e-01 -1.34066734e-02
1.90020040e-01 2.38496497e-01 3.95341158e-01 -6.14073873e-02
-1.62518954e+00 1.15984833e+00 6.37534285e+00 6.34050608e-01
-9.83096719e-01 4.21582311e-01 4.06080574e-01 -8.17050278e-01
4.80416268e-01 -1.62624970e-01 -6.91132784e-01 3.26071978e-01
1.09512854e+00 -1.19525298e-01 2.52612680e-01 1.05489326e+00
1.35247409e-01 -6.55917451e-02 -1.71892667e+00 1.12860119e+00
2.54086047e-01 -1.77593935e+00 -8.83369520e-02 -1.62371963e-01
9.20761049e-01 4.28269684e-01 -1.92885876e-01 7.40374684e-01
2.34502763e-01 -9.90507782e-01 5.93532443e-01 7.22939551e-01
9.57337916e-01 -3.82218748e-01 3.04600418e-01 2.24781096e-01
-1.56241381e+00 -5.15087485e-01 -2.96772420e-01 -3.76517236e-01
4.20024246e-01 9.26327482e-02 -3.74525398e-01 -4.39324789e-02
8.21283340e-01 1.83299518e+00 -5.77504039e-01 1.12201512e+00
-1.06491394e-01 6.33285463e-01 -1.25001460e-01 -7.78646842e-02
6.43549442e-01 1.90544844e-01 5.21102250e-01 1.19626784e+00
2.01625943e-01 4.03574973e-01 2.18021333e-01 6.05636299e-01
-3.42338294e-01 -3.38012934e-01 -1.09772456e+00 -6.47693157e-01
1.00613371e-01 1.07080555e+00 -8.15218925e-01 -8.13450098e-01
-7.24656522e-01 1.18676233e+00 5.66111207e-01 3.19857746e-01
-1.08408451e+00 -3.80878001e-01 8.85418832e-01 2.73426652e-01
7.32876241e-01 -2.59981632e-01 1.93764701e-01 -1.39295006e+00
-6.60936162e-02 -5.13720989e-01 8.11199129e-01 -6.66797400e-01
-1.21077132e+00 1.69724777e-01 1.48306191e-01 -1.45467043e+00
-3.29220504e-01 -5.97988784e-01 -7.94439495e-01 2.18767896e-02
-1.04186952e+00 -1.06995606e+00 -3.20965320e-01 6.16576254e-01
1.04180610e+00 -2.51784086e-01 6.86380625e-01 5.11406720e-01
-4.64536756e-01 2.71950632e-01 -8.10204670e-02 3.40640604e-01
5.76396465e-01 -1.11256468e+00 6.27765894e-01 6.36442363e-01
4.79987890e-01 2.58946776e-01 4.12751079e-01 -6.53182209e-01
-1.25561833e+00 -1.28739333e+00 8.61262023e-01 -8.16496372e-01
1.03912783e+00 -6.04486644e-01 -7.62588263e-01 1.11102557e+00
-3.97606157e-02 6.73963010e-01 6.09264493e-01 3.45360748e-02
-4.65924054e-01 4.61151004e-02 -8.32831025e-01 5.29817283e-01
1.79862523e+00 -8.19144428e-01 -7.72912383e-01 5.82629085e-01
6.89342678e-01 2.58122608e-02 -8.03486347e-01 1.91399083e-01
6.36307180e-01 -6.49120927e-01 1.02143025e+00 -1.50503933e+00
4.99067873e-01 -1.61729809e-02 -1.96676344e-01 -1.29001522e+00
-4.59646881e-01 -4.99400645e-01 -8.33500683e-01 6.50440931e-01
1.60613880e-01 -7.84297884e-02 9.07229245e-01 3.89266998e-01
-1.66837618e-01 -6.90278947e-01 -9.44163740e-01 -1.21506917e+00
-2.06729844e-01 -5.38507700e-01 2.38014683e-01 1.11575902e+00
3.63831580e-01 3.87705624e-01 -5.68078220e-01 -3.40320796e-01
4.87666219e-01 -1.19422361e-01 6.49744809e-01 -1.10234344e+00
-3.46536547e-01 -3.39406013e-01 -1.07346749e+00 -8.00879419e-01
3.76096338e-01 -7.68322706e-01 -6.50884137e-02 -1.59361541e+00
4.51583773e-01 3.95898998e-01 -5.64818084e-01 7.12766826e-01
2.08283037e-01 3.58402103e-01 6.64149225e-02 -6.83308691e-02
-1.43454790e+00 7.21670449e-01 7.66688526e-01 -5.28709948e-01
-4.67547588e-02 -1.78552046e-01 4.21251319e-02 8.66131902e-01
5.68028390e-01 -6.03480339e-01 -6.63599133e-01 -3.10556561e-01
9.09183025e-02 -6.00124756e-03 5.22011757e-01 -1.49141145e+00
2.67115682e-01 -2.13870272e-01 4.58118528e-01 -5.26618659e-01
7.61466682e-01 -5.78589976e-01 -2.36547932e-01 4.85702187e-01
-6.56628132e-01 -2.90300131e-01 3.10355313e-02 1.11099100e+00
-1.91569000e-01 5.75580262e-03 6.57746971e-01 -3.81173760e-01
-1.69035769e+00 9.29431200e-01 -6.10277593e-01 3.23672354e-01
1.53393400e+00 -4.05519634e-01 -4.05609399e-01 -3.85597587e-01
-1.16902745e+00 2.11424708e-01 2.16862440e-01 5.76907992e-01
7.37324774e-01 -1.59544456e+00 -3.16708475e-01 -3.33577633e-01
5.86902797e-01 -6.82109833e-01 5.15204966e-01 1.00256252e+00
-2.80258983e-01 3.65602404e-01 -4.96967703e-01 -3.58640671e-01
-1.20199370e+00 7.81151474e-01 4.46936727e-01 3.74884671e-03
-1.02587700e+00 9.33274448e-01 -2.76687704e-02 -1.10542029e-01
4.40101862e-01 -4.89962399e-01 -2.92252272e-01 4.68587637e-01
7.53920078e-01 7.51966715e-01 -4.92907912e-01 -4.88658696e-01
-5.44066846e-01 3.72667313e-01 6.79336935e-02 -5.35223931e-02
1.67464530e+00 2.07760230e-01 3.53652090e-01 9.75285769e-01
1.60167408e+00 -9.57298458e-01 -1.76308203e+00 -2.19330400e-01
1.62277117e-01 -5.90169191e-01 -1.45984128e-01 -4.95231241e-01
-8.35985243e-01 1.24794376e+00 6.72335923e-01 -1.39928116e-02
5.70913136e-01 3.21198255e-01 1.00173628e+00 7.95574129e-01
3.57764602e-01 -1.44951248e+00 7.64419138e-01 6.45084321e-01
4.99027550e-01 -1.45907581e+00 2.40529887e-02 1.77744672e-01
-5.95468700e-01 1.07182348e+00 6.52402341e-01 -2.77300775e-01
9.45495665e-01 -6.32564798e-02 -5.62918365e-01 -7.40743041e-01
-1.14221084e+00 -3.69011283e-01 3.41923445e-01 6.32340968e-01
3.05373400e-01 -5.32036237e-02 2.51861773e-02 4.18547034e-01
5.21179914e-01 5.28554142e-01 4.73169833e-01 1.28533900e+00
-5.61976314e-01 -4.16909724e-01 6.62183642e-01 8.36598277e-01
-3.92892480e-01 1.34564461e-02 -3.38828713e-01 8.52859259e-01
1.15808249e-01 4.84255552e-01 5.48074305e-01 -3.99775684e-01
1.88766479e-01 3.65848839e-01 5.27427554e-01 -1.02293265e+00
-2.62894034e-01 -4.72585797e-01 3.63790989e-01 -1.28720546e+00
-7.26192832e-01 -8.23939204e-01 -1.17001331e+00 -1.44303456e-01
1.23606101e-01 -3.25750202e-01 1.57189310e-01 1.06819987e+00
4.36312675e-01 5.63678503e-01 2.17499763e-01 -8.74521196e-01
-5.30017197e-01 -1.01504135e+00 -8.01891088e-01 6.85591877e-01
2.76954293e-01 -8.85114312e-01 -2.23757446e-01 2.16971099e-01] | [8.646477699279785, 0.8722423911094666] |
f18ded8f-1460-439b-ab16-96a9a745062f | using-explainable-ai-to-cross-validate-socio | 2302.08605 | null | https://arxiv.org/abs/2302.08605v1 | https://arxiv.org/pdf/2302.08605v1.pdf | Using Explainable AI to Cross-Validate Socio-economic Disparities Among Covid-19 Patient Mortality | This paper applies eXplainable Artificial Intelligence (XAI) methods to investigate the socioeconomic disparities in COVID patient mortality. An Extreme Gradient Boosting (XGBoost) prediction model is built based on a de-identified Austin area hospital dataset to predict the mortality of COVID-19 patients. We apply two XAI methods, Shapley Additive exPlanations (SHAP) and Locally Interpretable Model Agnostic Explanations (LIME), to compare the global and local interpretation of feature importance. This paper demonstrates the advantages of using XAI which shows the feature importance and decisive capability. Furthermore, we use the XAI methods to cross-validate their interpretations for individual patients. The XAI models reveal that Medicare financial class, older age, and gender have high impact on the mortality prediction. We find that LIME local interpretation does not show significant differences in feature importance comparing to SHAP, which suggests pattern confirmation. This paper demonstrates the importance of XAI methods in cross-validation of feature attributions. | ['Ying Ding', 'Justin F. Rousseau', 'Jacek Gwizdka', 'Esther Melamed', 'Redoan Rahman', 'Li Shi'] | 2023-02-16 | null | null | null | null | ['mortality-prediction'] | ['medical'] | [-3.03113639e-01 1.38194099e-01 -6.28657520e-01 -4.80654210e-01
-1.98656425e-01 6.41525015e-02 4.92864013e-01 6.79036021e-01
-1.97776467e-01 1.10013723e+00 8.88031960e-01 -1.00821221e+00
-1.01577079e+00 -6.52771294e-01 -4.35819119e-01 -3.69227320e-01
-2.69002408e-01 5.20241499e-01 -8.22953105e-01 -3.09357613e-01
3.13118368e-01 3.95625561e-01 -1.36764216e+00 5.75069070e-01
9.90620494e-01 5.56626320e-01 -2.81683534e-01 5.76830149e-01
1.86052725e-01 1.16648543e+00 -2.95031875e-01 -2.50396639e-01
4.13171291e-01 -6.26959682e-01 -6.49303138e-01 -7.93654025e-01
-6.52559102e-02 -2.47351870e-01 2.57470489e-01 2.82368362e-01
3.96540195e-01 -5.78601003e-01 1.18931806e+00 -1.81195486e+00
-5.48390090e-01 8.29074919e-01 -3.84370387e-01 2.96385825e-01
3.34183663e-01 -5.37755936e-02 8.33578527e-01 -7.70470679e-01
4.34600979e-01 1.35388935e+00 1.26353645e+00 3.65695804e-01
-7.20982611e-01 -9.09284294e-01 1.38029978e-02 2.96529353e-01
-8.31159949e-01 2.22092018e-01 6.61194921e-01 -7.93403864e-01
7.09995449e-01 9.70616043e-01 7.24845171e-01 9.09177542e-01
6.07112408e-01 1.49150148e-01 1.55577075e+00 -7.24100292e-01
2.90672004e-01 2.64069498e-01 7.96041608e-01 7.56264508e-01
9.17618096e-01 5.72850227e-01 -6.15675271e-01 -7.48241067e-01
2.86048412e-01 6.14769459e-01 -6.76982328e-02 1.20306946e-01
-1.43171799e+00 1.00815153e+00 6.02898717e-01 -1.71819359e-01
-5.95144093e-01 -1.49652697e-02 3.46453428e-01 2.45928600e-01
3.90039355e-01 7.39552796e-01 -1.05610907e+00 1.78027689e-01
-4.71850544e-01 6.43184409e-02 5.49554765e-01 3.72630060e-01
3.75206083e-01 -1.88133717e-01 -1.85615003e-01 3.24097365e-01
6.55265272e-01 5.18121183e-01 5.91287374e-01 -8.40119779e-01
3.85142386e-01 1.18131745e+00 5.91405891e-02 -9.33561981e-01
-8.99478972e-01 -5.67646980e-01 -8.08291376e-01 6.93158209e-01
3.81148815e-01 -4.08539653e-01 -6.82043374e-01 1.52101552e+00
1.91676635e-02 -1.02681033e-01 9.46912095e-02 7.98717618e-01
9.02408779e-01 1.36881694e-01 6.91353381e-01 -4.37828936e-02
1.50746560e+00 -6.37066364e-01 -4.81646299e-01 -3.21090631e-02
6.81298554e-01 -1.57806024e-01 6.94157481e-01 6.62938654e-02
-7.94444561e-01 -4.77387011e-01 -8.41765106e-01 3.79846215e-01
-4.01131779e-01 3.45875807e-02 8.39330971e-01 7.45264530e-01
-6.63175404e-01 6.42308593e-01 -5.93216002e-01 -5.73728561e-01
2.43460834e-01 4.09224927e-01 -2.63059378e-01 4.65401649e-01
-1.08197522e+00 9.05771255e-01 1.29439935e-01 -3.71017307e-01
-4.48556423e-01 -1.24277377e+00 -5.78086317e-01 3.11761320e-01
-5.44725597e-01 -1.37298548e+00 3.25202316e-01 -1.31849182e+00
-6.49699450e-01 6.26918435e-01 -1.42789051e-01 -5.30935943e-01
6.88250899e-01 -7.15393275e-02 -3.19464892e-01 -7.60108456e-02
3.51551831e-01 4.45336580e-01 2.45131671e-01 -1.33304667e+00
-7.80216157e-01 -8.70139241e-01 -2.70162255e-01 5.75916097e-03
-1.47236973e-01 1.53257415e-01 6.77383423e-01 -8.72654617e-01
1.80621713e-01 -7.90557563e-01 -2.16716319e-01 -3.38065356e-01
-3.40873897e-01 -6.89536408e-02 2.70708472e-01 -8.64123404e-01
1.27098048e+00 -2.03063035e+00 -3.87691855e-01 4.57992822e-01
3.58833462e-01 -4.47351426e-01 5.20936847e-01 3.24507475e-01
-5.92017591e-01 4.81337756e-01 -4.44723107e-02 4.00466397e-02
-3.16915363e-01 2.61144876e-01 -2.55145550e-01 1.30594805e-01
2.17440367e-01 8.33815038e-01 -6.85215712e-01 -6.28784657e-01
1.97635412e-01 1.73763990e-01 -7.23622084e-01 1.26776308e-01
5.00359356e-01 4.96210068e-01 -3.51502031e-01 7.31280863e-01
4.24079001e-01 -2.72126049e-01 2.06362978e-01 -1.64973345e-02
-1.56385988e-01 -2.12830724e-04 -5.05156457e-01 8.03154826e-01
5.57236373e-03 3.01104367e-01 -5.15153468e-01 -7.71789730e-01
1.04514968e+00 3.20924282e-01 4.70159680e-01 -2.51677245e-01
5.33132441e-02 4.20966536e-01 2.45195836e-01 -3.73071611e-01
-8.45256373e-02 -3.80530119e-01 1.28386497e-01 2.72725970e-01
-6.17254496e-01 6.78701520e-01 -4.95943248e-01 1.31739378e-01
8.84742200e-01 -8.88758376e-02 8.63164604e-01 -9.84488845e-01
4.09982532e-01 4.13637102e-01 8.53604257e-01 8.91906738e-01
-2.31695130e-01 7.96324611e-01 4.53090012e-01 -1.09692752e+00
-7.44832456e-01 -9.48080301e-01 -6.12088144e-01 8.44596088e-01
1.31672367e-01 -1.17756948e-01 -3.44539553e-01 -8.47261369e-01
5.66986740e-01 9.50726569e-01 -1.36608481e+00 -1.97525218e-01
-2.20261082e-01 -1.24611330e+00 8.99874568e-02 8.51808012e-01
3.73244554e-01 -8.13348353e-01 -1.23584342e+00 -2.62258165e-02
1.99748188e-01 -4.27685352e-03 4.07731503e-01 3.49279046e-01
-1.15346122e+00 -1.19287980e+00 -4.68789071e-01 -2.35089734e-01
7.11265624e-01 4.86096330e-02 1.46293938e+00 6.79318190e-01
-2.36471355e-01 2.82500267e-01 -3.38961661e-01 -1.12452221e+00
-5.30130923e-01 -7.15955123e-02 8.23161006e-02 -4.21105921e-01
6.11942768e-01 -2.70138621e-01 -9.91156697e-01 3.57859910e-01
-2.60397792e-01 3.44471604e-01 3.55896115e-01 8.38946879e-01
1.97069377e-01 -4.32396561e-01 9.55037117e-01 -9.97868538e-01
3.73641878e-01 -1.09657621e+00 1.38490200e-01 3.86349231e-01
-1.58959162e+00 7.19588995e-02 2.54651159e-01 1.24067284e-01
-9.54989910e-01 -3.43397290e-01 4.49367255e-01 3.69177870e-02
-1.24368206e-01 6.31619573e-01 2.37483248e-01 3.76152009e-01
9.07010138e-01 -6.58228457e-01 1.75043046e-01 -5.66336393e-01
-4.44135040e-01 6.77911103e-01 1.89833060e-01 -4.70908344e-01
2.96525478e-01 4.95941639e-01 2.48868242e-01 2.70507764e-02
-6.07472956e-01 -2.01529339e-01 -5.29410183e-01 -1.74725831e-01
1.05375898e+00 -9.69768703e-01 -8.09812248e-01 -2.19498992e-01
-5.99702299e-01 -1.03781097e-01 -2.19808728e-01 1.08035898e+00
-4.21538770e-01 -8.66634101e-02 -2.71057606e-01 -8.44466686e-01
-6.33462548e-01 -1.08264542e+00 6.68420136e-01 7.64481798e-02
-9.14759636e-01 -1.06183946e+00 2.16194749e-01 5.62656343e-01
3.42412084e-01 7.39337444e-01 1.51151240e+00 -8.04529369e-01
2.65786028e-03 -1.71066020e-02 -4.40214008e-01 -4.23204631e-01
1.58018991e-01 2.00791016e-01 -8.85085046e-01 3.75145264e-02
-3.47016066e-01 2.71764964e-01 6.90260053e-01 8.22979867e-01
9.03547347e-01 -4.39805418e-01 -5.21416008e-01 6.30031943e-01
1.48740482e+00 4.95442927e-01 2.72987396e-01 1.03063941e+00
4.95170087e-01 7.13929713e-01 4.75994378e-01 6.71666980e-01
5.09588838e-01 3.96333933e-01 5.53763270e-01 -6.60021067e-01
-6.98240995e-02 -1.72430545e-01 -6.37556911e-02 1.37696043e-01
-6.69398248e-01 3.98339242e-01 -1.44035327e+00 3.54802370e-01
-2.03341746e+00 -8.47578108e-01 -6.29544497e-01 1.99776912e+00
2.70357579e-01 -1.98658061e-04 9.86442789e-02 2.80636936e-01
3.23602885e-01 -6.05924964e-01 -3.25007111e-01 -8.24677765e-01
-1.00306287e-01 -1.83217600e-01 5.61410129e-01 3.79297674e-01
-6.85886562e-01 -1.59780219e-01 6.99057055e+00 -9.68830064e-02
-7.53775060e-01 8.58365446e-02 1.34412813e+00 4.04569060e-02
-4.46220309e-01 1.17069580e-01 -3.12328428e-01 6.95220351e-01
1.05253971e+00 -1.72465771e-01 6.57981914e-03 1.02279592e+00
4.70958710e-01 8.98169205e-02 -9.44637418e-01 6.11993253e-01
-1.68049589e-01 -1.30079317e+00 1.21711440e-01 1.34127468e-01
6.70639813e-01 -1.77625567e-01 -5.18104993e-02 1.75131902e-01
3.40382814e-01 -1.02320063e+00 8.82009625e-01 8.31196010e-01
6.63355529e-01 -7.08144188e-01 1.28854835e+00 -9.22754779e-02
-5.70954978e-01 -6.83872104e-01 -2.30063066e-01 -7.05919981e-01
-2.67936051e-01 4.33272541e-01 -1.11651087e+00 7.49360502e-01
1.29579604e+00 4.24928367e-01 -7.57209778e-01 7.52067804e-01
4.37370203e-02 1.06591845e+00 1.11122154e-01 2.31060937e-01
-1.15270294e-01 -2.62799323e-01 1.57816604e-01 1.00681865e+00
5.13045788e-01 3.08526546e-01 -3.74349952e-01 6.20558202e-01
2.83975661e-01 3.05861741e-01 -5.50617635e-01 7.72419333e-01
2.84246832e-01 6.23097837e-01 -5.64389229e-01 -7.01265514e-01
-5.31856120e-01 4.97979671e-01 -5.64930812e-02 1.61292881e-01
-7.05809772e-01 1.23920001e-01 7.33842015e-01 4.18856025e-01
-4.01170433e-01 6.27039909e-01 -1.50808096e+00 -9.23151553e-01
-6.58491015e-01 -8.38413656e-01 1.00285125e+00 -1.10501039e+00
-1.31766236e+00 3.74341100e-01 3.03791523e-01 -1.14798129e+00
-1.02150001e-01 -5.94307244e-01 -8.19297314e-01 1.06238663e+00
-1.29907155e+00 -1.08176017e+00 -7.77338028e-01 4.68074411e-01
2.53645539e-01 -4.71691400e-01 1.03302824e+00 -1.60623312e-01
-6.94950283e-01 3.85490119e-01 3.49325866e-01 -1.18935615e-01
6.13791883e-01 -1.34271979e+00 -1.69941604e-01 2.28374019e-01
-5.27058482e-01 7.14806199e-01 6.62191689e-01 -1.00283384e+00
-6.62475884e-01 -9.62285697e-01 1.09821999e+00 -8.53057265e-01
7.27004409e-02 4.56917107e-01 -5.39724886e-01 7.02460885e-01
2.60447562e-01 -4.95114118e-01 1.18792844e+00 4.90132451e-01
-1.66400641e-01 -1.87735587e-01 -1.37739050e+00 4.32371974e-01
8.17506552e-01 2.03476459e-01 -8.21267307e-01 2.46132225e-01
3.76129389e-01 4.64174598e-01 -1.03601646e+00 9.28917527e-01
9.22380090e-01 -1.27465224e+00 1.15425968e+00 -1.23350036e+00
7.67899632e-01 -5.36441691e-02 -1.38666406e-01 -9.08895254e-01
-8.56371105e-01 1.05861790e-01 4.96696383e-01 9.12160337e-01
1.04299974e+00 -1.02825630e+00 6.29335284e-01 1.38757908e+00
1.03914782e-01 -9.10500288e-01 -6.61239088e-01 -3.85616869e-01
1.97617456e-01 -1.50664762e-01 1.29383898e+00 1.46561730e+00
4.41304594e-01 -2.70960499e-02 -2.47339725e-01 1.38474688e-01
5.32867670e-01 3.35848719e-01 6.47380769e-01 -1.71659863e+00
-2.20976591e-01 -3.08970094e-01 -5.27335405e-01 6.39447868e-01
-1.71879753e-01 -8.38467658e-01 -8.62125337e-01 -1.53358507e+00
8.67669404e-01 -7.49443889e-01 -8.64268303e-01 8.00394058e-01
-8.21921587e-01 -1.32695604e-02 2.33066946e-01 7.83174932e-01
1.78421557e-01 9.20830220e-02 4.09037769e-01 -1.12733126e-01
-1.92505732e-01 -2.52372921e-02 -9.93041277e-01 8.14671516e-01
1.15143418e+00 -6.41107798e-01 -3.81152153e-01 -3.40284854e-01
2.45162651e-01 2.09471881e-01 7.03090608e-01 -9.31092083e-01
-4.52646464e-01 -4.26981211e-01 9.35778618e-01 -3.31779122e-01
-2.73924351e-01 -1.11604822e+00 7.09678113e-01 1.12480640e+00
-4.91190195e-01 6.59543276e-01 5.29110245e-02 3.34637970e-01
2.11349949e-01 1.16666257e-02 2.01339528e-01 -1.04048312e-01
-8.48788247e-02 -2.29734004e-01 -2.09125102e-01 -3.13975275e-01
1.02544641e+00 -4.05344963e-01 -3.88014823e-01 -3.13853115e-01
-6.30306721e-01 -1.07933566e-01 5.93920529e-01 1.14307210e-01
3.05124789e-01 -1.25019205e+00 -1.06584096e+00 2.18708575e-01
2.91659564e-01 -7.36732304e-01 -1.77532490e-02 1.04594088e+00
-8.93887758e-01 3.93767476e-01 -5.95673740e-01 -1.50885165e-01
-1.38308465e+00 9.22564089e-01 3.15635651e-01 -7.61650354e-02
-7.28246808e-01 4.58856374e-01 5.31814516e-01 -2.20170841e-01
-1.46486133e-01 -3.12395245e-01 -3.99315566e-01 -2.27309003e-01
3.85029882e-01 1.03610146e+00 -2.34125406e-01 -3.33924741e-01
-6.55781448e-01 4.61196154e-01 4.03499633e-01 1.27628192e-01
1.32076359e+00 -9.79147851e-02 -1.77748069e-01 2.86369979e-01
9.13212180e-01 6.86759576e-02 -7.37285316e-01 4.62815642e-01
1.81633145e-01 -3.90922695e-01 -2.12115541e-01 -1.40073001e+00
-7.76684701e-01 5.42947829e-01 9.45175529e-01 2.76423339e-02
1.18171203e+00 -1.73020795e-01 -1.46137312e-01 -8.38768482e-02
7.28944987e-02 -5.27707160e-01 -7.39199877e-01 -3.23489189e-01
9.26041722e-01 -1.50628960e+00 2.70211816e-01 -8.75690430e-02
-8.15760612e-01 1.15407586e+00 3.20831686e-01 1.88709587e-01
9.16756272e-01 -9.95143801e-02 4.32297319e-01 -3.30095917e-01
-7.67129004e-01 4.46689039e-01 2.90279090e-01 3.55759621e-01
4.85446930e-01 6.74128771e-01 -9.64882374e-01 8.99368763e-01
-5.05949378e-01 1.23095721e-01 2.76915222e-01 5.41687369e-01
-9.64004919e-02 -6.72234416e-01 -8.58342826e-01 9.48071241e-01
-6.60517693e-01 -1.77265286e-01 -6.62156224e-01 1.07994032e+00
2.89220065e-01 9.81904924e-01 7.28952512e-02 -3.83417368e-01
-5.57901114e-02 3.76334459e-01 -3.97133261e-01 1.67189240e-01
-9.87479866e-01 -5.34839511e-01 6.41987473e-02 -2.36267701e-01
-4.33418512e-01 -5.89301586e-01 -1.27004647e+00 -2.88438708e-01
-8.04956257e-02 5.19451976e-01 8.36308777e-01 6.54853761e-01
6.94604278e-01 5.10714650e-01 7.70075321e-01 -1.80944860e-01
-2.38344401e-01 -9.37858343e-01 -4.09630984e-01 8.18025351e-01
2.48395383e-01 -6.47236705e-01 -7.38354623e-01 -2.78404113e-02] | [8.237771987915039, 5.801236629486084] |
69dab021-d675-4325-8d7c-1c946615abd7 | warped-linear-models-for-time-series | 1711.09156 | null | http://arxiv.org/abs/1711.09156v1 | http://arxiv.org/pdf/1711.09156v1.pdf | Warped-Linear Models for Time Series Classification | This article proposes and studies warped-linear models for time series
classification. The proposed models are time-warp invariant analogues of linear
models. Their construction is in line with time series averaging and extensions
of k-means and learning vector quantization to dynamic time warping (DTW)
spaces. The main theoretical result is that warped-linear models correspond to
polyhedral classifiers in Euclidean spaces. This result simplifies the analysis
of time-warp invariant models by reducing to max-linear functions. We exploit
this relationship and derive solutions to the label-dependency problem and the
problem of learning warped-linear models. Empirical results on time series
classification suggest that warped-linear functions better trade solution
quality against computation time than nearest-neighbor and prototype-based
methods. | ['Brijnesh J. Jain'] | 2017-11-24 | null | null | null | null | ['time-series-averaging'] | ['time-series'] | [ 1.75444081e-01 -5.55685818e-01 -4.88697618e-01 -5.13513267e-01
-8.81745934e-01 -1.00455761e+00 8.48761678e-01 1.65226102e-01
-5.80914319e-01 5.84276140e-01 2.48883307e-01 -1.39595106e-01
-8.49148631e-01 -6.62205577e-01 -2.01701373e-01 -9.23458397e-01
-9.23715591e-01 3.92020643e-01 1.85720354e-01 -3.50377142e-01
5.03478587e-01 1.03311563e+00 -1.10603046e+00 1.92044586e-01
3.94271195e-01 8.51557374e-01 -4.20565397e-01 9.52389061e-01
4.89156991e-02 2.76905000e-01 -3.42513591e-01 1.35835409e-01
4.77673382e-01 -1.84117764e-01 -5.86790204e-01 1.82219557e-02
3.14509541e-01 2.38500059e-01 -6.73461497e-01 7.39534736e-01
2.97885090e-01 7.26066470e-01 9.15478289e-01 -1.92189646e+00
-7.85457075e-01 3.47157300e-01 -1.71645969e-01 3.92603904e-01
7.78724626e-02 -5.14252484e-01 8.29964697e-01 -8.10222745e-01
5.79108536e-01 1.37008536e+00 1.13391793e+00 3.59038040e-02
-1.43380117e+00 -3.99904102e-01 -1.25454769e-01 6.71318471e-01
-1.43838131e+00 -1.16572089e-01 7.35707998e-01 -5.10083795e-01
1.09860253e+00 6.06810451e-01 6.90119207e-01 8.73887122e-01
8.82937133e-01 3.61229867e-01 1.15670526e+00 -5.27799010e-01
3.79784286e-01 -4.67206895e-01 4.46014613e-01 4.99854416e-01
-2.13511318e-01 1.47128627e-01 -1.22740738e-01 -4.54778939e-01
6.42745674e-01 4.65719461e-01 -1.78735927e-01 -7.54077077e-01
-1.56426704e+00 1.05037785e+00 3.90784293e-01 8.77141535e-01
-2.84261554e-01 1.81079388e-01 7.31385529e-01 7.24367619e-01
7.27241457e-01 6.85028255e-01 -4.86294419e-01 -1.62445977e-01
-1.23876762e+00 3.02736014e-01 5.74326813e-01 7.89185107e-01
3.38545829e-01 4.06962812e-01 -9.61761624e-02 6.59088612e-01
-2.17252761e-01 5.88808596e-01 9.39643443e-01 -1.02434397e+00
1.58984270e-02 1.58463553e-01 -1.06157791e-02 -1.12659287e+00
-8.58843625e-01 -9.29477960e-02 -8.00760150e-01 -4.40293513e-02
3.39661241e-01 2.72223145e-01 -7.39728510e-01 1.27011323e+00
-4.88478690e-02 5.29713273e-01 -6.99549615e-02 6.03778303e-01
-1.32992238e-01 1.12699401e+00 -2.00544670e-01 -9.68193412e-01
8.59727919e-01 -9.87613380e-01 -9.48298097e-01 5.54334998e-01
7.31120944e-01 -7.38295615e-01 6.46000028e-01 5.48951089e-01
-7.34142303e-01 -6.35671198e-01 -1.15696347e+00 1.60572022e-01
-6.49011135e-01 -3.23741734e-01 4.36814696e-01 5.32675982e-01
-1.15064991e+00 1.20133841e+00 -1.00202179e+00 -5.96681118e-01
-1.02591909e-01 2.53577232e-01 -5.23889840e-01 1.01228490e-01
-1.03421211e+00 1.06555343e+00 4.51194853e-01 -1.15131304e-01
-5.65036416e-01 -5.41227996e-01 -7.25614071e-01 -4.51091439e-01
-2.16490492e-01 1.32475436e-01 1.13881373e+00 -7.64382660e-01
-1.31920505e+00 3.70986730e-01 -3.47216636e-01 -7.85707474e-01
4.72464383e-01 6.11979980e-03 -8.46749902e-01 8.73593986e-02
-1.17490999e-01 3.33198994e-01 1.17552292e+00 -7.82948971e-01
-4.50855970e-01 -2.97556758e-01 -2.60679901e-01 -2.47874677e-01
-5.58483303e-01 -5.89469522e-02 1.48565546e-01 -9.99504507e-01
3.89777362e-01 -1.33387613e+00 -4.45416838e-01 1.87263295e-01
1.56212524e-01 -3.87641758e-01 1.21161067e+00 -6.68435395e-01
1.40396392e+00 -1.88146973e+00 3.51973087e-01 5.30149519e-01
-1.89873166e-02 6.63597882e-02 -3.93804312e-01 7.81794071e-01
-6.64598346e-01 -1.54327929e-01 -3.27130377e-01 8.51692930e-02
-6.12571277e-02 6.95712388e-01 -7.93642104e-01 1.03753960e+00
-7.58985505e-02 7.95484841e-01 -9.67083335e-01 -6.76697910e-01
6.49380147e-01 5.87575674e-01 -5.40141389e-02 -2.90807992e-01
3.28666627e-01 6.65917695e-02 -2.04392318e-02 4.67756897e-01
5.09462535e-01 4.12307560e-01 -1.27191573e-01 -4.09823418e-01
-5.02578259e-01 -3.47109169e-01 -9.39706147e-01 1.64959025e+00
-5.61050475e-01 8.60929251e-01 -5.90939999e-01 -1.42975509e+00
1.18328369e+00 3.97036314e-01 1.10519171e+00 -3.93430859e-01
1.76002890e-01 1.02355234e-01 -3.74445140e-01 -4.88148361e-01
5.15241385e-01 -2.60671467e-01 -3.54060173e-01 1.21052958e-01
2.64145941e-01 -4.77201492e-01 9.41062346e-02 -3.53644282e-01
9.10443485e-01 2.08342671e-01 6.06012881e-01 -4.73342597e-01
4.75942105e-01 1.85222164e-01 4.07873690e-01 4.49945897e-01
-4.14384961e-01 3.82970929e-01 1.35523662e-01 -7.78077424e-01
-1.33787858e+00 -1.24444509e+00 -3.82111937e-01 1.25114608e+00
-2.85179347e-01 -5.37894726e-01 -4.05754298e-01 -3.39353412e-01
-7.46351704e-02 9.59322035e-01 -7.60437131e-01 -2.26961076e-01
-9.18820500e-01 -6.01541817e-01 5.18419683e-01 5.83746731e-01
-1.35388419e-01 -5.23347795e-01 -5.06542206e-01 5.78804553e-01
-1.21744350e-01 -6.23919189e-01 -7.98155665e-01 3.19603324e-01
-1.27713943e+00 -7.72037506e-01 -8.19910228e-01 -7.88890541e-01
4.16876376e-01 3.93915325e-01 4.37251121e-01 -9.45039809e-01
-2.44357005e-01 6.97622180e-01 -7.02967346e-01 -3.99277955e-01
-1.74610361e-01 -2.55486399e-01 4.80609953e-01 4.60673589e-03
4.11465257e-01 -6.71557128e-01 -2.11928442e-01 6.74156487e-01
-7.39270389e-01 -7.00709462e-01 6.54321164e-03 8.82124484e-01
9.25920963e-01 3.41412008e-01 3.11190367e-01 -8.53713751e-02
7.15521038e-01 -3.20165515e-01 -6.95336998e-01 4.98969048e-01
-8.42042983e-01 2.17165813e-01 7.47580647e-01 -9.49060082e-01
-2.16658041e-01 3.39917392e-01 3.09704036e-01 -1.00322509e+00
3.04643363e-01 1.40887082e-01 5.41239500e-01 -4.95166898e-01
7.80304074e-01 3.88234496e-01 -3.43140401e-02 -2.18231544e-01
7.52931237e-01 6.75093055e-01 6.04325294e-01 -3.63640457e-01
8.70896280e-01 7.22897708e-01 3.52787435e-01 -1.21542144e+00
-3.45252991e-01 -5.84316254e-01 -1.20892704e+00 -4.23652589e-01
9.11065161e-01 -4.20014709e-01 -3.23575795e-01 1.40910521e-01
-1.03731358e+00 7.55649805e-02 -6.61096573e-01 8.08084309e-01
-1.11204278e+00 5.06762683e-01 -4.06473577e-01 -8.11957777e-01
-1.19561903e-01 -7.59709537e-01 9.06921566e-01 -1.92185998e-01
-5.36401093e-01 -1.41762602e+00 5.74243188e-01 -3.24000120e-01
4.75284457e-01 6.77750349e-01 8.30321670e-01 -8.22595716e-01
4.06421334e-01 -7.43787706e-01 4.27616447e-01 3.39988917e-01
-7.13935941e-02 6.81551248e-02 -7.53593683e-01 -5.20413876e-01
3.14534664e-01 1.91765398e-01 5.34420550e-01 6.34592056e-01
1.24988329e+00 -3.48783046e-01 -4.43768173e-01 5.67581177e-01
1.38876891e+00 6.71117485e-01 2.59208798e-01 2.77139485e-01
5.90368271e-01 5.83223701e-01 8.34634423e-01 3.91776681e-01
1.03200659e-01 9.25707817e-01 9.27805528e-02 3.31445962e-01
4.32506621e-01 -4.68599536e-02 4.13294435e-01 1.54275465e+00
-2.44536966e-01 6.57161400e-02 -9.69674587e-01 7.21750557e-01
-1.93659115e+00 -1.20930493e+00 -2.98016906e-01 2.22883725e+00
5.90165854e-01 -1.35689721e-01 3.09811950e-01 7.21652091e-01
8.07327867e-01 2.86385357e-01 -2.71481574e-01 -8.40950251e-01
-2.05138862e-01 3.25657964e-01 8.81178319e-01 7.42187798e-01
-1.50466073e+00 5.37009120e-01 7.35061741e+00 1.01385069e+00
-1.40474081e+00 3.77685249e-01 -4.40908670e-02 -4.25166450e-02
1.49232805e-01 -1.54016405e-01 -3.08536112e-01 3.18594098e-01
1.44626164e+00 -9.36935902e-01 5.63116074e-01 8.97014797e-01
4.05884922e-01 6.81464195e-01 -1.30565977e+00 1.39762425e+00
2.96250373e-01 -1.06541419e+00 -5.10893688e-02 -1.22800395e-01
7.95950532e-01 -3.87428701e-02 7.48015121e-02 2.31736064e-01
1.39171347e-01 -8.40348005e-01 6.70738399e-01 7.91260660e-01
8.09484243e-01 -7.96286225e-01 5.10418236e-01 -8.12344924e-02
-1.64016747e+00 -4.99094009e-01 -4.99412745e-01 6.56229481e-02
3.33112597e-01 2.29928732e-01 -5.09867311e-01 6.34858072e-01
2.74935424e-01 1.14044130e+00 -3.14499229e-01 1.25412047e+00
3.37216079e-01 4.04462039e-01 -2.34496310e-01 1.32865757e-01
6.39850557e-01 -3.79529387e-01 5.62250912e-01 1.39233255e+00
7.09651411e-01 6.38058558e-02 3.40882778e-01 -1.29289562e-02
7.45079696e-01 1.36803165e-01 -8.01904857e-01 -1.51197761e-01
5.91701210e-01 1.00988007e+00 -8.01322103e-01 -3.40236038e-01
-2.82509834e-01 9.66047525e-01 -3.63849968e-01 7.18000755e-02
-8.46612751e-01 -7.77700484e-01 5.54141700e-01 -1.75919686e-03
1.73891813e-01 -9.45597112e-01 -2.45074127e-02 -8.72289658e-01
-2.39164740e-01 -5.19321740e-01 6.42206609e-01 -6.95154130e-01
-1.59440863e+00 6.97696090e-01 5.65067649e-01 -1.98507321e+00
-4.33569580e-01 -7.22535253e-01 -3.00242424e-01 3.89538288e-01
-1.06492317e+00 -1.01325846e+00 -5.54163828e-02 1.03978693e+00
8.05209339e-01 -5.14699295e-02 1.19238937e+00 1.34143472e-01
9.21670943e-02 3.70658845e-01 8.34965587e-01 -7.93474242e-02
7.07031965e-01 -1.28409564e+00 3.56509984e-01 6.95149779e-01
2.86192954e-01 2.77019471e-01 7.25409269e-01 -1.73384041e-01
-1.33703864e+00 -1.50660896e+00 9.88517702e-01 -5.13704658e-01
1.18994629e+00 4.39434908e-02 -9.53480005e-01 6.24600589e-01
-1.12203859e-01 2.26005480e-01 8.12884212e-01 -4.51819927e-01
-5.57014585e-01 -2.17697710e-01 -1.19482636e+00 3.61039132e-01
9.84819174e-01 -8.52962196e-01 -9.46867526e-01 7.44034827e-01
5.97630978e-01 9.97694433e-02 -1.31531239e+00 2.70510241e-02
7.42792845e-01 -3.49601716e-01 1.24766171e+00 -8.04074109e-01
-1.89809665e-01 -2.35016361e-01 -3.63518655e-01 -1.61383414e+00
-6.97803020e-01 -7.96958566e-01 1.30573705e-01 5.79852402e-01
-1.40216872e-01 -7.78249741e-01 2.30945811e-01 1.02339178e-01
-2.28308603e-01 -5.30452073e-01 -1.46955764e+00 -1.45755756e+00
4.20647591e-01 -5.04287124e-01 2.50746250e-01 1.45320368e+00
3.65547329e-01 -2.03682050e-01 -4.15319920e-01 -8.34476948e-02
7.35090077e-01 1.20012701e-01 9.61191356e-02 -1.35571682e+00
1.16660580e-01 -4.30621326e-01 -1.07865083e+00 -5.15868306e-01
4.86851156e-01 -1.08517432e+00 -1.82420850e-01 -1.11699796e+00
-6.43167973e-01 -3.13033938e-01 -4.94969517e-01 6.05912089e-01
5.64229548e-01 2.76079237e-01 2.33899534e-01 3.61593425e-01
-3.18418294e-01 3.01584244e-01 7.74319947e-01 -2.03428924e-01
-3.66022527e-01 -8.34291428e-02 4.40849930e-01 6.71238482e-01
8.39391708e-01 -4.30317879e-01 -5.44011712e-01 -9.00554806e-02
-1.86186597e-01 3.09729159e-01 1.20797262e-01 -1.21416557e+00
2.38442302e-01 -3.72117579e-01 1.56450942e-01 -5.54363906e-01
5.03051937e-01 -1.22147942e+00 2.68648177e-01 7.12319970e-01
-6.17294729e-01 7.70807564e-01 2.39364132e-02 6.16427720e-01
-2.82941610e-01 -8.53127763e-02 9.28035796e-01 6.39345199e-02
-8.76006305e-01 2.58118659e-01 -5.19614220e-01 -3.47846955e-01
1.39844263e+00 -4.83742595e-01 -1.54515818e-01 -3.67231339e-01
-1.08146524e+00 -4.60515022e-02 5.28610721e-02 8.03698063e-01
5.38882673e-01 -1.71642017e+00 -6.80104673e-01 8.85230750e-02
1.45537660e-01 -8.16630483e-01 1.50724808e-02 8.68011475e-01
-7.24297285e-01 8.11416447e-01 -7.95527518e-01 -6.56227231e-01
-1.46998072e+00 9.50085461e-01 3.66084844e-01 3.09577119e-02
-5.65166831e-01 5.10978639e-01 -5.85964203e-01 -3.56697768e-01
1.60753578e-01 -5.77234089e-01 -1.66128218e-01 6.12809718e-01
5.52340627e-01 9.91586208e-01 5.86476251e-02 -1.00372612e+00
-4.47531015e-01 1.16338348e+00 1.54643640e-01 -2.85279602e-01
1.51922452e+00 6.20926209e-02 -2.96055019e-01 1.15623975e+00
1.81853974e+00 -6.74832225e-01 -8.18998754e-01 -4.43312705e-01
2.06304148e-01 -6.08505428e-01 -6.72581652e-03 -1.93164229e-01
-4.64518875e-01 9.73421335e-01 1.00998724e+00 4.70151633e-01
1.11735272e+00 -4.63166356e-01 5.83976209e-01 6.69561386e-01
6.57561183e-01 -1.13750434e+00 -1.44720465e-01 5.03845811e-01
1.14893055e+00 -5.51002264e-01 -3.76245454e-02 2.30762847e-02
-3.79903644e-01 1.67240536e+00 -1.89464062e-01 -5.05514145e-01
1.03210080e+00 2.79885739e-01 -8.28729495e-02 3.70620131e-01
-7.82097459e-01 8.80407095e-02 3.07574719e-01 8.43554616e-01
2.04391211e-01 3.89297694e-01 -6.85303330e-01 8.75300840e-02
-5.48127234e-01 -5.45104183e-02 4.41447735e-01 9.56407726e-01
-4.08252895e-01 -8.94641221e-01 -7.34889746e-01 2.24053562e-01
4.67452928e-02 4.38740581e-01 -5.19161969e-02 7.10639954e-01
3.72671373e-02 9.89822805e-01 3.63124996e-01 -6.87756240e-01
5.05410969e-01 6.18308961e-01 5.72566450e-01 -1.56775247e-02
-3.67815554e-01 1.01630740e-01 -2.81222105e-01 -8.13987792e-01
-5.36215484e-01 -8.25572252e-01 -1.04329908e+00 -3.68880481e-01
-2.43167758e-01 2.06817284e-01 9.94321644e-01 6.15877569e-01
-8.49640369e-02 3.35960984e-01 9.58090484e-01 -1.21838176e+00
-9.79907453e-01 -6.32101417e-01 -6.36563480e-01 2.32851461e-01
4.80658293e-01 -4.86772001e-01 -3.86666030e-01 3.87948275e-01] | [7.298186779022217, 3.308553457260132] |
61098973-c06f-4892-96ea-2209b151f397 | dialog-simulation-with-realistic-variations | 2011.08243 | null | https://arxiv.org/abs/2011.08243v1 | https://arxiv.org/pdf/2011.08243v1.pdf | Dialog Simulation with Realistic Variations for Training Goal-Oriented Conversational Systems | Goal-oriented dialog systems enable users to complete specific goals like requesting information about a movie or booking a ticket. Typically the dialog system pipeline contains multiple ML models, including natural language understanding, state tracking and action prediction (policy learning). These models are trained through a combination of supervised or reinforcement learning methods and therefore require collection of labeled domain specific datasets. However, collecting annotated datasets with language and dialog-flow variations is expensive, time-consuming and scales poorly due to human involvement. In this paper, we propose an approach for automatically creating a large corpus of annotated dialogs from a few thoroughly annotated sample dialogs and the dialog schema. Our approach includes a novel goal-sampling technique for sampling plausible user goals and a dialog simulation technique that uses heuristic interplay between the user and the system (Alexa), where the user tries to achieve the sampled goal. We validate our approach by generating data and training three different downstream conversational ML models. We achieve 18 ? 50% relative accuracy improvements on a held-out test set compared to a baseline dialog generation approach that only samples natural language and entity value variations from existing catalogs but does not generate any novel dialog flow variations. We also qualitatively establish that the proposed approach is better than the baseline. Moreover, several different conversational experiences have been built using this method, which enables customers to have a wide variety of conversations with Alexa. | ['Dilek Hakkani-Tur', 'Suranjit Adhikari', 'Charlie Shucheng Zhu', 'Tagyoung Chung', 'Angeliki Metallinou', 'Matt Zhao', 'Shubhra Chandra', 'Nehal Belgamwar', 'Maryam Fazel-Zarandi', 'Arijit Biswas', 'Daniel Elkind', 'Vincent Auvray', 'Chien-Wei Lin'] | 2020-11-16 | null | null | null | null | ['goal-oriented-dialog'] | ['natural-language-processing'] | [-4.30257954e-02 6.96593344e-01 -4.64518443e-02 -7.58053303e-01
-8.12181175e-01 -9.29973364e-01 1.04434264e+00 1.01476684e-01
-1.44227222e-01 1.17767537e+00 5.25819540e-01 -2.67963558e-01
2.96722233e-01 -6.24198735e-01 -7.31540099e-02 -1.31225258e-01
2.56753594e-01 1.29494870e+00 3.84236455e-01 -7.34320879e-01
3.76471728e-01 2.26995498e-01 -1.20835137e+00 4.75767404e-01
8.52935314e-01 6.64551556e-01 3.32229257e-01 9.23232496e-01
-7.04048455e-01 1.15611196e+00 -9.50201213e-01 -3.66564065e-01
1.84680551e-01 -8.76960874e-01 -1.24339104e+00 5.02060056e-01
-4.28686067e-02 -6.85121059e-01 1.90243721e-01 5.95373333e-01
3.66738588e-01 4.29166347e-01 3.82859111e-01 -1.40189874e+00
-1.97332785e-01 7.43994117e-01 3.42538744e-01 -3.59996617e-01
9.10420179e-01 6.35320544e-01 1.00069547e+00 -3.45168948e-01
7.10163116e-01 1.64626086e+00 2.35020891e-01 1.01478219e+00
-1.29770648e+00 -2.54897326e-01 1.04833759e-01 -2.97427386e-01
-7.07095385e-01 -5.73297560e-01 6.86608791e-01 -3.77754152e-01
9.64949012e-01 1.66606680e-01 3.38224888e-01 1.31538749e+00
-1.66840747e-01 7.43242264e-01 1.17198801e+00 -5.98608673e-01
5.52992761e-01 9.16779518e-01 3.60477895e-01 4.75465894e-01
-4.98447120e-01 -2.54876502e-02 -3.56554329e-01 -5.03337145e-01
5.89579046e-01 -3.69074136e-01 -3.21523510e-02 -1.73069462e-01
-9.59770560e-01 1.06657755e+00 -1.64093420e-01 2.87927777e-01
-3.36293370e-01 -4.03112769e-01 3.62942725e-01 4.91863549e-01
3.90130132e-02 8.95254195e-01 -5.14219761e-01 -4.32975441e-01
-4.37444270e-01 7.23734438e-01 1.87090468e+00 1.29439485e+00
6.26519918e-01 -1.54974356e-01 -5.10287821e-01 9.28174794e-01
4.02649373e-01 1.55744240e-01 4.51256186e-01 -1.28911793e+00
4.30597663e-01 9.54674065e-01 8.99999201e-01 -2.54182994e-01
-3.81143928e-01 6.16156399e-01 -1.27829686e-01 1.68601200e-01
1.05025232e+00 -7.22846389e-01 -3.34235430e-01 1.62223470e+00
4.15721625e-01 -3.36148053e-01 4.49529290e-01 7.07641602e-01
8.97661388e-01 7.07235217e-01 2.32473612e-01 -4.73194480e-01
1.43172777e+00 -1.06246412e+00 -1.05098557e+00 -1.77988976e-01
5.88151336e-01 -6.65815115e-01 1.48541415e+00 4.58764941e-01
-1.31438851e+00 -5.06525278e-01 -5.79324901e-01 1.46380737e-01
-2.30989695e-01 -1.70015052e-01 5.94457150e-01 7.37395525e-01
-1.00630581e+00 3.38443160e-01 -3.14780861e-01 -5.65622449e-01
-3.53661180e-01 3.84700865e-01 4.43479829e-02 3.08165997e-01
-1.35211384e+00 9.09868062e-01 3.71491343e-01 -3.99762988e-01
-7.67616153e-01 -2.91082591e-01 -9.88653302e-01 1.53214380e-01
5.85787416e-01 -3.97425652e-01 2.06735539e+00 -7.72045910e-01
-2.38436246e+00 4.44236696e-01 -1.63342044e-01 -5.56353748e-01
5.64537883e-01 -2.05313429e-01 -4.71649095e-02 -1.24961853e-01
-1.92261949e-01 7.61788011e-01 3.73618633e-01 -1.36438870e+00
-7.23534644e-01 -2.07145307e-02 6.22748852e-01 3.81905466e-01
3.29451188e-02 1.49652794e-01 -1.78640828e-01 1.07911848e-01
-4.86647010e-01 -9.44562852e-01 -5.02065361e-01 -6.33696437e-01
-4.31503475e-01 -5.22349715e-01 6.95101321e-01 -4.82487947e-01
1.11408854e+00 -1.64992237e+00 -9.83943045e-03 -1.71409413e-01
-8.34328532e-02 2.08865270e-01 6.62807301e-02 9.77460682e-01
5.01039863e-01 3.07447519e-02 3.85275781e-02 -5.40098131e-01
3.68931055e-01 1.64556384e-01 -3.92641127e-01 -4.24994558e-01
5.88226579e-02 7.95900881e-01 -9.68106687e-01 -5.58272302e-01
5.24521768e-01 -1.83981374e-01 -7.35335350e-01 1.05688894e+00
-1.09521329e+00 9.28848028e-01 -4.41468596e-01 2.37825841e-01
2.84659207e-01 -1.96900025e-01 6.79854333e-01 2.68522799e-01
-8.34784284e-02 7.63647497e-01 -9.81654108e-01 1.59509075e+00
-8.17235589e-01 1.48791358e-01 1.28680304e-01 -4.20083821e-01
1.18655312e+00 5.71718752e-01 1.37870356e-01 -2.48768985e-01
3.27128738e-01 -2.09249392e-01 1.54542178e-01 -6.20886147e-01
6.16767704e-01 -2.21031621e-01 -4.66883868e-01 7.53247619e-01
2.33799934e-01 -5.42657971e-01 2.42551953e-01 3.62390816e-01
9.63908374e-01 -1.18825147e-02 5.16226292e-01 -3.59292962e-02
8.63739967e-01 3.95614922e-01 3.31426412e-01 8.38204682e-01
-4.83915985e-01 9.74507779e-02 7.72745371e-01 -2.49131814e-01
-8.50873530e-01 -7.40518510e-01 3.86713177e-01 1.24234474e+00
1.71426460e-01 -3.92745852e-01 -9.87922370e-01 -9.48792398e-01
-3.23696196e-01 1.43865681e+00 -3.92635204e-02 8.26272890e-02
-6.97919726e-01 -2.10208133e-01 3.25359702e-01 1.68806449e-01
8.11107159e-01 -1.56408143e+00 -3.17693025e-01 3.60607326e-01
-3.58002692e-01 -1.33624732e+00 -5.51341414e-01 -1.89057380e-01
-7.04936385e-01 -9.07738090e-01 -6.56713024e-02 -7.43861616e-01
2.75367022e-01 -1.88238457e-01 1.30146754e+00 -1.08323544e-01
2.12972850e-01 4.09389347e-01 -4.23367858e-01 -2.81552166e-01
-1.28677464e+00 1.19612820e-01 -1.33288458e-01 -1.64472297e-01
6.92860007e-01 -2.69077331e-01 -3.28188330e-01 5.90514183e-01
-4.58440781e-01 8.89128000e-02 2.33950466e-01 9.88261402e-01
-3.03888500e-01 -2.70190179e-01 9.67150331e-01 -1.23798573e+00
1.45629227e+00 -3.34756523e-01 -6.91785991e-01 3.08333099e-01
-4.67547864e-01 2.56967187e-01 9.23933089e-01 -4.86593574e-01
-1.67582858e+00 3.90606016e-01 -1.81581214e-01 8.57860073e-02
-8.67078185e-01 -6.09356724e-02 -4.45262343e-01 5.11842251e-01
7.27665663e-01 1.36220157e-01 2.68115044e-01 -3.56341124e-01
6.07748210e-01 9.71357405e-01 1.66079044e-01 -8.06690216e-01
4.46749091e-01 -1.47407085e-01 -7.20863283e-01 -6.99138343e-01
-4.79442030e-01 -4.07842010e-01 -6.03323758e-01 -3.41359526e-01
8.16559732e-01 -4.47753400e-01 -1.30375445e+00 2.96604067e-01
-1.24746943e+00 -9.00903702e-01 -2.28597715e-01 2.50036180e-01
-7.27340162e-01 7.15088621e-02 -7.61898100e-01 -1.33709753e+00
-2.85712808e-01 -1.28368580e+00 9.62091029e-01 4.63581324e-01
-8.12707603e-01 -1.17364037e+00 5.08684590e-02 5.98517358e-01
4.12395775e-01 -5.63524030e-02 8.11343789e-01 -1.41384709e+00
-5.32047927e-01 -1.16856828e-01 1.84275612e-01 1.51956841e-01
4.48902279e-01 -2.24599838e-01 -8.79261076e-01 5.76455751e-03
1.17646858e-01 -7.97131419e-01 -1.95367247e-01 2.33489629e-02
5.18795431e-01 -7.58684814e-01 -1.46746308e-01 -4.72364038e-01
7.73925543e-01 8.11699629e-01 4.67986465e-01 -4.03729267e-02
1.43203333e-01 1.20288062e+00 9.54411328e-01 5.28471351e-01
5.88348508e-01 9.97895479e-01 -4.30416409e-03 4.24321920e-01
2.17331007e-01 -3.04177582e-01 4.54152048e-01 3.10919791e-01
2.39089578e-01 -3.69908661e-01 -7.54868448e-01 4.73467261e-01
-1.90645576e+00 -9.33178484e-01 1.07157946e-01 2.09826517e+00
1.24167490e+00 3.96530986e-01 6.89676762e-01 -3.11896294e-01
6.89436853e-01 -8.37151334e-02 -4.49951828e-01 -7.19022274e-01
5.47180176e-01 1.79973468e-01 2.75241081e-02 1.22138655e+00
-6.83743834e-01 1.40111971e+00 6.10313606e+00 2.00070873e-01
-6.59904599e-01 3.52798291e-02 5.81564486e-01 1.63016379e-01
-1.50278375e-01 2.28905827e-01 -1.17541862e+00 3.20657045e-01
1.30566144e+00 -3.17894042e-01 5.51304638e-01 9.78943408e-01
6.03773236e-01 -2.05428362e-01 -1.43657708e+00 4.67157483e-01
-2.88866609e-01 -1.25307631e+00 -1.15360685e-01 2.12351624e-02
2.64563024e-01 -6.76184714e-01 -5.93997240e-01 8.09271753e-01
1.25663507e+00 -8.35657179e-01 1.61249861e-01 3.40537906e-01
3.16372752e-01 -4.54044521e-01 6.01336002e-01 8.62294972e-01
-7.42141366e-01 -1.45782232e-01 5.74456900e-02 -2.70704508e-01
6.00235581e-01 -1.38865769e-01 -1.88615060e+00 2.00802431e-01
1.08097248e-01 -7.42321759e-02 -6.06939048e-02 3.15806776e-01
-3.35623026e-01 5.58166623e-01 -6.94212765e-02 -6.77498162e-01
1.94286600e-01 -3.90113324e-01 5.23128808e-01 1.07812786e+00
-2.36140296e-01 3.30754727e-01 6.19433820e-01 1.01606369e+00
5.15577458e-02 1.41041338e-01 -7.46976018e-01 1.72072351e-02
7.37333000e-01 1.32682920e+00 -4.52329665e-01 -5.82056582e-01
-2.78241187e-01 8.51123393e-01 1.17928073e-01 7.72726759e-02
-6.53086364e-01 -2.15508640e-01 4.38957334e-01 8.47375467e-02
-2.17029378e-01 -1.15003757e-01 -2.08795235e-01 -8.77899230e-01
-3.89266104e-01 -1.26971424e+00 2.58942157e-01 -5.47501385e-01
-1.27984691e+00 8.44630718e-01 2.62943029e-01 -9.12434161e-01
-1.16855240e+00 -3.67800921e-01 -8.79247963e-01 9.84229088e-01
-7.73684859e-01 -8.07672322e-01 -5.13565063e-01 4.51243997e-01
1.33808005e+00 -3.82173926e-01 1.02672017e+00 -1.55974522e-01
-3.42944592e-01 4.50630963e-01 -6.13772094e-01 3.97764109e-02
8.90470028e-01 -1.57443655e+00 4.52246189e-01 1.25852153e-01
-3.00049603e-01 7.18681812e-01 9.80210423e-01 -7.16144860e-01
-1.06932640e+00 -8.23980868e-01 9.57180440e-01 -6.83599591e-01
6.21173382e-01 -7.75999486e-01 -9.04348075e-01 7.84695625e-01
7.49629557e-01 -8.54110062e-01 6.73048913e-01 1.07907392e-01
2.69430578e-01 2.09259138e-01 -1.50414860e+00 8.54177356e-01
7.16122627e-01 -1.76812127e-01 -9.75372195e-01 5.21665752e-01
8.72463167e-01 -4.85519201e-01 -8.23878288e-01 1.51741300e-02
2.76080191e-01 -1.03593254e+00 5.90632617e-01 -9.14128900e-01
2.10566744e-01 2.16511264e-02 -3.64286229e-02 -1.27215719e+00
2.86035657e-01 -1.12067366e+00 3.81946228e-02 1.58913171e+00
5.60789406e-01 -5.29533029e-01 7.98909664e-01 1.51169181e+00
-9.15575586e-03 -4.10603046e-01 -2.85387009e-01 -4.97146398e-01
1.69755090e-02 -1.08022250e-01 6.47590995e-01 7.11710513e-01
6.92413807e-01 1.09405172e+00 -4.97864872e-01 -1.07171476e-01
1.19983628e-01 1.66735873e-01 1.35834110e+00 -1.09641075e+00
-4.83252704e-01 -1.82662338e-01 3.69923592e-01 -1.49919200e+00
3.39956284e-01 -4.25944954e-01 2.93757588e-01 -1.34772015e+00
-3.37269455e-01 -4.81777817e-01 5.76064050e-01 2.90021956e-01
-4.47692871e-02 -7.21239388e-01 1.29500657e-01 2.69518584e-01
-5.02419770e-01 5.23718059e-01 1.16368973e+00 9.82064158e-02
-1.04888713e+00 5.10777473e-01 -6.64062977e-01 6.93370461e-01
9.47955251e-01 -2.23358888e-02 -7.57456303e-01 3.90929848e-01
-4.35946077e-01 8.12759042e-01 1.52065686e-03 -6.34002686e-01
3.02492768e-01 -4.84272808e-01 -1.89267442e-01 -3.46527040e-01
5.49912512e-01 -6.16084099e-01 -1.43820584e-01 3.44151109e-01
-8.41709137e-01 -1.27298653e-01 1.91541463e-01 3.37768674e-01
-1.41245753e-01 -5.30487359e-01 7.24368334e-01 -6.00068867e-01
-6.20972872e-01 -2.88939685e-01 -6.56623244e-01 3.63515131e-02
1.14898372e+00 9.04750675e-02 -3.79710138e-01 -1.15412700e+00
-8.71921539e-01 5.42090118e-01 3.00184488e-01 4.87632364e-01
1.93988368e-01 -9.60108042e-01 -4.24904823e-01 -3.20490785e-02
-5.64528145e-02 -9.80066359e-02 -1.36304125e-01 2.50521541e-01
-4.03571576e-01 6.00995421e-01 -1.59764737e-01 -5.13848364e-01
-1.13113165e+00 4.29547042e-01 3.53222847e-01 -7.25582838e-01
-3.13137442e-01 4.73423690e-01 1.68011814e-01 -9.75942731e-01
4.15628016e-01 -1.69649497e-01 -5.41661501e-01 -3.12867463e-02
4.82280821e-01 9.21796784e-02 -2.68719941e-01 -2.42679626e-01
-7.31146634e-02 -3.56603295e-01 -2.31831059e-01 -6.63950086e-01
7.90505826e-01 -2.65672177e-01 2.72239119e-01 6.97371244e-01
5.08762836e-01 6.70248717e-02 -1.34521604e+00 -2.11819306e-01
3.62687796e-01 -3.11003029e-01 -7.77788281e-01 -1.03793347e+00
-1.99171603e-01 5.73004663e-01 2.63918608e-01 8.04253459e-01
7.29097188e-01 6.73271641e-02 7.67301381e-01 7.11169899e-01
6.05679631e-01 -1.07156944e+00 5.07135212e-01 7.42238760e-01
1.00949514e+00 -1.55546498e+00 -5.17317474e-01 -5.76715708e-01
-1.34196508e+00 9.55536544e-01 1.17622316e+00 7.95235783e-02
2.59800375e-01 2.21423313e-01 4.47337598e-01 -6.97035342e-02
-1.25435936e+00 -1.28744856e-01 -1.93671808e-01 5.53014755e-01
5.84160149e-01 -1.72528073e-01 -4.49547708e-01 6.28510654e-01
-4.54758078e-01 -1.11889271e-02 6.52294695e-01 9.38528359e-01
-5.44466555e-01 -1.70098627e+00 -2.89027601e-01 1.75770819e-01
-4.74200249e-02 7.98329990e-03 -1.07734585e+00 8.96991074e-01
-4.97566044e-01 1.50794077e+00 -1.83799535e-01 -3.98934692e-01
5.18154025e-01 8.25015903e-01 9.90718305e-02 -1.10207903e+00
-1.06151688e+00 1.24164131e-02 8.18188190e-01 -4.03043360e-01
-2.44635329e-01 -6.30977213e-01 -1.24144852e+00 -1.83178052e-01
-3.53662342e-01 5.80436468e-01 3.71271044e-01 1.00301790e+00
2.71485001e-01 2.57901222e-01 8.77049923e-01 -7.02728152e-01
-8.23097527e-01 -1.51835561e+00 -1.99275732e-01 6.45442784e-01
-8.00932273e-02 -5.35560548e-01 -2.71277457e-01 1.87030882e-01] | [12.922344207763672, 7.994905948638916] |
d9f8cb9a-5984-4a5e-b941-6fae4a3ea227 | comparing-feature-based-classifiers-and | null | null | https://doi.org/10.22489/CinC.2017.360-239 | http://prucka.com/2017CinC/pdf/360-239.pdf | Comparing feature-based classifiers and convolutional neural networks to detect arrhythmia from short segments of ECG | The diagnosis of cardiovascular diseases such as atrial fibrillation (AF) is a lengthy and expensive procedure that often requires visual inspection of ECG signals by experts. In order to improve patient management and reduce healthcare costs, automated detection of these pathologies is of utmost importance. In this study, we classify short segments of ECG into four classes (AF, normal, other rhythms or noise) as part of the Physionet/Computing in Cardiology Challenge 2017. We compare a state-of-the-art feature-based classifier with a convolutional neural network approach. Both methods were trained using the challenge data, supplemented with an additional database derived from Physionet. The feature-based classifier obtained an F1 score of 72.0% on the training set (5-fold cross-validation), and 79% on the hidden test set. Similarly, the convolutional neural network scored 72.1% on the augmented database and 83% on the test set. The latter method resulted on a final score of 79% at the competition. Developed routines and pre-trained models are freely available under a GNU GPLv3 license. | ['Fernando Andreotti', 'Marco A. F. Pimentel', 'Adam Mahdi', 'Oliver Carr', 'Maarten De Vos'] | 2017-09-24 | null | null | null | 2017-computing-in-cardiology-cinc-2017-9 | ['arrhythmia-detection', 'electrocardiography-ecg'] | ['medical', 'methodology'] | [ 2.42731586e-01 -2.47751586e-02 2.39398554e-01 -2.36282527e-01
-8.24783862e-01 -7.20449269e-01 -8.16123113e-02 4.36970264e-01
-4.41471756e-01 7.85329342e-01 -9.99814272e-02 -5.82719862e-01
-1.83781832e-01 -4.35953200e-01 -2.70069659e-01 -5.49938321e-01
-4.79340553e-01 5.01871586e-01 -2.09972724e-01 2.86501348e-01
-7.25487024e-02 6.39484525e-01 -1.17924833e+00 6.76594198e-01
7.63708532e-01 1.32289004e+00 -4.24625725e-01 1.13265717e+00
5.19337296e-01 4.06806231e-01 -8.21312308e-01 -1.57962248e-01
2.62484193e-01 -6.18837178e-01 -6.58976674e-01 -2.09564999e-01
4.15350437e-01 2.00713769e-01 1.16493762e-01 4.81856704e-01
1.08714914e+00 -5.88666260e-01 4.14593399e-01 -7.71476150e-01
-8.11937377e-02 3.66377413e-01 -1.38243452e-01 5.78055024e-01
2.32183129e-01 2.14579120e-01 6.30973995e-01 -8.69929135e-01
4.23345387e-01 4.28008497e-01 9.30948317e-01 1.32942751e-01
-1.36904287e+00 -4.77954030e-01 -5.23765266e-01 1.35306224e-01
-1.53364980e+00 -3.81297082e-01 5.24823725e-01 -6.61973715e-01
7.61618972e-01 4.22307163e-01 1.07855487e+00 8.35755289e-01
2.61733979e-01 2.58094579e-01 1.25666142e+00 -6.59101069e-01
1.40358120e-01 3.04986108e-02 2.10501328e-01 6.73812449e-01
4.81595337e-01 2.96079069e-01 -2.48002619e-01 -5.51149607e-01
7.79898465e-01 -6.03750870e-02 -7.28363037e-01 -2.21617877e-01
-1.51106191e+00 5.63626945e-01 2.76476443e-01 4.80802894e-01
-6.50816381e-01 -2.77997494e-01 4.57019031e-01 5.54799676e-01
1.97877124e-01 6.52010381e-01 -7.54186928e-01 -3.99852157e-01
-9.32892799e-01 1.54934734e-01 9.10731077e-01 1.59769103e-01
1.75336584e-01 5.19095510e-02 -3.23794276e-01 6.66264772e-01
7.63197392e-02 5.43054938e-01 5.10922968e-01 -6.40226781e-01
1.10672213e-01 7.76601613e-01 -9.59530398e-02 -6.60895824e-01
-7.79625654e-01 -1.12689900e+00 -1.21228290e+00 5.63479722e-01
5.64097345e-01 -5.03809571e-01 -1.06872392e+00 1.35804176e+00
-5.64684682e-02 -2.30050720e-02 4.81032990e-02 8.29881847e-01
1.04224622e+00 8.83631110e-02 -1.06294103e-01 -3.17845255e-01
1.40149164e+00 -3.68873835e-01 -4.78317529e-01 -2.14209501e-03
7.80562222e-01 -6.29940987e-01 6.57241702e-01 8.46851826e-01
-1.02290642e+00 -7.85090506e-01 -1.26250839e+00 5.67236543e-01
-2.56381799e-02 7.40641356e-01 1.19704425e-01 1.07866633e+00
-1.11295187e+00 8.44761252e-01 -8.68237197e-01 -2.65230119e-01
6.64720953e-01 4.81079698e-01 -5.54640055e-01 6.66993931e-02
-1.08923686e+00 8.66194308e-01 2.14288190e-01 2.01587945e-01
-7.65884817e-01 -7.28851318e-01 -7.43213773e-01 -1.71448793e-02
-1.81233361e-01 -6.09836698e-01 9.89474773e-01 -9.65630949e-01
-9.18181002e-01 1.19780040e+00 1.45825550e-01 -6.34912074e-01
8.19467545e-01 -3.92598808e-01 -5.31088889e-01 2.29673803e-01
5.87556884e-02 3.90246371e-03 7.24837720e-01 -7.67654300e-01
-6.45593703e-01 -4.28705841e-01 -2.70409137e-01 -1.10297114e-01
1.66555837e-01 -1.76743582e-01 3.59385624e-03 -6.46286964e-01
-7.77026638e-02 -9.86665428e-01 -2.18755245e-01 2.29115831e-03
-2.11247236e-01 1.28004864e-01 5.55936217e-01 -9.86989498e-01
1.28935993e+00 -2.31720614e+00 5.00442609e-02 5.00207782e-01
6.51882887e-01 5.98513484e-01 3.19099963e-01 1.48523301e-01
-5.99902391e-01 1.40294746e-01 -4.26942348e-01 1.42971307e-01
-4.43882078e-01 -6.58236220e-02 1.85175970e-01 6.09239817e-01
2.52045840e-01 8.24113607e-01 -5.39682508e-01 -2.99925089e-01
2.07518175e-01 6.57794237e-01 -1.30167216e-01 1.30095840e-01
5.76967418e-01 7.91002810e-01 -9.06918272e-02 5.94156265e-01
3.97730082e-01 -3.78171206e-01 4.41221803e-01 -3.18574727e-01
1.68882459e-01 -3.63288037e-02 -1.19799340e+00 1.58149707e+00
-1.21341459e-01 6.03687525e-01 -2.97629267e-01 -1.11096573e+00
9.76294160e-01 8.22527468e-01 3.95166159e-01 -4.65864003e-01
7.48061910e-02 4.06049281e-01 7.13341415e-01 -4.65324193e-01
-4.77688253e-01 -9.74172726e-02 2.05533728e-01 4.66722697e-01
6.83113560e-02 3.92472267e-01 2.33717352e-01 -8.82304236e-02
1.32024550e+00 6.22691847e-02 7.35293567e-01 -3.76879871e-01
4.50212270e-01 -1.95641875e-01 5.21944165e-01 8.66830230e-01
-4.05379027e-01 1.05283880e+00 6.13486767e-01 -1.22562540e+00
-5.77886760e-01 -9.92713809e-01 -4.22535032e-01 1.57783374e-01
-6.46778703e-01 -5.45016110e-01 -6.66639745e-01 -9.09433424e-01
-4.88948114e-02 1.56310961e-01 -8.52267921e-01 -2.61403501e-01
-5.36702335e-01 -7.73775578e-01 7.88884640e-01 6.77508593e-01
3.07297468e-01 -1.38258123e+00 -1.37192476e+00 2.57857919e-01
-2.16098562e-01 -6.50862813e-01 1.65760890e-01 3.76120418e-01
-8.27215612e-01 -1.41458964e+00 -1.02648115e+00 -6.58689499e-01
3.30150425e-01 -5.78205407e-01 1.51518333e+00 3.39692771e-01
-8.50185156e-01 1.20711632e-01 -3.81283641e-01 -7.54280567e-01
-3.23421776e-01 2.78082415e-02 -4.96794097e-02 2.06951991e-01
2.69563913e-01 -5.97505271e-01 -9.60364759e-01 1.47594154e-01
-3.30148786e-01 -2.33655274e-01 8.31563354e-01 7.50962913e-01
5.43990433e-01 -5.10236919e-01 7.63102293e-01 -8.70401442e-01
6.05246603e-01 -2.95898635e-02 -3.19558233e-01 5.54293990e-02
-8.92485142e-01 -4.12000805e-01 3.36408556e-01 -1.24796726e-01
-2.32162207e-01 4.09257591e-01 -1.63513914e-01 -3.69228452e-01
-4.66873020e-01 6.84188783e-01 -1.23612862e-03 3.75954877e-03
1.19010997e+00 -5.65815791e-02 -1.71308517e-02 -2.88570374e-01
-2.51490742e-01 7.76621699e-01 5.88152528e-01 -2.34246582e-01
3.53128880e-01 1.52098030e-01 1.82892382e-01 -6.18411720e-01
-7.25102425e-01 -3.83305162e-01 -8.63956928e-01 -1.82096303e-01
8.40675533e-01 -9.05085862e-01 -3.18654150e-01 3.32223535e-01
-8.95671129e-01 -2.57375181e-01 -4.83796448e-01 6.12324655e-01
-2.11180925e-01 3.70450169e-01 -8.77319276e-02 -6.86904609e-01
-9.27702248e-01 -9.06934261e-01 7.87966549e-01 -1.32465541e-01
-7.41227806e-01 -7.15743899e-01 3.60426366e-01 -1.38010073e-03
4.68678713e-01 1.01579881e+00 7.95536518e-01 -9.60533023e-01
2.01158538e-01 -6.58645928e-01 -7.91310705e-03 5.27009130e-01
2.77671784e-01 -9.78113189e-02 -1.24730957e+00 -3.85437816e-01
-8.32466558e-02 -2.42254371e-03 8.29037309e-01 6.02008522e-01
9.52503026e-01 3.64059180e-01 -2.69615263e-01 4.81580049e-01
1.05351257e+00 4.45086420e-01 7.45941699e-01 2.00102910e-01
3.66870850e-01 5.23688495e-02 2.71810859e-01 3.69618148e-01
-1.87086493e-01 4.32797939e-01 3.02040905e-01 -5.49947560e-01
6.23694900e-03 5.74502170e-01 -1.37898967e-01 1.08994365e-01
-6.14469707e-01 2.58753985e-01 -1.41773152e+00 5.45995772e-01
-1.76932955e+00 -6.57677829e-01 -5.26638627e-01 2.48605108e+00
4.92067248e-01 3.79265875e-01 3.09958488e-01 9.56492603e-01
5.84879220e-01 -4.31857020e-01 -2.12900624e-01 -3.83659780e-01
-2.23824039e-01 8.13909054e-01 -5.06127290e-02 4.14441004e-02
-1.42999101e+00 -2.23039109e-02 6.01731014e+00 1.28915742e-01
-1.39791274e+00 -2.87132449e-02 9.38138545e-01 4.43114340e-02
5.97797751e-01 -3.20582300e-01 -1.56178325e-01 3.38527381e-01
1.35442030e+00 1.15558710e-02 -6.15011118e-02 6.39808595e-01
1.59589753e-01 1.34940416e-01 -1.04471326e+00 1.30200124e+00
-2.79228408e-02 -1.38776064e+00 -4.43975747e-01 6.44139796e-02
2.41249964e-01 9.00445282e-02 -2.84986347e-01 2.04476029e-01
-5.69382846e-01 -1.38387072e+00 4.85471874e-01 5.91109872e-01
1.44593167e+00 -7.48487413e-01 1.31880724e+00 7.12200776e-02
-1.12470686e+00 -7.74864340e-03 2.02185512e-01 -1.81204677e-01
6.75850809e-02 9.14413333e-01 -7.18952954e-01 6.49593830e-01
1.10589421e+00 8.73028219e-01 -7.88708091e-01 1.45359766e+00
-1.60290375e-01 9.39552784e-01 -1.86981544e-01 5.63405156e-01
-3.81540149e-01 3.43659103e-01 4.67161655e-01 1.30714512e+00
2.21401259e-01 -5.56534454e-02 3.01950663e-01 3.70201796e-01
5.46221025e-02 1.13754325e-01 -4.31639612e-01 4.68867719e-02
-1.64674997e-01 1.56171679e+00 -8.10900390e-01 -6.10003531e-01
-2.46501952e-01 8.16374123e-01 3.01511586e-02 1.78352728e-01
-6.66278005e-01 -7.36423194e-01 -3.10277194e-02 3.19300801e-01
8.60743597e-02 4.24668223e-01 -5.61825395e-01 -9.37450826e-01
2.34667301e-01 -1.17707503e+00 5.75163960e-01 -4.34095979e-01
-1.02886450e+00 1.15702212e+00 -3.44037473e-01 -1.48231685e+00
-2.29817450e-01 -6.90650463e-01 -6.41515315e-01 1.26128173e+00
-1.13056183e+00 -7.93391705e-01 -6.88470244e-01 4.49381739e-01
6.82314113e-03 -3.62125337e-01 1.50951397e+00 5.15989363e-01
-1.78931981e-01 4.80102241e-01 -3.64565074e-01 3.94253701e-01
8.47042322e-01 -1.51871705e+00 3.39702696e-01 6.96518362e-01
2.22097158e-01 4.58159208e-01 3.12009633e-01 -4.11457241e-01
-5.28520703e-01 -1.18221509e+00 1.13458276e+00 -5.16319752e-01
-4.91667986e-02 1.06776223e-01 -9.11182284e-01 4.10565197e-01
8.46227482e-02 4.93987769e-01 1.07177234e+00 1.24100544e-01
-7.07471371e-02 3.38457078e-02 -1.07240224e+00 -2.22401391e-03
6.00668192e-01 -2.74329871e-01 -4.86569703e-01 1.68123230e-01
-2.95946270e-01 -5.91991484e-01 -1.17671645e+00 8.73005807e-01
1.09676957e+00 -1.13677239e+00 7.87991405e-01 -7.13012815e-01
2.89771706e-01 -3.43084633e-01 1.84348345e-01 -1.22091007e+00
-3.98673415e-01 -6.33012474e-01 -3.97922955e-02 5.17941654e-01
7.75934458e-01 -6.42049611e-01 7.14263082e-01 6.87251762e-02
-1.82874173e-01 -1.09066033e+00 -9.09891248e-01 -4.01383877e-01
-7.73397833e-02 -3.10743779e-01 -2.65291855e-02 7.64642715e-01
-1.25424236e-01 4.12400633e-01 -8.80276859e-02 4.05341983e-02
3.52070212e-01 1.58383489e-01 5.23135006e-01 -1.94881046e+00
-3.24435711e-01 -2.27931336e-01 -7.82429516e-01 2.33926013e-01
-5.11556685e-01 -1.02517807e+00 -3.67671609e-01 -1.61887681e+00
1.49994060e-01 -2.78894931e-01 -7.88055241e-01 7.73160517e-01
-1.66795701e-01 7.38489985e-01 4.32731919e-02 1.98375657e-01
-1.35976702e-01 -2.47671768e-01 7.00935543e-01 -3.53026278e-02
-3.20816785e-01 3.80152851e-01 -7.35245287e-01 7.10440218e-01
1.04421270e+00 -4.88920510e-01 -1.79421470e-01 6.83508292e-02
9.61690918e-02 1.57148968e-02 6.03053093e-01 -1.47666323e+00
-4.01496738e-01 6.07213795e-01 1.06532168e+00 -3.39663118e-01
8.21135715e-02 -6.98130310e-01 4.75171179e-01 1.07453334e+00
-2.02245459e-01 3.42717409e-01 4.40765917e-01 2.48102814e-01
-1.83702826e-01 1.82221338e-01 7.28596509e-01 -2.14935213e-01
3.33895721e-02 6.11181967e-02 -4.22870547e-01 1.69557199e-01
8.82176757e-01 -3.24346751e-01 -1.95088927e-02 -3.31448704e-01
-1.42737758e+00 -1.96161628e-01 -2.30070557e-02 1.74657643e-01
6.58675611e-01 -1.02394783e+00 -1.19115603e+00 4.55736756e-01
3.65606129e-01 -1.64037481e-01 2.29507655e-01 1.33524287e+00
-7.81010866e-01 2.83783734e-01 -4.25547510e-01 -9.53374803e-01
-1.62399912e+00 3.69594879e-02 8.03459823e-01 -4.58807349e-01
-9.56560075e-01 5.66203773e-01 -1.17490791e-01 -9.38013643e-02
2.32135102e-01 -3.48241210e-01 -4.79292214e-01 5.78467548e-02
7.59051979e-01 3.49697471e-01 7.59654224e-01 -3.89757246e-01
-6.72469437e-01 6.07940555e-01 1.18255243e-01 3.95051315e-02
1.09198189e+00 5.27291596e-01 2.39911377e-02 3.69531333e-01
8.43149662e-01 -8.60117674e-02 -4.32076305e-01 1.38393432e-01
-1.29241690e-01 -2.15087742e-01 3.95240262e-02 -1.44130361e+00
-1.18583453e+00 1.19920564e+00 1.55081689e+00 4.29214269e-01
1.14612496e+00 -3.70027333e-01 2.43388698e-01 2.18557000e-01
-6.23531938e-02 -4.83155459e-01 -4.42836314e-01 -1.20415539e-02
9.86950517e-01 -1.04841721e+00 -4.58840653e-02 -2.49117658e-01
-6.54384673e-01 1.40938509e+00 9.42073092e-02 -1.33275837e-01
9.34899271e-01 8.62273499e-02 6.83506131e-01 -4.46424335e-01
-4.50368077e-01 3.26328464e-02 5.24955750e-01 7.91425586e-01
8.31452787e-01 1.65777490e-01 -5.27197540e-01 8.66109371e-01
-5.74027747e-03 4.08306092e-01 3.24318737e-01 1.21649659e+00
-7.44986013e-02 -9.76226151e-01 -1.84700876e-01 8.76552105e-01
-1.00620759e+00 -7.27207363e-02 -4.77946460e-01 6.43160045e-01
2.70774186e-01 8.93543482e-01 -3.12865049e-01 -2.71057397e-01
5.34040034e-01 4.74730909e-01 3.32404077e-01 -5.73019564e-01
-1.15314972e+00 2.26803198e-01 2.95977563e-01 -4.17977840e-01
-3.27418983e-01 -6.01354301e-01 -1.01114142e+00 4.30071503e-01
-1.07218444e-01 3.40734154e-01 6.07707739e-01 7.30586112e-01
6.61177516e-01 9.27717209e-01 2.00012818e-01 -6.60409570e-01
-2.15116918e-01 -1.19025671e+00 -5.47961056e-01 3.71826172e-01
5.56663871e-01 -4.42733407e-01 -3.73670816e-01 3.84982377e-01] | [14.325573921203613, 3.298762083053589] |
e008e407-3a3a-46dd-b471-185aa0662ee1 | context-faithful-prompting-for-large-language | 2303.11315 | null | https://arxiv.org/abs/2303.11315v1 | https://arxiv.org/pdf/2303.11315v1.pdf | Context-faithful Prompting for Large Language Models | Large language models (LLMs) encode parametric knowledge about world facts and have shown remarkable performance in knowledge-driven NLP tasks. However, their reliance on parametric knowledge may cause them to overlook contextual cues, leading to incorrect predictions in context-sensitive NLP tasks (e.g., knowledge acquisition tasks). In this paper, we seek to assess and enhance LLMs' contextual faithfulness in two aspects: knowledge conflict and prediction with abstention. We demonstrate that LLMs' faithfulness can be significantly improved using carefully designed prompting strategies. In particular, we identify opinion-based prompts and counterfactual demonstrations as the most effective methods. Opinion-based prompts reframe the context as a narrator's statement and inquire about the narrator's opinions, while counterfactual demonstrations use instances containing false facts to improve faithfulness in knowledge conflict situations. Neither technique requires additional training. We conduct experiments on three datasets of two standard NLP tasks, machine reading comprehension and relation extraction, and the results demonstrate significant improvement in faithfulness to contexts. | ['Muhao Chen', 'Hoifung Poon', 'Sheng Zhang', 'Wenxuan Zhou'] | 2023-03-20 | null | null | null | null | ['reading-comprehension', 'machine-reading-comprehension'] | ['natural-language-processing', 'natural-language-processing'] | [ 4.39292759e-01 7.66784310e-01 -5.88704050e-01 -5.65518081e-01
-9.39149082e-01 -6.19188607e-01 8.52816761e-01 5.21385133e-01
-4.92154300e-01 1.24236262e+00 6.44179583e-01 -6.86814308e-01
-5.51936589e-03 -1.04604197e+00 -8.35293412e-01 -3.09773684e-01
2.52009988e-01 4.39893007e-01 2.36757129e-01 -3.62373143e-01
5.43819487e-01 1.65623993e-01 -1.07820165e+00 4.79567975e-01
1.04310155e+00 6.91636086e-01 -1.30643800e-01 3.23825985e-01
-4.17662054e-01 1.52255166e+00 -8.31277907e-01 -7.42065787e-01
-1.41856223e-01 -3.34286988e-01 -1.16689920e+00 -5.63262761e-01
3.72702301e-01 -4.54772800e-01 1.02560199e-03 1.02194786e+00
2.93626964e-01 2.58632183e-01 6.39059424e-01 -1.09106243e+00
-9.05193925e-01 1.16134799e+00 -2.45954126e-01 5.14043510e-01
8.45380425e-01 2.40086332e-01 1.02246714e+00 -7.12869704e-01
4.68761384e-01 1.51649368e+00 5.47957540e-01 6.99025333e-01
-1.40511739e+00 -6.00056410e-01 2.83335119e-01 6.91104412e-01
-9.72410560e-01 -7.24909127e-01 8.81353736e-01 -3.02852273e-01
1.25522888e+00 3.14502031e-01 3.91261488e-01 1.50885761e+00
6.66033477e-02 8.02182674e-01 1.55261755e+00 -7.23906934e-01
3.84870440e-01 3.63832712e-01 3.40609312e-01 6.31280243e-01
4.16145056e-01 4.75096405e-01 -1.06664455e+00 -2.89376646e-01
5.20703673e-01 -5.97593606e-01 -4.68684465e-01 1.39081478e-01
-1.30088973e+00 1.04901707e+00 3.34802598e-01 1.78118184e-01
-5.04129946e-01 -3.69731039e-02 5.69595024e-02 3.71008992e-01
2.31675550e-01 1.07222557e+00 -6.71661496e-01 -2.69006968e-01
-6.12385333e-01 4.41734284e-01 1.21719396e+00 7.69153535e-01
5.78540027e-01 2.33938713e-02 -3.77520204e-01 7.04501987e-01
7.58015960e-02 5.10661840e-01 5.20233572e-01 -1.21331346e+00
5.80061853e-01 4.09423143e-01 4.80796814e-01 -1.21105802e+00
-3.42845440e-01 -1.35586396e-01 -4.44810182e-01 -2.52064466e-01
4.38869536e-01 -3.49044323e-01 -5.87673783e-01 2.01439190e+00
3.71536672e-01 -1.90875623e-02 4.72410381e-01 7.59020567e-01
7.26448834e-01 7.18152881e-01 4.43141133e-01 -6.74503326e-01
1.42194700e+00 -6.12028778e-01 -9.41494286e-01 -6.19350135e-01
7.05416024e-01 -4.05497372e-01 1.37992716e+00 2.60447055e-01
-1.10780787e+00 -7.58874342e-02 -1.00514019e+00 -1.72877103e-01
-3.16188037e-01 -4.73036945e-01 7.45706737e-01 2.54422158e-01
-5.13390899e-01 4.89315182e-01 -5.96227348e-01 -6.92004710e-03
4.02436286e-01 -4.92729694e-02 -1.91708416e-01 6.49367422e-02
-1.91709566e+00 1.46752930e+00 7.32946098e-01 7.03163221e-02
-6.80011868e-01 -7.86546767e-01 -1.07154524e+00 1.77958190e-01
9.59996581e-01 -6.84126556e-01 1.65915108e+00 -8.56902897e-01
-1.65009737e+00 7.64117658e-01 -3.74357313e-01 -7.40679502e-01
2.47299477e-01 -5.15605688e-01 -4.61382151e-01 1.78875878e-01
2.90607840e-01 5.27076840e-01 6.70597434e-01 -1.22059357e+00
-3.55042547e-01 -8.73087421e-02 4.18034434e-01 2.38316104e-01
-1.25136767e-02 7.73154607e-04 5.49415946e-01 -6.02660239e-01
7.91797042e-02 -4.38581079e-01 1.56423524e-02 -3.63636285e-01
-6.62853241e-01 -3.94794106e-01 1.94574431e-01 -5.43421149e-01
1.30133343e+00 -1.70773709e+00 -2.44298339e-01 5.89758009e-02
6.17819838e-02 2.73635387e-01 3.60883921e-02 3.15492779e-01
9.21017975e-02 3.92692268e-01 -2.11292997e-01 4.05530751e-01
1.81001544e-01 4.55243438e-01 -7.85784543e-01 -2.10610718e-01
3.10779363e-01 1.24202323e+00 -1.06503403e+00 -7.97324955e-01
6.01332150e-02 -6.96084797e-02 -3.80083501e-01 3.23003858e-01
-6.66219234e-01 -6.28716080e-03 -4.74905491e-01 5.10730803e-01
2.16724887e-01 -4.48682934e-01 2.75219053e-01 -1.08949758e-01
1.30491227e-01 1.03401768e+00 -7.03222036e-01 1.19915032e+00
-5.27849138e-01 4.50797677e-01 -5.09761810e-01 -8.04336667e-01
5.74644148e-01 3.35717559e-01 -5.56205034e-01 -5.94278872e-01
-1.26955122e-01 4.70395386e-02 5.41633330e-02 -8.88878167e-01
3.11306357e-01 -8.49590421e-01 -1.32170022e-01 4.69256282e-01
-2.24319696e-01 -4.37249213e-01 -5.59224747e-02 4.34465379e-01
7.62094736e-01 1.49801120e-01 1.06682599e+00 -1.78410128e-01
3.01520616e-01 3.00799042e-01 7.61637807e-01 1.00150871e+00
-2.17357531e-01 -2.52187364e-02 8.53583455e-01 -3.89934957e-01
-5.06995440e-01 -1.03982520e+00 6.62361160e-02 1.19276357e+00
-1.64245591e-02 -2.43146345e-01 -4.60392416e-01 -8.55824232e-01
-1.55458465e-01 1.89957023e+00 -6.77890837e-01 -4.47162271e-01
-6.61792576e-01 -6.41463757e-01 5.50609112e-01 7.32752979e-01
6.55217350e-01 -1.32567286e+00 -7.93482900e-01 2.47707188e-01
-7.98253417e-01 -1.08136857e+00 -2.05630749e-01 2.52934188e-01
-7.58073866e-01 -1.04995024e+00 -3.54705518e-03 -4.31006491e-01
4.82542425e-01 -1.36072248e-01 1.30023110e+00 -1.70935586e-01
4.34987575e-01 1.94298342e-01 -2.79629052e-01 -6.69496298e-01
-5.63266337e-01 -1.76808342e-01 1.20579936e-02 -6.43718183e-01
7.85770476e-01 -6.86002254e-01 -3.32126766e-01 4.70107161e-02
-6.05781674e-01 1.86112911e-01 5.56105018e-01 9.42239404e-01
3.16294640e-01 -3.19057614e-01 8.75588775e-01 -1.35519660e+00
1.12460613e+00 -5.69462180e-01 -3.70871812e-01 6.01659596e-01
-7.32499063e-01 4.38423991e-01 5.63021064e-01 -6.14919782e-01
-1.48351121e+00 -5.16655743e-01 1.81909069e-01 2.12363362e-01
-2.23840058e-01 8.87304544e-01 -2.86725104e-01 3.22152793e-01
1.18206644e+00 3.00895214e-01 -3.17507148e-01 -6.98926151e-02
3.89737904e-01 3.65671247e-01 6.56283319e-01 -1.23097599e+00
5.86881876e-01 -4.98396456e-02 -2.48474613e-01 -4.49354619e-01
-1.76625454e+00 1.96958154e-01 -1.58434615e-01 2.27432489e-01
4.70989615e-01 -6.44016206e-01 -8.10896158e-01 -3.99440862e-02
-1.25612676e+00 -4.71194178e-01 -3.90498549e-01 3.44221771e-01
-7.13644564e-01 2.45644048e-01 -4.82260138e-01 -8.98459971e-01
-2.46990442e-01 -5.11119127e-01 5.65864980e-01 3.58165383e-01
-7.74235129e-01 -1.01330113e+00 3.31737772e-02 3.97592992e-01
3.81679773e-01 2.45178640e-01 1.59525621e+00 -1.16215992e+00
-3.21519345e-01 5.07353172e-02 8.70170444e-02 2.05221608e-01
1.28012836e-01 -2.94666380e-01 -1.05232680e+00 3.79308134e-01
2.36453786e-01 -8.43660653e-01 6.49163306e-01 7.39712268e-02
1.03259671e+00 -1.18802154e+00 -1.68732733e-01 1.43557400e-01
1.04361188e+00 1.57421350e-01 5.30240655e-01 2.64314294e-01
1.63712904e-01 8.50948811e-01 6.78587735e-01 7.02233464e-02
6.47394061e-01 3.62902910e-01 2.16709636e-02 4.97746557e-01
1.48561165e-01 -8.96491170e-01 2.66955048e-01 3.51089209e-01
2.43990999e-02 -7.03362599e-02 -8.95347238e-01 6.08344078e-01
-1.79991424e+00 -1.09628749e+00 1.73034757e-01 1.94914937e+00
1.77152765e+00 3.83849889e-01 -3.61939073e-01 7.27527291e-02
3.44584584e-01 4.81189013e-01 -5.31972766e-01 -4.40690607e-01
-2.57723838e-01 1.00953460e-01 -1.59432322e-01 1.05600250e+00
-8.05962980e-01 1.34081554e+00 6.29032183e+00 6.10764444e-01
-7.35769629e-01 -3.04915421e-02 4.69619721e-01 -8.92123133e-02
-7.25567043e-01 2.78361350e-01 -4.76874828e-01 2.85248697e-01
9.17347968e-01 -3.87484193e-01 2.07788959e-01 6.98794842e-01
-8.78823921e-03 -5.83190501e-01 -1.56203973e+00 5.75197279e-01
-4.80499528e-02 -1.36103380e+00 3.57065499e-01 -4.81762916e-01
5.69506645e-01 -7.30201304e-01 -1.95810974e-01 5.88200510e-01
4.40032780e-01 -1.19156229e+00 6.26691639e-01 6.12210095e-01
4.33089942e-01 -6.53418481e-01 8.46616983e-01 7.53367960e-01
-2.64539063e-01 7.61335790e-02 -3.07953387e-01 -4.84871566e-01
3.03094655e-01 6.17403090e-01 -1.00673258e+00 5.86061180e-02
1.94382235e-01 7.98545331e-02 -2.70101964e-01 2.77859896e-01
-1.17074335e+00 1.03086627e+00 -3.32442373e-01 -4.84241337e-01
-3.17959227e-02 2.86830038e-01 5.20637035e-01 1.08721328e+00
-2.68111795e-01 8.47262740e-01 -1.30469918e-01 1.04945254e+00
-2.26314589e-01 -9.05689299e-02 -6.58169687e-01 -8.83317068e-02
9.07110989e-01 6.61989927e-01 -1.16171725e-01 -6.86942935e-01
-1.55820116e-01 5.81599891e-01 6.66850984e-01 6.21346056e-01
-5.32316029e-01 -2.15171576e-01 3.03358167e-01 5.24724014e-02
-1.86670959e-01 1.04887635e-01 -5.11521101e-01 -1.24693692e+00
9.25167203e-02 -1.02357066e+00 4.07269090e-01 -9.74919736e-01
-1.33884943e+00 1.15579836e-01 5.03583729e-01 -3.36931944e-01
-6.94702268e-01 -4.36177909e-01 -6.16795063e-01 7.15085924e-01
-1.61956811e+00 -7.87980556e-01 1.86475962e-01 3.99966300e-01
4.68228996e-01 3.28072399e-01 1.01925814e+00 -4.68538105e-01
-2.62634575e-01 5.33157051e-01 -7.43212402e-01 1.35707930e-01
6.76171064e-01 -1.35471499e+00 1.41399458e-01 7.20004141e-01
1.86523512e-01 1.02808058e+00 1.08044088e+00 -7.35897660e-01
-1.06139195e+00 -6.52075350e-01 1.54489076e+00 -7.69439816e-01
7.27921069e-01 -4.47261669e-02 -1.39783871e+00 1.01875198e+00
1.83793202e-01 -3.13811541e-01 9.23063934e-01 4.88511205e-01
-7.91274726e-01 1.58706918e-01 -1.27830017e+00 8.26191485e-01
9.69674408e-01 -6.54021680e-01 -1.73899007e+00 4.92848068e-01
9.31927919e-01 -5.10779560e-01 -5.22620976e-01 3.97090614e-01
3.56070220e-01 -7.85613716e-01 7.64896333e-01 -1.20518303e+00
6.78558707e-01 -7.54939765e-02 -1.97590441e-01 -1.41224504e+00
-9.85731483e-02 -6.49636209e-01 -6.92942142e-01 1.04098499e+00
7.96709299e-01 -8.35763574e-01 3.25036913e-01 1.16398728e+00
2.64650613e-01 -7.01111436e-01 -9.61410940e-01 -4.83571410e-01
2.80797392e-01 -4.18536574e-01 5.24789870e-01 1.13239729e+00
7.24702299e-01 8.41832399e-01 -2.31498808e-01 1.97052717e-01
3.80335242e-01 4.24702764e-01 3.75674516e-01 -9.56304789e-01
-1.74304843e-01 -1.82649300e-01 2.52108723e-01 -1.00620294e+00
4.70409155e-01 -6.21340513e-01 2.32725486e-01 -1.38667274e+00
3.29889566e-01 -6.05315494e-04 3.44242081e-02 8.56562614e-01
-6.76463068e-01 -5.04305899e-01 4.75889891e-02 -1.73021331e-02
-4.26139474e-01 5.03785610e-01 1.15311921e+00 -1.66055784e-02
-1.83815449e-01 -9.06364098e-02 -1.20017004e+00 1.06143761e+00
9.03395057e-01 -3.52444381e-01 -6.46747530e-01 -3.15012425e-01
6.31882191e-01 2.51384228e-01 6.48787200e-01 -2.96287179e-01
3.00164878e-01 -8.34345341e-01 3.09915632e-01 -1.50279030e-01
2.45628878e-01 -4.67334151e-01 -3.71539295e-01 3.11144739e-01
-9.83345091e-01 -1.65442362e-01 3.59437466e-01 5.48976660e-01
-1.91399500e-01 -2.38777757e-01 5.72536588e-01 -2.87495971e-01
-6.45328462e-01 -4.42675769e-01 -2.95769215e-01 5.96882343e-01
6.42815173e-01 1.11488590e-03 -7.40206540e-01 -6.32397175e-01
-7.17203856e-01 3.15867573e-01 1.06794685e-01 2.33168900e-01
9.14583921e-01 -1.02957892e+00 -7.64394283e-01 -1.08640119e-01
1.05476066e-01 2.12389499e-01 1.35971710e-01 6.83038771e-01
-1.16497502e-01 5.58306038e-01 1.58936203e-01 -1.22656621e-01
-7.71563053e-01 7.16649115e-01 3.91373545e-01 -2.55724102e-01
-2.84862280e-01 1.01644063e+00 9.64722782e-02 -3.00146580e-01
1.94539592e-01 -5.72199583e-01 -1.82930663e-01 -3.33628021e-02
6.23503804e-01 2.34466016e-01 -2.14381933e-01 1.15842605e-03
-4.57217336e-01 9.75198895e-02 -3.53234380e-01 -2.79963583e-01
1.00651014e+00 -5.58324531e-03 -6.64229318e-02 6.68927372e-01
5.84501863e-01 1.74653381e-01 -1.04239941e+00 -5.81810057e-01
4.45129097e-01 -4.04296488e-01 -1.80070847e-01 -1.62904000e+00
-1.37690231e-01 8.07114184e-01 -2.58486032e-01 1.05021089e-01
8.53846908e-01 1.73965558e-01 6.18940294e-01 9.19035733e-01
5.54477394e-01 -1.05610609e+00 -4.54315804e-02 5.20029366e-01
1.14061534e+00 -1.42411005e+00 7.22734481e-02 -4.75161225e-01
-8.23380351e-01 8.75280201e-01 9.71222103e-01 3.63401592e-01
2.64900923e-01 1.90128848e-01 1.96348190e-01 -1.86553642e-01
-1.29345226e+00 1.75124317e-01 8.80381986e-02 6.40527070e-01
3.88550788e-01 3.27408195e-01 -3.69698822e-01 8.98712575e-01
-7.56824434e-01 -1.30580604e-01 5.32766879e-01 8.78002644e-01
-6.44333422e-01 -4.87174124e-01 -4.05218482e-01 5.28235972e-01
-3.52170646e-01 -6.31559908e-01 -7.79521585e-01 7.95612276e-01
-3.34746420e-01 1.19101918e+00 -3.92309427e-01 -2.94730021e-03
3.34721655e-01 4.86462295e-01 6.64189756e-01 -6.79444075e-01
-4.03383344e-01 -5.13042331e-01 6.39659107e-01 -5.29931545e-01
-4.88461167e-01 -4.37927872e-01 -1.43526316e+00 -3.97130460e-01
-2.66987741e-01 3.64846110e-01 1.10989936e-01 1.29101491e+00
1.81197181e-01 1.34445354e-01 1.72207266e-01 4.85012271e-02
-1.18100250e+00 -1.19945180e+00 -3.08649242e-02 1.96686178e-01
4.84919310e-01 -5.51578045e-01 -5.34184515e-01 -1.77240241e-02] | [10.176116943359375, 7.9428558349609375] |
84cbfc1e-22d0-4336-a2fc-a0b91a9f21e1 | deep-face-image-retrieval-a-comparative-study | 1812.05490 | null | http://arxiv.org/abs/1812.05490v1 | http://arxiv.org/pdf/1812.05490v1.pdf | Deep Face Image Retrieval: a Comparative Study with Dictionary Learning | Facial image retrieval is a challenging task since faces have many similar
features (areas), which makes it difficult for the retrieval systems to
distinguish faces of different people. With the advent of deep learning, deep
networks are often applied to extract powerful features that are used in many
areas of computer vision. This paper investigates the application of different
deep learning models for face image retrieval, namely, Alexlayer6, Alexlayer7,
VGG16layer6, VGG16layer7, VGG19layer6, and VGG19layer7, with two types of
dictionary learning techniques, namely $K$-means and $K$-SVD. We also
investigate some coefficient learning techniques such as the Homotopy, Lasso,
Elastic Net and SSF and their effect on the face retrieval system. The
comparative results of the experiments conducted on three standard face image
datasets show that the best performers for face image retrieval are Alexlayer7
with $K$-means and SSF, Alexlayer6 with $K$-SVD and SSF, and Alexlayer6 with
$K$-means and SSF. The APR and ARR of these methods were further compared to
some of the state of the art methods based on local descriptors. The
experimental results show that deep learning outperforms most of those methods
and therefore can be recommended for use in practice of face image retrieval | ['M. Sohel Rahman', 'Ahmad S. Tarawneh', 'Dmitry Chetverikov', 'Chaman Verma', 'Ahmad B. A. Hassanat', 'Ceyhun Celik'] | 2018-12-13 | null | null | null | null | ['face-image-retrieval'] | ['computer-vision'] | [-5.38699448e-01 -5.00771999e-01 -2.82148421e-01 -4.92189348e-01
-2.06439674e-01 1.96880624e-02 6.11364961e-01 -4.89699692e-02
-3.75114530e-01 4.16861564e-01 3.73668107e-03 1.93648234e-01
-7.93053865e-01 -7.96862304e-01 -3.42355847e-01 -8.94911885e-01
-5.16369581e-01 4.59968656e-01 -5.73146045e-01 -2.46625066e-01
2.00034603e-01 9.98654544e-01 -1.91054392e+00 2.51635253e-01
6.71970472e-02 1.58846200e+00 -3.54414433e-01 1.53897390e-01
-1.46306962e-01 4.19267654e-01 -3.82526189e-01 -6.73210979e-01
7.48215377e-01 2.48209368e-02 -4.49278116e-01 -3.74461651e-01
8.29173625e-01 -5.86637914e-01 -6.72596574e-01 1.01737070e+00
9.69645321e-01 2.56044000e-01 1.10568094e+00 -1.34663701e+00
-8.34825635e-01 -1.84947178e-02 -7.79153526e-01 3.35901856e-01
3.73319745e-01 -2.92962432e-01 7.82994330e-01 -1.45136857e+00
6.62946641e-01 1.82063222e+00 6.99221313e-01 5.87089658e-01
-6.43322945e-01 -1.20022047e+00 -3.05792004e-01 5.26963115e-01
-1.96842802e+00 -8.17085445e-01 6.07565939e-01 -2.37561628e-01
1.09352064e+00 8.77156556e-02 3.71858567e-01 7.15642154e-01
3.63909334e-01 6.88673019e-01 1.04876840e+00 -3.56496900e-01
-6.77128410e-05 -7.80950626e-03 1.56545922e-01 1.16631567e+00
5.07827401e-02 3.91657501e-01 -6.95476830e-01 -3.77694190e-01
7.39414454e-01 4.92996484e-01 1.95391066e-02 -5.21040820e-02
-6.02980494e-01 1.08066571e+00 5.59254169e-01 2.89016247e-01
-4.73846555e-01 9.40233842e-02 5.73200881e-01 8.38833630e-01
6.24480665e-01 -1.69628352e-01 -2.99076706e-01 5.36948383e-01
-1.10265446e+00 3.12433660e-01 7.01206088e-01 8.34341884e-01
9.05596137e-01 1.45582959e-01 -6.16060607e-02 1.28533304e+00
6.96245611e-01 8.26173127e-01 6.04993999e-01 -9.00428593e-01
-6.72844946e-02 6.38936937e-01 -3.71198148e-01 -1.65251207e+00
-1.33798897e-01 1.10600755e-01 -1.18463421e+00 2.03384489e-01
1.19683735e-01 2.51446337e-01 -1.06938922e+00 1.27430475e+00
2.41188616e-01 3.08167100e-01 -2.01691668e-02 7.78733373e-01
1.51914465e+00 6.53760552e-01 1.22022159e-01 -2.42452294e-01
1.09691358e+00 -7.31549025e-01 -7.37190127e-01 1.89190254e-01
5.17615020e-01 -1.05817282e+00 5.69784224e-01 2.21791983e-01
-8.90856624e-01 -7.21618235e-01 -6.76844358e-01 7.89163783e-02
-7.49038517e-01 2.06004843e-01 9.15540874e-01 4.87947136e-01
-1.52555835e+00 6.88640475e-01 -2.01013371e-01 -4.34126496e-01
7.50897884e-01 9.49301481e-01 -7.97530174e-01 -4.44112539e-01
-1.16335273e+00 7.93341696e-01 2.14730650e-01 3.40979308e-01
-1.02044249e+00 -5.50632119e-01 -7.69910336e-01 1.70243889e-01
-1.26911163e-01 -2.91731894e-01 5.37939250e-01 -8.84731889e-01
-1.22976172e+00 1.36392391e+00 -1.79027706e-01 -2.37317309e-02
7.44894370e-02 -9.45138559e-03 -5.67904472e-01 4.29985136e-01
-1.50265440e-01 8.99420977e-01 1.45194995e+00 -8.74749184e-01
-2.11140990e-01 -8.55465710e-01 -1.90448135e-01 -5.66309690e-02
-5.51240802e-01 2.77272969e-01 -2.44692400e-01 -4.78380203e-01
8.55051428e-02 -8.40292037e-01 3.48055363e-01 3.48635674e-01
1.41031235e-01 -8.15788090e-01 1.25778759e+00 -6.51725531e-01
1.04805589e+00 -2.30370021e+00 5.09531647e-02 7.04337001e-01
1.49946690e-01 8.95060301e-01 -5.52375793e-01 3.84793401e-01
-2.94302344e-01 5.58573194e-02 1.87249899e-01 -4.13633108e-01
-2.62961034e-02 3.10416430e-01 -4.56999950e-02 6.90854907e-01
-6.79685548e-02 8.12026262e-01 -4.95836705e-01 -7.06885576e-01
2.32848778e-01 9.36349630e-01 -5.30238926e-01 1.77414402e-01
2.24214777e-01 -1.92078367e-01 -3.86129498e-01 1.24624789e+00
8.60336721e-01 1.51596993e-01 -4.55482826e-02 -2.70103872e-01
2.92870373e-01 -4.57043529e-01 -1.17126429e+00 1.25620925e+00
-3.42954367e-01 6.36115909e-01 3.40724051e-01 -1.24388289e+00
1.16471016e+00 4.28946853e-01 4.87947226e-01 -8.57983828e-01
4.19080406e-01 3.51464182e-01 -2.03118190e-01 -6.56865299e-01
-3.98100130e-02 -1.51966736e-01 4.76175040e-01 3.73040348e-01
3.90299022e-01 2.23594382e-01 5.01184948e-02 4.02101614e-02
6.43890977e-01 -4.38006371e-01 2.22011637e-02 -5.79438269e-01
1.08242702e+00 -7.88610458e-01 2.96741039e-01 4.48181808e-01
-2.92236894e-01 4.17461008e-01 1.86414227e-01 -9.09608126e-01
-6.54877782e-01 -7.26397157e-01 -2.62946665e-01 1.19645023e+00
-8.68714601e-02 -4.40913737e-01 -6.42011821e-01 -5.25687754e-01
5.53583026e-01 -3.01339567e-01 -6.65907204e-01 -4.28357959e-01
-4.19966310e-01 -6.38356686e-01 6.61883414e-01 3.86371642e-01
9.43358481e-01 -1.52913320e+00 -7.55629363e-03 -3.06550980e-01
2.65306234e-01 -7.20458627e-01 -3.84153247e-01 -3.08586210e-01
-7.68012166e-01 -1.08614051e+00 -1.14307773e+00 -1.15937352e+00
8.19406509e-01 2.97626376e-01 1.05048037e+00 8.68346512e-01
-7.50225008e-01 6.98489964e-01 -2.84317523e-01 -3.49915326e-01
9.48568732e-02 -3.11898530e-01 4.82906312e-01 3.09490114e-01
8.36607337e-01 -3.17068666e-01 -7.96212435e-01 1.69134483e-01
-8.61797631e-01 -1.18118608e+00 7.50293851e-01 9.84689236e-01
5.22650838e-01 1.78059831e-01 4.25613850e-01 -6.58503592e-01
7.72733390e-01 -3.62378180e-01 -3.80570590e-01 2.45969325e-01
-7.25522518e-01 -2.10603043e-01 3.09972852e-01 -4.13587153e-01
-6.04220867e-01 -2.58738637e-01 -1.21385552e-01 -9.84925687e-01
7.89740756e-02 5.73804021e-01 -1.05323844e-01 -7.53928363e-01
4.35145676e-01 2.27050692e-01 3.14968318e-01 -3.87691200e-01
6.50351793e-02 8.09433818e-01 -1.62565544e-01 -3.50063324e-01
6.50856793e-01 4.94347215e-01 6.46666661e-02 -1.37807083e+00
-2.58061886e-01 -4.86131370e-01 -4.59145129e-01 -2.61633217e-01
6.39065802e-01 -8.74150515e-01 -1.03024125e+00 8.63774896e-01
-8.95154595e-01 4.29823250e-02 1.65503532e-01 2.54197925e-01
-9.04447138e-02 4.59776253e-01 -6.82443857e-01 -6.66821659e-01
-7.66063452e-01 -1.06315625e+00 1.24048483e+00 3.20185065e-01
1.50734484e-01 -1.05627155e+00 -3.14403385e-01 3.57893080e-01
6.16659224e-01 5.82557358e-03 1.05004430e+00 -5.76834917e-01
-2.39157096e-01 -3.69617492e-01 -4.34753984e-01 6.86170161e-01
1.70907341e-02 1.31263644e-01 -1.04784393e+00 -8.76847446e-01
-1.42874554e-01 -5.00237405e-01 9.55539346e-01 4.51374829e-01
1.30382884e+00 -5.53186238e-01 -3.76309335e-01 8.02271068e-01
1.53113413e+00 3.86044055e-01 8.74445319e-01 1.99656695e-01
4.67535108e-01 5.65575421e-01 2.70059258e-01 2.81819046e-01
-3.63271683e-02 3.95597070e-01 3.21430832e-01 -2.81402115e-02
-1.07090108e-01 2.15279549e-01 1.68067798e-01 7.99704611e-01
-3.94705802e-01 8.64205286e-02 -7.26552188e-01 3.44675630e-01
-1.40040636e+00 -1.02429378e+00 5.15219033e-01 2.07218194e+00
3.52358043e-01 -4.69045401e-01 -1.33107543e-01 3.72636467e-01
4.41354215e-01 4.28943127e-01 -2.49537066e-01 -5.97195506e-01
7.17399223e-03 8.24604034e-01 2.07607925e-01 2.98286945e-01
-1.12480056e+00 1.14134777e+00 6.32980394e+00 9.35947776e-01
-1.40999055e+00 -6.67433292e-02 6.89564049e-01 4.96965013e-02
2.61883795e-01 -5.24361551e-01 -9.82680798e-01 4.17264938e-01
7.42966533e-01 1.71060041e-01 6.71292722e-01 8.67801547e-01
-3.09535414e-01 1.00633174e-01 -1.17498803e+00 1.77441931e+00
4.91324514e-01 -1.02585220e+00 5.19955039e-01 9.56738740e-02
7.14379489e-01 -2.05590323e-01 5.20917594e-01 4.45618510e-01
-2.06951469e-01 -1.45799971e+00 -1.03317991e-01 5.70841253e-01
8.01256120e-01 -9.57391441e-01 8.97729993e-01 -1.04229741e-01
-1.10162425e+00 -2.80963451e-01 -9.47913408e-01 1.97682500e-01
-5.53975582e-01 2.97117203e-01 -5.01564741e-01 2.52279431e-01
1.18168819e+00 9.54847038e-01 -4.43576783e-01 6.86788738e-01
2.55016595e-01 8.97005275e-02 -2.63850152e-01 -4.19525318e-02
2.59648025e-01 -3.15939605e-01 1.98722497e-01 1.17244506e+00
3.04687709e-01 4.25218523e-01 -5.74884824e-02 3.04755032e-01
-4.17396009e-01 3.25154454e-01 -1.00148082e+00 -1.61802575e-01
3.30626041e-01 1.26518631e+00 -3.53293598e-01 -3.69254380e-01
-2.55048305e-01 6.44482017e-01 -2.53885053e-02 4.45930362e-01
-3.96292299e-01 -5.53220868e-01 1.00381839e+00 1.67952210e-01
3.04216206e-01 -9.00927931e-02 4.86105621e-01 -7.89930284e-01
-2.04158321e-01 -1.14750004e+00 5.87379992e-01 -6.60123408e-01
-1.54231393e+00 6.83198214e-01 9.71728712e-02 -8.24569345e-01
-1.15424499e-01 -9.69610512e-01 -3.76306206e-01 8.42003763e-01
-1.78008747e+00 -1.00713444e+00 -3.37680399e-01 1.33606267e+00
3.55891407e-01 -1.14244425e+00 1.01941144e+00 8.17494512e-01
-3.48885000e-01 9.20614660e-01 3.68269354e-01 3.93542826e-01
7.59555578e-01 -4.03679013e-01 -3.54016662e-01 9.98846162e-03
1.48653179e-01 1.05942571e+00 3.30499746e-02 -2.07958370e-01
-1.78232145e+00 -1.00624192e+00 9.74490225e-01 1.20548308e-01
6.00588024e-02 -6.29529953e-02 -8.42212617e-01 4.31311429e-01
1.39895156e-01 3.82889479e-01 7.23833025e-01 5.83357885e-02
-4.53391045e-01 -7.50338435e-01 -1.69399297e+00 1.61175072e-01
8.83373439e-01 -9.41676557e-01 -3.84317040e-01 5.65684140e-01
-1.21395878e-01 -4.64278683e-02 -9.75885093e-01 5.36258698e-01
9.27391589e-01 -9.65416074e-01 1.46860576e+00 -6.43947840e-01
1.92013189e-01 1.64514393e-01 -4.03009981e-01 -9.81819153e-01
-4.67789918e-01 -7.28700235e-02 -1.87116638e-01 1.04870868e+00
-3.55394073e-02 -8.75582874e-01 7.00594544e-01 4.65137035e-01
4.29024428e-01 -6.80869818e-01 -1.02625799e+00 -6.37603998e-01
4.54103611e-02 2.94742525e-01 6.14241481e-01 1.04278505e+00
-6.39676630e-01 -6.58695549e-02 -3.40974569e-01 -1.08974576e-01
7.05041230e-01 -3.79885696e-02 5.37765980e-01 -1.51068807e+00
3.87594312e-01 -4.73584622e-01 -8.96577716e-01 -5.04054189e-01
6.90469027e-01 -1.13285768e+00 -6.08390152e-01 -1.29991627e+00
1.17500611e-01 -4.67771590e-01 -5.80709577e-01 5.44763029e-01
3.45866621e-01 7.39284396e-01 2.32925206e-01 2.70911038e-01
-3.82427603e-01 5.04179955e-01 1.23428679e+00 -5.60599864e-01
2.96458090e-03 -3.61753911e-01 -4.62882698e-01 7.70016849e-01
4.76244420e-01 -3.16340804e-01 -2.83778012e-01 -6.17990375e-01
-8.55426788e-02 -1.70368463e-01 2.10620508e-01 -9.09005284e-01
2.43547380e-01 3.07190537e-01 9.92982507e-01 -4.39196795e-01
6.89343810e-01 -9.52030063e-01 -1.31319627e-01 4.23244715e-01
-7.38460422e-02 2.03138158e-01 2.47625381e-01 2.52343178e-01
-6.04484499e-01 -1.33975640e-01 8.95664692e-01 -3.59567046e-01
-1.03591728e+00 9.26474750e-01 -1.44112125e-01 -2.90382832e-01
9.23037469e-01 -2.89785445e-01 4.18954492e-02 -3.96712631e-01
-9.54493582e-01 1.26124367e-01 -2.26267055e-01 5.96149683e-01
1.19429624e+00 -1.61981416e+00 -6.08700454e-01 6.72313452e-01
-1.17479265e-01 -4.03885335e-01 1.02260187e-01 5.00366688e-01
-7.61770308e-01 5.69398344e-01 -5.46195686e-01 -5.23220956e-01
-1.74708211e+00 5.33253849e-01 2.61451483e-01 2.11580172e-01
-2.52886295e-01 1.18861508e+00 1.68153256e-01 -3.61296833e-01
6.09238505e-01 2.13054582e-01 -4.18877333e-01 4.09864873e-01
7.58232832e-01 5.38167059e-01 1.75408810e-01 -1.07348228e+00
-5.86153150e-01 1.13935184e+00 -3.20582002e-01 4.48105633e-01
1.48591626e+00 2.63646156e-01 -7.42455721e-01 -2.72076279e-01
1.84583461e+00 -6.23799860e-01 -4.21318531e-01 -2.03248248e-01
-1.85773328e-01 -7.40877271e-01 1.74915209e-01 -4.85418648e-01
-1.80250955e+00 1.13101101e+00 1.19508016e+00 -2.40368798e-01
1.24844813e+00 -2.00860530e-01 7.46958971e-01 6.90021634e-01
5.71603239e-01 -1.14422250e+00 2.17317045e-01 5.06489635e-01
1.12706351e+00 -1.42018056e+00 1.84154913e-01 -6.57580420e-02
-6.24607988e-02 1.18113554e+00 4.83666956e-01 -5.46960235e-01
1.42471218e+00 -1.34706885e-01 1.52438432e-01 -5.07592380e-01
-4.11598682e-01 2.94595230e-02 4.39746380e-01 4.12907928e-01
4.35443074e-01 -2.32253492e-01 -3.91795576e-01 1.33265452e-02
1.81498826e-01 4.63614054e-02 -4.02864784e-01 9.90306675e-01
-1.55949160e-01 -1.19809616e+00 -4.93517429e-01 7.31465518e-01
-6.91158891e-01 -9.92874578e-02 -5.71198225e-01 9.12643909e-01
2.49273896e-01 7.16495633e-01 -1.20826371e-01 -3.47854823e-01
2.24187702e-01 2.40259245e-01 6.63021743e-01 -3.25099796e-01
-5.42412519e-01 -6.50127009e-02 -5.48929155e-01 -5.90352833e-01
-7.27185011e-01 -3.32001179e-01 -9.31854486e-01 -8.12001884e-01
-3.13960999e-01 1.75295338e-01 6.66882992e-01 7.37594962e-01
4.15774882e-01 -1.06564127e-01 6.30276322e-01 -1.02284467e+00
-5.07822752e-01 -9.53706086e-01 -8.75222802e-01 5.97997665e-01
3.36102933e-01 -1.02157521e+00 -4.36919540e-01 -1.96581036e-02] | [12.940523147583008, 0.7503724098205566] |
0867497e-4603-4499-9782-e7a451235afa | down-the-rabbit-hole-detecting-online | 2301.11579 | null | https://arxiv.org/abs/2301.11579v1 | https://arxiv.org/pdf/2301.11579v1.pdf | Down the Rabbit Hole: Detecting Online Extremism, Radicalisation, and Politicised Hate Speech | Social media is a modern person's digital voice to project and engage with new ideas and mobilise communities $\unicode{x2013}$ a power shared with extremists. Given the societal risks of unvetted content-moderating algorithms for Extremism, Radicalisation, and Hate speech (ERH) detection, responsible software engineering must understand the who, what, when, where, and why such models are necessary to protect user safety and free expression. Hence, we propose and examine the unique research field of ERH context mining to unify disjoint studies. Specifically, we evaluate the start-to-finish design process from socio-technical definition-building and dataset collection strategies to technical algorithm design and performance. Our 2015-2021 51-study Systematic Literature Review (SLR) provides the first cross-examination of textual, network, and visual approaches to detecting extremist affiliation, hateful content, and radicalisation towards groups and movements. We identify consensus-driven ERH definitions and propose solutions to existing ideological and geographic biases, particularly due to the lack of research in Oceania/Australasia. Our hybridised investigation on Natural Language Processing, Community Detection, and visual-text models demonstrates the dominating performance of textual transformer-based algorithms. We conclude with vital recommendations for ERH context mining researchers and propose an uptake roadmap with guidelines for researchers, industries, and governments to enable a safer cyberspace. | ['Panos Patros', 'Aaron Dant', 'Philip Feldman', 'Jarod Govers'] | 2023-01-27 | null | null | null | null | ['community-detection'] | ['graphs'] | [ 3.17930788e-01 2.32690737e-01 -3.20237637e-01 3.64116907e-01
-1.18324891e-01 -8.60013366e-01 9.13448334e-01 5.18802106e-01
-2.94031829e-01 5.80340810e-02 1.05744636e+00 -8.94688725e-01
-3.22307050e-01 -4.68885511e-01 -1.21530909e-02 -1.63763747e-01
1.93426788e-01 -1.97408184e-01 -2.15618312e-01 -4.49482709e-01
1.11015022e+00 2.80450433e-01 -1.35240853e+00 4.40133750e-01
8.44985664e-01 4.43373144e-01 -2.96616495e-01 6.05291605e-01
-3.59108269e-01 1.28231847e+00 -9.31443512e-01 -5.13223231e-01
-4.40414362e-02 -5.82378447e-01 -8.23562145e-01 -1.77172929e-01
4.39336628e-01 1.74874794e-02 1.88972026e-01 1.08237529e+00
7.86143064e-01 -4.12456691e-01 3.52830350e-01 -1.23553181e+00
-9.88741219e-01 1.00966668e+00 -9.63046908e-01 5.01027703e-01
5.07627606e-01 3.88465434e-01 7.34556317e-01 -7.64680386e-01
1.38562620e+00 1.23740673e+00 7.88074255e-01 1.47341445e-01
-9.74688470e-01 -8.70543480e-01 -2.31419913e-02 -5.62984198e-02
-1.30356371e+00 -5.21633267e-01 8.49174798e-01 -1.24266744e+00
8.24212492e-01 5.26389599e-01 8.86898339e-01 1.37101889e+00
1.30581021e-01 3.42355594e-02 1.08262551e+00 -4.84816611e-01
-6.21524714e-02 2.94676244e-01 4.10556905e-02 6.91688895e-01
3.77857357e-01 -4.93608683e-01 -4.91392344e-01 -6.52316451e-01
1.35416836e-01 -1.92942634e-01 3.24583203e-02 3.29747647e-01
-1.26339328e+00 9.90544736e-01 6.25117421e-02 6.72059417e-01
-1.89499646e-01 -1.42468616e-01 8.00675392e-01 2.61392951e-01
9.31772590e-01 6.20253265e-01 3.69692922e-01 -3.72667640e-01
-1.06953847e+00 -1.78236485e-01 6.84579432e-01 3.84665877e-01
1.98204473e-01 -5.08798286e-02 1.21733874e-01 7.38654912e-01
4.74264741e-01 5.71322680e-01 -7.98058882e-02 -8.43883276e-01
3.80565971e-01 9.35157180e-01 -2.93642074e-01 -2.01047564e+00
-4.62003320e-01 -4.36502635e-01 -6.34155273e-01 1.23861253e-01
8.97335708e-02 -3.89500260e-01 -9.81852710e-02 1.05652642e+00
3.64444494e-01 -3.42621297e-01 -5.85626245e-01 5.44539392e-01
6.84894979e-01 5.69939613e-01 3.28673840e-01 -4.10122424e-01
1.33365238e+00 -1.26561254e-01 -7.45611131e-01 -1.98322237e-01
9.06826854e-01 -1.27447760e+00 9.80760932e-01 3.16661865e-01
-8.52389574e-01 1.09699450e-01 -9.28047180e-01 1.37417251e-02
-4.47423100e-01 -3.42433929e-01 2.99217738e-02 1.01500165e+00
-6.13393724e-01 3.04032922e-01 -3.92129682e-02 -8.02803934e-01
6.70985997e-01 -4.65369046e-01 -2.18727533e-02 3.33692342e-01
-8.10324430e-01 1.02184415e+00 -8.54132697e-02 -5.60292639e-02
-2.11938247e-01 -6.21056855e-01 -2.97991812e-01 -6.99311554e-01
4.89653796e-01 -2.66210586e-01 4.75646228e-01 -1.12021661e+00
-6.52015686e-01 1.29379487e+00 3.39825302e-01 -2.18743280e-01
4.03015524e-01 -6.16263561e-02 -6.28230870e-01 1.32249251e-01
2.72766531e-01 -1.01934925e-01 8.00533175e-01 -1.23770678e+00
-3.97776216e-01 -3.94078702e-01 -1.91449210e-01 -3.19266319e-01
-8.16756308e-01 1.13145161e+00 4.86193717e-01 -7.33035564e-01
-1.71671033e-01 -7.14704871e-01 3.48502807e-02 -2.08771944e-01
-6.06377363e-01 5.95702045e-02 1.13941491e+00 -1.10072124e+00
2.17371011e+00 -2.33671641e+00 -1.48076579e-01 5.67295551e-01
8.67220879e-01 2.19140694e-01 2.36201808e-01 1.06072330e+00
1.08023919e-01 9.88030076e-01 1.33314297e-01 8.14570114e-02
6.47573322e-02 -4.31030303e-01 -2.86268115e-01 1.03888178e+00
-2.65123457e-01 4.73127931e-01 -9.69636500e-01 -6.52871072e-01
2.15411652e-02 4.75015461e-01 -3.31778169e-01 -5.84548533e-01
2.13653922e-01 -9.08000916e-02 -4.12915275e-02 6.42455161e-01
2.59792805e-01 -3.02951217e-01 6.21789932e-01 1.23172216e-01
-1.11815965e+00 2.70518184e-01 -8.33391011e-01 8.54533195e-01
-1.54373825e-01 1.41474831e+00 6.08990312e-01 -3.25415671e-01
1.05798984e+00 -4.12393138e-02 3.88611645e-01 -7.62770176e-01
5.17602026e-01 9.25215408e-02 4.53335904e-02 -7.65314281e-01
4.44158196e-01 9.83936936e-02 -2.94576705e-01 9.36774194e-01
-6.41040206e-01 2.98363090e-01 -2.09776580e-01 5.80258131e-01
1.08669949e+00 -1.75390303e-01 4.89948660e-01 -4.47329164e-01
3.05777043e-01 1.28672689e-01 4.28508282e-01 4.45118368e-01
-4.53408211e-01 2.50567913e-01 9.86792803e-01 -4.87930626e-01
-1.09468472e+00 -4.40124452e-01 1.63360044e-01 1.40219772e+00
-1.59124866e-01 -8.57812941e-01 -6.80956721e-01 -5.72797239e-01
-8.60965475e-02 8.43459964e-01 -7.02218950e-01 1.93605646e-02
-4.81410921e-01 -5.65259635e-01 6.86485291e-01 -3.89921814e-01
1.52443707e-01 -9.77679849e-01 -1.07719088e+00 2.48798996e-01
-4.65312988e-01 -7.21604109e-01 -2.01560482e-01 -5.92931032e-01
-2.48804688e-01 -1.37475395e+00 -1.35754094e-01 -3.64812791e-01
3.79041731e-01 3.27636659e-01 7.58150280e-01 3.98841977e-01
-4.23418045e-01 2.75105745e-01 -4.22520429e-01 -6.32714570e-01
-8.84194016e-01 -9.57167894e-02 -1.44097924e-01 -1.15202583e-01
4.48598415e-01 -5.17441392e-01 -5.91975987e-01 2.45033264e-01
-8.04965496e-01 1.14899918e-01 6.52336702e-02 1.26973301e-01
-3.92300189e-01 -3.15699250e-01 5.25554061e-01 -8.29283059e-01
9.71365809e-01 -9.96849656e-01 6.13542497e-02 2.36617640e-01
-7.95270920e-01 -7.79636502e-01 1.11062266e-01 -3.52739096e-01
-7.58779466e-01 -6.90477967e-01 4.23918128e-01 8.73937830e-02
1.72051966e-01 6.12938344e-01 4.52916592e-01 3.78942192e-02
1.29446399e+00 -4.74222749e-01 3.76631796e-01 -2.96418190e-01
4.07145739e-01 1.06753719e+00 1.38332233e-01 -2.23713920e-01
7.80646265e-01 7.48650074e-01 -5.66829741e-01 -1.29316676e+00
-2.43333399e-01 -6.80605650e-01 -2.19675556e-01 -1.11206377e+00
7.26090729e-01 -6.16872251e-01 -6.53995931e-01 2.23374695e-01
-1.26150167e+00 -7.00522065e-02 -1.25636254e-02 -5.01293689e-02
2.05496967e-01 5.74143708e-01 -3.21184814e-01 -1.12007022e+00
-6.44316137e-01 -4.42726910e-01 2.47099176e-01 -2.13781431e-01
-9.44311261e-01 -9.34263945e-01 2.52651095e-01 6.86059177e-01
5.87834477e-01 1.20168924e+00 8.12623382e-01 -5.64409196e-01
1.18232086e-01 1.38265237e-01 -3.98130506e-01 -1.53383892e-03
4.08131070e-03 9.75026608e-01 -5.12058258e-01 -3.69039625e-02
-1.00262374e-01 3.50809209e-02 1.06785312e-01 -1.39800310e-01
2.32608840e-01 -1.10952318e+00 -2.16550261e-01 -1.89181745e-01
1.22666228e+00 4.42055836e-02 6.61595225e-01 8.22711587e-01
7.20360279e-01 1.14775658e+00 3.30489218e-01 1.01060402e+00
2.65091300e-01 6.45586103e-02 2.28139386e-01 4.14154753e-02
-2.11910799e-01 -1.16988845e-01 6.25637650e-01 1.01392817e+00
-1.92970797e-01 1.12789750e-01 -1.70368958e+00 7.15686858e-01
-1.64722323e+00 -1.39552307e+00 -6.33044600e-01 1.57994437e+00
6.13031387e-01 3.55061978e-01 6.19117975e-01 2.10267246e-01
9.34265792e-01 5.43085158e-01 2.89365023e-01 -7.40287304e-01
-5.77529073e-02 -2.94527113e-01 4.57801908e-01 2.28686184e-01
-5.56845963e-01 4.76440758e-01 5.71438122e+00 6.87056184e-01
-1.21698332e+00 5.14999986e-01 3.96091431e-01 -3.96987766e-01
-6.82011843e-01 1.87317103e-01 -1.77201852e-01 5.53647876e-01
9.84230518e-01 -4.00464684e-01 3.06396842e-01 7.44877279e-01
7.27421165e-01 3.85559015e-02 -2.61958003e-01 7.55457759e-01
4.87317979e-01 -1.78707874e+00 -5.65684617e-01 3.34577650e-01
7.20417738e-01 1.46814615e-01 1.96613014e-01 -2.26811215e-01
2.08904997e-01 -7.61358798e-01 1.22737217e+00 3.71151596e-01
7.04126239e-01 -5.17126977e-01 2.90003747e-01 4.27549221e-02
-6.86119497e-01 -5.38778722e-01 1.90434292e-01 -4.10309672e-01
2.08813041e-01 6.46002412e-01 -7.12622762e-01 1.34781659e-01
6.44003808e-01 6.93146169e-01 -7.50234902e-01 5.37220597e-01
-3.37782353e-02 7.89756119e-01 -5.18332385e-02 -2.81567574e-01
2.33958364e-01 -8.94934908e-02 8.94466996e-01 1.71658313e+00
1.37442118e-02 -4.46990989e-02 -2.74121016e-01 7.08742499e-01
1.59168422e-01 4.16404426e-01 -9.07090247e-01 -5.99042118e-01
8.83862257e-01 1.33660674e+00 -1.16033912e+00 8.60635117e-02
-3.64289790e-01 1.39628291e-01 -2.62274090e-02 1.70236140e-01
-6.03406191e-01 -3.93758833e-01 3.78543556e-01 9.85948324e-01
-2.70061493e-01 -3.80841196e-01 -7.46093869e-01 -5.20932674e-01
-4.53188807e-01 -1.38466418e+00 4.23320442e-01 -3.60989839e-01
-1.12351930e+00 4.16632071e-02 4.08409573e-02 -7.51835883e-01
2.43636742e-01 -1.62442043e-01 -6.78422689e-01 3.08081746e-01
-5.79737425e-01 -1.28829813e+00 -8.33172873e-02 1.13505676e-01
1.75065780e-03 -5.56446500e-02 1.64776146e-01 2.28580907e-01
-6.26793087e-01 1.49439588e-01 -2.72917859e-02 2.39318907e-01
6.73329651e-01 -6.18242085e-01 3.50391120e-01 7.85995245e-01
-3.84963632e-01 9.68383551e-01 8.38511765e-01 -1.11322343e+00
-1.19341719e+00 -5.07538259e-01 1.13052189e+00 -7.27115393e-01
1.43394375e+00 -5.01190424e-01 -4.42939639e-01 2.69466221e-01
5.91102779e-01 -7.81881511e-01 1.23746443e+00 3.24497432e-01
-5.99067092e-01 3.12645674e-01 -1.21620035e+00 8.55157614e-01
1.21504939e+00 -8.32705379e-01 -7.25692749e-01 2.72448272e-01
4.09471720e-01 4.46781754e-01 -8.54500532e-01 -2.73365051e-01
6.98469639e-01 -9.86990869e-01 5.93171775e-01 -2.32430965e-01
5.75709999e-01 -1.90111279e-01 1.34151787e-01 -4.53115761e-01
-3.44219089e-01 -1.27639866e+00 2.61539847e-01 1.47945797e+00
2.70092905e-01 -5.02483368e-01 2.30009124e-01 5.57361424e-01
5.93900196e-02 -7.02905729e-02 -7.64537454e-01 -2.01353177e-01
7.99958110e-02 -5.47650993e-01 1.29668728e-01 1.87977433e+00
5.82902670e-01 2.84844041e-01 -4.10488337e-01 -2.09410235e-01
6.27295196e-01 -3.65247726e-01 7.46758521e-01 -1.31967318e+00
3.63383740e-01 -8.38949680e-01 -2.68971562e-01 3.09343576e-01
-3.22508216e-01 -6.63883865e-01 -6.14728630e-01 -1.66048229e+00
3.49405468e-01 -8.14732835e-02 2.85745531e-01 3.43843549e-01
5.24245143e-01 2.95947671e-01 5.12875319e-01 3.76745433e-01
-7.08875895e-01 -2.33043119e-01 8.65505457e-01 -3.80811132e-02
-4.00477976e-01 -7.88035154e-01 -1.17298806e+00 6.61938488e-01
6.78710699e-01 -6.62074685e-01 -1.67390347e-01 5.60371578e-02
1.41423273e+00 -5.58730841e-01 6.37224019e-01 -5.90960324e-01
3.26763779e-01 -5.14061928e-01 -1.21650271e-01 -4.32255387e-01
-5.72306037e-01 -5.51683426e-01 4.45845455e-01 6.65375471e-01
-3.22698951e-01 2.84927189e-01 2.59347949e-02 1.75698131e-01
3.35762382e-01 -5.43334074e-02 5.20635009e-01 3.20484373e-03
-2.29622126e-01 -4.16988462e-01 -1.02727580e+00 3.87737900e-01
1.00729144e+00 -4.59726900e-01 -1.12020206e+00 -4.59549010e-01
-4.04677480e-01 3.02198119e-02 7.71804035e-01 5.51816642e-01
4.11550343e-01 -9.30821419e-01 -1.14698243e+00 -4.02008951e-01
7.95361921e-02 -8.15582693e-01 1.82671577e-01 1.01115859e+00
-6.53048515e-01 -2.97873408e-01 -1.90330803e-01 7.55158020e-03
-1.47568905e+00 5.81243694e-01 -1.90834597e-01 4.08475280e-01
-6.14888310e-01 3.04032564e-01 -3.86548519e-01 -3.58057991e-02
7.86165148e-02 2.34564573e-01 -3.52398902e-01 8.37308586e-01
7.31318057e-01 1.13239956e+00 -4.10424531e-01 -1.05470383e+00
-6.25542700e-01 1.95923895e-01 2.58160383e-01 -1.67321503e-01
1.29973733e+00 -4.47990060e-01 -5.86819410e-01 4.36114013e-01
1.13613117e+00 5.27394712e-01 -3.53746474e-01 1.51752293e-01
4.74600703e-01 -5.55944860e-01 -1.55840755e-01 -9.02712166e-01
-4.21717316e-01 4.80407506e-01 2.17670292e-01 9.04427052e-01
5.60469031e-01 6.61910623e-02 3.06792051e-01 -2.96658039e-01
-2.24066705e-01 -1.68872607e+00 2.87219048e-01 2.24639043e-01
1.28355014e+00 -6.77129626e-01 5.87105036e-01 -2.98384786e-01
-5.95783532e-01 1.11549973e+00 2.42339075e-01 2.61801481e-01
7.26338863e-01 2.37132356e-01 6.62914217e-02 -7.38127947e-01
-4.70301598e-01 2.02779099e-01 -6.22528512e-03 6.10418677e-01
4.45987225e-01 1.88348014e-02 -7.84319103e-01 3.48455697e-01
-2.15360090e-01 -2.81400979e-01 7.71830380e-01 1.03509295e+00
-5.31367719e-01 -5.38251877e-01 -7.78769791e-01 1.05024308e-01
-8.06772292e-01 -3.67198810e-02 -1.26067305e+00 1.03751159e+00
4.95385736e-01 1.27767313e+00 3.99276540e-02 -6.85451865e-01
1.23792402e-01 -7.57165328e-02 -1.89014792e-01 -3.42130125e-01
-1.34598851e+00 4.43945304e-02 8.00146520e-01 -3.02698970e-01
-5.69630921e-01 -9.26650941e-01 -8.18119824e-01 -1.10997617e+00
-1.87372327e-01 -1.58825785e-01 9.91729617e-01 6.11015499e-01
7.53876269e-01 1.48886725e-01 6.24490023e-01 -2.26182789e-01
7.35348165e-02 -7.31322885e-01 -5.95601387e-02 2.32148647e-01
1.69535592e-01 -1.68281406e-01 -6.40145659e-01 -1.34411663e-01] | [8.729939460754395, 10.371533393859863] |
18135194-f161-4d18-a16d-d938797fb679 | deep-learning-for-medical-imaging-from | 2209.02929 | null | https://arxiv.org/abs/2209.02929v1 | https://arxiv.org/pdf/2209.02929v1.pdf | Deep Learning for Medical Imaging From Diagnosis Prediction to its Counterfactual Explanation | Deep neural networks (DNN) have achieved unprecedented performance in computer-vision tasks almost ubiquitously in business, technology, and science. While substantial efforts are made to engineer highly accurate architectures and provide usable model explanations, most state-of-the-art approaches are first designed for natural vision and then translated to the medical domain. This dissertation seeks to address this gap by proposing novel architectures that integrate the domain-specific constraints of medical imaging into the DNN model and explanation design. | ['Sumedha Singla'] | 2022-09-07 | null | null | null | null | ['counterfactual-explanation'] | ['miscellaneous'] | [ 1.09870046e-01 6.36777937e-01 -4.45960402e-01 -7.36946583e-01
-6.74975738e-02 -7.03791156e-02 4.56878424e-01 -2.51254559e-01
-3.09235025e-02 4.57211435e-01 3.51545841e-01 -6.49303257e-01
-3.40070844e-01 -3.15527380e-01 -2.84409225e-01 -7.29943365e-02
3.48076165e-01 5.24487913e-01 -1.96772516e-01 -5.16738221e-02
2.06266209e-01 8.06893587e-01 -1.17689478e+00 3.80802929e-01
7.39163697e-01 8.74988317e-01 3.24449502e-02 6.36886537e-01
-2.80412763e-01 1.10183680e+00 -3.82907331e-01 -3.49147499e-01
3.76741849e-02 -6.09299839e-01 -7.92892635e-01 3.04824769e-01
3.31699580e-01 -6.83577240e-01 -8.46578002e-01 8.55300486e-01
2.46032029e-01 -1.29482895e-01 5.67407787e-01 -1.24896121e+00
-1.37042701e+00 3.95645440e-01 8.37200209e-02 1.59996212e-01
8.85719284e-02 1.26211360e-01 5.55790484e-01 -7.30767071e-01
5.28525889e-01 1.22191322e+00 6.74653530e-01 1.32211089e+00
-8.62191498e-01 -5.04161000e-01 1.34451419e-01 4.51890826e-01
-7.98833251e-01 -2.58401781e-01 8.40613842e-01 -3.48044425e-01
1.13560534e+00 5.47043271e-02 8.19952667e-01 1.30620933e+00
4.72484827e-01 8.09118152e-01 5.75439930e-01 -2.83325970e-01
1.98032185e-01 1.55987695e-01 3.74439090e-01 6.73865795e-01
4.82648820e-01 2.66595572e-01 -4.64033246e-01 1.95046723e-01
1.32341754e+00 3.95529002e-01 -1.46976084e-01 -1.21729501e-01
-9.44198191e-01 8.48564744e-01 6.72755241e-01 4.40961480e-01
-7.59323299e-01 3.86928260e-01 1.98279649e-01 1.06184162e-01
6.58454522e-02 8.17377687e-01 -5.56615353e-01 -1.05329789e-01
-9.72607970e-01 2.63172418e-01 7.50841856e-01 8.74060810e-01
3.08309138e-01 6.91422462e-01 6.43627420e-02 4.71803010e-01
4.58577305e-01 -3.14845219e-02 7.67994642e-01 -1.28927004e+00
-2.24806681e-01 6.95699632e-01 -1.62368491e-01 -1.01369727e+00
-5.81717253e-01 -7.13551164e-01 -8.69714439e-01 1.22397006e-01
4.54564253e-03 5.05959615e-02 -1.50385940e+00 1.53613722e+00
2.38159718e-03 -3.25762630e-02 2.87230700e-01 1.03351355e+00
1.26309896e+00 2.74123222e-01 3.33136767e-01 8.57324302e-02
9.36199248e-01 -1.20044410e+00 -8.46235633e-01 -8.34417403e-01
2.48075947e-01 -2.76049078e-01 8.47719312e-01 5.29972255e-01
-1.19243026e+00 -6.88041449e-01 -1.07922947e+00 -3.87635380e-01
-1.34831980e-01 -1.11615315e-01 8.92640769e-01 4.02182132e-01
-1.02007771e+00 5.62865734e-01 -1.01286459e+00 -4.63163644e-01
6.88109756e-01 5.57579756e-01 -2.55039960e-01 -3.18446636e-01
-8.99715126e-01 1.45034695e+00 2.96257526e-01 2.46878579e-01
-1.17540097e+00 -7.88121104e-01 -8.58790874e-01 2.31147647e-01
3.77245210e-02 -1.18960810e+00 1.58485413e+00 -1.37730491e+00
-1.12074399e+00 6.45703256e-01 -1.45156637e-01 -7.89498925e-01
2.18318790e-01 -2.67667830e-01 -5.43242991e-01 1.95090041e-01
-2.50901967e-01 1.08201385e+00 5.21262109e-01 -1.40960991e+00
-4.04147387e-01 -3.61562639e-01 1.17903166e-01 2.44085081e-02
-3.61595869e-01 -2.13399932e-01 -3.90690625e-01 -6.08373463e-01
4.80391860e-01 -7.96443641e-01 -6.95133686e-01 3.74745309e-01
-3.59606922e-01 -5.11745960e-02 9.21318352e-01 -7.60958433e-01
8.77073765e-01 -2.03313398e+00 8.33465308e-02 -4.57393788e-02
6.47422791e-01 4.46228445e-01 -1.10944502e-01 -9.91532858e-03
-1.74834892e-01 6.22377582e-02 -5.91556914e-02 -3.22167814e-01
-1.00516818e-01 7.67798126e-01 -1.83159113e-01 5.84444404e-02
7.07049295e-02 1.17082047e+00 -7.02487886e-01 -3.61106157e-01
5.34565747e-01 8.39775026e-01 -6.38261557e-01 1.34223893e-01
-3.77249032e-01 4.23655957e-01 -6.23066723e-01 7.53791988e-01
1.72178894e-01 -3.79879624e-01 2.16686100e-01 -1.50063634e-01
3.16028893e-01 1.22782297e-01 -5.20637333e-01 1.64333916e+00
-2.64916569e-01 1.04260492e+00 1.83608513e-02 -1.04305053e+00
8.36728632e-01 5.49051106e-01 4.69219863e-01 -5.67502558e-01
3.02880496e-01 2.38692284e-01 3.12640756e-01 -8.94876599e-01
4.70431656e-01 -5.20217657e-01 4.59991157e-01 4.91631776e-03
-5.18151447e-02 -2.94178128e-01 -2.85557002e-01 -9.63301659e-02
1.01852500e+00 2.18514111e-02 3.26410830e-01 -1.08528174e-01
1.27383754e-01 4.68241096e-01 7.35404134e-01 6.68761671e-01
-3.00251424e-01 8.18360627e-01 1.45907253e-01 -1.18045092e+00
-1.08601689e+00 -9.16374445e-01 1.44154266e-01 5.64031124e-01
-8.39838944e-03 2.14968789e-02 -9.36563432e-01 -6.34892225e-01
-7.27130547e-02 1.24649405e+00 -7.78773546e-01 -2.94676661e-01
-2.86044329e-01 -5.88323437e-02 5.10369837e-01 1.09235883e+00
5.21697283e-01 -1.31564093e+00 -9.24017012e-01 4.07651395e-01
1.13206916e-01 -1.14270103e+00 1.68119390e-02 1.37238234e-01
-1.30116975e+00 -8.62243354e-01 -5.02924323e-01 -7.00034797e-01
7.92017102e-01 1.46274209e-01 1.20331812e+00 4.28872287e-01
-5.16594708e-01 5.72901905e-01 -1.22436024e-01 -7.41807878e-01
-5.24370909e-01 -1.00590512e-01 1.11891599e-02 -6.19223356e-01
6.39956594e-01 -5.74809194e-01 -7.05471277e-01 1.88837364e-01
-1.10230613e+00 2.40656525e-01 7.68894672e-01 8.44406962e-01
4.80433494e-01 -1.61749884e-01 5.51679432e-01 -8.58585298e-01
7.01966822e-01 -2.90802240e-01 -1.03885651e-01 2.25951716e-01
-8.46379578e-01 -1.11225106e-01 5.61771274e-01 -5.99518597e-01
-1.00847971e+00 1.07719496e-01 -3.04402173e-01 -7.08299577e-01
-5.18502355e-01 8.33712518e-01 -4.50722352e-02 -1.29307411e-03
8.58441412e-01 5.38874343e-02 1.14314303e-01 -2.83388287e-01
1.71245441e-01 4.19763684e-01 8.91621411e-01 -3.40931833e-01
2.97111213e-01 3.50962669e-01 -2.08798256e-02 -8.48813295e-01
-9.80246007e-01 2.15963677e-01 -4.00810629e-01 -7.91769400e-02
9.80598211e-01 -4.89591569e-01 -1.94389343e-01 1.11662008e-01
-1.45047438e+00 -1.33353636e-01 -4.93228614e-01 6.17369473e-01
-4.52265501e-01 -6.13513775e-02 -5.94000936e-01 -6.29398525e-01
-1.41736016e-01 -1.25709426e+00 5.28126478e-01 5.96606135e-01
-7.65823901e-01 -1.23966634e+00 -2.65127093e-01 3.88583034e-01
7.42310822e-01 1.76359773e-01 1.18439925e+00 -8.56421649e-01
-6.15173578e-01 -4.03902709e-01 -3.73279661e-01 4.50999737e-01
1.43776372e-01 -1.72101960e-01 -1.01032865e+00 2.76927680e-01
3.42929065e-01 -1.16198443e-01 5.92121065e-01 9.89770234e-01
1.52493370e+00 -2.42585689e-01 -4.60473210e-01 5.91649711e-01
1.32480943e+00 6.60475671e-01 7.53900588e-01 3.61149460e-01
6.60705745e-01 7.05418110e-01 7.55960271e-02 1.08550951e-01
4.94130552e-01 1.75877675e-01 8.23437929e-01 -3.50399017e-01
-2.09469795e-01 -2.39900261e-01 -2.78663874e-01 4.15158629e-01
1.09630674e-02 -2.23238438e-01 -1.18986809e+00 8.22002351e-01
-2.05492568e+00 -7.26305306e-01 -5.59243709e-02 1.30729580e+00
3.86352301e-01 3.26285779e-01 -6.20133758e-01 6.01275787e-02
2.12762475e-01 -3.13478261e-01 -1.03031909e+00 -6.25068843e-01
3.65407169e-02 5.15839271e-03 9.04277787e-02 3.81563544e-01
-5.83212852e-01 8.02130222e-01 8.26535988e+00 7.49726295e-02
-1.13882971e+00 -5.90408109e-02 6.56197190e-01 2.70489436e-02
-5.94082654e-01 -1.85876295e-01 -4.14192826e-01 -1.11518100e-01
9.91914630e-01 1.41635939e-01 2.51974076e-01 1.19924963e+00
3.85235190e-01 1.93368658e-01 -1.50319111e+00 1.02057588e+00
1.11059040e-01 -1.75074267e+00 3.07332784e-01 9.93657634e-02
6.32410705e-01 -1.16620325e-01 4.01358902e-01 -3.02735884e-02
2.79112607e-01 -1.64828134e+00 5.50241411e-01 6.13936901e-01
3.85463864e-01 -4.35712218e-01 7.71896660e-01 2.96013892e-01
-4.36506152e-01 -3.81038964e-01 -3.87422770e-01 -3.04721713e-01
9.52891335e-02 4.79097098e-01 -7.75626183e-01 2.19298184e-01
7.14031398e-01 6.35855019e-01 -1.53171003e-01 8.87603760e-01
-3.50543261e-01 6.28342509e-01 1.35880904e-02 1.17186159e-01
4.99976814e-01 3.24468762e-01 3.19149733e-01 1.01530004e+00
1.60763219e-01 3.25848341e-01 -1.06057823e-01 1.21111774e+00
7.28529841e-02 -6.06309831e-01 -7.22128034e-01 -1.86845392e-01
3.27025622e-01 9.44628179e-01 -5.65888643e-01 -4.55096811e-01
-4.20358002e-01 8.81174684e-01 -1.14388846e-01 4.53650326e-01
-8.41336727e-01 3.09624299e-02 9.56338584e-01 2.09212705e-01
-6.56539649e-02 -4.11330044e-01 -1.08651042e+00 -8.73975635e-01
-1.83163077e-01 -1.08324480e+00 9.34062228e-02 -1.18168497e+00
-1.42868912e+00 1.02412081e+00 3.70808691e-02 -7.40796387e-01
-5.64039528e-01 -1.02071774e+00 -5.12865245e-01 7.31487811e-01
-1.36222756e+00 -1.46370375e+00 -4.75642055e-01 4.54250425e-01
9.10364866e-01 -3.34990174e-01 1.13659918e+00 1.32148936e-01
-3.68408293e-01 1.95644692e-01 -2.57583648e-01 5.10444380e-02
8.38322863e-02 -9.52764690e-01 4.59706903e-01 6.62683606e-01
5.78680523e-02 1.04510379e+00 1.02436876e+00 -5.49271047e-01
-1.20888555e+00 -7.52292931e-01 9.57630217e-01 -3.39218020e-01
1.21142104e-01 1.85729876e-01 -8.60537171e-01 9.21343505e-01
3.93590331e-01 2.74734781e-03 6.99765801e-01 -1.03504173e-01
-1.98436037e-01 4.01693955e-02 -1.51990294e+00 7.38203883e-01
8.24063659e-01 -4.72010374e-01 -8.60606432e-01 9.74061713e-02
6.99533761e-01 -5.77740133e-01 -5.88666081e-01 4.12997603e-01
8.12776685e-01 -9.14813995e-01 8.68352711e-01 -1.05673790e+00
8.18074763e-01 -4.23681475e-02 -1.91839531e-01 -1.18961060e+00
-4.19535786e-01 -3.66342336e-01 -2.03915402e-01 6.08604491e-01
5.37406564e-01 -4.11091328e-01 1.27683866e+00 1.56855738e+00
-6.30584717e-01 -1.03516722e+00 -7.70035088e-01 -4.30064857e-01
3.82160619e-02 -6.35975718e-01 6.91875637e-01 9.62829649e-01
-2.08457887e-01 1.66179955e-01 -2.34119296e-01 7.88526163e-02
6.01903856e-01 -3.18554282e-01 2.63643622e-01 -1.43810749e+00
-1.14546075e-01 -5.63630164e-01 -5.05795717e-01 -1.11824226e+00
1.12287551e-01 -2.61069775e-01 1.26684308e-01 -2.37431669e+00
1.80141523e-01 4.18154560e-02 -2.20289171e-01 8.36330354e-01
2.77026385e-01 1.03318721e-01 1.20858543e-01 1.38004705e-01
-1.11631200e-01 2.81172961e-01 1.20377254e+00 -1.97556227e-01
-3.53255011e-02 -3.76339704e-01 -1.23100483e+00 1.16260779e+00
8.06923509e-01 -3.42951238e-01 -7.34507799e-01 -1.15763915e+00
2.76808441e-02 1.45710006e-01 5.51830828e-01 -1.01498151e+00
3.48916531e-01 -4.54548657e-01 7.59627581e-01 -4.44142401e-01
3.32182169e-01 -1.27139544e+00 3.40954840e-01 6.52072966e-01
-5.18215477e-01 1.35053709e-01 4.64387298e-01 2.40735382e-01
-3.20351243e-01 -3.15697432e-01 9.53006029e-01 -4.18997973e-01
-8.22884560e-01 2.80360907e-01 -5.73854923e-01 -3.57726812e-01
9.25360024e-01 -6.61173224e-01 -3.28395814e-01 -6.92480147e-01
-8.61806750e-01 1.13180563e-01 3.51558924e-01 4.91447300e-01
1.15527487e+00 -1.10357189e+00 -3.96617860e-01 1.93873346e-01
-2.67824233e-01 2.39485919e-01 3.34783524e-01 2.96597153e-01
-7.21547782e-01 8.40291262e-01 -6.10584676e-01 -2.99883187e-01
-8.69894981e-01 3.54701072e-01 7.62226641e-01 -1.77924916e-01
-6.75752163e-01 8.46592963e-01 2.45471299e-01 -2.61564851e-01
4.90928352e-01 -3.78064841e-01 4.29090559e-02 -6.51719928e-01
4.89223361e-01 3.03803682e-01 -2.38245323e-01 -1.83592811e-01
-4.07680422e-01 2.52327859e-01 -1.36894763e-01 -1.44592270e-01
1.55304730e+00 1.65359259e-01 1.97698146e-01 9.26541537e-02
6.21750355e-01 -9.81280267e-01 -1.18453658e+00 1.68194905e-01
1.52756097e-02 -1.06555618e-01 2.93656200e-01 -1.16130173e+00
-1.20925915e+00 1.15177727e+00 5.42445064e-01 7.61254355e-02
1.16105103e+00 -1.11287691e-01 8.23053956e-01 5.15523732e-01
1.17275171e-01 -8.15565467e-01 2.11479545e-01 4.42429364e-01
9.11400437e-01 -1.12978041e+00 1.09682031e-01 -2.41424460e-02
-7.93688536e-01 1.26393163e+00 8.80182087e-01 -9.40434784e-02
7.16658235e-01 3.36402893e-01 3.14037234e-01 -5.97098470e-01
-5.69266438e-01 3.55434895e-01 4.32590097e-01 9.38705444e-01
6.08062923e-01 -1.62790239e-01 -6.60276115e-02 8.44057500e-01
-6.21837974e-02 5.75374424e-01 4.04484898e-01 8.64585102e-01
-3.82123351e-01 -9.57407117e-01 -1.04764044e-01 5.54922283e-01
-5.08368373e-01 -9.25302505e-02 -5.81027448e-01 8.62822235e-01
1.86383501e-01 9.16692615e-01 -2.53723077e-02 -3.19916934e-01
3.00191194e-01 2.84902416e-02 4.53019440e-01 -5.35255134e-01
-7.41753697e-01 -2.55836040e-01 1.00697890e-01 -5.25354922e-01
-3.98997635e-01 -2.65109301e-01 -1.54221737e+00 -2.74937391e-01
4.04496528e-02 -1.18394196e-01 9.13725495e-01 1.10654259e+00
4.29719120e-01 9.14142132e-01 -1.95991889e-01 -7.97784626e-01
-4.96031314e-01 -9.39938128e-01 -4.15169477e-01 1.17528625e-01
4.97704953e-01 -4.85471964e-01 -2.10430175e-02 2.36626491e-01] | [8.974401473999023, 5.571403980255127] |
2d33a40c-f94b-4c0d-8f86-a030051b40cc | chatgpt-beyond-english-towards-a | 2304.05613 | null | https://arxiv.org/abs/2304.05613v1 | https://arxiv.org/pdf/2304.05613v1.pdf | ChatGPT Beyond English: Towards a Comprehensive Evaluation of Large Language Models in Multilingual Learning | Over the last few years, large language models (LLMs) have emerged as the most important breakthroughs in natural language processing (NLP) that fundamentally transform research and developments in the field. ChatGPT represents one of the most exciting LLM systems developed recently to showcase impressive skills for language generation and highly attract public attention. Among various exciting applications discovered for ChatGPT in English, the model can process and generate texts for multiple languages due to its multilingual training data. Given the broad adoption of ChatGPT for English in different problems and areas, a natural question is whether ChatGPT can also be applied effectively for other languages or it is necessary to develop more language-specific technologies. The answer to this question requires a thorough evaluation of ChatGPT over multiple tasks with diverse languages and large datasets (i.e., beyond reported anecdotes), which is still missing or limited in current research. Our work aims to fill this gap for the evaluation of ChatGPT and similar LLMs to provide more comprehensive information for multilingual NLP applications. While this work will be an ongoing effort to include additional experiments in the future, our current paper evaluates ChatGPT on 7 different tasks, covering 37 diverse languages with high, medium, low, and extremely low resources. We also focus on the zero-shot learning setting for ChatGPT to improve reproducibility and better simulate the interactions of general users. Compared to the performance of previous models, our extensive experimental results demonstrate a worse performance of ChatGPT for different NLP tasks and languages, calling for further research to develop better models and understanding for multilingual learning. | ['Thien Huu Nguyen', 'Trung Bui', 'Franck Dernoncourt', 'Hieu Man', 'Amir Pouran Ben Veyseh', 'Nghia Trung Ngo', 'Viet Dac Lai'] | 2023-04-12 | null | null | null | null | ['multilingual-nlp'] | ['natural-language-processing'] | [-9.41924974e-02 5.49187921e-02 -2.47042656e-01 -7.10568130e-02
-1.09066451e+00 -6.13494217e-01 9.46959794e-01 -1.60535853e-02
-3.69047344e-01 1.05628157e+00 4.46030736e-01 -6.35597050e-01
3.28467309e-01 -7.50720799e-01 -4.61017072e-01 -3.24079692e-01
6.67004213e-02 8.21423769e-01 2.24580050e-01 -5.69890380e-01
2.11783163e-02 1.15567878e-01 -1.20044661e+00 6.06308103e-01
1.06442153e+00 2.25035354e-01 5.55408359e-01 7.27205575e-01
-5.47380149e-01 8.74118507e-01 -7.13583052e-01 -6.13675416e-01
7.40021765e-02 -5.67918956e-01 -9.84750450e-01 -2.76929975e-01
1.05395615e-01 1.02568068e-01 9.27922577e-02 6.29885018e-01
8.62691104e-01 1.72682524e-01 3.95477861e-01 -1.14395857e+00
-9.30447221e-01 1.09964657e+00 -3.58887821e-01 8.50029662e-02
5.72292864e-01 3.31393450e-01 1.18789017e+00 -9.22808230e-01
7.83764780e-01 1.51798642e+00 5.11003494e-01 7.09258497e-01
-1.06136906e+00 -7.11531162e-01 4.49120030e-02 8.47049356e-02
-1.26966381e+00 -4.03317183e-01 3.48691672e-01 -1.27384260e-01
1.29253531e+00 1.68456063e-01 4.25665528e-01 1.67407227e+00
1.46630883e-01 1.21553731e+00 1.15566814e+00 -7.67491043e-01
2.75685526e-02 5.20309865e-01 -1.64375350e-01 5.15377998e-01
-4.40043435e-02 -8.74397084e-02 -5.61934888e-01 -9.06720012e-02
6.33751333e-01 -4.99030709e-01 -1.90172493e-01 3.59189361e-02
-1.48726225e+00 1.05923808e+00 4.94431937e-03 7.11945117e-01
-5.75821064e-02 -1.34876505e-01 3.79223257e-01 4.36853558e-01
7.53753543e-01 8.30690384e-01 -5.45855224e-01 -6.41437054e-01
-8.60811234e-01 4.78041261e-01 1.32409203e+00 1.14936185e+00
4.73333180e-01 1.05639778e-01 -3.65583867e-01 1.17956328e+00
4.57692258e-02 4.65337664e-01 7.05330372e-01 -7.45344520e-01
9.40979004e-01 4.07985270e-01 3.28794159e-02 -5.53455114e-01
-3.07283252e-01 -2.10920125e-01 -6.31908059e-01 -2.16537133e-01
6.61545813e-01 -5.79993367e-01 -3.12233597e-01 1.64508402e+00
-5.09772003e-02 2.14959636e-01 4.03485149e-01 6.85667217e-01
9.91508186e-01 1.23205471e+00 2.66142517e-01 -7.75226718e-03
1.32390928e+00 -1.07988286e+00 -5.26523530e-01 -4.61194873e-01
1.00700963e+00 -1.19802141e+00 1.51713574e+00 2.69638538e-01
-9.52805996e-01 -3.65144402e-01 -5.69214642e-01 -1.66424379e-01
-4.61390197e-01 2.46496081e-01 9.37362790e-01 6.41802847e-01
-9.75915432e-01 2.43857086e-01 -6.93739295e-01 -9.57255483e-01
-8.29911828e-02 -1.85526758e-01 -9.32551175e-02 -2.42588371e-01
-1.88245547e+00 1.08371723e+00 3.81930143e-01 -1.51524320e-01
-6.46129608e-01 -7.48991966e-01 -9.00366306e-01 -5.31249382e-02
4.97675121e-01 -5.82375765e-01 1.69085383e+00 -4.90397751e-01
-1.58811557e+00 7.46357918e-01 -9.12132040e-02 -5.98831713e-01
5.39319813e-01 -1.97983161e-01 -3.67259502e-01 -3.34354818e-01
1.90936536e-01 7.84235001e-01 1.89306736e-01 -9.67848182e-01
-6.57636642e-01 1.76604390e-01 5.04444167e-02 3.84612620e-01
-3.32866848e-01 3.37228775e-01 -4.83139187e-01 -6.48370504e-01
-5.20917833e-01 -1.02630186e+00 -3.45799893e-01 -5.37007689e-01
-2.76075065e-01 -5.20161033e-01 4.38477218e-01 -5.94767034e-01
1.13814557e+00 -1.79120779e+00 -4.60310541e-02 -4.39442188e-01
-2.65341192e-01 4.27571654e-01 -4.60659534e-01 1.00874603e+00
2.48138964e-01 4.44614112e-01 -3.29565033e-02 -4.98287886e-01
-3.30726244e-02 2.44902164e-01 -5.09185493e-01 -1.73502296e-01
2.66782314e-01 1.21963036e+00 -1.11379385e+00 -5.12922227e-01
1.85190558e-01 3.37385654e-01 -3.98911864e-01 1.36573613e-01
-5.23932993e-01 3.99859190e-01 -5.72085738e-01 4.04278517e-01
1.34687021e-01 -1.72051311e-01 -1.20566124e-02 4.30202693e-01
-3.66898954e-01 5.59793055e-01 -1.03168559e+00 1.72652507e+00
-9.39620197e-01 7.04694390e-01 -1.25927910e-01 -8.85517299e-01
8.77356768e-01 7.40430474e-01 2.13691175e-01 -6.26647234e-01
4.57645953e-02 3.58678907e-01 3.49305272e-02 -6.23347878e-01
6.99480295e-01 -2.34361544e-01 -5.01750827e-01 8.06521475e-01
2.72348136e-01 -3.96640658e-01 6.07494235e-01 4.14121419e-01
8.01034629e-01 2.26679593e-01 3.55603129e-01 -2.18993768e-01
2.93769956e-01 1.59711301e-01 2.25123033e-01 1.06550944e+00
-1.05839156e-01 5.04438519e-01 3.79370123e-01 -2.85847694e-01
-7.79719770e-01 -8.77211988e-01 9.12222713e-02 1.39905167e+00
-2.69809961e-01 -4.72886384e-01 -6.75765336e-01 -4.51985210e-01
-3.94623131e-01 1.21295774e+00 -3.80655117e-02 1.26460001e-01
-5.47229290e-01 -1.12994051e+00 8.38188231e-01 3.58080775e-01
4.30452645e-01 -1.65493155e+00 -1.92817926e-01 4.87726301e-01
-7.00493753e-01 -1.43225706e+00 -3.19508463e-01 -1.23987056e-01
-7.32415676e-01 -7.22661316e-01 -9.92572725e-01 -1.01182282e+00
1.71920553e-01 2.25688607e-01 1.34783602e+00 -3.48907709e-01
-4.82161529e-02 3.46133381e-01 -6.35597229e-01 -7.32141376e-01
-9.12650645e-01 6.93956077e-01 -3.07298750e-02 -3.91879499e-01
4.56368119e-01 -4.43034470e-01 -3.29293050e-02 1.31078824e-01
-6.55407012e-01 4.36679900e-01 7.39529312e-01 6.42212868e-01
6.92574680e-02 6.26140134e-03 8.60169291e-01 -1.05360639e+00
1.24963033e+00 -6.53176248e-01 -3.86986405e-01 4.10058051e-01
-2.62854427e-01 6.41243756e-02 7.28400826e-01 -5.61199665e-01
-1.52999294e+00 -4.62403715e-01 -3.46107274e-01 1.86034441e-01
-1.73165336e-01 8.51038694e-01 5.17008873e-03 3.92711729e-01
8.28164637e-01 4.27895039e-01 -3.61225039e-01 -5.54827929e-01
5.67195296e-01 7.39191711e-01 1.92515925e-01 -1.03873432e+00
6.10470951e-01 -2.87494082e-02 -6.21050954e-01 -1.06889284e+00
-8.20179820e-01 -4.06710863e-01 -2.23949462e-01 8.37936252e-02
6.81745768e-01 -1.12506950e+00 -3.27851415e-01 3.08002383e-01
-1.40521777e+00 -7.18565762e-01 -1.02849171e-01 4.93025482e-01
-3.15266103e-01 3.51282448e-01 -9.89636719e-01 -8.84719908e-01
-6.01780355e-01 -1.07083881e+00 1.14984286e+00 1.62151352e-01
-5.70147753e-01 -1.36639130e+00 3.63193363e-01 5.03487110e-01
4.28243667e-01 -2.71655321e-01 9.83705282e-01 -8.14293265e-01
-5.28031290e-01 -2.20716730e-01 -1.98099948e-03 1.58691242e-01
5.04462495e-02 -5.00466488e-02 -7.87097454e-01 -1.37920976e-01
-1.52543813e-01 -7.06355274e-01 6.54054701e-01 1.36496112e-01
6.44144237e-01 -3.22353452e-01 -2.56597936e-01 3.99689153e-02
1.04810190e+00 9.22557786e-02 5.49414158e-01 3.78722578e-01
5.64705312e-01 7.86703527e-01 5.82074940e-01 2.62324631e-01
5.68644524e-01 5.78731835e-01 -1.60142094e-01 -6.99259117e-02
-1.45745069e-01 -5.25616586e-01 7.21461296e-01 1.02742124e+00
-8.84668529e-02 -6.92685306e-01 -1.18329763e+00 6.87452018e-01
-1.74572980e+00 -1.10493088e+00 -1.59921139e-01 2.12212777e+00
1.16962302e+00 9.77658182e-02 -7.43053332e-02 -3.05250615e-01
5.24747431e-01 1.77160874e-01 -1.93014801e-01 -5.96760869e-01
-2.11122498e-01 2.95860946e-01 -9.77558941e-02 4.98003393e-01
-7.60343134e-01 1.42890227e+00 6.15195704e+00 1.00073719e+00
-1.10553515e+00 8.67357776e-02 6.98809206e-01 5.82902916e-02
-3.21268976e-01 1.33660346e-01 -1.13940895e+00 3.61684054e-01
1.14277256e+00 -5.46000600e-01 5.41722655e-01 8.24736893e-01
4.78271246e-01 6.45943135e-02 -1.01541138e+00 6.81643903e-01
6.82208911e-02 -1.23642969e+00 1.71598643e-01 -1.44241393e-01
9.02559698e-01 4.21142757e-01 -4.26801369e-02 1.05449080e+00
8.02296817e-01 -8.84432793e-01 4.59334910e-01 2.91427388e-03
5.79458952e-01 -6.00022078e-01 6.73521578e-01 8.92095566e-01
-1.10960340e+00 9.16473195e-02 -4.94480193e-01 -3.32345933e-01
5.74458539e-01 4.41029310e-01 -1.31105399e+00 6.01039469e-01
5.37035525e-01 6.41022205e-01 -4.22284216e-01 8.52876425e-01
-3.99832278e-01 9.26906228e-01 -1.61151737e-01 -3.63510370e-01
5.19466400e-01 -5.15772253e-02 5.75672209e-01 1.44182229e+00
5.30367911e-01 -2.17730016e-01 4.40398932e-01 9.16124403e-01
-1.64951012e-01 5.65160871e-01 -7.62301326e-01 -4.62118566e-01
5.15969753e-01 1.47436023e+00 -5.12868166e-01 -3.60456586e-01
-7.59094238e-01 8.10397804e-01 4.31368262e-01 4.38797951e-01
-6.19588971e-01 -1.80153742e-01 3.70078117e-01 1.65402502e-01
-2.86962509e-01 -3.84434551e-01 -4.67126556e-02 -1.36163676e+00
-2.76659161e-01 -1.11014163e+00 4.12869364e-01 -7.86474526e-01
-1.54137504e+00 7.19900072e-01 1.42524794e-01 -1.01451790e+00
-7.37962961e-01 -4.71884310e-01 -7.30218112e-01 1.09991610e+00
-1.44126511e+00 -1.22832036e+00 1.24073826e-01 3.61195415e-01
1.10558033e+00 -3.10525507e-01 1.01929188e+00 3.12160724e-03
-4.70588624e-01 4.47312653e-01 -2.28901561e-02 1.56644523e-01
1.00130558e+00 -1.11368287e+00 8.30143034e-01 7.40591168e-01
5.53826809e-01 5.51788568e-01 6.86144590e-01 -5.61623454e-01
-1.07783139e+00 -1.12040055e+00 1.56605732e+00 -8.06239903e-01
9.60415840e-01 -7.21565068e-01 -8.48284304e-01 7.82316566e-01
6.15539730e-01 -5.33909380e-01 7.17402697e-01 3.31447393e-01
-2.80687325e-02 2.62834728e-01 -6.91152155e-01 9.37963545e-01
8.80161226e-01 -2.98733473e-01 -6.79873109e-01 5.21824956e-01
9.85239446e-01 -3.69201422e-01 -7.33007491e-01 9.73378420e-02
3.31887931e-01 -6.71760380e-01 7.53273964e-01 -5.10058403e-01
6.14340365e-01 2.19866127e-01 1.48263842e-01 -1.58394396e+00
-3.11392918e-03 -8.77757370e-01 2.24883020e-01 1.60356212e+00
8.19398701e-01 -6.51055276e-01 5.84209740e-01 4.07783151e-01
-1.10629059e-01 -7.47836530e-01 -6.90931082e-01 -8.30925584e-01
5.27930021e-01 -6.74614370e-01 3.58642399e-01 9.20924962e-01
2.59627879e-01 8.46197009e-01 -6.40892982e-01 -2.70501912e-01
1.77710637e-01 7.38846138e-02 8.55019569e-01 -1.10863018e+00
-4.47640181e-01 -3.91911447e-01 3.31356317e-01 -1.06381786e+00
4.87878948e-01 -1.20269191e+00 7.60718882e-02 -1.72601557e+00
2.80362725e-01 -4.99348104e-01 1.25753030e-01 6.71970427e-01
-2.94117481e-01 8.25510547e-02 2.70687312e-01 5.19213155e-02
-5.16868651e-01 6.42201424e-01 1.43703127e+00 -7.14569986e-02
-5.46072423e-01 2.15549618e-01 -1.04973602e+00 5.57061315e-01
9.51685965e-01 -3.74905139e-01 -6.49332464e-01 -5.94408929e-01
2.85067230e-01 1.28681332e-01 -5.92676252e-02 -8.63050103e-01
1.04916550e-01 -2.56875187e-01 7.52745662e-03 -1.02964260e-01
1.39924243e-01 -2.81951934e-01 -2.53914148e-01 1.49261639e-01
-4.19019759e-01 5.27998491e-04 3.43090147e-01 2.22772747e-01
-1.99990258e-01 -2.14312464e-01 3.06074858e-01 -5.53920805e-01
-8.16381752e-01 2.70409971e-01 -7.71660864e-01 4.15999502e-01
7.53457963e-01 1.37390986e-01 -2.61271060e-01 -7.69833088e-01
-4.25527155e-01 4.00097907e-01 4.82959002e-02 1.01649737e+00
3.15122664e-01 -1.18902755e+00 -1.17510986e+00 -1.06781594e-01
2.49747559e-01 -1.55581772e-01 1.07964404e-01 5.74543953e-01
-2.87266612e-01 8.04562688e-01 1.46304503e-01 -3.09661806e-01
-1.11474264e+00 3.20938528e-01 3.26367766e-02 -8.81196082e-01
-5.66114008e-01 6.72047377e-01 9.96462330e-02 -8.31313968e-01
-3.06785852e-02 -1.90233171e-01 -2.46808380e-01 -4.49229479e-02
6.06033325e-01 2.21917212e-01 -1.12617835e-01 -5.23124516e-01
-1.39302546e-02 4.76268642e-02 -2.95606554e-01 -3.87226194e-01
1.18993461e+00 -8.86862949e-02 -7.13860104e-03 7.32784510e-01
6.81700110e-01 1.14207380e-01 -7.91487038e-01 -3.25336754e-01
9.12790671e-02 -2.17014784e-03 -4.60266471e-01 -9.80703831e-01
-5.94159424e-01 1.01737452e+00 -2.77783666e-02 1.03697650e-01
6.84842169e-01 2.00819954e-01 9.77332175e-01 5.56223214e-01
7.92887270e-01 -9.60774481e-01 1.70939103e-01 9.55002546e-01
1.11068070e+00 -1.26718497e+00 -3.65383953e-01 -2.39257768e-01
-1.02792048e+00 8.66404951e-01 7.12229609e-01 3.57120246e-01
2.56011814e-01 2.86887974e-01 2.17637360e-01 2.06464529e-01
-1.04858804e+00 -1.06142096e-01 7.18285516e-02 4.90862608e-01
1.06655884e+00 7.92698562e-02 -5.59720159e-01 8.22733879e-01
-5.72905838e-01 7.51694515e-02 5.03924251e-01 6.95771754e-01
-3.64899278e-01 -1.64473128e+00 -2.54028618e-01 3.09747845e-01
-4.24715161e-01 -5.66537797e-01 -3.69574368e-01 1.03704560e+00
-1.28168970e-01 1.19055712e+00 -2.84192443e-01 -4.49815728e-02
2.85489142e-01 3.81561548e-01 3.87860149e-01 -1.13196111e+00
-5.71601272e-01 -3.13942060e-02 3.84952098e-01 -2.01549008e-01
-1.66411206e-01 -6.33798182e-01 -1.11750913e+00 -3.51708770e-01
-1.13759160e-01 3.40279162e-01 4.10635620e-01 8.90084743e-01
2.34264046e-01 1.51906267e-01 1.83295444e-01 -7.68467009e-01
-4.50704426e-01 -1.28338718e+00 -3.66226882e-01 2.02494070e-01
-3.60225856e-01 -2.48683318e-01 -1.57669395e-01 -4.92792614e-02] | [11.42872142791748, 8.988014221191406] |
e5efbcdf-554f-40b0-a973-b89238a77520 | haloc-hardware-aware-automatic-low-rank | 2301.09422 | null | https://arxiv.org/abs/2301.09422v2 | https://arxiv.org/pdf/2301.09422v2.pdf | HALOC: Hardware-Aware Automatic Low-Rank Compression for Compact Neural Networks | Low-rank compression is an important model compression strategy for obtaining compact neural network models. In general, because the rank values directly determine the model complexity and model accuracy, proper selection of layer-wise rank is very critical and desired. To date, though many low-rank compression approaches, either selecting the ranks in a manual or automatic way, have been proposed, they suffer from costly manual trials or unsatisfied compression performance. In addition, all of the existing works are not designed in a hardware-aware way, limiting the practical performance of the compressed models on real-world hardware platforms. To address these challenges, in this paper we propose HALOC, a hardware-aware automatic low-rank compression framework. By interpreting automatic rank selection from an architecture search perspective, we develop an end-to-end solution to determine the suitable layer-wise ranks in a differentiable and hardware-aware way. We further propose design principles and mitigation strategy to efficiently explore the rank space and reduce the potential interference problem. Experimental results on different datasets and hardware platforms demonstrate the effectiveness of our proposed approach. On CIFAR-10 dataset, HALOC enables 0.07% and 0.38% accuracy increase over the uncompressed ResNet-20 and VGG-16 models with 72.20% and 86.44% fewer FLOPs, respectively. On ImageNet dataset, HALOC achieves 0.9% higher top-1 accuracy than the original ResNet-18 model with 66.16% fewer FLOPs. HALOC also shows 0.66% higher top-1 accuracy increase than the state-of-the-art automatic low-rank compression solution with fewer computational and memory costs. In addition, HALOC demonstrates the practical speedups on different hardware platforms, verified by the measurement results on desktop GPU, embedded GPU and ASIC accelerator. | ['Bo Yuan', 'Dingwen Tao', 'Lizhi Xiang', 'Yang Sui', 'Miao Yin', 'Yu Gong', 'Chengming Zhang', 'Jinqi Xiao'] | 2023-01-20 | null | null | null | null | ['low-rank-compression'] | ['computer-code'] | [ 2.19227523e-01 -4.80670989e-01 -1.97476685e-01 -4.91346240e-01
-6.26031220e-01 -9.55774188e-02 1.65791765e-01 4.62233275e-02
-7.07993448e-01 5.31979620e-01 5.72007410e-02 -4.54896063e-01
-4.64213639e-01 -7.27522790e-01 -6.96934640e-01 -6.69551611e-01
-2.41264198e-02 3.99595588e-01 3.58760148e-01 5.03410809e-02
4.14189458e-01 2.12666750e-01 -1.76150239e+00 3.74832779e-01
9.00988877e-01 1.24284804e+00 6.48346126e-01 5.23426890e-01
3.20126042e-02 6.07369065e-01 -4.39856708e-01 -1.06219649e-01
4.15917575e-01 -1.16633466e-02 -4.27443355e-01 -3.98244143e-01
6.55668974e-01 -5.47111630e-01 -4.32429850e-01 1.21257281e+00
5.53599060e-01 -2.87687927e-01 3.34105730e-01 -9.30216789e-01
-5.41337430e-02 9.50757205e-01 -9.03794527e-01 3.53640437e-01
-3.42263818e-01 -3.23581472e-02 8.02899480e-01 -6.63695335e-01
2.30310604e-01 9.60264146e-01 4.66908783e-01 3.59254330e-01
-1.04713964e+00 -1.21417093e+00 3.03337201e-02 4.48165715e-01
-1.67662859e+00 -3.76604855e-01 6.62103891e-01 -1.74853444e-01
9.40613627e-01 4.01562601e-01 6.34559870e-01 4.87941742e-01
3.61833423e-02 4.98979598e-01 1.17766190e+00 -3.08058560e-01
2.62578547e-01 -1.93335518e-01 5.20378113e-01 5.94572186e-01
7.49038219e-01 -1.98568683e-03 -5.55476487e-01 -1.11120395e-01
6.18250430e-01 -6.23395555e-02 -4.13376957e-01 8.49236473e-02
-9.27252233e-01 4.75757003e-01 4.85507339e-01 2.53761768e-01
-3.28273028e-01 4.57024872e-01 5.21893978e-01 8.75807256e-02
-3.88344191e-02 7.07532763e-02 -5.54447412e-01 -2.35765934e-01
-1.31312704e+00 2.22479969e-01 3.43227476e-01 8.61528933e-01
3.44103009e-01 3.25562716e-01 -3.40861902e-02 8.77029717e-01
4.23837394e-01 4.56079513e-01 7.59992003e-01 -5.36870718e-01
7.44069815e-01 5.66807687e-01 -5.02848089e-01 -1.21459794e+00
-3.97002518e-01 -1.19174993e+00 -1.17956054e+00 1.99413504e-02
-1.33719683e-01 3.08899343e-01 -8.78139734e-01 1.73626912e+00
5.88175934e-03 6.13611639e-01 1.55785635e-01 8.77067924e-01
8.10273111e-01 7.22561300e-01 2.03067377e-01 -2.99509615e-01
1.56067908e+00 -8.27193737e-01 -4.73598808e-01 -2.16186404e-01
6.59346104e-01 -7.83091724e-01 1.07155228e+00 5.37128508e-01
-1.08644891e+00 -6.13763452e-01 -1.54647648e+00 9.61535871e-02
3.20405066e-02 5.81895053e-01 6.45823419e-01 6.59430206e-01
-9.26010489e-01 6.57355011e-01 -8.78986776e-01 1.52738810e-01
4.32460219e-01 7.17625380e-01 -3.56659363e-03 -7.58344680e-02
-1.01453865e+00 3.58709663e-01 5.41615784e-01 -7.97908530e-02
-5.72628677e-01 -9.14399683e-01 -2.53573000e-01 3.64843518e-01
6.19747043e-02 -3.14679414e-01 1.01791310e+00 -6.15477860e-01
-9.98650670e-01 4.19901490e-01 4.98237961e-04 -1.05004334e+00
2.35433996e-01 -1.98785007e-01 -4.93662298e-01 1.88368917e-01
-3.56131613e-01 5.65132678e-01 4.74094719e-01 -8.60374808e-01
-7.84750700e-01 -4.69914287e-01 -9.95141640e-02 2.98755229e-01
-7.73768783e-01 -1.76197544e-01 -6.47031426e-01 -6.44716620e-01
4.62477326e-01 -8.76838326e-01 -2.22446427e-01 -1.75791830e-01
-6.94972947e-02 1.51999101e-01 8.28618586e-01 -5.83582282e-01
1.71941841e+00 -2.07450747e+00 -2.58198351e-01 2.54079908e-01
4.08266753e-01 8.20683241e-01 3.24270576e-02 -2.79338807e-01
1.28926292e-01 1.01447076e-01 3.90567258e-02 -1.40377194e-01
-2.27855116e-01 -2.85266880e-02 -2.30102375e-01 4.07550484e-01
-2.88673162e-01 4.58989680e-01 -3.93545479e-01 -5.81107259e-01
1.31568104e-01 7.44143128e-01 -8.82984936e-01 -1.69105172e-01
2.34724715e-01 -1.37480155e-01 -4.37253296e-01 4.71829504e-01
1.11338127e+00 -2.75220573e-01 1.71333626e-01 -7.45467365e-01
-9.58175883e-02 3.04031819e-01 -1.18210006e+00 1.34754896e+00
-4.63477314e-01 5.38537741e-01 -3.99273559e-02 -9.78585422e-01
9.30523992e-01 7.17698038e-02 2.46200085e-01 -9.15111601e-01
4.27607268e-01 5.12172878e-01 2.69017696e-01 7.00820833e-02
6.34958804e-01 2.84392327e-01 1.47539243e-01 1.35718331e-01
-1.92457989e-01 5.09580433e-01 2.10557297e-01 1.45293593e-01
1.07288396e+00 -4.01872873e-01 9.19308886e-02 -4.99092907e-01
6.79732680e-01 -7.66240135e-02 5.66431761e-01 5.56945622e-01
-1.68475419e-01 5.95364273e-01 1.47268727e-01 -3.62749785e-01
-8.32737684e-01 -6.15664244e-01 -5.45032144e-01 7.41592765e-01
1.22226857e-01 -6.83276236e-01 -9.06669617e-01 -7.11548477e-02
-3.32004696e-01 4.77408648e-01 -6.36883546e-03 -1.26614168e-01
-7.26982951e-01 -9.31237936e-01 6.10713780e-01 4.12244797e-01
1.01112747e+00 -6.10648274e-01 -9.86177385e-01 2.06630781e-01
1.31502375e-01 -1.25151610e+00 -2.25374937e-01 6.68649450e-02
-1.28634667e+00 -8.31778526e-01 -3.00923079e-01 -9.15930331e-01
8.02188039e-01 4.87521738e-01 9.75249588e-01 4.19030666e-01
-2.28065655e-01 -4.66997683e-01 -2.31747806e-01 -6.23076558e-02
1.78560708e-02 3.08558136e-01 1.06783427e-01 -2.06221372e-01
1.88359126e-01 -8.00578713e-01 -8.80642176e-01 1.24926180e-01
-6.92793131e-01 5.55919170e-01 8.73587072e-01 7.32150614e-01
1.04932427e+00 4.01381493e-01 4.78910565e-01 -9.69632268e-01
4.98572290e-01 -7.42503777e-02 -1.02414548e+00 1.13077447e-01
-1.14270651e+00 2.06813827e-01 9.66249883e-01 -4.99055684e-01
-5.07787883e-01 1.62001342e-01 -2.33974800e-01 -6.72680259e-01
2.14620382e-01 6.14175260e-01 -1.35153815e-01 -1.39182076e-01
5.51024079e-01 4.30981696e-01 -3.19633245e-01 -5.01028597e-01
-1.39018416e-01 7.35927165e-01 6.62906170e-01 -4.47257340e-01
7.16327369e-01 9.19816643e-02 1.89196780e-01 -7.12599516e-01
-4.05440956e-01 -3.19917440e-01 -8.72134492e-02 -3.69056966e-03
5.53927064e-01 -1.26079571e+00 -6.44737899e-01 3.78950030e-01
-8.73240590e-01 -3.01702116e-02 2.95900613e-01 6.23844862e-01
-1.77621648e-01 1.72970787e-01 -5.86621404e-01 -5.63605785e-01
-1.07428622e+00 -1.51396513e+00 8.05742204e-01 2.97737807e-01
2.05948204e-01 -2.66380131e-01 -5.43936670e-01 3.45036983e-01
6.33519709e-01 2.18674317e-02 1.02920222e+00 -4.53999788e-01
-8.70248258e-01 -2.70300031e-01 -6.15107477e-01 8.56605023e-02
-4.16906387e-01 -2.32671127e-01 -8.46983850e-01 -4.68453914e-01
-1.01476677e-01 -1.88225172e-02 8.66020322e-01 3.98717850e-01
1.55261302e+00 -4.18980718e-01 -2.58402169e-01 9.65476274e-01
1.89239419e+00 3.76165211e-01 6.25016272e-01 2.27105156e-01
6.90424323e-01 -5.15743494e-02 6.01794958e-01 5.41208804e-01
6.46398813e-02 8.02195549e-01 5.67297876e-01 7.32508302e-02
-2.86307037e-01 -2.34695762e-01 1.31079599e-01 1.36822855e+00
7.63036907e-02 -4.52752449e-02 -7.76775420e-01 2.77590007e-01
-1.62271941e+00 -7.16753721e-01 -3.03799510e-01 2.37576890e+00
8.65687072e-01 5.23691952e-01 -3.68381232e-01 5.55106461e-01
7.03417957e-01 9.88843441e-02 -4.70762253e-01 -1.25694886e-01
-1.24996826e-02 3.84855509e-01 9.33653116e-01 3.44929427e-01
-9.58457530e-01 7.40535080e-01 5.15679407e+00 1.23735452e+00
-1.35177290e+00 1.03122607e-01 8.88088346e-01 -4.80426282e-01
2.00143810e-02 -7.64791481e-03 -1.18911934e+00 5.33881605e-01
1.11906016e+00 -2.24667355e-01 4.67672139e-01 9.81405973e-01
2.01001003e-01 2.19708845e-01 -7.64352918e-01 1.38238645e+00
-1.38481870e-01 -1.47010398e+00 1.16961107e-01 3.22149873e-01
6.20154321e-01 1.22474797e-01 8.75800103e-02 2.03189969e-01
-3.23741227e-01 -8.88676941e-01 8.15870106e-01 2.27133352e-02
9.75444257e-01 -1.12556374e+00 8.05455327e-01 3.70194346e-01
-1.47678244e+00 -8.39812532e-02 -7.93054581e-01 7.25095347e-02
-1.73945427e-02 7.40284741e-01 -4.50687647e-01 1.39141589e-01
6.88101828e-01 4.20612127e-01 -7.20954955e-01 1.00819623e+00
2.42697835e-01 9.42833364e-01 -4.21226770e-01 -1.17030695e-01
7.36793354e-02 -6.77435324e-02 2.03916892e-01 1.13427484e+00
4.72263604e-01 4.00772721e-01 -3.86990532e-02 4.45538431e-01
-1.59585819e-01 2.67118156e-01 -1.51028916e-01 2.21351534e-01
8.11705709e-01 1.19809294e+00 -7.41527021e-01 -3.89577061e-01
-2.31386140e-01 4.50444013e-01 2.49863639e-01 1.45769026e-02
-1.22811449e+00 -4.17579710e-01 5.97296059e-01 3.95820439e-01
1.08506672e-01 -2.23209709e-01 -4.64320719e-01 -8.73839855e-01
7.93227255e-02 -9.49601829e-01 6.34663403e-02 -5.22311151e-01
-4.33992535e-01 9.93511438e-01 3.79703268e-02 -1.40673530e+00
1.04039833e-01 -4.90061015e-01 -3.87281179e-01 6.74142361e-01
-1.50746119e+00 -7.04905510e-01 -4.67440397e-01 2.70565182e-01
6.55913830e-01 -3.96222860e-01 5.31634688e-01 7.89966464e-01
-6.37089193e-01 1.16909325e+00 1.76719919e-01 -2.17843577e-01
2.89956987e-01 -5.43517113e-01 2.90326923e-01 9.76591766e-01
9.28400666e-04 8.39921474e-01 6.07403100e-01 -3.48937899e-01
-1.68067527e+00 -1.25120592e+00 7.55378425e-01 6.05904639e-01
3.38130236e-01 -2.61275619e-01 -9.49876904e-01 -1.06615499e-02
-1.38972044e-01 4.96365204e-02 3.44464898e-01 -2.12454677e-01
-2.46374384e-01 -6.71725273e-01 -1.19167721e+00 7.34780669e-01
1.18117833e+00 -2.24198490e-01 -3.09604220e-03 2.82901883e-01
8.99878263e-01 -5.23501575e-01 -6.73018813e-01 7.05656052e-01
6.41288638e-01 -9.73428845e-01 1.04676580e+00 2.82890759e-02
4.48094964e-01 -5.02824426e-01 -6.24523580e-01 -5.31025290e-01
-3.83585930e-01 -1.60097092e-01 -2.16351822e-01 1.15179706e+00
3.91250312e-01 -5.25591850e-01 1.01077795e+00 4.27965105e-01
-2.13385612e-01 -1.17157543e+00 -8.78403902e-01 -6.38306499e-01
-3.01173478e-01 -5.68659902e-01 7.99264729e-01 6.33645415e-01
-4.28352714e-01 2.62875885e-01 -2.72168905e-01 2.47709006e-01
7.46643722e-01 -2.09590998e-02 5.44777334e-01 -1.14281404e+00
-4.36683625e-01 -5.85167766e-01 -6.53760910e-01 -1.02465248e+00
-9.59691182e-02 -9.08289135e-01 -3.00378948e-01 -1.33801067e+00
3.29022467e-01 -8.15730214e-01 -5.33513129e-01 4.65938359e-01
1.73096895e-01 4.40529555e-01 3.17428589e-01 3.94686639e-01
-4.98515964e-01 2.61609465e-01 8.70199978e-01 -1.52188554e-01
-1.16346799e-01 -3.83688092e-01 -7.44756281e-01 5.97395957e-01
1.13107145e+00 -4.42697912e-01 -7.88598478e-01 -7.14975476e-01
7.61963576e-02 -4.26353328e-02 1.02103874e-01 -1.66092324e+00
3.45645040e-01 1.28755316e-01 2.78180838e-01 -9.05825913e-01
2.19385654e-01 -8.53153706e-01 3.84750724e-01 8.10125768e-01
-3.38314742e-01 3.23348343e-01 4.28409576e-01 4.15382624e-01
-2.46976241e-01 -3.75340998e-01 9.96642053e-01 2.74694741e-01
-7.64904559e-01 3.20312828e-01 1.33591458e-01 -2.56783247e-01
7.54487157e-01 -2.68343955e-01 -5.00599623e-01 1.51880388e-03
-1.77198127e-01 -4.67387252e-02 2.50614285e-01 1.51347771e-01
7.65968204e-01 -1.33561838e+00 -7.12727427e-01 5.25406361e-01
-1.33769810e-01 -4.07537445e-02 4.26400661e-01 6.20832503e-01
-8.97415936e-01 5.64328849e-01 -1.38452053e-01 -6.99283183e-01
-1.69352508e+00 2.31323078e-01 -3.55013423e-02 -5.36062896e-01
-5.28241098e-01 7.30090320e-01 2.04600140e-01 2.43236989e-01
3.66785794e-01 -3.82432193e-01 -2.74947017e-01 -4.25006717e-01
7.18561709e-01 4.00399566e-01 4.39525396e-01 -6.73822880e-01
-3.17907661e-01 7.91754901e-01 -2.51510412e-01 5.87002449e-02
1.21321905e+00 1.86107248e-01 -1.09522091e-02 -2.63941854e-01
1.40200996e+00 -2.39986926e-01 -1.01183844e+00 9.51965339e-03
-2.27034941e-01 -4.78713214e-01 6.21523499e-01 -6.52709842e-01
-1.62848413e+00 7.72001922e-01 1.17054236e+00 -3.34285200e-01
1.65854621e+00 -4.77658063e-01 9.24473107e-01 1.71546459e-01
6.79938197e-01 -9.33204710e-01 -3.59324396e-01 4.98983622e-01
6.64012849e-01 -7.54400790e-01 4.39980537e-01 -4.36440170e-01
-3.11671048e-01 7.36508191e-01 8.47815931e-01 -1.49350822e-01
7.64065385e-01 4.90275651e-01 -2.76611000e-01 -1.37232440e-02
-9.78472531e-01 2.03889877e-01 2.03680709e-01 2.34225243e-01
4.65774179e-01 2.55337656e-01 -8.68012726e-01 8.63930285e-01
-6.16080701e-01 -4.70269211e-02 3.13778311e-01 6.94031656e-01
-6.20367348e-01 -9.71068025e-01 -2.20498130e-01 8.48923624e-01
-6.03348672e-01 -3.75456363e-01 2.92705625e-01 4.80035543e-01
1.71955094e-01 8.48501563e-01 8.16890001e-02 -8.97496760e-01
2.06237242e-01 -2.73235500e-01 3.18988532e-01 -1.86645925e-01
-5.72250545e-01 2.77074635e-01 -3.56273353e-02 -5.02406657e-01
1.06797209e-02 -2.84278899e-01 -1.53253055e+00 -5.51923394e-01
-4.91257906e-01 2.19643731e-02 9.54744160e-01 4.78358328e-01
6.56590343e-01 6.99596524e-01 3.72570515e-01 -5.14793158e-01
-6.50610209e-01 -7.46315062e-01 -3.97125900e-01 1.91286892e-01
-6.65091202e-02 -4.39533263e-01 -2.25948632e-01 -2.69066125e-01] | [8.575440406799316, 3.0162808895111084] |
f4997b8c-068a-4787-83c2-f7d05456efa8 | learning-to-segment-with-limited-annotations | 2205.13109 | null | https://arxiv.org/abs/2205.13109v1 | https://arxiv.org/pdf/2205.13109v1.pdf | Learning to segment with limited annotations: Self-supervised pretraining with regression and contrastive loss in MRI | Obtaining manual annotations for large datasets for supervised training of deep learning (DL) models is challenging. The availability of large unlabeled datasets compared to labeled ones motivate the use of self-supervised pretraining to initialize DL models for subsequent segmentation tasks. In this work, we consider two pre-training approaches for driving a DL model to learn different representations using: a) regression loss that exploits spatial dependencies within an image and b) contrastive loss that exploits semantic similarity between pairs of images. The effect of pretraining techniques is evaluated in two downstream segmentation applications using Magnetic Resonance (MR) images: a) liver segmentation in abdominal T2-weighted MR images and b) prostate segmentation in T2-weighted MR images of the prostate. We observed that DL models pretrained using self-supervision can be finetuned for comparable performance with fewer labeled datasets. Additionally, we also observed that initializing the DL model using contrastive loss based pretraining performed better than the regression loss. | ['Ali Bilgin', 'Maria Altbach', 'Diego Martin', 'Rohit Philip', 'Zhiyang Fu', 'Lavanya Umapathy'] | 2022-05-26 | null | null | null | null | ['liver-segmentation'] | ['medical'] | [ 5.67091227e-01 7.43973017e-01 -3.72131675e-01 -1.01706970e+00
-9.04577196e-01 -5.03467381e-01 5.25391340e-01 5.26412010e-01
-9.18056548e-01 7.05913186e-01 5.12741096e-02 -2.64637649e-01
4.65449840e-02 -5.06233513e-01 -8.74212384e-01 -4.93313909e-01
-3.23921561e-01 7.76596844e-01 2.68497676e-01 1.31312147e-01
-4.12895391e-03 6.46267772e-01 -8.57848108e-01 3.27464849e-01
8.54588807e-01 7.26885259e-01 4.78291035e-01 6.29846513e-01
-7.08133429e-02 8.62873852e-01 -3.37952524e-01 1.78673808e-02
4.39289868e-01 -5.05607247e-01 -1.16452074e+00 2.49313250e-01
6.20989263e-01 -1.47390068e-01 5.71865179e-02 1.06105614e+00
4.61041063e-01 2.01351181e-01 6.59265518e-01 -8.56802583e-01
-4.28434938e-01 9.34050024e-01 -4.61919963e-01 7.04535544e-01
-2.21288458e-01 2.28738919e-01 6.65766895e-01 -4.73626643e-01
7.96985805e-01 1.05037618e+00 8.68883014e-01 6.64047182e-01
-1.70196831e+00 -4.06280875e-01 1.54900085e-02 -3.26270580e-01
-1.05730677e+00 -1.12586558e-01 5.29469192e-01 -5.18749356e-01
7.00680673e-01 -7.78127015e-02 3.52155477e-01 7.76007891e-01
1.83728680e-01 7.51558423e-01 1.38661551e+00 -4.53284949e-01
1.70357645e-01 2.68878043e-01 3.91260952e-01 8.70326221e-01
2.33691290e-01 3.94754887e-01 7.75814876e-02 9.45437234e-04
9.84489560e-01 -1.28431350e-01 6.75038099e-02 -6.36497200e-01
-1.06171441e+00 1.00066388e+00 9.33674037e-01 5.34381807e-01
-3.04919243e-01 2.20340043e-01 6.28501475e-01 1.95253536e-01
3.27289134e-01 9.08121109e-01 -5.80105662e-01 4.66146618e-01
-1.24409759e+00 -2.16374576e-01 5.66218734e-01 7.40426540e-01
9.71195161e-01 -9.13541317e-02 -2.85351694e-01 7.28455067e-01
3.03101242e-01 1.65435765e-02 8.03335786e-01 -7.67050385e-01
3.07921171e-01 4.14868951e-01 -3.48616153e-01 -5.25320768e-01
-7.71183610e-01 -5.06920159e-01 -5.29153585e-01 1.52231991e-01
6.37652993e-01 -4.37094986e-01 -1.46089339e+00 1.71790302e+00
1.13267086e-01 -3.58479396e-02 8.13884139e-02 9.16259766e-01
9.06448305e-01 2.04488158e-01 5.61582148e-01 -1.22416489e-01
9.60873663e-01 -1.19315088e+00 -5.10751247e-01 -4.23196167e-01
9.24558580e-01 -5.33225775e-01 1.09019458e+00 3.21856476e-02
-1.09987819e+00 -6.44341230e-01 -1.00745320e+00 -3.60358767e-02
-3.02326053e-01 1.24655083e-01 6.87433481e-01 6.41936004e-01
-1.13716257e+00 9.53749299e-01 -1.31393194e+00 -3.36990863e-01
6.95637524e-01 7.44835258e-01 -2.48548999e-01 -1.13405392e-01
-9.44438756e-01 9.29663718e-01 5.38056254e-01 -1.74852535e-01
-1.27911282e+00 -8.67804289e-01 -1.10274231e+00 -1.76676244e-01
2.78259009e-01 -3.11114788e-01 1.30526149e+00 -1.53594673e+00
-1.26085365e+00 1.41933155e+00 3.70820105e-01 -7.37431765e-01
7.20054030e-01 -9.00713610e-04 -7.46502206e-02 3.14820319e-01
2.14187458e-01 1.27636647e+00 8.28070402e-01 -1.51044559e+00
-3.20162445e-01 -3.96430373e-01 -4.74301912e-02 3.54751498e-01
2.40312174e-01 -1.53822035e-01 -1.70352489e-01 -6.78933978e-01
1.61294177e-01 -9.84456837e-01 -7.50583410e-01 -2.55595028e-01
-6.01537228e-01 6.72800913e-02 6.36581421e-01 -6.15118146e-01
6.59307837e-01 -1.89594722e+00 -9.05873999e-02 2.69920319e-01
5.69577590e-02 1.35912418e-01 -2.79130459e-01 -3.27676088e-01
-3.12233448e-01 1.54904947e-01 -4.95121449e-01 -2.91917503e-01
-3.56977463e-01 4.27289873e-01 1.16408296e-01 4.78980988e-01
2.44576439e-01 9.65350032e-01 -1.02435255e+00 -7.28068829e-01
1.60711870e-01 2.94528484e-01 -5.54326892e-01 3.34441721e-01
-2.13548362e-01 1.03422713e+00 -1.93698883e-01 4.47076797e-01
3.73275518e-01 -3.88992399e-01 3.23565394e-01 -3.88250113e-01
1.83961257e-01 1.17586710e-01 -6.31617904e-01 1.96373165e+00
-4.44081575e-01 4.03540343e-01 2.11932763e-01 -1.17267656e+00
8.43981683e-01 8.42591822e-02 6.11616969e-01 -7.21067309e-01
2.20448300e-01 9.09768641e-02 2.44666934e-01 -4.16456789e-01
1.80777758e-01 -4.15783077e-01 -6.99890703e-02 5.31113148e-01
5.31977892e-01 -4.47630346e-01 3.06903273e-01 -5.05607128e-02
9.53756392e-01 1.81859061e-01 6.07136935e-02 -7.75687993e-01
4.35625285e-01 3.36354107e-01 3.93690377e-01 8.87933671e-01
-4.36763972e-01 7.26740420e-01 2.62850076e-01 -7.37774014e-01
-1.03911126e+00 -1.04469025e+00 -3.95300716e-01 1.14063311e+00
1.37106359e-01 8.69259238e-02 -9.08224463e-01 -1.29449666e+00
9.64770652e-03 6.99812353e-01 -7.70953059e-01 -1.69602260e-01
-9.08466578e-01 -9.90502715e-01 6.47423923e-01 7.70223618e-01
4.41070646e-01 -1.16100121e+00 -6.72010660e-01 1.62680790e-01
1.79250419e-01 -1.07167411e+00 -5.11599422e-01 8.08409691e-01
-1.43728328e+00 -1.17817700e+00 -7.71000028e-01 -1.01592934e+00
1.26416028e+00 -3.57081711e-01 1.40265727e+00 1.10743262e-01
-6.01485372e-01 5.58564246e-01 1.35353627e-02 -7.79957771e-02
-7.00386345e-01 4.07105237e-01 -1.61810264e-01 -4.52394545e-01
1.18305109e-01 -2.29616955e-01 -5.47548473e-01 3.46700996e-01
-1.06094897e+00 -6.72530904e-02 7.64382362e-01 9.40322876e-01
1.00901413e+00 -2.61732638e-01 4.44186747e-01 -1.41375136e+00
5.63766062e-01 -4.01941538e-01 -4.99224782e-01 2.21366808e-01
-6.78183854e-01 3.11194718e-01 4.94201005e-01 -5.40016532e-01
-1.04733157e+00 4.96672004e-01 -7.86832795e-02 -2.31739730e-01
-3.53302360e-01 4.45457101e-01 3.56560737e-01 -1.89753562e-01
9.19996619e-01 -7.50068650e-02 1.29575297e-01 -3.65660727e-01
2.42645428e-01 -6.14652829e-03 6.17630363e-01 -6.66980863e-01
3.82208496e-01 4.18789893e-01 -9.94469747e-02 -4.78433579e-01
-9.51192915e-01 -1.51064262e-01 -1.08936143e+00 8.83839130e-02
1.15503943e+00 -6.75751805e-01 -4.48913164e-02 7.44119287e-03
-4.91377115e-01 -1.18675113e+00 -7.18269765e-01 6.65832341e-01
-6.49865031e-01 1.65798947e-01 -8.39557350e-01 -9.05666426e-02
-1.85848534e-01 -1.52000034e+00 7.78065205e-01 3.81146640e-01
-2.24510580e-01 -1.47135615e+00 1.72871962e-01 4.64016795e-01
4.99311686e-01 2.05458641e-01 9.48270202e-01 -1.12084317e+00
-2.38764793e-01 -1.40643865e-02 -2.86897242e-01 4.87545162e-01
3.95676911e-01 -4.50785697e-01 -8.84852946e-01 -5.51857710e-01
-1.51362970e-01 -7.93691158e-01 8.83404315e-01 7.43405402e-01
1.40345955e+00 -3.59090418e-02 -2.97830582e-01 7.62639225e-01
1.32956696e+00 6.76888824e-02 4.82473761e-01 3.17892104e-01
6.88626468e-01 5.53934395e-01 4.33432788e-01 -1.17096826e-01
1.59973249e-01 3.10086131e-01 2.07636833e-01 -7.71361649e-01
-3.23930055e-01 -8.33534375e-02 -3.17920968e-02 3.58068138e-01
1.54078677e-01 3.20669293e-01 -1.13655257e+00 5.54162621e-01
-1.44684231e+00 -4.20476794e-01 2.26599887e-01 2.16361427e+00
1.19264114e+00 2.59273559e-01 5.41495383e-02 -4.35370356e-01
7.87532389e-01 -1.01512969e-01 -8.03882360e-01 -3.42340708e-01
3.16448182e-01 3.05379182e-01 9.51517284e-01 7.17792988e-01
-1.48300588e+00 9.96455669e-01 7.04591608e+00 4.37717080e-01
-1.51414549e+00 2.91039348e-01 1.22675753e+00 2.83331305e-01
-1.00922324e-01 -1.61181659e-01 -7.55489111e-01 2.68930703e-01
8.61807168e-01 2.15599164e-01 1.06373668e-01 8.67594481e-01
8.13056454e-02 -3.09026361e-01 -1.38936210e+00 6.25972152e-01
-3.60378623e-02 -1.31424534e+00 -1.07225172e-01 -2.20655546e-01
1.00659788e+00 4.49776977e-01 -3.55846086e-03 4.86858040e-01
7.46224582e-01 -1.32625389e+00 2.08742544e-01 1.90654874e-01
8.21938694e-01 -5.46274185e-01 8.43009889e-01 1.44327611e-01
-6.47267640e-01 2.09160119e-01 -1.31236419e-01 4.85544890e-01
-1.00656942e-01 4.72907752e-01 -1.37324238e+00 1.65901497e-01
5.47107697e-01 3.79054755e-01 -8.42556238e-01 9.83470261e-01
-3.59534502e-01 7.45046616e-01 -2.23606795e-01 5.10001302e-01
6.92981422e-01 -2.14469079e-02 1.65056512e-01 1.42832363e+00
-2.84944862e-01 -1.55103967e-01 7.12697744e-01 1.01717949e+00
-1.64993122e-01 1.38167394e-02 -3.58289272e-01 5.25587276e-02
-8.89732838e-02 1.38377750e+00 -1.33248663e+00 -2.80403763e-01
-1.00369364e-01 8.88266921e-01 3.40827852e-01 2.12408483e-01
-6.48003042e-01 7.27921426e-02 -8.27130526e-02 2.97395945e-01
1.30613059e-01 -4.37362678e-02 -3.39902133e-01 -8.12578082e-01
-5.78541994e-01 -6.49599373e-01 5.93675673e-01 -4.85916615e-01
-1.33150864e+00 8.45538020e-01 2.11277246e-01 -6.71443999e-01
-2.99119711e-01 -4.27930504e-01 -4.80779648e-01 6.80185080e-01
-1.71819556e+00 -9.79248047e-01 -2.75975436e-01 4.65712368e-01
4.47021067e-01 -1.61512420e-02 7.36597717e-01 2.10498378e-01
-4.44166452e-01 3.88207376e-01 -1.79358006e-01 3.94840717e-01
7.85029352e-01 -1.67171252e+00 8.51891339e-02 5.87079585e-01
-4.01754640e-02 4.27169412e-01 4.70583200e-01 -7.91595519e-01
-6.97957039e-01 -1.31523025e+00 3.51591617e-01 -1.48850560e-01
2.74784356e-01 -6.70440048e-02 -9.18870509e-01 1.09422839e+00
2.44536206e-01 5.59659541e-01 8.53408754e-01 -1.43038914e-01
7.72447735e-02 7.46948645e-02 -1.62611771e+00 2.97852725e-01
5.02770364e-01 -3.32373172e-01 -7.07498968e-01 5.41498601e-01
4.47632700e-01 -7.20400393e-01 -1.05902255e+00 5.47411382e-01
8.99330154e-02 -6.69475436e-01 9.85368609e-01 -7.07205892e-01
4.08639967e-01 2.12091152e-02 1.95715666e-01 -1.44534719e+00
-1.51067093e-01 -3.87713850e-01 4.18325752e-01 7.16527224e-01
5.74686944e-01 -2.57351011e-01 1.03893983e+00 9.12083864e-01
-2.63548344e-01 -6.05878890e-01 -5.14623344e-01 -5.74570775e-01
3.96566302e-01 6.06046915e-02 -1.46741308e-02 1.24513292e+00
-2.32404634e-01 3.05350274e-01 2.34887496e-01 4.79403362e-02
7.10762143e-01 -1.11777045e-01 3.63071591e-01 -1.05098176e+00
-1.83973074e-01 -2.65766442e-01 -3.59823376e-01 -8.07456672e-01
2.50040591e-01 -1.36767375e+00 1.62366375e-01 -1.39701760e+00
3.69757384e-01 -1.02208030e+00 -4.17135060e-01 7.93927550e-01
-1.81874245e-01 3.08720201e-01 3.94255109e-02 2.17999861e-01
-5.43551862e-01 6.92892149e-02 1.46590590e+00 -2.87254423e-01
-3.81204456e-01 7.77452216e-02 -4.28279310e-01 8.02613378e-01
7.88801014e-01 -7.11801052e-01 -5.95647097e-01 -4.69192833e-01
-3.49505216e-01 1.55263796e-01 1.90371394e-01 -8.91876221e-01
1.39963627e-01 1.00505479e-01 7.30628431e-01 -4.00522619e-01
-1.82160616e-01 -7.56357968e-01 -3.08270305e-01 7.81640470e-01
-9.18950379e-01 -2.03096103e-02 5.10409594e-01 2.38243461e-01
-1.99653581e-01 -6.27237976e-01 1.29672873e+00 -6.60308182e-01
-8.11662257e-01 1.96883604e-01 -3.37357402e-01 3.13899875e-01
1.06101906e+00 -1.64134398e-01 4.70011830e-02 -1.19870774e-01
-1.40551245e+00 3.29217613e-01 2.98534334e-01 8.99336189e-02
4.33364838e-01 -1.02088797e+00 -5.32494962e-01 2.47457534e-01
-2.41236031e-01 2.28496164e-01 3.58963269e-03 9.76143718e-01
-7.78192163e-01 3.21511775e-01 -5.17081380e-01 -9.89616692e-01
-1.12521100e+00 5.07047594e-01 7.06119537e-01 -8.79841506e-01
-5.35094261e-01 9.88810539e-01 3.31920326e-01 -1.14211738e+00
3.88926268e-01 -6.20578051e-01 -1.62889123e-01 -1.49509937e-01
2.25661471e-01 -1.81958303e-02 2.65933812e-01 -4.92313474e-01
-2.07247704e-01 7.46094212e-02 -4.97745574e-01 -1.32461153e-02
1.37906742e+00 6.26807213e-02 1.07700884e-01 2.48019218e-01
1.37249696e+00 -4.02645469e-01 -1.51741004e+00 -3.42613608e-01
3.95937026e-01 -7.79826939e-02 4.32774276e-01 -1.10689902e+00
-1.44767690e+00 7.08119810e-01 9.01485503e-01 8.94672889e-03
8.12472820e-01 9.30487830e-03 5.20523131e-01 3.97290587e-01
1.85365289e-01 -1.01889443e+00 2.39782795e-01 2.29491398e-01
5.80619156e-01 -1.60274315e+00 9.00073498e-02 -2.21818537e-01
-9.29765105e-01 1.11450398e+00 7.97010005e-01 -3.58037710e-01
5.79814911e-01 4.83071208e-01 3.45144540e-01 -3.53278846e-01
-1.89020008e-01 -1.73026651e-01 3.76632929e-01 5.57394981e-01
6.19404197e-01 -7.62456730e-02 -1.24899909e-01 3.64275575e-01
1.39076963e-01 7.29393065e-02 4.00563866e-01 1.02408445e+00
-1.18005909e-01 -1.07305682e+00 -1.85141340e-01 6.95182621e-01
-6.14142716e-01 -1.56939477e-01 -2.68214911e-01 9.09102678e-01
1.59618892e-02 4.29217190e-01 2.79222906e-01 1.34688005e-01
8.29948783e-02 1.28238425e-02 7.63260722e-01 -1.10008216e+00
-1.07829916e+00 1.16853714e-01 -2.55183969e-02 -4.19819206e-01
-5.23052156e-01 -6.23267591e-01 -1.66304028e+00 4.43439275e-01
-3.46942216e-01 2.02634722e-01 5.91532052e-01 9.73066807e-01
6.78625181e-02 7.18385220e-01 4.76911843e-01 -1.04429436e+00
-6.40336514e-01 -8.79033625e-01 -5.04588604e-01 6.33899033e-01
2.37228096e-01 -5.39636910e-01 -2.30803013e-01 4.33792740e-01] | [14.739346504211426, -2.226040840148926] |
d4857e67-53dd-476c-a042-ef04a144c85d | segsalsa-str-a-convex-formulation-to | 1504.07028 | null | http://arxiv.org/abs/1504.07028v1 | http://arxiv.org/pdf/1504.07028v1.pdf | SegSALSA-STR: A convex formulation to supervised hyperspectral image segmentation using hidden fields and structure tensor regularization | We present a supervised hyperspectral image segmentation algorithm based on a
convex formulation of a marginal maximum a posteriori segmentation with hidden
fields and structure tensor regularization: Segmentation via the Constraint
Split Augmented Lagrangian Shrinkage by Structure Tensor Regularization
(SegSALSA-STR). This formulation avoids the generally discrete nature of
segmentation problems and the inherent NP-hardness of the integer optimization
associated.
We extend the Segmentation via the Constraint Split Augmented Lagrangian
Shrinkage (SegSALSA) algorithm by generalizing the vectorial total variation
prior using a structure tensor prior constructed from a patch-based Jacobian.
The resulting algorithm is convex, time-efficient and highly parallelizable.
This shows the potential of combining hidden fields with convex optimization
through the inclusion of different regularizers. The SegSALSA-STR algorithm is
validated in the segmentation of real hyperspectral images. | ['Jelena Kovacevic', 'Jose Bioucas-Dias', 'Filipe Condessa'] | 2015-04-27 | null | null | null | null | ['hyperspectral-image-segmentation'] | ['computer-vision'] | [ 8.59081328e-01 2.26208881e-01 2.23308265e-01 -2.14513063e-01
-6.77421570e-01 -5.46046436e-01 2.96480477e-01 -2.24260874e-02
-3.79258692e-01 5.86560011e-01 -1.62818506e-01 -2.21272379e-01
-6.67509794e-01 -6.17192209e-01 -5.21776140e-01 -1.21140957e+00
4.15665023e-02 5.03384054e-01 -1.43228188e-01 -1.55508533e-01
3.47234398e-01 7.29040444e-01 -1.10740256e+00 -2.27400705e-01
1.05458069e+00 9.20617998e-01 3.50768328e-01 4.74695534e-01
1.16479009e-01 2.70562142e-01 7.39969909e-02 1.34450421e-01
7.47246206e-01 -2.71841377e-01 -1.12344623e+00 1.23725140e+00
7.14974523e-01 -8.65497962e-02 5.05388319e-01 1.12822640e+00
1.30657200e-02 3.61149937e-01 4.11109179e-01 -9.71396863e-01
-2.70587772e-01 1.25598580e-01 -9.90983248e-01 -4.04345125e-01
9.19470638e-02 -6.94435537e-02 1.05955517e+00 -8.49771321e-01
7.28227377e-01 1.05393958e+00 7.78676450e-01 -2.89977759e-01
-1.88037801e+00 3.04641217e-01 1.89503342e-01 -3.33532333e-01
-1.44514525e+00 -1.52638396e-02 6.28979385e-01 -9.34160471e-01
9.29987967e-01 5.29487252e-01 5.17400682e-01 3.22437100e-02
-2.08413348e-01 5.79727054e-01 1.48332262e+00 -5.08587062e-01
3.86370540e-01 -7.46453926e-02 3.90718460e-01 9.11980450e-01
3.94896328e-01 -3.69785763e-02 -6.48376122e-02 -3.07169497e-01
6.16525650e-01 -1.72044456e-01 -4.27782953e-01 -5.01096845e-01
-1.20827389e+00 9.80807841e-01 2.94858724e-01 9.11357254e-02
-4.95716780e-01 -8.96678269e-02 5.42601794e-02 1.73223689e-01
8.89100075e-01 4.81848419e-01 -3.66916925e-01 7.60542393e-01
-1.59058607e+00 1.40464842e-01 8.64738762e-01 8.21098089e-01
1.39009261e+00 1.64682046e-01 -3.93094048e-02 5.93322814e-01
8.08852136e-01 5.32022417e-01 -3.16646934e-01 -1.51092017e+00
8.68802369e-02 6.01294219e-01 4.63990793e-02 -8.31775486e-01
-4.73970294e-01 -5.98397017e-01 -6.35449767e-01 5.81915915e-01
4.53632742e-01 -1.26918718e-01 -9.97788131e-01 1.29397154e+00
5.22187412e-01 -1.09589934e-01 -1.50492519e-01 9.14160609e-01
1.12718649e-01 7.73259342e-01 -8.96950439e-02 -6.83553398e-01
9.36327875e-01 -9.55953956e-01 -5.26827395e-01 -6.63300380e-02
7.28950739e-01 -8.26438487e-01 5.20101607e-01 7.25916207e-01
-9.91503179e-01 -1.46058142e-01 -1.05021453e+00 -1.69925258e-01
-1.58823237e-01 3.53577495e-01 6.99352324e-01 8.41019332e-01
-1.13275898e+00 7.29918063e-01 -1.04993081e+00 -5.68376156e-03
2.33734652e-01 7.73778141e-01 -2.58474469e-01 1.55236199e-01
-3.67285192e-01 5.96397281e-01 6.60758555e-01 3.54840457e-01
-2.78211474e-01 -7.61793196e-01 -8.64938200e-01 -2.31907293e-01
5.40750802e-01 -4.98170346e-01 6.41257584e-01 -1.11802876e+00
-1.65868485e+00 9.86944735e-01 -1.75783843e-01 -2.09494755e-01
3.90599430e-01 7.63110667e-02 3.35573733e-01 4.01882321e-01
3.37794721e-02 6.02253258e-01 9.20397818e-01 -1.51038146e+00
-1.97112873e-01 -4.90106970e-01 -7.47690499e-02 1.83979914e-01
-2.41372772e-02 -1.77725062e-01 -2.98570562e-02 -4.43411618e-01
8.41530263e-01 -1.38581598e+00 -6.81249976e-01 -9.97698605e-02
-6.02352858e-01 3.34377587e-01 8.54340434e-01 -1.13635206e+00
7.42391646e-01 -2.00994802e+00 7.38742292e-01 6.78352952e-01
2.52253890e-01 4.28733639e-02 -1.27259418e-01 4.34034139e-01
-2.05232516e-01 9.17753875e-02 -1.24204731e+00 -3.65641683e-01
-1.40324518e-01 3.97723228e-01 -1.13200113e-01 8.51839662e-01
1.09951183e-01 5.38596451e-01 -7.71995246e-01 -1.63885131e-01
3.64112020e-01 4.61097389e-01 -6.45891130e-01 -1.10711657e-01
-4.75382239e-01 5.27655125e-01 -3.40155959e-01 5.31363666e-01
1.18738556e+00 -1.14001453e-01 1.36536568e-01 -3.86673123e-01
-9.22847450e-01 -1.58552110e-01 -1.58749497e+00 1.76714599e+00
3.41890901e-02 3.09529632e-01 1.03756905e+00 -1.23292065e+00
7.78115451e-01 7.44273886e-02 9.68825102e-01 1.77377269e-01
-1.28278568e-01 4.12366658e-01 -4.40661103e-01 -3.17679673e-01
4.67918336e-01 -4.47748631e-01 5.23193121e-01 3.08386773e-01
8.97624046e-02 -6.77840352e-01 4.55345392e-01 2.08587348e-01
6.67374074e-01 6.30995691e-01 1.08871147e-01 -8.86046171e-01
8.07442307e-01 3.37647527e-01 4.65625882e-01 3.38701576e-01
2.76867449e-01 6.68161690e-01 5.11313498e-01 -8.20627585e-02
-9.30427611e-01 -6.92040205e-01 -6.28426373e-01 8.28303635e-01
-2.48387903e-01 -1.38252378e-01 -7.48596251e-01 -3.27371627e-01
-7.49991983e-02 5.08603036e-01 -3.17966312e-01 4.90896970e-01
-1.99784666e-01 -1.49083924e+00 7.28863478e-02 -1.34262130e-01
4.05775845e-01 -4.23180997e-01 -4.04110074e-01 3.53332251e-01
-5.19382954e-02 -1.19928265e+00 -7.55198747e-02 4.07177120e-01
-1.16713166e+00 -7.47964203e-01 -6.47708476e-01 -6.00207031e-01
9.65116143e-01 1.34872168e-01 7.68157542e-01 -9.63956565e-02
-3.10120463e-01 7.44413197e-01 -9.79670659e-02 -1.49543807e-01
-3.27771574e-01 -1.69201940e-01 -3.42484981e-01 3.88808489e-01
-2.05087557e-01 -4.80218768e-01 -1.59102157e-01 -3.75227034e-02
-1.19321620e+00 2.11348414e-01 3.87080878e-01 1.00381851e+00
9.52328920e-01 2.01061606e-01 -2.00929046e-01 -9.61041868e-01
2.33202413e-01 -1.47135377e-01 -1.21411681e+00 2.03509882e-01
-8.46364796e-01 2.02663802e-02 1.36009887e-01 9.92137492e-02
-1.11661160e+00 7.66640365e-01 7.93978721e-02 -3.21170129e-02
9.82588739e-04 8.78714323e-01 -2.35599577e-02 -8.89144421e-01
4.29802984e-01 2.10996941e-01 3.00913006e-01 -4.69248116e-01
2.89360762e-01 2.45267883e-01 2.81118065e-01 -6.03766441e-01
9.59678769e-01 8.50743473e-01 7.55682528e-01 -1.50566685e+00
-8.35940123e-01 -8.26517403e-01 -9.45713937e-01 -1.57682300e-01
1.03388703e+00 -7.97611713e-01 -3.14640850e-01 5.02864778e-01
-1.03301442e+00 -4.75907505e-01 -4.90475714e-01 5.18479586e-01
-5.73589206e-01 1.03434432e+00 -3.69526535e-01 -8.83237958e-01
-4.20866311e-02 -1.43396485e+00 1.24733913e+00 -2.12084815e-01
4.46380377e-01 -1.29432631e+00 1.72608972e-01 6.97972655e-01
1.85322329e-01 6.89382136e-01 6.02394760e-01 1.70893017e-02
-9.89049733e-01 1.42478794e-01 -3.67541999e-01 7.28831053e-01
-2.82636821e-01 2.29376718e-01 -8.49126995e-01 -4.41681325e-01
3.88249725e-01 -9.03275609e-02 9.94241714e-01 8.35070252e-01
6.84543967e-01 -2.51857013e-01 1.66097749e-02 1.10901010e+00
2.17027617e+00 -2.74079591e-01 5.62665105e-01 2.76854515e-01
1.01248717e+00 8.78796935e-01 2.88038373e-01 4.52674687e-01
3.93924378e-02 4.72874045e-01 7.10029662e-01 -5.30860960e-01
1.74596921e-01 5.55514991e-01 2.69115120e-01 7.60057688e-01
-6.06987715e-01 8.39565173e-02 -9.02003825e-01 5.74702024e-01
-1.71784592e+00 -6.56071484e-01 -9.49661911e-01 2.07535839e+00
6.62873805e-01 -4.25636441e-01 -2.87901700e-01 3.25561881e-01
4.64875579e-01 2.80394912e-01 -1.77352950e-01 -3.14264596e-01
-2.70269424e-01 3.18444371e-01 1.13469386e+00 1.14402199e+00
-1.22594166e+00 7.60275543e-01 5.51213360e+00 5.51427782e-01
-8.88865769e-01 2.10735127e-01 2.73356020e-01 3.50753039e-01
-2.58431524e-01 6.07111096e-01 -7.21780300e-01 -1.22987423e-02
4.73814040e-01 2.06108004e-01 8.90046835e-01 2.57589012e-01
4.15653616e-01 -4.59809601e-01 -5.47640920e-01 5.78402281e-01
-6.69542477e-02 -1.17146099e+00 -1.53166905e-01 5.15084863e-01
1.13434780e+00 1.81810036e-01 -1.28877744e-01 -4.61008519e-01
4.81101759e-02 -6.45344257e-01 5.75160027e-01 4.19595301e-01
4.84417170e-01 -2.19702169e-01 3.53353530e-01 1.79955319e-01
-9.81300056e-01 -5.60262725e-02 -2.23200604e-01 4.14008647e-02
3.60518932e-01 9.67560589e-01 -5.97471476e-01 7.76046872e-01
2.85145551e-01 7.38868535e-01 -3.07116181e-01 1.04297984e+00
-9.56987515e-02 7.74237454e-01 -7.12491691e-01 6.41553283e-01
6.48419261e-01 -1.39851534e+00 1.23915327e+00 1.20396364e+00
1.07230648e-01 1.81108013e-01 6.62307978e-01 1.02778089e+00
4.22764361e-01 3.06992501e-01 -2.38726661e-01 -5.55858798e-02
-4.38517720e-01 1.47298491e+00 -1.12158632e+00 4.91311355e-03
-3.90853405e-01 7.44771183e-01 -1.56559646e-01 6.78461313e-01
-1.58800736e-01 2.02578768e-01 7.63808414e-02 6.45532459e-02
5.13658881e-01 -4.60992634e-01 -6.03793681e-01 -1.05263007e+00
1.01515446e-02 -7.02821910e-01 2.30587631e-01 -8.40399802e-01
-9.85529602e-01 3.51847053e-01 2.97091335e-01 -8.68907511e-01
3.15120757e-01 -1.03387249e+00 -1.74607992e-01 9.16752934e-01
-1.70158410e+00 -1.39087212e+00 -3.19395274e-01 3.70355815e-01
1.75200805e-01 2.78308064e-01 7.30387032e-01 -4.29639630e-02
-6.49069786e-01 -3.39645356e-01 3.65481943e-01 -4.02840465e-01
2.88970638e-02 -1.54663932e+00 -4.13052142e-01 1.15024626e+00
-3.11216354e-01 3.77291948e-01 7.33692288e-01 -6.63171589e-01
-1.29861081e+00 -1.10589516e+00 5.95479608e-01 1.89466715e-01
7.14764714e-01 -8.11992213e-02 -9.23737228e-01 8.47434402e-01
3.09433699e-01 -1.42064959e-01 6.80577636e-01 -2.81560183e-01
8.04472193e-02 1.75268147e-02 -1.17868936e+00 1.25985369e-01
5.56077063e-01 -3.10051203e-01 -2.18655646e-01 9.93081748e-01
3.11688900e-01 -2.40345940e-01 -1.08108187e+00 5.49364924e-01
7.79063106e-02 -7.71241009e-01 8.80050182e-01 1.19325649e-02
9.50462520e-02 -4.98508841e-01 -2.30675861e-01 -9.60699737e-01
-4.33767676e-01 -1.05937040e+00 3.89529794e-01 8.93941104e-01
4.96363670e-01 -8.09767067e-01 5.81936300e-01 7.31395900e-01
-3.73605072e-01 -4.29924071e-01 -7.58957684e-01 -8.40416908e-01
-1.18978955e-01 -8.12648237e-02 -6.52032718e-02 1.22315979e+00
-2.35473022e-01 5.42787947e-02 -1.36203468e-01 5.29554844e-01
1.14780676e+00 3.99653204e-02 3.65342885e-01 -1.49475300e+00
-4.98445749e-01 -1.90187126e-01 -2.97706872e-01 -8.65014195e-01
1.20308086e-01 -1.15258360e+00 1.60677418e-01 -1.59403849e+00
-3.27684879e-02 -3.89113992e-01 1.66579902e-01 6.87331676e-01
2.64943868e-01 3.04595143e-01 1.81906968e-01 2.74348617e-01
-1.64968163e-01 3.38642836e-01 1.41581368e+00 -1.96428314e-01
-5.31586230e-01 -1.60179034e-01 -3.06375504e-01 8.03658724e-01
6.08312130e-01 -6.15435183e-01 -4.07069564e-01 -3.67121816e-01
5.06805599e-01 4.92559634e-02 5.38784444e-01 -7.84428000e-01
7.03091025e-02 -3.05754691e-01 -2.37959832e-01 -5.62645733e-01
3.39345723e-01 -8.72809827e-01 4.68583971e-01 3.84490132e-01
-6.20110705e-02 -5.58330297e-01 8.34624693e-02 4.49549019e-01
-8.45391452e-02 -7.50329912e-01 8.52289319e-01 -2.90445775e-01
-5.63609421e-01 1.96928203e-01 -4.55771238e-01 -3.47532332e-01
8.02899599e-01 -4.36123341e-01 8.07358995e-02 1.39866278e-01
-9.97611403e-01 -4.23098840e-02 5.04538178e-01 -6.00461245e-01
3.23109269e-01 -6.52408600e-01 -9.32443261e-01 1.82720140e-01
-3.77453566e-01 2.97870725e-01 -2.28873529e-02 1.55016863e+00
-9.98422265e-01 2.23678678e-01 -8.77919123e-02 -8.97829235e-01
-1.22029090e+00 9.63107646e-02 6.49772763e-01 -3.54260713e-01
-5.66358328e-01 8.25867355e-01 9.92349014e-02 -5.31310856e-01
-3.43956411e-01 -3.09529305e-01 2.06939876e-02 8.54206234e-02
-1.62769452e-01 8.51415932e-01 8.86619687e-02 -1.00278997e+00
-1.24962918e-01 7.37464368e-01 5.48004568e-01 -3.43538523e-01
1.67912400e+00 -3.65835041e-01 -1.02487338e+00 1.01054229e-01
1.10438931e+00 2.19044685e-01 -1.52124536e+00 2.52063368e-02
4.68109362e-02 -2.11151868e-01 9.67314005e-01 -6.71117425e-01
-1.23268497e+00 5.35947740e-01 4.28800911e-01 3.54864389e-01
1.20462358e+00 -4.68035877e-01 3.06569606e-01 4.20822531e-01
2.17535987e-01 -1.19160223e+00 -6.00884914e-01 4.91905868e-01
1.07677805e+00 -1.03253007e+00 8.53121698e-01 -1.13495028e+00
-2.40773812e-01 1.15899718e+00 -1.19358785e-01 -2.57296145e-01
8.82069051e-01 1.09644696e-01 -6.19054325e-02 -3.47559452e-01
8.29291418e-02 -6.61875725e-01 4.31879252e-01 3.70722473e-01
3.20975006e-01 1.24192797e-01 -8.09220672e-01 -2.02016264e-01
1.38666227e-01 -2.11364046e-01 5.72497964e-01 8.96255791e-01
-2.76333958e-01 -1.24175489e+00 -6.83393061e-01 2.29132161e-01
-2.76035756e-01 -2.12090567e-01 -3.04732144e-01 5.60778856e-01
4.22661871e-01 9.38657820e-01 -4.04377192e-01 1.77418172e-01
-1.19841069e-01 1.93188280e-01 6.16600156e-01 -7.27486193e-01
-5.38620770e-01 8.04565609e-01 -5.89671619e-02 -3.20306122e-01
-1.09494781e+00 -1.18399858e+00 -1.18500412e+00 3.04876983e-01
-7.84522533e-01 6.36059791e-02 1.00061750e+00 1.21310771e+00
4.53895107e-02 2.07564235e-01 6.66749835e-01 -1.27838552e+00
-5.23615718e-01 -6.76185608e-01 -1.08354342e+00 6.12701774e-02
3.22085053e-01 -4.80875671e-01 -7.33680964e-01 4.42254096e-01] | [10.067710876464844, -2.0222268104553223] |
374ccc26-020a-4321-9889-e489fb706c8e | memobert-pre-training-model-with-prompt-based | 2111.00865 | null | https://arxiv.org/abs/2111.00865v1 | https://arxiv.org/pdf/2111.00865v1.pdf | MEmoBERT: Pre-training Model with Prompt-based Learning for Multimodal Emotion Recognition | Multimodal emotion recognition study is hindered by the lack of labelled corpora in terms of scale and diversity, due to the high annotation cost and label ambiguity. In this paper, we propose a pre-training model \textbf{MEmoBERT} for multimodal emotion recognition, which learns multimodal joint representations through self-supervised learning from large-scale unlabeled video data that come in sheer volume. Furthermore, unlike the conventional "pre-train, finetune" paradigm, we propose a prompt-based method that reformulates the downstream emotion classification task as a masked text prediction one, bringing the downstream task closer to the pre-training. Extensive experiments on two benchmark datasets, IEMOCAP and MSP-IMPROV, show that our proposed MEmoBERT significantly enhances emotion recognition performance. | ['Haizhou Li', 'Xinchao Wang', 'Qin Jin', 'Ruichen Li', 'Jinming Zhao'] | 2021-10-27 | null | null | null | null | ['multimodal-emotion-recognition', 'multimodal-emotion-recognition'] | ['computer-vision', 'speech'] | [ 3.28818232e-01 -2.13781908e-01 -7.23273605e-02 -7.75443375e-01
-9.23268199e-01 -6.84782088e-01 3.90717000e-01 -1.23801529e-01
-6.74375057e-01 6.19611621e-01 3.15453887e-01 4.05688188e-04
1.99903652e-01 -7.55432770e-02 -5.24895668e-01 -7.00152636e-01
1.79592356e-01 1.50017247e-01 -3.90625656e-01 -7.75896013e-02
3.19229923e-02 -7.61540011e-02 -1.55308509e+00 8.19195390e-01
7.97118723e-01 1.41660058e+00 -2.04459459e-01 7.51352012e-01
-2.26780251e-01 1.06140769e+00 -2.56667018e-01 -7.93379545e-01
-9.55092758e-02 -4.70797777e-01 -9.06931698e-01 4.09938842e-01
2.80666530e-01 -2.72127807e-01 -4.04742599e-01 8.06868732e-01
5.83951890e-01 3.21332991e-01 7.65508175e-01 -1.62206066e+00
-4.92502093e-01 1.28919140e-01 -5.84447861e-01 -1.19718432e-01
4.05250102e-01 -5.23388125e-02 9.07725334e-01 -1.14833212e+00
5.69801331e-01 1.00693679e+00 5.54519236e-01 8.15772772e-01
-1.13551033e+00 -5.30995846e-01 3.71435434e-01 3.77668113e-01
-1.26638699e+00 -6.88658714e-01 7.94903398e-01 -4.06869769e-01
1.00262535e+00 3.03031832e-01 1.68466240e-01 1.73483026e+00
-1.71138540e-01 1.10863400e+00 1.06453514e+00 -3.89779598e-01
2.64240861e-01 2.41059154e-01 1.35384813e-01 7.40861177e-01
-5.76410115e-01 -3.49196464e-01 -8.21166039e-01 -2.09899515e-01
4.06747133e-01 5.99539690e-02 -3.45409214e-01 -3.63313943e-01
-1.11096036e+00 7.25811839e-01 -9.36959982e-02 1.49085313e-01
-2.92246729e-01 -3.41051184e-02 9.98266518e-01 6.33691669e-01
7.12058067e-01 1.74394175e-02 -7.00617373e-01 -5.49848795e-01
-6.92210317e-01 -3.44506234e-01 8.61368954e-01 8.01358461e-01
8.29181254e-01 -2.83432864e-02 -1.37345135e-01 1.02334869e+00
3.59919280e-01 1.96141154e-01 7.59094119e-01 -6.96668565e-01
6.42819405e-01 6.13211989e-01 8.64647254e-02 -8.64818215e-01
-3.84469241e-01 1.48536190e-01 -9.03569341e-01 -1.26730487e-01
2.44547904e-01 -5.67813993e-01 -7.53487945e-01 1.60290956e+00
2.38094687e-01 2.23852143e-01 4.90404546e-01 1.12572503e+00
1.11407173e+00 8.67971063e-01 4.72690850e-01 -2.03842700e-01
1.42998052e+00 -1.30752695e+00 -8.78627717e-01 -1.61817998e-01
8.56290102e-01 -5.61589420e-01 1.09574080e+00 7.62769520e-01
-5.79745889e-01 -2.94768304e-01 -7.20687270e-01 -7.00879619e-02
-5.39199233e-01 5.11165023e-01 6.09204412e-01 7.27555633e-01
-8.45177710e-01 9.04600918e-02 -4.60848391e-01 -4.42008048e-01
3.79653245e-01 2.64519811e-01 -8.06167126e-01 -7.67339021e-02
-9.89084899e-01 6.69331908e-01 4.51124877e-01 5.90945899e-01
-7.97828972e-01 -1.68469131e-01 -1.04685342e+00 3.12794745e-02
6.75347626e-01 -2.66273290e-01 1.19297349e+00 -1.70353436e+00
-1.92278337e+00 9.34203327e-01 -3.55429858e-01 1.23047158e-01
5.56105018e-01 -3.43159050e-01 -5.92760801e-01 3.55304748e-01
-3.23389322e-01 8.88799191e-01 1.09972167e+00 -1.33283615e+00
-4.33478832e-01 -4.30240691e-01 -2.66069144e-01 4.61464942e-01
-7.76490033e-01 2.57391840e-01 -7.55642712e-01 -4.39138621e-01
-2.30398118e-01 -8.76532316e-01 -3.22215974e-01 -4.54138577e-01
-4.23652887e-01 -2.60317355e-01 8.02661896e-01 -5.36753595e-01
1.06399333e+00 -2.36005092e+00 2.93347806e-01 7.64986426e-02
1.25208452e-01 -7.15927184e-02 -5.50303936e-01 5.09857476e-01
-3.28121722e-01 3.32306065e-02 -5.56772128e-02 -6.78915799e-01
3.50963444e-01 2.08411872e-01 -3.23408782e-01 2.06974298e-01
4.28844720e-01 9.82322574e-01 -7.07782805e-01 -6.74993694e-01
-3.74335237e-02 4.11329150e-01 -2.65822262e-01 5.58500767e-01
-1.91410229e-01 4.68476176e-01 -3.53656828e-01 8.92835855e-01
5.37049294e-01 -1.38377219e-01 3.05095792e-01 -3.59628797e-01
3.52931432e-02 -3.03962022e-01 -1.14396846e+00 1.93538630e+00
-2.77591079e-01 5.32038510e-01 1.66418433e-01 -1.09954524e+00
8.27462077e-01 6.86605930e-01 5.75119793e-01 -6.88494205e-01
4.30330902e-01 -1.37363270e-01 -5.92316985e-01 -1.12522721e+00
8.04473460e-01 -1.23085849e-01 -5.84775686e-01 3.59631747e-01
4.55303401e-01 3.60435426e-01 1.04668111e-01 1.86485067e-01
9.69164610e-01 3.27255547e-01 8.19899049e-03 3.35348010e-01
4.35610622e-01 -1.55058429e-01 6.77310944e-01 5.03877163e-01
-5.43361902e-01 5.97582459e-01 7.96773314e-01 -2.71840930e-01
-7.04216719e-01 -7.22526312e-01 9.03288573e-02 1.74759936e+00
-7.16845542e-02 -4.98252541e-01 -6.85324013e-01 -1.14365447e+00
-3.82731497e-01 9.06725228e-02 -7.33976781e-01 -1.51322424e-01
1.55716250e-02 -8.60840321e-01 7.53748238e-01 5.12027979e-01
3.81721944e-01 -1.07287788e+00 -4.42331105e-01 8.52286741e-02
-5.62586784e-01 -1.31158710e+00 -4.27695394e-01 4.25031632e-01
-6.25839531e-01 -7.47340620e-01 -6.39984906e-01 -8.35844219e-01
6.90227687e-01 -1.43510029e-01 9.57060039e-01 -2.23082677e-01
-2.62584209e-01 8.10626447e-01 -8.44384253e-01 -2.34227911e-01
6.59485385e-02 -4.02281247e-02 -1.15730919e-01 6.82229280e-01
6.21081889e-01 -1.50737509e-01 -4.01273012e-01 2.49699488e-01
-1.12886941e+00 9.60552320e-02 6.32020295e-01 1.17497051e+00
4.43067610e-01 -1.26612887e-01 7.62282312e-01 -6.18756413e-01
7.12626755e-01 -6.34064913e-01 -1.45316064e-01 4.75035220e-01
-3.04890782e-01 -2.24260032e-01 3.66743684e-01 -7.89810479e-01
-1.27512038e+00 4.32009608e-01 4.32701558e-02 -6.08419895e-01
-6.95902944e-01 7.34133601e-01 -3.71913850e-01 6.71794936e-02
2.02630982e-01 3.17040801e-01 1.24867164e-01 -3.14648032e-01
4.69448537e-01 8.51035953e-01 6.11673832e-01 -5.59450507e-01
2.43671924e-01 2.62952507e-01 -5.15816152e-01 -6.68079615e-01
-6.46995366e-01 -4.44042921e-01 -6.93393707e-01 -5.57526112e-01
1.04871106e+00 -1.15213823e+00 -1.09444237e+00 3.75592262e-01
-1.06296062e+00 -3.51361513e-01 7.42011070e-02 5.04651427e-01
-5.97046316e-01 4.86801356e-01 -7.90091217e-01 -1.03234494e+00
-2.47135058e-01 -1.08510089e+00 1.15541661e+00 2.38489911e-01
-9.62045342e-02 -9.11857486e-01 1.01404980e-01 6.24310970e-01
1.11101650e-01 1.45011842e-01 6.91621542e-01 -9.02382314e-01
-1.51844978e-01 -2.54710943e-01 -3.78186643e-01 4.90956068e-01
-2.85850227e-01 -8.05150047e-02 -1.15738308e+00 -5.63460067e-02
-1.91805288e-01 -1.28141284e+00 8.36812377e-01 -1.44193083e-01
1.20178771e+00 -2.30291381e-01 -6.59030229e-02 3.75557691e-01
1.33272755e+00 3.65036577e-02 6.97102726e-01 1.45026714e-01
7.71216333e-01 1.01768625e+00 7.05555320e-01 7.72898316e-01
6.38633668e-01 3.45151693e-01 2.81321824e-01 -1.63139969e-01
6.00294590e-01 5.01643084e-02 5.75578153e-01 9.08508480e-01
1.67046458e-01 -4.45182413e-01 -7.99341083e-01 2.82767028e-01
-2.42939472e+00 -8.44190836e-01 1.88614186e-02 1.71781814e+00
6.75665617e-01 -5.44429660e-01 1.36731610e-01 -1.61071181e-01
5.87614954e-01 4.43913937e-02 -4.11406994e-01 -5.09843349e-01
-1.89617231e-01 -2.54548371e-01 -2.41659880e-02 1.81653991e-01
-1.40293181e+00 1.12603307e+00 6.16418552e+00 8.89457166e-01
-1.16442263e+00 2.25334063e-01 9.32283223e-01 -1.95873678e-01
8.97271857e-02 -3.15627784e-01 -4.69341248e-01 4.08291519e-01
9.80558753e-01 3.28030497e-01 3.22980940e-01 8.15472662e-01
2.48703659e-02 -1.77818328e-01 -1.26333797e+00 1.45840001e+00
5.51660120e-01 -7.76293635e-01 -9.08315629e-02 -2.10736394e-01
4.41245198e-01 -5.81713505e-02 7.13302344e-02 8.40676069e-01
-7.24436641e-02 -1.05398643e+00 5.64411521e-01 6.32743120e-01
6.86732054e-01 -8.25734019e-01 7.81713009e-01 1.97926879e-01
-9.66950417e-01 -1.22209951e-01 -2.07339674e-01 1.22823730e-01
7.20322877e-02 2.20562980e-01 -5.49847484e-01 6.31507933e-01
8.86285067e-01 6.04984283e-01 -5.69850564e-01 6.60684884e-01
3.61223333e-02 5.42186499e-01 7.92608112e-02 -1.51394188e-01
1.75788105e-01 -2.94211119e-01 5.31913377e-02 1.62445343e+00
3.80588584e-02 1.31834850e-01 3.40909421e-01 5.00195742e-01
-4.01111811e-01 5.38274884e-01 -5.29565811e-01 -4.35440540e-01
9.25606936e-02 1.71195018e+00 -4.82852668e-01 -3.59046489e-01
-5.70406497e-01 1.54906964e+00 5.35051584e-01 7.05482483e-01
-8.58694255e-01 -3.79704386e-01 3.68428707e-01 -9.18334782e-01
3.54644656e-01 -1.45014925e-02 -4.45875004e-02 -1.59266067e+00
1.33929926e-03 -1.06198990e+00 5.66980422e-01 -8.39405239e-01
-1.70872474e+00 7.95122087e-01 -3.93149555e-01 -1.14790559e+00
-2.38089353e-01 -8.00536335e-01 -4.83126253e-01 3.73093039e-01
-1.39381528e+00 -1.30940711e+00 -3.26608717e-01 8.84833455e-01
5.88467836e-01 -1.77395478e-01 1.10107040e+00 4.98506576e-01
-1.01179540e+00 8.23454261e-01 1.32416725e-01 3.73604685e-01
1.07467651e+00 -1.10150290e+00 -6.06842637e-01 3.35780114e-01
6.07765540e-02 6.94778413e-02 4.66038793e-01 -3.71541679e-01
-1.76245725e+00 -8.73026490e-01 7.06391335e-01 -5.86636364e-01
8.31845820e-01 -7.13573694e-01 -9.66005921e-01 6.96415484e-01
5.21012127e-01 -1.00502847e-02 1.34346735e+00 3.85670245e-01
-6.11593246e-01 -7.03985170e-02 -9.25563037e-01 5.26078701e-01
6.16525233e-01 -8.03875566e-01 -5.15775919e-01 7.72867650e-02
5.41137397e-01 -6.74625114e-02 -1.08546102e+00 5.23783863e-01
7.14048803e-01 -5.63553035e-01 5.66515565e-01 -9.06928897e-01
6.40555024e-01 7.80942887e-02 -3.02792966e-01 -1.21434319e+00
1.28835902e-01 -5.56840420e-01 3.34977210e-02 1.75686955e+00
4.76166219e-01 -2.35917822e-01 8.89581561e-01 9.33413923e-01
5.86485155e-02 -5.43665469e-01 -1.01210415e+00 -4.47852790e-01
-2.57738978e-01 -5.75034440e-01 1.44086480e-01 1.36133313e+00
6.00493133e-01 6.02374613e-01 -8.39999437e-01 -3.34844459e-03
2.99765706e-01 6.09707758e-02 8.62935841e-01 -9.27056074e-01
-2.63980746e-01 -3.23104799e-01 -3.13281029e-01 -8.74507427e-01
5.53918839e-01 -6.59604430e-01 4.08216149e-01 -9.60372925e-01
5.26150763e-01 7.85314068e-02 -5.33117771e-01 7.06756175e-01
-1.99811131e-01 4.84946370e-01 3.32722738e-02 4.67518605e-02
-1.50134671e+00 1.00539815e+00 6.57920420e-01 -1.28887057e-01
-1.52141452e-01 -6.42926276e-01 -4.71465141e-01 6.08475566e-01
4.90430415e-01 -3.28813285e-01 -2.75603086e-01 -3.53589088e-01
2.89276838e-01 2.00950265e-01 1.98321864e-01 -5.36835790e-01
1.60331056e-01 7.78617989e-03 3.86693567e-01 -4.40468460e-01
6.04572296e-01 -1.00795984e+00 -3.61707956e-01 -3.01548660e-01
-6.13348722e-01 1.04329638e-01 3.71655852e-01 6.63221836e-01
-4.50141370e-01 -2.21775681e-01 1.81558445e-01 2.21762449e-01
-1.07859719e+00 3.18432093e-01 -7.93932617e-01 -1.92391366e-01
9.23039019e-01 -9.18770358e-02 -4.46754605e-01 -4.83612776e-01
-9.74697649e-01 5.30650258e-01 1.99106380e-01 6.11676753e-01
7.74635732e-01 -1.35691476e+00 -5.07422388e-01 -1.11838192e-01
7.00040519e-01 -4.49217767e-01 7.20310390e-01 1.06296432e+00
1.43103540e-01 6.60692435e-03 -2.47962132e-01 -5.34583986e-01
-1.48362768e+00 6.49265170e-01 1.04426838e-01 -1.67533144e-01
-1.67279378e-01 7.28910327e-01 1.50099948e-01 -7.84232855e-01
6.63877547e-01 1.45964831e-01 -2.64332235e-01 3.39622527e-01
4.88520533e-01 3.51771228e-02 -1.19896419e-01 -7.51165450e-01
-2.78309762e-01 1.42775372e-01 -1.39168277e-01 -3.30843121e-01
1.28759706e+00 -5.69367886e-01 -1.29811093e-01 7.27949619e-01
1.44647026e+00 -3.69046897e-01 -1.28313434e+00 -2.20007807e-01
2.44807974e-01 -2.34069094e-01 -7.33961388e-02 -9.37384069e-01
-8.17085207e-01 8.68352890e-01 7.80136883e-01 1.99053846e-02
1.31534195e+00 -6.79355860e-02 5.17589092e-01 8.34268928e-01
-1.24582406e-02 -1.62664974e+00 6.52488589e-01 6.07073367e-01
7.33151793e-01 -1.67888117e+00 -5.77398717e-01 -1.72749668e-01
-1.36085021e+00 1.09193134e+00 1.01020503e+00 3.00022393e-01
3.57206076e-01 1.03632934e-01 4.35657173e-01 -2.21314743e-01
-1.09899712e+00 -1.43702447e-01 2.08299413e-01 3.21278185e-01
5.44445574e-01 -2.21429661e-01 -9.09338146e-02 1.08039176e+00
3.57965857e-01 1.82472095e-01 2.08902344e-01 1.15296912e+00
-1.64607376e-01 -9.74282205e-01 -4.33037639e-01 2.44595766e-01
-4.11801666e-01 -8.49134624e-02 -6.63630307e-01 4.41919982e-01
1.73826925e-02 1.15348577e+00 7.58112743e-02 -7.73441494e-01
1.41866595e-01 7.27369487e-01 3.44069809e-01 -2.07916304e-01
-5.82037151e-01 4.46367502e-01 4.37419295e-01 -5.62501967e-01
-4.64067250e-01 -3.96081388e-01 -1.18183303e+00 1.11522496e-01
-2.24816874e-01 3.07620138e-01 7.58145690e-01 1.05094790e+00
6.09332919e-01 3.56306642e-01 7.87482619e-01 -9.70792055e-01
-3.31369370e-01 -9.84196246e-01 -5.51177680e-01 7.28552341e-01
8.56241584e-02 -4.07256991e-01 -1.02158442e-01 3.10319692e-01] | [13.22551441192627, 5.182877063751221] |
d3a544a2-a050-4e5b-9116-833dc464899f | embarrassingly-simple-mixup-for-time-series | 2304.04271 | null | https://arxiv.org/abs/2304.04271v1 | https://arxiv.org/pdf/2304.04271v1.pdf | Embarrassingly Simple MixUp for Time-series | Labeling time series data is an expensive task because of domain expertise and dynamic nature of the data. Hence, we often have to deal with limited labeled data settings. Data augmentation techniques have been successfully deployed in domains like computer vision to exploit the use of existing labeled data. We adapt one of the most commonly used technique called MixUp, in the time series domain. Our proposed, MixUp++ and LatentMixUp++, use simple modifications to perform interpolation in raw time series and classification model's latent space, respectively. We also extend these methods with semi-supervised learning to exploit unlabeled data. We observe significant improvements of 1\% - 15\% on time series classification on two public datasets, for both low labeled data as well as high labeled data regimes, with LatentMixUp++. | ['Jaideep Srivastava', 'Karan Aggarwal'] | 2023-04-09 | null | null | null | null | ['time-series-classification'] | ['time-series'] | [ 1.99190512e-01 -1.66785419e-01 -5.92985034e-01 -3.67782176e-01
-6.54514432e-01 -7.74622858e-01 7.62488484e-01 2.51826108e-01
-4.19704705e-01 6.13876998e-01 1.69983685e-01 -3.94387960e-01
1.52262568e-01 -5.65242112e-01 -2.81385899e-01 -6.51063800e-01
-2.45861188e-01 3.44265640e-01 -2.48902157e-01 -8.99141952e-02
-1.37545004e-01 1.63830578e-01 -1.27297187e+00 6.48296997e-02
8.05213273e-01 1.01361811e+00 -2.92687058e-01 3.13434273e-01
-2.75732934e-01 7.95176327e-01 -2.70031542e-01 5.00111580e-02
6.20148182e-01 -1.62733987e-01 -3.95705581e-01 2.09538668e-01
-1.85403407e-01 -3.26044532e-03 -3.84624153e-01 6.32772148e-01
1.67201281e-01 1.89635947e-01 5.83771288e-01 -1.67329633e+00
-3.77261996e-01 7.16806412e-01 -7.84515202e-01 4.17279094e-01
-6.30327985e-02 6.44201338e-02 5.56484997e-01 -5.62193573e-01
4.83938366e-01 7.98333704e-01 9.41812396e-01 2.60525584e-01
-1.71545744e+00 -8.81555498e-01 2.10557848e-01 -9.19814967e-03
-1.15502489e+00 -2.51941234e-01 1.18342626e+00 -6.84134483e-01
8.04058373e-01 1.72081664e-01 3.29135656e-01 1.23761237e+00
-3.98094833e-01 7.45771527e-01 1.65199685e+00 -4.94111270e-01
3.69828016e-01 7.26402327e-02 5.67554593e-01 1.46667033e-01
-4.75612640e-01 1.93196177e-01 -2.41556168e-01 -4.57006872e-01
7.14194894e-01 4.87288326e-01 1.78146780e-01 -5.79779260e-02
-1.42553329e+00 7.68945456e-01 1.38174713e-01 3.62621307e-01
-5.27109265e-01 5.31660765e-02 7.51235068e-01 6.64145052e-01
1.13687456e+00 4.35740501e-01 -9.27510619e-01 -3.79755974e-01
-9.68366683e-01 1.17225848e-01 5.00936806e-01 8.78131032e-01
6.74789548e-01 2.66354501e-01 -1.53107405e-01 8.45297456e-01
-1.97805688e-01 3.07557076e-01 1.02478683e+00 -8.44361305e-01
5.08941352e-01 5.88480473e-01 1.27300516e-01 -6.52644277e-01
-4.70383704e-01 -5.31454206e-01 -9.79137540e-01 -2.31496319e-01
4.16386127e-01 -3.84814173e-01 -1.23711622e+00 2.07608747e+00
4.23292398e-01 7.63393342e-01 1.90214410e-01 2.79547513e-01
2.22758174e-01 8.56030881e-01 3.50501090e-01 -8.45028222e-01
1.04461694e+00 -9.65741038e-01 -9.87264335e-01 -1.41469330e-01
9.34952438e-01 -6.37861669e-01 1.22817087e+00 2.77568489e-01
-5.57845354e-01 -6.94077194e-01 -6.68997765e-01 2.31779650e-01
-6.88076675e-01 1.88893870e-01 9.24368858e-01 4.69508559e-01
-5.34063458e-01 8.31022978e-01 -1.15248454e+00 -2.73961604e-01
2.91473031e-01 2.28256673e-01 -3.30932975e-01 9.35386419e-02
-1.10544193e+00 4.98358011e-01 6.68033004e-01 -3.93125027e-01
-7.10174322e-01 -9.23531890e-01 -7.10079312e-01 -3.80840421e-01
3.33568603e-01 -6.29342627e-03 1.29422069e+00 -6.97957635e-01
-1.05337358e+00 7.50468910e-01 -6.39549568e-02 -8.07479322e-01
5.89279056e-01 -4.06780988e-01 -7.51256883e-01 -2.56583482e-01
1.82684749e-01 4.96904790e-01 9.39577639e-01 -8.45870614e-01
-3.20856005e-01 -2.66466737e-01 -3.85475099e-01 -3.11340332e-01
-7.80339658e-01 -9.32113230e-02 -1.09153070e-01 -1.30807936e+00
2.27954447e-01 -1.19495070e+00 -5.00456810e-01 -2.57761300e-01
-2.92891771e-01 -3.46394420e-01 1.39257753e+00 -7.75895774e-01
1.42074454e+00 -2.28547215e+00 -2.64061540e-01 1.95980951e-01
1.78225443e-01 1.48746371e-01 -7.39536360e-02 6.86799467e-01
-6.88362598e-01 4.11655791e-02 -3.41510952e-01 -6.37107313e-01
-6.85089529e-02 4.76089120e-01 -8.76920819e-01 3.54064852e-01
8.80029947e-02 6.78972363e-01 -9.80462611e-01 -1.79542094e-01
3.68567944e-01 4.50873971e-02 -3.17074150e-01 1.17922761e-01
-4.36113417e-01 9.80159640e-01 -3.05406839e-01 5.01214564e-01
3.88835549e-01 -5.94362020e-01 2.74851695e-02 -3.57577763e-02
-2.86988318e-01 7.10888728e-02 -9.08635020e-01 1.82361031e+00
-5.10075271e-01 5.52499652e-01 -5.95720053e-01 -1.19103515e+00
9.15331185e-01 7.64317811e-01 1.14507556e+00 -4.56775576e-01
6.57996461e-02 3.42120647e-01 -1.44403905e-01 -5.69038630e-01
1.81343347e-01 -2.22920701e-01 -2.07585916e-01 6.98408186e-01
-6.48862571e-02 1.13597780e-01 2.45066151e-01 3.01817041e-02
9.71514463e-01 4.36683327e-01 3.28959584e-01 -9.54729393e-02
7.65269324e-02 2.96846628e-01 6.37790203e-01 4.02558863e-01
-2.30951935e-01 3.66413206e-01 4.34313625e-01 -5.01952648e-01
-1.52251768e+00 -6.14209056e-01 -2.11927503e-01 1.24869430e+00
-6.00370347e-01 -4.09179419e-01 -3.23060304e-01 -7.43977547e-01
-7.50315413e-02 6.15166843e-01 -7.16848910e-01 -2.29529932e-01
-6.20330691e-01 -1.12602091e+00 4.70018536e-01 8.62796605e-01
4.49060857e-01 -8.74843895e-01 -3.12672287e-01 4.15553957e-01
-2.89740920e-01 -1.38311529e+00 -3.95517707e-01 5.17654061e-01
-1.15673053e+00 -6.68564558e-01 -6.53695762e-01 -4.30188805e-01
6.03567362e-01 9.67540368e-02 9.16418016e-01 -3.96856219e-01
-5.57724573e-02 1.35312125e-01 -7.09078550e-01 -6.55632317e-01
-3.94728214e-01 2.15184048e-01 4.06199604e-01 2.13399276e-01
5.31830907e-01 -9.95653689e-01 -3.63195837e-01 4.46116567e-01
-9.27528679e-01 2.40450040e-01 1.72239646e-01 8.18683803e-01
5.56412876e-01 1.10385917e-01 1.01678896e+00 -9.32098031e-01
2.99074769e-01 -8.62272024e-01 -5.61625481e-01 -1.99372936e-02
-9.44415927e-01 5.16209379e-02 6.67096913e-01 -1.10495675e+00
-5.60084939e-01 1.80666849e-01 4.56452034e-02 -9.71661448e-01
-2.25419715e-01 7.92024016e-01 4.26875234e-01 2.96645850e-01
7.95147955e-01 1.19118549e-01 -2.67243087e-02 -8.23182464e-01
3.30064267e-01 7.29154170e-01 6.41720414e-01 -5.75364351e-01
1.06579638e+00 7.94907749e-01 -1.45262510e-01 -4.40723479e-01
-1.08179581e+00 -7.78413713e-01 -7.69416928e-01 9.60591212e-02
6.30141199e-01 -1.09886956e+00 -7.02940077e-02 4.93258327e-01
-6.01288676e-01 -5.84870636e-01 -8.17281783e-01 7.47575819e-01
-5.70431232e-01 2.69387782e-01 -4.32476699e-01 -9.29947853e-01
-2.10589439e-01 -6.13153338e-01 9.12691712e-01 -1.08227730e-01
-4.12872791e-01 -1.22782624e+00 3.09420288e-01 1.06546775e-01
4.06682789e-01 8.70904207e-01 9.04480696e-01 -1.06976557e+00
8.16524923e-02 -5.40703475e-01 -1.97540149e-02 3.51828039e-01
5.40852904e-01 -2.11952239e-01 -1.29149592e+00 -3.92827481e-01
1.11603804e-01 -4.30168241e-01 6.45478725e-01 2.98052311e-01
1.33242667e+00 -2.84763396e-01 -3.05817395e-01 5.21174729e-01
1.21605766e+00 3.34540337e-01 2.58024275e-01 1.28352821e-01
6.62041366e-01 5.90388179e-01 8.57645094e-01 7.33185053e-01
1.49918810e-01 5.32197893e-01 8.42453688e-02 -3.36195797e-01
3.08559895e-01 -3.52993697e-01 1.69864953e-01 1.14993489e+00
-6.18506409e-02 1.45860702e-01 -1.18703139e+00 7.58874118e-01
-1.99893153e+00 -8.26924741e-01 -1.87098563e-01 2.26449442e+00
1.07090199e+00 1.97561875e-01 4.41651314e-01 8.14568520e-01
4.74646300e-01 7.64695704e-02 -7.96093702e-01 1.83080927e-01
-8.23727809e-03 2.32262686e-01 6.23665333e-01 -1.02742895e-01
-1.55188286e+00 7.06119001e-01 6.64867735e+00 7.38973856e-01
-1.39230883e+00 3.52764487e-01 6.87253177e-01 -5.34407794e-02
6.48585036e-02 1.38301209e-01 -3.15082818e-01 7.21548915e-01
1.34000158e+00 -1.49311393e-01 1.84976324e-01 7.30800807e-01
5.36643326e-01 3.81760240e-01 -1.21070719e+00 1.14555466e+00
-4.37700510e-01 -1.13495612e+00 -5.64721167e-01 8.67562070e-02
9.57771420e-01 2.27429479e-01 2.59469241e-01 7.12742805e-01
3.20943177e-01 -7.76237786e-01 5.26780903e-01 2.25008667e-01
9.95680451e-01 -3.85950774e-01 5.00616431e-01 6.30877018e-01
-1.31917000e+00 -1.83174282e-01 1.88308761e-01 -3.00347596e-01
2.72077829e-01 7.11747825e-01 -7.35550463e-01 4.61642742e-01
5.70028603e-01 1.13420331e+00 -4.91111487e-01 7.14838326e-01
8.23295414e-02 1.03392994e+00 -6.30155802e-01 6.50350273e-01
2.00040102e-01 -1.70312583e-01 3.02362651e-01 7.05883205e-01
3.34458411e-01 1.02525875e-01 6.81606770e-01 4.71083015e-01
4.94437814e-02 2.50176229e-02 -6.15006208e-01 -5.50937772e-01
5.22967517e-01 1.16599512e+00 -6.38575256e-01 -7.18678594e-01
-4.39207882e-01 6.58029854e-01 -1.02133073e-01 4.27060753e-01
-9.58724260e-01 -4.81329262e-02 3.32298279e-01 1.94515064e-02
-2.66708523e-01 -5.56428194e-01 -3.33100706e-01 -1.44103503e+00
-2.16138922e-02 -7.94602156e-01 7.35283256e-01 -5.96582472e-01
-1.43950081e+00 4.78093594e-01 3.26771349e-01 -1.89101982e+00
-5.95481157e-01 -2.03734517e-01 -4.03648019e-01 8.13168705e-01
-1.46802485e+00 -1.39206839e+00 -2.58223146e-01 7.05012143e-01
7.38844991e-01 -1.36165112e-01 8.19004536e-01 7.04418302e-01
-7.25913882e-01 2.71824092e-01 3.32922071e-01 2.07411468e-01
7.09956408e-01 -1.25623131e+00 5.63180864e-01 8.47552598e-01
1.64584711e-01 4.57500547e-01 8.16780090e-01 -5.78973353e-01
-1.03992784e+00 -1.40399671e+00 7.04223514e-01 -4.14918542e-01
9.58369374e-01 -3.58796000e-01 -1.15952528e+00 1.01259744e+00
-2.12035142e-02 3.00539434e-01 1.07705295e+00 2.10712925e-01
-5.57011664e-01 -6.08292557e-02 -1.18420708e+00 4.17077959e-01
8.74362826e-01 -7.44358778e-01 -5.89833796e-01 6.43659890e-01
1.01336336e+00 -2.17529580e-01 -1.22594130e+00 5.28171837e-01
1.94752738e-01 -2.66848624e-01 9.28532481e-01 -8.36194456e-01
5.73436804e-02 -1.20265521e-01 -2.04655439e-01 -1.35766232e+00
-2.54011936e-02 -9.28680956e-01 -4.13303107e-01 1.45806289e+00
2.08696410e-01 -7.90994048e-01 8.57030928e-01 6.86847210e-01
-7.32202902e-02 -2.11197183e-01 -7.70367265e-01 -1.04192162e+00
-7.35248104e-02 -6.62741244e-01 6.99697614e-01 1.72587049e+00
4.48458306e-02 7.62611404e-02 -6.63932979e-01 1.19151222e-02
5.89624584e-01 2.12904692e-01 7.75160253e-01 -1.38957942e+00
-1.84701845e-01 -8.39489847e-02 -1.23381078e-01 -5.97595036e-01
1.09724231e-01 -8.37775826e-01 -4.66328144e-01 -8.99716258e-01
-2.88628429e-01 -8.67962420e-01 -6.59278512e-01 1.01713622e+00
-4.77678217e-02 4.64964896e-01 -3.87292542e-02 8.46690178e-01
-2.71541309e-02 5.17843127e-01 7.11131215e-01 2.53758840e-02
-6.18875563e-01 -1.77747998e-02 -1.90744832e-01 7.83593833e-01
9.49355721e-01 -5.53586006e-01 -6.92693174e-01 -2.05155954e-01
-2.36735847e-02 1.62470415e-01 2.43025899e-01 -1.12248373e+00
-1.98549688e-01 -3.91335189e-01 1.17868550e-01 -7.64297485e-01
3.31127107e-01 -1.13016617e+00 3.29606503e-01 4.69043076e-01
-6.05076551e-01 3.25002313e-01 2.22250223e-01 6.14378154e-01
-4.10782427e-01 2.64671028e-01 7.72101521e-01 7.69736841e-02
-8.23724508e-01 5.58739662e-01 -1.12232380e-02 -8.16627815e-02
1.27859151e+00 3.38503458e-02 -2.04478174e-01 -2.88752705e-01
-1.15195560e+00 4.07463014e-01 1.07905388e-01 5.61066091e-01
8.11908245e-02 -1.66391027e+00 -4.66860414e-01 2.69701928e-01
4.22138482e-01 -1.99496269e-01 1.54968709e-01 1.10597718e+00
-3.92857268e-02 2.93162972e-01 -1.75241709e-01 -7.15446949e-01
-8.25099826e-01 1.15345538e+00 -1.20014690e-01 -4.06731963e-01
-6.47700667e-01 2.38737613e-01 -1.23527408e-01 -4.57807481e-01
1.05061568e-01 -6.23383105e-01 -1.90876931e-01 3.96862268e-01
2.84998804e-01 4.01224643e-01 -2.49414444e-02 -4.04546499e-01
6.99476302e-02 4.55158830e-01 -3.65773737e-02 -2.45343328e-01
1.40956914e+00 -1.33155316e-01 8.36331025e-02 9.98104930e-01
1.29664814e+00 -2.73778319e-01 -1.32339001e+00 -7.23136544e-01
2.83625215e-01 -3.66199613e-01 -1.93182364e-01 -5.59171259e-01
-7.91635275e-01 7.97835767e-01 8.06691229e-01 7.30606616e-01
1.45080531e+00 -2.19242811e-01 7.51843274e-01 1.65921777e-01
4.48345095e-01 -9.88687277e-01 1.58181936e-02 3.10527951e-01
5.23205876e-01 -1.54991007e+00 -1.71645477e-01 -2.89843410e-01
-5.26452720e-01 6.09415531e-01 5.27296484e-01 -5.53861186e-02
8.84952843e-01 1.26431346e-01 3.07381034e-01 8.83136690e-02
-1.01059699e+00 -2.64118195e-01 1.26655623e-01 4.38817501e-01
4.11721706e-01 -4.97877598e-04 -2.39402771e-01 3.82358789e-01
-2.06536055e-02 2.18447655e-01 2.08550483e-01 1.17334223e+00
1.07222095e-01 -1.37058794e+00 -4.92312819e-01 6.34910941e-01
-5.28234899e-01 1.07126758e-01 1.45956457e-01 6.87948942e-01
1.53365210e-01 8.90214145e-01 2.26123855e-01 -3.66141051e-01
1.27756089e-01 4.92114455e-01 -2.35026851e-01 -5.95671594e-01
-5.88166416e-01 4.81883734e-01 -8.21911618e-02 -2.78095096e-01
-8.28388512e-01 -7.66352117e-01 -1.16131556e+00 -4.12148312e-02
-1.80838332e-01 1.70403734e-01 7.24248350e-01 9.16211665e-01
3.46760452e-01 5.73807538e-01 1.07113922e+00 -7.55040705e-01
-7.09336281e-01 -1.22574556e+00 -4.54595000e-01 7.95754015e-01
5.35310447e-01 -6.05413139e-01 -2.01587349e-01 5.10147870e-01] | [7.281513214111328, 2.9338676929473877] |
208e0314-e648-4aac-ac19-dcf4d1f3af40 | promptfusion-decoupling-stability-and | 2303.07223 | null | https://arxiv.org/abs/2303.07223v1 | https://arxiv.org/pdf/2303.07223v1.pdf | PromptFusion: Decoupling Stability and Plasticity for Continual Learning | Continual learning refers to the capability of continuously learning from a stream of data. Current research mainly focuses on relieving catastrophic forgetting, and most of their success is at the cost of limiting the performance of newly incoming tasks. Such a trade-off is referred to as the stabilityplasticity dilemma and is a more general and challenging problem for continual learning. However, the inherent conflict between these two concepts makes it seemingly impossible to devise a satisfactory solution to both of them simultaneously. Therefore, we ask, "is it possible to divide them into two problems to conquer independently?" To this end, we propose a prompt-tuning-based method termed PromptFusion to enable the decoupling of stability and plasticity. Specifically, PromptFusion consists of a carefully designed Stabilizer module that deals with catastrophic forgetting and a Booster module to learn new knowledge concurrently. During training, PromptFusion first passes an input image to the two modules separately. Then the resulting logits are further fused with a learnable weight parameter. Finally, a weight mask is applied to the derived logits to balance between old and new classes. Extensive experiments show that our method achieves promising results on popular continual learning datasets for both class-incremental and domain incremental settings. Especially on Split-Imagenet-R, one of the most challenging datasets for class-incremental learning, our method exceeds state-of-the-art prompt-based methods L2P and DualPrompt by more than 10%. | ['Yu-Gang Jiang', 'Menglin Jia', 'Xintong Han', 'Zuxuan Wu', 'Haoran Chen'] | 2023-03-13 | null | null | null | null | ['class-incremental-learning'] | ['computer-vision'] | [ 5.15587665e-02 -1.81720689e-01 -2.15530708e-01 -2.35670313e-01
-6.08694613e-01 -3.30186367e-01 4.55272287e-01 2.03859597e-01
-7.40232646e-01 9.10142243e-01 -2.45868087e-01 -1.11653000e-01
-2.36590028e-01 -5.65434396e-01 -8.42876792e-01 -8.94658029e-01
7.69834071e-02 3.90218079e-01 6.19232059e-01 -1.92943543e-01
1.01970106e-01 1.81589171e-01 -1.62992620e+00 9.17612910e-02
1.18370974e+00 1.12276089e+00 4.92860526e-01 1.48291230e-01
-2.22308472e-01 6.68521523e-01 -3.73476774e-01 -2.98369408e-01
1.02359585e-01 -3.85207862e-01 -6.03939891e-01 2.63239071e-02
2.89333254e-01 -1.21544451e-01 -6.91266954e-02 9.72885847e-01
5.64683139e-01 2.83404350e-01 4.46920395e-01 -1.23702383e+00
-6.59125388e-01 6.93996429e-01 -7.74483025e-01 5.95461726e-01
-4.04799134e-02 4.18637097e-01 7.33990073e-01 -1.01353109e+00
2.97698051e-01 1.01910985e+00 5.41033983e-01 6.10431254e-01
-1.44321883e+00 -8.37750912e-01 6.42004251e-01 2.32734188e-01
-1.22132897e+00 -4.04637545e-01 8.07925105e-01 -2.95939296e-01
7.93992221e-01 -6.14678301e-02 7.95157731e-01 1.12051845e+00
2.08122849e-01 9.67720270e-01 1.12036133e+00 -8.84035975e-02
4.69463587e-01 2.93112248e-01 9.33975279e-02 5.36804557e-01
1.64886430e-01 -1.26305558e-02 -8.66057754e-01 1.33981183e-01
5.37059069e-01 1.31854296e-01 -3.64999354e-01 -3.40817899e-01
-9.94848967e-01 4.73081350e-01 5.51772773e-01 2.55529374e-01
-3.67052436e-01 -5.78757115e-02 1.94701284e-01 6.03797138e-01
5.85103869e-01 4.38242912e-01 -5.70116520e-01 -6.68882877e-02
-9.69336689e-01 2.85912454e-01 3.67661834e-01 5.52366316e-01
7.81176984e-01 1.10155262e-01 -6.08160049e-02 8.94424140e-01
3.07846032e-02 2.02246264e-01 7.63113797e-01 -5.66895247e-01
3.74514997e-01 8.30419660e-01 -8.59321132e-02 -6.58530653e-01
-3.19183052e-01 -8.54884028e-01 -9.29265559e-01 2.14659855e-01
3.67367595e-01 -7.36757219e-02 -9.10359204e-01 2.16445088e+00
5.15213192e-01 3.56861234e-01 -1.12038195e-01 7.88071215e-01
4.72761840e-01 5.62465489e-01 3.98989797e-01 -5.23177505e-01
1.02842975e+00 -1.05887210e+00 -5.07702589e-01 -5.00931144e-01
1.05201572e-01 -4.35150862e-01 1.53961837e+00 4.50881153e-01
-1.17028868e+00 -5.19076467e-01 -1.19981182e+00 9.82083157e-02
-2.76731551e-01 -1.95844501e-01 4.05218393e-01 1.72325820e-01
-8.02468538e-01 7.51348197e-01 -9.77350831e-01 -3.86082381e-02
4.98788506e-01 3.44995856e-01 -1.46571890e-01 1.13468282e-02
-1.13063943e+00 8.21115077e-01 4.81622368e-01 -1.40748516e-01
-9.28523421e-01 -1.00782466e+00 -4.13502783e-01 2.01025844e-01
5.78080118e-01 -8.67987692e-01 1.19717538e+00 -1.04769576e+00
-1.40030706e+00 5.12461722e-01 -9.13428813e-02 -5.59751153e-01
7.95488119e-01 -4.42464292e-01 -1.82541803e-01 -1.84472308e-01
-3.23079410e-03 6.97876930e-01 1.22920120e+00 -1.15025520e+00
-6.25491858e-01 -4.05068547e-01 3.98573317e-02 4.39022899e-01
-8.16553295e-01 -4.52573627e-01 -4.14924502e-01 -7.51873970e-01
1.96297556e-01 -8.02834630e-01 6.04331009e-02 6.86766952e-02
-7.88523722e-03 -2.80547470e-01 8.59048843e-01 -2.33286768e-01
1.40446532e+00 -2.28315187e+00 4.01531965e-01 -3.20691288e-01
2.64953047e-01 3.99421751e-01 -1.71484366e-01 -4.26320396e-02
-2.50555098e-01 -3.61184254e-02 -3.62401247e-01 -5.79452217e-01
-4.05798137e-01 1.28428280e-01 -5.88885486e-01 2.47411489e-01
3.13425869e-01 9.85472500e-01 -1.12598777e+00 -2.97551930e-01
-1.88669249e-01 4.18941736e-01 -4.50521290e-01 1.18840367e-01
-4.78254497e-01 4.35061961e-01 -7.14136958e-02 6.02683961e-01
7.20505059e-01 -3.90037805e-01 6.83575124e-02 1.23530060e-01
-7.41570741e-02 3.50124575e-02 -1.08335912e+00 1.68393207e+00
-2.83381373e-01 1.83052167e-01 -1.40957579e-01 -9.23642635e-01
7.08403528e-01 1.29199550e-01 2.78176636e-01 -8.28281701e-01
-7.45936111e-02 2.13793308e-01 -1.39962703e-01 -3.41134518e-01
3.44034761e-01 -5.34735620e-01 1.23694196e-01 4.92319047e-01
1.52192250e-01 4.23140414e-02 1.00193664e-01 1.83919072e-01
1.16822350e+00 5.47547303e-02 4.30231616e-02 -2.38174409e-01
4.87153649e-01 -3.40086460e-01 9.05194759e-01 5.56706548e-01
-4.14590150e-01 5.26928484e-01 5.05583584e-01 -5.43763816e-01
-7.30647564e-01 -1.32704568e+00 1.05145521e-01 1.23652685e+00
2.64404774e-01 -2.02487573e-01 -5.48052013e-01 -8.85534525e-01
1.65791243e-01 5.87855875e-01 -7.01069534e-01 -6.10221505e-01
-5.94899654e-01 -9.75799978e-01 1.85846031e-01 4.51194197e-01
7.78046668e-01 -9.63810563e-01 -6.39037967e-01 2.90627658e-01
-1.59261048e-01 -5.91229916e-01 -5.64791918e-01 5.40279388e-01
-1.23449540e+00 -9.16251183e-01 -7.10090816e-01 -6.86459541e-01
6.31081164e-01 6.27673149e-01 1.02624202e+00 8.32169652e-02
-5.92814200e-02 6.71087205e-02 -1.47512525e-01 -2.06424102e-01
-3.51370350e-02 4.80369925e-01 3.15184146e-01 1.36670217e-01
-1.06891366e-02 -1.03633332e+00 -7.33561397e-01 3.04142892e-01
-1.17765462e+00 8.41538757e-02 7.24960625e-01 1.02863860e+00
6.86881423e-01 2.41392761e-01 1.01325917e+00 -7.71797538e-01
6.88575685e-01 -6.81925833e-01 -4.28915143e-01 4.26292479e-01
-9.65738177e-01 2.21073180e-01 8.28999341e-01 -1.07538748e+00
-1.16233027e+00 8.10433999e-02 4.85333689e-02 -6.59214616e-01
2.99590439e-01 5.16793072e-01 -2.61717081e-01 1.24562792e-01
7.07270563e-01 4.56274867e-01 8.14894140e-02 -4.42350954e-01
4.39735353e-01 2.02892229e-01 6.09384656e-01 -6.38194621e-01
8.46891582e-01 3.71487051e-01 -4.23054099e-01 -5.73149741e-01
-1.10170019e+00 -1.58968061e-01 -4.86472011e-01 -3.01526815e-01
3.85095656e-01 -9.12205100e-01 -4.29528117e-01 9.08428311e-01
-7.52817929e-01 -5.26103973e-01 -7.03800976e-01 7.82122090e-02
-4.39911246e-01 7.03604817e-02 -4.67371643e-01 -6.51382267e-01
-3.98176283e-01 -8.91270697e-01 5.48219502e-01 5.34956276e-01
6.63183033e-02 -8.38765323e-01 1.64514527e-01 3.09622169e-01
7.16437221e-01 -6.39024228e-02 1.04422855e+00 -3.00402045e-01
-6.01408899e-01 8.53600260e-03 -7.73844719e-02 5.12423217e-01
1.38879269e-01 -3.63976926e-01 -9.50592041e-01 -6.39222920e-01
2.09170520e-01 -7.06076682e-01 1.39270401e+00 3.73959206e-02
1.09022176e+00 -3.45181614e-01 -1.75359666e-01 5.94726801e-01
1.31491756e+00 6.44396618e-02 4.88476634e-01 3.48332852e-01
3.32735151e-01 3.38506550e-01 5.82988739e-01 3.98326784e-01
4.11987305e-01 3.58742207e-01 4.23603892e-01 2.39617422e-01
-2.77794480e-01 -4.54587936e-01 3.05576682e-01 8.78794193e-01
2.16640472e-01 7.16021890e-03 -8.19098949e-01 4.95714009e-01
-1.99352825e+00 -8.81624579e-01 4.92836714e-01 2.32178926e+00
1.31614697e+00 5.18866181e-01 3.46024632e-02 2.01849177e-01
5.23093224e-01 2.58972228e-01 -1.14939916e+00 2.76081890e-01
-2.70084530e-01 -1.35633662e-01 4.65796702e-02 3.14975411e-01
-1.06363034e+00 1.05237174e+00 5.48735285e+00 1.03861892e+00
-1.50357258e+00 3.53267193e-01 9.00253177e-01 -4.90130693e-01
-3.11069727e-01 1.42968208e-01 -8.77902269e-01 6.14822268e-01
8.00275326e-01 -3.30180645e-01 5.62833667e-01 8.11901212e-01
-1.00901254e-01 -2.86911458e-01 -8.72270644e-01 8.02146137e-01
-7.29695633e-02 -1.11169171e+00 -1.59597546e-02 -3.42221707e-01
7.54009068e-01 2.28077676e-02 5.75732112e-01 7.13491201e-01
3.38345468e-02 -6.47516131e-01 9.61237371e-01 5.49232543e-01
7.42003798e-01 -6.56017303e-01 2.68692672e-01 7.49741971e-01
-1.05256951e+00 -4.39611495e-01 -2.42335483e-01 4.83162850e-02
6.67089075e-02 8.54583323e-01 -4.56151009e-01 2.38567069e-01
7.09457397e-01 6.30308926e-01 -7.98912108e-01 1.20724738e+00
-2.94687480e-01 6.38301015e-01 -3.18364263e-01 2.83746183e-01
-2.49129776e-02 8.58665034e-02 5.37413299e-01 7.62517810e-01
1.86772525e-01 -2.15593465e-02 -6.38684956e-03 8.86592865e-01
-3.27577949e-01 -8.42959508e-02 -3.45296830e-01 1.33373186e-01
5.69777906e-01 1.16958451e+00 -8.03180337e-01 -3.19453508e-01
-1.56311184e-01 1.01070940e+00 8.80865932e-01 2.90451974e-01
-8.07246685e-01 -5.36074080e-02 3.80614221e-01 4.27023351e-01
2.35679835e-01 -2.13638559e-01 -4.35004175e-01 -1.47926176e+00
2.30424002e-01 -7.82360077e-01 4.75902736e-01 -4.83421236e-01
-1.29434085e+00 5.92827737e-01 -8.86954889e-02 -1.10117853e+00
1.47499651e-01 -3.44391353e-02 -6.08677149e-01 6.22019649e-01
-1.93637025e+00 -9.43389297e-01 -4.63968873e-01 5.62578082e-01
7.88964570e-01 -1.21252224e-01 4.41039115e-01 4.01974678e-01
-8.78321528e-01 7.21565068e-01 9.17161107e-02 -5.61286688e-01
9.09254014e-01 -1.24622619e+00 2.11205706e-01 7.15613246e-01
-1.11213751e-01 7.07781315e-01 5.51527798e-01 -7.82360077e-01
-1.31861925e+00 -1.14658034e+00 7.09731340e-01 -3.14751059e-01
5.71923077e-01 -3.51787448e-01 -1.44650126e+00 3.95762920e-01
-3.63983475e-02 1.32480949e-01 3.67128551e-01 1.69810709e-02
-5.76653302e-01 -5.35331786e-01 -9.37957942e-01 4.00163800e-01
1.10754514e+00 -2.02333137e-01 -6.75242901e-01 2.33073756e-01
1.05892849e+00 -2.58744001e-01 -4.00933862e-01 4.58084524e-01
3.56824160e-01 -1.01799297e+00 8.78615201e-01 -4.62308139e-01
3.58619630e-01 -2.06435010e-01 1.24381915e-01 -1.40697765e+00
-2.89243191e-01 -6.04714632e-01 -5.06119728e-01 1.28931689e+00
3.34253401e-01 -6.63912892e-01 7.62756765e-01 3.96944225e-01
5.74646518e-02 -1.14292204e+00 -1.09125113e+00 -1.07051849e+00
3.20684075e-01 -2.58674145e-01 4.86219883e-01 8.62878799e-01
-1.45378068e-01 5.54283440e-01 -2.61021256e-01 -2.00383738e-01
6.47301376e-01 6.81857020e-02 5.17470777e-01 -1.22283673e+00
-3.87039989e-01 -4.74159151e-01 1.61444426e-01 -1.14325464e+00
8.54575038e-02 -7.99539208e-01 6.27092645e-02 -9.99512315e-01
3.27447802e-01 -6.68201566e-01 -6.17448509e-01 7.28051484e-01
-5.58747947e-01 -9.03060138e-02 2.16504142e-01 5.93673527e-01
-8.20902705e-01 1.07693994e+00 1.05683005e+00 -1.75653592e-01
-6.25888288e-01 -2.02461425e-02 -8.67993832e-01 5.78715265e-01
7.55998433e-01 -4.99895126e-01 -7.00598836e-01 -4.70529437e-01
3.62619698e-01 -4.82877009e-02 2.04896212e-01 -1.19996214e+00
5.42006254e-01 -3.47779870e-01 1.99294135e-01 -4.69205618e-01
1.74305558e-01 -5.11498570e-01 -6.90176338e-02 5.28351367e-01
-2.93194264e-01 1.25700936e-01 3.13965917e-01 8.65634978e-01
-1.38135254e-01 9.20712352e-02 1.14378941e+00 -1.59024358e-01
-4.91167605e-01 4.61588800e-01 -9.47903097e-02 3.21035206e-01
1.08823633e+00 6.89420998e-02 -4.45339650e-01 -1.40740678e-01
-6.40670240e-01 5.49809575e-01 3.38421494e-01 6.19923115e-01
8.16292524e-01 -1.16369081e+00 -4.54115391e-01 3.12468469e-01
-1.24039473e-02 2.46164724e-01 4.33879167e-01 1.08227026e+00
7.29464814e-02 2.64486242e-02 -1.35999739e-01 -4.93213385e-01
-9.77787912e-01 7.51087487e-01 3.05806935e-01 -3.98152173e-01
-5.43197930e-01 1.06651223e+00 2.54789591e-01 -1.55073762e-01
5.58945596e-01 -1.05314389e-01 -1.13973580e-01 3.99504244e-01
6.48210764e-01 2.65521288e-01 9.77633223e-02 -9.18996185e-02
-3.47484350e-01 1.63910866e-01 -4.35121357e-01 -8.70938897e-02
1.56366932e+00 -2.33228743e-01 -3.57879177e-02 7.16705799e-01
8.64562869e-01 -4.75926042e-01 -1.82764328e+00 -6.15524948e-01
-7.06272349e-02 -2.65104145e-01 -1.12695858e-01 -9.77396905e-01
-1.13034368e+00 9.13052320e-01 7.14161336e-01 6.39296845e-02
1.37586439e+00 -5.09630516e-02 9.34673429e-01 3.25311661e-01
3.08551669e-01 -1.20161963e+00 7.58384943e-01 6.63222849e-01
8.62546146e-01 -1.24408126e+00 -1.14950716e-01 1.46311373e-01
-5.19839048e-01 8.96758437e-01 9.73811626e-01 -1.62491377e-03
7.00269461e-01 2.71915317e-01 -1.52166560e-01 9.92366299e-02
-1.14692140e+00 6.16347156e-02 1.17619917e-01 1.76343724e-01
8.71465281e-02 -2.89064795e-01 -3.33066404e-01 8.71985734e-01
-2.87776534e-02 1.44938678e-01 2.17222691e-01 1.08211720e+00
-7.89954603e-01 -9.59481537e-01 -1.00361481e-01 3.52285087e-01
-2.38786936e-01 -5.85015193e-02 -2.07473248e-01 4.84886795e-01
3.64884317e-01 5.17786503e-01 -1.79845184e-01 -6.01147234e-01
3.52442265e-01 2.48500541e-01 2.83146530e-01 -5.11766613e-01
-5.76171160e-01 -6.92189066e-03 -6.26668453e-01 -3.16108257e-01
-7.02465922e-02 -7.14375973e-01 -1.17742801e+00 -2.38665134e-01
-3.74253899e-01 7.80820032e-04 4.43059623e-01 8.41944933e-01
4.55858678e-01 5.53209960e-01 6.49000704e-01 -5.78704298e-01
-9.61357474e-01 -7.60267496e-01 -4.41678137e-01 3.26549292e-01
4.76893485e-01 -8.67497563e-01 -5.71129799e-01 -7.45825917e-02] | [9.845757484436035, 3.4486210346221924] |
fd013cb7-36f0-4dd8-82c3-0144810649f5 | segmentation-of-argumentative-texts-with | null | null | https://aclanthology.org/W19-4501 | https://aclanthology.org/W19-4501.pdf | Segmentation of Argumentative Texts with Contextualised Word Representations | The segmentation of argumentative units is an important subtask of argument mining, which is frequently addressed at a coarse granularity, usually assuming argumentative units to be no smaller than sentences. Approaches focusing at the clause-level granularity, typically address the task as sequence labeling at the token level, aiming to classify whether a token begins, is inside, or is outside of an argumentative unit. Most approaches exploit highly engineered, manually constructed features, and algorithms typically used in sequential tagging {--} such as Conditional Random Fields, while more recent approaches try to exploit manually constructed features in the context of deep neural networks. In this context, we examined to what extend recent advances in sequential labelling allow to reduce the need for highly sophisticated, manually constructed features, and whether limiting features to embeddings, pre-trained on large corpora is a promising approach. Evaluation results suggest the examined models and approaches can exhibit comparable performance, minimising the need for feature engineering. | ['Georgios Petasis'] | 2019-08-01 | null | null | null | ws-2019-8 | ['contextualised-word-representations'] | ['natural-language-processing'] | [ 4.14071083e-01 6.73917890e-01 -5.04484594e-01 -4.26025480e-01
-7.65878141e-01 -8.88459682e-01 8.80153239e-01 1.00232208e+00
-7.67667294e-01 8.84528697e-01 4.03347343e-01 -7.45902121e-01
-4.92613800e-02 -7.57916987e-01 -5.18755257e-01 -4.14727956e-01
-1.09544548e-03 6.07328653e-01 2.66107023e-01 1.43686101e-01
4.85096186e-01 2.70329356e-01 -1.42983139e+00 5.06465256e-01
5.96319795e-01 8.83992910e-01 -1.06523551e-01 5.28266609e-01
-6.80401266e-01 9.62077260e-01 -8.30474198e-01 -4.63228047e-01
-1.37889177e-01 -3.30880702e-01 -1.40225923e+00 -2.23629639e-01
1.60589486e-01 9.22470689e-02 5.88988364e-01 9.13318336e-01
1.15308590e-01 -1.98754266e-01 6.25054359e-01 -4.89729315e-01
-1.97806135e-01 1.21308744e+00 -9.78831574e-02 2.86130756e-01
2.64164180e-01 -3.79590094e-01 1.42330468e+00 -4.67471600e-01
6.90703392e-01 9.84837234e-01 6.69159889e-01 3.27346355e-01
-1.34804714e+00 -1.36998594e-01 2.09837377e-01 -1.01982960e-02
-6.77708030e-01 -2.51886964e-01 7.36242354e-01 -6.22928619e-01
9.83538687e-01 2.54349172e-01 4.14292276e-01 9.26619291e-01
-2.10695729e-01 8.83897960e-01 1.19699943e+00 -8.42242599e-01
4.73177403e-01 1.33167654e-01 5.88672817e-01 2.52554536e-01
3.62301648e-01 -3.44810814e-01 -6.90792352e-02 -4.48177099e-01
2.89915472e-01 -2.63252378e-01 6.93093538e-02 -1.25236571e-01
-9.39168513e-01 1.28197956e+00 6.10534064e-02 1.02555585e+00
-3.59468490e-01 1.82329398e-02 8.61747205e-01 2.76054174e-01
5.56698620e-01 7.77256191e-01 -8.21284473e-01 -4.97938097e-01
-1.00213885e+00 1.73331693e-01 1.00417459e+00 6.08939409e-01
8.23153377e-01 -2.66185582e-01 -1.93385735e-01 5.90375066e-01
2.09821403e-01 -2.89098680e-01 5.15425503e-01 -6.84258878e-01
6.35355592e-01 1.06844866e+00 -1.09809591e-02 -3.22663963e-01
-3.61156791e-01 -3.77015084e-01 -1.60891399e-01 -1.95862493e-03
6.61354601e-01 -3.53694648e-01 -7.47731984e-01 1.63259923e+00
5.44264555e-01 -3.60292852e-01 1.19805977e-01 5.48177123e-01
3.83842707e-01 2.79961765e-01 4.24736798e-01 -3.40060174e-01
1.62530422e+00 -7.21051037e-01 -4.61541742e-01 -2.21399724e-01
1.18216407e+00 -8.04685473e-01 9.11087036e-01 3.34119946e-01
-9.63608742e-01 -3.65741998e-01 -8.05475950e-01 -4.49939724e-03
-5.35866916e-01 1.63368046e-01 7.46841490e-01 7.82916129e-01
-6.69763029e-01 8.85487795e-01 -6.62046254e-01 6.12811011e-04
2.86682844e-01 4.49981600e-01 -3.00501913e-01 2.50884742e-01
-1.34034169e+00 7.12984085e-01 6.98421896e-01 7.89767429e-02
-1.85100228e-01 -2.96944678e-01 -1.05958211e+00 3.31238121e-01
4.75448161e-01 -1.58422768e-01 1.41674292e+00 -1.23338580e+00
-1.38323200e+00 9.64010298e-01 -5.41795194e-02 -8.67018998e-01
3.84884238e-01 -4.46200073e-01 -7.52217025e-02 9.03003290e-02
1.45123184e-01 3.75952631e-01 7.95773983e-01 -8.98467362e-01
-6.73099458e-01 -2.52782077e-01 3.90657127e-01 1.30735775e-02
-3.86463732e-01 4.44839329e-01 2.93588400e-01 -3.79492640e-01
-1.08201832e-01 -7.43840933e-01 -3.94372344e-01 -6.85557663e-01
-4.67265397e-01 -9.37962115e-01 7.12729275e-01 -4.70490158e-01
1.37046182e+00 -1.88518167e+00 -9.86379534e-02 1.42019689e-01
2.52471678e-02 6.78607106e-01 3.32843065e-01 4.65878546e-01
-2.31459245e-01 4.90991861e-01 -1.89962804e-01 -4.41590808e-02
1.42099202e-01 2.47183710e-01 -1.53858989e-01 2.70755321e-01
3.55846971e-01 6.28710806e-01 -8.98025036e-01 -6.03584588e-01
8.18536952e-02 2.77256761e-02 -5.05685925e-01 -4.68653925e-02
-4.67206806e-01 8.59304965e-02 -7.96455801e-01 3.11202407e-01
2.51467139e-01 -9.51513872e-02 7.50134110e-01 5.59733175e-02
-4.91740733e-01 1.09118354e+00 -1.01346850e+00 1.25034392e+00
-4.37437117e-01 5.06764710e-01 3.29900123e-02 -1.58559525e+00
7.93451548e-01 5.03511310e-01 2.37815216e-01 -4.33267623e-01
4.34372544e-01 5.42268634e-01 2.01258674e-01 -2.73971826e-01
4.19325501e-01 3.78120616e-02 -3.88643354e-01 3.33234906e-01
-2.00008461e-03 2.73765206e-01 3.92355680e-01 -2.00767636e-01
1.31526399e+00 2.18336210e-01 8.06543827e-01 -3.28185707e-01
6.60676479e-01 2.57862866e-01 5.40829241e-01 6.11421943e-01
2.39008125e-02 2.96966761e-01 9.07509923e-01 -5.29336810e-01
-1.05963719e+00 -4.90577847e-01 -3.32464933e-01 1.02837908e+00
-4.09073591e-01 -5.64573050e-01 -1.10350931e+00 -1.25734353e+00
-1.92187354e-01 5.04983664e-01 -6.87290072e-01 4.20773685e-01
-1.03962755e+00 -5.46103895e-01 4.29177701e-01 4.99385118e-01
1.18008025e-01 -1.27679777e+00 -1.00233936e+00 6.59906328e-01
1.08925760e-01 -9.27318633e-01 1.39943451e-01 9.75331366e-01
-9.92984951e-01 -1.04526269e+00 -4.56006259e-01 -7.86592007e-01
5.04824698e-01 -5.10536015e-01 1.21819639e+00 2.66726911e-01
-1.59095302e-01 -8.02660659e-02 -7.92331576e-01 -3.83965433e-01
-3.12335819e-01 4.47676301e-01 -3.51983279e-01 -2.26063132e-01
6.86801016e-01 -4.10118341e-01 -2.65466362e-01 -6.89965039e-02
-8.28088164e-01 -4.12704438e-01 6.86894000e-01 1.01757944e+00
4.06729370e-01 -2.97052324e-01 4.16224152e-01 -1.55238450e+00
6.90021634e-01 -3.34691733e-01 -2.45013207e-01 1.78995401e-01
-5.10208428e-01 4.43361253e-01 8.67310464e-01 -4.58256543e-01
-8.49366367e-01 -1.14092134e-01 -4.39042777e-01 2.30892286e-01
-7.68311322e-01 6.66031539e-01 -1.40762463e-01 3.90307724e-01
6.51252985e-01 -2.54721671e-01 -3.10785621e-01 -6.38265967e-01
2.80603856e-01 7.28626132e-01 -1.54409530e-02 -9.62597549e-01
4.55068409e-01 2.24667192e-01 -1.19325317e-01 -5.96909642e-01
-1.09402251e+00 -6.15489244e-01 -7.64012218e-01 3.21557343e-01
8.95968199e-01 -4.74031478e-01 -4.50099140e-01 -1.77305877e-01
-1.13792658e+00 -3.58294517e-01 -5.48224509e-01 4.94468302e-01
-5.27645230e-01 4.40367281e-01 -6.76767826e-01 -8.24280679e-01
-2.90566564e-01 -7.85096526e-01 9.26036656e-01 1.87228277e-01
-8.81640613e-01 -1.08096337e+00 1.20111816e-01 3.95527899e-01
5.18295430e-02 1.82842731e-01 1.17058027e+00 -1.52677548e+00
-7.74967074e-02 -3.10667813e-01 -8.69934261e-02 3.55990767e-01
9.30053145e-02 -1.22007698e-01 -8.95816207e-01 7.98803270e-02
7.53992423e-02 -4.76521701e-01 7.33865380e-01 1.12502292e-01
9.74696994e-01 -4.28116411e-01 -3.40697795e-01 -1.01955034e-01
1.13997805e+00 2.78452903e-01 4.41456616e-01 7.16921270e-01
2.66242653e-01 9.22410846e-01 8.89970183e-01 1.68195039e-01
3.32699120e-02 4.96198803e-01 3.39819670e-01 -1.58017501e-02
1.88912362e-01 -8.06021467e-02 2.09627017e-01 5.78898549e-01
-8.64647105e-02 -2.19048828e-01 -7.45621920e-01 7.69422174e-01
-1.71913946e+00 -9.69238818e-01 -3.70013863e-01 1.78853202e+00
9.98023927e-01 7.75711358e-01 5.57307452e-02 6.81871116e-01
7.09942400e-01 1.39232680e-01 1.81352392e-01 -1.00826859e+00
2.85205305e-01 6.09821856e-01 4.37447309e-01 3.90830606e-01
-1.28353143e+00 8.93108785e-01 5.54381943e+00 9.10929799e-01
-9.12995934e-01 8.01740214e-02 6.23535931e-01 4.04524297e-01
-3.66895288e-01 6.52595878e-01 -1.04917967e+00 5.39417624e-01
1.10635555e+00 5.97743154e-01 -2.31560156e-01 1.02837062e+00
1.05641358e-01 -2.84524202e-01 -1.15575469e+00 1.56117111e-01
-1.79517180e-01 -1.36868072e+00 -2.50784576e-01 2.98970371e-01
3.43551368e-01 -1.39364704e-01 -2.30447665e-01 4.19254571e-01
2.73210436e-01 -8.24477494e-01 7.80905902e-01 -2.13405211e-03
3.03844571e-01 -5.07786453e-01 1.12342322e+00 3.61309826e-01
-1.04107702e+00 -3.99616593e-03 -1.90243125e-01 -4.38892007e-01
1.92911968e-01 6.89364731e-01 -1.07549870e+00 5.01511931e-01
4.49418396e-01 2.63633966e-01 -2.90306300e-01 7.39772916e-01
-4.81351584e-01 1.15089667e+00 -2.93807626e-01 -4.87178206e-01
7.09067404e-01 -7.49040991e-02 3.22813869e-01 1.49463177e+00
-1.02868475e-01 -2.43697181e-01 3.24195147e-01 4.30016845e-01
1.87457681e-01 4.30048019e-01 -2.16270968e-01 -2.13936597e-01
2.24167556e-01 1.29714239e+00 -1.04469705e+00 -4.78616863e-01
-4.05498385e-01 4.93795365e-01 5.92413962e-01 -9.04383883e-02
-6.13580227e-01 -2.94672608e-01 3.02301675e-01 3.09339225e-01
6.68918192e-01 -7.99170136e-02 -3.19275051e-01 -6.25944495e-01
2.39782214e-01 -7.01320767e-01 5.56528091e-01 7.11381296e-03
-9.30995047e-01 6.59166574e-01 3.62404212e-02 -9.93987858e-01
-6.39325440e-01 -8.14495504e-01 -7.98037469e-01 6.73294961e-01
-1.49639809e+00 -1.01186454e+00 4.00459290e-01 1.02192119e-01
6.24052286e-01 1.60062402e-01 8.01316440e-01 1.32141650e-01
-2.49940544e-01 3.28164488e-01 -1.19893946e-01 3.22814971e-01
3.31909209e-01 -1.44410980e+00 1.42494425e-01 5.74035168e-01
3.69898558e-01 7.00975597e-01 6.77519321e-01 -5.02551734e-01
-6.25925958e-01 -6.51736796e-01 1.58439541e+00 -2.44233027e-01
6.68458641e-01 -4.24602240e-01 -9.08181310e-01 4.23695147e-01
1.79172590e-01 -1.44706722e-02 8.09922159e-01 6.60196900e-01
-2.16038436e-01 3.65594834e-01 -1.06495535e+00 2.25918904e-01
7.31897295e-01 -5.39176762e-01 -1.07516336e+00 -6.09284686e-03
5.59926152e-01 -1.59957096e-01 -8.74581277e-01 3.33203346e-01
2.67396092e-01 -7.34638572e-01 5.99392116e-01 -7.34770834e-01
3.92624497e-01 -5.81565052e-02 8.22751820e-02 -8.66540849e-01
-5.87783121e-02 -5.27585804e-01 -9.50994194e-02 1.60203886e+00
7.63180375e-01 -5.25882781e-01 1.05734789e+00 4.31418329e-01
-2.67853945e-01 -7.95561969e-01 -1.04636240e+00 -4.90598857e-01
2.31041878e-01 -2.47956887e-01 5.22833943e-01 9.71582532e-01
2.59959042e-01 5.63122869e-01 9.18775350e-02 -2.91689903e-01
1.15725972e-01 3.11701685e-01 4.94356573e-01 -1.66591954e+00
-3.38502169e-01 -4.78214711e-01 -3.94183576e-01 -8.61983597e-01
4.43157464e-01 -7.59227097e-01 -8.61632600e-02 -1.48145652e+00
-2.22649008e-01 -6.16138279e-01 -2.17267856e-01 4.27251488e-01
-1.23529918e-01 -1.12499677e-01 -1.09221525e-01 -5.95638789e-02
-4.44802493e-01 7.55807236e-02 7.51034081e-01 -9.52531919e-02
-1.44945249e-01 7.61763602e-02 -5.28488696e-01 1.02808774e+00
7.77353883e-01 -7.28175104e-01 -7.40176886e-02 -1.83377802e-01
4.21302468e-01 -1.99571624e-01 -6.76334091e-03 -7.16196477e-01
1.07748441e-01 -1.42095700e-01 6.52124584e-02 -4.35851395e-01
6.25414327e-02 -7.28217006e-01 -2.37718359e-01 2.82286882e-01
-5.03219068e-01 -1.48695251e-02 -5.24306446e-02 3.55642259e-01
-6.41895592e-01 -1.02980626e+00 3.83053094e-01 -3.30523193e-01
-4.75285530e-01 -2.25548267e-01 -4.15366054e-01 2.16240793e-01
9.55201328e-01 -3.91261786e-01 -2.21598893e-02 2.22790852e-01
-8.45061362e-01 -8.41944143e-02 1.04276545e-01 2.17241682e-02
2.43663937e-02 -7.15736151e-01 -3.38306665e-01 -2.12412223e-01
-7.71854892e-02 1.38379008e-01 -3.20743710e-01 7.53964722e-01
-4.90986794e-01 6.64330423e-01 2.51301974e-01 -3.92384946e-01
-1.40997350e+00 5.38449347e-01 -6.63064122e-02 -9.36094165e-01
-5.27983546e-01 7.13889658e-01 -1.52916655e-01 -3.68060410e-01
3.26473378e-02 -6.64581597e-01 -7.40119696e-01 6.25597775e-01
1.27256006e-01 1.49062835e-02 3.81703824e-01 -5.08441210e-01
-2.92091340e-01 2.77186036e-01 -2.59314716e-01 4.34300713e-02
1.30675769e+00 2.18870044e-01 -6.48021325e-02 3.48790616e-01
1.01313913e+00 4.03231412e-01 -9.66963828e-01 -9.84021053e-02
8.64083230e-01 2.72281538e-03 -2.18043938e-01 -4.56180692e-01
-3.38051319e-01 8.94566655e-01 1.06741637e-01 8.65885615e-01
6.12954974e-01 2.26968482e-01 5.86124122e-01 3.28225076e-01
3.85526329e-01 -1.28547704e+00 -2.54345655e-01 7.06496239e-01
1.62164763e-01 -8.91616642e-01 -1.49473771e-02 -5.11167765e-01
-3.20966423e-01 1.33293283e+00 2.81018019e-01 -2.32641846e-01
4.71116632e-01 4.08326209e-01 -2.28832319e-01 -2.08431974e-01
-5.12503445e-01 -4.04054135e-01 1.08339049e-01 1.92242429e-01
9.63255882e-01 1.22818649e-01 -1.01786494e+00 5.65234601e-01
-2.07891583e-01 -1.63070753e-01 2.39116713e-01 1.28541505e+00
-5.93440950e-01 -1.75311005e+00 -2.12046206e-01 6.69714570e-01
-1.04589224e+00 -2.96553373e-01 -6.31150782e-01 9.17071402e-01
4.76181120e-01 7.66181052e-01 2.41059214e-01 9.89618823e-02
1.69539616e-01 2.86018997e-01 3.85798335e-01 -1.04173326e+00
-1.18601978e+00 -1.00830160e-01 8.31696689e-01 -1.62842780e-01
-7.47951448e-01 -1.04935718e+00 -1.29165244e+00 2.86635756e-01
-6.84693515e-01 8.08431208e-01 6.65080726e-01 1.49906051e+00
-5.10922186e-02 3.05829674e-01 3.43412220e-01 -5.87604582e-01
-6.62813783e-01 -1.11665344e+00 -2.66824543e-01 3.83702278e-01
2.50806421e-01 -6.38372719e-01 -3.86748374e-01 -1.12133972e-01] | [10.186649322509766, 9.563204765319824] |
bd3745f5-151c-4d17-81b9-2f7d7206bbd0 | prevention-of-cyberattacks-in-wsn-and-packet | 2306.09448 | null | https://arxiv.org/abs/2306.09448v1 | https://arxiv.org/pdf/2306.09448v1.pdf | Prevention of cyberattacks in WSN and packet drop by CI framework and information processing protocol using AI and Big Data | As the reliance on wireless sensor networks (WSNs) rises in numerous sectors, cyberattack prevention and data transmission integrity become essential problems. This study provides a complete framework to handle these difficulties by integrating a cognitive intelligence (CI) framework, an information processing protocol, and sophisticated artificial intelligence (AI) and big data analytics approaches. The CI architecture is intended to improve WSN security by dynamically reacting to an evolving threat scenario. It employs artificial intelligence algorithms to continuously monitor and analyze network behavior, identifying and mitigating any intrusions in real time. Anomaly detection algorithms are also included in the framework to identify packet drop instances caused by attacks or network congestion. To support the CI architecture, an information processing protocol focusing on efficient and secure data transfer within the WSN is introduced. To protect data integrity and prevent unwanted access, this protocol includes encryption and authentication techniques. Furthermore, it enhances the routing process with the use of AI and big data approaches, providing reliable and timely packet delivery. Extensive simulations and tests are carried out to assess the efficiency of the suggested framework. The findings show that it is capable of detecting and preventing several forms of assaults, including as denial-of-service (DoS) attacks, node compromise, and data tampering. Furthermore, the framework is highly resilient to packet drop occurrences, which improves the WSN's overall reliability and performance | ['Shreyanth S'] | 2023-06-15 | null | null | null | null | ['anomaly-detection'] | ['methodology'] | [ 2.68879116e-01 1.18735790e-01 -5.23952860e-03 -1.15914047e-01
2.89359629e-01 -5.08472979e-01 6.04396880e-01 7.64326394e-01
-5.31193197e-01 6.14709079e-01 -3.17625940e-01 -4.15786028e-01
-4.73644555e-01 -1.52881646e+00 -5.48652187e-02 -9.57524538e-01
-4.90876257e-01 -1.37168150e-02 4.31376666e-01 -2.60777056e-01
4.18868273e-01 9.70018446e-01 -1.53773439e+00 -2.76550531e-01
4.84179199e-01 1.69441605e+00 -7.79567659e-02 4.71910983e-01
-2.05200359e-01 5.97515643e-01 -1.12982893e+00 -8.50462243e-02
3.91042558e-03 3.18669558e-01 -5.48494279e-01 -3.01082969e-01
-9.34481204e-01 -7.40446448e-02 2.74804294e-01 1.10612082e+00
3.66137743e-01 -3.95149082e-01 1.82245031e-01 -1.81536114e+00
-1.98203310e-01 4.47171777e-01 -2.96015918e-01 1.72758400e-01
5.68196356e-01 -2.16059640e-01 -2.71634832e-02 -4.70161974e-01
2.15523884e-01 1.04935062e+00 4.52120841e-01 3.07719260e-01
-6.15793645e-01 -1.01826990e+00 -3.89466248e-02 5.26634455e-01
-1.26192224e+00 -2.85964727e-01 8.60468924e-01 8.68200734e-02
5.10111570e-01 4.96271372e-01 3.16958427e-01 8.00296783e-01
5.87939143e-01 -6.91421237e-03 6.30223930e-01 -6.15207851e-01
1.13243735e+00 3.41389515e-02 2.92727053e-01 1.30757824e-01
7.42251515e-01 4.33265388e-01 1.68404740e-03 -3.27496976e-01
-3.04392762e-02 4.03480321e-01 1.96220979e-01 1.53866217e-01
-9.51574743e-01 5.72592854e-01 4.97584760e-01 7.21953809e-01
-1.07751107e+00 -3.14316332e-01 6.88134730e-01 4.52933311e-01
3.40374224e-02 2.77206898e-02 -3.96864444e-01 2.11267769e-01
-1.34248421e-01 -3.62360209e-01 1.12636113e+00 7.92912185e-01
5.91755271e-01 2.53075063e-01 4.24189121e-01 2.40416840e-01
5.33076227e-01 7.49963999e-01 3.04324746e-01 -8.95911276e-01
-4.54052538e-03 9.83999968e-01 -2.74013937e-01 -1.50848758e+00
-9.52376485e-01 -4.62150052e-02 -1.48511457e+00 6.20044649e-01
-1.42902926e-01 -2.62213558e-01 -6.04938328e-01 1.45085466e+00
6.06733263e-01 -2.40285676e-02 7.27781773e-01 7.18058571e-02
4.59647238e-01 8.23405743e-01 5.89866638e-01 -5.93308091e-01
1.53212965e+00 1.49598762e-01 -1.18735278e+00 4.28831577e-02
2.51114398e-01 -4.82979029e-01 1.75356984e-01 3.76573324e-01
-8.84967327e-01 -4.18803543e-01 -1.24077117e+00 7.46138930e-01
-1.16101205e+00 -7.91977406e-01 4.52716917e-01 1.10143757e+00
-9.77600336e-01 9.27556381e-02 -6.99143469e-01 -5.60305953e-01
1.29292816e-01 6.36398375e-01 -2.97394753e-01 1.56064695e-02
-1.30230010e+00 6.52060568e-01 9.53852475e-01 1.45603105e-01
-6.30297482e-01 -5.30738711e-01 -1.03234684e+00 3.93368630e-03
2.60445774e-01 -2.51348883e-01 5.99405408e-01 -5.14343381e-01
-1.06377268e+00 1.44570813e-01 2.46032685e-01 -6.34267449e-01
-1.74951628e-01 2.42085129e-01 -1.20516419e+00 4.95984077e-01
-1.80056334e-01 -6.20091408e-02 3.89089108e-01 -1.19127417e+00
-7.24332035e-01 -7.52790987e-01 -1.20265134e-01 -5.95017552e-01
-7.17486680e-01 3.84785146e-01 3.03112835e-01 -5.02145648e-01
1.07635163e-01 -5.28830290e-01 -3.83630216e-01 1.17359005e-01
-2.40105927e-01 2.92816795e-02 1.51985824e+00 -1.53283611e-01
1.25981259e+00 -2.06650996e+00 -5.09391487e-01 1.05747986e+00
1.40521303e-01 7.11678505e-01 5.10487556e-02 7.50672579e-01
1.85986862e-01 3.02282065e-01 -2.17557102e-01 2.21658587e-01
-3.86514515e-01 6.77675605e-01 2.92371493e-02 3.07436705e-01
3.71357128e-02 1.01974577e-01 -5.86391091e-01 -2.86580622e-01
4.50238645e-01 6.06342554e-01 3.11867502e-02 2.36715525e-01
1.55906066e-01 2.10377246e-01 -9.52946246e-01 1.04747081e+00
8.29606652e-01 1.34225920e-01 1.53007701e-01 -1.93368912e-01
-3.80280226e-01 -3.60638648e-01 -1.32850909e+00 8.33725452e-01
-3.03630829e-01 -3.49440239e-02 8.19146335e-01 -1.26084793e+00
1.19925189e+00 6.18026555e-01 8.79905045e-01 -1.08311880e+00
5.19491613e-01 2.06871793e-01 -3.34427893e-01 -7.44859636e-01
-2.14716624e-02 2.93849170e-01 -1.94080472e-01 6.12100661e-01
-3.96614045e-01 5.12819529e-01 9.22789276e-02 1.10290863e-01
1.59435582e+00 -7.28042841e-01 5.41979432e-01 -2.49814570e-01
1.14829993e+00 -2.32840180e-01 5.66842079e-01 4.47699010e-01
-2.87588418e-01 -7.04532743e-01 -1.01519085e-01 -5.93547761e-01
-4.94504273e-01 -1.02270019e+00 8.33976269e-02 6.71789825e-01
3.52332562e-01 -3.02550256e-01 -5.89915931e-01 -2.63872176e-01
-2.79182643e-01 6.74897254e-01 -3.65265131e-01 -6.29457474e-01
-2.40622416e-01 -8.69548976e-01 6.50582492e-01 2.08699539e-01
9.01662707e-01 -1.41095424e+00 -8.92194033e-01 5.79233766e-01
1.11682221e-01 -1.19486511e+00 5.56231797e-01 2.47758359e-01
-5.53667128e-01 -1.49509835e+00 2.86899239e-01 -5.09864450e-01
7.57552922e-01 3.84294629e-01 6.76858187e-01 5.59328854e-01
-4.14404809e-01 5.85969925e-01 -8.00606549e-01 -1.01301372e+00
-9.66642261e-01 -4.11751240e-01 2.51010627e-01 -5.34171127e-02
5.19388378e-01 -7.51667500e-01 -4.94274646e-01 4.63616729e-01
-1.55180943e+00 -6.89624250e-01 6.50394440e-01 9.47050229e-02
1.02637924e-01 8.85321856e-01 1.06710589e+00 -4.77876365e-01
8.13138425e-01 -9.43568885e-01 -6.45287812e-01 2.13873252e-01
-8.97787929e-01 -2.07013980e-01 8.46755266e-01 -1.86133862e-01
-7.79290199e-01 -3.48066658e-01 -2.90392041e-01 3.26845556e-01
-6.91706598e-01 3.10963959e-01 -5.32265306e-01 -6.67362332e-01
2.11964399e-01 5.64248748e-02 5.36180198e-01 -2.44389072e-01
-3.24851513e-01 9.51389968e-01 3.86718988e-01 -1.49469599e-01
1.21236241e+00 6.52937293e-01 3.34638089e-01 -1.19822097e+00
-2.85580635e-01 -7.39044622e-02 -3.48631799e-01 -4.70878184e-01
7.48211741e-01 -3.33220363e-01 -1.41472232e+00 6.64349616e-01
-9.95900810e-01 4.27597225e-01 -1.78935919e-02 2.43200228e-01
1.91954166e-01 4.71378535e-01 -3.79960358e-01 -1.26149189e+00
-8.45062613e-01 -8.01478982e-01 3.04754734e-01 4.61452484e-01
1.21632516e-01 -1.03675663e+00 -3.38065177e-01 7.50899687e-02
1.04177332e+00 8.53748381e-01 8.46381903e-01 -9.83919919e-01
-2.45315999e-01 -6.74755156e-01 -2.73723662e-01 2.70099163e-01
4.28360552e-01 2.96883851e-01 -6.09445572e-01 -2.64010400e-01
2.98322439e-01 7.58077353e-02 -7.18938261e-02 -1.36890024e-01
1.07526755e+00 -6.61684394e-01 -3.41176689e-01 2.72862792e-01
1.36412466e+00 9.99368608e-01 8.17779005e-01 5.45631468e-01
7.72056282e-02 9.02775943e-01 6.91715479e-01 7.30337858e-01
6.40173376e-01 5.11538275e-02 1.44332671e+00 1.89485233e-02
5.79349577e-01 3.54718447e-01 3.85441482e-01 3.23076218e-01
1.35392189e-01 -3.47092092e-01 -6.71747386e-01 -3.05718537e-02
-1.29065120e+00 -1.19606042e+00 8.37713033e-02 2.05289030e+00
1.67597950e-01 5.44593513e-01 1.00666493e-01 1.22421944e+00
8.18376422e-01 2.77980156e-02 -2.63791740e-01 -5.48607111e-01
1.29618526e-01 9.43865702e-02 4.99412179e-01 1.09330297e-01
-8.07871044e-01 2.61785150e-01 5.19136429e+00 2.94050455e-01
-1.00213349e+00 -1.24178529e-01 5.46918586e-02 8.18890154e-01
1.02010936e-01 -1.93449587e-01 -3.57896686e-01 6.17547572e-01
1.38468063e+00 7.64521509e-02 1.34663895e-01 6.28132582e-01
6.29281104e-01 -1.18801072e-01 -1.75945610e-01 2.39540517e-01
-2.30539337e-01 -1.13682628e+00 -1.54604688e-01 2.28631273e-01
-2.48805121e-01 -3.33647519e-01 -6.65899456e-01 -1.47208333e-01
1.87124372e-01 -4.82382417e-01 3.81426990e-01 3.50650102e-01
1.16090737e-01 -1.04606438e+00 1.06746960e+00 4.04357791e-01
-1.29048395e+00 -7.38074005e-01 -9.18829516e-02 -2.61853397e-01
2.67516315e-01 5.79137027e-01 -4.69469517e-01 7.91072249e-01
7.65595317e-01 1.43801481e-01 -5.31098247e-01 9.22377110e-01
7.65865743e-02 3.27769428e-01 -6.72993183e-01 -1.94528714e-01
-1.31642539e-03 2.24643610e-02 5.95792413e-01 5.99309027e-01
1.59400105e-01 3.15606564e-01 3.47513467e-01 2.66822696e-01
2.96638012e-01 -1.05975151e-01 -7.29576468e-01 2.81263232e-01
1.12318778e+00 1.27123022e+00 -7.06015706e-01 3.25144380e-02
-1.96943119e-01 3.36606026e-01 -5.02893567e-01 1.21780790e-01
-2.78011471e-01 -8.05563569e-01 7.49229431e-01 -3.12804990e-02
-2.13873670e-01 -2.68798590e-01 -2.52016902e-01 -2.75199026e-01
-7.64876157e-02 -4.90069270e-01 8.86532843e-01 -3.95450294e-01
-1.05059147e+00 7.03395426e-01 4.78620231e-02 -9.55764174e-01
1.73514930e-03 -3.90665710e-01 -9.97830689e-01 -4.78098728e-02
-1.82471645e+00 -1.09651899e+00 -6.75089300e-01 1.38255596e+00
1.19061314e-01 -2.68237263e-01 1.39261901e+00 5.56495190e-01
-7.31164336e-01 3.69429767e-01 -2.68582404e-01 9.00737345e-02
-4.13891003e-02 -5.52906811e-01 -2.93825716e-01 1.24315381e+00
-5.47354996e-01 3.72133613e-01 7.14786530e-01 -6.36643946e-01
-1.67617714e+00 -9.15524900e-01 3.81213129e-01 4.44590747e-01
5.90434611e-01 -1.65117368e-01 -7.83997059e-01 1.48271248e-01
-6.06635176e-02 9.07422826e-02 9.08469796e-01 -6.10818148e-01
-1.84778258e-01 -6.74228191e-01 -1.97375762e+00 2.04706475e-01
3.18040222e-01 1.16067819e-01 -2.02409551e-01 7.53843933e-02
6.88749075e-01 5.45756698e-01 -1.12484777e+00 6.72261059e-01
2.84591496e-01 -8.38074505e-01 1.05887341e+00 -2.85525084e-01
-6.44347250e-01 -4.28901136e-01 -3.14762771e-01 -7.36918807e-01
-4.85127140e-03 -7.49424934e-01 -2.71189153e-01 1.64646339e+00
2.74876673e-02 -9.92209435e-01 4.62115288e-01 7.37627327e-01
5.73583394e-02 -6.19621128e-02 -1.08792377e+00 -6.19556665e-01
-5.86155236e-01 -7.61504471e-01 7.99704731e-01 6.47215784e-01
5.99961206e-02 -1.51677743e-01 9.38702524e-02 9.17631149e-01
1.26170981e+00 -6.46992981e-01 5.12564659e-01 -1.88188696e+00
8.42751980e-01 -1.42207697e-01 -9.56260800e-01 4.49142784e-01
-4.57976246e-03 -1.81703866e-01 -3.45422506e-01 -1.27812004e+00
-7.95891881e-01 -5.75269759e-01 -6.27341390e-01 6.43177807e-01
5.88754654e-01 2.32293010e-01 1.51146322e-01 4.82637696e-02
-3.64533633e-01 2.76138693e-01 3.41164112e-01 1.16502449e-01
-9.80510637e-02 5.89715779e-01 -6.74841225e-01 9.48126912e-01
1.30831265e+00 -5.18277764e-01 -4.64496642e-01 4.04558815e-02
2.68127420e-03 3.36481392e-01 4.81899500e-01 -1.48111105e+00
6.99061692e-01 -2.72496372e-01 1.69685841e-01 -5.67926526e-01
2.83797812e-02 -2.02254295e+00 2.82379150e-01 1.16894710e+00
6.31626099e-02 4.72700566e-01 1.97812885e-01 6.95207357e-01
-1.27116308e-01 1.29455224e-01 1.00352526e+00 3.35434228e-01
-7.70108163e-01 1.15884408e-01 -7.71681190e-01 -5.68306029e-01
1.89219177e+00 -3.86818796e-01 -4.33100373e-01 -1.83204427e-01
-4.76508617e-01 4.53318179e-01 1.69604167e-01 5.00677884e-01
9.53043640e-01 -9.13874805e-01 -1.94891080e-01 6.79727077e-01
1.90033764e-01 -1.63708448e-01 6.03533303e-03 5.35825372e-01
-5.69934487e-01 3.24252918e-02 -5.86820126e-01 -2.27874294e-02
-1.33005583e+00 8.83909047e-01 -4.46020402e-02 -7.40064457e-02
-4.53676790e-01 1.04889207e-01 -8.27848077e-01 -4.91894893e-02
7.72676289e-01 1.91732943e-01 -7.32816041e-01 2.05192678e-02
1.29433763e+00 7.22346961e-01 6.77678064e-02 -4.93622154e-01
-7.71980464e-01 3.47583473e-01 4.63168733e-02 2.04269573e-01
1.28089297e+00 -6.20620251e-01 -4.93260413e-01 -2.30855465e-01
7.80194521e-01 -2.92242706e-01 -3.51311415e-01 -1.39616609e-01
6.97347879e-01 -7.89055601e-02 1.44592822e-01 -8.74231696e-01
-1.31870341e+00 4.14297253e-01 6.67151511e-01 1.16121256e+00
1.71462727e+00 -4.25143629e-01 7.18273818e-01 4.45896894e-01
7.13444948e-01 -8.47538292e-01 -8.85451064e-02 2.48131499e-01
3.51443380e-01 -9.23372686e-01 -1.68894380e-01 -4.32856768e-01
-1.41951725e-01 1.38873541e+00 4.57380176e-01 1.08395584e-01
1.31268060e+00 9.96134996e-01 5.14129689e-03 -4.86300856e-01
-4.34567302e-01 1.82878882e-01 -7.49790490e-01 1.12759161e+00
-3.93144041e-01 -2.08831295e-01 -2.05153763e-01 3.45099926e-01
1.67871937e-01 -1.00750461e-01 8.42718780e-01 1.46918380e+00
-1.06402230e+00 -1.25701487e+00 -8.14115405e-01 4.39441204e-02
-8.09735835e-01 4.37874943e-01 -2.69569486e-01 7.11252093e-01
-2.25442667e-02 1.69873154e+00 -7.79064968e-02 -6.99805319e-01
2.34511316e-01 -3.05215746e-01 -5.98865688e-01 -3.52640562e-02
-5.31898081e-01 -4.91985619e-01 -1.57458588e-01 -7.07275391e-01
-7.80845463e-01 -3.88781689e-02 -1.53520226e+00 -6.89887166e-01
-7.62113631e-02 5.37801564e-01 1.49780405e+00 9.47608054e-01
4.75986123e-01 5.85007787e-01 1.12643576e+00 -4.65447307e-01
-2.53084004e-01 -5.75275660e-01 -5.27694404e-01 -1.99167375e-02
4.90452141e-01 -6.24416053e-01 -4.71190393e-01 -3.68494332e-01] | [5.210951805114746, 7.12512731552124] |
e04d4631-6d93-4e38-84df-548a54ef0416 | synthetic-alone-exploring-the-dark-side-of | 2306.14377 | null | https://arxiv.org/abs/2306.14377v1 | https://arxiv.org/pdf/2306.14377v1.pdf | Synthetic Alone: Exploring the Dark Side of Synthetic Data for Grammatical Error Correction | Data-centric AI approach aims to enhance the model performance without modifying the model and has been shown to impact model performance positively. While recent attention has been given to data-centric AI based on synthetic data, due to its potential for performance improvement, data-centric AI has long been exclusively validated using real-world data and publicly available benchmark datasets. In respect of this, data-centric AI still highly depends on real-world data, and the verification of models using synthetic data has not yet been thoroughly carried out. Given the challenges above, we ask the question: Does data quality control (noise injection and balanced data), a data-centric AI methodology acclaimed to have a positive impact, exhibit the same positive impact in models trained solely with synthetic data? To address this question, we conducted comparative analyses between models trained on synthetic and real-world data based on grammatical error correction (GEC) task. Our experimental results reveal that the data quality control method has a positive impact on models trained with real-world data, as previously reported in existing studies, while a negative impact is observed in models trained solely on synthetic data. | ['Heuiseok Lim', 'Hyeonseok Moon', 'Sugyeong Eo', 'Jaehyung Seo', 'Seolhwa Lee', 'Seonmin Koo', 'Chanjun Park'] | 2023-06-26 | null | null | null | null | ['grammatical-error-correction'] | ['natural-language-processing'] | [ 4.35198575e-01 2.49994203e-01 1.40346631e-01 -3.64054054e-01
-5.93235493e-01 -3.09972137e-01 8.18280995e-01 4.95627671e-01
-8.37882638e-01 6.85932934e-01 -1.66168734e-02 -1.63223878e-01
-9.23828036e-02 -8.44898999e-01 -1.14439857e+00 -3.32937062e-01
1.31280273e-01 5.53348958e-01 -4.08094227e-02 -3.78184408e-01
3.63740563e-01 2.09378719e-01 -1.89951098e+00 2.99826235e-01
1.38658917e+00 6.50414288e-01 -2.64741983e-02 4.51296389e-01
1.64494872e-01 5.15082121e-01 -9.61360335e-01 -8.12648296e-01
2.70413637e-01 -3.74206156e-01 -5.24419248e-01 -1.99848101e-01
4.31862712e-01 1.69771075e-01 7.66688213e-02 1.18734074e+00
6.02372587e-01 -5.47861814e-01 4.36471224e-01 -1.45183909e+00
-7.99609780e-01 1.02826440e+00 9.59150419e-02 7.01968446e-02
2.25836858e-01 6.39481425e-01 9.91969049e-01 -9.01268899e-01
9.56185997e-01 1.10159492e+00 5.82946897e-01 5.97787678e-01
-1.61506629e+00 -6.08651161e-01 4.52480800e-02 7.80385286e-02
-1.16531527e+00 -2.92480022e-01 8.78572047e-01 -2.93669164e-01
9.15875375e-01 2.99876869e-01 8.10781896e-01 1.45145154e+00
2.32079905e-02 9.36212599e-01 1.16663170e+00 -7.01929986e-01
3.06357324e-01 3.50594878e-01 9.64820087e-02 3.15750897e-01
7.71768451e-01 5.14489710e-01 -6.08735144e-01 2.65511423e-01
-9.22903866e-02 -6.76196814e-01 -3.08855295e-01 -3.71285170e-01
-1.22419810e+00 6.51900113e-01 3.71893793e-01 5.42897105e-01
-4.77126926e-01 9.51142833e-02 5.61544836e-01 5.55061281e-01
5.40240049e-01 1.27393055e+00 -7.44042695e-01 -7.11085916e-01
-9.28777575e-01 7.26156175e-01 6.38335168e-01 9.37635124e-01
5.02975464e-01 4.19763088e-01 -2.12597132e-01 7.20026791e-01
2.04661593e-01 6.80520713e-01 7.02539563e-01 -4.09892708e-01
6.28719628e-01 1.45352173e+00 -1.52643800e-01 -1.18715477e+00
-3.46632093e-01 -6.77952290e-01 -6.79639459e-01 2.09421486e-01
5.95769107e-01 -6.03630058e-02 -6.81965470e-01 2.00260878e+00
3.83267412e-03 -3.65421563e-01 3.16844016e-01 7.90751457e-01
7.66195416e-01 1.91940859e-01 1.01744503e-01 1.23770982e-01
8.70992959e-01 -5.09873033e-01 -7.62130082e-01 -1.89916179e-01
1.03684700e+00 -4.74518746e-01 1.61243594e+00 6.85854793e-01
-8.51829767e-01 -5.45939744e-01 -1.26764703e+00 6.88410923e-02
-6.94217503e-01 1.83248580e-01 2.23376438e-01 9.46618080e-01
-7.97929585e-01 5.53081393e-01 -6.59469306e-01 -3.72276217e-01
3.67089540e-01 1.36943996e-01 -5.59773684e-01 -2.21302882e-01
-1.32912338e+00 1.17231965e+00 6.94410503e-01 3.84363197e-02
-8.68753314e-01 -9.69921410e-01 -7.70768046e-01 -1.12631150e-01
4.38056469e-01 -3.83089662e-01 1.05049431e+00 -1.09080184e+00
-1.09963298e+00 8.11102867e-01 2.32646763e-01 -8.40454876e-01
7.29737759e-01 -2.23510459e-01 -7.07927346e-01 -5.34784913e-01
-1.77104160e-01 7.11745977e-01 3.83407265e-01 -1.49033153e+00
-3.44813764e-01 -4.89537895e-01 -2.50207204e-02 -3.34252477e-01
-3.73292208e-01 -1.45134315e-01 -7.53389150e-02 -8.58195662e-01
-1.68357000e-01 -1.11387503e+00 2.09155545e-01 -2.89365739e-01
-4.00572300e-01 -7.56505579e-02 5.59243798e-01 -5.86233735e-01
1.50615239e+00 -1.93493187e+00 2.07285911e-01 1.11625828e-01
-9.37813446e-02 7.93196857e-01 -2.80541986e-01 4.21853215e-01
-7.75157288e-02 5.88519156e-01 -6.19155109e-01 -3.21046233e-01
1.42093375e-01 1.63887367e-01 1.51866049e-01 6.56373501e-02
7.19510734e-01 9.22000706e-01 -8.16520154e-01 -3.18789810e-01
-1.31343871e-01 2.51373917e-01 -9.32101786e-01 1.56196848e-01
-4.90549415e-01 2.85335064e-01 2.18393393e-02 6.03387475e-01
6.04681790e-01 6.54523820e-02 -1.02329962e-02 -6.59144372e-02
2.34670099e-02 1.26579240e-01 -1.02938342e+00 1.52965581e+00
-2.44043633e-01 7.26092517e-01 -4.61608887e-01 -7.45711505e-01
1.13213003e+00 2.72679508e-01 -2.31121518e-02 -1.36769557e+00
8.17301497e-02 5.06985307e-01 7.41492450e-01 -5.65429091e-01
7.55350769e-01 8.05123448e-02 -9.63465571e-02 2.96499938e-01
-7.98454043e-03 -3.26461643e-01 5.98171711e-01 1.86071873e-01
1.27275932e+00 1.99361473e-01 3.66758257e-02 -1.90149158e-01
4.10650760e-01 5.26813388e-01 7.28391409e-01 6.08938038e-01
-1.56818539e-01 6.18061364e-01 4.55429405e-01 -1.21518575e-01
-1.21457136e+00 -5.19533694e-01 -1.38870269e-01 5.75860679e-01
-9.59107280e-02 -7.29808271e-01 -1.01084280e+00 -8.94380808e-01
8.09261501e-02 1.22334421e+00 -8.81839812e-01 -6.16957605e-01
-4.45856363e-01 -8.71508062e-01 9.64524031e-01 2.91002512e-01
4.57000345e-01 -1.34563863e+00 -8.88666689e-01 2.66411453e-01
-1.36424629e-02 -9.20574486e-01 9.24920365e-02 4.66201790e-02
-8.71938288e-01 -1.29111433e+00 -3.11245531e-01 -1.09196313e-01
6.95385575e-01 -3.97278875e-01 1.44943583e+00 4.11603093e-01
-1.45898744e-01 1.23485059e-01 -8.60618055e-01 -8.82685602e-01
-1.11707699e+00 2.80196488e-01 -6.90330938e-02 -8.59606490e-02
5.42801917e-01 -2.47844294e-01 -1.16409659e-01 2.83299059e-01
-1.45402992e+00 1.63176015e-01 6.54415429e-01 9.00762200e-01
2.39592254e-01 -2.34842017e-01 7.80370295e-01 -1.29540741e+00
8.03614855e-01 -3.18393111e-01 -5.02353013e-01 3.85308653e-01
-1.14731288e+00 3.90420079e-01 5.74456275e-01 -3.81710351e-01
-8.31027389e-01 -2.17281327e-01 -2.14087114e-01 -1.17759883e-01
-3.15466374e-01 8.37015748e-01 -4.54538733e-01 3.17873865e-01
1.01422346e+00 9.33379233e-02 1.55855611e-01 -4.27891582e-01
2.44149655e-01 6.38368249e-01 2.81501144e-01 -6.91911221e-01
7.14578688e-01 1.43539235e-02 -3.06403637e-01 -4.64864403e-01
-4.03084487e-01 3.22930574e-01 -6.25880480e-01 -1.18595213e-01
5.26331186e-01 -6.48240268e-01 -2.97916740e-01 5.52613974e-01
-1.12332690e+00 -2.89156586e-01 -3.74906957e-01 4.18456048e-01
-3.66359621e-01 -6.25792146e-02 4.41741608e-02 -6.10561669e-01
-2.86593139e-01 -1.32045627e+00 6.88547671e-01 -3.18236947e-02
-5.39729238e-01 -8.40994716e-01 1.57783613e-01 3.46803188e-01
6.59961045e-01 5.17639399e-01 1.32649755e+00 -1.03413355e+00
-3.76351655e-01 -5.14163852e-01 -1.71077661e-02 5.57334006e-01
-4.39803377e-02 3.36591303e-01 -1.01663327e+00 -1.74632162e-01
-1.93763673e-01 -3.24718267e-01 3.55167776e-01 -2.02288076e-01
9.78359222e-01 -2.03280613e-01 9.74061489e-02 8.28420445e-02
1.28309953e+00 1.95008844e-01 6.41494513e-01 4.83318001e-01
4.97313291e-01 6.17625535e-01 9.34996545e-01 2.36981913e-01
3.47477406e-01 8.61121774e-01 7.65373111e-01 1.22905485e-01
-2.24747658e-01 -4.77434576e-01 2.72982866e-01 6.70687795e-01
2.53893644e-01 -5.32352567e-01 -1.26254857e+00 6.81185842e-01
-1.63764441e+00 -5.34184515e-01 -3.49359006e-01 2.29587293e+00
9.66834247e-01 4.37881231e-01 2.64975205e-02 7.15049505e-01
3.38809401e-01 -2.70911843e-01 -5.66546440e-01 -5.95529616e-01
-5.62001824e-01 -1.37375295e-01 1.50327638e-01 4.07358445e-02
-6.07149422e-01 7.83325732e-01 5.50828791e+00 6.02322757e-01
-1.32068419e+00 -3.33296172e-02 5.50358951e-01 -1.12248130e-01
-7.55890071e-01 3.37145142e-02 -5.93098462e-01 7.26574242e-01
1.14135754e+00 -1.73358455e-01 2.22260579e-01 8.07299018e-01
3.32587570e-01 -2.17398465e-01 -1.50604713e+00 7.64861643e-01
1.35887563e-01 -1.11608005e+00 2.38815859e-01 3.42547029e-01
6.05361819e-01 -1.68202460e-01 -4.98406440e-02 6.92911625e-01
2.48386592e-01 -1.17536104e+00 9.74454939e-01 5.03467321e-01
4.87530500e-01 -7.30975628e-01 9.48363960e-01 5.77735782e-01
-2.16462255e-01 1.12590119e-02 -1.54915929e-01 -3.20583694e-02
-1.85173959e-01 7.25808561e-01 -1.15166247e+00 5.59107900e-01
5.90175688e-01 5.22184253e-01 -1.33198118e+00 8.56591821e-01
-1.92520484e-01 9.99730885e-01 -4.94296551e-02 -5.70610166e-02
-1.24277331e-01 1.23292036e-01 6.29757822e-01 1.13917732e+00
1.55926421e-01 -3.93596560e-01 -5.48564136e-01 1.06951129e+00
-8.83869380e-02 2.82914430e-01 -8.53084445e-01 -4.32344913e-01
5.69192171e-01 8.13807428e-01 -2.10630909e-01 -1.99004844e-01
-2.40853950e-01 5.81124485e-01 3.80989373e-01 -4.66625057e-02
-7.08553612e-01 -2.82834500e-01 2.78625816e-01 3.50704491e-01
-1.81217015e-01 1.12159729e-01 -6.94931448e-01 -9.31674004e-01
1.92303017e-01 -1.59995902e+00 1.54835150e-01 -5.68827510e-01
-1.15955508e+00 8.22495103e-01 -1.45673349e-01 -1.24701416e+00
-2.52846718e-01 -4.31294739e-01 -2.90908873e-01 6.10059261e-01
-1.22806489e+00 -1.02402604e+00 -3.06159645e-01 8.52253214e-02
3.76460403e-01 -4.22368616e-01 7.84352064e-01 4.72311467e-01
-7.20335245e-01 1.10858238e+00 7.52447499e-03 1.83975175e-01
6.88867331e-01 -1.22716641e+00 6.37179434e-01 1.10588872e+00
1.65481672e-01 6.87306643e-01 1.11648965e+00 -7.21203148e-01
-1.62872577e+00 -1.02671337e+00 1.01853156e+00 -6.72128737e-01
2.21729770e-01 -4.21544880e-01 -1.22857034e+00 3.34101200e-01
-4.07115929e-02 -1.12631634e-01 4.68587935e-01 9.66623053e-02
-4.27704185e-01 -1.54341802e-01 -1.35441709e+00 4.83486772e-01
9.24013436e-01 -1.75919369e-01 -6.42673194e-01 -2.00253949e-01
7.45593429e-01 -2.80526459e-01 -9.46229875e-01 8.40751588e-01
3.07229578e-01 -8.80717516e-01 4.14320201e-01 -8.53551090e-01
7.38341451e-01 -2.05356061e-01 -3.42410594e-01 -1.77043045e+00
3.54266673e-01 4.42077452e-03 2.10616782e-01 1.64140224e+00
1.01021683e+00 -4.57694530e-01 6.90235436e-01 1.00708997e+00
-2.90164471e-01 -5.88729799e-01 -8.61721218e-01 -9.24033284e-01
2.07061440e-01 -7.89477646e-01 8.39912355e-01 1.00342321e+00
-2.24189818e-01 3.07445624e-03 -2.83108026e-01 -2.87082009e-02
1.06153630e-01 -2.16444165e-01 1.32564676e+00 -1.35247231e+00
-5.13489917e-03 -3.64016533e-01 -4.78536963e-01 -1.79818302e-01
4.48332019e-02 -9.99169409e-01 6.60002157e-02 -1.28848648e+00
-1.41065359e-01 -7.13810980e-01 -3.82198729e-02 3.86782765e-01
-3.29670638e-01 1.62278920e-01 4.93893504e-01 -6.66355118e-02
-3.31774682e-01 8.97211492e-01 1.02713692e+00 -8.40553641e-02
-5.32156192e-02 -3.13069165e-01 -7.26233304e-01 3.17021608e-01
5.51004589e-01 -7.98362792e-01 -2.89086461e-01 -5.81858635e-01
7.48669505e-01 -2.99522817e-01 1.52575672e-01 -1.31002080e+00
2.01584883e-02 6.62454292e-02 -8.89892951e-02 -1.59356594e-01
1.57862633e-01 -9.83689010e-01 2.14866892e-01 6.78822339e-01
-5.78006685e-01 4.56968874e-01 4.30385828e-01 2.85090923e-01
-4.57983345e-01 -2.82053381e-01 6.16104424e-01 2.57473201e-01
-4.91008639e-01 -2.48063296e-01 -1.11927740e-01 3.51068646e-01
9.05180991e-01 -3.87932241e-01 -4.22720462e-01 -4.66975197e-02
-3.96444350e-01 1.16549879e-01 6.82076037e-01 9.44488585e-01
3.06036890e-01 -1.12753177e+00 -1.06188989e+00 4.04213756e-01
7.58278072e-01 -1.17634423e-01 4.14503776e-02 6.04894638e-01
-3.81249279e-01 4.77132916e-01 -2.29131788e-01 -6.23591721e-01
-1.27356791e+00 4.97215629e-01 2.17484027e-01 -2.12674886e-01
-9.66492444e-02 5.45224965e-01 -5.11795819e-01 -9.66704607e-01
3.07049721e-01 -5.51647782e-01 -1.60027042e-01 1.59354970e-01
5.31072319e-02 3.34283769e-01 6.03944480e-01 -5.15397787e-01
-2.48694241e-01 -2.32423786e-02 -1.19957075e-01 -2.21466143e-02
1.45502985e+00 3.12588125e-01 2.33841851e-01 6.05737746e-01
7.40999937e-01 -3.60472091e-02 -8.43148887e-01 -1.97181955e-01
4.62235868e-01 -5.98745942e-01 -1.91029370e-01 -1.47324085e+00
-1.00328708e+00 9.10638094e-01 7.60115743e-01 2.09669277e-01
9.65036571e-01 -5.58982968e-01 2.33792856e-01 2.21721530e-01
4.78416741e-01 -1.15187323e+00 1.37350172e-01 4.81137991e-01
1.11912215e+00 -1.58696651e+00 -3.79700869e-01 -2.11559281e-01
-9.25151169e-01 4.93288070e-01 9.97477531e-01 1.30552083e-01
4.20273781e-01 9.73717868e-02 2.12393515e-02 -1.20135389e-01
-9.80257094e-01 4.24008071e-02 1.64314494e-01 5.18186450e-01
6.40763700e-01 9.12415702e-03 -6.14571035e-01 7.47653961e-01
-4.40156817e-01 2.47759536e-01 6.70867443e-01 9.55110133e-01
6.64102472e-03 -1.53446615e+00 -2.25030884e-01 4.19813514e-01
-3.15752089e-01 -6.46754950e-02 -7.26402700e-01 1.26458013e+00
2.91206270e-01 1.01405549e+00 9.24772304e-03 -5.25918245e-01
8.04315627e-01 3.16255778e-01 1.66427284e-01 -6.14013135e-01
-1.15485334e+00 -7.64515996e-01 2.80794054e-01 -2.92079002e-01
-2.15706527e-01 -5.98947406e-01 -9.55248237e-01 -1.54038966e-01
-2.91826487e-01 1.55487165e-01 1.28649092e+00 9.54363644e-01
7.83090413e-01 5.88806689e-01 1.75737634e-01 -1.26958564e-01
-5.15882075e-01 -1.36385763e+00 -1.24705471e-01 8.71247768e-01
4.19406779e-02 -5.00433505e-01 -2.58680403e-01 -2.17499331e-01] | [11.23824691772461, 9.085269927978516] |
4aab54d5-4a52-47b0-b65a-4bdfafee247f | utility-assessment-of-synthetic-data | 2211.14428 | null | https://arxiv.org/abs/2211.14428v1 | https://arxiv.org/pdf/2211.14428v1.pdf | Utility Assessment of Synthetic Data Generation Methods | Big data analysis poses the dual problem of privacy preservation and utility, i.e., how accurate data analyses remain after transforming original data in order to protect the privacy of the individuals that the data is about - and whether they are accurate enough to be meaningful. In this paper, we thus investigate across several datasets whether different methods of generating fully synthetic data vary in their utility a priori (when the specific analyses to be performed on the data are not known yet), how closely their results conform to analyses on original data a posteriori, and whether these two effects are correlated. We find some methods (decision-tree based) to perform better than others across the board, sizeable effects of some choices of imputation parameters (notably the number of released datasets), no correlation between broad utility metrics and analysis accuracy, and varying correlations for narrow metrics. We did get promising findings for classification tasks when using synthetic data for training machine learning models, which we consider worth exploring further also in terms of mitigating privacy attacks against ML models such as membership inference and model inversion. | ['Sonja Buchegger', 'Niklas Reje', 'Md Sakib Nizam Khan'] | 2022-11-23 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [ 3.20385128e-01 3.75117689e-01 -2.48502456e-02 -5.70881605e-01
-8.53721201e-01 -9.78469491e-01 6.47269070e-01 5.56112409e-01
-6.28294170e-01 1.06548524e+00 2.70176858e-01 -3.27298790e-01
-4.46765661e-01 -9.24628437e-01 -9.90747452e-01 -6.36324108e-01
-4.88762259e-02 4.44547653e-01 -2.42783561e-01 1.45342469e-01
2.03761131e-01 2.34635144e-01 -1.19283807e+00 1.84709758e-01
8.55678201e-01 6.53273523e-01 -9.64272380e-01 4.41786706e-01
4.64888394e-01 6.58128560e-01 -6.63235068e-01 -9.45756674e-01
4.93986726e-01 -2.30316684e-01 -6.51839554e-01 -2.13729709e-01
2.32105166e-01 -3.28801215e-01 1.74520463e-01 1.10845256e+00
4.13742930e-01 -3.09201986e-01 5.39679587e-01 -1.30679584e+00
-4.89958882e-01 7.13728666e-01 -3.54058892e-01 -2.59405971e-01
4.63458091e-01 2.55710244e-01 7.01801956e-01 -1.39682204e-01
6.13741636e-01 1.04006934e+00 9.33229804e-01 1.68638662e-01
-1.72788596e+00 -8.01066518e-01 -3.87145162e-01 -1.99842289e-01
-1.38177860e+00 -8.90924215e-01 1.42030194e-01 -5.02602458e-01
1.94503248e-01 6.95823729e-01 1.58584058e-01 1.17048538e+00
1.32393256e-01 -1.33888155e-01 1.22595990e+00 -2.64786810e-01
5.51980734e-01 6.12235129e-01 3.50995481e-01 1.15752950e-01
1.01538074e+00 1.13853857e-01 -3.47859710e-01 -8.11376810e-01
8.95928591e-02 -3.00314128e-01 -2.32930034e-01 -5.34018338e-01
-1.08693993e+00 8.63238990e-01 -1.11560263e-01 8.98643136e-02
-2.90010214e-01 -4.48332019e-02 4.31350917e-01 3.91006261e-01
3.15565348e-01 6.20622516e-01 -5.43233037e-01 7.63886434e-04
-8.56728435e-01 5.79893529e-01 9.02985394e-01 7.79253900e-01
6.97329223e-01 -3.01397681e-01 -2.43612692e-01 3.96205992e-01
-6.52064383e-02 3.58794302e-01 1.36135712e-01 -1.06226993e+00
7.52397537e-01 3.83889318e-01 5.21972477e-01 -1.19002604e+00
-2.60019988e-01 -1.96851403e-01 -8.31401527e-01 2.29671195e-01
1.10989189e+00 -5.43716490e-01 -2.67298102e-01 2.09294844e+00
1.46641955e-01 -3.19135398e-01 1.61686674e-01 6.09289885e-01
3.66323113e-01 3.72435749e-01 1.80650458e-01 -1.03616476e-01
1.38438892e+00 3.65113094e-02 -6.41571105e-01 -1.92180037e-01
7.97698975e-01 -4.90706146e-01 9.32970881e-01 4.99271482e-01
-1.08057761e+00 -2.63272434e-01 -1.00902236e+00 -8.07157233e-02
-4.01066035e-01 -1.31672278e-01 6.35063052e-01 1.31251979e+00
-7.27995098e-01 6.73408449e-01 -5.75236261e-01 -3.80721480e-01
6.46478295e-01 4.22188759e-01 -7.11060703e-01 1.22604948e-02
-1.18994009e+00 7.07104325e-01 2.47359633e-01 1.71029456e-02
-3.84014726e-01 -8.15570652e-01 -6.38865948e-01 3.22775662e-01
2.58231550e-01 -5.97970366e-01 6.81521118e-01 -7.65818715e-01
-6.36678874e-01 9.42000985e-01 1.02449946e-01 -7.66266465e-01
9.58293498e-01 5.75226322e-02 -3.88976753e-01 -2.74238735e-01
2.10532710e-01 3.87056410e-01 5.16189456e-01 -1.17883861e+00
-5.65084696e-01 -9.09415781e-01 -1.58901643e-02 -2.73071289e-01
-1.70187116e-01 1.92163456e-02 3.31121087e-02 -5.82877576e-01
-1.31919324e-01 -1.04698741e+00 -2.13171080e-01 8.62772241e-02
-5.94831765e-01 3.55713248e-01 4.04664278e-01 -8.19173336e-01
1.02893269e+00 -2.19737554e+00 -2.70013392e-01 5.52684069e-01
-7.34960362e-02 1.66193575e-01 2.81922016e-02 2.50569314e-01
-6.96009248e-02 5.73705912e-01 -3.35565478e-01 -1.24954835e-01
9.94905084e-02 7.90777281e-02 -3.97271782e-01 7.81523585e-01
7.01174811e-02 7.01388478e-01 -2.55618513e-01 -1.65530980e-01
-1.10457592e-01 3.84945482e-01 -6.32174194e-01 -4.05971631e-02
-2.12887302e-03 6.96804345e-01 -3.25457960e-01 3.39647174e-01
9.85554457e-01 4.51799482e-02 2.97237247e-01 1.57511503e-01
2.82715186e-02 1.70859724e-01 -1.28827393e+00 1.01860905e+00
-8.58878270e-02 5.27843833e-01 2.63159394e-01 -8.14098239e-01
8.14656317e-01 1.50429413e-01 1.16085999e-01 -4.17736292e-01
-8.21398944e-02 4.27358299e-02 1.24793448e-01 -3.69142473e-01
3.76692861e-01 -1.18614964e-01 -2.19874471e-01 6.21810496e-01
-4.73598540e-01 2.27965429e-01 -4.22517471e-02 1.40508572e-02
1.02775228e+00 -2.85600364e-01 1.38220266e-01 -3.61861616e-01
2.40346760e-01 -8.47343169e-03 6.34980738e-01 8.59936476e-01
-3.92592661e-02 7.86300600e-01 1.07168210e+00 -2.34440893e-01
-1.10648274e+00 -8.36690068e-01 -5.35903871e-01 5.66361487e-01
-1.38073623e-01 -3.59997243e-01 -8.07778299e-01 -5.55056632e-01
3.28376949e-01 1.12987196e+00 -7.96445131e-01 -1.37196705e-01
-3.27624904e-04 -9.81133819e-01 8.85026217e-01 1.26613840e-01
2.48288438e-01 -6.93934023e-01 -7.51087189e-01 -1.94952175e-01
-1.14733934e-01 -1.00164270e+00 -1.20749377e-01 7.09797069e-02
-7.18804717e-01 -1.03771651e+00 -2.13723764e-01 5.61716370e-02
7.13789701e-01 -3.07846993e-01 7.64407456e-01 1.01314962e-01
1.72806699e-02 3.89474183e-02 -1.64902881e-01 -5.58699191e-01
-6.14008784e-01 2.79081408e-02 -1.86501935e-01 1.94016829e-01
4.13450539e-01 -6.12715483e-01 -1.46523446e-01 1.94006801e-01
-1.08025539e+00 -3.92389238e-01 3.33425969e-01 4.07363862e-01
2.12661639e-01 3.11136663e-01 3.72398973e-01 -1.61751652e+00
6.74742520e-01 -5.49107432e-01 -5.46232462e-01 5.14109313e-01
-7.55600750e-01 1.74739271e-01 5.71804941e-01 -1.83527052e-01
-8.30294251e-01 -2.61304259e-01 -2.73264814e-02 7.03078955e-02
-3.45289648e-01 2.44732797e-01 -4.20439959e-01 -5.73725775e-02
6.16069674e-01 -1.44876152e-01 3.12040925e-01 -4.35919672e-01
1.28151819e-01 6.03456318e-01 3.10722500e-01 -6.55050516e-01
7.48060584e-01 5.13777018e-01 1.08253434e-01 -5.73828340e-01
-4.87477899e-01 2.00958684e-01 -5.51177084e-01 4.64397162e-01
7.73616076e-01 -7.33692586e-01 -9.63182926e-01 4.51348245e-01
-7.15292633e-01 -1.53085813e-01 -4.20936078e-01 4.48433667e-01
-2.71195441e-01 5.10653615e-01 -2.49859840e-01 -8.07143688e-01
-1.31490141e-01 -1.21436059e+00 6.85015321e-01 -1.07545733e-01
-7.26682723e-01 -8.68763983e-01 -1.11796185e-01 6.44129872e-01
3.85541320e-01 7.41149426e-01 1.09227073e+00 -8.59548986e-01
-4.52273816e-01 -6.06002569e-01 -1.69525132e-01 1.86362296e-01
9.72419009e-02 1.03702940e-01 -1.19226766e+00 -4.76574749e-01
3.85855287e-02 -2.20672458e-01 6.28773749e-01 4.05809075e-01
1.21619880e+00 -9.60393488e-01 -1.72412142e-01 7.36896932e-01
1.19093883e+00 -2.38321647e-01 1.03460014e+00 2.58024454e-01
2.95732021e-01 1.08506370e+00 3.51394325e-01 5.18573105e-01
2.40870476e-01 7.37006128e-01 1.72118053e-01 1.29081905e-01
5.58772922e-01 -3.33514005e-01 2.34104946e-01 -1.08337834e-01
2.82246113e-01 -1.95981741e-01 -6.49769366e-01 3.26810569e-01
-1.72472656e+00 -1.15118873e+00 -3.13326389e-01 2.74089599e+00
7.25222468e-01 2.88929254e-01 3.43957931e-01 3.31308655e-02
6.78795636e-01 1.44849062e-01 -4.72400010e-01 -6.96795106e-01
-2.82777369e-01 -3.41426916e-02 8.79918635e-01 4.00482833e-01
-8.44582856e-01 1.92547977e-01 6.21492958e+00 4.43596631e-01
-8.15787733e-01 1.11972531e-02 1.11263192e+00 -1.81171224e-01
-6.79450810e-01 5.29923975e-01 -5.23647428e-01 7.49210417e-01
1.17551160e+00 -2.58756131e-01 4.22318101e-01 6.33045077e-01
1.55242130e-01 -2.37250417e-01 -1.30726433e+00 5.26047885e-01
-2.52592206e-01 -1.24465561e+00 -6.41495585e-02 5.77326715e-01
4.19824868e-01 -4.16725844e-01 1.30297244e-01 1.09708920e-01
2.70770073e-01 -1.23609221e+00 8.25242579e-01 5.80636919e-01
6.64058268e-01 -7.69279242e-01 8.77013862e-01 5.22584677e-01
-2.60163456e-01 -1.18442118e-01 -4.02852297e-01 -6.99133649e-02
-1.57387212e-01 7.38272488e-01 -6.72947049e-01 4.70907301e-01
6.67869031e-01 1.61273941e-01 -9.23120081e-01 5.92025399e-01
-3.60384732e-02 7.53785253e-01 -3.55765790e-01 1.80992573e-01
-1.94793314e-01 -4.96144801e-01 4.21697199e-01 8.75658631e-01
1.86657608e-01 -5.93382008e-02 -3.96411836e-01 1.10910583e+00
-6.16968051e-02 -1.07214898e-01 -5.90599954e-01 -7.53184110e-02
6.56659842e-01 1.02687967e+00 -3.13227832e-01 8.18164274e-02
-2.62197793e-01 3.56267869e-01 7.37406537e-02 3.14861417e-01
-5.35713434e-01 -1.89944312e-01 8.52722168e-01 5.48271000e-01
4.32170220e-02 1.79465026e-01 -7.62071073e-01 -1.10421944e+00
3.96170527e-01 -1.03810048e+00 7.37997830e-01 -6.88141465e-01
-1.24736285e+00 7.53646716e-02 -7.93858469e-02 -8.62244666e-01
-3.97577912e-01 -9.04777125e-02 -5.29225409e-01 9.55277681e-01
-1.04241157e+00 -1.00525343e+00 7.89563209e-02 3.17881912e-01
-4.58447963e-01 6.21811673e-03 9.49463546e-01 -6.14293898e-03
-4.75772917e-01 1.09085500e+00 3.99354160e-01 2.64538944e-01
7.84523845e-01 -8.79322767e-01 3.22200269e-01 9.78913903e-01
1.86438411e-02 9.58883286e-01 8.80564451e-01 -7.42184520e-01
-1.32417095e+00 -9.40451741e-01 8.88597667e-01 -8.63284528e-01
2.25375533e-01 -5.95789731e-01 -9.96710718e-01 9.95789170e-01
-1.59868017e-01 -1.17955156e-01 7.87512362e-01 4.47716773e-01
-5.06370664e-01 -2.07929119e-01 -1.81675696e+00 3.93688291e-01
5.64248681e-01 -3.02902281e-01 -1.60011292e-01 6.54709265e-02
5.75728953e-01 2.55249697e-03 -1.04419744e+00 2.91495562e-01
7.57196426e-01 -1.32582748e+00 8.90260100e-01 -9.14308190e-01
1.56870767e-01 -1.00921281e-01 -4.17477310e-01 -9.04711723e-01
-2.09742546e-01 -4.86489594e-01 3.00340861e-01 1.84630656e+00
7.21805751e-01 -1.08985531e+00 8.74117017e-01 1.53339398e+00
7.55913258e-01 -3.60486180e-01 -9.19958472e-01 -5.28389812e-01
2.91564405e-01 -3.63786161e-01 1.10317600e+00 1.26046085e+00
-2.19860151e-01 -5.06157130e-02 -4.01299566e-01 4.72836405e-01
9.31399405e-01 5.53504862e-02 1.22267771e+00 -1.19594133e+00
-1.81938916e-01 8.83538090e-03 -4.17787284e-01 2.71390639e-02
1.30827427e-01 -7.64369428e-01 -6.75925553e-01 -9.35242116e-01
2.58792311e-01 -6.99462771e-01 1.20029710e-02 5.82251191e-01
-1.82690874e-01 3.32226381e-02 1.61494732e-01 1.69939294e-01
-1.35452285e-01 4.87654544e-02 5.03373563e-01 1.06618643e-01
-1.79992110e-01 3.62398684e-01 -1.30102825e+00 3.80892336e-01
7.73579657e-01 -6.97889447e-01 -2.63236642e-01 -3.44201088e-01
3.39987606e-01 2.26072699e-01 6.95513964e-01 -1.04382980e+00
1.89840734e-01 -8.42548683e-02 7.06473291e-01 -2.12860629e-01
1.07620917e-01 -9.85454857e-01 8.39252174e-01 4.09508765e-01
-7.69606471e-01 -2.04508901e-01 1.14449196e-01 4.64417994e-01
1.78424671e-01 -1.77418023e-01 7.07399487e-01 8.94590020e-02
-4.92190793e-02 1.39378369e-01 -1.05450623e-01 3.84315491e-01
1.08636034e+00 -2.71035075e-01 -4.49703872e-01 -7.05800354e-01
-6.42773688e-01 1.66854233e-01 9.43178296e-01 2.67292053e-01
-6.19349740e-02 -9.52981293e-01 -1.01436758e+00 5.26222527e-01
2.20913231e-01 -3.07299346e-01 8.86573717e-02 6.81346476e-01
-1.93718389e-01 3.31973732e-01 -2.52281398e-01 -1.37735665e-01
-1.25401747e+00 5.70926666e-01 3.57315615e-02 -5.99012151e-02
-3.57340664e-01 3.72901678e-01 6.44260645e-02 -6.56347871e-01
1.79406613e-01 -1.38873577e-01 3.36057574e-01 2.69562781e-01
4.30788726e-01 5.09577036e-01 2.40786776e-01 -5.39441884e-01
-4.42516029e-01 2.14490548e-01 -4.63051200e-02 -5.80038987e-02
1.56306136e+00 -1.10361725e-01 -2.43306592e-01 2.20605597e-01
1.18173993e+00 4.40960318e-01 -1.06214774e+00 6.31918460e-02
-1.94725133e-02 -8.46898973e-01 -3.66756171e-01 -8.48311663e-01
-8.94421339e-01 6.39847755e-01 4.17917877e-01 4.27541554e-01
7.96831727e-01 -2.95659214e-01 3.99345994e-01 8.38706940e-02
4.66966450e-01 -8.62592816e-01 -8.65049422e-01 -3.13773900e-01
6.60947144e-01 -1.06847024e+00 5.07151544e-01 -2.44946256e-01
-6.01841986e-01 7.08909392e-01 1.27359763e-01 1.71162158e-01
4.62635130e-01 2.05910012e-01 -1.60219401e-01 6.45974949e-02
-5.76187789e-01 4.27892447e-01 -1.88223392e-01 8.03643465e-01
8.07336420e-02 1.13214329e-01 -2.75137931e-01 7.47716844e-01
-5.93394876e-01 9.39619960e-04 7.87829459e-01 5.28141260e-01
5.66293076e-02 -1.14106452e+00 -6.24951124e-01 7.40809321e-01
-7.26028502e-01 1.78966537e-01 -7.07016945e-01 8.96887004e-01
1.58115208e-01 1.04532707e+00 8.33110884e-02 -2.38595009e-01
3.18575501e-01 1.70103118e-01 5.96938208e-02 -3.73099983e-01
-6.77476466e-01 -5.39844990e-01 5.41975558e-01 -3.28831643e-01
-4.07812260e-02 -1.19995546e+00 -7.69389212e-01 -8.83050740e-01
-2.20566973e-01 3.65830600e-01 6.33915722e-01 7.79598057e-01
4.88389522e-01 -2.50334769e-01 5.26908636e-01 -1.99970528e-01
-1.01918161e+00 -6.19278252e-01 -7.60449409e-01 7.31644809e-01
3.08391362e-01 -4.32852050e-03 -4.24690723e-01 -1.15591399e-01] | [6.177578449249268, 6.893073558807373] |
0aab189c-eabe-4754-a5e4-ed1006e63eb0 | automated-identication-of-atrial-fibrillation | 2306.15096 | null | https://arxiv.org/abs/2306.15096v1 | https://arxiv.org/pdf/2306.15096v1.pdf | Automated Identication of Atrial Fibrillation from Single-lead ECGs Using Multi-branching ResNet | Atrial fibrillation (AF) is the most common cardiac arrhythmia, which is clinically identified with irregular and rapid heartbeat rhythm. AF puts a patient at risk of forming blood clots, which can eventually lead to heart failure, stroke, or even sudden death. It is of critical importance to develop an advanced analytical model that can effectively interpret the electrocardiography (ECG) signals and provide decision support for accurate AF diagnostics. In this paper, we propose an innovative deep-learning method for automated AF identification from single-lead ECGs. We first engage the continuous wavelet transform (CWT) to extract time-frequency features from ECG signals. Then, we develop a convolutional neural network (CNN) structure that incorporates ResNet for effective network training and multi-branching architectures for addressing the imbalanced data issue to process the 2D time-frequency features for AF classification. We evaluate the proposed methodology using two real-world ECG databases. The experimental results show a superior performance of our method compared with traditional deep learning models. | ['Bing Yao', 'Stavros Stavrakis', 'Jianxin Xie'] | 2023-06-26 | null | null | null | null | ['electrocardiography-ecg'] | ['methodology'] | [ 2.60652810e-01 -5.12599349e-01 2.59578049e-01 -2.62413830e-01
-4.34743583e-01 -4.38981086e-01 -1.98366940e-01 2.32940599e-01
-1.25827223e-01 7.65878737e-01 8.24200958e-02 -6.67012870e-01
-2.59108216e-01 -6.89195931e-01 -4.53051776e-02 -6.16483569e-01
-6.26699626e-01 3.15125972e-01 -4.96685296e-01 1.47997350e-01
1.41328797e-01 7.98254013e-01 -8.86266828e-01 3.26251239e-01
1.13204515e+00 1.33573031e+00 -4.07746255e-01 8.16814721e-01
3.75214666e-02 8.05054545e-01 -9.67577577e-01 1.56232074e-01
1.28879249e-01 -7.86174178e-01 -6.02026105e-01 -3.61229807e-01
-3.06955993e-01 -1.86041921e-01 -1.66776460e-02 6.46541357e-01
1.22245610e+00 -4.81298000e-01 5.44308305e-01 -8.36957932e-01
-1.25897303e-01 4.60232168e-01 -4.22699869e-01 1.02038634e+00
1.18240334e-01 1.81297898e-01 3.45045686e-01 -8.55158925e-01
-2.62190811e-02 8.26352417e-01 1.18430495e+00 1.36254027e-01
-7.60228336e-01 -8.21743667e-01 -4.24931049e-01 2.91957587e-01
-1.16223788e+00 -2.96861351e-01 1.12000418e+00 -3.57907534e-01
6.25008881e-01 1.81813717e-01 1.23423231e+00 8.23534429e-01
7.64663339e-01 5.33434749e-01 1.09082484e+00 -2.48659253e-01
4.87351641e-02 -7.41904140e-01 3.22841108e-01 4.72760290e-01
2.61450261e-01 9.75913033e-02 -2.72150636e-01 -3.99520725e-01
8.57103169e-01 2.34619305e-01 -5.86105883e-01 2.25414202e-01
-1.39351332e+00 5.31301975e-01 4.69447106e-01 4.95542109e-01
-7.23997116e-01 7.51497075e-02 6.59720898e-01 5.91578007e-01
4.24527556e-01 5.32945693e-01 -4.98584777e-01 -3.56467962e-01
-8.55508268e-01 1.47841379e-01 5.60218394e-01 6.63919225e-02
2.51737386e-01 4.70125556e-01 -3.17174524e-01 5.62693477e-01
3.17832947e-01 4.51413006e-01 8.53152275e-01 -6.25698388e-01
2.87959814e-01 6.97173119e-01 4.94270623e-02 -1.16670096e+00
-6.90575302e-01 -1.18482184e+00 -1.66437280e+00 3.40513214e-02
3.12337935e-01 -3.35318506e-01 -6.31498158e-01 1.21375299e+00
-2.87980009e-02 6.01979315e-01 1.32013157e-01 9.05162215e-01
9.52918887e-01 4.59444016e-01 2.87481546e-02 -7.68737495e-01
1.30626917e+00 -3.93504888e-01 -1.06285512e+00 1.80078521e-02
3.23450625e-01 -4.38470930e-01 8.75041544e-01 5.75625837e-01
-9.30498898e-01 -8.33348334e-01 -1.28254366e+00 2.28702039e-01
3.95510718e-02 2.54305661e-01 4.91698980e-01 6.58581197e-01
-6.91056490e-01 7.35240459e-01 -8.26548338e-01 9.99535397e-02
8.36005747e-01 4.07576233e-01 -8.38484336e-03 2.80497909e-01
-1.54746795e+00 7.50176251e-01 2.08085164e-01 7.60755241e-01
-6.95117533e-01 -7.59835660e-01 -5.93437612e-01 1.40530495e-02
-4.00983900e-01 -8.49135816e-01 7.51856327e-01 -1.01684594e+00
-1.21933722e+00 5.70038378e-01 -4.36682068e-02 -6.18087113e-01
4.30158079e-01 -4.44193572e-01 -6.00064874e-01 1.50247276e-01
-1.29490599e-01 -2.91624665e-01 1.11578655e+00 -6.40185297e-01
-1.42824054e-01 -6.16788149e-01 -5.12466490e-01 1.47068247e-01
-2.35892162e-01 -5.61764836e-03 4.31279451e-01 -1.04141545e+00
3.22206229e-01 -4.04662520e-01 -3.07232141e-01 -3.48255664e-01
-2.28655592e-01 -3.55967693e-02 6.41902983e-01 -8.95561337e-01
1.60729456e+00 -2.17292857e+00 9.07197520e-02 3.12076509e-01
6.48715138e-01 6.19717896e-01 2.82640457e-01 5.95340021e-02
-3.17917913e-01 2.21655712e-01 -4.38443929e-01 1.82273373e-01
-7.53105938e-01 -7.81058567e-03 -3.52619201e-01 5.17862558e-01
3.55900317e-01 1.03622365e+00 -6.11504614e-01 -2.73580581e-01
6.72383532e-02 5.53841710e-01 -5.98703846e-02 4.20427531e-01
3.10099661e-01 9.58315074e-01 -5.58285773e-01 4.55183446e-01
7.59458721e-01 -3.24002951e-01 7.12994561e-02 -5.47303975e-01
1.15351371e-01 2.07773298e-01 -9.53535795e-01 1.57474971e+00
-1.79868594e-01 3.07420284e-01 -3.15460056e-01 -1.36998236e+00
1.19052029e+00 5.42837858e-01 7.10710704e-01 -5.36206901e-01
4.48891401e-01 4.47957069e-01 3.31531882e-01 -5.77258646e-01
-6.60532296e-01 -2.24615470e-01 1.72943994e-01 6.35256290e-01
-2.55475581e-01 3.22548747e-01 -3.06867629e-01 -4.31260288e-01
1.10599411e+00 -1.87702984e-01 4.69172925e-01 -5.80592811e-01
9.15987432e-01 -4.07588869e-01 9.45907652e-01 4.85079885e-01
-2.40135580e-01 6.62145019e-01 4.30573970e-01 -1.60280025e+00
-5.51331758e-01 -9.27187562e-01 -3.22868109e-01 7.79989436e-02
-5.16179241e-02 -3.07746649e-01 -5.52101016e-01 -5.40644407e-01
-1.23756416e-01 3.47273238e-02 -4.59578484e-01 -4.66600686e-01
-8.84005368e-01 -1.24043083e+00 9.39772010e-01 6.97834015e-01
9.36275184e-01 -1.34533727e+00 -1.16947842e+00 6.00105464e-01
-4.66312319e-01 -4.81349319e-01 -1.03444189e-01 1.43660337e-01
-1.23001385e+00 -1.14751780e+00 -7.37833798e-01 -6.77848339e-01
2.60246038e-01 -3.90899986e-01 1.23034763e+00 5.74774444e-01
-7.58554220e-01 -3.45441848e-01 -2.65973598e-01 -7.49417067e-01
-4.00320858e-01 1.83221009e-02 1.27428189e-01 3.98073643e-01
3.74808699e-01 -9.19979453e-01 -1.16585076e+00 -2.35429164e-02
-4.99497652e-01 -7.17205927e-02 6.63527191e-01 8.66310656e-01
6.46063745e-01 9.02737379e-02 1.35903919e+00 -6.88208222e-01
1.11409235e+00 -2.43440300e-01 -9.85421538e-02 7.55481655e-03
-7.90586174e-01 -3.17564040e-01 8.29252362e-01 -1.39799237e-01
-5.98395944e-01 -2.11842820e-01 -2.74693519e-01 -2.74996966e-01
-1.48849174e-01 7.58472145e-01 -8.96083564e-02 9.60322917e-02
6.86085820e-01 4.76919770e-01 8.88014138e-02 -3.35467607e-01
-1.76814690e-01 8.67070377e-01 5.89946032e-01 -4.18793678e-01
4.37922508e-01 3.04593533e-01 6.45295577e-03 -6.74818218e-01
-7.31318414e-01 -2.42167830e-01 -6.64556324e-01 -1.37606516e-01
8.75620127e-01 -9.76744652e-01 -8.24739814e-01 6.79447114e-01
-1.23267663e+00 -1.92094900e-04 -2.31954560e-01 4.59366471e-01
-4.22734439e-01 5.21692395e-01 -7.82463014e-01 -7.55836964e-01
-1.34418023e+00 -7.65390575e-01 7.36742139e-01 6.38549700e-02
-3.20185930e-01 -8.80326152e-01 1.94204435e-01 2.65833130e-03
6.01081908e-01 8.15380991e-01 1.03879845e+00 -5.72546601e-01
-6.14900813e-02 -2.03439251e-01 -4.09010015e-02 5.88990510e-01
4.40065503e-01 -2.50330001e-01 -9.44124818e-01 -3.83558154e-01
5.49853563e-01 7.27559477e-02 7.13525951e-01 6.36583745e-01
1.44250214e+00 -9.06189382e-02 -2.46909544e-01 1.10495520e+00
1.12753654e+00 5.92957795e-01 8.36660028e-01 9.62001532e-02
7.20571876e-01 -1.18648641e-01 3.09541970e-01 5.30964434e-01
1.24411769e-01 7.10452348e-02 1.48887513e-02 -5.69851518e-01
6.11154400e-02 2.46255934e-01 -5.52229397e-02 1.06284738e+00
-6.48243666e-01 6.46739006e-02 -1.13272226e+00 3.64026368e-01
-1.72172844e+00 -8.55061293e-01 -3.61514807e-01 1.99926519e+00
9.02032733e-01 4.11200486e-02 8.87283124e-03 9.10843134e-01
5.29356241e-01 -1.93189934e-01 -6.93594813e-01 -2.04229385e-01
-2.10148588e-01 7.94389784e-01 -2.44473234e-01 9.78668183e-02
-1.26582146e+00 2.18611956e-01 6.22926140e+00 3.31888311e-02
-1.56516254e+00 6.09199405e-02 1.03377187e+00 3.49433511e-01
7.83507600e-02 -3.71817976e-01 1.11768255e-02 5.71230888e-01
8.20687115e-01 -2.69939750e-01 1.38373002e-01 2.32910454e-01
3.51642698e-01 5.65533459e-01 -1.02994740e+00 1.44049549e+00
-9.53601375e-02 -1.28178942e+00 4.78407480e-02 -3.05974394e-01
1.50638163e-01 -2.54465401e-01 -2.13816717e-01 -1.06920868e-01
-5.56787312e-01 -1.47071278e+00 1.92098737e-01 8.31178010e-01
1.12584114e+00 -9.34462368e-01 1.12467277e+00 2.25886002e-01
-1.29411185e+00 -4.89044935e-01 -2.84771144e-01 -4.01934743e-01
1.24080218e-02 1.10345697e+00 -5.69610715e-01 7.87979722e-01
8.91169071e-01 1.31556952e+00 -4.57498908e-01 1.11418223e+00
1.41885541e-02 9.39713180e-01 -1.15142412e-01 3.59208047e-01
-5.23017824e-01 -1.86060056e-01 4.98012483e-01 9.06107962e-01
4.70183909e-01 3.14571559e-01 1.97542205e-01 7.67782629e-01
7.70332366e-02 2.14223698e-01 -5.14157057e-01 1.82990313e-01
3.37663382e-01 1.14012325e+00 -8.55087519e-01 -4.10471350e-01
-3.09694782e-02 6.84676170e-01 -1.19331263e-01 3.31234455e-01
-6.90482616e-01 -6.62964702e-01 2.67745018e-01 1.72893658e-01
-2.95267820e-01 -7.10537378e-03 -6.84439242e-01 -1.27086496e+00
2.18428344e-01 -9.95382845e-01 5.78748345e-01 -4.30894285e-01
-1.38771367e+00 9.56841767e-01 -6.10337853e-01 -1.36609304e+00
-1.13484517e-01 -3.07205737e-01 -1.06590021e+00 1.29441917e+00
-1.59567392e+00 -8.74463737e-01 -6.70980215e-01 8.17011774e-01
2.76884228e-01 -3.16644371e-01 1.23574960e+00 6.10243678e-01
-3.62160742e-01 2.97090322e-01 -5.79039037e-01 3.56306374e-01
5.37602067e-01 -1.21572828e+00 4.84087229e-01 8.64579797e-01
-1.21730953e-01 6.38081372e-01 3.45959514e-01 -6.30693734e-01
-1.12065196e+00 -1.20434093e+00 7.82585323e-01 -3.34718615e-01
2.09688842e-02 -5.68938628e-03 -8.47831070e-01 3.09957325e-01
6.90812618e-02 4.47704077e-01 9.11999881e-01 -2.23929852e-01
1.40159488e-01 -6.41910672e-01 -9.92263198e-01 7.54424036e-02
7.75936186e-01 -2.96030492e-01 -7.66355634e-01 1.84067145e-01
3.14011157e-01 -2.55836159e-01 -1.09593451e+00 9.39462423e-01
7.37547696e-01 -1.03140128e+00 1.01403105e+00 -6.73049390e-01
2.28232548e-01 -4.41725850e-01 5.99039435e-01 -1.37064230e+00
-4.86449063e-01 -9.23947692e-01 -3.93107295e-01 6.66780353e-01
1.04796201e-01 -7.82576680e-01 5.35146654e-01 -1.57977834e-01
-2.11349145e-01 -1.02006292e+00 -8.14306676e-01 -2.06829280e-01
-1.78369984e-04 -1.96959853e-01 8.01376283e-01 1.02808297e+00
-1.50402650e-01 4.92478997e-01 -2.30751708e-01 3.18911821e-02
6.01891518e-01 3.59158337e-01 2.78888553e-01 -1.91353321e+00
2.54715141e-02 -2.30770960e-01 -4.88823205e-01 -3.91284585e-01
-2.54421145e-01 -8.81201982e-01 -3.59725237e-01 -1.48789024e+00
-1.59596562e-01 -5.59137881e-01 -9.78568077e-01 3.87373924e-01
-4.03878808e-01 3.28920662e-01 -3.28599036e-01 2.20180348e-01
-3.45939919e-02 3.19173962e-01 1.21120119e+00 -1.50465205e-01
-4.04165030e-01 4.14261192e-01 -6.70348942e-01 6.44931495e-01
1.05718553e+00 -2.34354213e-01 -5.12278795e-01 -2.72917807e-01
3.55412275e-01 4.66837913e-01 3.11571121e-01 -1.35958469e+00
-1.01278059e-01 4.06368703e-01 1.09504056e+00 -3.98625940e-01
-2.82137454e-01 -7.53865898e-01 2.27153227e-01 1.03624201e+00
-2.05356598e-01 5.25731266e-01 1.58877626e-01 3.00828338e-01
-4.06372637e-01 4.01511967e-01 5.75370252e-01 -7.16836229e-02
1.31609187e-01 6.28723323e-01 -5.04938543e-01 1.26272827e-01
8.14285219e-01 -1.34391740e-01 -1.83450654e-02 -1.35471255e-01
-9.59734440e-01 1.36677492e-02 -2.24617645e-01 2.89706022e-01
1.10791671e+00 -1.48186862e+00 -1.06108451e+00 7.39732862e-01
-6.02867045e-02 2.74918854e-01 2.05622777e-01 1.07495117e+00
-1.08380949e+00 -8.57260227e-02 -5.47753274e-01 -7.51097441e-01
-1.09118021e+00 2.45194003e-01 7.78182089e-01 -2.76337981e-01
-1.08481312e+00 5.66259921e-01 -2.86854178e-01 5.02672978e-02
1.74256235e-01 -5.64929068e-01 -6.48549438e-01 -5.29599749e-02
6.98329151e-01 3.29899102e-01 2.32169092e-01 -1.68522894e-01
-5.74208200e-01 7.84681797e-01 2.97360599e-01 3.72887105e-01
1.25476170e+00 2.15204075e-01 -6.49531901e-01 4.40075427e-01
7.23791778e-01 -3.07932287e-01 -7.57448614e-01 2.41695851e-01
-1.60664394e-01 -9.25309211e-02 2.57110577e-02 -9.06611860e-01
-1.31805873e+00 1.34554589e+00 1.17065012e+00 3.00923467e-01
1.52172220e+00 -6.24505281e-01 1.08936346e+00 3.42552304e-01
2.43158937e-01 -5.74215293e-01 5.13022803e-02 1.32399565e-02
9.46879208e-01 -6.65591240e-01 2.39171814e-02 -4.72769886e-02
-1.60796821e-01 1.63000357e+00 2.27525905e-01 -3.01018119e-01
1.09937513e+00 3.40840697e-01 6.62966251e-01 -3.24881554e-01
-3.08831543e-01 5.05562305e-01 1.02098942e-01 6.16234601e-01
6.16108596e-01 1.94095429e-02 -4.26103503e-01 1.08961475e+00
-8.33339244e-02 5.27138948e-01 2.58702219e-01 7.78583407e-01
-2.39253730e-01 -9.94301021e-01 -1.67369679e-01 8.69352758e-01
-8.71709943e-01 -8.70425031e-02 -2.51004308e-01 5.81731014e-02
3.75804275e-01 8.40719163e-01 9.53623094e-03 -2.66896755e-01
1.25477836e-01 4.28698123e-01 2.72582680e-01 -3.26086253e-01
-7.96580195e-01 4.70620096e-02 -2.49181509e-01 -3.11120719e-01
-6.15867019e-01 -3.15808594e-01 -1.26275575e+00 3.28624278e-01
-1.92124806e-02 2.72966236e-01 1.91326201e-01 8.34230781e-01
6.69555962e-01 9.90422487e-01 7.66698837e-01 -3.26098591e-01
-3.79627854e-01 -1.20008624e+00 -5.54552972e-01 4.22912210e-01
8.02088261e-01 -3.12038660e-01 -3.25005293e-01 2.58210570e-01] | [14.281596183776855, 3.264986038208008] |
6d13eaf4-0da2-43a3-8010-da21182d8640 | graph-convolutional-transformer-learning-the | 1906.04716 | null | https://arxiv.org/abs/1906.04716v3 | https://arxiv.org/pdf/1906.04716v3.pdf | Learning the Graphical Structure of Electronic Health Records with Graph Convolutional Transformer | Effective modeling of electronic health records (EHR) is rapidly becoming an important topic in both academia and industry. A recent study showed that using the graphical structure underlying EHR data (e.g. relationship between diagnoses and treatments) improves the performance of prediction tasks such as heart failure prediction. However, EHR data do not always contain complete structure information. Moreover, when it comes to claims data, structure information is completely unavailable to begin with. Under such circumstances, can we still do better than just treating EHR data as a flat-structured bag-of-features? In this paper, we study the possibility of jointly learning the hidden structure of EHR while performing supervised prediction tasks on EHR data. Specifically, we discuss that Transformer is a suitable basis model to learn the hidden EHR structure, and propose Graph Convolutional Transformer, which uses data statistics to guide the structure learning process. The proposed model consistently outperformed previous approaches empirically, on both synthetic data and publicly available EHR data, for various prediction tasks such as graph reconstruction and readmission prediction, indicating that it can serve as an effective general-purpose representation learning algorithm for EHR data. | ['Zhen Xu', 'Yujia Li', 'Yuan Xue', 'Gerardo Flores', 'Michael W. Dusenberry', 'Edward Choi', 'Andrew M. Dai'] | 2019-06-11 | null | null | null | null | ['graph-reconstruction', 'readmission-prediction'] | ['graphs', 'medical'] | [ 2.06991509e-01 6.30745351e-01 -1.89290136e-01 -5.16777694e-01
-6.07937813e-01 -1.62506312e-01 9.22640637e-02 7.82816410e-01
2.02902868e-01 6.56506240e-01 8.33219945e-01 -6.41732574e-01
-2.08814040e-01 -1.03492343e+00 -8.35178494e-01 -3.60004187e-01
-2.42648855e-01 6.64615273e-01 -4.12887901e-01 -1.63665693e-02
-3.60341966e-01 1.09158449e-01 -8.69948268e-01 4.21872556e-01
7.33946681e-01 9.09008503e-01 7.47296736e-02 2.14929745e-01
-6.89560995e-02 1.32994568e+00 -1.57510489e-01 -6.10915005e-01
6.42912760e-02 -4.32344645e-01 -7.96600819e-01 2.10901424e-01
1.14016928e-01 -1.75823614e-01 -8.62437427e-01 7.54393518e-01
4.15307462e-01 -5.14367580e-01 5.91068625e-01 -7.83680499e-01
-6.76465333e-01 1.11691105e+00 -4.19821322e-01 -1.15715638e-01
3.54239076e-01 3.84884663e-02 1.38877046e+00 -4.25651819e-01
5.91577113e-01 9.78746235e-01 8.73540521e-01 3.32258672e-01
-1.25078797e+00 -5.02381027e-01 5.35184033e-02 1.69553831e-01
-1.34919798e+00 -3.73378664e-01 8.75786185e-01 -6.08233809e-01
6.25779748e-01 3.23927552e-01 8.69894564e-01 9.99659181e-01
3.72615486e-01 9.51197445e-01 7.64591515e-01 -4.27343361e-02
1.41638881e-02 -7.82753453e-02 5.51041067e-01 9.85736191e-01
7.51537025e-01 -1.72024965e-01 -9.28151142e-03 -7.51182675e-01
6.84025347e-01 6.72137260e-01 -5.24087429e-01 -4.26632673e-01
-1.28568554e+00 6.62564218e-01 7.06509709e-01 1.13639131e-03
-7.83075690e-01 4.45882007e-02 3.50186348e-01 2.14392319e-01
3.14799041e-01 2.20187202e-01 -5.22819638e-01 2.19680309e-01
-4.66637522e-01 2.40514483e-02 8.42401981e-01 8.30876648e-01
5.04411757e-01 1.10747144e-01 -3.07086021e-01 5.32411695e-01
5.21610200e-01 2.06904575e-01 7.84201771e-02 -2.34863997e-01
7.27093697e-01 1.05312049e+00 -1.24291524e-01 -1.11146450e+00
-7.29856014e-01 -7.14055717e-01 -1.47578645e+00 -7.21152723e-01
3.58724535e-01 -2.08352879e-01 -7.93034256e-01 1.59393990e+00
3.29401940e-01 5.57930112e-01 8.53280053e-02 5.22865176e-01
1.21619284e+00 3.52947682e-01 1.83780462e-01 -1.85861692e-01
1.47714067e+00 -5.22551537e-01 -9.96754050e-01 1.29198998e-01
1.31922615e+00 -2.38950089e-01 7.66367972e-01 1.47460192e-01
-6.79205477e-01 -2.42141172e-01 -5.84695935e-01 6.85312152e-02
9.39153507e-02 2.14523584e-01 9.15321887e-01 4.09032077e-01
-7.15618968e-01 5.30485511e-01 -9.78402138e-01 -2.71782964e-01
7.94992745e-01 2.15081558e-01 -5.29857099e-01 -6.54697061e-01
-1.19342613e+00 4.49676365e-01 2.82357246e-01 1.29293293e-01
-7.33061433e-01 -6.62351072e-01 -1.06648362e+00 5.12725770e-01
3.65013868e-01 -1.24912453e+00 8.86538684e-01 -2.32941285e-01
-8.62695038e-01 5.91912806e-01 -1.38142928e-02 -7.80235112e-01
2.67941535e-01 -1.85219541e-01 -4.56283242e-01 -1.76341057e-01
-4.06442970e-01 -5.51692164e-03 6.54079258e-01 -1.09446597e+00
-6.01789989e-02 -7.05645204e-01 -2.64650751e-02 -7.55520090e-02
-4.08446968e-01 -4.49377984e-01 -2.11628571e-01 -4.61318880e-01
1.08539939e-01 -9.62690294e-01 -5.75080216e-01 -3.21464211e-01
-9.03278470e-01 -2.24356651e-01 5.65909028e-01 -1.09544170e+00
1.62818933e+00 -2.10638309e+00 -1.05311526e-02 2.84579217e-01
1.00868642e+00 -3.04287579e-02 8.52723345e-02 7.16366589e-01
-1.02876581e-01 6.62623271e-02 -2.50699580e-01 -1.79180235e-01
-2.40822956e-01 3.16689819e-01 -1.49412662e-01 5.08365810e-01
6.49418384e-02 1.39631426e+00 -7.79708862e-01 -5.51159143e-01
-3.84603254e-02 7.22203135e-01 -6.24349296e-01 3.25667560e-01
-8.54353141e-03 6.39935493e-01 -7.65312612e-01 5.37450075e-01
4.81760561e-01 -1.26378226e+00 7.18698621e-01 -3.26001614e-01
6.28798366e-01 6.24308527e-01 -8.61726165e-01 1.48173404e+00
-1.93998590e-01 1.76641569e-01 -3.84496719e-01 -1.26142037e+00
7.83405304e-01 5.21671236e-01 9.28330302e-01 -3.70575160e-01
-1.26597017e-01 -2.64261097e-01 2.76238650e-01 -4.46251720e-01
-7.99621716e-02 1.22863874e-01 4.86377962e-02 3.19114417e-01
-2.83402652e-01 3.62027645e-01 -3.40301841e-01 3.66559446e-01
1.64535236e+00 -4.43893701e-01 7.32515335e-01 3.28384750e-02
1.09301962e-01 -7.90046528e-02 5.86875141e-01 5.63684881e-01
2.81663597e-01 5.67971766e-01 6.22040808e-01 -6.71380997e-01
-8.61054838e-01 -7.61678815e-01 -1.84381351e-01 4.37708169e-01
-2.06073597e-01 -8.94895613e-01 -4.39548701e-01 -9.00877297e-01
1.82835951e-01 5.65067351e-01 -5.88471115e-01 -2.88916051e-01
-3.16547483e-01 -9.13930237e-01 3.79155934e-01 4.91735309e-01
2.45368928e-01 -8.25293124e-01 -2.81035274e-01 3.45116585e-01
-2.95483977e-01 -1.25005639e+00 -3.33619118e-01 -1.53782889e-02
-1.28707957e+00 -1.43569016e+00 -4.12312776e-01 -6.41429901e-01
7.24342227e-01 2.86657959e-01 1.38377106e+00 4.33722347e-01
-3.02042216e-01 4.15443897e-01 -3.40046734e-01 -2.69860536e-01
-4.10253853e-01 8.98982733e-02 -3.97054732e-01 2.85137445e-01
2.46436566e-01 -4.67742503e-01 -7.48978913e-01 5.03734946e-02
-7.71849334e-01 3.42276633e-01 7.34430671e-01 1.00996029e+00
6.74144745e-01 1.68540984e-01 6.58916771e-01 -1.63604641e+00
5.76042175e-01 -7.26592362e-01 -1.53521180e-01 3.82375240e-01
-7.26357877e-01 9.35887545e-02 6.00078106e-01 -5.67515520e-03
-5.66803992e-01 3.33051592e-01 -2.47689143e-01 -3.92374307e-01
-1.72341522e-02 1.12037408e+00 -1.41338363e-01 4.00652230e-01
4.43379730e-01 2.84327388e-01 -2.09286306e-02 -8.62536848e-01
4.20240834e-02 6.97711587e-01 -2.17239708e-01 -1.67094871e-01
5.14772475e-01 3.27679336e-01 1.54809698e-01 -6.70874178e-01
-1.07049060e+00 -4.63648111e-01 -3.92867237e-01 2.87739605e-01
8.76832664e-01 -1.15487039e+00 -7.14751244e-01 2.41731852e-02
-9.59489286e-01 -3.65172103e-02 -2.32495666e-01 5.50469100e-01
-1.81583136e-01 5.45882821e-01 -8.31772804e-01 -6.67411923e-01
-5.02951622e-01 -8.38427007e-01 1.29594970e+00 -4.56811368e-01
-7.73403496e-02 -1.23877811e+00 9.39115044e-03 5.32097876e-01
1.45562604e-01 4.33999866e-01 1.39603674e+00 -9.51250374e-01
-6.77472949e-01 -4.87112075e-01 -2.98103005e-01 1.66019365e-01
3.94212991e-01 -4.21981066e-01 -7.67417431e-01 -4.36475307e-01
-4.91780080e-02 -1.69966802e-01 7.81912684e-01 5.59605300e-01
1.57022667e+00 -7.79913962e-01 -5.84642470e-01 5.95655322e-01
1.25753713e+00 -2.06166938e-01 7.32550502e-01 -2.95564801e-01
1.21415246e+00 4.56477016e-01 -1.75572988e-02 7.69968450e-01
9.76289272e-01 6.48774147e-01 5.17822981e-01 -4.42635298e-01
-3.78448470e-03 -5.78520894e-01 -8.33915994e-02 1.19874156e+00
-1.30806848e-01 -1.47910282e-01 -1.27084351e+00 2.14061126e-01
-2.12966323e+00 -7.64746368e-01 -4.15651530e-01 2.22062898e+00
5.62729239e-01 -5.36146807e-03 8.88942853e-02 1.06012419e-01
5.75403750e-01 -2.99361497e-02 -4.67204750e-01 3.78634840e-01
-2.35202047e-03 -7.98224099e-03 4.25697803e-01 6.39866944e-03
-9.85479414e-01 4.32374954e-01 5.73876429e+00 2.74843216e-01
-8.54605794e-01 1.02525339e-01 7.64989555e-01 4.41104412e-01
-5.25170326e-01 6.70467690e-02 -6.10688627e-01 4.02409226e-01
9.86157537e-01 -6.70505390e-02 1.10566273e-01 7.40399480e-01
2.29324311e-01 6.99356914e-01 -1.34478760e+00 1.20145631e+00
-2.78694093e-01 -1.56699967e+00 3.16978276e-01 5.24799705e-01
4.39603329e-01 1.64790988e-01 -2.08512709e-01 4.16082531e-01
4.39171463e-01 -1.29709649e+00 -5.77420369e-02 6.47755980e-01
5.97426593e-01 -2.61767536e-01 9.91138458e-01 6.12959862e-01
-1.30088317e+00 -1.52497157e-01 -4.35886204e-01 1.98830664e-02
1.27906566e-02 8.92254472e-01 -1.26651311e+00 9.71128941e-01
4.36802655e-01 1.26309705e+00 -7.14767396e-01 1.25292695e+00
5.47616743e-02 1.07569969e+00 -6.78724572e-02 2.67588586e-01
-1.11787729e-01 -2.37625867e-01 -1.37100890e-02 1.05101240e+00
4.30724055e-01 1.81123450e-01 6.22610390e-01 4.78790104e-01
-4.07735974e-01 4.04398531e-01 -1.09378242e+00 -3.95515978e-01
1.36409670e-01 1.08235180e+00 -3.85326952e-01 -4.19728011e-01
-7.11854041e-01 4.35843468e-01 5.04859805e-01 3.69740754e-01
-7.65835464e-01 2.12227911e-01 4.33013231e-01 6.67874336e-01
1.98758051e-01 5.12351617e-02 -2.94933230e-01 -1.59491372e+00
-1.73370451e-01 -9.04112160e-01 8.00149322e-01 -4.68173653e-01
-1.53509116e+00 5.64742625e-01 -3.54610294e-01 -1.52848899e+00
-1.71037227e-01 -1.89953774e-01 -2.69620806e-01 4.70388472e-01
-1.64268005e+00 -1.28647447e+00 -2.97804505e-01 7.34170496e-01
1.03333369e-01 -9.67930406e-02 1.07668304e+00 4.80414540e-01
-6.20528340e-01 5.26697576e-01 2.12312803e-01 5.49821436e-01
3.43995154e-01 -1.06004488e+00 2.69600391e-01 1.87918529e-01
5.09533644e-01 7.56124377e-01 4.02072847e-01 -8.28499496e-01
-1.93419671e+00 -1.43451488e+00 8.74399722e-01 -4.26197350e-01
5.30972242e-01 -3.29967082e-01 -1.19520867e+00 1.04184234e+00
-1.27020344e-01 2.89164960e-01 1.04797828e+00 4.28952366e-01
-2.64349133e-01 -1.14359438e-01 -8.82479310e-01 3.81344587e-01
1.25952041e+00 -5.11265337e-01 -4.98483181e-01 7.83316672e-01
8.35526049e-01 -3.51726443e-01 -1.28736103e+00 7.35593379e-01
2.78402776e-01 -4.83333826e-01 9.08676088e-01 -1.06606948e+00
4.20060903e-01 -9.58653614e-02 -9.56914797e-02 -1.24987626e+00
-5.72552681e-01 -5.09540379e-01 -7.25605547e-01 7.73901522e-01
5.08875787e-01 -6.60261929e-01 8.70919526e-01 5.31526923e-01
-1.08530864e-01 -1.02875018e+00 -5.95730901e-01 -3.90536308e-01
-4.69488621e-01 -2.36579314e-01 7.36330926e-01 1.31009495e+00
1.18486481e-02 9.09946382e-01 -6.80302918e-01 2.01321498e-01
6.72481120e-01 3.91854972e-01 9.81738269e-01 -1.74259913e+00
-6.27134860e-01 4.03167419e-02 -7.11806595e-01 -1.18387413e+00
5.16243279e-02 -1.28553355e+00 -6.51211858e-01 -2.12572098e+00
5.47118306e-01 -5.66331685e-01 -6.40264034e-01 7.36621737e-01
-3.46492857e-01 -3.06581050e-01 8.76250863e-02 3.88742268e-01
-5.75544894e-01 6.02235794e-01 1.49943852e+00 -3.39555711e-01
-7.34319314e-02 2.94558674e-01 -9.33013380e-01 4.09994036e-01
5.91998696e-01 -6.10249996e-01 -5.98880410e-01 -2.84300983e-01
3.39125693e-01 7.70048440e-01 3.31388474e-01 -6.29289448e-01
6.56329244e-02 1.35160789e-01 2.81490535e-01 -4.19100910e-01
1.96893528e-01 -1.18447363e+00 6.44784331e-01 6.31135404e-01
-5.05301416e-01 1.37322411e-01 -2.11594552e-01 1.05551422e+00
-3.45397383e-01 2.62255609e-01 1.93124324e-01 -1.29813537e-01
-1.31987765e-01 8.12424123e-01 1.63401678e-01 2.19857972e-02
7.11979270e-01 4.01718430e-02 -1.51356891e-01 -6.27153814e-01
-8.32099259e-01 2.51567841e-01 2.32535273e-01 1.60475880e-01
8.44673276e-01 -1.21715891e+00 -9.81513143e-01 1.74573973e-01
4.37736064e-01 1.08779892e-01 3.63885403e-01 1.03231645e+00
-2.53682315e-01 4.25300926e-01 2.58904248e-01 -7.71518171e-01
-1.24922192e+00 8.91745448e-01 -8.53175588e-04 -7.76902318e-01
-1.30255485e+00 2.67722934e-01 5.93141198e-01 -4.85496193e-01
3.29917967e-01 -5.62020957e-01 -5.22606850e-01 -3.56104881e-01
3.19105595e-01 -1.57671317e-01 1.52532920e-01 -4.11324054e-01
-1.40523463e-01 1.96884260e-01 -1.29977509e-01 7.85156131e-01
1.78103554e+00 2.36219931e-02 3.34181823e-02 4.68791276e-01
1.05290282e+00 -1.93133354e-01 -7.91662633e-01 -5.80212295e-01
2.45472118e-01 -3.14782470e-01 -3.98617126e-02 -4.64729041e-01
-1.23653114e+00 9.44828093e-01 2.74767458e-01 5.98701596e-01
9.45954740e-01 1.41792402e-01 9.23634470e-01 7.59196341e-01
4.21063423e-01 -4.10367489e-01 -2.54096240e-01 2.72600234e-01
6.45517886e-01 -1.31852853e+00 1.12857141e-01 -7.26476669e-01
-8.65750313e-01 8.67641330e-01 2.00724438e-01 -1.91322677e-02
1.18004417e+00 -4.38833944e-02 -3.10730249e-01 -6.00424588e-01
-1.12006009e+00 -4.70567159e-02 5.61481476e-01 4.76044714e-01
7.81738460e-01 4.15699273e-01 -8.27288702e-02 6.08337045e-01
-8.73521157e-03 1.87233910e-01 4.26454216e-01 6.82286739e-01
-1.50055975e-01 -9.86350060e-01 -5.50940447e-02 1.26059341e+00
-5.56524456e-01 -2.53251106e-01 -2.38710433e-01 6.08429074e-01
-3.67790848e-01 6.61792099e-01 -3.59782219e-01 -5.44536829e-01
3.42273235e-01 1.31822407e-01 2.04551116e-01 -9.30566490e-01
-5.34974396e-01 4.46564518e-02 3.17510545e-01 -3.82418096e-01
-1.82576075e-01 -4.77210104e-01 -1.19841611e+00 -1.64784878e-01
-3.16334575e-01 1.55581087e-01 3.13216269e-01 9.07498002e-01
7.79995501e-01 7.63865411e-01 5.76353908e-01 -5.46399765e-02
-7.51303554e-01 -8.37175786e-01 -7.07182944e-01 7.48997152e-01
2.87857383e-01 -3.88139784e-01 1.12881780e-01 5.45170568e-02] | [7.758249282836914, 6.451534271240234] |
bb96c04f-2f03-4f82-b55c-bf3ea857c05f | denseran-for-offline-handwritten-chinese | 1808.04134 | null | http://arxiv.org/abs/1808.04134v1 | http://arxiv.org/pdf/1808.04134v1.pdf | DenseRAN for Offline Handwritten Chinese Character Recognition | Recently, great success has been achieved in offline handwritten Chinese
character recognition by using deep learning methods. Chinese characters are
mainly logographic and consist of basic radicals, however, previous research
mostly treated each Chinese character as a whole without explicitly considering
its internal two-dimensional structure and radicals. In this study, we propose
a novel radical analysis network with densely connected architecture (DenseRAN)
to analyze Chinese character radicals and its two-dimensional structures
simultaneously. DenseRAN first encodes input image to high-level visual
features by employing DenseNet as an encoder. Then a decoder based on recurrent
neural networks is employed, aiming at generating captions of Chinese
characters by detecting radicals and two-dimensional structures through
attention mechanism. The manner of treating a Chinese character as a
composition of two-dimensional structures and radicals can reduce the size of
vocabulary and enable DenseRAN to possess the capability of recognizing unseen
Chinese character classes, only if the corresponding radicals have been seen in
training set. Evaluated on ICDAR-2013 competition database, the proposed
approach significantly outperforms whole-character modeling approach with a
relative character error rate (CER) reduction of 18.54%. Meanwhile, for the
case of recognizing 3277 unseen Chinese characters in CASIA-HWDB1.2 database,
DenseRAN can achieve a character accuracy of about 41% while the traditional
whole-character method has no capability to handle them. | ['Zi-Rui Wang', 'Yixing Zhu', 'Wenchao Wang', 'Jun Du', 'Jianshu Zhang'] | 2018-08-13 | null | null | null | null | ['offline-handwritten-chinese-character', 'offline-handwritten-chinese-character'] | ['computer-vision', 'natural-language-processing'] | [ 2.26845562e-01 -3.22521806e-01 2.20351554e-02 2.01252148e-01
-3.17596316e-01 -4.27699715e-01 4.99375731e-01 -9.72212628e-02
-3.80787373e-01 4.28010076e-01 1.68898031e-01 -2.35238060e-01
6.92889750e-01 -1.05540776e+00 -6.53220952e-01 -8.98496091e-01
2.34196097e-01 2.03094363e-01 1.18460674e-02 -9.78260040e-02
7.05995560e-01 8.70585322e-01 -1.25414503e+00 4.61094499e-01
9.56867158e-01 7.17656136e-01 5.46736121e-01 8.28609765e-01
-5.62769353e-01 1.42795551e+00 -1.08668160e+00 -5.18768728e-01
-2.17404719e-02 -5.32358110e-01 -6.21437907e-01 2.70020068e-01
5.27732968e-01 -4.86630827e-01 -7.49660373e-01 1.02744317e+00
7.28738785e-01 -2.65162528e-01 9.06433344e-01 -3.96147698e-01
-1.47402108e+00 7.60485232e-01 -4.66030061e-01 3.25391918e-01
-4.83816676e-02 -1.00179844e-01 7.94305027e-01 -1.00346422e+00
6.49065852e-01 1.16249955e+00 3.98960233e-01 9.15362656e-01
-6.81321919e-01 -5.55685163e-01 1.83722183e-01 5.06962776e-01
-1.29043889e+00 1.00828275e-01 1.04305422e+00 -4.01088029e-01
1.28138340e+00 2.17781126e-01 2.77472496e-01 1.14066696e+00
2.96172768e-01 1.17997789e+00 4.62118655e-01 -5.32646120e-01
-1.55964419e-01 -6.48835674e-02 5.12189984e-01 8.07083964e-01
2.12279439e-01 -3.54209006e-01 -2.63352662e-01 3.44923645e-01
7.54064381e-01 -3.19589339e-02 -1.90559283e-01 3.41061980e-01
-1.09479904e+00 7.05598831e-01 4.75473434e-01 2.73574948e-01
-4.66176122e-01 1.69355303e-01 7.70413041e-01 2.13662093e-03
-1.13064952e-01 2.62039125e-01 9.37888548e-02 7.82757252e-02
-4.97070879e-01 -1.19592473e-01 3.06641072e-01 1.07733309e+00
4.61692750e-01 7.71238565e-01 -9.16833282e-02 8.85418594e-01
-7.74637982e-02 5.51580131e-01 1.03098655e+00 -1.50569633e-01
7.54997373e-01 6.99078321e-01 -2.94790983e-01 -9.16553020e-01
-1.85988516e-01 -4.07666653e-01 -1.23521292e+00 -1.28619984e-01
2.62812823e-02 -8.90109986e-02 -1.37232172e+00 1.22006333e+00
-2.14665622e-01 -1.47996768e-01 4.68824476e-01 9.53616381e-01
9.80413914e-01 1.27643359e+00 -1.12155212e-04 1.33118862e-02
1.42321825e+00 -1.27141738e+00 -7.02303469e-01 -1.86598852e-01
5.85162163e-01 -7.76399016e-01 1.08615816e+00 5.84822178e-01
-7.72564113e-01 -7.35864639e-01 -1.31689656e+00 -5.59454858e-01
-4.10442352e-01 8.71463358e-01 4.89106119e-01 6.75893784e-01
-4.95518237e-01 2.28084013e-01 -5.21468878e-01 8.96712616e-02
4.04819101e-01 9.29138213e-02 -1.04808718e-01 -2.91319072e-01
-1.06902838e+00 6.86334848e-01 4.55058068e-01 5.22452235e-01
-1.16989493e+00 -3.44754606e-01 -7.40413368e-01 3.12481910e-01
3.54829654e-02 7.72125227e-03 8.33994329e-01 -1.05323589e+00
-1.46706688e+00 4.28349763e-01 1.39306024e-01 -4.45033848e-01
4.24041778e-01 -1.54715940e-01 -7.30747700e-01 5.98067701e-01
-2.80799389e-01 7.15352476e-01 7.69528687e-01 -9.45479453e-01
-3.78435582e-01 -2.53035612e-02 -3.62704515e-01 2.29618162e-01
-7.66218305e-01 5.81136569e-02 -5.24861991e-01 -8.73341084e-01
8.35848525e-02 -7.03595698e-01 9.07321498e-02 -1.92640111e-01
-6.35389447e-01 -1.60925627e-01 9.53809083e-01 -9.55319166e-01
1.19942403e+00 -2.00517082e+00 7.04139918e-02 -6.02553412e-02
1.17333990e-03 6.88823879e-01 -2.74677336e-01 3.38595003e-01
-1.44253880e-01 1.45253196e-01 -2.34257355e-01 -2.75025427e-01
-1.76646218e-01 -7.22246841e-02 -7.08419502e-01 4.44393724e-01
6.92461491e-01 7.00000644e-01 -5.38949192e-01 -4.66837347e-01
4.72437143e-02 5.89267015e-01 -1.99122354e-01 -3.09104510e-02
-1.08734488e-01 -2.50269562e-01 -4.22238111e-01 1.02316177e+00
8.09603512e-01 -1.57801345e-01 2.34476864e-01 -1.23538584e-01
-3.63530964e-01 1.72660872e-02 -6.31633580e-01 1.07727611e+00
-3.75765651e-01 1.00066042e+00 -6.21149063e-01 -8.64916801e-01
1.26647079e+00 8.63287225e-02 -1.48400784e-01 -1.11234462e+00
1.24720201e-01 1.47760257e-01 4.24525626e-02 -5.81674397e-01
8.93063843e-01 5.70042506e-02 -5.96416183e-02 7.26288930e-02
-3.06436151e-01 4.12187427e-01 3.58347356e-01 2.36711830e-01
1.02056348e+00 -3.40773284e-01 -4.94872719e-01 9.93651524e-03
8.69164944e-01 9.21738520e-02 5.01090646e-01 6.46317720e-01
2.93694258e-01 9.70181763e-01 6.19626939e-01 -6.17870271e-01
-1.37774920e+00 -7.89857090e-01 -1.42962664e-01 4.86158550e-01
1.60302877e-01 -1.51462585e-01 -4.91875023e-01 -3.86183649e-01
-2.46626809e-01 6.81033790e-01 -6.52754724e-01 -1.17484644e-01
-8.50979209e-01 -8.49757314e-01 1.15342402e+00 6.39252067e-01
8.85215163e-01 -1.48424780e+00 -5.78069270e-01 2.76769817e-01
-2.66330801e-02 -9.71341133e-01 -5.71416557e-01 2.53914982e-01
-7.76257694e-01 -9.41042244e-01 -1.36250305e+00 -1.34112442e+00
8.29815328e-01 1.29653916e-01 6.34353220e-01 1.77882016e-01
-6.93665802e-01 -2.16032058e-01 -5.91069639e-01 -1.02036417e-01
-4.58025992e-01 7.19356686e-02 -4.02756840e-01 8.23856294e-02
1.99780554e-01 7.58934990e-02 -4.38032508e-01 4.88316007e-02
-1.04147029e+00 1.30563840e-01 9.33403432e-01 9.33633566e-01
3.87090117e-01 1.10800918e-02 5.17583668e-01 -7.59379983e-01
4.61177647e-01 -2.31868982e-01 -5.53805113e-01 4.96881783e-01
-3.68082285e-01 8.05160552e-02 1.14433086e+00 -8.38919461e-01
-1.17847478e+00 9.75264178e-04 5.51866367e-02 -6.10415280e-01
1.56553566e-01 7.32967019e-01 -2.46441111e-01 1.86182231e-01
5.48677623e-01 9.14079189e-01 -2.49763265e-01 -6.01414144e-01
-1.52621046e-02 8.87886167e-01 7.49362409e-01 -4.70546663e-01
3.99359494e-01 2.43470848e-01 -1.63893625e-02 -1.22983277e+00
-1.49215922e-01 -1.37444973e-01 -3.43610615e-01 -7.00561553e-02
1.08609307e+00 -1.15199637e+00 -5.55568159e-01 1.30419743e+00
-1.23602819e+00 -1.81857765e-01 2.51389474e-01 3.14586639e-01
6.80719838e-02 1.16025913e+00 -1.23581135e+00 -6.68442249e-01
-6.73276305e-01 -1.09570074e+00 7.47698903e-01 2.56030858e-01
5.04435003e-01 -7.20813394e-01 -1.27156094e-01 1.00963414e-01
1.31284520e-01 -3.93836200e-02 1.29724872e+00 -3.65046233e-01
-5.90334713e-01 -5.41876853e-01 -3.70643675e-01 6.88437879e-01
-2.88558286e-02 1.92260221e-01 -7.13300645e-01 -1.05261929e-01
-3.15645844e-01 -4.55208778e-01 1.17350721e+00 -1.61229387e-01
1.32160878e+00 -1.24104649e-01 1.78305641e-01 3.25938284e-01
1.55432177e+00 4.01049525e-01 1.44320571e+00 2.70363301e-01
1.27423799e+00 2.61221647e-01 3.53742421e-01 4.84848231e-01
2.23525509e-01 2.57336468e-01 3.84167999e-01 -5.15894890e-02
-5.60071766e-01 -2.14134082e-01 7.74810076e-01 9.92996573e-01
2.04732627e-01 -7.20946848e-01 -1.06623662e+00 5.66445947e-01
-1.27320099e+00 -1.05666125e+00 -6.10251904e-01 1.86500394e+00
7.09455729e-01 1.11963786e-01 -2.70245522e-01 1.78619355e-01
1.16652071e+00 8.31163675e-02 -7.15492487e-01 -7.93444991e-01
-8.96551073e-01 3.64265032e-02 6.57231152e-01 6.86997697e-02
-8.11131477e-01 1.24763215e+00 5.25368071e+00 9.53485668e-01
-1.81481612e+00 -6.08479083e-01 6.96406603e-01 2.97350556e-01
-2.45355606e-01 -1.10036582e-01 -1.09357107e+00 6.03229821e-01
7.11012781e-01 2.31357157e-01 1.96131438e-01 8.77378643e-01
-1.72339082e-01 1.21461861e-02 -6.92632675e-01 1.14634836e+00
5.92651367e-01 -1.68306041e+00 7.49359429e-01 1.90242022e-01
7.61626244e-01 -1.92797765e-01 2.78394341e-01 2.32601419e-01
-3.58559489e-01 -1.04395211e+00 1.04287386e+00 6.88585877e-01
9.20643985e-01 -8.45640361e-01 5.61980188e-01 3.26962471e-01
-1.01320922e+00 -3.71319532e-01 -8.96090388e-01 8.25101882e-02
-1.15756609e-01 3.93012404e-01 -5.92594087e-01 4.83617157e-01
2.38378823e-01 8.11773598e-01 -7.65295029e-01 9.18231130e-01
-1.41895786e-01 6.17521763e-01 5.67835756e-02 -2.46886209e-01
4.88136947e-01 -1.80742174e-01 2.35300362e-02 1.39543033e+00
3.87438983e-01 2.69978493e-01 -1.90997243e-01 7.34284341e-01
-2.85512745e-01 1.50304198e-01 -1.15475707e-01 -5.89966834e-01
3.75621378e-01 1.03172684e+00 -8.65805387e-01 -7.54684925e-01
-3.36419642e-01 1.35632634e+00 2.73543745e-01 1.35777563e-01
-8.92491519e-01 -1.00526035e+00 9.88155827e-02 -4.65815455e-01
7.57950187e-01 -2.15288892e-01 -2.89034516e-01 -1.37319481e+00
3.75615060e-02 -1.11604750e+00 1.84135273e-01 -9.84982908e-01
-9.79825199e-01 9.15413380e-01 -7.85756290e-01 -1.23233283e+00
2.54578084e-01 -1.03121340e+00 -8.43366444e-01 1.02334952e+00
-1.44450283e+00 -1.16373539e+00 -2.21519560e-01 4.75808084e-01
9.72953379e-01 -3.84679824e-01 6.63337111e-01 4.97986048e-01
-1.22424495e+00 8.60602438e-01 3.80061865e-01 6.18264675e-01
4.07271385e-01 -1.04329181e+00 4.86377329e-01 1.18258762e+00
-8.53738189e-03 3.30137104e-01 3.13439071e-01 -1.01786983e+00
-1.70896709e+00 -1.10703838e+00 8.21528971e-01 1.69565789e-02
4.15547311e-01 -4.35421556e-01 -9.57577884e-01 2.33637154e-01
-3.88071537e-02 -3.01300228e-01 3.63156319e-01 -5.80781937e-01
-5.61416984e-01 2.09433481e-01 -7.74953008e-01 6.63241386e-01
5.66697478e-01 -6.71971500e-01 -4.45106298e-01 2.94095039e-01
3.79532337e-01 -2.63639003e-01 -4.43779826e-01 -1.53930187e-02
2.73880631e-01 -5.68795264e-01 7.44046748e-01 -4.69118208e-01
9.07454550e-01 -3.46628964e-01 -1.53515488e-01 -5.45537591e-01
-2.79229552e-01 -1.38557464e-01 -8.82094949e-02 1.11071932e+00
3.32432210e-01 -6.13430962e-02 8.28247726e-01 1.40798092e-01
-6.45670772e-01 -4.77302343e-01 -4.21363652e-01 -7.61630654e-01
1.90571412e-01 -1.05449818e-01 3.37813079e-01 8.35547090e-01
-3.55087996e-01 4.46293473e-01 -6.76147461e-01 3.99816692e-01
4.51738179e-01 1.45521879e-01 4.01671380e-01 -6.30669773e-01
4.35595438e-02 -4.00885403e-01 -5.54702461e-01 -1.35203576e+00
2.90191203e-01 -7.64100790e-01 -3.37693542e-02 -1.34457958e+00
2.89170265e-01 -2.94222504e-01 -9.76721570e-02 3.66364032e-01
2.63818335e-02 4.14402395e-01 2.44342744e-01 2.99952894e-01
-2.68294692e-01 7.06515968e-01 1.09656382e+00 -4.69101399e-01
5.71424328e-02 -3.50289702e-01 -3.51437151e-01 2.64498711e-01
7.44805872e-01 -6.87608570e-02 -6.60930127e-02 -8.54230046e-01
7.27846771e-02 6.85067624e-02 9.06662047e-02 -1.02012706e+00
3.06947857e-01 2.14168772e-01 8.22746038e-01 -1.18149817e+00
2.18809023e-01 -3.35306585e-01 9.79879126e-03 9.02331591e-01
-2.95041233e-01 2.89837897e-01 2.91076154e-01 4.49149162e-01
-2.52363384e-01 -6.51726425e-01 6.35308027e-01 -8.88120010e-02
-1.06175435e+00 4.10921350e-02 -9.01727974e-01 -3.94419014e-01
7.66419113e-01 -5.17613053e-01 -7.07178175e-01 4.38194871e-02
-3.61029834e-01 -9.05408040e-02 2.43758604e-01 3.49916279e-01
1.00296330e+00 -1.08426940e+00 -7.17067361e-01 9.98520702e-02
-1.80533733e-02 -1.22632153e-01 6.46695018e-01 2.04183429e-01
-1.44663608e+00 4.41819310e-01 -4.24595654e-01 -1.78856552e-01
-1.32256722e+00 9.71266508e-01 2.50096411e-01 -1.20934166e-01
-8.41244876e-01 7.23918021e-01 2.18055829e-01 1.72971278e-01
4.05707359e-01 -2.72388726e-01 -3.97248209e-01 2.03399971e-01
6.69595897e-01 3.92523825e-01 4.94909547e-02 -7.69870162e-01
-1.74901471e-01 7.71728337e-01 -5.38148761e-01 3.25611740e-01
1.20789254e+00 1.96464211e-01 -1.30653441e-01 1.72413588e-01
1.33800554e+00 1.66571289e-01 -1.16295040e+00 4.95412126e-02
-1.46568343e-01 -1.11759774e-01 4.00391892e-02 -7.81238735e-01
-1.23826063e+00 1.29188824e+00 6.00201130e-01 -1.84339926e-01
1.05396676e+00 -5.16823888e-01 8.84713233e-01 8.13730419e-01
-1.97055876e-01 -1.04697943e+00 4.79997009e-01 7.50383377e-01
7.47708321e-01 -8.56091082e-01 -1.12259842e-01 -2.87515581e-01
-1.02165282e+00 1.65258038e+00 6.04836464e-01 -3.66956353e-01
5.59907369e-02 -7.56231919e-02 1.57901540e-01 -1.95602864e-01
-6.00104749e-01 7.72225335e-02 9.41341966e-02 2.28802741e-01
1.12379216e-01 -1.81457505e-01 -1.32298261e-01 3.89181495e-01
1.18665725e-01 -4.13948566e-01 1.03500462e+00 8.78754020e-01
-4.92339641e-01 -1.05309701e+00 -3.12817067e-01 1.82700664e-01
-4.93659645e-01 -5.00203371e-01 -5.06888211e-01 4.97383893e-01
-1.09127045e-01 5.39157569e-01 4.60919946e-01 -3.20953935e-01
2.56144404e-02 -1.32805124e-01 4.51613456e-01 -4.77010310e-01
-6.39341056e-01 3.49868909e-02 -8.00022706e-02 1.75828487e-01
1.55138209e-01 -4.72636342e-01 -1.41457438e+00 -2.22356483e-01
-4.88398343e-01 2.88426261e-02 7.79871047e-01 6.01326585e-01
2.25252062e-01 5.96890986e-01 8.73975277e-01 -4.15861577e-01
-5.59208870e-01 -7.42045283e-01 -5.76176643e-01 1.73909098e-01
2.22862720e-01 -9.08288639e-04 -1.08649544e-01 1.45668685e-01] | [11.913246154785156, 2.191969394683838] |
f153e4b5-6f83-4c21-990c-23eb769fd0d1 | palette-image-to-image-diffusion-models-1 | 2111.05826 | null | https://arxiv.org/abs/2111.05826v2 | https://arxiv.org/pdf/2111.05826v2.pdf | Palette: Image-to-Image Diffusion Models | This paper develops a unified framework for image-to-image translation based on conditional diffusion models and evaluates this framework on four challenging image-to-image translation tasks, namely colorization, inpainting, uncropping, and JPEG restoration. Our simple implementation of image-to-image diffusion models outperforms strong GAN and regression baselines on all tasks, without task-specific hyper-parameter tuning, architecture customization, or any auxiliary loss or sophisticated new techniques needed. We uncover the impact of an L2 vs. L1 loss in the denoising diffusion objective on sample diversity, and demonstrate the importance of self-attention in the neural architecture through empirical studies. Importantly, we advocate a unified evaluation protocol based on ImageNet, with human evaluation and sample quality scores (FID, Inception Score, Classification Accuracy of a pre-trained ResNet-50, and Perceptual Distance against original images). We expect this standardized evaluation protocol to play a role in advancing image-to-image translation research. Finally, we show that a generalist, multi-task diffusion model performs as well or better than task-specific specialist counterparts. Check out https://diffusion-palette.github.io for an overview of the results. | ['Mohammad Norouzi', 'David J. Fleet', 'Tim Salimans', 'Jonathan Ho', 'Chris A. Lee', 'Huiwen Chang', 'William Chan', 'Chitwan Saharia'] | 2021-11-10 | palette-image-to-image-diffusion-models | https://openreview.net/forum?id=FPGs276lUeq | https://openreview.net/pdf?id=FPGs276lUeq | null | ['jpeg-decompression', 'uncropping'] | ['computer-vision', 'computer-vision'] | [ 4.97122526e-01 -6.57417178e-02 -2.18004078e-01 -2.21362844e-01
-1.27076197e+00 -5.66585720e-01 8.12169969e-01 -3.82634103e-01
-5.64665318e-01 6.08096898e-01 4.14831847e-01 -2.57900447e-01
3.11680615e-01 -3.27051520e-01 -8.93097878e-01 -7.35757768e-01
2.66316265e-01 2.82236010e-01 -5.22835255e-02 -1.20595090e-01
1.91431835e-01 1.08816244e-01 -9.15675879e-01 2.55731165e-01
8.98318350e-01 9.70405400e-01 2.77588993e-01 1.00089669e+00
3.97702336e-01 8.14062119e-01 -6.86243057e-01 -7.70574808e-01
4.74836618e-01 -8.39438736e-01 -6.40432894e-01 2.29771718e-01
7.70557642e-01 -5.93838930e-01 -3.27384979e-01 1.13085973e+00
9.58741426e-01 -4.39089574e-02 7.96663821e-01 -1.15901005e+00
-1.44753206e+00 2.94854283e-01 -6.37866616e-01 3.20744157e-01
6.09799810e-02 9.52659011e-01 8.23227108e-01 -7.90982723e-01
9.01331246e-01 1.21040225e+00 6.96716428e-01 6.56168640e-01
-1.48392463e+00 -6.19817913e-01 -1.87346593e-01 1.06472716e-01
-1.05014443e+00 -5.67891538e-01 6.09613538e-01 -4.42322373e-01
7.25789487e-01 8.09357464e-02 4.69529420e-01 1.61693263e+00
3.66351306e-01 6.45435989e-01 1.42987132e+00 -3.42061013e-01
5.11163287e-02 1.16191693e-01 -4.57935959e-01 6.78912878e-01
1.21207885e-01 2.08732113e-01 -4.26312953e-01 1.83732554e-01
1.09685349e+00 -4.38185930e-01 -4.23121661e-01 -2.03478724e-01
-1.21723223e+00 8.39602649e-01 4.84721452e-01 1.54929072e-01
-3.92652690e-01 6.79968953e-01 4.19950813e-01 5.37309527e-01
7.46008635e-01 4.39745098e-01 -3.00858289e-01 -1.58240363e-01
-1.11868107e+00 1.29213154e-01 4.28685635e-01 8.12215865e-01
6.18608415e-01 3.29672724e-01 -5.34567833e-01 9.02213514e-01
-1.31513581e-01 5.91869295e-01 4.48292226e-01 -1.46573865e+00
2.71941990e-01 -6.02732748e-02 -8.98095369e-02 -7.26441205e-01
-1.57960057e-02 -6.13988698e-01 -1.07593715e+00 3.90311480e-01
2.21297145e-01 -2.21788064e-01 -1.07049143e+00 1.86764228e+00
-8.21596663e-03 -1.46312816e-02 -1.59245148e-01 1.05248451e+00
5.20872056e-01 4.48693931e-01 3.40301767e-02 2.05172654e-02
1.24923706e+00 -1.34614956e+00 -4.97632980e-01 -2.52666682e-01
3.39935452e-01 -9.94529426e-01 1.57174551e+00 3.40015978e-01
-1.47853982e+00 -6.70887113e-01 -9.08159494e-01 -4.35454458e-01
-8.48974288e-03 2.41625965e-01 4.43986148e-01 7.14426816e-01
-1.56032121e+00 5.12979805e-01 -7.22333908e-01 -4.27294075e-01
6.38682067e-01 1.10025860e-01 -1.87345222e-01 -2.70730734e-01
-8.88266146e-01 1.03564799e+00 -2.45303847e-03 -1.74046651e-01
-1.29712105e+00 -8.92435610e-01 -6.14729166e-01 -2.11094737e-01
-3.37426132e-03 -1.35095215e+00 1.13239598e+00 -1.46754706e+00
-1.66243541e+00 1.07057703e+00 1.32095426e-01 -5.45544028e-01
1.11498976e+00 -1.01399139e-01 -2.72172481e-01 2.61512548e-01
2.35441893e-01 1.16672504e+00 1.20939362e+00 -1.27708316e+00
-2.49007508e-01 -7.28448555e-02 7.93588459e-02 2.64723331e-01
-2.98353702e-01 -1.34387597e-01 -8.20459962e-01 -1.04711330e+00
-4.62093413e-01 -1.07291031e+00 -1.04517147e-01 3.30094963e-01
-3.40189427e-01 2.91946948e-01 5.05910337e-01 -1.04360890e+00
7.21613228e-01 -2.25312877e+00 2.67072171e-01 -1.23868510e-01
2.07132876e-01 -2.80381199e-02 -7.57260799e-01 1.32696256e-01
5.30562811e-02 1.98641598e-01 -4.40489680e-01 -6.62327886e-01
-8.16529244e-02 3.00813410e-02 -6.48590475e-02 6.04965866e-01
3.32073539e-01 1.19596100e+00 -6.63376093e-01 -3.34860235e-01
1.74687371e-01 8.37097406e-01 -8.18569899e-01 8.09718892e-02
-2.14881197e-01 8.50945234e-01 1.41618937e-01 5.08082747e-01
6.34279311e-01 -4.32747990e-01 -2.04021454e-01 -3.84742230e-01
2.65160561e-01 8.68447311e-03 -4.84978080e-01 2.13223529e+00
-8.63698065e-01 1.00890446e+00 1.76887557e-01 -5.58846295e-01
3.99216920e-01 9.70299914e-02 2.81407744e-01 -1.01700294e+00
-1.23851486e-02 3.28031927e-01 8.37973226e-03 -3.25679958e-01
4.72890675e-01 5.86406840e-03 2.29065344e-01 4.86361176e-01
2.43510917e-01 -4.05937493e-01 3.25086385e-01 2.98197478e-01
1.00128746e+00 1.96104988e-01 -2.04645708e-01 -2.65052766e-01
1.94813699e-01 -4.35131155e-02 1.20828405e-01 7.68552601e-01
-1.78547546e-01 1.15467668e+00 4.30914700e-01 -1.02293804e-01
-1.57490492e+00 -1.04401779e+00 7.00231269e-02 1.14001000e+00
-1.67020485e-02 -6.31160289e-02 -9.92004752e-01 -5.38077593e-01
-1.62816495e-01 8.27944756e-01 -9.41854417e-01 -2.41938531e-01
-3.25699836e-01 -8.40797782e-01 7.82913744e-01 2.16245711e-01
6.65417552e-01 -9.45270598e-01 -3.30450416e-01 -1.15593225e-01
-2.96679437e-01 -1.13686442e+00 -9.98670280e-01 -1.62952736e-01
-7.85198450e-01 -7.13248253e-01 -1.37332761e+00 -7.19056368e-01
5.73706150e-01 2.33155429e-01 1.29673135e+00 5.64743504e-02
-1.78531170e-01 6.94684863e-01 -1.52008116e-01 1.36970922e-01
-7.69445002e-01 2.48383820e-01 -2.99119323e-01 -2.04810485e-01
-3.81325603e-01 -6.18413329e-01 -1.26571119e+00 3.17482322e-01
-1.09877324e+00 2.25534260e-01 8.15615773e-01 8.82743001e-01
4.51912433e-01 -3.82081926e-01 2.30405897e-01 -6.95586681e-01
1.06092811e+00 -2.41091818e-01 -3.37177843e-01 2.24258333e-01
-9.05965388e-01 5.49392179e-02 2.81585455e-01 -6.40983880e-01
-8.70943308e-01 -3.52364779e-01 -2.51888365e-01 -7.00398207e-01
2.41582841e-01 2.69294024e-01 7.54788611e-03 -2.65669256e-01
8.58610332e-01 5.44581831e-01 1.21316113e-01 -2.02282399e-01
7.92698324e-01 3.17216367e-01 6.47941768e-01 -5.31761050e-01
7.14179635e-01 4.63183492e-01 -3.10502350e-01 -5.28020978e-01
-4.26551849e-01 9.91419852e-02 -3.16184014e-01 -1.58858284e-01
1.07128012e+00 -1.23255801e+00 -3.78521919e-01 4.93297607e-01
-1.16078258e+00 -9.03595448e-01 -4.69260335e-01 2.46007696e-01
-7.16118097e-01 3.51189166e-01 -8.42980981e-01 -2.70721912e-01
-5.50878406e-01 -1.42619431e+00 1.08174992e+00 -8.24322477e-02
-1.15830980e-01 -1.16358614e+00 -2.09788582e-03 6.30710840e-01
1.04720736e+00 1.96944028e-01 8.12832713e-01 -6.43796772e-02
-6.62683547e-01 1.25138998e-01 -4.63194907e-01 7.77823269e-01
-1.09396510e-01 4.58916649e-03 -8.53677332e-01 -5.48943639e-01
-2.97218710e-02 -5.77231109e-01 1.23205221e+00 6.48510337e-01
1.03232896e+00 -4.22273904e-01 1.93400383e-01 9.78212476e-01
1.31765783e+00 -2.55933464e-01 8.68162990e-01 3.67726117e-01
7.37563789e-01 1.91429660e-01 -7.36790672e-02 1.56801537e-01
4.45386380e-01 4.95447189e-01 2.92758226e-01 -5.13109326e-01
-1.03081858e+00 -1.98600292e-01 6.34434760e-01 6.65028989e-01
-1.13047734e-01 -3.83719832e-01 -5.59836805e-01 4.89500672e-01
-1.27469289e+00 -7.12486327e-01 9.39300731e-02 1.98493087e+00
9.29681003e-01 1.41373023e-01 1.92210436e-01 -4.90303606e-01
6.78906083e-01 2.70066947e-01 -8.55297744e-01 -2.80823767e-01
-4.96588230e-01 1.70780450e-01 7.49632001e-01 6.68190539e-01
-7.91490316e-01 1.05280852e+00 6.35283852e+00 1.13422692e+00
-1.39155936e+00 7.17285037e-01 1.43541241e+00 -2.93691784e-01
-5.63212812e-01 -1.20117724e-01 -1.23865984e-01 4.63472396e-01
7.76404262e-01 -5.18573113e-02 8.67623866e-01 5.87464094e-01
2.12359473e-01 -3.43004358e-03 -9.61144209e-01 1.06802452e+00
1.79371253e-01 -1.46529531e+00 2.42022723e-01 1.86157629e-01
1.09170520e+00 4.81189162e-01 7.18805909e-01 1.31854728e-01
4.68220383e-01 -9.68743443e-01 1.02718306e+00 3.15979362e-01
1.40981221e+00 -2.99071163e-01 2.41637647e-01 -2.54502743e-01
-5.87271214e-01 1.82474837e-01 -2.67663985e-01 4.07932162e-01
2.02805355e-01 6.59486473e-01 -3.88420254e-01 3.45312536e-01
5.60747147e-01 6.79266572e-01 -7.65067995e-01 7.33286500e-01
-3.17955643e-01 6.20828032e-01 -4.59042750e-02 4.73954767e-01
1.83587998e-01 -2.69763827e-01 5.02619326e-01 1.38835180e+00
4.59588379e-01 -3.55044931e-01 -2.61089921e-01 1.02901816e+00
-5.09727597e-01 -3.52375470e-02 -2.96437711e-01 6.09515756e-02
-9.12135746e-03 1.21836030e+00 -5.97958744e-01 -3.54841113e-01
-2.65319973e-01 1.74981153e+00 7.17737973e-02 7.82831430e-01
-1.01832211e+00 4.38261740e-02 6.08618855e-01 2.09655181e-01
3.66144747e-01 -2.31300697e-01 -3.98030192e-01 -1.25153315e+00
-5.64219728e-02 -1.21309173e+00 5.61512858e-02 -1.10406303e+00
-1.43840480e+00 8.45112205e-01 -2.86626637e-01 -1.04807067e+00
-2.02266112e-01 -2.53733814e-01 -5.14681518e-01 9.09222603e-01
-1.45712507e+00 -1.33897471e+00 -3.05306852e-01 6.17723227e-01
6.52573287e-01 -1.48213759e-01 4.00948763e-01 4.51110601e-01
-4.62644458e-01 8.82188380e-01 2.45434120e-01 5.21973185e-02
1.04720616e+00 -1.02099228e+00 6.96999133e-01 9.13952887e-01
-2.46264525e-02 3.32255095e-01 7.79414177e-01 -5.53981185e-01
-1.37962770e+00 -1.17258346e+00 2.99357921e-01 -5.21536112e-01
6.81286693e-01 -2.56837875e-01 -5.04308462e-01 5.50507128e-01
8.67700338e-01 -3.31847258e-02 3.11866015e-01 -2.55782038e-01
-4.76279080e-01 -3.73693891e-02 -1.20020354e+00 7.35325933e-01
1.16505003e+00 -5.72353542e-01 2.88406074e-01 4.96610492e-01
8.81904900e-01 -4.15620834e-01 -9.16293323e-01 5.18779084e-02
5.71220577e-01 -1.10599422e+00 1.06977451e+00 -2.16133818e-01
9.80523467e-01 4.27046940e-02 -2.49498039e-01 -1.42863476e+00
-4.37472731e-01 -7.74101734e-01 2.24686325e-01 1.07210398e+00
5.31705618e-01 -5.21673203e-01 6.21109545e-01 4.17118698e-01
-8.44152942e-02 -5.93336642e-01 -8.89651835e-01 -7.56182075e-01
5.00681639e-01 -3.05519015e-01 1.38393477e-01 9.56651926e-01
-7.62280345e-01 4.29027885e-01 -7.11469769e-01 -2.43446201e-01
7.05726922e-01 -2.48702571e-01 9.06641483e-01 -4.41123128e-01
-6.61856592e-01 -9.31733966e-01 -7.69308805e-02 -1.14228225e+00
-5.86610325e-02 -9.29366767e-01 -2.95433402e-01 -1.57517660e+00
2.50338316e-01 -2.04584837e-01 -1.28086224e-01 2.07962573e-01
-2.07606673e-01 8.25785577e-01 4.75182921e-01 6.01483822e-01
-5.45416057e-01 7.33911812e-01 1.70327091e+00 -4.42573696e-01
-1.34963999e-02 -3.83238494e-01 -9.03244197e-01 2.56053597e-01
7.52994120e-01 -5.23879349e-01 -4.31031764e-01 -9.81228113e-01
1.23281986e-01 -1.78669374e-02 4.91741538e-01 -1.02028120e+00
-1.37400357e-02 4.86943685e-02 4.73731250e-01 2.15010777e-01
4.28049386e-01 -4.83571500e-01 3.05771828e-01 6.31458282e-01
-5.77374101e-01 3.64951134e-01 1.40372038e-01 4.37137753e-01
-1.96802039e-02 6.89123571e-02 9.46661413e-01 -8.72583613e-02
-4.69266027e-01 2.84758866e-01 -1.22040480e-01 3.21409047e-01
7.14488089e-01 -1.90563455e-01 -4.59308743e-01 -8.55207980e-01
-6.11757934e-01 -1.12665072e-01 8.63973558e-01 4.14586097e-01
4.42853779e-01 -1.22023797e+00 -1.21922684e+00 6.26982078e-02
2.64928304e-02 -5.30038536e-01 3.85188729e-01 1.06214857e+00
-7.39633143e-01 3.77248763e-03 -3.89604390e-01 -5.90111971e-01
-8.80806029e-01 3.71573657e-01 4.26401883e-01 -2.95877397e-01
-5.30265927e-01 9.28447723e-01 3.52113962e-01 -2.86305159e-01
1.35544285e-01 -1.86102122e-01 6.23930156e-01 -2.79102802e-01
2.06869990e-01 2.09552586e-01 -2.16784049e-03 -5.32979786e-01
-1.04813822e-01 5.76918185e-01 -1.80772200e-01 -5.62056124e-01
1.09655178e+00 -2.95664698e-01 5.58338910e-02 6.73113614e-02
1.25855589e+00 -1.78363726e-01 -1.68832576e+00 -8.09529573e-02
-5.72557569e-01 -4.60590124e-01 1.91458181e-01 -1.15520656e+00
-1.49382746e+00 8.59762251e-01 1.03287315e+00 -1.36553824e-01
1.33526099e+00 -8.78045633e-02 8.22248459e-01 -1.22490466e-01
2.03353129e-02 -8.47697854e-01 4.73661304e-01 2.04347894e-01
1.24457181e+00 -1.36502945e+00 -4.93453667e-02 -2.21555885e-02
-1.02515888e+00 6.51472986e-01 3.55355233e-01 -2.19486400e-01
4.30580288e-01 2.09670231e-01 3.00786674e-01 1.97995976e-01
-7.29717910e-01 -9.01078880e-02 2.94842929e-01 7.22426713e-01
5.16823888e-01 -1.68539897e-01 -2.54693329e-01 3.08316923e-03
-2.36374170e-01 9.30364057e-02 4.36607987e-01 4.50266451e-01
-1.05838953e-02 -1.01028907e+00 -1.82590783e-01 2.63934523e-01
-5.02454877e-01 -5.14884055e-01 -2.14202389e-01 6.70183539e-01
8.13997909e-03 7.82975018e-01 8.27058405e-02 -3.51179212e-01
7.98625350e-02 -2.88973004e-01 8.01583290e-01 -1.15706995e-01
-7.82333970e-01 2.46885687e-01 -1.06325537e-01 -5.00252962e-01
-3.67054850e-01 -4.71130192e-01 -5.80436945e-01 -6.40072644e-01
3.30079198e-02 -3.22536737e-01 9.74960744e-01 5.06714582e-01
7.39347517e-01 5.56201816e-01 3.51715833e-01 -1.01803064e+00
-4.83458281e-01 -1.05345798e+00 -1.90761611e-01 6.52005017e-01
3.20696861e-01 -1.98115245e-01 -2.43933290e-01 2.89129853e-01] | [11.466636657714844, -0.26934704184532166] |
4b85e02f-df5e-407a-ba64-f36868b8c08c | accurate-airway-tree-segmentation-in-ct-scans | 2306.09116 | null | https://arxiv.org/abs/2306.09116v1 | https://arxiv.org/pdf/2306.09116v1.pdf | Accurate Airway Tree Segmentation in CT Scans via Anatomy-aware Multi-class Segmentation and Topology-guided Iterative Learning | Intrathoracic airway segmentation in computed tomography (CT) is a prerequisite for various respiratory disease analyses such as chronic obstructive pulmonary disease (COPD), asthma and lung cancer. Unlike other organs with simpler shapes or topology, the airway's complex tree structure imposes an unbearable burden to generate the "ground truth" label (up to 7 or 3 hours of manual or semi-automatic annotation on each case). Most of the existing airway datasets are incompletely labeled/annotated, thus limiting the completeness of computer-segmented airway. In this paper, we propose a new anatomy-aware multi-class airway segmentation method enhanced by topology-guided iterative self-learning. Based on the natural airway anatomy, we formulate a simple yet highly effective anatomy-aware multi-class segmentation task to intuitively handle the severe intra-class imbalance of the airway. To solve the incomplete labeling issue, we propose a tailored self-iterative learning scheme to segment toward the complete airway tree. For generating pseudo-labels to achieve higher sensitivity , we introduce a novel breakage attention map and design a topology-guided pseudo-label refinement method by iteratively connecting breaking branches commonly existed from initial pseudo-labels. Extensive experiments have been conducted on four datasets including two public challenges. The proposed method ranked 1st in both EXACT'09 challenge using average score and ATM'22 challenge on weighted average score. In a public BAS dataset and a private lung cancer dataset, our method significantly improves previous leading approaches by extracting at least (absolute) 7.5% more detected tree length and 4.0% more tree branches, while maintaining similar precision. | ['Dakai Jin', 'Xianghua Ye', 'Le Lu', 'Yun Gu', 'Jia Ge', 'Xin Sun', 'Haogang Yu', 'Minghui Zhang', 'Dandan Zheng', 'Dazhou Guo', 'Puyang Wang'] | 2023-06-15 | null | null | null | null | ['computed-tomography-ct', 'anatomy', 'pseudo-label', 'self-learning'] | ['methodology', 'miscellaneous', 'miscellaneous', 'natural-language-processing'] | [ 5.14485061e-01 4.09875512e-01 -2.83844024e-01 -3.14444929e-01
-1.24514699e+00 -7.55614638e-01 -8.06108303e-03 1.53687567e-01
-1.83724746e-01 6.44108891e-01 9.18329805e-02 -6.39611781e-01
-2.08449185e-01 -6.33079171e-01 -4.92486238e-01 -6.55656397e-01
1.74344227e-01 1.02629995e+00 5.55984616e-01 2.23498523e-01
-1.91944078e-01 3.56269330e-01 -9.88750041e-01 3.10336500e-01
1.11596668e+00 9.12475467e-01 1.10221036e-01 8.12684417e-01
-5.47651410e-01 4.41792339e-01 -2.96335101e-01 -3.96043658e-01
4.67853665e-01 -3.75431448e-01 -1.24581969e+00 1.91900387e-01
6.89385831e-01 -1.62802830e-01 1.62186727e-01 6.72634184e-01
6.77580833e-01 -1.84271380e-01 7.03639567e-01 -9.88886476e-01
4.48226593e-02 5.84361792e-01 -7.10705757e-01 4.31890070e-01
-1.13502458e-01 2.38896117e-01 1.10427356e+00 -6.33635998e-01
3.33104163e-01 8.41617465e-01 1.00796533e+00 6.95820630e-01
-9.99922693e-01 -6.79538131e-01 9.92696509e-02 -2.05695510e-01
-1.22585750e+00 7.48110414e-02 6.50036454e-01 -4.69177395e-01
4.51379687e-01 5.47318280e-01 6.60305381e-01 7.05139041e-01
-1.31928191e-01 4.39393282e-01 1.07623184e+00 -1.11068055e-01
-6.82146326e-02 7.35078529e-02 -9.83887911e-02 1.11041605e+00
2.88000375e-01 -5.88150509e-02 2.33895510e-01 -3.44668418e-01
6.38939917e-01 -6.24996945e-02 -2.33203426e-01 -4.11698759e-01
-1.25998998e+00 6.65031970e-01 7.04130471e-01 2.83847213e-01
-2.39469647e-01 5.82118668e-02 3.48727256e-01 -3.63190502e-01
3.51830304e-01 3.54774952e-01 -6.07900202e-01 1.74293160e-01
-8.61863673e-01 2.58455183e-02 6.58082664e-01 8.24487627e-01
4.49593455e-01 -2.79470354e-01 -3.64275008e-01 7.75545657e-01
3.95975322e-01 4.32549536e-01 5.96826136e-01 -8.31019640e-01
5.90885580e-01 7.65979946e-01 -2.19099835e-01 -4.64619786e-01
-7.32592940e-01 -8.47679377e-01 -1.10826063e+00 -2.38718554e-01
5.66880286e-01 -1.24935478e-01 -1.26695752e+00 1.49459755e+00
1.01492059e+00 4.33330655e-01 -4.75019991e-01 8.03701878e-01
9.76542115e-01 2.51867652e-01 3.54481876e-01 -2.30548963e-01
1.58810997e+00 -1.20420885e+00 -3.20260018e-01 4.76462729e-02
8.45858634e-01 -6.82601750e-01 1.14365566e+00 1.30090833e-01
-9.60424840e-01 -4.12184298e-01 -7.18980610e-01 7.23384321e-02
-1.73190124e-02 -1.26352236e-01 4.03244674e-01 7.75060058e-01
-8.51399422e-01 5.07253528e-01 -8.05764616e-01 8.36687721e-03
8.92472684e-01 5.21029234e-01 -1.09280333e-01 4.50462885e-02
-7.98842311e-01 5.80902219e-01 2.78606176e-01 -1.58041745e-01
-5.92898428e-01 -1.23211181e+00 -4.97784346e-01 -1.97786897e-01
6.26257896e-01 -1.31510353e+00 1.46098793e+00 -3.66036713e-01
-1.18497050e+00 9.71056461e-01 -1.94466516e-01 -3.80999833e-01
8.85838211e-01 -1.55345546e-02 -4.24299464e-02 4.28581506e-01
7.79821277e-02 8.07007313e-01 6.35085225e-01 -1.38628626e+00
-8.21294546e-01 -4.16167915e-01 -2.08382860e-01 2.90406317e-01
-5.28262183e-02 -3.43754053e-01 -2.97332913e-01 -7.13679492e-01
3.15713525e-01 -1.15641510e+00 -7.21067429e-01 1.82787746e-01
-5.15616477e-01 -3.02016705e-01 8.86748612e-01 -4.97823507e-01
1.35583806e+00 -1.66954136e+00 -1.26091704e-01 1.25137642e-01
6.71044171e-01 3.76369506e-01 1.26115397e-01 -3.28478962e-01
-5.45687005e-02 8.51503432e-01 -8.51595104e-01 -4.10504222e-01
-4.32480663e-01 4.14557397e-01 4.81787361e-02 2.41625100e-01
1.08020015e-01 9.87500072e-01 -1.01246643e+00 -1.22886205e+00
2.29634166e-01 4.58590299e-01 -6.90402865e-01 1.39338210e-01
-2.57073134e-01 1.01563668e+00 -7.94314742e-01 7.21552730e-01
5.14194191e-01 -6.14964545e-01 -1.29858866e-01 -3.18326652e-01
1.14735417e-01 3.81475240e-01 -8.78148615e-01 1.61327314e+00
-7.32196331e-01 -8.22594836e-02 1.26471192e-01 -6.31889045e-01
4.94234741e-01 6.59936130e-01 1.08169997e+00 -2.63070881e-01
2.56286383e-01 3.10059100e-01 1.15326025e-01 -5.69860518e-01
-2.15125799e-01 -5.78122795e-01 1.62025094e-01 4.63428319e-01
-4.23338890e-01 -6.31553829e-01 3.81037928e-02 -2.80112606e-02
1.14357400e+00 -1.18267544e-01 2.62758374e-01 -2.12289318e-01
8.13287199e-01 2.20236003e-01 6.20744288e-01 4.36171919e-01
-5.49010754e-01 1.00818801e+00 7.58067891e-02 -5.86197913e-01
-1.00226867e+00 -1.13021994e+00 -4.94462669e-01 7.33499527e-01
-3.52259353e-02 -1.14294745e-01 -8.42785716e-01 -1.30543733e+00
-9.29260105e-02 3.81032854e-01 -3.58680665e-01 1.36465326e-01
-9.99263525e-01 -1.05108964e+00 6.03883743e-01 5.26520133e-01
6.05790257e-01 -9.74578559e-01 -5.13804197e-01 2.56491780e-01
-6.11110866e-01 -1.23005533e+00 -6.96198463e-01 -3.29866298e-02
-1.25003612e+00 -1.17409694e+00 -7.12142289e-01 -7.24785507e-01
8.69807482e-01 1.56600967e-01 1.23698378e+00 4.22667027e-01
-7.18451321e-01 1.39130458e-01 -1.42117411e-01 -3.72631073e-01
-5.06170571e-01 4.75320727e-01 -2.73063451e-01 -1.34620130e-01
-3.03959042e-01 -5.43415248e-01 -9.51907516e-01 5.65765738e-01
-6.64186358e-01 1.05353318e-01 7.06653774e-01 8.11613977e-01
1.10475361e+00 2.21890166e-01 6.99661672e-01 -1.25967121e+00
2.60040939e-01 -5.19955754e-01 -3.65761161e-01 1.87839493e-01
-7.96444237e-01 -1.26339674e-01 6.55796289e-01 -2.65652031e-01
-1.07523429e+00 3.25902194e-01 -4.85743225e-01 -3.77150953e-01
-2.53934383e-01 -6.25795051e-02 1.36577189e-01 -1.15255587e-01
5.49328387e-01 -6.00560084e-02 -2.34707713e-01 -5.23262262e-01
3.48461360e-01 5.60728788e-01 5.49109042e-01 -7.07770824e-01
1.04257238e+00 5.24309397e-01 3.78141284e-01 -3.67861480e-01
-1.32635653e+00 -8.24419618e-01 -9.96034324e-01 7.63420910e-02
1.02509964e+00 -6.23084784e-01 -4.02054727e-01 1.23733595e-01
-8.32964182e-01 -3.84888232e-01 -6.47940516e-01 2.73930281e-01
-3.49968493e-01 6.36089027e-01 -4.45485294e-01 -5.23631752e-01
-9.63102520e-01 -1.20500255e+00 1.34959674e+00 7.40139037e-02
-3.20322245e-01 -1.01929736e+00 3.06779653e-01 8.88481379e-01
3.86592716e-01 4.42836434e-01 1.12905025e+00 -6.60895467e-01
-5.82714140e-01 9.82513800e-02 -3.72282863e-01 2.35420600e-01
3.82931590e-01 -6.47381395e-02 -8.57850850e-01 -1.47745639e-01
-2.60569565e-02 -1.07294925e-01 8.01855683e-01 3.41653913e-01
1.68293202e+00 -2.57047266e-01 -5.02871573e-01 7.03131139e-01
1.25576305e+00 1.23462804e-01 7.76562607e-03 -1.63539320e-01
1.05967426e+00 7.07340419e-01 5.71419537e-01 2.38379985e-01
4.10376549e-01 3.95241618e-01 5.84497333e-01 -4.04081702e-01
-5.17542958e-01 -1.53583437e-01 -3.79079878e-01 8.54528368e-01
4.58818860e-02 -2.93807030e-01 -1.20518887e+00 6.77791655e-01
-1.22659743e+00 -5.12190878e-01 -5.65492272e-01 2.10350728e+00
1.05624914e+00 1.91833869e-01 2.52766103e-01 1.05612054e-01
7.40455747e-01 -1.38244748e-01 -6.31066680e-01 -6.78085759e-02
4.71135765e-01 6.42176270e-01 4.29819673e-01 4.40484285e-01
-1.28958285e+00 6.08438671e-01 5.95033312e+00 8.18564475e-01
-9.94302750e-01 3.26552451e-01 9.21522796e-01 -5.73835010e-03
-3.98149669e-01 -2.17068180e-01 -8.06800365e-01 4.50647324e-01
7.63628006e-01 1.47348857e-02 -8.07912350e-02 8.17163587e-01
8.91825557e-02 1.68041810e-01 -7.57075906e-01 6.73195720e-01
-3.65877002e-01 -1.11934817e+00 1.30349353e-01 1.95394292e-01
8.80733907e-01 2.61236012e-01 -6.66751526e-03 1.46624207e-01
2.09335327e-01 -9.25984204e-01 9.88036618e-02 -4.39907722e-02
1.23523033e+00 -4.15931523e-01 6.64725602e-01 3.73103946e-01
-1.51913106e+00 5.47715509e-03 3.39883454e-02 6.79393768e-01
3.49808037e-01 7.84284592e-01 -1.41471958e+00 6.79152310e-01
6.21299446e-01 1.22373171e-01 -5.64039588e-01 1.35975218e+00
-6.73219711e-02 8.97306561e-01 -6.94154620e-01 3.26215893e-01
2.57550031e-01 2.57705245e-02 6.52033567e-01 1.00905156e+00
3.34506810e-01 2.94573635e-01 3.91346604e-01 6.44459903e-01
-3.44921291e-01 4.98182505e-01 -4.01401758e-01 3.32138211e-01
4.49552834e-01 1.67840183e+00 -1.00423479e+00 -4.95219499e-01
-2.16226459e-01 5.23808181e-01 4.08603251e-03 -1.91565201e-01
-1.18050706e+00 3.33264500e-01 6.20359518e-02 5.78354001e-01
8.42184648e-02 3.01432639e-01 -5.74527204e-01 -6.22456193e-01
-1.56896189e-01 -5.62671661e-01 7.02491760e-01 -3.90849531e-01
-1.19398069e+00 7.37103939e-01 -4.98225316e-02 -1.20332837e+00
-1.38162915e-02 -3.35692137e-01 -7.86509991e-01 5.22044718e-01
-1.65603077e+00 -1.28365636e+00 -6.88844562e-01 2.39776850e-01
7.54373550e-01 3.65761429e-01 7.32601225e-01 6.34085774e-01
-6.63553059e-01 5.79485238e-01 -5.26347876e-01 -1.18752412e-01
6.13121092e-01 -1.68205404e+00 3.51098120e-01 3.11831027e-01
-1.39779598e-02 2.66944110e-01 2.27083445e-01 -7.18976378e-01
-6.59631312e-01 -1.59957421e+00 6.45585179e-01 -7.35757530e-01
4.14795786e-01 1.21055730e-01 -1.03134179e+00 5.32581508e-01
-1.96433589e-01 4.59535778e-01 7.58108318e-01 -3.10410976e-01
-3.54215093e-02 -6.30567297e-02 -1.50304461e+00 4.02643532e-01
1.32782876e+00 4.23679464e-02 -5.17967880e-01 6.79676712e-01
1.01655257e+00 -6.40433669e-01 -1.13792729e+00 1.18827391e+00
4.65401620e-01 -7.43678391e-01 1.19010198e+00 -4.34149206e-01
2.15858787e-01 -3.95268053e-01 2.76592940e-01 -7.68934011e-01
-2.52968103e-01 -6.17740095e-01 -6.05185479e-02 1.10087526e+00
5.47758818e-01 -5.99769413e-01 1.27639544e+00 3.81510884e-01
-5.34714639e-01 -1.40431583e+00 -1.08697391e+00 -3.82721454e-01
3.11093032e-01 -2.36990541e-01 5.41345298e-01 8.55483234e-01
-7.57914424e-01 2.39312917e-01 2.18800828e-01 8.50231946e-02
6.28011525e-01 1.30329490e-01 6.10200584e-01 -1.42799556e+00
-2.13606894e-01 -6.16301179e-01 1.56622216e-01 -8.60834062e-01
-1.65065601e-01 -1.15571928e+00 1.10482210e-02 -1.65077078e+00
2.46792153e-01 -1.04931056e+00 -3.74913573e-01 4.30626094e-01
-4.79125500e-01 3.11895788e-01 -2.76150078e-01 8.65849778e-02
-4.30320621e-01 2.05911621e-01 1.91386938e+00 -5.84167279e-02
-6.85680881e-02 5.15000343e-01 -6.87344611e-01 1.00953722e+00
9.20575857e-01 -7.68099427e-01 -6.36629105e-01 -1.72568053e-01
-1.02049269e-01 3.63325588e-02 2.73826569e-01 -8.80552113e-01
2.14927495e-02 5.47427200e-02 1.23435222e-01 -9.68209684e-01
-4.12816331e-02 -9.61650252e-01 8.35384130e-02 8.64332736e-01
-1.49296179e-01 -1.16955684e-02 1.33679703e-01 6.31788671e-01
6.99539036e-02 -3.12588394e-01 1.05978370e+00 -4.47208077e-01
-1.09809332e-01 7.46622026e-01 4.09067981e-02 6.19994998e-01
1.13424885e+00 -1.44266769e-01 -1.59268245e-01 7.56435841e-02
-8.83508801e-01 6.36164486e-01 9.95197445e-02 3.06052491e-02
3.48161221e-01 -9.00764287e-01 -7.21688211e-01 7.06158951e-02
-5.90933226e-02 1.01695549e+00 4.05354410e-01 1.11215866e+00
-7.51430273e-01 4.32380497e-01 3.79196227e-01 -9.88386810e-01
-1.34035003e+00 3.38170469e-01 6.66486204e-01 -9.61384892e-01
-6.95463181e-01 1.12755644e+00 4.65647519e-01 -7.53746390e-01
-6.66865474e-03 -7.02255964e-01 8.08678418e-02 -2.02500969e-01
-9.42075774e-02 5.00299513e-01 1.58251449e-01 -5.94266891e-01
-2.89174616e-01 1.12585032e+00 -8.75412300e-02 3.09529394e-01
7.74146318e-01 -1.09054923e-01 7.82831609e-02 -1.81516516e-03
1.10046256e+00 1.61403999e-01 -9.11536515e-01 -2.06648245e-01
-1.22676883e-02 -1.81540847e-01 -2.70069931e-02 -9.35926497e-01
-1.20540202e+00 1.02814925e+00 6.47976339e-01 5.35647795e-02
1.07286537e+00 9.66066793e-02 1.46928012e+00 9.00757611e-02
2.09671393e-01 -6.84997737e-01 1.93245649e-01 -4.45822440e-02
7.16425836e-01 -1.45641828e+00 1.74810812e-01 -1.09643567e+00
-3.73887420e-01 7.42571414e-01 8.66214693e-01 3.38536263e-01
7.40854025e-01 2.71049440e-01 2.25236729e-01 -2.84230709e-01
-6.09073460e-01 -2.29168922e-01 4.95887011e-01 3.57746124e-01
3.12306255e-01 6.02781624e-02 -1.95512086e-01 4.90874380e-01
-4.05906379e-01 -3.02527815e-01 -1.67354550e-02 5.92080295e-01
-4.86401618e-01 -1.23325598e+00 -3.18937659e-01 7.09106147e-01
-8.68952155e-01 -2.95564923e-02 -3.26815903e-01 9.06852067e-01
4.07641917e-01 6.13750577e-01 -4.36149575e-02 -5.40046692e-02
2.80134439e-01 1.82761967e-01 3.64044577e-01 -9.16451931e-01
-8.31372797e-01 2.90209889e-01 1.33581206e-01 -3.34852934e-01
-2.71330237e-01 -5.17851353e-01 -1.82519984e+00 1.90299496e-01
-5.78668714e-01 -7.74946017e-03 4.58515972e-01 7.72806525e-01
2.66003102e-01 8.90915453e-01 7.24244416e-01 -2.68663257e-01
-6.40339494e-01 -8.24530721e-01 6.08086679e-03 6.03803813e-01
2.18137190e-01 -5.91404796e-01 -5.54228783e-01 2.28523649e-02] | [14.975198745727539, -2.184054374694824] |
ddfa30c7-af9e-44aa-9176-4a88785f3e46 | u-net-fixed-point-quantization-for-medical | 1908.01073 | null | https://arxiv.org/abs/1908.01073v2 | https://arxiv.org/pdf/1908.01073v2.pdf | U-Net Fixed-Point Quantization for Medical Image Segmentation | Model quantization is leveraged to reduce the memory consumption and the computation time of deep neural networks. This is achieved by representing weights and activations with a lower bit resolution when compared to their high precision floating point counterparts. The suitable level of quantization is directly related to the model performance. Lowering the quantization precision (e.g. 2 bits), reduces the amount of memory required to store model parameters and the amount of logic required to implement computational blocks, which contributes to reducing the power consumption of the entire system. These benefits typically come at the cost of reduced accuracy. The main challenge is to quantize a network as much as possible, while maintaining the performance accuracy. In this work, we present a quantization method for the U-Net architecture, a popular model in medical image segmentation. We then apply our quantization algorithm to three datasets: (1) the Spinal Cord Gray Matter Segmentation (GM), (2) the ISBI challenge for segmentation of neuronal structures in Electron Microscopic (EM), and (3) the public National Institute of Health (NIH) dataset for pancreas segmentation in abdominal CT scans. The reported results demonstrate that with only 4 bits for weights and 6 bits for activations, we obtain 8 fold reduction in memory requirements while loosing only 2.21%, 0.57% and 2.09% dice overlap score for EM, GM and NIH datasets respectively. Our fixed point quantization provides a flexible trade off between accuracy and memory requirement which is not provided by previous quantization methods for U-Net such as TernaryNet. | ['Jean-Pierre David', 'Julien Cohen-Adad', 'Yvon Savaria', 'Lucas Rouhier', 'Sina Honari', 'MohammadHossein AskariHemmat', 'Christian S. Perone'] | 2019-08-02 | null | null | null | null | ['unet-quantization', 'pancreas-segmentation'] | ['computer-vision', 'medical'] | [ 4.83575821e-01 9.70810354e-02 -1.48137659e-01 -2.89568275e-01
-6.70868993e-01 -4.15940374e-01 1.01818211e-01 6.32771850e-01
-9.71091926e-01 7.08936751e-01 -3.57029349e-01 -4.15387630e-01
1.80016682e-02 -9.08182800e-01 -6.97770655e-01 -7.56734848e-01
-1.54235493e-02 2.48210207e-01 4.65604126e-01 6.87951893e-02
2.56587356e-01 5.08857727e-01 -1.05734515e+00 2.60591269e-01
6.83303237e-01 1.54337728e+00 1.77505955e-01 5.67747056e-01
-7.29759261e-02 4.62733924e-01 -5.08286059e-01 -2.62774140e-01
3.62292349e-01 -1.01238579e-01 -6.92813694e-01 -2.36530289e-01
3.93201202e-01 -3.38707894e-01 -8.11362267e-02 1.43484986e+00
4.75176960e-01 -1.49479993e-02 6.86149120e-01 -1.00696337e+00
-1.25679255e-01 7.97743618e-01 -5.81815183e-01 4.38773841e-01
-4.11007315e-01 8.69357437e-02 5.89889884e-01 -6.49202317e-02
4.46486503e-01 1.09850895e+00 7.89981663e-01 5.65131605e-01
-1.31400120e+00 -7.26293206e-01 -2.08405450e-01 5.43154329e-02
-1.47353554e+00 -2.69070536e-01 2.65036821e-01 -3.50964189e-01
1.20220387e+00 1.44582674e-01 8.52009714e-01 4.96329546e-01
6.76398098e-01 1.97058633e-01 1.02318907e+00 -2.18158782e-01
6.64719939e-01 -2.35319212e-01 3.76687199e-01 7.20247209e-01
4.46949482e-01 -3.46567750e-01 -2.87328690e-01 -8.61702785e-02
9.71724749e-01 5.81830107e-02 -2.65972376e-01 1.33760661e-01
-9.19725060e-01 8.88879657e-01 7.35687912e-01 3.15363675e-01
-4.20126796e-01 7.54770339e-01 6.38048112e-01 1.08268075e-01
-9.89510790e-02 2.08857566e-01 -3.39912295e-01 -1.21782348e-01
-1.23684716e+00 8.93136188e-02 5.04192770e-01 6.80901587e-01
5.41516542e-01 1.87392294e-01 -7.86823556e-02 5.65106511e-01
5.16448319e-01 1.86693370e-01 8.01248431e-01 -1.10164452e+00
2.97619015e-01 8.02540421e-01 -2.84506291e-01 -8.22934687e-01
-6.71553910e-01 -3.63258749e-01 -1.17903149e+00 3.36629242e-01
3.89189750e-01 -4.23252620e-02 -1.33794391e+00 1.74526262e+00
2.53090620e-01 -1.51257202e-01 -2.01873168e-01 7.48065114e-01
7.28484809e-01 4.52183992e-01 1.47104993e-01 -5.69414310e-02
1.65631533e+00 -6.26440942e-01 -6.28119290e-01 -4.67805266e-02
7.02933431e-01 -5.79923034e-01 8.18941951e-01 3.75835449e-01
-1.27927411e+00 -1.54025003e-01 -1.51577306e+00 -2.66301334e-01
-4.31492269e-01 -7.85791203e-02 4.85176146e-01 8.55199695e-01
-1.21659839e+00 8.04277718e-01 -1.42590010e+00 -2.11147810e-04
7.35624909e-01 1.05343068e+00 -1.03250377e-01 3.26224327e-01
-1.05311441e+00 8.18672776e-01 5.05532444e-01 -1.36631191e-01
-4.66049045e-01 -8.69399905e-01 -6.22548997e-01 2.90825099e-01
-3.39245200e-01 -4.60594803e-01 1.10888362e+00 -7.50304282e-01
-1.41049480e+00 6.81287825e-01 1.22076742e-01 -1.08675063e+00
3.69423509e-01 3.74159604e-01 -2.81742699e-02 4.24082041e-01
-2.16736138e-01 1.42931402e+00 4.99642789e-01 -6.62701368e-01
-4.91135478e-01 -5.05228162e-01 -5.74565381e-02 -5.99300042e-02
-3.32418054e-01 -3.09239417e-01 -2.82788217e-01 -5.30198753e-01
3.72916400e-01 -9.98277187e-01 -4.01834965e-01 3.41598362e-01
-9.18848813e-02 1.74064547e-01 5.99095225e-01 -6.07603669e-01
1.07327414e+00 -1.96211159e+00 -2.32318565e-01 3.97361010e-01
2.03358397e-01 3.39059740e-01 3.26848596e-01 -2.89499640e-01
9.91043150e-02 2.38567606e-01 -4.44593817e-01 -2.93616345e-03
-2.42416337e-01 3.28961372e-01 2.01041564e-01 6.33048952e-01
2.13530064e-02 9.19521987e-01 -3.71443212e-01 -6.30902112e-01
2.07606837e-01 7.96653926e-01 -8.56470704e-01 -4.32903856e-01
9.59054902e-02 4.20408659e-02 -1.21528327e-01 4.57976401e-01
6.99405909e-01 -3.55601311e-01 2.41271153e-01 -6.59174204e-01
1.24996351e-02 2.97227174e-01 -1.21422625e+00 1.74802971e+00
-2.24948183e-01 4.17671710e-01 3.11872244e-01 -9.71031189e-01
7.56493032e-01 4.37621713e-01 3.74575078e-01 -9.87617075e-01
5.90122879e-01 1.86652347e-01 3.30588490e-01 1.28381059e-01
2.11041987e-01 -3.32541555e-01 -3.89680825e-02 1.39921144e-01
1.45196915e-01 2.91107949e-02 3.10927540e-01 -1.50224432e-01
9.69055951e-01 -3.98911089e-01 3.68340611e-01 -6.57536805e-01
2.28271589e-01 1.64044768e-01 6.29525840e-01 3.93555492e-01
-4.75833833e-01 4.40720260e-01 6.09380543e-01 -2.89965928e-01
-1.01832175e+00 -1.00100815e+00 -4.89138067e-01 5.72627544e-01
-2.82799313e-03 -1.48823053e-01 -1.20554233e+00 -2.66160071e-02
-2.18603835e-01 3.00420970e-01 -5.52326262e-01 -8.66892263e-02
-7.48763204e-01 -9.95995998e-01 9.35133576e-01 6.31213188e-01
5.76190889e-01 -6.74112737e-01 -1.41368461e+00 2.68708169e-01
1.34203613e-01 -1.05646718e+00 -2.72201002e-01 8.01273108e-01
-1.42520225e+00 -7.20065713e-01 -5.80986917e-01 -7.76315928e-01
8.39055061e-01 -3.57173443e-01 8.75165701e-01 1.36172861e-01
-3.88665617e-01 -3.01043659e-01 1.29797935e-01 -3.31643671e-01
-3.20665985e-01 1.85425937e-01 -1.11888163e-02 -5.22954047e-01
2.63757348e-01 -6.02525532e-01 -9.24268186e-01 -8.17688331e-02
-1.23149860e+00 1.24350548e-01 6.33604765e-01 1.00651956e+00
9.68274653e-01 7.43098110e-02 5.35083354e-01 -6.54996157e-01
5.09069562e-01 -2.10989505e-01 -8.00951481e-01 -9.65441689e-02
-7.62753665e-01 2.14502335e-01 6.41496897e-01 -4.65211749e-01
-4.30778474e-01 2.19366029e-01 -2.63059825e-01 -3.43485415e-01
1.95424885e-01 2.29045734e-01 3.80496323e-01 -4.30412799e-01
5.91249406e-01 -9.93425474e-02 2.03413159e-01 -2.18946829e-01
5.22332862e-02 5.47217607e-01 5.59280157e-01 -2.26145744e-01
9.18613896e-02 4.78917092e-01 4.38428372e-01 -4.93492842e-01
-9.52490270e-02 -1.39027566e-01 -6.27049446e-01 1.63116202e-01
1.17547882e+00 -7.74968445e-01 -9.78457451e-01 7.22520798e-02
-9.82839644e-01 -3.02280456e-01 -9.48277712e-02 3.72299403e-01
-4.81101632e-01 2.08734512e-01 -1.01464224e+00 -3.88876766e-01
-1.05521834e+00 -1.63258934e+00 6.77062333e-01 3.35004985e-01
-4.10600811e-01 -8.32784414e-01 -4.62694287e-01 2.07434103e-01
5.97032845e-01 4.84823376e-01 1.40216517e+00 -2.42014363e-01
-3.47502142e-01 1.15892570e-03 -3.73569131e-01 3.03633898e-01
-1.99903492e-02 -3.10593337e-01 -7.24197805e-01 -2.53254861e-01
1.84388757e-01 -2.46761456e-01 6.77810967e-01 7.21736550e-01
1.12304485e+00 -3.37792218e-01 -1.44060224e-01 7.20656753e-01
1.68297410e+00 4.72306937e-01 5.16547620e-01 2.06471771e-01
4.83059824e-01 1.11849532e-01 1.58454984e-01 2.27271959e-01
7.89344907e-02 5.07485807e-01 4.67238635e-01 -1.78066164e-01
-2.67237872e-01 1.76866382e-01 -4.24988866e-02 9.57078040e-01
2.88074493e-01 3.07692364e-02 -1.17105937e+00 5.21958768e-01
-1.48658240e+00 -4.12325263e-01 -5.76520674e-02 1.99755371e+00
1.21883869e+00 3.29415023e-01 -1.41734913e-01 5.25549948e-01
5.05955100e-01 -3.46225709e-01 -5.48708498e-01 -8.49541903e-01
3.34929049e-01 5.50904214e-01 1.03163707e+00 4.07660633e-01
-8.57201636e-01 5.30108333e-01 6.26215744e+00 9.00982022e-01
-1.51457846e+00 2.32182473e-01 1.02980864e+00 -5.26639104e-01
1.87048599e-01 -4.33081359e-01 -8.05587351e-01 6.48730755e-01
1.41176641e+00 9.73794535e-02 4.59320128e-01 5.62236249e-01
3.38795520e-02 -2.06318408e-01 -1.13077903e+00 1.16545880e+00
-3.46107155e-01 -1.60939598e+00 1.66438848e-01 5.19974232e-02
6.20740175e-01 2.91725278e-01 1.33846059e-01 -3.46888863e-02
-1.52057558e-01 -1.46047950e+00 6.66656077e-01 5.03012314e-02
8.90848577e-01 -9.62476432e-01 8.54741633e-01 1.46529943e-01
-1.02859080e+00 -1.21305196e-03 -4.65189457e-01 1.60046548e-01
1.27209067e-01 4.51400310e-01 -8.30375969e-01 -6.28580600e-02
9.09636855e-01 -1.25774279e-01 -2.73465008e-01 7.93766379e-01
4.18083996e-01 5.41805923e-01 -7.57703781e-01 -7.45047629e-02
7.07559526e-01 -1.04682207e-01 7.04957079e-03 1.33057213e+00
1.79585800e-01 8.11048672e-02 -1.87857687e-01 9.17231381e-01
-3.77366483e-01 -6.30364940e-02 -3.81455682e-02 1.53920665e-01
5.47383308e-01 1.32899630e+00 -1.25073123e+00 -4.17318702e-01
-1.43182054e-01 6.43503904e-01 2.26226166e-01 -6.79939687e-02
-7.61922002e-01 -5.14285564e-01 4.54598874e-01 2.51616150e-01
2.80751437e-01 -1.66526601e-01 -8.50058556e-01 -3.57093692e-01
-1.03063479e-01 -7.77435422e-01 1.29787549e-01 -2.03418583e-01
-6.67823493e-01 5.39253771e-01 -1.04749069e-01 -8.92193317e-01
-2.34792814e-01 -7.41633832e-01 -2.61892408e-01 9.87181723e-01
-1.46431792e+00 -6.45575047e-01 -2.82937616e-01 3.62917125e-01
1.53322786e-01 2.94874012e-01 7.65499592e-01 6.92272484e-01
-4.13867593e-01 7.82159209e-01 9.34243351e-02 1.66405082e-01
1.90613508e-01 -1.04554641e+00 1.83310613e-01 4.53668743e-01
-3.27633291e-01 7.75522411e-01 4.96604353e-01 -3.48373055e-01
-1.45483947e+00 -1.21411479e+00 4.80289996e-01 1.57794058e-01
3.38977396e-01 -2.75886059e-01 -8.47512364e-01 4.17340904e-01
-5.56015968e-02 1.87016755e-01 6.94386065e-01 -6.94463849e-01
1.21308222e-01 -1.24360763e-01 -1.88190866e+00 6.37781322e-01
4.11203563e-01 -1.71648368e-01 -2.22063571e-01 8.52599069e-02
6.57357335e-01 -4.97030109e-01 -1.34074175e+00 3.40740710e-01
6.51950002e-01 -6.68814242e-01 9.77375090e-01 -4.49095815e-02
3.27664912e-01 -2.61090219e-01 -3.65550697e-01 -9.17836189e-01
-2.42820069e-01 -2.05649763e-01 -6.06456250e-02 8.15592647e-01
3.04860622e-01 -4.73686844e-01 9.71214116e-01 7.31797814e-01
-2.35468335e-02 -1.30165672e+00 -1.44807148e+00 -5.11354327e-01
3.83116454e-01 -1.91662595e-01 4.74921167e-01 6.14297152e-01
-9.07067508e-02 8.45196992e-02 3.43144327e-01 -1.43563122e-01
6.88446939e-01 -3.23812097e-01 -9.54260826e-02 -9.72912073e-01
-1.16924621e-01 -6.68447375e-01 -8.41005027e-01 -7.84294009e-01
-2.97670662e-01 -1.08444214e+00 -1.86968341e-01 -1.47831094e+00
3.65083069e-02 -4.75521535e-01 -4.62345541e-01 6.09024763e-01
2.89477050e-01 7.15898693e-01 1.82181850e-01 1.36685371e-02
-2.49911115e-01 1.06167063e-01 1.15763485e+00 -6.40327036e-02
9.87289567e-03 -5.50131440e-01 -5.07718801e-01 8.19830000e-01
9.81042445e-01 -6.84743345e-01 -3.73619348e-01 -6.78612590e-01
-6.31540269e-02 3.79242487e-02 1.47649094e-01 -1.23937488e+00
4.26482707e-01 1.83815002e-01 7.94624031e-01 -5.23838580e-01
4.08530533e-01 -8.35473120e-01 1.40692398e-01 1.08893359e+00
-1.93088889e-01 2.62925565e-01 4.57243443e-01 1.39307559e-01
-1.38904452e-01 -4.22328383e-01 1.46809912e+00 -1.85345188e-01
-3.42771798e-01 1.71989307e-01 -4.58176315e-01 -1.67889416e-01
8.94380748e-01 -4.35009748e-01 -1.32797346e-01 2.78537184e-01
-5.54585695e-01 1.23212256e-01 3.51351947e-01 -1.35513574e-01
3.47958475e-01 -1.25400484e+00 -2.00503573e-01 1.85104877e-01
-4.86738741e-01 1.94244042e-01 1.35632008e-01 1.00684834e+00
-1.07728863e+00 7.84482598e-01 -7.06685603e-01 -7.54921556e-01
-1.22769845e+00 1.05392434e-01 5.46855509e-01 -4.06873167e-01
-6.62588954e-01 8.28825712e-01 -8.95596072e-02 -7.57559612e-02
3.62309903e-01 -1.04931319e+00 -1.16979182e-01 -1.04094893e-01
4.95621771e-01 5.63861489e-01 4.29484338e-01 -3.66067559e-01
-5.60245156e-01 4.70311731e-01 -1.35063455e-01 -1.31072894e-01
1.14083397e+00 7.20800534e-02 -1.97106302e-01 2.25738481e-01
1.37445319e+00 -8.05045784e-01 -1.08156312e+00 1.62170321e-01
-1.12622663e-01 1.57057256e-01 6.94129705e-01 -7.36104906e-01
-1.28702152e+00 1.11249256e+00 1.30852759e+00 1.49507690e-02
1.17016244e+00 -4.20606554e-01 1.15898418e+00 2.90600836e-01
3.81280690e-01 -1.13710308e+00 -4.17476773e-01 3.52788806e-01
3.79807621e-01 -8.59839678e-01 7.14061633e-02 -1.52420193e-01
-2.06819028e-01 1.08365321e+00 4.09083307e-01 -2.90774703e-01
6.57107651e-01 8.83844197e-01 -1.41947167e-02 -1.70784220e-01
-6.19396091e-01 3.93674135e-01 9.27997604e-02 3.24581504e-01
6.34451866e-01 1.61414489e-01 -5.88626623e-01 7.20716476e-01
-3.48302633e-01 2.21517712e-01 2.78372169e-01 1.03593111e+00
-6.18142903e-01 -8.47143292e-01 -3.09503019e-01 8.41272533e-01
-1.04784548e+00 -2.46288911e-01 4.34096009e-02 3.42891246e-01
2.88990319e-01 6.55465841e-01 4.05070156e-01 -2.38282755e-02
-8.41662064e-02 -2.83058863e-02 5.78485489e-01 -2.95182943e-01
-1.02508759e+00 -1.55758334e-03 -2.73227423e-01 -4.54076320e-01
-8.00259933e-02 -3.32858562e-02 -2.04681849e+00 -4.19990331e-01
-8.52488056e-02 -1.28286108e-01 1.00157917e+00 5.87295592e-01
3.52594554e-01 9.11076725e-01 -2.93015450e-01 -7.91609406e-01
-5.91775477e-01 -7.92594910e-01 -4.80363756e-01 1.99178353e-01
1.88521788e-01 -5.83942652e-01 -3.57574713e-03 2.33487319e-02] | [8.595264434814453, 3.013394594192505] |
2b24f98d-0240-4825-89cc-3567d0b148c7 | attention-based-convolutional-neural-network-3 | 2303.02518 | null | https://arxiv.org/abs/2303.02518v1 | https://arxiv.org/pdf/2303.02518v1.pdf | Attention-based convolutional neural network for perfusion T2-weighted MR images preprocessing | Accurate skull-stripping is crucial preprocessing in dynamic susceptibility contrast-enhanced perfusion magnetic resonance data analysis. The presence of non-brain tissues impacts the perfusion parameters assessment. In this study, we propose different integration strategies for the spatial and channel squeeze and excitation attention mechanism into the baseline U-Net+ResNet neural network architecture to provide automatic skull-striping i.e., Standard scSE, scSE-PRE, scSE-POST, and scSE Identity strategies of plugging of scSE block into the ResNet backbone. We comprehensively investigate the performance of skull-stripping in T2-star weighted MR images with abnormal brain anatomy. The comparison that utilizing any of the proposed strategies provides the robustness of skull-stripping. However, the scSE-POST integration strategy provides the best result with an average Dice Coefficient of 0.9810. | ['Oleksii Diumin', 'Svitlana Alkhimova'] | 2023-03-04 | null | null | null | null | ['skull-stripping', 'anatomy'] | ['medical', 'miscellaneous'] | [ 2.27088526e-01 -1.23170100e-01 3.11813146e-01 -3.38015586e-01
-2.61890382e-01 -1.68963715e-01 3.35739106e-01 -2.42602751e-02
-7.72510469e-01 8.93176794e-01 4.36837047e-01 -3.71075183e-01
-4.56611127e-01 -3.71599764e-01 -4.23247576e-01 -9.91902709e-01
-2.76228815e-01 1.71149522e-01 4.88054693e-01 -2.13013321e-01
5.40275037e-01 7.20928907e-01 -7.58790374e-01 5.16218022e-02
1.06145835e+00 8.50913823e-01 5.34307837e-01 3.95313472e-01
-4.01598169e-03 8.76455843e-01 -3.84135693e-01 -4.48467918e-02
4.04004902e-01 -4.85774279e-01 -9.51137781e-01 -8.37483287e-01
1.29079312e-01 -4.91851926e-01 -4.37882751e-01 1.26628327e+00
1.17780674e+00 4.50179785e-01 5.05329192e-01 -9.12988126e-01
-5.45723915e-01 8.52523088e-01 -7.63497710e-01 1.50146985e+00
-2.62689829e-01 1.64935976e-01 -4.07394283e-02 -7.58381724e-01
7.11931288e-01 4.45497751e-01 8.20257127e-01 4.75437045e-01
-9.74812567e-01 -9.09609735e-01 -2.36032337e-01 6.20999038e-01
-1.30083764e+00 -4.98483390e-01 7.47011065e-01 -5.22700548e-01
1.16418362e+00 2.13263720e-01 6.68839872e-01 6.35765910e-01
7.87728131e-01 2.40071267e-01 1.45198452e+00 -1.07866757e-01
-1.63922701e-02 -2.48457387e-01 5.45618832e-01 3.39758962e-01
2.35477984e-01 1.22631760e-02 -1.50854051e-01 8.37675035e-02
1.18834102e+00 -1.75537169e-01 -6.46032512e-01 1.47004291e-01
-1.30553019e+00 3.77419114e-01 7.23006010e-01 7.11113155e-01
-9.46721971e-01 -1.97362825e-02 7.48865128e-01 1.03001744e-01
3.75124216e-01 4.33710068e-01 -2.08386153e-01 3.39199990e-01
-1.35689294e+00 -1.51226148e-01 9.61667597e-02 6.24357581e-01
-3.79170701e-02 4.09763455e-01 -4.77363497e-01 7.20883906e-01
-1.54469401e-01 1.34401143e-01 1.06396341e+00 -6.77123964e-01
1.03700362e-01 -2.95076426e-02 -1.93197310e-01 -8.35855186e-01
-7.66605258e-01 -8.85724545e-01 -1.18953836e+00 1.01193778e-01
8.75433609e-02 -4.31346506e-01 -9.90410268e-01 1.53177047e+00
2.88938105e-01 3.95589411e-01 -9.62972939e-02 1.21905601e+00
1.04319572e+00 3.66510123e-01 4.73380685e-01 -4.01212543e-01
1.01759851e+00 -1.24709487e+00 -1.08512664e+00 1.22411855e-01
3.90065879e-01 -4.65984464e-01 6.31961703e-01 6.50515556e-02
-1.34802389e+00 -2.92132288e-01 -1.18096769e+00 2.00403124e-01
-1.47849545e-01 -2.20397085e-01 3.98048192e-01 4.58421856e-01
-1.56908667e+00 1.04049122e+00 -1.04009569e+00 -4.16474566e-02
5.02541482e-01 6.55037940e-01 -7.68129528e-01 5.47005944e-02
-1.29176056e+00 1.45997787e+00 3.79538924e-01 3.51814359e-01
-7.60582209e-01 -1.17126465e+00 -3.29573274e-01 1.76137164e-01
2.90355682e-02 -7.73884118e-01 8.63921404e-01 -9.62189555e-01
-1.34462321e+00 5.41800141e-01 8.00029607e-04 -6.56305909e-01
6.13348246e-01 2.25690246e-01 -3.70267838e-01 4.29534703e-01
-3.65575519e-03 6.05221033e-01 3.33573699e-01 -9.12543297e-01
2.91057918e-02 -5.06476045e-01 -3.82990420e-01 4.54952031e-01
2.17277735e-01 6.38157248e-01 1.87394857e-01 -6.81107283e-01
5.23818791e-01 -4.35090274e-01 -3.11035246e-01 -3.03251684e-01
-4.77356762e-01 2.21592546e-01 4.54064190e-01 -1.46323049e+00
1.19272721e+00 -1.87596011e+00 -2.06258968e-01 5.08366942e-01
5.57955563e-01 1.39727578e-01 -1.87258065e-01 -3.35076421e-01
-9.56524134e-01 2.33846769e-01 -2.62203395e-01 3.23312879e-01
-5.91530085e-01 -2.61371523e-01 3.30682904e-01 8.99208486e-01
-2.78513849e-01 9.15434659e-01 -7.31884837e-01 -4.91539508e-01
1.33961841e-01 6.38401330e-01 -4.47925657e-01 1.81608391e-03
8.21885467e-01 7.61367798e-01 -1.13380998e-01 3.35245579e-01
9.95739818e-01 9.97660533e-02 2.58386275e-03 -5.36079228e-01
-3.60129029e-01 1.12519443e-01 -8.40681553e-01 1.58336544e+00
-4.33410406e-02 3.63568604e-01 4.10792828e-01 -1.00432920e+00
7.41456270e-01 5.78127682e-01 9.09060895e-01 -9.70953584e-01
5.31369865e-01 3.44197035e-01 7.76906371e-01 -5.15136242e-01
2.07696393e-01 -5.68126380e-01 7.64516234e-01 4.95481342e-01
2.56509274e-01 2.54385531e-01 5.90672065e-03 6.78815991e-02
1.00357807e+00 -7.13268518e-02 2.05287382e-01 -1.19178879e+00
5.65930128e-01 -4.19720113e-01 4.24003512e-01 6.48774028e-01
-7.97171414e-01 6.97658181e-01 3.32170337e-01 -3.38506192e-01
-1.02549076e+00 -8.77387762e-01 -3.69282961e-01 6.49108946e-01
-6.70708418e-02 3.80371004e-01 -1.19353890e+00 -5.00705004e-01
-5.97534060e-01 6.81047678e-01 -7.04917431e-01 -2.09878936e-01
-7.42147803e-01 -1.20603502e+00 4.99461532e-01 6.70919120e-01
7.53955603e-01 -1.23485494e+00 -8.09224904e-01 3.74186814e-01
-2.54345298e-01 -5.80145359e-01 -6.72841012e-01 4.91165757e-01
-1.01552248e+00 -8.70313168e-01 -1.20146108e+00 -6.60117686e-01
7.96845019e-01 8.60083327e-02 7.44625032e-01 3.19288135e-01
-1.27384007e-01 -1.26043260e-01 -1.15089312e-01 -4.48396616e-02
-6.95132464e-02 -2.34911200e-02 -2.96565834e-02 -2.03035355e-01
1.31609812e-02 -9.57493484e-01 -1.12907290e+00 2.30718672e-01
-7.48419285e-01 3.55141871e-02 4.63627577e-01 6.42492533e-01
6.75080478e-01 -1.21914104e-01 7.92779326e-01 -4.33925658e-01
9.00090098e-01 -7.11424947e-01 -2.26031423e-01 2.19759658e-01
-6.82911992e-01 -2.57248491e-01 4.97859240e-01 -3.30694824e-01
-9.91455793e-01 -3.40482622e-01 -2.91361958e-01 -4.32834923e-01
-6.90593794e-02 4.94160175e-01 1.87858716e-01 -6.16763353e-01
6.41882241e-01 5.09862185e-01 1.39791414e-01 -1.56233311e-01
-2.02887073e-01 2.26772159e-01 7.67900646e-01 -2.39377394e-01
6.08126707e-02 2.26827428e-01 -1.04309104e-01 -4.97237444e-01
-1.15455724e-01 -2.44793415e-01 -9.13041711e-01 -4.53359276e-01
1.10337818e+00 -3.67175251e-01 -3.50806773e-01 5.86710334e-01
-1.10381520e+00 -3.07169288e-01 -1.90981045e-01 7.32229829e-01
-1.80270180e-01 4.01118338e-01 -6.65955961e-01 -2.63018191e-01
-1.16970992e+00 -1.57068348e+00 -5.26080579e-02 3.33700210e-01
-4.37399559e-02 -8.61878991e-01 -9.58591923e-02 1.18213974e-01
1.16301823e+00 3.32189739e-01 9.27589238e-01 -9.88746881e-01
-7.38481283e-02 1.24218337e-01 -3.28224212e-01 4.47210252e-01
-1.50543347e-01 -3.58819634e-01 -7.16726601e-01 -2.28646696e-01
2.64147758e-01 3.00925642e-01 6.68893635e-01 1.23064756e+00
1.39382148e+00 -1.42436296e-01 1.07707538e-01 8.45192373e-01
1.47803903e+00 7.64696717e-01 1.01649725e+00 4.36952055e-01
3.21229607e-01 5.40295303e-01 -1.03398599e-01 9.52142254e-02
2.19385475e-01 8.64768177e-02 4.01551783e-01 -3.56537431e-01
-5.17814040e-01 2.23995596e-01 -1.49727419e-01 1.16199934e+00
-5.61766803e-01 -3.15685049e-02 -1.15255392e+00 5.45800984e-01
-1.21053135e+00 -9.12244797e-01 -3.44081938e-01 2.13705659e+00
6.74234211e-01 -1.54657394e-01 -2.08703697e-01 -8.04121867e-02
1.09169209e+00 -2.25724373e-02 -6.25521123e-01 -2.14044616e-01
-1.55626208e-01 5.28868079e-01 7.50678241e-01 6.40608191e-01
-8.58096004e-01 6.16371810e-01 6.69893074e+00 6.92523301e-01
-1.52868974e+00 7.76413321e-01 6.58334672e-01 -8.62899646e-02
-5.52789085e-02 -3.24728251e-01 -1.27652526e-01 6.94731712e-01
1.01164627e+00 -2.74591863e-01 5.78208208e-01 2.27616087e-01
5.18433630e-01 -2.73099244e-01 -5.77076435e-01 7.89159298e-01
1.19529463e-01 -1.25717819e+00 -1.68613642e-01 -2.78524220e-01
5.35613000e-01 4.61332947e-01 -2.18445003e-01 -1.37044802e-01
-2.60672718e-01 -1.00264883e+00 6.34946525e-01 6.37508690e-01
9.98813570e-01 -5.28907835e-01 1.03730428e+00 -5.36833070e-02
-6.60920143e-01 5.21909110e-02 -9.76777971e-02 3.64970088e-01
4.43533808e-01 4.40408260e-01 -4.98855978e-01 6.68914020e-01
6.58090949e-01 1.74409494e-01 -3.43435019e-01 1.55972850e+00
-5.00336066e-02 6.07102156e-01 -2.51845062e-01 5.01561940e-01
2.13668481e-01 -2.99032241e-01 6.25152647e-01 1.10656428e+00
3.71518373e-01 4.95833069e-01 -2.66888529e-01 9.15925086e-01
1.00127466e-01 5.78721821e-01 -8.66270587e-02 6.60921335e-01
3.28630149e-01 1.23974252e+00 -1.12304461e+00 -5.55448294e-01
-6.40878305e-02 7.06488669e-01 -8.01363289e-02 4.10860300e-01
-7.97856033e-01 -3.39858055e-01 -1.97876260e-01 4.31921422e-01
-2.01107726e-01 9.04241279e-02 -5.43140829e-01 -8.15110862e-01
-2.50980377e-01 -6.51902199e-01 1.14739440e-01 -1.13667023e+00
-1.00440443e+00 1.01167822e+00 2.64848322e-01 -4.85864192e-01
4.63590741e-01 -2.32076287e-01 -1.16062880e+00 1.42831624e+00
-1.72881222e+00 -4.97545660e-01 -3.37792009e-01 7.96860278e-01
1.83191597e-01 6.52346909e-02 4.93324697e-01 7.25519955e-01
-7.59696901e-01 3.59987229e-01 -7.66363069e-02 4.74252701e-02
4.74244893e-01 -9.39562678e-01 -1.09583393e-01 1.10362780e+00
-8.81820619e-01 8.74716401e-01 7.16554463e-01 -9.19967294e-01
-6.01217330e-01 -8.17823946e-01 6.51205063e-01 3.70108604e-01
5.04921079e-01 3.88611436e-01 -1.23104167e+00 5.95477760e-01
6.32907331e-01 3.80809665e-01 5.00122190e-01 -6.08845770e-01
2.19556794e-01 3.66859250e-02 -1.61793447e+00 2.54708737e-01
5.74775517e-01 -1.11592561e-01 -8.00429583e-01 2.22604364e-01
3.01588416e-01 -4.36084032e-01 -1.17289960e+00 4.62475479e-01
5.32377839e-01 -7.70899475e-01 1.00228345e+00 -5.93506098e-01
6.08240008e-01 -1.50328279e-01 3.00702602e-01 -1.53153336e+00
-7.58159995e-01 -4.51254517e-01 4.25519198e-01 6.57107055e-01
3.45777810e-01 -1.01642501e+00 5.37308455e-01 7.43178606e-01
-7.55243897e-01 -8.09726834e-01 -1.26594901e+00 -3.95271480e-01
5.57618916e-01 2.27486193e-01 5.75746536e-01 1.36612248e+00
-5.70123903e-02 -9.64958519e-02 -2.02483252e-01 -1.11409500e-01
6.61515296e-01 -6.04521096e-01 -3.14437509e-01 -1.02198136e+00
2.87171364e-01 -7.85705566e-01 -1.78832829e-01 -4.35897827e-01
1.34185538e-01 -1.26281941e+00 -1.17723294e-01 -1.76283967e+00
5.61403811e-01 -5.03702044e-01 -8.46563935e-01 5.36028564e-01
-1.80485249e-01 2.04990566e-01 -8.45860923e-04 1.70636833e-01
1.18218735e-01 3.39838415e-01 1.85390651e+00 8.00602287e-02
-2.23307326e-01 -4.68501061e-01 -6.22659922e-01 5.47399223e-01
1.15494967e+00 -5.45797467e-01 -4.42430615e-01 -4.64901537e-01
-3.56623650e-01 3.73825341e-01 2.86254615e-01 -1.02790546e+00
4.82761115e-01 9.66692269e-02 5.29142141e-01 -6.15335286e-01
-2.60546923e-01 -7.93837667e-01 2.37304628e-01 7.00663388e-01
-3.49564970e-01 5.82862198e-01 3.03678125e-01 -1.81609809e-01
-3.96077447e-02 -4.99296993e-01 1.16085160e+00 -5.19049287e-01
-3.08395863e-01 4.96332228e-01 -5.44357240e-01 -1.06432177e-01
9.27465737e-01 -4.58414435e-01 -4.86162156e-01 6.34776475e-03
-1.14222765e+00 7.34415948e-02 -2.09849030e-01 -1.02488637e-01
6.66774869e-01 -1.10500050e+00 -7.23190546e-01 1.91746831e-01
-6.36638343e-01 -3.48262906e-01 1.03313923e+00 1.97377574e+00
-9.26492214e-01 3.75453949e-01 -1.01322246e+00 -1.27036199e-01
-1.09179008e+00 3.61139417e-01 1.04877996e+00 -4.04632747e-01
-8.81887197e-01 8.70836318e-01 1.26947105e-01 -6.92631751e-02
-8.55206698e-02 -2.20481694e-01 -6.75124645e-01 -1.14234515e-01
5.48450053e-01 6.43664956e-01 4.01529640e-01 -9.08084333e-01
-4.92077351e-01 1.70934886e-01 -1.12173513e-01 -8.02280754e-02
1.50664353e+00 -1.06909461e-01 -4.98976469e-01 -3.50572258e-01
9.30331886e-01 -3.58859330e-01 -8.24235320e-01 1.39926046e-01
-3.94009888e-01 -1.22632377e-01 9.50834572e-01 -1.38379848e+00
-1.57345200e+00 9.78224277e-01 1.06117475e+00 -3.10971588e-01
1.11118245e+00 -6.08324826e-01 6.44140542e-01 -2.46416986e-01
1.46875635e-01 -7.34219849e-01 -5.33323526e-01 2.78852195e-01
1.17997968e+00 -7.05032051e-01 -2.24579927e-02 -4.15464453e-02
-6.97482407e-01 9.51115131e-01 7.81418025e-01 -2.54412830e-01
8.27406645e-01 4.96731967e-01 -2.06470832e-01 -4.29296851e-01
-2.27649778e-01 4.16768700e-01 1.44617736e-01 5.26089132e-01
5.17796695e-01 -1.23098075e-01 -7.94504464e-01 8.09850812e-01
-4.84258235e-02 1.27462596e-01 6.99517846e-01 6.97618008e-01
-2.82041430e-01 -4.25016940e-01 -2.82951325e-01 7.63803720e-01
-8.80926847e-01 -7.26134896e-01 1.38331443e-01 6.95018530e-01
4.71405126e-02 3.73604089e-01 3.89719084e-02 4.24239086e-03
4.04340863e-01 3.29906762e-01 5.69635451e-01 -1.47145718e-01
-1.18465114e+00 2.61174113e-01 -2.24846140e-01 -4.82539237e-01
-5.09448349e-01 -6.46772861e-01 -1.44619858e+00 -2.66510040e-01
-4.66309249e-01 1.62695810e-01 8.39449227e-01 7.62533486e-01
4.06507999e-01 9.24583077e-01 2.92006493e-01 -8.43002379e-01
-4.12284285e-01 -1.34850240e+00 -6.54109895e-01 9.64391604e-02
2.88239568e-01 -8.34355414e-01 -2.91903406e-01 -2.69979328e-01] | [14.137109756469727, -2.3563241958618164] |
4499eb7a-37a7-4be9-be27-22d18eae21fe | arnet-automatic-refinement-network-for-noisy | 2211.04774 | null | https://arxiv.org/abs/2211.04774v5 | https://arxiv.org/pdf/2211.04774v5.pdf | IRNet: Iterative Refinement Network for Noisy Partial Label Learning | Partial label learning (PLL) is a typical weakly supervised learning, where each sample is associated with a set of candidate labels. The basic assumption of PLL is that the ground-truth label must reside in the candidate set. However, this assumption may not be satisfied due to the unprofessional judgment of the annotators, thus limiting the practical application of PLL. In this paper, we relax this assumption and focus on a more general problem, noisy PLL, where the ground-truth label may not exist in the candidate set. To address this challenging problem, we propose a novel framework called "Iterative Refinement Network (IRNet)". It aims to purify the noisy samples by two key modules, i.e., noisy sample detection and label correction. Ideally, we can convert noisy PLL into traditional PLL if all noisy samples are corrected. To guarantee the performance of these modules, we start with warm-up training and exploit data augmentation to reduce prediction errors. Through theoretical analysis, we prove that IRNet is able to reduce the noise level of the dataset and eventually approximate the Bayes optimal classifier. Experimental results on multiple benchmark datasets demonstrate the effectiveness of our method. IRNet is superior to existing state-of-the-art approaches on noisy PLL. | ['JianHua Tao', 'Bin Liu', 'Licai Sun', 'Lan Chen', 'Mingyu Xu', 'Zheng Lian'] | 2022-11-09 | null | null | null | null | ['partial-label-learning'] | ['methodology'] | [ 3.52075368e-01 2.43141696e-01 -3.63992006e-01 -3.61449122e-01
-9.66133952e-01 -3.32542628e-01 4.25753668e-02 2.22105950e-01
-4.69761252e-01 9.45737958e-01 -1.12669110e-01 -2.58768424e-02
-1.33982822e-01 -7.16369212e-01 -6.40884161e-01 -9.09852684e-01
4.85403776e-01 3.76151145e-01 1.72030881e-01 2.48428226e-01
-8.11848566e-02 1.55972391e-02 -1.55858386e+00 2.57637084e-01
8.75297427e-01 1.19359112e+00 1.20428026e-01 3.35065164e-02
-1.35884047e-01 9.18500543e-01 -6.07096314e-01 -1.49171472e-01
2.08896488e-01 -4.67830360e-01 -8.85769546e-01 2.41952434e-01
7.37749264e-02 -8.34506974e-02 1.42617002e-01 1.58718383e+00
4.37299281e-01 5.82880974e-02 3.65715444e-01 -1.32531714e+00
-2.42713436e-01 1.01385605e+00 -5.06648242e-01 -2.31479421e-01
1.81838691e-01 -1.68269023e-01 1.22341990e+00 -9.95176435e-01
4.86146778e-01 9.63229001e-01 1.00228512e+00 6.93826437e-01
-1.28106713e+00 -8.67363095e-01 4.45807755e-01 -4.05104086e-02
-1.67027318e+00 -1.75709724e-01 7.79751301e-01 -2.07863674e-01
2.03654468e-02 2.80035585e-01 3.86352658e-01 1.00152218e+00
-3.38224500e-01 9.59863365e-01 1.45266044e+00 -4.23764974e-01
5.09536147e-01 3.73220056e-01 5.10561764e-01 5.46541870e-01
2.96843201e-01 -4.16502766e-02 -6.53323948e-01 -3.78734082e-01
2.41425231e-01 -2.52060331e-02 -6.08956575e-01 -1.90702751e-01
-1.04832566e+00 5.63406944e-01 2.43078291e-01 2.77962893e-01
-2.09979296e-01 -1.48884237e-01 3.53562385e-01 2.02079624e-01
6.49316728e-01 2.76322454e-01 -4.98651981e-01 2.30069324e-01
-8.94443214e-01 1.35736376e-01 8.53452981e-01 9.63414490e-01
8.57391238e-01 -3.47856671e-01 -3.15952212e-01 8.08571696e-01
3.19927305e-01 2.51676947e-01 4.01875466e-01 -8.52524519e-01
1.43954068e-01 5.71339905e-01 3.58113140e-01 -7.43529022e-01
-3.76601070e-01 -8.06876540e-01 -1.13665724e+00 -1.19099148e-01
5.57804406e-01 -9.61466879e-02 -7.88647473e-01 1.76460528e+00
5.70897520e-01 5.73074937e-01 -3.81817669e-02 9.97252464e-01
7.92182684e-01 5.90557635e-01 1.42436206e-01 -7.41829455e-01
1.04797339e+00 -7.28640735e-01 -9.46958423e-01 -9.20262262e-02
8.22165728e-01 -5.56672454e-01 1.17627442e+00 7.83795536e-01
-4.69062388e-01 -5.35633922e-01 -9.88789022e-01 4.17665958e-01
1.16612755e-01 2.49963030e-01 3.09380919e-01 6.82411790e-01
-5.14577329e-01 6.41341329e-01 -5.34936488e-01 -2.77455617e-02
4.61537719e-01 3.08722466e-01 -1.97360590e-01 -1.82222873e-01
-1.15323138e+00 3.83468807e-01 5.95115721e-01 4.09292161e-01
-7.25238681e-01 -5.74379086e-01 -6.30176365e-01 -1.11454338e-01
8.82959247e-01 -3.42325181e-01 1.49287164e+00 -1.03022838e+00
-1.15898037e+00 6.20893121e-01 -2.58905858e-01 -2.64527768e-01
6.47357404e-01 -2.76506454e-01 -2.88736552e-01 -2.09338516e-01
2.70434171e-01 2.08761707e-01 6.55179143e-01 -1.67033482e+00
-9.69605923e-01 -5.38836047e-02 -1.80173084e-01 1.38949184e-02
-5.23110747e-01 -1.51437417e-01 -4.13127422e-01 -5.47641873e-01
4.60760653e-01 -8.43031645e-01 -3.37834507e-01 -1.55078158e-01
-6.51932716e-01 -6.80697441e-01 8.74304771e-01 -2.41785064e-01
1.25465691e+00 -2.17957854e+00 -4.42056954e-01 5.57043016e-01
3.56434554e-01 3.48576456e-01 3.81276272e-02 -9.93799195e-02
-6.02426901e-02 3.17572206e-01 -2.38908887e-01 -5.16433477e-01
-4.02516201e-02 3.81604165e-01 -3.30590755e-01 6.65004432e-01
2.07198393e-02 5.41250587e-01 -1.35233104e+00 -5.83274722e-01
-4.90458868e-02 1.63033172e-01 -3.32027704e-01 2.55120158e-01
-3.83384347e-01 6.24227107e-01 -4.55483198e-01 6.66423142e-01
8.30967486e-01 -3.93621057e-01 1.88734785e-01 -3.47729146e-01
1.73719347e-01 1.81673899e-01 -1.84430802e+00 1.31150186e+00
-2.26307467e-01 5.99857345e-02 9.64888185e-02 -1.08592546e+00
9.98180449e-01 4.15217996e-01 6.30702317e-01 -2.99417287e-01
3.46704274e-01 4.66581285e-01 -2.99681544e-01 -4.88744140e-01
1.43110022e-01 -3.99906695e-01 -4.51323837e-02 4.83683437e-01
-1.34994447e-01 2.24362597e-01 1.39411300e-01 -1.02272458e-01
9.29277718e-01 6.23180158e-02 2.97102481e-01 -3.25722486e-01
6.32879376e-01 -1.48834929e-01 1.24790144e+00 1.06944978e+00
-2.45186552e-01 5.70634723e-01 5.40957749e-01 -3.61810654e-01
-7.70143807e-01 -7.02270865e-01 -2.28533208e-01 9.30209100e-01
3.85538429e-01 -5.31913400e-01 -9.08787429e-01 -1.13050342e+00
-2.89823532e-01 4.36564684e-01 -5.77649653e-01 -1.33990943e-01
-4.21658069e-01 -9.69800413e-01 4.27994072e-01 4.37008053e-01
5.72330773e-01 -9.63270545e-01 -1.48884922e-01 2.11214915e-01
-5.06530941e-01 -1.15190434e+00 -3.80079359e-01 4.49884742e-01
-5.45293033e-01 -1.11108005e+00 -1.29373997e-01 -9.02669072e-01
9.23306882e-01 1.60948068e-01 9.26963747e-01 3.97764683e-01
3.88220936e-01 -8.76876339e-02 -5.97143292e-01 -3.68408024e-01
-5.83346367e-01 2.16358006e-01 2.29650229e-01 3.03947687e-01
4.99231309e-01 -2.93023944e-01 -2.95635223e-01 5.36240876e-01
-9.36790705e-01 -5.07720634e-02 6.15574121e-01 1.08083177e+00
1.16414654e+00 6.31708503e-01 8.08099926e-01 -1.32449365e+00
5.76270580e-01 -4.63434011e-01 -5.47264874e-01 3.64422560e-01
-7.71865547e-01 7.83955008e-02 7.84629583e-01 -7.17690408e-01
-8.16637218e-01 3.86129886e-01 -2.76437670e-01 -2.59011596e-01
-2.12874517e-01 7.07509220e-01 -5.33050537e-01 1.04622461e-01
7.03888118e-01 7.78816268e-02 -2.78026104e-01 -6.41714275e-01
-4.67612967e-03 8.97897243e-01 5.89384317e-01 -6.59186244e-01
8.39223206e-01 4.41282660e-01 -1.65996216e-02 -4.25296277e-01
-1.80481470e+00 -6.70233846e-01 -5.36620855e-01 -3.21289122e-01
2.85429150e-01 -8.03765416e-01 -7.75316656e-01 3.13755929e-01
-9.83794928e-01 -2.84167677e-01 -5.17194152e-01 3.37668359e-01
-2.14740321e-01 3.30650181e-01 -4.50478286e-01 -9.80489075e-01
-3.89606535e-01 -1.13260198e+00 8.97342801e-01 1.25090852e-01
-2.10808828e-01 -5.96290946e-01 -1.71265468e-01 1.98773891e-01
-1.61274076e-01 2.01649889e-01 4.80544567e-01 -8.92917156e-01
-3.22626173e-01 -4.42698449e-01 -2.11873218e-01 6.51244283e-01
1.51494473e-01 -4.07839209e-01 -1.12337470e+00 -2.96518058e-01
3.13962370e-01 -4.91248578e-01 9.51159060e-01 -7.61756627e-03
1.41215253e+00 -3.23674142e-01 -3.14044237e-01 2.62980968e-01
1.38626802e+00 -1.58978149e-01 2.50055820e-01 1.47709385e-01
7.55618632e-01 4.82362360e-01 9.51454639e-01 5.25048375e-01
1.84167862e-01 4.67389852e-01 3.33915442e-01 3.09528336e-02
5.43543510e-02 -4.22121257e-01 1.80837050e-01 8.28704596e-01
2.26126358e-01 -2.71685064e-01 -8.78736615e-01 3.74612302e-01
-2.11206460e+00 -5.50003171e-01 -4.02251035e-01 2.20170164e+00
1.12793934e+00 3.67443681e-01 -1.06201053e-01 7.83702791e-01
9.17663097e-01 -3.04729700e-01 -4.96227652e-01 3.21947157e-01
8.78214277e-03 1.94742028e-02 3.97755861e-01 4.11328882e-01
-1.33409178e+00 7.83573866e-01 5.48639870e+00 1.18111646e+00
-9.56384301e-01 4.67199057e-01 8.18977058e-01 1.27676561e-01
-1.91410825e-01 -5.29744588e-02 -1.10593235e+00 5.28260648e-01
6.09866440e-01 1.07959762e-01 2.01070651e-01 9.28425372e-01
2.26843700e-01 -1.79506034e-01 -1.14618576e+00 8.56167376e-01
-3.85407060e-02 -1.01898956e+00 -2.92421758e-01 -1.69441953e-01
1.00245953e+00 -2.87246525e-01 -2.40828499e-01 4.89784271e-01
3.92586917e-01 -8.80181909e-01 8.86713326e-01 5.34637094e-01
7.26089716e-01 -9.16198194e-01 1.25723958e+00 8.78389895e-01
-1.08115089e+00 -1.31738096e-01 -4.44967389e-01 1.53064076e-02
-4.33249325e-02 1.31865609e+00 -5.74932158e-01 4.66699421e-01
7.99661279e-01 5.92084110e-01 -2.67380655e-01 1.11327326e+00
-4.89254534e-01 8.52237582e-01 -4.98395950e-01 3.91021818e-02
1.53632881e-02 5.39748967e-02 2.69160211e-01 9.87742662e-01
1.96607471e-01 1.70237079e-01 7.35463798e-01 7.79317200e-01
-4.38934714e-01 2.88842380e-01 -2.50028223e-01 3.95743459e-01
6.57066166e-01 1.29786849e+00 -8.62838209e-01 -3.11702430e-01
-3.16812366e-01 6.08318865e-01 4.03067768e-01 2.13790163e-01
-6.46298587e-01 -4.32476178e-02 1.60725385e-01 4.90560532e-02
3.36863697e-02 3.67794633e-01 -5.35026431e-01 -1.07857668e+00
2.20747218e-01 -9.84196067e-01 4.05503929e-01 -2.87813723e-01
-1.56176364e+00 4.69592214e-01 -3.53673220e-01 -1.35441172e+00
2.62358278e-01 -2.79478699e-01 -2.62086421e-01 7.16640592e-01
-1.40999401e+00 -9.52461243e-01 -6.00014091e-01 4.03921813e-01
3.10512334e-01 2.78185517e-01 6.86060965e-01 3.64424288e-01
-6.51920557e-01 7.29259133e-01 7.17402846e-02 1.86834723e-01
7.28182018e-01 -1.23033845e+00 -2.71891654e-01 8.91164482e-01
9.65701044e-02 4.37215716e-01 7.59660304e-01 -8.01250875e-01
-7.76143134e-01 -1.55628407e+00 8.88570368e-01 -1.66431487e-01
5.32252431e-01 -5.04296243e-01 -1.22385573e+00 4.89916205e-01
-3.24456930e-01 4.84691650e-01 6.94450676e-01 1.08453356e-01
-2.71345377e-01 -3.63237798e-01 -1.14818668e+00 3.36709470e-01
9.67151284e-01 -4.02206361e-01 -2.67025530e-01 5.59675395e-01
6.39247715e-01 -4.66734946e-01 -7.56743968e-01 8.49536359e-01
2.57473469e-01 -6.81766093e-01 5.38833976e-01 -2.57963270e-01
8.07337537e-02 -6.83971345e-01 -7.37207010e-03 -1.14588594e+00
-1.43683091e-01 -5.38384557e-01 -1.97219461e-01 1.44901776e+00
3.60049009e-01 -3.54274511e-01 9.62072313e-01 3.56792778e-01
-5.10141663e-02 -9.40605521e-01 -9.48348165e-01 -8.36585760e-01
-2.40487859e-01 -8.71111274e-01 5.79716146e-01 1.06380689e+00
3.61784957e-02 2.76198268e-01 -6.29974782e-01 4.68916625e-01
7.42222011e-01 4.45437878e-02 4.89251286e-01 -1.56658471e+00
-2.37827897e-01 -1.32523760e-01 2.95754652e-02 -1.12293673e+00
3.58666062e-01 -8.46339524e-01 7.49391139e-01 -1.34326315e+00
3.08601797e-01 -9.34308887e-01 -5.63697577e-01 6.32575750e-01
-5.01747370e-01 5.24857521e-01 -1.85742944e-01 4.57940727e-01
-9.87087667e-01 5.26301622e-01 1.20618522e+00 1.96217257e-03
-2.17635155e-01 3.01091760e-01 -8.03796291e-01 9.43436325e-01
8.84264171e-01 -9.44820523e-01 -3.22210878e-01 1.10217601e-01
4.77386832e-01 -2.61949152e-01 1.64370820e-01 -1.01568401e+00
3.07659030e-01 -1.77049667e-01 1.34522274e-01 -5.97401083e-01
-1.87744632e-01 -9.10544753e-01 6.23085350e-02 4.17766988e-01
-5.30717611e-01 -6.80560410e-01 -3.24443817e-01 7.15823114e-01
-1.42509401e-01 -6.34338260e-01 9.23582792e-01 5.04985265e-02
-4.58515286e-01 3.17683190e-01 -1.90262720e-01 -1.38550717e-02
9.53943312e-01 4.56697904e-02 -1.53530523e-01 -1.21674746e-01
-8.57011557e-01 3.64078194e-01 2.51314610e-01 6.91518933e-02
4.67339039e-01 -1.39430547e+00 -6.86832368e-01 1.34337887e-01
2.86987901e-01 4.65284735e-01 -9.53542143e-02 1.00092435e+00
-6.15697801e-02 2.38968898e-02 6.57434821e-01 -6.43591940e-01
-1.13700831e+00 6.79658651e-01 3.21244836e-01 -4.59171027e-01
-6.30734384e-01 7.47159362e-01 -7.24529177e-02 -5.22517264e-01
6.58576548e-01 -2.30300874e-01 -3.42368662e-01 -1.92634296e-02
6.13243043e-01 1.48427114e-01 1.46407023e-01 -5.15574574e-01
-1.41926900e-01 3.06149572e-01 -1.09550105e-02 2.45702475e-01
1.16777110e+00 -1.88993275e-01 -2.73295522e-01 6.81562901e-01
8.50820363e-01 -1.71865930e-03 -1.06574678e+00 -8.55948806e-01
3.38582844e-01 -3.03151459e-01 1.53199241e-01 -5.95688939e-01
-1.07608700e+00 4.25537974e-01 5.87102115e-01 3.31762820e-01
1.22983491e+00 8.77208784e-02 7.88182378e-01 5.24732471e-01
4.52779204e-01 -1.29832625e+00 3.11545040e-02 4.68927979e-01
3.72682482e-01 -1.30155408e+00 -4.81675155e-02 -7.83936262e-01
-2.86801487e-01 8.49725962e-01 7.42078900e-01 6.33698478e-02
8.05254638e-01 3.57206255e-01 6.96711168e-02 4.25389931e-02
-5.85864782e-01 -2.63788402e-01 2.83403128e-01 3.23434383e-01
2.79425532e-01 -3.65442634e-02 -5.21963954e-01 1.07144320e+00
-8.53176117e-02 1.09590419e-01 3.40920538e-01 8.02024424e-01
-6.27855062e-01 -1.16080582e+00 -5.97309053e-01 6.29748285e-01
-6.42917573e-01 9.64610055e-02 -1.66823104e-01 4.17902440e-01
7.96192288e-01 1.18735635e+00 -3.01778823e-01 -5.14324725e-01
3.93502116e-01 1.16751537e-01 7.29071647e-02 -8.65915835e-01
-4.00694370e-01 3.38552982e-01 4.09963280e-02 -3.91962498e-01
-6.55013144e-01 -6.14978015e-01 -1.21771598e+00 -7.09718242e-02
-7.98361182e-01 3.71220946e-01 2.03267604e-01 1.28732908e+00
-2.02951521e-01 4.51612890e-01 6.22906089e-01 -3.25527281e-01
-6.91569388e-01 -1.06739223e+00 -6.89325809e-01 4.73484188e-01
2.97204763e-01 -6.17138207e-01 -4.89814639e-01 1.56951174e-01] | [9.442492485046387, 3.9844112396240234] |
24bb9d5b-0ee2-4a80-81f2-4edb71a884f3 | inferring-implicit-relations-with-language | 2204.13778 | null | https://arxiv.org/abs/2204.13778v2 | https://arxiv.org/pdf/2204.13778v2.pdf | Inferring Implicit Relations in Complex Questions with Language Models | A prominent challenge for modern language understanding systems is the ability to answer implicit reasoning questions, where the required reasoning steps for answering the question are not mentioned in the text explicitly. In this work, we investigate why current models struggle with implicit reasoning question answering (QA) tasks, by decoupling inference of reasoning steps from their execution. We define a new task of implicit relation inference and construct a benchmark, IMPLICITRELATIONS, where given a question, a model should output a list of concept-relation pairs, where the relations describe the implicit reasoning steps required for answering the question. Using IMPLICITRELATIONS, we evaluate models from the GPT-3 family and find that, while these models struggle on the implicit reasoning QA task, they often succeed at inferring implicit relations. This suggests that the challenge in implicit reasoning questions does not stem from the need to plan a reasoning strategy alone, but to do it while also retrieving and reasoning over relevant information. | ['Jonathan Berant', 'Mor Geva', 'Uri Katz'] | 2022-04-28 | null | null | null | null | ['implicit-relations'] | ['natural-language-processing'] | [ 4.41424578e-01 1.05871654e+00 1.78374872e-02 -3.86740983e-01
-7.53333628e-01 -7.89372504e-01 9.73710120e-01 5.96336424e-01
-6.95653036e-02 6.96947336e-01 3.89210075e-01 -9.79078472e-01
-5.39930284e-01 -1.31561053e+00 -4.24757272e-01 -2.55441461e-02
3.00575554e-01 9.89437461e-01 5.38663805e-01 -5.90698957e-01
3.57133836e-01 3.55192989e-01 -1.42317164e+00 7.74585843e-01
1.04991686e+00 6.43273890e-01 -1.90414697e-01 9.23429251e-01
-6.89848065e-01 2.14978814e+00 -5.85659087e-01 -7.61090994e-01
-3.18132311e-01 -5.91891468e-01 -1.87797952e+00 -6.07766509e-01
4.77547765e-01 -3.68212372e-01 -7.43150413e-02 7.73887217e-01
-3.53885554e-02 1.60166919e-01 8.46100450e-01 -1.37243569e+00
-3.67179871e-01 6.94328487e-01 3.60507339e-01 2.72191107e-01
9.45868433e-01 1.51293650e-01 1.58012426e+00 -4.86190468e-01
6.58554614e-01 1.39985955e+00 6.82772756e-01 8.13860357e-01
-1.29865861e+00 4.59729023e-02 1.39011964e-01 7.33620763e-01
-1.07334149e+00 -6.06515110e-01 2.08918005e-01 -3.15699667e-01
1.54928195e+00 8.47427785e-01 2.16734648e-01 3.87239188e-01
-1.87422452e-03 6.16692245e-01 1.04304779e+00 -6.81048810e-01
2.16227084e-01 1.89818993e-01 9.03030038e-01 8.35962415e-01
1.83563605e-01 -3.95859778e-02 -4.02400523e-01 -2.86917329e-01
4.01941270e-01 -4.64749068e-01 -1.77684411e-01 -1.11438194e-02
-8.04016054e-01 8.13598156e-01 2.57316709e-01 4.05964106e-01
-7.75685087e-02 1.94247723e-01 1.95423275e-01 6.20924830e-01
-1.76204234e-01 9.49685931e-01 -8.70759845e-01 -1.63443357e-01
-3.11638296e-01 6.63992822e-01 1.74672508e+00 8.55748177e-01
1.07362819e+00 -6.95949078e-01 -3.66110474e-01 2.68790692e-01
2.68102169e-01 2.15506151e-01 4.27719615e-02 -1.63196504e+00
4.58943605e-01 1.22457600e+00 2.09606990e-01 -8.14137459e-01
-2.55362839e-01 4.20858897e-02 -1.78614244e-01 4.48315628e-02
8.18931162e-01 1.17598914e-01 -4.92795318e-01 1.70654988e+00
3.49667281e-01 2.51506679e-02 4.74425048e-01 4.77906287e-01
1.12837517e+00 4.80965078e-01 4.48065221e-01 -1.28455460e-01
1.67996705e+00 -8.99384856e-01 -6.71489120e-01 -4.96873379e-01
1.27073967e+00 -3.10570598e-01 1.01769626e+00 9.78916660e-02
-1.18918908e+00 -2.35572264e-01 -7.01252282e-01 -7.69810855e-01
-4.97490674e-01 -3.55102450e-01 7.19687819e-01 8.24933276e-02
-9.22464907e-01 3.52891773e-01 -2.52868861e-01 -4.19092596e-01
-5.78695834e-02 1.64105788e-01 -1.72330707e-01 -3.66928428e-01
-1.43807197e+00 1.54090452e+00 6.05384111e-01 -7.51043931e-02
-6.20873809e-01 -7.35974371e-01 -9.39034522e-01 3.61124307e-01
1.10151947e+00 -1.21756113e+00 1.79692805e+00 -7.63334274e-01
-1.17463470e+00 9.63926256e-01 -5.29209852e-01 -7.63723552e-01
3.87049109e-01 -5.88473082e-01 -2.70131886e-01 2.89736390e-01
2.45222583e-01 3.13720167e-01 3.41126263e-01 -1.07718039e+00
-4.89075333e-01 -3.89136910e-01 1.15915644e+00 2.21722394e-01
5.25419414e-01 -7.63837695e-02 -2.47556090e-01 8.22579041e-02
2.55948216e-01 -5.31807363e-01 -1.65542096e-01 -5.50485328e-02
-3.27755541e-01 -6.82434797e-01 5.95838606e-01 -6.25883162e-01
1.09787107e+00 -1.64098763e+00 2.33836755e-01 4.07322086e-02
5.63984752e-01 1.70007959e-01 -5.51372394e-03 5.67981958e-01
-1.14452742e-01 2.22016409e-01 -3.37819815e-01 2.41509557e-01
4.78397906e-01 7.98481584e-01 -7.74658561e-01 -3.78380716e-01
4.80213702e-01 1.41570687e+00 -9.14174139e-01 -6.59824312e-01
-6.80655688e-02 -2.87572801e-01 -7.06999421e-01 5.43599129e-01
-1.22532105e+00 1.41260885e-02 -5.76126933e-01 3.52085650e-01
3.81652832e-01 -5.20377576e-01 5.47555745e-01 -1.67615175e-01
2.65787959e-01 1.18823218e+00 -8.57423902e-01 1.43046486e+00
-4.84326154e-01 5.02545059e-01 -4.57230136e-02 -9.60770130e-01
3.89603317e-01 4.00304496e-01 -1.81473598e-01 -8.66301835e-01
-2.64695644e-01 1.75536960e-01 2.69493848e-01 -1.00197494e+00
1.87236577e-01 -4.97211695e-01 7.37419948e-02 7.65973270e-01
-3.47217284e-02 -7.53838181e-01 5.67794323e-01 6.01816833e-01
1.42646754e+00 4.29462552e-01 5.38669944e-01 -2.34483376e-01
1.03275275e+00 6.39500201e-01 1.90583840e-01 9.16114628e-01
3.57850641e-01 1.01679494e-03 1.01726997e+00 -6.83793426e-01
-6.53982997e-01 -9.88300622e-01 2.00902611e-01 1.03584719e+00
1.92288626e-02 -9.17696238e-01 -6.12767994e-01 -9.46819186e-01
-1.73311546e-01 1.43441975e+00 -6.60624921e-01 -1.54954955e-01
-8.63839746e-01 -4.60294969e-02 7.18486488e-01 4.02581334e-01
4.45732892e-01 -1.25044179e+00 -9.16001678e-01 1.42953619e-01
-7.09073782e-01 -1.12480807e+00 3.88317078e-01 1.75249532e-01
-1.01748037e+00 -1.81247330e+00 1.62562087e-01 -4.06204045e-01
5.61172843e-01 -1.36499956e-01 1.94028962e+00 9.60716188e-01
-3.21793742e-02 6.89652085e-01 -3.44822377e-01 -2.46034741e-01
-4.98884678e-01 2.12508678e-01 -8.90336812e-01 -6.86371863e-01
7.35477626e-01 -4.77827013e-01 -2.12324783e-01 3.16728294e-01
-7.29567230e-01 2.69727230e-01 4.11310345e-01 5.27070403e-01
3.19557846e-01 1.47495747e-01 3.58264714e-01 -1.29225814e+00
7.97964633e-01 -5.68966448e-01 -4.03359085e-01 7.14469492e-01
-4.76825595e-01 5.66582203e-01 4.09544736e-01 1.18169107e-01
-1.17112434e+00 -5.16734600e-01 -3.25911283e-01 4.39758629e-01
-3.78981709e-01 7.93322265e-01 -3.16947699e-01 4.03065920e-01
8.90991390e-01 1.76625624e-01 -7.53282160e-02 -2.04307735e-01
7.02320039e-01 -7.36527517e-02 3.73066723e-01 -1.49451804e+00
1.02349758e+00 3.28892954e-02 2.39021286e-01 -5.81920922e-01
-1.50183284e+00 -3.45972687e-01 -4.76883024e-01 3.94821852e-01
8.18696082e-01 -4.81520534e-01 -1.14436734e+00 -3.94926459e-01
-1.39967525e+00 -7.76180804e-01 -8.55589509e-01 -1.96932346e-01
-9.90271151e-01 3.77898157e-01 -6.26903176e-01 -9.49306190e-01
-4.01746035e-01 -8.74623775e-01 8.69595408e-01 -8.22165832e-02
-1.01816368e+00 -1.02243376e+00 1.66176721e-01 6.21003449e-01
3.84637654e-01 -2.29335740e-01 1.93923986e+00 -9.51881826e-01
-1.09058321e+00 1.98424682e-01 -4.13396150e-01 1.36878327e-01
6.42126277e-02 -5.15879631e-01 -8.76525044e-01 5.51622808e-01
9.40978229e-02 -5.67365885e-01 2.29852885e-01 -3.10244381e-01
9.89630103e-01 -6.77155554e-01 -3.28498445e-02 5.87480180e-02
1.15161037e+00 -8.28068629e-02 1.15597153e+00 2.83723146e-01
1.06800154e-01 1.06635404e+00 4.16105896e-01 -5.19718349e-01
7.49590158e-01 4.42284703e-01 1.43733814e-01 3.23356420e-01
-1.03363685e-01 -3.07628661e-01 -8.16962421e-02 3.52326512e-01
-3.03338319e-01 2.72042304e-01 -1.29562330e+00 3.59646291e-01
-2.06299782e+00 -1.18500853e+00 -5.09162307e-01 1.93475032e+00
1.24721873e+00 6.35009333e-02 -2.65056729e-01 3.12786251e-01
-2.01548919e-01 -1.52169645e-01 -5.64911246e-01 -4.60041344e-01
6.85881525e-02 6.28157854e-01 -4.70995188e-01 1.08159757e+00
-5.42862773e-01 9.74030018e-01 6.53998280e+00 4.23532099e-01
-3.40281576e-01 -5.68085127e-02 2.29859635e-01 4.53100145e-01
-6.68403685e-01 5.22819579e-01 -5.37138760e-01 -4.69582438e-01
1.16036654e+00 -2.67964840e-01 4.96351659e-01 4.81699049e-01
-3.50610048e-01 -4.67000067e-01 -1.79400325e+00 2.79143453e-01
-1.70637608e-01 -1.27008986e+00 5.01835942e-01 -2.79266030e-01
8.49319249e-02 -6.28072321e-01 -4.42650408e-01 7.88262129e-01
5.80521822e-01 -1.45456398e+00 4.97703552e-01 7.41826355e-01
2.02317357e-01 -3.08497041e-01 5.88580251e-01 8.25828969e-01
-1.09453046e+00 -1.65319085e-01 -1.46644816e-01 -4.93227810e-01
5.00810444e-02 1.95421115e-01 -8.10209036e-01 7.09054768e-01
3.32923561e-01 3.52023721e-01 -6.88281059e-01 6.73236132e-01
-1.00546777e+00 4.10923868e-01 -1.50561005e-01 -5.89012317e-02
-7.31373727e-02 -1.72737494e-01 2.81235516e-01 9.11376774e-01
9.81084350e-03 8.95003617e-01 -2.04534382e-01 1.13666832e+00
1.76849857e-01 -1.31842002e-01 -4.69162315e-01 -1.86061412e-01
5.79211228e-02 9.05406833e-01 -1.67623177e-01 -7.92342126e-01
-3.69125038e-01 5.47886670e-01 6.85830712e-01 5.35376012e-01
-5.14727294e-01 -1.07454672e-01 5.19190788e-01 2.57120699e-01
-1.90400764e-01 -2.03671485e-01 -2.66383886e-01 -1.11160874e+00
6.26123995e-02 -1.20930362e+00 7.69261181e-01 -1.52439189e+00
-1.17325139e+00 3.44324172e-01 3.01362693e-01 -3.17132533e-01
-5.90236962e-01 -7.66842425e-01 -5.34142017e-01 1.20744860e+00
-1.65711355e+00 -9.18765843e-01 -3.91831517e-01 6.58998430e-01
4.17616159e-01 5.54623604e-01 1.28488958e+00 -5.65808341e-02
1.64326534e-01 -6.20613545e-02 -1.27509511e+00 2.01370895e-01
2.03578010e-01 -1.49204218e+00 3.92537594e-01 5.73424637e-01
9.66909230e-02 1.28940380e+00 7.96743810e-01 -3.49958479e-01
-1.63993275e+00 -6.21533573e-01 1.26627409e+00 -1.17804587e+00
8.65853667e-01 8.69310200e-02 -1.28274834e+00 1.21510196e+00
4.58227307e-01 -2.54695833e-01 6.62158132e-01 5.28185606e-01
-8.13553989e-01 2.31499329e-01 -9.44967151e-01 7.59389818e-01
1.16762054e+00 -1.10844254e+00 -1.80325246e+00 3.29891622e-01
9.93381858e-01 -3.29737723e-01 -5.92034936e-01 3.60125959e-01
3.02469224e-01 -8.30179036e-01 1.13521934e+00 -1.41135108e+00
8.20998251e-01 -5.34470975e-01 -3.24841946e-01 -8.25600386e-01
-1.34326398e-01 -2.60447830e-01 -8.26439559e-01 8.81715775e-01
8.22279334e-01 -9.17690635e-01 5.94449222e-01 1.30061150e+00
1.94242209e-01 -6.17224336e-01 -5.02367139e-01 -2.67736167e-01
2.21133545e-01 -6.04270756e-01 6.22225165e-01 7.87438512e-01
1.82214200e-01 1.03565586e+00 4.54131246e-01 2.69421250e-01
4.20071095e-01 5.67266881e-01 8.68543088e-01 -1.52673721e+00
-5.44457853e-01 -2.32595995e-01 1.62106827e-01 -1.31285214e+00
3.31196815e-01 -8.47747326e-01 1.71232335e-02 -2.10807514e+00
7.47427717e-02 -7.39256799e-01 1.75178275e-01 7.15033114e-01
-1.82155862e-01 -4.71778244e-01 1.06456451e-01 1.29473004e-02
-7.65956521e-01 1.17107823e-01 1.29793930e+00 -3.90180826e-01
4.11987513e-01 2.76602851e-03 -9.06278968e-01 8.50067794e-01
6.06707752e-01 -3.64192635e-01 -9.08431947e-01 -5.13099372e-01
1.01078355e+00 4.96714264e-01 6.58462882e-01 -8.09005082e-01
6.17163360e-01 -5.23303270e-01 -2.93018699e-01 -3.19911689e-01
5.30084550e-01 -1.04506576e+00 8.39111060e-02 5.50730109e-01
-6.56834900e-01 1.08655840e-01 3.23737502e-01 1.84408858e-01
-3.42169762e-01 -7.84443617e-01 1.49155870e-01 -5.07305443e-01
-7.81974435e-01 -5.84800765e-02 -4.15178269e-01 7.32100427e-01
6.73249424e-01 3.71450074e-02 -8.55355859e-01 -2.72030890e-01
-1.05579758e+00 2.21196949e-01 2.32179299e-01 1.56885404e-02
4.53072459e-01 -6.99356675e-01 -5.68661451e-01 -2.72272497e-01
2.91032404e-01 3.15905005e-01 1.45153459e-02 8.79229009e-01
-8.06787729e-01 6.01436913e-01 2.43978783e-01 -2.29695469e-01
-1.15203023e+00 5.84043682e-01 8.10655653e-01 -6.25124693e-01
-5.69683254e-01 6.78169072e-01 3.82544160e-01 -9.03390288e-01
-8.40235874e-02 -4.86948401e-01 -3.75135481e-01 -1.61453888e-01
7.77427256e-01 -5.69190308e-02 1.21500745e-01 -1.66363209e-01
-3.35903198e-01 5.63663304e-01 -2.09540711e-03 -3.14187035e-02
1.03690565e+00 8.79085287e-02 -6.60085678e-01 3.57822061e-01
4.62560028e-01 7.21195340e-02 -3.32831055e-01 -5.27921617e-01
5.22763371e-01 -2.01370329e-01 -6.03601515e-01 -1.24437201e+00
-3.34308237e-01 1.04242432e+00 -2.57908046e-01 6.86268866e-01
8.00627649e-01 2.63228625e-01 3.22549164e-01 1.15079689e+00
5.93242466e-01 -6.37099862e-01 2.70797405e-02 1.15450716e+00
9.86924350e-01 -8.01539600e-01 2.00518727e-01 -7.90557384e-01
-2.73351252e-01 1.17656016e+00 8.65107119e-01 3.23546857e-01
1.03238277e-01 2.68702775e-01 8.38389769e-02 -7.24335551e-01
-1.30706990e+00 -3.06464970e-01 1.75262704e-01 4.56294060e-01
5.96822858e-01 -1.24950521e-01 -1.32043719e-01 2.16802374e-01
-3.99910867e-01 1.36391953e-01 1.54669747e-01 1.14622188e+00
-5.02613723e-01 -1.30307519e+00 -2.23193035e-01 3.57358754e-01
-8.18983838e-02 -3.66041034e-01 -7.85529435e-01 8.90290618e-01
8.79011210e-03 1.11400104e+00 -1.94873765e-01 2.04886079e-01
4.71111059e-01 4.89566714e-01 8.65803957e-01 -1.04799545e+00
-7.04884827e-01 -1.07607830e+00 9.00384903e-01 -6.76175177e-01
-4.33699071e-01 -1.87297866e-01 -1.58244860e+00 -3.96333337e-01
-1.04354650e-01 9.15402651e-01 7.66498074e-02 1.63917816e+00
1.20529793e-01 6.19720459e-01 -4.50430423e-01 1.62947804e-01
-6.08177900e-01 -7.03946531e-01 1.01505369e-01 4.47042078e-01
2.26706311e-01 -2.45082870e-01 -3.48254979e-01 6.69795200e-02] | [9.9564790725708, 7.592714786529541] |
4c3b7442-4079-4088-b3e3-9c80a7f416c6 | tensor-networks-meet-neural-networks-a-survey | 2302.09019 | null | https://arxiv.org/abs/2302.09019v2 | https://arxiv.org/pdf/2302.09019v2.pdf | Tensor Networks Meet Neural Networks: A Survey and Future Perspectives | Tensor networks (TNs) and neural networks (NNs) are two fundamental data modeling approaches. TNs were introduced to solve the curse of dimensionality in large-scale tensors by converting an exponential number of dimensions to polynomial complexity. As a result, they have attracted significant attention in the fields of quantum physics and machine learning. Meanwhile, NNs have displayed exceptional performance in various applications, e.g., computer vision, natural language processing, and robotics research. Interestingly, although these two types of networks originate from different observations, they are inherently linked through the common multilinearity structure underlying both TNs and NNs, thereby motivating a significant number of intellectual developments regarding combinations of TNs and NNs. In this paper, we refer to these combinations as tensorial neural networks (TNNs), and present an introduction to TNNs in three primary aspects: network compression, information fusion, and quantum circuit simulation. Furthermore, this survey also explores methods for improving TNNs, examines flexible toolboxes for implementing TNNs, and documents TNN development while highlighting potential future directions. To the best of our knowledge, this is the first comprehensive survey that bridges the connections among NNs, TNs, and quantum circuits. We provide a curated list of TNNs at \url{https://github.com/tnbar/awesome-tensorial-neural-networks}. | ['Andrzej Cichocki', 'Zenglin Xu', 'Guangxi Li', 'Xiangli Yang', 'Yu Pan', 'Maolin Wang'] | 2023-01-22 | null | null | null | null | ['tensor-networks'] | ['methodology'] | [ 3.24096292e-01 -6.89064711e-02 -3.90889376e-01 -1.27584621e-01
-3.04364026e-01 -5.71418762e-01 4.04577464e-01 -2.44871713e-02
-1.83295503e-01 5.28732479e-01 -1.51009917e-01 -5.70379436e-01
-6.78430080e-01 -8.07674348e-01 -5.95815003e-01 -8.01734328e-01
-2.14000642e-01 9.50219110e-02 -2.66763151e-01 -5.58905602e-01
3.33808780e-01 5.24074078e-01 -1.24478364e+00 2.27668434e-02
7.74667919e-01 1.48467183e+00 -3.27392578e-01 4.69970375e-01
1.00814454e-01 6.74666643e-01 -1.32134706e-01 -7.32141733e-01
3.19603950e-01 -1.33309558e-01 -8.87554169e-01 -5.76866865e-01
2.27694139e-01 -7.15442896e-02 -1.08695543e+00 1.34752524e+00
2.44685799e-01 7.19562471e-02 3.71022761e-01 -1.57055819e+00
-9.14626718e-01 7.84883440e-01 -5.79250418e-02 2.19902977e-01
-1.38982892e-01 2.71031447e-02 1.23609245e+00 -7.48290062e-01
3.90050292e-01 9.90674257e-01 7.71872520e-01 5.68777323e-01
-1.15349054e+00 -6.00151062e-01 -4.23011512e-01 5.79896748e-01
-1.31800354e+00 -4.43153590e-01 9.34168994e-01 -3.02181870e-01
8.00740361e-01 4.47607428e-01 7.10694313e-01 8.97042453e-01
4.41443652e-01 7.69347191e-01 9.50633705e-01 -3.55784595e-01
2.62847513e-01 -1.95461512e-01 3.01060081e-01 7.08143294e-01
3.91327560e-01 2.38655224e-01 -8.20502043e-01 -2.71562636e-01
8.39526176e-01 5.31509742e-02 -1.35934502e-01 -3.25144619e-01
-1.77521670e+00 8.40565562e-01 6.94434464e-01 5.04706562e-01
-2.33811542e-01 5.39614618e-01 6.85948193e-01 3.46985072e-01
6.84192479e-02 5.84497452e-01 -3.79886836e-01 -2.45081127e-01
-3.83467555e-01 3.75891715e-01 7.70478487e-01 1.06272721e+00
7.69543529e-01 2.16732815e-01 1.15097933e-01 6.76143289e-01
-6.14151545e-02 5.85836768e-01 2.30255976e-01 -1.42501307e+00
3.54068637e-01 4.55687404e-01 -3.99639666e-01 -1.29620957e+00
-5.96759260e-01 -4.51046050e-01 -1.65366495e+00 -4.68392044e-01
2.83339977e-01 6.64240494e-02 -4.61622953e-01 1.72137964e+00
6.88455552e-02 -1.72367081e-01 1.52963204e-02 7.54962981e-01
7.79559433e-01 5.27195036e-01 -5.36461413e-01 7.00936243e-02
1.35637593e+00 -6.93702042e-01 -8.63943338e-01 3.32758874e-01
8.15823317e-01 -4.78974462e-01 3.38048130e-01 5.16604960e-01
-1.10381484e+00 -1.79538503e-01 -1.08863330e+00 -4.09984916e-01
-5.34145772e-01 -4.39121649e-02 1.34521568e+00 5.61092138e-01
-1.23357749e+00 1.05598104e+00 -9.78819966e-01 -2.91127115e-01
3.82101536e-01 6.26841664e-01 -2.19827205e-01 -8.39209184e-02
-1.57069266e+00 7.63416588e-01 3.08730811e-01 5.89346707e-01
-4.71013963e-01 -5.09146571e-01 -5.23864865e-01 -1.66258380e-01
3.35669219e-01 -8.68410647e-01 1.26197791e+00 -1.65939927e-01
-1.37773073e+00 3.00307155e-01 -2.12093830e-01 -5.75261772e-01
-2.72335052e-01 3.53997737e-01 -7.20362127e-01 1.98854268e-01
-4.10979949e-02 6.28558159e-01 4.35663819e-01 -6.21128976e-01
-1.81337923e-01 -4.31743473e-01 3.65491718e-01 -1.22132443e-01
-5.80872655e-01 -1.50031477e-01 -9.77059677e-02 -4.90070432e-01
7.47555912e-01 -1.16769493e+00 -5.15602827e-01 -2.50126850e-02
-7.47554958e-01 -4.53209698e-01 6.51141942e-01 -1.23082854e-01
1.20165050e+00 -1.86040878e+00 6.06217861e-01 3.70155752e-01
1.01432037e+00 1.79415166e-01 -6.06566183e-02 6.14012539e-01
-1.47304848e-01 3.47557038e-01 -2.43740037e-01 -2.52044387e-02
1.86748207e-01 3.49856555e-01 -2.79331565e-01 4.06236827e-01
4.10800315e-02 1.19910550e+00 -9.07504439e-01 -1.79217368e-01
1.11925386e-01 5.32549918e-01 -5.32583654e-01 -5.03220379e-01
-1.21584199e-01 5.14016569e-01 -4.71970737e-01 8.05520892e-01
5.01705170e-01 -4.10869181e-01 4.50124852e-02 -8.66652966e-01
-1.00467078e-01 4.46443200e-01 -8.40389431e-01 1.61283803e+00
4.54840288e-02 8.67283762e-01 1.81015357e-02 -1.44873607e+00
6.36103928e-01 3.70799005e-01 8.25623512e-01 -8.34051490e-01
2.68721730e-01 6.27682686e-01 3.82028639e-01 -3.06371480e-01
8.88710022e-01 -8.70502591e-02 -1.93239212e-01 2.87311316e-01
2.62144119e-01 -2.18714043e-01 4.76256251e-01 4.94872898e-01
9.82359886e-01 -4.29598570e-01 -8.25369209e-02 -8.97092447e-02
3.77679467e-01 -1.41379863e-01 5.32270551e-01 7.07896948e-01
-4.05958951e-01 7.19450265e-02 6.43721759e-01 -6.50752306e-01
-1.20450509e+00 -9.05778766e-01 -4.32418436e-01 6.37541950e-01
7.22292736e-02 -8.39669764e-01 -6.41053498e-01 1.09830275e-02
-2.37134695e-01 1.82953387e-01 -4.31502193e-01 -2.91811556e-01
-5.43190300e-01 -9.23486888e-01 1.17893624e+00 3.24364096e-01
6.25928819e-01 -6.71924353e-01 1.62416454e-02 -1.14082053e-01
-4.64136750e-01 -1.28658676e+00 -1.46050468e-01 2.16725782e-01
-1.23572958e+00 -1.00370967e+00 -3.33958507e-01 -4.27936852e-01
3.61435413e-01 3.61835629e-01 8.10835302e-01 -6.78416863e-02
-1.20220266e-01 2.47163147e-01 -2.12575138e-01 -1.91948324e-01
-3.08004826e-01 2.04409689e-01 6.64724469e-01 -1.69275522e-01
3.24500144e-01 -8.34129035e-01 -3.50778788e-01 1.06081985e-01
-1.16399467e+00 1.22324370e-01 6.37398601e-01 8.34311962e-01
4.99166608e-01 2.21636564e-01 1.47391111e-01 -3.84418756e-01
6.92829132e-01 -3.89346510e-01 -6.21525049e-01 1.94235593e-01
-5.53330123e-01 3.18590283e-01 7.00766027e-01 -6.70892373e-02
-1.89808205e-01 -5.37611008e-01 -4.96505648e-02 -4.02662218e-01
2.42894173e-01 8.74006331e-01 2.48254240e-01 -6.41470492e-01
5.02224684e-01 2.10422188e-01 2.72988230e-02 -2.58353561e-01
4.58315820e-01 5.93183994e-01 3.38285416e-01 -6.98059499e-01
9.14703608e-01 5.02301693e-01 7.07199872e-01 -8.75041127e-01
-8.50892425e-01 -1.89912662e-01 -6.87791705e-01 -1.94333643e-01
5.50286651e-01 -4.17170703e-01 -1.24998987e+00 7.09032953e-01
-1.45658815e+00 2.27592602e-01 -3.01248968e-01 6.83998525e-01
-4.44060594e-01 3.75800580e-01 -9.82521474e-01 -7.37754762e-01
-3.44391137e-01 -1.25291610e+00 6.00994885e-01 1.52369604e-01
2.04433993e-01 -1.19208288e+00 -1.95130646e-01 4.44003940e-01
7.11436808e-01 2.12329879e-01 1.05092752e+00 -3.05080235e-01
-8.88067722e-01 -2.26088643e-01 -3.28705162e-01 5.70115447e-01
-2.73641169e-01 -1.78955961e-02 -7.67372072e-01 -1.66899160e-01
6.03948087e-02 -2.77019352e-01 8.03360462e-01 1.92327723e-01
1.37738931e+00 -3.61548454e-01 -1.30719259e-01 8.65300477e-01
1.17476583e+00 2.31436402e-01 5.11100352e-01 1.03880307e-02
9.94719684e-01 3.90262216e-01 -1.19755864e-01 1.58715576e-01
5.95648587e-01 5.94207466e-01 5.48672974e-01 2.20209643e-01
2.32592806e-01 1.53964963e-02 2.39846796e-01 1.83593214e+00
-5.17130613e-01 1.28436550e-01 -9.31566954e-01 9.27900076e-02
-1.64092183e+00 -9.48244214e-01 -5.27396739e-01 1.94214642e+00
6.54836774e-01 -4.36298512e-02 -1.07211731e-01 3.61376852e-01
7.73411274e-01 1.63854346e-01 -9.42694843e-01 -4.51851279e-01
-5.31951785e-01 2.48675779e-01 8.47744167e-01 1.84774131e-01
-9.31294322e-01 7.27900088e-01 6.13440514e+00 9.79335725e-01
-1.13935161e+00 2.02785075e-01 4.20448005e-01 5.39895594e-02
-4.56208199e-01 1.58929229e-01 -5.77762663e-01 6.84443265e-02
1.20168209e+00 -2.34083802e-01 1.01399076e+00 3.94880921e-01
-5.07767275e-02 2.42468476e-01 -1.03548157e+00 1.30524051e+00
-2.07018420e-01 -1.80849183e+00 2.03775316e-01 3.03673059e-01
8.23463857e-01 5.59308946e-01 3.26410979e-01 2.51423568e-01
-1.05926700e-01 -1.03322148e+00 4.67968702e-01 5.84556639e-01
9.00420189e-01 -5.71537793e-01 5.93193710e-01 1.48169231e-02
-1.17397380e+00 -1.74799249e-01 -3.60800236e-01 -2.75992751e-01
7.18924776e-02 9.00517941e-01 -3.74556407e-02 8.55323076e-01
5.78380704e-01 9.31786895e-01 -4.31501091e-01 9.76569772e-01
2.85768777e-01 5.17788589e-01 -4.39552933e-01 -4.02331263e-01
3.88868660e-01 -5.66391587e-01 6.80556059e-01 6.04764760e-01
2.15710998e-01 2.03326747e-01 -3.25978637e-01 1.10810316e+00
-2.85009325e-01 -2.77941376e-01 -6.32818937e-01 -6.45841241e-01
5.64284563e-01 1.38762176e+00 -4.42376882e-01 -1.92027286e-01
-1.80942014e-01 4.28831726e-01 1.33311495e-01 3.88959438e-01
-6.94310784e-01 -7.66734183e-01 5.96368313e-01 -2.90180236e-01
-1.91438854e-01 -6.16582811e-01 -5.85159957e-01 -1.36904669e+00
1.81400791e-01 -7.07370877e-01 -1.61068708e-01 -5.87760687e-01
-1.30685866e+00 5.78581810e-01 -6.75894413e-03 -1.28723514e+00
2.85220165e-02 -1.09467971e+00 -6.98368326e-02 4.87514615e-01
-1.29245043e+00 -7.21282721e-01 4.90667187e-02 5.22851467e-01
-4.32978153e-01 -1.56086935e-02 9.44413245e-01 6.09951496e-01
-8.17007720e-01 4.47570205e-01 7.06197739e-01 3.43662947e-01
1.21256977e-01 -9.62306857e-01 5.63756526e-01 5.00539124e-01
5.24255028e-03 1.02137887e+00 4.49601442e-01 -1.76462889e-01
-2.06241608e+00 -9.51630414e-01 8.06099892e-01 -2.20922858e-01
1.13026619e+00 -5.09610772e-01 -5.48581064e-01 5.73911726e-01
-6.71753362e-02 1.38266953e-02 6.97820187e-01 2.20944703e-01
-5.97322345e-01 -3.20726126e-01 -8.58393967e-01 6.96913123e-01
1.14584970e+00 -9.88898695e-01 6.32506087e-02 8.39394629e-01
7.71226645e-01 -4.56418335e-01 -1.33453619e+00 4.55149651e-01
7.57232547e-01 -1.19454467e+00 8.86652052e-01 -4.72909272e-01
4.05023932e-01 -1.66428387e-01 -4.49683040e-01 -1.05641055e+00
-1.89099908e-01 -9.66067791e-01 -3.57238919e-01 6.26035035e-01
2.30035350e-01 -9.22254086e-01 7.16498673e-01 6.24639809e-01
-4.04793978e-01 -1.09456396e+00 -1.35517538e+00 -8.42178106e-01
2.48630196e-01 -8.32920969e-01 5.60176730e-01 1.04120278e+00
3.46089005e-01 2.81056464e-01 -3.79425198e-01 -1.87358901e-01
8.20945680e-01 -2.51174450e-01 3.71351451e-01 -1.29683661e+00
6.19986020e-02 -7.24280000e-01 -7.59957492e-01 -1.20304787e+00
-1.53369501e-01 -1.34876299e+00 -4.68071908e-01 -1.22333765e+00
1.18906014e-01 -6.01644814e-01 -3.47740352e-01 4.05601025e-01
6.02547228e-01 5.23263812e-01 2.00852513e-01 4.34827685e-01
-7.95970261e-01 7.07799375e-01 1.45265794e+00 -7.92845562e-02
1.02903321e-01 -9.96800810e-02 -7.69066930e-01 6.52783632e-01
1.11926866e+00 -1.75889924e-01 -7.71229863e-02 -5.59358537e-01
8.52835774e-01 2.20610157e-01 6.52246535e-01 -1.08224607e+00
7.09753454e-01 -1.12573402e-02 -1.09686911e-01 -6.71387613e-01
6.23383939e-01 -4.59438831e-01 -4.85897157e-03 4.68535781e-01
-2.56369084e-01 3.27221781e-01 -1.18250906e-01 3.48998666e-01
-2.75090635e-01 -3.25586684e-02 4.25175935e-01 -1.43736243e-01
-3.73700768e-01 6.38394177e-01 -3.96355510e-01 -9.08775851e-02
5.38702667e-01 -1.15133278e-01 -6.51431143e-01 -3.39659721e-01
-5.15976131e-01 1.37085244e-01 -3.73274088e-02 2.48774096e-01
6.13770962e-01 -1.48798335e+00 -1.58323571e-01 1.49960414e-01
2.76936907e-02 1.52731702e-01 4.90246892e-01 1.44858313e+00
-2.77823478e-01 1.19102073e+00 -7.05366954e-02 -5.68671703e-01
-5.60141623e-01 3.88328582e-01 4.56457585e-01 -1.10280208e-01
-1.69946656e-01 7.62644470e-01 -2.43456930e-01 -8.46210122e-01
2.03218594e-01 -6.73410058e-01 1.03773274e-01 -3.01680833e-01
2.38683596e-01 7.29067326e-01 1.57900169e-01 -7.63007224e-01
-2.47032270e-01 5.57055652e-01 1.97070325e-03 -5.52180363e-03
1.22065091e+00 1.81303956e-02 -9.42596853e-01 7.31618881e-01
1.44866407e+00 -3.99502993e-01 -3.91115785e-01 -5.56890309e-01
-1.05965766e-03 1.90884635e-01 -1.95003729e-02 -4.31322545e-01
-1.21381629e+00 1.07089102e+00 1.61004558e-01 5.62039554e-01
1.09899569e+00 3.64042930e-02 1.25257862e+00 1.03218734e+00
5.67195654e-01 -9.89136934e-01 -1.76068619e-02 9.91786838e-01
6.14787936e-01 -8.43168318e-01 -9.35056582e-02 -3.60186309e-01
-1.39741570e-01 1.43447614e+00 1.85640842e-01 -3.51589881e-02
9.19049799e-01 3.68541991e-03 -4.44720685e-01 -4.21037316e-01
-7.44461119e-01 1.57686368e-01 2.57932484e-01 3.47428232e-01
3.62729818e-01 2.08214656e-01 -7.92151317e-02 2.46102825e-01
-7.30399489e-01 8.47226381e-03 6.97337806e-01 6.83656931e-01
-4.99507561e-02 -1.19131863e+00 -3.29411864e-01 7.88877964e-01
-2.75729895e-01 -5.05760312e-01 -1.91325426e-01 3.98889184e-01
2.11084798e-01 9.00281310e-01 -2.18110055e-01 -1.00917923e+00
1.67959526e-01 -1.85024247e-01 6.08445346e-01 -3.65455560e-02
-2.91787773e-01 -4.47686166e-01 -5.53972647e-02 -6.82108223e-01
-6.42451167e-01 -6.19521976e-01 -9.96962190e-01 -1.05368185e+00
-4.85744536e-01 2.74475396e-01 1.14240420e+00 1.04629457e+00
4.78469253e-01 5.19510150e-01 5.07815301e-01 -8.34204376e-01
-6.94682896e-01 -8.09929907e-01 -6.49015903e-01 -2.52213746e-01
4.39557850e-01 -6.38561130e-01 -3.46532404e-01 -5.02338290e-01] | [5.872995853424072, 5.042004585266113] |
252faeb3-a9d3-4848-af32-cf0cb36989c5 | incorporating-joint-embeddings-into-goal | 2001.10468 | null | https://arxiv.org/abs/2001.10468v1 | https://arxiv.org/pdf/2001.10468v1.pdf | Incorporating Joint Embeddings into Goal-Oriented Dialogues with Multi-Task Learning | Attention-based encoder-decoder neural network models have recently shown promising results in goal-oriented dialogue systems. However, these models struggle to reason over and incorporate state-full knowledge while preserving their end-to-end text generation functionality. Since such models can greatly benefit from user intent and knowledge graph integration, in this paper we propose an RNN-based end-to-end encoder-decoder architecture which is trained with joint embeddings of the knowledge graph and the corpus as input. The model provides an additional integration of user intent along with text generation, trained with a multi-task learning paradigm along with an additional regularization technique to penalize generating the wrong entity as output. The model further incorporates a Knowledge Graph entity lookup during inference to guarantee the generated output is state-full based on the local knowledge graph provided. We finally evaluated the model using the BLEU score, empirical evaluation depicts that our proposed architecture can aid in the betterment of task-oriented dialogue system`s performance. | ['Firas Kassawat', 'Debanjan Chaudhuri', 'Jens Lehmann'] | 2020-01-28 | null | null | null | null | ['goal-oriented-dialog', 'goal-oriented-dialogue-systems'] | ['natural-language-processing', 'natural-language-processing'] | [ 1.61893055e-01 1.19265473e+00 -3.09305917e-02 -4.57763791e-01
-8.81890059e-01 -4.27245170e-01 1.00829399e+00 3.81160006e-02
-3.71338725e-01 9.41537559e-01 8.23819578e-01 -2.85653621e-01
2.07038641e-01 -8.54495585e-01 -6.73548520e-01 -1.87230468e-01
2.67059624e-01 7.53418267e-01 -1.01677753e-01 -6.58172131e-01
5.28853200e-02 -4.43917930e-01 -9.50781941e-01 5.46379566e-01
1.14381921e+00 6.42677248e-01 2.14680329e-01 1.21081543e+00
-1.32685900e-01 1.27783012e+00 -7.44760394e-01 -9.04338777e-01
-6.34442344e-02 -7.03266978e-01 -1.20497072e+00 -2.42667973e-01
1.48973271e-01 -5.62639832e-01 -4.34759200e-01 7.52510011e-01
7.04714239e-01 2.53017873e-01 7.24799097e-01 -1.12054193e+00
-1.05432832e+00 9.60894942e-01 1.78480163e-01 -2.54710257e-01
5.97428083e-01 4.01082844e-01 1.32170320e+00 -7.26563334e-01
7.35876441e-01 1.27558696e+00 2.98930287e-01 9.43216980e-01
-1.03471732e+00 7.02766469e-04 -9.00037065e-02 5.92209026e-02
-9.03644145e-01 -6.36420429e-01 8.03446114e-01 -1.55660108e-01
1.73275542e+00 -2.62064058e-02 2.68234789e-01 1.22417474e+00
3.29164296e-01 9.65295672e-01 4.45166916e-01 -5.54509580e-01
-4.26071249e-02 3.42256486e-01 7.43251964e-02 1.00716925e+00
-7.65775517e-02 -8.80535692e-02 -6.94404960e-01 6.79503381e-02
5.35199583e-01 -6.75068080e-01 -2.43899509e-01 -2.24544778e-02
-1.01682198e+00 9.06922936e-01 3.75967592e-01 8.23108628e-02
-4.82419223e-01 2.38178059e-01 6.04047894e-01 3.15774292e-01
5.05288005e-01 5.56508303e-01 -4.58655566e-01 -2.79140443e-01
-7.57501900e-01 4.57418114e-01 1.27185988e+00 1.16799235e+00
5.77318788e-01 2.72304118e-01 -9.03301060e-01 8.40917587e-01
5.72919309e-01 3.00945252e-01 6.02880597e-01 -7.35923529e-01
9.35441554e-01 8.79969597e-01 1.53785840e-01 -6.51462615e-01
-2.73848921e-01 -3.18107247e-01 -7.84126639e-01 -1.77296828e-02
4.77255732e-01 -6.98105335e-01 -6.30411923e-01 1.95651662e+00
9.12768841e-02 -2.12239161e-01 6.97281063e-01 7.11250544e-01
8.90044987e-01 7.96255887e-01 2.62547433e-01 8.81574973e-02
1.30704880e+00 -1.28605020e+00 -1.04513526e+00 -3.83964151e-01
8.58770907e-01 -4.95554835e-01 9.04112458e-01 5.18113784e-02
-1.38442361e+00 -5.68487763e-01 -9.16787326e-01 -5.03387749e-01
-3.74185652e-01 4.51959312e-01 4.02672082e-01 4.04098749e-01
-1.25697768e+00 5.42919219e-01 -5.77722132e-01 -3.05110186e-01
9.81271733e-03 2.95200706e-01 -2.22725987e-01 2.31266901e-01
-1.61078107e+00 1.34807980e+00 7.93950140e-01 2.71608680e-01
-7.89860129e-01 -3.43192965e-01 -1.12571490e+00 2.87962645e-01
1.86556712e-01 -1.20322299e+00 1.58396709e+00 -7.00727046e-01
-1.95854664e+00 3.19015652e-01 -5.73499314e-02 -7.32618988e-01
5.77728510e-01 -3.29584837e-01 -1.83912084e-01 -4.01586853e-02
-2.27410328e-02 8.33397090e-01 4.92258668e-01 -8.45807552e-01
-5.47597349e-01 -1.01094410e-01 3.32158476e-01 6.56853259e-01
-3.39731038e-01 -3.70159417e-01 -3.26367885e-01 -3.94050509e-01
-5.72995663e-01 -6.91307366e-01 -1.89200640e-01 -5.07785857e-01
-7.10293651e-01 -5.77070534e-01 5.91754735e-01 -1.08884084e+00
1.44250917e+00 -1.45648086e+00 3.42565209e-01 -2.98490584e-01
-4.78498675e-02 4.10836369e-01 -2.64164627e-01 8.64677131e-01
3.31799775e-01 -6.55568764e-02 -1.46528125e-01 -6.10899031e-01
3.31915349e-01 1.15662701e-01 -1.41215861e-01 -8.95309225e-02
7.32727408e-01 1.25737047e+00 -1.15189397e+00 -3.73580307e-01
1.90455228e-01 4.87817407e-01 -7.22553909e-01 7.78852761e-01
-8.48577499e-01 1.85872361e-01 -3.04089427e-01 2.36443412e-02
1.35573030e-01 -2.95092136e-01 2.90936917e-01 -1.25761509e-01
2.28893980e-01 7.14003444e-01 -7.16709256e-01 2.11101270e+00
-8.07177007e-01 5.04003346e-01 1.55315837e-02 -6.36868536e-01
8.70543301e-01 8.07429433e-01 -3.17292809e-01 -6.15623951e-01
1.20455787e-01 -9.65557098e-02 1.32983655e-01 -4.20697451e-01
1.06551051e+00 -1.02920365e-03 -2.23966062e-01 6.86033487e-01
5.91160893e-01 5.90009540e-02 2.49510333e-01 6.42901480e-01
9.12138522e-01 4.93307471e-01 1.59183502e-01 -4.29450255e-03
6.03778720e-01 -5.39627438e-03 9.86814499e-02 6.94569349e-01
1.94880188e-01 3.82345766e-01 5.73968828e-01 2.90097669e-02
-1.13848841e+00 -6.48478091e-01 5.78823030e-01 1.25793302e+00
-3.20290893e-01 -4.63392705e-01 -1.01395392e+00 -1.06406498e+00
-3.70339364e-01 1.47937894e+00 -4.02391851e-01 -5.33383429e-01
-7.14120746e-01 -4.89416033e-01 1.01537776e+00 4.71690655e-01
5.41659057e-01 -1.26424599e+00 -3.12527597e-01 5.88085473e-01
-5.29961169e-01 -1.09652638e+00 -4.60151851e-01 3.00705582e-02
-5.97875416e-01 -7.71045387e-01 -7.73785889e-01 -8.09345424e-01
6.84567511e-01 -5.14961958e-01 1.11040688e+00 -8.00570101e-02
1.49424449e-01 2.91135669e-01 -3.39978367e-01 -1.27753302e-01
-1.07626700e+00 4.38694268e-01 -3.36745441e-01 3.44044506e-03
1.31276548e-01 -3.23010176e-01 -4.10097182e-01 -2.07491711e-01
-7.13828802e-01 6.60839558e-01 5.52160382e-01 1.11504495e+00
-1.22408122e-01 -4.43614304e-01 1.10170138e+00 -8.89033616e-01
1.30791771e+00 -4.98844624e-01 -3.00949067e-01 4.69519109e-01
-6.71139359e-01 8.27350199e-01 9.46697414e-01 1.53802137e-03
-1.64368165e+00 -1.08667977e-01 -4.86775547e-01 3.10253911e-02
7.48225972e-02 6.33720756e-01 -2.87067443e-01 6.33770287e-01
6.53431773e-01 4.46645230e-01 -3.32641113e-03 -3.56926382e-01
9.87639427e-01 9.27455842e-01 5.03863037e-01 -5.39262414e-01
2.93730408e-01 -3.25309187e-01 -3.12883228e-01 -3.94769311e-01
-8.56380105e-01 -1.58594504e-01 -4.43617761e-01 -1.35063156e-01
1.03734910e+00 -9.90136981e-01 -6.74015522e-01 2.88551569e-01
-1.73300600e+00 -5.05827904e-01 -1.35596171e-01 1.97412014e-01
-7.33870387e-01 3.59354109e-01 -6.51186705e-01 -1.08683443e+00
-9.63640630e-01 -8.25368941e-01 9.58019912e-01 2.35367551e-01
-2.98918277e-01 -1.42821705e+00 1.83687121e-01 4.51899081e-01
5.31173408e-01 -8.78011882e-02 1.01273119e+00 -1.24881017e+00
-4.69185382e-01 -1.85842916e-01 -3.05946857e-01 3.96532476e-01
7.04177842e-02 -3.36347252e-01 -9.97171521e-01 -9.56021994e-02
-3.32366109e-01 -6.75295413e-01 6.72609568e-01 -4.92748097e-02
3.32878113e-01 -8.80416274e-01 -3.49931717e-02 2.48664021e-02
1.29380536e+00 -1.57588646e-01 6.64061010e-01 -1.03586458e-01
7.51358330e-01 7.11890578e-01 3.31913114e-01 4.07735616e-01
8.60134602e-01 6.78240240e-01 3.28180254e-01 1.52651519e-01
-5.93966067e-01 -7.91551054e-01 6.69252634e-01 9.42481637e-01
8.75363871e-02 -8.47193837e-01 -6.67211652e-01 7.87554979e-01
-2.15493751e+00 -9.13638175e-01 -2.36314666e-02 1.99677157e+00
1.27102935e+00 1.95446089e-01 -1.09619237e-01 -3.69101524e-01
6.17369115e-01 1.90251768e-01 -3.61852556e-01 -7.66200542e-01
9.42732990e-02 9.16787386e-02 3.47197205e-01 9.95892286e-01
-7.49463618e-01 1.30059040e+00 5.69562864e+00 6.80574536e-01
-7.90163755e-01 1.00465797e-01 3.42202783e-01 1.90910682e-01
-3.39558572e-01 4.15479504e-02 -1.09261310e+00 3.31422210e-01
1.37817264e+00 -3.83545607e-01 1.83357701e-01 8.06273282e-01
2.29607686e-01 1.65143564e-01 -1.12353027e+00 3.90210241e-01
2.00157687e-01 -1.24043632e+00 3.13409209e-01 -1.84462965e-01
4.71876591e-01 -1.74671873e-01 -3.89373243e-01 7.96036363e-01
8.67402077e-01 -9.44854915e-01 4.56671834e-01 6.88397408e-01
5.89106083e-01 -6.49838090e-01 8.45383167e-01 6.63961709e-01
-7.73197770e-01 1.34744510e-01 -2.09601030e-01 -2.40911111e-01
5.69920957e-01 2.03134909e-01 -1.72371948e+00 8.69539201e-01
-3.23405743e-01 3.89070928e-01 -4.29594874e-01 4.20536786e-01
-6.87629461e-01 4.00866419e-01 6.19899817e-02 -3.81778210e-01
3.08259040e-01 6.22365950e-03 5.33839464e-01 1.52271986e+00
1.52182981e-01 -1.74685001e-01 4.40992117e-02 1.09793520e+00
-3.69322419e-01 6.58271909e-02 -6.96577609e-01 -3.96255672e-01
1.61095858e-01 1.23542011e+00 4.05014828e-02 -5.40286183e-01
-3.18664551e-01 1.41300082e+00 7.52850175e-01 2.56863475e-01
-6.62745953e-01 -7.02384353e-01 3.11688542e-01 -2.00847551e-01
2.40770519e-01 -1.62554726e-01 -9.66585726e-02 -1.13359165e+00
-1.40656024e-01 -6.16041958e-01 2.66520351e-01 -7.72363544e-01
-1.02066243e+00 8.01711738e-01 -2.51787782e-01 -6.36901975e-01
-1.07911754e+00 -4.23183322e-01 -7.07210124e-01 1.14432728e+00
-1.63106370e+00 -1.48165119e+00 6.30725101e-02 3.77405643e-01
8.66217554e-01 -3.03184718e-01 1.09892464e+00 5.70414998e-02
-5.69500625e-01 8.67390931e-01 -1.10745884e-01 4.10312027e-01
7.66577899e-01 -1.54987538e+00 5.89137316e-01 8.25383365e-01
-6.76287711e-02 5.69933295e-01 7.17512369e-01 -8.62191916e-01
-1.32570732e+00 -1.35725260e+00 1.50382638e+00 -4.53651816e-01
4.37914789e-01 -3.86989683e-01 -6.35174215e-01 7.63099611e-01
9.47036505e-01 -5.86940050e-01 4.99052793e-01 1.24745637e-01
-2.84522157e-02 4.57726955e-01 -9.00896966e-01 6.79741263e-01
7.09998369e-01 -4.94026005e-01 -8.95362020e-01 4.39858377e-01
8.43411386e-01 -4.97502476e-01 -9.28742468e-01 6.86661601e-02
2.35809237e-01 -6.38161838e-01 5.23199201e-01 -8.68252456e-01
8.42908740e-01 -6.31407574e-02 1.67663261e-01 -1.58110273e+00
-1.40405059e-01 -9.07271862e-01 -6.30392849e-01 1.33361042e+00
1.01070666e+00 -1.40800133e-01 7.32510448e-01 9.31509078e-01
-5.38331032e-01 -5.56391954e-01 -8.30650866e-01 -4.78426784e-01
4.56849113e-02 -6.44661719e-03 3.27144355e-01 5.58863997e-01
5.78774333e-01 1.17758250e+00 -8.39699149e-01 -6.93961978e-02
3.11621398e-01 -2.17715144e-01 7.92692840e-01 -8.24527621e-01
-4.02435601e-01 -2.97910064e-01 1.32209852e-01 -1.32809496e+00
3.75097871e-01 -1.08061910e+00 1.93060756e-01 -2.20552754e+00
-5.28477281e-02 8.45917407e-03 1.78048491e-01 6.77117765e-01
-4.53050762e-01 -2.96915412e-01 1.78294722e-02 -2.39181250e-01
-8.42533767e-01 9.60246205e-01 1.26390457e+00 -3.09153479e-02
-2.73893833e-01 -1.21384293e-01 -7.69180119e-01 2.58433133e-01
7.51474380e-01 -2.24429116e-01 -5.98433197e-01 -5.96498191e-01
3.38278562e-01 4.90941852e-01 1.23217285e-01 -6.55828297e-01
5.43064475e-01 1.68884948e-01 8.25185701e-02 -2.51264900e-01
5.16691267e-01 -3.96025002e-01 -3.34407389e-01 4.08872515e-01
-7.99342752e-01 -1.13737211e-01 1.07088260e-01 5.49567878e-01
-1.94899738e-01 -4.06728506e-01 4.59589869e-01 -2.74473697e-01
-4.69697326e-01 -1.01991139e-01 -4.25619572e-01 2.83684075e-01
5.88405073e-01 3.70019488e-02 -4.53185797e-01 -7.94347882e-01
-7.58433044e-01 4.02796954e-01 3.97732481e-02 5.42830050e-01
4.85966235e-01 -1.17938375e+00 -1.01311409e+00 6.14044480e-02
3.25087123e-02 -9.82899591e-02 1.54279247e-01 3.83863360e-01
-4.00683790e-01 8.83324504e-01 -1.88108996e-01 -1.84238460e-02
-1.03466392e+00 2.10683450e-01 3.37868661e-01 -9.79691625e-01
-3.05268377e-01 7.32297003e-01 -1.93007171e-01 -7.41588414e-01
2.87789196e-01 -1.71796590e-01 -3.55703026e-01 -9.81409773e-02
3.48184615e-01 2.46874243e-01 8.00039619e-02 -4.23336685e-01
1.11109510e-01 -2.30213255e-01 -2.76147902e-01 -3.66520822e-01
1.00690758e+00 -2.15045795e-01 1.81456491e-01 1.05889022e-01
1.02317369e+00 -2.87032396e-01 -1.18301117e+00 -2.29281083e-01
-1.81501061e-02 9.44176465e-02 7.45059038e-03 -1.37708855e+00
-6.35969222e-01 1.06719363e+00 -6.26726523e-02 1.08861320e-01
5.91152430e-01 -3.00682247e-01 1.03963399e+00 5.70859015e-01
5.39747290e-02 -1.25319648e+00 2.07457885e-01 1.03382778e+00
9.48934674e-01 -1.33040237e+00 -3.59586000e-01 -1.02255754e-01
-1.09041786e+00 1.05572677e+00 9.02371347e-01 4.90452833e-02
7.56811053e-02 -9.79588851e-02 -5.24827316e-02 -1.82756726e-02
-1.31514549e+00 -1.44511208e-01 3.74514967e-01 5.75721204e-01
8.26494873e-01 -1.24065518e-01 -3.50204498e-01 8.47519696e-01
-4.18540448e-01 1.17746880e-02 4.46715236e-01 6.66779160e-01
-4.50250477e-01 -1.27555311e+00 3.26754868e-01 3.17536801e-01
-5.06500185e-01 -3.99529457e-01 -5.53694725e-01 3.31644118e-01
-4.21653360e-01 1.19084740e+00 -1.17965639e-01 -2.63653606e-01
4.02659267e-01 6.87052846e-01 3.82727116e-01 -9.22322512e-01
-1.09930909e+00 -3.03959072e-01 9.28590238e-01 -1.59629479e-01
1.52524173e-01 -1.73934400e-02 -1.24426055e+00 -2.18744501e-02
-5.52731276e-01 4.72462505e-01 6.16048276e-01 8.97809088e-01
6.91884756e-01 7.76852846e-01 3.68751377e-01 -3.67936790e-01
-8.00960660e-01 -1.46227038e+00 -1.42582327e-01 3.84533197e-01
1.64782509e-01 6.06099097e-03 -5.47686219e-02 2.12159991e-01] | [12.336528778076172, 8.533047676086426] |
ac07df4f-6e0c-49b2-b810-ec709d386ac4 | assessing-bias-in-face-image-quality | 2211.15265 | null | https://arxiv.org/abs/2211.15265v1 | https://arxiv.org/pdf/2211.15265v1.pdf | Assessing Bias in Face Image Quality Assessment | Face image quality assessment (FIQA) attempts to improve face recognition (FR) performance by providing additional information about sample quality. Because FIQA methods attempt to estimate the utility of a sample for face recognition, it is reasonable to assume that these methods are heavily influenced by the underlying face recognition system. Although modern face recognition systems are known to perform well, several studies have found that such systems often exhibit problems with demographic bias. It is therefore likely that such problems are also present with FIQA techniques. To investigate the demographic biases associated with FIQA approaches, this paper presents a comprehensive study involving a variety of quality assessment methods (general-purpose image quality assessment, supervised face quality assessment, and unsupervised face quality assessment methods) and three diverse state-of-theart FR models. Our analysis on the Balanced Faces in the Wild (BFW) dataset shows that all techniques considered are affected more by variations in race than sex. While the general-purpose image quality assessment methods appear to be less biased with respect to the two demographic factors considered, the supervised and unsupervised face image quality assessment methods both show strong bias with a tendency to favor white individuals (of either sex). In addition, we found that methods that are less racially biased perform worse overall. This suggests that the observed bias in FIQA methods is to a significant extent related to the underlying face recognition system. | ['Vitomir Štruc', 'Žiga Babnik'] | 2022-11-28 | null | null | null | null | ['face-image-quality', 'face-image-quality-assessment'] | ['computer-vision', 'computer-vision'] | [ 3.18919793e-02 -1.30566105e-01 1.32373618e-02 -8.96451116e-01
-7.19520271e-01 -4.12844419e-01 7.14996457e-01 -3.99134248e-01
-4.24749434e-01 5.89160800e-01 3.52628738e-01 -1.11029018e-02
-1.16821162e-01 -9.01136160e-01 -3.54230672e-01 -7.60474980e-01
1.25582546e-01 5.16012728e-01 -6.24391079e-01 -1.73447937e-01
4.47279900e-01 8.38678896e-01 -1.92077863e+00 2.06945673e-01
6.64056659e-01 8.95704210e-01 -7.00257957e-01 6.17783844e-01
-1.96235716e-01 7.19290078e-01 -9.69037652e-01 -8.91635001e-01
3.99882764e-01 -5.91102600e-01 -8.18593502e-01 2.09464222e-01
1.49438453e+00 -5.54138064e-01 -2.12684810e-01 1.07461584e+00
7.81144023e-01 -6.73302636e-02 8.27025533e-01 -1.43407154e+00
-6.12532437e-01 1.36138409e-01 -6.00177646e-01 3.37677598e-01
4.80580598e-01 4.44556653e-01 7.09484756e-01 -9.95775700e-01
6.15750849e-01 1.99476647e+00 7.64581621e-01 7.44632781e-01
-1.48465288e+00 -1.13312078e+00 -1.61799237e-01 -9.10404101e-02
-1.43086123e+00 -1.28206563e+00 5.46280384e-01 -4.79415447e-01
4.50376213e-01 1.20244831e-01 2.84350753e-01 8.48496437e-01
4.12129983e-02 2.05941811e-01 1.58960867e+00 -2.80718207e-01
9.04564485e-02 2.47722283e-01 6.76651597e-02 7.96095729e-01
3.34866762e-01 4.14136857e-01 -7.46551454e-01 -6.46108508e-01
5.56999445e-01 -4.24648434e-01 -7.30500892e-02 -2.53723294e-01
-7.99332023e-01 9.22619522e-01 2.27763161e-01 3.19896102e-01
-3.10963362e-01 -9.82153043e-03 2.51100183e-01 5.81135392e-01
4.70684588e-01 5.06286263e-01 -1.54519454e-01 -2.29675435e-02
-1.24886048e+00 2.52947122e-01 7.97289491e-01 3.27949554e-01
9.23424780e-01 4.12034422e-01 -2.93257177e-01 1.08052540e+00
3.72515887e-01 9.50414956e-01 2.11625472e-01 -1.39235604e+00
8.07237774e-02 4.27155465e-01 -6.98144287e-02 -1.25920737e+00
-1.67419150e-01 -2.43669171e-02 -4.28031653e-01 8.43324900e-01
8.83549571e-01 1.63469926e-01 -8.84952784e-01 1.84637642e+00
6.31815270e-02 -4.82375890e-01 -1.03546642e-01 6.69032693e-01
8.64712536e-01 2.07721636e-01 4.06183720e-01 -3.53632271e-01
1.39355052e+00 -3.07188213e-01 -5.61595440e-01 -9.95688438e-02
2.40687951e-01 -9.25779045e-01 1.07298374e+00 4.15262550e-01
-1.08838320e+00 -7.48312533e-01 -8.32533538e-01 2.90812850e-01
-9.36946049e-02 3.84300910e-02 5.33201575e-01 1.85822952e+00
-1.37688756e+00 4.45581853e-01 -1.83068410e-01 -5.53386152e-01
9.43633676e-01 5.87769330e-01 -9.71801758e-01 -4.62868035e-01
-5.68651497e-01 8.81076992e-01 -1.94214672e-01 -1.23573288e-01
-1.02692759e+00 -7.51664639e-01 -7.28619814e-01 -2.32327923e-01
9.13188532e-02 -3.94432813e-01 9.28275585e-01 -1.74067688e+00
-1.15578616e+00 1.22285664e+00 -6.36898398e-01 -9.84390676e-02
2.99812913e-01 9.77709964e-02 -7.77939141e-01 4.79635745e-01
-9.10349637e-02 8.34548354e-01 1.29178691e+00 -1.27077091e+00
-2.31718197e-01 -8.46205592e-01 -9.10489187e-02 -4.91141854e-03
-2.64682561e-01 5.44115901e-01 2.01539889e-01 -4.45540547e-01
-2.79286146e-01 -6.70466363e-01 4.29845572e-01 3.54482949e-01
3.08164150e-01 -2.60584980e-01 6.88857079e-01 -6.90784872e-01
1.02356267e+00 -2.18943167e+00 -2.92146713e-01 3.09260190e-01
2.36559838e-01 2.80031502e-01 -4.03245032e-01 6.36340901e-02
-2.30200410e-01 4.64878529e-01 -5.11558913e-02 -2.88184166e-01
-9.41587090e-02 1.93086684e-01 1.45107627e-01 7.46328652e-01
5.48011720e-01 5.53532183e-01 -5.74827075e-01 -6.98585093e-01
-6.67115971e-02 4.89434212e-01 -7.16936946e-01 1.36691347e-01
3.56790364e-01 1.58831716e-01 1.12936124e-01 9.85431492e-01
9.81104195e-01 4.15123969e-01 8.32386408e-03 -3.24831069e-01
1.11968368e-01 -8.91670808e-02 -1.05016601e+00 9.21721280e-01
-1.79976776e-01 6.16016686e-01 3.34502965e-01 -6.30969644e-01
9.72312689e-01 3.44336092e-01 3.66522938e-01 -8.21845472e-01
1.41206831e-01 3.32806170e-01 5.47109604e-01 -2.54470766e-01
4.57684815e-01 -6.94685817e-01 5.68973303e-01 6.07396126e-01
3.26424569e-01 -1.44405290e-01 9.53929201e-02 6.63855895e-02
8.25166345e-01 -2.69038975e-01 -9.59702022e-03 -5.85370898e-01
6.62171721e-01 -1.85945570e-01 5.73454022e-01 5.74605763e-01
-8.40629935e-01 7.35402882e-01 5.37801206e-01 -2.81863242e-01
-9.41312611e-01 -1.12198162e+00 -4.19112325e-01 9.27704751e-01
-5.24638534e-01 -2.10205317e-01 -9.26173031e-01 -8.94991159e-01
1.60733312e-01 3.63173932e-01 -8.85549843e-01 -2.26679102e-01
-1.16703935e-01 -8.58844757e-01 8.56015205e-01 3.24967772e-01
5.93614399e-01 -9.96129036e-01 -2.53574312e-01 -3.62733513e-01
1.33865789e-01 -6.32509649e-01 -3.05672884e-01 -6.45560503e-01
-9.71557856e-01 -1.32957220e+00 -8.64533901e-01 -4.03002203e-01
8.46412539e-01 1.21925704e-01 1.57653081e+00 4.69264805e-01
-2.99651802e-01 8.92801583e-01 -2.25479409e-01 -3.33906740e-01
-6.45635962e-01 -3.05405676e-01 3.15564841e-01 1.92289412e-01
8.23326230e-01 -5.85799990e-03 -6.11735523e-01 6.97651803e-01
-8.31419706e-01 -8.93343389e-01 4.19406116e-01 8.44774008e-01
-9.49690212e-03 9.87384692e-02 6.84949398e-01 -9.60510254e-01
5.40552378e-01 -1.64666280e-01 -2.06031159e-01 2.83500910e-01
-7.97696769e-01 -3.54455076e-02 2.13594288e-02 -2.46177152e-01
-1.34748197e+00 -2.00951368e-01 -2.15407744e-01 -2.64649838e-01
-4.60849911e-01 9.70387906e-02 -4.29534137e-01 -5.19659102e-01
1.12711215e+00 -3.04575711e-01 7.00886667e-01 -1.38332859e-01
-1.70276940e-01 7.91687131e-01 1.87580675e-01 -7.92035699e-01
9.13122714e-01 5.99996448e-01 -1.21023700e-01 -1.22220874e+00
-3.36559892e-01 -9.83514413e-02 -4.05033588e-01 -6.51368320e-01
5.61376750e-01 -8.93469453e-01 -7.94281006e-01 7.39967048e-01
-4.03253615e-01 -2.11442128e-01 -1.09999761e-01 4.06471521e-01
-2.86885798e-01 3.24272454e-01 -2.74277747e-01 -1.16800487e+00
-2.00564638e-01 -1.32510340e+00 9.07313585e-01 1.81779921e-01
-3.74127626e-01 -7.21521139e-01 -1.64702028e-01 7.03949988e-01
5.30731976e-01 -2.23502833e-02 8.57948482e-01 -2.38596603e-01
-5.48976287e-02 -2.61022220e-03 -4.18291330e-01 5.85113645e-01
3.95281523e-01 4.75371599e-01 -1.36714602e+00 -5.68105042e-01
2.54483055e-03 -5.65710008e-01 6.65069044e-01 4.05987620e-01
8.08638215e-01 -1.45655826e-01 2.98164874e-01 1.97528884e-01
1.38972461e+00 9.07439739e-02 1.12793767e+00 -2.37648368e-01
3.26762259e-01 1.11855090e+00 5.71840405e-01 7.27683678e-02
3.05976206e-03 5.13150871e-01 1.20612390e-01 -2.54846253e-02
-2.66600221e-01 -5.81438765e-02 6.94335997e-01 1.36640340e-01
-1.99093401e-01 1.89053625e-01 -9.43789065e-01 3.40146542e-01
-9.29050148e-01 -1.33416903e+00 1.05318524e-01 2.41233706e+00
6.05276823e-01 -2.54223645e-01 7.03511536e-01 2.77013987e-01
6.42765760e-01 6.31398112e-02 -4.20824617e-01 -5.84846020e-01
-2.81856060e-01 4.20553595e-01 2.10437924e-01 3.17007720e-01
-6.11827612e-01 5.47363460e-01 7.98077393e+00 6.14085495e-01
-1.03773570e+00 -2.17305377e-01 1.20791876e+00 -2.11936563e-01
-2.75973618e-01 -2.42560923e-01 -6.46871090e-01 3.15278918e-01
1.10586917e+00 -1.50231449e-02 4.95531231e-01 5.81499696e-01
1.87140051e-02 -4.52008069e-01 -1.03285003e+00 1.07865286e+00
4.55779582e-01 -7.14831531e-01 2.35686541e-01 3.75389427e-01
5.49400508e-01 -4.09011364e-01 4.74081576e-01 1.99616820e-01
7.52033368e-02 -1.56986129e+00 6.90321922e-01 5.32923222e-01
1.15731621e+00 -8.99910867e-01 7.02019930e-01 -4.14839149e-01
-7.80057967e-01 -1.73906311e-01 -4.94831890e-01 -8.04174915e-02
-4.12456512e-01 5.79380453e-01 -3.60319912e-01 8.74123499e-02
9.72428322e-01 3.24256301e-01 -1.03538156e+00 6.84782267e-01
2.75858015e-01 7.60535896e-01 -1.31207764e-01 3.64832699e-01
-1.93400323e-01 -1.38748527e-01 2.01192006e-01 7.07597077e-01
2.43704140e-01 2.12688953e-01 -4.47369784e-01 7.13373363e-01
-1.24513976e-01 2.79295325e-01 -6.54873729e-01 -4.61139917e-01
3.03692192e-01 1.18076909e+00 -5.59216261e-01 -1.72721535e-01
-6.91100717e-01 5.58709204e-01 9.19798166e-02 1.84994772e-01
-2.17800722e-01 -1.02431765e-02 1.01561034e+00 2.66213268e-01
-6.34163395e-02 1.60780009e-02 -2.37197161e-01 -8.94174457e-01
-1.92995563e-01 -1.61389208e+00 4.91122603e-01 -7.02960789e-01
-1.44105208e+00 6.05924964e-01 -1.34789407e-01 -6.71866596e-01
1.00072874e-02 -8.22740614e-01 -3.54701549e-01 1.14624488e+00
-1.25050259e+00 -8.45045626e-01 -3.62047791e-01 7.43913114e-01
2.01842606e-01 -5.98182380e-01 7.25316942e-01 3.73567611e-01
-4.05440181e-01 9.76969779e-01 -2.84226120e-01 2.83748865e-01
1.19441569e+00 -8.96818161e-01 -8.56984258e-02 6.96512878e-01
1.62812799e-01 1.01382613e+00 4.89882261e-01 -5.84042311e-01
-1.31991899e+00 -4.96328801e-01 5.87151527e-01 -6.15582407e-01
-3.24501544e-02 1.00952424e-01 -7.59728312e-01 1.70582011e-01
1.86747804e-01 -1.41873911e-01 9.86748040e-01 4.98243123e-01
-8.78639162e-01 -4.45792854e-01 -1.92149258e+00 3.58945549e-01
1.03720641e+00 -8.84636223e-01 -4.05349135e-01 -1.60769999e-01
-3.02502394e-01 4.70838189e-01 -1.08079517e+00 3.18205357e-01
8.26960921e-01 -1.42938054e+00 9.54478204e-01 -4.21015412e-01
3.89452875e-01 -2.27769300e-01 -2.12534681e-01 -1.21599829e+00
-4.25600708e-01 -1.19096942e-01 2.80617058e-01 1.54567063e+00
3.59242320e-01 -7.50029445e-01 9.54234719e-01 1.01501751e+00
4.88519847e-01 -1.28366530e-01 -8.21390808e-01 -7.23863661e-01
4.27072763e-01 -1.55901626e-01 8.41998637e-01 9.11303222e-01
-4.22809005e-01 1.14938095e-01 -9.15075913e-02 -6.82404637e-02
8.04542899e-01 -3.04929852e-01 8.98896158e-01 -1.45240438e+00
1.18674763e-01 -6.80615544e-01 -6.53207362e-01 -1.88962277e-02
2.95811236e-01 -5.52926898e-01 -1.08266644e-01 -1.00535858e+00
3.47782105e-01 -4.81371462e-01 -6.31290898e-02 2.23495200e-01
-2.31137246e-01 8.06946397e-01 1.61964387e-01 1.11011691e-01
9.94423777e-02 3.22688788e-01 9.78995264e-01 -1.67132705e-01
2.92635590e-01 -1.49750888e-01 -9.15006220e-01 4.21990454e-01
8.26755583e-01 -2.44050905e-01 -2.77547657e-01 -3.24148238e-01
1.77911431e-01 -2.73619890e-01 3.35949838e-01 -1.20739841e+00
-3.02968621e-01 -7.49418810e-02 8.30413580e-01 -4.94232439e-02
2.58090436e-01 -7.86072791e-01 2.11903736e-01 4.70044166e-01
-1.39844328e-01 1.65231273e-01 2.52184272e-01 6.44014701e-02
-2.68776089e-01 -3.64159018e-01 1.24158645e+00 -2.41529822e-01
-4.11105335e-01 4.00481194e-01 -5.20391405e-01 1.19873226e-01
5.16679049e-01 -5.77815294e-01 -3.63007933e-01 -5.22967637e-01
-3.84088337e-01 -1.80009365e-01 9.47028339e-01 3.24984550e-01
4.09502327e-01 -1.33983779e+00 -9.96202052e-01 4.79521364e-01
3.82958859e-01 -9.44009662e-01 2.03999028e-01 7.79102206e-01
-5.42840600e-01 3.61729898e-02 -6.62289262e-01 -3.55753005e-01
-1.66237116e+00 4.55262393e-01 5.38048089e-01 3.95352215e-01
2.11205304e-01 9.58954096e-01 1.66560665e-01 -4.49101776e-01
1.25342935e-01 3.89539301e-01 -1.14356361e-01 5.67566633e-01
1.00516808e+00 7.57949471e-01 2.16798838e-02 -1.27444470e+00
-3.96325469e-01 5.56459248e-01 -6.73555881e-02 -2.53904760e-01
9.74404156e-01 1.37038678e-01 -3.11777443e-01 9.24713016e-02
1.13420129e+00 1.63660645e-01 -8.95046592e-01 2.10881680e-02
-1.85064822e-01 -1.07345951e+00 3.29874456e-02 -8.70296299e-01
-1.45883214e+00 9.82618868e-01 1.08287942e+00 -1.31901175e-01
1.25535989e+00 -3.05387855e-01 1.19316941e-02 6.12983964e-02
6.03510976e-01 -1.09338355e+00 5.82572259e-02 1.89626962e-02
8.91626298e-01 -1.49125183e+00 2.23999351e-01 -2.83677459e-01
-5.22522986e-01 9.98999238e-01 6.84738159e-01 1.95419967e-01
6.17070556e-01 -5.92957102e-02 4.20472026e-01 -2.90883392e-01
-3.25219780e-01 -1.55505449e-01 2.77999312e-01 1.04949474e+00
6.39290094e-01 -1.08445786e-01 -2.96943575e-01 -1.02729440e-01
-2.48495072e-01 5.14846928e-02 3.55959475e-01 6.18150830e-01
-2.30095461e-01 -1.32228291e+00 -8.93337607e-01 9.27231610e-01
-6.18974924e-01 2.34158576e-01 -7.90519238e-01 5.42793512e-01
1.79515496e-01 1.26970863e+00 1.97627917e-01 -1.75317183e-01
2.74841875e-01 4.54403549e-01 8.21781337e-01 -4.44537938e-01
-9.96592999e-01 -3.89329433e-01 2.59823740e-01 -7.12185860e-01
-7.79964924e-01 -1.01685417e+00 -5.07243693e-01 -8.83741796e-01
-2.99142241e-01 1.06224477e-01 4.01537359e-01 4.87274170e-01
1.20585218e-01 -5.90872318e-02 5.46462953e-01 -5.47782481e-01
-3.03380132e-01 -9.61268187e-01 -8.16214383e-01 6.34403944e-01
1.64628670e-01 -8.38757873e-01 -4.74052191e-01 -5.32912686e-02] | [13.053277969360352, 1.1629164218902588] |
fcbbed36-7d90-4b2f-b881-208d5863a25e | face-shape-guided-deep-feature-alignment-for | 2209.07220 | null | https://arxiv.org/abs/2209.07220v1 | https://arxiv.org/pdf/2209.07220v1.pdf | Face Shape-Guided Deep Feature Alignment for Face Recognition Robust to Face Misalignment | For the past decades, face recognition (FR) has been actively studied in computer vision and pattern recognition society. Recently, due to the advances in deep learning, the FR technology shows high performance for most of the benchmark datasets. However, when the FR algorithm is applied to a real-world scenario, the performance has been known to be still unsatisfactory. This is mainly attributed to the mismatch between training and testing sets. Among such mismatches, face misalignment between training and testing faces is one of the factors that hinder successful FR. To address this limitation, we propose a face shape-guided deep feature alignment framework for FR robust to the face misalignment. Based on a face shape prior (e.g., face keypoints), we train the proposed deep network by introducing alignment processes, i.e., pixel and feature alignments, between well-aligned and misaligned face images. Through the pixel alignment process that decodes the aggregated feature extracted from a face image and face shape prior, we add the auxiliary task to reconstruct the well-aligned face image. Since the aggregated features are linked to the face feature extraction network as a guide via the feature alignment process, we train the robust face feature to the face misalignment. Even if the face shape estimation is required in the training stage, the additional face alignment process, which is usually incorporated in the conventional FR pipeline, is not necessarily needed in the testing phase. Through the comparative experiments, we validate the effectiveness of the proposed method for the face misalignment with the FR datasets. | ['Yong Man Ro', 'Kimin Yun', 'Hyung-Il Kim'] | 2022-09-15 | null | null | null | null | ['face-alignment'] | ['computer-vision'] | [ 3.49044025e-01 -1.29345939e-01 3.28970760e-01 -6.63175821e-01
-4.50093180e-01 -3.51186305e-01 4.42726284e-01 -6.06228948e-01
-1.11965612e-02 3.88771594e-01 -1.97913170e-01 1.50940254e-01
-4.25252728e-02 -6.51789486e-01 -8.67349923e-01 -8.53851140e-01
4.04879749e-01 2.23224252e-01 -2.44622752e-01 1.05914529e-02
2.38259986e-01 8.89819622e-01 -1.83100975e+00 7.33702779e-02
7.30486572e-01 1.40414703e+00 1.38056472e-01 1.10324875e-01
-2.46372893e-01 1.09954074e-01 -4.65114266e-01 -4.15158570e-01
4.70499188e-01 -3.95347744e-01 -4.85314041e-01 4.13176209e-01
6.58119559e-01 -5.68093181e-01 -1.81861043e-01 1.13613009e+00
5.08883357e-01 -2.15714708e-01 3.80855203e-01 -1.29086614e+00
-3.23214293e-01 1.16724726e-02 -9.20481503e-01 -5.78022860e-02
4.39566880e-01 1.61545768e-01 5.34241974e-01 -1.21290278e+00
3.74757916e-01 1.43900406e+00 5.64800560e-01 5.08820057e-01
-8.97932768e-01 -9.32015300e-01 7.05156475e-02 1.76257372e-01
-1.50751102e+00 -1.00308776e+00 1.05278945e+00 -3.15829754e-01
5.82973778e-01 -1.28771355e-02 6.01137936e-01 7.65182257e-01
6.43903166e-02 5.25896311e-01 9.06872571e-01 -2.86228448e-01
-1.88805535e-01 -1.00829072e-01 -1.60516635e-01 6.96721613e-01
8.79696533e-02 2.49503300e-01 -1.63935527e-01 2.16965809e-01
9.34048176e-01 2.05979750e-01 -5.04451334e-01 -2.04547182e-01
-7.19190180e-01 4.82255906e-01 5.50289214e-01 2.70111740e-01
-3.44232202e-01 -1.76768377e-01 5.30578382e-02 1.11557119e-01
3.64198357e-01 -6.45437986e-02 -3.88860643e-01 2.46783197e-01
-1.16794884e+00 -7.84329623e-02 4.75325406e-01 8.06913435e-01
9.60253119e-01 2.60928303e-01 -8.83204639e-02 8.26586545e-01
7.32238591e-01 6.76516116e-01 2.03281045e-01 -4.82532144e-01
4.52064872e-01 9.00178373e-01 2.74053086e-02 -1.32673013e+00
-3.55488032e-01 -5.39722741e-01 -9.24334705e-01 9.33792964e-02
4.86084610e-01 -8.59183911e-03 -9.82059300e-01 1.74166727e+00
6.17691815e-01 5.28049171e-01 1.76024526e-01 9.47606802e-01
9.84100640e-01 2.99133599e-01 -3.16819817e-01 -3.93679500e-01
1.40925801e+00 -7.22974300e-01 -7.42436349e-01 -2.08317116e-01
2.04979926e-01 -1.19455659e+00 8.51137221e-01 1.52080759e-01
-8.54057252e-01 -8.90241683e-01 -1.09428120e+00 5.24656475e-02
3.89357582e-02 6.52130008e-01 2.51111269e-01 6.86077952e-01
-8.87280643e-01 2.89059401e-01 -6.99722707e-01 -3.91078264e-01
6.95254624e-01 6.19506419e-01 -7.91524053e-01 -3.45860541e-01
-8.30820084e-01 6.25820637e-01 1.45282924e-01 9.76123631e-01
-8.98392022e-01 -3.56154203e-01 -8.52888763e-01 1.79285735e-01
2.95338184e-01 -4.29042399e-01 9.22795475e-01 -1.36849320e+00
-1.52358460e+00 6.97540760e-01 -2.53458023e-01 6.72066808e-02
5.15690327e-01 -2.84207851e-01 -6.15784645e-01 6.13465533e-03
-1.08937219e-01 5.01716435e-01 1.20888352e+00 -1.16509581e+00
-4.89377409e-01 -7.05018640e-01 -1.16503797e-01 1.63585350e-01
-1.75606087e-01 1.95278794e-01 -5.90674400e-01 -3.77243817e-01
2.67267674e-01 -7.60460377e-01 3.07502955e-01 8.11246783e-03
-2.81659961e-01 -9.76514742e-02 1.08948994e+00 -7.44474947e-01
8.49375188e-01 -2.32838774e+00 1.36593962e-03 3.67051065e-01
-2.38854975e-01 3.23511839e-01 -1.98676750e-01 1.47576528e-02
-2.83862054e-01 -1.54546529e-01 -2.27178738e-01 -3.09855431e-01
-2.77956158e-01 5.43602780e-02 -1.49311453e-01 7.01296210e-01
7.43212640e-01 7.42571831e-01 -4.30940360e-01 -3.75420302e-01
1.26898214e-01 6.61682844e-01 -5.37804604e-01 5.46038628e-01
-1.02785736e-01 8.44828427e-01 -4.21433568e-01 7.42017627e-01
1.23926878e+00 -1.97795928e-02 1.22872680e-01 -7.36973941e-01
8.88904463e-03 -2.84710735e-01 -1.42604721e+00 1.56116712e+00
-2.57214904e-01 1.85631052e-01 2.87629098e-01 -9.90454078e-01
1.17776465e+00 5.39198816e-01 4.59156513e-01 -6.14410520e-01
3.47405881e-01 3.69360596e-01 2.31334612e-01 -6.60465896e-01
-2.42299382e-02 -4.63241860e-02 5.10624230e-01 3.91575396e-01
1.58275008e-01 1.10427953e-01 -4.56282161e-02 -4.12890881e-01
6.68042183e-01 4.97574747e-01 1.39407530e-01 1.32273799e-02
1.14867747e+00 -4.56772149e-01 6.45826757e-01 -5.48321307e-02
-2.54996642e-02 7.75643289e-01 2.16481999e-01 -6.83090091e-01
-8.17362249e-01 -7.25783944e-01 -3.08599144e-01 5.26486993e-01
5.98819964e-02 -1.22122735e-01 -1.00391197e+00 -9.05934393e-01
-1.57564506e-01 -9.04511511e-02 -3.69059354e-01 -1.63688704e-01
-7.41856575e-01 -6.75240457e-01 3.54994953e-01 4.40880030e-01
9.30741251e-01 -1.20741498e+00 -3.53993803e-01 2.78564878e-02
5.17764501e-02 -1.20326447e+00 -5.21432281e-01 -3.50547880e-01
-5.78263104e-01 -1.34986532e+00 -5.98415971e-01 -9.28006053e-01
1.17199767e+00 2.80430049e-01 6.11455977e-01 6.95513129e-01
-2.20792174e-01 1.02757014e-01 -1.29912496e-01 -1.46112684e-02
-6.97651878e-02 -1.80976823e-01 7.86023512e-02 5.90695262e-01
3.01018745e-01 -4.10385638e-01 -8.35587025e-01 5.39841354e-01
-9.34577227e-01 -3.30491513e-02 8.77303600e-01 9.19289351e-01
4.73387331e-01 1.88046783e-01 7.28006721e-01 -6.23050749e-01
1.44943133e-01 -2.43549123e-01 -7.73738444e-01 2.96553046e-01
-2.82367915e-01 -1.32898301e-01 6.95424795e-01 -2.73478270e-01
-1.13747752e+00 4.08396065e-01 -5.26797473e-01 -4.17151183e-01
-3.21540743e-01 4.12636995e-01 -8.34842324e-01 -3.26091588e-01
2.67145842e-01 3.10284555e-01 2.61379004e-01 -4.22488242e-01
9.31911319e-02 7.79815972e-01 5.06920576e-01 -5.29151976e-01
1.07907951e+00 3.52107197e-01 1.20340018e-02 -6.47492766e-01
-6.16641223e-01 -1.91536620e-02 -8.01604211e-01 -3.13985825e-01
7.28465140e-01 -8.30261827e-01 -7.67900407e-01 6.88561261e-01
-1.38003266e+00 2.85743505e-01 2.66836375e-01 3.81913155e-01
-3.57769579e-01 4.44254667e-01 -1.91577196e-01 -7.47169375e-01
-4.81695563e-01 -1.56901658e+00 1.15628874e+00 5.47362030e-01
3.31006169e-01 -5.87459028e-01 -4.08868760e-01 2.80049920e-01
2.37721860e-01 1.72549561e-01 7.52394140e-01 -5.79336882e-01
-6.88767910e-01 -2.48969808e-01 -3.98161620e-01 5.68802059e-01
5.61373889e-01 2.22215071e-01 -1.25986099e+00 -5.36304712e-01
2.76824027e-01 -2.02036723e-01 3.94768894e-01 1.39288723e-01
1.03661418e+00 -2.29652032e-01 -2.50204563e-01 7.24259496e-01
1.31828141e+00 3.15853894e-01 7.68692195e-01 6.76814392e-02
7.41848826e-01 6.71296239e-01 7.45869458e-01 2.89693266e-01
1.76056415e-01 7.02107787e-01 4.61330146e-01 -1.43156290e-01
-1.18401781e-01 -1.72494397e-01 3.79448563e-01 6.23099208e-01
4.58585806e-02 7.79236332e-02 -7.06193805e-01 2.66589597e-02
-1.49141192e+00 -7.37548590e-01 1.20708637e-01 2.26698661e+00
5.21743655e-01 -4.11042795e-02 -2.22909227e-01 2.74847955e-01
9.37877953e-01 -4.58906591e-02 -6.06291175e-01 1.96170546e-02
-3.98084186e-02 1.88886881e-01 -1.74238667e-01 2.86750197e-01
-9.69033837e-01 8.69866669e-01 5.11419392e+00 6.27438843e-01
-1.56164956e+00 -1.77516401e-01 7.49019206e-01 2.61364996e-01
-1.18425667e-01 -7.46561289e-02 -9.85808253e-01 4.25786078e-01
3.75376463e-01 2.84791112e-01 5.58923960e-01 6.50215983e-01
8.93227607e-02 9.65734497e-02 -1.28769195e+00 1.27856874e+00
2.90994793e-01 -8.80400836e-01 -2.61750165e-02 1.77659560e-02
5.80704153e-01 -4.72269356e-01 1.08189456e-01 1.81804284e-01
-4.68292147e-01 -1.39204156e+00 5.33808947e-01 5.27020872e-01
9.34590995e-01 -8.85728717e-01 9.95629609e-01 2.40374953e-01
-1.41775262e+00 -1.46280006e-01 -3.93550932e-01 1.35823548e-01
-1.26598515e-02 6.36001110e-01 -8.32272530e-01 8.16906393e-01
6.37084842e-01 6.67781055e-01 -4.94222343e-01 8.16558778e-01
-2.17032507e-01 1.56461701e-01 -3.37271899e-01 6.73351824e-01
-1.71790361e-01 -3.61995906e-01 2.22411051e-01 5.57616472e-01
5.68653107e-01 -1.95083722e-01 8.86249691e-02 8.73565376e-01
-2.22408667e-01 1.80183530e-01 -5.62474906e-01 -6.81188926e-02
5.37442207e-01 1.72528017e+00 -6.24787271e-01 1.14658490e-01
-5.53463459e-01 8.85427535e-01 2.72650748e-01 3.27870190e-01
-7.98175514e-01 -4.63232473e-02 5.38990617e-01 1.28185019e-01
4.04002219e-01 9.58439521e-03 7.66379088e-02 -8.95862520e-01
4.35363263e-01 -1.04830050e+00 5.79126291e-02 -5.50181568e-01
-1.26091874e+00 9.94654477e-01 -3.05986315e-01 -1.24214971e+00
-1.54894799e-01 -4.58261341e-01 -7.91813016e-01 1.10973132e+00
-1.78112006e+00 -1.33542323e+00 -4.85211492e-01 7.16580093e-01
2.69944727e-01 -3.35778683e-01 6.53766334e-01 6.50815010e-01
-9.88898516e-01 8.81851077e-01 -2.67152846e-01 2.03353763e-01
7.38851368e-01 -5.53803623e-01 2.45789066e-01 8.55675459e-01
-2.96965688e-02 7.17879474e-01 4.05998349e-01 -6.27009213e-01
-1.57809424e+00 -1.10165799e+00 4.94356215e-01 -4.18588445e-02
1.29197106e-01 -3.57476562e-01 -9.51070666e-01 6.98264658e-01
3.79587188e-02 5.48880696e-01 4.17309970e-01 -3.06113660e-01
-2.01955974e-01 -5.77077091e-01 -1.44629943e+00 2.59355068e-01
1.03768981e+00 -4.03052479e-01 -4.90705222e-01 6.15432486e-02
2.51175523e-01 -4.34866816e-01 -9.13828671e-01 8.17385375e-01
7.53824532e-01 -8.15765381e-01 7.88463295e-01 -2.85235018e-01
1.88461721e-01 -7.09456503e-01 -1.71117812e-01 -1.11024272e+00
-4.71117869e-02 -3.37522537e-01 1.97225660e-01 1.78406632e+00
1.56570766e-02 -7.07589746e-01 7.45595157e-01 3.70003134e-01
-3.09481658e-02 -7.88598299e-01 -1.11254525e+00 -3.99567157e-01
-3.52154136e-01 6.22658022e-02 1.02529001e+00 6.84875965e-01
-7.34279633e-01 2.57815540e-01 -2.46169403e-01 4.02754545e-01
3.57824177e-01 3.63650858e-01 9.74852800e-01 -1.36429262e+00
-9.44366902e-02 -2.05250502e-01 -4.08535957e-01 -9.94580090e-01
4.38319832e-01 -8.05345356e-01 1.42139107e-01 -1.23840189e+00
8.08592364e-02 -5.08256197e-01 -1.27596170e-01 4.72955436e-01
-2.89565802e-01 4.29371893e-01 1.04152471e-01 1.71342418e-01
-7.31859952e-02 7.66124606e-01 1.39168251e+00 6.15782142e-02
1.66073460e-02 -7.95185491e-02 -5.46564341e-01 7.40458012e-01
6.37953579e-01 -2.18151957e-01 -3.42066467e-01 -4.95239496e-01
-3.16150580e-03 1.22906929e-02 2.47532308e-01 -1.03292334e+00
1.98164612e-01 1.06578901e-01 9.46639121e-01 -6.11900747e-01
3.14484984e-01 -1.32765591e+00 3.84309232e-01 3.45684618e-01
2.93504447e-01 1.95198879e-01 1.59086958e-01 2.63253838e-01
-3.69826078e-01 -1.14186049e-01 8.69905531e-01 1.01776615e-01
-4.32248861e-01 6.69704974e-01 2.95897305e-01 -2.99309224e-01
9.43688154e-01 -4.58037943e-01 -2.77765125e-01 1.18699996e-03
-2.81573594e-01 1.71276048e-01 4.62185264e-01 4.01064962e-01
8.36420357e-01 -1.54316664e+00 -7.48711109e-01 9.91930485e-01
-7.69336224e-02 3.13927174e-01 5.18904686e-01 9.73284304e-01
-3.30570161e-01 1.83199957e-01 -3.45710665e-01 -7.93255329e-01
-1.35750055e+00 5.32428443e-01 4.60291207e-01 1.17229857e-01
-3.45119894e-01 6.14066243e-01 3.40171725e-01 -4.03483659e-01
1.98997572e-01 -5.52665666e-02 -4.21400994e-01 -5.65890148e-02
6.38458908e-01 -5.75911999e-02 2.92772084e-01 -1.06495583e+00
-4.24864233e-01 1.12091208e+00 -1.24174401e-01 3.92976046e-01
1.34378779e+00 -1.43686473e-01 -2.96395242e-01 -7.81238303e-02
1.22279382e+00 1.30639430e-02 -1.33637536e+00 -1.79665104e-01
-1.11070223e-01 -6.44174814e-01 -1.21131897e-01 -5.33201694e-01
-1.67731380e+00 9.75762963e-01 8.85937691e-01 -2.35766351e-01
1.35867846e+00 -2.52781540e-01 6.72270715e-01 4.60857190e-02
2.56250054e-01 -6.92944527e-01 4.49242815e-02 3.43076855e-01
1.06602013e+00 -1.33361912e+00 -8.78631324e-02 -5.48011839e-01
-1.94117323e-01 1.38976777e+00 9.25617754e-01 -6.57051504e-02
7.15134561e-01 2.09250227e-01 1.13122709e-01 -1.31886065e-01
-2.90563852e-01 4.03226726e-02 4.35775757e-01 3.27910781e-01
4.70465094e-01 -3.81085932e-01 -9.88996178e-02 6.57083511e-01
-1.63692653e-01 1.25865340e-01 -1.03608385e-01 6.65500581e-01
-3.06674182e-01 -1.21326077e+00 -6.00460768e-01 3.18813115e-01
-4.14808959e-01 1.18323840e-01 -1.15737297e-01 6.84002757e-01
3.47283334e-01 9.80570138e-01 1.11595400e-01 -3.25360835e-01
3.48354667e-01 5.09144813e-02 7.72170305e-01 -5.67491353e-01
-5.33721626e-01 2.17989922e-01 -3.24082851e-01 -4.32110786e-01
-5.50622582e-01 -4.14883316e-01 -1.33942986e+00 -2.24967137e-01
-5.25953114e-01 -8.66912007e-02 5.51238596e-01 1.26372588e+00
3.98231447e-01 2.91640848e-01 9.95710731e-01 -9.29970205e-01
-4.99251276e-01 -9.78529394e-01 -3.79254460e-01 4.59200621e-01
2.67808408e-01 -9.78849769e-01 -3.44052613e-01 -1.92256957e-01] | [13.227621078491211, 0.43013837933540344] |
b2e32d41-79cb-4c66-a7b4-f3d88bbe5567 | data-driven-prediction-of-battery-cycle-life | 2110.09687 | null | https://arxiv.org/abs/2110.09687v1 | https://arxiv.org/pdf/2110.09687v1.pdf | Data Driven Prediction of Battery Cycle Life Before Capacity Degradation | Ubiquitous use of lithium-ion batteries across multiple industries presents an opportunity to explore cost saving initiatives as the price to performance ratio continually decreases in a competitive environment. Manufacturers using lithium-ion batteries ranging in applications from mobile phones to electric vehicles need to know how long batteries will last for a given service life. To understand this, expensive testing is required. This paper utilizes the data and methods implemented by Kristen A. Severson, et al, to explore the methodologies that the research team used and presents another method to compare predicted results vs. actual test data for battery capacity fade. The fundamental effort is to find out if machine learning techniques may be trained to use early life cycle data in order to accurately predict battery capacity over the battery life cycle. Results show comparison of methods between Gaussian Process Regression (GPR) and Elastic Net Regression (ENR) and highlight key data features used from the extensive dataset found in the work of Severson, et al. | ['Kurt I. Kuhn', 'Jamie Peck', 'Caitlin Feltner', 'Anmol Singh'] | 2021-10-19 | null | null | null | null | ['gpr', 'gpr'] | ['computer-vision', 'miscellaneous'] | [-3.69004793e-02 -4.78896946e-01 -4.15840536e-01 -4.21971947e-01
-4.64787155e-01 -3.45802039e-01 3.77506137e-01 5.26552380e-04
-2.92532057e-01 1.01486003e+00 -1.61754951e-01 -9.97625589e-01
-4.12810177e-01 -7.50603497e-01 -4.66018915e-01 -6.45843327e-01
1.72664523e-01 9.52639639e-01 5.50908335e-05 5.30984513e-02
5.17502010e-01 7.36403942e-01 -1.59992278e+00 -1.31199256e-01
6.29555941e-01 1.08076489e+00 4.66892123e-01 5.93017399e-01
-2.90555321e-03 4.53813791e-01 -1.76326737e-01 -1.31389080e-02
6.71405206e-03 1.86560750e-01 -1.99697867e-01 -4.03198451e-01
-5.73809743e-01 -1.15803726e-01 -2.07172826e-01 5.04700005e-01
5.90710163e-01 7.86067098e-02 9.82024193e-01 -1.68376172e+00
-1.85450092e-01 3.23385477e-01 -1.80764318e-01 5.28078318e-01
3.46332163e-01 1.15133934e-01 4.65176195e-01 -8.40539932e-01
7.01005280e-04 1.01316214e+00 8.38627517e-01 3.17275584e-01
-1.09262240e+00 -1.06733358e+00 -3.90867263e-01 3.43068957e-01
-1.27112257e+00 -3.42652082e-01 5.90497613e-01 -3.63406211e-01
1.45623791e+00 4.39455390e-01 7.86565781e-01 8.88679445e-01
7.06866086e-01 4.04361486e-01 1.48691523e+00 -3.87122780e-01
2.64788538e-01 4.29882079e-01 1.63830683e-01 -1.67206228e-01
7.09342718e-01 1.43608868e-01 -2.19470695e-01 -2.70860344e-01
2.81026185e-01 1.06260307e-01 1.27814397e-01 -1.09365903e-01
-5.22395730e-01 5.32656372e-01 -4.80271339e-01 5.22814453e-01
-3.51861835e-01 3.43474865e-01 3.44092220e-01 2.17509672e-01
4.20708388e-01 -1.75493166e-01 -8.12543631e-01 -9.22439337e-01
-8.79093587e-01 2.28819445e-01 1.25651348e+00 1.05956745e+00
4.45901245e-01 2.26832211e-01 2.28167534e-01 6.93005860e-01
9.31473017e-01 4.08084720e-01 6.66304708e-01 -5.56172550e-01
1.61090925e-01 2.85751224e-01 2.44148806e-01 -2.88992792e-01
-5.05619109e-01 -4.98925596e-02 -3.14796418e-01 -4.77001667e-02
1.16935400e-02 -6.54201582e-02 -7.37873554e-01 5.66940427e-01
-1.19200476e-01 5.65497428e-02 2.72903144e-02 1.45115986e-01
4.58672971e-01 7.57665575e-01 3.51463884e-01 -3.40205103e-01
9.87228692e-01 -3.58712345e-01 -6.32458746e-01 -1.46502689e-01
5.16593993e-01 -5.30529618e-01 8.47357750e-01 5.88799536e-01
-1.23681939e+00 -3.62319708e-01 -1.24249947e+00 3.62036377e-01
-5.21551132e-01 -3.22520286e-01 6.08894885e-01 1.27010882e+00
-1.09533584e+00 8.40578854e-01 -1.03752267e+00 -4.52753156e-01
3.78620654e-01 9.24903631e-01 2.10182387e-02 -2.39285659e-02
-6.83248580e-01 1.20836246e+00 2.80202389e-01 2.45635509e-02
-7.24112809e-01 -7.29952812e-01 -4.59033102e-01 -1.65804580e-01
-8.39299932e-02 -1.46164641e-01 1.30015099e+00 -4.25800323e-01
-1.03331852e+00 4.04704660e-01 -2.15787277e-01 -3.47057492e-01
5.70337713e-01 1.99483931e-02 -6.68877840e-01 -4.70951945e-01
-9.95540991e-02 1.28471807e-01 2.16672063e-01 -1.36493301e+00
-9.59875584e-01 -2.35782936e-01 -7.16838837e-01 -1.76850244e-01
-2.32076928e-01 1.06307745e-01 -1.83523893e-01 -2.32533768e-01
-4.63302247e-02 -1.27349007e+00 -2.47717965e-02 -8.58698845e-01
6.77916035e-02 -5.49065650e-01 1.10105288e+00 -1.08782053e+00
1.18827152e+00 -1.68411696e+00 -5.17550409e-01 5.52955687e-01
-5.12019873e-01 -2.93092161e-01 6.70205295e-01 5.14968872e-01
-1.84451416e-01 5.23574352e-01 -2.02895869e-02 -3.68490338e-01
9.28568617e-02 5.93241692e-01 4.36421335e-02 6.16077721e-01
1.00770788e-02 8.08320522e-01 -7.99893677e-01 -1.62693650e-01
-4.43412811e-02 4.09570098e-01 1.76142573e-01 -3.87372039e-02
1.88538119e-01 4.30302262e-01 -2.75896579e-01 1.24777877e+00
4.68375772e-01 3.12951833e-01 5.89409955e-02 -4.97481674e-02
-3.34055424e-01 2.64220923e-01 -8.02818239e-01 8.49901021e-01
-5.79045773e-01 6.23363078e-01 -3.84942144e-01 -9.85890210e-01
1.13998616e+00 4.07555908e-01 8.03842127e-01 -9.81508970e-01
2.93907762e-01 7.69079864e-01 -1.62852302e-01 -4.85809237e-01
3.74368042e-01 -6.68101966e-01 1.27781928e-01 1.78702757e-01
-4.38553184e-01 -3.04331571e-01 1.14754535e-01 -4.84609395e-01
1.03858149e+00 4.02521104e-01 -2.56066591e-01 -5.65347672e-01
2.52847046e-01 -2.19439510e-02 5.96683621e-01 3.38530868e-01
-1.41772047e-01 2.41583988e-01 2.41944179e-01 -4.02024053e-02
-1.44528997e+00 -8.18785191e-01 -3.49682361e-01 6.55203879e-01
-4.08037193e-02 1.58636913e-01 -1.07711308e-01 -1.90547064e-01
4.47169036e-01 1.19091570e+00 -2.17174053e-01 -3.88162047e-01
-1.68062091e-01 -9.53576505e-01 2.29569510e-01 9.45843279e-01
4.44095172e-02 -7.94719458e-01 -5.83237410e-01 4.14982557e-01
4.03349549e-01 -6.25558138e-01 2.02324554e-01 6.78198159e-01
-1.28273988e+00 -6.93582237e-01 -3.58925670e-01 -4.40600485e-01
2.75776297e-01 -2.59144694e-01 1.32440948e+00 -3.23702581e-03
-3.59330326e-01 7.49464035e-01 -6.36595264e-02 -1.08663559e+00
-5.06857216e-01 -1.60610542e-01 4.18888897e-01 -7.85268605e-01
8.63979161e-01 -6.23948038e-01 -5.97584784e-01 7.37268448e-01
-6.00980699e-01 -5.44490814e-01 7.22449541e-01 4.76784617e-01
7.33036995e-01 7.00107753e-01 1.23232162e+00 -4.80022669e-01
8.12679589e-01 -1.03244483e+00 -4.62781757e-01 5.36939390e-02
-1.60901737e+00 -3.61900181e-01 -1.46439418e-01 -4.41443771e-01
-8.12540829e-01 -2.15538025e-01 -2.79136091e-01 -2.55927533e-01
6.78905919e-02 3.50874990e-01 -2.96329141e-01 5.24312854e-02
-1.93723999e-02 7.49427378e-02 2.17560694e-01 -5.07444143e-01
-4.35530514e-01 1.09217799e+00 2.89677143e-01 -4.42577928e-01
5.06075442e-01 3.67271334e-01 2.25879982e-01 -5.16916692e-01
1.21775746e-01 -7.97364056e-01 -4.17822868e-01 -4.90136385e-01
4.40374523e-01 -1.07353020e+00 -9.24635231e-01 2.56996602e-01
-5.96243083e-01 -6.01670563e-01 -2.48273566e-01 3.91477823e-01
-8.72449875e-01 -6.13208562e-02 -1.53203681e-01 -1.43148184e+00
-3.08848858e-01 -9.16142941e-01 5.75020790e-01 2.85155147e-01
-4.09618378e-01 -1.22051513e+00 -2.02006102e-01 4.04735714e-01
6.08916581e-01 9.70902070e-02 8.50993216e-01 -8.62645924e-01
-2.45803416e-01 -7.44410813e-01 5.81156574e-02 5.96523583e-01
3.29773575e-01 -2.81974929e-03 -9.03176129e-01 -6.50044978e-01
2.65824348e-01 3.13094497e-01 9.64275226e-02 3.48204374e-01
1.23309839e+00 -1.70733985e-02 -6.26356184e-01 -1.56377718e-01
1.65849817e+00 9.77566421e-01 1.07616389e+00 7.84771681e-01
4.10106391e-01 4.91086483e-01 9.87239540e-01 1.64152443e-01
4.46950108e-01 1.91381216e-01 3.42264742e-01 3.34253699e-01
5.96302092e-01 -2.06438154e-02 4.24840003e-01 5.33245623e-01
-1.92861915e-01 -2.41170868e-01 -1.02149403e+00 6.50773644e-01
-1.67039716e+00 -6.86046779e-01 -3.64404202e-01 2.27068639e+00
4.63533789e-01 8.10581625e-01 4.52827692e-01 6.83408260e-01
3.19559306e-01 -8.64929676e-01 -6.57318771e-01 -7.25717843e-01
1.06083192e-01 -6.25179410e-02 1.30856276e+00 -1.10290209e-02
-3.61032993e-01 1.82442024e-01 7.04749441e+00 6.87681437e-01
-8.15222263e-01 2.93029934e-01 1.04534042e+00 -7.71892443e-02
-5.38384259e-01 3.29703003e-01 -8.75063419e-01 8.66390646e-01
1.77870023e+00 -3.61749500e-01 1.54449746e-01 6.47264898e-01
8.09907436e-01 -7.01803029e-01 -1.14457762e+00 9.58558381e-01
-6.99577555e-02 -7.99358845e-01 -4.46552753e-01 3.62029880e-01
4.87659574e-01 -2.70062461e-02 -6.24582656e-02 3.55781823e-01
-1.42581299e-01 -1.26863766e+00 7.46284187e-01 1.31345594e+00
5.86469233e-01 -1.04691660e+00 9.74877596e-01 2.76152253e-01
-9.15335655e-01 -6.69356823e-01 -2.28596851e-01 -1.19891889e-01
2.75269359e-01 5.81421077e-01 -1.20161307e+00 3.65178138e-01
1.18811357e+00 4.03570652e-01 -3.56079102e-01 1.28686237e+00
6.15246117e-01 1.00616884e+00 -5.03304780e-01 -3.84649426e-01
-2.06481457e-01 -5.34661651e-01 2.05830112e-01 1.18820798e+00
8.38685811e-01 -3.37647587e-01 -3.18125039e-01 5.10820508e-01
3.31804752e-01 -1.03374369e-01 -6.58856690e-01 -3.30006480e-01
4.50872660e-01 1.08522856e+00 -9.76574481e-01 7.67529681e-02
-6.95300579e-01 4.94171470e-01 -8.53874028e-01 9.15598199e-02
-5.32458127e-01 -4.49622385e-02 4.39849466e-01 9.31494951e-01
2.52265465e-02 -3.69316190e-01 -7.29755163e-01 1.12552524e-01
-1.14138752e-01 -3.52337152e-01 -7.10338205e-02 -9.75283802e-01
-1.41484845e+00 -1.75525919e-01 3.75157505e-01 -9.20928478e-01
-1.18182115e-01 -8.22786629e-01 -9.64426756e-01 1.19945848e+00
-1.37055337e+00 -1.13653982e+00 -1.28802404e-01 -1.30619943e-01
7.52410829e-01 -2.14282274e-01 2.13680908e-01 6.16326809e-01
-3.50575119e-01 1.76210284e-01 4.61927891e-01 -7.47917652e-01
3.69590908e-01 -1.41988850e+00 3.04427296e-01 1.00328095e-01
-5.80978215e-01 2.91035682e-01 1.07526457e+00 -1.20102561e+00
-1.87377179e+00 -6.81985319e-01 6.45980060e-01 -7.60682881e-01
8.52158546e-01 -8.65238607e-02 -9.36943054e-01 5.17267346e-01
3.25918831e-02 -6.97803676e-01 8.01850796e-01 -3.14187467e-01
8.05184543e-01 -4.42733675e-01 -1.26376343e+00 8.99324790e-02
5.65512478e-01 -4.24553782e-01 -1.69276789e-01 7.62256533e-02
5.98632731e-02 4.10053581e-02 -1.27664661e+00 7.26776302e-01
9.60982144e-01 -7.24403679e-01 7.12869465e-01 -2.17617035e-01
-2.29253426e-01 -7.30040893e-02 -1.99188307e-01 -8.73281419e-01
4.99873050e-02 -5.56159437e-01 -1.58887878e-01 1.69706666e+00
6.82794273e-01 -8.29385996e-01 9.32716370e-01 1.25310481e+00
-2.73564428e-01 -1.22533762e+00 -1.22723782e+00 -9.82925713e-01
2.98979700e-01 -9.18952227e-01 7.78402030e-01 1.94614902e-01
-3.33970457e-01 -1.47801891e-01 -4.94481921e-02 -1.79494217e-01
3.26155633e-01 -7.55654156e-01 4.87304360e-01 -1.68384826e+00
1.76086918e-01 -2.65490472e-01 -5.94323814e-01 -3.85555655e-01
2.39718035e-01 -4.76930887e-01 7.94236436e-02 -1.93735826e+00
3.52094978e-01 -1.25077915e+00 -5.28511286e-01 1.96424201e-01
2.29299664e-01 2.16106966e-01 -9.40643027e-02 3.23874801e-01
1.07611507e-01 2.00901210e-01 5.26797175e-01 -1.26317337e-01
-1.62167370e-01 6.99901998e-01 -7.44058311e-01 4.54842329e-01
1.01080513e+00 -4.68585402e-01 -5.65776885e-01 3.85643870e-01
7.74280727e-01 -1.81269109e-01 3.49852920e-01 -1.08054900e+00
2.14422628e-01 -9.54394713e-02 6.20162308e-01 -1.00439811e+00
2.13686153e-01 -1.36876774e+00 1.05449057e+00 3.60067725e-01
6.86291516e-01 6.93235576e-01 5.00514269e-01 8.55785072e-01
3.74927521e-01 -6.93630517e-01 3.10612857e-01 1.80338725e-01
-5.93080759e-01 5.61945774e-02 -6.28776968e-01 -9.41168725e-01
1.32164526e+00 -9.96218383e-01 2.25765467e-01 -4.97105807e-01
-7.78426349e-01 4.45020646e-01 7.32477486e-01 5.50527632e-01
4.07816201e-01 -1.27521408e+00 -1.39665127e-01 -1.54161021e-01
-2.21155882e-01 -1.03525966e-01 -1.45252898e-01 1.04533613e+00
-3.92256975e-01 3.88624430e-01 6.33558556e-02 -5.73061168e-01
-1.16282070e+00 3.49625021e-01 4.13855404e-01 2.37826463e-02
-3.44116449e-01 4.67681497e-01 -9.10592854e-01 1.17740177e-01
-2.77091712e-02 -2.46697530e-01 -1.14523232e-01 2.67267615e-01
-3.52444872e-02 1.02764332e+00 5.68675816e-01 -5.85078537e-01
-2.47249320e-01 2.87587434e-01 4.20466900e-01 -1.37193352e-01
1.43563592e+00 -4.14781332e-01 -1.63706347e-01 1.39680767e+00
1.09647357e+00 -2.71777868e-01 -1.27273095e+00 7.90438235e-01
7.52002060e-01 -2.53702849e-01 2.01847807e-01 -6.98823631e-01
-6.38393462e-01 3.63977045e-01 1.39658546e+00 4.54500526e-01
9.74437177e-01 -1.29944542e-02 5.32078207e-01 7.28158504e-02
3.62766653e-01 -1.67938387e+00 -2.42811561e-01 2.99486145e-03
6.63940191e-01 -1.12789738e+00 2.46693686e-01 -8.19185674e-02
-3.96972626e-01 1.04808033e+00 2.18739316e-01 2.34613702e-01
1.29503560e+00 5.29157221e-01 -4.55609322e-01 -6.12570457e-02
-4.55217212e-01 2.78546542e-01 -2.00465441e-01 6.61723316e-01
3.58408839e-01 1.47405788e-01 -5.42875707e-01 7.72030592e-01
-1.70636922e-03 2.54418403e-01 4.15414840e-01 1.48150122e+00
-3.76957893e-01 -1.36560380e+00 -4.93279606e-01 1.34875739e+00
-6.69073880e-01 2.49231458e-01 1.31860852e-01 6.94296598e-01
9.76296365e-02 1.26809394e+00 3.88562083e-01 -1.74889266e-01
5.89183807e-01 4.24308211e-01 2.69560546e-01 -3.08531970e-01
-1.83308288e-01 -4.27301973e-03 3.42682838e-01 3.44359465e-02
-4.60937351e-01 -1.05331683e+00 -1.29970014e+00 -3.22517097e-01
-6.31920278e-01 1.40286803e-01 1.62557280e+00 9.65413511e-01
-8.31743702e-02 7.31817484e-01 6.13188684e-01 -1.20875120e+00
-3.44074726e-01 -1.16502607e+00 -9.05626357e-01 -1.50396243e-01
-1.37784660e-01 -9.96526241e-01 -4.72293139e-01 -6.42728284e-02] | [6.354377269744873, 2.740863561630249] |
bc9a05cb-2f80-43e3-9040-9d672be4d589 | cop-factual-inconsistency-detection-by | 2212.01611 | null | https://arxiv.org/abs/2212.01611v2 | https://arxiv.org/pdf/2212.01611v2.pdf | CoP: Factual Inconsistency Detection by Controlling the Preference | Abstractive summarization is the process of generating a summary given a document as input. Although significant progress has been made, the factual inconsistency between the document and the generated summary still limits its practical applications. Previous work found that the probabilities assigned by the generation model reflect its preferences for the generated summary, including the preference for factual consistency, and the preference for the language or knowledge prior as well. To separate the preference for factual consistency, we propose an unsupervised framework named CoP by controlling the preference of the generation model with the help of prompt. More specifically, the framework performs an extra inference step in which a text prompt is introduced as an additional input. In this way, another preference is described by the generation probability of this extra inference process. The difference between the above two preferences, i.e. the difference between the probabilities, could be used as measurements for detecting factual inconsistencies. Interestingly, we found that with the properly designed prompt, our framework could evaluate specific preferences and serve as measurements for fine-grained categories of inconsistency, such as entity-related inconsistency, coreference-related inconsistency, etc. Moreover, our framework could also be extended to the supervised setting to learn better prompt from the labeled data as well. Experiments show that our framework achieves new SOTA results on three factual inconsistency detection tasks. | ['Jiajun Chen', 'ShuJian Huang', 'Xiang Geng', 'Shuaijie She'] | 2022-12-03 | null | null | null | null | ['abstractive-text-summarization'] | ['natural-language-processing'] | [ 6.50651082e-02 6.29405379e-01 -4.84210521e-01 -5.95577180e-01
-9.28285420e-01 -7.19742715e-01 9.66599047e-01 6.39277816e-01
-2.56258190e-01 8.25363576e-01 5.78420341e-01 -1.91635620e-02
-1.03901483e-01 -7.08917975e-01 -5.88552058e-01 -5.53728521e-01
5.54639101e-01 6.45013809e-01 2.93179631e-01 -1.91925354e-02
5.39395094e-01 -5.06184325e-02 -1.33560002e+00 5.47454298e-01
1.35126030e+00 7.98332632e-01 2.18305781e-01 2.98059434e-01
-1.94753900e-01 8.20874810e-01 -9.27874863e-01 -4.82820839e-01
-1.44641370e-01 -6.29292250e-01 -1.04002964e+00 7.60198757e-02
3.57515275e-01 -2.24433035e-01 2.27981150e-01 1.12550187e+00
3.59167606e-01 2.78742816e-02 5.56232572e-01 -1.29904950e+00
-1.99860811e-01 1.04123139e+00 -1.68234065e-01 1.26969963e-01
7.10858881e-01 -1.11643247e-01 1.16896927e+00 -5.59300363e-01
6.24642968e-01 1.14121830e+00 2.78970152e-01 2.61290759e-01
-8.24668705e-01 -2.80111074e-01 3.65428150e-01 3.82401884e-01
-9.35066581e-01 -5.40548623e-01 9.62083936e-01 -2.95736432e-01
6.19334936e-01 3.88486028e-01 3.27443957e-01 1.04264867e+00
1.52298361e-01 8.62518787e-01 9.10749197e-01 -5.60943842e-01
5.01391590e-01 3.56981784e-01 3.50832582e-01 4.47900683e-01
4.93840069e-01 -4.44897383e-01 -4.62759882e-01 -3.41405511e-01
2.62708396e-01 -4.59574759e-01 -5.67660809e-01 -9.64049846e-02
-1.25703394e+00 5.70643425e-01 1.11937396e-01 3.95699382e-01
-3.62511009e-01 -2.99815506e-01 4.04235661e-01 1.21688128e-01
4.49348629e-01 5.91004014e-01 -3.93396616e-01 -1.00806519e-01
-1.06468928e+00 3.23871613e-01 9.92166936e-01 7.46671140e-01
6.99125230e-01 -4.36935872e-01 -7.38794506e-01 8.03542912e-01
1.33471459e-01 1.94276065e-01 5.08551121e-01 -1.14872444e+00
9.06329572e-01 8.45991015e-01 4.95098442e-01 -1.07427859e+00
-5.07506728e-01 -4.17688757e-01 -6.39475167e-01 -4.16643739e-01
5.17222762e-01 -1.05478577e-01 -3.64145696e-01 1.93825614e+00
2.58819312e-01 -1.65394649e-01 8.59209225e-02 9.16481018e-01
7.06907451e-01 4.64454174e-01 -1.33042127e-01 -6.43635809e-01
1.50908482e+00 -8.57087076e-01 -9.64654565e-01 -1.90331846e-01
5.99277198e-01 -7.17612386e-01 8.60226333e-01 3.04569155e-01
-1.17219388e+00 -3.92820537e-01 -9.41495419e-01 5.07269800e-02
-5.55826537e-03 3.85194600e-01 3.19435477e-01 2.52544284e-01
-7.42973745e-01 4.94037688e-01 -6.61528409e-01 -4.78101581e-01
-1.49199411e-01 9.87543166e-02 -1.38797849e-01 -1.87826063e-02
-1.23015857e+00 9.74068999e-01 7.29383826e-01 6.59662709e-02
-1.45111308e-01 -3.24415863e-01 -7.24984050e-01 2.37879395e-01
7.76151896e-01 -1.00931895e+00 1.32693565e+00 -8.24765325e-01
-1.37377691e+00 6.60267830e-01 -3.54212046e-01 -4.33155507e-01
6.43280923e-01 -1.06206998e-01 -2.46321946e-01 1.73815355e-01
3.78947914e-01 2.46093124e-01 5.17764688e-01 -1.18016481e+00
-7.04674900e-01 -3.70578408e-01 4.58258182e-01 3.75014573e-01
4.94351350e-02 -4.90653329e-02 -5.61535954e-01 -4.91407067e-01
2.82158256e-01 -8.41645777e-01 5.88694513e-02 -4.95844632e-01
-8.33151281e-01 -6.51359022e-01 4.74328190e-01 -6.25816226e-01
1.43045962e+00 -1.91315389e+00 5.99253066e-02 1.01093203e-01
-8.10633600e-02 7.85487983e-03 -5.11066280e-02 5.68623185e-01
7.25159943e-02 3.22457194e-01 -3.72103781e-01 -1.43364489e-01
6.28657788e-02 7.61553347e-02 -2.88894713e-01 9.47925672e-02
2.33407244e-01 5.78290284e-01 -1.09266078e+00 -5.75086296e-01
-8.20634142e-02 -3.07982694e-02 -5.79304099e-01 3.50328952e-01
-3.78912240e-01 4.22061920e-01 -6.39505684e-01 2.01399788e-01
6.07368946e-01 -3.33554506e-01 3.95909607e-01 -4.83716071e-01
-1.74917921e-01 6.78229630e-01 -1.24871707e+00 1.51404929e+00
-3.25355500e-01 3.16369146e-01 -1.24467641e-01 -1.09038866e+00
8.16735864e-01 4.09352690e-01 1.77573934e-01 -5.27788818e-01
-3.42186503e-02 2.44928136e-01 5.70416152e-02 -7.46111155e-01
6.97216570e-01 5.96184358e-02 -1.91794887e-01 6.08023643e-01
-2.50505339e-02 -1.25368118e-01 4.46470082e-01 5.42431593e-01
8.87037456e-01 2.76772631e-03 6.97672725e-01 -2.59719163e-01
7.26071000e-01 7.24808034e-03 7.35109270e-01 9.55124438e-01
3.89658846e-02 6.67585194e-01 8.75034273e-01 -1.47940665e-01
-5.46675205e-01 -7.54200637e-01 -3.87170874e-02 6.83168590e-01
2.69323170e-01 -7.00843811e-01 -7.89792538e-01 -9.89509165e-01
-3.47176254e-01 1.22109401e+00 -3.40931326e-01 -2.99640685e-01
-6.44567013e-01 -7.05230892e-01 3.74940455e-01 3.80016774e-01
6.55743778e-01 -9.71679926e-01 -6.13461733e-01 2.05728367e-01
-1.02826595e+00 -1.13003504e+00 -5.13176918e-01 1.66041274e-02
-7.61353672e-01 -1.28633559e+00 -1.85772166e-01 -3.32344860e-01
7.00873435e-01 7.58986920e-02 1.02907503e+00 9.52830687e-02
5.12871563e-01 4.00319934e-01 -3.96441340e-01 -1.41379088e-01
-6.21738732e-01 -3.54620591e-02 -3.01806591e-02 2.00039875e-02
-6.50203750e-02 -3.37230921e-01 -5.44043958e-01 2.37669230e-01
-9.92417336e-01 1.84088320e-01 4.02219713e-01 7.46249914e-01
3.12973827e-01 2.06579536e-01 6.79012001e-01 -1.09194481e+00
9.72165883e-01 -5.75850725e-01 -2.37215966e-01 6.48923814e-01
-6.13410234e-01 5.08781612e-01 8.16970706e-01 -1.03310943e-01
-1.51642132e+00 -5.04245043e-01 -5.31260334e-02 1.92179367e-01
-1.64729178e-01 8.99897218e-01 -5.62671900e-01 7.74740398e-01
4.64949638e-01 6.20150417e-02 -4.84065950e-01 -2.44269103e-01
2.61749327e-01 6.92250311e-01 6.70737922e-01 -9.26522136e-01
4.21232790e-01 2.40399390e-01 -3.91845465e-01 -5.50095737e-01
-1.12996542e+00 -3.66102248e-01 -4.93308663e-01 -1.78997964e-02
4.99786615e-01 -5.66864848e-01 -3.85435045e-01 1.79661453e-01
-1.57049680e+00 8.31356123e-02 -1.18619971e-01 4.13743556e-01
-5.64564228e-01 7.54756391e-01 -3.33727747e-01 -8.54921579e-01
-3.53510469e-01 -9.59996462e-01 9.03554261e-01 2.89417237e-01
-6.04333103e-01 -9.96665835e-01 1.17079057e-01 4.42540258e-01
6.88475817e-02 5.26061207e-02 1.19066632e+00 -1.10176361e+00
-4.30092454e-01 -3.41726840e-01 -1.14695460e-01 2.27257341e-01
2.49742091e-01 2.29597792e-01 -7.15842009e-01 -9.95250139e-03
3.28784496e-01 -1.28842443e-01 7.49906719e-01 2.93475538e-01
1.01384807e+00 -8.48532915e-01 -2.51179129e-01 5.99136017e-02
1.14504158e+00 1.11563921e-01 5.86895645e-01 3.73181462e-01
3.19919080e-01 7.26088941e-01 7.95929074e-01 5.12759089e-01
6.20465577e-01 9.50785577e-01 1.44466937e-01 4.27817523e-01
3.38685103e-02 -3.50866646e-01 3.34929377e-01 8.30383420e-01
-1.93673745e-01 -5.95664740e-01 -6.62819505e-01 4.03517157e-01
-2.21705413e+00 -1.15405798e+00 -7.36493841e-02 2.47406459e+00
1.00080919e+00 2.15890631e-01 -1.35672957e-01 2.72184402e-01
8.44592094e-01 3.34939182e-01 -2.74800271e-01 -3.75916928e-01
-9.13355872e-02 -2.95963109e-01 -1.33780763e-01 5.84183514e-01
-7.55818009e-01 6.63883984e-01 5.56324387e+00 6.55877113e-01
-1.11897194e+00 -1.74899697e-01 3.80521566e-01 1.69163346e-01
-7.06587195e-01 3.95520091e-01 -6.55643761e-01 7.38262177e-01
7.41515040e-01 -3.88901114e-01 2.79145464e-02 5.39213121e-01
5.39202154e-01 -4.84061480e-01 -1.30879533e+00 5.12096167e-01
1.86659485e-01 -1.17305326e+00 2.85774082e-01 -1.91978931e-01
3.99190426e-01 -4.92934436e-01 -3.02827001e-01 1.44466385e-01
1.08022839e-01 -3.47354293e-01 8.66004169e-01 5.14701962e-01
3.28628629e-01 -4.63307410e-01 1.12961149e+00 7.93131292e-01
-6.78263247e-01 2.52893511e-02 -1.04313143e-01 -5.35380132e-02
1.35768756e-01 1.02255976e+00 -1.02563894e+00 9.61465120e-01
2.17672691e-01 4.88089919e-01 -6.06645823e-01 9.55793262e-01
-6.20071232e-01 6.19446516e-01 -2.40601808e-01 1.41031772e-01
-1.56537905e-01 -1.70431435e-01 7.46753693e-01 1.16593909e+00
3.14875841e-01 1.28149033e-01 2.35596076e-01 9.13541794e-01
-9.59681273e-02 1.43008918e-01 -2.77911305e-01 1.36530399e-01
7.75802076e-01 1.18540764e+00 -5.99769473e-01 -6.03906691e-01
-5.31899408e-02 8.98501217e-01 4.68155235e-01 4.16428357e-01
-7.49749422e-01 -1.14076450e-01 3.55547853e-02 -4.93525676e-02
5.06472327e-02 2.54951775e-01 -3.79059076e-01 -1.42057991e+00
5.10132909e-01 -7.58442640e-01 5.84609687e-01 -8.52098465e-01
-1.09990692e+00 3.71669292e-01 2.25181863e-01 -1.10001707e+00
-4.33500856e-01 -8.03359374e-02 -1.08187735e+00 6.07664645e-01
-1.50329792e+00 -6.83148563e-01 -3.48578393e-01 2.32255191e-01
6.18485034e-01 2.11647525e-01 6.61159456e-01 -1.31968722e-01
-8.95729482e-01 5.08244097e-01 -4.23733473e-01 -1.03692450e-02
1.04588246e+00 -1.32705450e+00 -2.45841488e-01 1.04007220e+00
8.85245055e-02 7.78435767e-01 1.09770429e+00 -6.79885209e-01
-8.65855157e-01 -8.23597431e-01 1.38951850e+00 -3.03091407e-01
3.94024074e-01 2.28813022e-01 -1.03049743e+00 5.34389257e-01
3.62606943e-01 -3.69774491e-01 5.03550291e-01 2.18910813e-01
-4.05822188e-01 -6.94445148e-02 -1.07200551e+00 4.89103466e-01
9.08320665e-01 -3.17431092e-01 -1.07682061e+00 2.39349842e-01
5.32733142e-01 -5.17226219e-01 -5.74864447e-01 4.29990113e-01
3.69147629e-01 -1.10557044e+00 4.17938292e-01 -4.41134453e-01
7.34263480e-01 -3.81941497e-01 -8.03788975e-02 -1.49876714e+00
-2.91399777e-01 -3.22324008e-01 -1.82572350e-01 1.65729272e+00
4.83621269e-01 -4.98878270e-01 2.86052525e-01 7.97883451e-01
-1.91197887e-01 -4.47340369e-01 -8.21404576e-01 -4.90981013e-01
-1.51349962e-01 -2.99858242e-01 6.33256018e-01 8.43235075e-01
4.98877704e-01 5.21701992e-01 -1.96206629e-01 2.92615384e-01
3.14377397e-01 3.98600250e-01 6.60993934e-01 -1.23595774e+00
-2.17923775e-01 -3.80145520e-01 1.39633924e-01 -9.83644187e-01
3.16404432e-01 -7.55006492e-01 1.35759309e-01 -1.85601497e+00
3.52093101e-01 -3.38598043e-01 -2.17567265e-01 3.64858687e-01
-6.22525215e-01 -5.19795060e-01 1.62602052e-01 3.97106409e-01
-8.16390276e-01 4.48768377e-01 1.07730865e+00 -1.71397075e-01
-3.72371078e-01 2.40752637e-01 -8.71919215e-01 8.86611402e-01
8.89297307e-01 -4.46489841e-01 -3.82113695e-01 -3.02569717e-01
3.18975449e-01 4.24528629e-01 2.68324107e-01 -6.77672863e-01
4.93209600e-01 -3.45352292e-01 -1.09579824e-01 -4.98410970e-01
-8.41948390e-02 -5.45898080e-01 2.31797192e-02 1.44454375e-01
-6.06236100e-01 -1.46413341e-01 -1.90348208e-01 5.47319353e-01
-4.01392967e-01 -4.71062303e-01 3.07070613e-01 -2.38172412e-01
-3.84797901e-01 -3.13729376e-01 -2.16246903e-01 3.12011153e-01
5.13052702e-01 -1.05531469e-01 -7.73478627e-01 -5.48128366e-01
-4.45615262e-01 3.41597378e-01 4.34295774e-01 3.90426248e-01
3.44111294e-01 -9.63813245e-01 -7.87842512e-01 -2.32478976e-01
3.00707340e-01 1.04184866e-01 1.78851530e-01 1.01056004e+00
1.75018713e-01 6.09651804e-01 3.49738114e-02 -4.90792215e-01
-1.00890696e+00 2.88167208e-01 1.51571825e-01 -7.07094550e-01
-3.00306559e-01 1.34103298e-01 7.38991499e-02 -2.39983827e-01
7.80980512e-02 -5.33481002e-01 -3.83216292e-01 3.41044605e-01
5.20956039e-01 3.84980291e-01 2.72807658e-01 -2.79037595e-01
-3.79289955e-01 3.25848132e-01 -9.95645821e-02 -9.69469845e-02
9.72069800e-01 -3.67599279e-01 -3.41263115e-01 6.00300491e-01
6.97511137e-01 3.92344952e-01 -1.01284385e+00 -1.62859395e-01
4.08750355e-01 -1.94624707e-01 -2.88641989e-01 -1.06802583e+00
-7.02339709e-01 4.28124487e-01 -2.57575631e-01 3.75807464e-01
9.38123524e-01 1.63813159e-01 3.30297410e-01 4.20396775e-01
3.85248095e-01 -1.09759820e+00 -2.07169633e-02 4.53570813e-01
1.11365628e+00 -1.31598818e+00 -3.05184647e-02 -6.38879001e-01
-7.12069571e-01 1.19300258e+00 5.91206849e-01 2.55642325e-01
1.10717155e-02 4.48908359e-02 -5.96594289e-02 -1.28826410e-01
-8.61780405e-01 8.30769613e-02 4.33770806e-01 2.54518151e-01
5.79770565e-01 1.22318767e-01 -8.80966067e-01 1.08618963e+00
-3.64916712e-01 6.17485493e-02 7.60306478e-01 5.04337966e-01
-4.55140948e-01 -1.01334178e+00 -3.46160144e-01 3.83735627e-01
-2.36996651e-01 1.56159014e-01 -5.54106474e-01 5.60451567e-01
7.43370280e-02 1.27562261e+00 -2.56385710e-02 -9.60721299e-02
5.60418725e-01 2.29125410e-01 4.55709845e-01 -7.52888560e-01
-5.70823848e-01 -1.47544488e-01 5.33018172e-01 -5.05793095e-01
-6.09511971e-01 -5.55184066e-01 -1.31868696e+00 8.39100555e-02
-4.55557674e-01 5.82309663e-01 2.99453944e-01 1.38167870e+00
3.08865607e-01 2.43284732e-01 5.18114984e-01 -4.50393081e-01
-6.19060576e-01 -1.01606250e+00 -2.04453379e-01 5.59077740e-01
2.29821563e-01 -4.57120150e-01 -6.21313035e-01 -9.82629135e-02] | [12.146148681640625, 9.292502403259277] |
af88d7f1-e37d-4b7b-af1a-bbe8cd943640 | the-limitations-of-limited-context-for | 2106.01580 | null | https://arxiv.org/abs/2106.01580v1 | https://arxiv.org/pdf/2106.01580v1.pdf | The Limitations of Limited Context for Constituency Parsing | Incorporating syntax into neural approaches in NLP has a multitude of practical and scientific benefits. For instance, a language model that is syntax-aware is likely to be able to produce better samples; even a discriminative model like BERT with a syntax module could be used for core NLP tasks like unsupervised syntactic parsing. Rapid progress in recent years was arguably spurred on by the empirical success of the Parsing-Reading-Predict architecture of (Shen et al., 2018a), later simplified by the Order Neuron LSTM of (Shen et al., 2019). Most notably, this is the first time neural approaches were able to successfully perform unsupervised syntactic parsing (evaluated by various metrics like F-1 score). However, even heuristic (much less fully mathematical) understanding of why and when these architectures work is lagging severely behind. In this work, we answer representational questions raised by the architectures in (Shen et al., 2018a, 2019), as well as some transition-based syntax-aware language models (Dyer et al., 2016): what kind of syntactic structure can current neural approaches to syntax represent? Concretely, we ground this question in the sandbox of probabilistic context-free-grammars (PCFGs), and identify a key aspect of the representational power of these approaches: the amount and directionality of context that the predictor has access to when forced to make parsing decision. We show that with limited context (either bounded, or unidirectional), there are PCFGs, for which these approaches cannot represent the max-likelihood parse; conversely, if the context is unlimited, they can represent the max-likelihood parse of any PCFG. | ['Andrej Risteski', 'Yuchen Li'] | 2021-06-03 | null | https://aclanthology.org/2021.acl-long.208 | https://aclanthology.org/2021.acl-long.208.pdf | acl-2021-5 | ['constituency-parsing'] | ['natural-language-processing'] | [ 4.10729975e-01 5.82057476e-01 -1.43538073e-01 -5.54933429e-01
-7.82847285e-01 -8.79898846e-01 5.50931811e-01 2.65280932e-01
-4.39383984e-01 5.61352670e-01 5.91443181e-01 -1.00490046e+00
-1.69714212e-01 -9.42277133e-01 -7.57027984e-01 -5.79168200e-01
-7.28416368e-02 3.63659114e-01 -3.15619558e-02 -1.27001807e-01
1.53792100e-02 2.71514177e-01 -1.41436183e+00 3.17716688e-01
6.95212066e-01 6.12122893e-01 6.08074546e-01 6.62755668e-01
-4.23499048e-01 8.12429667e-01 -3.03892583e-01 -3.54597807e-01
-3.26511115e-02 -3.83298278e-01 -9.05550063e-01 -5.50642729e-01
2.82993704e-01 -1.14092238e-01 -2.17334088e-02 1.05963349e+00
1.22023284e-01 -4.74634208e-02 4.53534126e-01 -5.67146897e-01
-7.71009505e-01 1.15409017e+00 8.88180733e-03 1.47000000e-01
1.43898830e-01 2.43622050e-01 1.44375038e+00 -5.48310280e-01
6.76111639e-01 1.40519583e+00 4.41307962e-01 6.14137053e-01
-1.49052954e+00 -4.09503251e-01 4.57705587e-01 -2.84948558e-01
-7.94841230e-01 -2.31856987e-01 6.65565729e-01 -3.12565804e-01
1.11738110e+00 3.36418673e-02 6.26268864e-01 1.28123772e+00
1.45007625e-01 1.07431233e+00 1.36837673e+00 -6.19829774e-01
3.92487198e-01 -2.58870989e-01 4.11244273e-01 6.54214621e-01
3.16396266e-01 3.33263516e-01 -4.62734163e-01 -1.23210169e-01
5.88600457e-01 -4.35608715e-01 -1.63060322e-03 -8.00344273e-02
-9.68415797e-01 9.69315708e-01 1.82023630e-01 5.94098270e-01
-2.03459963e-01 2.55479604e-01 4.01300043e-01 2.52961218e-01
1.75995067e-01 6.42766833e-01 -7.39388704e-01 -3.00558358e-01
-1.00431132e+00 4.37132359e-01 9.11232948e-01 6.06479585e-01
6.41721487e-01 1.70752883e-01 4.84381430e-02 5.84119081e-01
4.69769090e-01 2.40903661e-01 5.14284968e-01 -9.82181311e-01
4.02312130e-01 2.34850705e-01 -1.97183356e-01 -6.36308968e-01
-6.56440914e-01 -1.83956847e-01 -4.60375696e-01 8.31604749e-02
9.79161501e-01 -3.49141210e-01 -7.97716439e-01 2.31322265e+00
-1.86148994e-02 -3.10059786e-01 3.90149988e-02 7.61651456e-01
3.89367878e-01 7.49542296e-01 6.80129409e-01 -8.05471241e-02
1.44617212e+00 -2.26566181e-01 -3.55683565e-01 -7.00197160e-01
9.14460003e-01 -5.14909327e-01 1.26251853e+00 4.05255824e-01
-1.03748357e+00 -3.30995202e-01 -8.46634686e-01 -1.77535415e-01
-2.87856072e-01 5.26805073e-02 1.12405574e+00 7.80154228e-01
-1.21189713e+00 7.45058537e-01 -9.97760773e-01 -4.98844475e-01
3.21144372e-01 2.77937710e-01 -4.12473440e-01 -9.51198488e-03
-1.30244422e+00 9.09476757e-01 6.33663774e-01 1.31906852e-01
-6.44821942e-01 -2.93721646e-01 -8.02480936e-01 2.36469999e-01
4.12070394e-01 -6.09814525e-01 1.22566652e+00 -1.32269514e+00
-1.43876338e+00 8.75037551e-01 -4.48230416e-01 -5.21212101e-01
2.21967883e-02 -2.05507249e-01 1.11005217e-01 -1.52737021e-01
-2.23505035e-01 6.30115151e-01 3.31043541e-01 -1.08523488e+00
-4.57507968e-01 -5.64996839e-01 3.89195770e-01 4.38897945e-02
-1.19322632e-02 4.00823772e-01 1.54269800e-01 -6.34249330e-01
3.31589550e-01 -8.41279149e-01 -2.41999850e-01 -2.06135884e-01
-4.52000052e-01 -5.14833212e-01 2.10325003e-01 -8.34802568e-01
1.00347066e+00 -2.10514879e+00 6.05338365e-02 4.08465527e-02
1.41082749e-01 2.20363602e-01 -1.70576647e-01 3.86636376e-01
-1.71359554e-01 6.97527528e-01 -5.42597353e-01 -3.74129266e-01
3.19187820e-01 4.25592065e-01 -5.58851540e-01 1.83825642e-01
5.31285524e-01 1.02443802e+00 -8.11913908e-01 -2.69126683e-01
-1.55737745e-02 3.71377856e-01 -6.15471661e-01 -6.02604337e-02
-5.87819219e-01 2.68503726e-01 -4.59755331e-01 5.19600332e-01
2.92374760e-01 -8.74643959e-03 6.73055828e-01 3.07758272e-01
-3.13068807e-01 9.73109663e-01 -1.01749265e+00 1.48341572e+00
-3.88338149e-01 6.40267491e-01 3.30858380e-01 -1.09754133e+00
7.36681402e-01 1.91081136e-01 -6.93646222e-02 -4.59786922e-01
1.56266615e-01 3.06827933e-01 4.84289438e-01 -1.15745030e-01
4.21344787e-01 -5.02783716e-01 -2.12663263e-01 6.66263044e-01
1.64500654e-01 -2.72907373e-02 6.49539754e-02 9.72619727e-02
1.29729438e+00 5.90778649e-01 3.17431390e-01 -4.44596231e-01
1.19788744e-01 1.72058679e-02 7.21895874e-01 1.07201421e+00
-2.80743778e-01 5.33087134e-01 9.17148709e-01 -4.34125185e-01
-9.93173599e-01 -1.17596424e+00 -2.83721060e-01 1.33573055e+00
-5.93186319e-01 -4.46733534e-01 -7.64969587e-01 -5.92906833e-01
-2.17544913e-01 9.81216848e-01 -5.89128315e-01 7.34694824e-02
-9.98056710e-01 -8.02477181e-01 8.56850266e-01 6.43430769e-01
-3.80490236e-02 -1.35260665e+00 -6.54017150e-01 3.65774870e-01
1.14554957e-01 -8.34938884e-01 2.08768606e-01 8.43173265e-01
-1.17442727e+00 -7.80958295e-01 -3.38262647e-01 -8.28540087e-01
3.41612071e-01 -4.00724083e-01 1.13895440e+00 -3.89528275e-02
1.31307304e-01 7.97887444e-02 -2.75101364e-01 -4.27238852e-01
-5.59113562e-01 7.00405762e-02 -2.33638391e-01 -5.12263238e-01
5.14290869e-01 -8.16992342e-01 -2.39038274e-01 -3.48479867e-01
-9.44828510e-01 -2.39144992e-02 6.63309395e-01 8.38791966e-01
5.47607660e-01 -2.51427680e-01 6.25308216e-01 -1.10763538e+00
5.78203797e-01 -5.67637086e-01 -5.80762982e-01 1.00540496e-01
-3.95866185e-01 3.83476138e-01 8.42275321e-01 -2.38882363e-01
-1.04096007e+00 5.15684113e-02 -4.89641786e-01 1.96050525e-01
-5.43688834e-01 7.92525649e-01 -3.48652720e-01 4.22716379e-01
4.97744381e-01 3.29019010e-01 -2.70158738e-01 -6.07454240e-01
5.89783490e-01 2.33960599e-01 1.97126135e-01 -1.17250264e+00
5.90210915e-01 1.53422982e-01 -1.89037025e-02 -7.08880305e-01
-7.76728928e-01 1.69248194e-01 -6.08429492e-01 2.10945487e-01
9.19702888e-01 -6.55300498e-01 -4.96279180e-01 1.23090751e-01
-1.28106272e+00 -6.27351046e-01 -3.35872054e-01 4.35030133e-01
-6.03214741e-01 3.85494471e-01 -7.31711447e-01 -1.10302484e+00
5.72107993e-02 -1.17626703e+00 8.19173872e-01 4.16152664e-02
-5.08957624e-01 -1.13054156e+00 -1.22718364e-01 -2.82504261e-02
5.24913311e-01 6.61241636e-02 1.72815311e+00 -1.08644223e+00
-4.74789679e-01 4.56428081e-02 -9.71142501e-02 2.99763352e-01
-3.10505748e-01 -1.39102815e-02 -1.08169842e+00 1.29988760e-01
2.07289666e-01 -3.66956562e-01 9.31375444e-01 5.48942149e-01
1.07987964e+00 -4.36831564e-01 -1.54441297e-01 4.36353087e-01
1.31551576e+00 1.83750197e-01 4.65584725e-01 2.95251697e-01
4.18783844e-01 1.06914842e+00 1.58410862e-01 -1.85292751e-01
6.06926799e-01 3.41873199e-01 2.65416116e-01 3.62190485e-01
-3.15076411e-02 -5.69276094e-01 6.82124734e-01 8.11897576e-01
2.05324903e-01 -3.05165082e-01 -1.06726229e+00 5.45697331e-01
-1.54785597e+00 -8.47713113e-01 -6.15958385e-02 2.21722341e+00
9.45945680e-01 3.73261362e-01 -1.02401346e-01 -2.97517441e-02
5.17865121e-01 4.58892554e-01 -3.60569805e-01 -1.01435053e+00
-3.31908017e-01 6.70982122e-01 1.46606416e-01 6.11334920e-01
-9.81651723e-01 1.17564201e+00 6.23003674e+00 6.35973036e-01
-9.78227854e-01 -4.04229164e-02 8.40999484e-01 1.33842500e-02
-6.80671215e-01 5.57465672e-01 -8.38870287e-01 4.54531580e-01
1.42322028e+00 4.82960008e-02 5.43977320e-01 8.27116668e-01
3.17531154e-02 -3.53805751e-01 -1.32122076e+00 4.35702592e-01
-1.34747669e-01 -1.04519796e+00 -9.08038840e-02 2.25484431e-01
1.32670000e-01 2.71344453e-01 -1.05672807e-01 5.24288654e-01
7.88633943e-01 -1.30102944e+00 8.05692077e-01 3.48585606e-01
3.22697073e-01 -4.95080680e-01 5.01242638e-01 7.35025764e-01
-7.97239244e-01 -7.64217377e-02 -4.97983783e-01 -3.38602602e-01
3.01088244e-01 8.33385825e-01 -4.15251225e-01 1.88967481e-01
4.27975476e-01 2.54068375e-01 -4.29027796e-01 4.45457757e-01
-7.09770679e-01 1.19987357e+00 -6.35798216e-01 -3.58208358e-01
3.06064785e-01 -1.88139930e-01 6.86511636e-01 1.32721674e+00
1.53272912e-01 1.45978913e-01 5.86318411e-03 1.14726591e+00
8.70373920e-02 1.46249533e-01 -7.32495010e-01 -3.90195876e-01
5.32475948e-01 1.02145243e+00 -6.91220224e-01 -7.00391829e-02
-2.93628663e-01 3.20034206e-01 6.86070502e-01 3.55097264e-01
-3.19637060e-01 -8.74937549e-02 4.77440804e-01 -5.60572669e-02
2.18719840e-01 -5.19189417e-01 -6.47690177e-01 -1.03181303e+00
9.34102833e-02 -7.33941436e-01 2.10210696e-01 -6.24266982e-01
-1.16449404e+00 3.23674619e-01 -2.42961780e-03 -2.08314151e-01
-3.95519048e-01 -1.07790136e+00 -6.06388986e-01 1.09439635e+00
-1.22529256e+00 -1.01944065e+00 6.74623847e-01 -3.15813497e-02
3.42762202e-01 8.63184482e-02 1.06581581e+00 -1.95758536e-01
-4.34124768e-01 3.25130969e-01 3.86354588e-02 4.36686546e-01
1.91457063e-01 -1.49408674e+00 5.96605003e-01 9.24618185e-01
5.13176143e-01 1.14402139e+00 7.00405598e-01 -4.92774308e-01
-1.65793395e+00 -7.96785593e-01 1.35802031e+00 -6.17512584e-01
7.61720717e-01 -6.47915781e-01 -9.68509793e-01 1.03655779e+00
9.53301340e-02 -1.48613706e-01 6.60244405e-01 4.70230728e-01
-4.77635443e-01 3.02816391e-01 -8.94227684e-01 7.37118542e-01
8.80178690e-01 -5.79776943e-01 -9.70093727e-01 7.42260814e-02
7.37861037e-01 -1.27724022e-01 -6.99504256e-01 2.90789008e-01
6.80137694e-01 -1.08549798e+00 5.49029171e-01 -7.48967767e-01
5.92791438e-01 1.67058129e-02 -4.53610331e-01 -1.08091533e+00
-1.86584726e-01 -4.73744839e-01 8.56886283e-02 1.41163623e+00
6.37414575e-01 -7.83493996e-01 7.59213090e-01 8.15426469e-01
-2.30332226e-01 -1.06067729e+00 -1.10094845e+00 -5.93100011e-01
9.04731810e-01 -9.14482951e-01 3.67705822e-01 5.93390822e-01
6.98259100e-02 2.39330798e-01 2.89224088e-01 -3.19599509e-02
3.04213494e-01 -1.67594422e-02 4.07297581e-01 -1.32295477e+00
-5.59664786e-01 -7.42946148e-01 -2.05418631e-01 -1.04487193e+00
4.76083189e-01 -1.27677703e+00 2.40860656e-01 -1.57072008e+00
8.11879933e-02 -5.32839835e-01 -2.90873557e-01 6.77652538e-01
-1.10961884e-01 -4.86683011e-01 5.04686117e-01 1.06068663e-01
-1.96993664e-01 2.18917578e-01 6.77515388e-01 3.22175711e-01
-2.36318242e-02 -3.36204410e-01 -9.67323422e-01 1.02652800e+00
8.86353314e-01 -6.56289399e-01 -1.37870461e-01 -5.87513626e-01
6.64199054e-01 4.60872836e-02 3.62753004e-01 -6.59960449e-01
5.30615672e-02 -2.15843916e-01 2.50038683e-01 -2.69946814e-01
8.12121332e-02 -4.08514917e-01 -2.53846526e-01 2.74664670e-01
-5.99817455e-01 1.75805971e-01 3.18126768e-01 4.52280223e-01
-7.42284581e-02 -5.48529923e-01 5.87907195e-01 -4.24308389e-01
-4.31830615e-01 -1.72259510e-01 -7.41729319e-01 3.17978442e-01
4.58934575e-01 -3.02384757e-02 -3.75948101e-01 -2.24050328e-01
-7.99258590e-01 -1.90629035e-01 3.60474378e-01 8.99909139e-02
2.88697869e-01 -6.73629522e-01 -5.55830419e-01 1.03241380e-03
-2.31019810e-01 1.23333380e-01 -1.61757141e-01 6.06887579e-01
-2.34454215e-01 4.85113770e-01 2.37366542e-01 -2.37090036e-01
-5.50026178e-01 2.81775802e-01 1.74874470e-01 -5.65964520e-01
-5.59330046e-01 9.62827802e-01 4.38794315e-01 -7.69517660e-01
-3.73948179e-02 -7.05037475e-01 -4.91461381e-02 -8.31818804e-02
4.11493212e-01 -9.46677253e-02 -1.50902212e-01 -4.82897490e-01
-3.42178255e-01 2.51413554e-01 -1.24154679e-01 -2.68769860e-01
1.43952012e+00 1.71336621e-01 -1.86638340e-01 7.19618618e-01
8.22991788e-01 1.60243988e-01 -1.15245152e+00 -5.07189110e-02
4.49024856e-01 2.32386231e-01 -6.92433938e-02 -1.16355610e+00
-6.48572505e-01 1.31182253e+00 2.10564107e-01 2.81574667e-01
8.10690880e-01 1.75984874e-01 6.54590428e-01 4.79014248e-01
3.95960480e-01 -9.16593850e-01 -6.16788626e-01 1.00449872e+00
4.89752829e-01 -9.56246257e-01 -3.24220985e-01 -7.15377880e-03
-3.85250151e-01 1.33015609e+00 2.81197339e-01 -2.50587761e-01
4.69475359e-01 4.79808152e-01 -2.81259000e-01 -1.34013444e-01
-9.46625829e-01 -2.77120441e-01 -1.40668884e-01 6.48426473e-01
8.84401083e-01 2.57296056e-01 -3.88983577e-01 9.87945914e-01
-5.44983983e-01 -2.37943679e-01 3.85412842e-01 8.75151992e-01
-6.32853806e-01 -1.29856634e+00 -2.14681342e-01 5.92417300e-01
-6.81871831e-01 -5.39899647e-01 -4.82975304e-01 8.30608785e-01
3.18333916e-02 9.14990902e-01 -9.63446591e-03 -5.23245297e-02
2.52886377e-02 7.24540770e-01 3.36394340e-01 -9.89290655e-01
-7.52239048e-01 -5.13437279e-02 2.13501379e-01 -5.28251588e-01
-1.24346994e-01 -9.33829129e-01 -1.39006162e+00 -1.17731011e-02
-7.68255293e-02 -4.57901582e-02 8.58551919e-01 1.26182961e+00
2.20041603e-01 2.52512932e-01 7.33034760e-02 -6.56161249e-01
-6.55753374e-01 -8.70060980e-01 -4.99022186e-01 -8.86548460e-02
1.26722142e-01 -3.03942323e-01 -5.39817750e-01 -1.70526206e-01] | [10.435135841369629, 9.510151863098145] |
627b51ab-8110-43c4-8d53-1ebaf5ff75bc | ds4dh-at-semeval-2022-task-11-multilingual | null | null | https://aclanthology.org/2022.semeval-1.212 | https://aclanthology.org/2022.semeval-1.212.pdf | DS4DH at SemEval-2022 Task 11: Multilingual Named Entity Recognition Using an Ensemble of Transformer-based Language Models | In this paper, we describe our proposed method for the SemEval 2022 Task 11: Multilingual Complex Named Entity Recognition (MultiCoNER). The goal of this task is to locate and classify named entities in unstructured short complex texts in 11 different languages.After training a variety of contextual language models on the NER dataset, we used an ensemble strategy based on a majority vote to finalize our model. We evaluated our proposed approach on the multilingual NER dataset at SemEval-2022. The ensemble model provided consistent improvements against the individual models on the multilingual track, achieving a macro F1 performance of 65.2%. However, our results were significantly outperformed by the top ranking systems, achieving thus a baseline performance. | ['Douglas Teodoro', 'Hossein Rouhizadeh'] | null | null | null | null | semeval-naacl-2022-7 | ['multilingual-named-entity-recognition'] | ['natural-language-processing'] | [-4.73223239e-01 5.28321788e-02 4.15565558e-02 -3.21601629e-01
-1.45289958e+00 -8.69744837e-01 9.52883303e-01 3.30964059e-01
-1.18848670e+00 1.10144258e+00 3.68197888e-01 -3.58779639e-01
3.20161998e-01 -3.01386267e-01 -5.21114945e-01 -1.03498027e-01
8.71824473e-03 7.09869146e-01 2.22781926e-01 -3.67549747e-01
9.43439081e-02 6.29973412e-01 -8.51263463e-01 3.98411483e-01
1.12611854e+00 5.45021236e-01 1.34674534e-01 6.66693270e-01
-2.29796901e-01 1.04240477e+00 -7.77623892e-01 -7.70305634e-01
-5.78084402e-02 2.90018730e-02 -1.03161621e+00 -6.46931350e-01
4.92319465e-01 4.68254834e-01 -1.17003741e-02 9.71522927e-01
7.70985484e-01 2.89370030e-01 7.43297219e-01 -5.99744618e-01
-4.05509949e-01 1.26703548e+00 -1.23278715e-01 2.13027403e-01
3.95520508e-01 -5.67648590e-01 1.19509792e+00 -1.59393311e+00
1.09583688e+00 9.32538986e-01 8.44773233e-01 4.70382333e-01
-9.96250987e-01 -7.25022793e-01 2.17776336e-02 1.69548124e-01
-1.82028842e+00 -8.97971451e-01 -4.63638417e-02 -2.65623480e-01
1.51373792e+00 5.54668158e-02 -6.71826303e-02 7.89609373e-01
1.30810305e-01 7.86770225e-01 1.30408561e+00 -6.74208581e-01
2.07429342e-02 2.71609187e-01 4.26380903e-01 3.51325244e-01
4.01376992e-01 -2.11351767e-01 -3.71839941e-01 -1.97298333e-01
3.38539295e-02 -6.30728424e-01 -1.41994461e-01 3.23868215e-01
-1.36374748e+00 5.56540072e-01 1.39361799e-01 9.21808898e-01
-6.22690380e-01 -3.57028872e-01 4.87363309e-01 3.19992900e-01
3.95958692e-01 8.44622493e-01 -1.18211567e+00 7.00243004e-03
-9.80252504e-01 -9.51315388e-02 1.17954481e+00 1.05129802e+00
3.74804705e-01 -1.54629886e-01 -2.31200710e-01 1.03444219e+00
2.74097025e-01 7.52972543e-01 2.74710208e-01 -3.40214163e-01
9.01138008e-01 3.71081740e-01 3.52625698e-01 -6.38969898e-01
-6.61356151e-01 -5.04291952e-01 -6.59425259e-01 -2.97458678e-01
4.22852457e-01 -7.08368659e-01 -9.46768224e-01 1.46516991e+00
8.50976482e-02 -2.30399385e-01 5.78124702e-01 2.33343512e-01
9.05907631e-01 7.42613733e-01 7.25642622e-01 -1.38883382e-01
1.30015063e+00 -1.21617782e+00 -9.55986977e-01 -4.28060405e-02
1.00662196e+00 -1.15349758e+00 2.67186075e-01 1.30343646e-01
-7.75375843e-01 -4.75409091e-01 -7.87112176e-01 -4.70786393e-02
-7.63377428e-01 6.43935680e-01 1.23650849e-01 5.97030938e-01
-1.10246670e+00 3.99442852e-01 -7.60566890e-01 -5.40422261e-01
-3.10769588e-01 1.64418742e-01 -8.29044700e-01 1.15228547e-02
-1.48303986e+00 1.32302010e+00 9.70020533e-01 1.68306813e-01
-6.70423090e-01 -4.63237137e-01 -9.12451327e-01 -1.27783805e-01
1.14144102e-01 -7.34275505e-02 1.21829724e+00 -2.14522079e-01
-1.17924750e+00 8.30944598e-01 -1.91983879e-01 -5.33129275e-01
4.36075479e-01 -6.68842673e-01 -1.20922673e+00 -2.46009976e-01
4.14810181e-01 4.05960590e-01 -1.59699291e-01 -1.01804686e+00
-1.01559210e+00 -1.08005725e-01 -3.34607661e-01 1.50931701e-01
6.10980298e-03 7.25102484e-01 -4.04969394e-01 -6.83780909e-01
-2.59796262e-01 -1.04992378e+00 -4.33293730e-01 -1.26295269e+00
-5.84416270e-01 -4.64810103e-01 3.16208929e-01 -1.24013352e+00
1.58221912e+00 -1.88082540e+00 -7.03683048e-02 3.98731828e-02
-8.07017237e-02 4.46183473e-01 -1.89297467e-01 7.34305859e-01
-2.73573905e-01 4.77322549e-01 -2.27478668e-02 -4.30214167e-01
-1.37561545e-01 -2.86144167e-01 -5.51584139e-02 1.95351660e-01
1.51123151e-01 7.01853871e-01 -9.22473907e-01 -5.51410258e-01
-5.93627542e-02 6.15778029e-01 -8.94880146e-02 1.12995818e-01
1.22129753e-01 4.24110651e-01 -3.34756911e-01 6.09845281e-01
5.30654848e-01 1.30930483e-01 6.03030562e-01 -2.63826728e-01
-6.35930061e-01 4.89792734e-01 -1.32197845e+00 1.59727204e+00
-7.29874492e-01 5.23462534e-01 -1.88401952e-01 -3.48484784e-01
9.20916855e-01 8.08444321e-01 8.49107932e-03 -5.50775349e-01
-1.51088685e-01 8.36490095e-01 -1.29833296e-01 -3.63076515e-02
7.83477962e-01 1.09579556e-01 -6.78544700e-01 -1.85834207e-02
5.77724755e-01 2.86030799e-01 3.70637059e-01 1.42950520e-01
1.11208153e+00 1.19029596e-01 9.40500021e-01 -5.13173401e-01
8.77613187e-01 1.81784630e-01 7.86765575e-01 9.01964903e-01
-2.38701016e-01 2.57289916e-01 -1.37300804e-01 -4.71501678e-01
-9.46139693e-01 -9.76789892e-01 -4.11428779e-01 1.23602378e+00
-3.57216984e-01 -5.75043619e-01 -5.41398585e-01 -1.19496119e+00
-5.10429323e-01 1.11484551e+00 -2.21897155e-01 6.58694208e-01
-9.45579708e-01 -6.68240011e-01 1.24473989e+00 2.63221532e-01
5.32867074e-01 -1.25473070e+00 1.80902481e-01 5.38101792e-01
-4.93729860e-01 -1.64311647e+00 -5.31043828e-01 4.73225564e-01
-3.93464506e-01 -9.63181376e-01 -7.52927899e-01 -1.10003436e+00
3.27173859e-01 -4.99234647e-01 1.38183355e+00 -4.67081696e-01
2.68808335e-01 6.89129755e-02 -4.02350068e-01 -2.77091861e-01
-6.37153447e-01 7.32032478e-01 1.21390313e-01 -2.46331975e-01
2.76214182e-01 3.16077210e-02 -2.91652810e-02 1.12991333e-01
-4.49728668e-01 -3.29787225e-01 7.11205721e-01 7.52486765e-01
6.29502773e-01 -9.99202356e-02 7.71090090e-01 -1.13456404e+00
4.10286695e-01 -4.78939146e-01 -4.46680933e-01 8.10833395e-01
-5.77999949e-01 2.62026459e-01 8.33018243e-01 -1.34410396e-01
-1.45852792e+00 2.05805391e-01 -5.36240101e-01 3.81835014e-01
-3.60678315e-01 6.22868598e-01 -4.61268961e-01 7.60944337e-02
5.31040013e-01 4.08593230e-02 -1.33945131e+00 -9.08277094e-01
5.61787188e-01 8.96690905e-01 5.81658363e-01 -6.24733090e-01
4.59280014e-01 -3.53110343e-01 -5.07174671e-01 -8.47603679e-01
-9.90412652e-01 -6.15503788e-01 -8.94946396e-01 4.39145155e-02
1.12206864e+00 -1.37448180e+00 -1.50476042e-02 6.07350290e-01
-1.52720773e+00 1.16590895e-01 1.31504864e-01 7.88051009e-01
-1.36981765e-02 5.60430065e-02 -9.69300508e-01 -7.63258040e-01
-6.06254935e-01 -7.90102899e-01 8.97432625e-01 1.53132141e-01
-1.05204865e-01 -1.27056849e+00 5.46602488e-01 3.39623541e-01
3.20106238e-01 2.48691529e-01 6.07545078e-01 -1.70625472e+00
1.23743147e-01 -1.18912779e-01 6.93991259e-02 1.53257892e-01
-7.91321397e-02 -2.46658027e-01 -7.97505498e-01 -2.34705672e-01
-5.20010293e-01 -2.16688320e-01 6.40338302e-01 -3.18435460e-01
1.82070449e-01 -1.26174718e-01 -3.87922078e-01 8.79688114e-02
1.50749886e+00 2.71091223e-01 3.28818172e-01 5.84157228e-01
6.04958713e-01 3.62064451e-01 4.56799626e-01 1.26445353e-01
7.25399017e-01 5.90512156e-01 -3.33491921e-01 1.23535572e-02
-2.24275202e-01 -3.63780290e-01 4.26548541e-01 1.50583720e+00
-3.84040400e-02 -3.04570049e-01 -1.62577426e+00 8.34062457e-01
-1.55595005e+00 -8.72206926e-01 -3.56192261e-01 1.99281728e+00
9.57005084e-01 1.71331558e-02 -3.48384947e-01 -5.31931996e-01
1.12581098e+00 4.79857530e-03 -1.61668677e-02 -3.95045161e-01
-6.68699145e-01 4.80377048e-01 7.29359150e-01 4.82088655e-01
-1.50570643e+00 1.42907012e+00 6.68256044e+00 1.04710543e+00
-7.72327304e-01 3.80690813e-01 3.22571546e-01 5.86190760e-01
3.38676460e-02 1.25596434e-01 -1.49847317e+00 1.18479379e-01
1.62101185e+00 -3.16815525e-01 6.01812871e-03 6.25005960e-01
-1.82631642e-01 1.91249996e-01 -7.78814971e-01 5.22944152e-01
3.05884443e-02 -9.92565513e-01 -5.52389622e-02 -1.82817698e-01
1.12508011e+00 8.53788435e-01 -6.23465300e-01 8.42527568e-01
9.30422068e-01 -9.12335932e-01 5.55398166e-01 6.85865164e-01
7.19927788e-01 -8.80599558e-01 1.05571783e+00 5.32135069e-01
-1.48290849e+00 2.83633918e-01 -7.76592689e-03 5.11632860e-01
4.18315798e-01 6.22634292e-01 -8.00933003e-01 1.04849291e+00
5.86445212e-01 3.48661333e-01 -5.83023906e-01 1.12314594e+00
-5.16842008e-01 6.95956349e-01 -2.78579950e-01 4.75281440e-02
1.99662417e-01 3.26878250e-01 5.53124309e-01 1.89705372e+00
2.54754364e-01 -1.54762909e-01 2.23247960e-01 -6.91030100e-02
-6.04167998e-01 9.19692099e-01 -4.87460017e-01 -9.74398032e-02
7.54953325e-01 1.31728518e+00 -5.52472055e-01 -5.96370161e-01
-4.51832861e-01 9.09995377e-01 8.23183715e-01 1.97882071e-01
-6.04206026e-01 -8.49503040e-01 7.59633556e-02 -4.44254905e-01
5.44707775e-01 -3.09289277e-01 1.04598708e-01 -1.61001551e+00
-2.31649533e-01 -9.17192459e-01 6.56473756e-01 -3.38342339e-01
-1.25099945e+00 1.32690394e+00 -2.66832441e-01 -7.73466408e-01
-2.85697341e-01 -5.57819426e-01 -2.58393019e-01 8.37792933e-01
-1.43790686e+00 -1.18196225e+00 4.37808096e-01 4.58793432e-01
3.82606715e-01 -4.67331856e-01 1.20010424e+00 7.98623681e-01
-5.42615235e-01 8.21073294e-01 4.70681489e-01 6.63191915e-01
1.15047371e+00 -1.34840953e+00 5.85866511e-01 1.12187433e+00
4.58670825e-01 7.37444818e-01 4.87599403e-01 -8.67508411e-01
-7.55475640e-01 -1.28673422e+00 2.18581057e+00 -6.16279840e-01
1.01798236e+00 -1.83520824e-01 -5.90780675e-01 8.94656539e-01
6.31953478e-01 -6.19607279e-03 8.00455272e-01 3.43349725e-01
-4.51413333e-01 3.05382848e-01 -1.21423435e+00 3.47451866e-01
8.29380035e-01 -6.98909044e-01 -1.06006598e+00 2.04805449e-01
7.01786399e-01 -2.31942281e-01 -1.57943165e+00 5.45218587e-01
3.58624876e-01 -3.48909289e-01 5.65258026e-01 -8.53570104e-01
-1.81663737e-01 -4.34640646e-01 -5.30535460e-01 -1.48960757e+00
-1.62001893e-01 -3.01898658e-01 2.08025277e-01 1.67471921e+00
1.13688743e+00 -4.91292626e-01 -7.16948509e-02 3.55785340e-01
-1.41115710e-01 -4.08026017e-02 -9.68642771e-01 -7.38911510e-01
2.64539003e-01 -4.95160937e-01 2.68089384e-01 1.25540268e+00
-2.08319768e-01 7.96848655e-01 -4.09519166e-01 5.17309129e-01
6.46472931e-01 -3.41678709e-01 2.33222604e-01 -1.10578382e+00
8.62129033e-02 -5.72580248e-02 -3.17854643e-01 -5.94651759e-01
5.51974893e-01 -1.09503925e+00 2.02879950e-01 -1.44064951e+00
2.09960267e-01 -3.96131545e-01 -7.18653500e-01 6.75851941e-01
-3.57084602e-01 1.61446363e-01 4.74450231e-01 3.33396673e-01
-1.08929455e+00 2.07176179e-01 4.27783787e-01 7.85700604e-03
-5.73746637e-02 -1.65479034e-01 -4.78722274e-01 6.90207779e-01
6.51052117e-01 -8.54198515e-01 6.41364396e-01 -5.23313046e-01
2.14921907e-01 1.39944166e-01 -4.65049684e-01 -1.19066536e+00
4.78331596e-01 2.43890315e-01 3.86172980e-01 -8.64996910e-01
-2.93733865e-01 -7.10626662e-01 3.56998265e-01 3.79289806e-01
-3.60008150e-01 3.81609261e-01 2.48972759e-01 1.82113856e-01
-5.14096141e-01 -1.43511817e-01 6.27189815e-01 3.52013931e-02
-1.06829214e+00 -8.15738663e-02 -5.04350185e-01 4.57358092e-01
8.07519138e-01 6.98665619e-01 -1.79505885e-01 2.80331939e-01
-1.28031361e+00 2.85403490e-01 1.35232314e-01 6.09646022e-01
2.67014652e-02 -1.42394257e+00 -1.20832813e+00 -3.56552094e-01
4.06632304e-01 -6.14207804e-01 -1.41259879e-02 4.98615593e-01
-5.30448079e-01 8.90352368e-01 1.02417758e-02 -3.85868326e-02
-1.00416255e+00 1.45043373e-01 5.01213908e-01 -1.10219646e+00
-1.18902519e-01 4.66067880e-01 -1.96526080e-01 -1.29068160e+00
-1.80008225e-02 2.16996655e-01 -7.48211682e-01 1.03447080e-01
6.20882392e-01 5.56256533e-01 4.40166891e-01 -1.24309206e+00
-8.37490141e-01 4.00835305e-01 -2.47678116e-01 -4.11151439e-01
1.42156279e+00 -1.79067776e-01 -1.35682561e-02 5.72060466e-01
1.16324079e+00 8.33709836e-01 -1.68260530e-01 -3.31224591e-01
9.63529468e-01 4.36925024e-01 -7.03957528e-02 -1.37196600e+00
-6.86795056e-01 3.85322094e-01 5.05329669e-01 -2.14884922e-01
7.64276743e-01 -2.57189739e-02 6.05772436e-01 9.29809570e-01
7.07325101e-01 -1.24525177e+00 -8.63186240e-01 1.22596204e+00
6.12366617e-01 -1.35184479e+00 -3.55541795e-01 -1.07883401e-01
-7.23835528e-01 1.06745481e+00 3.75878960e-01 8.97393748e-02
8.03972423e-01 3.25107425e-01 3.73432487e-01 1.17981201e-02
-6.72132254e-01 -2.27530599e-01 6.01140201e-01 2.57837951e-01
9.65882361e-01 3.51536751e-01 -9.29857969e-01 8.23263109e-01
-1.75042413e-02 -2.31570154e-01 2.12471306e-01 7.27109969e-01
-2.97159076e-01 -1.36609018e+00 -2.71989107e-01 -1.03565030e-01
-1.24338257e+00 -6.13468111e-01 -2.95985579e-01 9.81971383e-01
1.67269960e-01 1.26052797e+00 -4.76111710e-01 -2.98954546e-01
7.53822863e-01 4.62802261e-01 -3.37005258e-02 -6.14709854e-01
-1.03422379e+00 -1.92188080e-02 8.81503880e-01 -3.28019798e-01
-5.16719282e-01 -7.71739602e-01 -1.25834072e+00 4.98799793e-02
-3.35063308e-01 8.82466078e-01 8.90049815e-01 9.35738087e-01
3.43831658e-01 1.54126599e-01 6.70704484e-01 -2.56726891e-01
-3.95964295e-01 -1.21761560e+00 -3.95851374e-01 2.10757419e-01
-1.52571350e-01 -2.08078250e-01 -3.28074604e-01 -1.35649312e-02] | [9.829479217529297, 9.750079154968262] |
f27b8599-664e-4204-ae06-f66a28a4a881 | a-generalised-deep-meta-learning-model-for | 2303.13324 | null | https://arxiv.org/abs/2303.13324v1 | https://arxiv.org/pdf/2303.13324v1.pdf | A Generalised Deep Meta-Learning Model for Automated Quality Control of Cardiovascular Magnetic Resonance Images | Background and Objectives: Cardiovascular magnetic resonance (CMR) imaging is a powerful modality in functional and anatomical assessment for various cardiovascular diseases. Sufficient image quality is essential to achieve proper diagnosis and treatment. A large number of medical images, the variety of imaging artefacts, and the workload of imaging centres are among the things that reveal the necessity of automatic image quality assessment (IQA). However, automated IQA requires access to bulk annotated datasets for training deep learning (DL) models. Labelling medical images is a tedious, costly and time-consuming process, which creates a fundamental challenge in proposing DL-based methods for medical applications. This study aims to present a new method for CMR IQA when there is limited access to annotated datasets. Methods: The proposed generalised deep meta-learning model can evaluate the quality by learning tasks in the prior stage and then fine-tuning the resulting model on a small labelled dataset of the desired tasks. This model was evaluated on the data of over 6,000 subjects from the UK Biobank for five defined tasks, including detecting respiratory motion, cardiac motion, Aliasing and Gibbs ringing artefacts and images without artefacts. Results: The results of extensive experiments show the superiority of the proposed model. Besides, comparing the model's accuracy with the domain adaptation model indicates a significant difference by using only 64 annotated images related to the desired tasks. Conclusion: The proposed model can identify unknown artefacts in images with acceptable accuracy, which makes it suitable for medical applications and quality assessment of large cohorts. | ['Alejandro F. Frangi', 'Ahmad Ali Abin', 'Mohsen Ebrahimi Moghaddam', 'Hossein Simchi', 'Shahabedin Nabavi'] | 2023-03-23 | null | null | null | null | ['image-quality-assessment'] | ['computer-vision'] | [ 3.25328410e-01 -4.98393215e-02 5.48401177e-02 -3.42847168e-01
-1.12296295e+00 -3.72462600e-01 3.57641906e-01 1.13708571e-01
-6.36497021e-01 8.46445978e-01 1.58550054e-01 -1.87809870e-01
-3.88815701e-01 -2.91365266e-01 -2.58999556e-01 -7.90392339e-01
-2.90557027e-01 6.50735915e-01 2.69080192e-01 3.18017960e-01
2.58037388e-01 5.28507710e-01 -9.65645909e-01 5.02034485e-01
7.86901712e-01 9.51652706e-01 5.69505274e-01 8.44174743e-01
1.65571809e-01 1.09281218e+00 -6.71785116e-01 -2.19029590e-01
5.38337082e-02 -7.40865648e-01 -1.14361370e+00 1.96038708e-01
2.50195771e-01 -2.84955353e-01 1.29663544e-02 7.83836365e-01
1.34604526e+00 1.95409164e-01 6.45974278e-01 -6.25738680e-01
-2.92197973e-01 4.02934551e-01 -2.62445867e-01 1.09515190e+00
-1.47230849e-01 4.11290944e-01 3.72816026e-01 -6.67722404e-01
4.95718330e-01 6.89589560e-01 8.62964988e-01 5.89042902e-01
-1.06884873e+00 -3.53392899e-01 -5.81854105e-01 4.21562761e-01
-1.05342901e+00 -5.30809224e-01 5.84344447e-01 -7.02575862e-01
8.51864219e-01 2.13903666e-01 4.24416691e-01 7.66464472e-01
3.22594374e-01 3.88426483e-01 1.36641490e+00 -2.37022042e-01
2.22454622e-01 1.17409766e-01 -1.63873389e-01 4.74576473e-01
4.58746254e-02 1.83929190e-01 -1.14285387e-01 -1.70286551e-01
7.95243323e-01 -3.94307524e-01 -2.82045692e-01 -3.60441893e-01
-1.42366338e+00 6.66157842e-01 4.11224276e-01 7.20176220e-01
-5.77597082e-01 2.26312540e-02 8.30557406e-01 2.07911581e-01
4.73174304e-01 4.51637119e-01 -4.90924060e-01 -1.77995101e-01
-1.30100095e+00 -1.32588625e-01 4.20728207e-01 2.66157001e-01
3.49828973e-02 2.87165511e-02 -3.15962583e-01 1.08473611e+00
1.67176992e-01 3.31308722e-01 1.04529369e+00 -1.10922444e+00
2.45747775e-01 2.32061043e-01 -3.84886153e-02 -8.53302181e-01
-8.04128051e-01 -7.60899901e-01 -1.26488781e+00 1.51650831e-01
6.21290803e-01 -1.34765990e-02 -8.83319855e-01 1.42534137e+00
2.97129542e-01 1.29251644e-01 -1.44516662e-01 1.13379133e+00
9.68802452e-01 3.30057502e-01 3.61623704e-01 -4.60242659e-01
1.42812967e+00 -7.32996941e-01 -7.24938989e-01 -2.24872708e-01
6.17784202e-01 -7.94104755e-01 9.11914945e-01 3.87696683e-01
-1.30158889e+00 -8.93397450e-01 -8.48070741e-01 2.29366228e-01
1.40909091e-01 3.63029897e-01 4.68972206e-01 9.02164280e-01
-1.10048199e+00 7.21877456e-01 -9.65425193e-01 -6.51158020e-02
6.42512619e-01 3.91853988e-01 -3.78644437e-01 -2.61156738e-01
-1.30853367e+00 1.22464359e+00 4.65974629e-01 4.25668985e-01
-9.08642292e-01 -9.36493099e-01 -5.09563208e-01 -2.97494441e-01
1.13686480e-01 -8.36970448e-01 1.00306463e+00 -9.55668449e-01
-1.27498841e+00 1.21003568e+00 1.62994921e-01 -5.83583772e-01
8.82727802e-01 3.12338513e-03 -4.81214404e-01 4.41072375e-01
1.15456030e-01 3.47928137e-01 9.08754587e-01 -7.66814888e-01
-3.10541511e-01 -4.55505371e-01 -4.48007494e-01 1.22253522e-01
3.23818475e-01 5.93298525e-02 -1.93134516e-01 -7.27972209e-01
4.26881053e-02 -7.93213725e-01 -3.29530776e-01 -2.99850613e-01
4.47564274e-02 3.14207487e-02 1.97732583e-01 -1.14769959e+00
1.19908643e+00 -2.15621877e+00 -4.68202820e-03 7.45735019e-02
3.66614878e-01 4.33918983e-01 4.93744984e-02 -2.86594480e-01
-2.83080399e-01 1.37316287e-01 -4.76337314e-01 2.03092337e-01
-3.03010553e-01 -1.45246116e-02 5.96995294e-01 6.72261953e-01
2.07577378e-01 8.44040811e-01 -7.32151926e-01 -7.77405560e-01
5.29703736e-01 5.65787017e-01 -3.30123842e-01 3.97242129e-01
2.81375051e-01 1.25191426e+00 -2.39868715e-01 2.26673201e-01
5.46638429e-01 -4.47553217e-01 1.90734074e-01 -4.74586725e-01
2.17916250e-01 1.92228347e-01 -1.21234858e+00 1.76938295e+00
-5.71527362e-01 3.22145641e-01 -1.68454006e-01 -1.39625907e+00
6.77398801e-01 7.98948526e-01 7.93052375e-01 -1.07019055e+00
1.74290568e-01 4.87410694e-01 4.38214600e-01 -8.99627686e-01
-3.09858501e-01 -6.33602142e-01 2.22450048e-01 3.17093730e-01
2.99717098e-01 -1.57052446e-02 9.38205570e-02 -4.36496437e-02
1.15155602e+00 2.02093478e-02 3.67879629e-01 -3.71394992e-01
8.51202846e-01 -1.54359818e-01 5.35483181e-01 8.54489386e-01
-7.34591842e-01 7.66293883e-01 6.33923113e-02 -8.17328990e-01
-1.15795577e+00 -8.19847286e-01 -5.04989207e-01 5.52395880e-01
-4.18124914e-01 2.18930483e-01 -6.59163713e-01 -5.14585078e-01
-5.48843086e-01 2.45474502e-01 -5.65321922e-01 -1.92245424e-01
-7.99282134e-01 -1.40416026e+00 4.08910006e-01 3.96763563e-01
6.63394332e-01 -1.41448426e+00 -1.13618720e+00 4.76849735e-01
-7.02956855e-01 -1.16425204e+00 -2.92025954e-01 9.66332573e-03
-1.32316065e+00 -1.07602048e+00 -1.26051283e+00 -5.65679312e-01
2.10986719e-01 -3.09156716e-01 1.56151378e+00 7.67723843e-02
-7.59996593e-01 1.50346041e-01 -1.74450129e-01 -1.27466992e-01
-6.61472023e-01 1.00735232e-01 -3.79725039e-01 -1.62065610e-01
-7.93716460e-02 -3.96726310e-01 -1.07289660e+00 2.06362054e-01
-8.59364569e-01 -2.36336559e-01 9.02801633e-01 1.03835809e+00
5.01484990e-01 1.87299788e-01 8.58887553e-01 -1.00051963e+00
4.84820843e-01 -2.48884588e-01 -3.10251921e-01 1.82253480e-01
-5.44010520e-01 -2.84184720e-02 1.69342458e-01 -2.37400129e-01
-1.06734836e+00 -3.61684188e-02 -2.93817282e-01 -3.88565697e-02
-2.37896666e-01 4.42430317e-01 2.85156444e-02 -3.71382385e-02
9.61241841e-01 8.73050541e-02 2.66332533e-02 -2.82833219e-01
8.53537619e-02 4.34301704e-01 6.81065977e-01 -3.79532963e-01
1.19981706e-01 3.39601457e-01 3.18038940e-01 -7.16631472e-01
-8.86316240e-01 -6.07599497e-01 -9.74626839e-01 -2.91637748e-01
1.02224290e+00 -7.43075669e-01 -2.57439047e-01 4.25206959e-01
-8.57984126e-01 -3.49823713e-01 -2.61528611e-01 6.79365337e-01
-5.89213789e-01 6.00248098e-01 -6.21778667e-01 -7.44657636e-01
-7.16512501e-01 -1.43965471e+00 7.29186475e-01 -1.12730920e-01
-1.93009987e-01 -1.20171118e+00 2.80343682e-01 6.82721496e-01
7.35070646e-01 5.14920592e-01 9.32092309e-01 -4.84522641e-01
-2.90424526e-01 -6.80715889e-02 -2.49482542e-01 6.00002170e-01
1.78763613e-01 -7.03771174e-01 -1.00424802e+00 -2.33930632e-01
3.79213661e-01 -3.68439972e-01 7.55530894e-01 9.94839728e-01
1.09317386e+00 1.29352897e-01 2.78301090e-01 5.08641124e-01
1.32628167e+00 3.20514530e-01 8.50365222e-01 4.50183600e-01
4.09862101e-01 4.26283538e-01 3.98438513e-01 3.65680456e-01
1.03678346e-01 5.14227569e-01 2.49189094e-01 -4.04020995e-01
-3.63047272e-01 6.09801948e-01 -1.57242700e-01 7.51432240e-01
-4.29397821e-01 2.04638720e-01 -1.03860569e+00 7.19345868e-01
-1.49484777e+00 -1.01746678e+00 -3.37025434e-01 2.26337886e+00
9.44761336e-01 7.84793273e-02 2.06949189e-01 4.23997700e-01
7.02764452e-01 -5.54562993e-02 -4.64559793e-01 -2.78248101e-01
6.38324618e-02 6.58327341e-01 3.32496464e-01 2.13671848e-01
-1.23725784e+00 3.06433052e-01 6.12061167e+00 5.74267507e-01
-9.80951905e-01 6.80363834e-01 1.01158321e+00 1.82480082e-01
2.74624646e-01 -4.20511901e-01 -1.64539695e-01 4.94195402e-01
1.37835991e+00 3.70552778e-01 2.33720243e-01 3.86105120e-01
5.70715368e-01 -2.72681981e-01 -7.96501100e-01 1.08503687e+00
8.14044401e-02 -1.16479874e+00 -3.31740916e-01 -2.40524948e-01
5.67661405e-01 1.65536046e-01 -2.51712501e-02 1.22302935e-01
-2.61516094e-01 -1.16865480e+00 3.14827889e-01 7.40241826e-01
9.41328228e-01 -6.21360600e-01 1.24173415e+00 3.15297097e-01
-6.52969301e-01 4.91300784e-02 -3.82949263e-01 4.72938299e-01
1.90984219e-01 1.02288151e+00 -8.48025203e-01 6.27651572e-01
7.36026764e-01 3.77593338e-01 -6.32802069e-01 1.22950196e+00
1.69981211e-01 6.79767013e-01 1.10764183e-01 5.79877019e-01
8.67048725e-02 6.31701201e-02 2.94884115e-01 1.20885742e+00
3.26997608e-01 1.46668687e-01 -7.05605969e-02 6.62905455e-01
9.08897594e-02 2.70942837e-01 -2.27957264e-01 1.73868030e-01
-8.80144462e-02 1.38723183e+00 -9.02238250e-01 -5.32208383e-01
-3.37078869e-01 1.02443767e+00 1.20899931e-03 -1.15949869e-01
-8.15430403e-01 4.06683721e-02 -3.35428417e-01 3.57876778e-01
-8.40297937e-02 1.54739752e-01 -2.27595910e-01 -8.91880691e-01
-1.88151762e-01 -1.10690987e+00 7.20329225e-01 -7.03709960e-01
-1.29786587e+00 8.05076778e-01 -1.36386082e-01 -9.83865082e-01
-2.48729274e-01 -3.41538042e-01 -1.79794997e-01 1.21476054e+00
-1.64220667e+00 -7.79388249e-01 -4.19070005e-01 5.79034150e-01
5.06192088e-01 -4.16149870e-02 8.36865604e-01 8.20806265e-01
-1.93222076e-01 1.96282417e-01 -2.45365441e-01 1.19183436e-01
8.77069652e-01 -1.41471183e+00 -2.47524846e-02 6.45649314e-01
-1.62526861e-01 2.96027154e-01 3.75431150e-01 -5.03668904e-01
-6.95861816e-01 -1.02332127e+00 8.23152244e-01 -4.14567053e-01
1.58097953e-01 4.54309821e-01 -1.02309048e+00 1.63893417e-01
1.12013258e-01 5.28062403e-01 7.80299902e-01 -3.93452674e-01
2.69213766e-01 -1.57855339e-02 -1.38576508e+00 -1.99359462e-01
6.78229332e-01 -3.62320185e-01 -6.76426947e-01 3.34054559e-01
2.92840954e-02 -4.71441865e-01 -1.25508463e+00 6.49457335e-01
5.29855728e-01 -1.08943892e+00 1.10272789e+00 -5.14103293e-01
2.58467764e-01 -2.00672939e-01 1.44655213e-01 -1.17230296e+00
-4.09455717e-01 -2.14813769e-01 1.51921406e-01 8.01215827e-01
2.97505498e-01 -3.11849385e-01 5.72542131e-01 5.60793519e-01
-1.28361881e-01 -3.29143137e-01 -1.08510232e+00 -3.67830455e-01
1.53372765e-01 -4.69545484e-01 1.12584054e-01 1.09330595e+00
-5.29581428e-01 2.11574078e-01 -3.62435132e-01 7.51543567e-02
8.17314804e-01 -2.90611774e-01 1.71502367e-01 -1.31971371e+00
-5.18078446e-01 -2.58017749e-01 -4.76654053e-01 -1.33295760e-01
-1.56959683e-01 -1.02985632e+00 -1.16578273e-01 -1.63511610e+00
5.05856037e-01 -6.18133068e-01 -6.37857676e-01 -1.16711333e-02
-3.43620718e-01 6.31947696e-01 -5.53298593e-02 3.14104736e-01
-4.21790153e-01 3.29524018e-02 1.53502452e+00 -9.36283171e-02
-6.88751414e-02 2.82431751e-01 -3.28975141e-01 4.76688951e-01
8.71412337e-01 -5.25094211e-01 -4.46033388e-01 -2.71744907e-01
2.13075191e-01 5.41515708e-01 6.13315701e-01 -1.24474871e+00
-2.29528889e-01 3.76824826e-01 6.96741700e-01 -4.01110709e-01
-1.18805893e-01 -8.13711524e-01 4.28006470e-01 8.48461628e-01
-4.41824764e-01 2.05555514e-01 1.54072821e-01 2.39593178e-01
-2.22083136e-01 -3.84756595e-01 1.35026717e+00 -8.08023453e-01
-5.76143205e-01 2.43495151e-01 -2.64177918e-01 4.73462135e-01
6.87522948e-01 -6.30002841e-02 2.20351338e-01 -1.71782047e-01
-1.42896199e+00 -2.57449836e-01 -2.64646411e-01 9.50366333e-02
5.53134918e-01 -1.11481643e+00 -9.35954034e-01 -1.48156866e-01
-1.55143052e-01 1.85784940e-02 7.63310194e-01 1.41980958e+00
-6.92402124e-01 5.65207303e-01 -3.82586181e-01 -9.46627915e-01
-1.05658185e+00 5.11403799e-01 8.85738552e-01 -8.13026071e-01
-8.34292412e-01 6.37292922e-01 7.19090402e-02 -3.68504137e-01
-6.28855228e-02 -3.16645384e-01 -3.70506167e-01 -4.29984517e-02
5.89487851e-01 4.33358490e-01 6.65453196e-01 -6.52945518e-01
-3.80006641e-01 4.60952818e-01 1.85859427e-01 -1.16200693e-01
1.28232861e+00 -2.73938626e-01 -8.43366534e-02 3.79122168e-01
1.02523577e+00 -5.30676305e-01 -8.93522322e-01 -1.58310637e-01
2.23658249e-01 -2.76495010e-01 4.75381464e-01 -1.21310198e+00
-1.30235255e+00 1.28799152e+00 1.46407962e+00 -1.49612173e-01
1.27266312e+00 -1.72436461e-01 5.44443488e-01 -2.22065628e-01
1.46322370e-01 -9.73330855e-01 1.52235776e-01 -8.98318738e-02
7.58291304e-01 -1.42203057e+00 -2.14536916e-02 -5.76506043e-03
-8.92506957e-01 1.00703526e+00 1.26051813e-01 1.35955930e-01
6.11241281e-01 -2.62826569e-02 2.79319346e-01 -3.53860021e-01
-3.24719250e-01 1.69221126e-02 3.79655600e-01 7.05325007e-01
7.36592174e-01 -9.80334356e-02 -5.39617538e-01 3.96160394e-01
3.66968036e-01 3.48417163e-01 3.38879913e-01 7.26385057e-01
-1.98004276e-01 -9.93946016e-01 -3.94633353e-01 5.13670385e-01
-1.17861700e+00 -2.38150284e-01 2.38842741e-01 6.77101970e-01
2.83033878e-01 7.50041306e-01 -4.00321364e-01 3.51575077e-01
3.52777690e-01 3.32579166e-01 6.37176514e-01 -4.87746119e-01
-7.98157036e-01 2.97361702e-01 1.85895376e-02 -4.66738671e-01
-1.04174399e+00 -6.05367959e-01 -1.03030157e+00 2.07988188e-01
-2.64726341e-01 5.49241193e-02 5.97691953e-01 9.94087696e-01
1.72604918e-01 1.06321549e+00 5.30728281e-01 -6.03473604e-01
-5.17203152e-01 -1.19600570e+00 -4.93424833e-01 7.31368840e-01
2.23037422e-01 -6.24523342e-01 -2.84556448e-01 3.90395522e-01] | [14.15661334991455, -2.4412853717803955] |
a2d653ac-3032-430c-b8ac-d6b11350333e | unsupervised-high-impedance-fault-detection | 2301.01867 | null | https://arxiv.org/abs/2301.01867v1 | https://arxiv.org/pdf/2301.01867v1.pdf | Unsupervised High Impedance Fault Detection Using Autoencoder and Principal Component Analysis | Detection of high impedance faults (HIF) has been one of the biggest challenges in the power distribution network. The low current magnitude and diverse characteristics of HIFs make them difficult to be detected by over-current relays. Recently, data-driven methods based on machine learning models are gaining popularity in HIF detection due to their capability to learn complex patterns from data. Most machine learning-based detection methods adopt supervised learning techniques to distinguish HIFs from normal load conditions by performing classifications, which rely on a large amount of data collected during HIF. However, measurements of HIF are difficult to acquire in the real world. As a result, the reliability and generalization of the classification methods are limited when the load profiles and faults are not present in the training data. Consequently, this paper proposes an unsupervised HIF detection framework using the autoencoder and principal component analysis-based monitoring techniques. The proposed fault detection method detects the HIF by monitoring the changes in correlation structure within the current waveforms that are different from the normal loads. The performance of the proposed HIF detection method is tested using real data collected from a 4.16 kV distribution system and compared with results from a commercially available solution for HIF detection. The numerical results demonstrate that the proposed method outperforms the commercially available HIF detection technique while maintaining high security by not falsely detecting during load conditions. | ['James Stoupis', 'Mohammad Razeghi-Jahromi', 'Yingxiang Liu'] | 2023-01-05 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [-1.80048034e-01 -4.91104603e-01 1.85366675e-01 -2.12239027e-01
-3.19694728e-01 -3.55350733e-01 2.94326156e-01 3.77482086e-01
2.69411653e-01 6.85866833e-01 -1.08100891e-01 1.16523961e-03
-6.59172595e-01 -8.57345343e-01 -1.33573279e-01 -1.20065057e+00
-5.03091514e-01 3.63510638e-01 8.16518441e-02 -3.86394322e-01
3.90960187e-01 8.19117248e-01 -1.85087132e+00 1.39303878e-01
1.04299486e+00 9.61985826e-01 3.48474741e-01 3.02529931e-01
2.77238369e-01 8.04401278e-01 -1.36450410e+00 5.30510545e-01
-8.44833627e-02 -3.41992944e-01 -1.86532259e-01 9.15913135e-02
-5.59326291e-01 -3.21090817e-01 -5.57903290e-01 1.01257002e+00
6.92324579e-01 1.67479128e-01 9.53314841e-01 -1.45585275e+00
-5.30006737e-02 4.62477356e-01 -4.84730333e-01 6.95565283e-01
6.73144042e-01 -3.17066252e-01 5.31430542e-01 -6.30275726e-01
-3.33734453e-02 6.68905377e-01 6.07193947e-01 -1.17603831e-01
-8.35384369e-01 -3.82978618e-01 -3.70994389e-01 9.04222727e-01
-1.39555085e+00 1.12993762e-01 1.36365020e+00 -5.33632934e-01
1.10808551e+00 3.35466623e-01 5.27609408e-01 6.11442387e-01
4.29261297e-01 6.42364979e-01 1.06571472e+00 -5.03973603e-01
5.31817377e-01 1.98500365e-01 1.65344924e-01 1.82137012e-01
2.47865647e-01 3.03947423e-02 -1.10993035e-01 -2.01271579e-01
1.67714864e-01 -3.42730917e-02 -7.61369824e-01 3.62515077e-02
-4.61928099e-01 8.86878848e-01 3.23780566e-01 1.03248322e+00
-5.74627876e-01 -7.97722518e-01 4.66354519e-01 3.22338402e-01
3.49241734e-01 3.57996404e-01 -2.72723883e-01 -2.61654258e-01
-9.01871681e-01 -1.32037446e-01 9.21878695e-01 5.74375987e-01
3.53185028e-01 7.50309944e-01 4.56352055e-01 7.90591240e-01
8.31089616e-02 6.07929587e-01 9.24560308e-01 -7.60568976e-02
1.81188866e-01 7.47951269e-01 1.03175193e-02 -1.37753546e+00
-6.75891995e-01 -7.26096570e-01 -8.37052643e-01 2.53582627e-01
-6.30509779e-02 -5.11193275e-01 -2.70658374e-01 1.00968504e+00
1.87427193e-01 -3.63880433e-02 3.17904949e-01 7.58088112e-01
5.23329437e-01 1.07516670e+00 -4.66109961e-01 -4.30095166e-01
9.00269330e-01 -2.41390258e-01 -1.12576640e+00 1.95141077e-01
5.29417098e-01 -6.14479482e-01 4.19441223e-01 8.14149559e-01
-5.03704488e-01 -5.67668080e-01 -1.65429938e+00 9.29958463e-01
-3.91911447e-01 3.87840569e-01 3.22150320e-01 6.55261993e-01
-5.02733648e-01 7.24975765e-01 -6.28251970e-01 -3.35834503e-01
1.80065483e-02 2.11791709e-01 -2.60891229e-01 2.50297248e-01
-1.48289776e+00 1.23366106e+00 5.94674528e-01 4.95248497e-01
-7.62646556e-01 -4.65892553e-01 -6.85035050e-01 2.25422710e-01
-5.59082348e-03 4.59420532e-01 8.42306912e-01 -9.04910922e-01
-1.28525090e+00 -1.14313573e-01 2.66748130e-01 -4.71390486e-01
2.50480354e-01 -1.37748383e-02 -1.13253617e+00 6.43914640e-01
-2.65637130e-01 -6.78401351e-01 9.20488060e-01 -1.06256795e+00
-6.40756786e-01 -1.43866941e-01 -3.83147418e-01 1.79856701e-03
-5.37660956e-01 -1.62797749e-01 6.72604978e-01 -3.49477828e-01
1.86548695e-01 -2.24431247e-01 1.95132568e-01 -6.76871777e-01
-3.85191023e-01 -3.24742705e-01 1.58545923e+00 -8.71136785e-01
1.11950648e+00 -1.93034911e+00 -1.21301711e-01 5.61210632e-01
-1.45629734e-01 3.90667796e-01 5.17786682e-01 1.00440454e+00
-2.98350066e-01 -6.00792885e-01 -2.24494472e-01 5.25311351e-01
-2.41227850e-01 3.42617750e-01 -5.05807996e-02 9.20328796e-01
2.47222960e-01 1.00226521e-01 -6.93083704e-01 -1.15098683e-02
6.74316287e-01 4.76253986e-01 -6.58384934e-02 7.00510800e-01
2.90091842e-01 4.01687056e-01 -4.47464615e-01 4.50049996e-01
6.42475605e-01 2.35060886e-01 8.90097171e-02 -6.99067235e-01
-8.13152418e-02 -9.14105177e-02 -1.36754704e+00 8.90106916e-01
-3.30474347e-01 6.16188288e-01 -7.00121522e-02 -1.94871795e+00
1.32751608e+00 7.68008947e-01 8.64930451e-01 -6.32008493e-01
4.04799074e-01 1.18703090e-01 2.75706708e-01 -1.01026487e+00
-1.12604387e-01 3.68173569e-02 1.21882930e-01 3.60393941e-01
3.36195976e-01 -6.48192465e-02 3.86244237e-01 -4.44154926e-02
1.01541328e+00 -4.40685362e-01 2.49968499e-01 -5.05142570e-01
8.82616997e-01 -1.75410435e-01 8.88201714e-01 3.67015153e-01
9.84443575e-02 1.23420902e-01 3.96865129e-01 -1.89177722e-01
-7.95318961e-01 -9.62040186e-01 -5.69920659e-01 1.13984138e-01
6.45404980e-02 7.67696425e-02 -6.10929072e-01 -3.68684560e-01
1.39423862e-01 7.00522661e-01 -3.35932255e-01 -6.26312971e-01
-4.27024156e-01 -1.26256037e+00 3.79629433e-01 5.23496032e-01
4.52928782e-01 -9.00131047e-01 -8.21822226e-01 4.79804128e-01
7.83350393e-02 -8.92484128e-01 6.21229410e-01 4.18011516e-01
-6.19903684e-01 -1.30018258e+00 -4.82650250e-01 -8.18288863e-01
7.10414648e-01 -1.07379593e-01 6.09388769e-01 2.98932523e-01
-6.48901224e-01 1.38156191e-01 -6.73406363e-01 -4.14164335e-01
-4.69850540e-01 -1.57544240e-01 2.47643009e-01 1.46825030e-01
2.66460240e-01 -8.85128498e-01 -2.82825053e-01 3.38214427e-01
-7.12723494e-01 -6.13947093e-01 6.52526319e-01 9.72787797e-01
-2.33806223e-01 1.24173224e+00 1.23098600e+00 -4.87913638e-01
7.41142631e-01 -1.04046726e+00 -7.14848280e-01 -2.77589317e-02
-7.30610371e-01 -3.16666454e-01 1.05742896e+00 -3.62558544e-01
-1.19643831e+00 -4.85150814e-01 -2.46593654e-01 -9.66324434e-02
-5.75221658e-01 7.33287513e-01 -5.78353524e-01 -1.36665374e-01
5.77212036e-01 2.76073724e-01 1.81435913e-01 -5.76988459e-01
-4.08286989e-01 1.01398456e+00 5.49551904e-01 -3.74516726e-01
9.94415760e-01 1.44051120e-01 -2.75342483e-02 -1.09658194e+00
-3.49823296e-01 -6.06221855e-01 -5.28824449e-01 -5.65652311e-01
3.75348479e-01 -7.68141091e-01 -6.89081013e-01 8.48637879e-01
-7.27143347e-01 2.79644251e-01 7.32882097e-02 6.88936055e-01
-1.50883391e-01 5.79881072e-01 -9.25942183e-01 -1.00418115e+00
-4.33969527e-01 -1.06857002e+00 3.30079466e-01 4.40209746e-01
9.31331590e-02 -1.06739509e+00 -2.04111502e-01 -6.09255470e-02
6.57384098e-01 4.52906758e-01 1.10821390e+00 -8.29826415e-01
1.30692363e-01 -7.97248721e-01 4.11304593e-01 7.50271797e-01
5.17014921e-01 1.54655561e-01 -8.89780998e-01 -4.94669020e-01
7.00552166e-01 -2.38711853e-02 -1.18119188e-01 -5.43015860e-02
8.03339183e-01 -1.53666928e-01 -1.68250337e-01 2.00697467e-01
1.70229220e+00 9.21612918e-01 5.54881573e-01 2.97094315e-01
4.00401682e-01 4.82188404e-01 5.93960643e-01 7.88577080e-01
-8.02781805e-02 2.99650311e-01 3.16394687e-01 -1.50005236e-01
5.31327546e-01 1.01427436e-01 6.91530332e-02 1.25914884e+00
1.19121499e-01 -2.15875313e-01 -8.25690329e-01 4.61889744e-01
-1.61929178e+00 -1.07693291e+00 -3.06894660e-01 1.83976793e+00
3.96925330e-01 2.61913419e-01 -4.89032678e-02 1.46854794e+00
9.02206182e-01 -1.06431819e-01 -4.11343724e-01 -4.94519681e-01
-4.22607988e-01 1.79791778e-01 1.25637084e-01 1.85445055e-01
-9.88234878e-01 -1.37503043e-01 5.42688465e+00 6.83685005e-01
-1.32713664e+00 -1.51095256e-01 1.47073627e-01 3.16979527e-01
1.84348717e-01 -3.23503882e-01 -2.02115774e-01 6.77914798e-01
1.02442563e+00 -3.59883845e-01 -5.72261252e-02 9.26015258e-01
4.12434459e-01 -4.69742209e-01 -8.03508401e-01 9.77176547e-01
3.23089123e-01 -6.59101248e-01 -2.26134717e-01 -3.87538016e-01
5.52676201e-01 -2.50372350e-01 -3.66086274e-01 8.49483684e-02
-4.70736116e-01 -7.18843222e-01 3.07631671e-01 4.59903091e-01
5.59715964e-02 -1.10068285e+00 1.28108466e+00 5.21800280e-01
-8.99312973e-01 -6.17084146e-01 -3.46470296e-01 -3.30505699e-01
3.63494247e-01 1.19876909e+00 -7.48750210e-01 9.21629906e-01
7.56157339e-01 8.39368641e-01 -3.47929567e-01 1.29479647e+00
-4.21277583e-01 1.01823103e+00 -4.08941567e-01 -1.14989765e-01
-2.16618389e-01 -2.23170578e-01 5.21532357e-01 8.66404712e-01
4.04044628e-01 -6.03027716e-02 1.44104704e-01 3.42531949e-01
5.12306273e-01 7.28713647e-02 -4.90454018e-01 2.65318364e-01
6.90456390e-01 1.27116227e+00 -5.02944946e-01 -2.75377005e-01
-3.51554155e-01 4.78330612e-01 -1.97400585e-01 3.02409738e-01
-6.77699387e-01 -7.32107341e-01 2.31378466e-01 -4.41372246e-02
8.13823715e-02 4.18596864e-02 2.36369148e-02 -9.07959580e-01
1.48280546e-01 -7.03383923e-01 5.33548713e-01 -8.12756538e-01
-1.53472066e+00 7.12819159e-01 2.57166535e-01 -1.54449952e+00
-5.18989623e-01 -4.21639919e-01 -1.03091431e+00 8.02178860e-01
-1.25820649e+00 -4.44910079e-01 -3.65803063e-01 7.02919304e-01
4.51362789e-01 -4.15884525e-01 9.06765103e-01 6.52260780e-01
-7.22391665e-01 3.41002941e-01 4.71747011e-01 2.71113902e-01
2.30733007e-02 -1.17354417e+00 -6.14466608e-01 1.01644325e+00
-2.41462067e-01 -2.18013022e-02 8.37193668e-01 -5.16051233e-01
-1.55604005e+00 -5.01219630e-01 3.18021148e-01 5.90267599e-01
6.76015258e-01 -3.05247337e-01 -1.30520201e+00 2.79084355e-01
4.24925834e-01 1.03564914e-02 5.72322488e-01 -3.41128796e-01
1.94117546e-01 -3.80465597e-01 -1.64771497e+00 -2.89638545e-02
-1.75424647e-02 -5.66473186e-01 -1.02358711e+00 2.17163369e-01
-4.02030140e-01 1.99636891e-02 -1.43725979e+00 7.16792822e-01
1.65856451e-01 -8.76474500e-01 5.97293019e-01 8.75980686e-03
-1.13372691e-01 -5.09561241e-01 1.08059578e-01 -1.74492621e+00
-3.18743974e-01 -1.86888114e-01 -1.91921771e-01 1.35416114e+00
1.16909035e-01 -9.69467640e-01 4.90138501e-01 -5.07231914e-02
-6.53605908e-02 -6.34303927e-01 -9.58219409e-01 -5.38829803e-01
-1.77935451e-01 -1.43879682e-01 6.50252700e-01 1.13561869e+00
6.65335298e-01 7.26782391e-03 -3.25173214e-02 6.59554780e-01
7.42953956e-01 1.18624702e-01 1.25907227e-01 -1.27872109e+00
-2.97452867e-01 -6.89965561e-02 -1.06447101e+00 2.31903717e-02
1.12011120e-01 -5.05665541e-01 4.89033163e-02 -1.40958416e+00
-3.82795393e-01 -3.61434132e-01 -3.71198237e-01 2.08330125e-01
-1.27981156e-01 -9.88151804e-02 -1.95398986e-01 1.81280777e-01
2.50469118e-01 7.53545523e-01 5.24563074e-01 -2.49306306e-01
3.16701829e-01 1.64037384e-02 1.81122571e-01 6.93301439e-01
1.32013357e+00 -1.52658552e-01 -7.33662367e-01 -2.04898581e-01
-3.05223137e-01 3.64898384e-01 -2.08174124e-01 -1.68722689e+00
2.02544421e-01 3.34589571e-01 7.06996083e-01 -8.47195446e-01
-5.42999385e-03 -1.07023466e+00 1.68773398e-01 5.54842591e-01
3.64064932e-01 4.30914313e-01 1.06720388e-01 2.14790538e-01
-5.36021233e-01 -4.26348358e-01 6.11409783e-01 2.18942732e-01
-8.71132076e-01 -2.78688759e-01 -1.17801499e+00 -3.95220250e-01
1.23232007e+00 -3.45502310e-02 -2.96613663e-01 -3.93237114e-01
-4.79628175e-01 1.63040068e-02 -5.29968478e-02 4.98429686e-01
7.68303752e-01 -1.09781647e+00 -6.08471751e-01 5.46076298e-01
-4.66336496e-02 -2.72766739e-01 4.26144570e-01 8.18391621e-01
-4.16550785e-01 2.14440033e-01 -3.85259330e-01 -7.48526394e-01
-9.06661510e-01 5.08514404e-01 4.63923097e-01 2.08175480e-01
-7.82193244e-01 9.44268033e-02 -6.86431408e-01 1.21976972e-01
2.05008566e-01 6.26871511e-02 -8.53649616e-01 1.50577202e-01
5.93733788e-01 5.85662127e-01 6.60808921e-01 -7.19122469e-01
-2.39878401e-01 3.93478572e-01 1.91059157e-01 4.03984964e-01
1.41420341e+00 1.15794927e-01 -1.50712267e-01 5.44612527e-01
1.30633080e+00 -3.12372446e-01 -7.86646783e-01 2.84542263e-01
2.45131239e-01 -2.81394035e-01 1.58117533e-01 -8.11273634e-01
-1.45501482e+00 6.86762691e-01 8.76403213e-01 5.97931921e-01
1.37757051e+00 -4.29559022e-01 4.27253485e-01 2.60180622e-01
6.24973476e-01 -1.23384655e+00 -2.90331364e-01 -2.45675836e-02
7.16290295e-01 -7.61097193e-01 -1.50481565e-02 -3.49567682e-01
-1.16443478e-01 1.35857439e+00 4.74250257e-01 -4.12159085e-01
9.85370398e-01 8.91564369e-01 1.96463689e-01 -9.55041423e-02
-5.57694852e-01 2.35457823e-01 -2.60427445e-02 8.22105706e-01
1.66735426e-01 1.03624174e-02 -5.75245857e-01 5.76373458e-01
-1.59119353e-01 -3.11453164e-01 8.11910927e-01 1.42747200e+00
-4.33364093e-01 -9.22511876e-01 -8.49993169e-01 7.73431361e-01
-6.54281020e-01 5.21700144e-01 3.26509356e-01 6.50055945e-01
1.17111355e-01 1.50482512e+00 -7.12732645e-03 -4.31274831e-01
5.54466963e-01 -1.11161202e-01 3.00013393e-01 -6.92178085e-02
-4.34891224e-01 -9.48290899e-02 -7.64517412e-02 -2.00446323e-01
-2.61290222e-01 -6.63084984e-01 -1.47122240e+00 5.92398643e-02
-7.40527749e-01 9.86144304e-01 9.02146041e-01 1.17756939e+00
-7.05322623e-02 6.59563959e-01 1.33151090e+00 -8.35956693e-01
-5.61293721e-01 -1.38763797e+00 -1.02114427e+00 6.64306819e-01
9.87144783e-02 -1.11729801e+00 -9.72007275e-01 -3.29603016e-01] | [6.3098673820495605, 2.500185012817383] |
72cf61ac-bb13-401f-9ee3-b0235e700f5c | ensembles-of-vision-transformers-as-a-new | 2203.01726 | null | https://arxiv.org/abs/2203.01726v3 | https://arxiv.org/pdf/2203.01726v3.pdf | Ensembles of Vision Transformers as a New Paradigm for Automated Classification in Ecology | Monitoring biodiversity is paramount to manage and protect natural resources. Collecting images of organisms over large temporal or spatial scales is a promising practice to monitor the biodiversity of natural ecosystems, providing large amounts of data with minimal interference with the environment. Deep learning models are currently used to automate classification of organisms into taxonomic units. However, imprecision in these classifiers introduces a measurement noise that is difficult to control and can significantly hinder the analysis and interpretation of data. {We overcome this limitation through ensembles of Data-efficient image Transformers (DeiTs), which not only are easy to train and implement, but also significantly outperform} the previous state of the art (SOTA). We validate our results on ten ecological imaging datasets of diverse origin, ranging from plankton to birds. On all the datasets, we achieve a new SOTA, with a reduction of the error with respect to the previous SOTA ranging from 29.35% to 100.00%, and often achieving performances very close to perfect classification. Ensembles of DeiTs perform better not because of superior single-model performances but rather due to smaller overlaps in the predictions by independent models and lower top-1 probabilities. This increases the benefit of ensembling, especially when using geometric averages to combine individual learners. While we only test our approach on biodiversity image datasets, our approach is generic and can be applied to any kind of images. | ['F. Pomati', 'P. Brun', 'M. Baity-Jesi', 'T. Bulas', 'E. Merz', 'M. Reyes', 'T. Hardeman', 'S. Kyathanahally'] | 2022-03-03 | null | null | null | null | ['fine-grained-image-classification'] | ['computer-vision'] | [ 1.60665959e-01 -5.14930785e-01 2.58004487e-01 -6.89486712e-02
-4.40627849e-03 -7.15934873e-01 6.23971641e-01 4.73552495e-01
-9.90409195e-01 9.78873372e-01 -1.75314277e-01 -1.09943636e-01
-2.84464598e-01 -1.05761302e+00 -7.44285643e-01 -9.02650654e-01
-5.22491217e-01 3.77967983e-01 3.90877634e-01 -1.68899268e-01
1.81119457e-01 6.87183380e-01 -2.01798820e+00 2.05081314e-01
9.53102529e-01 9.63861227e-01 4.28047836e-01 6.39515817e-01
-2.46139333e-01 4.27976847e-01 -7.51781464e-01 -4.23450410e-01
3.10260445e-01 -8.62957612e-02 -4.73809153e-01 -2.58813411e-01
6.12008154e-01 -3.30786526e-01 3.80974352e-01 1.17256963e+00
5.24715364e-01 -8.07564482e-02 8.45009804e-01 -7.72411942e-01
-4.47791129e-01 5.51432490e-01 -5.84121168e-01 1.67127624e-01
-2.46605769e-01 3.49118829e-01 9.03157234e-01 -4.12422121e-01
2.92346299e-01 1.15889633e+00 1.15002239e+00 3.02926779e-01
-1.62322891e+00 -7.01584876e-01 1.94783226e-01 2.75908321e-01
-1.48994315e+00 -3.08888733e-01 4.59331498e-02 -7.65036583e-01
1.11044014e+00 3.56974512e-01 1.00425804e+00 8.62740576e-01
1.30472198e-01 3.54838610e-01 1.29584253e+00 -1.15502901e-01
3.11769515e-01 1.39179185e-01 -2.06817389e-01 3.83707315e-01
5.04963994e-01 1.32382438e-01 -3.16408664e-01 -3.34919281e-02
8.41417015e-01 2.84983456e-01 -2.73967952e-01 -1.13015078e-01
-1.20104909e+00 7.24466980e-01 8.85049641e-01 2.86809206e-01
-6.10372722e-01 -2.90855337e-02 3.69982183e-01 5.89381382e-02
5.70392728e-01 6.47879601e-01 -4.46062386e-01 8.72657895e-02
-1.04064322e+00 5.58282323e-02 8.00554156e-01 2.64228791e-01
6.25573635e-01 5.34982793e-02 3.26092213e-01 1.12668645e+00
1.54095575e-01 9.19547975e-01 4.11301643e-01 -7.60773659e-01
-1.63756415e-01 7.49844015e-01 1.12819038e-01 -9.02298748e-01
-5.57125628e-01 -8.80896270e-01 -1.25760293e+00 5.87540746e-01
5.89634955e-01 8.32584128e-02 -7.74651289e-01 1.60568798e+00
2.23954674e-02 -3.60194445e-02 -8.35206881e-02 5.80887914e-01
6.56968653e-01 8.02048981e-01 3.14654350e-01 -2.21539468e-01
1.19840324e+00 -4.86335516e-01 -2.68286049e-01 -1.95839226e-01
1.04991831e-01 -3.35079372e-01 9.32052910e-01 7.32043445e-01
-5.91925144e-01 -3.81741911e-01 -9.62706208e-01 4.70048606e-01
-7.50362992e-01 1.77646708e-02 6.09541714e-01 6.37252450e-01
-1.22500324e+00 7.94679582e-01 -7.12343872e-01 -5.07965982e-01
5.73758662e-01 3.99636000e-01 -4.81345326e-01 5.32302797e-01
-7.49309778e-01 9.41558063e-01 6.17111146e-01 4.39021997e-02
-8.58500361e-01 -7.00674176e-01 -3.93528074e-01 2.08095789e-01
7.89179467e-03 -4.18035269e-01 9.82091248e-01 -1.24879372e+00
-1.25783408e+00 8.72684658e-01 3.12324435e-01 -7.79880881e-01
5.03801584e-01 -2.24791780e-01 -7.46930912e-02 -1.72920644e-01
-4.20248210e-02 1.11330903e+00 3.74917865e-01 -1.26225948e+00
-8.43506515e-01 -3.60859871e-01 1.83031540e-02 -7.91310593e-02
-8.57830584e-01 -1.39367014e-01 2.96234190e-01 -5.71026623e-01
-2.34492719e-01 -7.54146397e-01 -3.14642698e-01 3.85979265e-01
1.55004337e-01 8.80536661e-02 2.62842894e-01 -6.12118185e-01
1.02017689e+00 -1.87920928e+00 3.28743219e-01 -1.32283077e-01
-5.52769825e-02 4.85181898e-01 -1.17282577e-01 3.92190307e-01
2.56585509e-01 3.59188914e-01 -7.02616572e-01 -3.98164876e-02
-2.28051499e-01 4.15533274e-01 -1.11023374e-01 3.43211681e-01
1.66932374e-01 3.93921673e-01 -9.10482645e-01 -2.66686022e-01
5.37859082e-01 4.92112428e-01 -4.96266991e-01 1.10461734e-01
-3.06113094e-01 3.54842305e-01 1.41451016e-01 4.93448317e-01
6.38791323e-01 -3.88828106e-02 1.18636422e-01 1.16708636e-01
-5.18543005e-01 -8.52911361e-03 -1.05773795e+00 1.17937326e+00
-6.96491897e-01 6.23521626e-01 8.84023905e-02 -1.01370895e+00
9.96822000e-01 -6.20995536e-02 3.71401519e-01 -5.63653409e-01
-4.20365520e-02 4.48255002e-01 2.04526052e-01 -4.04440999e-01
1.85617030e-01 -3.89284581e-01 2.81326920e-01 9.31326225e-02
2.88416505e-01 -1.69631764e-02 2.04020947e-01 -5.05666375e-01
6.27193272e-01 9.17460546e-02 6.43320739e-01 -7.61937797e-01
6.17235839e-01 -1.33094797e-02 3.84973139e-01 7.58477032e-01
-1.46148652e-01 2.59775698e-01 2.30413631e-01 -8.33209276e-01
-1.14144552e+00 -9.70921814e-01 -6.20160162e-01 1.16363454e+00
-3.09854429e-02 9.94695649e-02 -5.05760789e-01 -3.43675405e-01
2.39556760e-01 6.09803021e-01 -7.49831319e-01 4.08861367e-03
-9.43323895e-02 -1.21265459e+00 5.97452879e-01 4.25696075e-01
6.59305036e-01 -1.21708679e+00 -9.46882248e-01 2.64116406e-01
-6.24387600e-02 -6.81967616e-01 3.42103451e-01 3.83161932e-01
-1.01716232e+00 -8.09846103e-01 -8.16680968e-01 -3.04037899e-01
1.99061289e-01 2.26206988e-01 1.24697495e+00 -4.98677120e-02
-3.14911425e-01 -9.30575207e-02 -2.73420036e-01 -7.71447599e-01
-3.94807905e-01 7.96252713e-02 1.54664576e-01 -1.73327222e-01
3.60235482e-01 -8.87573779e-01 -7.19759166e-01 4.85970736e-01
-1.04915512e+00 -1.77629337e-01 6.89019203e-01 8.62144947e-01
6.03602946e-01 7.93151408e-02 6.46703064e-01 -3.46365809e-01
1.80644095e-01 -4.58859295e-01 -8.05365324e-01 2.53120124e-01
-3.98460925e-01 2.99199764e-02 9.04828668e-01 -4.34098274e-01
-6.86682224e-01 1.54498160e-01 -3.77642602e-01 -1.53583303e-01
-3.87773842e-01 4.66159105e-01 1.09231733e-01 -3.18269372e-01
9.56521690e-01 2.97042578e-01 1.43292591e-01 -5.68900168e-01
-1.19995639e-01 8.51297081e-01 5.22236466e-01 -3.00823778e-01
2.69157231e-01 5.33729374e-01 1.06748171e-01 -1.32975030e+00
-6.94415689e-01 -5.16955912e-01 -7.27232158e-01 -4.15907502e-01
6.96697593e-01 -9.25242305e-01 -7.76849985e-01 6.58153176e-01
-8.55407655e-01 -3.89173120e-01 -3.40281785e-01 5.38882315e-01
-2.09860757e-01 3.31138730e-01 -2.46530637e-01 -1.08392489e+00
-6.05763257e-01 -1.11054528e+00 9.46192086e-01 1.61280662e-01
9.24061313e-02 -1.02800059e+00 2.18240529e-01 -1.20963223e-01
6.60927176e-01 4.21048075e-01 6.02939308e-01 -5.23437619e-01
-2.37529963e-01 -1.11891143e-01 -2.49609545e-01 7.03296125e-01
2.63275295e-01 4.18658942e-01 -1.20765638e+00 -2.57237613e-01
-2.15580598e-01 -2.29693800e-01 1.29233575e+00 4.66591448e-01
1.17105031e+00 -2.48023793e-01 -4.33893278e-02 5.46450913e-01
1.69081581e+00 -1.15563977e-03 5.16950309e-01 6.70589626e-01
2.15718314e-01 8.17718923e-01 2.09462821e-01 5.78052938e-01
9.25547704e-02 6.40537620e-01 1.10710573e+00 7.00803325e-02
1.35488957e-01 1.76540166e-01 3.99226010e-01 2.70236313e-01
-5.76209545e-01 -3.12681735e-01 -1.13821220e+00 6.18911147e-01
-1.72078240e+00 -1.22155631e+00 -3.32904220e-01 2.70463514e+00
6.92226589e-01 -6.06242828e-02 4.59164292e-01 3.11802123e-02
5.38150012e-01 -1.78646818e-01 -5.40072620e-01 -1.96656778e-01
-2.88516700e-01 2.37075195e-01 7.37831473e-01 3.63289833e-01
-1.39172876e+00 7.17230320e-01 6.69293928e+00 6.35930061e-01
-1.43572271e+00 -1.26400277e-01 3.81812364e-01 -1.85747802e-01
5.66616468e-02 -4.68394905e-01 -6.95892751e-01 4.36742455e-01
1.05182540e+00 1.42955944e-01 4.65674102e-01 8.43820870e-01
2.17458010e-01 -2.13669553e-01 -7.12875843e-01 7.67635405e-01
-1.91180915e-01 -9.90069330e-01 1.42073974e-01 2.88009644e-01
4.44849014e-01 5.08734941e-01 -2.39669532e-01 2.02337161e-01
4.37980086e-01 -1.10485041e+00 7.99418509e-01 5.93066454e-01
6.22920692e-01 -7.15414464e-01 1.04989922e+00 6.41952813e-01
-9.97076690e-01 -2.97290117e-01 -8.79473090e-01 -4.75646526e-01
-2.61299372e-01 5.99713206e-01 -6.73169434e-01 3.28873605e-01
1.26255977e+00 7.37490833e-01 -7.16020346e-01 1.57279932e+00
-4.05779481e-02 6.04882002e-01 -7.71974623e-01 -4.22546417e-01
1.82050735e-01 -2.50851899e-01 4.25524503e-01 1.42997062e+00
3.91681850e-01 -2.57316440e-01 1.20065376e-01 5.69200158e-01
1.22783110e-01 2.69859284e-01 -5.72492659e-01 1.80659771e-01
2.69585222e-01 1.32139134e+00 -6.98957860e-01 -3.19551468e-01
-6.18431345e-02 6.65047884e-01 3.88452083e-01 -1.89171612e-01
-6.38184845e-01 -1.28896475e-01 7.63753891e-01 -4.62733917e-02
3.47722083e-01 -2.18333215e-01 -1.03467010e-01 -1.03799653e+00
-6.95160106e-02 -8.48918736e-01 4.08089578e-01 -3.95126969e-01
-1.56871521e+00 7.94224143e-01 1.01866208e-01 -1.39080656e+00
1.01223648e-01 -9.31230783e-01 -2.58493632e-01 6.90579712e-01
-1.56515515e+00 -1.00194776e+00 -5.78516126e-01 7.02534020e-02
3.96355301e-01 1.54323667e-01 1.12798285e+00 3.20123076e-01
-3.73513937e-01 1.43447846e-01 5.61626136e-01 -3.41449112e-01
6.03388250e-01 -1.37180972e+00 9.99822617e-02 5.66824853e-01
1.39746964e-01 4.26480830e-01 6.76006556e-01 -4.12259042e-01
-7.58164525e-01 -1.08737147e+00 3.54970336e-01 -1.68873534e-01
8.22680652e-01 -1.59776568e-01 -1.18242788e+00 2.53830492e-01
1.20627776e-01 -1.40392825e-01 9.91397023e-01 7.54365548e-02
-3.26507777e-01 -4.64011282e-01 -1.32446730e+00 3.23495865e-01
9.64643002e-01 -5.67039028e-02 -3.30418378e-01 2.57983774e-01
1.48142070e-01 2.96033204e-01 -1.13648891e+00 5.27571559e-01
1.11607957e+00 -1.37424314e+00 1.11629796e+00 -4.96246219e-01
4.75777209e-01 -4.18819577e-01 -2.85010934e-01 -1.37608528e+00
-4.93776113e-01 1.89945146e-01 3.62583220e-01 1.10348225e+00
4.65693235e-01 -8.04812193e-01 4.01802480e-01 -1.51655287e-01
-2.63027586e-02 -2.09872723e-01 -8.96359146e-01 -1.08675563e+00
3.27225924e-01 -1.83837071e-01 5.11338890e-01 7.55643904e-01
-3.20940733e-01 1.33338928e-01 -3.17044526e-01 4.54256147e-01
8.45808864e-01 6.03237376e-02 7.83324540e-01 -1.83003473e+00
-3.67433697e-01 -9.73128200e-01 -6.48275852e-01 -5.38126588e-01
-2.02095553e-01 -6.63299799e-01 2.15858370e-01 -1.45057678e+00
3.65374118e-01 -2.38668114e-01 -3.53453308e-01 6.71601534e-01
-1.38797909e-01 5.80863059e-01 -1.09121963e-01 4.12507713e-01
-2.47703165e-01 5.40490806e-01 8.29958797e-01 -1.80984080e-01
-7.76730031e-02 2.97044460e-02 -3.57862294e-01 8.07402790e-01
8.58754098e-01 -3.50871563e-01 5.27916737e-02 -5.26665628e-01
7.07273483e-02 -4.99213159e-01 6.28902137e-01 -1.42589247e+00
-2.37181574e-01 -2.45861232e-01 5.44065952e-01 -3.21732074e-01
1.92311868e-01 -9.83450651e-01 5.54955125e-01 8.35417986e-01
4.02880609e-02 -2.72021741e-01 5.60316682e-01 4.61910903e-01
-6.14617839e-02 -4.26064253e-01 1.11015189e+00 -4.76306677e-01
-6.50260627e-01 8.84151235e-02 -4.90128487e-01 -6.35108113e-01
7.86890924e-01 -1.77383199e-01 -3.45782876e-01 -4.48857062e-02
-5.31128407e-01 6.94343150e-02 6.71633601e-01 1.97020188e-01
1.25890583e-01 -6.85683250e-01 -9.81737733e-01 -4.17531356e-02
2.62426764e-01 3.08369864e-02 9.18574706e-02 7.26727307e-01
-1.04614329e+00 2.96162188e-01 -6.87887371e-01 -9.69081044e-01
-1.28257430e+00 3.63303602e-01 7.16889918e-01 -1.94991499e-01
-3.25568616e-01 8.45172524e-01 2.02771962e-01 -6.41629040e-01
2.82527179e-01 -2.87274688e-01 -6.16086423e-01 2.68807352e-01
7.18218863e-01 2.97492713e-01 1.33242220e-01 -5.11091113e-01
-3.73487473e-01 5.78221202e-01 1.76560849e-01 1.33442849e-01
1.86183536e+00 1.46942526e-01 -4.33307856e-01 5.80369174e-01
5.93107939e-01 -2.75365800e-01 -1.38923192e+00 1.12923555e-01
-1.43091436e-02 -5.50756812e-01 1.25531003e-01 -9.82671082e-01
-9.10459995e-01 1.10410297e+00 9.30328727e-01 6.94164097e-01
1.19782877e+00 -3.66048157e-01 5.34082484e-03 6.10246897e-01
5.82505345e-01 -4.72533643e-01 -1.95734367e-01 2.70337850e-01
1.04684210e+00 -1.32270181e+00 1.39368385e-01 1.64013222e-01
-2.85524398e-01 1.23990834e+00 5.34616172e-01 -7.77706057e-02
3.86478424e-01 1.46483138e-01 -4.23966115e-03 3.73393930e-02
-5.28646946e-01 -4.95346338e-01 1.64330695e-02 6.90200686e-01
4.88989919e-01 1.61345080e-01 -3.34689587e-01 1.82960048e-01
-2.16912314e-01 -2.18593678e-03 2.87701756e-01 5.79678833e-01
-9.13270175e-01 -1.02582693e+00 -4.57113773e-01 5.60593784e-01
-4.26172525e-01 -2.06721172e-01 -6.75254226e-01 7.55686820e-01
2.42177308e-01 7.15932071e-01 1.79511726e-01 -3.29016358e-01
3.97390991e-01 -1.01563931e-01 3.39522213e-01 -3.30515146e-01
-7.94767976e-01 -2.70507578e-02 5.79145388e-04 -8.73101354e-02
-7.56245434e-01 -8.95720005e-01 -6.19010746e-01 -4.66923356e-01
-5.11266053e-01 1.38104603e-01 8.57996702e-01 5.22914708e-01
1.92223728e-01 4.39700186e-01 5.35873234e-01 -1.23910344e+00
-5.33169210e-01 -1.32988667e+00 -4.93251979e-01 2.49682590e-01
1.84253603e-01 -7.60947287e-01 -4.92371589e-01 -2.92399880e-02] | [9.142046928405762, -1.3071895837783813] |
bc878668-c6fe-4748-ac3b-afbb7f783845 | swintextspotter-scene-text-spotting-via | 2203.10209 | null | https://arxiv.org/abs/2203.10209v1 | https://arxiv.org/pdf/2203.10209v1.pdf | SwinTextSpotter: Scene Text Spotting via Better Synergy between Text Detection and Text Recognition | End-to-end scene text spotting has attracted great attention in recent years due to the success of excavating the intrinsic synergy of the scene text detection and recognition. However, recent state-of-the-art methods usually incorporate detection and recognition simply by sharing the backbone, which does not directly take advantage of the feature interaction between the two tasks. In this paper, we propose a new end-to-end scene text spotting framework termed SwinTextSpotter. Using a transformer encoder with dynamic head as the detector, we unify the two tasks with a novel Recognition Conversion mechanism to explicitly guide text localization through recognition loss. The straightforward design results in a concise framework that requires neither additional rectification module nor character-level annotation for the arbitrarily-shaped text. Qualitative and quantitative experiments on multi-oriented datasets RoIC13 and ICDAR 2015, arbitrarily-shaped datasets Total-Text and CTW1500, and multi-lingual datasets ReCTS (Chinese) and VinText (Vietnamese) demonstrate SwinTextSpotter significantly outperforms existing methods. Code is available at https://github.com/mxin262/SwinTextSpotter. | ['Lianwen Jin', 'Kai Ding', 'Nicholas Yuan', 'Shenggao Zhu', 'Dahua Lin', 'Chongyu Liu', 'Zhenghao Peng', 'Yuliang Liu', 'Mingxin Huang'] | 2022-03-19 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Huang_SwinTextSpotter_Scene_Text_Spotting_via_Better_Synergy_Between_Text_Detection_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Huang_SwinTextSpotter_Scene_Text_Spotting_via_Better_Synergy_Between_Text_Detection_CVPR_2022_paper.pdf | cvpr-2022-1 | ['text-spotting', 'scene-text-detection'] | ['computer-vision', 'computer-vision'] | [ 2.44498715e-01 -3.70878220e-01 -6.64603114e-02 -3.93830985e-01
-8.38690519e-01 -5.76756120e-01 8.01429570e-01 -2.42031693e-01
-5.42089403e-01 1.57067776e-01 5.54209411e-01 -2.66579151e-01
1.87068358e-01 -3.93962681e-01 -6.41372979e-01 -4.21361148e-01
8.07300985e-01 4.07267183e-01 2.98959136e-01 -1.56312868e-01
5.10828674e-01 2.90174200e-03 -1.17066133e+00 6.51246190e-01
9.75206912e-01 7.53718436e-01 6.86330259e-01 7.34569609e-01
-3.35655570e-01 6.03673398e-01 -3.90095860e-01 -5.83447695e-01
2.70749956e-01 -3.53040189e-01 -5.74340701e-01 2.71854460e-01
6.13286018e-01 -3.94721329e-01 -6.81120157e-01 8.72514546e-01
8.47376764e-01 -2.41440296e-01 3.37887615e-01 -8.64891589e-01
-7.36840904e-01 8.29603910e-01 -8.31892073e-01 -6.01083077e-02
3.25202405e-01 2.05531567e-01 1.15648270e+00 -1.33014333e+00
6.08292937e-01 1.00680351e+00 7.45695531e-01 2.74923474e-01
-9.47558224e-01 -6.12604439e-01 1.62388146e-01 1.94763213e-01
-1.54638040e+00 -6.43472195e-01 6.89638376e-01 -3.27324629e-01
9.46423411e-01 2.80134171e-01 2.12171927e-01 1.31718826e+00
-1.96917266e-01 1.32455504e+00 8.40445220e-01 -6.89924419e-01
-2.14247793e-01 1.17878430e-01 -7.80663686e-03 6.54346287e-01
1.65608257e-01 -1.95286244e-01 -7.61694670e-01 4.90870118e-01
4.46144402e-01 1.74224228e-01 -3.87791991e-01 -1.35552719e-01
-1.53507078e+00 5.50030291e-01 2.09960312e-01 3.56405646e-01
1.34270519e-01 -3.76561545e-02 7.49989629e-01 6.83673024e-02
4.79651928e-01 9.59268659e-02 -2.53213912e-01 -4.40437675e-01
-1.20064449e+00 -1.74185842e-01 5.31737983e-01 1.29926610e+00
4.83919740e-01 8.75523016e-02 -4.08073902e-01 1.02480984e+00
2.86795646e-01 8.22183847e-01 7.99436331e-01 -9.78557393e-02
1.05970371e+00 8.60531628e-01 -1.85095921e-01 -6.57754362e-01
-3.17399055e-01 -3.31247598e-01 -7.71505177e-01 -2.65274644e-01
4.10157591e-01 -2.10758775e-01 -1.03260565e+00 1.07540500e+00
2.08574861e-01 -2.38448113e-01 -2.81173199e-01 9.21852052e-01
7.15729475e-01 7.28574216e-01 -3.56219202e-01 4.50803906e-01
1.44164026e+00 -1.24531674e+00 -7.02176690e-01 -3.72210681e-01
1.01112640e+00 -1.11118913e+00 1.50566053e+00 1.96532980e-01
-7.42649734e-01 -4.06947404e-01 -1.05940890e+00 -6.71104133e-01
-5.74547350e-01 9.50136304e-01 6.28574491e-02 6.85126722e-01
-8.40599358e-01 9.48026925e-02 -7.13629484e-01 -7.06129253e-01
3.88269067e-01 1.74205884e-01 -1.46357387e-01 -1.58905536e-02
-8.26601505e-01 6.36571229e-01 4.25745964e-01 1.66822836e-01
-4.32583243e-01 -3.87146443e-01 -7.10870445e-01 5.19657992e-02
7.09366620e-01 -3.76766473e-01 1.19347060e+00 -7.65334368e-01
-1.68003726e+00 7.59867787e-01 -2.16046214e-01 -2.27527887e-01
1.03261924e+00 -4.98909026e-01 -4.59394842e-01 4.17999178e-02
3.23206693e-01 4.91069824e-01 9.44046974e-01 -7.82937825e-01
-7.76466846e-01 -2.82560885e-01 -5.23662329e-01 4.96009231e-01
-6.08781636e-01 2.36887217e-01 -8.36764991e-01 -1.02272320e+00
1.81831457e-02 -6.64437950e-01 3.99918526e-01 1.67155981e-01
-1.00874579e+00 -9.44652036e-02 1.29978001e+00 -6.74616635e-01
1.27079940e+00 -2.26463032e+00 -1.79424435e-02 -2.24597663e-01
1.62242532e-01 2.37351701e-01 -1.24684691e-01 6.11835659e-01
1.00466326e-01 7.92072788e-02 -1.58899233e-01 -8.45971346e-01
3.55282485e-01 -1.50480852e-01 -6.48907304e-01 7.12328851e-01
-7.75363147e-02 1.06402516e+00 -5.36869645e-01 -6.89361811e-01
5.82694352e-01 3.90882313e-01 -3.88654470e-01 -8.47463161e-02
-1.88148975e-01 1.84094422e-02 -4.78887886e-01 8.25350165e-01
5.95824957e-01 -3.63177866e-01 9.47486609e-02 -1.03555724e-01
-4.75176990e-01 4.06966686e-01 -1.08419883e+00 1.86699915e+00
-2.35625565e-01 1.14567816e+00 8.80265888e-03 -7.85832822e-01
7.46365786e-01 1.75504699e-01 2.59036422e-01 -1.09920335e+00
3.01644742e-01 1.85987651e-01 -4.05816555e-01 -3.89727145e-01
9.98299599e-01 2.05473244e-01 -1.55620322e-01 4.67613190e-01
-1.90077603e-01 4.84296158e-02 1.61569178e-01 2.44668439e-01
8.78807664e-01 3.53332400e-01 9.13676172e-02 -4.77004498e-02
4.18313771e-01 3.69250178e-02 2.51401216e-01 8.00266862e-01
-9.93295163e-02 9.02642488e-01 2.26812303e-01 -1.59485951e-01
-1.24038374e+00 -6.68027580e-01 -1.82369158e-01 1.32679188e+00
2.02456325e-01 -8.16372812e-01 -7.12868094e-01 -7.99194157e-01
-3.15352939e-02 5.64210713e-01 -6.25632107e-01 2.87622154e-01
-7.14227855e-01 -5.23674786e-01 1.05874062e+00 4.75371659e-01
7.97388732e-01 -7.93044448e-01 -5.85995913e-01 -1.25253886e-01
-3.98828149e-01 -1.39719725e+00 -1.12431550e+00 1.81063995e-01
-5.65994442e-01 -6.94209814e-01 -9.61583316e-01 -7.64637649e-01
5.19341767e-01 6.18793666e-01 6.07639551e-01 -1.42937377e-01
-4.48518246e-01 2.62347609e-01 -5.79932988e-01 -6.58721179e-02
-1.23781063e-01 3.69394243e-01 -3.05160493e-01 2.56460160e-01
2.86989719e-01 -4.29316834e-02 -5.06040454e-01 5.81767797e-01
-8.00279617e-01 7.17738926e-01 6.70187414e-01 8.41796935e-01
3.32075119e-01 -1.17935181e-01 1.04120456e-01 -7.68563986e-01
2.81659126e-01 -6.34395480e-02 -6.19178951e-01 3.93958956e-01
-5.55897832e-01 -7.77811557e-02 8.82027805e-01 -2.95507610e-01
-1.15504670e+00 2.15109110e-01 -7.68181309e-02 -3.58197033e-01
-1.08269751e-01 3.91720146e-01 -3.28104526e-01 2.86151528e-01
5.54616451e-01 7.53210306e-01 -4.22253489e-01 -6.43665314e-01
4.76975232e-01 1.11369681e+00 5.50776482e-01 -3.66117835e-01
7.73600280e-01 6.26498222e-01 -6.07822657e-01 -1.18827963e+00
-5.64406633e-01 -6.03561699e-01 -7.96906531e-01 -4.02299650e-02
7.90930450e-01 -9.89020884e-01 -6.12604320e-01 7.25976467e-01
-1.08911228e+00 -5.67575812e-01 -1.79207727e-01 2.04673499e-01
-3.30260277e-01 6.50513828e-01 -4.62821603e-01 -6.16256535e-01
-4.69089121e-01 -1.06041443e+00 1.73892450e+00 2.53872555e-02
-2.26815082e-02 -6.79325044e-01 -1.63999170e-01 4.40866232e-01
3.40200424e-01 -2.88810164e-01 5.51639915e-01 -5.97188234e-01
-6.35079503e-01 -3.48302573e-01 -6.49922431e-01 -1.62337627e-02
5.85211925e-02 -2.96322852e-02 -1.02260816e+00 -2.63067603e-01
-3.31818193e-01 -1.78708375e-01 9.85321820e-01 1.62357375e-01
1.00590122e+00 -1.90205961e-01 -2.66564012e-01 8.26897502e-01
1.37624872e+00 6.44858554e-02 7.09288895e-01 4.73369807e-01
1.25261188e+00 3.62550318e-01 3.83505732e-01 6.73059464e-01
5.03142715e-01 9.07417119e-01 1.52378514e-01 -2.22671255e-01
-6.03968620e-01 -5.28123856e-01 6.05357170e-01 7.53215730e-01
5.20591795e-01 -6.64460719e-01 -1.03430319e+00 3.91855508e-01
-1.83823025e+00 -8.72401118e-01 -3.12720209e-01 2.05014229e+00
8.09551358e-01 1.15805559e-01 1.30536437e-01 1.78172573e-01
9.31159616e-01 2.85576671e-01 -5.57006776e-01 -1.26549676e-02
-4.59909081e-01 -1.89831108e-01 7.13134825e-01 3.92031580e-01
-1.15889657e+00 1.28961027e+00 5.17977905e+00 1.21190178e+00
-1.41162562e+00 2.39966325e-02 4.71406281e-01 -1.45653218e-01
1.11451431e-03 2.46581007e-02 -1.18396425e+00 6.75544381e-01
5.36604881e-01 -4.79281088e-03 3.75342697e-01 6.35167658e-01
3.03570509e-01 3.05938125e-02 -1.04239035e+00 1.23454833e+00
3.04407567e-01 -1.30875182e+00 -1.02288760e-01 -1.60690144e-01
4.05097872e-01 3.71656984e-01 2.15946212e-01 1.23655736e-01
-2.13742331e-02 -8.27673554e-01 1.24549222e+00 1.81186736e-01
1.30847907e+00 -1.50842339e-01 1.93170920e-01 2.41134956e-01
-1.39115953e+00 9.07693133e-02 -1.44760296e-01 1.48567900e-01
1.84399724e-01 5.17243505e-01 -1.02413321e+00 5.28073728e-01
5.52038968e-01 1.15199447e+00 -9.03850973e-01 9.53790605e-01
-2.32207432e-01 7.31040359e-01 -4.93420243e-01 -2.47762144e-01
1.80566773e-01 4.95866127e-02 6.14717185e-01 1.72331691e+00
3.15796345e-01 -4.37447190e-01 1.37277186e-01 9.60092068e-01
-2.73075223e-01 2.87058085e-01 -4.38161135e-01 -1.83255687e-01
4.75951970e-01 1.17121530e+00 -1.05558944e+00 -4.98796433e-01
-5.31577587e-01 1.46800923e+00 7.47833028e-02 2.79200882e-01
-1.02344298e+00 -8.35787714e-01 2.62676179e-02 -6.57652447e-04
6.33315146e-01 -2.12190151e-01 -7.52219915e-01 -1.62821257e+00
3.28392744e-01 -9.74978805e-01 1.76268756e-01 -8.06982636e-01
-9.80571270e-01 4.73148078e-01 -4.64040339e-01 -1.41019571e+00
2.29426816e-01 -7.08870769e-01 -4.79846001e-01 8.22238624e-01
-1.40250778e+00 -1.53159773e+00 -4.36774671e-01 6.60777092e-01
1.12183464e+00 -4.28715684e-02 3.86119276e-01 5.61426878e-01
-1.12080204e+00 1.01111853e+00 5.19618571e-01 4.93610442e-01
1.05736864e+00 -1.11270916e+00 7.45603085e-01 1.21280050e+00
2.28027403e-01 2.32981652e-01 4.69029576e-01 -7.80938029e-01
-1.64084220e+00 -1.17801750e+00 9.49388444e-01 -5.56460023e-01
6.29815340e-01 -9.92867172e-01 -6.90231323e-01 6.92815006e-01
3.18198055e-01 -1.73553437e-01 1.05913244e-01 -8.72998536e-02
-5.07350862e-01 5.84817529e-02 -6.99261665e-01 1.00867689e+00
1.01207209e+00 -7.67062783e-01 -3.57007354e-01 3.18365574e-01
5.84317863e-01 -4.94605303e-01 -2.86957949e-01 5.37148153e-04
5.70076287e-01 -9.09167051e-01 7.22981989e-01 9.77101326e-02
5.25517642e-01 -4.04411674e-01 -2.41135046e-01 -7.74315417e-01
-4.77446131e-02 -7.92110085e-01 1.48110420e-01 1.24045253e+00
4.12156761e-01 -6.44691527e-01 8.56881559e-01 2.30041504e-01
-5.02263010e-01 -3.66173714e-01 -8.91205072e-01 -7.31366813e-01
-3.35204192e-02 -5.23551822e-01 3.55398685e-01 9.92568374e-01
2.23289236e-01 5.63143492e-01 -6.37083590e-01 -5.84131293e-02
3.95950377e-01 1.50649056e-01 9.66406941e-01 -7.42243230e-01
-3.22028697e-01 -6.97281599e-01 -1.49992436e-01 -1.77472627e+00
-2.16877148e-01 -9.98064101e-01 2.77424723e-01 -1.35385776e+00
2.57654756e-01 -2.94884205e-01 2.65262693e-01 8.17193508e-01
-1.21680729e-03 1.94271639e-01 3.41675311e-01 3.33614290e-01
-8.58399928e-01 8.86730254e-01 1.03324556e+00 -1.35542259e-01
-7.67352059e-02 -4.15311873e-01 -6.02708876e-01 4.93754953e-01
6.83776855e-01 -3.20850939e-01 -1.84986949e-01 -9.61494029e-01
1.53208926e-01 -1.46019578e-01 2.47951359e-01 -6.91591680e-01
6.42581046e-01 2.08810836e-01 4.31209654e-01 -1.00534332e+00
1.76781148e-01 -7.49042153e-01 -3.30436856e-01 1.46968335e-01
-4.11836863e-01 6.12660497e-02 2.29684591e-01 4.38942105e-01
-1.50382519e-04 -1.87343076e-01 6.10861421e-01 3.73057842e-01
-7.26368129e-01 2.76210140e-02 -3.19379061e-01 2.33196661e-01
6.63885474e-01 -6.16589367e-01 -7.36372709e-01 -1.41870335e-01
-8.72019753e-02 2.00360924e-01 6.46148205e-01 6.20038271e-01
6.64081991e-01 -9.80699062e-01 -7.83997476e-01 3.03268433e-01
2.66160935e-01 -1.16588743e-02 2.40798026e-01 1.03226388e+00
-4.79193717e-01 6.96618855e-01 1.85057774e-01 -7.22200990e-01
-1.28712225e+00 4.15831268e-01 3.12764376e-01 -1.70271814e-01
-1.07749450e+00 5.94906747e-01 3.89986902e-01 -3.97456318e-01
3.95581841e-01 -3.92396867e-01 3.46210450e-01 -5.76596409e-02
4.56940442e-01 5.04043996e-01 1.19782664e-01 -6.53204441e-01
-2.97712535e-01 6.84463441e-01 -4.54839736e-01 -1.73952371e-01
9.80803847e-01 -5.45007646e-01 2.64827460e-01 2.74731278e-01
1.18961799e+00 2.19863564e-01 -1.27183962e+00 -4.28138703e-01
1.97447971e-01 -5.39805293e-01 9.16601941e-02 -1.06203818e+00
-8.79776597e-01 9.46270466e-01 4.86180931e-01 -1.18795015e-01
9.61053431e-01 -2.02492431e-01 8.81351709e-01 4.42289114e-01
-5.44745252e-02 -1.26953673e+00 1.71943709e-01 7.94432700e-01
7.36645579e-01 -1.31545925e+00 -4.32177000e-02 -4.29862916e-01
-7.66657889e-01 1.14068258e+00 4.23986912e-01 2.55121022e-01
2.16117799e-01 4.74792272e-01 -2.80545745e-02 -6.71448261e-02
-5.25531769e-01 -2.55876273e-01 1.58695400e-01 1.57239139e-01
5.91855764e-01 6.42968016e-03 1.24458969e-01 4.42901760e-01
-1.33648857e-01 -1.91907540e-01 4.96921182e-01 1.01200283e+00
-3.86314094e-01 -1.04564178e+00 -4.22463685e-01 4.10354644e-01
-4.50601488e-01 -5.05646229e-01 -4.60160255e-01 7.93857813e-01
-2.64886677e-01 7.91009068e-01 -2.51358468e-02 -4.40778673e-01
3.28283548e-01 -1.31575875e-02 8.01203251e-02 -4.57296222e-01
-4.61062163e-01 5.95316529e-01 3.04788761e-02 -3.04120868e-01
1.71499193e-01 -7.18325078e-01 -1.25582671e+00 -3.75758708e-01
-6.74548268e-01 -3.75761360e-01 7.31269181e-01 8.19644272e-01
6.84708476e-01 4.78884429e-01 5.34782171e-01 -7.62288690e-01
-3.98749799e-01 -1.05862415e+00 -5.39443493e-01 1.36273414e-01
3.38931262e-01 -2.81403929e-01 -1.32308051e-01 2.49462858e-01] | [11.920682907104492, 2.2241227626800537] |
1485b4f7-6f2a-4156-9e40-e6a2576341c4 | aspera-aspect-based-rating-prediction-based | null | null | https://aclanthology.org/W19-3605 | https://aclanthology.org/W19-3605.pdf | AspeRa: Aspect-Based Rating Prediction Based on User Reviews | We propose a novel Aspect-based Rating Prediction model (AspeRa) that estimates user rating based on review texts for the items. It is based on aspect extraction with neural networks and combines the advantages of deep learning and topic modeling. It is mainly designed for recommendations, but an important secondary goal of AspeRa is to discover coherent aspects of reviews that can be used to explain predictions or for user profiling. We conduct a comprehensive empirical study of AspeRa, showing that it outperforms state-of-the-art models in terms of recommendation quality and produces interpretable aspects. This paper is an abridged version of our work (Nikolenko et al., 2019) | ['Elena Tutubalina', 'Sergey Nikolenko', 'Ilya Shenbin', 'Valentin Malykh', 'Anton Alekseev'] | 2019-08-01 | null | null | null | ws-2019-8 | ['aspect-extraction'] | ['natural-language-processing'] | [-3.37077707e-01 5.64586341e-01 -9.52060223e-01 -5.88676572e-01
-4.58143264e-01 -2.28347868e-01 8.67885053e-01 3.31132978e-01
2.12390665e-02 4.68405962e-01 7.76583552e-01 -6.08501077e-01
-4.46295261e-01 -8.75456750e-01 -5.03063560e-01 -7.11515825e-03
1.40179008e-01 7.52025306e-01 -3.49953651e-01 -5.57911754e-01
5.54028988e-01 2.93910056e-02 -1.53526330e+00 8.50645065e-01
6.27491653e-01 1.18919408e+00 -2.61472523e-01 3.92859131e-01
-4.58144873e-01 7.14637280e-01 -7.22892225e-01 -1.02367640e+00
-1.37517154e-01 -1.51685312e-01 -6.00165188e-01 -3.83649357e-02
8.59779269e-02 -2.77682602e-01 -2.95792110e-02 6.18121922e-01
1.90986842e-01 5.36905378e-02 1.10874975e+00 -1.13884532e+00
-1.25490844e+00 1.14665508e+00 -2.19689101e-01 -1.61730781e-01
1.10816054e-01 -5.41295052e-01 1.73806846e+00 -1.18955112e+00
4.05030727e-01 9.16353703e-01 8.66736174e-01 3.61025155e-01
-7.41879761e-01 -1.33072823e-01 5.04986763e-01 1.99193552e-01
-5.69006205e-01 2.72557186e-03 6.31823242e-01 -3.60207826e-01
1.37095249e+00 2.87978679e-01 1.05493605e+00 9.72107708e-01
4.31587279e-01 9.89602149e-01 7.10126579e-01 -2.40816981e-01
4.74202037e-01 4.26405668e-01 6.73830748e-01 2.44269714e-01
6.33092701e-01 -2.19579533e-01 -6.79678023e-01 -2.52756327e-01
3.22994560e-01 2.56556690e-01 1.57348678e-01 -2.00182736e-01
-8.52879822e-01 1.22005916e+00 2.00156450e-01 1.21802732e-01
-7.84612656e-01 -2.49011993e-01 2.89667845e-01 3.72399151e-01
1.16610920e+00 8.55450571e-01 -1.06662679e+00 -2.17928380e-01
-8.08745444e-01 4.68347311e-01 1.16616201e+00 6.03325665e-01
2.62557238e-01 1.91608727e-01 -3.61575723e-01 9.84315932e-01
5.99232316e-01 3.64335179e-01 6.43549442e-01 -6.78400099e-01
1.04882009e-01 8.11359107e-01 1.16483286e-01 -8.28676701e-01
-5.78292310e-01 -7.24277377e-01 -6.02447152e-01 1.18659630e-01
-2.59440005e-01 -2.76178926e-01 -8.94650221e-01 8.94715965e-01
-1.21422075e-01 -3.95677716e-01 9.45231691e-02 6.65661573e-01
1.53359437e+00 6.55840755e-01 1.05041936e-01 -1.95287794e-01
1.53063858e+00 -1.25775576e+00 -9.37853634e-01 -2.08583280e-01
1.99458197e-01 -5.67522228e-01 9.86988366e-01 7.48222768e-01
-1.14019382e+00 -5.93685210e-01 -1.16776133e+00 1.30830824e-01
-7.44047642e-01 4.56264764e-01 1.30728316e+00 6.97653890e-01
-1.03578544e+00 5.84302485e-01 -3.77281368e-01 -2.12159723e-01
4.73848969e-01 2.59716064e-01 -8.54951367e-02 2.37511545e-01
-1.18863225e+00 8.33077729e-01 -1.63283080e-01 -2.37200037e-01
-3.83341283e-01 -7.15856791e-01 -8.54071140e-01 2.66367316e-01
3.03210407e-01 -8.99625659e-01 1.54935896e+00 -8.57398450e-01
-1.63537860e+00 3.32431972e-01 -2.25259990e-01 -5.42611480e-01
-3.04389745e-01 -5.29862642e-01 -7.55368888e-01 -2.31383309e-01
-2.49106996e-02 1.90983757e-01 6.40500605e-01 -1.07354653e+00
-6.71973526e-01 -4.22563404e-01 3.54433626e-01 3.02322865e-01
-5.25171340e-01 1.74344391e-01 -5.22706866e-01 -8.58018398e-01
-5.56605995e-01 -6.41439795e-01 -4.35594380e-01 -6.48261487e-01
-3.08646619e-01 -5.29449582e-01 3.01745474e-01 -6.17520213e-01
1.46962035e+00 -1.38900685e+00 -1.64195910e-01 1.22316927e-01
4.19436872e-01 3.83415729e-01 -2.66089052e-01 5.73064744e-01
1.49282592e-03 3.25318873e-01 1.63004339e-01 -5.70971966e-01
1.83329672e-01 -1.37391984e-01 -4.81937230e-01 -2.00684145e-02
2.40492281e-02 1.19924569e+00 -6.44393265e-01 6.29786775e-02
1.88097823e-02 6.09950006e-01 -5.30160546e-01 1.79281607e-01
-4.45624709e-01 -2.66679138e-01 -3.97707611e-01 7.12234735e-01
2.89052486e-01 -4.62404937e-01 -3.54356878e-03 -1.69841766e-01
1.37144595e-01 1.05857730e+00 -6.47391140e-01 1.10183299e+00
-7.75923789e-01 6.59973264e-01 -5.76323450e-01 -7.86993027e-01
1.06869590e+00 5.44231951e-01 4.21135992e-01 -4.06357855e-01
3.80201191e-01 -1.41776651e-01 -3.55950505e-01 -2.18066201e-01
1.12330222e+00 -2.51871161e-02 -7.37831593e-02 9.32988584e-01
1.09504804e-01 -1.54280961e-01 2.26184621e-01 3.23370367e-01
7.59714961e-01 5.50026298e-02 8.58414114e-01 -3.26400623e-02
2.88687885e-01 -7.19523281e-02 2.55225986e-01 7.09593356e-01
2.35560104e-01 6.62272215e-01 6.95606947e-01 -9.55885947e-01
-1.17527175e+00 -7.72655785e-01 -8.34516734e-02 1.18395710e+00
-3.67521048e-01 -1.01506662e+00 -3.17572474e-01 -1.05189812e+00
-1.58866882e-01 1.05112910e+00 -1.17686594e+00 9.39498283e-03
1.06770374e-01 -8.32938552e-01 -2.77062356e-01 6.85272992e-01
-9.23226308e-03 -1.35418427e+00 -7.52376765e-02 1.11740828e-01
1.03818804e-01 -7.59426653e-01 -8.63473341e-02 2.14030609e-01
-8.74750376e-01 -9.28210139e-01 -5.49118280e-01 -3.69075924e-01
3.82847518e-01 2.93163508e-01 1.92543006e+00 1.23790562e-01
4.55777287e-01 3.06423455e-01 -7.90854871e-01 -9.19702947e-01
-2.45560139e-01 4.58265722e-01 5.78238517e-02 -2.50332981e-01
9.88681078e-01 -6.14483237e-01 -5.81835628e-01 2.21094266e-01
-6.67854071e-01 -9.34666991e-02 7.69766092e-01 6.02364600e-01
7.10894465e-01 -7.03471676e-02 8.91139328e-01 -1.56373811e+00
1.39028454e+00 -7.89303780e-01 -1.36447117e-01 1.09435603e-01
-1.47815430e+00 -3.10358226e-01 5.59860945e-01 -2.25562394e-01
-9.27016079e-01 -5.61532080e-01 -4.38736230e-01 2.91625172e-01
-2.43033335e-01 9.57025826e-01 8.22347552e-02 5.85323632e-01
6.44437134e-01 -4.30528969e-02 -3.47571433e-01 -6.05429053e-01
6.01098895e-01 9.28644180e-01 -5.72444359e-03 -8.80067982e-03
2.60121435e-01 3.44165981e-01 -2.76553869e-01 -5.32023311e-01
-1.55769849e+00 -5.43567598e-01 -4.46407050e-01 -2.40488827e-01
4.62883443e-01 -9.68235195e-01 -2.90085644e-01 5.99521287e-02
-9.87250865e-01 8.26886073e-02 -5.52889049e-01 3.49055409e-01
-4.77759868e-01 -1.60968632e-01 -7.51659691e-01 -7.87983239e-01
-1.01560748e+00 -7.31789649e-01 9.00062203e-01 3.44859689e-01
-6.21879220e-01 -1.10206032e+00 4.11985368e-01 5.53435206e-01
7.82798886e-01 -3.00299734e-01 9.28519189e-01 -1.20505595e+00
3.90249118e-02 -4.50730711e-01 -1.07865473e-02 3.65449011e-01
-2.21189093e-02 -1.13699950e-01 -9.86814737e-01 2.03104988e-01
-1.56217769e-01 -1.56160802e-01 9.50041831e-01 6.94182932e-01
1.10480058e+00 -5.14776945e-01 1.00925134e-03 2.44257718e-01
1.21640849e+00 1.72732785e-01 6.65088356e-01 7.50228465e-01
5.39207995e-01 5.42053878e-01 8.11436057e-01 5.44771254e-01
7.91002572e-01 5.27119756e-01 4.63099957e-01 1.30099237e-01
-8.00941139e-03 -2.41106674e-01 4.19798851e-01 1.17309034e+00
-3.05528194e-01 -2.74070144e-01 -3.82847160e-01 7.86657989e-01
-1.98544538e+00 -7.93043375e-01 -3.02620709e-01 1.69883502e+00
3.75304371e-01 3.35609823e-01 3.50814432e-01 6.98825791e-02
2.80241370e-01 3.59461308e-01 -4.99312133e-01 -1.21724832e+00
4.79292087e-02 2.73294300e-01 -1.05312318e-02 2.77153283e-01
-1.04648006e+00 6.96893036e-01 7.29215384e+00 5.93977153e-01
-4.07877237e-01 1.35571688e-01 9.33799207e-01 -1.48302302e-01
-8.00701797e-01 -2.72710562e-01 -7.41287768e-01 7.26961046e-02
1.24873090e+00 -2.24736065e-01 1.88714519e-01 1.27012444e+00
1.82952702e-01 2.33296156e-01 -7.97007978e-01 3.03811371e-01
4.65819418e-01 -1.44083214e+00 4.77365792e-01 2.01195672e-01
1.19105196e+00 2.12424323e-01 3.88014853e-01 6.94594443e-01
5.20516336e-01 -9.90965247e-01 2.81389266e-01 6.08431697e-01
2.79329002e-01 -1.05405641e+00 1.42118764e+00 1.18991723e-02
-9.01580393e-01 -9.34150741e-02 -7.33344078e-01 -3.85854572e-01
9.95075479e-02 1.06752551e+00 -5.40223837e-01 4.73707020e-01
7.75177777e-01 1.17171156e+00 -4.80742306e-01 9.13992822e-01
-6.92001045e-01 8.99557829e-01 4.12342191e-01 -3.69918555e-01
8.25428888e-02 -2.97124118e-01 3.96031260e-01 9.65402901e-01
4.29447174e-01 -1.94995046e-01 -2.18828484e-01 5.99348605e-01
-2.97251284e-01 5.16651988e-01 -6.04246080e-01 -9.95481461e-02
9.03739110e-02 1.65270269e+00 -5.93839228e-01 -6.40375435e-01
-6.72949076e-01 5.04323244e-01 2.76141286e-01 2.13440940e-01
-4.32916671e-01 -3.21043372e-01 8.29978824e-01 9.08576027e-02
9.52626765e-01 3.47493380e-01 -9.87754047e-01 -1.28421891e+00
-1.66304603e-01 -1.01386464e+00 3.10681939e-01 -9.00114000e-01
-1.53991044e+00 1.00501776e+00 -2.90789694e-01 -1.33087325e+00
-5.36561966e-01 -6.84748948e-01 -9.75172937e-01 7.25332975e-01
-1.62129653e+00 -1.34555936e+00 6.67774603e-02 1.34830028e-01
8.16073298e-01 -7.20794439e-01 9.68491912e-01 -9.47173461e-02
-1.32062852e-01 4.61888701e-01 2.33695880e-01 -2.83516139e-01
3.99575055e-01 -1.65259194e+00 9.90518212e-01 4.29663837e-01
5.17269969e-01 7.80288398e-01 7.64576435e-01 -5.92444241e-01
-8.98443222e-01 -1.03112769e+00 1.48701918e+00 -7.65688717e-01
6.86474681e-01 7.73406774e-02 -5.95942974e-01 5.77269554e-01
6.44862831e-01 -6.10066295e-01 1.37518263e+00 9.99029279e-01
-4.60875452e-01 -1.18611097e-01 -7.20851719e-01 6.32522941e-01
3.73203844e-01 -3.66764843e-01 -9.88079429e-01 1.24669284e-01
9.04825509e-01 1.80818647e-01 -1.10133779e+00 4.02940869e-01
9.66422021e-01 -1.15607226e+00 8.49697948e-01 -8.46689939e-01
1.13669217e+00 9.44680944e-02 5.77351861e-02 -1.81476665e+00
-5.05204260e-01 -1.18647233e-01 -7.73313165e-01 1.18988705e+00
1.03649175e+00 -2.56769001e-01 8.53009164e-01 5.00234723e-01
-8.59711990e-02 -1.41562712e+00 -2.65503526e-01 -2.25341946e-01
5.68868928e-02 -8.46883774e-01 9.99794424e-01 6.52148187e-01
4.59570438e-01 1.07321799e+00 -5.70284724e-01 -3.24438989e-01
2.57356763e-01 4.84985799e-01 6.61156893e-01 -1.55583763e+00
-5.45210719e-01 -6.55148804e-01 2.20586807e-01 -8.81921768e-01
2.35931464e-02 -5.27407169e-01 -2.49316722e-01 -2.27690315e+00
3.70419294e-01 -2.52567418e-02 -6.52889073e-01 3.75948757e-01
-2.65398204e-01 4.98990476e-01 -2.57767439e-02 6.37169853e-02
-9.54604864e-01 6.38469815e-01 1.05334556e+00 -1.51863068e-01
-7.82365799e-02 9.25452828e-01 -1.56426430e+00 8.06850791e-01
1.12131643e+00 -3.46522450e-01 -5.90369821e-01 -1.89699784e-01
1.15961289e+00 -2.13280886e-01 -4.33324844e-01 -6.01607680e-01
-3.23797315e-01 7.40479603e-02 4.01518673e-01 -1.02560771e+00
2.27505669e-01 -6.74515724e-01 -4.20266867e-01 -3.55644301e-02
-7.75287271e-01 3.54594022e-01 -5.20553365e-02 7.13298440e-01
-2.39592075e-01 -3.77067059e-01 1.34660397e-02 -1.60025880e-01
-3.44247162e-01 4.74570781e-01 -9.07109559e-01 -3.44992161e-01
2.81041324e-01 2.92152345e-01 -4.82065111e-01 -1.18376172e+00
-4.03662026e-01 -2.38681018e-01 1.33818462e-01 7.44928241e-01
7.25779116e-01 -1.38090181e+00 -6.10567868e-01 -2.39468336e-01
4.40232188e-01 -5.07349253e-01 -5.30870399e-03 3.24999988e-01
1.12421557e-01 9.12085950e-01 -7.62916505e-02 2.25528613e-01
-1.05059624e+00 7.99312413e-01 -1.71041012e-01 -9.64995921e-01
-4.95045990e-01 4.36997443e-01 -1.18564606e-01 -7.69391716e-01
1.70790449e-01 -2.53281474e-01 -1.42550588e+00 2.08225012e-01
9.98799860e-01 2.58053839e-01 2.65320331e-01 -4.42087799e-01
8.34521055e-02 2.00972095e-01 -3.14947903e-01 -5.48497550e-02
1.88352036e+00 -2.05498084e-01 -1.02460116e-01 7.70968199e-01
8.46392453e-01 3.17660533e-02 -6.10872567e-01 -2.16716424e-01
-1.72963813e-01 -1.40476868e-01 4.83102113e-01 -1.33314800e+00
-1.16310227e+00 8.16755116e-01 7.24738538e-02 7.05260873e-01
8.84314120e-01 1.21886946e-01 9.43465054e-01 6.03326559e-01
-1.11159422e-01 -1.12212980e+00 4.24672626e-02 7.27544844e-01
8.12885940e-01 -1.37988281e+00 4.33481574e-01 -1.50811495e-02
-9.98268008e-01 1.04883993e+00 5.51736236e-01 -5.25371023e-02
1.01101089e+00 -2.31005549e-02 2.83005059e-01 -5.63199043e-01
-1.38295972e+00 -1.50999933e-01 1.01431346e+00 7.84321189e-01
7.34540284e-01 2.45590538e-01 -7.19282508e-01 1.56023777e+00
-3.88554901e-01 -1.36319503e-01 7.02529669e-01 3.88239086e-01
-4.54457432e-01 -1.10866535e+00 1.13914385e-01 1.17297280e+00
-9.37413990e-01 -5.32149613e-01 -5.11900425e-01 4.16209668e-01
-1.23233356e-01 1.22131658e+00 -1.72438070e-01 -7.25990415e-01
3.16563994e-01 -4.62705903e-02 -3.83884490e-01 -7.13425994e-01
-9.88312006e-01 -3.04386322e-03 5.58511257e-01 -3.30956340e-01
-5.76449931e-01 -5.88079691e-01 -4.81795818e-01 -2.17245281e-01
-2.77320415e-01 5.93340993e-01 1.13480914e+00 1.13058054e+00
5.09382188e-01 7.66277492e-01 6.81335449e-01 -5.87008953e-01
-1.35007769e-01 -1.39049876e+00 -7.15212524e-01 -7.85042066e-03
1.38444319e-01 -4.28090125e-01 -2.02140525e-01 -2.73888946e-01] | [11.34514331817627, 6.624413967132568] |
a856aadc-4e3d-4e07-ab86-b64dd883e892 | normalizing-flow-based-neural-process-for-few | 2304.08183 | null | https://arxiv.org/abs/2304.08183v1 | https://arxiv.org/pdf/2304.08183v1.pdf | Normalizing Flow-based Neural Process for Few-Shot Knowledge Graph Completion | Knowledge graphs (KGs), as a structured form of knowledge representation, have been widely applied in the real world. Recently, few-shot knowledge graph completion (FKGC), which aims to predict missing facts for unseen relations with few-shot associated facts, has attracted increasing attention from practitioners and researchers. However, existing FKGC methods are based on metric learning or meta-learning, which often suffer from the out-of-distribution and overfitting problems. Meanwhile, they are incompetent at estimating uncertainties in predictions, which is critically important as model predictions could be very unreliable in few-shot settings. Furthermore, most of them cannot handle complex relations and ignore path information in KGs, which largely limits their performance. In this paper, we propose a normalizing flow-based neural process for few-shot knowledge graph completion (NP-FKGC). Specifically, we unify normalizing flows and neural processes to model a complex distribution of KG completion functions. This offers a novel way to predict facts for few-shot relations while estimating the uncertainty. Then, we propose a stochastic ManifoldE decoder to incorporate the neural process and handle complex relations in few-shot settings. To further improve performance, we introduce an attentive relation path-based graph neural network to capture path information in KGs. Extensive experiments on three public datasets demonstrate that our method significantly outperforms the existing FKGC methods and achieves state-of-the-art performance. Code is available at https://github.com/RManLuo/NP-FKGC.git. | ['Shirui Pan', 'Gholamreza Haffari', 'Yuan-Fang Li', 'Linhao Luo'] | 2023-04-17 | null | null | null | null | ['metric-learning', 'knowledge-graph-completion', 'metric-learning'] | ['computer-vision', 'knowledge-base', 'methodology'] | [-3.97726238e-01 2.63840973e-01 -4.08279121e-01 -3.43982011e-01
-6.63680553e-01 -1.91608235e-01 3.55606586e-01 9.39428732e-02
1.03860654e-01 8.63309085e-01 3.39254081e-01 -2.18710989e-01
-4.27220285e-01 -1.09140825e+00 -7.80139208e-01 -4.26507384e-01
1.60832465e-01 5.24241328e-01 2.64405936e-01 -3.95879507e-01
-1.24914110e-01 -7.30574951e-02 -1.04593742e+00 -3.06241103e-02
1.23000348e+00 9.15351808e-01 -3.21504823e-03 3.41717273e-01
-2.63964683e-01 1.17791259e+00 -3.49214792e-01 -1.08775294e+00
1.67550314e-02 -2.60905057e-01 -8.50159466e-01 -2.31048986e-01
4.91449311e-02 -3.15608412e-01 -1.05798471e+00 1.30728292e+00
4.46362376e-01 4.84059215e-01 6.88130200e-01 -1.38705719e+00
-1.18381429e+00 1.00889170e+00 -4.46207970e-01 4.31441247e-01
3.27911615e-01 8.38065818e-02 1.17687082e+00 -9.81501937e-01
4.06900078e-01 1.31265426e+00 5.85181057e-01 2.82123268e-01
-9.03153837e-01 -6.82164252e-01 3.32587868e-01 7.25065768e-01
-1.68035948e+00 -3.75878960e-01 7.02406526e-01 -2.73076087e-01
9.20312941e-01 -4.34959717e-02 4.84710544e-01 1.05348873e+00
6.49294257e-02 9.19251084e-01 3.96115452e-01 -2.16302369e-02
2.58628607e-01 -2.40602463e-01 1.02234058e-01 7.85971344e-01
4.45167124e-01 -6.49421588e-02 -6.69217944e-01 2.32379697e-02
7.00932801e-01 1.04539141e-01 -5.22490203e-01 -2.07369864e-01
-1.09677923e+00 7.46430039e-01 6.03961587e-01 7.87409544e-02
-1.75820217e-01 2.36508206e-01 3.20325971e-01 2.99115360e-01
8.13227057e-01 1.36986285e-01 -4.27098691e-01 -2.90747643e-01
-5.58881760e-01 1.02989241e-01 9.42099035e-01 1.39968061e+00
6.95882320e-01 5.00170290e-02 -4.12168533e-01 8.68159831e-01
2.16515690e-01 2.72441745e-01 2.41265923e-01 -6.38749063e-01
8.81487191e-01 8.21130514e-01 -1.08853452e-01 -1.32172000e+00
-4.27786678e-01 -5.84607661e-01 -1.11513126e+00 -5.31009197e-01
1.65057421e-01 -1.58458591e-01 -9.69012082e-01 1.57045722e+00
3.72936934e-01 7.91009963e-01 2.69074529e-01 9.27229404e-01
1.21088791e+00 7.45402038e-01 -2.33348459e-02 -1.89584076e-01
1.10129964e+00 -1.12170947e+00 -8.26786280e-01 -2.08798841e-01
7.31675208e-01 -3.19884241e-01 1.14113498e+00 1.36873871e-01
-8.00025344e-01 -3.23830068e-01 -9.41095531e-01 -2.71621764e-01
-4.88857061e-01 -6.95947707e-02 9.24833119e-01 3.97540480e-01
-7.08966017e-01 6.86504543e-01 -9.11905527e-01 -4.67398353e-02
7.44036913e-01 1.02876984e-01 -9.14002061e-02 -5.26765645e-01
-1.72637033e+00 7.84750760e-01 9.02165234e-01 2.66250223e-01
-7.23278999e-01 -8.14262033e-01 -1.18420696e+00 4.94985223e-01
1.31719375e+00 -7.70932853e-01 1.11597431e+00 -2.34072525e-02
-1.35499752e+00 2.06767261e-01 -4.46159393e-02 -3.08405608e-01
3.77864629e-01 -1.54686555e-01 -8.36454391e-01 -3.63171510e-02
1.13383442e-01 1.18637443e-01 5.58449030e-01 -9.39395905e-01
-4.88473564e-01 -1.98357761e-01 3.65986973e-01 2.81093925e-01
-3.12180996e-01 -4.43971753e-01 -8.77030849e-01 -7.66979218e-01
1.31774470e-01 -4.42230135e-01 -9.00458246e-02 -4.40510392e-01
-7.83015013e-01 -3.51352215e-01 4.38119799e-01 -5.07995009e-01
1.46466374e+00 -1.95505071e+00 2.33882234e-01 -5.43096475e-03
4.62553769e-01 4.69099313e-01 -3.46928909e-02 3.94755393e-01
2.03532934e-01 1.43206775e-01 -3.01672280e-01 -2.82829285e-01
1.58488587e-01 4.17303026e-01 -3.34239066e-01 1.32855222e-01
3.21091741e-01 1.43487108e+00 -1.27074039e+00 -4.69361514e-01
8.31837654e-02 4.48909581e-01 -1.95796877e-01 3.04407831e-02
-5.01200676e-01 1.63627774e-01 -6.25711739e-01 8.92513990e-01
6.53978884e-01 -6.59582198e-01 2.37575117e-02 -2.10902229e-01
5.77304780e-01 7.26437345e-02 -1.08210194e+00 1.92522538e+00
-1.76078886e-01 1.43786684e-01 -5.38262248e-01 -1.09827912e+00
8.53903413e-01 2.25444704e-01 1.35865316e-01 -3.30437720e-01
3.12985688e-01 7.67950863e-02 -1.05756529e-01 -4.93273467e-01
5.43477356e-01 -2.20522583e-01 4.38359864e-02 5.19462079e-02
4.27096397e-01 -7.50101451e-03 3.58503103e-01 7.57101953e-01
1.14343572e+00 4.82040904e-02 2.36942843e-01 1.93782747e-01
4.75025445e-01 -1.72782034e-01 9.95584190e-01 4.70615983e-01
-1.83044881e-01 5.41969597e-01 8.40988517e-01 -2.11354509e-01
-4.75828290e-01 -1.08783174e+00 2.47876167e-01 7.55849600e-01
5.52218556e-01 -7.47309566e-01 -2.73312926e-01 -8.45554829e-01
1.38420805e-01 9.91087794e-01 -4.89609718e-01 -6.28755271e-01
-5.78044541e-02 -8.05710793e-01 4.51892555e-01 6.70828581e-01
5.70042372e-01 -7.42750704e-01 3.50643575e-01 3.82145345e-01
-4.52185869e-01 -1.36382544e+00 -5.25803804e-01 -1.73853040e-01
-6.18024409e-01 -1.29565251e+00 -6.23763561e-01 -5.28024256e-01
4.74708021e-01 3.36536169e-01 1.13315928e+00 -4.69588824e-02
-8.75564590e-02 2.83764631e-01 -6.03017807e-01 -3.04918170e-01
-3.99888344e-02 1.52832583e-01 7.08699822e-02 -9.51978043e-02
6.31565332e-01 -7.92634130e-01 -4.47028399e-01 2.53093541e-01
-8.11954737e-01 5.42643555e-02 4.97438163e-01 9.33030009e-01
6.69979393e-01 4.79272664e-01 9.48059857e-01 -1.17583156e+00
8.12095404e-01 -8.77773941e-01 -5.26764989e-01 6.06998265e-01
-7.39015043e-01 6.23417795e-02 7.72475243e-01 -3.07516724e-01
-1.20538318e+00 -3.77717763e-01 7.00782463e-02 -9.05208766e-01
2.65736848e-01 1.02516580e+00 -4.25398380e-01 4.18636091e-02
5.39560437e-01 1.61781520e-01 -3.75704318e-01 -2.61698604e-01
7.13995695e-01 2.86284208e-01 6.14427745e-01 -5.64833462e-01
8.12866569e-01 2.04853028e-01 -2.87555959e-02 -2.82227993e-01
-1.45452225e+00 -4.93900031e-01 -3.80144149e-01 -1.07150003e-01
4.46689636e-01 -1.06757247e+00 -6.48881078e-01 3.16415399e-01
-1.05181038e+00 -1.56041324e-01 -3.00076216e-01 6.62445903e-01
-4.45933610e-01 4.69856203e-01 -7.25522995e-01 -7.25467026e-01
-3.59475106e-01 -8.61763239e-01 7.25423157e-01 5.54113030e-01
3.51233870e-01 -1.11147451e+00 -7.51309693e-02 5.28027654e-01
2.30822295e-01 1.75153211e-01 8.10451210e-01 -6.10537469e-01
-7.06509471e-01 -1.26443520e-01 -5.05818725e-01 2.29404539e-01
1.88997403e-01 -2.80145556e-01 -7.34223723e-01 1.00119682e-02
-2.85739928e-01 -4.89016175e-01 1.11839020e+00 2.42739990e-01
1.13841987e+00 -2.37425134e-01 -3.91167969e-01 6.94368064e-01
1.34811819e+00 -5.65805621e-02 6.67631149e-01 -1.34508222e-01
1.17519343e+00 3.25604171e-01 7.40510762e-01 4.72095132e-01
9.52577055e-01 2.61863738e-01 2.85801291e-01 4.85939085e-01
-9.58372504e-02 -5.79225004e-01 1.37949765e-01 1.19488525e+00
-2.18861684e-01 -4.89721328e-01 -9.96769965e-01 5.91453612e-01
-2.31456017e+00 -8.54628742e-01 7.55977165e-03 1.81590950e+00
8.62841129e-01 2.96078235e-01 -1.40544727e-01 -1.49436370e-01
9.69131351e-01 2.60429442e-01 -8.67822409e-01 9.85972285e-02
-1.32449746e-01 -1.23863198e-01 2.60502845e-01 3.95739198e-01
-9.24859703e-01 1.30545866e+00 4.81047869e+00 1.26089311e+00
-4.46978718e-01 2.41889998e-01 6.09026253e-01 -1.84538618e-01
-4.27314937e-01 1.04636125e-01 -7.88107336e-01 6.59029067e-01
7.91425288e-01 -5.30141890e-01 4.44477767e-01 8.22507381e-01
-1.85796216e-01 -2.25657262e-02 -8.94381464e-01 1.19991684e+00
1.10768326e-01 -1.33580065e+00 2.56447103e-02 -2.77524829e-01
9.05060410e-01 -7.52497837e-02 -2.17584759e-01 9.16262090e-01
5.40227473e-01 -9.36898470e-01 3.05714548e-01 8.30999970e-01
7.82947481e-01 -9.20523405e-01 8.73324811e-01 3.91914427e-01
-1.46085477e+00 3.96001292e-03 -8.16615462e-01 -7.43169412e-02
4.78778273e-01 1.06105494e+00 -6.45964444e-01 1.16548026e+00
4.39617962e-01 9.90324378e-01 -3.46383423e-01 1.04925323e+00
-6.14912093e-01 6.45376027e-01 -1.01920389e-01 -6.82254732e-02
1.04598545e-01 -1.60002902e-01 4.71281707e-01 8.29746723e-01
2.83861786e-01 5.33380806e-01 1.89553395e-01 1.07573545e+00
-4.32365835e-01 1.31179867e-02 -4.67624098e-01 -3.34710419e-01
7.09235013e-01 1.18884802e+00 -6.68006241e-01 -4.52996492e-01
-6.18055880e-01 8.51839244e-01 7.43833244e-01 5.27276576e-01
-8.64580929e-01 -5.63019276e-01 4.53021556e-01 -3.13264698e-01
3.17846626e-01 -9.11393017e-02 -8.68872181e-02 -1.78839338e+00
2.90814102e-01 -4.68264550e-01 6.96754813e-01 -7.35996366e-01
-1.67008066e+00 3.99636030e-01 3.05357464e-02 -1.05482852e+00
-9.07368436e-02 -3.04514110e-01 -6.65623963e-01 6.57540560e-01
-1.64509881e+00 -1.23591149e+00 -3.95662189e-01 6.39141321e-01
2.95512974e-01 -2.15254590e-01 4.60558236e-01 3.95990998e-01
-8.15756917e-01 6.13613605e-01 6.65899590e-02 1.82152063e-01
4.88156587e-01 -1.24686003e+00 5.50043821e-01 9.32204127e-01
1.84246659e-01 4.49315637e-01 3.85702819e-01 -1.05552113e+00
-1.45105052e+00 -1.35101616e+00 5.77288985e-01 -3.07059914e-01
9.51579928e-01 -1.67551756e-01 -1.18787289e+00 7.58370042e-01
-3.46217960e-01 3.53619426e-01 6.44262075e-01 4.80213344e-01
-4.82881278e-01 -1.35021180e-01 -8.88858438e-01 6.08620405e-01
1.37336278e+00 -5.59951305e-01 -7.16893137e-01 5.89469373e-01
1.18431258e+00 -5.73203325e-01 -1.04665244e+00 4.96202707e-01
1.03058390e-01 -7.33669877e-01 7.95698702e-01 -6.27695739e-01
5.59104562e-01 -2.75478095e-01 -6.45560771e-02 -1.61236989e+00
-3.78418535e-01 -5.70146799e-01 -1.02174413e+00 1.44810891e+00
4.99689221e-01 -7.82507718e-01 8.26436281e-01 6.91108584e-01
-2.33260363e-01 -1.13461053e+00 -7.06348121e-01 -1.03854597e+00
-1.24027565e-01 -6.19996250e-01 8.64925921e-01 1.05659962e+00
2.79009372e-01 5.94314933e-01 -5.84559023e-01 2.96209753e-01
5.71333587e-01 1.46376982e-01 6.03406131e-01 -1.15569043e+00
-3.24837387e-01 -2.21522942e-01 -7.19495356e-01 -9.03211892e-01
2.67972738e-01 -1.10203278e+00 -6.15373030e-02 -1.94274056e+00
2.48438224e-01 -3.48767549e-01 -3.42580676e-01 6.37323618e-01
-7.53410697e-01 -2.10398883e-01 8.46374966e-03 1.67887539e-01
-9.39998388e-01 1.12518418e+00 1.35546017e+00 -2.03324139e-01
-1.00257948e-01 -4.95382100e-02 -1.01028430e+00 5.43758392e-01
8.90848577e-01 -3.70383650e-01 -8.76705348e-01 -3.37349802e-01
6.27159238e-01 1.67025656e-01 2.65608042e-01 -9.20996308e-01
6.95995331e-01 -2.79835090e-02 1.10841341e-01 -6.80534780e-01
4.33587521e-01 -4.32223529e-01 -3.78703959e-02 -1.09230196e-02
-1.34226764e-02 -3.89553577e-01 -7.87539035e-03 1.20345235e+00
-3.69988263e-01 -4.31930795e-02 2.15890080e-01 -9.98293832e-02
-9.94964242e-01 8.73516321e-01 3.26025814e-01 5.13765574e-01
9.28714991e-01 1.71456754e-01 -7.74478674e-01 -5.78397512e-01
-7.09528506e-01 8.56959522e-01 -1.88394468e-02 3.56134325e-01
8.79429102e-01 -1.56161201e+00 -6.66634738e-01 -2.98529774e-01
3.93435508e-01 6.05485559e-01 6.73215926e-01 8.96802306e-01
-1.74012274e-01 3.18371117e-01 3.74408185e-01 -5.70877790e-02
-5.18889785e-01 8.28719020e-01 9.64555889e-02 -3.36035281e-01
-8.38446558e-01 1.16315556e+00 2.24681315e-03 -4.66983914e-01
1.83426484e-01 -3.40486467e-01 -1.14827752e-01 -9.76019651e-02
5.63185036e-01 5.31232178e-01 -4.16481011e-02 -3.74963105e-01
-1.62380308e-01 1.84269413e-01 -3.18098634e-01 3.57021242e-01
1.23038864e+00 -2.39424869e-01 6.66294768e-02 5.63771367e-01
9.17327702e-01 -3.37931335e-01 -1.18203616e+00 -8.03624451e-01
-1.65143516e-02 -5.39836884e-01 1.00986555e-01 -7.51539767e-01
-1.25668848e+00 7.17746079e-01 -1.98231533e-01 1.71173096e-01
9.79303420e-01 2.56033391e-01 1.04032922e+00 5.60840845e-01
5.17286301e-01 -1.05727899e+00 -4.05749194e-02 5.99881709e-01
6.70073211e-01 -1.45198917e+00 1.17419764e-01 -9.36172366e-01
-8.11092257e-01 9.82565999e-01 7.88267434e-01 1.44091025e-01
7.73322940e-01 -8.89998227e-02 -3.74155611e-01 -5.39982557e-01
-8.72894764e-01 -3.78945351e-01 4.18180823e-01 5.18019557e-01
1.97187886e-02 2.28538603e-01 -1.37117594e-01 1.16915596e+00
-1.64425522e-01 2.10812405e-01 5.04445195e-01 6.97162926e-01
-2.81120121e-01 -7.02403307e-01 2.51937564e-02 8.77483606e-01
-1.82574764e-01 -4.35652763e-01 -1.38050795e-01 5.22621572e-01
-3.98779921e-02 1.12299871e+00 -4.11249787e-01 -7.05090582e-01
3.08067977e-01 -2.27049049e-02 2.36891255e-01 -7.51632869e-01
7.90653080e-02 -2.67346710e-01 9.06639025e-02 -4.73791927e-01
-1.77560925e-01 -2.40891948e-01 -1.29066646e+00 -5.99132001e-01
-6.98138714e-01 1.73042029e-01 3.13240737e-02 1.10055256e+00
5.82757652e-01 8.02332819e-01 2.60907501e-01 -3.01653594e-01
-5.08484781e-01 -1.01144922e+00 -9.99160647e-01 1.48743689e-01
-1.35013729e-01 -1.15527475e+00 -2.54722476e-01 -3.41442287e-01] | [8.794621467590332, 7.955242156982422] |
283ca44c-4281-45d3-980a-0019b82b2208 | multi-scale-prediction-for-robust-hand | 1804.08220 | null | http://arxiv.org/abs/1804.08220v1 | http://arxiv.org/pdf/1804.08220v1.pdf | Multi-scale prediction for robust hand detection and classification | In this paper, we present a multi-scale Fully Convolutional Networks
(MSP-RFCN) to robustly detect and classify human hands under various
challenging conditions. In our approach, the input image is passed through the
proposed network to generate score maps, based on multi-scale predictions. The
network has been specifically designed to deal with small objects. It uses an
architecture based on region proposals generated at multiple scales. Our method
is evaluated on challenging hand datasets, namely the Vision for Intelligent
Vehicles and Applications (VIVA) Challenge and the Oxford hand dataset. It is
compared against recent hand detection algorithms. The experimental results
demonstrate that our proposed method achieves state-of-the-art detection for
hands of various sizes. | ['Robert Laganiere', 'Yong Wang', 'Xinbin Luo', 'Ding Lu', 'Shan Fu'] | 2018-04-23 | null | null | null | null | ['hand-detection'] | ['computer-vision'] | [ 9.22990069e-02 -3.54341000e-01 6.31992072e-02 -6.09642938e-02
-2.40174875e-01 -5.06032467e-01 4.78000522e-01 -7.43550956e-01
-6.87724113e-01 4.31718498e-01 -1.98807344e-01 1.07941590e-01
2.41837949e-01 -5.84697425e-01 -4.87646520e-01 -5.01488924e-01
2.40767822e-02 5.36667168e-01 8.35631073e-01 -3.25931579e-01
1.99180707e-01 1.08754468e+00 -1.51083517e+00 5.85501552e-01
2.71422178e-01 9.28600430e-01 5.52708924e-01 1.10867310e+00
4.65139061e-01 7.07549274e-01 -7.67323256e-01 -3.00802380e-01
6.19500458e-01 1.39640540e-01 -7.85449743e-01 -1.61108091e-01
6.97397768e-01 -7.08803475e-01 -6.79129839e-01 8.18082929e-01
1.00533152e+00 -7.12259114e-02 8.12056124e-01 -1.21968460e+00
-3.15421224e-01 3.35925430e-01 -7.37332523e-01 5.59124112e-01
3.78022075e-01 5.74360669e-01 7.12068141e-01 -1.14726949e+00
8.85674894e-01 1.70707452e+00 8.54555607e-01 5.14151275e-01
-8.41105998e-01 -6.51726723e-01 1.29719689e-01 3.57704371e-01
-1.77866006e+00 -2.22299825e-02 5.02886057e-01 -5.45481026e-01
1.14258933e+00 -1.72201768e-01 7.10959673e-01 1.41513658e+00
3.82366598e-01 1.29553223e+00 1.03651786e+00 -4.43007261e-01
-4.12507914e-02 -2.01649461e-02 -7.69846812e-02 1.11230469e+00
1.76706344e-01 2.65560895e-01 -2.12386310e-01 -2.74866093e-02
1.23032415e+00 -1.17854781e-01 1.17740661e-01 -4.20322061e-01
-1.36700523e+00 6.82783961e-01 7.72780299e-01 3.48060578e-01
-7.66025066e-01 1.69550493e-01 4.85044897e-01 -2.37625182e-01
2.05086134e-02 -3.01268697e-02 -4.38415170e-01 4.50980663e-01
-1.18784928e+00 6.25534892e-01 5.91440916e-01 9.62653875e-01
-6.13690093e-02 -7.38988072e-02 -8.41070175e-01 5.82093060e-01
5.14422655e-01 7.76628375e-01 4.98372391e-02 -6.34993613e-01
5.81631422e-01 5.18145323e-01 3.10426831e-01 -9.04486895e-01
-8.44227433e-01 -4.50310171e-01 -8.03911805e-01 8.82421613e-01
8.36926043e-01 -1.84829816e-01 -1.38629138e+00 1.23662329e+00
2.55910456e-01 -1.89712942e-01 -1.91160619e-01 1.34790742e+00
1.09491789e+00 3.41496468e-01 2.14059070e-01 4.85660136e-01
1.42266381e+00 -1.30375814e+00 -3.31014842e-01 -3.02414864e-01
-1.44088164e-01 -8.99372756e-01 8.28576326e-01 6.63793981e-01
-9.07684565e-01 -1.20816886e+00 -9.69184220e-01 1.82732955e-01
-5.69046617e-01 9.91965592e-01 2.30344981e-01 7.37736702e-01
-1.19651222e+00 4.01630998e-01 -6.92839742e-01 -8.58529627e-01
7.08223104e-01 3.69620264e-01 -2.29342431e-01 2.28270829e-01
-8.80046546e-01 1.09273612e+00 5.35114527e-01 4.81548309e-01
-1.12657690e+00 2.44868658e-02 -4.28254455e-01 -4.85826703e-03
3.16290528e-01 -7.21740723e-01 1.04221392e+00 -6.65572286e-01
-1.22418046e+00 9.06131744e-01 1.36883631e-01 -4.44812924e-01
1.06387627e+00 -2.43026197e-01 -1.76963225e-01 2.51469284e-01
1.06738739e-01 9.39456701e-01 1.11819685e+00 -9.40526009e-01
-9.90284085e-01 -4.88649786e-01 8.63522962e-02 -1.71525523e-01
1.67290047e-01 6.18334889e-01 -6.02291405e-01 -8.38791847e-01
-3.95199031e-01 -9.71288800e-01 -2.80214518e-01 2.66299665e-01
-6.52134776e-01 -7.02394307e-01 9.16214228e-01 -8.36762667e-01
7.48730600e-01 -1.43796277e+00 4.18310761e-01 3.99167597e-01
1.97756469e-01 7.62832701e-01 -4.46987987e-01 -3.49558890e-03
4.13540184e-01 -5.05916476e-01 6.96803182e-02 -2.46657103e-01
-3.14580649e-02 -1.87931895e-01 3.85868661e-02 5.32069147e-01
3.21571350e-01 1.03525674e+00 -6.29619181e-01 -8.55935812e-01
7.27934420e-01 8.89453471e-01 -2.06081629e-01 3.08244824e-01
-8.22371095e-02 1.01075239e-01 -4.18209851e-01 9.84387517e-01
6.58401966e-01 -1.79357603e-01 -3.74513380e-02 -3.66145611e-01
-1.25346124e-01 -6.01811290e-01 -1.46249878e+00 1.64159584e+00
-2.28628054e-01 6.92802191e-01 1.02047697e-01 -3.74486297e-01
5.61832309e-01 2.33146638e-01 -1.08427823e-01 -1.65000498e-01
5.89030981e-01 -2.98537239e-02 1.62610278e-01 -3.63891333e-01
3.02103937e-01 5.74470758e-01 3.19364429e-01 2.46237278e-01
4.58757281e-01 3.63404602e-01 3.71102929e-01 1.18666500e-01
1.00025284e+00 4.54542786e-01 3.90628427e-01 -1.74637049e-01
8.62259150e-01 -1.21882729e-01 1.06143296e-01 1.07658744e+00
-7.12750077e-01 7.34383643e-01 1.89515382e-01 -7.33627081e-01
-1.24953020e+00 -1.10940313e+00 -1.62099302e-02 1.51082790e+00
-1.17085300e-01 5.86960576e-02 -1.05289507e+00 -1.01758194e+00
2.36209497e-01 -1.55866683e-01 -9.36383247e-01 4.83538002e-01
-7.36035585e-01 -3.20232242e-01 7.93742836e-01 1.33347642e+00
6.77751958e-01 -1.84316635e+00 -1.03608763e+00 2.51386464e-01
1.81536809e-01 -1.11240828e+00 -4.49825436e-01 -1.76805660e-01
-9.99671817e-02 -1.41678941e+00 -1.47149646e+00 -1.03336275e+00
3.71563315e-01 1.87662438e-01 7.80897260e-01 -1.75255805e-01
-1.29118145e+00 4.79012221e-01 -3.25315684e-01 -5.89085281e-01
-2.68919259e-01 3.77871424e-01 2.89197862e-01 -1.74000427e-01
2.90868849e-01 3.24024097e-03 -1.02785087e+00 5.71919620e-01
-2.77516603e-01 -4.22967315e-01 1.07395506e+00 7.92778969e-01
3.26631546e-01 -4.65952039e-01 4.97333646e-01 -3.61755848e-01
5.74715853e-01 1.53761715e-01 -6.99341238e-01 6.63571417e-01
-2.76576400e-01 -8.24472383e-02 3.47273201e-01 -5.29388130e-01
-1.19450712e+00 7.64987826e-01 -2.44468957e-01 -5.02490282e-01
-3.35454315e-01 -5.72335958e-01 1.56915933e-01 -6.26247585e-01
9.49991107e-01 1.78852066e-01 -2.73353815e-01 -3.37186128e-01
6.42478943e-01 1.06955528e+00 6.87311530e-01 -2.48653352e-01
4.82683778e-01 5.41467488e-01 -7.83483461e-02 -6.89333081e-01
-5.72811782e-01 -6.19196713e-01 -1.37557840e+00 -7.19026566e-01
1.07650292e+00 -8.02818000e-01 -6.83444738e-01 8.73890758e-01
-1.51350808e+00 -3.07087719e-01 1.69149458e-01 2.29748487e-01
-5.46683609e-01 2.60052621e-01 -6.63766682e-01 -1.12074065e+00
-1.03158081e+00 -1.15307462e+00 1.33752203e+00 3.31129283e-01
-8.48621428e-02 -6.08492076e-01 -8.50229785e-02 9.05175358e-02
4.69303817e-01 1.97517261e-01 2.59003192e-02 -4.37110633e-01
-4.67094421e-01 -3.55337679e-01 -6.97079122e-01 2.86033630e-01
-1.50649250e-01 1.96961358e-01 -1.27493238e+00 -7.46627927e-01
-9.59719658e-01 -4.97897208e-01 1.09145725e+00 5.86099029e-01
1.10855103e+00 1.35292396e-01 -6.71157658e-01 3.15763593e-01
1.04565811e+00 -1.39205419e-02 2.85428852e-01 2.81665802e-01
6.90646291e-01 4.18519020e-01 6.08339846e-01 7.44262278e-01
1.09378763e-01 6.61789536e-01 3.24805945e-01 -3.15727830e-01
-6.53936148e-01 1.71125814e-01 3.20808101e-03 -3.23838115e-01
-8.83404255e-01 -8.95716324e-02 -9.06834722e-01 5.76325417e-01
-1.87203813e+00 -1.01607156e+00 9.84263718e-02 1.58983445e+00
2.46180028e-01 2.00039923e-01 9.13056254e-01 6.55622333e-02
1.07097888e+00 1.33162916e-01 -6.13237619e-01 4.51411083e-02
-1.15020066e-01 6.91719055e-01 5.86758614e-01 7.20930919e-02
-1.61977565e+00 1.26284933e+00 7.00727177e+00 5.26393533e-01
-8.98551702e-01 2.15430990e-01 5.32877855e-02 -1.11365266e-01
1.06321847e+00 -7.90908575e-01 -9.30183172e-01 5.22657484e-02
1.67298853e-01 3.29464227e-01 4.04344350e-01 1.28518403e+00
-1.46455094e-01 2.23098725e-01 -9.20277715e-01 9.60878253e-01
2.99014956e-01 -1.03445709e+00 2.63962969e-02 -3.24683368e-01
5.16066492e-01 2.99428880e-01 1.67657912e-01 2.25659400e-01
2.58564532e-01 -9.84640360e-01 1.03008485e+00 4.42109853e-01
9.13375854e-01 -7.94543326e-01 9.01468813e-01 8.90083984e-02
-1.67585504e+00 -3.35353136e-01 -5.34281850e-01 1.12178057e-01
-5.58473989e-02 -2.69120425e-01 -1.08071351e+00 -6.52216598e-02
1.01173413e+00 2.10669234e-01 -8.20221066e-01 1.11062479e+00
-3.92537922e-01 7.42463842e-02 -1.28596619e-01 -3.57051015e-01
1.99083298e-01 5.20577550e-01 4.26085263e-01 1.86926901e+00
-2.32006118e-04 -7.07626343e-04 3.55024934e-01 9.55557108e-01
1.65379047e-01 8.06152821e-03 -3.93638253e-01 3.46733421e-01
3.09333444e-01 1.58285904e+00 -1.09790480e+00 -5.54263651e-01
-2.58703858e-01 1.29836321e+00 2.93939084e-01 2.62608945e-01
-7.30560958e-01 -8.59699190e-01 2.89547592e-01 1.03058983e-02
8.70139658e-01 -7.96164498e-02 5.50062023e-02 -8.32154036e-01
1.98447526e-01 -4.34757650e-01 4.18998539e-01 -8.05481672e-01
-1.29511428e+00 9.24738169e-01 -3.51472110e-01 -7.43985593e-01
-2.76616812e-01 -1.26091099e+00 -5.73815048e-01 8.54496300e-01
-1.26895118e+00 -1.79874635e+00 -6.32156074e-01 9.10687745e-01
7.98168480e-01 -6.33837402e-01 6.09211922e-01 1.61600381e-01
-3.70341003e-01 8.42445016e-01 -2.20209524e-01 6.98393583e-01
6.85771108e-01 -1.28451216e+00 8.08656394e-01 6.99872673e-01
-3.83961275e-02 4.36755598e-01 5.00831485e-01 -6.69920743e-01
-1.06111288e+00 -1.26919055e+00 1.95072964e-01 -7.41952658e-01
2.72640467e-01 -3.09746742e-01 -4.24051762e-01 6.01345837e-01
2.80940324e-01 4.01104420e-01 2.11923942e-02 -7.26287514e-02
-5.39301038e-01 1.22882992e-01 -1.57423043e+00 3.69495377e-02
1.29230201e+00 -2.84155160e-01 -6.81745410e-01 4.45983618e-01
6.17500395e-02 -4.70462918e-01 -6.27066731e-01 4.67286319e-01
1.30688250e+00 -8.80205870e-01 1.26141608e+00 -8.18428874e-01
1.20437950e-01 -4.05817389e-01 -5.05966600e-03 -1.01021361e+00
-6.61875725e-01 -9.07532498e-02 -3.51130694e-01 6.40779734e-01
1.13764815e-01 -1.28742784e-01 1.03135729e+00 -1.16108522e-01
4.78860021e-01 -4.87631589e-01 -9.31412160e-01 -7.75927544e-01
-1.99773371e-01 -8.95140767e-02 2.35215828e-01 1.73112884e-01
-4.37960029e-01 1.08037621e-01 -3.09632033e-01 4.20910478e-01
1.15654480e+00 4.27330472e-02 8.18115354e-01 -1.36967504e+00
-1.35676488e-01 -5.34893692e-01 -8.22104692e-01 -7.13921607e-01
2.05240399e-02 -4.89070058e-01 1.83809653e-01 -1.71130908e+00
5.35180807e-01 -8.90117586e-02 -4.35794979e-01 4.76549983e-01
-1.77500829e-01 7.11904824e-01 5.39758980e-01 5.41538075e-02
-8.41168702e-01 -4.66662794e-02 1.02945828e+00 -3.86598855e-01
-1.78057060e-01 3.52450579e-01 7.97073171e-02 8.29873085e-01
7.75884569e-01 4.35118005e-02 2.38433838e-01 7.64394226e-03
-4.02547956e-01 -2.38175228e-01 8.09552312e-01 -1.50457788e+00
1.98516577e-01 1.51230484e-01 1.12341452e+00 -1.15959692e+00
2.89090693e-01 -5.29461801e-01 -3.62806380e-01 1.01183617e+00
-2.08825409e-01 -1.16167255e-01 1.31254703e-01 5.11456370e-01
3.18576515e-01 2.20903188e-01 1.12179244e+00 -2.46196270e-01
-1.01518714e+00 1.62771463e-01 -2.42910311e-01 -4.01301205e-01
1.19356048e+00 -6.27984107e-02 -1.94032583e-02 2.22723380e-01
-8.95319521e-01 1.76990539e-01 6.07076026e-02 8.19615304e-01
8.83938074e-01 -1.24990141e+00 -9.32148218e-01 1.70022979e-01
2.16929510e-01 -1.83526114e-01 2.03726932e-01 3.65619808e-01
-7.05097616e-01 8.68008673e-01 -6.81102872e-01 -7.03398049e-01
-1.80557716e+00 6.43077075e-01 4.22777772e-01 -3.93714160e-02
-6.88785315e-01 1.10487711e+00 1.61534082e-02 -5.55380642e-01
8.17404211e-01 -5.47931552e-01 -2.88035750e-01 -1.36868030e-01
8.86622667e-01 7.76091933e-01 -1.33439228e-01 -8.01367104e-01
-6.42627358e-01 7.73884475e-01 -1.33375257e-01 1.97322816e-01
1.08062041e+00 4.08642560e-01 1.78483039e-01 -1.77842319e-01
7.31658399e-01 -5.07579982e-01 -1.46158302e+00 -3.24288905e-01
-3.37718189e-01 -4.52039570e-01 -1.59076825e-01 -1.18661380e+00
-1.11059928e+00 1.01423323e+00 1.49449193e+00 -2.87527412e-01
6.87206626e-01 2.64965236e-01 6.63726568e-01 8.20078313e-01
6.47473574e-01 -1.28330946e+00 2.14510649e-01 3.36839795e-01
1.19692183e+00 -1.34511197e+00 -8.28235447e-02 -1.49341181e-01
-4.75670069e-01 1.32523596e+00 9.44847107e-01 -2.57629722e-01
3.24446619e-01 2.51614362e-01 5.95069267e-02 -1.45860970e-01
-4.53718230e-02 -6.81639433e-01 4.13376510e-01 8.47151935e-01
8.44974220e-02 2.79465258e-01 -6.29511178e-02 4.14473236e-01
2.42316887e-01 3.88640612e-01 -1.06914835e-02 1.00332248e+00
-7.35620618e-01 -6.82484448e-01 -8.66543531e-01 4.73072052e-01
-2.73353904e-01 2.78432667e-01 -8.29918444e-01 8.87542307e-01
5.07993698e-01 7.69507289e-01 -3.54969382e-01 -5.05730510e-01
6.26377642e-01 -2.95422003e-02 1.00062895e+00 -3.49800348e-01
-6.05363131e-01 -2.34955043e-01 -3.20999414e-01 -7.82025337e-01
-5.47682047e-01 -4.70918983e-01 -9.96494949e-01 -9.72523764e-02
-4.41882789e-01 -6.10832453e-01 6.59781039e-01 7.72188425e-01
1.84140846e-01 7.94034660e-01 2.38755643e-01 -1.58380198e+00
-8.39689255e-01 -1.40118158e+00 -7.30554223e-01 2.93586822e-03
2.75277406e-01 -1.12115240e+00 1.49906471e-01 -2.46902734e-01] | [6.577028274536133, -0.6516555547714233] |
a7532e54-b0ac-4b6b-aa0a-0b128c3d75e8 | convergence-and-complexity-of-stochastic | 2201.01652 | null | https://arxiv.org/abs/2201.01652v3 | https://arxiv.org/pdf/2201.01652v3.pdf | Stochastic regularized majorization-minimization with weakly convex and multi-convex surrogates | Stochastic majorization-minimization (SMM) is a class of stochastic optimization algorithms that proceed by sampling new data points and minimizing a recursive average of surrogate functions of an objective function. The surrogates are required to be strongly convex and convergence rate analysis for the general non-convex setting was not available. In this paper, we propose an extension of SMM where surrogates are allowed to be only weakly convex or block multi-convex, and the averaged surrogates are approximately minimized with proximal regularization or block-minimized within diminishing radii, respectively. For the general nonconvex constrained setting with non-i.i.d. data samples, we show that the first-order optimality gap of the proposed algorithm decays at the rate $O((\log n)^{1+\epsilon}/n^{1/2})$ for the empirical loss and $O((\log n)^{1+\epsilon}/n^{1/4})$ for the expected loss, where $n$ denotes the number of data samples processed. Under some additional assumption, the latter convergence rate can be improved to $O((\log n)^{1+\epsilon}/n^{1/2})$. As a corollary, we obtain the first convergence rate bounds for various optimization methods under general nonconvex dependent data setting: Double-averaging projected gradient descent and its generalizations, proximal point empirical risk minimization, and online matrix/tensor decomposition algorithms. We also provide experimental validation of our results. | ['Hanbaek Lyu'] | 2022-01-05 | null | null | null | null | ['image-deep-networks'] | ['computer-vision'] | [-7.93758966e-03 9.47860107e-02 -4.41670902e-02 -2.79324949e-01
-1.22895110e+00 -4.73504990e-01 -2.40561202e-01 1.06120229e-01
-7.87964582e-01 9.32162225e-01 -1.33092269e-01 -3.51964295e-01
-5.00710726e-01 -3.95790994e-01 -1.05766332e+00 -9.65930521e-01
-4.51805085e-01 3.35087299e-01 -2.08030120e-01 6.48847222e-02
2.77045280e-01 3.79589111e-01 -1.04682732e+00 -4.36066240e-01
1.17780387e+00 1.45197666e+00 1.76210571e-02 5.99731326e-01
2.24856243e-01 3.18222940e-01 -1.24771595e-01 -4.44664747e-01
6.91307068e-01 -2.05282658e-01 -7.26943851e-01 3.46751601e-01
3.48787636e-01 -1.09224759e-01 -8.81251022e-02 1.57789290e+00
3.84271950e-01 6.18460894e-01 4.90999043e-01 -1.05779088e+00
-3.29803020e-01 1.97349027e-01 -1.06025720e+00 8.27878863e-02
-8.26660767e-02 -6.61788434e-02 9.06078756e-01 -1.35178733e+00
2.56068587e-01 8.79792094e-01 5.92558563e-01 3.54972154e-01
-1.10707080e+00 -5.20101309e-01 4.41455513e-01 -3.32980275e-01
-1.44000947e+00 -1.68344200e-01 6.31873250e-01 -4.48780745e-01
3.31320673e-01 5.65276325e-01 2.80353308e-01 2.77905494e-01
-1.16039746e-01 8.42739642e-01 9.03828502e-01 -3.08524191e-01
2.80831248e-01 3.01679522e-01 2.27302119e-01 9.51899707e-01
5.51935017e-01 -8.48573446e-02 -1.72811568e-01 -4.28852439e-01
5.72038531e-01 1.63731217e-01 -4.57345575e-01 -1.73022345e-01
-9.92302895e-01 1.00359297e+00 2.72395968e-01 -2.89767366e-02
-3.93846840e-01 2.61113167e-01 4.50921416e-01 1.38355166e-01
8.81865680e-01 1.31847739e-01 -5.14933109e-01 -2.26419881e-01
-1.23977268e+00 1.83756113e-01 6.84404969e-01 1.20899355e+00
6.43233359e-01 2.28201598e-01 -4.61078659e-02 9.34849083e-01
3.44070345e-01 9.15568471e-01 5.64156286e-02 -1.19745684e+00
1.01779568e+00 2.86367834e-01 5.41635811e-01 -9.99443769e-01
-2.58346468e-01 -7.49579430e-01 -1.11051917e+00 9.43132490e-02
5.98052621e-01 -4.96741712e-01 -3.69212747e-01 1.83039212e+00
3.81396443e-01 -1.54648885e-01 -3.63457143e-01 1.05496943e+00
2.62324363e-02 4.85737771e-01 -4.94921058e-01 -8.30604672e-01
1.05493927e+00 -8.35335732e-01 -4.18620676e-01 2.48731468e-02
6.31788611e-01 -8.44094574e-01 1.28345382e+00 4.88349140e-01
-1.57851624e+00 -1.88842729e-01 -7.65725195e-01 2.73948908e-01
2.31395423e-01 1.68978155e-01 3.59148592e-01 6.01529419e-01
-6.58140779e-01 6.68428302e-01 -8.91949296e-01 3.03025335e-01
4.59506720e-01 5.12199759e-01 -1.98006898e-01 -1.60382122e-01
-4.84281957e-01 1.61976561e-01 -1.91737428e-01 6.21961236e-01
-8.63907099e-01 -9.58643436e-01 -5.53530693e-01 -9.23150256e-02
5.06333411e-01 -4.26463157e-01 6.41431928e-01 -5.44088542e-01
-1.16165459e+00 6.30463243e-01 -3.47039431e-01 -2.54657030e-01
7.74596214e-01 -3.55493903e-01 1.47572950e-01 -8.57408624e-03
1.59748450e-01 -1.89150944e-01 8.23935568e-01 -1.07181621e+00
-5.86207628e-01 -6.64249003e-01 1.15524277e-01 2.38007352e-01
-4.61977094e-01 7.06500262e-02 -4.31762367e-01 -6.88637495e-01
1.83441699e-01 -1.17709959e+00 -6.15432143e-01 6.41162694e-02
-4.38395530e-01 2.04090253e-01 3.99296880e-01 -8.56057405e-01
1.26784527e+00 -2.32192016e+00 2.73027390e-01 4.53762293e-01
2.81888992e-01 1.09248824e-01 1.95876002e-01 1.15046985e-01
6.72793016e-02 2.78811008e-01 -6.07185066e-01 -6.73595846e-01
-2.04705261e-02 -6.25479445e-02 -1.14109829e-01 9.52989995e-01
-4.66388047e-01 3.28151107e-01 -8.80106270e-01 -1.98552087e-01
-7.61821941e-02 3.19651097e-01 -6.84962451e-01 -1.85009941e-01
2.36039162e-01 3.20631027e-01 -5.38520396e-01 6.95602655e-01
1.11129868e+00 -4.49172169e-01 -4.73613702e-02 -2.14374632e-01
-5.31160645e-02 -3.25612366e-01 -1.67929602e+00 1.52886987e+00
-7.63042748e-01 3.74496758e-01 8.56483579e-01 -1.37873447e+00
6.01656258e-01 1.90202653e-01 6.99066758e-01 -2.27926914e-02
6.02233075e-02 5.82298279e-01 -4.71266687e-01 -3.13322574e-01
1.60149649e-01 -4.87294883e-01 6.45966306e-02 3.98278326e-01
-4.62677538e-01 2.54462183e-01 2.97555447e-01 -3.57404025e-03
9.65096712e-01 -2.55199552e-01 -2.68435404e-02 -5.83786428e-01
6.42032921e-01 -2.13582635e-01 8.53108644e-01 6.47816896e-01
-2.28516385e-01 5.48947990e-01 5.60057640e-01 -2.11327896e-01
-9.71528709e-01 -9.67305243e-01 -3.26740652e-01 1.05851650e+00
1.46148041e-01 7.89795816e-03 -6.78935826e-01 -7.92587638e-01
2.72882613e-03 5.92044294e-01 -5.71002424e-01 6.53229877e-02
-5.95889747e-01 -1.38010836e+00 1.59734309e-01 4.16455507e-01
4.92377698e-01 -4.81505573e-01 -1.35337800e-01 2.61966288e-01
-1.67622268e-01 -9.26799536e-01 -1.14663839e+00 4.55288291e-02
-1.24544704e+00 -8.88681054e-01 -1.10513222e+00 -6.37284815e-01
1.38264871e+00 3.06138843e-01 7.39292920e-01 -8.21777806e-02
-1.90809727e-01 4.25979137e-01 -8.64644200e-02 -2.21042499e-01
2.42192045e-01 -2.42073655e-01 3.60951304e-01 4.00963008e-01
-2.34184816e-01 -5.88711023e-01 -9.90764380e-01 4.05935168e-01
-8.87718678e-01 -3.46504420e-01 3.61268759e-01 9.92772043e-01
1.06181371e+00 1.06108591e-01 1.34643450e-01 -8.39595795e-01
6.86643004e-01 -4.17888582e-01 -1.00149703e+00 2.20090747e-01
-7.77448475e-01 1.06448449e-01 8.09797347e-01 -5.05770981e-01
-8.60392988e-01 -1.06475532e-01 2.71663189e-01 -8.23435724e-01
5.84080219e-01 7.41457462e-01 9.55788232e-03 -3.03210109e-01
4.93843436e-01 2.95711845e-01 -6.82855323e-02 -7.29650855e-01
2.75685936e-01 3.55525762e-01 2.32298940e-01 -8.84203672e-01
8.67900312e-01 9.07864630e-01 3.80075067e-01 -6.94190145e-01
-1.01935124e+00 -4.59287673e-01 -2.49860451e-01 7.26156086e-02
3.89807612e-01 -6.05230510e-01 -1.10232806e+00 -1.32495770e-02
-6.79900050e-01 -8.12716708e-02 -5.28579593e-01 8.10904860e-01
-5.96954763e-01 6.04167044e-01 -3.66774261e-01 -1.25021255e+00
-6.80793583e-01 -1.16679788e+00 6.22277021e-01 -2.34822091e-02
1.92606077e-01 -1.04763806e+00 -2.33538136e-01 5.14246464e-01
3.59987378e-01 3.25240314e-01 6.42132819e-01 -2.85140902e-01
-3.74576926e-01 -5.45067728e-01 -4.01161015e-01 9.41925704e-01
-4.47644368e-02 -2.97720313e-01 -1.71775296e-01 -6.82703257e-01
5.17832339e-01 2.01265067e-02 5.95557034e-01 6.69844985e-01
1.39110577e+00 -8.97141397e-01 -1.37414753e-01 9.12481904e-01
1.63118613e+00 -3.59690152e-02 2.10522577e-01 -1.49870396e-01
5.00230432e-01 3.30773026e-01 6.88755214e-01 8.24397564e-01
7.79508129e-02 2.59615272e-01 5.77555418e-01 -5.94040491e-02
2.64930904e-01 2.14208260e-01 3.72057766e-01 9.92158294e-01
-3.27322364e-01 8.47796872e-02 -5.97768962e-01 8.15866113e-01
-1.85346162e+00 -6.40464485e-01 -2.61093140e-01 2.74045396e+00
7.19534695e-01 -3.34234014e-02 9.33440998e-02 -1.34758037e-02
8.05717230e-01 -1.14394471e-01 -7.48000920e-01 -4.68534797e-01
-8.93433169e-02 4.00019765e-01 9.06624913e-01 6.49999321e-01
-9.09354568e-01 3.30519259e-01 4.70508957e+00 1.15180254e+00
-8.54530632e-01 2.41348177e-01 6.85544848e-01 -6.94250822e-01
-2.12619096e-01 2.01106533e-01 -5.98173022e-01 7.26641119e-01
6.69843018e-01 -3.49212795e-01 8.26483667e-01 1.00329685e+00
4.32587624e-01 -2.12196901e-01 -9.30223048e-01 1.17659175e+00
-2.15662703e-01 -1.14303172e+00 -3.71696025e-01 4.89254236e-01
1.05481946e+00 4.67316099e-02 2.49626085e-01 -8.82640332e-02
1.24862082e-01 -7.89169073e-01 4.17557955e-01 4.52702284e-01
8.41437340e-01 -8.76322687e-01 7.18715012e-01 4.13120747e-01
-1.19239926e+00 -1.86553672e-01 -3.61813426e-01 1.63022369e-01
3.92074317e-01 1.16532671e+00 -3.03124279e-01 5.90462923e-01
8.98869932e-01 2.69174367e-01 2.11240813e-01 1.02067280e+00
1.66525796e-01 5.44475317e-01 -8.49761307e-01 -2.93844554e-04
2.89069653e-01 -7.52428234e-01 8.73760283e-01 1.01949954e+00
4.58475947e-01 2.59946346e-01 2.81332940e-01 5.44350863e-01
-3.12332958e-01 4.37437713e-01 2.46110298e-02 2.57987142e-01
2.96234727e-01 1.10162354e+00 -4.71756995e-01 -1.56025603e-01
-4.34742987e-01 9.41914618e-01 3.21969450e-01 6.48897767e-01
-8.89816225e-01 -6.98978245e-01 7.10449994e-01 2.58830041e-01
4.28479344e-01 -4.17646170e-01 -5.65922558e-01 -1.23721921e+00
7.35280871e-01 -4.17662412e-01 4.02107656e-01 -1.21337436e-01
-1.38872731e+00 3.47520202e-01 3.01791145e-03 -1.18637848e+00
1.08689599e-01 -4.45678800e-01 -4.21982706e-01 9.49340522e-01
-1.03985059e+00 -5.32147884e-01 5.06841838e-02 6.54600918e-01
1.57910645e-01 -8.27094838e-02 4.70628589e-01 5.08538425e-01
-8.02725971e-01 8.71392131e-01 7.42519557e-01 3.98042127e-02
1.74558342e-01 -9.77454126e-01 -4.00283664e-01 9.24812734e-01
-3.03179592e-01 7.05464602e-01 7.78790891e-01 -3.02233726e-01
-1.33414960e+00 -1.10007358e+00 6.84250712e-01 1.42679915e-01
6.00431144e-01 -1.96292669e-01 -7.21868396e-01 5.90713382e-01
-3.15146983e-01 5.74495971e-01 7.20771194e-01 5.36582917e-02
-1.33897774e-02 -4.31097478e-01 -1.49717808e+00 6.02332115e-01
1.00714254e+00 -3.17262232e-01 1.58898324e-01 7.44738579e-01
3.40952396e-01 -3.72633100e-01 -1.22134233e+00 5.91668367e-01
3.34902853e-01 -7.02579677e-01 9.02126968e-01 -6.48984253e-01
9.75025222e-02 -2.67746687e-01 -6.81974649e-01 -8.76543403e-01
-7.98800774e-03 -1.13474798e+00 -3.73554170e-01 7.64118254e-01
7.00369358e-01 -7.15564728e-01 1.00021875e+00 8.15013528e-01
-1.09215789e-01 -1.27580857e+00 -1.42354977e+00 -1.00099766e+00
2.85488635e-01 -5.14952958e-01 -4.63143215e-02 6.51812613e-01
-8.77369493e-02 -1.34515464e-01 -6.28211558e-01 1.87280566e-01
1.04975379e+00 4.41050716e-02 4.79859054e-01 -6.85641468e-01
-5.80284655e-01 -3.42578381e-01 2.21077334e-02 -1.24933243e+00
-2.09706411e-01 -8.96287680e-01 -1.14868283e-01 -1.21079743e+00
3.76005411e-01 -7.25332141e-01 -4.17437047e-01 1.11350089e-01
-1.76968157e-01 -3.36671919e-02 1.20953254e-01 2.18459532e-01
-4.23107266e-01 7.65149117e-01 1.26079750e+00 -1.77575238e-02
-2.58817017e-01 4.49797571e-01 -6.75794244e-01 8.51042092e-01
5.00994623e-01 -5.17762721e-01 -4.86394525e-01 -6.29722118e-01
4.42237675e-01 4.58738595e-01 1.03337899e-01 -6.40450120e-01
6.24485537e-02 -3.08507413e-01 -1.23974651e-01 -5.62105358e-01
3.63601625e-01 -7.15697408e-01 -8.59091952e-02 3.46265614e-01
-1.78791389e-01 -1.32772177e-01 -8.11446980e-02 8.10485661e-01
-4.96933647e-02 -4.67540383e-01 9.58405375e-01 -1.26324788e-01
2.37561986e-01 7.25832045e-01 8.55383426e-02 4.40285325e-01
1.04282808e+00 -9.97441486e-02 1.02382123e-01 -5.36602795e-01
-9.64040697e-01 2.66033649e-01 -3.44394502e-04 -3.07187378e-01
5.42632461e-01 -1.38965416e+00 -8.36785495e-01 -1.16047375e-01
-4.53810066e-01 3.98341030e-01 4.46080238e-01 1.44382322e+00
-4.67684090e-01 2.79608369e-01 6.14281654e-01 -5.56907356e-01
-9.05825555e-01 6.00868285e-01 2.74597704e-01 -4.82248336e-01
-2.48331130e-01 1.28276885e+00 2.20101193e-01 -2.36704573e-01
2.87634254e-01 -2.43065953e-01 6.88415170e-01 -2.16037169e-01
2.81754315e-01 1.09865153e+00 3.62861939e-02 -3.20576817e-01
-3.37206334e-01 5.32098293e-01 -1.20065555e-01 -3.41417283e-01
1.35546255e+00 -3.35503906e-01 -3.79743904e-01 1.97687492e-01
1.84911990e+00 2.17753112e-01 -1.37795126e+00 -3.11645538e-01
-2.30829701e-01 -4.89083707e-01 6.04166128e-02 -3.23253095e-01
-1.33793008e+00 7.16370940e-01 5.40579200e-01 -6.54610991e-02
1.27056062e+00 -1.95705384e-01 9.69419479e-01 4.79538552e-02
4.19745862e-01 -1.47144938e+00 -4.15998287e-02 3.80049437e-01
1.07239401e+00 -1.22823143e+00 3.31367314e-01 -4.16656166e-01
-3.73038143e-01 8.42509806e-01 2.58129627e-01 -4.70794678e-01
1.19406688e+00 -6.38884455e-02 -4.92393196e-01 -4.69631553e-02
-2.63029248e-01 2.54600734e-01 4.63083535e-01 2.34436244e-02
3.94527048e-01 1.96103588e-01 -8.21105123e-01 6.61590576e-01
-2.44181696e-02 -1.11729167e-01 3.02808195e-01 8.55195165e-01
-2.58261710e-01 -8.23896229e-01 -3.97525936e-01 6.70921385e-01
-7.45375395e-01 -1.45341277e-01 4.12956417e-01 5.21751940e-01
-1.49605349e-01 9.28838372e-01 -2.25278899e-01 -4.04021377e-03
3.20157260e-01 -1.84033602e-01 4.93565768e-01 -3.14459890e-01
-2.52528101e-01 2.49877289e-01 -1.90368474e-01 -5.22103131e-01
-1.53584227e-01 -8.31342638e-01 -1.27509809e+00 -4.41522717e-01
-4.30062979e-01 3.65486443e-01 7.44674087e-01 8.80920053e-01
3.32391709e-01 9.32014361e-02 1.00645828e+00 -5.71497738e-01
-9.87987578e-01 -8.31656992e-01 -8.45526218e-01 3.27620894e-01
2.51897156e-01 -4.20355827e-01 -7.71317840e-01 -5.89190982e-02] | [6.533515453338623, 4.534720420837402] |
b6a3b307-3887-433b-9c7c-7852f53e97cb | integrating-markov-processes-with-structural | 1911.02175 | null | https://arxiv.org/abs/1911.02175v1 | https://arxiv.org/pdf/1911.02175v1.pdf | Integrating Markov processes with structural causal modeling enables counterfactual inference in complex systems | This manuscript contributes a general and practical framework for casting a Markov process model of a system at equilibrium as a structural causal model, and carrying out counterfactual inference. Markov processes mathematically describe the mechanisms in the system, and predict the system's equilibrium behavior upon intervention, but do not support counterfactual inference. In contrast, structural causal models support counterfactual inference, but do not identify the mechanisms. This manuscript leverages the benefits of both approaches. We define the structural causal models in terms of the parameters and the equilibrium dynamics of the Markov process models, and counterfactual inference flows from these settings. The proposed approach alleviates the identifiability drawback of the structural causal models, in that the counterfactual inference is consistent with the counterfactual trajectories simulated from the Markov process model. We showcase the benefits of this framework in case studies of complex biomolecular systems with nonlinear dynamics. We illustrate that, in presence of Markov process model misspecification, counterfactual inference leverages prior data, and therefore estimates the outcome of an intervention more accurately than a direct simulation. | ['Olga Vitek', 'Kaushal Paneri', 'Robert Osazuwa Ness'] | 2019-11-06 | integrating-markov-processes-with-structural-1 | http://papers.nips.cc/paper/9569-integrating-markov-processes-with-structural-causal-modeling-enables-counterfactual-inference-in-complex-systems | http://papers.nips.cc/paper/9569-integrating-markov-processes-with-structural-causal-modeling-enables-counterfactual-inference-in-complex-systems.pdf | neurips-2019-12 | ['counterfactual-inference'] | ['miscellaneous'] | [ 4.66458261e-01 4.17249292e-01 -3.57400715e-01 3.25751156e-01
-1.51967660e-01 -6.23979211e-01 9.73530650e-01 1.66436240e-01
-1.40923962e-01 1.29542994e+00 2.72329777e-01 -9.14534390e-01
-4.43585962e-01 -7.68221915e-01 -1.04468465e+00 -7.63886750e-01
-4.09883112e-01 4.21690702e-01 -2.86348701e-01 2.51376629e-01
3.42454880e-01 3.91748399e-01 -1.19722009e+00 -1.89981878e-01
7.62068152e-01 1.96072146e-01 1.50753215e-01 8.34600210e-01
-7.77604133e-02 5.37731647e-01 -2.97140062e-01 -2.45729998e-01
-1.42645454e-02 -3.89335126e-01 -5.78222394e-01 -4.10151839e-01
-2.78887212e-01 -3.29511553e-01 -3.05503070e-01 8.05304527e-01
2.62214571e-01 -1.61878943e-01 9.32813883e-01 -1.48576927e+00
-2.94228047e-01 5.74539602e-01 -5.53278506e-01 8.44192952e-02
4.74118710e-01 6.06228292e-01 7.47083187e-01 -2.63222426e-01
4.22128022e-01 1.67697799e+00 8.31614852e-01 3.41459990e-01
-1.89684868e+00 -6.55530453e-01 4.54702407e-01 -4.34399009e-01
-1.05061364e+00 -3.39424819e-01 3.57573181e-01 -6.34247780e-01
6.66435778e-01 2.24303916e-01 9.52185035e-01 1.28758967e+00
9.51316774e-01 5.58909953e-01 1.28660274e+00 -1.13358878e-01
6.37020290e-01 -4.39738601e-01 1.62782043e-01 2.76124477e-01
7.41294563e-01 9.12239373e-01 -1.62882701e-01 -8.72593522e-01
9.07342136e-01 4.04457122e-01 -4.13359791e-01 -5.27220905e-01
-1.11730719e+00 6.23500466e-01 -2.83343017e-01 -5.09470850e-02
-9.23991859e-01 5.85278988e-01 1.55682787e-01 1.63309231e-01
2.07556218e-01 4.25344080e-01 -5.05369365e-01 -1.13418199e-01
-8.82747173e-01 6.62906229e-01 1.14704859e+00 5.76508701e-01
4.58501637e-01 -1.89788923e-01 -4.00217563e-01 -1.83043823e-01
5.38007021e-01 9.74374831e-01 -5.82962893e-02 -9.20670271e-01
-1.38683319e-02 1.30240738e-01 8.57081473e-01 -3.99775982e-01
-9.08680484e-02 -5.15660495e-02 -6.98140919e-01 4.32379507e-02
7.24052966e-01 -4.35307026e-01 -9.48560417e-01 2.00131655e+00
3.78882885e-01 6.94790483e-01 1.22662872e-01 4.65866864e-01
5.58104180e-03 7.62887657e-01 5.35552680e-01 -1.05185854e+00
1.02783871e+00 3.67782228e-02 -1.06472576e+00 3.06209266e-01
3.26774895e-01 -6.75805390e-01 7.56707370e-01 7.44674401e-03
-1.28795993e+00 -1.74711365e-02 -6.82823718e-01 8.30715179e-01
4.95407246e-02 -5.67882895e-01 7.52475441e-01 7.58964658e-01
-9.69781518e-01 8.58081400e-01 -1.14766955e+00 -3.80705684e-01
2.27102458e-01 3.87271613e-01 2.29655594e-01 1.08332232e-01
-1.19833732e+00 8.06228340e-01 1.57843709e-01 -1.52320042e-01
-1.46189308e+00 -1.40052378e+00 -2.38098547e-01 2.13999033e-01
5.24552524e-01 -1.17469037e+00 1.24555862e+00 -7.25271583e-01
-1.53095257e+00 2.20911548e-01 -5.06308854e-01 -6.38935983e-01
8.67667317e-01 7.42964670e-02 -1.19723164e-01 -1.05561607e-01
1.91927284e-01 1.49982005e-01 6.83968008e-01 -1.28953385e+00
-2.62159824e-01 -2.30102643e-01 2.64156580e-01 9.42248665e-03
6.43612862e-01 -4.41956550e-01 2.02224955e-01 -2.75045305e-01
-1.34498268e-01 -1.00125778e+00 -5.67633629e-01 -2.10033655e-01
-6.81586087e-01 1.33790806e-01 7.21336663e-01 -1.87153593e-01
1.11972845e+00 -1.64293659e+00 -2.39150658e-01 1.71765581e-01
1.58342019e-01 -9.80729163e-02 9.58054811e-02 8.64962220e-01
-4.41258937e-01 4.55611974e-01 -4.23608840e-01 3.50734919e-01
-5.24628982e-02 1.38982177e-01 -7.95725405e-01 5.75708568e-01
3.87394801e-02 9.87360477e-01 -9.67740357e-01 -1.67554677e-01
1.47563696e-01 1.58965349e-01 -4.38763350e-01 1.90980732e-01
-3.45591903e-01 5.83125710e-01 -5.64700246e-01 -4.19186167e-02
8.40592623e-01 -3.19579065e-01 8.97285759e-01 1.10319190e-01
-4.49825317e-01 1.95050806e-01 -1.32797658e+00 9.50525999e-01
-3.65502805e-01 5.25342785e-02 -1.84362411e-01 -9.92347538e-01
3.42367411e-01 7.07966745e-01 7.34308064e-01 -1.74603164e-01
-9.33248848e-02 -1.80933908e-01 3.13071847e-01 -4.72104222e-01
8.53525195e-03 -7.94387102e-01 -6.39597839e-03 1.09434855e+00
-1.92545339e-01 -2.40982212e-02 -8.32185000e-02 2.98383266e-01
1.09856832e+00 3.08372706e-01 8.33736062e-01 -5.87275982e-01
1.18179180e-01 -1.17633134e-01 5.48240542e-01 1.30084646e+00
-2.10537344e-01 -8.49065110e-02 8.16869676e-01 -4.07378048e-01
-1.08782971e+00 -1.45753205e+00 -1.40079081e-01 1.95267171e-01
1.80757135e-01 -9.57512334e-02 -6.52442038e-01 -4.34526384e-01
3.38867247e-01 8.43461633e-01 -9.15260613e-01 -3.11997801e-01
-4.00225818e-01 -1.32073224e+00 3.64468426e-01 7.62087181e-02
7.39334598e-02 -5.05524278e-01 -7.01449335e-01 4.19662952e-01
-1.74065560e-01 -4.28763032e-01 -1.73697770e-01 -2.91453898e-01
-9.91356611e-01 -1.50610089e+00 -5.09286821e-01 3.15831840e-01
3.83405387e-01 1.88797116e-01 8.69158149e-01 -1.17877558e-01
-6.13920502e-02 4.86633033e-01 3.77246708e-01 -5.41986823e-01
-8.32258344e-01 -7.28400469e-01 3.01159471e-01 -1.45961478e-01
1.82958655e-02 -7.36382484e-01 -8.63316536e-01 6.22476749e-02
-7.68119693e-01 1.68158934e-01 4.19597030e-01 8.95240784e-01
5.57617664e-01 5.99109419e-02 5.99905550e-01 -9.14186656e-01
7.52123535e-01 -7.26060331e-01 -7.35202610e-01 2.98339427e-01
-8.83827090e-01 3.70602250e-01 5.09476960e-01 -6.55904412e-01
-1.61783409e+00 -1.90995857e-01 3.21077168e-01 -1.52975380e-01
-1.89040095e-01 5.66706717e-01 -3.58051777e-01 6.89385712e-01
2.09079221e-01 9.63371620e-02 1.84045643e-01 -3.04827958e-01
4.10892665e-01 4.16194871e-02 8.13179240e-02 -1.05672204e+00
4.68761086e-01 1.00164199e+00 5.77004194e-01 -5.77555656e-01
-3.57420683e-01 3.06339301e-02 -4.27411795e-01 -1.74503922e-01
5.88521421e-01 -7.10795462e-01 -1.46689153e+00 2.03467906e-01
-1.25897861e+00 -3.93322825e-01 -4.64649856e-01 6.82724833e-01
-9.88120675e-01 1.43318608e-01 -4.05147940e-01 -1.55175090e+00
2.61513263e-01 -9.50929821e-01 8.05735767e-01 5.01863658e-02
-4.26100820e-01 -1.34481668e+00 4.12042439e-01 -4.85084474e-01
4.82089743e-02 6.24353707e-01 1.07565880e+00 -2.34648988e-01
-7.40273833e-01 -1.58807784e-01 5.66257443e-03 -6.09931946e-01
6.69159144e-02 3.32736611e-01 -8.12885225e-01 -2.38704070e-01
8.33031014e-02 5.15007377e-01 5.14417589e-01 1.27526581e+00
5.80724120e-01 -7.73824453e-01 -9.80094790e-01 -1.79290920e-02
1.47881973e+00 4.41102386e-01 6.25015259e-01 -1.79325864e-01
2.69961417e-01 6.41161621e-01 5.74994624e-01 6.17744386e-01
1.72704905e-01 4.82301205e-01 2.35022530e-01 4.74091098e-02
3.73807490e-01 -8.49693179e-01 4.66613472e-01 -2.84288675e-02
-1.80412322e-01 -1.29077911e-01 -9.02867973e-01 5.18980443e-01
-1.98792779e+00 -1.42417264e+00 -5.96983194e-01 2.52500391e+00
7.25614727e-01 6.13192804e-02 3.99785072e-01 -2.74223626e-01
9.59954977e-01 -3.40670764e-01 -8.53670835e-01 -3.79769146e-01
1.14729181e-01 1.23801818e-02 8.58016253e-01 7.63604045e-01
-5.88749528e-01 4.70353782e-01 7.80216837e+00 5.03400743e-01
-9.12169039e-01 2.40798190e-01 6.32178962e-01 -9.52756107e-02
-5.75540662e-01 7.77822852e-01 -6.30954206e-01 6.65746331e-01
1.51066256e+00 -8.78659964e-01 1.05775200e-01 4.72743250e-02
1.26781917e+00 -3.89766991e-01 -1.40910923e+00 2.48075381e-01
-9.37923253e-01 -1.62732470e+00 1.48709550e-01 6.00278735e-01
6.94923699e-01 -4.95650142e-01 -1.88803289e-03 4.21856456e-02
1.20116186e+00 -1.05756199e+00 7.68446743e-01 1.12559021e+00
6.16077900e-01 -4.73261327e-01 5.25889993e-01 5.64797699e-01
-7.58497596e-01 -5.39326072e-02 5.11318929e-02 -6.14748716e-01
5.65598309e-01 9.16357636e-01 -8.98702979e-01 5.18854141e-01
1.72514245e-01 4.55319494e-01 4.83080149e-01 9.40587401e-01
-1.69285133e-01 9.52974081e-01 -3.32762063e-01 4.80312258e-02
-9.97341871e-02 -4.01142478e-01 8.68217349e-01 9.71641839e-01
2.32884258e-01 1.10769749e-01 -2.24133208e-02 1.32391357e+00
4.77341324e-01 -3.52212727e-01 -8.90039504e-01 -1.66759178e-01
6.57344103e-01 3.90704274e-01 -8.10256004e-01 -4.08182144e-01
-1.44054815e-01 3.62713337e-01 -2.41320387e-01 8.21399152e-01
-9.73687351e-01 3.49483848e-01 8.36589217e-01 2.23723620e-01
1.96833387e-02 7.30535686e-02 -4.55171078e-01 -1.10903752e+00
-3.87478471e-01 -6.31331682e-01 3.54482949e-01 -4.90747362e-01
-1.09649992e+00 -5.48771739e-01 4.22078490e-01 -8.37332904e-01
-3.09825748e-01 -3.42945129e-01 -9.02553856e-01 1.12779880e+00
-1.19876444e+00 -8.28218341e-01 6.57009184e-01 1.85600787e-01
2.30913870e-02 6.61990225e-01 6.31418288e-01 -4.58630174e-01
-7.45690405e-01 -5.84321208e-02 3.58758569e-01 -6.12847686e-01
1.99159324e-01 -1.23242795e+00 3.41636509e-01 8.26743245e-01
-6.33348763e-01 1.08959877e+00 1.42558110e+00 -1.10515141e+00
-1.41765082e+00 -9.53144193e-01 5.22607028e-01 -5.08903801e-01
1.00854862e+00 -1.58898905e-01 -7.44242013e-01 5.65288842e-01
-9.71206203e-02 -5.72031617e-01 5.46271324e-01 2.80761659e-01
1.97001323e-02 4.01222259e-01 -1.15345013e+00 8.99400175e-01
9.28889453e-01 -2.44062215e-01 -5.86501598e-01 2.39155650e-01
6.43591285e-01 1.57471269e-01 -8.51371229e-01 1.28655151e-01
8.97562146e-01 -6.20944560e-01 1.04942560e+00 -9.31279659e-01
4.57056344e-01 -3.95860851e-01 1.24288999e-01 -1.24018347e+00
-9.11573991e-02 -1.02495885e+00 -3.21566433e-01 1.02478194e+00
3.66629541e-01 -8.24158311e-01 4.31888431e-01 8.49679887e-01
4.22228575e-01 -3.48060340e-01 -1.02678657e+00 -9.81338441e-01
4.46642071e-01 -4.05288219e-01 1.00591195e+00 9.44876254e-01
1.19157672e-01 1.62427172e-01 -2.44677082e-01 3.45882058e-01
8.65682960e-01 4.15265352e-01 7.73209572e-01 -1.02970946e+00
-5.38876951e-01 -2.78487742e-01 9.48927775e-02 -7.97460675e-01
2.08852783e-01 -3.38674098e-01 -7.15113431e-02 -1.18758166e+00
7.04951823e-01 -4.16157544e-01 5.66566512e-02 -1.67991072e-02
-4.80248749e-01 -5.97993672e-01 1.11670561e-01 3.93447876e-01
-7.27109760e-02 5.64525127e-01 1.26804233e+00 1.21108033e-01
-2.37099618e-01 2.14261711e-01 -6.92733288e-01 6.75043583e-01
4.87325847e-01 -5.90097070e-01 -5.45895934e-01 5.17608821e-01
1.85632080e-01 8.81860435e-01 8.35805535e-01 -1.66848272e-01
-9.26442370e-02 -8.31078291e-01 1.67471901e-01 -3.34747612e-01
-4.30433452e-02 -6.03263855e-01 9.34531868e-01 1.14906549e+00
-3.68708283e-01 -1.85701758e-01 3.04079443e-01 1.29656374e+00
3.42183024e-01 -1.85941570e-02 5.78232408e-01 -1.90641373e-01
-6.63436875e-02 3.48206490e-01 -9.20937300e-01 -1.33760959e-01
1.03162992e+00 -3.36749762e-01 -2.66130865e-01 -6.39667630e-01
-1.02254677e+00 1.34081349e-01 5.78059316e-01 -2.32613459e-01
2.97678083e-01 -9.19151604e-01 -4.18318868e-01 -2.42692605e-01
-3.50660026e-01 -5.43965757e-01 4.15198296e-01 1.12748075e+00
1.28892079e-01 6.31436169e-01 -2.23186631e-02 -5.06071091e-01
-8.32404554e-01 8.51447344e-01 5.79884589e-01 -5.72346926e-01
-2.64377356e-01 -6.33748397e-02 8.27392042e-01 -2.16824308e-01
-6.94094658e-01 -1.44819498e-01 1.08701877e-01 -2.46874064e-01
3.68743122e-01 5.25483549e-01 -5.54057240e-01 -8.68712142e-02
-1.72914058e-01 -1.52761983e-02 9.36455280e-02 -4.87925589e-01
1.09404862e+00 -5.18149436e-01 5.15905693e-02 7.11808622e-01
5.07064581e-01 -2.10974693e-01 -1.78170776e+00 2.55490959e-01
6.34494349e-02 -2.19937950e-01 -1.15246559e-03 -7.99452126e-01
-2.31565133e-01 8.24311197e-01 4.67282355e-01 4.19779509e-01
6.70815825e-01 -2.14847982e-01 7.46054202e-02 -8.86703655e-02
3.55019927e-01 -6.29985034e-01 -4.70280379e-01 1.19847648e-01
6.48011863e-01 -6.41493201e-01 5.74254990e-02 -5.60636818e-01
-1.36354357e-01 7.11513460e-01 9.81789306e-02 -1.70475110e-01
8.89748454e-01 4.39591467e-01 -5.65942705e-01 -2.05459729e-01
-1.23809719e+00 -1.06310323e-02 -2.95554429e-01 7.31999874e-01
5.49224079e-01 5.13111889e-01 -6.61036611e-01 5.74498713e-01
1.38899922e-01 3.66471529e-01 9.13721144e-01 9.33741152e-01
-4.56306571e-03 -8.61531913e-01 -6.11467421e-01 3.55385810e-01
-4.74102676e-01 -2.25570545e-01 -1.02566630e-01 1.08100605e+00
-3.45815033e-01 1.01438534e+00 1.77450806e-01 3.70990366e-01
2.37450406e-01 1.78609312e-01 5.08889019e-01 -3.27693760e-01
5.11426665e-02 1.75516024e-01 2.87933443e-02 -5.21823585e-01
-4.88330662e-01 -1.19600093e+00 -1.06257176e+00 -8.23376596e-01
-3.68938178e-01 2.77359098e-01 3.20984751e-01 1.29050183e+00
5.07994473e-01 5.32297075e-01 4.73656178e-01 -5.70205748e-01
-6.60870612e-01 -7.23841310e-01 -4.36780304e-01 2.07917407e-01
3.48899126e-01 -8.08120012e-01 -2.99479008e-01 3.04706186e-01] | [8.014569282531738, 5.371291637420654] |
e18f3cab-d061-4d16-8c93-57d719f7ed35 | multimodal-prompt-learning-for-product-title | 2307.01969 | null | https://arxiv.org/abs/2307.01969v1 | https://arxiv.org/pdf/2307.01969v1.pdf | Multimodal Prompt Learning for Product Title Generation with Extremely Limited Labels | Generating an informative and attractive title for the product is a crucial task for e-commerce. Most existing works follow the standard multimodal natural language generation approaches, e.g., image captioning, and employ the large scale of human-labelled datasets to train desirable models. However, for novel products, especially in a different domain, there are few existing labelled data. In this paper, we propose a prompt-based approach, i.e., the Multimodal Prompt Learning framework, to accurately and efficiently generate titles for novel products with limited labels. We observe that the core challenges of novel product title generation are the understanding of novel product characteristics and the generation of titles in a novel writing style. To this end, we build a set of multimodal prompts from different modalities to preserve the corresponding characteristics and writing styles of novel products. As a result, with extremely limited labels for training, the proposed method can retrieve the multimodal prompts to generate desirable titles for novel products. The experiments and analyses are conducted on five novel product categories under both the in-domain and out-of-domain experimental settings. The results show that, with only 1% of downstream labelled data for training, our proposed approach achieves the best few-shot results and even achieves competitive results with fully-supervised methods trained on 100% of training data; With the full labelled data for training, our method achieves state-of-the-art results. | ['Yuexian Zou', 'Bing Yin', 'Chenyu You', 'Qingyu Yin', 'Zheng Li', 'Fenglin Liu', 'Bang Yang'] | 2023-07-05 | null | null | null | null | ['image-captioning', 'text-generation'] | ['computer-vision', 'natural-language-processing'] | [ 7.31258810e-01 2.17237458e-01 -4.11144644e-01 -5.51636994e-01
-9.95458722e-01 -6.98263228e-01 6.82277679e-01 -5.62375747e-02
-5.61028719e-02 5.06168783e-01 1.91485241e-01 7.62844533e-02
1.90626234e-01 -4.56768960e-01 -8.12096417e-01 -6.66963458e-01
5.81457496e-01 5.13171136e-01 -2.46578187e-01 -4.08751726e-01
2.64107823e-01 -1.03849217e-01 -1.72896361e+00 4.93038058e-01
9.72759783e-01 1.21140242e+00 5.40661871e-01 6.55480444e-01
-3.93326700e-01 5.10103345e-01 -4.67374086e-01 -8.98502827e-01
1.18600354e-01 -6.30886734e-01 -7.11035907e-01 7.42951691e-01
3.00118506e-01 -3.26548159e-01 -2.28450559e-02 9.89582837e-01
6.19987309e-01 1.49986297e-01 7.21492112e-01 -1.24973285e+00
-1.37048161e+00 7.38569140e-01 -5.01975358e-01 -4.57842529e-01
6.67695701e-01 2.84913123e-01 9.39519942e-01 -1.13963759e+00
9.07460272e-01 1.14214110e+00 7.79830739e-02 8.58363867e-01
-1.23059213e+00 -4.03318554e-01 2.73826659e-01 5.02983574e-04
-1.25574636e+00 -5.19572377e-01 1.05141294e+00 -2.71663040e-01
3.36486876e-01 5.01745846e-03 1.22763537e-01 1.47775221e+00
-1.94568127e-01 1.12481368e+00 9.32512164e-01 -4.69101578e-01
1.69696361e-01 6.49351239e-01 -1.83429778e-01 3.13102484e-01
-7.63647184e-02 -2.46888846e-02 -7.18732893e-01 8.71595666e-02
4.57258493e-01 -1.38569041e-03 -2.14425728e-01 -3.86326671e-01
-1.51527619e+00 9.21635747e-01 5.98803684e-02 4.96884249e-02
-4.50040042e-01 -3.78404021e-01 4.53913003e-01 1.86374843e-01
5.03327310e-01 5.87349951e-01 -4.39096451e-01 -1.26768604e-01
-7.04376638e-01 4.13865626e-01 7.69689739e-01 1.73498809e+00
6.06569052e-01 -1.62832454e-01 -5.41419685e-01 9.55007672e-01
3.29607517e-01 8.31482053e-01 5.56154788e-01 -6.82912230e-01
7.19609380e-01 5.43049395e-01 4.98635978e-01 -1.02058661e+00
-5.49991652e-02 -2.75193840e-01 -7.72145212e-01 -3.54362071e-01
1.23602778e-01 -2.90246606e-01 -9.93778050e-01 1.58343017e+00
2.01439247e-01 -2.26314232e-01 5.10665596e-01 8.47637951e-01
1.15719914e+00 1.09416211e+00 2.51901418e-01 -4.71949428e-01
1.37543023e+00 -1.40204930e+00 -9.93014276e-01 -2.54040390e-01
5.06321549e-01 -1.20218241e+00 1.46600676e+00 3.00226063e-01
-1.06202126e+00 -9.03351903e-01 -9.12750900e-01 -1.56990637e-03
-5.06383359e-01 3.23128074e-01 5.28780758e-01 3.73524100e-01
-5.48044682e-01 2.27375507e-01 -4.23119105e-02 -4.80022073e-01
2.56404132e-01 3.69940139e-02 -1.37047932e-01 -5.68565607e-01
-1.24691248e+00 5.61610162e-01 5.60271621e-01 2.08696932e-01
-9.61181939e-01 -6.23145819e-01 -1.11150920e+00 -1.46351203e-01
4.52267468e-01 -7.60799885e-01 1.71769130e+00 -1.29014063e+00
-1.43567944e+00 7.28932321e-01 -1.58328593e-01 1.69268362e-02
3.86638016e-01 -1.09928608e-01 -6.31009161e-01 2.24275947e-01
2.37358809e-01 1.28027511e+00 1.02732909e+00 -1.85720479e+00
-8.59574616e-01 -8.75503570e-02 1.21762469e-01 3.90106499e-01
-5.19747138e-01 8.87070969e-03 -4.37342674e-01 -6.08571708e-01
-2.25684106e-01 -8.85513961e-01 -1.77323163e-01 -2.33370081e-01
-3.06717694e-01 -4.90385741e-01 6.30379498e-01 -6.01136804e-01
8.80774319e-01 -2.19215488e+00 1.16479859e-01 -2.41474286e-01
-4.44136038e-02 1.54974654e-01 -5.51054001e-01 4.50845242e-01
2.64937848e-01 4.16293219e-02 -3.21904309e-02 -4.81210232e-01
3.88374656e-01 3.03042382e-02 -3.66674036e-01 -2.89591849e-01
4.35044199e-01 1.23324335e+00 -1.15062213e+00 -6.14939809e-01
8.30997378e-02 2.47532591e-01 -3.65448184e-02 4.98470217e-01
-5.08702397e-01 5.00102997e-01 -5.18479466e-01 8.67029905e-01
7.78416872e-01 -3.48172754e-01 1.51731903e-02 -3.70044053e-01
2.38836989e-01 -3.75387460e-01 -9.59919095e-01 2.13584828e+00
-5.87575495e-01 1.72462210e-01 -2.43315190e-01 -5.66765606e-01
1.27454340e+00 4.86935973e-01 1.59687355e-01 -1.08575201e+00
1.63004696e-01 4.10507351e-01 -4.47169155e-01 -8.71409833e-01
9.43732202e-01 -2.17569888e-01 -4.33680505e-01 4.08409715e-01
1.64354458e-01 -9.27642584e-02 4.84713405e-01 2.29405358e-01
5.06412864e-01 3.85328352e-01 -7.07423761e-02 2.66930252e-01
3.64966959e-01 1.75787032e-01 3.33133191e-01 7.70143747e-01
-6.84851110e-02 7.87631333e-01 2.02932015e-01 -1.43447250e-01
-1.23324156e+00 -9.71523821e-01 1.58033907e-01 1.33003759e+00
5.99359870e-01 -9.94897559e-02 -6.43551052e-01 -8.09415758e-01
-3.54514867e-01 8.33992362e-01 -3.79901648e-01 -2.33121216e-01
-1.68892354e-01 -4.23770130e-01 1.83164507e-01 3.71954650e-01
5.15934885e-01 -1.53695190e+00 -2.33691052e-01 2.62118429e-01
-4.53772068e-01 -1.30275762e+00 -7.50221789e-01 -2.45939448e-01
-7.05131173e-01 -6.83112204e-01 -1.40319467e+00 -1.44175851e+00
8.97061110e-01 4.20536429e-01 1.06533313e+00 -4.42927688e-01
-1.16170058e-02 4.38690424e-01 -8.06790411e-01 -3.78081709e-01
-5.76293528e-01 2.87705034e-01 -2.75787026e-01 2.78367519e-01
4.81843740e-01 -9.53847319e-02 -5.46975255e-01 3.24941546e-01
-1.07342362e+00 4.37442213e-01 1.01681006e+00 9.86683369e-01
5.31857729e-01 -1.63636860e-02 1.00085974e+00 -8.51186514e-01
9.05542731e-01 -4.86635387e-01 -2.62015134e-01 5.48693061e-01
-6.35303259e-01 -3.30287106e-02 7.00878203e-01 -8.62174630e-01
-1.67648482e+00 2.55410314e-01 1.29551202e-01 -2.87012488e-01
-5.17029583e-01 5.99709690e-01 -3.27839077e-01 3.16542059e-01
5.34550667e-01 4.37294066e-01 -1.54078022e-01 -6.33613527e-01
8.05162072e-01 9.85440254e-01 5.30284882e-01 -4.79031652e-01
7.15871334e-01 5.18544316e-02 -4.30663258e-01 -3.61510009e-01
-9.75085378e-01 -5.70053160e-01 -5.93811810e-01 -2.86495715e-01
6.94946945e-01 -9.40651655e-01 -4.61472034e-01 6.25642061e-01
-1.36718869e+00 -6.57353131e-03 -4.31213707e-01 2.83504039e-01
-6.41777277e-01 2.58739978e-01 -4.91832435e-01 -8.36419940e-01
-5.01805961e-01 -1.04836369e+00 1.42728484e+00 5.94030619e-01
-8.07583779e-02 -8.62242281e-01 -1.02842689e-01 6.25267804e-01
2.94150651e-01 9.43894684e-02 9.43631589e-01 -6.70715034e-01
-5.15550256e-01 -1.60028785e-01 -4.36591685e-01 3.17412466e-01
2.98157126e-01 -4.10321534e-01 -9.78495777e-01 4.36573401e-02
-1.42465234e-01 -7.09767997e-01 5.67126095e-01 -1.57024078e-02
6.67313039e-01 -4.48164552e-01 -1.78525105e-01 1.01081498e-01
1.23862910e+00 4.89154100e-01 5.90016723e-01 1.83529809e-01
6.97496772e-01 1.17044985e+00 1.38695419e+00 6.68002069e-01
4.89025950e-01 5.75877368e-01 1.78311303e-01 -2.25101575e-01
-1.85514111e-02 -5.99806905e-01 4.89448547e-01 7.29336083e-01
3.32257360e-01 -4.71290857e-01 -4.23801661e-01 6.82543993e-01
-2.08734632e+00 -7.80953884e-01 7.38073662e-02 1.78722584e+00
1.10877752e+00 -9.92327705e-02 9.16680135e-03 -3.26531947e-01
9.41832423e-01 9.90279838e-02 -7.32198358e-01 -5.27467847e-01
-2.49602690e-01 -1.34749800e-01 1.58694714e-01 1.47153083e-02
-1.08081436e+00 9.53585088e-01 5.74383926e+00 9.47372317e-01
-7.61172533e-01 6.36159256e-02 7.02423334e-01 1.50263731e-04
-5.79474211e-01 -2.95460671e-01 -9.73086059e-01 5.34468412e-01
8.13312829e-01 -2.75760770e-01 3.72521669e-01 9.56079245e-01
3.42531502e-01 3.96989137e-02 -1.25594676e+00 1.29502225e+00
6.06894732e-01 -1.07712436e+00 3.65076512e-01 -1.39915273e-01
1.07816529e+00 -7.19882190e-01 3.67972314e-01 4.62400854e-01
-2.13872176e-02 -7.75173306e-01 8.16541851e-01 4.40338582e-01
8.25729132e-01 -7.76193917e-01 9.90890205e-01 3.09628725e-01
-9.19897676e-01 -6.48674928e-03 -2.82861978e-01 2.00789094e-01
6.91356242e-01 4.02643323e-01 -7.39737689e-01 6.72374189e-01
3.95660281e-01 8.17857563e-01 -7.32831955e-01 8.29396307e-01
-1.61393389e-01 1.46900609e-01 2.45483771e-01 -2.32650951e-01
2.45802894e-01 -3.39180306e-02 1.18044607e-01 1.03007782e+00
5.32764435e-01 2.30940040e-02 2.42706165e-01 9.86744821e-01
-3.86976689e-01 3.30698073e-01 -6.33908868e-01 -4.78263378e-01
1.32066667e-01 1.54276276e+00 -4.39566672e-01 -3.95339906e-01
-4.99732554e-01 1.14749432e+00 -1.11768745e-01 5.80621958e-01
-8.34346652e-01 -5.82583368e-01 5.02961949e-02 -1.82314545e-01
4.71187294e-01 1.17433771e-01 -1.57517865e-01 -1.01218569e+00
2.74624705e-01 -1.05118799e+00 2.05789879e-01 -1.13973129e+00
-1.69838166e+00 6.11402988e-01 3.19257565e-02 -1.44434249e+00
-4.22473550e-01 -4.95393872e-01 -4.07919943e-01 6.64130449e-01
-1.67987704e+00 -1.61218429e+00 -3.56845915e-01 3.96821380e-01
1.08980811e+00 -2.69139171e-01 6.41141236e-01 2.82445967e-01
-3.08471203e-01 5.84182203e-01 1.18646115e-01 -1.64651990e-01
9.90779936e-01 -8.85689378e-01 1.84217170e-01 6.59312129e-01
-8.95293057e-03 4.35913414e-01 7.52501190e-01 -6.83221757e-01
-1.41216111e+00 -1.11659801e+00 1.12476552e+00 -1.41518384e-01
4.71921951e-01 -4.54425633e-01 -6.40790284e-01 4.25401300e-01
7.06847370e-01 -5.15865088e-01 8.03950906e-01 -1.04050517e-01
-1.25287458e-01 -1.27136543e-01 -1.04307270e+00 6.68560565e-01
1.04897511e+00 -1.88181609e-01 -5.58648765e-01 5.84330976e-01
1.18476021e+00 -2.15997353e-01 -9.23282743e-01 2.64373392e-01
5.31224549e-01 -3.90493542e-01 6.24107182e-01 -5.99457860e-01
8.02423000e-01 -2.49550447e-01 -8.33974853e-02 -1.31549323e+00
-1.48809254e-01 -7.32300222e-01 -8.73188451e-02 2.00162101e+00
7.76837111e-01 -1.97026610e-01 5.85165143e-01 8.71934652e-01
-1.79454252e-01 -5.87904215e-01 -2.99685210e-01 -6.63671196e-01
-4.70958352e-01 -1.21790133e-01 7.21414089e-01 6.18454814e-01
-4.59777638e-02 8.37218583e-01 -7.01313794e-01 -1.77275419e-01
5.27832150e-01 5.15945315e-01 7.98775136e-01 -7.65068948e-01
-5.56340329e-02 -1.54726133e-01 1.34292850e-02 -1.36037481e+00
1.83498383e-01 -6.34052098e-01 4.86432165e-01 -1.54644120e+00
5.11937261e-01 -3.30698520e-01 -1.56879261e-01 2.76977509e-01
-2.19473884e-01 2.03203857e-01 3.93181980e-01 2.11646572e-01
-9.98212278e-01 7.63448954e-01 1.73939872e+00 -5.33159196e-01
-4.52259183e-01 -5.01171872e-02 -1.09970200e+00 3.69249731e-01
6.61286652e-01 -2.91031361e-01 -6.39216065e-01 -4.15403187e-01
3.35525364e-01 -1.17060177e-01 1.74355090e-01 -4.52458471e-01
1.13690168e-01 -3.58380765e-01 3.81669044e-01 -7.01157093e-01
3.17018747e-01 -7.27251291e-01 -1.21472046e-01 6.51105121e-03
-7.72979498e-01 -1.28463414e-02 -5.15070148e-02 6.66528225e-01
-3.64028126e-01 -5.48489630e-01 3.10620993e-01 -1.08947299e-01
-1.20518124e+00 2.30322987e-01 -3.56224813e-02 4.07696664e-02
1.13949478e+00 -1.29383653e-01 -4.60866749e-01 -5.38424432e-01
-6.83705389e-01 3.81715745e-01 4.67889309e-01 9.75363255e-01
8.17717075e-01 -1.71135902e+00 -7.80478895e-01 -5.71902022e-02
8.27329040e-01 4.04629298e-02 5.38831174e-01 5.49209297e-01
-3.78949530e-02 5.06498694e-01 -1.11443236e-01 -6.43444300e-01
-1.01400387e+00 1.01666307e+00 -2.18511388e-01 -2.40890205e-01
-1.18359067e-01 5.89838386e-01 4.66423839e-01 -3.32639307e-01
1.20474555e-01 1.03281431e-01 -3.35803211e-01 3.55183542e-01
6.31929755e-01 4.58451593e-03 -1.70934290e-01 -8.12199950e-01
1.68042108e-01 5.37799299e-01 -3.83800834e-01 -1.90780789e-01
9.77594852e-01 -5.96618414e-01 2.46733546e-01 4.85746592e-01
1.20179486e+00 -4.80569482e-01 -1.32839632e+00 -3.67572188e-01
-4.80185263e-02 -5.04024029e-01 -2.98939496e-01 -1.01902270e+00
-8.97998750e-01 9.25104618e-01 7.09577441e-01 5.21605574e-02
1.21888554e+00 3.40883732e-01 1.23756969e+00 2.54088223e-01
3.49253744e-01 -1.32951057e+00 3.27211380e-01 1.81104720e-01
8.98256302e-01 -1.65093756e+00 -4.83035684e-01 -5.53281009e-01
-1.33029687e+00 8.89939189e-01 7.82153070e-01 5.06103396e-01
4.90169153e-02 -3.70063573e-01 3.27718079e-01 -6.45329282e-02
-6.05854452e-01 -2.34866396e-01 3.85810405e-01 8.97519290e-01
3.43899012e-01 -8.84042457e-02 -1.73895866e-01 9.20981228e-01
1.13535807e-01 1.08463377e-01 4.44219112e-01 8.99176240e-01
-2.86135197e-01 -1.11831784e+00 -2.04226673e-01 2.93928295e-01
-3.39935869e-02 -1.51194319e-01 -4.12594825e-01 5.11648774e-01
1.94114715e-01 1.33357894e+00 -2.00249210e-01 -2.66285270e-01
6.66855097e-01 9.73466039e-02 3.97773325e-01 -8.08276355e-01
-3.64758223e-01 2.06042692e-01 1.56948835e-01 -1.36126027e-01
-7.17648149e-01 -5.85679591e-01 -8.93216789e-01 -1.57258496e-01
-3.85010779e-01 3.47516723e-02 6.91938818e-01 7.79616237e-01
6.14374995e-01 3.92143071e-01 1.00211442e+00 -9.54957426e-01
-6.61405861e-01 -1.05189419e+00 -6.85202003e-01 9.02667940e-01
-3.70419282e-03 -4.97395068e-01 -1.47903204e-01 6.76143825e-01] | [11.0477933883667, 0.9930055737495422] |
ccd322a7-35e0-4d40-b8f4-12b12fbdee1e | implicit-identity-driven-deepfake-face | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Huang_Implicit_Identity_Driven_Deepfake_Face_Swapping_Detection_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Huang_Implicit_Identity_Driven_Deepfake_Face_Swapping_Detection_CVPR_2023_paper.pdf | Implicit Identity Driven Deepfake Face Swapping Detection | In this paper, we consider the face swapping detection from the perspective of face identity. Face swapping aims to replace the target face with the source face and generate the fake face that the human cannot distinguish between real and fake. We argue that the fake face contains the explicit identity and implicit identity, which respectively corresponds to the identity of the source face and target face during face swapping. Note that the explicit identities of faces can be extracted by regular face recognizers. Particularly, the implicit identity of real face is consistent with the its explicit identity. Thus the difference between explicit and implicit identity of face facilitates face swapping detection. Following this idea, we propose a novel implicit identity driven framework for face swapping detection. Specifically, we design an explicit identity contrast (EIC) loss and an implicit identity exploration (IIE) loss, which supervises a CNN backbone to embed face images into the implicit identity space. Under the guidance of EIC, real samples are pulled closer to their explicit identities, while fake samples are pushed away from their explicit identities. Moreover, IIE is derived from the margin-based classification loss function, which encourages the fake faces with known target identities to enjoy intra-class compactness and inter-class diversity. Extensive experiments and visualizations on several datasets demonstrate the generalization of our method against the state-of-the-art counterparts. | ['Dengpan Ye', 'Qian Wang', 'Qin Zou', 'Jiaxin Ai', 'Jifan Yang', 'Zhongyuan Wang', 'Baojin Huang'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['face-swapping'] | ['computer-vision'] | [ 2.83471406e-01 5.14324307e-01 -1.71758588e-02 -3.60608786e-01
-2.55432576e-01 -7.08162010e-01 4.23777729e-01 -6.70963228e-01
-2.76639108e-02 5.58837235e-01 -1.54027283e-01 1.48497030e-01
2.06011340e-01 -7.08179235e-01 -6.46825969e-01 -8.90916109e-01
1.91841424e-01 1.96288183e-01 -2.57594138e-01 -5.76691441e-02
9.81115624e-02 5.21935463e-01 -1.61301744e+00 3.38502616e-01
7.42083549e-01 1.26143837e+00 -3.20473462e-01 -8.35381076e-02
-1.14859650e-02 4.33997035e-01 -6.35460794e-01 -6.08258724e-01
6.62928224e-01 -7.13202775e-01 -4.94123608e-01 1.06282860e-01
7.77777135e-01 -4.13803428e-01 -3.90748501e-01 1.39502525e+00
3.24397296e-01 -3.02565813e-01 4.33063656e-01 -1.98729944e+00
-1.09906507e+00 -1.20004797e-02 -7.55759895e-01 -2.51877904e-01
3.30301166e-01 3.54245424e-01 5.46040058e-01 -1.20748127e+00
5.94367564e-01 1.44064689e+00 7.86532819e-01 9.01434779e-01
-1.06117094e+00 -1.31290102e+00 2.25668773e-01 -5.01051953e-04
-1.68212759e+00 -7.79895544e-01 9.36461508e-01 -4.83008623e-01
-9.65949744e-02 4.05428201e-01 4.85282034e-01 1.03816330e+00
-3.47981930e-01 6.10549748e-01 1.12698162e+00 -2.32466668e-01
-1.28119498e-01 7.27516711e-01 -1.79036528e-01 7.78778732e-01
2.73274571e-01 5.72190285e-01 -4.77756828e-01 -2.64803141e-01
8.48040581e-01 4.29668538e-02 -6.86971962e-01 -4.84440565e-01
-7.04180300e-01 7.26878941e-01 5.50896883e-01 -2.55034510e-02
-1.22668497e-01 -9.12543312e-02 1.66904643e-01 3.30550998e-01
3.42783272e-01 1.93949148e-01 -1.02645587e-02 4.49553639e-01
-8.23727906e-01 1.42799154e-01 4.95409817e-01 1.08059037e+00
8.73191833e-01 -2.13067710e-01 -1.82934627e-01 6.72061920e-01
3.89509946e-01 3.74803066e-01 4.09613878e-01 -8.02384734e-01
1.78031847e-01 8.69182050e-01 2.81494558e-01 -1.41039860e+00
4.46840018e-01 -2.29810447e-01 -8.40930045e-01 4.40663904e-01
4.19862479e-01 1.63707405e-01 -6.30305469e-01 1.88652146e+00
6.14630699e-01 6.26093924e-01 -1.11474328e-01 1.02742255e+00
7.24523723e-01 2.69386083e-01 -1.37348041e-01 -2.87923872e-01
1.39954400e+00 -9.59121466e-01 -9.31042016e-01 -1.32862031e-01
3.62070739e-01 -6.09897912e-01 1.03488493e+00 -5.53149432e-02
-8.76249015e-01 -3.52196515e-01 -1.08528841e+00 1.47466436e-01
-2.63468891e-01 4.14277256e-01 3.21918637e-01 9.09667015e-01
-8.41800928e-01 5.40584326e-01 -3.18703026e-01 3.48452702e-02
9.23532963e-01 4.75255370e-01 -8.13894331e-01 2.43669614e-01
-1.23459816e+00 6.28649771e-01 1.88707620e-01 4.87177342e-01
-7.32004166e-01 -7.87838578e-01 -9.20570374e-01 -5.81174344e-03
4.33163077e-01 -4.35791492e-01 7.42738128e-01 -1.90857804e+00
-1.40119600e+00 1.29843986e+00 -1.55468106e-01 -1.45868495e-01
9.56538618e-01 -5.50043816e-03 -5.54878116e-01 7.11593106e-02
2.41625071e-01 5.60405612e-01 1.46997488e+00 -1.54487062e+00
-3.36954266e-01 -5.43365598e-01 -2.61570781e-01 -4.33296477e-03
-4.32179689e-01 -1.02095669e-02 -3.00987929e-01 -8.12080264e-01
2.12775126e-01 -9.56869423e-01 6.08509183e-01 7.67832994e-01
-6.24474645e-01 6.68489560e-02 1.36575496e+00 -7.35185206e-01
1.03848314e+00 -2.39761233e+00 -1.42477721e-01 3.43014687e-01
4.77897763e-01 4.52368349e-01 -2.00591743e-01 -8.65077823e-02
-6.07135713e-01 1.24076352e-01 -2.82807231e-01 -3.45534921e-01
-1.15869820e-01 1.23123666e-02 -6.32877052e-01 7.87288725e-01
6.28494620e-01 8.46521199e-01 -1.00379288e+00 -3.71775657e-01
-2.06921116e-01 5.01731992e-01 -2.67337859e-01 1.75035521e-01
3.45146030e-01 5.36715508e-01 -3.36088568e-01 7.42697239e-01
1.39630485e+00 2.63968781e-02 3.46068799e-01 -4.18255329e-01
6.63969442e-02 1.04395159e-01 -1.27030540e+00 8.59184802e-01
-4.73667607e-02 2.79818356e-01 4.67865646e-01 -6.17142737e-01
9.92610812e-01 2.89071649e-01 9.69539508e-02 -5.37066698e-01
1.93556398e-01 3.46771121e-01 -1.99288264e-01 -1.52016252e-01
1.43424749e-01 -2.66880393e-01 4.69661176e-01 5.79268813e-01
-2.84508497e-01 2.91014552e-01 -4.33532000e-01 -6.37993366e-02
5.60470521e-01 1.11824051e-01 6.17006496e-02 -3.54503065e-01
8.18814099e-01 -5.71412981e-01 7.61439025e-01 2.39133269e-01
-6.06768370e-01 5.32282293e-01 5.49930692e-01 -4.66558844e-01
-6.71360254e-01 -1.10207236e+00 -1.91666380e-01 7.09152758e-01
5.20015419e-01 1.10744677e-01 -9.68178034e-01 -1.23871803e+00
3.85950744e-01 2.51595795e-01 -9.95886862e-01 -4.71293479e-01
-7.36976087e-01 -2.52097994e-01 7.60560036e-01 2.13437960e-01
7.21054673e-01 -1.05772614e+00 -1.42444745e-01 -2.33713344e-01
-1.22946456e-01 -7.76677608e-01 -1.01750290e+00 -5.48367202e-01
-5.43618858e-01 -1.49204421e+00 -6.60996020e-01 -1.01641798e+00
1.11314416e+00 3.94877583e-01 7.19704449e-01 4.81699467e-01
-2.32911050e-01 -1.27576850e-02 -1.24901623e-01 -1.32505313e-01
-5.50524116e-01 -5.64390898e-01 3.36647719e-01 7.98850775e-01
1.64584309e-01 -4.58423495e-01 -6.72213674e-01 6.46617353e-01
-7.95688093e-01 2.38103941e-02 1.97084695e-01 1.03217840e+00
4.73332822e-01 5.50556667e-02 7.09923208e-01 -9.59386647e-01
3.40441287e-01 -4.15211916e-01 -5.24158061e-01 3.38666230e-01
-6.22535467e-01 -1.53617278e-01 4.40103292e-01 -7.11416543e-01
-9.12005961e-01 6.68009669e-02 3.86858195e-01 -9.11121249e-01
1.46757916e-01 -2.93365777e-01 -6.49445295e-01 -4.57064778e-01
2.64139116e-01 3.66093636e-01 3.75062346e-01 -2.65850067e-01
4.23670471e-01 7.02747881e-01 6.95208728e-01 -3.39147389e-01
1.21367502e+00 7.74618566e-01 -2.46657670e-01 -3.92011106e-01
-5.20379841e-01 2.91942023e-02 -4.48201925e-01 -2.54662335e-01
3.78229618e-01 -7.77270734e-01 -8.74865472e-01 8.44501555e-01
-1.20069051e+00 2.02551350e-01 -3.03999066e-01 -5.58848083e-02
-3.11635673e-01 5.38435817e-01 -3.44396383e-01 -9.36984718e-01
-2.93139458e-01 -1.07556653e+00 1.25857985e+00 2.53908366e-01
-1.97881669e-01 -7.59447575e-01 -2.25221902e-01 2.56828725e-01
2.12161094e-01 5.79340219e-01 7.99863219e-01 -3.73406917e-01
-5.77287972e-01 -4.23081875e-01 -4.45678204e-01 5.12943983e-01
6.33455455e-01 2.31355876e-02 -1.16389441e+00 -6.02756023e-01
2.78427392e-01 -2.53221393e-01 6.30280197e-01 -1.84760347e-01
1.07228148e+00 -7.63139725e-01 -4.89387959e-01 7.12784886e-01
1.10236263e+00 2.27085829e-01 8.20099533e-01 1.48482949e-01
7.30593979e-01 9.83324349e-01 5.71293712e-01 2.72473216e-01
5.50141074e-02 6.49307013e-01 5.54049134e-01 -1.35251120e-01
-5.00157289e-02 -5.02376854e-01 3.82614464e-01 1.78015053e-01
3.78158957e-01 9.80073959e-02 -4.01323169e-01 4.13398325e-01
-1.68341196e+00 -1.02534723e+00 1.32826254e-01 2.53079867e+00
8.94671082e-01 -1.53756872e-01 -1.53417857e-02 2.56891586e-02
1.37421477e+00 -8.91758427e-02 -7.54376411e-01 -6.78221928e-03
-2.52441198e-01 1.29134029e-01 2.23986715e-01 3.18983078e-01
-1.20447862e+00 8.59627068e-01 5.29399347e+00 8.64959955e-01
-1.30160165e+00 1.03448302e-01 9.37291145e-01 -2.36209109e-01
-2.45615721e-01 -8.91616791e-02 -6.14067376e-01 9.01315093e-01
1.42108023e-01 -1.98564008e-01 6.51997924e-01 8.70891988e-01
2.65942104e-02 3.52162123e-01 -1.35268593e+00 1.23204863e+00
-6.04797527e-03 -1.26933658e+00 7.16171861e-02 1.94718525e-01
5.22896349e-01 -8.10586095e-01 4.80213046e-01 9.39092785e-02
-1.73442438e-01 -1.21916914e+00 1.05777133e+00 3.80328506e-01
1.27553177e+00 -6.64381206e-01 5.39137185e-01 1.08434565e-01
-1.36469841e+00 -2.49508154e-02 -4.77535315e-02 1.01656832e-01
-9.94485691e-02 2.51812756e-01 -4.49928999e-01 3.79684567e-01
5.88202894e-01 6.07004821e-01 -3.35506916e-01 3.87188613e-01
-1.43564507e-01 2.17950847e-02 -1.95062026e-01 6.47175908e-01
-6.97857961e-02 -5.73746622e-01 6.61655962e-01 6.83011115e-01
3.00056994e-01 1.86312664e-02 -2.09758669e-01 1.47237027e+00
-5.45296252e-01 -1.53277162e-02 -7.18625426e-01 -9.28684846e-02
9.30876136e-01 1.26301837e+00 -3.62439305e-01 -2.93810219e-01
1.08382795e-02 1.23883200e+00 3.55092287e-01 1.93310022e-01
-9.24914122e-01 -2.95842201e-01 1.10235572e+00 2.51554966e-01
6.74793348e-02 6.27361119e-01 -1.81901500e-01 -1.02338767e+00
5.83329320e-01 -9.41113293e-01 1.23327963e-01 -6.06185138e-01
-1.50513256e+00 6.87475920e-01 -3.24555427e-01 -1.43487751e+00
2.46195167e-01 -4.37430859e-01 -9.33910668e-01 1.21487451e+00
-1.33428872e+00 -1.11597788e+00 -5.39472222e-01 5.04080057e-01
-3.64647172e-02 -2.09631994e-01 6.06731117e-01 3.31408471e-01
-6.54839993e-01 1.19771123e+00 -1.41744062e-01 4.88730609e-01
7.44660318e-01 -6.03130877e-01 3.67037565e-01 5.12419283e-01
-2.03153789e-01 8.59955370e-01 3.47444683e-01 -7.09413588e-01
-1.22285175e+00 -1.28557980e+00 6.26216590e-01 -3.73250812e-01
4.50593084e-01 -4.48274821e-01 -1.11748326e+00 6.36116505e-01
-2.52003521e-01 4.45703596e-01 4.75470394e-01 -6.00481749e-01
-8.92756999e-01 -5.97385392e-02 -1.73911202e+00 5.86047828e-01
1.12357736e+00 -8.68261278e-01 -3.52468640e-01 3.38119298e-01
4.70900685e-01 -6.44235611e-02 -4.06579882e-01 4.81725305e-01
8.84494364e-01 -1.06585479e+00 1.00054777e+00 -5.41523933e-01
1.92448273e-01 -5.32238960e-01 3.18251044e-01 -8.87234151e-01
-6.79626316e-02 -6.40160263e-01 -2.04785377e-01 1.48554099e+00
-1.79846480e-01 -1.01190209e+00 9.97844577e-01 5.69748700e-01
2.98571408e-01 -8.29966843e-01 -1.15581787e+00 -1.07156217e+00
7.28146583e-02 2.51328707e-01 8.71950150e-01 1.43484783e+00
-3.41691107e-01 -1.40477017e-01 -4.81524318e-01 2.69849449e-01
6.58762753e-01 1.73055723e-01 6.07302010e-01 -1.08010960e+00
-4.63714525e-02 -5.52908778e-01 -3.99798363e-01 -6.90846264e-01
4.79082227e-01 -7.93677628e-01 -1.61141559e-01 -4.93980289e-01
1.26600295e-01 -7.45554328e-01 -2.07093179e-01 6.06003344e-01
-2.12635964e-01 5.55024505e-01 1.36429220e-01 4.09586459e-01
7.29369745e-02 9.37334359e-01 1.41098893e+00 -2.67035246e-01
-9.28654969e-02 -2.34823853e-01 -8.34377408e-01 8.42783034e-01
5.62387526e-01 -5.77038229e-01 -2.67398179e-01 1.19918641e-02
-2.15344176e-01 -1.16602384e-01 8.22060943e-01 -5.80200374e-01
-1.00272968e-01 -9.92788896e-02 2.93421328e-01 -4.55301814e-02
3.44517201e-01 -8.59765708e-01 3.95991087e-01 6.16304874e-01
-1.17509872e-01 -1.86441973e-01 8.65873992e-02 5.83789587e-01
-2.09024563e-01 -9.89438221e-02 1.14309371e+00 -1.94985177e-02
-3.48709494e-01 5.21972120e-01 3.31948698e-01 6.06124997e-02
1.24684227e+00 -6.84774101e-01 -3.85021150e-01 -1.48031637e-01
-6.08645320e-01 1.60325930e-01 7.75843799e-01 5.17403603e-01
8.70856881e-01 -1.84649312e+00 -6.58159196e-01 7.18500614e-01
2.89663464e-01 -2.79932290e-01 6.79375753e-02 8.70914996e-01
-3.45632762e-01 -1.41634718e-01 -2.30208367e-01 -2.17111021e-01
-1.58247197e+00 7.87069738e-01 7.11395323e-01 2.32120454e-01
-3.82959336e-01 1.03137255e+00 8.79815638e-01 -6.11118436e-01
7.49569945e-03 2.28128538e-01 8.35361630e-02 7.72614554e-02
7.85433948e-01 4.11108911e-01 -2.36577660e-01 -1.10518813e+00
-6.30479574e-01 4.13169265e-01 -1.29141808e-01 2.38205448e-01
8.21360469e-01 -7.81411529e-02 -4.92652237e-01 -1.32172287e-01
1.42731655e+00 1.60014138e-01 -1.42163336e+00 -2.39129096e-01
-1.74577489e-01 -1.11735868e+00 -2.77841717e-01 -6.36868298e-01
-1.36038959e+00 6.11221492e-01 8.35475087e-01 -5.08167893e-02
1.12545204e+00 -1.60302341e-01 7.03679979e-01 -1.67452663e-01
3.98384273e-01 -8.92187715e-01 1.14886910e-01 8.70648921e-02
1.02722692e+00 -1.28782856e+00 -2.43661448e-01 -8.64094257e-01
-4.81375009e-01 8.98079038e-01 7.84883976e-01 -1.42135382e-01
6.11914635e-01 -4.36435342e-02 3.49482372e-02 -1.57467812e-01
-2.45741785e-01 3.24595541e-01 2.85951644e-01 5.77945113e-01
-7.37377331e-02 1.26291856e-01 1.37759238e-01 5.84715545e-01
-1.49076104e-01 -7.22839832e-02 1.48042277e-01 6.89107239e-01
-9.49704200e-02 -1.09586275e+00 -6.01428449e-01 1.08754829e-01
8.25495645e-02 -9.63562056e-02 -7.11336195e-01 8.15445244e-01
6.86416924e-01 6.41907156e-01 2.29868367e-01 -4.12893623e-01
8.44257623e-02 8.76160115e-02 5.60809791e-01 -4.87174988e-01
-5.71388364e-01 -3.18159670e-01 -4.48703766e-01 -5.18731058e-01
-9.37919468e-02 -3.80370736e-01 -1.13984227e+00 -6.30974472e-01
-5.64278185e-01 8.71142000e-02 3.13310862e-01 7.19364464e-01
5.13533115e-01 -6.71903603e-04 1.11074245e+00 -7.90705442e-01
-6.43436134e-01 -6.26236916e-01 -7.31432438e-01 7.52635658e-01
5.21880567e-01 -9.42155659e-01 -6.21552527e-01 8.09198767e-02] | [12.764129638671875, 0.22504787147045135] |
c8d6efd4-862f-49ab-8eee-588165227b68 | towards-real-world-hdrtv-reconstruction-a | 2211.03058 | null | https://arxiv.org/abs/2211.03058v1 | https://arxiv.org/pdf/2211.03058v1.pdf | Towards Real World HDRTV Reconstruction: A Data Synthesis-based Approach | Existing deep learning based HDRTV reconstruction methods assume one kind of tone mapping operators (TMOs) as the degradation procedure to synthesize SDRTV-HDRTV pairs for supervised training. In this paper, we argue that, although traditional TMOs exploit efficient dynamic range compression priors, they have several drawbacks on modeling the realistic degradation: information over-preservation, color bias and possible artifacts, making the trained reconstruction networks hard to generalize well to real-world cases. To solve this problem, we propose a learning-based data synthesis approach to learn the properties of real-world SDRTVs by integrating several tone mapping priors into both network structures and loss functions. In specific, we design a conditioned two-stream network with prior tone mapping results as a guidance to synthesize SDRTVs by both global and local transformations. To train the data synthesis network, we form a novel self-supervised content loss to constraint different aspects of the synthesized SDRTVs at regions with different brightness distributions and an adversarial loss to emphasize the details to be more realistic. To validate the effectiveness of our approach, we synthesize SDRTV-HDRTV pairs with our method and use them to train several HDRTV reconstruction networks. Then we collect two inference datasets containing both labeled and unlabeled real-world SDRTVs, respectively. Experimental results demonstrate that, the networks trained with our synthesized data generalize significantly better to these two real-world datasets than existing solutions. | ['Zhiwei Xiong', 'Chang Chen', 'Fenglong Song', 'Yong Li', 'Tao Wang', 'Zhen Cheng'] | 2022-11-06 | null | null | null | null | ['tone-mapping'] | ['computer-vision'] | [ 4.00826067e-01 -9.73611251e-02 -2.92692363e-01 -5.83557844e-01
-7.45851219e-01 -3.54555726e-01 4.82927620e-01 -6.68938220e-01
-5.24290614e-02 8.01097810e-01 2.32979611e-01 -1.60186738e-01
2.48357475e-01 -9.36813056e-01 -1.10329056e+00 -7.72132397e-01
3.48204553e-01 2.64965385e-01 2.05982938e-01 -3.06803048e-01
-1.01752423e-01 4.87324089e-01 -1.38405740e+00 2.43817553e-01
1.24200785e+00 1.12169945e+00 3.88768435e-01 5.11541784e-01
-3.21577303e-02 1.14423180e+00 -7.29514718e-01 -3.28897268e-01
7.46923387e-01 -6.97177649e-01 -4.61454719e-01 1.24339081e-01
4.11810607e-01 -7.44316757e-01 -1.06194627e+00 1.03507960e+00
6.36785269e-01 2.34016284e-01 6.27181351e-01 -1.30199838e+00
-9.64451313e-01 3.95380825e-01 -5.55109918e-01 -8.75380784e-02
1.04835376e-01 5.51597297e-01 8.35271895e-01 -5.58159113e-01
7.77879894e-01 1.16744232e+00 6.90400243e-01 5.27728617e-01
-1.24681747e+00 -7.35448718e-01 -1.99813806e-02 2.49426320e-01
-1.28496277e+00 -4.13779974e-01 1.05424356e+00 1.85955107e-01
6.57009542e-01 3.41581821e-01 5.47390640e-01 1.35028839e+00
1.77489281e-01 8.68990898e-01 1.38769507e+00 -1.34364337e-01
1.68427512e-01 2.12562785e-01 -5.91361403e-01 4.91808236e-01
-2.41799161e-01 5.31468272e-01 -1.42346159e-01 3.79973531e-01
1.00023913e+00 1.30193150e-02 -7.36482978e-01 -4.19506490e-01
-1.01205254e+00 7.39529729e-01 6.98417008e-01 -3.40257511e-02
6.11457676e-02 2.92771935e-01 3.93703908e-01 5.77686787e-01
3.84550840e-01 9.39016640e-02 -2.49147877e-01 2.93540716e-01
-8.24569046e-01 4.28924412e-02 4.25368726e-01 1.10891593e+00
8.49958897e-01 4.40977514e-01 -4.67227906e-01 1.15110385e+00
4.22901034e-01 8.48718882e-01 4.38162029e-01 -9.78565812e-01
8.17001045e-01 6.60412759e-02 8.49821512e-03 -6.99240208e-01
8.84748399e-02 -5.54177880e-01 -1.00609767e+00 4.93656456e-01
8.10078308e-02 1.10590681e-01 -1.35988200e+00 1.77150369e+00
6.29070178e-02 4.46150005e-01 2.55989909e-01 1.22370434e+00
9.69267309e-01 1.12893450e+00 -2.62353331e-01 -2.66862422e-01
7.60199308e-01 -1.03322077e+00 -1.00187361e+00 -1.15412533e-01
1.58356816e-01 -7.67356157e-01 1.53441060e+00 4.24373955e-01
-1.10688162e+00 -7.63580501e-01 -1.44045424e+00 -3.24348748e-01
-6.48279041e-02 -1.06303647e-01 2.95345485e-01 5.05176604e-01
-9.54725087e-01 6.82215571e-01 -4.37230080e-01 8.53052065e-02
4.63224530e-01 -3.68567891e-02 3.58454622e-02 -5.00840902e-01
-1.55126286e+00 8.13429296e-01 1.79500848e-01 1.96552426e-01
-1.49516654e+00 -7.49285340e-01 -7.58906662e-01 -4.17064689e-02
2.92537034e-01 -6.34353399e-01 1.01115727e+00 -1.21283424e+00
-1.81138027e+00 7.94143021e-01 3.09013724e-01 -3.20602298e-01
7.91677952e-01 1.38201311e-01 -5.93279600e-01 2.46639233e-02
-7.26802424e-02 5.05769193e-01 1.15823436e+00 -1.78951287e+00
-3.32894295e-01 7.83106983e-02 -4.00680006e-02 2.89353251e-01
2.82411650e-03 -2.54839659e-01 -5.71288645e-01 -1.07355833e+00
-2.74973903e-02 -5.17093182e-01 9.18744281e-02 3.98245424e-01
-5.33423305e-01 3.12024355e-01 1.04998994e+00 -7.22769558e-01
8.33344638e-01 -2.10841393e+00 6.48350716e-02 2.61560798e-01
5.04196770e-02 2.51227200e-01 -3.37031901e-01 1.83487460e-01
-1.49429977e-01 -1.09679848e-01 -5.42887747e-01 -4.03528988e-01
1.20885573e-01 5.32674670e-01 -8.10465872e-01 4.90207642e-01
4.08010632e-02 9.41823840e-01 -6.65800214e-01 -4.24988717e-01
3.90420586e-01 7.41696894e-01 -1.93237633e-01 6.22492015e-01
-4.61588353e-01 6.45887613e-01 -1.99264154e-01 5.69770217e-01
1.07689238e+00 4.04744856e-02 -4.62220386e-02 -5.81968307e-01
2.15151772e-01 3.11200265e-02 -1.02040112e+00 1.81011033e+00
-6.82734609e-01 7.31080055e-01 -1.69202611e-01 -1.02174425e+00
1.05398953e+00 -4.40959266e-04 3.20312977e-01 -1.22625017e+00
2.66917408e-01 1.45350829e-01 -3.16492438e-01 -4.62499619e-01
3.10502917e-01 -5.54272115e-01 1.64982110e-01 4.32903618e-01
-5.32280933e-03 -5.20706952e-01 -3.63415748e-01 6.46760389e-02
8.84535491e-01 3.83756429e-01 -3.10212642e-01 1.41187623e-01
3.93215507e-01 -4.64937836e-01 8.49718094e-01 5.42914271e-01
-1.32823184e-01 1.19996154e+00 2.87463337e-01 -2.20894933e-01
-1.48723507e+00 -1.52264118e+00 -1.86615586e-01 6.06588483e-01
5.15019834e-01 7.83619881e-02 -5.59932828e-01 -6.57267272e-01
-2.84658253e-01 8.81854534e-01 -5.15092731e-01 -3.74073595e-01
-5.27907193e-01 -6.20376170e-01 6.88931823e-01 4.65999216e-01
1.02252162e+00 -1.10324907e+00 -1.69255137e-01 -3.28610986e-02
-4.28767264e-01 -1.31446767e+00 -6.70496225e-01 2.31990784e-01
-6.97951138e-01 -6.52574956e-01 -8.85053217e-01 -7.76886284e-01
4.85744804e-01 1.55109137e-01 1.22900105e+00 6.38038665e-02
-2.30353519e-01 9.86025333e-02 -3.41571569e-01 -3.59904021e-03
-6.78671658e-01 -3.55853319e-01 -2.21749589e-01 1.72024816e-02
-3.50454777e-01 -7.11156189e-01 -7.56785929e-01 5.24003804e-01
-1.28814387e+00 2.13991016e-01 4.74762797e-01 7.60870576e-01
8.96872878e-01 2.56792486e-01 3.54275316e-01 -7.44786739e-01
2.93593198e-01 -3.10418487e-01 -5.64885855e-01 4.16698962e-01
-5.85377514e-01 5.60886897e-02 1.06135333e+00 -4.47895974e-01
-1.40847898e+00 -6.40691072e-02 -3.43866765e-01 -9.82458472e-01
7.86131471e-02 -4.46901135e-02 -5.89474678e-01 -1.72890544e-01
6.08775914e-01 6.84910357e-01 -1.01514280e-01 -1.27897099e-01
5.72857380e-01 4.99143511e-01 6.76596582e-01 -6.09369874e-01
1.36862588e+00 6.33428931e-01 -1.81324854e-01 -4.26177472e-01
-6.76822424e-01 1.97656482e-01 1.29965581e-02 -1.99950367e-01
8.55601251e-01 -1.02285492e+00 -3.74432832e-01 6.17072403e-01
-8.37741196e-01 -9.12598193e-01 -5.52006245e-01 1.77879184e-01
-1.04317391e+00 5.62268198e-01 -6.50256217e-01 -3.72570246e-01
-2.95316786e-01 -1.22167552e+00 1.01198053e+00 1.67912841e-01
5.70541799e-01 -8.85864675e-01 1.50521517e-01 2.77641088e-01
4.13434148e-01 3.71481180e-01 9.34229136e-01 9.93495807e-02
-1.00683451e+00 3.40992749e-01 -5.08566678e-01 7.70283341e-01
2.35077739e-01 -7.03255236e-02 -1.13196015e+00 -3.88396025e-01
1.96662396e-01 -6.35634601e-01 9.35765207e-01 3.88090700e-01
1.62716758e+00 -1.64200217e-01 1.33923842e-02 1.39407158e+00
1.79746473e+00 1.74429312e-01 1.38393474e+00 1.87702879e-01
8.70321155e-01 3.01268578e-01 6.98421896e-01 3.71630222e-01
2.45747164e-01 7.35245943e-01 6.60525680e-01 -3.66717309e-01
-6.29819810e-01 -5.98436832e-01 5.25412440e-01 5.91245294e-01
1.27488151e-01 -6.90169156e-01 -3.47264469e-01 1.91646650e-01
-1.52585983e+00 -7.79787719e-01 1.12159252e-01 2.12796974e+00
9.93881106e-01 -3.29916994e-03 -2.26584122e-01 3.73833515e-02
6.27004027e-01 5.62540829e-01 -9.03666973e-01 -1.06239818e-01
-5.59658587e-01 2.55391181e-01 6.28549278e-01 4.08305526e-01
-8.47971022e-01 9.19793725e-01 5.89242697e+00 1.05500472e+00
-1.38032222e+00 1.63332283e-01 8.69777501e-01 7.78500317e-03
-9.13310826e-01 -2.47014061e-01 -2.73337454e-01 6.57458305e-01
5.23124099e-01 3.47828925e-01 8.93083692e-01 4.10341531e-01
3.95455092e-01 3.18441242e-01 -8.28050733e-01 1.13608849e+00
9.83518362e-02 -1.21720707e+00 6.23205192e-02 -2.00197279e-01
9.36222255e-01 5.81660978e-02 3.72459024e-01 2.20370129e-01
5.46902716e-01 -1.09835768e+00 7.90979564e-01 4.64582026e-01
1.24985731e+00 -6.81350648e-01 4.89007145e-01 8.64125043e-02
-1.09239507e+00 -8.29081610e-02 -6.00875735e-01 5.45366764e-01
2.99158186e-01 4.02668327e-01 -4.10469979e-01 7.19102859e-01
7.98001409e-01 8.92068803e-01 -4.29190874e-01 7.18379498e-01
-4.61815625e-01 4.44545865e-01 -1.62448749e-01 4.44285572e-01
-1.01496484e-02 -2.68308342e-01 3.20147246e-01 7.19230711e-01
4.43154067e-01 -1.69665124e-02 -2.26077616e-01 1.16239738e+00
-3.18688869e-01 -1.33332253e-01 -6.47854745e-01 2.25918099e-01
3.51262122e-01 1.01630247e+00 -5.31931043e-01 -1.58491656e-01
-4.07192230e-01 1.37934780e+00 1.72461838e-01 7.99346626e-01
-1.37426531e+00 -4.59356576e-01 5.76406717e-01 1.77713796e-01
4.67589796e-01 1.29891723e-01 -3.73807192e-01 -1.45948589e+00
-7.28469193e-02 -1.01255262e+00 8.44163224e-02 -1.30752254e+00
-1.55507600e+00 6.45357132e-01 -1.40766859e-01 -1.63506699e+00
8.24536532e-02 -3.52582991e-01 -9.25957859e-01 7.45614529e-01
-1.95469177e+00 -1.27044463e+00 -6.62641883e-01 9.07766700e-01
4.80775237e-01 -2.80605614e-01 2.23986968e-01 5.56274414e-01
-4.76407707e-01 8.34347129e-01 2.75752515e-01 -1.93146635e-02
9.39987898e-01 -1.02392888e+00 3.17429692e-01 9.31782484e-01
-8.70363265e-02 1.14854157e-01 8.91367078e-01 -4.84421670e-01
-1.34666157e+00 -1.33424258e+00 -6.30852357e-02 1.59368571e-02
3.26632798e-01 -4.51634169e-01 -1.10883641e+00 7.69251943e-01
4.16593313e-01 3.04415256e-01 1.32377937e-01 -6.00400388e-01
-4.58502710e-01 -5.72703600e-01 -1.47436571e+00 5.30050218e-01
1.40172327e+00 -5.17892957e-01 -3.62028688e-01 2.39952818e-01
1.12973523e+00 -5.20838678e-01 -9.25642848e-01 5.06606877e-01
1.88947111e-01 -1.14166307e+00 1.23929739e+00 -1.55023664e-01
7.73893774e-01 -6.85487688e-01 -4.36514884e-01 -1.54355717e+00
2.16763373e-02 -4.69856381e-01 -3.03395535e-03 1.36202705e+00
1.11849882e-01 -7.40931988e-01 7.42165625e-01 2.00005248e-01
-5.09716809e-01 -5.77829838e-01 -8.02298486e-01 -7.07703292e-01
2.39478037e-01 -4.36535597e-01 7.94070005e-01 9.88452017e-01
-7.60674655e-01 1.60345197e-01 -8.66399288e-01 8.47139657e-02
8.29766750e-01 1.87138602e-01 7.97394335e-01 -6.66859746e-01
-5.32142818e-01 -1.11187272e-01 -8.88891295e-02 -1.29230201e+00
2.48168781e-01 -7.35463619e-01 3.42556775e-01 -1.46384871e+00
2.41640173e-02 -8.04878891e-01 -2.01792940e-01 2.07269758e-01
4.81384024e-02 3.81326765e-01 2.06034198e-01 1.80804491e-01
-3.07013631e-01 1.20002091e+00 1.85115588e+00 -3.22640359e-01
-3.92558500e-02 -1.51757881e-01 -4.84959304e-01 2.70301551e-01
6.28828168e-01 -2.71723330e-01 -1.06653416e+00 -5.65116286e-01
1.37650266e-01 1.48260698e-01 4.13955122e-01 -9.52013552e-01
-1.56923801e-01 -2.33151451e-01 5.48233449e-01 -5.15920162e-01
3.96002889e-01 -1.05433428e+00 3.50490898e-01 2.29558349e-01
-2.66470969e-01 -4.50999200e-01 -4.07511815e-02 5.79797983e-01
-2.57670552e-01 -8.84896144e-02 1.10847080e+00 4.21849033e-03
-7.26994038e-01 7.12948084e-01 6.79683164e-02 2.61202902e-01
9.04269278e-01 -2.84476221e-01 -5.22586465e-01 -8.12119067e-01
-3.80317479e-01 1.59901381e-01 7.07962513e-01 4.07476813e-01
1.05489564e+00 -1.60811746e+00 -5.50732374e-01 4.40379769e-01
1.22219615e-01 1.71930134e-01 5.11197746e-01 3.68429661e-01
-8.70317340e-01 -2.20733508e-01 -4.83740330e-01 -6.12650275e-01
-7.32595503e-01 7.57652640e-01 6.37083948e-01 -1.46558732e-01
-9.68585074e-01 5.40903449e-01 4.36722338e-01 -7.04362631e-01
4.27574456e-01 -4.57770348e-01 6.23981655e-02 -4.92651165e-01
1.73711926e-01 4.14449684e-02 -2.04070389e-01 -4.16485429e-01
1.51927829e-01 6.59307122e-01 1.32121667e-01 -4.55217622e-02
1.24166334e+00 -4.81823623e-01 2.37298980e-01 1.65938526e-01
1.51514304e+00 -2.23566115e-01 -1.79523170e+00 -3.58186275e-01
-7.46628463e-01 -7.80449927e-01 1.07432306e-01 -9.24752891e-01
-1.68188667e+00 7.72394001e-01 1.08999431e+00 -3.58014517e-02
1.59697914e+00 -1.49794534e-01 1.28022432e+00 1.16843045e-01
2.10136294e-01 -1.12654269e+00 4.30222392e-01 1.60008460e-01
1.03122640e+00 -1.30458283e+00 -1.34708434e-01 -4.97951001e-01
-8.75522435e-01 8.49901319e-01 7.55886972e-01 -2.55687594e-01
4.20721591e-01 2.91124821e-01 3.30345869e-01 2.36925080e-01
-5.29723942e-01 6.93035964e-03 4.72399592e-02 8.53128076e-01
-6.55384883e-02 -2.59327263e-01 8.24307278e-02 1.62304312e-01
-7.22563416e-02 6.03479370e-02 5.82552612e-01 5.04035830e-01
-2.50781208e-01 -1.02877045e+00 -2.25559056e-01 2.41163790e-01
-1.09877512e-01 -2.50082090e-02 3.01396139e-02 7.49282122e-01
1.67007625e-01 5.54295957e-01 7.42650107e-02 -5.97084105e-01
4.57596034e-01 -5.64537704e-01 5.92690945e-01 -1.72430396e-01
-2.36290425e-01 5.74793071e-02 -1.09070309e-01 -5.86300850e-01
-2.32210994e-01 -2.08956361e-01 -1.36478961e+00 -4.85437423e-01
7.31016174e-02 -3.66716176e-01 4.25663322e-01 8.00306141e-01
-2.56234165e-02 8.79743040e-01 1.21061277e+00 -6.13100231e-01
-5.77799618e-01 -6.67385340e-01 -8.31137419e-01 6.79491043e-01
4.98343319e-01 -5.95057487e-01 -4.82211411e-01 -3.62246297e-02] | [11.014676094055176, -2.026592493057251] |
a605ff1d-58ea-4e46-a945-5cbbffce6c88 | second-order-information-in-first-order | 1912.09926 | null | https://arxiv.org/abs/1912.09926v1 | https://arxiv.org/pdf/1912.09926v1.pdf | Second-order Information in First-order Optimization Methods | In this paper, we try to uncover the second-order essence of several first-order optimization methods. For Nesterov Accelerated Gradient, we rigorously prove that the algorithm makes use of the difference between past and current gradients, thus approximates the Hessian and accelerates the training. For adaptive methods, we related Adam and Adagrad to a powerful technique in computation statistics---Natural Gradient Descent. These adaptive methods can in fact be treated as relaxations of NGD with only a slight difference lying in the square root of the denominator in the update rules. Skeptical about the effect of such difference, we design a new algorithm---AdaSqrt, which removes the square root in the denominator and scales the learning rate by sqrt(T). Surprisingly, our new algorithm is comparable to various first-order methods(such as SGD and Adam) on MNIST and even beats Adam on CIFAR-10! This phenomenon casts doubt on the convention view that the square root is crucial and training without it will lead to terrible performance. As far as we have concerned, so long as the algorithm tries to explore second or even higher information of the loss surface, then proper scaling of the learning rate alone will guarantee fast training and good generalization performance. To the best of our knowledge, this is the first paper that seriously considers the necessity of square root among all adaptive methods. We believe that our work can shed light on the importance of higher-order information and inspire the design of more powerful algorithms in the future. | ['Shange Tang', 'Licong Lin', 'Yuzheng Hu'] | 2019-12-20 | null | null | null | null | ['2d-human-pose-estimation'] | ['computer-vision'] | [-2.06666812e-01 -4.65092584e-02 -3.68987955e-02 -5.13743877e-01
-4.51522082e-01 -3.66371363e-01 3.39476317e-01 2.10348666e-01
-9.10940886e-01 7.49394655e-01 -2.46255789e-02 -7.33503163e-01
-1.23943366e-01 -5.95008492e-01 -6.99666381e-01 -9.07029331e-01
-2.86510944e-01 3.60480487e-01 2.96717405e-01 -5.19885421e-01
4.51239884e-01 4.13880289e-01 -1.19761670e+00 -2.17289910e-01
8.76407623e-01 9.23080981e-01 -5.66811375e-02 6.47072256e-01
-1.32556222e-02 7.64994621e-01 -3.34344387e-01 -5.61097980e-01
5.18177152e-01 -5.64877510e-01 -7.88837433e-01 -1.53338641e-01
5.04381895e-01 -4.80263025e-01 -2.71001011e-01 1.07850158e+00
4.89250600e-01 3.54285002e-01 5.74066401e-01 -8.09620440e-01
-3.60728800e-01 4.57587570e-01 -5.49002409e-01 5.51014662e-01
-1.18775956e-01 4.55114432e-02 1.16171014e+00 -9.11988199e-01
4.95586812e-01 9.44027841e-01 9.93880689e-01 4.88385618e-01
-1.00710475e+00 -2.73084670e-01 3.28155160e-01 1.22369677e-01
-1.20904338e+00 -2.69845992e-01 7.48077691e-01 -3.07771415e-01
8.41891944e-01 3.94014597e-01 6.51203930e-01 5.34245610e-01
7.79774934e-02 9.73640859e-01 1.12013078e+00 -6.37044072e-01
3.99952829e-01 3.13024759e-01 2.66708136e-01 9.76810098e-01
3.55914652e-01 2.77226835e-01 -4.48155701e-01 -2.24311724e-01
6.92433834e-01 -2.62373507e-01 -3.16160381e-01 -2.52134949e-01
-8.61086071e-01 1.07805145e+00 3.70556563e-01 5.39072514e-01
-2.60755181e-01 2.21554577e-01 4.41693664e-01 5.06344736e-01
6.62469923e-01 6.02184772e-01 -5.77034414e-01 -3.21271211e-01
-1.05419624e+00 3.85260642e-01 9.63274896e-01 2.26426274e-01
6.27327204e-01 2.16637805e-01 2.30336919e-01 6.69916630e-01
1.82440430e-01 2.24730074e-01 5.42492330e-01 -9.73356605e-01
2.13353202e-01 2.23780438e-01 1.72345966e-01 -9.27342653e-01
-4.63418514e-01 -7.84043789e-01 -6.82846725e-01 4.34580415e-01
8.48280072e-01 -2.93184966e-01 -4.70322579e-01 1.65338194e+00
3.48557442e-01 -6.65022284e-02 -3.00113976e-01 1.14912653e+00
7.40222037e-02 4.44250077e-01 -2.93925107e-01 -3.39389205e-01
9.42643642e-01 -7.11750031e-01 -4.66175377e-01 -4.37445939e-02
7.86164641e-01 -5.43123424e-01 1.26178372e+00 6.07116282e-01
-1.14412975e+00 -2.71021128e-01 -1.24930120e+00 1.96489375e-02
-2.15974227e-01 1.96452085e-02 1.11072004e+00 5.75320959e-01
-9.68977928e-01 1.24520671e+00 -1.02023232e+00 -5.64396493e-02
6.05147704e-02 3.00595701e-01 7.38883391e-03 3.22160035e-01
-1.10434198e+00 9.04748738e-01 1.45231664e-01 3.01166832e-01
-3.88830453e-01 -7.14482665e-01 -4.61292237e-01 -1.27642229e-01
1.96587473e-01 -5.17713070e-01 1.28570402e+00 -1.01097691e+00
-1.48839998e+00 7.00685382e-01 -3.03088456e-01 -8.07297707e-01
8.55892777e-01 -4.29265171e-01 -5.51047176e-02 -1.13044932e-01
-1.55921593e-01 2.21956745e-01 8.69111598e-01 -8.69190276e-01
-4.66575682e-01 -5.58451116e-01 1.47302740e-03 2.73450434e-01
-4.17742133e-01 -2.84176141e-01 5.01428992e-02 -6.21313632e-01
2.79012054e-01 -8.61021399e-01 -3.42823118e-01 8.56952295e-02
1.25083998e-01 -2.75855452e-01 2.91703850e-01 -4.77423668e-01
1.25479162e+00 -2.19209933e+00 -1.05321415e-01 3.28308523e-01
2.68406034e-01 2.92243183e-01 1.38228461e-01 2.87966907e-01
2.10691355e-02 8.32819380e-03 -2.60175884e-01 -2.80510843e-01
7.24292547e-02 2.92635709e-01 -4.57633793e-01 8.40329587e-01
-1.13753945e-01 6.93851650e-01 -1.02087677e+00 -2.22889766e-01
-1.69215932e-01 4.63227093e-01 -6.45350397e-01 -2.38591135e-01
-5.65881766e-02 1.35634735e-01 -5.35865366e-01 5.95592484e-02
5.52621663e-01 -1.89498380e-01 -5.52588478e-02 -9.50777903e-02
-3.98124278e-01 5.54884672e-01 -1.24382854e+00 1.29592872e+00
-2.51633435e-01 6.98328376e-01 -4.53162715e-02 -1.31025171e+00
6.77131474e-01 8.77731144e-02 4.98403072e-01 -6.83912098e-01
2.51068491e-02 6.14927530e-01 5.90875037e-02 -2.85831362e-01
3.87344450e-01 -3.29720885e-01 4.41323936e-01 3.22046101e-01
-1.67037085e-01 -2.11026251e-01 4.27295059e-01 9.80434343e-02
8.43791783e-01 -5.15399203e-02 3.73466238e-02 -4.89270657e-01
4.80906785e-01 -8.90385434e-02 5.11917472e-01 9.46638107e-01
-4.41991985e-02 3.67269576e-01 4.91607368e-01 -9.09794331e-01
-1.09695017e+00 -8.65910888e-01 -2.39893481e-01 1.17729223e+00
-2.48323515e-01 -3.43210489e-01 -6.23650908e-01 -6.73757792e-01
1.35223150e-01 7.30711639e-01 -6.06741369e-01 -1.13871612e-01
-6.70892417e-01 -1.31220031e+00 4.16473985e-01 3.71195197e-01
5.57056725e-01 -6.04487002e-01 -5.21650672e-01 1.89051464e-01
1.80609524e-01 -7.21393228e-01 -5.05518794e-01 4.71896350e-01
-1.33013904e+00 -8.28145385e-01 -6.73785269e-01 -5.66258490e-01
5.39157748e-01 -6.51080981e-02 1.02123964e+00 2.31539756e-01
-4.02824134e-02 1.75536349e-01 -1.67087153e-01 -4.33007896e-01
-2.33979404e-01 5.15825637e-02 1.55480623e-01 -1.69691771e-01
2.98280954e-01 -8.53402317e-01 -6.38114274e-01 2.79226340e-02
-5.96942544e-01 -3.19098532e-01 4.92672205e-01 9.04086888e-01
4.71582681e-01 1.31026626e-01 3.07431251e-01 -9.26066101e-01
7.87166953e-01 -1.60053760e-01 -8.11844468e-01 2.31205542e-02
-1.03997886e+00 6.10912204e-01 7.55352080e-01 -4.38616484e-01
-7.98256397e-01 -1.65708825e-01 -3.26139420e-01 -1.12950660e-01
3.87033939e-01 4.43482339e-01 4.70188349e-01 -4.14790899e-01
8.69525969e-01 3.68796587e-01 -2.23730937e-01 -6.32658780e-01
2.72978067e-01 2.15032741e-01 3.90039772e-01 -6.30753696e-01
6.96868718e-01 6.36964381e-01 2.29246527e-01 -8.88974130e-01
-1.10406137e+00 -4.18928385e-01 -2.31603637e-01 7.07979351e-02
4.08014864e-01 -5.20525753e-01 -8.11324894e-01 4.36739177e-01
-9.91734147e-01 -3.90268356e-01 -5.75498819e-01 6.87612355e-01
-4.66955453e-01 4.16478306e-01 -9.36238587e-01 -1.03039718e+00
-4.21226770e-01 -7.20267415e-01 4.78486419e-01 6.47548139e-02
8.52659997e-03 -1.21602809e+00 2.15219185e-01 5.26915421e-04
5.32207310e-01 -4.17198017e-02 6.18741751e-01 -6.28633499e-01
-1.25832990e-01 -1.98301896e-01 2.54470184e-02 7.12794006e-01
-2.63360262e-01 -4.86552268e-02 -7.85354257e-01 -2.64201760e-01
6.45060658e-01 -2.07449049e-01 1.07751048e+00 4.66351777e-01
8.40700984e-01 -5.22330105e-01 1.02947123e-01 6.67660892e-01
1.39742017e+00 8.64887051e-03 3.38538408e-01 4.54836249e-01
5.15995324e-01 2.65514851e-01 4.22714293e-01 6.32417560e-01
3.60127479e-01 4.07430142e-01 1.79049298e-01 -1.62860841e-01
6.03501685e-02 -1.01016507e-01 3.28749150e-01 1.13873029e+00
-3.27045441e-01 2.48823673e-01 -7.38439500e-01 4.12095010e-01
-1.67849839e+00 -8.67369831e-01 -3.00466120e-01 2.37343478e+00
1.07853317e+00 4.63484496e-01 1.29444182e-01 2.75040984e-01
2.55712956e-01 1.65316612e-01 -6.40563846e-01 -6.23842955e-01
-8.44906047e-02 2.16508895e-01 7.74946570e-01 8.71150076e-01
-9.14777875e-01 7.29271412e-01 6.62613916e+00 9.30196881e-01
-1.19903696e+00 1.19041957e-01 6.44702435e-01 -1.37482002e-01
-2.21025005e-01 2.09505334e-01 -8.42553437e-01 3.32541406e-01
6.70210660e-01 2.76019070e-02 6.48554206e-01 9.59946513e-01
1.52264550e-01 -2.52215952e-01 -9.58227992e-01 7.40143299e-01
-5.46202697e-02 -9.91151690e-01 -3.06568980e-01 7.12764114e-02
7.59430051e-01 3.26527208e-01 1.40983135e-01 1.89095676e-01
2.25187391e-01 -7.69903898e-01 5.77611387e-01 3.25534821e-01
1.24108590e-01 -7.23233104e-01 7.02270865e-01 4.51174915e-01
-7.81425536e-01 9.64014325e-03 -5.72494447e-01 -4.09481466e-01
2.84725353e-02 1.11336589e+00 -5.86548746e-01 1.00532964e-01
4.41437244e-01 3.69462639e-01 -4.25371319e-01 1.06432188e+00
-2.61007518e-01 8.69674027e-01 -7.63479054e-01 -3.45255286e-01
5.15937626e-01 -4.87059474e-01 6.42792523e-01 1.03812349e+00
3.54837887e-02 1.21336311e-01 -8.68047494e-03 4.86894220e-01
1.74739033e-01 2.67019272e-01 -2.27613449e-01 -1.06368065e-01
-3.57437320e-02 9.74429190e-01 -6.21002793e-01 -2.82035619e-01
-4.46677655e-01 9.44554150e-01 5.32448947e-01 3.74671310e-01
-5.88400602e-01 -3.94310117e-01 3.59357297e-01 2.78142512e-01
3.40879172e-01 -4.79950190e-01 -5.71015596e-01 -1.19599092e+00
3.95444751e-01 -7.58894145e-01 3.20855290e-01 -1.16120562e-01
-1.18383551e+00 5.80520928e-01 -3.31225812e-01 -6.73217893e-01
-2.77410090e-01 -7.00329363e-01 -5.00773013e-01 5.68226993e-01
-1.47211730e+00 -4.51197445e-01 3.71006012e-01 3.71934325e-01
2.62279838e-01 1.11032814e-01 5.57395279e-01 2.86934882e-01
-2.91721195e-01 7.30317771e-01 5.18911481e-01 8.57950673e-02
4.16380078e-01 -1.28034472e+00 2.68931299e-01 7.07554042e-01
5.25029540e-01 6.97737813e-01 1.12722218e+00 -3.62467229e-01
-1.13251889e+00 -4.13949996e-01 1.07341743e+00 -2.42978156e-01
8.41597676e-01 -1.35266468e-01 -9.90289807e-01 4.23563272e-01
-1.67284220e-01 1.08238243e-01 2.87197828e-01 3.43279809e-01
-2.25641817e-01 -2.98908681e-01 -8.46118510e-01 5.23824155e-01
7.87417531e-01 -4.01639849e-01 -4.86749917e-01 5.63995659e-01
4.63572860e-01 -3.53457540e-01 -7.28844643e-01 3.56093228e-01
6.02318585e-01 -1.11699808e+00 6.96849763e-01 -5.80458760e-01
4.29608859e-02 1.45802265e-02 -5.09278290e-02 -1.11908400e+00
-5.67214936e-02 -6.73458993e-01 -2.08310828e-01 8.18735182e-01
4.51088130e-01 -8.70835483e-01 9.14045811e-01 5.01777112e-01
-6.36486113e-02 -1.36957920e+00 -8.47169220e-01 -1.03786373e+00
4.70894486e-01 -4.15409923e-01 1.02955736e-01 8.79972100e-01
1.04915336e-01 2.12614894e-01 -2.36831278e-01 -9.11427289e-02
6.07685685e-01 -2.28360780e-02 4.18633163e-01 -1.04187620e+00
-6.22350097e-01 -6.18380725e-01 -2.17180312e-01 -1.29596317e+00
-6.50639832e-02 -6.07459962e-01 -7.70309046e-02 -1.02481437e+00
-9.03918669e-02 -6.36707187e-01 -3.52913469e-01 2.60576010e-01
-2.03816056e-01 1.07376568e-01 1.19331223e-03 1.24471560e-01
-4.63328272e-01 4.66798306e-01 1.16337228e+00 2.28571311e-01
-4.20452386e-01 1.88358903e-01 -5.94240725e-01 9.95388389e-01
9.75972474e-01 -5.64651549e-01 -3.30391139e-01 -4.51209247e-01
6.26493573e-01 -2.42344946e-01 2.92075962e-01 -8.72555673e-01
3.01832289e-01 -1.43313259e-02 2.42623240e-01 -2.55210310e-01
2.46375814e-01 -3.72855037e-01 -4.14709836e-01 5.76056838e-01
-4.19362843e-01 2.50535637e-01 4.74791527e-02 2.01137826e-01
-1.48922145e-01 -5.37030518e-01 7.31927872e-01 -2.82633632e-01
-4.65485364e-01 1.58121333e-01 -2.86544889e-01 3.61955404e-01
3.60746086e-01 3.90412360e-02 3.06999423e-02 -4.92736131e-01
-6.27768636e-01 4.85740006e-02 4.06210631e-01 -1.18213274e-01
4.29685205e-01 -1.03904390e+00 -7.56301641e-01 4.28732224e-02
-5.69283903e-01 -2.22365469e-01 -1.05081629e-02 1.16558254e+00
-5.62966287e-01 2.22436100e-01 3.04586053e-01 -2.37757623e-01
-8.56143653e-01 4.91764754e-01 4.44857925e-01 -5.14642954e-01
-5.48672318e-01 1.13232994e+00 -2.30553895e-02 -2.14423597e-01
3.96837890e-01 -1.49447933e-01 2.92126268e-01 -4.60677929e-02
5.16597927e-01 4.32308584e-01 1.17588691e-01 -2.12315127e-01
-4.09829795e-01 5.37768722e-01 -3.48002672e-01 -4.24372524e-01
1.37189353e+00 4.19577323e-02 -1.38838813e-01 6.86824620e-01
1.40687573e+00 1.36011466e-01 -1.28492737e+00 -1.82953537e-01
6.93010539e-02 -3.52335781e-01 2.75239140e-01 -6.36328161e-01
-9.66076493e-01 9.43073809e-01 6.02034509e-01 2.08852977e-01
1.09971452e+00 -2.62863070e-01 8.06719303e-01 6.32812381e-01
2.63739437e-01 -1.31074584e+00 -3.42648732e-03 6.59277618e-01
5.61343074e-01 -1.20439339e+00 4.31715250e-01 3.83428298e-02
-4.97037590e-01 1.07627690e+00 3.14519018e-01 -5.66835225e-01
5.99157929e-01 1.80997089e-01 -7.16987401e-02 -6.03187717e-02
-7.05041409e-01 -1.48602083e-01 4.13853973e-01 8.18069652e-02
6.11399829e-01 -1.70053899e-01 -9.27679241e-01 1.62240162e-01
-5.94746649e-01 -2.74978243e-02 1.85721353e-01 7.96188712e-01
-6.21363699e-01 -1.02798474e+00 -8.48325640e-02 2.96771914e-01
-7.76464343e-01 -3.89704376e-01 -1.34534344e-01 7.96335697e-01
1.27261117e-01 6.59861743e-01 -1.73000604e-01 -3.00351512e-02
1.29254028e-01 5.79518825e-02 6.96109295e-01 -2.31422056e-02
-4.97182518e-01 -7.87349120e-02 -2.74700634e-02 -5.41928232e-01
-5.21479286e-02 -5.83008170e-01 -1.23764455e+00 -5.12482584e-01
-4.13952500e-01 5.93327522e-01 8.64783406e-01 1.05007577e+00
-5.87180033e-02 2.47663502e-02 5.90068460e-01 -6.22089565e-01
-1.28075755e+00 -6.45071566e-01 -7.41922498e-01 3.46898526e-01
4.97057348e-01 -3.33872110e-01 -7.96562850e-01 -3.78450364e-01] | [7.621766567230225, 3.8066296577453613] |
537b36d2-cabc-40da-8546-a108d91754ed | wallpaper-texture-generation-and-style | 2106.11482 | null | https://arxiv.org/abs/2106.11482v1 | https://arxiv.org/pdf/2106.11482v1.pdf | Wallpaper Texture Generation and Style Transfer Based on Multi-label Semantics | Textures contain a wealth of image information and are widely used in various fields such as computer graphics and computer vision. With the development of machine learning, the texture synthesis and generation have been greatly improved. As a very common element in everyday life, wallpapers contain a wealth of texture information, making it difficult to annotate with a simple single label. Moreover, wallpaper designers spend significant time to create different styles of wallpaper. For this purpose, this paper proposes to describe wallpaper texture images by using multi-label semantics. Based on these labels and generative adversarial networks, we present a framework for perception driven wallpaper texture generation and style transfer. In this framework, a perceptual model is trained to recognize whether the wallpapers produced by the generator network are sufficiently realistic and have the attribute designated by given perceptual description; these multi-label semantic attributes are treated as condition variables to generate wallpaper images. The generated wallpaper images can be converted to those with well-known artist styles using CycleGAN. Finally, using the aesthetic evaluation method, the generated wallpaper images are quantitatively measured. The experimental results demonstrate that the proposed method can generate wallpaper textures conforming to human aesthetics and have artistic characteristics. | ['Junyu Dong', 'Lin Qi', 'Huiyu Zhou', 'Eric Rigall', 'Tiange Zhang', 'Xiaohan Feng', 'Ying Gao'] | 2021-06-22 | null | null | null | null | ['texture-synthesis'] | ['computer-vision'] | [ 5.32312572e-01 3.73676457e-02 1.06237076e-01 -3.14725846e-01
-1.46559924e-01 -6.64349616e-01 6.27707183e-01 -3.83722544e-01
2.66974062e-01 7.44506121e-01 4.58818004e-02 1.21156245e-01
1.05487334e-03 -1.30405223e+00 -6.22173011e-01 -7.23498464e-01
6.84839904e-01 3.26504171e-01 -1.23209238e-01 -3.14367205e-01
1.93943411e-01 3.02164018e-01 -1.57808399e+00 5.90780079e-01
5.90673566e-01 1.20035541e+00 2.36054450e-01 4.83859152e-01
-2.75981337e-01 6.06045604e-01 -7.08120286e-01 -8.07020307e-01
3.60703737e-01 -7.94775784e-01 -5.79536080e-01 5.40230691e-01
-2.21676044e-02 -1.89526543e-01 2.06238218e-02 1.42448592e+00
2.95597494e-01 2.64894664e-01 9.38190043e-01 -1.07077301e+00
-1.31709802e+00 4.76394981e-01 -4.78408307e-01 -9.40991163e-01
1.42040566e-01 9.97948125e-02 7.06825018e-01 -5.95545232e-01
8.28580320e-01 1.51085711e+00 2.13580236e-01 6.98375106e-01
-1.23655164e+00 -7.10085988e-01 -6.57421350e-02 6.52695820e-02
-1.23354948e+00 -4.66250256e-02 1.09891820e+00 -3.57595056e-01
2.61682961e-02 4.83678967e-01 7.52067149e-01 1.37580562e+00
2.39500999e-01 4.23194557e-01 1.58462286e+00 -4.72175360e-01
3.29813868e-01 4.55646396e-01 -6.82071924e-01 6.12619877e-01
9.89598557e-02 1.20801158e-01 2.79116593e-02 4.44583505e-01
1.11898458e+00 -1.12474538e-01 -8.30638856e-02 -2.59327948e-01
-1.02033651e+00 7.94201672e-01 2.47298658e-01 7.41630867e-02
-6.46293014e-02 -4.49277982e-02 3.97521257e-01 2.83028483e-01
4.01218981e-01 5.45860231e-01 1.46047860e-01 5.62730245e-02
-4.94932413e-01 1.58039927e-01 7.03438103e-01 1.19126189e+00
5.56701064e-01 2.70699620e-01 -1.56145066e-01 1.15483296e+00
3.79089236e-01 8.58844459e-01 4.86957788e-01 -9.05395389e-01
1.67257518e-01 5.22253752e-01 2.37327307e-01 -1.40116656e+00
2.58486688e-01 8.03697556e-02 -1.14647472e+00 7.94207394e-01
8.30662400e-02 2.45452508e-01 -9.50272918e-01 1.37179410e+00
7.21205547e-02 -1.29226640e-01 8.73869434e-02 1.09311020e+00
7.91382909e-01 7.75583386e-01 2.06365228e-01 2.42852345e-01
1.36678672e+00 -9.21106219e-01 -8.66966426e-01 3.97482030e-02
-5.57733439e-02 -1.36125124e+00 1.58797288e+00 6.18697047e-01
-8.42827737e-01 -1.04795778e+00 -1.25626588e+00 3.10855210e-01
-2.85201520e-01 3.15120280e-01 5.87381721e-01 9.88345981e-01
-5.79439461e-01 5.77755272e-01 -2.12486610e-01 -1.76484272e-01
3.99367362e-01 -1.96373910e-01 -1.83800250e-01 9.10852626e-02
-1.11524916e+00 6.17259264e-01 6.68718934e-01 1.75682940e-02
-6.62887454e-01 -2.59590503e-02 -5.54399073e-01 -6.53581098e-02
1.12090483e-01 -5.78196526e-01 1.04860485e+00 -1.38694930e+00
-1.92336059e+00 9.93848622e-01 4.50284779e-01 1.52966321e-01
8.19639027e-01 1.45931423e-01 -9.19868231e-01 -7.48045146e-02
1.26384467e-01 6.28365636e-01 1.12861753e+00 -1.70176244e+00
-4.48102623e-01 1.73035815e-01 3.06583708e-03 1.65864695e-02
-2.33348772e-01 -2.45311171e-01 -3.19574744e-01 -1.10529459e+00
-3.23090851e-02 -8.81421924e-01 3.09024975e-02 -4.27623279e-02
-6.67848945e-01 2.14539558e-01 8.59411716e-01 -3.51819813e-01
6.29114628e-01 -2.16372418e+00 -1.05427504e-01 5.80657423e-01
-5.30114509e-02 1.87808558e-01 -2.45947704e-01 2.70611018e-01
1.11146748e-01 3.61395538e-01 -7.83235580e-02 -3.22602838e-02
3.75670314e-01 2.18185648e-01 -5.53557992e-01 -1.59935370e-01
2.39460692e-01 8.24414968e-01 -8.73584509e-01 -3.58467311e-01
5.96159399e-01 3.94589871e-01 -3.73826057e-01 4.08881754e-01
-4.47729021e-01 7.36603379e-01 -6.81594849e-01 6.26470387e-01
7.13332653e-01 -3.16402949e-02 1.03394113e-01 -3.13878715e-01
3.11262310e-01 -6.04296207e-01 -1.10186148e+00 1.29474938e+00
-7.50628650e-01 3.85757595e-01 -5.10105789e-01 -5.59092522e-01
1.64962852e+00 1.10299051e-01 -1.12634160e-01 -7.58473516e-01
5.09717584e-01 3.03870618e-01 -2.51477718e-01 -6.08657777e-01
5.31851590e-01 -4.29793298e-01 -1.83307797e-01 4.16142553e-01
-4.59385246e-01 -9.16315377e-01 1.34641305e-01 -3.77909541e-01
5.37514210e-01 5.21533132e-01 -1.82380199e-01 1.52725339e-01
5.88363767e-01 -2.21938729e-01 4.36450481e-01 5.03425181e-01
4.27582532e-01 8.84245217e-01 3.09538752e-01 -7.40750074e-01
-1.69659972e+00 -1.21447825e+00 2.88239457e-02 5.40166557e-01
3.55818391e-01 2.30478168e-01 -1.06749761e+00 -2.97243446e-01
-2.47073397e-01 6.77949667e-01 -6.49608314e-01 -2.68157274e-01
-1.70433402e-01 -3.71366262e-01 3.98798019e-01 3.27344328e-01
1.14447069e+00 -1.66151679e+00 -2.61934221e-01 2.15686232e-01
-1.80628389e-01 -1.13171732e+00 -3.26390624e-01 -3.43930662e-01
-3.66953135e-01 -9.43386674e-01 -8.45287621e-01 -1.01912832e+00
9.35102820e-01 -3.66786346e-02 8.51212859e-01 -1.71919882e-01
-4.47200745e-01 -1.08794030e-02 -7.42512465e-01 -4.23307657e-01
-9.47997451e-01 -2.66988158e-01 -1.55371964e-01 5.53010166e-01
1.19216874e-01 -3.10982734e-01 -5.82118988e-01 4.68380034e-01
-1.26347315e+00 6.13419890e-01 6.78869843e-01 8.87202024e-01
8.47932518e-01 5.32006204e-01 5.67684889e-01 -1.25859725e+00
8.47531319e-01 -3.47318575e-02 -5.35735190e-01 3.18584621e-01
-4.46373522e-01 -1.11268483e-01 1.08021128e+00 -6.11617386e-01
-1.48173440e+00 -1.92487746e-01 5.81477396e-02 -4.40256357e-01
-5.05275428e-01 1.33747414e-01 -5.61282396e-01 -7.49345198e-02
4.41041142e-01 1.00776196e-01 -1.19037196e-01 -2.24538967e-01
6.02242887e-01 6.77527368e-01 5.53773284e-01 -8.31192315e-01
8.61074328e-01 4.78521109e-01 -2.21508667e-02 -5.43711066e-01
-7.07938373e-01 4.29086715e-01 -1.48173958e-01 -5.71165979e-01
1.14057088e+00 -3.67881477e-01 -6.88932836e-01 8.66553366e-01
-1.19916821e+00 -3.70129675e-01 -5.15045583e-01 2.57297903e-01
-8.37761521e-01 2.57798702e-01 -4.44003880e-01 -6.88219905e-01
-1.07853942e-01 -1.24309015e+00 9.75065768e-01 2.71777391e-01
-3.48421186e-01 -9.09953415e-01 -1.54287919e-01 3.91696602e-01
2.39618123e-01 6.77973568e-01 1.36297333e+00 1.38252839e-01
-4.14014369e-01 -2.91204304e-01 -5.65921068e-01 6.85627103e-01
5.01926839e-01 1.12384530e-02 -9.84464169e-01 7.89950565e-02
1.47946924e-02 -3.11093420e-01 4.74732995e-01 2.15058774e-02
1.56153750e+00 -3.76796275e-01 1.83990210e-01 7.13109136e-01
1.48994696e+00 7.03640223e-01 8.29259098e-01 4.46168721e-01
7.72746980e-01 8.03822398e-01 5.90171218e-01 3.91408980e-01
-7.88599551e-02 3.90724629e-01 3.05944711e-01 -2.67157495e-01
-3.77714574e-01 -6.74979925e-01 -7.89932236e-02 8.50858510e-01
-4.79885697e-01 -6.14638031e-01 -4.33214456e-01 1.88546386e-02
-1.65940142e+00 -1.05320704e+00 1.19769804e-01 1.92008102e+00
7.66613722e-01 2.01616064e-01 -4.23898339e-01 3.97092789e-01
1.13005948e+00 7.72886351e-02 -3.49625975e-01 -8.26269448e-01
-3.86598051e-01 5.07226646e-01 3.44384253e-01 1.40657738e-01
-1.06055152e+00 1.17201054e+00 5.43061304e+00 1.08635664e+00
-1.11541128e+00 -1.85002103e-01 8.86161268e-01 4.05510902e-01
-6.22669399e-01 -1.25204876e-01 -2.02875763e-01 7.02741265e-01
7.74311274e-02 -2.48812214e-02 7.24696219e-01 8.01507890e-01
2.81186014e-01 1.51194647e-01 -7.59452462e-01 1.05575716e+00
1.80971194e-02 -1.20688260e+00 4.94346350e-01 -5.90902083e-02
1.37152648e+00 -1.09560108e+00 5.69730580e-01 -1.37776092e-01
5.64192235e-01 -1.29946911e+00 9.39777434e-01 7.82927036e-01
1.28682327e+00 -8.78539920e-01 8.17589998e-01 -2.82870024e-01
-1.06860173e+00 2.05871701e-01 -5.44450700e-01 4.23132665e-02
5.30925542e-02 5.73026597e-01 -5.58321357e-01 4.61779326e-01
4.80764508e-01 4.04117316e-01 -1.94116727e-01 6.72618926e-01
-6.87550306e-01 4.41988587e-01 2.22174153e-01 -4.28261518e-01
1.47750869e-01 -6.51207566e-01 2.18942687e-01 7.73701847e-01
6.28055394e-01 3.94588932e-02 1.64995134e-01 1.21716607e+00
-1.90617606e-01 4.04354244e-01 -5.73015094e-01 -2.97883987e-01
3.66455168e-01 1.13694596e+00 -1.09284770e+00 -3.28992277e-01
-8.56473371e-02 1.18121088e+00 -3.43225032e-01 4.09481287e-01
-9.00727987e-01 -7.61759996e-01 9.61749628e-02 7.82167092e-02
-1.48415761e-02 9.77240652e-02 -5.92300773e-01 -7.86715984e-01
7.67457634e-02 -8.61412287e-01 -4.27730590e-01 -1.20971787e+00
-1.56092215e+00 8.01428676e-01 -4.16622937e-01 -1.54065359e+00
1.26581505e-01 -7.05597401e-01 -6.93117142e-01 8.79657209e-01
-1.07427084e+00 -1.54214346e+00 -7.43202567e-01 3.36371034e-01
4.95935947e-01 -6.24650538e-01 1.05679226e+00 9.78285447e-02
-9.55499709e-02 3.05050939e-01 1.50516957e-01 1.98013410e-01
6.29014134e-01 -9.83859062e-01 4.27572012e-01 4.20241237e-01
-1.16225677e-02 2.33256102e-01 6.37123406e-01 -7.23848760e-01
-9.13712859e-01 -1.34099889e+00 4.77455914e-01 1.02092527e-01
4.50338811e-01 -2.43334129e-01 -6.30383611e-01 7.74035305e-02
1.28854215e-01 -4.02003258e-01 5.93735516e-01 -4.58065271e-01
-3.10216486e-01 -2.42521092e-02 -1.26099920e+00 8.97635758e-01
9.38640475e-01 -3.15785468e-01 -4.26443130e-01 2.07359910e-01
3.38192403e-01 -2.99041837e-01 -7.43130028e-01 2.57998288e-01
6.55446887e-01 -8.80948067e-01 7.11462796e-01 -3.29800725e-01
1.04300976e+00 -3.07568789e-01 -1.09605797e-01 -1.37623167e+00
-3.31222236e-01 -4.65466917e-01 7.42769778e-01 1.47026086e+00
5.70848845e-02 -5.87483346e-01 8.04382682e-01 3.21005851e-01
-5.07365838e-02 -2.07723022e-01 -4.10870075e-01 -8.82223666e-01
-1.59062147e-01 -1.69370145e-01 1.21913040e+00 1.06587493e+00
-4.79342550e-01 -3.18685994e-02 -5.87814212e-01 -2.75834292e-01
6.26761913e-01 3.69190365e-01 7.87928224e-01 -1.16330576e+00
-3.22283655e-01 -7.40287542e-01 -5.95951974e-01 -4.32031244e-01
1.52735531e-01 -1.00135469e+00 3.61566991e-02 -1.47825098e+00
2.52105091e-02 -8.75019252e-01 -2.18451023e-01 2.98452288e-01
1.35032564e-01 7.50774145e-01 2.40257546e-01 -2.04673894e-02
-1.06785640e-01 6.31211817e-01 1.92340147e+00 -5.30767262e-01
-1.97734889e-02 -1.41917422e-01 -5.75866461e-01 8.54290545e-01
1.10501885e+00 -1.15537599e-01 -6.00859344e-01 -2.76334286e-01
2.96892047e-01 -8.63895342e-02 4.53820407e-01 -1.01542866e+00
-5.78791082e-01 -4.36056226e-01 6.76575303e-01 7.71942213e-02
2.78605789e-01 -8.50184739e-01 5.71888566e-01 2.17384860e-01
-3.10214579e-01 -3.54441911e-01 -7.63109103e-02 3.42590690e-01
-3.54689389e-01 -3.34042370e-01 9.53585088e-01 -1.46573439e-01
-9.24258292e-01 1.74817309e-01 -2.24436507e-01 -2.62175873e-02
1.13057804e+00 -4.14735585e-01 -1.57430008e-01 -4.49237347e-01
-7.42553830e-01 -4.03890491e-01 7.57829309e-01 7.18155324e-01
8.43882322e-01 -1.84117937e+00 -5.97615957e-01 4.71908391e-01
1.64215818e-01 -1.42016545e-01 4.56236929e-01 -2.80937791e-01
-9.63961720e-01 -2.37816796e-02 -8.57771099e-01 -2.20731273e-01
-8.59989703e-01 4.38188612e-01 -1.72186896e-01 -1.66439246e-02
-4.76711184e-01 5.67595959e-01 5.50480008e-01 -2.76803911e-01
-2.47488886e-01 -1.49215935e-02 -1.86640397e-01 -2.97365576e-01
1.96784943e-01 2.64166325e-01 -1.24935612e-01 -4.57593918e-01
3.69321704e-01 9.46403563e-01 2.51849324e-01 -1.85165182e-01
9.63754773e-01 1.07192181e-01 -9.65215415e-02 2.17964634e-01
9.03236032e-01 1.20626919e-01 -1.18827248e+00 2.20857695e-01
-4.35214847e-01 -7.91004658e-01 -2.25840226e-01 -8.37079227e-01
-1.25077963e+00 9.28355992e-01 6.89923346e-01 2.31605679e-01
1.15144515e+00 -3.62325102e-01 8.86170149e-01 1.56236723e-01
7.09055126e-01 -1.52509212e+00 3.75216752e-01 4.50313576e-02
9.96971786e-01 -1.05871999e+00 -3.46752018e-01 -5.62276363e-01
-9.37001228e-01 1.12298810e+00 5.40520608e-01 -3.11649680e-01
1.53123960e-01 2.25062415e-01 4.45647389e-01 1.12914927e-01
-5.15035316e-02 7.50320107e-02 3.37296277e-01 9.17363763e-01
2.92475313e-01 5.22973239e-01 -2.09451318e-01 7.96358645e-01
-6.05475485e-01 -3.87553312e-02 3.94852698e-01 4.60046202e-01
-3.98006558e-01 -1.50288987e+00 -4.86165494e-01 3.28758597e-01
-2.25633964e-01 1.60218805e-01 -3.49380791e-01 4.84695107e-01
4.92245942e-01 9.77328718e-01 1.92795340e-02 -6.82411969e-01
3.47468346e-01 -1.68699160e-01 5.94685853e-01 -4.83543843e-01
-1.73316494e-01 -3.05915647e-03 -2.65928600e-02 -1.45492271e-01
-3.62009883e-01 -8.78026262e-02 -1.01912689e+00 -4.32644546e-01
-4.65445183e-02 -7.51957623e-03 6.63755178e-01 5.99912167e-01
-1.24630906e-01 8.58801365e-01 9.25723195e-01 -6.35319650e-01
-2.36462355e-01 -6.13347709e-01 -7.79123664e-01 1.02041745e+00
-5.12437224e-01 -6.87533975e-01 1.59313437e-02 5.32702327e-01] | [11.698387145996094, -0.4594956934452057] |
c0946a4d-0471-4de3-8b1c-54dba6e3cf7e | neural-language-taskonomy-which-nlp-tasks-are | 2205.01404 | null | https://arxiv.org/abs/2205.01404v1 | https://arxiv.org/pdf/2205.01404v1.pdf | Neural Language Taskonomy: Which NLP Tasks are the most Predictive of fMRI Brain Activity? | Several popular Transformer based language models have been found to be successful for text-driven brain encoding. However, existing literature leverages only pretrained text Transformer models and has not explored the efficacy of task-specific learned Transformer representations. In this work, we explore transfer learning from representations learned for ten popular natural language processing tasks (two syntactic and eight semantic) for predicting brain responses from two diverse datasets: Pereira (subjects reading sentences from paragraphs) and Narratives (subjects listening to the spoken stories). Encoding models based on task features are used to predict activity in different regions across the whole brain. Features from coreference resolution, NER, and shallow syntax parsing explain greater variance for the reading activity. On the other hand, for the listening activity, tasks such as paraphrase generation, summarization, and natural language inference show better encoding performance. Experiments across all 10 task representations provide the following cognitive insights: (i) language left hemisphere has higher predictive brain activity versus language right hemisphere, (ii) posterior medial cortex, temporo-parieto-occipital junction, dorsal frontal lobe have higher correlation versus early auditory and auditory association cortex, (iii) syntactic and semantic tasks display a good predictive performance across brain regions for reading and listening stimuli resp. | ['Bapi Raju Surampudi', 'Manish Gupta', 'Mounika Marreddy', 'Veeral Agarwal', 'Jashn Arora', 'Subba Reddy Oota'] | 2022-05-03 | null | https://aclanthology.org/2022.naacl-main.235 | https://aclanthology.org/2022.naacl-main.235.pdf | naacl-2022-7 | ['paraphrase-generation', 'paraphrase-generation'] | ['computer-code', 'natural-language-processing'] | [ 4.13204253e-01 4.81934875e-01 -6.85245469e-02 -5.07308364e-01
-6.17176771e-01 -2.14626729e-01 1.26119649e+00 3.38223100e-01
-3.92555326e-01 5.60944676e-01 1.35965550e+00 1.09067475e-02
-3.19471389e-01 -7.95308769e-01 -4.24943358e-01 -3.83472383e-01
3.77065279e-02 3.81615520e-01 3.18722688e-02 -2.42938161e-01
7.64161825e-01 1.21402644e-01 -1.07286167e+00 1.00857222e+00
6.46612167e-01 9.99945939e-01 4.10921812e-01 4.11774337e-01
-1.46089926e-01 1.28244877e+00 -4.92719471e-01 -2.98994780e-01
-3.98567051e-01 -1.00265980e+00 -1.24457371e+00 -4.98533547e-01
2.08890557e-01 -2.61566285e-02 -3.89671832e-01 8.36026669e-01
7.60447860e-01 2.69336432e-01 1.04756987e+00 -7.11434126e-01
-9.08358693e-01 1.15681827e+00 -3.32902282e-01 9.95596945e-01
6.21137857e-01 4.62923851e-03 1.21426642e+00 -1.03265309e+00
7.20099330e-01 1.43836725e+00 4.13637012e-01 4.50764507e-01
-1.19448876e+00 -7.10471928e-01 7.50900432e-02 3.08622897e-01
-9.34630275e-01 -7.04007566e-01 5.32853961e-01 -4.84889179e-01
1.73046148e+00 -9.45502073e-02 6.02815151e-01 1.82919395e+00
1.05032992e+00 6.84102893e-01 1.50749242e+00 -4.46655080e-02
1.10561281e-01 1.06527001e-01 6.11782610e-01 4.92796808e-01
-7.24412724e-02 1.73241898e-01 -1.21825099e+00 -1.60917628e-03
3.46412718e-01 -2.38446668e-01 -4.07146871e-01 3.60079139e-01
-1.65204322e+00 1.13532865e+00 4.30904299e-01 6.62020266e-01
-7.81634688e-01 -6.53500408e-02 8.69284868e-01 4.62446719e-01
3.77057433e-01 5.83514690e-01 -4.89412516e-01 -6.65865019e-02
-1.07950759e+00 -6.27561212e-02 6.39310837e-01 4.46686894e-01
3.83418471e-01 3.58194351e-01 -7.64827788e-01 1.04747736e+00
3.90886307e-01 3.45609009e-01 1.28015137e+00 -4.08043623e-01
8.33988130e-01 2.09271774e-01 -8.07203114e-01 -1.17482460e+00
-8.50648940e-01 -3.82038921e-01 -1.00243866e+00 -4.82912332e-01
-3.04046478e-02 1.12729609e-01 -4.80010182e-01 1.80894661e+00
-5.72861493e-01 -4.55738991e-01 2.40529016e-01 7.55612254e-01
1.24163008e+00 7.07408011e-01 5.52467227e-01 -3.50081682e-01
1.92230034e+00 -6.59423947e-01 -6.69072628e-01 -9.39260900e-01
3.02910715e-01 -4.00221169e-01 7.62217522e-01 2.28970334e-01
-1.31876087e+00 -6.79443896e-01 -9.17818248e-01 -2.76168406e-01
-4.27824944e-01 -1.12324975e-01 4.66576368e-01 1.98464438e-01
-1.10928392e+00 2.10456550e-01 -3.90981346e-01 -8.05376887e-01
5.60119212e-01 8.10494721e-02 -4.96540636e-01 1.38347611e-01
-1.39379418e+00 1.40001655e+00 6.87569678e-01 -3.06993574e-01
-1.24502504e+00 -7.47667491e-01 -8.55957747e-01 4.36195403e-01
-1.26626909e-01 -9.02900577e-01 9.29040134e-01 -1.01950622e+00
-1.52024686e+00 1.27322459e+00 -2.03216463e-01 -6.72516823e-01
-3.86079937e-01 -5.08719198e-02 -3.40695471e-01 3.61715436e-01
2.81340837e-01 8.36458325e-01 6.20708227e-01 -3.80446941e-01
8.95043164e-02 -6.41941428e-01 -3.25157255e-01 3.78393888e-01
-2.70510852e-01 4.19001073e-01 6.56192958e-01 -1.09752381e+00
2.47055754e-01 -2.93956071e-01 4.46437687e-01 -4.27068233e-01
-2.73493767e-01 -5.70734382e-01 1.89005390e-01 -1.10752392e+00
8.16429377e-01 -2.08357430e+00 3.03876072e-01 -1.55882180e-01
1.42149001e-01 -4.59301651e-01 -2.63634026e-01 3.90432060e-01
-4.59516853e-01 3.18015397e-01 -9.01341885e-02 1.66888297e-01
2.14902297e-01 -1.18231639e-01 -5.56073785e-01 4.49386418e-01
2.93354124e-01 1.10278690e+00 -7.85939455e-01 -4.76430148e-01
-2.62108982e-01 1.62847966e-01 -6.54395700e-01 1.24161489e-01
-3.00287399e-02 3.44393790e-01 -3.31554830e-01 2.50197262e-01
1.37556002e-01 3.43919806e-02 -4.61411290e-02 -3.39788973e-01
5.85808381e-02 1.18795323e+00 -2.94614822e-01 1.82129860e+00
-3.81270736e-01 1.01594770e+00 -4.02252227e-02 -1.41747415e+00
1.05565643e+00 5.85835338e-01 -5.04246764e-02 -1.55610907e+00
5.71007371e-01 -3.24210107e-01 5.81409693e-01 -5.37728608e-01
-4.29563783e-02 -4.89910483e-01 -2.41108105e-01 9.46442962e-01
7.52936363e-01 1.28098913e-02 -2.01336354e-01 3.08131725e-01
1.04847229e+00 -1.87738389e-01 7.10176289e-01 -5.82874537e-01
4.81372774e-01 -4.21239495e-01 2.71934569e-01 5.73958814e-01
-2.66534507e-01 3.78877580e-01 6.84647799e-01 -1.46795288e-01
-4.66761440e-01 -1.31119883e+00 -6.58530295e-02 1.77408791e+00
-5.56090772e-01 -3.48995507e-01 -7.71744549e-01 -8.95909965e-03
-5.91738760e-01 1.22094381e+00 -8.51678014e-01 -7.80239165e-01
-4.17818248e-01 -7.62293160e-01 7.81349242e-01 8.96520853e-01
8.87101710e-01 -1.58745360e+00 -8.02192509e-01 -2.49137413e-02
-5.60596585e-01 -1.11716151e+00 -3.31525385e-01 3.77755821e-01
-9.70319748e-01 -7.54503727e-01 -3.36532235e-01 -9.63217914e-01
1.77093297e-01 -2.38343075e-01 1.14784348e+00 -4.65345591e-01
-3.37692261e-01 4.92565095e-01 -1.99870139e-01 -3.30627918e-01
-2.25371331e-01 1.30664244e-01 -1.36090532e-01 -2.76951164e-01
3.37411255e-01 -5.68866372e-01 -4.10007268e-01 3.52707207e-02
-4.56304729e-01 1.70380130e-01 7.99520075e-01 9.75667000e-01
2.44041309e-01 -5.43831706e-01 9.13545012e-01 -5.69879055e-01
1.25334871e+00 -1.02574813e+00 3.99634093e-01 9.76199880e-02
-4.58413541e-01 2.58205026e-01 4.75760669e-01 -3.70241016e-01
-1.38718712e+00 -7.40878344e-01 5.66646317e-03 3.41298342e-01
-3.68197381e-01 7.51184702e-01 7.90626109e-02 5.02148807e-01
1.00918984e+00 6.75205767e-01 -8.62420276e-02 -7.29458183e-02
1.47683948e-01 3.07454199e-01 4.95751709e-01 -7.93533146e-01
1.15871668e-01 -9.29695964e-02 -3.39531183e-01 -9.12952423e-01
-8.20061326e-01 1.00315601e-01 -5.85798025e-01 1.24545209e-02
1.55569327e+00 -8.53891671e-01 -3.76396447e-01 1.17453486e-02
-1.26262844e+00 -3.26528817e-01 -8.03157911e-02 6.47063851e-01
-7.50995159e-01 2.57336292e-02 -7.23330796e-01 -4.44583476e-01
-7.84395754e-01 -7.15972424e-01 9.44929361e-01 -2.12904741e-03
-8.37670982e-01 -1.05003524e+00 4.89569232e-02 3.42316002e-01
7.49533772e-01 6.74917474e-02 1.49125338e+00 -1.18138385e+00
1.68648824e-01 3.79318655e-01 -2.82505155e-01 -1.80882186e-01
-2.31037199e-01 -5.53930402e-01 -1.14289522e+00 1.20959625e-01
3.34203809e-01 -6.47331417e-01 1.08188736e+00 4.42781001e-01
1.12457645e+00 -3.95256996e-01 -2.42828518e-01 2.52551228e-01
9.10891652e-01 -6.62467554e-02 6.93445444e-01 -6.69298023e-02
1.37081603e-02 9.49936092e-01 6.97417278e-03 1.83683738e-01
3.45103145e-01 2.63178796e-01 -4.82381918e-02 5.45320034e-01
-2.13057980e-01 -2.99134642e-01 1.13261569e+00 8.42125893e-01
1.13598250e-01 -1.47690818e-01 -1.03048277e+00 5.76111794e-01
-1.32006490e+00 -1.34990597e+00 1.62900418e-01 1.79106736e+00
7.94411898e-01 1.92948300e-02 -1.96820438e-01 -6.97200224e-02
2.75744647e-01 4.28053081e-01 -4.07901973e-01 -9.31891263e-01
-1.96444705e-01 5.84930420e-01 -1.20956033e-01 -8.27660505e-03
-6.69704258e-01 9.53257263e-01 6.13681984e+00 4.50962752e-01
-9.20244515e-01 6.03841782e-01 6.51511133e-01 -2.10547656e-01
-2.20607877e-01 -3.61177802e-01 -3.44906390e-01 4.52299751e-02
1.63814247e+00 -4.06463325e-01 5.84383905e-01 3.42353314e-01
2.05450401e-01 -1.60149932e-01 -1.40097475e+00 9.92308497e-01
6.25144839e-01 -1.14329088e+00 3.14201057e-01 -1.49937123e-01
2.40121797e-01 2.70173639e-01 -3.40608023e-02 7.17534661e-01
-3.33042219e-02 -1.42672396e+00 7.90068984e-01 8.79055500e-01
5.53888261e-01 -3.15246999e-01 5.24448216e-01 5.87461710e-01
-9.51456368e-01 -1.93741679e-01 -5.48852801e-01 -8.85516256e-02
9.90197957e-02 1.19117141e-01 -6.06774330e-01 4.07863706e-02
7.56821811e-01 7.63046086e-01 -6.64781034e-01 3.77132773e-01
-4.22757447e-01 9.67996180e-01 2.50384241e-01 -1.50221556e-01
1.11996762e-01 3.45889747e-01 4.16888744e-01 1.62269723e+00
2.91867614e-01 4.94570225e-01 -1.68591991e-01 1.17619741e+00
-2.13877738e-01 6.19664669e-01 -7.72165716e-01 -2.75218576e-01
3.42252314e-01 1.06906664e+00 -7.26934195e-01 -1.20651990e-01
-3.43111843e-01 8.14968050e-01 3.95119607e-01 3.56294423e-01
-7.15170562e-01 9.58699360e-02 2.61686176e-01 -2.48069107e-03
-7.31516257e-03 -1.67121887e-01 -4.44458187e-01 -8.64127100e-01
-4.93614584e-01 -6.72391474e-01 5.19221008e-01 -1.33217847e+00
-1.35876536e+00 6.16610229e-01 1.62639424e-01 -3.43797296e-01
-3.69517207e-01 -7.58791983e-01 -9.12060857e-01 1.06399751e+00
-1.13364613e+00 -6.68618858e-01 -1.79370254e-01 8.64652932e-01
1.09883511e+00 -6.07574940e-01 1.27836251e+00 -9.27038044e-02
-3.45970273e-01 4.39184785e-01 -6.95479751e-01 2.76544750e-01
7.45670259e-01 -8.19347024e-01 -5.32435328e-02 4.70138639e-01
1.30260393e-01 8.35210562e-01 1.13125868e-01 -4.60615486e-01
-1.18636835e+00 -9.27047431e-01 1.04319596e+00 -2.02975050e-01
6.10174775e-01 -3.31004739e-01 -8.07903886e-01 7.72462130e-01
1.01293063e+00 -5.53616881e-01 9.81605172e-01 7.16473833e-02
-5.11366069e-01 2.01722324e-01 -1.16243231e+00 3.61570597e-01
9.88338709e-01 -1.00752330e+00 -1.35747337e+00 4.94886667e-01
6.01323247e-01 2.41210982e-01 -1.06708491e+00 -1.78204119e-01
2.57749379e-01 -6.78226531e-01 9.58455861e-01 -8.87645960e-01
8.43772590e-01 3.53076518e-01 -2.96411991e-01 -1.43746364e+00
-9.51077759e-01 9.46929399e-03 5.14999442e-02 1.30231178e+00
4.50127840e-01 -8.56584013e-01 -1.55672505e-01 1.78124115e-01
-2.04051927e-01 -2.88368821e-01 -9.35901165e-01 -2.90501863e-01
4.06504720e-01 -1.79735750e-01 -1.91686936e-02 1.06437123e+00
4.12668109e-01 1.19078743e+00 1.26816630e-01 -2.51814961e-01
3.11755408e-02 -2.66256660e-01 6.11915626e-03 -1.38526535e+00
-8.22816268e-02 -7.87451863e-01 -1.29883513e-01 -5.75489283e-01
8.18429828e-01 -1.72071624e+00 -8.66195261e-02 -1.84769964e+00
8.40578973e-01 5.76084197e-01 -2.34117359e-01 7.11241841e-01
1.47731915e-01 -2.44155586e-01 -1.63812395e-02 -9.04459730e-02
-1.75811946e-01 6.27042115e-01 9.76438105e-01 -2.71666318e-01
4.48656410e-01 -4.68229800e-01 -1.24926829e+00 7.61559367e-01
1.01268744e+00 -4.01878387e-01 -5.96963644e-01 -5.19362271e-01
4.59206760e-01 2.85091311e-01 4.16719168e-01 -9.74768341e-01
1.79771125e-01 9.41190869e-02 7.74887621e-01 -1.15163796e-01
1.25070900e-01 -2.08392307e-01 -4.60906386e-01 3.83491337e-01
-8.73301804e-01 4.12933111e-01 3.43546718e-01 4.10488993e-01
-3.05983186e-01 -3.05766732e-01 1.00865042e+00 -3.94232452e-01
-6.33516908e-01 -6.36757165e-02 -1.04961729e+00 5.94785511e-01
5.02724826e-01 -2.38413140e-01 -6.23693347e-01 -4.40415710e-01
-8.20652246e-01 -1.88279510e-01 -5.01223326e-01 8.01574707e-01
8.48768175e-01 -1.13872147e+00 -1.14955103e+00 1.00894913e-01
-5.95283099e-02 -7.05019176e-01 2.13353503e-02 1.31689382e+00
2.27927584e-02 7.17328429e-01 -6.08630657e-01 -3.18696231e-01
-8.39887559e-01 4.40723121e-01 4.50594962e-01 -2.39466906e-01
-5.71561396e-01 9.18127775e-01 5.91553986e-01 -2.01216787e-01
-3.56208682e-02 -5.79544961e-01 -6.69829488e-01 7.14522898e-01
7.26207733e-01 2.40457848e-01 3.70637588e-02 -8.04037631e-01
-4.87221152e-01 1.97696924e-01 -1.80852473e-01 -2.37951323e-01
1.52564180e+00 2.21803129e-01 -4.49154645e-01 7.07628250e-01
1.10696077e+00 -3.50366235e-01 -2.95414865e-01 -2.49891385e-01
1.37148485e-01 3.80056590e-01 2.60806799e-01 -1.04775178e+00
-9.91712868e-01 1.31708944e+00 5.14612615e-01 -1.63613647e-01
1.04451025e+00 2.29881480e-01 4.66839969e-01 2.78206766e-01
3.34865332e-01 -1.32927656e+00 3.95533353e-01 8.73174727e-01
1.42271662e+00 -8.19351614e-01 -1.47442535e-01 9.90856662e-02
-9.86667931e-01 1.24166775e+00 6.13404393e-01 -2.34862864e-01
6.08589113e-01 4.06714119e-02 -7.88751781e-01 -5.72674632e-01
-1.14299452e+00 2.63679832e-01 3.67621839e-01 4.18350726e-01
1.09815884e+00 5.47414683e-02 -5.59491396e-01 1.32966387e+00
-5.40945590e-01 -4.87888217e-01 2.19508916e-01 5.01984835e-01
-6.56827211e-01 -3.40673566e-01 -3.19733530e-01 8.86894464e-01
-4.60271925e-01 -5.72392225e-01 -7.03694999e-01 4.94109303e-01
-1.23418413e-01 8.04895341e-01 2.34494060e-01 4.45179492e-02
2.00492978e-01 8.41915369e-01 5.13767838e-01 -9.80416179e-01
-1.09357584e+00 -2.68422425e-01 3.38312387e-01 -7.47552037e-01
-4.28746521e-01 -9.79543447e-01 -1.57794666e+00 -3.01759457e-03
2.63551801e-01 -6.30084351e-02 2.50464827e-01 1.26160645e+00
3.00893337e-01 8.52844954e-01 1.65138133e-02 -6.74465001e-01
-3.69838119e-01 -1.62886834e+00 -3.43492806e-01 2.24545240e-01
-1.03371263e-01 -8.25824142e-01 -1.67383745e-01 1.78448260e-01] | [10.394036293029785, 8.420644760131836] |
2e6c3d81-9911-4105-8f1b-9151c0f85d87 | object-propagation-via-inter-frame-attentions | 2111.07529 | null | https://arxiv.org/abs/2111.07529v3 | https://arxiv.org/pdf/2111.07529v3.pdf | Object Propagation via Inter-Frame Attentions for Temporally Stable Video Instance Segmentation | Video instance segmentation aims to detect, segment, and track objects in a video. Current approaches extend image-level segmentation algorithms to the temporal domain. However, this results in temporally inconsistent masks. In this work, we identify the mask quality due to temporal stability as a performance bottleneck. Motivated by this, we propose a video instance segmentation method that alleviates the problem due to missing detections. Since this cannot be solved simply using spatial information, we leverage temporal context using inter-frame attentions. This allows our network to refocus on missing objects using box predictions from the neighbouring frame, thereby overcoming missing detections. Our method significantly outperforms previous state-of-the-art algorithms using the Mask R-CNN backbone, by achieving 36.0% mAP on the YouTube-VIS benchmark. Additionally, our method is completely online and requires no future frames. Our code is publicly available at https://github.com/anirudh-chakravarthy/ObjProp. | ['Hanspeter Pfister', 'Song Bai', 'Donglai Wei', 'Zudi Lin', 'Won-Dong Jang', 'Anirudh S Chakravarthy'] | 2021-11-15 | null | null | null | null | ['video-instance-segmentation'] | ['computer-vision'] | [ 1.38089284e-01 6.54044747e-02 -3.95739496e-01 -1.35959119e-01
-6.26646042e-01 -6.38100863e-01 1.92967132e-01 -3.28089714e-01
-4.15963292e-01 6.05379045e-01 -2.41395133e-03 -1.09366290e-01
3.31118822e-01 -4.18779820e-01 -8.29695344e-01 -4.98118848e-01
1.33880787e-02 5.51729240e-02 9.07039106e-01 1.27207324e-01
2.23739564e-01 2.81610578e-01 -1.39446473e+00 3.44749480e-01
6.70406461e-01 9.73846257e-01 4.67411786e-01 5.59669912e-01
-5.95047027e-02 9.02450919e-01 -4.06043291e-01 -1.61580235e-01
5.80572605e-01 -3.62699986e-01 -8.44837487e-01 3.99044633e-01
6.34553075e-01 -7.82774031e-01 -5.32248557e-01 9.14788842e-01
1.87992170e-01 2.70827919e-01 6.97808415e-02 -1.24710095e+00
-2.06868470e-01 6.62054420e-01 -8.58353198e-01 5.06401360e-01
1.59343511e-01 2.70741254e-01 8.55106473e-01 -9.25909519e-01
7.88972020e-01 9.49826837e-01 5.33644557e-01 5.84165752e-01
-1.27179074e+00 -7.26301074e-01 6.75760984e-01 2.43608296e-01
-1.43902302e+00 -6.27925158e-01 6.32607162e-01 -4.60359722e-01
9.27763224e-01 1.43636152e-01 8.08329940e-01 9.44144487e-01
-1.05275251e-01 1.07914841e+00 8.50443602e-01 -8.34316239e-02
1.20675281e-01 -3.42821717e-01 9.03703123e-02 7.23749280e-01
-6.14212081e-02 6.58334345e-02 -4.66539443e-01 4.25193608e-01
1.03364825e+00 1.78258419e-01 -3.64901781e-01 -3.69067401e-01
-1.19684517e+00 3.76863092e-01 4.37420279e-01 2.23836124e-01
-4.12258089e-01 4.74812597e-01 1.12628579e-01 4.21367884e-02
6.32906199e-01 2.15219691e-01 -6.63816690e-01 -3.29917371e-01
-1.42568839e+00 9.64162201e-02 3.18209171e-01 1.03846502e+00
6.68260694e-01 1.11257926e-01 -2.62934715e-01 5.35617590e-01
1.28422737e-01 1.05020761e-01 7.15382025e-02 -1.53020024e+00
2.44514614e-01 4.20759737e-01 2.21218273e-01 -6.88113391e-01
-3.89390707e-01 -3.13596040e-01 -3.07732046e-01 2.51798153e-01
5.93499839e-01 -2.30414465e-01 -1.26969445e+00 1.55096722e+00
5.22542536e-01 9.05402660e-01 -4.13847953e-01 1.15618598e+00
7.76250720e-01 5.69375813e-01 5.50500005e-02 -2.59455442e-01
9.92735386e-01 -1.45452988e+00 -6.89278722e-01 -2.94987768e-01
4.63845730e-01 -6.10644341e-01 7.75992930e-01 5.34603953e-01
-1.33365643e+00 -5.79423249e-01 -8.24542582e-01 -9.72448140e-02
-5.95227070e-03 4.26204354e-02 4.60583359e-01 3.77279073e-01
-1.16729844e+00 5.83481252e-01 -1.27405155e+00 -1.84239894e-01
6.48859382e-01 4.43732500e-01 -1.79415345e-01 -7.75557756e-02
-7.32407689e-01 4.50589955e-01 5.00740767e-01 7.08366036e-02
-9.09507632e-01 -7.15651035e-01 -7.46328175e-01 -8.88187364e-02
8.33532333e-01 -4.18278277e-01 1.42435956e+00 -1.17692447e+00
-1.25666344e+00 6.23337448e-01 -5.08703232e-01 -5.77864766e-01
8.54976535e-01 -4.62237597e-01 -1.61486119e-02 2.87908703e-01
2.10340559e-01 1.16666651e+00 7.85453081e-01 -1.25610304e+00
-8.57148290e-01 -8.65567029e-02 3.25387895e-01 2.06003368e-01
1.69993304e-02 9.86681506e-02 -1.36300480e+00 -6.79631650e-01
1.29630566e-01 -9.56572890e-01 -4.01121199e-01 -2.43936796e-02
-4.40088958e-01 7.27044940e-02 9.11885858e-01 -7.40610659e-01
1.26117051e+00 -2.27676535e+00 1.55891344e-01 -8.06505382e-02
3.66563261e-01 1.91491470e-01 -1.12397775e-01 -1.09958410e-01
4.52517457e-02 1.66390449e-01 -2.87546843e-01 -6.66664839e-01
-4.14369732e-01 1.47321060e-01 -5.61280400e-02 5.07227540e-01
1.82939976e-01 9.78046656e-01 -8.29143643e-01 -6.82902038e-01
4.92102772e-01 5.25418282e-01 -7.39333093e-01 -5.41949421e-02
-4.77749169e-01 6.82389379e-01 -3.10542732e-01 8.20421815e-01
7.45932937e-01 -4.12434757e-01 4.03403416e-02 -1.83814466e-01
-3.98506343e-01 1.62880212e-01 -1.15792131e+00 1.98342669e+00
1.10907458e-01 7.72463500e-01 3.08483452e-01 -8.09477985e-01
3.25864643e-01 1.72915250e-01 9.78910685e-01 -8.09611619e-01
1.36448860e-01 3.43052149e-02 -7.31713995e-02 -3.55320036e-01
6.13250852e-01 4.69776601e-01 3.38467270e-01 1.67754233e-01
-6.40430897e-02 2.52233952e-01 5.83793819e-01 2.38295600e-01
1.10773778e+00 6.25186205e-01 -1.06720671e-01 -2.33221669e-02
1.27028435e-01 1.80547491e-01 9.56905246e-01 8.45373213e-01
-5.24133921e-01 1.08977318e+00 3.93933415e-01 -4.39213008e-01
-8.60065699e-01 -9.69260931e-01 -4.88763303e-02 9.67185616e-01
6.78340912e-01 -4.29580182e-01 -9.10015643e-01 -7.30161428e-01
-1.76787078e-01 4.53132927e-01 -5.51505566e-01 3.85654211e-01
-8.70403707e-01 -3.10924202e-01 1.51514739e-01 6.09028816e-01
4.29381818e-01 -1.08157539e+00 -8.54733169e-01 5.05482852e-01
-3.89395118e-01 -1.45514071e+00 -6.86450958e-01 1.85393821e-02
-9.05001819e-01 -1.18218279e+00 -7.95629382e-01 -5.92658281e-01
4.96213466e-01 6.37760818e-01 1.06095624e+00 4.02466655e-01
-3.71540099e-01 2.72843897e-01 -4.60754633e-01 8.01392365e-03
2.03770492e-03 1.31292880e-01 -2.50163376e-01 -8.51141438e-02
1.64291859e-01 -3.03953856e-01 -9.89570439e-01 5.04968941e-01
-8.66251469e-01 4.51751471e-01 2.67778963e-01 4.55380559e-01
8.24622929e-01 -1.19826958e-01 3.18168074e-01 -6.30705476e-01
-1.28028810e-01 -3.99199694e-01 -7.92496502e-01 8.29255302e-03
-2.64444560e-01 -3.58943701e-01 1.22468144e-01 -4.55468863e-01
-9.32370603e-01 4.40165520e-01 9.61484089e-02 -8.10715795e-01
-2.18391567e-01 5.97794130e-02 1.89213634e-01 -1.01126090e-01
1.92295760e-01 7.58988112e-02 -1.01069055e-01 -3.24164331e-01
3.43447328e-01 2.29838923e-01 4.76770192e-01 -3.67498070e-01
4.80990142e-01 7.53293037e-01 -3.06255132e-01 -5.32469034e-01
-1.03235447e+00 -7.73332596e-01 -7.73570359e-01 -5.62887669e-01
1.15743685e+00 -1.12550461e+00 -4.95749354e-01 2.88373291e-01
-1.26104856e+00 -7.88302839e-01 -1.25948831e-01 3.27705503e-01
-5.26539147e-01 3.40299696e-01 -8.09808016e-01 -7.67505407e-01
3.57572250e-02 -1.35765231e+00 1.10269487e+00 1.88810602e-01
-2.27281064e-01 -4.94270295e-01 -2.96765208e-01 5.77681601e-01
2.40193456e-01 3.42620939e-01 -7.21707344e-02 -2.35254481e-01
-1.26134276e+00 1.89896122e-01 -3.28596383e-01 1.27754167e-01
-2.00605206e-02 1.07007645e-01 -8.50270271e-01 -1.76019430e-01
-1.79368019e-01 7.05433711e-02 1.19285512e+00 7.73430824e-01
1.39386964e+00 -1.74133718e-01 -3.62660348e-01 7.06439734e-01
1.23325372e+00 2.89193988e-01 6.28047705e-01 4.19285446e-01
9.77127373e-01 5.10500908e-01 8.29311132e-01 3.85332555e-01
4.18218672e-01 8.87047172e-01 6.17458045e-01 -3.04403186e-01
-3.88729870e-01 4.29769009e-02 2.65885144e-01 2.62194425e-01
-8.84078220e-02 -3.51804972e-01 -9.16205943e-01 7.67831624e-01
-2.17762542e+00 -1.08655596e+00 -2.52019048e-01 1.98046672e+00
6.30707920e-01 3.21827799e-01 3.23533773e-01 -2.05223143e-01
8.63545120e-01 2.08823264e-01 -6.42903030e-01 2.08126992e-01
-4.09697928e-02 -1.67684965e-02 7.50398874e-01 6.34404898e-01
-1.32016170e+00 1.39345479e+00 5.36759615e+00 5.37551343e-01
-9.72436726e-01 3.32692385e-01 8.79210353e-01 -6.40387297e-01
1.54438779e-01 7.94939175e-02 -7.58841276e-01 6.12053812e-01
6.01362467e-01 2.63517499e-01 5.30414999e-01 4.84922916e-01
6.20107174e-01 -3.13043356e-01 -1.12598860e+00 9.38887417e-01
-7.36602545e-02 -1.51553774e+00 -4.02289003e-01 7.85515606e-02
8.97429585e-01 4.61792082e-01 2.67272778e-02 1.06486052e-01
2.10594926e-02 -8.88692975e-01 1.04972661e+00 2.46729359e-01
5.27960956e-01 -5.82264006e-01 4.29821312e-01 1.22020453e-01
-1.28871477e+00 -3.80741805e-02 8.59181397e-03 -5.87143609e-03
4.97716755e-01 4.92761880e-01 -5.43986142e-01 2.88158953e-01
9.17204738e-01 9.20206487e-01 -5.05433440e-01 1.33237195e+00
-2.60011852e-02 5.79355955e-01 -4.10598248e-01 5.68500340e-01
3.70651543e-01 -1.82417512e-01 4.19847548e-01 1.24622190e+00
2.50151008e-01 2.26798326e-01 6.08223021e-01 9.27628696e-01
-7.78696537e-02 -2.16218501e-01 -3.38030577e-01 1.98896363e-01
4.68365788e-01 1.12970781e+00 -1.16257060e+00 -3.80037457e-01
-5.56888938e-01 1.14722586e+00 1.55725002e-01 5.97584605e-01
-1.24269116e+00 2.44476199e-01 6.24518752e-01 2.65111148e-01
6.86156571e-01 -4.24682707e-01 -3.04134816e-01 -1.21683717e+00
1.32504731e-01 -6.18909597e-01 2.29438156e-01 -7.35053062e-01
-6.85033977e-01 4.15580720e-01 -4.52945903e-02 -1.19152772e+00
-1.20371422e-02 -2.35887021e-01 -3.07367414e-01 3.75563174e-01
-1.56523335e+00 -9.08694744e-01 -4.21532512e-01 5.49712121e-01
1.04432046e+00 4.59723324e-01 1.26398385e-01 5.16470253e-01
-7.43912578e-01 3.24821293e-01 -2.33445659e-01 2.58991361e-01
6.21578693e-01 -1.01583827e+00 5.19817054e-01 1.22810459e+00
2.55077660e-01 2.60020733e-01 6.37252569e-01 -7.99732566e-01
-1.05288422e+00 -1.25984585e+00 4.79530662e-01 -5.38790107e-01
5.25883675e-01 -3.11391532e-01 -9.65222120e-01 7.64447868e-01
1.86341465e-01 3.56597245e-01 2.49508679e-01 -3.43005270e-01
-1.69829905e-01 1.18978947e-01 -9.78046477e-01 7.72540033e-01
1.30819106e+00 -1.18340753e-01 -7.99881890e-02 3.73814583e-01
1.03333509e+00 -7.89705098e-01 -6.21853948e-01 4.12513465e-01
4.17590857e-01 -1.15217471e+00 9.60903168e-01 -1.62244767e-01
3.94929558e-01 -6.22976243e-01 5.84535226e-02 -6.51394129e-01
-2.21975580e-01 -6.90004408e-01 -3.60385686e-01 1.00080550e+00
3.71780723e-01 -1.66437104e-01 1.11379373e+00 7.52146304e-01
-2.68311262e-01 -6.90430045e-01 -1.02153802e+00 -7.19847262e-01
-2.95820802e-01 -6.11668706e-01 2.70348102e-01 8.78702343e-01
-2.87407935e-01 -1.52070284e-01 -4.43201751e-01 2.94307351e-01
6.54386938e-01 1.34616718e-01 6.41271234e-01 -9.16969240e-01
-3.49428922e-01 -5.66991389e-01 -2.61926085e-01 -1.41934574e+00
1.97735161e-01 -5.51716447e-01 2.09468633e-01 -1.50619519e+00
1.26145959e-01 -3.82397860e-01 -3.52330148e-01 6.42598331e-01
-1.04020834e-01 6.42242789e-01 4.84866381e-01 3.34441870e-01
-1.18892109e+00 2.35540792e-01 1.20450032e+00 -4.11871113e-02
-3.99765760e-01 -1.63869977e-01 -4.04872537e-01 7.50050247e-01
1.19807565e+00 -5.83860040e-01 -1.92464948e-01 -6.52068198e-01
-2.16763631e-01 1.65487409e-01 4.81528282e-01 -1.03099847e+00
2.74171770e-01 -3.02735031e-01 3.90172601e-01 -7.67495453e-01
5.68529725e-01 -7.95941532e-01 2.07521662e-01 4.49251622e-01
-1.62608460e-01 4.85623665e-02 3.53764534e-01 5.41755080e-01
-6.19938113e-02 -1.16898589e-01 7.80784428e-01 -2.06001922e-01
-1.14051116e+00 4.86198515e-01 -2.47361109e-01 9.53539163e-02
1.27426279e+00 -5.44160664e-01 -2.89199650e-01 -2.19650880e-01
-7.72723138e-01 5.45409083e-01 7.98247695e-01 5.18593907e-01
5.38433790e-01 -8.60825419e-01 -4.97364819e-01 3.35401110e-02
-1.39451161e-01 2.43707716e-01 4.05763239e-01 1.19569254e+00
-5.93655169e-01 3.24615091e-01 -2.31881160e-03 -9.24482822e-01
-1.38024962e+00 4.93644983e-01 3.46893013e-01 1.61719456e-01
-8.50524604e-01 8.50856066e-01 3.76125991e-01 1.64016008e-01
4.94547844e-01 -4.06144708e-01 -1.28891384e-02 -1.29517272e-01
4.80125517e-01 2.03137606e-01 -1.69414937e-01 -7.06697285e-01
-4.23339158e-01 5.11797726e-01 -2.68597841e-01 -2.47943625e-01
1.12053764e+00 -3.81945491e-01 1.27894983e-01 2.62369424e-01
1.00283098e+00 -1.28956124e-01 -1.93200648e+00 -1.50933206e-01
-9.38198417e-02 -6.43544436e-01 1.55476645e-01 -6.37491882e-01
-1.36544788e+00 4.58009720e-01 4.81830418e-01 1.69247791e-01
1.13218665e+00 1.22742414e-01 8.67041767e-01 -6.06889874e-02
3.07245404e-01 -1.11152077e+00 8.24410915e-02 5.03194630e-01
4.94944721e-01 -1.38678443e+00 -4.69826497e-02 -6.15050495e-01
-5.76618910e-01 8.10721159e-01 9.89938617e-01 -6.04898296e-02
4.75998521e-01 3.98630500e-01 -1.06181316e-02 -4.56617512e-02
-7.14493454e-01 -3.95360947e-01 1.86769113e-01 3.29168171e-01
4.76556569e-01 -1.26538917e-01 -2.48543397e-01 1.16152994e-01
1.45541355e-01 1.58723399e-01 6.99666619e-01 8.98269415e-01
-3.43885273e-01 -9.37309146e-01 -2.47826234e-01 2.67886668e-01
-7.62740493e-01 -1.99016169e-01 -2.07951635e-01 7.39233494e-01
1.81694791e-01 1.13428223e+00 2.52064794e-01 -1.73769176e-01
4.43955027e-02 -2.81910121e-01 4.40907329e-01 -5.76170743e-01
-4.28127468e-01 5.32014072e-01 6.97858073e-03 -1.13897109e+00
-7.95192659e-01 -7.54053652e-01 -1.52959478e+00 -1.70817196e-01
-3.08031201e-01 -1.36311099e-01 4.50269133e-01 6.99368894e-01
5.55674076e-01 6.83431745e-01 2.66988754e-01 -1.19133174e+00
1.74340412e-01 -5.36723256e-01 -8.36918578e-02 2.42126167e-01
4.66500551e-01 -6.67136252e-01 -2.53361881e-01 2.79053539e-01] | [9.130178451538086, -0.15287771821022034] |
c4f5c712-8031-4143-a9ed-a04ea2059cce | experimental-estimation-of-number-of-clusters | 1503.03168 | null | http://arxiv.org/abs/1503.03168v1 | http://arxiv.org/pdf/1503.03168v1.pdf | Experimental Estimation of Number of Clusters Based on Cluster Quality | Text Clustering is a text mining technique which divides the given set of
text documents into significant clusters. It is used for organizing a huge
number of text documents into a well-organized form. In the majority of the
clustering algorithms, the number of clusters must be specified apriori, which
is a drawback of these algorithms. The aim of this paper is to show
experimentally how to determine the number of clusters based on cluster
quality. Since partitional clustering algorithms are well-suited for clustering
large document datasets, we have confined our analysis to a partitional
clustering algorithm. | ['Desikan Kalyani', 'Grace G. Hannah'] | 2015-03-10 | null | null | null | null | ['text-clustering'] | ['natural-language-processing'] | [-3.60471696e-01 -8.24892893e-02 -1.65657505e-01 -4.84503716e-01
-1.34098884e-02 -5.74505448e-01 5.27384222e-01 8.92757893e-01
-4.47203666e-01 3.76049846e-01 1.27457753e-02 -6.14076972e-01
-4.29375976e-01 -1.00201929e+00 2.20351234e-01 -7.00873792e-01
-3.34934354e-01 1.32169235e+00 4.11825955e-01 1.47531360e-01
5.62391341e-01 5.25332868e-01 -1.94101274e+00 5.80547154e-01
8.54577959e-01 4.43573028e-01 1.39402330e-01 6.57307684e-01
-6.79427028e-01 5.22619128e-01 -7.12842464e-01 1.20093592e-01
1.41501511e-02 -7.47307360e-01 -1.16732740e+00 3.56461585e-01
-5.55302620e-01 2.83568174e-01 5.37174821e-01 9.45795715e-01
3.85515802e-02 3.67030919e-01 9.61369157e-01 -1.39171922e+00
3.79353166e-01 9.26880002e-01 -6.17701232e-01 1.69740841e-01
2.71068186e-01 -8.63826692e-01 9.95354891e-01 -6.86433315e-01
6.33029461e-01 1.11661267e+00 2.88297892e-01 -2.45766006e-02
-1.12524569e+00 -3.41100574e-01 8.93107429e-02 1.71574503e-01
-1.58156431e+00 -1.80558622e-01 5.85280061e-01 -5.65798700e-01
8.38799775e-01 5.78011215e-01 6.17090404e-01 -1.20980084e-01
8.90974775e-02 6.27107322e-01 8.94323647e-01 -9.96830046e-01
8.08833897e-01 4.89476502e-01 7.10421085e-01 8.64612982e-02
6.75805211e-01 -5.62374711e-01 1.20832562e-01 -3.41634452e-01
2.43130233e-02 1.20811962e-01 1.45913452e-01 -5.81699729e-01
-1.00504172e+00 1.20948541e+00 -2.28242427e-02 1.01233065e+00
-3.31211895e-01 -3.51906091e-01 5.41696548e-01 2.96642005e-01
3.38771194e-01 4.73877221e-01 -2.86876559e-01 1.06510203e-02
-1.42588055e+00 -2.66707316e-02 1.00015211e+00 8.52238417e-01
7.13132203e-01 -4.86720324e-01 4.54762906e-01 5.40888965e-01
4.07986075e-01 -5.72135225e-02 8.08574736e-01 -5.29236913e-01
1.86572462e-01 1.19625843e+00 6.15008287e-02 -1.10269058e+00
-6.68336689e-01 1.07130021e-01 -8.55457902e-01 8.28629881e-02
3.74927014e-01 -8.64468738e-02 -7.44903445e-01 8.17988217e-01
4.51498330e-01 -8.70690703e-01 2.05230966e-01 4.75590944e-01
4.95489150e-01 7.54005849e-01 -1.53821900e-01 -8.79622340e-01
1.25649512e+00 -5.06649554e-01 -8.46509993e-01 6.10307574e-01
7.73003936e-01 -8.45844030e-01 6.44672930e-01 8.21084559e-01
-1.03480697e+00 -4.67221171e-01 -7.31109381e-01 4.64357078e-01
-6.63501382e-01 -3.45540680e-02 6.34510756e-01 9.02988613e-01
-1.16474879e+00 2.98592955e-01 -7.30450153e-01 -8.34572554e-01
-2.07560554e-01 6.76018536e-01 -1.16089098e-01 1.60465911e-01
-8.94152641e-01 4.23942298e-01 1.15115333e+00 -3.78148735e-01
1.49095386e-01 5.64482845e-02 -3.30494851e-01 2.87729084e-01
1.13759182e-01 -2.79992700e-01 8.84984732e-01 -7.53882587e-01
-7.96186090e-01 9.87547874e-01 -2.89640427e-01 -3.79182726e-01
3.29347372e-01 5.27974427e-01 -5.77539325e-01 4.84616071e-01
1.06313519e-01 4.28057969e-01 5.91518641e-01 -1.65813243e+00
-1.17182088e+00 -5.63381016e-01 -6.90420985e-01 1.16390795e-01
-5.79742551e-01 4.38897640e-01 -5.70421636e-01 -3.78804654e-01
4.90760058e-01 -7.95190096e-01 -6.05084062e-01 -1.09183586e+00
-5.42322218e-01 -8.34761262e-01 1.12952006e+00 1.01136252e-01
1.69722116e+00 -1.89455068e+00 -1.58340216e-01 9.99860823e-01
5.53383470e-01 -4.25693542e-02 7.14630246e-01 9.07939017e-01
-1.13503955e-01 3.82924914e-01 -1.18113920e-01 -1.27038389e-01
-7.83121362e-02 2.13934049e-01 -2.19734356e-01 4.86202270e-01
-4.82330710e-01 2.82367647e-01 -6.91813529e-01 -1.16525280e+00
3.90734613e-01 1.71776731e-02 -4.33325320e-01 1.16450056e-01
-1.31440789e-01 -9.28091705e-02 -6.41323447e-01 2.10406110e-01
5.04140913e-01 -1.45746902e-01 5.51485896e-01 2.83982337e-01
-6.56650364e-01 8.89502540e-02 -1.57485926e+00 8.20819438e-01
4.73266095e-01 4.50010806e-01 6.98506013e-02 -1.42508388e+00
1.04585576e+00 4.53257143e-01 1.29701018e+00 2.50511654e-02
4.90691751e-01 -2.71288212e-02 1.08833812e-01 -2.42014244e-01
7.61592925e-01 -2.41667807e-01 -1.52114853e-01 1.06973517e+00
-3.98012340e-01 1.01314574e-01 9.29796696e-01 4.38061506e-01
8.21943581e-01 -9.38103795e-01 4.82995480e-01 -9.79201615e-01
5.27681530e-01 7.70705223e-01 1.46442428e-01 4.96421307e-01
1.79761291e-01 5.51375687e-01 5.24594069e-01 -5.60337424e-01
-9.93315697e-01 -5.15893996e-01 -4.05103356e-01 9.85590219e-01
7.47949854e-02 -8.38623703e-01 -1.11056995e+00 -4.36880827e-01
-1.61541268e-01 4.43319052e-01 -6.60962284e-01 3.36855561e-01
-2.64329255e-01 -1.09301209e+00 9.87643227e-02 1.06773913e-01
1.93648234e-01 -8.99870098e-01 -7.07170963e-01 3.87031674e-01
-2.01503754e-01 -4.19120252e-01 -1.34788617e-01 6.74829185e-01
-9.97665226e-01 -1.34348118e+00 -3.16844970e-01 -1.15098763e+00
1.00634432e+00 7.00638890e-01 1.07815647e+00 4.56049472e-01
-5.82185537e-02 3.45761895e-01 -1.10447812e+00 -6.09406650e-01
-7.47655928e-01 5.35149872e-01 2.20339090e-01 2.57043652e-02
1.06515276e+00 -2.84885675e-01 -1.89442277e-01 7.28627443e-01
-1.08111346e+00 -2.26500079e-01 2.04528660e-01 2.85670310e-01
6.38247013e-01 1.50650895e+00 2.35172763e-01 -1.10109735e+00
1.04717875e+00 -4.96131331e-01 -4.56082493e-01 2.38947853e-01
-1.17669117e+00 -8.07474032e-02 8.70653808e-01 -1.70871109e-01
-5.95094264e-01 3.82931352e-01 1.55666187e-01 1.82482943e-01
-4.81128365e-01 7.28480339e-01 -2.17802227e-01 4.28025693e-01
6.98389947e-01 1.32570550e-01 -1.00845337e-01 -4.36004460e-01
1.35146797e-01 1.18083835e+00 1.49635553e-01 -4.27623540e-01
5.61651289e-01 4.86949176e-01 -3.85505445e-02 -9.94289279e-01
-1.77501440e-01 -1.56496978e+00 -1.32264817e+00 -3.56300086e-01
7.17076540e-01 -4.73015308e-01 -8.13656867e-01 -1.04808927e-01
-7.43450344e-01 5.77804446e-02 -2.55668551e-01 4.68421012e-01
-3.69848520e-01 6.41686380e-01 -8.82976204e-02 -8.78863394e-01
-2.68510073e-01 -8.01244259e-01 4.06802744e-01 -2.63191879e-01
-6.02089882e-01 -1.08199978e+00 3.11893135e-01 9.74758416e-02
-3.06551278e-01 1.62839651e-01 9.50069308e-01 -1.20752215e+00
1.96489632e-01 -5.87635040e-01 2.56821245e-01 -1.96747959e-01
2.80512333e-01 2.58699745e-01 -5.06343484e-01 -4.98782307e-01
1.97874337e-01 1.33140177e-01 8.34890962e-01 5.16642928e-01
8.97447765e-01 -2.53785998e-01 -8.26584458e-01 5.91650419e-02
1.40368891e+00 9.49392557e-01 6.59494102e-01 5.02550244e-01
4.45827127e-01 1.05257428e+00 8.62850904e-01 6.28186285e-01
1.45306453e-01 2.63754964e-01 -5.45411333e-02 -1.57737970e-01
7.38623083e-01 1.56943440e-01 -3.01063478e-01 9.57714677e-01
2.84334049e-02 -5.30084014e-01 -1.36049557e+00 5.61257422e-01
-1.90591502e+00 -1.09256029e+00 -8.85213733e-01 1.96006799e+00
7.06853449e-01 3.12049270e-01 8.57168794e-01 1.22847867e+00
1.02964664e+00 -5.73213160e-01 2.07410865e-02 -5.61116755e-01
2.14456409e-01 -2.92820781e-01 2.10765585e-01 3.58864009e-01
-1.05315483e+00 7.95094967e-01 6.98023701e+00 5.72015524e-01
-7.24321723e-01 -4.54825968e-01 5.98377705e-01 1.77078649e-01
-1.23275453e-02 3.04117128e-02 -9.48652565e-01 5.51845968e-01
1.07239902e+00 -4.42890823e-01 -9.77135822e-02 7.94360101e-01
5.38359702e-01 -5.88418543e-01 -8.96406949e-01 6.25312567e-01
-1.71166942e-01 -9.60114479e-01 2.01954961e-01 4.18106437e-01
7.72110760e-01 -4.69275415e-01 -2.49913916e-01 -1.11359030e-01
7.12468863e-01 -9.91985679e-01 2.35054255e-01 -3.05316925e-01
4.53343034e-01 -1.33239532e+00 6.76775634e-01 5.91399729e-01
-1.16712368e+00 -4.54867259e-02 -5.66837251e-01 -1.19796336e-01
-6.25032783e-02 9.14539874e-01 -1.05926049e+00 3.98020804e-01
1.05452073e+00 4.26204473e-01 -4.46894526e-01 1.15836370e+00
3.58742893e-01 7.24365652e-01 -6.11303926e-01 -3.86419803e-01
4.60478067e-01 -4.56996113e-01 2.52773520e-02 1.41705728e+00
7.31143951e-02 3.08492690e-01 5.05717516e-01 2.56269783e-01
3.93228978e-01 6.99396729e-01 -3.75419229e-01 -2.36228481e-03
6.66318774e-01 1.13823378e+00 -1.92001581e+00 -4.86489892e-01
2.72194073e-02 4.23173666e-01 -2.16431156e-01 -1.32575542e-01
-3.37848552e-02 -7.47573733e-01 3.14260498e-02 5.49466193e-01
3.42891812e-01 -3.35264921e-01 -7.09641278e-01 -7.27448046e-01
-3.95930409e-01 -6.12390995e-01 9.50868785e-01 -2.36138433e-01
-8.53758454e-01 7.91451693e-01 2.92037308e-01 -1.44648695e+00
-5.83648682e-01 -3.54951352e-01 -6.25244319e-01 4.12770718e-01
-6.08633637e-01 -6.03577793e-01 -5.79408929e-02 9.13284063e-01
3.58919322e-01 -3.77833217e-01 8.30052018e-01 -1.70269415e-01
-3.96668732e-01 1.98237136e-01 8.03575218e-01 2.49733210e-01
5.66981554e-01 -1.60223734e+00 1.42587153e-02 7.61134088e-01
1.09459780e-01 8.31865788e-01 1.01308465e+00 -6.60492599e-01
-7.83485830e-01 -8.92881632e-01 1.26433361e+00 -3.96613598e-01
4.81755018e-01 -4.04336631e-01 -9.56012905e-01 2.83290803e-01
2.32885197e-01 -6.05901778e-01 1.15310585e+00 4.38692570e-01
1.24237292e-01 6.41478673e-02 -1.16009951e+00 3.21838915e-01
1.26431793e-01 2.38377318e-01 -6.53053820e-01 2.96159506e-01
1.51105344e-01 3.20877880e-01 -1.07550740e+00 -2.15481639e-01
1.42167866e-01 -1.17051435e+00 5.76162279e-01 -4.16200608e-01
1.50960460e-01 -5.50083995e-01 2.15402231e-01 -1.11836743e+00
-3.19135696e-01 -6.97876096e-01 1.77504629e-01 1.31308734e+00
2.98140287e-01 -2.80805796e-01 1.18837690e+00 4.14189130e-01
3.42291594e-01 -2.68734694e-01 -7.05214143e-01 -7.82512486e-01
2.24082828e-01 -3.18458050e-01 5.49986482e-01 1.45093763e+00
8.85758519e-01 5.58385193e-01 5.60814105e-02 -2.42696181e-01
7.72884488e-01 5.11225164e-01 6.43370807e-01 -2.06995368e+00
3.01674396e-01 -7.48274505e-01 -3.29315722e-01 -4.22711223e-01
-2.24774197e-01 -8.24681044e-01 9.33206379e-02 -1.71842670e+00
2.34232053e-01 -7.17903852e-01 -1.65995844e-02 1.68920785e-01
3.50435562e-02 -8.50376412e-02 -9.84401926e-02 8.47628295e-01
-9.20344353e-01 -8.45636055e-02 5.28244257e-01 3.81322317e-02
-8.25877130e-01 4.24881339e-01 -6.60107791e-01 5.07438719e-01
1.22366512e+00 -6.82406843e-01 -6.82492554e-01 2.62342691e-01
2.60298282e-01 -2.92963028e-01 -7.03112543e-01 -7.92330921e-01
8.26738298e-01 -2.16472313e-01 3.34201753e-01 -1.32044876e+00
-4.94366020e-01 -1.24410558e+00 2.99763501e-01 5.17311752e-01
-3.86448562e-01 1.25299498e-01 -7.29353204e-02 3.87968153e-01
-5.48367262e-01 -5.14065921e-01 9.64566708e-01 -2.43889287e-01
-3.89206976e-01 -7.64558613e-02 -1.24708796e+00 -2.40508795e-01
1.33084452e+00 -7.30816364e-01 4.80724216e-01 -2.33787075e-01
-7.44963348e-01 4.98467356e-01 7.64584839e-01 4.17917520e-02
4.02862012e-01 -1.09049213e+00 -4.87295032e-01 2.65862104e-02
3.14254731e-01 3.72320533e-01 -3.38590443e-01 4.13183719e-01
-8.28026831e-01 8.35431099e-01 -1.82604641e-02 -5.66919088e-01
-1.70738268e+00 1.31094038e+00 -2.01415688e-01 -2.56009609e-01
-7.68906951e-01 1.01670891e-01 -2.39748493e-01 -1.36529893e-01
3.29062551e-01 -1.04744732e-01 -8.10797393e-01 5.39821506e-01
7.86465883e-01 5.30120492e-01 1.85928911e-01 -6.00079060e-01
-2.73834944e-01 2.68479258e-01 -2.49058619e-01 -7.41008297e-02
1.29404867e+00 -3.16509217e-01 -7.12254465e-01 5.73075294e-01
1.09296083e+00 -1.36730894e-01 -3.94828528e-01 1.15438260e-01
8.48427057e-01 -2.20176458e-01 -9.57298651e-03 -2.17602432e-01
-5.71004391e-01 5.46826720e-01 1.31572351e-01 1.25096905e+00
1.30099487e+00 2.63370991e-01 2.26858914e-01 6.95828915e-01
2.41471201e-01 -1.57648671e+00 -1.61945120e-01 4.25054997e-01
2.25058377e-01 -1.12902510e+00 3.44979674e-01 -5.10054708e-01
-4.56150711e-01 1.15819621e+00 1.86754331e-01 1.01021565e-01
1.06888986e+00 6.50720477e-01 -7.78155401e-02 -4.31262225e-01
-5.94603837e-01 -3.04953933e-01 2.44198851e-02 6.60188079e-01
4.95012373e-01 1.03645325e-01 -8.25239420e-01 4.98692632e-01
-5.53086400e-01 -4.27374274e-01 7.53290117e-01 1.09077179e+00
-1.25102735e+00 -1.22408521e+00 -9.85304594e-01 6.32134438e-01
-6.47212505e-01 2.14447692e-01 -9.72533107e-01 9.74460781e-01
5.81172407e-02 1.60004544e+00 3.38640988e-01 -3.89219850e-01
2.04166219e-01 -5.77965863e-02 -1.43301800e-01 -7.19025314e-01
-7.31839776e-01 3.98898214e-01 -2.11193144e-01 2.07108378e-01
-6.72742248e-01 -8.07832837e-01 -1.77501380e+00 -6.74606919e-01
-3.64450842e-01 1.30402780e+00 6.74881995e-01 6.40868723e-01
-1.74448714e-01 2.08804775e-02 1.04879212e+00 -3.76703829e-01
1.58831641e-01 -8.30495715e-01 -1.23407054e+00 3.31692278e-01
-1.94580942e-01 -3.11245173e-01 -3.34828049e-01 5.55354834e-01] | [10.264469146728516, 7.0787129402160645] |
6e84c35f-6702-4bc4-9c8b-c9ffb08b4755 | sidenoter-scholarly-paper-browsing-system | null | null | https://aclanthology.org/C16-2029 | https://aclanthology.org/C16-2029.pdf | SideNoter: Scholarly Paper Browsing System based on PDF Restructuring and Text Annotation | In this paper, we discuss our ongoing efforts to construct a scientific paper browsing system that helps users to read and understand advanced technical content distributed in PDF. Since PDF is a format specifically designed for printing, layout and logical structures of documents are indistinguishably embedded in the file. It requires much effort to extract natural language text from PDF files, and reversely, display semantic annotations produced by NLP tools on the original page layout. In our browsing system, we tackle these issues caused by the gap between printable document and plain text. Our system provides ways to extract natural language sentences from PDF files together with their logical structures, and also to map arbitrary textual spans to their corresponding regions on page images. We setup a demonstration system using papers published in ACL anthology and demonstrate the enhanced search and refined recommendation functions which we plan to make widely available to NLP researchers. | ['Akiko Aizawa', 'Takeshi Abekawa'] | 2016-12-01 | sidenoter-scholarly-paper-browsing-system-1 | https://aclanthology.org/C16-2029 | https://aclanthology.org/C16-2029.pdf | coling-2016-12 | ['text-annotation'] | ['natural-language-processing'] | [ 3.65451783e-01 3.67160171e-01 4.66260090e-02 -3.52239490e-01
-9.19203997e-01 -1.28629291e+00 5.46591818e-01 2.21186340e-01
-1.39703110e-01 1.09345925e+00 2.81847239e-01 -7.24705160e-01
-3.32269460e-01 -4.76848572e-01 -8.14547718e-01 2.25263730e-01
2.66254425e-01 7.27738500e-01 4.84140068e-01 1.04147427e-01
8.97195637e-01 7.39391446e-01 -1.27826083e+00 6.82884157e-01
1.02416360e+00 2.60585815e-01 6.87360644e-01 1.07107723e+00
-1.19760633e+00 4.45924312e-01 -1.22628438e+00 -6.84464753e-01
5.04372269e-02 -4.93213683e-01 -1.11888504e+00 -5.42026013e-02
1.00827336e+00 -2.30099335e-01 -3.04998010e-01 1.17187679e+00
1.56546950e-01 -4.68272537e-01 6.72517657e-01 -1.14721191e+00
-9.10597026e-01 9.98477995e-01 -2.97825068e-01 1.01479419e-01
1.00101817e+00 -3.93728226e-01 6.81850493e-01 -7.88909912e-01
1.34720588e+00 1.38408160e+00 4.10835296e-01 4.53316599e-01
-1.00019002e+00 -3.27926755e-01 -1.58812210e-01 8.49510729e-02
-8.86229694e-01 -5.29484868e-01 5.26998878e-01 -4.06186759e-01
6.24151886e-01 6.79239571e-01 5.97920239e-01 1.31612325e+00
5.19707322e-01 8.59072089e-01 1.14407933e+00 -8.73316526e-01
-1.46135828e-02 4.64514911e-01 4.00758713e-01 9.10904646e-01
3.61931413e-01 -6.16608500e-01 -7.65432596e-01 -1.26620963e-01
9.49083507e-01 -5.94256580e-01 -4.26161051e-01 -1.77476168e-01
-1.29440355e+00 -6.96610194e-03 -5.29769897e-01 2.41621450e-01
1.55651346e-01 -1.13013215e-01 3.11619997e-01 2.14450836e-01
-1.57038718e-02 7.32142806e-01 -4.31018531e-01 -6.01882041e-01
-1.12375391e+00 4.04925287e-01 1.36768150e+00 1.64347422e+00
9.53402221e-02 -7.10415423e-01 -2.55836368e-01 6.93878531e-01
3.21932077e-01 5.69259584e-01 6.75120205e-02 -1.33855796e+00
6.53671443e-01 3.25606674e-01 3.45104396e-01 -9.64773953e-01
2.20428668e-02 -1.98794797e-01 -9.60606188e-02 2.06990778e-01
8.07983935e-01 2.54695505e-01 -5.89155197e-01 1.01572859e+00
-7.02165812e-02 -6.26634479e-01 -2.21607596e-01 4.70638305e-01
7.69955158e-01 5.27738094e-01 -1.32376105e-01 -1.25755876e-01
1.59905815e+00 -6.74210668e-01 -1.33701980e+00 1.08934291e-01
3.96239460e-01 -1.28621805e+00 1.33959889e+00 7.61657894e-01
-1.49980807e+00 -3.26546341e-01 -1.20565414e+00 -4.51551974e-01
-7.35852301e-01 5.10564089e-01 9.77335274e-02 5.00429988e-01
-1.01887763e+00 1.02601171e+00 -5.24865925e-01 -6.87804103e-01
3.79798681e-01 -1.50040403e-01 -2.31290385e-01 1.04732670e-01
-8.37503195e-01 7.33820796e-01 2.88464427e-01 -1.31996721e-01
6.66845888e-02 -9.08442676e-01 -2.85360128e-01 2.44799167e-01
5.46825588e-01 -6.77061796e-01 1.46541178e+00 -3.46955627e-01
-1.57220948e+00 8.43991518e-01 -1.48527235e-01 -2.08422780e-01
8.29940736e-01 -7.20650077e-01 -5.68787634e-01 4.66145396e-01
1.28214374e-01 3.85063469e-01 6.66561425e-01 -1.12148881e+00
-5.57340264e-01 -2.16576457e-01 -8.52753222e-02 -1.59748346e-01
-3.40545744e-01 3.39082688e-01 -9.70012069e-01 -9.60342944e-01
-1.53757185e-02 -5.41526616e-01 4.10113722e-01 5.39513111e-01
-7.78914511e-01 -2.25245789e-01 1.09508598e+00 -1.13889146e+00
1.38772094e+00 -1.80349743e+00 -1.60945561e-02 4.91434544e-01
1.83404163e-01 8.99639875e-02 9.85035151e-02 7.52982736e-01
3.31755400e-01 6.62874699e-01 8.03122818e-02 1.45587847e-01
5.37814498e-01 -6.96023554e-02 -7.01153040e-01 1.84419490e-02
-3.60571206e-01 7.41736233e-01 -7.92824388e-01 -1.06706226e+00
-1.46301106e-01 2.38579035e-01 -1.89781576e-01 6.22864366e-02
-5.10831177e-01 1.12493671e-01 -6.46468222e-01 5.91307998e-01
7.58493364e-01 -2.46255234e-01 4.17390496e-01 -2.46946514e-01
-2.73167700e-01 6.02244914e-01 -1.17604959e+00 1.89761984e+00
-1.87132955e-01 1.03269541e+00 2.30410799e-01 -1.69607803e-01
7.87501693e-01 8.36674720e-02 -1.46956993e-02 -5.34764469e-01
-3.04939985e-01 1.98035359e-01 -5.67430735e-01 -6.92828476e-01
9.28536594e-01 6.75824940e-01 -1.44870743e-01 7.86939085e-01
-6.28660768e-02 -2.69158125e-01 7.45211303e-01 7.98072278e-01
1.01890814e+00 8.02246094e-01 -6.76892176e-02 -5.08492112e-01
4.98702794e-01 1.75633579e-01 -7.65534565e-02 1.22997272e+00
4.05945271e-01 5.22768974e-01 8.11353743e-01 -8.35463032e-02
-1.39031982e+00 -1.42068231e+00 -2.81358778e-01 7.06526637e-01
-1.07870191e-01 -1.13737178e+00 -1.13064778e+00 -9.16436315e-01
-2.84431845e-01 8.97788107e-01 -1.63666412e-01 6.04198158e-01
-6.81955695e-01 2.29220539e-01 6.29592597e-01 5.23371994e-01
3.32530916e-01 -9.82318640e-01 -4.60295767e-01 1.13183528e-01
-1.31260604e-01 -1.03860688e+00 -6.40412450e-01 -6.78325817e-02
-7.54476786e-01 -9.36101258e-01 -7.16960490e-01 -6.69875264e-01
8.22304666e-01 -2.07566187e-01 1.00639379e+00 1.56071350e-01
-4.16334540e-01 4.69820470e-01 -3.33550781e-01 -5.28009057e-01
-6.16538644e-01 1.25322416e-01 -3.68011862e-01 -9.02012885e-01
-1.38579784e-02 -3.01868439e-01 -2.15241686e-03 -3.58681828e-02
-1.05843055e+00 3.70889544e-01 4.19112861e-01 4.08541530e-01
4.18185920e-01 -3.06549929e-02 -3.12290117e-02 -1.17343497e+00
1.03545725e+00 1.21468149e-01 -7.18254089e-01 8.26837182e-01
-4.35272038e-01 4.14595842e-01 5.99582374e-01 3.80427986e-02
-1.30533648e+00 -2.59338558e-01 -9.61866900e-02 5.85476272e-02
-2.98523843e-01 4.41230595e-01 -4.53822970e-01 2.37190306e-01
3.37237090e-01 -4.15275358e-02 -1.93490431e-01 -1.16891360e+00
6.05719268e-01 9.39069748e-01 9.38244283e-01 -8.55216026e-01
9.74889278e-01 3.16259302e-02 -1.35442942e-01 -7.52649963e-01
-6.62466645e-01 -3.65810320e-02 -9.07429397e-01 -2.04942927e-01
5.05315065e-01 -1.29525125e-01 -5.63607156e-01 -6.22842498e-02
-1.50509930e+00 1.01544611e-01 -2.56098092e-01 2.82314420e-01
-5.03657103e-01 8.07421625e-01 -6.97053850e-01 -4.39205915e-01
-2.39577278e-01 -6.21093631e-01 1.03890073e+00 4.31205571e-01
-6.82284415e-01 -8.86798024e-01 -1.19803704e-01 2.90369600e-01
8.17271508e-03 -1.66377440e-01 1.23671591e+00 -6.23384595e-01
-7.88897574e-01 1.09503977e-01 -3.70450944e-01 -1.88625529e-01
-1.19763471e-01 5.83264530e-01 -7.05160022e-01 1.00623056e-01
-5.48245251e-01 7.26376995e-02 4.13755745e-01 -9.23250020e-02
1.62636411e+00 -6.87979937e-01 -7.73010015e-01 3.61190826e-01
1.14268970e+00 2.85669446e-01 9.38766301e-01 7.20265090e-01
4.31534082e-01 6.80814385e-01 6.05227172e-01 1.39260992e-01
-2.06267163e-01 5.89853287e-01 -3.96680206e-01 3.28677922e-01
-3.41832310e-01 -8.46084654e-01 1.68819934e-01 9.82517838e-01
2.35854089e-01 -6.80047214e-01 -7.40588844e-01 9.72230211e-02
-1.54307747e+00 -1.08089578e+00 -4.19755340e-01 1.94132173e+00
9.94996667e-01 5.22885263e-01 -1.97803810e-01 -1.62391394e-01
6.99444890e-01 -2.74745226e-01 8.46627280e-02 -7.20846832e-01
-1.61316618e-01 3.73993635e-01 4.52265769e-01 4.67316449e-01
-7.60674834e-01 8.35789144e-01 7.19928265e+00 1.05266106e+00
-7.17355728e-01 -2.89198041e-01 -7.42433546e-03 4.90510762e-02
-5.30382216e-01 9.84924808e-02 -9.89773989e-01 5.17900407e-01
1.10511303e+00 -4.81531769e-01 2.56566942e-01 6.06553674e-01
4.64357585e-01 1.21962493e-02 -1.27599788e+00 6.95611954e-01
-9.78568643e-02 -1.87931108e+00 3.96858126e-01 1.55183943e-02
2.18787625e-01 -7.42399395e-01 -1.39975294e-01 -1.37810722e-01
2.68387139e-01 -1.02866757e+00 1.14130509e+00 1.08723366e+00
8.74170780e-01 -3.57601523e-01 4.28014249e-01 7.85891712e-02
-7.00409234e-01 4.07501578e-01 -3.96784991e-01 3.26719344e-01
1.45338848e-01 5.90893626e-01 -9.08089101e-01 5.53050160e-01
8.03167939e-01 4.98843312e-01 -8.54415119e-01 1.27385437e+00
-4.83269185e-01 2.51960486e-01 -1.61585167e-01 -4.97663796e-01
-3.19594949e-01 -2.17367232e-01 7.48018622e-01 1.66248631e+00
7.24895895e-01 -1.08169895e-02 -2.05917642e-01 1.27267563e+00
-3.50676119e-01 2.42248356e-01 -5.22902012e-01 -7.54790664e-01
7.13231325e-01 1.27732933e+00 -1.15477109e+00 -4.55502331e-01
-1.34775102e-01 1.12099576e+00 1.05634451e-01 4.30885881e-01
-7.94268548e-01 -1.25518787e+00 6.66555464e-02 3.52440834e-01
3.13537866e-01 -4.72955912e-01 -5.04966378e-01 -1.09530616e+00
4.43294317e-01 -8.42586458e-01 2.60635628e-03 -1.44121230e+00
-1.04185534e+00 5.73028922e-01 5.10916114e-02 -1.12105107e+00
1.52415512e-02 -7.10738599e-01 -4.09426868e-01 1.00180328e+00
-7.29206145e-01 -6.17882013e-01 4.04869393e-02 1.15645900e-01
6.59873605e-01 3.33433561e-02 7.82706082e-01 1.18098795e-01
-2.45090306e-01 4.11704272e-01 5.87935150e-01 2.70976692e-01
9.65042651e-01 -1.48517549e+00 4.96614158e-01 5.72030544e-01
4.59114730e-01 1.38255739e+00 1.06048250e+00 -1.09768224e+00
-1.58551025e+00 -5.84121704e-01 1.26690078e+00 -6.10488653e-01
7.19084620e-01 -6.78753436e-01 -9.01325285e-01 5.38791001e-01
6.60939217e-01 -7.75505066e-01 4.73993331e-01 -1.92108527e-01
-2.50359148e-01 2.17007950e-01 -9.86486137e-01 7.77455986e-01
1.11668491e+00 -6.69400096e-01 -1.30243599e+00 5.93759537e-01
4.95324224e-01 -6.87773466e-01 -8.48806798e-01 -1.49268836e-01
9.27124023e-01 -5.07616520e-01 7.99040318e-01 -5.45698583e-01
5.92789352e-01 -4.55172271e-01 3.59893262e-01 -1.02142787e+00
-3.12396307e-02 -9.23555017e-01 -2.26193368e-02 1.55712605e+00
8.04943919e-01 -2.46365622e-01 8.19758534e-01 5.55982590e-01
-1.25561759e-01 -2.86531389e-01 -4.19321001e-01 -7.48383641e-01
-1.61861498e-02 -3.53068501e-01 4.38379824e-01 5.40111959e-01
6.75117075e-01 3.04127812e-01 9.20519307e-02 -1.53730780e-01
6.57033741e-01 1.86231107e-01 5.26172876e-01 -1.38837004e+00
-3.35903913e-01 -3.88667703e-01 1.36798993e-01 -1.34556711e+00
-5.96597157e-02 -8.40613961e-01 -4.03814316e-02 -1.95257819e+00
1.58368126e-01 -2.47215658e-01 3.21460813e-01 3.11442226e-01
3.59547943e-01 1.47247929e-02 3.41223210e-01 4.56654996e-01
-6.50175989e-01 -6.69956654e-02 1.30144477e+00 -1.52473092e-01
7.03383684e-02 -3.29334944e-01 -6.48650229e-01 8.59490216e-01
4.88669336e-01 -5.06582618e-01 -3.06233644e-01 -3.42908353e-01
2.87061453e-01 1.68885723e-01 2.47413829e-01 -1.00764763e+00
4.12969977e-01 -2.78723627e-01 1.03787851e+00 -1.07914472e+00
-2.62315795e-02 -6.67272210e-01 6.55061007e-02 2.87475921e-02
-7.95321584e-01 1.03239320e-01 4.01540458e-01 3.23191196e-01
2.74328113e-01 -7.53620028e-01 1.59593552e-01 -3.88445616e-01
-3.54604185e-01 -4.38660771e-01 -7.99623311e-01 1.29197627e-01
5.90794504e-01 -3.33249241e-01 -1.03235638e+00 -2.35036433e-01
-7.58452117e-01 9.11760405e-02 7.84637809e-01 4.18247849e-01
6.36413455e-01 -9.33640957e-01 -1.47259399e-01 8.38211644e-03
-2.82844990e-01 -6.53562009e-01 -1.28029644e-01 5.24722517e-01
-1.29905665e+00 9.22345221e-01 -5.67555547e-01 -2.75551006e-02
-1.46649373e+00 4.41036880e-01 -3.20517510e-01 -9.11002010e-02
-9.81927991e-01 4.18016106e-01 -3.51402670e-01 -2.85792351e-01
6.34615242e-01 -3.85440350e-01 -2.33996928e-01 -2.51144439e-01
9.04877245e-01 3.99068803e-01 7.90605024e-02 2.21362650e-01
-2.18814671e-01 4.60662067e-01 -2.53502190e-01 -5.43767273e-01
1.14161754e+00 -3.70341510e-01 -1.27368987e-01 4.34901983e-01
1.00405598e+00 1.04231310e+00 -8.82162333e-01 2.34332338e-01
5.71889877e-01 -4.62305605e-01 -3.19748253e-01 -1.39875352e+00
-2.49203160e-01 7.38092542e-01 1.53824627e-01 3.41809362e-01
6.16985440e-01 1.66833401e-01 8.04152191e-01 7.04303682e-01
2.72824943e-01 -1.39693511e+00 -2.69935876e-01 3.61021399e-01
9.59572792e-01 -2.84593403e-01 3.89251441e-01 -8.47417474e-01
-7.67819360e-02 2.00771809e+00 3.70367289e-01 3.70707601e-01
2.89641410e-01 9.16089654e-01 -8.66220817e-02 -3.54509294e-01
-5.81161082e-01 4.45223838e-01 2.58453727e-01 6.64716423e-01
6.67648256e-01 -4.64683354e-01 -6.73148513e-01 5.98817587e-01
-5.69170475e-01 2.60790169e-01 1.05165219e+00 1.43476558e+00
-5.87816000e-01 -1.59326053e+00 -6.49617136e-01 4.44006205e-01
-7.18517780e-01 -1.38872907e-01 -9.89112139e-01 8.08408380e-01
-2.43610606e-01 5.39765239e-01 9.09522176e-02 -2.57259514e-02
2.30713323e-01 3.17691684e-01 7.93714583e-01 -5.58207929e-01
-3.07510465e-01 1.71081021e-01 5.25021017e-01 -4.93379265e-01
7.70132104e-03 -7.34258652e-01 -1.46507299e+00 -3.79308164e-01
1.35187134e-01 7.16352284e-01 9.82160211e-01 8.00476611e-01
4.35792685e-01 6.13128483e-01 -4.07677472e-01 -3.27084124e-01
-4.46044743e-01 -7.59689331e-01 -5.90533137e-01 -6.44981489e-02
-1.42527968e-01 -2.82332420e-01 -1.48635320e-02 6.37378454e-01] | [9.881878852844238, 8.121674537658691] |
8f1111c8-d398-4d0f-9b72-abbb8a8c471e | hierarchical-pre-training-for-sequence | 2009.11152 | null | https://arxiv.org/abs/2009.11152v3 | https://arxiv.org/pdf/2009.11152v3.pdf | Hierarchical Pre-training for Sequence Labelling in Spoken Dialog | Sequence labelling tasks like Dialog Act and Emotion/Sentiment identification are a key component of spoken dialog systems. In this work, we propose a new approach to learn generic representations adapted to spoken dialog, which we evaluate on a new benchmark we call Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE benchmark (\texttt{SILICONE}). \texttt{SILICONE} is model-agnostic and contains 10 different datasets of various sizes. We obtain our representations with a hierarchical encoder based on transformer architectures, for which we extend two well-known pre-training objectives. Pre-training is performed on OpenSubtitles: a large corpus of spoken dialog containing over $2.3$ billion of tokens. We demonstrate how hierarchical encoders achieve competitive results with consistently fewer parameters compared to state-of-the-art models and we show their importance for both pre-training and fine-tuning. | ['Matthieu Labeau', 'Pierre Colombo', 'Matteo Manica', 'Emile Chapuis', 'Chloe Clavel'] | 2020-09-23 | null | https://aclanthology.org/2020.findings-emnlp.239 | https://aclanthology.org/2020.findings-emnlp.239.pdf | findings-of-the-association-for-computational | ['emotion-recognition-in-conversation', 'dialogue-act-classification'] | ['natural-language-processing', 'natural-language-processing'] | [ 1.53291553e-01 4.40830946e-01 1.83748871e-01 -1.02269888e+00
-9.79938984e-01 -7.68634260e-01 7.55031943e-01 -2.04783008e-01
-6.34011090e-01 7.65846014e-01 7.64945805e-01 -2.38426894e-01
6.18292332e-01 -8.25226307e-02 -3.35257620e-01 -3.16227198e-01
-5.85831888e-02 1.02320969e+00 5.40084876e-02 -7.96248674e-01
-8.35793689e-02 3.32170129e-02 -1.24278963e+00 5.73457956e-01
4.09463525e-01 1.20583296e+00 -1.15331873e-01 1.17339325e+00
-3.47661823e-01 1.14516962e+00 -1.03297007e+00 -5.75174510e-01
-9.34382901e-02 -4.11803126e-01 -1.36063170e+00 3.29838008e-01
1.72267452e-01 -6.34826303e-01 -3.40812206e-01 6.33239806e-01
8.29330266e-01 2.41984382e-01 6.25950634e-01 -1.13172495e+00
-4.72943604e-01 7.57774174e-01 7.88637549e-02 9.73472148e-02
3.40602785e-01 3.30538094e-01 1.41332352e+00 -8.76209199e-01
4.88162100e-01 1.71216381e+00 4.70690340e-01 1.17016673e+00
-1.10076916e+00 -4.47130531e-01 2.47161523e-01 -1.40335597e-02
-8.79444957e-01 -9.69470739e-01 3.00564617e-01 -2.43205473e-01
1.50238729e+00 2.58102983e-01 7.35343099e-02 1.51382637e+00
-1.83713123e-01 1.11400163e+00 9.80310142e-01 -3.35199594e-01
3.10747296e-01 2.75183678e-01 4.58834201e-01 7.69301593e-01
-8.89006197e-01 -1.48871690e-01 -5.81383348e-01 -2.94009477e-01
3.26237798e-01 -5.12159824e-01 -2.80575454e-01 2.22504772e-02
-9.73674536e-01 1.33887684e+00 1.41375676e-01 1.39121860e-01
1.78036019e-02 1.29928932e-01 8.77702355e-01 5.36823750e-01
4.58578676e-01 2.63533533e-01 -8.71688783e-01 -5.99614680e-01
-3.59861404e-01 2.12016761e-01 1.55226362e+00 1.07856035e+00
4.39327121e-01 1.43310606e-01 -3.61357599e-01 1.31675303e+00
2.82748818e-01 9.35975760e-02 7.77968645e-01 -1.14160156e+00
5.08364320e-01 4.13058370e-01 9.73651633e-02 1.90822855e-01
-5.63125432e-01 1.85567752e-01 -8.25957894e-01 -5.14646284e-02
2.25578353e-01 -7.36497879e-01 -8.70507658e-01 1.87349164e+00
1.47251144e-01 5.14929593e-02 5.20266473e-01 5.99674404e-01
1.26630974e+00 8.07251334e-01 2.02066526e-01 -1.39273498e-02
1.54145980e+00 -1.33648801e+00 -7.43060768e-01 -5.08937359e-01
6.27867103e-01 -5.09282470e-01 1.37169981e+00 1.80351689e-01
-1.14368582e+00 -3.98886949e-01 -8.93124163e-01 -1.22695513e-01
-4.18487847e-01 6.80261031e-02 4.83467132e-01 7.20447481e-01
-1.37090480e+00 2.73608655e-01 -5.89538038e-01 -3.52086663e-01
-1.01511087e-02 5.18852770e-01 -1.91020578e-01 4.43054378e-01
-1.40170276e+00 1.13625097e+00 -1.06639154e-02 -1.22517951e-01
-1.12173152e+00 -4.41829890e-01 -1.03289139e+00 3.27086449e-01
1.25544429e-01 -3.35631788e-01 2.30701876e+00 -7.16691375e-01
-2.40651298e+00 1.09078968e+00 -4.64288175e-01 -7.76066840e-01
2.14922577e-01 -2.33697593e-01 -3.03507715e-01 5.81897013e-02
-1.67396247e-01 9.65021729e-01 7.45532632e-01 -1.12963676e+00
-4.96073782e-01 1.12525985e-01 2.45403931e-01 3.19076627e-01
-4.10949796e-01 4.98525888e-01 -1.62930325e-01 -3.26016277e-01
-6.20360374e-01 -8.68782103e-01 -4.04937148e-01 -5.59056222e-01
-3.59931678e-01 -6.91330552e-01 7.59733915e-01 -5.00769138e-01
9.20959711e-01 -1.92944372e+00 1.36209473e-01 -4.07789171e-01
1.94600359e-01 3.00492436e-01 -3.85269850e-01 8.16461623e-01
7.40174949e-02 2.04314455e-01 -4.43699449e-01 -1.02699649e+00
5.63541770e-01 3.75682801e-01 -4.94157970e-01 -7.24900141e-02
4.76954758e-01 9.15195942e-01 -5.65484524e-01 -2.56605506e-01
2.85245091e-01 3.77945453e-01 -3.94023865e-01 8.58193636e-01
-5.94680369e-01 6.09904170e-01 -1.87711179e-01 3.21282953e-01
2.71244377e-01 -5.10830700e-01 2.40392193e-01 2.22549185e-01
5.22388183e-02 1.01237440e+00 -5.92644155e-01 1.86364543e+00
-7.14091718e-01 5.36831439e-01 4.90338922e-01 -8.45061481e-01
9.30267513e-01 8.12380135e-01 2.50412732e-01 -4.37107474e-01
3.56422186e-01 -4.71826605e-02 -4.86397259e-02 -3.73490423e-01
5.83447218e-01 -1.11464672e-01 -5.39805174e-01 7.72238553e-01
6.64617121e-01 -2.88915575e-01 1.50700271e-01 4.03780341e-01
1.17322433e+00 -4.45059955e-01 3.39382380e-01 -1.43707678e-01
7.95336664e-01 -2.83953160e-01 2.01655522e-01 5.73398530e-01
-3.52489650e-01 3.88747007e-01 7.27142692e-01 -3.13999891e-01
-7.73870945e-01 -9.25196588e-01 1.65012479e-02 1.87902880e+00
-3.56517643e-01 -5.05111337e-01 -8.98769855e-01 -9.29373085e-01
-3.21691304e-01 6.99081242e-01 -5.79403162e-01 -7.58284628e-02
-5.62010765e-01 -6.88555539e-01 9.08608675e-01 7.05431819e-01
6.43041015e-01 -1.51075900e+00 -5.26871085e-01 4.08751369e-01
-1.19952150e-01 -1.39542282e+00 -6.24130249e-01 5.86116076e-01
-2.85831243e-01 -6.78515911e-01 -6.75896645e-01 -9.72744107e-01
1.15238212e-01 -3.26243550e-01 1.71635008e+00 -1.55949607e-01
-6.14342541e-02 5.74286878e-01 -5.56139946e-01 -5.38534760e-01
-9.63117421e-01 4.85501409e-01 -2.19437346e-01 -1.34964079e-01
5.12221932e-01 -2.28101969e-01 -3.11094046e-01 3.11072052e-01
-6.87606812e-01 -2.92924732e-01 3.55170697e-01 1.25687265e+00
-1.91613719e-01 -7.12695062e-01 9.37333167e-01 -9.15701210e-01
9.98005748e-01 -1.94801465e-01 -5.19402206e-01 3.25283527e-01
-9.88136157e-02 2.01513439e-01 4.23379779e-01 -1.14223383e-01
-1.24670756e+00 2.76634097e-01 -8.32961082e-01 -1.23906963e-01
-5.22916019e-01 3.36191505e-01 -1.76264539e-01 2.40646318e-01
3.44108939e-01 2.76691049e-01 1.50982723e-01 -5.12764156e-01
5.90799749e-01 1.11554146e+00 4.34587210e-01 -4.57448810e-01
1.17184669e-01 2.08404120e-02 -7.19823837e-01 -9.86633718e-01
-7.80411005e-01 -5.88460624e-01 -6.13558412e-01 1.24216296e-01
1.03962123e+00 -9.61004972e-01 -1.40744627e+00 3.87641221e-01
-1.32858896e+00 -9.40066576e-01 -1.99361265e-01 -3.29760625e-03
-6.50432765e-01 3.56601000e-01 -1.07707143e+00 -1.04847586e+00
-7.71276414e-01 -1.25705850e+00 1.44867969e+00 4.67542037e-02
-5.28811634e-01 -1.17608607e+00 3.34832102e-01 3.96742404e-01
5.58087289e-01 -4.72262532e-01 7.20907986e-01 -1.18988168e+00
-7.64522478e-02 1.59022689e-01 -3.67786437e-02 8.04946542e-01
1.62022226e-02 -3.65086108e-01 -1.60453355e+00 -4.06360418e-01
1.31490752e-01 -1.27139485e+00 8.23604643e-01 1.48244932e-01
8.78852427e-01 -5.11257410e-01 9.01785642e-02 1.48721367e-01
6.88281119e-01 3.24850172e-01 4.21633512e-01 3.62719223e-02
3.42115313e-01 9.44453716e-01 3.41695905e-01 7.57490337e-01
7.64888644e-01 6.44161344e-01 2.56562382e-01 -7.30092973e-02
2.77547747e-01 1.02971338e-01 5.24909735e-01 9.44002211e-01
4.82683629e-01 -6.50479496e-01 -8.10288846e-01 5.26865005e-01
-1.86531579e+00 -6.67409956e-01 3.92533213e-01 1.73348796e+00
1.29766822e+00 -1.40137061e-01 3.32415968e-01 -1.39386311e-01
2.24821225e-01 4.55800742e-01 -5.85582197e-01 -7.49593437e-01
-2.31691282e-02 6.25181019e-01 1.65943876e-01 8.82408082e-01
-1.33285797e+00 1.36693919e+00 6.78619719e+00 4.37992662e-01
-9.26836848e-01 2.13531360e-01 8.67440701e-01 5.56326546e-02
-4.44895402e-02 -3.44507486e-01 -1.18240368e+00 2.19719365e-01
1.63901722e+00 -1.30512327e-01 2.32954517e-01 8.76688778e-01
-1.30455151e-01 3.16167861e-01 -1.31364155e+00 6.89306378e-01
1.11746475e-01 -1.32762682e+00 -2.46643841e-01 -1.87570766e-01
4.34594989e-01 3.67919147e-01 -2.93028215e-03 8.97253394e-01
1.09097219e+00 -1.43346822e+00 4.01251405e-01 -6.74361736e-02
6.84845209e-01 -5.73564827e-01 9.16126907e-01 1.42515168e-01
-1.14309442e+00 2.75278315e-02 -2.72362858e-01 8.58979970e-02
5.25098264e-01 1.31511893e-02 -1.48411989e+00 2.02018380e-01
7.49845445e-01 7.04071581e-01 -2.13459119e-01 1.95038468e-01
-2.30986714e-01 7.70420372e-01 -1.78434253e-01 -2.44037881e-01
4.83247787e-01 1.55286193e-01 8.43013451e-02 1.61810851e+00
-1.95397839e-01 9.07747373e-02 2.87994802e-01 5.67807674e-01
-4.97020245e-01 -1.45373136e-01 -4.26513553e-01 -1.55259222e-01
5.30726790e-01 1.24757683e+00 -1.11825094e-01 -7.15393305e-01
-6.54370487e-01 1.08938527e+00 4.20574605e-01 3.01481664e-01
-6.62655532e-01 1.16784172e-02 1.01812899e+00 -8.33125949e-01
4.66634303e-01 -1.56531408e-01 7.97390640e-02 -1.21382236e+00
-3.54257941e-01 -1.25370288e+00 3.29703987e-01 -2.78041184e-01
-1.67163634e+00 1.23833609e+00 -1.76101550e-01 -7.70033836e-01
-1.09036183e+00 -8.34427714e-01 -5.63275397e-01 7.65728891e-01
-1.41635180e+00 -1.15493298e+00 -2.49217629e-01 6.14561677e-01
1.22666335e+00 -3.99514884e-01 1.48750353e+00 6.87447861e-02
-4.83004957e-01 7.90210068e-01 -6.47289529e-02 4.66260076e-01
8.04228902e-01 -1.67391145e+00 9.46160674e-01 7.93294236e-02
8.88783261e-02 2.49865025e-01 4.92772341e-01 -1.25537261e-01
-9.14103627e-01 -9.13877428e-01 1.02918470e+00 -7.24626064e-01
7.68374860e-01 -9.88113403e-01 -9.58142042e-01 1.13127530e+00
9.37534451e-01 -4.88064796e-01 8.98115158e-01 2.52735496e-01
-3.51502806e-01 2.47816280e-01 -8.62612069e-01 3.98909152e-01
8.23398352e-01 -7.26214468e-01 -6.81787312e-01 3.50666165e-01
1.08706582e+00 -4.84497756e-01 -1.02170599e+00 3.46766740e-01
3.19152653e-01 -1.05237865e+00 9.43180740e-01 -6.99508369e-01
2.56208926e-01 5.23094773e-01 -2.18747750e-01 -1.44482589e+00
1.90089583e-01 -9.90558922e-01 5.82920574e-02 1.48023450e+00
5.82765996e-01 -6.36893988e-01 8.70249808e-01 7.07436323e-01
-5.30463040e-01 -6.59888446e-01 -1.04147267e+00 -5.15981257e-01
3.44022185e-01 -1.80676412e-02 6.09563053e-01 7.46957004e-01
4.01769310e-01 1.12312508e+00 -5.24513602e-01 -2.31655315e-01
1.18212976e-01 -2.01128069e-02 9.01392877e-01 -1.08289289e+00
-2.73887634e-01 -4.92765099e-01 -2.17477493e-02 -1.62142801e+00
7.19598472e-01 -4.48452175e-01 4.87572312e-01 -1.23250365e+00
-2.10199416e-01 -1.61615402e-01 3.32199894e-02 5.98467886e-01
-1.16831072e-01 -1.71882212e-01 6.96031302e-02 -1.25669479e-01
-7.80181527e-01 1.12706006e+00 7.24765897e-01 -2.89124876e-01
-1.73607007e-01 -6.33445084e-02 -4.46141630e-01 5.19592464e-01
7.82933295e-01 -8.78346190e-02 -4.45050657e-01 -3.53125721e-01
-4.04761314e-01 3.12746227e-01 1.60806924e-01 -6.79067254e-01
1.29204929e-01 2.54937112e-01 -3.02288413e-01 -4.51518536e-01
9.98704016e-01 -3.81408691e-01 -7.67348528e-01 2.06129611e-01
-1.02034974e+00 9.12661627e-02 4.21358734e-01 2.48623833e-01
-5.97057700e-01 -2.85544366e-01 7.61713862e-01 -3.14043581e-01
-7.64818847e-01 5.07496409e-02 -7.04315305e-01 2.78495699e-01
5.47629595e-01 5.86871028e-01 -5.32634377e-01 -1.15634298e+00
-7.84813702e-01 5.12454927e-01 -1.23989038e-01 5.14766455e-01
5.05992413e-01 -7.52259493e-01 -8.29793930e-01 1.31026998e-01
1.50929436e-01 -2.13063866e-01 1.64740890e-01 5.01708448e-01
-1.50179878e-01 7.31673002e-01 -2.25324798e-02 -5.51633239e-01
-1.49734735e+00 5.98051101e-02 4.17096525e-01 -6.22361004e-01
-2.68012315e-01 1.31517172e+00 4.91096526e-01 -9.34561253e-01
7.56565928e-01 -4.94860351e-01 -3.25505972e-01 -5.88178076e-03
5.18080890e-01 -1.47216499e-01 1.19517893e-01 -6.10328317e-01
-3.31939399e-01 -1.11103870e-01 -4.30199146e-01 -5.69695532e-01
1.13829720e+00 -2.74349481e-01 9.53768268e-02 5.83214343e-01
1.23619223e+00 -3.73068482e-01 -1.47551978e+00 -3.69559646e-01
2.17099354e-01 2.44585365e-01 -3.07477236e-01 -9.86605883e-01
-7.57415175e-01 1.23472047e+00 6.40766919e-01 4.05718505e-01
8.17974925e-01 2.53391623e-01 9.52597141e-01 9.12643790e-01
1.33357823e-01 -8.88426542e-01 4.91152853e-01 1.22349322e+00
1.07854950e+00 -1.58189988e+00 -6.62631929e-01 -3.55870098e-01
-1.47441912e+00 7.69723952e-01 7.09071398e-01 7.16575310e-02
4.64398712e-01 4.92175251e-01 5.47174096e-01 -2.36877978e-01
-1.36139536e+00 -4.78788912e-01 -1.63292754e-02 5.09746850e-01
8.94487858e-01 -1.61462724e-01 2.28919879e-01 7.37903059e-01
-4.58362699e-01 -2.10259750e-01 3.64867210e-01 9.53516901e-01
-3.49027961e-01 -1.41507030e+00 1.36212453e-01 1.96072266e-01
-3.98091435e-01 -3.93247485e-01 -1.00163436e+00 7.07087159e-01
-6.23094440e-01 1.30295062e+00 7.47710913e-02 -4.72352833e-01
3.24957401e-01 6.51161015e-01 6.77468404e-02 -8.97335887e-01
-1.14571369e+00 1.26610622e-01 8.29925358e-01 -3.81573617e-01
-4.34392869e-01 -4.00722116e-01 -1.37219346e+00 -7.40142316e-02
7.90791512e-02 5.48102140e-01 4.41464126e-01 9.61185992e-01
4.49289888e-01 5.84556997e-01 7.30047822e-01 -9.22647476e-01
-1.00579667e+00 -1.62823761e+00 -3.21373701e-01 2.82389343e-01
3.50576878e-01 -4.44386244e-01 -3.28259856e-01 1.86445683e-01] | [12.807656288146973, 7.788235187530518] |
d4e062bb-d9d1-48d7-a1ea-ce950ff78cfc | utilizing-automated-breast-cancer-detection | 1905.10841 | null | https://arxiv.org/abs/1905.10841v3 | https://arxiv.org/pdf/1905.10841v3.pdf | Utilizing Automated Breast Cancer Detection to Identify Spatial Distributions of Tumor Infiltrating Lymphocytes in Invasive Breast Cancer | Quantitative assessment of Tumor-TIL spatial relationships is increasingly important in both basic science and clinical aspects of breast cancer research. We have developed and evaluated convolutional neural network (CNN) analysis pipelines to generate combined maps of cancer regions and tumor infiltrating lymphocytes (TILs) in routine diagnostic breast cancer whole slide tissue images (WSIs). We produce interactive whole slide maps that provide 1) insight about the structural patterns and spatial distribution of lymphocytic infiltrates and 2) facilitate improved quantification of TILs. We evaluated both tumor and TIL analyses using three CNN networks - Resnet-34, VGG16 and Inception v4, and demonstrated that the results compared favorably to those obtained by what believe are the best published methods. We have produced open-source tools and generated a public dataset consisting of tumor/TIL maps for 1,015 TCGA breast cancer images. We also present a customized web-based interface that enables easy visualization and interactive exploration of high-resolution combined Tumor-TIL maps for 1,015TCGA invasive breast cancer cases that can be downloaded for further downstream analyses. | ['Jonas S. Almeida', 'ASHISH SHARMA', 'Tahsin Kurc', 'Shahira Abousamra', 'Rebecca Batiste', 'Han Le', 'Erich Bremer', 'Alison L. Van Dyke', 'Danielle Fassler', 'Arvind Rao', 'Tianhao Zhao', 'Rajarsi Gupta', 'Joel Saltz', 'Le Hou', 'Dimitris Samaras'] | 2019-05-26 | null | null | null | null | ['breast-cancer-detection', 'breast-cancer-detection'] | ['knowledge-base', 'medical'] | [ 3.37353408e-01 6.84107393e-02 -2.76920915e-01 -3.27585250e-01
-1.14774477e+00 -3.66920739e-01 5.01176894e-01 9.37508583e-01
-6.32013381e-01 5.92682719e-01 4.04487073e-01 -1.03308320e+00
-1.48405060e-01 -8.56795847e-01 -3.38963509e-01 -9.16526616e-01
-2.59105980e-01 7.98735678e-01 2.28896201e-01 -1.56424940e-01
-8.85633379e-02 8.66034210e-01 -8.59789193e-01 8.37313771e-01
4.91417021e-01 9.21311140e-01 -3.33045460e-02 1.15860975e+00
-3.44806373e-01 6.02305651e-01 -4.05019492e-01 1.45243898e-01
-3.64336163e-01 -2.19510019e-01 -5.84701240e-01 -5.81894815e-01
4.07799691e-01 -1.34947479e-01 -3.32151316e-02 6.92090333e-01
5.23929000e-01 -6.99651361e-01 6.10298455e-01 -8.65158498e-01
2.44586878e-02 3.61825198e-01 -7.01975584e-01 7.16365874e-01
-4.16098505e-01 2.99862146e-01 7.18796670e-01 -6.21492863e-01
1.26658833e+00 8.19560766e-01 9.93844807e-01 2.91047394e-01
-1.40927565e+00 -7.62921572e-01 -7.16021836e-01 -6.30566031e-02
-1.30108070e+00 -1.03710413e-01 7.63838142e-02 -6.62691832e-01
1.08540010e+00 6.57015502e-01 1.01730895e+00 7.57765293e-01
8.40609968e-01 4.94336009e-01 1.21479917e+00 -4.69046503e-01
8.65871161e-02 6.71656290e-03 3.71331424e-01 5.95530927e-01
3.69616479e-01 -1.43734172e-01 -3.17898095e-01 -2.98184633e-01
8.31711769e-01 1.62907973e-01 6.09512851e-02 -6.88713938e-02
-1.36232746e+00 8.82321119e-01 6.52738988e-01 6.55286610e-01
-1.03990264e-01 4.03333604e-01 7.21782386e-01 8.03600773e-02
7.69123673e-01 -3.00307968e-03 -8.78561288e-02 1.56874612e-01
-8.38837683e-01 2.15608820e-01 2.61179507e-01 1.09783508e-01
5.20338953e-01 -4.50675786e-01 -1.81499273e-01 6.16396666e-01
2.43866146e-01 2.98689842e-01 3.85272443e-01 -4.04451698e-01
-1.52075008e-01 8.80530298e-01 -2.11159855e-01 -9.88669991e-01
-9.91210818e-01 -6.02980256e-01 -1.18677175e+00 4.41058606e-01
7.55516887e-01 2.05414146e-01 -1.05137444e+00 1.17839670e+00
3.39896649e-01 -2.76446581e-01 -2.97792852e-01 6.91182315e-01
1.13706779e+00 7.03709607e-04 4.34051782e-01 3.54931027e-01
1.67630613e+00 -3.87908667e-01 -9.13413525e-01 1.67831987e-01
1.57964039e+00 -5.25414228e-01 7.88549602e-01 -2.57755250e-01
-7.91141927e-01 -1.75052565e-02 -9.40324605e-01 -1.31886214e-01
-8.85668397e-01 2.48267770e-01 7.94142425e-01 5.36137223e-01
-1.55037403e+00 3.45942259e-01 -1.23010588e+00 -1.00146103e+00
1.15452182e+00 4.35745478e-01 -7.49164224e-01 1.81503057e-01
-6.86685801e-01 8.35214019e-01 6.50437623e-02 1.20026745e-01
-7.58405685e-01 -1.16699839e+00 -5.71809053e-01 -1.92044750e-01
-2.57475615e-01 -5.54870486e-01 1.07198322e+00 -7.48158634e-01
-8.60400975e-01 1.28819311e+00 -1.86432466e-01 -3.74037206e-01
4.05652791e-01 6.11069918e-01 -3.16585511e-01 2.02660248e-01
6.40544742e-02 7.65658677e-01 -4.43963826e-01 -7.55009413e-01
-7.99354672e-01 -6.03382409e-01 -5.61777472e-01 -1.14811637e-01
-2.05628857e-01 -3.14090401e-02 -2.39547625e-01 -3.88469517e-01
-1.19899008e-02 -6.67225480e-01 -4.08518642e-01 7.62967169e-01
-4.37746257e-01 2.10049778e-01 8.12709391e-01 -7.66600311e-01
8.20224404e-01 -1.86781347e+00 -3.56562883e-01 4.32848126e-01
5.23042858e-01 1.66080728e-01 -1.43120363e-01 2.26053491e-01
-3.37118566e-01 5.94483793e-01 7.40773156e-02 -5.06225862e-02
-2.14213535e-01 -1.14815339e-01 3.07735056e-01 8.76132369e-01
1.45533487e-01 1.45553613e+00 -6.81098461e-01 -6.38418198e-01
2.07660228e-01 6.16638303e-01 -2.21092150e-01 -5.26489057e-02
-1.20934941e-01 4.59621519e-01 -2.61770636e-01 8.19385648e-01
7.75053263e-01 -5.17137885e-01 5.17075062e-01 -2.67039984e-01
-7.31264427e-02 -1.10981483e-02 -2.55614281e-01 1.30749667e+00
-2.02326670e-01 1.22578502e+00 4.35558021e-01 -4.31134820e-01
6.88287795e-01 4.86969575e-02 4.36408639e-01 -7.81582296e-01
3.80378574e-01 9.98345912e-02 1.66517824e-01 -2.83313096e-01
1.26293376e-01 -2.93655008e-01 2.12340087e-01 6.66242421e-01
-8.80968794e-02 1.94194898e-01 1.50667176e-01 2.19913259e-01
1.48690522e+00 -4.45641398e-01 4.43731964e-01 -7.73070574e-01
1.20161943e-01 7.06614733e-01 1.43237606e-01 4.67732459e-01
-2.76743740e-01 5.42500794e-01 1.21026909e+00 -9.17642295e-01
-1.11665988e+00 -1.06758726e+00 -3.92686099e-01 9.52985764e-01
-5.45235038e-01 -5.56955673e-02 -2.01635316e-01 -4.71540064e-01
6.20379709e-02 2.29086146e-01 -1.53754997e+00 5.67992449e-01
-5.83725512e-01 -1.18412340e+00 9.57227111e-01 6.65136516e-01
1.71976835e-01 -7.97151923e-01 -4.78470653e-01 -1.83808114e-02
3.11135873e-02 -4.19672668e-01 3.16977724e-02 5.66267431e-01
-6.40748441e-01 -1.40345335e+00 -7.86288977e-01 -8.15482676e-01
1.07402289e+00 -1.03708945e-01 1.14274728e+00 3.80619079e-01
-1.17184460e+00 -2.31053665e-01 -5.82214482e-02 -8.26976717e-01
-6.88383758e-01 1.88730061e-01 -6.44402444e-01 -5.02894044e-01
6.26178503e-01 -2.08342507e-01 -7.54167140e-01 7.94396922e-02
-9.61535692e-01 3.93732935e-01 8.25610995e-01 9.12729979e-01
9.55527186e-01 -3.13693821e-01 3.47564220e-01 -1.19309378e+00
4.76696044e-01 -5.19726992e-01 -4.85819370e-01 6.17313832e-02
-1.90142602e-01 -4.97278839e-01 2.08881751e-01 1.02429636e-01
-8.80643964e-01 -9.72427502e-02 -4.66904640e-01 2.79806793e-01
-4.13811356e-01 7.49250233e-01 3.96871597e-01 -1.86946049e-01
8.43725741e-01 -1.07724003e-01 6.71242177e-01 5.63855283e-02
-1.09107964e-01 5.94576120e-01 4.55709845e-01 1.84476361e-01
2.43383050e-01 1.04927766e+00 3.35069537e-01 -7.65327394e-01
-2.90514171e-01 -5.19597173e-01 -5.85468173e-01 -2.59505719e-01
1.05804098e+00 -5.37151814e-01 -1.08137584e+00 5.38173258e-01
-5.51601350e-01 -9.52972889e-01 -1.95198298e-01 9.73908305e-02
-2.08979845e-01 -2.31684640e-01 -9.66025770e-01 -1.66170716e-01
-6.33530974e-01 -1.00245416e+00 1.20511055e+00 1.67171299e-01
-5.78519523e-01 -1.50188911e+00 6.15560293e-01 -1.00605078e-01
1.10487866e+00 9.37471986e-01 1.28161156e+00 -8.38823140e-01
-3.40828419e-01 -4.90968049e-01 -6.56183481e-01 -4.97073382e-01
1.31267369e-01 3.88071895e-01 -8.68815958e-01 -2.37970620e-01
-9.95009959e-01 -1.44876704e-01 1.06124187e+00 7.42015123e-01
1.07258391e+00 -7.88002461e-02 -1.15201890e+00 8.10346842e-01
1.56330764e+00 1.33552969e-01 7.79994845e-01 5.04549742e-01
4.04931068e-01 7.44663119e-01 2.71018267e-01 1.06736936e-01
-7.89829437e-03 3.41748416e-01 4.01258051e-01 -1.10810220e+00
-1.40785098e-01 1.64187863e-01 -6.18083239e-01 1.03081092e-01
1.37768552e-01 -2.95452565e-01 -1.39668310e+00 7.64132798e-01
-1.37908053e+00 -5.17090023e-01 -3.65631610e-01 1.57051969e+00
6.92596495e-01 -1.52632687e-02 -6.89099953e-02 -2.92010278e-01
5.29681087e-01 6.64701089e-02 -4.95999247e-01 -2.18616903e-01
-2.00823680e-01 4.51916665e-01 5.80532193e-01 3.39817077e-01
-9.39710736e-01 4.63672191e-01 7.45932817e+00 9.30589020e-01
-1.44082856e+00 2.31154025e-01 1.55073714e+00 -3.38606894e-01
-4.09162492e-01 -3.64907861e-01 -4.86075193e-01 -8.56753662e-02
9.62544739e-01 -3.12310830e-02 -3.79432887e-01 4.15014744e-01
2.60422558e-01 -3.88032585e-01 -1.04450870e+00 4.45284814e-01
-4.82564449e-01 -2.33621883e+00 -1.14089929e-01 5.82983851e-01
3.66450995e-01 4.36762363e-01 8.04800317e-02 -6.21397421e-02
6.27618611e-01 -1.54373002e+00 -1.93624899e-01 7.72654831e-01
1.58995068e+00 -7.05273211e-01 1.53137779e+00 -2.02075034e-01
-8.61935735e-01 3.84165794e-01 -9.00286958e-02 5.07760942e-01
-3.38587224e-01 5.05495787e-01 -1.75095427e+00 1.67745128e-01
8.26887548e-01 4.94507343e-01 -9.62807596e-01 6.03558362e-01
6.58706665e-01 5.29222190e-01 -2.17059210e-01 -3.35881084e-01
1.94378626e-02 4.77438778e-01 -4.95149866e-02 1.66144729e+00
2.63545245e-01 -9.41988826e-02 -4.90708262e-01 8.59367013e-01
3.48861724e-01 2.90248126e-01 -1.96440414e-01 -2.61942685e-01
2.24169672e-01 2.06102395e+00 -1.46030903e+00 -3.10084850e-01
-1.05057716e-01 7.55059943e-02 2.46427581e-01 1.80272654e-01
-4.77165580e-01 -4.75910723e-01 6.16689026e-01 6.67024672e-01
-1.60303861e-01 1.43885478e-01 -5.62486589e-01 -3.34077179e-01
-8.13937366e-01 -5.66548705e-01 6.98076844e-01 -7.33043373e-01
-1.07370305e+00 6.51728868e-01 -2.72669256e-01 -7.22624779e-01
-1.21400043e-01 -8.58327925e-01 -9.58007276e-01 9.48132873e-01
-1.32611060e+00 -1.38552594e+00 -7.69749343e-01 7.59146065e-02
-4.36424576e-02 2.53425799e-02 1.22038901e+00 -1.25456974e-01
-5.10584533e-01 6.29852712e-01 2.63606042e-01 2.69419998e-01
5.97792506e-01 -1.30654764e+00 3.15717191e-01 1.68196356e-03
-7.55628109e-01 4.96813178e-01 4.47766721e-01 -6.86957300e-01
-1.11803246e+00 -1.40555966e+00 6.11049891e-01 -3.84448141e-01
9.34596598e-01 -4.30123270e-01 -6.98024035e-01 9.70838368e-01
4.40876991e-01 2.54713327e-01 1.23967946e+00 -7.38595352e-02
1.30478337e-01 -2.01625545e-02 -1.30657601e+00 7.35654414e-01
3.64323050e-01 -1.89383805e-01 3.70293707e-01 4.87378508e-01
1.80119425e-01 -6.23037338e-01 -1.19535303e+00 4.09174889e-01
7.99890757e-01 -1.12967312e+00 7.25578427e-01 -3.37020636e-01
3.63149822e-01 -1.57150537e-01 7.85075203e-02 -1.20542908e+00
-5.13542235e-01 1.77932605e-01 6.77439392e-01 5.29955268e-01
6.58974886e-01 -9.77874696e-01 1.09790909e+00 1.24740355e-01
-2.95705050e-01 -1.22162151e+00 -1.08740032e+00 3.07223555e-02
4.49033648e-01 -6.55153692e-02 6.44317269e-01 6.62584364e-01
2.73113400e-01 -3.94636989e-01 8.16787899e-01 -2.44961023e-01
3.45822930e-01 -3.73838753e-01 7.74870574e-01 -9.60984230e-01
1.64407507e-01 -8.64086092e-01 -6.85857832e-01 3.35797220e-02
-1.80751830e-01 -1.09998548e+00 -2.84627646e-01 -1.69091165e+00
6.59799397e-01 -6.45380795e-01 -4.34004545e-01 8.58702898e-01
1.91209137e-01 8.44073236e-01 -5.18365264e-01 2.37531587e-01
-4.39756513e-01 -2.41481677e-01 1.26266694e+00 -3.10544997e-01
1.06326774e-01 -6.94749832e-01 -7.79121339e-01 4.47667211e-01
7.00882435e-01 -2.60136753e-01 2.29960993e-01 7.33931130e-03
2.08382964e-01 2.07260624e-01 6.11951947e-01 -1.00084269e+00
2.05867738e-01 -1.49262041e-01 1.10799527e+00 -1.28297496e+00
8.83530173e-03 -3.40312988e-01 4.89242494e-01 7.57991135e-01
-7.00334966e-01 -1.85018525e-01 7.10073411e-01 2.02904403e-01
-4.16440032e-02 3.97243887e-01 7.63797700e-01 -1.65169120e-01
-9.74770114e-02 2.14384988e-01 -9.76035118e-01 -7.26367414e-01
1.11941636e+00 -4.88325417e-01 -1.11547661e+00 -1.44885272e-01
-8.74711990e-01 1.64252743e-01 6.45480096e-01 -1.75822899e-01
3.24329406e-01 -1.26389873e+00 -9.79611933e-01 2.78112710e-01
5.25938332e-01 -3.65062580e-02 8.14712405e-01 1.32695115e+00
-1.37922025e+00 7.94491053e-01 -5.57257116e-01 -9.17906702e-01
-1.47478914e+00 5.88668063e-02 6.82274044e-01 -7.70295918e-01
-5.26085854e-01 9.83435869e-01 6.14731252e-01 -5.24253309e-01
-9.95144024e-02 -3.12202960e-01 -2.79325604e-01 1.12131462e-02
7.68501878e-01 1.52456596e-01 3.14098895e-01 -5.98061442e-01
-4.75326270e-01 8.35498720e-02 -4.66699958e-01 1.46852508e-01
1.36868227e+00 3.18698794e-01 -5.17725408e-01 4.31560367e-01
1.40994918e+00 -1.60180688e-01 -6.94407821e-01 -3.91584588e-03
-1.84580177e-01 -1.58981174e-01 5.59967309e-02 -1.02338493e+00
-1.11238897e+00 6.65954173e-01 1.12944913e+00 5.08826487e-02
9.34617281e-01 1.87407315e-01 4.25990760e-01 -8.56252760e-02
-1.41651124e-01 -6.16018891e-01 -7.32104108e-03 1.29486904e-01
5.70622504e-01 -1.06745481e+00 -1.72980875e-02 -4.51193005e-01
-1.32946344e-02 1.43523037e+00 5.39208472e-01 -1.08990744e-01
8.19975853e-01 1.01511979e+00 7.03198135e-01 -7.53810048e-01
-1.13050413e+00 1.11998312e-01 3.89332771e-02 6.95709944e-01
9.43377316e-01 2.52674282e-01 -6.54653609e-02 5.27686536e-01
-2.39025086e-01 1.05738357e-01 3.40646774e-01 9.61211205e-01
-3.95178258e-01 -7.17232764e-01 -3.39850992e-01 9.82999206e-01
-6.95678830e-01 -3.09051462e-02 -5.69622576e-01 1.28215098e+00
-1.17933065e-01 2.66695231e-01 7.66350389e-01 -1.21456973e-01
-1.82294503e-01 -2.25945890e-01 4.40900065e-02 -3.50347608e-01
-7.23330796e-01 2.26075813e-01 2.87689030e-01 -4.45006430e-01
-1.18302200e-02 -4.56395090e-01 -1.17881072e+00 -4.45201784e-01
-1.63994461e-01 -2.12245896e-01 9.12621915e-01 5.19059658e-01
4.74745154e-01 1.04325342e+00 -1.01114616e-01 -7.15605855e-01
4.95908916e-01 -1.21695328e+00 -7.90981412e-01 -1.20587498e-01
3.63153189e-01 -1.66091621e-01 -3.25866014e-01 -1.11444153e-01] | [15.134242057800293, -3.089184522628784] |
69b502ae-bd99-4e25-97c7-9d08f128c05b | knowledge-base-relation-detection-via-multi | 1803.00612 | null | http://arxiv.org/abs/1803.00612v2 | http://arxiv.org/pdf/1803.00612v2.pdf | Knowledge Base Relation Detection via Multi-View Matching | Relation detection is a core component for Knowledge Base Question Answering
(KBQA). In this paper, we propose a KB relation detection model via multi-view
matching which utilizes more useful information extracted from question and KB.
The matching inside each view is through multiple perspectives to compare two
input texts thoroughly. All these components are designed in an end-to-end
trainable neural network model. Experiments on SimpleQuestions and WebQSP yield
state-of-the-art results. | ['Wei zhang', 'Yang Yu', 'Mo Yu', 'Zhiguo Wang', 'Kazi Saidul Hasan'] | 2018-03-01 | null | null | null | null | ['knowledge-base-question-answering'] | ['natural-language-processing'] | [-2.91173160e-01 4.72503871e-01 -2.39417955e-01 -3.14826936e-01
-1.32717335e+00 -6.32745743e-01 2.92324811e-01 2.68423587e-01
-2.00887740e-01 7.35687077e-01 3.19259018e-01 -4.58916008e-01
-1.60531268e-01 -1.07315767e+00 -6.57292128e-01 3.81521322e-02
3.61085594e-01 8.70343208e-01 9.49591637e-01 -8.96808803e-01
-3.77023332e-02 1.50524378e-02 -1.12450945e+00 1.24568486e+00
6.83127582e-01 1.09672201e+00 -7.85082653e-02 1.06528699e+00
-5.36139607e-01 1.43779337e+00 -5.93886137e-01 -1.11803472e+00
-1.92076385e-01 -1.86589375e-01 -1.78341675e+00 -4.48460490e-01
7.75957942e-01 -2.58019030e-01 -3.98500949e-01 7.80238867e-01
6.30813479e-01 1.24057107e-01 4.19998974e-01 -1.03298151e+00
-9.96965230e-01 7.27955401e-01 -8.16149786e-02 6.52020633e-01
1.17362964e+00 -2.72976965e-01 1.61847591e+00 -9.50790465e-01
7.21037090e-01 1.25560236e+00 3.64725471e-01 4.27859455e-01
-6.53227031e-01 -1.85656488e-01 1.00194179e-01 1.11584854e+00
-9.31349874e-01 -4.03702259e-01 7.90794075e-01 -1.21154345e-03
1.57060719e+00 5.71254015e-01 3.68685126e-01 8.22161138e-01
8.85477290e-02 1.17378759e+00 8.38276923e-01 -5.61418712e-01
-2.34340817e-01 3.64841193e-01 8.50217879e-01 8.34614813e-01
7.01872259e-02 -6.70891821e-01 -8.64640296e-01 -4.82928246e-01
1.84227690e-01 -4.78964478e-01 -6.51551008e-01 -2.80451685e-01
-9.95499909e-01 7.32899427e-01 3.07915956e-01 1.30915076e-01
-1.72767252e-01 -1.14270851e-01 5.29218495e-01 7.81683445e-01
4.03207034e-01 5.47654152e-01 -8.94780219e-01 -8.39518234e-02
-1.93257824e-01 3.26181471e-01 1.31283164e+00 1.05790520e+00
6.37730062e-01 -7.55525172e-01 -2.81789362e-01 8.65522087e-01
4.20654088e-01 4.09697741e-01 3.54711652e-01 -6.69556499e-01
1.29290438e+00 1.29781890e+00 -1.34594977e-01 -1.04205632e+00
-4.56398249e-01 -1.42949209e-01 -5.42057753e-01 -5.94001472e-01
1.72581345e-01 1.45891622e-01 -5.38000464e-01 1.08622503e+00
7.63105750e-01 -1.54312521e-01 7.21268117e-01 6.98286951e-01
1.89278424e+00 6.25391662e-01 -1.18691005e-01 -3.19017768e-02
1.97325432e+00 -1.55218792e+00 -1.03812468e+00 -1.95018500e-01
9.31259096e-01 -5.43169081e-01 1.14976501e+00 2.03173339e-01
-1.10868943e+00 -4.77361679e-01 -9.26345825e-01 -6.86685383e-01
-6.86934352e-01 -1.42122477e-01 2.89244264e-01 3.23514402e-01
-9.10986125e-01 -4.12861183e-02 -4.20522958e-01 -4.03461069e-01
1.81218386e-01 1.74736202e-01 -6.03546917e-01 -1.07541665e-01
-1.82929146e+00 1.23486757e+00 7.29322076e-01 4.56403568e-02
-4.07774597e-01 -6.67047322e-01 -8.61457646e-01 1.56130999e-01
9.56023216e-01 -1.30011308e+00 1.67844689e+00 -2.21005499e-01
-1.60627341e+00 8.82936418e-01 -2.53847718e-01 -3.33243161e-01
1.37206726e-02 -6.27996147e-01 -8.83306444e-01 4.42097366e-01
8.19754079e-02 -1.45342853e-02 3.83581519e-01 -1.18737364e+00
-4.81028974e-01 -5.64461708e-01 5.02807140e-01 5.53760707e-01
-1.33257464e-01 3.17390561e-01 -7.19005644e-01 -1.75262600e-01
6.83622435e-02 -3.82779062e-01 2.68007189e-01 -3.91902208e-01
-3.88690233e-01 -6.77005827e-01 8.08373570e-01 -1.28793478e+00
1.28892207e+00 -1.36102295e+00 2.24270448e-01 -1.54829815e-01
4.72611874e-01 3.68817657e-01 1.90853197e-02 8.57995152e-01
9.20577720e-02 -1.40698642e-01 8.65433142e-02 2.19749853e-01
-1.13502398e-01 2.02321280e-02 -4.13939506e-01 -1.36602715e-01
3.01775843e-01 1.45948291e+00 -9.59867597e-01 -7.48493910e-01
-3.61405909e-01 -1.15805097e-01 -1.70626834e-01 6.66347444e-01
-5.55958569e-01 -2.17115253e-01 -5.67956150e-01 7.54591346e-01
4.05269712e-01 -6.82174981e-01 2.37608001e-01 -4.10061926e-01
7.17307270e-01 8.66452456e-01 -1.08240724e+00 1.75188732e+00
-3.58008683e-01 4.56637859e-01 -2.46870667e-01 -8.18092823e-01
9.11314607e-01 7.31096208e-01 -2.79551238e-01 -7.45095611e-01
-1.19329490e-01 1.19627610e-01 -8.48724023e-02 -1.13908589e+00
6.93058848e-01 -4.07943688e-02 9.15841982e-02 5.73123455e-01
5.32856941e-01 -1.56826407e-01 1.81193531e-01 5.79818130e-01
1.06991613e+00 -5.44330431e-03 9.45980906e-01 7.98440948e-02
9.90212798e-01 1.30368397e-01 -8.73848349e-02 5.13411582e-01
-9.77372658e-03 3.09224665e-01 6.66310787e-01 -4.59830135e-01
-3.65339011e-01 -1.09015298e+00 3.02449465e-01 1.24479330e+00
1.86294511e-01 -5.22830248e-01 -3.18455249e-01 -1.44786727e+00
-7.47560561e-02 7.68947065e-01 -7.08686054e-01 -2.48270392e-01
-7.28743196e-01 -1.32354885e-01 6.31869435e-01 8.02427709e-01
6.91093147e-01 -1.15581870e+00 -3.59647393e-01 3.31053287e-01
-7.02844918e-01 -1.32781327e+00 -1.00616418e-01 -5.76719157e-02
-6.96202338e-01 -1.46288347e+00 -3.10428798e-01 -7.64366448e-01
4.48943637e-02 4.70688254e-01 1.89733768e+00 9.09420103e-02
1.49644196e-01 5.58847427e-01 -7.05031097e-01 -4.51912075e-01
-3.20583940e-01 4.48250175e-01 -4.17696416e-01 -3.48611861e-01
6.80666327e-01 -4.58191723e-01 -3.90419364e-01 1.77267849e-01
-6.60442591e-01 -8.92955437e-02 4.24194485e-01 5.89504480e-01
3.61315876e-01 -5.10198176e-01 7.31895149e-01 -9.73025560e-01
1.13828731e+00 -5.48744142e-01 -3.37102890e-01 1.30871260e+00
-3.60438317e-01 9.24848467e-02 3.99487793e-01 -1.15288764e-01
-1.26809132e+00 -6.18996918e-01 -2.41735145e-01 -4.39089537e-03
-1.73208341e-01 9.15213108e-01 -5.38190782e-01 5.13043590e-02
7.13210106e-01 3.51804107e-01 -2.60838836e-01 -5.45043647e-01
9.45345223e-01 8.35528612e-01 5.14083445e-01 -5.36206484e-01
4.82053965e-01 2.22504631e-01 -5.31282604e-01 -3.13584864e-01
-1.34579432e+00 -9.56809521e-01 -7.36283302e-01 -2.47241825e-01
7.69053459e-01 -8.80937099e-01 -1.07649779e+00 6.25122488e-02
-1.45726919e+00 1.29162326e-01 -9.53350812e-02 1.20020941e-01
-3.83068293e-01 3.59621853e-01 -6.28679872e-01 -6.41994894e-01
-9.79327440e-01 -6.53367162e-01 1.06761980e+00 4.01177943e-01
-5.59629314e-02 -9.59377825e-01 7.61084080e-01 1.33026850e+00
8.81234184e-02 -4.23685670e-01 9.33191717e-01 -1.38508058e+00
-5.40021777e-01 -1.00691393e-01 -3.33341211e-01 2.14966521e-01
9.35968459e-02 -1.82288677e-01 -1.05698919e+00 1.08989112e-01
-2.92498735e-03 -9.61981952e-01 9.34945941e-01 -4.62026328e-01
7.35600591e-01 -4.77119923e-01 -3.52980673e-01 1.77511096e-01
1.27958906e+00 1.97827667e-01 6.22188151e-01 6.13732636e-01
8.02833498e-01 8.20941389e-01 4.74005938e-01 -2.36307189e-01
1.09964073e+00 6.44611537e-01 2.91251510e-01 2.45180830e-01
-3.99603024e-02 -8.84594619e-02 7.39185885e-02 1.28708744e+00
2.19996348e-01 -5.24422586e-01 -1.33011055e+00 6.11623287e-01
-2.00272131e+00 -1.14632320e+00 -2.74764061e-01 1.35598004e+00
1.10154784e+00 5.15535995e-02 -4.37113345e-02 -1.29803985e-01
2.58389235e-01 2.22588331e-01 -3.68872553e-01 -5.17128646e-01
-1.33231863e-01 1.58548161e-01 -2.44500458e-01 6.03688359e-01
-1.13853538e+00 1.02231240e+00 6.21880102e+00 7.35350311e-01
-4.89685833e-01 9.91932452e-02 4.40287292e-01 3.88109863e-01
-4.98898268e-01 -1.67051673e-01 -9.70735073e-01 -3.08778495e-01
7.55146503e-01 -1.32286429e-01 -1.69929042e-02 8.80647779e-01
-8.14253986e-01 -2.06059828e-01 -9.97185826e-01 7.32334912e-01
5.55273712e-01 -1.63086712e+00 4.00315583e-01 -6.11896396e-01
2.08926857e-01 -5.93790263e-02 -3.22154105e-01 7.06255674e-01
2.40927443e-01 -8.00036430e-01 8.20538774e-02 6.49064958e-01
2.38015980e-01 -7.11151361e-01 1.05739689e+00 5.04703641e-01
-1.26339042e+00 2.77065247e-01 -3.89114618e-01 4.04051319e-02
2.03763530e-01 3.00287485e-01 -1.23726749e+00 1.43852270e+00
8.19601536e-01 3.34370762e-01 -8.22967052e-01 5.25873780e-01
-4.98300791e-01 5.65699458e-01 -1.49190156e-02 -3.27689826e-01
-7.06482753e-02 1.77284721e-02 3.16364706e-01 1.12052751e+00
-3.11812192e-01 4.23468620e-01 -2.21040726e-01 5.75780272e-01
-3.57013851e-01 3.62472147e-01 -6.34455204e-01 1.53960571e-01
3.01768452e-01 1.40414977e+00 -2.70837933e-01 -6.59986854e-01
-8.72441947e-01 8.26469123e-01 1.18513119e+00 2.29093626e-01
-4.70177889e-01 -5.09446144e-01 2.96142280e-01 -5.30422926e-01
3.67033750e-01 2.18347237e-01 2.60707643e-02 -1.46798801e+00
4.90794361e-01 -1.17224669e+00 1.06545293e+00 -9.04151499e-01
-1.71528184e+00 7.70970583e-01 1.18703119e-01 -7.86511958e-01
-2.66564488e-01 -6.74506843e-01 -4.39305454e-01 8.19966376e-01
-1.82000446e+00 -1.27815950e+00 -3.59158933e-01 7.43131042e-01
5.08158684e-01 -1.69004560e-01 1.04695129e+00 1.01266786e-01
-4.46672738e-01 5.59908330e-01 -1.93433389e-01 5.50753772e-01
7.50348210e-01 -1.40622365e+00 6.17664158e-01 5.36569178e-01
6.40766144e-01 5.91562927e-01 2.64698267e-01 -7.27529466e-01
-1.49979448e+00 -7.47731745e-01 1.13591313e+00 -1.07683337e+00
7.94763625e-01 -1.28120273e-01 -1.39341688e+00 8.70239437e-01
7.40501821e-01 4.37314473e-02 1.13224649e+00 5.23522258e-01
-8.89766276e-01 -1.53566882e-01 -8.79048109e-01 4.85872388e-01
7.26261199e-01 -9.95800555e-01 -1.57491136e+00 3.58598113e-01
1.30294013e+00 -6.65825486e-01 -1.15901244e+00 7.35109389e-01
4.24698591e-01 -8.16981733e-01 1.16223383e+00 -1.39565039e+00
4.28289682e-01 -2.15077221e-01 -2.43929207e-01 -1.19095051e+00
-1.32964835e-01 -3.48163217e-01 -1.05102801e+00 1.22774494e+00
8.39070082e-01 -4.95196521e-01 7.81193197e-01 4.64044869e-01
4.56449715e-03 -1.14874327e+00 -1.10035360e+00 -5.65962672e-01
-2.28521600e-01 -3.04143131e-01 6.16382480e-01 1.17208815e+00
6.81032121e-01 1.26226258e+00 7.98664689e-02 3.43812525e-01
3.44883017e-02 5.37884593e-01 8.20856690e-01 -9.38698292e-01
-4.91942346e-01 -2.19996706e-01 -1.87335834e-01 -1.09040177e+00
1.33669361e-01 -7.56520629e-01 -2.82476693e-01 -2.11081409e+00
5.24327397e-01 2.99287647e-01 -2.25349620e-01 2.28983581e-01
-9.72924709e-01 -2.95342892e-01 6.31435812e-02 6.38875589e-02
-1.17265749e+00 6.48145258e-01 1.26745725e+00 -3.24591309e-01
2.12996006e-02 -6.46636412e-02 -6.29382253e-01 5.74503779e-01
8.76751661e-01 -3.47792387e-01 -6.34807289e-01 -4.77956802e-01
7.94981658e-01 5.49875855e-01 6.60455972e-02 -6.24839067e-01
4.93236512e-01 -1.75645962e-01 6.32990450e-02 -9.33101952e-01
5.51422179e-01 -7.31963694e-01 -4.99586493e-01 1.76683828e-01
-5.75811684e-01 4.12546456e-01 1.57813907e-01 6.33865595e-01
-6.12558126e-01 -1.73477545e-01 3.03760283e-02 -3.41084301e-01
-6.92994237e-01 2.07560305e-02 -6.91227540e-02 7.30556965e-01
7.82678545e-01 1.60001546e-01 -1.05228579e+00 -5.82432091e-01
-4.90160793e-01 5.61922908e-01 -4.44665790e-01 4.56931949e-01
8.39465857e-01 -1.26465309e+00 -7.52957940e-01 -3.11887354e-01
7.27783978e-01 1.78759202e-01 2.20037848e-01 5.55423379e-01
-5.79556704e-01 8.31171453e-01 1.31700635e-01 -1.83089003e-01
-1.54433382e+00 7.17621684e-01 4.35456932e-01 -9.15262580e-01
-3.64840657e-01 1.37050056e+00 -2.78018862e-01 -9.66081560e-01
1.81805596e-01 -3.77600610e-01 -9.37482238e-01 2.71464318e-01
7.62986779e-01 3.34719121e-01 4.65296149e-01 -2.85537839e-01
-4.23843086e-01 2.08116934e-01 -4.75284308e-01 -9.65324491e-02
1.08048439e+00 -4.19932194e-02 -4.20275480e-01 6.10167325e-01
1.19208837e+00 -1.71859637e-01 -1.79707319e-01 -6.66722059e-01
4.32189018e-01 -2.91729420e-02 -4.29758161e-01 -1.06144524e+00
-5.79305708e-01 8.56242418e-01 2.02160701e-01 5.15375316e-01
8.21957350e-01 4.82928991e-01 1.10935688e+00 1.41547287e+00
8.97665769e-02 -9.41249549e-01 4.61390465e-01 7.75273681e-01
1.19783902e+00 -1.36015224e+00 5.44360839e-02 -6.21445537e-01
-6.62562490e-01 1.27367795e+00 1.24391294e+00 3.17557931e-01
6.23018086e-01 -6.81129992e-02 4.74632293e-01 -8.77421916e-01
-1.27245140e+00 -3.17715019e-01 6.90660298e-01 5.15707672e-01
3.77056390e-01 -2.03834832e-01 -1.78844661e-01 8.29258561e-01
-2.13686362e-01 -2.19413638e-01 1.82759479e-01 1.13973904e+00
-6.40220940e-01 -9.51385319e-01 -1.70376375e-01 4.72400755e-01
-4.04798627e-01 -4.23093438e-01 -8.47650647e-01 9.31683838e-01
-5.45395613e-01 9.99389648e-01 -2.92215765e-01 -5.20262599e-01
7.79526055e-01 6.65097892e-01 4.57361996e-01 -6.89758778e-01
-1.01659048e+00 -6.27841115e-01 8.34692895e-01 -5.98163366e-01
-5.06213546e-01 -9.75385085e-02 -1.01943648e+00 3.20955142e-02
-7.50835478e-01 3.34098548e-01 6.53051138e-02 1.23048866e+00
4.99092102e-01 5.39583385e-01 2.72051513e-01 4.59166944e-01
-5.22869527e-01 -1.19790769e+00 -1.53365180e-01 3.72275174e-01
2.06083849e-01 -3.14292252e-01 7.38918483e-02 -5.46782129e-02] | [10.61073112487793, 7.932063102722168] |
5a9cec87-c7a5-4642-a122-bd422e677959 | cmb-ai-lab-at-semeval-2022-task-11-a-two | null | null | https://aclanthology.org/2022.semeval-1.221 | https://aclanthology.org/2022.semeval-1.221.pdf | CMB AI Lab at SemEval-2022 Task 11: A Two-Stage Approach for Complex Named Entity Recognition via Span Boundary Detection and Span Classification | This paper presents a solution for the SemEval-2022 Task 11 Multilingual Complex Named Entity Recognition. What is challenging in this task is detecting semantically ambiguous and complex entities in short and low-context settings. Our team (CMB AI Lab) propose a two-stage method to recognize the named entities: first, a model based on biaffine layer is built to predict span boundaries, and then a span classification model based on pooling layer is built to predict semantic tags of the spans. The basic pre-trained models we choose are XLM-RoBERTa and mT5. The evaluation result of our approach achieves an F1 score of 84.62 on sub-task 13, which ranks the third on the learder board. | ['Yaohan He', 'Wenyi Lv', 'Jiangzhou Ji', 'Yixiao Yang', 'Hongyi Liu', 'Keyu Pu'] | null | null | null | null | semeval-naacl-2022-7 | ['boundary-detection'] | ['computer-vision'] | [-3.93938631e-01 3.22417796e-01 -5.14363609e-02 -6.38879359e-01
-1.13451076e+00 -7.92208135e-01 3.96555126e-01 1.13129899e-01
-9.55020905e-01 9.71446872e-01 5.11329830e-01 -2.83228427e-01
4.71786141e-01 -2.89834976e-01 -7.17202067e-01 1.48697734e-01
-1.71621636e-01 5.11038840e-01 2.26180196e-01 -5.51170334e-02
3.92278545e-02 5.08457184e-01 -8.21571827e-01 6.24539196e-01
1.10691762e+00 1.13851190e+00 2.60585826e-03 5.54630101e-01
-4.93289322e-01 9.42285180e-01 -6.66020751e-01 -6.41235530e-01
-5.48859350e-02 -1.33732501e-02 -1.49751997e+00 -4.34534043e-01
7.61869431e-01 1.56349182e-01 -4.02530059e-02 7.70414293e-01
6.87544584e-01 2.88504958e-01 6.46300256e-01 -8.98970246e-01
-5.44233799e-01 1.25543773e+00 -7.43803382e-02 4.87427473e-01
4.20760185e-01 -2.64418095e-01 1.39080715e+00 -1.43781281e+00
1.00325954e+00 1.11370373e+00 9.76328969e-01 4.38111961e-01
-1.09258223e+00 -6.46286190e-01 2.63786435e-01 4.06833977e-01
-1.69891882e+00 -7.38546610e-01 1.00515999e-01 -3.05260450e-01
1.52019978e+00 -1.51137188e-02 2.05047980e-01 1.05244553e+00
8.33616257e-02 1.16721427e+00 1.00895560e+00 -3.75289917e-01
1.27137169e-01 -1.09293237e-01 4.49740618e-01 5.74230671e-01
3.36458266e-01 -2.25800067e-01 -6.74612463e-01 1.14347763e-01
2.46691242e-01 -5.75253427e-01 -1.03070475e-01 2.83016503e-01
-1.16564155e+00 4.93269622e-01 5.74249148e-01 6.44637346e-01
-6.20103836e-01 -7.15568140e-02 5.44661880e-01 2.29560778e-01
3.30669940e-01 9.32295024e-01 -9.97792840e-01 -2.49682162e-02
-1.30029094e+00 1.77412599e-01 1.09207976e+00 1.27701652e+00
4.89845842e-01 -1.59915134e-01 -5.71149945e-01 7.74922132e-01
3.53311300e-01 2.54279941e-01 2.78622240e-01 -4.56928223e-01
8.10526609e-01 4.34493423e-01 3.27792674e-01 -4.66141552e-01
-9.50416386e-01 -4.11528289e-01 -3.35517198e-01 -3.39632869e-01
4.73161906e-01 -5.02222180e-01 -1.13139844e+00 1.72655606e+00
2.40979120e-01 -2.74910145e-02 1.80346236e-01 6.86996579e-01
1.03920949e+00 5.99047840e-01 8.29506457e-01 1.48960337e-01
1.82178903e+00 -1.22749960e+00 -7.73779929e-01 -4.65160996e-01
9.10493433e-01 -6.98496938e-01 7.44007945e-01 -1.61206573e-01
-1.16478932e+00 -5.44434965e-01 -9.65317249e-01 -5.16938984e-01
-8.81462276e-01 5.07959127e-01 3.75101566e-01 2.82221705e-01
-9.56646025e-01 3.27091992e-01 -6.42545283e-01 -5.73957741e-01
1.48265585e-01 1.34591395e-02 -5.40978611e-01 -3.42831202e-02
-1.81351864e+00 1.39212751e+00 1.03617287e+00 2.24426165e-01
-6.15373850e-01 -7.71025479e-01 -1.30408752e+00 2.28455544e-01
1.62937105e-01 -4.31454241e-01 1.20012498e+00 -4.68765616e-01
-9.98766959e-01 1.33294845e+00 -2.28989184e-01 -6.91757977e-01
3.07209760e-01 -6.39343679e-01 -9.62343335e-01 -2.79600150e-03
4.94673759e-01 7.85074055e-01 1.06111236e-01 -6.99032903e-01
-8.65808368e-01 -2.63618886e-01 -1.90115616e-01 1.99622110e-01
1.22384265e-01 5.16727388e-01 -2.96003699e-01 -6.69855058e-01
-9.62018743e-02 -7.37475991e-01 -3.03617120e-01 -8.35456014e-01
-6.25688314e-01 -7.91443050e-01 1.98475599e-01 -1.37249207e+00
1.64081562e+00 -1.90461254e+00 -2.69159973e-01 9.79697797e-03
9.86643042e-03 2.71859318e-01 -2.33935848e-01 4.75249887e-01
-2.57954419e-01 2.97604442e-01 2.21011974e-03 -3.71570110e-01
3.18061113e-01 -3.90868723e-01 -3.56824517e-01 9.10525396e-02
5.98731399e-01 1.04221225e+00 -7.92505980e-01 -6.35958254e-01
-4.28516656e-01 3.02058645e-03 -1.75148442e-01 3.16743314e-01
-1.74838230e-01 2.11607695e-01 -2.55354822e-01 5.42769670e-01
6.79186702e-01 -2.76205391e-01 2.43657351e-01 -2.85480976e-01
-4.64092731e-01 9.20548201e-01 -1.18944550e+00 1.94554389e+00
-3.78591746e-01 4.97188449e-01 -4.31447327e-02 -4.55369651e-01
6.68632865e-01 5.82040966e-01 6.15868233e-02 -5.64626694e-01
-8.23206380e-02 3.98609102e-01 -4.04498816e-01 -5.54188728e-01
8.03989947e-01 -1.48532063e-01 -8.84070575e-01 -3.32233198e-02
5.04280567e-01 2.34781072e-01 4.64446783e-01 1.52718857e-01
1.32484436e+00 2.35181168e-01 5.62481225e-01 -4.65632975e-01
5.42244315e-01 -1.61099844e-02 1.08885300e+00 6.29944980e-01
-5.21751881e-01 3.46826911e-01 3.18151593e-01 -4.95398909e-01
-9.54423130e-01 -1.12988222e+00 -1.55395553e-01 1.33143342e+00
-1.92683518e-01 -4.39007372e-01 -7.40172982e-01 -1.18215454e+00
-1.03377119e-01 1.01071823e+00 -4.57744569e-01 3.80510509e-01
-8.64643037e-01 -1.09748736e-01 1.13128889e+00 8.97632718e-01
5.50016880e-01 -1.47042048e+00 -2.12979406e-01 4.57173437e-01
-4.98706073e-01 -1.69312954e+00 -8.32386732e-01 3.24754328e-01
-3.81636113e-01 -8.36721241e-01 -4.36437339e-01 -1.22367454e+00
1.60487309e-01 -6.04433060e-01 1.46327746e+00 -4.47971612e-01
-1.22361571e-01 -1.00441709e-01 -2.19848379e-01 -4.23267573e-01
8.58011022e-02 7.42711425e-01 -6.56461641e-02 -2.63068378e-01
7.10063219e-01 -4.38102819e-02 -4.09191698e-01 3.50695923e-02
-4.99041945e-01 -6.95839301e-02 8.40216517e-01 7.19794691e-01
5.54249823e-01 -4.93276685e-01 7.44496822e-01 -9.67023551e-01
2.43255869e-01 -4.69928205e-01 -3.58599424e-01 6.58905506e-01
-4.24678475e-01 1.57794759e-01 6.26214564e-01 -2.05832273e-01
-1.34357452e+00 1.98396489e-01 -4.20513064e-01 3.44102569e-02
-3.49973261e-01 6.51893675e-01 -5.93189776e-01 4.98701751e-01
6.07320905e-01 -2.23055452e-01 -1.04009664e+00 -7.23883569e-01
6.37824833e-01 6.90581679e-01 5.04436314e-01 -5.75607955e-01
4.65993255e-01 -1.73282802e-01 -5.76173723e-01 -8.53142142e-01
-1.43065846e+00 -5.99349141e-01 -7.28182673e-01 6.09467998e-02
1.47794044e+00 -1.32501221e+00 -2.94925392e-01 4.47559446e-01
-1.44000030e+00 -1.73156425e-01 -1.90237299e-01 6.70027971e-01
-2.75846511e-01 -7.35544860e-02 -1.13429713e+00 -4.94870305e-01
-5.79522312e-01 -6.02163851e-01 9.85092878e-01 5.49838960e-01
-4.01666373e-01 -8.84438217e-01 6.22292943e-02 4.99127924e-01
4.26430821e-01 2.10188016e-01 7.31961787e-01 -1.45405424e+00
-2.74613500e-01 -9.98215675e-02 -2.50427455e-01 2.01590508e-01
-4.87618655e-01 -4.53554481e-01 -9.10440803e-01 -9.26149413e-02
-4.51663405e-01 -4.70617831e-01 9.06593144e-01 8.78814161e-02
7.88030088e-01 -1.57355100e-01 -4.50484723e-01 3.73772413e-01
1.36055410e+00 1.87482145e-02 5.32049417e-01 3.37148100e-01
7.29127109e-01 4.40856844e-01 6.24552667e-01 1.07933227e-02
9.37494516e-01 4.52402443e-01 -2.22961277e-01 -5.03948629e-02
-1.79031909e-01 -4.79115427e-01 5.42844951e-01 9.60122287e-01
2.47900799e-01 -1.09763846e-01 -1.31335080e+00 8.67527544e-01
-1.65908813e+00 -9.63550508e-01 -1.02892064e-01 1.79772055e+00
1.10903275e+00 2.38019273e-01 -8.93226042e-02 -5.92001140e-01
9.25338864e-01 2.47015849e-01 -2.63224274e-01 -5.19644856e-01
-2.95610428e-01 4.32355553e-01 3.28626156e-01 2.48437494e-01
-1.57163131e+00 1.65563011e+00 6.94937563e+00 8.80678296e-01
-8.88758898e-01 1.99875355e-01 3.70107383e-01 2.39890873e-01
1.48122041e-02 7.99873397e-02 -1.42169166e+00 3.57117623e-01
1.43739629e+00 -4.10088748e-02 -1.08686469e-01 8.62181485e-01
-2.34027132e-01 1.27052248e-01 -1.14955151e+00 6.36780083e-01
4.18835804e-02 -1.04291379e+00 -1.77968279e-01 -3.12645197e-01
7.24069536e-01 4.15861249e-01 -4.54264969e-01 9.45807159e-01
5.23501992e-01 -1.11804223e+00 9.17000592e-01 9.32520092e-01
7.17988908e-01 -6.74222231e-01 7.99388945e-01 1.37213171e-01
-1.41051054e+00 1.17047496e-01 -1.42030478e-01 3.63764286e-01
5.13686359e-01 6.61067724e-01 -9.63430047e-01 4.63645995e-01
7.00084269e-01 3.41562837e-01 -5.42201042e-01 1.23903263e+00
-7.19239056e-01 7.97881961e-01 -2.18410775e-01 1.00524304e-02
1.85341015e-01 3.80211771e-01 4.85453129e-01 1.92497885e+00
9.10849795e-02 8.00716802e-02 2.81040937e-01 6.81172550e-01
-6.63022161e-01 5.39792180e-01 -3.09041381e-01 -3.04289788e-01
7.62715638e-01 1.36771178e+00 -5.07611811e-01 -5.41221619e-01
-3.77993435e-01 1.04405665e+00 1.10625196e+00 3.64792824e-01
-7.78949082e-01 -8.19297552e-01 4.12504703e-01 -3.15091223e-01
7.32001781e-01 -3.52432460e-01 -3.68782401e-01 -1.41340137e+00
6.86628819e-02 -6.80613518e-01 9.35438335e-01 -6.52695417e-01
-1.48156178e+00 8.01193058e-01 -3.18900853e-01 -7.24906802e-01
-2.25753710e-02 -7.72527933e-01 -5.52408397e-01 9.15639520e-01
-1.45436907e+00 -1.39439762e+00 1.57342896e-01 1.81028366e-01
3.99362683e-01 2.34593004e-02 9.00688708e-01 4.42738920e-01
-6.57918453e-01 8.26552749e-01 -2.04653651e-01 7.97204256e-01
9.86641586e-01 -1.45242608e+00 6.55390739e-01 1.13679409e+00
8.21133256e-02 6.89553082e-01 3.96546036e-01 -7.70888865e-01
-7.16828942e-01 -1.19036722e+00 2.11460638e+00 -6.33110046e-01
6.77734733e-01 -4.19832408e-01 -8.29801321e-01 1.07414019e+00
3.78594965e-01 2.55523715e-02 8.19229662e-01 4.86130178e-01
-7.16448247e-01 1.38910785e-01 -1.01438057e+00 3.56018096e-01
1.09607458e+00 -8.07955623e-01 -1.18303835e+00 1.09468520e-01
1.03010988e+00 -4.75799710e-01 -1.44875598e+00 4.69368428e-01
3.00326765e-01 -3.67119491e-01 6.10470891e-01 -1.11964607e+00
1.52456746e-01 -3.07462573e-01 -1.74110875e-01 -1.40138662e+00
-3.09964478e-01 -3.82480532e-01 -4.34722491e-02 1.56190872e+00
1.08474147e+00 -2.20473871e-01 3.22444499e-01 5.76518357e-01
-2.96590596e-01 -5.81056535e-01 -8.86345863e-01 -8.13519716e-01
1.58028454e-01 -3.32772523e-01 4.81268823e-01 1.32317400e+00
7.55081773e-02 9.52982903e-01 5.98794855e-02 3.12473267e-01
2.49649152e-01 5.23672402e-02 7.66453966e-02 -9.36011493e-01
2.76813924e-01 -3.22376102e-01 -3.04518580e-01 -8.74106705e-01
5.06420135e-01 -1.09623551e+00 3.97855788e-01 -1.69242752e+00
-3.96677367e-02 -3.80749226e-01 -5.68864822e-01 6.55566454e-01
-5.57060421e-01 -2.09922910e-01 2.66677380e-01 -1.01064704e-01
-1.15102518e+00 2.80324966e-01 5.89723170e-01 5.28558642e-02
-5.62254302e-02 -2.77507544e-01 -5.12501121e-01 5.39051056e-01
5.80765069e-01 -4.36220288e-01 4.86215770e-01 -5.53239882e-01
3.11110377e-01 1.10238967e-02 -2.40871251e-01 -1.06531191e+00
2.87269503e-01 -3.48335095e-02 6.60118520e-01 -8.52397203e-01
-1.89732477e-01 -4.90356594e-01 -7.45554045e-02 1.82044238e-01
-5.25785029e-01 2.43601188e-01 2.48580828e-01 -4.41394784e-02
-3.40657741e-01 -2.21400201e-01 5.90478361e-01 -2.46998370e-01
-1.13368034e+00 2.47683942e-01 -3.62718254e-01 7.45852828e-01
8.32625270e-01 5.20705402e-01 -3.29779714e-01 1.59446910e-01
-1.16182745e+00 6.60695970e-01 1.03799641e-01 5.82716227e-01
2.29804635e-01 -1.60413504e+00 -9.95943487e-01 -2.73154855e-01
3.40933502e-01 -1.33085623e-01 2.12822869e-01 6.86370850e-01
-5.22078395e-01 6.74512744e-01 -1.04420558e-01 6.23186268e-02
-8.98873448e-01 4.21907037e-01 3.02812934e-01 -9.52203393e-01
-2.30466425e-01 1.14387548e+00 -1.48071826e-01 -7.10779190e-01
1.13447666e-01 -1.99251220e-01 -5.35394311e-01 4.50766444e-01
6.83944583e-01 5.07722795e-01 2.64395446e-01 -8.65968227e-01
-7.62975216e-01 6.65567294e-02 -2.83422470e-01 -2.54430294e-01
1.22636116e+00 -1.41261801e-01 -1.12465762e-01 5.28288007e-01
1.22362852e+00 1.84348941e-01 -5.97270966e-01 -4.54541594e-01
1.05531275e+00 4.86081362e-01 -3.21698874e-01 -1.13064015e+00
-4.61254388e-01 7.12134063e-01 3.37526530e-01 5.25755025e-02
7.46321917e-01 1.48810759e-01 1.08690965e+00 4.81900156e-01
3.01173210e-01 -1.66281068e+00 -4.12169456e-01 1.35566294e+00
9.11702514e-01 -1.06186342e+00 -4.19346690e-01 -2.00920731e-01
-7.38808811e-01 8.88697565e-01 8.51910233e-01 4.87007946e-02
5.23815870e-01 9.56651568e-02 2.07050890e-01 -2.31976897e-01
-6.75109386e-01 -4.64814842e-01 7.26119816e-01 1.49526760e-01
7.95219898e-01 2.86514699e-01 -7.84234643e-01 1.08634412e+00
-3.39861542e-01 -2.87704378e-01 -7.28913620e-02 8.54424477e-01
-5.10704517e-01 -7.20360577e-01 -1.50246859e-01 1.86360672e-01
-8.89649272e-01 -4.77049559e-01 -2.20108584e-01 6.04367256e-01
2.71484613e-01 8.84878814e-01 1.04969889e-02 -2.41204321e-01
7.00815678e-01 8.79709840e-01 1.67601839e-01 -6.74752653e-01
-9.02612925e-01 -3.33794057e-01 7.93065071e-01 -7.86543310e-01
-1.28857896e-01 -7.01557398e-01 -1.68360806e+00 1.53189808e-01
-2.44994894e-01 5.10149717e-01 7.93561041e-01 1.06000483e+00
6.40496492e-01 3.07804197e-01 3.62287343e-01 -3.99891883e-01
-4.39774662e-01 -1.31918585e+00 -4.54092562e-01 5.77931762e-01
-8.26977566e-03 -3.00918281e-01 -7.58866444e-02 -2.57602980e-04] | [9.673691749572754, 9.553919792175293] |
4a521970-61b1-452c-8117-4cd86c4677ea | the-chinese-causative-passive-homonymy | null | null | https://aclanthology.org/2022.lrec-1.460 | https://aclanthology.org/2022.lrec-1.460.pdf | The Chinese Causative-Passive Homonymy Disambiguation: an adversarial Dataset for NLI and a Probing Task | The disambiguation of causative-passive homonymy (CPH) is potentially tricky for machines, as the causative and the passive are not distinguished by the sentences’ syntactic structure. By transforming CPH disambiguation to a challenging natural language inference (NLI) task, we present the first Chinese Adversarial NLI challenge set (CANLI). We show that the pretrained transformer model RoBERTa, fine-tuned on an existing large-scale Chinese NLI benchmark dataset, performs poorly on CANLI. We also employ Word Sense Disambiguation as a probing task to investigate to what extent the CPH feature is captured in the model’s internal representation. We find that the model’s performance on CANLI does not correspond to its internal representation of CPH, which is the crucial linguistic ability central to the CANLI dataset. CANLI is available on Hugging Face Datasets (Lhoest et al., 2021) at https://huggingface.co/datasets/sxu/CANLI | ['Katja Markert', 'Shanshan Xu'] | null | null | null | null | lrec-2022-6 | ['word-sense-disambiguation'] | ['natural-language-processing'] | [ 1.91574246e-01 6.10383272e-01 -2.77319700e-01 -2.33014032e-01
-7.87698746e-01 -9.68675911e-01 1.05593002e+00 -1.19413927e-01
-3.96363884e-01 7.31676638e-01 6.44465804e-01 -4.79818791e-01
-2.07208134e-02 -8.52057219e-01 -6.00331843e-01 -3.49246919e-01
2.53259301e-01 7.78955877e-01 -1.25278428e-01 -4.89914864e-01
-2.85565406e-01 2.96695173e-01 -1.10990310e+00 3.77223969e-01
7.56467104e-01 7.21025169e-01 -8.05527791e-02 3.96985531e-01
8.45538452e-02 8.06677282e-01 -6.16461813e-01 -8.57258737e-01
2.78686970e-01 -3.75411958e-01 -1.23543549e+00 -7.83764303e-01
6.02451742e-01 1.30667826e-02 -2.44107142e-01 1.04168153e+00
4.43918228e-01 -6.80450872e-02 6.19500041e-01 -1.42943311e+00
-6.56661749e-01 1.16816831e+00 -1.12601057e-01 3.24410915e-01
5.20507276e-01 5.72769195e-02 1.75794685e+00 -6.55015886e-01
1.12267458e+00 1.52212441e+00 7.33895302e-01 7.69370317e-01
-1.05896270e+00 -8.80882204e-01 9.62908268e-02 2.79919922e-01
-1.41714263e+00 -4.47594464e-01 8.25997293e-01 -2.50662476e-01
1.32515514e+00 4.38759029e-01 3.28300744e-01 1.45863104e+00
8.78485814e-02 9.88882601e-01 1.25180078e+00 -5.13980806e-01
-1.08470447e-01 -2.70714223e-01 3.41417611e-01 6.72751725e-01
1.49720386e-01 3.64875019e-01 -5.34499884e-01 -2.46337160e-01
2.20486656e-01 -5.29989600e-01 -3.32415611e-01 1.53641254e-01
-1.14856339e+00 1.01086664e+00 3.89878958e-01 6.48514688e-01
1.16204061e-01 1.00407414e-01 3.45027924e-01 3.83707553e-01
8.30079243e-02 9.04928803e-01 -7.70565927e-01 -8.72210562e-02
-4.48485523e-01 4.08081502e-01 1.02971280e+00 8.66362453e-01
4.97901261e-01 -2.94560790e-01 -2.27619961e-01 7.27881670e-01
-3.24516967e-02 4.01622444e-01 6.09816968e-01 -1.07870626e+00
5.39148927e-01 3.14976662e-01 -2.01367557e-01 -8.54242921e-01
-3.84703755e-01 7.55319512e-03 -5.28845429e-01 -2.15970144e-01
7.11574852e-01 -1.25091136e-01 -5.04851818e-01 2.39914203e+00
1.00659564e-01 -1.69586986e-01 1.59144655e-01 5.89057148e-01
1.03310668e+00 4.83890116e-01 4.12287086e-01 4.26968858e-02
1.54743171e+00 -6.25225127e-01 -5.10912955e-01 -6.88589811e-01
8.02043021e-01 -6.32013440e-01 1.25782645e+00 1.56952754e-01
-9.91489708e-01 -2.64725238e-01 -9.82247114e-01 -4.57817525e-01
-6.76106095e-01 3.70120257e-02 9.36205804e-01 4.14677292e-01
-8.53514910e-01 3.24528545e-01 -6.02009714e-01 -4.96798009e-01
4.15732026e-01 8.29058886e-02 -6.53487444e-01 -1.04059130e-01
-2.00178456e+00 1.31808043e+00 5.42008281e-01 -1.37517184e-01
-5.67252398e-01 -9.43015158e-01 -1.00974035e+00 -1.05572887e-01
4.42088723e-01 -5.32086134e-01 1.14732695e+00 -7.27132082e-01
-1.12299550e+00 1.39357996e+00 -1.87535062e-01 -6.16160035e-01
4.19137508e-01 -1.08464919e-01 -5.50857723e-01 3.01215779e-02
3.01830649e-01 7.79695451e-01 5.32120645e-01 -1.03501236e+00
-3.82088870e-01 -4.60205823e-01 1.92492694e-01 1.60814393e-02
1.30047172e-01 3.82681072e-01 -4.97044856e-03 -7.62049973e-01
-3.68601024e-01 -9.45441008e-01 2.95824349e-01 -4.26535904e-01
-8.88143241e-01 -5.69159448e-01 5.60586154e-01 -7.46302962e-01
1.27458298e+00 -2.01027274e+00 -9.50162932e-02 1.69455081e-01
1.13652810e-01 5.46340644e-01 -3.34012836e-01 4.07930017e-01
-5.13316989e-01 4.47647214e-01 -4.11797106e-01 -2.68702984e-01
4.27831948e-01 4.45850313e-01 -6.28332555e-01 2.94668496e-01
4.59624887e-01 1.40606928e+00 -8.69948804e-01 -3.28218460e-01
-1.32118046e-01 1.83393404e-01 -6.55294716e-01 2.25985453e-01
-3.28846008e-01 3.51482332e-01 -1.04244262e-01 6.87710762e-01
4.88213897e-01 1.06281906e-01 4.46380913e-01 -3.16867322e-01
1.53229102e-01 8.03176641e-01 -6.30017638e-01 1.55665886e+00
-3.98761392e-01 6.01429641e-01 9.64316130e-02 -1.00246346e+00
5.46145737e-01 2.81965047e-01 9.09325182e-02 -7.83130050e-01
5.71204089e-02 1.26120955e-01 4.00278091e-01 -5.91530576e-02
1.65816784e-01 -3.22995156e-01 -6.52996778e-01 4.06837642e-01
2.62150168e-01 -2.56030768e-01 3.06597948e-01 3.94968927e-01
1.17435217e+00 8.36135447e-02 7.50162244e-01 -5.59167743e-01
4.37837005e-01 8.39931443e-02 9.82549787e-01 7.90579796e-01
-2.97848076e-01 3.78587514e-01 7.51803458e-01 -2.32149929e-01
-6.26152754e-01 -1.37302780e+00 -3.84422302e-01 1.01695645e+00
-2.13890389e-01 -7.76427865e-01 -6.53074384e-01 -8.55820417e-01
-6.99974373e-02 1.09528315e+00 -7.33494401e-01 -2.62662381e-01
-8.33905458e-01 -6.10519111e-01 1.06601930e+00 3.74054283e-01
3.94243419e-01 -1.38385463e+00 -3.38295311e-01 8.86418968e-02
-3.87508839e-01 -1.37724304e+00 -3.41433585e-01 1.24656551e-01
-6.02761172e-02 -1.19812715e+00 1.42974392e-01 -6.32230163e-01
2.93738723e-01 -5.54475725e-01 1.59177351e+00 2.07605749e-01
-4.59268987e-01 2.25113690e-01 -2.49046296e-01 -5.95482111e-01
-7.31701016e-01 2.22034484e-01 2.31743325e-03 -6.42382622e-01
6.53371155e-01 -7.02841818e-01 -2.87093103e-01 8.12437162e-02
-7.93098867e-01 -1.63607195e-01 1.26102597e-01 9.95846331e-01
3.17171931e-01 -2.80884374e-02 4.15186256e-01 -1.49533916e+00
5.28035820e-01 -4.29752946e-01 -6.77422762e-01 3.48410875e-01
-3.85859430e-01 1.63673842e-03 5.81558585e-01 -2.55427867e-01
-1.15764856e+00 -2.58289129e-01 -5.15070796e-01 3.63993011e-02
-2.95743018e-01 5.19398510e-01 -8.11156154e-01 1.95843145e-01
4.48405206e-01 -2.54558504e-01 -1.44747868e-01 -3.81881565e-01
7.02416480e-01 3.40246260e-01 7.60151565e-01 -9.64812815e-01
9.46096420e-01 2.83250213e-01 -1.36786729e-01 -6.18238330e-01
-1.26326668e+00 -1.03719875e-01 -7.86258698e-01 4.30595130e-01
9.15781319e-01 -7.49961317e-01 -8.01099479e-01 2.13426799e-01
-1.40031934e+00 -6.33372366e-01 -2.86191434e-01 8.62009972e-02
-4.73928243e-01 3.40999246e-01 -7.74166703e-01 -3.47331852e-01
-6.41241908e-01 -9.62615550e-01 9.78486598e-01 -2.02042982e-01
-8.89089167e-01 -1.13014412e+00 1.29951775e-01 5.19031525e-01
1.74943760e-01 2.38129124e-01 1.42912054e+00 -1.14238381e+00
-1.92031458e-01 -1.06563531e-01 -1.55917674e-01 3.09048980e-01
-4.51418385e-02 -2.74099298e-02 -1.09567273e+00 -6.21829703e-02
4.26674858e-02 -5.22773445e-01 8.62327456e-01 -5.19477800e-02
8.90476644e-01 -7.14977503e-01 -2.72328228e-01 6.76845491e-01
1.31398129e+00 -3.93811651e-02 7.88354635e-01 2.93799818e-01
7.46058226e-01 5.54413617e-01 5.82572103e-01 -5.51362224e-02
4.51288164e-01 5.38672149e-01 1.89911216e-01 1.21577248e-01
-3.85880172e-01 -6.76301360e-01 4.92374390e-01 6.08814299e-01
3.76446933e-01 -1.93281978e-01 -1.16172326e+00 4.98124063e-01
-1.52052426e+00 -9.39095795e-01 3.58396858e-01 1.83654141e+00
1.39157760e+00 1.60077721e-01 -1.79907054e-01 -3.85869411e-03
3.89910042e-01 3.03862870e-01 -4.44933847e-02 -5.39357543e-01
-4.89230633e-01 8.02043319e-01 3.85761708e-01 1.01486194e+00
-1.21327043e+00 1.53144252e+00 6.01513767e+00 1.14057839e+00
-8.10933471e-01 3.24547231e-01 4.54294503e-01 1.58564016e-01
-5.64922929e-01 2.57540584e-01 -9.48898435e-01 4.49583054e-01
8.13909531e-01 -1.76593125e-01 5.41872025e-01 4.89003003e-01
-3.97092193e-01 9.09833163e-02 -1.31833291e+00 4.73934025e-01
7.99265429e-02 -1.19182777e+00 5.51758558e-02 -1.83458090e-01
3.88338596e-01 1.81026384e-01 2.57071145e-02 4.65427369e-01
7.92051971e-01 -1.40198469e+00 6.72140419e-01 9.58913863e-02
9.15176988e-01 -5.53056836e-01 5.18713832e-01 2.13846266e-01
-9.32381094e-01 -6.26871288e-02 -2.50495166e-01 -1.18771657e-01
7.04157203e-02 3.03805858e-01 -6.71897173e-01 2.89605409e-01
4.87285972e-01 6.20142281e-01 -6.95620298e-01 1.96769446e-01
-1.00364339e+00 9.77061272e-01 -3.78529400e-01 2.13448375e-01
1.16367653e-01 -9.62494612e-02 9.85161841e-01 1.23981261e+00
-3.75621840e-02 2.13858709e-01 1.38649091e-01 1.23878860e+00
-2.77553201e-01 -1.00521654e-01 -7.76374936e-01 -3.25258285e-01
7.73610473e-01 1.15329194e+00 -2.64382988e-01 -1.64541066e-01
-3.28470081e-01 8.57637405e-01 6.50229037e-01 2.32813284e-01
-6.64531827e-01 -3.55519831e-01 9.96199965e-01 -7.79118165e-02
5.98074943e-02 -9.70644411e-03 -8.35778713e-02 -1.26330507e+00
-1.55858174e-01 -9.50805366e-01 8.35894048e-01 -7.56317496e-01
-1.54263687e+00 4.88544494e-01 2.44343989e-02 -5.91322422e-01
-6.93083704e-01 -9.45536375e-01 -8.09920728e-01 7.70007014e-01
-1.49789405e+00 -1.64501011e+00 1.75535783e-01 7.01136947e-01
7.20884874e-02 -1.64595947e-01 1.31520724e+00 9.81148407e-02
-5.53026259e-01 1.16018403e+00 -5.22406518e-01 7.61501312e-01
7.87736952e-01 -1.25242376e+00 7.07592964e-01 9.67720926e-01
5.07628858e-01 8.40205312e-01 6.16842687e-01 -6.15067184e-01
-1.19993424e+00 -9.68355060e-01 1.58941913e+00 -1.07438684e+00
1.19504595e+00 -7.28798628e-01 -7.81836748e-01 1.13073146e+00
3.49559337e-01 8.75077695e-02 7.71314502e-01 2.76189268e-01
-8.81036758e-01 1.03488006e-01 -1.13820457e+00 6.36515200e-01
1.46480036e+00 -8.63222778e-01 -1.09100938e+00 3.76918674e-01
1.05961359e+00 -3.28876346e-01 -8.72764468e-01 4.56975728e-01
1.71152264e-01 -6.08218551e-01 1.10759282e+00 -1.06147981e+00
6.11375690e-01 3.43235508e-02 -2.90093213e-01 -1.30992889e+00
-3.10658276e-01 -6.66365981e-01 2.04009995e-01 1.61500025e+00
6.25981331e-01 -8.25219870e-01 3.02985877e-01 7.65718758e-01
1.81369394e-01 -5.54860711e-01 -1.11229789e+00 -8.55360150e-01
6.33334696e-01 -7.27502942e-01 6.32841527e-01 1.11116612e+00
2.52739757e-01 5.71240723e-01 -2.00999249e-02 -1.40040472e-01
4.65975314e-01 3.13163340e-01 4.83943135e-01 -1.21485341e+00
-2.97461301e-01 -4.78380948e-01 -3.59596372e-01 -3.10779124e-01
1.12187159e+00 -1.56053686e+00 -1.42519772e-01 -1.11549485e+00
2.59678178e-02 -4.21786308e-01 -7.78095201e-02 9.85051453e-01
-2.67768025e-01 2.49731302e-01 3.75575423e-01 -4.70686965e-02
-2.62843966e-01 2.67138243e-01 8.92900765e-01 -2.23056018e-01
3.82459491e-01 -3.43177795e-01 -1.02047718e+00 8.38889658e-01
7.42422760e-01 -3.52329314e-01 -2.97376990e-01 -5.18063962e-01
3.40047091e-01 -3.33030879e-01 4.47773486e-01 -4.71076369e-01
1.23526096e-01 -5.90682663e-02 6.74307793e-02 -5.46388887e-02
2.44689018e-01 -4.90804374e-01 -1.32523298e-01 3.71339381e-01
-5.72125971e-01 6.07544417e-03 3.23548049e-01 1.03607535e-01
-2.08070934e-01 -1.80756018e-01 7.58572161e-01 -2.84524024e-01
-9.46655154e-01 2.69598991e-01 -2.58485347e-01 9.78318870e-01
4.87849563e-01 3.15227300e-01 -5.78535676e-01 -1.34757906e-01
-6.77853942e-01 2.03004368e-02 3.98987144e-01 4.48393673e-01
2.05212325e-01 -1.22727787e+00 -6.15911603e-01 2.04121217e-01
2.59307384e-01 -2.70369411e-01 -1.59235209e-01 5.19382596e-01
-2.11590290e-01 5.01686871e-01 -7.37930611e-02 2.08077103e-01
-1.00133240e+00 6.32620573e-01 2.68039942e-01 -5.77487051e-01
-4.43612218e-01 1.18845856e+00 1.95052072e-01 -8.14133406e-01
2.85658352e-02 -2.09113747e-01 -4.59434167e-02 4.64791581e-02
3.93590331e-01 -1.19466325e-02 -1.16471641e-01 -8.89570951e-01
-7.14796901e-01 2.02390745e-01 2.56667798e-03 -1.22114934e-01
1.11078942e+00 7.55503327e-02 -5.68142474e-01 1.60031453e-01
1.41398180e+00 3.31680089e-01 -3.01656485e-01 -3.66702080e-01
2.39238665e-01 -2.22228423e-01 -3.65812182e-01 -1.05244684e+00
-9.09025073e-01 7.81650424e-01 -2.86745336e-02 -2.82789409e-01
7.75345266e-01 3.34841520e-01 9.90260959e-01 4.62365836e-01
3.72706294e-01 -1.07540166e+00 -4.86646801e-01 8.40275466e-01
1.05290675e+00 -1.14061952e+00 -3.24848056e-01 -5.86612999e-01
-7.15304136e-01 6.15069926e-01 6.62900031e-01 -1.70702532e-01
7.52149642e-01 6.32654250e-01 1.49676874e-01 -2.00730294e-01
-7.31144011e-01 -3.08043599e-01 1.63417205e-01 6.60314560e-01
4.19654697e-01 3.37567657e-01 -4.00713921e-01 1.19808292e+00
-1.05576968e+00 -4.41957325e-01 1.05139248e-01 5.62547326e-01
1.46470487e-01 -1.26912093e+00 5.86866699e-02 3.10228676e-01
-6.39445662e-01 -5.97251952e-01 -7.62206197e-01 1.07494009e+00
1.28726453e-01 8.87533069e-01 5.13935052e-02 -1.80028841e-01
1.04802497e-01 3.35312545e-01 4.95820642e-01 -7.36935675e-01
-5.46303391e-01 -5.64145029e-01 5.09732962e-01 -8.61863077e-01
-1.31442070e-01 -5.46481490e-01 -1.34855604e+00 -2.51313120e-01
1.74885035e-01 2.15277568e-01 1.38242468e-01 1.22228193e+00
1.26568526e-01 2.58164015e-02 1.86282158e-01 -2.50955552e-01
-5.36960959e-01 -8.02958250e-01 -4.43076313e-01 8.95927429e-01
-3.33677828e-02 -5.38637519e-01 -4.95653361e-01 -3.19856852e-01] | [10.617630958557129, 9.27602481842041] |
6828bf48-5905-4340-856b-a97f6658698d | forecasting-irregularly-sampled-time-series | 2305.12932 | null | https://arxiv.org/abs/2305.12932v1 | https://arxiv.org/pdf/2305.12932v1.pdf | Forecasting Irregularly Sampled Time Series using Graphs | Forecasting irregularly sampled time series with missing values is a crucial task for numerous real-world applications such as healthcare, astronomy, and climate sciences. State-of-the-art approaches to this problem rely on Ordinary Differential Equations (ODEs) but are known to be slow and to require additional features to handle missing values. To address this issue, we propose a novel model using Graphs for Forecasting Irregularly Sampled Time Series with missing values which we call GraFITi. GraFITi first converts the time series to a Sparsity Structure Graph which is a sparse bipartite graph, and then reformulates the forecasting problem as the edge weight prediction task in the graph. It uses the power of Graph Neural Networks to learn the graph and predict the target edge weights. We show that GraFITi can be used not only for our Sparsity Structure Graph but also for alternative graph representations of time series. GraFITi has been tested on 3 real-world and 1 synthetic irregularly sampled time series dataset with missing values and compared with various state-of-the-art models. The experimental results demonstrate that GraFITi improves the forecasting accuracy by up to 17% and reduces the run time up to 5 times compared to the state-of-the-art forecasting models. | ['Lars Schmidt-Thieme', 'Stefan Born', 'Shayan Javed', 'Johannes Burchert', 'Nourhan Ahmed', 'Randolf Sholz', 'Kiran Madusudanan', 'Vijaya Krishna Yalavarthi'] | 2023-05-22 | null | null | null | null | ['astronomy'] | ['miscellaneous'] | [ 1.65232196e-01 5.05189896e-02 1.43702656e-01 -2.10261457e-02
6.60721809e-02 -3.02168161e-01 2.37772390e-01 1.36430010e-01
3.71691883e-01 7.41344929e-01 -1.62505601e-02 -5.69880188e-01
-3.76971364e-01 -9.85843480e-01 -6.36096895e-01 -6.86551154e-01
-8.95984530e-01 4.56671178e-01 -8.70233309e-03 -6.35402739e-01
6.26661554e-02 4.85075027e-01 -1.31459153e+00 -2.22342387e-02
9.16955173e-01 1.15582883e+00 -3.09353679e-01 5.09492457e-01
-2.14036286e-01 8.14666152e-01 -2.98698038e-01 1.34771364e-02
4.37833190e-01 -2.68806577e-01 -3.29844177e-01 8.89501125e-02
7.15301484e-02 1.55011326e-01 -3.99923950e-01 8.49822879e-01
3.00110906e-01 2.64967859e-01 5.58889925e-01 -1.69908512e+00
-5.17167032e-01 4.23919767e-01 -6.50167048e-01 2.78197020e-01
1.80596694e-01 -2.26118460e-01 6.46543086e-01 -5.87761641e-01
4.32682365e-01 9.55493569e-01 1.19154751e+00 1.01448804e-01
-1.50028694e+00 -6.83768928e-01 2.70053536e-01 1.87611580e-03
-1.22097886e+00 -9.21892375e-03 1.27763152e+00 -6.37212574e-01
1.07486010e+00 3.29409689e-01 7.91248381e-01 6.19820833e-01
4.58349526e-01 2.02083468e-01 9.19639587e-01 3.19897048e-02
1.77287564e-01 -7.06050575e-01 2.47114614e-01 7.65492082e-01
3.98801088e-01 1.54981703e-01 -1.31644338e-01 -5.35129488e-01
7.45768666e-01 4.20972139e-01 -4.31066424e-01 -1.90422207e-01
-1.23769367e+00 1.07753670e+00 4.10767138e-01 4.38249916e-01
-6.83853626e-01 3.96808609e-02 4.64347303e-01 7.59272933e-01
1.09018290e+00 2.40791216e-01 -4.72844720e-01 9.88253504e-02
-7.88318336e-01 -2.34044180e-03 1.27076650e+00 5.34930527e-01
6.49819911e-01 8.10465395e-01 1.94778964e-01 5.12402773e-01
6.74516708e-02 4.72068369e-01 4.29172635e-01 -4.31200087e-01
5.27625978e-01 8.86649966e-01 -1.77868813e-01 -1.69536233e+00
-9.85141873e-01 -5.86577654e-01 -1.65208185e+00 -1.40166804e-01
3.20705414e-01 -4.18948829e-01 -7.97240973e-01 1.46767223e+00
3.10537517e-01 8.67135823e-01 -1.52527750e-01 7.59414792e-01
6.92779422e-01 1.09150767e+00 -2.93475807e-01 -6.55736029e-01
7.29902446e-01 -6.74793124e-01 -6.94974840e-01 1.60803854e-01
7.19897687e-01 -5.98101854e-01 4.85004574e-01 2.62314230e-01
-7.70998776e-01 -4.44268942e-01 -8.17341685e-01 4.64458972e-01
-3.13306868e-01 -1.15696520e-01 7.22612262e-01 3.28760028e-01
-1.18996060e+00 9.41312730e-01 -9.49019372e-01 -1.63015455e-01
-1.95154771e-01 4.01278764e-01 -3.40887696e-01 1.59957990e-01
-1.36430311e+00 6.07335389e-01 6.09191507e-02 2.84641623e-01
-2.24985883e-01 -1.00950480e+00 -1.05696225e+00 2.76302874e-01
2.49236718e-01 -6.63726866e-01 5.03542423e-01 -8.72771621e-01
-1.26117969e+00 3.77350420e-01 -1.90007146e-02 -6.93794250e-01
1.77047119e-01 5.12175977e-01 -9.57088232e-01 -1.59700379e-01
-1.42165065e-01 -2.58186460e-01 1.05624533e+00 -8.25891554e-01
-1.05867013e-01 -3.29080582e-01 -3.22260827e-01 -2.92181700e-01
-1.51206419e-01 -4.71100360e-01 2.73097485e-01 -1.13408148e+00
3.47147822e-01 -1.15192091e+00 -4.14957941e-01 -3.37243170e-01
-1.57004029e-01 -1.50182635e-01 1.16187453e+00 -8.88305306e-01
1.45205247e+00 -1.89543378e+00 2.14878097e-01 6.65149689e-01
4.37464833e-01 8.22027549e-02 -1.89325944e-01 9.19359863e-01
-3.97830933e-01 -2.05132291e-01 -5.11528611e-01 -1.11958064e-01
-4.01244521e-01 3.58284116e-01 -4.65942442e-01 6.97468162e-01
6.05374835e-02 5.44793665e-01 -8.63268673e-01 1.75055102e-01
1.33193851e-01 7.25249410e-01 -2.88920581e-01 2.39058003e-01
-1.46412924e-01 5.96647561e-01 -3.61203104e-01 3.21075350e-01
7.59040177e-01 -6.18229628e-01 3.21294159e-01 -1.88086763e-01
-5.43353856e-02 -1.50627941e-01 -1.28636289e+00 1.27555060e+00
-3.98724228e-01 3.51530343e-01 -2.01382592e-01 -1.61045468e+00
1.43423975e+00 3.99309158e-01 1.01344287e+00 -5.21410167e-01
4.32282463e-02 1.84863135e-01 -1.30783170e-01 -4.16413426e-01
1.23703189e-01 -2.44432613e-02 6.44405335e-02 3.46722513e-01
-2.06845284e-01 -1.13253362e-01 2.46216685e-01 5.06549589e-02
1.17430437e+00 -9.04916227e-02 2.55445153e-01 -5.18139899e-01
6.51157022e-01 1.04391808e-02 5.91049373e-01 2.07837716e-01
4.04041290e-01 3.91714722e-01 7.17936933e-01 -1.08432078e+00
-8.23593140e-01 -5.07047415e-01 3.67825359e-01 6.82528615e-01
-2.99636871e-01 -2.62470305e-01 -2.99990863e-01 -3.54408264e-01
3.53011727e-01 5.62975705e-01 -7.03457892e-01 -7.44629577e-02
-6.00205839e-01 -8.81398380e-01 4.11553122e-02 2.97235817e-01
1.31540373e-01 -8.38319421e-01 -1.77499756e-01 4.81523812e-01
-7.19766691e-02 -1.12675452e+00 -5.63848555e-01 1.10596858e-01
-1.24471962e+00 -1.02880800e+00 -5.88591695e-01 -6.64990962e-01
8.37481201e-01 2.65350938e-01 1.39864349e+00 3.03456217e-01
-1.93492606e-01 4.04246449e-01 -4.29218024e-01 -2.61500031e-01
-3.62786740e-01 -1.12056443e-02 -2.65980698e-02 5.60502172e-01
-1.38471916e-01 -1.11561608e+00 -5.26720524e-01 2.32599095e-01
-9.99531329e-01 1.07387178e-01 7.88422376e-02 9.22932684e-01
5.87593794e-01 1.67662017e-02 1.03923476e+00 -1.11502767e+00
6.49617970e-01 -9.92682517e-01 -1.12769198e+00 1.44375399e-01
-7.38359153e-01 2.40174219e-01 1.22659338e+00 -5.91552734e-01
-2.55483598e-01 2.68636703e-01 1.65018290e-01 -6.09549344e-01
3.72419268e-01 1.10333502e+00 6.06207132e-01 -5.15831709e-01
3.73087615e-01 2.34706476e-01 2.92135924e-01 -4.51857656e-01
-1.72192901e-02 2.79670030e-01 2.69043237e-01 -2.57337719e-01
6.39050126e-01 4.34348822e-01 8.18738282e-01 -9.34327424e-01
-6.61975503e-01 -5.69147110e-01 -2.70274282e-01 -3.68795395e-01
4.28233981e-01 -8.71442437e-01 -7.92667449e-01 4.75618660e-01
-9.20744359e-01 -3.18770826e-01 -1.22863285e-01 3.65037560e-01
-4.65945691e-01 3.86033088e-01 -5.83229482e-01 -7.91242421e-01
-5.45929074e-01 -5.84933758e-01 9.37646747e-01 -2.71929830e-01
1.55074179e-01 -1.54615581e+00 2.98953623e-01 -3.24613780e-01
7.36361444e-01 1.10810435e+00 1.02219570e+00 -5.21629274e-01
-2.23392278e-01 -3.90366673e-01 -6.75694495e-02 7.81572238e-02
1.72559589e-01 -1.36400566e-01 -4.65418607e-01 -5.33056498e-01
1.91035122e-01 2.12820798e-01 6.34793520e-01 4.45568919e-01
1.08447587e+00 -4.83044356e-01 -1.42789409e-01 7.23999500e-01
1.59544933e+00 7.32699856e-02 4.81762618e-01 -2.06292093e-01
8.66537750e-01 4.44872141e-01 6.77557439e-02 7.56986320e-01
5.51018596e-01 4.57215488e-01 6.19238675e-01 -1.51269794e-01
1.58884153e-01 -1.06823452e-01 1.81838393e-01 1.59189749e+00
-3.30492437e-01 -1.83280736e-01 -1.02679455e+00 6.73855424e-01
-2.20902824e+00 -9.55621839e-01 -5.21360695e-01 2.20005345e+00
4.11659837e-01 -1.21302173e-01 1.99685037e-01 4.61919904e-01
6.61127508e-01 3.43109637e-01 -4.32683825e-01 -3.69716197e-01
2.16124747e-02 2.00671479e-01 5.65712869e-01 4.58267212e-01
-9.94229913e-01 3.87189656e-01 6.01895428e+00 2.89464265e-01
-1.57808709e+00 -2.57988572e-01 5.33832073e-01 4.33008373e-01
-2.77266592e-01 -3.89587507e-02 -2.54574358e-01 5.28391480e-01
1.16315210e+00 -5.23845315e-01 8.62828851e-01 6.10817313e-01
3.20975900e-01 4.27462399e-01 -8.00426006e-01 1.16767776e+00
1.96968600e-01 -1.33771050e+00 2.37124576e-03 -7.76883736e-02
9.60577488e-01 1.22049622e-01 -3.29813570e-01 2.92819858e-01
1.88768730e-01 -9.56672192e-01 1.77655201e-02 8.56910050e-01
6.13099515e-01 -7.62102187e-01 7.36504614e-01 4.23115700e-01
-1.66502059e+00 1.29531547e-01 -3.80802333e-01 -5.67833662e-01
2.41738334e-01 1.05633521e+00 -5.84272087e-01 9.62585807e-01
3.62049103e-01 1.09282410e+00 -2.29175508e-01 1.06586254e+00
3.63971144e-01 8.45809579e-01 -5.03416240e-01 1.03180828e-02
2.68371344e-01 -5.36032915e-01 5.84715545e-01 7.75165796e-01
9.22181487e-01 2.21729249e-01 6.26464486e-01 3.56874168e-01
8.77751410e-02 1.63095787e-01 -9.16562736e-01 -1.28030121e-01
-5.37327342e-02 1.07510507e+00 -6.67889237e-01 -2.68102348e-01
-6.44005835e-01 7.03353047e-01 1.89712033e-01 5.71287632e-01
-6.04067922e-01 -4.12679613e-01 3.45275134e-01 4.41100985e-01
3.70786607e-01 -4.60109413e-01 -1.32971674e-01 -1.36984944e+00
-9.90532711e-02 -7.44207501e-01 7.22431600e-01 -7.28486180e-01
-1.63890958e+00 9.41935599e-01 -1.62935719e-01 -1.65882730e+00
-5.83477020e-01 -4.64161605e-01 -6.18962169e-01 8.64850163e-01
-1.58542645e+00 -9.62990344e-01 -5.06815374e-01 9.01988745e-01
1.01168431e-01 -2.04175323e-01 1.04565382e+00 3.12782526e-01
-3.30051810e-01 -4.66557778e-02 3.37536424e-01 -1.17345238e-02
1.43027917e-01 -9.42350924e-01 7.37830043e-01 6.64964378e-01
8.74925554e-02 1.46288490e-02 7.73301780e-01 -8.75025332e-01
-1.84957421e+00 -1.30107641e+00 1.08554149e+00 1.86722264e-01
1.15908051e+00 -1.81952536e-01 -1.07912076e+00 5.93156338e-01
7.91481212e-02 6.49879217e-01 4.10429031e-01 -1.26461357e-01
-1.99069858e-01 -2.62188017e-01 -1.08162594e+00 2.49506354e-01
8.53222191e-01 -1.37476578e-01 -3.28917056e-01 5.24978817e-01
6.42234087e-01 -4.11066800e-01 -1.08238649e+00 4.74476725e-01
2.65709847e-01 -6.72068536e-01 8.09793413e-01 -7.29171634e-01
2.05959156e-02 -2.25488007e-01 1.00324102e-01 -1.78932106e+00
-4.09221530e-01 -1.02987754e+00 -4.88942027e-01 6.56009853e-01
2.55715638e-01 -1.17841434e+00 4.20237839e-01 3.75602067e-01
-5.11339940e-02 -7.58134425e-01 -8.55264127e-01 -1.05036938e+00
-3.54268342e-01 -3.32966179e-01 7.49210000e-01 1.51658702e+00
3.98033671e-02 2.68409222e-01 -6.76042914e-01 2.57519215e-01
6.61957324e-01 5.82261384e-01 5.67399323e-01 -1.78800976e+00
-1.68684959e-01 -1.55446380e-01 -7.42100120e-01 -5.86537719e-01
3.15420985e-01 -1.00999129e+00 -4.61263895e-01 -1.55151343e+00
-5.08275151e-01 -4.52309877e-01 -2.96937138e-01 4.15941626e-01
1.55663595e-01 1.83030769e-01 1.06385358e-01 2.37797368e-02
-7.05564171e-02 4.51574266e-01 1.11615849e+00 -1.30782738e-01
-3.67998868e-01 2.02634320e-01 -1.50565818e-01 5.10496914e-01
8.59596968e-01 -4.60139960e-01 -5.14187992e-01 -2.04657376e-01
6.07826829e-01 8.23029697e-01 2.74591297e-01 -1.01164854e+00
1.37043357e-01 -1.82442725e-01 -7.19570890e-02 -5.35530031e-01
-1.69739034e-02 -1.24752200e+00 7.29024351e-01 5.02369463e-01
1.09617695e-01 7.22488046e-01 3.09114397e-01 6.38483703e-01
-3.82235885e-01 2.41425827e-01 2.67162561e-01 1.82217509e-01
-4.70433831e-01 6.45092666e-01 -7.62366354e-02 1.07384130e-01
9.21088576e-01 1.09739490e-01 -2.95203418e-01 -8.12317491e-01
-1.05509365e+00 2.62378365e-01 5.05279414e-02 2.61293650e-01
5.19235134e-01 -1.49120343e+00 -9.43048775e-01 5.92992127e-01
-1.28877446e-01 -3.89761835e-01 2.44402677e-01 1.14463902e+00
-5.72551489e-01 1.99510083e-01 -1.33106381e-01 -5.67680836e-01
-9.35604751e-01 8.37614179e-01 1.91731423e-01 -6.35004163e-01
-8.88606787e-01 7.63325542e-02 -1.27042934e-01 -3.41369987e-01
7.23275095e-02 -6.61469579e-01 -2.08872855e-01 1.12236906e-02
1.57951847e-01 5.42682409e-01 1.96174994e-01 -6.64045274e-01
-2.69866198e-01 8.38970125e-01 6.82361364e-01 5.89331210e-01
1.76826179e+00 7.98216760e-02 -3.54660422e-01 6.23136044e-01
1.22507775e+00 -2.37649322e-01 -8.88082683e-01 -1.91955253e-01
-6.40959069e-02 -2.80757099e-01 -2.41070706e-02 -3.52116823e-01
-1.54584157e+00 6.69681549e-01 2.99323231e-01 1.18596184e+00
1.31933868e+00 -6.42164767e-01 1.02020633e+00 3.43225837e-01
4.93177801e-01 -6.10241413e-01 -4.32784677e-01 7.01480865e-01
1.20766401e+00 -1.04262161e+00 2.00033505e-02 -6.57985389e-01
-4.81287032e-01 1.38443685e+00 -4.22487445e-02 -7.02290177e-01
1.30860388e+00 4.19953346e-01 -7.13575855e-02 -1.58747613e-01
-8.25064242e-01 -1.45317242e-01 6.11085594e-01 4.01298285e-01
5.94176412e-01 2.14135289e-01 -4.25659388e-01 4.40860748e-01
4.48906887e-03 1.59685716e-01 3.67155731e-01 5.82337916e-01
3.63122262e-02 -7.73185611e-01 -3.78880709e-01 8.66240144e-01
-3.56034040e-01 -3.61686796e-02 2.48316606e-03 6.50900185e-01
-4.32475448e-01 8.73327553e-01 7.02311695e-02 -3.99673283e-01
3.87029499e-01 1.77091852e-01 3.13116126e-02 -1.88886657e-01
-7.49622107e-01 1.03989109e-01 8.12223777e-02 -4.75453615e-01
-6.17595911e-01 -2.98479259e-01 -1.18409753e+00 -5.57589293e-01
-6.63317218e-02 2.28221610e-01 4.59777951e-01 6.50672495e-01
6.92406058e-01 6.65045321e-01 9.37128007e-01 -9.58507419e-01
-3.66122127e-01 -7.53439009e-01 -7.42359757e-01 5.34163177e-01
5.37744105e-01 -6.70190871e-01 -6.00096524e-01 -3.41342355e-04] | [6.8328633308410645, 2.9107511043548584] |
d84c0bd3-de97-4530-90bd-e30fb39b627f | improved-cross-view-completion-pre-training | 2211.10408 | null | https://arxiv.org/abs/2211.10408v2 | https://arxiv.org/pdf/2211.10408v2.pdf | Improved Cross-view Completion Pre-training for Stereo Matching and Optical Flow | Despite impressive performance for high-level downstream tasks, self-supervised pre-training methods have not yet fully delivered on dense geometric vision tasks such as stereo matching or optical flow. The application of selfsupervised concepts, such as instance discrimination or masked image modeling, to geometric tasks is an active area of research. In this work, we build on the recent crossview completion framework, a variation of masked image modeling that leverages a second view from the same scene which makes it well suited for binocular downstream tasks. The applicability of this concept has so far been limited in at least two ways: (a) by the difficulty of collecting realworld image pairs -- in practice only synthetic data have been used -- and (b) by the lack of generalization of vanilla transformers to dense downstream tasks for which relative position is more meaningful than absolute position. We explore three avenues of improvement: first, we introduce a method to collect suitable real-world image pairs at large scale. Second, we experiment with relative positional embeddings and show that they enable vision transformers to perform substantially better. Third, we scale up vision transformer based cross-completion architectures, which is made possible by the use of large amounts of data. With these improvements, we show for the first time that stateof-the-art results on stereo matching and optical flow can be reached without using any classical task-specific techniques like correlation volume, iterative estimation, image warping or multi-scale reasoning, thus paving the way towards universal vision models. | ['Romain Brégier', 'Vaibhav Arora', 'Yohann Cabon', 'Vincent Leroy', 'Thomas Lucas', 'Jérôme Revaud', 'Boris Chidlovskii', 'Leonid Antsfeld', 'Gabriela Csurka', 'Philippe Weinzaepfel'] | 2022-11-18 | null | null | null | null | ['stereo-matching-1'] | ['computer-vision'] | [ 2.86599904e-01 2.58714110e-01 8.70631412e-02 -2.64526308e-01
-5.00395596e-01 -5.10936141e-01 1.09413600e+00 -5.16592758e-03
-4.39977348e-01 4.89258915e-01 3.96229714e-01 -2.70669132e-01
-1.09069809e-01 -6.40370309e-01 -7.71828234e-01 -4.29267377e-01
-5.52089773e-02 4.61162955e-01 4.10845369e-01 -3.96324396e-01
3.84078085e-01 5.10265768e-01 -1.86838818e+00 2.18534276e-01
7.03055203e-01 7.99777389e-01 2.03942254e-01 5.61514854e-01
7.29829147e-02 6.23928189e-01 9.57292039e-03 -3.19461077e-01
7.55445480e-01 -2.68816501e-01 -9.59826410e-01 2.06636205e-01
9.66959178e-01 -3.34439576e-01 -3.67754161e-01 7.85154164e-01
2.12546527e-01 2.26226702e-01 5.28156877e-01 -1.12320614e+00
-6.56694591e-01 1.57961696e-01 -4.81058717e-01 1.19423844e-01
3.21697086e-01 4.90642846e-01 1.09132385e+00 -9.09639478e-01
8.70369434e-01 1.22933376e+00 7.20852792e-01 4.97384131e-01
-1.56260526e+00 -2.66991407e-01 1.22369930e-01 1.35509893e-01
-9.90621448e-01 -5.88304400e-01 7.26046145e-01 -6.80373073e-01
1.05518043e+00 -4.96072881e-02 7.05456018e-01 9.91334796e-01
-1.51974544e-01 5.80138028e-01 1.63055706e+00 -5.64950049e-01
3.20420377e-02 1.56322420e-01 -2.03608379e-01 7.68956542e-01
1.52538419e-01 6.47184074e-01 -4.65462536e-01 3.01315069e-01
9.95733559e-01 -1.10878214e-01 -5.31658709e-01 -9.80430603e-01
-1.63926613e+00 8.36844265e-01 8.09557080e-01 4.75982755e-01
-6.10116161e-02 1.02876000e-01 2.89450496e-01 3.54671538e-01
6.23716831e-01 8.03898036e-01 -3.58669579e-01 -3.44386995e-02
-1.17066634e+00 2.16657609e-01 5.62988758e-01 7.39317238e-01
1.03515697e+00 2.77199093e-02 4.09227684e-02 4.94960129e-01
1.61008999e-01 2.26840675e-01 4.19449568e-01 -1.18804729e+00
4.73471403e-01 4.33985710e-01 1.02688238e-01 -8.86454344e-01
-3.77049327e-01 -3.34044456e-01 -6.64328933e-01 7.97213495e-01
8.86782408e-01 1.75008476e-01 -9.19936538e-01 1.87304449e+00
2.66869720e-02 2.88458824e-01 8.83925706e-02 9.95774090e-01
4.18920696e-01 1.05156191e-01 -3.77236962e-01 2.30907295e-02
1.25800586e+00 -1.07640231e+00 -8.35443139e-02 -3.14815968e-01
5.54369688e-01 -7.94008255e-01 1.16920614e+00 4.15376455e-01
-1.07051432e+00 -7.68506587e-01 -9.92569149e-01 -4.30141747e-01
-5.16012132e-01 -2.33592585e-01 9.20935869e-01 5.65006912e-01
-1.33718085e+00 8.18877339e-01 -7.34614611e-01 -7.15993464e-01
4.87359911e-01 2.20750153e-01 -8.72050822e-01 -2.37317935e-01
-8.72681856e-01 1.06764781e+00 3.00985761e-02 -7.46785700e-02
-8.02512348e-01 -8.75636637e-01 -1.16492438e+00 -8.79649296e-02
1.54137626e-01 -1.02428067e+00 8.65922928e-01 -9.04686689e-01
-1.35925937e+00 1.27603531e+00 -1.96056888e-01 -6.02436185e-01
8.39847267e-01 -1.41398475e-01 1.07955076e-01 1.81410357e-01
1.04381762e-01 1.03434825e+00 1.04230118e+00 -1.37113035e+00
-3.90506655e-01 -6.36516869e-01 3.36343378e-01 1.60716817e-01
-2.30531469e-01 -3.38743716e-01 -2.24720448e-01 -4.98187125e-01
3.30676138e-02 -9.53383684e-01 -3.80550712e-01 3.44335616e-01
-1.88754529e-01 4.85412069e-02 6.26397371e-01 -3.10083151e-01
4.20053095e-01 -1.95056033e+00 3.10467213e-01 -9.44985077e-02
2.93823093e-01 3.27143818e-01 -2.07423061e-01 3.82696509e-01
-3.08687061e-01 -1.36940762e-01 -3.00145239e-01 -7.44032383e-01
-1.36669785e-01 2.10674539e-01 -5.79227030e-01 5.71893275e-01
3.88940752e-01 1.09165609e+00 -9.80886579e-01 -3.29542905e-01
7.32905865e-01 4.90406841e-01 -7.02121973e-01 1.64606228e-01
-1.45257458e-01 7.47269630e-01 6.04186319e-02 1.68914586e-01
6.13535106e-01 -5.27496822e-02 -2.75754035e-01 -4.23693001e-01
-3.83384466e-01 3.03693712e-01 -1.11889386e+00 2.15989208e+00
-6.24714136e-01 9.10347700e-01 1.53556792e-02 -1.23017991e+00
6.23328269e-01 -4.95827524e-03 4.24276501e-01 -6.99773312e-01
-6.17893003e-02 1.50621593e-01 -1.37361810e-01 -3.87289405e-01
4.06103134e-01 -2.77688205e-01 3.82236838e-01 3.12151760e-01
2.22595066e-01 -6.11001849e-01 1.86762959e-01 1.19529799e-01
9.55154002e-01 5.39993465e-01 1.80937111e-01 -2.70906210e-01
5.41965663e-01 2.69878238e-01 1.43436044e-01 6.32538974e-01
-2.04846010e-01 1.05314052e+00 1.83829114e-01 -2.97774792e-01
-1.31129599e+00 -1.20547533e+00 -3.55966061e-01 6.30070448e-01
6.63345084e-02 -2.67126590e-01 -6.37332201e-01 -3.65355879e-01
1.19364589e-01 4.02202994e-01 -8.15961361e-01 6.73113670e-03
-4.54993844e-01 -3.79768908e-01 3.09307963e-01 6.12431109e-01
5.36954105e-01 -8.56133282e-01 -6.35907650e-01 -6.51926873e-03
1.28129482e-01 -1.46650743e+00 -1.37288287e-01 7.31510371e-02
-1.06979692e+00 -1.13050258e+00 -8.24895084e-01 -4.95607674e-01
5.37646711e-01 6.79359615e-01 1.10355091e+00 -2.87822429e-02
-4.87259716e-01 5.99351466e-01 -1.40829891e-01 -3.81173380e-02
-1.29334748e-01 -1.84929475e-01 8.16180781e-02 9.21441428e-03
2.02798285e-02 -9.36921656e-01 -8.44268024e-01 2.35302165e-01
-8.44984949e-01 1.43516913e-01 6.66796684e-01 9.15799379e-01
2.72504002e-01 -4.44476277e-01 2.53712423e-02 -8.48293602e-01
3.42183560e-01 -4.62277755e-02 -6.66732013e-01 -7.72850737e-02
-7.13728428e-01 3.13507468e-01 6.01672351e-01 -1.51162505e-01
-8.34828317e-01 6.92267343e-02 -3.03766597e-03 -7.18757033e-01
-3.15657020e-01 1.03045560e-01 2.06060916e-01 -3.04811329e-01
8.32673967e-01 9.65219662e-02 5.59516728e-01 -3.68187785e-01
8.27407539e-01 2.65070826e-01 7.43205190e-01 -4.75859404e-01
1.17719507e+00 1.00268126e+00 2.80899674e-01 -7.35498786e-01
-7.52947211e-01 -6.23803377e-01 -8.64842236e-01 1.22858509e-01
1.11122668e+00 -1.02057660e+00 -6.66172743e-01 3.10407549e-01
-9.41467881e-01 -4.56290305e-01 -4.85283703e-01 6.27624333e-01
-8.62494230e-01 6.72264695e-01 -4.63448375e-01 -6.27991796e-01
1.22732200e-01 -1.20656824e+00 1.23596668e+00 -8.86064693e-02
-4.13498171e-02 -1.13208735e+00 8.10189098e-02 6.94760978e-01
4.65738088e-01 1.31862268e-01 5.92444062e-01 -3.00809115e-01
-8.46688330e-01 1.09536812e-01 -5.62453806e-01 4.50272322e-01
5.31751104e-03 -1.34208947e-01 -1.33489382e+00 -3.84202152e-01
2.97847670e-02 -5.64201295e-01 1.17048311e+00 1.94590271e-01
8.95452619e-01 3.67178798e-01 -2.11766168e-01 8.10785115e-01
1.46696496e+00 -4.60619628e-01 7.91392863e-01 3.92601430e-01
8.86435151e-01 9.87545431e-01 3.93626750e-01 -1.93875714e-03
5.00672400e-01 8.91419768e-01 5.89987457e-01 -4.02811140e-01
-5.17831504e-01 -3.10433775e-01 1.64523646e-01 3.80313724e-01
-2.83690155e-01 2.17884511e-01 -7.40868628e-01 7.31994808e-01
-1.90724790e+00 -1.03851032e+00 -1.52867466e-01 2.41086769e+00
4.01306838e-01 1.67364806e-01 6.94795996e-02 2.35958129e-01
2.40935147e-01 2.40828663e-01 -3.43232155e-01 -1.78418115e-01
-9.62457135e-02 3.76489073e-01 5.20559847e-01 6.98983192e-01
-1.00160241e+00 9.96256649e-01 5.75200462e+00 4.32420075e-01
-1.16010249e+00 1.55238092e-01 4.62759912e-01 1.47322088e-01
-3.20968628e-01 4.85372603e-01 -4.58897024e-01 1.33952767e-01
3.33518654e-01 3.50257963e-01 5.23769677e-01 6.25516891e-01
2.32484043e-01 -2.32109621e-01 -1.45306170e+00 1.09410095e+00
2.03097016e-01 -1.38846493e+00 -3.74728888e-02 3.25139582e-01
6.39725268e-01 3.52711827e-01 1.53568938e-01 9.48895141e-02
2.81624407e-01 -1.04553354e+00 5.16819715e-01 3.24511588e-01
7.25099385e-01 -8.45560282e-02 2.54352361e-01 2.05610722e-01
-8.76693726e-01 -4.15540822e-02 -4.35529858e-01 -2.64339119e-01
1.25420004e-01 5.93148708e-01 -4.92004991e-01 6.56165898e-01
6.21978283e-01 9.18620288e-01 -6.86776936e-01 9.30365086e-01
-8.91841874e-02 1.86465740e-01 -3.66975516e-01 4.96891737e-01
2.83218414e-01 -3.58724505e-01 5.12173176e-01 9.59591806e-01
1.01238787e-01 -4.89466518e-01 -5.27729057e-02 1.04117227e+00
3.30371469e-01 -9.76593047e-02 -1.02870083e+00 4.20315832e-01
3.87300178e-02 1.32849824e+00 -4.69626784e-01 -6.09174669e-02
-6.64496362e-01 9.70921576e-01 5.89737594e-01 3.02180618e-01
-4.32007730e-01 8.95142779e-02 7.51341462e-01 3.47015887e-01
4.06373113e-01 -6.05482280e-01 -4.29216474e-01 -1.56155849e+00
1.29426941e-01 -6.11910045e-01 1.07481256e-01 -1.00558054e+00
-1.34448397e+00 4.99905199e-01 4.33274172e-02 -1.25830781e+00
-4.02140856e-01 -8.70330811e-01 -6.85596168e-01 8.51809859e-01
-1.85490000e+00 -1.26313746e+00 -5.18899143e-01 8.44053686e-01
4.01795745e-01 -1.76829711e-01 7.49469638e-01 -3.50565114e-03
-1.59018114e-01 2.58967042e-01 -1.61573008e-01 3.58202755e-02
8.70518446e-01 -1.36020875e+00 5.49111187e-01 1.01012301e+00
6.63903415e-01 6.68841183e-01 7.36108482e-01 1.54084722e-02
-1.32160854e+00 -7.15910733e-01 7.13189483e-01 -7.93604374e-01
7.59525359e-01 -6.48078799e-01 -7.97625601e-01 5.40885985e-01
2.58901924e-01 4.25352603e-01 1.26467615e-01 1.87115490e-01
-6.40610754e-01 -2.79118627e-01 -9.99586821e-01 6.54265881e-01
1.42812502e+00 -7.51540661e-01 -7.14779615e-01 4.25766498e-01
5.84081829e-01 -2.20295578e-01 -6.83944166e-01 4.43539262e-01
3.38988900e-01 -1.64509177e+00 1.21010745e+00 -5.19323349e-01
6.11417651e-01 -2.92851835e-01 -1.00243382e-01 -1.30105555e+00
-5.36705703e-02 -7.00892806e-01 2.54424691e-01 1.23417294e+00
2.63958991e-01 -8.33676815e-01 7.96168983e-01 5.46969950e-01
-1.97724640e-01 -6.78316534e-01 -9.22632039e-01 -7.78972447e-01
2.28209943e-01 -5.53337872e-01 2.00949550e-01 9.30170655e-01
-1.12587340e-01 3.56174618e-01 -2.23300964e-01 -2.28930444e-01
7.00827718e-01 3.46565336e-01 1.04499698e+00 -1.21636355e+00
-6.00665629e-01 -5.06664157e-01 -7.65540123e-01 -1.23070467e+00
2.63007730e-01 -7.61655152e-01 -1.49481326e-01 -1.37009811e+00
-4.91498299e-02 -4.41905767e-01 9.48800296e-02 2.91566551e-01
8.97805095e-02 6.10661983e-01 3.02619696e-01 4.36351687e-01
-1.65124536e-01 4.66947049e-01 1.34125793e+00 5.27994521e-02
7.12015405e-02 -2.46678874e-01 -6.96008503e-01 5.43720067e-01
5.62551022e-01 1.01249970e-01 -6.90575957e-01 -5.43965340e-01
-1.98232587e-02 -2.93244496e-02 7.74913967e-01 -1.09479892e+00
1.16638459e-01 1.23348661e-01 2.70046324e-01 7.04930425e-02
6.03565514e-01 -7.75269270e-01 -2.67339736e-01 3.44383776e-01
-1.94747329e-01 1.01612665e-01 6.14953339e-02 6.41603351e-01
-2.95941740e-01 -3.96214090e-02 8.20320785e-01 -3.18992049e-01
-8.91046226e-01 2.50719637e-01 -2.11123582e-02 2.24254280e-01
9.08949018e-01 -5.72583675e-01 -3.38115484e-01 -3.91432166e-01
-5.52405834e-01 2.36407351e-02 8.93467963e-01 5.57731152e-01
4.95644778e-01 -9.80619729e-01 -6.23450220e-01 3.60521466e-01
3.77831787e-01 6.24105223e-02 -8.89067596e-04 1.06219220e+00
-4.67369884e-01 5.36177695e-01 -3.82241040e-01 -9.08242166e-01
-1.07755506e+00 8.63256335e-01 3.33998948e-01 -2.15116277e-01
-8.20906937e-01 6.49050653e-01 5.50934434e-01 -3.47850144e-01
-6.93766475e-02 -4.36787546e-01 -9.61567182e-03 -1.35351464e-01
2.45759562e-01 -4.74803383e-03 2.41169423e-01 -7.50712991e-01
-2.07877010e-01 9.93883312e-01 5.53691499e-02 -2.97287554e-01
1.32620418e+00 -2.03583404e-01 -5.22158891e-02 1.80511296e-01
1.26290119e+00 -3.81002128e-02 -1.59602976e+00 -1.70683816e-01
-1.77802667e-01 -7.43742585e-01 1.25953972e-01 -3.39652598e-01
-9.39994633e-01 1.27771020e+00 6.01902544e-01 1.24599732e-01
9.74599957e-01 -4.88102296e-03 6.10842049e-01 1.30336985e-01
4.46822107e-01 -7.26161420e-01 1.68004468e-01 3.57516050e-01
7.64004469e-01 -1.68281698e+00 -1.74493358e-01 -3.85839850e-01
-4.28586245e-01 1.07541502e+00 5.49774647e-01 -4.32737648e-01
5.40695965e-01 5.80916107e-02 1.62256762e-01 -2.68816054e-01
-6.41960144e-01 -7.40407526e-01 3.77194345e-01 9.63653564e-01
3.86851311e-01 -3.71247381e-01 -3.74525450e-02 -4.28224355e-01
-2.10547954e-01 -4.15549688e-02 4.72491503e-01 6.23098314e-01
-1.41342103e-01 -1.32595289e+00 -2.33147398e-01 1.71218589e-01
-1.12009145e-01 -1.86144605e-01 -3.93019766e-01 9.57192600e-01
5.59930056e-02 8.55464458e-01 1.52279377e-01 -2.27196649e-01
2.67508596e-01 -1.52740851e-01 1.02669370e+00 -7.35603929e-01
-3.85674119e-01 -2.47261450e-01 -5.71586601e-02 -7.11031437e-01
-7.42538035e-01 -6.32507741e-01 -8.05724561e-01 -1.55114979e-01
-8.61651823e-02 -2.74896055e-01 6.44485772e-01 1.04467189e+00
3.42186928e-01 7.52159134e-02 3.77452940e-01 -1.45365083e+00
-4.67717886e-01 -6.90926850e-01 -3.05046380e-01 8.28554332e-01
4.54475164e-01 -9.67294693e-01 -4.13651109e-01 -4.92211431e-02] | [8.650961875915527, -2.315253734588623] |
19f8f0e8-7cb8-4179-811b-f8dec360d023 | robust-object-detection-via-instance-level | 2104.08381 | null | https://arxiv.org/abs/2104.08381v2 | https://arxiv.org/pdf/2104.08381v2.pdf | Robust Object Detection via Instance-Level Temporal Cycle Confusion | Building reliable object detectors that are robust to domain shifts, such as various changes in context, viewpoint, and object appearances, is critical for real-world applications. In this work, we study the effectiveness of auxiliary self-supervised tasks to improve the out-of-distribution generalization of object detectors. Inspired by the principle of maximum entropy, we introduce a novel self-supervised task, instance-level temporal cycle confusion (CycConf), which operates on the region features of the object detectors. For each object, the task is to find the most different object proposals in the adjacent frame in a video and then cycle back to itself for self-supervision. CycConf encourages the object detector to explore invariant structures across instances under various motions, which leads to improved model robustness in unseen domains at test time. We observe consistent out-of-domain performance improvements when training object detectors in tandem with self-supervised tasks on large-scale video datasets (BDD100K and Waymo open data). The joint training framework also establishes a new state-of-the-art on standard unsupervised domain adaptative detection benchmarks (Cityscapes, Foggy Cityscapes, and Sim10K). The code and models are available at https://github.com/xinw1012/cycle-confusion. | ['Trevor Darrell', 'Joseph E. Gonzalez', 'Xiaolong Wang', 'Fisher Yu', 'Benlin Liu', 'Thomas E. Huang', 'Xin Wang'] | 2021-04-16 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Wang_Robust_Object_Detection_via_Instance-Level_Temporal_Cycle_Confusion_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Wang_Robust_Object_Detection_via_Instance-Level_Temporal_Cycle_Confusion_ICCV_2021_paper.pdf | iccv-2021-1 | ['robust-object-detection'] | ['computer-vision'] | [-0.05404698 -0.36887953 -0.23872696 -0.32657912 -0.71755284 -0.60548437
0.5684946 -0.19181702 -0.4334521 0.42420977 -0.02835524 0.2255503
0.19997011 -0.38383794 -0.7862049 -0.79254544 -0.14558223 0.33908227
1.0683919 -0.10697916 0.0291035 0.36080799 -1.6019633 0.64201313
0.45842552 0.9652832 0.35868305 0.6102018 0.34402022 0.5276595
-0.41443723 -0.22615762 0.53091633 -0.19368675 -0.6535834 0.4754688
0.5897834 -0.46230775 -0.5152647 1.2186679 0.43038672 0.18507135
0.91248804 -1.3798014 -0.5803683 0.09603514 -0.71265495 0.8295726
0.17455852 0.65167487 0.76308227 -1.2226686 0.85508996 1.3168817
0.5101943 0.777659 -1.1750412 -0.8393566 0.34036756 0.5161962
-1.4157138 -0.40509275 0.7356763 -0.47174767 0.871272 -0.08843443
0.372097 1.4745556 -0.02375794 1.1231852 0.902794 -0.27510843
0.41371968 0.28345975 0.11829104 0.4855847 0.4279962 0.36821797
-0.4431122 -0.01329291 0.71597224 0.0296832 -0.12431487 -0.7271467
-1.1234359 0.65249604 0.5401179 0.09328957 -0.02615602 -0.18258978
0.5011442 0.30862832 0.45418745 0.14866394 -0.610076 0.09568708
-0.6223553 0.23961869 0.46518427 1.2639422 0.58171374 -0.07791804
-0.25292894 0.68179697 0.3181509 0.57842547 0.6611342 -0.8413542
0.5166641 0.63952357 0.02130911 -0.7051824 -0.28061637 -0.52078015
-0.6591177 0.22105889 0.4577649 0.09587008 -1.1113551 1.5704267
0.5967523 0.32302418 0.10449109 0.9933291 0.96922505 0.41415572
0.23473388 -0.09633931 1.376011 -1.1537493 -0.26727688 -0.45050582
0.64766586 -0.71461195 0.87544316 0.10233359 -0.74462515 -0.9718351
-0.84442675 0.20958556 -0.4571228 0.16178426 0.17443952 0.3282159
-0.7534856 0.31418633 -0.95258355 -0.6319814 0.8046085 0.05969013
-0.45157278 -0.25373593 -0.9644652 0.85949147 0.79906017 -0.29691765
-1.2817509 -0.47259343 -0.81943154 -0.27536017 0.5843536 -0.4445707
1.3553303 -1.0927798 -0.9944319 1.1812105 -0.00969972 -0.58107644
0.7585632 -0.22188123 -0.71803236 0.24420464 0.37863928 1.0504181
1.2290033 -1.1427717 -0.9472127 -0.40243238 -0.22427098 0.20099905
-0.24767104 0.18782556 -0.83126944 -0.81847966 -0.01025603 -1.1376376
-0.00838015 0.14449322 -0.27298746 -0.5459066 1.2757777 -0.29902902
0.93698955 -2.5520697 -0.1270001 -0.20455703 0.05887427 0.44192663
-0.36229432 -0.12261526 -0.20809577 -0.25880432 -0.21825276 -0.34162825
-0.311151 0.18020023 -0.21655953 0.56713116 0.61202276 0.8573349
-0.9531968 -0.6535543 0.2035929 0.01271035 -0.5405395 0.20844229
-0.2388546 0.39613596 -0.4930999 0.89229673 0.72978663 -0.43549073
-0.15250736 -0.17247978 0.0731009 0.02075322 -1.3853775 1.5260687
0.1501386 0.7119508 -0.35070282 -0.96451885 0.70829266 0.05148517
0.26333383 -0.7233527 -0.0054694 0.09073392 0.02953842 -0.43833956
0.23548485 0.44279397 0.0214496 0.12816857 0.4449161 0.08150299
0.4582605 0.2742842 1.0242923 0.16563149 0.45331836 -0.3557581
0.3459093 0.11820835 0.6517309 0.86941934 -0.7842514 0.72659194
0.1603014 -0.5357225 -0.943994 -1.2737154 -0.38267696 1.2821466
0.55566126 -0.13025306 -0.51813394 -1.0969704 0.13305892 0.52448004
-0.65074664 -0.3419231 -0.37969846 -0.7008216 0.4124471 0.78789866
0.8554524 -1.1037221 -0.71999955 0.15022297 -0.20979744 -1.4338338
-0.6324304 0.30840722 -0.9130063 -1.215125 -0.76716375 -0.86118656
0.60683274 0.6491099 1.1738143 -0.30090445 -0.59759563 0.451757
-0.46923092 -0.38591945 -0.47139385 -0.1153291 0.3580343 -0.03704375
0.6270232 -0.18186311 -0.6898845 0.8715203 -0.84808403 -0.1916126
0.45733166 0.6585588 0.7573018 0.05496784 0.4013547 -0.5802629
-0.05297098 -0.5221729 -0.6989783 0.12974145 -0.31093165 -0.04905742
0.24644798 -0.7485429 -1.1247123 0.38653615 0.3399051 -0.75921756
-0.4674367 -0.29639944 -0.09094631 0.09668288 1.0596598 0.32721117
-0.22544591 -0.35607004 0.20732166 0.505788 0.73114413 -0.288242
1.0296735 0.7293554 -0.405013 -0.83940446 -0.90622586 -1.0233208
-0.78872794 -0.28993112 1.054759 -1.4028233 -0.08471508 0.57243586
-0.98203504 -0.3803353 -0.38225156 0.44116628 -0.5283047 0.38415742
-0.27081853 -0.6436298 -0.02819082 -0.9944 1.191748 0.3444865
-0.03032742 -0.76481867 0.00754264 0.20539811 -0.0259463 0.07836416
0.34398776 -0.95547473 -0.64694023 -0.05450437 -0.32464623 0.658566
0.19665489 -0.0902428 -1.0428253 -0.48203832 -0.0648514 -0.54024625
1.1881332 0.35665107 1.0406736 -0.05411306 -0.57495284 0.6057433
1.2014171 0.08094161 0.5553613 0.4889634 0.44180202 0.2526928
0.93248713 0.4181616 0.17228691 0.8841675 0.39370817 -0.12561414
-0.3552668 -0.15096025 0.61687076 0.11273252 -0.02687254 -0.19096315
-0.82591057 0.69496274 -1.888356 -1.1266255 -0.06608022 2.038643
0.61617386 0.5064673 0.45547095 -0.18096718 0.93231887 0.03481999
-0.8781249 0.27275124 -0.381336 -0.22426562 0.57886285 -0.11069703
-1.7063057 1.0107356 5.330649 0.89668363 -0.81433195 0.33761153
0.69486076 -0.07347161 0.57387674 -0.14719471 -1.1993917 0.56730855
0.3604206 0.13033155 0.04185342 1.2827061 0.01958246 -0.11564487
-1.311753 0.9389276 0.12886988 -1.1780355 -0.1172753 -0.23251216
1.0013075 0.45522436 0.17466912 0.45473355 0.31948072 -0.35771054
0.62462103 -0.10589148 0.67481005 -0.28569147 0.5435653 0.31028143
-1.3921705 -0.30927938 -0.5405538 0.27614003 -0.1240982 0.23479585
-0.9419561 0.28092837 1.2935313 0.9606689 -0.8589537 1.1117538
-0.2539929 0.5876279 -0.39651135 0.26984182 0.23210894 0.38218102
0.8987802 1.4826146 -0.03272041 0.10565206 0.3493952 0.62926346
0.06003467 -0.30107588 -0.5248557 0.37504992 0.45000172 1.0530832
-0.91951 -0.55149615 -0.46166426 1.1186255 0.19603944 0.49861616
-1.0343175 0.19693193 0.79640836 0.2325418 0.78220505 -0.08135724
0.08034674 -1.3280357 0.1400331 -0.9141201 0.9615658 -0.56963784
-1.5240302 0.45420995 0.17337252 -1.8254972 -0.04102964 -0.6730274
-0.687773 0.20273158 -1.521315 -0.94666076 -0.42645988 0.92222077
1.1153226 -0.4269873 0.3947282 0.17557722 -0.62711257 0.69962174
0.30912367 0.4254972 0.96730477 -1.0285013 0.58877033 0.9793602
0.26546884 0.14797935 0.5811868 -0.6663596 -1.0304537 -1.6367983
0.33082563 -0.65749896 0.6029489 -0.3687235 -1.0289713 0.5763117
-0.16852891 0.69714737 0.14719276 -0.27357388 -0.6706167 -0.03877833
-1.0821632 0.43844914 1.4787816 -0.38793385 -0.8295283 0.6409416
0.7929103 -0.44991678 -0.3728515 0.5100633 0.312166 -0.9882854
1.1355759 -0.8528768 0.21631874 -0.53860915 -0.21315339 -1.0444372
-0.63074136 -0.35028967 -0.35079834 0.9914523 0.2770696 -0.5165309
0.7112328 0.32842165 -0.01795384 -0.34966975 -1.0520641 -1.3028331
-0.29116988 -0.44159365 0.05504312 0.8490586 -0.34560204 0.11237448
-0.06910516 0.5598533 0.8498918 0.09498847 0.86793005 -1.0064653
-0.3817706 -0.28084654 -0.8849794 -1.1104925 0.03743327 -0.7384474
0.1079332 -0.9935323 0.47274724 -0.16325246 -0.5072187 0.46111703
-0.32761532 0.40298477 0.34380218 0.56743306 -1.2360476 0.50991696
1.1387985 -0.29247063 -0.36190516 0.25258875 -0.30324623 0.89355135
0.6952951 -0.71199036 -0.2473363 -0.35973695 -0.570011 -0.310414
0.742121 -1.3191733 0.17012274 0.012842 0.6908888 -0.5909046
0.1729837 -0.8324448 -0.2770825 0.70070904 -0.14943765 0.01765597
0.44190148 0.9549034 -0.19894883 -0.02173555 1.1967218 -0.10100736
-1.4363594 0.321646 -0.1156251 0.40270668 1.3645042 -0.5311386
-0.3614172 -0.02237857 -0.7610957 0.38289306 0.38021538 0.8239137
0.63991445 -1.2550529 -0.8004466 0.21667556 0.6992236 0.16283801
0.37867433 0.62913716 -0.0748541 0.28545025 -0.21865006 -1.2565973
-1.3617238 0.7007742 0.31496063 -0.11133569 -0.56479603 1.0297191
0.7828147 -0.01621398 0.33099142 -0.38193518 0.02996829 0.01038063
0.6136829 0.32136855 0.00911169 -0.5599159 -0.6845533 0.41018173
-0.36299542 0.20576894 1.0411881 -0.06776828 0.57663983 0.15006709
1.1952668 -0.60190815 -1.8620244 -0.6671072 0.05123382 -0.5117462
-0.23355214 -0.5881194 -0.9175968 0.54277456 1.1742591 -0.13112634
1.1314348 0.5052988 0.33883402 0.5868296 0.29708537 -1.3791922
0.6257043 0.54797566 0.8383034 -1.8107951 -0.09998461 -0.32313183
-0.9869578 0.82513434 1.0390056 -0.16208401 0.64931464 0.0980308
0.01636245 -0.03664232 -0.76109266 -0.48476905 0.421898 0.8457892
-0.0891213 -0.03278654 0.30753165 0.39751962 0.40830112 -0.1291256
0.19868843 0.90793276 -0.5779836 -0.5104832 -0.4408829 0.3381804
-0.0473876 0.13502975 -0.30560884 0.9466199 0.29797992 0.7801526
0.20307523 -0.28134477 0.4261364 0.05898113 0.26205695 -0.7459519
-0.12468237 0.09383859 -0.15468019 -0.5701015 -0.55042505 -0.955019
-1.1480935 0.26268202 -0.43562725 -0.35541672 0.20133959 0.8523967
0.4320604 0.33899334 0.67825675 -0.8412385 -0.7561545 -0.9454929
-0.31489953 0.78661156 0.35383478 -0.99930483 -0.3648738 0.46975702] | [9.521862983703613, 1.6226415634155273] |
5d7ea6da-9a66-48a7-af46-b49d2b99fe2f | scibert-pretrained-contextualized-embeddings | 1903.10676 | null | https://arxiv.org/abs/1903.10676v3 | https://arxiv.org/pdf/1903.10676v3.pdf | SciBERT: A Pretrained Language Model for Scientific Text | Obtaining large-scale annotated data for NLP tasks in the scientific domain is challenging and expensive. We release SciBERT, a pretrained language model based on BERT (Devlin et al., 2018) to address the lack of high-quality, large-scale labeled scientific data. SciBERT leverages unsupervised pretraining on a large multi-domain corpus of scientific publications to improve performance on downstream scientific NLP tasks. We evaluate on a suite of tasks including sequence tagging, sentence classification and dependency parsing, with datasets from a variety of scientific domains. We demonstrate statistically significant improvements over BERT and achieve new state-of-the-art results on several of these tasks. The code and pretrained models are available at https://github.com/allenai/scibert/. | ['Kyle Lo', 'Iz Beltagy', 'Arman Cohan'] | 2019-03-26 | scibert-a-pretrained-language-model-for | https://aclanthology.org/D19-1371 | https://aclanthology.org/D19-1371.pdf | ijcnlp-2019-11 | ['participant-intervention-comparison-outcome', 'medical-named-entity-recognition', 'citation-intent-classification'] | ['medical', 'natural-language-processing', 'natural-language-processing'] | [ 1.23645151e-02 -2.50663664e-02 -2.81680554e-01 -7.43026316e-01
-1.48870575e+00 -1.16233957e+00 4.71779913e-01 5.64186931e-01
-4.71217215e-01 1.17374659e+00 8.84931013e-02 -5.43532789e-01
2.48529553e-01 -3.31350684e-01 -1.00497174e+00 -3.76706898e-01
8.56809467e-02 6.16513968e-01 1.29004437e-02 4.29720640e-01
3.83211493e-01 4.89879102e-01 -6.49610162e-01 5.71078002e-01
8.45624924e-01 3.04772139e-01 4.67442214e-01 9.25318599e-01
-5.08961201e-01 4.91294861e-01 -6.36109233e-01 -6.58471107e-01
-7.37737343e-02 -3.88774097e-01 -1.08339763e+00 -5.67244470e-01
3.70992482e-01 3.02141458e-01 -7.18826149e-03 7.96018064e-01
5.79068065e-01 -2.49283239e-01 5.35391629e-01 -8.35738361e-01
-7.58991361e-01 8.56570959e-01 -4.37682062e-01 5.08164585e-01
5.42761795e-02 2.21892893e-01 1.01869130e+00 -7.56543517e-01
1.32313335e+00 9.61835682e-01 5.96750975e-01 6.61938488e-01
-1.26026309e+00 -8.89005303e-01 -1.91347003e-01 -2.25989252e-01
-1.00721502e+00 -6.77000344e-01 2.16934159e-01 -5.21566331e-01
9.97473061e-01 -4.53692786e-02 2.61152256e-02 1.45848143e+00
2.99259722e-01 6.03397429e-01 1.22570825e+00 -3.01290333e-01
1.63955450e-01 -3.09341937e-01 5.21199524e-01 7.57912934e-01
2.82408953e-01 -3.75817895e-01 -7.45460510e-01 -3.32509279e-01
3.34108919e-01 -3.88570189e-01 3.64246033e-02 1.76801488e-01
-1.32088387e+00 6.12530768e-01 -1.14227951e-01 4.49319810e-01
-1.86726674e-01 5.10669164e-02 5.28502345e-01 1.60150677e-01
8.90168190e-01 1.02016699e+00 -1.24160242e+00 -5.75408399e-01
-6.56664133e-01 -1.02038672e-02 1.28900635e+00 1.11448371e+00
3.08482230e-01 -2.87243605e-01 -2.72743404e-01 7.66059339e-01
3.23557220e-02 4.01059002e-01 1.87220514e-01 -1.24404895e+00
3.92277509e-01 1.28068909e-01 8.07459950e-02 -2.12267399e-01
-4.96607184e-01 -4.91677552e-01 -4.11801010e-01 -2.03693390e-01
6.97939098e-01 -5.49607635e-01 -1.04472172e+00 1.71828055e+00
3.14944774e-01 4.37641978e-01 2.16782764e-01 6.19278729e-01
1.27020752e+00 8.00624311e-01 7.67655134e-01 -2.38738898e-02
1.50810218e+00 -1.09325039e+00 -5.38995922e-01 -2.56701708e-01
9.61477101e-01 -1.03123081e+00 8.16258729e-01 3.49097729e-01
-9.80838180e-01 -3.22589904e-01 -5.42476773e-01 -6.22099757e-01
-6.16231620e-01 4.30757821e-01 7.76685238e-01 1.06493443e-01
-9.54603434e-01 8.13779652e-01 -1.08999336e+00 -4.33757842e-01
6.70854688e-01 -1.48782833e-03 -5.13954520e-01 -1.43422469e-01
-8.93349051e-01 6.44382060e-01 2.42968529e-01 -1.43907130e-01
-1.11625874e+00 -1.24365628e+00 -5.78570127e-01 9.76964459e-02
-6.14977302e-03 -4.62073058e-01 1.53367364e+00 -2.45042682e-01
-1.33775151e+00 1.47342610e+00 -2.88548619e-01 -3.61886770e-01
2.19883770e-01 -4.04014587e-01 -2.85833418e-01 1.40428245e-01
2.71069139e-01 7.32779324e-01 -5.17558232e-02 -9.92760181e-01
-3.97100419e-01 -3.18660289e-01 -4.03948426e-01 -2.76358575e-01
-8.67233053e-02 3.14363122e-01 -5.56129575e-01 -5.07310390e-01
-6.79990292e-01 -8.16690147e-01 -2.24632904e-01 3.44250957e-03
-4.74058598e-01 -4.57144707e-01 4.18565810e-01 -1.00306737e+00
4.78376985e-01 -1.85630512e+00 9.29077268e-02 -4.81539965e-01
2.73202181e-01 3.12721968e-01 -6.59680843e-01 3.85453552e-01
-8.49498585e-02 5.12852073e-01 -3.86681557e-01 -7.41179407e-01
-1.97774366e-01 1.60152569e-01 -1.17230356e-01 4.07096356e-01
4.00739372e-01 1.11337626e+00 -1.16152430e+00 -4.91601437e-01
-4.31568980e-01 3.51088732e-01 -3.79354693e-02 4.85639215e-01
-6.59963667e-01 9.89950478e-01 -8.26908469e-01 7.29048729e-01
5.70651889e-01 -5.57492673e-01 2.98511982e-01 1.52312577e-01
-4.24630910e-01 6.98304534e-01 -1.63135782e-01 2.39139533e+00
-3.26217830e-01 5.23220360e-01 4.94008660e-01 -1.05270505e+00
8.92008066e-01 2.47006446e-01 4.19148475e-01 -1.29758954e-01
2.05584150e-02 1.98841751e-01 -5.37961815e-03 -7.12946832e-01
3.31227072e-02 1.64792761e-01 -2.09174514e-01 4.10522640e-01
5.52868903e-01 -1.36974007e-01 6.11844122e-01 4.97909069e-01
1.56658328e+00 6.84623718e-01 5.64930998e-02 -5.33729434e-01
1.92276821e-01 5.40443778e-01 9.73717391e-01 7.15661108e-01
-2.17353344e-01 4.54216421e-01 8.30165148e-01 -5.88330105e-02
-1.15338886e+00 -8.98131549e-01 -4.36370105e-01 1.18306708e+00
-5.82675159e-01 -4.75436538e-01 -3.60077739e-01 -1.03115988e+00
1.87538877e-01 7.69106150e-01 -4.36946154e-01 2.74914175e-01
-3.81557196e-01 -1.00897408e+00 8.79687071e-01 2.73157388e-01
1.50794834e-02 -9.76235569e-01 1.40906528e-01 6.82576075e-02
-5.26315309e-02 -1.38440275e+00 -3.02950352e-01 4.89241660e-01
-7.30762661e-01 -1.09017515e+00 -5.82972646e-01 -9.21220779e-01
5.78801513e-01 -1.90031216e-01 1.50481164e+00 -3.32211286e-01
-6.02336943e-01 1.18534069e-03 -3.30176651e-01 -7.68481970e-01
-5.82666457e-01 3.93859476e-01 -4.08726931e-01 -9.29002464e-01
4.79491442e-01 -3.50102991e-01 -2.42115542e-01 -3.26208949e-01
-5.41046917e-01 1.61854252e-01 7.34882176e-01 7.61548996e-01
8.08107734e-01 -7.17618108e-01 9.69068766e-01 -1.47177911e+00
3.76323104e-01 -7.75530040e-01 -8.56609762e-01 3.63219082e-01
-3.28320175e-01 1.61055431e-01 7.89525926e-01 -1.20970830e-01
-9.70799744e-01 -5.15750423e-02 -5.37973225e-01 -1.84722841e-01
-5.15955806e-01 7.19784677e-01 6.44609928e-02 -5.63635752e-02
3.24402034e-01 -1.50863240e-02 -2.34795749e-01 -1.05395663e+00
4.59407628e-01 7.37026691e-01 5.04840076e-01 -9.16596949e-01
3.49104285e-01 2.69984249e-02 1.55796528e-01 -4.03926581e-01
-1.51868069e+00 -5.16231775e-01 -7.76384413e-01 3.84355754e-01
1.10180211e+00 -1.06713438e+00 -6.92182899e-01 2.44199917e-01
-1.32284784e+00 -7.92802215e-01 -8.47146884e-02 4.31234807e-01
-1.10538863e-01 2.69080400e-01 -1.32518792e+00 -3.50645512e-01
-7.12609351e-01 -8.38864148e-01 1.21065283e+00 1.87083513e-01
-1.75220981e-01 -9.71413970e-01 3.26827437e-01 5.17011702e-01
-6.31172359e-02 4.79168408e-02 9.29074883e-01 -1.23538339e+00
-2.81335115e-01 -1.09275859e-02 -2.51958787e-01 1.54606551e-01
-7.25918859e-02 1.70854971e-01 -9.13913667e-01 -1.17191806e-01
-4.04170752e-01 -7.81485081e-01 9.73051369e-01 1.64736599e-01
1.67503202e+00 2.33871713e-02 -5.40829241e-01 7.33837485e-01
1.06963360e+00 5.98920882e-02 5.56934848e-02 3.01080167e-01
6.66413128e-01 4.93420124e-01 5.98594725e-01 2.76015252e-01
2.45227084e-01 6.23637028e-02 -2.94872403e-01 -1.64940476e-01
-1.45927683e-01 -1.20767049e-01 1.72367115e-02 9.09201741e-01
2.46398315e-01 -6.63719177e-01 -1.29174864e+00 6.52041972e-01
-1.63472080e+00 -5.85078657e-01 -3.21782649e-01 1.81251049e+00
1.56448078e+00 2.17138436e-02 -2.65445888e-01 -9.29297149e-01
5.32665610e-01 -7.73311406e-02 -6.69099152e-01 -4.36145395e-01
-2.06815466e-01 5.32700062e-01 6.74136579e-01 5.21249354e-01
-1.28333950e+00 1.33819687e+00 6.45392227e+00 8.58557224e-01
-1.03552437e+00 4.32163477e-01 1.12084985e+00 -1.40028134e-01
-2.72669911e-01 1.73230261e-01 -1.11568713e+00 5.74265003e-01
1.56008840e+00 -2.61264205e-01 2.04567071e-02 6.81438386e-01
2.62916505e-01 1.31297916e-01 -1.05046129e+00 3.42780888e-01
-2.71001607e-01 -1.71953702e+00 -3.56708735e-01 -1.87155381e-01
6.16573274e-01 7.54709840e-01 -3.31036121e-01 4.37321335e-01
1.05619931e+00 -1.10903239e+00 1.01503819e-01 4.64547098e-01
8.49281013e-01 -4.12601531e-01 7.64448166e-01 2.20186934e-01
-5.03371716e-01 6.45467579e-01 -4.87366110e-01 1.60815120e-01
1.10949218e-01 1.26646531e+00 -6.80142164e-01 5.43779433e-01
7.17529297e-01 1.20597970e+00 -5.81940711e-01 8.33226681e-01
-6.25786483e-01 1.29813802e+00 -2.34363079e-01 3.83250229e-02
-8.84754304e-03 -2.59103894e-01 4.59678322e-01 1.74481428e+00
1.77537382e-01 3.45457733e-01 2.47666657e-01 1.09315777e+00
-9.44992065e-01 -2.81031281e-02 -4.57695007e-01 -8.86419237e-01
5.39885342e-01 1.89641094e+00 -6.61866069e-01 -4.65655625e-01
-4.69682097e-01 8.17186713e-01 8.66117299e-01 4.28876817e-01
-5.56958020e-01 -4.04183745e-01 7.28123486e-01 -5.15562356e-01
1.65653259e-01 -4.73866731e-01 -6.04100704e-01 -1.25806177e+00
-3.64130467e-01 -6.09550834e-01 5.97887874e-01 -7.77007639e-01
-1.67323983e+00 2.75591910e-01 -4.94033813e-01 -6.16191983e-01
2.10768566e-01 -7.11235821e-01 -5.59349835e-01 1.01713717e+00
-1.62038469e+00 -1.20349610e+00 6.30613640e-02 -1.50964096e-01
5.61502993e-01 -2.09684316e-02 1.13625073e+00 1.70637503e-01
-9.30073142e-01 2.30954722e-01 4.36367750e-01 2.02760577e-01
1.31784999e+00 -1.31623185e+00 8.28755677e-01 6.82478607e-01
-1.42166495e-01 8.11151028e-01 5.54163158e-01 -9.75752115e-01
-1.75507724e+00 -1.36427140e+00 1.39665437e+00 -7.96272755e-01
1.00581837e+00 -8.81576836e-01 -1.00939560e+00 9.49755073e-01
4.17106390e-01 1.61525095e-03 1.15286684e+00 4.78995860e-01
-3.58162194e-01 3.41867536e-01 -1.07072651e+00 1.52462825e-01
1.19440901e+00 -3.46897274e-01 -4.16574866e-01 9.78051484e-01
7.94284701e-01 -5.83186567e-01 -1.51908541e+00 1.74274847e-01
1.57804608e-01 -4.23114076e-02 9.27243471e-01 -1.02570724e+00
1.10750449e+00 1.90560028e-01 4.92405832e-01 -1.30635977e+00
-4.83206093e-01 -5.01621008e-01 1.42684788e-01 1.61355031e+00
9.05524254e-01 -5.23751736e-01 7.21279263e-01 5.33682764e-01
-4.86089528e-01 -6.35133922e-01 -6.71709716e-01 -6.93835080e-01
6.19600832e-01 5.71948513e-02 1.20330148e-03 1.19587290e+00
2.72471040e-01 7.24783540e-01 3.67018133e-02 6.51016682e-02
6.54703736e-01 4.81466740e-01 4.30196404e-01 -1.12674916e+00
-2.93219775e-01 -1.34510353e-01 1.17292389e-01 -7.04361081e-01
8.14659894e-01 -1.44755447e+00 2.09086165e-01 -1.87675905e+00
4.64214087e-01 -3.94863516e-01 -3.52218360e-01 1.02294135e+00
-4.43239659e-01 4.79552113e-02 -6.83102682e-02 3.86491954e-01
-9.56832707e-01 3.24789971e-01 1.17163384e+00 -2.96423268e-02
2.56654173e-01 -5.08775711e-01 -7.70208180e-01 4.14727867e-01
1.08206499e+00 -8.46990347e-01 1.09644480e-01 -7.61423349e-01
-1.70088455e-01 2.53292713e-02 9.97897703e-03 -5.41741192e-01
8.35278407e-02 -3.45831752e-01 6.46022081e-01 -2.80269116e-01
6.59346431e-02 -7.27705732e-02 -2.26103559e-01 3.46750706e-01
-9.47150528e-01 2.00530328e-02 7.51686096e-01 3.38288665e-01
8.17355961e-02 -3.13865602e-01 7.53254652e-01 -2.72811055e-01
-5.32163620e-01 3.31820071e-01 -2.65332788e-01 4.32916105e-01
7.93124914e-01 1.01072145e+00 -9.82232273e-01 9.37618911e-02
-5.78465343e-01 5.96627474e-01 3.58342350e-01 2.85260618e-01
-2.73644924e-02 -4.33816135e-01 -1.00145733e+00 -3.06237936e-01
-5.84143810e-02 1.35686010e-01 -4.77098897e-02 7.28292227e-01
-6.35960996e-01 7.22095370e-01 -1.18629023e-01 -1.31816298e-01
-1.20084727e+00 5.03306329e-01 -1.23427913e-01 -5.79342902e-01
-4.06585068e-01 1.17721403e+00 1.95677176e-01 -8.55542600e-01
-1.20042957e-01 -8.28071609e-02 -1.32661328e-01 -4.34360564e-01
2.82515287e-01 6.93159029e-02 -3.75837162e-02 4.99724224e-02
-4.03356940e-01 1.47313878e-01 -2.14348599e-01 7.22443834e-02
1.62402046e+00 1.42512962e-01 -4.66613203e-01 4.78912354e-01
1.30678570e+00 1.74943432e-01 -9.86268759e-01 -7.42555708e-02
3.24064314e-01 -1.29904553e-01 3.20758633e-02 -1.33691120e+00
-8.55999053e-01 8.61707807e-01 -2.46374793e-02 -5.14101982e-01
4.97753799e-01 2.34181151e-01 8.01953614e-01 4.89181131e-01
1.44629285e-01 -8.21860313e-01 -4.26563740e-01 7.04411685e-01
6.10445082e-01 -1.38533211e+00 9.16229561e-02 -6.01460218e-01
-5.92384517e-01 9.34477091e-01 6.38659656e-01 1.95145726e-01
5.30527651e-01 9.27399457e-01 2.58640528e-01 -3.24834645e-01
-1.07695925e+00 8.53070766e-02 -1.11789219e-01 3.84096026e-01
1.20634687e+00 -7.50277638e-02 -8.55557084e-01 9.28451359e-01
-2.97237821e-02 5.43512404e-01 4.13613677e-01 1.17267478e+00
-1.94982097e-01 -1.45796442e+00 1.17503956e-01 5.88819981e-01
-1.04305112e+00 -4.55632776e-01 -6.90651298e-01 2.87513018e-01
-1.85646906e-01 9.59089100e-01 -2.30379449e-03 2.04853460e-01
-5.09853959e-02 3.35933447e-01 3.06414366e-01 -9.76787031e-01
-6.06636524e-01 1.84510812e-01 6.10371053e-01 -3.24911803e-01
-2.97259003e-01 -7.68654883e-01 -1.59717584e+00 -2.47244194e-01
3.79140347e-01 7.46274114e-01 8.74307513e-01 7.07250893e-01
1.10939968e+00 7.37315536e-01 1.50655843e-02 -4.98759091e-01
-3.01151276e-01 -1.15067661e+00 -3.14887792e-01 3.91422510e-01
-8.03409964e-02 -6.98681548e-02 -9.64486450e-02 4.70537066e-01] | [9.114622116088867, 8.54814338684082] |
2a7100b2-bc14-45cd-beec-9a39521cdf60 | a-projective-geometric-view-for-6d-pose | 2302.00227 | null | https://arxiv.org/abs/2302.00227v2 | https://arxiv.org/pdf/2302.00227v2.pdf | A Projective Geometric View for 6D Pose Estimation in mmWave MIMO Systems | Millimeter-wave (mmWave) systems in the 30--300 GHz bands are among the fundamental enabling technologies of 5G and beyond 5G, providing large bandwidths, not only for high data rate communication, but also for precise positioning services, in support of high accuracy demanding applications such as vehicle positioning. With the possibility to introduce relatively large arrays on user devices with a small footprint, the ability to determine the user orientation becomes unlocked. The estimation of the full user pose (joint 3D position and 3D orientation) is referred to as 6D localization. Conventionally, the problem of 6D localization using antenna arrays has been considered difficult and was solved through a combination of heuristics and optimization. In this paper, we reveal a close connection between the AoA and AoD and the well-studied perspective projection model from computer vision. This connection allows us to solve the 6D localization problem, by adapting state-of-the-art methods from computer vision. More specifically, two problems, namely 6D pose estimation from AoA from multiple single-antenna base stations and 6D SLAM based on single-BS mmWave communication, are first modeled with the perspective projection model, and then solved. Numerical simulations show that the proposed estimators operate close to the theoretical performance bounds. Moreover, the proposed SLAM method is effective even in the absence of the LoS path, or knowledge of the LoS/NLoS condition. | ['Henk Wymeersch', 'Shengqiang Shen'] | 2023-02-01 | null | null | null | null | ['6d-pose-estimation-1'] | ['computer-vision'] | [-2.76447982e-01 2.72167940e-02 -1.62367299e-01 -9.41791162e-02
-5.40980756e-01 -7.02163041e-01 2.29746729e-01 8.95446818e-03
-1.43893272e-01 6.84494913e-01 -2.60403454e-01 -6.39774799e-01
-5.86399615e-01 -8.02964032e-01 -4.07842994e-01 -9.96219635e-01
-2.46693730e-01 6.22185886e-01 -5.42952120e-01 -5.08840233e-02
2.98167616e-02 8.78590226e-01 -9.13002074e-01 -9.12160516e-01
9.10883546e-01 1.30358160e+00 2.08873555e-01 3.64771664e-01
-2.87697595e-02 -2.38228589e-01 -6.10980034e-01 -2.36938208e-01
2.75516003e-01 -2.23272637e-01 1.32966608e-01 2.90233433e-01
1.45009741e-01 -8.90130624e-02 -1.82074800e-01 1.01382172e+00
6.82633698e-01 -2.04778388e-01 3.43684167e-01 -1.18131661e+00
1.03612833e-01 -1.00295253e-01 -4.64230478e-01 5.73755540e-02
4.28001285e-01 -3.15690428e-01 5.87109149e-01 -8.25639606e-01
4.60838377e-01 5.84876835e-01 5.97895503e-01 -1.05183825e-01
-7.24725246e-01 -5.41129708e-01 -2.48316139e-01 1.01752065e-01
-1.67850053e+00 -5.27821600e-01 7.13067591e-01 -3.26691628e-01
1.83622286e-01 4.21660781e-01 8.67554486e-01 5.54009557e-01
2.74422169e-01 7.60592744e-02 6.99606836e-01 -6.53527081e-01
3.92457128e-01 1.30595610e-01 -2.50571091e-02 7.57683516e-01
1.06946659e+00 -1.24578759e-01 -2.72752941e-01 -2.33070061e-01
7.65841782e-01 3.27271149e-02 -6.08462811e-01 -8.40988696e-01
-1.25578761e+00 5.22976637e-01 5.62860429e-01 6.57299817e-01
-4.79938686e-01 8.43970329e-02 -7.07340598e-01 1.89790905e-01
3.73928696e-01 6.34573340e-01 -1.05088830e-01 -1.04417410e-02
-6.83011472e-01 -9.90622118e-02 8.75127792e-01 1.42704487e+00
6.76886737e-01 1.66945904e-01 7.17917904e-02 3.08289587e-01
7.59941220e-01 1.13229048e+00 6.61938125e-03 -4.32465076e-01
7.68717229e-01 1.34163037e-01 6.02078557e-01 -1.21190250e+00
-8.03326786e-01 -1.76316309e+00 -1.05463672e+00 -2.96206117e-01
5.51579952e-01 -5.00147104e-01 -6.56310797e-01 1.60730791e+00
4.28805798e-01 4.11080301e-01 1.53682426e-01 7.47286141e-01
4.62578416e-01 4.87369239e-01 -7.85594940e-01 -7.71998167e-01
1.22519946e+00 -4.10307676e-01 -7.10646808e-01 -5.63515484e-01
5.11664987e-01 -7.10179806e-01 1.28736302e-01 4.71685976e-01
-7.42703199e-01 -1.42323762e-01 -1.32327735e+00 6.53002918e-01
6.16535693e-02 4.54170346e-01 5.10591686e-01 1.19568372e+00
-8.50896478e-01 -3.04720849e-01 -6.10350668e-01 -5.13561904e-01
2.16345027e-01 3.20420623e-01 -7.77549446e-02 -3.38141412e-01
-7.09577978e-01 7.79688776e-01 -2.56519943e-01 5.57799041e-01
-1.67077444e-02 -5.63406050e-01 -4.93481606e-01 2.36831352e-01
3.59456420e-01 -7.84895241e-01 9.31949437e-01 -1.42543867e-01
-1.37940729e+00 4.16670114e-01 -5.58859885e-01 -4.17480409e-01
4.06583905e-01 1.27055375e-02 -5.52506149e-01 -1.24723338e-01
1.56895190e-01 -4.99377698e-01 6.09098315e-01 -1.29828799e+00
-5.51063418e-01 -8.90857041e-01 -9.61838961e-02 7.99324214e-02
3.61066982e-02 -7.26170897e-01 -5.12028933e-01 -2.50651658e-01
1.15295601e+00 -1.18479037e+00 -5.89333475e-01 -3.13545108e-01
-4.25459743e-01 4.28931087e-01 3.48697752e-01 -3.64312083e-01
9.39745486e-01 -2.16052508e+00 1.13144577e-01 7.65550196e-01
2.43846968e-01 -7.49944374e-02 2.53852099e-01 4.47533518e-01
1.82610124e-01 -1.64596930e-01 1.59156129e-01 -3.66646409e-01
-1.91153944e-01 -1.54143915e-01 -1.53531328e-01 8.68613482e-01
-8.20686162e-01 5.39225459e-01 -7.67069340e-01 2.10700616e-01
2.30945140e-01 2.67728537e-01 -3.23284924e-01 -4.96855900e-02
3.58338058e-01 7.56382704e-01 -8.57168794e-01 6.38206482e-01
1.10579312e+00 -1.30087495e-01 3.42630863e-01 -2.31549084e-01
-4.17530715e-01 -1.78039238e-01 -1.42496538e+00 1.64288890e+00
-8.36917639e-01 4.07356083e-01 3.19646299e-01 -1.00672412e+00
1.13644075e+00 3.34567100e-01 6.23178363e-01 -6.49790645e-01
7.63838440e-02 5.93126297e-01 2.31246781e-02 -2.94990152e-01
-7.18741417e-02 1.96086406e-03 -3.20760012e-01 2.29473084e-01
-5.37139736e-02 -4.23979968e-01 -1.36078253e-01 -1.29018545e-01
8.64425421e-01 -4.38040137e-01 7.54883111e-01 -3.05449694e-01
7.26168931e-01 -3.27761948e-01 5.49227655e-01 8.01691353e-01
2.60988683e-01 2.57741749e-01 2.30386674e-01 -7.79635757e-02
-4.11396682e-01 -1.10858560e+00 -2.15355560e-01 -1.46924078e-01
6.02485299e-01 -2.41054833e-01 -3.26970637e-01 -3.21045160e-01
2.45419905e-01 6.34366333e-01 5.69205657e-02 1.21688068e-01
-4.42011058e-01 -8.91529441e-01 -6.92848787e-02 -1.76793754e-01
6.57688320e-01 2.52340049e-01 -3.12223822e-01 3.37496817e-01
-1.42543718e-01 -1.35931909e+00 1.44169331e-01 -2.24411301e-02
-7.08247542e-01 -7.95217335e-01 -7.99149811e-01 -5.42757392e-01
7.18446076e-01 8.71328592e-01 7.16499090e-01 -1.77369311e-01
-1.55309448e-02 5.93972504e-01 -2.06696287e-01 -4.73863393e-01
1.46882370e-01 9.27717909e-02 3.66435438e-01 4.72721756e-01
-7.48630613e-02 -8.97795796e-01 -5.93281269e-01 5.78928649e-01
9.60991308e-02 -1.00027181e-01 8.61964524e-01 2.10482270e-01
4.73181933e-01 1.30683839e-01 6.00438178e-01 -4.56497192e-01
1.13935418e-01 -5.55022955e-01 -9.94572759e-01 1.37815937e-01
-4.76702392e-01 -2.59900182e-01 3.98277134e-01 2.90359646e-01
-7.41986930e-01 1.67803511e-01 -2.95661211e-01 1.62561655e-01
8.17094520e-02 6.29674554e-01 -6.89566433e-01 -6.81232929e-01
3.17414224e-01 3.31083715e-01 -3.32522482e-01 -4.48183298e-01
3.23845208e-01 8.40801954e-01 1.15936548e-01 -5.85652255e-02
1.35137188e+00 6.89278901e-01 6.24237835e-01 -1.31700015e+00
-7.53771424e-01 -6.28589094e-01 -4.66072768e-01 -2.82252073e-01
4.46073681e-01 -9.90529954e-01 -5.27152121e-01 9.63316187e-02
-1.13341725e+00 4.17205393e-01 3.06249380e-01 9.64770019e-01
-3.25674087e-01 3.19182307e-01 2.15531707e-01 -9.10794973e-01
-2.59770285e-02 -1.07888985e+00 7.29988992e-01 2.00391993e-01
6.30673170e-02 -7.30778754e-01 -2.21191019e-01 3.54830980e-01
7.35371649e-01 2.09035054e-01 7.49425054e-01 -1.69590339e-01
-1.05500841e+00 -6.24846756e-01 1.74996350e-02 -5.16744494e-01
2.90485829e-01 -9.71639454e-01 -5.19372642e-01 -4.57910717e-01
3.93688321e-01 8.80338550e-01 1.35002300e-01 7.48815954e-01
6.36888206e-01 -2.84824464e-02 -7.78432369e-01 1.02103925e+00
1.52730429e+00 3.53892326e-01 4.11559463e-01 1.53404057e-01
3.42112869e-01 -3.01251784e-02 8.12884569e-01 6.53412879e-01
2.16946781e-01 1.08023751e+00 8.39641213e-01 1.89783692e-01
1.88606769e-01 2.57773638e-01 -5.07927537e-01 5.95091283e-01
-1.43043101e-01 -7.34611154e-01 -6.97381616e-01 2.43961904e-02
-1.75035918e+00 -5.53147316e-01 -3.18810016e-01 2.51604199e+00
-1.98343799e-01 8.31254795e-02 -3.99294436e-01 -4.52924557e-02
5.04675567e-01 1.95255369e-01 -2.64349073e-01 2.16449991e-01
1.18499860e-01 -4.35772091e-02 9.55009401e-01 8.16907227e-01
-8.68526638e-01 3.31660539e-01 4.46631908e+00 6.52033627e-01
-1.08426332e+00 2.08181545e-01 2.05302969e-01 7.24424645e-02
-3.71086866e-01 -2.71605216e-02 -9.17699635e-01 3.75550896e-01
3.62140447e-01 -4.28769886e-02 1.86970115e-01 6.31622612e-01
3.75040293e-01 -3.83653194e-01 -6.79640472e-01 1.55858696e+00
1.95535019e-01 -1.32428300e+00 -4.09264028e-01 5.70258141e-01
5.87696671e-01 -3.20631504e-01 1.13819093e-01 -1.56383887e-01
-5.52794516e-01 -4.69277292e-01 4.72117186e-01 5.54832816e-01
4.98104870e-01 -7.47794688e-01 8.93604219e-01 6.48856878e-01
-1.21023512e+00 -2.16704726e-01 4.62497249e-02 2.72899624e-02
5.18020093e-01 1.42489958e+00 -1.00039506e+00 1.01224101e+00
1.87377453e-01 4.68840927e-01 -7.32198730e-02 1.64226127e+00
-4.49297071e-01 3.46709430e-01 -6.27092421e-01 -2.38570720e-01
8.79205838e-02 -6.51172757e-01 9.42036211e-01 6.44793332e-01
1.32308733e+00 3.02479416e-02 -9.38016996e-02 4.50634718e-01
1.59899920e-01 1.86984167e-02 -6.13655627e-01 4.15253162e-01
7.62980163e-01 1.16417086e+00 -6.74170554e-01 2.83916712e-01
-3.33482563e-01 6.11306131e-01 -3.84931684e-01 7.38264799e-01
-6.39716923e-01 -3.90438110e-01 7.69849658e-01 2.94888973e-01
3.17046791e-01 -1.12942660e+00 -5.04561126e-01 -1.02322543e+00
2.76176721e-01 -2.76772946e-01 -1.05941720e-01 -4.44873720e-01
-5.90681314e-01 2.68832356e-01 -4.26332802e-01 -1.72585571e+00
-4.07028317e-01 -3.84543210e-01 -3.44909996e-01 7.08802462e-01
-1.50292516e+00 -8.08595121e-01 -6.08754218e-01 2.12094188e-01
1.03993848e-01 -1.71736971e-01 7.74941087e-01 5.98282635e-01
-3.26153934e-01 3.59943002e-01 6.48917496e-01 -1.43010929e-01
3.41779709e-01 -8.66894782e-01 -8.25910305e-04 9.47306514e-01
2.85719275e-01 6.22697890e-01 9.18120086e-01 -3.69573325e-01
-1.96725917e+00 -7.61173844e-01 9.06320512e-01 -1.23369120e-01
2.10274234e-01 -6.44301951e-01 1.33754626e-01 4.82039005e-01
-2.25159675e-01 2.16990560e-01 5.54347396e-01 1.36938810e-01
2.95575261e-01 -5.66804588e-01 -1.04431915e+00 4.70785648e-01
1.24492121e+00 1.76119700e-01 5.78242429e-02 5.58638275e-01
1.02237962e-01 -6.47493482e-01 -4.35998589e-01 4.54036266e-01
5.13569653e-01 -7.43678689e-01 1.28671885e+00 1.70041516e-01
-7.15243697e-01 -4.69992429e-01 -4.82316405e-01 -1.40011346e+00
-3.08301777e-01 -8.03769648e-01 -2.72485912e-01 1.06769013e+00
4.10709262e-01 -1.10823405e+00 1.08392572e+00 -1.40466124e-01
1.86196752e-02 -5.03565013e-01 -1.53840458e+00 -8.61867726e-01
-6.87962472e-01 -4.80656326e-01 6.76791728e-01 6.25227809e-01
-2.94253677e-01 3.85607511e-01 -2.45559886e-01 1.17245066e+00
9.74226117e-01 4.39880759e-01 9.92582083e-01 -1.39228046e+00
-3.86128366e-01 -2.76944339e-01 -7.77050614e-01 -1.66360664e+00
-1.91575944e-01 -6.87353671e-01 -1.50602117e-01 -1.82495368e+00
-5.69084823e-01 -1.06061709e+00 1.03216916e-01 -3.35922718e-01
4.59958255e-01 2.45866582e-01 -8.44112635e-02 2.97664870e-02
-3.05208355e-01 3.33300143e-01 9.35714185e-01 -4.28413339e-02
-3.48578453e-01 1.21856582e+00 -6.12658441e-01 6.46839797e-01
6.60711229e-01 -2.51795530e-01 -3.63072902e-01 -5.81514239e-01
5.23331583e-01 4.66932297e-01 1.72889337e-01 -1.35833085e+00
3.60263586e-01 1.92690659e-02 3.43966633e-01 -5.77113330e-01
7.96108902e-01 -1.24261069e+00 2.63253927e-01 4.47616518e-01
5.58796823e-01 -5.23584485e-01 -2.21521631e-01 8.72795045e-01
-4.53908957e-04 -3.53266537e-01 4.45002109e-01 6.67380616e-02
-5.16710401e-01 4.90864903e-01 -5.44237614e-01 -5.04319847e-01
1.24322081e+00 -2.60874689e-01 -1.30171001e-01 -1.00090933e+00
-8.09976816e-01 1.18722022e-01 -3.19759659e-02 6.83761463e-02
3.83152962e-01 -1.29744124e+00 -4.58413929e-01 2.91228235e-01
1.09882019e-01 -2.49373004e-01 2.99587637e-01 1.04652798e+00
-4.85616714e-01 8.91998172e-01 2.57540613e-01 -8.62955153e-01
-1.08196974e+00 5.61242253e-02 5.07559955e-01 1.05104730e-01
-1.97488695e-01 5.46796203e-01 -1.81454927e-01 -2.96046436e-01
1.17251188e-01 -2.35541493e-01 -1.23114832e-01 7.82609805e-02
1.91048905e-01 2.27038831e-01 4.51658934e-01 -5.69405973e-01
-6.24352455e-01 1.21237016e+00 6.95898712e-01 -1.57572925e-01
1.19941020e+00 -6.49380147e-01 -2.01985962e-03 -8.44467282e-02
9.34676349e-01 9.01735663e-01 -4.34597135e-01 -2.46312872e-01
-2.88044033e-03 -6.79828644e-01 1.98742762e-01 -4.25629884e-01
-8.54999125e-01 6.31526232e-01 7.92612851e-01 4.50560570e-01
9.09994245e-01 1.23007886e-01 4.72464621e-01 6.38833284e-01
1.20272076e+00 -4.49790329e-01 -3.32344830e-01 3.01486790e-01
5.48175752e-01 -9.38656032e-01 3.09151590e-01 -8.65514994e-01
2.32916713e-01 1.24315727e+00 2.23096937e-01 2.44953498e-01
9.27096486e-01 1.74428016e-01 5.49194813e-02 -1.86891370e-02
1.74330413e-01 -3.31000835e-01 1.71130911e-01 5.82668900e-01
5.18936962e-02 1.59027413e-01 -3.91190708e-01 6.60885394e-01
-2.82823920e-01 -4.79299814e-01 4.29103374e-01 7.00611115e-01
-7.48328328e-01 -1.05992627e+00 -6.52036786e-01 3.46331239e-01
1.25324456e-02 2.77139634e-01 2.78171510e-01 8.27349603e-01
2.79334366e-01 1.10330558e+00 -1.35514855e-01 -3.28687802e-02
4.28329855e-01 -2.66281873e-01 6.32072091e-01 -4.30568367e-01
5.62328994e-01 8.49567652e-02 1.67695493e-01 -3.83531153e-01
-1.38850689e-01 -6.75009310e-01 -7.94916332e-01 2.57000208e-01
-5.36893070e-01 5.68001151e-01 1.27490664e+00 1.17543042e+00
5.53372979e-01 2.32278645e-01 9.97285187e-01 -6.39338434e-01
-3.13251853e-01 -5.77302754e-01 -8.47643793e-01 -5.84425092e-01
4.69979018e-01 -6.02456987e-01 -4.63376760e-01 -8.41932297e-01] | [6.290822505950928, 1.1947555541992188] |
909edb6d-4c41-4167-a868-ae9a275cb3a2 | practical-privacy-preserving-gaussian-process | 2306.14498 | null | https://arxiv.org/abs/2306.14498v1 | https://arxiv.org/pdf/2306.14498v1.pdf | Practical Privacy-Preserving Gaussian Process Regression via Secret Sharing | Gaussian process regression (GPR) is a non-parametric model that has been used in many real-world applications that involve sensitive personal data (e.g., healthcare, finance, etc.) from multiple data owners. To fully and securely exploit the value of different data sources, this paper proposes a privacy-preserving GPR method based on secret sharing (SS), a secure multi-party computation (SMPC) technique. In contrast to existing studies that protect the data privacy of GPR via homomorphic encryption, differential privacy, or federated learning, our proposed method is more practical and can be used to preserve the data privacy of both the model inputs and outputs for various data-sharing scenarios (e.g., horizontally/vertically-partitioned data). However, it is non-trivial to directly apply SS on the conventional GPR algorithm, as it includes some operations whose accuracy and/or efficiency have not been well-enhanced in the current SMPC protocol. To address this issue, we derive a new SS-based exponentiation operation through the idea of 'confusion-correction' and construct an SS-based matrix inversion algorithm based on Cholesky decomposition. More importantly, we theoretically analyze the communication cost and the security of the proposed SS-based operations. Empirical results show that our proposed method can achieve reasonable accuracy and efficiency under the premise of preserving data privacy. | ['Zenglin Xu', 'Yue Yu', 'Hui Wang', 'Shuang Qin', 'JiaQi Zhang', 'Yehong Zhang', 'Jinglong Luo'] | 2023-06-26 | null | null | null | null | ['gpr', 'gpr'] | ['computer-vision', 'miscellaneous'] | [ 3.77228409e-01 -3.51197660e-01 1.71879619e-01 -2.97430009e-01
-7.73135841e-01 -5.69608390e-01 2.01836422e-01 2.55985230e-01
-4.99497056e-01 6.51066720e-01 -1.08422048e-01 -5.17386138e-01
-2.46501744e-01 -9.94446874e-01 -5.29782116e-01 -1.24598420e+00
2.87839547e-02 -9.47013944e-02 -9.77339521e-02 -1.11064404e-01
2.78953820e-01 3.37939531e-01 -1.02458727e+00 -2.04754565e-02
8.06533754e-01 1.08911479e+00 -2.36672893e-01 2.72078365e-01
2.16157258e-01 1.74159586e-01 -1.99716911e-01 -9.03209448e-01
7.35675812e-01 -2.34365970e-01 -3.79721105e-01 -3.73539716e-01
-4.66665626e-01 -3.24251175e-01 -5.35061181e-01 1.31444180e+00
5.57067752e-01 -1.76263407e-01 3.45115542e-01 -1.60289478e+00
-4.99908775e-01 2.83432901e-01 -8.44848752e-01 -3.87890458e-01
2.46960167e-02 -1.03924870e-01 5.68847179e-01 -5.10746896e-01
1.47473961e-01 1.06330013e+00 7.20544517e-01 1.94931880e-01
-1.20935190e+00 -1.34635389e+00 -3.61606002e-01 1.77724749e-01
-1.89191663e+00 -1.74377218e-01 6.44830167e-01 -1.07357807e-01
5.85109973e-03 6.61047041e-01 3.49389076e-01 6.38413846e-01
5.89413464e-01 5.06473184e-01 1.55092216e+00 -5.38302325e-02
2.41359428e-01 5.13016820e-01 1.09816073e-02 2.83693969e-01
8.70948792e-01 2.57405728e-01 -3.40456516e-01 -8.99750054e-01
6.77581549e-01 4.30267006e-01 -6.43088162e-01 -5.86050093e-01
-1.03961909e+00 8.73399973e-01 5.47334142e-02 -8.67110789e-02
-2.70724535e-01 4.84675169e-02 3.30111593e-01 4.10274148e-01
2.22148195e-01 -1.84604481e-01 -5.07547140e-01 2.66773224e-01
-6.02651834e-01 3.15963358e-01 1.14122808e+00 1.00269675e+00
7.78812885e-01 -3.71838599e-01 -1.85315926e-02 2.09974125e-01
5.47846317e-01 7.09910870e-01 3.11705321e-01 -4.49415505e-01
6.82370961e-01 1.73994005e-01 1.69068143e-01 -1.39396894e+00
1.64468251e-02 -1.69111639e-01 -1.36239767e+00 -1.57559499e-01
2.66129285e-01 -4.86254960e-01 -5.51217608e-02 1.69680035e+00
4.60025787e-01 3.69930655e-01 6.05368435e-01 5.14239371e-01
2.91940004e-01 7.51669645e-01 1.27924666e-01 -4.70402330e-01
1.57650971e+00 -4.31744218e-01 -8.44739556e-01 3.13763559e-01
5.18763721e-01 -6.27033532e-01 4.63172168e-01 3.83345723e-01
-7.56893873e-01 -1.92861422e-03 -1.07915962e+00 2.34729536e-02
-2.87408143e-01 -5.94457388e-02 9.67205942e-01 1.27370834e+00
-6.24009430e-01 4.24424708e-01 -6.20414317e-01 1.06013775e-01
5.88844299e-01 7.01673865e-01 -7.43888855e-01 -9.48446393e-02
-1.46990728e+00 3.94550562e-01 3.65181804e-01 5.89370467e-02
-3.95411223e-01 -7.11747825e-01 -7.69258916e-01 2.45121121e-01
3.05456132e-01 -7.39463985e-01 8.52795899e-01 -3.28004241e-01
-1.40080953e+00 6.19110763e-01 -1.24559179e-01 -5.64918339e-01
5.46384215e-01 2.10090190e-01 -5.59179246e-01 -3.79953943e-02
-2.88930088e-01 -9.50221866e-02 7.10074008e-01 -1.06650817e+00
-7.43860483e-01 -8.65719497e-01 -3.07327926e-01 9.76950005e-02
-4.37326998e-01 1.84056327e-01 -1.49616688e-01 -7.14780450e-01
3.02945822e-01 -1.22230625e+00 -5.04398048e-01 1.65030599e-01
-4.47489142e-01 3.40653718e-01 1.01321733e+00 -9.94010925e-01
1.22000420e+00 -2.51330972e+00 -1.92439780e-01 7.15877295e-01
-8.35648365e-03 1.79621845e-01 4.47539330e-01 6.37186229e-01
1.94637939e-01 1.99937984e-01 -6.61523581e-01 -4.16139662e-01
4.09383625e-02 6.61003962e-02 -3.80448639e-01 9.29218233e-01
-4.99518722e-01 7.46219039e-01 -6.33123279e-01 -3.08759987e-01
-1.87317338e-02 6.23841524e-01 -2.46124268e-01 -5.75711913e-02
4.76116210e-01 5.71309507e-01 -5.31870365e-01 5.93411386e-01
1.51881397e+00 -1.38326198e-01 4.24788743e-01 -1.97019741e-01
-1.12856980e-02 -2.84213871e-01 -1.71360219e+00 1.31302392e+00
-3.55457067e-01 3.12629975e-02 3.47033381e-01 -6.31251097e-01
8.83427858e-01 5.20211399e-01 3.79562169e-01 -1.33258238e-01
3.29457611e-01 3.47822040e-01 -2.74244726e-01 -8.43364447e-02
5.17809749e-01 -1.37760639e-01 -4.04371679e-01 6.39307261e-01
-4.82698828e-01 2.06992418e-01 -6.42449141e-01 -1.65159225e-01
7.54857659e-01 -2.73156166e-01 5.85513711e-01 -3.95096213e-01
9.06881154e-01 -5.11791825e-01 8.53221059e-01 4.38356519e-01
-1.00405119e-01 3.89463037e-01 2.95625746e-01 2.83393692e-02
-6.75627053e-01 -7.50064135e-01 -2.82443792e-01 5.33324957e-01
5.20995557e-01 -3.22649628e-01 -4.52374846e-01 -5.11096716e-01
4.22121495e-01 5.13233423e-01 -1.89344853e-01 -2.92623103e-01
-1.19776614e-01 -1.04542398e+00 8.23105335e-01 1.15240529e-01
1.07838404e+00 -3.72693121e-01 -2.12856084e-01 1.53182715e-01
-3.40217585e-03 -9.24381793e-01 -6.02982819e-01 -1.50363833e-01
-7.77625084e-01 -9.03283477e-01 -3.23506951e-01 -4.67169106e-01
7.10315824e-01 6.53905809e-01 -1.54265268e-02 -1.87391222e-01
2.23933131e-01 1.67680636e-01 -4.62278277e-02 -6.60944998e-01
-2.77218133e-01 -1.38991073e-01 1.28560767e-01 6.52638197e-01
4.28102344e-01 -6.93376303e-01 -7.07229555e-01 2.70047069e-01
-1.14675140e+00 -3.02560460e-02 7.19877124e-01 5.88214636e-01
5.88632464e-01 6.54082060e-01 4.48063225e-01 -1.07002831e+00
8.17756593e-01 -6.45678043e-01 -7.13253856e-01 4.82294291e-01
-8.40586126e-01 1.39050797e-01 6.22498333e-01 -2.26563424e-01
-1.10436201e+00 1.25632167e-01 -1.16586730e-01 -2.21327782e-01
2.00093612e-01 3.27523619e-01 -7.50268638e-01 -6.46767020e-01
2.82498524e-02 6.11862242e-01 2.95346767e-01 -4.59259540e-01
2.27197975e-01 1.15798199e+00 4.73247617e-01 -4.89506751e-01
9.64416265e-01 7.48149216e-01 6.18043542e-01 -3.52985531e-01
-3.06662861e-02 -3.91484737e-01 -6.30845949e-02 5.95855296e-01
5.63551664e-01 -1.09824300e+00 -1.24975061e+00 8.73452246e-01
-1.01867831e+00 5.18622816e-01 2.22815186e-01 6.60867155e-01
-2.93005675e-01 9.69660699e-01 -5.12604654e-01 -9.10809457e-01
-7.90651143e-01 -1.18652439e+00 6.92155957e-01 1.30934730e-01
2.54730165e-01 -8.02752852e-01 -1.23111494e-01 1.97989851e-01
4.36686426e-01 3.95426810e-01 7.04714119e-01 -8.84753108e-01
-6.57841444e-01 -4.29975271e-01 -2.47784227e-01 4.36303258e-01
2.48119965e-01 -5.23838282e-01 -8.95812750e-01 -6.54499352e-01
5.61905026e-01 2.60555029e-01 3.22751760e-01 -1.48936033e-01
1.54003179e+00 -6.99243665e-01 -3.53653222e-01 8.98918211e-01
1.62681377e+00 1.90333262e-01 9.35156941e-01 1.29437298e-01
5.41039705e-01 6.48999810e-01 5.51298499e-01 9.29084599e-01
5.92181861e-01 3.54968011e-01 2.15563834e-01 1.63555965e-01
9.00392950e-01 -3.97645444e-01 2.36909866e-01 5.72887599e-01
4.16980647e-02 -9.15510133e-02 -3.71964097e-01 8.10432434e-02
-2.01565218e+00 -7.36313522e-01 -3.91861886e-01 2.72298241e+00
9.11696136e-01 -4.40578103e-01 -5.06998122e-01 1.48832291e-01
7.68032789e-01 -1.32399229e-02 -3.55130911e-01 -4.38225299e-01
-1.68824956e-01 2.19402432e-01 1.32880628e+00 1.00097023e-02
-1.14655530e+00 3.31085265e-01 4.86660480e+00 9.77387011e-01
-9.35139120e-01 3.96362603e-01 5.13141155e-01 3.48173410e-01
-5.36829412e-01 2.73568511e-01 -7.66552031e-01 8.86846662e-01
8.19209516e-01 -5.45735538e-01 4.41574305e-01 7.89172411e-01
-2.45000757e-02 7.91483745e-02 -8.90250564e-01 1.27714801e+00
-5.52986115e-02 -1.05409551e+00 -1.49442717e-01 5.34916580e-01
6.07411146e-01 -7.99593806e-01 2.15030789e-01 9.63346437e-02
1.57462418e-01 -7.96355128e-01 1.48430467e-01 4.11381334e-01
6.74578905e-01 -1.16611266e+00 8.28602672e-01 4.22397196e-01
-1.06294787e+00 -1.54123634e-01 -4.30203676e-01 2.23546147e-01
1.86632544e-01 5.08497953e-01 -3.60405207e-01 1.13828123e+00
5.88698387e-01 7.83209130e-02 -1.61508307e-01 8.67613316e-01
-2.44226933e-01 3.78243893e-01 -5.33137619e-01 1.55492738e-01
-8.49630013e-02 -5.73647499e-01 4.64701504e-01 9.42080140e-01
7.73470521e-01 3.67097110e-01 -1.28714114e-01 4.43210959e-01
-1.85421467e-01 4.64876086e-01 -6.03840470e-01 2.60695279e-01
8.06801617e-01 1.10425711e+00 -2.20272504e-02 1.25720706e-02
-4.50120211e-01 1.09467888e+00 -3.34018588e-01 1.27106398e-01
-5.93453348e-01 -5.74099660e-01 8.90991569e-01 -6.18177839e-02
2.32318237e-01 -2.98097074e-01 -4.65606689e-01 -1.17827296e+00
-1.04217567e-02 -8.73727441e-01 5.12474895e-01 -1.58808246e-01
-1.32642341e+00 -8.06552172e-03 -8.44628662e-02 -1.40999782e+00
1.32790133e-01 -4.36186567e-02 -4.80818003e-01 1.16044021e+00
-1.58448863e+00 -1.44815660e+00 3.95677686e-02 9.96532142e-01
-5.25979996e-01 -1.11630976e-01 1.06119490e+00 2.86407948e-01
-4.80085433e-01 9.44704771e-01 7.45695829e-01 -2.06737250e-01
7.13801146e-01 -7.31813848e-01 1.70334488e-01 9.29580092e-01
-2.59943247e-01 1.11641049e+00 5.72667181e-01 -6.92264140e-01
-1.81213665e+00 -1.13860190e+00 7.96988547e-01 1.27397016e-01
3.43548805e-01 -4.36380625e-01 -9.90879714e-01 7.96360135e-01
2.92217154e-02 1.48495277e-02 1.22800016e+00 -3.36241633e-01
-2.87596226e-01 -4.22589868e-01 -1.81983316e+00 6.02564692e-01
5.47526360e-01 -5.28921783e-01 -2.00520307e-01 2.21423984e-01
7.00214505e-01 -3.23862404e-01 -1.05666244e+00 4.00349259e-01
6.52553201e-01 -6.12748981e-01 1.05148852e+00 -1.78206310e-01
-2.08772942e-01 -5.26902676e-01 -4.14118111e-01 -8.88230085e-01
-3.15978825e-02 -1.03719270e+00 -1.91244811e-01 1.37234759e+00
2.13061795e-02 -1.35673892e+00 6.71247363e-01 1.11864662e+00
4.92951214e-01 -5.23474514e-01 -1.08453977e+00 -6.24461174e-01
-1.79306850e-01 -2.85380214e-01 1.29210556e+00 9.90590274e-01
-1.53928101e-01 -3.66232365e-01 -8.60388756e-01 6.97805762e-01
9.66412067e-01 1.75938293e-01 1.03419602e+00 -1.18923986e+00
-3.39373350e-01 2.60944217e-01 -3.95589232e-01 -7.65136063e-01
3.26266959e-02 -8.28198731e-01 -4.37011808e-01 -9.61999595e-01
2.84412146e-01 -8.20205331e-01 -5.00993371e-01 2.78242886e-01
-1.41285971e-01 -1.36526767e-03 1.50362074e-01 3.97032857e-01
-1.36798546e-01 5.15122831e-01 1.00348175e+00 2.40475267e-01
-2.20931083e-01 4.11951512e-01 -1.25891793e+00 2.87153244e-01
1.06669521e+00 -6.05504870e-01 -4.39596951e-01 4.62340228e-02
1.85486227e-01 2.40469113e-01 4.81869102e-01 -7.61946499e-01
5.89568794e-01 -3.36670130e-01 -1.29662582e-03 -5.05436599e-01
2.67162859e-01 -1.38866901e+00 9.06730413e-01 6.57566011e-01
1.07412517e-01 -7.33235702e-02 -4.31623273e-02 9.53328013e-01
-5.41257560e-02 -1.39439911e-01 5.10170758e-01 8.81242976e-02
-4.93844867e-01 6.35532737e-01 -1.08804859e-01 -6.54453814e-01
1.48641396e+00 -1.25333548e-01 -2.12364569e-01 -3.34795207e-01
-2.25179881e-01 2.48806909e-01 6.47907555e-01 -9.87090450e-03
6.09878600e-01 -1.28453326e+00 -6.08359933e-01 5.31083047e-01
5.29903248e-02 -2.21316013e-02 6.18297279e-01 8.81186128e-01
-3.65088493e-01 4.89372551e-01 -3.76612023e-02 -7.65116289e-02
-1.46777558e+00 5.72056532e-01 -1.49700582e-01 -4.84325945e-01
-4.19040650e-01 2.88887799e-01 1.56009465e-01 -3.98364365e-01
-6.59778565e-02 3.40452567e-02 1.99125960e-01 -2.57874250e-01
6.05292559e-01 5.31734645e-01 -9.43681821e-02 -5.07332921e-01
-1.52006760e-01 3.59836251e-01 3.80427726e-02 -1.40281707e-01
9.97937500e-01 -4.93261844e-01 -5.68602264e-01 -8.12862739e-02
1.55309784e+00 3.16213340e-01 -8.78370225e-01 -4.56547946e-01
-2.94020236e-01 -8.57111335e-01 -1.60295740e-02 -7.55383447e-02
-1.02209675e+00 6.15111887e-01 6.64101303e-01 -1.67032003e-01
1.11668301e+00 -6.12821519e-01 1.22691965e+00 2.26446256e-01
9.18215275e-01 -7.93163180e-01 -8.28292429e-01 -9.84738916e-02
4.58075702e-01 -1.02142406e+00 2.17712536e-01 -8.04718435e-01
-7.35922933e-01 7.37175524e-01 5.22385575e-02 2.85740674e-01
1.02404308e+00 1.17343150e-01 -4.50141847e-01 2.48013347e-01
-2.91838229e-01 4.84401047e-01 -2.14931160e-01 5.66391587e-01
-3.52632791e-01 3.48800629e-01 -8.57363880e-01 1.30453420e+00
-1.35524645e-01 1.76891074e-01 5.13565361e-01 1.26762497e+00
6.59231916e-02 -1.51035631e+00 -6.25140667e-01 4.27089065e-01
-8.67549896e-01 -3.03494688e-02 2.48551205e-01 4.02447879e-01
9.46442410e-02 8.71242702e-01 -3.59284073e-01 -4.30542529e-01
6.57464415e-02 -1.28098100e-01 1.87285598e-02 -1.32596746e-01
-5.58875442e-01 -3.75930995e-01 -2.76765674e-01 -2.05086783e-01
-3.27168703e-01 -8.11552405e-01 -1.16042018e+00 -8.53095055e-01
-3.77216667e-01 2.22281843e-01 9.71404612e-01 5.41198075e-01
5.47965884e-01 -1.21334076e-01 1.11085773e+00 -1.34893253e-01
-1.00048494e+00 -3.21938455e-01 -1.12606120e+00 2.60497093e-01
8.71586427e-03 -2.71684051e-01 -4.24118191e-01 -3.02587360e-01] | [5.920133113861084, 6.67006778717041] |
95c4084f-4735-4d3e-a748-a66a71ce4f9a | halsie-hybrid-approach-to-learning | 2211.10754 | null | https://arxiv.org/abs/2211.10754v3 | https://arxiv.org/pdf/2211.10754v3.pdf | HALSIE: Hybrid Approach to Learning Segmentation by Simultaneously Exploiting Image and Event Modalities | We present HALSIE, a novel hybrid approach for semantic segmentation by simultaneously leveraging image and event modalities. Event cameras are vision sensors that detect changes in per-pixel intensity to generate asynchronous 'event streams'. They offer significant advantages over standard frame-based cameras due to their higher dynamic range, higher temporal resolution, and lack of motion blur. However, events only measure the varying component of the visual signal, which limits their ability to encode scene context. To augment the missing contextual information, we postulate that fusing spatially dense frames with temporally dense events can generate semantic maps with fine-grained predictions. Prior work in event-based vision has achieved outstanding performance but with substantial inference cost, typically beyond 50 mJ per cycle. By redesigning the end-to-end learning framework, we reduce inference cost by up to $\sim 20\times$ while retaining similar performance. To achieve this, our method efficiently extracts and fuses the complementary features, exploiting the best of both modalities. In particular, HALSIE comprises dual-encoders with a Spiking Neural Network (SNN) branch to provide rich temporal cues from asynchronous events, and a standard Artificial Neural Network (ANN) branch for extracting spatial information from regular frame data to enable cross-domain learning. Our hybrid network reaches state-of-the-art performance on real-world DDD-17, MVSEC and DSEC-Semantic datasets with up to $\sim 33\times$ higher parameter efficiency and favorable inference cost (17.9mJ per cycle), making it suitable for resource-constrained edge applications. Further, the effectiveness of design choices in our approach is evidenced by our thorough ablation study. | ['Kaushik Roy', 'Marco Apolinario', 'Chamika Liyanagedera', 'Adarsh Kosta', 'Shristi Das Biswas'] | 2022-11-19 | null | null | null | null | ['event-based-vision'] | ['computer-vision'] | [ 7.03115284e-01 -3.19423527e-01 -1.45066708e-01 -3.05027992e-01
-9.18583989e-01 -5.11610687e-01 6.15044951e-01 -4.76850830e-02
-7.61002123e-01 7.51238108e-01 -3.60135995e-02 -1.13808408e-01
2.80027449e-01 -6.94100201e-01 -8.67617428e-01 -7.40898907e-01
2.00747792e-02 -1.05701335e-01 7.63358831e-01 2.83173919e-01
1.47371322e-01 3.85024041e-01 -1.66221297e+00 5.09209454e-01
5.13771653e-01 1.35841537e+00 3.66813421e-01 8.24980974e-01
-4.21971567e-02 1.04111838e+00 -3.78046006e-01 -7.90698677e-02
1.54679254e-01 -5.13512790e-01 -5.29181778e-01 -1.72338088e-03
2.82475680e-01 -5.22794485e-01 -7.48873651e-01 6.02176368e-01
5.74614584e-01 -9.55501795e-02 2.34590784e-01 -1.28318524e+00
-1.49020374e-01 2.57846117e-01 -7.26246774e-01 7.68573105e-01
2.53260821e-01 5.93248785e-01 8.68660808e-01 -6.44818306e-01
8.24071229e-01 7.63245165e-01 5.91964245e-01 5.41121721e-01
-1.41709137e+00 -6.38903677e-01 1.43535942e-01 2.12883607e-01
-1.00689709e+00 -7.59247839e-01 6.05076730e-01 -7.98131749e-02
1.57466698e+00 4.01928686e-02 8.81244123e-01 1.32328534e+00
1.93412572e-01 9.99143064e-01 1.12181592e+00 -9.60370004e-02
4.35283780e-01 -6.58670008e-01 -1.63334608e-01 7.07850814e-01
-3.14207189e-02 3.62653643e-01 -1.02790070e+00 1.75029293e-01
1.00855517e+00 7.51460493e-02 -3.10497880e-01 1.29539669e-01
-1.33498240e+00 3.36244017e-01 4.74187851e-01 -8.21234584e-02
-3.31590086e-01 8.28238070e-01 3.11210573e-01 -1.38078913e-01
1.55435488e-01 1.43561885e-01 -3.81067812e-01 -4.84148294e-01
-1.03592348e+00 8.25923830e-02 6.01633310e-01 8.77452254e-01
5.96889377e-01 2.32081547e-01 -2.57672906e-01 4.12074298e-01
2.25182071e-01 5.73092103e-01 2.08082572e-01 -1.49419427e+00
2.41569072e-01 2.72360444e-01 9.08241328e-03 -4.90535945e-01
-4.92010206e-01 -2.17267051e-01 -6.11430526e-01 3.94592375e-01
6.05973065e-01 -1.25839189e-01 -1.19719863e+00 1.92071211e+00
1.04774974e-01 8.40539157e-01 -1.00869415e-02 9.51349199e-01
6.37762785e-01 7.42209315e-01 2.22805604e-01 -2.97929347e-01
1.58059454e+00 -6.22684538e-01 -4.35554415e-01 -5.57212949e-01
9.36092436e-02 -5.60967505e-01 8.15902591e-01 2.88402855e-01
-1.27826917e+00 -1.91777974e-01 -1.12928867e+00 -2.89051533e-01
-1.01244472e-01 -2.17288315e-01 9.89914060e-01 3.14299911e-01
-9.73158002e-01 4.71739680e-01 -1.32600737e+00 -1.53719038e-01
8.49304378e-01 5.19026697e-01 -3.44837345e-02 -5.73944673e-02
-8.15085232e-01 3.83915454e-01 2.78437138e-01 -2.26285174e-01
-9.65043128e-01 -9.17567849e-01 -8.36959183e-01 6.25708997e-02
3.30221117e-01 -8.47322881e-01 1.35411155e+00 -8.09547186e-01
-1.66361034e+00 8.19496453e-01 -5.24978757e-01 -9.35876966e-01
2.66916633e-01 2.08322465e-01 -2.87757456e-01 7.11061656e-01
1.68402255e-01 1.29047918e+00 6.30219042e-01 -8.19860518e-01
-9.15220618e-01 -2.14927226e-01 9.31196380e-03 1.96460009e-01
-4.80352305e-02 3.51799391e-02 -9.85319257e-01 -7.25738227e-01
4.67863753e-02 -9.41681266e-01 -8.36839303e-02 3.29593331e-01
6.35418445e-02 1.73329979e-01 8.78965676e-01 -2.72387475e-01
8.80125225e-01 -2.16670203e+00 -1.12681709e-01 -2.17987686e-01
3.39088053e-01 4.86750081e-02 2.87408121e-02 -2.30308995e-02
3.84406716e-01 -1.77273288e-01 -4.88685519e-01 -4.40524668e-01
-3.60030085e-01 3.45894367e-01 -2.86024272e-01 3.11519921e-01
5.49215317e-01 1.16559160e+00 -9.91995573e-01 -5.05544782e-01
4.69620645e-01 5.52729547e-01 -4.79083896e-01 -1.78906813e-01
-4.76089984e-01 4.94789273e-01 -2.60986567e-01 8.50245357e-01
3.83901656e-01 -6.05885088e-01 5.18522561e-02 -3.84789467e-01
-6.83564767e-02 2.92800814e-01 -9.53919470e-01 2.10576868e+00
-3.51968020e-01 9.58073616e-01 2.93197967e-02 -7.92812347e-01
5.14041424e-01 1.86190650e-01 7.55381942e-01 -1.10378599e+00
1.71512082e-01 1.19482219e-01 -1.90800324e-01 -2.28289023e-01
3.31886530e-01 3.72140519e-02 -1.30261421e-01 3.80357832e-01
2.30145738e-01 -4.18397225e-02 2.47215301e-01 2.72230804e-01
1.57692671e+00 3.94111484e-01 -3.39463465e-02 9.73201394e-02
-1.60024598e-01 4.29984145e-02 8.46130550e-01 7.39676058e-01
-4.22028452e-01 7.92476296e-01 4.05066580e-01 -1.85610309e-01
-9.81954157e-01 -1.37951493e+00 -2.55027443e-01 8.34310114e-01
6.07998848e-01 -2.90693879e-01 -5.67279875e-01 -2.58725196e-01
-2.13082075e-01 4.85968649e-01 -2.86442667e-01 -1.19761810e-01
-4.91094679e-01 -9.40104544e-01 8.88855457e-01 9.36659932e-01
8.09824467e-01 -9.65737998e-01 -1.33885956e+00 5.57033718e-01
-2.92348891e-01 -1.63769102e+00 -3.73321176e-01 5.20854235e-01
-8.39243889e-01 -8.74104977e-01 -4.31136906e-01 -6.08741879e-01
3.09738129e-01 2.79947847e-01 1.20604980e+00 -4.07378197e-01
-7.08413303e-01 3.70517582e-01 -1.01911388e-01 -3.34192276e-01
1.59245029e-01 -3.43059868e-01 -3.25456798e-01 -2.07520932e-01
3.75804543e-01 -8.08600903e-01 -9.79772449e-01 1.62807122e-01
-1.02856660e+00 4.39154893e-01 4.43143904e-01 9.69681621e-01
9.30415928e-01 -1.74772441e-01 5.49216270e-01 -6.16108716e-01
-1.00729361e-01 -2.88646907e-01 -7.61422515e-01 -1.37489095e-01
-5.62598050e-01 3.65972482e-02 4.43199843e-01 -2.78393686e-01
-1.31997049e+00 3.83028060e-01 9.23986509e-02 -4.67999607e-01
-1.43783793e-01 9.24821571e-02 9.53234136e-02 -9.73382071e-02
6.07162178e-01 2.64626950e-01 -1.07949361e-01 1.06846422e-01
2.86154121e-01 3.03240627e-01 1.15898919e+00 -6.22073233e-01
1.46950334e-01 1.06270695e+00 4.75503244e-02 -4.54021037e-01
-5.36562681e-01 -4.24458802e-01 -1.04133889e-01 -2.43464902e-01
1.07116508e+00 -1.32042849e+00 -9.81246650e-01 8.06836545e-01
-1.08162642e+00 -5.49647212e-01 -3.11397016e-01 5.18291652e-01
-8.23624134e-01 4.19313312e-02 -1.07180393e+00 -5.69121122e-01
-3.45846176e-01 -1.21638393e+00 1.36563599e+00 6.68601394e-01
3.28912288e-02 -6.97867095e-01 -4.84875560e-01 3.66507500e-01
2.97266632e-01 5.01567304e-01 5.44040680e-01 4.66468185e-03
-1.10285854e+00 2.21723899e-01 -5.89257896e-01 -1.08442001e-01
-1.72884524e-01 -1.30498171e-01 -1.32669902e+00 -1.38973594e-02
-1.57611787e-01 -2.86428720e-01 1.16128826e+00 6.70957923e-01
1.38050544e+00 2.94359118e-01 -5.26289701e-01 9.49117422e-01
1.61188304e+00 4.48089838e-01 7.73322463e-01 1.68088868e-01
5.65886915e-01 -5.53649254e-02 3.57662708e-01 6.35593116e-01
3.66009325e-01 6.39862180e-01 4.69675094e-01 -1.66398674e-01
-5.02327025e-01 -4.30089161e-02 4.12725091e-01 2.09568411e-01
1.59064353e-01 -5.40632546e-01 -8.15108895e-01 6.21556401e-01
-1.89980590e+00 -1.16535807e+00 4.69970778e-02 2.01352572e+00
9.43820477e-01 5.33545375e-01 -4.79840115e-02 -3.03916428e-02
5.44723451e-01 2.98275590e-01 -1.00980365e+00 -1.09819770e-01
-4.56209540e-01 5.57095051e-01 8.13639462e-01 1.62394509e-01
-1.02748764e+00 9.39993799e-01 5.88409233e+00 8.06048334e-01
-1.23899543e+00 -4.76222970e-02 7.95559645e-01 -7.06388950e-01
-3.06096673e-01 1.40596986e-01 -7.69089103e-01 7.42076337e-01
1.10407758e+00 1.47907928e-01 6.16034269e-01 3.17703754e-01
2.71328390e-01 -6.40016019e-01 -1.12585509e+00 1.29377711e+00
-1.14930145e-01 -1.90264857e+00 -4.65058804e-01 -1.65056193e-03
6.62634313e-01 5.21907270e-01 -9.78147611e-02 -1.87393412e-01
4.70570773e-01 -9.71181750e-01 8.20359170e-01 4.28640693e-01
1.04639375e+00 -6.16357565e-01 2.75087088e-01 6.43478259e-02
-1.29300988e+00 -3.11134160e-02 -1.90670253e-03 -1.10447362e-01
6.08489275e-01 8.24564755e-01 -5.97784281e-01 2.44160086e-01
9.75910902e-01 9.48812425e-01 -3.07639390e-01 9.21311319e-01
3.51388901e-02 6.84849501e-01 -8.26417029e-01 2.46496081e-01
2.40136772e-01 1.39114425e-01 4.59829926e-01 1.24136424e+00
2.38426238e-01 3.74464422e-01 1.02354988e-01 9.30394173e-01
-2.89766401e-01 -8.73660326e-01 -4.61687028e-01 -8.60391930e-02
7.65610993e-01 1.15144515e+00 -1.18114805e+00 -3.16028893e-01
-6.41754448e-01 1.35935771e+00 7.93638751e-02 3.78989011e-01
-1.16337907e+00 -3.16347957e-01 6.47195518e-01 -1.32384121e-01
5.95043778e-01 -2.84430504e-01 -7.16922522e-01 -1.11436832e+00
1.65274069e-01 -4.38140512e-01 2.98854172e-01 -9.48872387e-01
-9.09999490e-01 2.92944252e-01 -2.35907376e-01 -1.03615308e+00
-2.06911892e-01 -5.90073824e-01 -3.84367555e-01 6.69778109e-01
-1.67198908e+00 -1.17769086e+00 -4.76327211e-01 6.90515399e-01
6.82263255e-01 1.23493209e-01 5.59310734e-01 3.86219084e-01
-5.90846658e-01 3.89839321e-01 -3.51433679e-02 5.97187802e-02
5.45658171e-01 -1.23957622e+00 4.96920556e-01 1.08555615e+00
2.58571237e-01 3.84356938e-02 4.10448790e-01 -5.14237404e-01
-1.77394569e+00 -1.12213814e+00 3.66795093e-01 -3.05679142e-01
5.01250505e-01 -2.67868757e-01 -5.77276945e-01 4.73946124e-01
1.46508679e-01 5.17489433e-01 4.16962177e-01 -4.04421836e-01
-3.46920073e-01 -2.15593547e-01 -1.00320375e+00 6.76554561e-01
1.39338827e+00 -7.78989017e-01 -1.78441465e-01 -1.92831848e-02
7.66549170e-01 -6.57148540e-01 -8.41752708e-01 4.69917715e-01
4.54041511e-01 -1.11178768e+00 1.13549352e+00 6.70930818e-02
5.46824455e-01 -6.10576987e-01 -1.44605801e-01 -6.32820368e-01
-1.35906845e-01 -7.17324197e-01 -4.12133425e-01 1.29119313e+00
2.18001321e-01 -4.87747967e-01 8.89167130e-01 5.75116158e-01
-2.53736287e-01 -7.99834311e-01 -1.05956113e+00 -5.39280713e-01
-5.08565187e-01 -8.03941548e-01 1.77210674e-01 4.44045037e-01
-2.09318146e-01 3.39043885e-01 -7.27024153e-02 1.91437691e-01
6.56568110e-01 1.77644148e-01 1.71413213e-01 -7.88178742e-01
-5.55387497e-01 -4.58183140e-01 -5.47435224e-01 -1.25525701e+00
3.51081081e-02 -7.42901087e-01 1.78399459e-01 -1.32200301e+00
1.57031983e-01 -4.15583640e-01 -3.54787290e-01 5.61330318e-01
-1.50565892e-01 6.35191083e-01 1.67635322e-01 1.49586171e-01
-8.25164974e-01 2.84687221e-01 9.03582335e-01 5.15849479e-02
-1.26230448e-01 -5.53868949e-01 -5.35270810e-01 7.22694874e-01
5.08222938e-01 -3.16537410e-01 -5.39828539e-01 -7.99290299e-01
5.84497815e-03 3.20692390e-01 8.92532468e-01 -1.30163753e+00
4.75804359e-01 -9.92519483e-02 7.08031774e-01 -4.24984723e-01
6.53796017e-01 -5.58942556e-01 2.03878149e-01 1.96561307e-01
-1.09832942e-01 -1.23449266e-01 5.70882559e-01 9.01679575e-01
-2.57148623e-01 2.27960184e-01 8.77945125e-01 -1.12637334e-01
-1.32052779e+00 3.85607839e-01 -4.87653553e-01 2.44755939e-01
1.14526796e+00 -4.52551842e-01 -5.77537775e-01 -1.37552589e-01
-3.49389255e-01 1.62938535e-01 6.34442329e-01 1.30215093e-01
5.45526147e-01 -1.05633688e+00 -2.57349283e-01 1.23126060e-01
2.12477800e-02 2.83919483e-01 3.62132221e-01 8.81397784e-01
-5.54481328e-01 2.01574743e-01 -2.43083596e-01 -1.08208525e+00
-8.22646260e-01 2.49036551e-01 2.19136119e-01 5.55919483e-02
-8.05480957e-01 9.78271365e-01 1.67160392e-01 4.49774295e-01
3.00414711e-02 -3.75864297e-01 3.64026845e-01 -1.24637872e-01
5.22216916e-01 2.52768785e-01 -8.06182623e-03 -1.67832926e-01
-4.76248831e-01 3.68796051e-01 2.05976386e-02 -4.40849662e-01
1.28839755e+00 -2.51761705e-01 2.99536586e-01 2.83877373e-01
1.07623053e+00 -3.45521301e-01 -2.17436147e+00 -2.74994045e-01
-1.71795830e-01 -1.64765731e-01 4.75044101e-01 -9.09782529e-01
-1.12354374e+00 7.48892963e-01 6.73767209e-01 -2.10101292e-01
1.63469124e+00 1.56066775e-01 1.30718541e+00 3.32044326e-02
4.38062847e-01 -1.07352197e+00 1.51877195e-01 2.77151823e-01
9.42472219e-02 -1.26455963e+00 -2.74376243e-01 -4.42429006e-01
-6.01481378e-01 1.04114687e+00 5.81187606e-01 -6.40944317e-02
3.39207113e-01 9.41594064e-01 -2.06588224e-01 -1.61035702e-01
-1.04087806e+00 -3.39595556e-01 -2.12994903e-01 5.65747321e-01
8.22067857e-02 -9.13104415e-02 3.79736871e-02 4.16639984e-01
3.18029374e-01 5.12063384e-01 4.30983156e-01 1.01640069e+00
-3.18520308e-01 -8.66525114e-01 -8.44361261e-03 6.36625230e-01
-4.82809246e-01 -3.13945293e-01 2.29477108e-01 4.28673029e-01
1.47209570e-01 9.72378314e-01 5.06533086e-01 -1.64744541e-01
1.90779101e-02 -7.32249394e-02 6.50689542e-01 -2.61990756e-01
-2.31601983e-01 2.82124877e-01 1.22721724e-01 -1.06278372e+00
-6.66362226e-01 -7.54923463e-01 -1.90196419e+00 -2.89499670e-01
6.10504560e-02 -5.79031944e-01 5.42315900e-01 1.04858971e+00
7.91133702e-01 7.58136988e-01 2.78880358e-01 -8.87723505e-01
6.96341321e-02 -2.11118802e-01 -2.67050534e-01 3.06581616e-01
2.75659740e-01 -5.22722125e-01 -1.34822801e-01 5.96001804e-01] | [8.647269248962402, -1.108654499053955] |
de052560-cfb0-43db-8356-d7d4d2b360d6 | friend-or-foe-exploring-the-implications-of | 2306.09928 | null | https://arxiv.org/abs/2306.09928v1 | https://arxiv.org/pdf/2306.09928v1.pdf | Friend or Foe? Exploring the Implications of Large Language Models on the Science System | The advent of ChatGPT by OpenAI has prompted extensive discourse on its potential implications for science and higher education. While the impact on education has been a primary focus, there is limited empirical research on the effects of large language models (LLMs) and LLM-based chatbots on science and scientific practice. To investigate this further, we conducted a Delphi study involving 72 experts specialising in research and AI. The study focused on applications and limitations of LLMs, their effects on the science system, ethical and legal considerations, and the required competencies for their effective use. Our findings highlight the transformative potential of LLMs in science, particularly in administrative, creative, and analytical tasks. However, risks related to bias, misinformation, and quality assurance need to be addressed through proactive regulation and science education. This research contributes to informed discussions on the impact of generative AI in science and helps identify areas for future action. | ['Fabian Sofsky', 'Jörg Pohle', 'Melissa Laufer', 'Marcel Hebing', 'Benedikt Fecher'] | 2023-06-16 | null | null | null | null | ['misinformation'] | ['miscellaneous'] | [ 7.66905546e-02 8.64234447e-01 -3.62069428e-01 2.34687049e-02
-4.51374292e-01 -6.47622705e-01 7.25401878e-01 3.61981481e-01
-4.56498176e-01 3.77239525e-01 7.40989208e-01 -1.06543720e+00
-1.21398218e-01 -4.44181561e-01 -8.66282225e-01 -3.81282687e-01
8.27192843e-01 5.38119301e-02 -8.14584643e-02 1.84796765e-01
9.93673861e-01 3.53774697e-01 -9.47389007e-01 -9.31093618e-02
1.20724225e+00 -2.25580558e-02 1.47114679e-01 2.48126924e-01
-6.28468394e-01 1.47127903e+00 -1.12762105e+00 -6.93445742e-01
-1.84171408e-01 -5.77286005e-01 -9.26974654e-01 -8.34458470e-02
3.89417768e-01 -5.63364863e-01 2.65373588e-01 8.57080638e-01
7.39889085e-01 2.34996244e-01 2.00169712e-01 -8.98697615e-01
-9.85468447e-01 1.01055360e+00 -1.23633526e-01 3.22801620e-01
4.26646739e-01 3.94899577e-01 3.80183935e-01 -2.92898536e-01
9.60688353e-01 1.40737998e+00 2.71768332e-01 1.68422803e-01
-1.04827917e+00 -1.32600987e+00 -1.49552852e-01 -4.41500768e-02
-6.83711588e-01 -4.59865153e-01 6.95486590e-02 -7.67533422e-01
5.87901473e-01 -2.51813494e-02 9.95051920e-01 9.61670578e-01
5.51889598e-01 2.41737496e-02 1.48525357e+00 -7.06890345e-01
3.54954332e-01 7.90050030e-01 2.19671294e-01 1.55893909e-02
8.56739938e-01 -2.93620914e-01 -4.01067138e-01 -4.31986824e-02
7.52484262e-01 -4.25494999e-01 5.70540130e-02 3.78403127e-01
-1.04686594e+00 8.28647554e-01 4.59899940e-02 8.70027840e-01
-3.67546082e-01 6.45069852e-02 1.50348693e-01 1.85702473e-01
5.57277381e-01 8.85502458e-01 3.26558262e-01 -8.61563802e-01
-3.90322179e-01 -1.49662644e-01 1.06796384e+00 3.04085344e-01
2.24362582e-01 4.47731540e-02 1.39021799e-02 4.13257152e-01
2.75126159e-01 4.46219355e-01 -2.53332436e-01 -1.35450697e+00
8.30521882e-02 5.98193526e-01 -5.67163602e-02 -9.65336323e-01
2.22198907e-02 -5.35483837e-01 1.83827177e-01 -7.85617623e-03
5.38518190e-01 -4.56363797e-01 -4.76415873e-01 1.15333104e+00
2.68743932e-01 1.37080491e-01 -2.05939919e-01 8.75118554e-01
9.16935563e-01 7.62050688e-01 6.15081728e-01 -3.66821177e-02
9.04433012e-01 -3.23947877e-01 -9.47674215e-01 -4.87866133e-01
7.26440430e-01 -9.35031474e-01 1.01733756e+00 1.33005142e-01
-1.24291599e+00 -7.94204790e-03 -6.09639943e-01 -2.69049823e-01
-2.79451400e-01 -2.69732386e-01 3.92745405e-01 1.28467035e+00
-8.06197286e-01 2.79398054e-01 -7.57684588e-01 -6.61505401e-01
5.14482379e-01 -2.40543664e-01 1.16744451e-02 -3.37599099e-01
-7.26240218e-01 1.40017664e+00 -1.92306086e-01 3.01390111e-01
-2.76003391e-01 -1.02407217e+00 -4.82995093e-01 -1.99751779e-02
3.89330894e-01 -3.70231658e-01 1.12531197e+00 -9.52841580e-01
-1.50964618e+00 3.38676155e-01 1.99204266e-01 -1.78804919e-01
1.51278481e-01 -1.94988295e-01 -6.17204569e-02 2.01334044e-01
4.80004579e-01 5.50490201e-01 -1.10498324e-01 -1.10209489e+00
-2.21959263e-01 -3.05749208e-01 1.20285163e-02 -5.93284853e-02
-4.08297926e-01 5.14125824e-01 3.59508216e-01 -6.18928485e-02
-2.96717912e-01 -6.74886465e-01 -2.36903708e-02 -2.71980286e-01
1.20116882e-01 -2.49463186e-01 4.94470745e-01 -8.03578556e-01
8.86659980e-01 -1.67803240e+00 -6.41768396e-01 1.74907699e-01
1.47808313e-01 2.71327645e-01 -5.16366214e-03 1.10305464e+00
7.70589769e-01 8.09202492e-01 5.55047154e-01 3.54766607e-01
-8.91633779e-02 1.04668759e-01 -7.57931918e-02 4.18384254e-01
-2.26898015e-01 8.19743693e-01 -1.05273438e+00 -3.76424342e-01
1.25000402e-01 7.41714060e-01 4.89894263e-02 -1.15572110e-01
-4.59653102e-02 5.87460458e-01 -4.09354866e-01 4.10457850e-01
4.37897027e-01 -3.84011686e-01 2.94876575e-01 6.36108279e-01
-1.07544518e+00 8.62685561e-01 -4.16952103e-01 9.64282155e-01
-5.84705830e-01 1.13891101e+00 2.24747643e-01 -4.53354418e-01
8.59522879e-01 5.58276236e-01 -4.69662435e-02 -8.01245511e-01
5.09311974e-01 1.72046900e-01 8.87340128e-01 -8.14134657e-01
8.13403875e-02 1.24851845e-01 6.41458511e-01 9.41778600e-01
-2.14407980e-01 -3.39734524e-01 -1.29667148e-01 2.75448143e-01
8.93188536e-01 3.12545270e-01 -1.82178736e-01 -4.23103243e-01
-2.38212302e-01 2.35089526e-01 4.43965226e-01 8.34690571e-01
-4.06937413e-02 -2.53773570e-01 4.96398836e-01 8.83341059e-02
-7.17568278e-01 9.33088139e-02 -5.61405234e-02 9.49011147e-01
-1.62920952e-02 -2.68334538e-01 -6.71757400e-01 -2.83541828e-02
-1.23480171e-01 1.44065630e+00 -1.67719260e-01 -1.16119854e-01
-8.24941508e-03 -6.98092394e-03 4.04938549e-01 9.64561328e-02
4.42743748e-01 -1.14277840e+00 -1.14527655e+00 1.47987410e-01
-1.65168688e-01 -1.12818587e+00 -1.30013868e-01 -6.52558565e-01
-7.50219941e-01 -9.69002843e-01 -5.72623610e-01 -5.04220545e-01
5.42155027e-01 5.06272674e-01 5.27982891e-01 2.21221568e-03
-1.27100691e-01 7.71634221e-01 -2.10943505e-01 -1.05809915e+00
-1.13397563e+00 -9.49895754e-02 -7.48046696e-01 -5.57964921e-01
5.07554233e-01 -3.98079455e-01 -2.27228567e-01 3.01738083e-02
-8.49874794e-01 1.91429347e-01 8.13153863e-01 1.52484670e-01
-2.65909344e-01 -5.25731862e-01 9.48514402e-01 -1.07655644e+00
9.69310105e-01 -5.32195628e-01 -2.53383636e-01 3.45530808e-01
-8.46199453e-01 -5.44359326e-01 1.45572960e-01 -4.75934446e-01
-1.69523966e+00 -1.01632857e+00 1.03833288e-01 2.56851643e-01
-1.87830701e-01 1.06865060e+00 3.41646403e-01 -6.70509756e-01
8.27250242e-01 -5.18054545e-01 4.61990833e-01 -2.43707210e-01
-4.55650687e-02 6.19937956e-01 -1.97224379e-01 -4.36164200e-01
2.35053867e-01 4.88306955e-02 -2.62456119e-01 -1.32821536e+00
-4.86201912e-01 -1.17177129e-01 2.27254763e-01 -6.64710045e-01
8.96886766e-01 -8.50177705e-01 -1.04305804e+00 -2.22168699e-01
-7.38599598e-01 -5.93762159e-01 6.74673468e-02 1.23051190e+00
2.34728068e-01 2.96872973e-01 -6.20251596e-01 -1.24571908e+00
-2.64525026e-01 -8.45571935e-01 8.80666003e-02 6.37206137e-01
-6.27959251e-01 -9.30855095e-01 -6.92867786e-02 1.17918384e+00
7.45002568e-01 1.28534526e-01 1.10989499e+00 -6.29523516e-01
-7.52046287e-01 -1.75503910e-01 1.05479382e-01 2.39602048e-02
4.97786701e-02 1.40655994e-01 -6.85287535e-01 1.98148265e-01
1.53224841e-01 -6.94411993e-01 1.17750447e-02 3.69660407e-01
3.62405568e-01 -7.25998878e-01 -2.16935412e-03 -2.49441355e-01
1.24515927e+00 6.07826889e-01 3.52042675e-01 9.04512465e-01
4.00823832e-01 1.01511574e+00 6.16883218e-01 2.99188346e-01
4.48357433e-01 -2.88155615e-01 -6.05280362e-02 4.21599329e-01
4.03404459e-02 -3.55508029e-01 3.00091714e-01 8.38560045e-01
-1.53398663e-02 -8.72314423e-02 -1.18111682e+00 6.81247056e-01
-1.36189497e+00 -8.11302185e-01 -3.65952790e-01 2.02725744e+00
5.51798940e-01 1.86961159e-01 -7.08125010e-02 -4.01007712e-01
6.75393105e-01 -3.69514078e-02 -2.40469322e-01 -8.87653768e-01
2.64495164e-01 2.13584319e-01 3.52297395e-01 3.80092502e-01
3.10893133e-02 5.12646139e-01 5.97900009e+00 4.08712924e-01
-9.90732610e-01 1.48625299e-01 6.32798612e-01 6.54189959e-02
-1.21042788e+00 6.43710673e-01 -3.38688403e-01 1.72178388e-01
1.00798738e+00 -5.56248367e-01 1.17787257e-01 2.88363874e-01
6.15843415e-01 -5.84589541e-01 -4.38027978e-01 2.11947970e-02
-1.07147582e-01 -1.47572494e+00 -4.75379050e-01 2.10714906e-01
8.71208012e-01 -9.10620729e-04 -6.43379018e-02 -3.92966308e-02
3.10181499e-01 -7.79899597e-01 5.48646390e-01 2.27723077e-01
2.67945349e-01 -4.77752358e-01 7.41585672e-01 4.87037659e-01
9.94738340e-02 1.10848546e-01 -2.44453385e-01 -7.60130703e-01
1.25921480e-02 1.47333339e-01 -1.15891349e+00 -5.51525466e-02
1.51928455e-01 5.92278019e-02 -1.44592002e-01 9.05702412e-01
-1.68309107e-01 1.52058911e+00 -3.33077461e-01 -4.74259317e-01
2.68499076e-01 -6.54874086e-01 4.52238858e-01 9.57380593e-01
2.02216171e-02 2.69772232e-01 -4.53886002e-01 1.14385450e+00
1.96500659e-01 -2.74395221e-03 -7.86263347e-01 -1.12943232e+00
1.08448708e+00 1.08714879e+00 -1.03442657e+00 9.91891921e-02
-6.65730357e-01 -1.19235292e-01 -2.83559442e-01 5.43374777e-01
-1.54920697e-01 -2.46184364e-01 2.79153407e-01 6.87628210e-01
-9.91844311e-02 -5.46704493e-02 -7.86768496e-01 -4.96167362e-01
-4.29526478e-01 -1.00682950e+00 -2.31086463e-01 -3.23416114e-01
-8.14439476e-01 -1.63312867e-01 -1.09135009e-01 -3.23701322e-01
-1.38721615e-01 2.20740423e-01 -6.35815561e-01 8.46824586e-01
-8.89500260e-01 -1.07740474e+00 8.46033096e-02 -3.62958610e-01
1.67056352e-01 4.30910796e-01 5.78780830e-01 -2.29964554e-01
-4.29379970e-01 2.36482546e-01 2.92579502e-01 -2.58094966e-01
6.74615681e-01 -4.12477583e-01 2.78847337e-01 7.07447946e-01
-3.69627923e-01 8.02784204e-01 7.44695306e-01 -1.22150230e+00
-1.69080436e+00 -4.51904893e-01 1.31962085e+00 -1.78242058e-01
6.35318995e-01 6.56836340e-03 -7.41624117e-01 9.68254387e-01
7.04856932e-01 -9.73555803e-01 9.00609612e-01 2.37892166e-01
2.43134603e-01 2.82774687e-01 -1.06128514e+00 7.74493933e-01
8.28021944e-01 -6.00598335e-01 -2.45003536e-01 3.03140610e-01
3.98054898e-01 -3.75117481e-01 -1.13796365e+00 -3.55517536e-01
6.70866072e-01 -7.34196126e-01 3.50105137e-01 -9.88375992e-02
7.88512945e-01 4.98233646e-01 7.23403096e-01 -9.04886544e-01
-2.00913385e-01 -7.49836445e-01 7.71095991e-01 1.33586311e+00
3.27863753e-01 -9.88577843e-01 6.53895676e-01 1.64253068e+00
-4.12257552e-01 -4.28763300e-01 -7.64685094e-01 -2.75907099e-01
3.24051499e-01 -2.95623571e-01 1.60262600e-01 1.17929208e+00
4.02589202e-01 2.82265455e-01 -1.71092018e-01 -8.93824548e-02
1.72473803e-01 8.14314932e-02 8.62055182e-01 -1.03933382e+00
7.73655251e-02 -3.50142747e-01 -7.66971633e-02 -2.34123215e-01
-1.02322638e-01 -4.97155726e-01 -3.40945542e-01 -2.13571334e+00
5.04486859e-02 -1.40739784e-01 7.77578294e-01 1.54349923e-01
1.81789234e-01 -5.08498728e-01 8.44846666e-01 1.24083318e-01
-1.46260440e-01 2.93779671e-01 1.50265920e+00 2.80335426e-01
-4.82226282e-01 -1.48016378e-01 -1.27539456e+00 3.64297986e-01
9.40672517e-01 -4.64775622e-01 -4.14611578e-01 -3.76573086e-01
6.01818085e-01 3.47046196e-01 2.56269962e-01 -4.68097448e-01
4.72205132e-01 -6.61620021e-01 7.62345716e-02 7.77883008e-02
-1.80387236e-02 -8.70024920e-01 4.91762280e-01 3.09171081e-01
-6.93380296e-01 -9.32177901e-02 2.94359177e-01 -2.13800356e-01
6.63029775e-02 -5.48918843e-01 -4.50597107e-02 -1.94715902e-01
2.22187027e-01 -5.95092356e-01 -1.05288684e+00 1.96895435e-01
8.70149434e-01 -3.76921922e-01 -8.42866302e-01 -1.01760018e+00
-1.49640039e-01 2.91623056e-01 5.49330354e-01 1.87124476e-01
2.55676240e-01 -5.15031278e-01 -3.73925388e-01 -3.08194131e-01
-5.05394638e-01 -1.07612304e-01 2.63152421e-01 9.39735234e-01
-5.36851168e-01 6.96781874e-01 -2.36852154e-01 1.68478504e-01
-7.24947512e-01 -1.50857821e-01 -5.23182675e-02 4.41814661e-01
-6.58299148e-01 6.18367612e-01 1.07971812e-02 -9.93533134e-02
4.23370928e-01 -1.05431043e-01 -1.89608306e-01 -1.22951113e-01
2.19880402e-01 1.25183272e+00 -3.12558800e-01 -2.07498431e-01
-4.21720073e-02 -1.00736983e-01 6.64853975e-02 -5.58512330e-01
1.09670246e+00 -6.63038194e-02 -4.25693750e-01 8.75406444e-01
5.55726409e-01 2.28149369e-01 -8.51320982e-01 1.62440464e-01
2.36053038e-02 -9.50151503e-01 -8.97894129e-02 -1.15749919e+00
-2.95885772e-01 7.29245424e-01 5.94696589e-02 2.79949546e-01
4.19307470e-01 4.06574272e-03 4.95075136e-01 2.31113836e-01
-1.05507240e-01 -1.23395193e+00 -1.00401320e-01 2.16967594e-02
8.66123259e-01 -7.23452806e-01 -1.88093916e-01 -3.60457957e-01
-8.89222682e-01 9.75595593e-01 7.45591044e-01 4.72367793e-01
6.05613768e-01 -2.78398278e-03 3.77576232e-01 -2.73751110e-01
-7.02730536e-01 2.80508757e-01 -3.16093028e-01 4.83150482e-01
9.82659698e-01 3.38097066e-02 -1.18636405e+00 2.79369026e-01
-1.40230775e-01 7.35439718e-01 1.03585505e+00 1.01485848e+00
-4.80289310e-01 -8.52425277e-01 -6.92389846e-01 5.33761084e-01
-9.07078266e-01 2.32419476e-01 -9.45699155e-01 8.15901399e-01
1.09618068e-01 1.42753279e+00 1.05339721e-01 5.05846087e-03
-2.03363463e-01 1.03611112e-01 1.59395337e-01 -5.55138588e-01
-1.01714158e+00 1.57347754e-01 6.07997358e-01 -3.47109400e-02
-4.98046517e-01 -7.74650753e-01 -8.65143955e-01 -7.68500626e-01
-6.17335021e-01 2.04171568e-01 1.05332589e+00 9.63738978e-01
9.11880493e-01 1.86526358e-01 -6.73999265e-02 -4.86060195e-02
-5.34008622e-01 -1.24612653e+00 -1.87653109e-01 -3.66739273e-01
-5.77919520e-02 2.52262652e-02 -8.71126503e-02 -6.22013569e-01] | [10.19016170501709, 7.292768478393555] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.