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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
cdb4f94b-8d73-4180-995d-7bfeb0b0891c | svvad-personal-voice-activity-detection-for | 2305.19581 | null | https://arxiv.org/abs/2305.19581v1 | https://arxiv.org/pdf/2305.19581v1.pdf | SVVAD: Personal Voice Activity Detection for Speaker Verification | Voice activity detection (VAD) improves the performance of speaker verification (SV) by preserving speech segments and attenuating the effects of non-speech. However, this scheme is not ideal: (1) it fails in noisy environments or multi-speaker conversations; (2) it is trained based on inaccurate non-SV sensitive labels. To address this, we propose a speaker verification-based voice activity detection (SVVAD) framework that can adapt the speech features according to which are most informative for SV. To achieve this, we introduce a label-free training method with triplet-like losses that completely avoids the performance degradation of SV due to incorrect labeling. Extensive experiments show that SVVAD significantly outperforms the baseline in terms of equal error rate (EER) under conditions where other speakers are mixed at different ratios. Moreover, the decision boundaries reveal the importance of the different parts of speech, which are largely consistent with human judgments. | ['Jing Xiao', 'Junqing Peng', 'Jianzong Wang', 'Zuheng Kang'] | 2023-05-31 | null | null | null | null | ['action-detection', 'activity-detection', 'speaker-verification'] | ['computer-vision', 'computer-vision', 'speech'] | [ 2.08457544e-01 2.30083335e-02 -2.20826447e-01 -4.17854726e-01
-1.16194725e+00 -6.44595444e-01 5.41660905e-01 2.98094023e-02
-2.50398606e-01 5.27838111e-01 6.65917337e-01 -5.06762862e-01
3.94893140e-01 -9.78100672e-02 -2.23657355e-01 -7.65984178e-01
4.86198217e-01 8.01162645e-02 1.92767650e-01 -2.50879675e-02
9.08608511e-02 4.83087033e-01 -1.67351937e+00 1.86461017e-01
8.41659307e-01 8.63782763e-01 1.63350061e-01 5.35252213e-01
-1.58194944e-01 6.41890407e-01 -8.85410666e-01 -1.73024207e-01
1.51568174e-01 -7.79937029e-01 -5.87650776e-01 1.90455541e-01
4.59115744e-01 -1.28190815e-01 -1.69145912e-01 1.06546092e+00
8.30432415e-01 2.32408345e-01 4.76310819e-01 -1.08476317e+00
-1.81224436e-01 6.27643466e-01 -2.67085999e-01 2.10189819e-01
3.37843806e-01 1.17665529e-01 1.09823918e+00 -1.10763383e+00
2.96522826e-01 1.53064156e+00 6.60965919e-01 9.86343622e-01
-1.31263304e+00 -6.85222328e-01 2.81871498e-01 1.01961985e-01
-1.42833090e+00 -1.12341762e+00 7.35185921e-01 -3.53079706e-01
5.71692348e-01 6.04806423e-01 1.81065053e-01 1.38306987e+00
-4.18608844e-01 7.85433829e-01 9.00066614e-01 -6.15776837e-01
3.12528580e-01 3.75050694e-01 2.70866424e-01 3.87163520e-01
-2.28439614e-01 2.58267432e-01 -7.62062252e-01 -1.95122123e-01
8.35209563e-02 -4.75887895e-01 -6.26978338e-01 -9.73369032e-02
-1.12433076e+00 6.59726083e-01 -6.97874874e-02 5.49625754e-01
-2.51405150e-01 -3.65384698e-01 5.21279931e-01 2.96792120e-01
4.50368643e-01 2.73057967e-01 -3.09122205e-01 -1.51248470e-01
-1.00111699e+00 -2.95802634e-02 5.96416175e-01 6.05246484e-01
5.33212423e-01 5.27775049e-01 -7.94658780e-01 1.24977040e+00
5.30298889e-01 5.19323051e-01 6.76011086e-01 -9.84079540e-01
4.18341100e-01 2.39888757e-01 1.04479253e-01 -3.72822106e-01
-1.46367326e-01 -6.30509377e-01 -5.30610800e-01 1.57027215e-01
4.01069760e-01 -6.89545739e-03 -8.83202195e-01 1.98214316e+00
2.55375326e-01 1.69723049e-01 8.73493329e-02 8.67648184e-01
7.91958690e-01 5.61536014e-01 1.92834213e-01 -7.30528653e-01
1.31618261e+00 -9.90161240e-01 -1.21227300e+00 -3.45070332e-01
3.91793758e-01 -7.40670204e-01 1.40438211e+00 1.53434590e-01
-7.30047047e-01 -5.99833965e-01 -1.05246985e+00 2.29386464e-01
-7.53947794e-02 -4.97573987e-02 2.74535338e-03 1.33736503e+00
-1.06648993e+00 1.29689664e-01 -6.09806955e-01 -3.68182026e-02
9.94614810e-02 1.76161200e-01 -1.65857509e-01 9.08859074e-02
-1.22061074e+00 7.97305584e-01 -1.14893138e-01 -1.65709518e-02
-1.14326441e+00 -5.54063797e-01 -1.09649622e+00 1.28505707e-01
1.65492564e-01 -2.01830342e-01 1.52028275e+00 -1.04406273e+00
-1.92602551e+00 6.43240333e-01 -6.84390187e-01 -2.75618523e-01
5.93376458e-01 7.06352293e-02 -6.77218020e-01 -7.63414288e-03
-4.12172116e-02 3.86051089e-01 1.09469163e+00 -1.34297419e+00
-4.09565508e-01 -2.50624835e-01 -4.91558969e-01 2.47195572e-01
-4.38487619e-01 3.88283908e-01 -2.87365735e-01 -7.62645125e-01
1.35773867e-01 -8.91269743e-01 1.95670500e-01 -2.84148633e-01
-5.35532475e-01 -3.17786098e-01 9.95888174e-01 -9.22274649e-01
1.25495923e+00 -2.58192158e+00 -3.26566130e-01 7.65444189e-02
-1.36369377e-01 6.70671582e-01 -2.02631027e-01 2.16482833e-01
8.73951614e-02 2.02402711e-01 -3.03022772e-01 -9.27197158e-01
1.85660958e-01 1.46599799e-01 -2.85067171e-01 3.39756221e-01
-3.59123722e-02 3.82392764e-01 -7.11508155e-01 -5.60055673e-01
2.76465148e-01 5.74413180e-01 -3.12375903e-01 2.68511653e-01
1.33689955e-01 7.12464809e-01 4.36351933e-02 7.38946199e-01
7.01293230e-01 3.11792642e-01 2.08992898e-01 2.64480524e-02
-1.66694596e-01 8.67617965e-01 -1.05198455e+00 1.14376915e+00
-4.26093638e-01 7.60543525e-01 5.15170872e-01 -5.62494099e-01
8.21811676e-01 7.93197095e-01 1.20978475e-01 -5.24072051e-01
3.02194208e-02 2.08392024e-01 1.16600434e-03 -2.46350437e-01
3.12644541e-01 -2.61207819e-01 1.64394274e-01 2.96225041e-01
-1.46073215e-02 4.18551713e-02 -1.19694360e-01 -3.95272300e-02
9.28875804e-01 -5.12191832e-01 2.93856174e-01 -1.86397433e-01
9.21223819e-01 -7.46482015e-01 1.10056138e+00 6.64980292e-01
-9.21357512e-01 6.23279333e-01 2.14720070e-01 3.68077993e-01
-5.54890275e-01 -1.12858760e+00 -3.28764468e-01 1.20190489e+00
1.05859809e-01 -2.65422583e-01 -9.37175155e-01 -7.08948374e-01
-3.02606016e-01 1.11646056e+00 -3.38122785e-01 -2.25734606e-01
-4.91809458e-01 -3.93699288e-01 7.48113215e-01 4.45798934e-01
2.92751461e-01 -9.50268924e-01 1.04293905e-01 7.59534165e-02
-7.03169227e-01 -1.23836565e+00 -9.63695049e-01 9.22651514e-02
-4.61053282e-01 -7.16878235e-01 -5.89373171e-01 -7.99082696e-01
4.67403769e-01 4.47617948e-01 5.78870177e-01 -1.24677978e-01
3.48821402e-01 1.55231029e-01 -3.44078094e-01 -2.99016595e-01
-1.17852008e+00 -4.72612716e-02 3.50293636e-01 4.00568873e-01
4.29304451e-01 -2.01380700e-01 -2.37444028e-01 6.71237946e-01
-4.60797399e-01 -5.37240744e-01 2.61480868e-01 9.36335146e-01
3.35218370e-01 4.37723659e-03 7.54535079e-01 -4.98897493e-01
6.73536837e-01 -4.86525819e-02 -4.30036753e-01 2.25419566e-01
-7.49601841e-01 -1.58841640e-01 2.44540185e-01 -5.91098428e-01
-1.26720953e+00 -2.21772883e-02 -5.34153521e-01 -2.64626652e-01
-2.73980856e-01 -5.31260483e-02 -6.26359224e-01 -1.35170817e-01
6.14831984e-01 3.04372609e-01 1.10424750e-01 -6.20051861e-01
2.00694520e-02 1.20427704e+00 5.18811345e-01 -1.22873098e-01
6.23667240e-01 2.87589170e-02 -4.61987436e-01 -1.05967236e+00
-6.98478997e-01 -6.36645496e-01 -2.74058700e-01 -2.88735002e-01
5.70193231e-01 -1.07107532e+00 -5.42447209e-01 5.30672133e-01
-1.08613110e+00 -2.96194196e-01 -2.60992825e-01 7.37937987e-01
-1.51230112e-01 7.74885058e-01 -7.30383754e-01 -1.26364768e+00
-2.98561573e-01 -1.36523402e+00 8.81640613e-01 7.01534003e-02
-3.61411929e-01 -5.87087870e-01 -6.71848562e-03 6.82443798e-01
4.61996824e-01 -3.07015181e-01 7.22903788e-01 -8.20632041e-01
-1.09920777e-01 -4.29060347e-02 2.35911876e-01 8.40459824e-01
3.91066849e-01 -2.82123219e-02 -1.65548599e+00 -2.67925531e-01
9.43122953e-02 -4.25987840e-02 7.69504428e-01 5.04746795e-01
8.35751653e-01 -3.19966406e-01 -1.73025563e-01 1.96468771e-01
6.36608839e-01 2.39828408e-01 5.08689642e-01 -6.59098383e-03
4.09006834e-01 8.00557554e-01 4.73023951e-01 2.53054857e-01
2.78389871e-01 1.03687358e+00 2.90287495e-01 -1.35298995e-02
-5.98955095e-01 -3.74383271e-01 9.28815067e-01 9.56407011e-01
3.30508888e-01 -4.60891157e-01 -6.20585501e-01 6.60115778e-01
-1.44467604e+00 -9.96785820e-01 -1.28389820e-01 2.55244231e+00
9.01580334e-01 8.05116221e-02 3.90187532e-01 6.47966027e-01
1.06845272e+00 2.68934339e-01 -2.75660932e-01 -3.16270828e-01
-2.76649356e-01 -2.03481436e-01 1.21581487e-01 8.96251202e-01
-9.72336769e-01 8.54527116e-01 6.74720573e+00 7.63080001e-01
-1.20331192e+00 5.14946997e-01 3.19805741e-01 9.94461328e-02
-3.31855088e-01 -3.14845085e-01 -9.98010516e-01 6.26932323e-01
9.84874010e-01 1.01845436e-01 2.78604150e-01 7.76043952e-01
5.01299620e-01 2.17751250e-01 -9.04823720e-01 8.88255298e-01
2.20070556e-01 -5.79909980e-01 -2.41276741e-01 6.14620149e-02
3.91079038e-01 -4.30629812e-02 8.58319402e-02 3.77962470e-01
6.03331812e-02 -5.95261574e-01 1.08244026e+00 -8.66843164e-02
6.24240577e-01 -5.02120018e-01 7.70786166e-01 3.87182117e-01
-1.01040149e+00 -1.46761522e-01 1.76706053e-02 2.85601377e-01
2.48906478e-01 6.35919034e-01 -1.31307566e+00 1.49288803e-01
3.86966348e-01 2.46237405e-02 -2.99458206e-01 9.14937675e-01
-6.27945662e-01 1.11619651e+00 -1.16068885e-01 1.72181889e-01
-1.80054277e-01 1.88293293e-01 9.37151730e-01 1.37227523e+00
2.77102083e-01 -3.51746023e-01 -7.60388896e-02 6.01201653e-01
-1.41874492e-01 1.37784868e-01 -3.34868550e-01 1.43062010e-01
1.03307414e+00 8.71972680e-01 -3.15097243e-01 -1.66132420e-01
-2.51446187e-01 9.05957758e-01 -6.60502017e-02 5.54937065e-01
-5.48871338e-01 -9.04822350e-02 9.69743669e-01 4.14475612e-02
2.93025345e-01 -3.46686058e-02 -2.15470850e-01 -9.91255820e-01
2.89845560e-02 -8.83402050e-01 3.15540582e-01 -3.60229582e-01
-1.04146111e+00 7.21262872e-01 -3.75229865e-01 -1.26545203e+00
-4.32638735e-01 -1.41889960e-01 -4.81224597e-01 7.53512919e-01
-1.66733253e+00 -8.32681477e-01 -1.68316677e-01 4.90265667e-01
7.34570026e-01 -7.72421882e-02 8.81169140e-01 5.92377007e-01
-7.46773720e-01 1.18556738e+00 -1.55729055e-01 1.62192941e-01
9.32599425e-01 -1.10344648e+00 3.80613059e-01 1.14763415e+00
1.87521115e-01 3.90928686e-01 8.67893815e-01 -1.89712241e-01
-6.79395080e-01 -1.01084912e+00 1.45824599e+00 -3.21339965e-01
6.12780824e-02 -5.48769593e-01 -1.04807162e+00 4.30846572e-01
-5.34736440e-02 -7.71854520e-02 9.64907706e-01 1.95804328e-01
-5.31862795e-01 -2.36382663e-01 -1.22455943e+00 4.19484138e-01
9.17360008e-01 -9.25145686e-01 -6.00478768e-01 1.58192694e-01
7.61387885e-01 -7.12953657e-02 -4.53031927e-01 3.82546067e-01
1.86897546e-01 -9.30369616e-01 8.05746555e-01 -2.16559440e-01
-5.52121997e-01 -5.44185519e-01 -3.47901911e-01 -1.40447211e+00
-3.43574792e-01 -7.33996749e-01 -8.89357850e-02 1.76221895e+00
5.29072762e-01 -7.08734512e-01 3.57657790e-01 4.38086867e-01
-4.52784061e-01 3.76979895e-02 -1.33677840e+00 -1.08560765e+00
-3.95682842e-01 -4.42440450e-01 7.13646054e-01 1.07678759e+00
7.12147132e-02 4.11227703e-01 -4.40036744e-01 4.03603971e-01
3.91803116e-01 -2.12134421e-01 4.80295897e-01 -1.16321671e+00
-1.22606024e-01 -5.86118281e-01 -1.49310738e-01 -9.60444868e-01
4.17300224e-01 -5.51225007e-01 5.67991495e-01 -1.04839885e+00
-1.36526883e-01 -2.15411723e-01 -5.03214896e-01 6.02871358e-01
-4.53730315e-01 4.96666431e-02 1.21399201e-01 2.58748978e-01
-2.61059433e-01 7.52640426e-01 9.24232900e-01 -2.33152196e-01
-4.16269183e-01 4.45358753e-01 -6.54214859e-01 5.08124888e-01
6.91603899e-01 -4.64475721e-01 -2.01277956e-01 -2.47583259e-03
-6.11025393e-01 5.74404895e-02 7.85740763e-02 -8.86968195e-01
-1.34548116e-02 -4.03393582e-02 -2.21834257e-01 -3.55876982e-01
4.56525356e-01 -5.66833913e-01 -1.03838757e-01 4.55359310e-01
-4.95633692e-01 -4.86013591e-01 6.38454854e-02 5.47958016e-01
-4.35600847e-01 -2.81238526e-01 1.13211751e+00 2.90010899e-01
-3.58839005e-01 -1.61064155e-02 -7.41034389e-01 -1.39596015e-01
6.66719675e-01 -8.24204013e-02 -2.35170230e-01 -5.86083293e-01
-4.43367600e-01 2.11570095e-02 3.34181279e-01 5.15982330e-01
1.44001558e-01 -1.20354962e+00 -7.71189392e-01 3.17442507e-01
2.47986600e-01 -4.35512722e-01 2.01821551e-01 9.07552898e-01
4.92881611e-02 3.34508866e-01 3.47171187e-01 -6.00553989e-01
-1.66827476e+00 3.98357302e-01 5.37456810e-01 5.74350997e-04
-4.03554648e-01 9.29887235e-01 3.24526548e-01 -5.54379284e-01
6.97018921e-01 -1.82158902e-01 -2.11229712e-01 3.33232358e-02
7.92315602e-01 3.74108046e-01 5.08379638e-01 -9.98933434e-01
-5.74819267e-01 1.05118388e-02 2.64705298e-03 -3.88447881e-01
7.09965110e-01 -4.14186865e-01 3.69540393e-01 4.41848040e-01
9.24703717e-01 5.11322677e-01 -1.35172832e+00 -3.68598789e-01
-3.88828591e-02 -4.88983154e-01 3.16918790e-01 -8.66901696e-01
-7.76217818e-01 9.94921982e-01 7.93998599e-01 1.82033494e-01
9.77873564e-01 -1.45548180e-01 7.83483028e-01 -6.54689521e-02
-5.60839735e-02 -1.18145585e+00 -3.43618244e-02 4.28422421e-01
7.57224441e-01 -1.11668289e+00 -4.33770567e-01 -6.39088213e-01
-6.65792167e-01 7.62606323e-01 3.94753635e-01 8.76385093e-01
4.31911975e-01 1.50755979e-02 5.89456081e-01 2.97945321e-01
-5.36665857e-01 -3.84133071e-01 2.69603699e-01 7.30614066e-01
2.79076427e-01 1.15221873e-01 -2.65648574e-01 5.24453521e-01
-1.88826337e-01 -4.80802149e-01 2.19617426e-01 5.66820323e-01
-5.42655110e-01 -1.14358330e+00 -6.14257574e-01 -6.43453887e-03
-4.11647409e-01 -1.02752261e-01 -5.68694532e-01 2.24756792e-01
-8.04323852e-02 1.57390440e+00 -2.06298187e-01 -4.60921973e-01
3.73424053e-01 5.23189247e-01 8.34616795e-02 -5.76802671e-01
-6.34768128e-01 3.00714165e-01 2.96440542e-01 -2.29165301e-01
-3.72712314e-01 -9.22188997e-01 -1.05575550e+00 -1.59419104e-01
-7.36465931e-01 3.09989899e-01 8.66796732e-01 9.88986731e-01
3.92274737e-01 5.44886947e-01 9.88698542e-01 -4.77392048e-01
-8.42292726e-01 -1.26169169e+00 -6.49651110e-01 3.77348155e-01
7.77069330e-01 -7.24415362e-01 -9.41085100e-01 -1.79064237e-02] | [14.52081298828125, 6.120588779449463] |
d8737fe7-b4f9-4855-911b-c19abf4fa2f1 | hyrr-hybrid-infused-reranking-for-passage | 2212.10528 | null | https://arxiv.org/abs/2212.10528v1 | https://arxiv.org/pdf/2212.10528v1.pdf | HYRR: Hybrid Infused Reranking for Passage Retrieval | We present Hybrid Infused Reranking for Passages Retrieval (HYRR), a framework for training rerankers based on a hybrid of BM25 and neural retrieval models. Retrievers based on hybrid models have been shown to outperform both BM25 and neural models alone. Our approach exploits this improved performance when training a reranker, leading to a robust reranking model. The reranker, a cross-attention neural model, is shown to be robust to different first-stage retrieval systems, achieving better performance than rerankers simply trained upon the first-stage retrievers in the multi-stage systems. We present evaluations on a supervised passage retrieval task using MS MARCO and zero-shot retrieval tasks using BEIR. The empirical results show strong performance on both evaluations. | ['Jianmo Ni', 'Ji Ma', 'Keith Hall', 'Jing Lu'] | 2022-12-20 | null | null | null | null | ['passage-retrieval'] | ['natural-language-processing'] | [ 3.64415310e-02 -2.49187857e-01 -4.45877165e-01 1.00186564e-01
-1.65086424e+00 -3.87797832e-01 1.05583525e+00 3.47660601e-01
-8.76720965e-01 6.36701167e-01 9.99563277e-01 -4.42962125e-02
-6.02350414e-01 -5.59291363e-01 -5.05823851e-01 -1.43798932e-01
-1.42538458e-01 9.07689512e-01 4.64239180e-01 -9.13915813e-01
6.55660331e-01 3.54029834e-01 -1.48909891e+00 7.40239263e-01
4.58198160e-01 7.35594153e-01 1.34825721e-01 1.25099981e+00
-6.56669736e-02 9.44261074e-01 -8.68158638e-01 -1.17629327e-01
2.36367490e-02 -1.41615734e-01 -1.25026178e+00 -7.76305497e-01
8.38859856e-01 -7.61791945e-01 -1.01844168e+00 4.54286784e-01
8.42193007e-01 7.76429653e-01 9.98005152e-01 -4.50802773e-01
-1.28314388e+00 9.25041497e-01 -6.18200861e-02 7.71743298e-01
7.74844229e-01 -5.45531452e-01 1.70429254e+00 -9.89060521e-01
7.35402465e-01 1.16928601e+00 5.53826749e-01 6.57880902e-01
-9.73208845e-01 -1.24964654e-01 -3.02386016e-01 3.08244377e-01
-1.21099079e+00 -5.04838288e-01 1.52120501e-01 -8.94574672e-02
1.62351024e+00 4.01482940e-01 2.26305082e-01 8.86756003e-01
-7.60385171e-02 1.39834118e+00 3.20855498e-01 -6.02286994e-01
-1.65937588e-01 -1.65295258e-01 9.38205659e-01 3.94764036e-01
9.83605534e-02 1.17002137e-01 -4.53136742e-01 -5.23063004e-01
5.35233080e-01 2.77008742e-01 -4.47155565e-01 2.31917635e-01
-1.12225831e+00 7.01769650e-01 7.56151438e-01 7.78576851e-01
-4.40252662e-01 2.59368837e-01 5.89092612e-01 7.83146441e-01
5.44027090e-01 1.37198317e+00 -3.55719209e-01 1.63111746e-01
-1.43849003e+00 2.54486412e-01 7.27914095e-01 7.34784305e-01
4.95742023e-01 -2.33300805e-01 -1.38704884e+00 1.40560818e+00
3.35552543e-01 3.44342202e-01 1.18075514e+00 -6.19244516e-01
1.58317566e-01 2.22441077e-01 3.30271631e-01 -6.49859071e-01
-3.03366423e-01 -4.60539222e-01 -5.37147701e-01 -4.10766602e-01
-2.04393297e-01 2.86541432e-01 -1.31117320e+00 1.15813875e+00
-5.27186215e-01 6.84161633e-02 4.70128894e-01 7.13083863e-01
1.44162571e+00 9.84848320e-01 1.07805833e-01 5.83201796e-02
1.04947686e+00 -1.59749615e+00 -5.23877561e-01 2.37328559e-01
8.53955209e-01 -8.50375473e-01 9.43792701e-01 -9.53534897e-03
-1.35460401e+00 -6.03775024e-01 -1.17772305e+00 -5.06593764e-01
-6.34635746e-01 2.36096561e-01 3.51392210e-01 2.89385796e-01
-1.66643286e+00 1.07917655e+00 -2.05317393e-01 -5.32751918e-01
-1.01036087e-01 3.14500064e-01 -7.91774690e-02 -2.68318862e-01
-1.75386012e+00 1.13203633e+00 4.17066932e-01 -2.51866877e-01
-1.22205210e+00 -7.89332628e-01 -4.98058379e-01 4.74979550e-01
-2.09729925e-01 -8.18897903e-01 1.68659759e+00 -5.90479314e-01
-1.50577617e+00 7.00406730e-01 -1.21976539e-01 -6.95269525e-01
1.59022406e-01 -5.85507810e-01 -5.03463268e-01 6.62821352e-01
6.57915026e-02 6.51446402e-01 5.40964425e-01 -1.16242206e+00
-5.18019676e-01 2.89546311e-01 3.16868603e-01 6.35674119e-01
-3.93307447e-01 2.25120068e-01 -6.11864507e-01 -5.84723949e-01
-3.71930718e-01 -5.28116763e-01 -8.42689350e-02 -8.31220627e-01
-2.71809489e-01 -7.71260679e-01 3.97418827e-01 -6.56306803e-01
1.53048515e+00 -1.57142007e+00 -5.69324121e-02 7.97547549e-02
1.98298931e-01 8.36500943e-01 -8.54310930e-01 8.24919224e-01
6.75422922e-02 4.13813233e-01 5.80957055e-01 -2.98474073e-01
2.10086390e-01 -2.25952968e-01 -7.10608006e-01 -1.32455165e-02
-2.33250871e-01 1.26838660e+00 -1.15494776e+00 -3.77104998e-01
-8.32541361e-02 3.56233656e-01 -3.25200558e-01 5.19821644e-01
-1.98967218e-01 -4.39310193e-01 -5.63651383e-01 5.04633009e-01
-9.34358835e-02 -5.60231626e-01 -5.10126054e-01 1.30987585e-01
3.78651440e-01 5.73369205e-01 -3.49692374e-01 1.65293169e+00
-4.22800124e-01 7.82515407e-01 -6.58829391e-01 -4.33267474e-01
5.07687509e-01 7.85534561e-01 3.72576594e-01 -7.92980433e-01
-1.05269551e-01 2.09548324e-01 -4.37746793e-01 -3.13228220e-01
1.54832006e+00 1.33296564e-01 -3.75439599e-02 8.72919798e-01
3.60636592e-01 4.98698801e-02 7.16232002e-01 9.74015415e-01
1.54164004e+00 -1.51318729e-01 1.08525664e-01 -3.42539102e-02
3.84858757e-01 1.41356632e-01 -3.26023012e-01 1.84168935e+00
1.05723649e-01 9.08609629e-01 -3.57079327e-01 -1.09040886e-01
-1.00949132e+00 -9.24402177e-01 7.55788237e-02 1.81244051e+00
4.82976176e-02 -5.05274713e-01 -1.94923699e-01 -5.44767261e-01
2.10226141e-02 6.32394850e-01 -7.31567144e-01 -4.10751969e-01
-3.30384761e-01 -5.66589952e-01 9.37027574e-01 4.33767110e-01
1.55491933e-01 -1.35790801e+00 1.06414713e-01 3.60193372e-01
-2.57528573e-01 -4.89310890e-01 -6.26048386e-01 -5.83579503e-02
-8.14925194e-01 -7.19556808e-01 -1.73307621e+00 -8.66487741e-01
2.77333111e-01 8.11742842e-01 1.69285405e+00 8.71452153e-01
-3.17770019e-02 1.12139392e+00 -8.44006717e-01 -1.49207208e-02
-3.68906289e-01 6.88399732e-01 1.49097502e-01 -7.13520288e-01
3.38687330e-01 -7.79166594e-02 -8.16940904e-01 1.52972773e-01
-1.31294119e+00 -5.65919697e-01 7.47432172e-01 1.15852201e+00
2.14527667e-01 -4.19538766e-01 7.75217056e-01 -6.91816807e-01
1.42672443e+00 -5.53171098e-01 -2.58030463e-02 9.21280086e-01
-9.03208494e-01 6.09257631e-02 2.38053113e-01 -5.75878382e-01
-8.91935587e-01 -8.82726133e-01 -1.82402596e-01 -3.19854051e-01
2.11721107e-01 7.81527102e-01 7.73604214e-01 -1.01146437e-02
1.01377177e+00 2.25261852e-01 -5.38068652e-01 -7.03217030e-01
6.13899589e-01 9.23866391e-01 1.52847603e-01 -5.92058063e-01
6.43729866e-01 -4.91119288e-02 -6.36125803e-01 -9.17811453e-01
-1.02322912e+00 -1.36529183e+00 -4.16955978e-01 -1.86735496e-01
5.64420998e-01 -1.06391716e+00 -6.26706481e-02 2.17345521e-01
-1.16791379e+00 -2.01803625e-01 -4.77389276e-01 4.72205430e-01
-2.81433165e-01 5.69527268e-01 -1.30161107e+00 -4.13282216e-01
-9.33498144e-01 -7.26093829e-01 1.16357768e+00 4.35827047e-01
-2.44313270e-01 -1.07381618e+00 7.79864371e-01 3.72980803e-01
8.73213649e-01 -7.30888486e-01 9.62644637e-01 -1.46662307e+00
-3.72883588e-01 -7.94638336e-01 -3.64342958e-01 1.19582102e-01
-2.24978775e-01 -1.00366630e-01 -9.45570171e-01 -6.06401861e-01
-7.21099734e-01 -8.96598518e-01 1.55511248e+00 2.30832443e-01
6.04888439e-01 -2.58484304e-01 -2.46780634e-01 -1.12934954e-01
1.33695710e+00 -9.30599496e-02 8.62974942e-01 5.67992091e-01
4.93307084e-01 8.07919651e-02 2.50102818e-01 4.77529839e-02
1.00266851e-01 4.33506012e-01 -2.59733647e-01 -9.75048095e-02
-3.84110570e-01 -2.91894168e-01 3.93036097e-01 1.24351633e+00
-1.11916117e-01 -6.63388968e-01 -7.18650937e-01 5.67744672e-01
-1.71591568e+00 -1.19281185e+00 2.58708358e-01 2.01827502e+00
8.95628452e-01 -1.65684983e-01 -7.28707910e-02 -3.60158980e-01
6.36586070e-01 4.04574722e-01 -3.26841861e-01 -2.30943650e-01
-1.13101214e-01 4.82075572e-01 3.76407802e-01 6.99460268e-01
-9.49820995e-01 1.17832363e+00 8.27121353e+00 1.10455143e+00
-4.81310099e-01 -8.69636387e-02 2.22346395e-01 -2.41517469e-01
-3.12498957e-01 -2.96542972e-01 -9.56711769e-01 -1.87214930e-02
1.13871801e+00 -6.49544299e-01 5.15344083e-01 7.22255349e-01
-3.88890535e-01 1.46187231e-01 -1.19335413e+00 4.94215041e-01
5.96013546e-01 -1.57176530e+00 8.17344487e-01 -4.42968577e-01
9.75051522e-01 6.26430392e-01 -7.40370825e-02 1.29156983e+00
8.25683951e-01 -1.03257346e+00 1.06483385e-01 8.01463068e-01
3.94649208e-01 -3.28441978e-01 1.00174177e+00 1.51642095e-02
-6.90738678e-01 1.83606684e-01 -8.25094581e-01 2.65241265e-01
3.12813334e-02 1.28928468e-01 -8.48864615e-01 6.41293883e-01
6.18495524e-01 6.49148464e-01 -8.18916023e-01 1.63978350e+00
1.00323789e-01 4.61768150e-01 -8.40500444e-02 -5.22019207e-01
4.35582608e-01 5.07992327e-01 6.80956483e-01 1.58593321e+00
6.92279935e-02 8.76471028e-02 1.13370098e-01 4.38254047e-03
-5.31409502e-01 4.13506895e-01 -3.67200524e-01 -4.27225083e-01
2.45624304e-01 1.00228393e+00 -3.44799250e-01 -1.00686955e+00
-4.70547751e-03 9.52173829e-01 5.25751829e-01 8.09231043e-01
-2.95989990e-01 -7.74010599e-01 3.04116756e-01 -2.91187346e-01
1.39980957e-01 1.43562838e-01 4.18099701e-01 -1.40025520e+00
-5.15618503e-01 -7.35637069e-01 8.33449721e-01 -1.05979204e+00
-1.63120699e+00 8.77819419e-01 1.37389556e-01 -1.20187008e+00
-4.51101422e-01 -4.34542269e-01 -5.82015216e-01 8.11745167e-01
-1.89900637e+00 -7.04785228e-01 1.21558771e-01 3.68470609e-01
7.19421387e-01 -4.18154806e-01 1.10819733e+00 3.95166069e-01
-2.20157385e-01 7.18396842e-01 8.48212421e-01 3.03272545e-01
9.35922742e-01 -1.39327443e+00 2.39205524e-01 4.75190938e-01
5.25443256e-01 1.11495912e+00 4.23182279e-01 -5.37099659e-01
-1.15203023e+00 -7.66671479e-01 9.81749415e-01 -5.73332906e-01
7.23429263e-01 4.81576651e-01 -1.05510986e+00 5.76272964e-01
7.15304196e-01 -6.29530549e-01 9.60796893e-01 5.45430720e-01
-5.50320327e-01 2.33569950e-01 -7.56848276e-01 6.66655421e-01
6.18679166e-01 -9.61598098e-01 -1.42258835e+00 7.50957131e-01
8.99052083e-01 -1.74075842e-01 -1.06054866e+00 4.12365019e-01
7.46162117e-01 -4.31352645e-01 1.46476007e+00 -1.08930373e+00
4.97824848e-01 8.96198396e-03 -1.27925068e-01 -1.46218562e+00
-7.88690567e-01 -4.36157674e-01 -4.02089804e-01 1.00868309e+00
6.77709341e-01 -3.73909146e-01 4.83579040e-01 4.86319005e-01
-2.46806093e-03 -4.16518301e-01 -6.81813121e-01 -7.35831320e-01
3.38613749e-01 9.60348770e-02 4.22440827e-01 7.16804147e-01
1.89590856e-01 7.44311392e-01 -2.91255027e-01 -2.33810350e-01
1.12332866e-01 -8.67799222e-02 6.07389212e-01 -1.13150311e+00
-3.83654863e-01 -9.21602488e-01 -4.85987440e-02 -1.58871257e+00
1.11632869e-01 -1.22342885e+00 4.40777123e-01 -2.08101630e+00
4.80647266e-01 -2.23273695e-01 -1.20323312e+00 3.16457957e-01
-4.66888338e-01 6.26746356e-01 2.31595393e-02 7.64704466e-01
-1.38439310e+00 6.17350638e-01 9.80566561e-01 -6.04172349e-01
-1.22024462e-01 -1.23549253e-01 -6.67930007e-01 2.76485205e-01
5.78018606e-01 -5.74681401e-01 -3.53545815e-01 -6.59345865e-01
4.25149560e-01 1.72634572e-01 -2.86098331e-01 -1.02385914e+00
4.91961867e-01 4.52474862e-01 5.02543986e-01 -8.47302616e-01
3.34663302e-01 -2.04302773e-01 -5.24627507e-01 1.39955029e-01
-1.31096268e+00 2.52362132e-01 -7.92726700e-04 6.66960776e-01
-3.67307663e-01 -8.08230221e-01 3.95336926e-01 -3.85867327e-01
-8.05306375e-01 3.63915950e-01 -7.27304220e-01 1.63362280e-01
4.54994440e-02 1.00446247e-01 -6.84716761e-01 -8.09516788e-01
-6.53346479e-01 4.66782004e-01 1.70138955e-01 6.75457835e-01
7.91367471e-01 -1.41048944e+00 -9.99558866e-01 -4.27750111e-01
4.66871053e-01 -7.61657596e-01 1.83236986e-01 3.37512881e-01
-5.94853461e-01 9.30039406e-01 1.47378445e-01 -1.36084393e-01
-1.25836349e+00 4.88306463e-01 1.12558581e-01 -1.12423933e+00
-4.43569124e-01 8.91435564e-01 -3.48351628e-01 -3.16328108e-01
5.32992184e-01 5.38260788e-02 -9.52241898e-01 3.76263738e-01
1.07817698e+00 5.34011364e-01 2.77498722e-01 -5.29761553e-01
1.39293864e-01 3.34601730e-01 -9.17939007e-01 -3.75492692e-01
1.05049109e+00 -1.80806145e-01 -1.06332660e-01 4.05893743e-01
1.55049062e+00 -2.00582340e-01 -2.34523535e-01 -6.51515365e-01
2.62371540e-01 -1.42429635e-01 3.75675052e-01 -1.07671428e+00
-3.49672645e-01 6.37952089e-01 4.40246820e-01 3.74381930e-01
6.91273510e-01 -1.66126594e-01 1.09977293e+00 1.37077057e+00
2.01114595e-01 -1.15210950e+00 4.03933257e-01 1.09693873e+00
9.71800029e-01 -1.07865047e+00 2.85497218e-01 7.09231317e-01
-4.25613403e-01 1.10180533e+00 2.59328932e-01 -4.40928787e-01
5.52824378e-01 -6.63155377e-01 2.88233370e-01 -2.42823422e-01
-9.10701096e-01 -5.49641788e-01 1.02411103e+00 9.22517776e-02
5.64270377e-01 -2.26113990e-01 -1.97411239e-01 7.51631781e-02
9.52735990e-02 3.55284810e-02 3.64657909e-01 8.41197729e-01
-7.51804411e-01 -8.83259475e-01 -7.25885108e-02 8.39926124e-01
-8.34912956e-01 -7.15045094e-01 -5.53867221e-01 4.43567961e-01
-8.70451927e-01 1.10894752e+00 -4.07776274e-02 -3.99072230e-01
1.03167146e-01 2.02836439e-01 3.58274668e-01 -8.43406022e-01
-1.14460874e+00 -2.03801945e-01 3.90459567e-01 -3.28423947e-01
-3.86694700e-01 -1.69541270e-01 -6.45694852e-01 3.55647653e-02
-7.02913582e-01 9.60045934e-01 3.12685966e-01 8.44721615e-01
3.11238021e-01 4.21877176e-01 6.15564585e-01 -1.13901901e+00
-6.54035330e-01 -1.49672925e+00 -5.44891238e-01 4.21919525e-01
5.51365852e-01 -2.73119152e-01 -5.11752069e-01 -4.89110589e-01] | [11.508254051208496, 7.660204887390137] |
de7811e8-8a8f-42b0-b6aa-6d62ca1e2a4f | efficient-semantic-segmentation-by-altering | 2303.07224 | null | https://arxiv.org/abs/2303.07224v1 | https://arxiv.org/pdf/2303.07224v1.pdf | Efficient Semantic Segmentation by Altering Resolutions for Compressed Videos | Video semantic segmentation (VSS) is a computationally expensive task due to the per-frame prediction for videos of high frame rates. In recent work, compact models or adaptive network strategies have been proposed for efficient VSS. However, they did not consider a crucial factor that affects the computational cost from the input side: the input resolution. In this paper, we propose an altering resolution framework called AR-Seg for compressed videos to achieve efficient VSS. AR-Seg aims to reduce the computational cost by using low resolution for non-keyframes. To prevent the performance degradation caused by downsampling, we design a Cross Resolution Feature Fusion (CReFF) module, and supervise it with a novel Feature Similarity Training (FST) strategy. Specifically, CReFF first makes use of motion vectors stored in a compressed video to warp features from high-resolution keyframes to low-resolution non-keyframes for better spatial alignment, and then selectively aggregates the warped features with local attention mechanism. Furthermore, the proposed FST supervises the aggregated features with high-resolution features through an explicit similarity loss and an implicit constraint from the shared decoding layer. Extensive experiments on CamVid and Cityscapes show that AR-Seg achieves state-of-the-art performance and is compatible with different segmentation backbones. On CamVid, AR-Seg saves 67% computational cost (measured in GFLOPs) with the PSPNet18 backbone while maintaining high segmentation accuracy. Code: https://github.com/THU-LYJ-Lab/AR-Seg. | ['Yong-Jin Liu', 'Jiangtao Wen', 'Yuxing Han', 'Jisheng Li', 'Yanghao Li', 'Yuze He', 'Yubin Hu'] | 2023-03-13 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Hu_Efficient_Semantic_Segmentation_by_Altering_Resolutions_for_Compressed_Videos_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Hu_Efficient_Semantic_Segmentation_by_Altering_Resolutions_for_Compressed_Videos_CVPR_2023_paper.pdf | cvpr-2023-1 | ['video-semantic-segmentation'] | ['computer-vision'] | [ 2.40776494e-01 -2.36835569e-01 -2.25678369e-01 -3.00631523e-01
-7.28464127e-01 -7.71787390e-02 2.45517641e-01 -2.33222961e-01
-6.45441532e-01 4.07246888e-01 6.44731671e-02 -7.07562715e-02
8.94284397e-02 -9.03673053e-01 -7.01001704e-01 -6.35394692e-01
-8.27436298e-02 -1.38552904e-01 9.51827645e-01 -1.76322177e-01
2.34849736e-01 4.65435594e-01 -1.63561404e+00 4.00857449e-01
8.85450840e-01 1.20802641e+00 7.89327979e-01 7.43041039e-01
-1.00434661e-01 6.11615658e-01 -3.16531003e-01 -1.12852693e-01
6.42189026e-01 -4.56129879e-01 -7.99897730e-01 3.22921053e-02
3.84488940e-01 -5.08181930e-01 -4.69878018e-01 9.93595898e-01
4.47565466e-01 3.49681437e-01 2.69113272e-01 -1.14419568e+00
-7.56299496e-02 2.69464582e-01 -8.02745640e-01 6.12118840e-01
3.09339195e-01 4.00186479e-02 7.15238810e-01 -8.26979280e-01
7.22905695e-01 1.27771175e+00 6.52752221e-01 3.72495234e-01
-8.79164994e-01 -7.48205006e-01 1.57665402e-01 4.26917404e-01
-1.57664096e+00 -4.75983530e-01 4.38422680e-01 -1.72554299e-01
9.70119119e-01 2.95664787e-01 6.89579129e-01 6.74193263e-01
2.34476715e-01 6.23566329e-01 7.28933811e-01 -1.60342053e-01
1.52284622e-01 -3.11517477e-01 -9.99435797e-05 8.11571240e-01
1.62312482e-02 -1.22075722e-01 -6.19806945e-01 2.29284078e-01
1.23159683e+00 -3.29298861e-02 -5.08623004e-01 -2.19653975e-02
-1.09279382e+00 7.99147248e-01 4.87560034e-01 3.91095906e-01
-4.37077194e-01 4.39104587e-02 4.38453048e-01 2.77892172e-01
4.15840328e-01 3.37696373e-02 -6.17943406e-01 -1.61256939e-01
-1.37143600e+00 1.17724560e-01 4.21757996e-01 1.05077875e+00
8.46220195e-01 1.00091314e-02 -1.07384279e-01 8.08022678e-01
2.33511761e-01 2.27778703e-01 5.63760459e-01 -1.40474391e+00
6.70222461e-01 3.08646500e-01 -2.11201906e-01 -1.20637977e+00
-3.69678468e-01 -1.54968083e-01 -1.02718580e+00 5.10364771e-02
1.50773093e-01 -8.04599673e-02 -1.05690789e+00 1.40582728e+00
4.29258257e-01 5.82686245e-01 5.74505702e-02 1.28387511e+00
8.30072999e-01 1.03704453e+00 9.61869806e-02 -3.81440580e-01
1.31799924e+00 -1.25207663e+00 -4.27368790e-01 -4.95209806e-02
6.09522045e-01 -8.62055779e-01 7.91474044e-01 2.80375540e-01
-1.20209730e+00 -9.03212368e-01 -1.02746105e+00 -2.72324622e-01
-1.26158446e-01 -7.85059631e-02 2.49312490e-01 2.40372807e-01
-1.11556745e+00 8.28953147e-01 -1.06483781e+00 -3.97159159e-01
4.67473418e-01 4.93673146e-01 -3.26222211e-01 -1.26091214e-02
-1.22637498e+00 4.87437546e-01 6.27210319e-01 -5.89221828e-02
-5.04359782e-01 -6.45631194e-01 -9.48685169e-01 1.02877855e-01
4.47358519e-01 -6.70920372e-01 1.02921367e+00 -1.31661761e+00
-1.55094707e+00 3.73266399e-01 -2.60334551e-01 -6.33033633e-01
5.34338593e-01 -3.29050779e-01 -2.82765567e-01 5.86184502e-01
1.28453344e-01 1.02125478e+00 9.53664780e-01 -7.81989932e-01
-1.08448541e+00 -2.30587181e-02 -1.54481772e-02 4.89826679e-01
-1.70642778e-01 2.50303388e-01 -9.46286500e-01 -8.45957577e-01
3.20502937e-01 -8.00172448e-01 -3.43928754e-01 -8.06811452e-03
-2.81129275e-02 1.64429732e-02 1.06102777e+00 -8.93899918e-01
1.25194633e+00 -2.31522155e+00 1.62334323e-01 1.03471085e-01
-3.65785742e-03 6.09586418e-01 -1.60876393e-01 -7.59607181e-02
9.68468711e-02 7.26505518e-02 -3.17013204e-01 -1.82621866e-01
-4.34614956e-01 2.41900206e-01 -8.73877853e-02 3.57452661e-01
2.29211256e-01 5.49377620e-01 -6.41930163e-01 -8.56059372e-01
5.29399216e-01 7.00869203e-01 -7.79661357e-01 1.46547362e-01
3.66182961e-02 3.80808264e-01 -5.23358941e-01 6.24803185e-01
9.12309289e-01 -1.60277411e-01 -8.81653950e-02 -5.24251759e-01
-4.75203216e-01 7.40507171e-02 -1.45705569e+00 1.90183043e+00
-1.72412604e-01 4.33642775e-01 2.48460889e-01 -1.03451002e+00
6.94049597e-01 1.13666154e-01 5.78046679e-01 -9.37264085e-01
3.51558328e-01 2.26893872e-01 -2.29136139e-01 -4.34884727e-01
7.21747696e-01 3.21208566e-01 2.03931838e-01 -1.93852901e-01
1.40482336e-01 1.35200053e-01 2.80825973e-01 1.24225117e-01
9.57508922e-01 4.81636345e-01 2.86711007e-01 -1.97562829e-01
8.24066997e-01 -5.84228337e-02 1.02110612e+00 3.04296732e-01
-2.16287404e-01 9.32800412e-01 2.51990169e-01 -4.25027132e-01
-1.07553828e+00 -7.93827891e-01 -6.17123656e-02 1.05356824e+00
6.25021517e-01 -5.57914734e-01 -9.73429859e-01 -4.78815973e-01
-3.71644795e-01 4.13911998e-01 -1.73888355e-01 5.36905713e-02
-8.76828969e-01 -4.88990694e-01 3.82554442e-01 5.11137366e-01
1.13591075e+00 -9.28763509e-01 -9.20107841e-01 3.65325540e-01
-3.09993088e-01 -1.34748733e+00 -5.77701807e-01 -1.04801215e-01
-1.00573850e+00 -9.57725406e-01 -9.38694358e-01 -8.49991798e-01
4.35749829e-01 5.94168782e-01 6.85307324e-01 1.52603775e-01
-1.66754574e-01 -1.22915924e-01 -6.49537921e-01 3.06751966e-01
-4.92047295e-02 2.82073200e-01 -2.09098905e-01 -4.56620976e-02
3.92787047e-02 -4.86021191e-01 -8.33163261e-01 4.83736336e-01
-9.12107170e-01 5.15640140e-01 5.40386379e-01 5.98916650e-01
9.48359489e-01 -3.99869401e-03 2.69811362e-01 -4.93198067e-01
4.77447510e-02 -3.88634920e-01 -5.50584495e-01 -1.04927301e-01
-3.06098640e-01 -2.02324107e-01 7.54263043e-01 -1.86873227e-01
-9.69179153e-01 1.38411403e-01 -3.57395947e-01 -7.31470704e-01
-1.52539924e-01 8.20920840e-02 -3.89326923e-02 -9.89236832e-02
1.55346930e-01 3.77489746e-01 -7.64705092e-02 -4.25321817e-01
9.85713825e-02 6.86810911e-01 4.58395660e-01 -3.28645647e-01
5.16479850e-01 4.68610674e-01 -8.25623721e-02 -9.43710029e-01
-5.79159737e-01 -4.46791172e-01 -6.18463337e-01 -9.90341157e-02
1.20338953e+00 -1.20618129e+00 -3.64355326e-01 5.09648800e-01
-1.08288932e+00 -2.58377790e-01 -8.35754648e-02 6.90042675e-01
-6.86904550e-01 4.92480099e-01 -7.46849656e-01 -4.01290566e-01
-5.42216122e-01 -1.40198803e+00 1.06592333e+00 6.22065544e-01
-8.13201219e-02 -4.68419462e-01 -3.70684206e-01 3.96340430e-01
3.90211970e-01 1.89655602e-01 3.71536076e-01 -3.05697441e-01
-8.35226715e-01 3.87916118e-01 -5.96254110e-01 5.22516131e-01
-1.24928854e-01 9.29135457e-02 -6.41993880e-01 -2.94884861e-01
-1.17711619e-01 1.64496511e-01 1.01640630e+00 5.89968324e-01
1.30765784e+00 -7.10960552e-02 -2.39014432e-01 1.15068710e+00
1.56173134e+00 3.92239183e-01 8.05131733e-01 5.74955940e-01
7.73832262e-01 2.03219220e-01 1.08930802e+00 3.62682104e-01
4.32659775e-01 6.49713159e-01 3.59269142e-01 -2.34232232e-01
-2.61689276e-01 1.78321786e-02 4.53399658e-01 9.07364070e-01
-4.07700449e-01 -2.38531232e-01 -5.72512925e-01 4.42929596e-01
-2.01207113e+00 -1.00157750e+00 -1.32799938e-01 1.98957884e+00
5.59901416e-01 1.45812839e-01 1.85790695e-02 1.43293768e-01
9.91797566e-01 3.98186058e-01 -3.64481449e-01 -4.07192409e-01
-1.54483974e-01 2.46704862e-01 7.36346245e-01 5.10123849e-01
-1.22880137e+00 1.15477157e+00 4.76642656e+00 1.24863672e+00
-1.12964499e+00 2.68119186e-01 7.29207158e-01 -2.21572444e-01
3.23456854e-01 -1.04049839e-01 -7.96994805e-01 7.90937126e-01
9.40427184e-01 1.83114365e-01 2.71619976e-01 7.37992883e-01
4.05944526e-01 -3.32013279e-01 -5.62204361e-01 1.10232866e+00
-6.25213832e-02 -1.40746427e+00 5.95884249e-02 -1.13018930e-01
5.17442048e-01 2.18893677e-01 -1.94309145e-01 8.41088444e-02
-2.84704059e-01 -6.79570615e-01 8.08866680e-01 4.69290912e-01
9.21052158e-01 -1.04541695e+00 8.43672276e-01 7.56080374e-02
-1.65709531e+00 -3.27924676e-02 -5.28174758e-01 1.08863831e-01
3.69008064e-01 3.29972953e-01 -3.21684837e-01 7.18436837e-01
1.06397986e+00 8.15359712e-01 -3.43808711e-01 9.43548024e-01
8.43665153e-02 4.54944432e-01 -4.59185213e-01 4.96580601e-01
5.13501704e-01 -2.77259976e-01 4.66340244e-01 1.40970707e+00
6.85828209e-01 4.35687065e-01 2.34793469e-01 4.39722866e-01
1.65512964e-01 1.11888602e-01 -1.63029313e-01 3.69246483e-01
4.35281515e-01 1.26399112e+00 -1.01749575e+00 -5.66600800e-01
-5.49429953e-01 1.23552871e+00 -9.82698053e-04 2.00152591e-01
-1.21122634e+00 -4.53255832e-01 7.66767859e-01 2.85384536e-01
7.00638115e-01 -2.13570446e-01 2.75305212e-02 -1.07358003e+00
-7.25171342e-02 -6.15858436e-01 3.12187254e-01 -8.06663692e-01
-5.57984114e-01 7.82821536e-01 -9.66618359e-02 -1.31945109e+00
-1.12187266e-01 -1.09251857e-01 -2.62403846e-01 6.96463585e-01
-1.71081316e+00 -8.79507422e-01 -4.57769454e-01 8.98280799e-01
1.13229215e+00 1.91716794e-02 3.21230710e-01 5.49244583e-01
-7.28060603e-01 4.88092095e-01 -1.09368958e-01 3.25964689e-02
5.94105721e-01 -7.83932507e-01 4.38823789e-01 9.62466121e-01
-2.75765240e-01 1.22249171e-01 5.27880132e-01 -5.93266547e-01
-1.06899476e+00 -1.36654198e+00 6.38548017e-01 4.54126358e-01
2.37824291e-01 3.64395864e-02 -1.01597440e+00 3.77104193e-01
1.31727383e-01 3.05417717e-01 2.67245203e-01 -6.69533372e-01
-1.04428241e-02 -9.42696705e-02 -1.29038870e+00 4.81374562e-01
1.29196584e+00 -1.25303462e-01 -4.75895226e-01 -5.48681244e-02
1.08561099e+00 -5.64601243e-01 -8.75075698e-01 4.92121994e-01
4.39665407e-01 -1.23171067e+00 1.00644994e+00 8.72897357e-02
5.26139498e-01 -6.51321769e-01 -3.75797421e-01 -8.56107354e-01
-5.20043433e-01 -5.61668277e-01 -7.20503107e-02 1.08601534e+00
-3.12252883e-02 -4.82970953e-01 6.56275392e-01 2.30950922e-01
-3.85383755e-01 -8.74679089e-01 -1.10010600e+00 -5.65164745e-01
-4.20191586e-01 -3.79412323e-01 4.24882799e-01 8.26269507e-01
-4.43211168e-01 8.61855298e-02 -2.03240395e-01 3.74083698e-01
4.73751158e-01 -6.98689595e-02 5.11053920e-01 -8.80402923e-01
-4.22764495e-02 -3.17695171e-01 -6.80299401e-01 -1.29467046e+00
-1.75398201e-01 -6.25981331e-01 -1.56789392e-01 -1.46855986e+00
-8.85280445e-02 -2.75291741e-01 -1.67412668e-01 3.94723147e-01
2.33041737e-02 5.20923972e-01 5.49366713e-01 2.54882365e-01
-8.26474309e-01 3.86536121e-01 1.34878993e+00 3.05726618e-01
-3.66937220e-01 -3.01811725e-01 -3.33785444e-01 1.05319369e+00
9.01230633e-01 -3.34123671e-01 -3.25905025e-01 -5.01692295e-01
-2.59691149e-01 2.84282982e-01 3.45060259e-01 -1.41232026e+00
2.75273234e-01 -1.66251257e-01 5.23897886e-01 -8.35801482e-01
5.41953444e-01 -8.83600056e-01 1.67123288e-01 5.05675077e-01
1.17368244e-01 1.43998682e-01 1.39419556e-01 2.99931705e-01
-3.93374681e-01 -4.44495231e-02 1.10010529e+00 -1.09746292e-01
-1.20654249e+00 4.03876722e-01 -2.67170608e-01 -7.25664496e-02
1.25679994e+00 -6.91983223e-01 -6.72416240e-02 1.03129052e-01
-6.54637754e-01 2.68277705e-01 5.25724411e-01 4.13881391e-01
7.94278502e-01 -1.13130629e+00 -5.98254859e-01 4.46047843e-01
-4.91784096e-01 3.61514032e-01 6.31103218e-01 9.02853191e-01
-1.10534239e+00 3.12102169e-01 -3.26183468e-01 -6.99964404e-01
-1.50426376e+00 2.77202994e-01 2.11768553e-01 -1.29738599e-01
-8.77937138e-01 8.35082352e-01 1.82653710e-01 1.66143030e-01
2.00941078e-02 -3.84463847e-01 -2.53854454e-01 1.71716958e-01
6.27239108e-01 5.77427626e-01 8.33880901e-03 -1.08773482e+00
-3.55631202e-01 9.31222320e-01 -1.57699198e-01 3.44089419e-02
1.37286496e+00 -4.41221237e-01 1.44217998e-01 -3.53583917e-02
1.34972131e+00 -4.12290275e-01 -1.56037652e+00 -1.69621497e-01
-3.67249757e-01 -6.85114741e-01 3.29899073e-01 -2.96033710e-01
-1.55701995e+00 5.83954871e-01 7.74068475e-01 -5.06440774e-02
1.48866427e+00 -2.31420010e-01 1.24313426e+00 -7.38859326e-02
4.07516062e-01 -1.23792100e+00 -2.35103562e-01 6.22896850e-01
6.37642503e-01 -9.55281854e-01 2.20467485e-02 -6.69101894e-01
-6.27618253e-01 1.11759496e+00 7.19255686e-01 -4.68784511e-01
5.86547256e-01 2.12730125e-01 -2.08953410e-01 1.54663563e-01
-5.23921847e-01 -2.11559817e-01 9.51018780e-02 3.80883247e-01
1.78106442e-01 -2.70671487e-01 -5.92765152e-01 4.97365862e-01
-1.31171748e-01 9.35109630e-02 4.18483913e-01 8.58142853e-01
-6.70249283e-01 -6.83670342e-01 -4.53165054e-01 3.94081742e-01
-5.74386954e-01 -1.34185210e-01 3.21636736e-01 7.50299454e-01
6.30970955e-01 7.87629008e-01 4.89604056e-01 -5.85033119e-01
3.03220838e-01 -2.05348879e-01 1.96694285e-01 -2.66430765e-01
-5.93818784e-01 4.45706844e-01 -6.37749955e-02 -1.14273775e+00
-7.14027345e-01 -6.35935605e-01 -1.65263712e+00 -4.36577022e-01
-2.06910893e-01 1.35503802e-02 4.87756312e-01 7.86170602e-01
6.45684838e-01 6.81972861e-01 6.14874363e-01 -1.15699446e+00
9.68510509e-02 -7.13997960e-01 -4.70641047e-01 1.55049384e-01
2.15693802e-01 -5.41866541e-01 -3.59079778e-01 1.19884416e-01] | [9.194637298583984, -0.21430885791778564] |
50f6529f-f745-4e83-9408-03c9221240b0 | seq3-differentiable-sequence-to-sequence-to-1 | null | null | https://aclanthology.org/N19-1071 | https://aclanthology.org/N19-1071.pdf | SEQ\^3: Differentiable Sequence-to-Sequence-to-Sequence Autoencoder for Unsupervised Abstractive Sentence Compression | Neural sequence-to-sequence models are currently the dominant approach in several natural language processing tasks, but require large parallel corpora. We present a sequence-to-sequence-to-sequence autoencoder (SEQ{\^{}}3), consisting of two chained encoder-decoder pairs, with words used as a sequence of discrete latent variables. We apply the proposed model to unsupervised abstractive sentence compression, where the first and last sequences are the input and reconstructed sentences, respectively, while the middle sequence is the compressed sentence. Constraining the length of the latent word sequences forces the model to distill important information from the input. A pretrained language model, acting as a prior over the latent sequences, encourages the compressed sentences to be human-readable. Continuous relaxations enable us to sample from categorical distributions, allowing gradient-based optimization, unlike alternatives that rely on reinforcement learning. The proposed model does not require parallel text-summary pairs, achieving promising results in unsupervised sentence compression on benchmark datasets. | ['ros', 'Ioannis Konstas', 'Christos Baziotis', 'Ion Androutsopoulos', 'Alex Potamianos'] | 2019-06-01 | null | null | null | naacl-2019-6 | ['sentence-compression', 'unsupervised-abstractive-sentence-compression'] | ['natural-language-processing', 'natural-language-processing'] | [ 7.07201004e-01 4.00399119e-01 -2.02720493e-01 -4.54121739e-01
-6.93961322e-01 -2.24380523e-01 6.41021609e-01 4.49896157e-01
-9.28304017e-01 8.85230660e-01 6.24348104e-01 -1.79668888e-01
2.14432448e-01 -7.87032545e-01 -1.02875876e+00 -7.39081860e-01
1.06001318e-01 6.61072671e-01 -2.84169555e-01 -1.02692164e-01
4.13183302e-01 -6.10499866e-02 -1.46730983e+00 4.23669726e-01
9.70169187e-01 6.65882409e-01 9.60213482e-01 9.95734036e-01
-3.89055938e-01 8.72139633e-01 -5.50380111e-01 -5.10921657e-01
1.06316298e-01 -8.26610565e-01 -7.24946082e-01 1.83621436e-01
3.74939330e-02 -3.45731080e-01 -4.15715575e-01 1.16067564e+00
3.68309051e-01 3.06926608e-01 8.01873386e-01 -4.86841857e-01
-6.82271838e-01 9.30978179e-01 -2.23021790e-01 -6.29623011e-02
4.56119359e-01 1.45434305e-01 1.29028273e+00 -6.61440194e-01
7.09933400e-01 1.22716069e+00 1.24431804e-01 6.23650134e-01
-1.35970628e+00 1.16966898e-02 3.36076654e-02 2.64961451e-01
-1.09236276e+00 -5.87485433e-01 6.18962944e-01 -3.17902803e-01
1.42815256e+00 3.07652891e-01 7.25852907e-01 1.08215773e+00
4.38301414e-01 9.76550162e-01 5.06949306e-01 -8.26102197e-01
6.54287815e-01 1.57493189e-01 -8.17957819e-02 5.42878330e-01
2.36147214e-02 -1.92153454e-01 -4.88022000e-01 -1.11047797e-01
3.76790017e-01 1.25080779e-01 -2.20579863e-01 -2.29453132e-01
-7.84057558e-01 1.04601228e+00 -1.80500690e-02 1.83087096e-01
-7.39140928e-01 2.70368345e-02 5.93371928e-01 3.55445683e-01
4.55887884e-01 3.02863479e-01 -2.05763578e-01 -3.57717365e-01
-1.35153615e+00 4.13833588e-01 8.34038854e-01 9.25422609e-01
5.93098462e-01 2.02564374e-01 -3.08437049e-01 9.55998540e-01
2.02537715e-01 3.40450048e-01 8.83595705e-01 -8.72897804e-01
7.68022239e-01 1.36643201e-01 -1.24511182e-01 -7.68963754e-01
2.21781999e-01 -3.09374601e-01 -1.18473530e+00 -3.10105294e-01
-2.12523248e-02 -1.24208570e-01 -7.09337771e-01 1.72710383e+00
-2.44643882e-01 -8.96006599e-02 2.96783656e-01 5.80265522e-01
8.52487162e-02 1.03146982e+00 2.69308407e-02 -6.94475889e-01
1.10903227e+00 -1.02633953e+00 -8.87327611e-01 -4.51189935e-01
5.55465281e-01 -4.40998137e-01 1.16577613e+00 4.27546054e-01
-1.63570261e+00 -5.64128280e-01 -9.87707078e-01 -2.33527303e-01
-8.77014175e-02 1.09186225e-01 -6.12742454e-02 2.04544440e-01
-1.07961380e+00 8.95169735e-01 -8.97952497e-01 -5.90458997e-02
2.31259748e-01 2.29218423e-01 -2.09583744e-01 -9.51301828e-02
-1.12321222e+00 8.47939193e-01 9.29769397e-01 1.88280195e-02
-9.35409904e-01 -1.15673736e-01 -1.04729497e+00 5.15617788e-01
1.49629042e-01 -8.49985838e-01 1.18513238e+00 -9.88421619e-01
-1.72402406e+00 6.45734489e-01 -4.63238329e-01 -1.13776541e+00
3.99405032e-01 -3.93311411e-01 -5.51027358e-02 4.49735016e-01
-1.65632412e-01 8.09070289e-01 1.14956081e+00 -9.27590549e-01
-2.98327237e-01 4.08861563e-02 -3.87263387e-01 3.59257549e-01
-6.40833735e-01 -1.91411495e-01 -4.25686717e-01 -8.25487256e-01
-3.18328440e-01 -6.07588828e-01 -4.16154325e-01 -3.14045876e-01
-4.95971829e-01 -1.67878941e-01 2.24549264e-01 -1.03236222e+00
1.49840724e+00 -2.16030192e+00 8.23204398e-01 -1.72545165e-01
-2.06761435e-02 1.40786111e-01 -3.71043086e-01 7.48770058e-01
1.87364575e-02 -8.13873857e-02 -8.51721168e-01 -9.94423926e-01
3.72810313e-03 4.82182354e-01 -4.00429785e-01 2.86448300e-01
2.71179676e-01 7.67246723e-01 -8.55565667e-01 -5.46755910e-01
7.49164000e-02 2.60357201e-01 -8.89025211e-01 6.42432332e-01
-6.59334064e-01 1.78824678e-01 -5.63447736e-02 -1.92863926e-01
3.85188997e-01 -7.42369369e-02 2.72839278e-01 3.95659596e-01
1.17683925e-01 5.41305482e-01 -7.68602490e-01 1.98999405e+00
-5.23221195e-01 4.60100353e-01 -3.24626788e-02 -1.27647376e+00
1.02315247e+00 3.04755777e-01 1.82428494e-01 -6.54578209e-01
1.41844869e-01 1.47342887e-02 -9.81261730e-02 -8.30428958e-01
8.21331859e-01 -3.58520061e-01 -1.07542224e-01 5.46169996e-01
2.41875455e-01 -3.70361775e-01 7.54323363e-01 4.62070882e-01
8.39792967e-01 1.53293563e-02 2.46383056e-01 -9.40104388e-03
6.02736294e-01 -4.77326542e-01 3.40279996e-01 8.67866337e-01
2.33403578e-01 6.52536690e-01 7.85912931e-01 -2.73828566e-01
-1.59909892e+00 -9.24052298e-01 3.47664207e-01 8.60438049e-01
-3.66371483e-01 -5.49019098e-01 -1.09202313e+00 -3.84418994e-01
-2.80425221e-01 1.27133822e+00 -3.65984648e-01 -4.48716730e-01
-6.96871877e-01 -3.05457294e-01 4.67050880e-01 3.28742266e-01
-5.58242798e-02 -1.35127461e+00 -7.03115463e-01 4.53007579e-01
-4.19558465e-01 -9.60415363e-01 -6.53801918e-01 4.76275384e-01
-1.14948595e+00 -3.25409502e-01 -8.62716079e-01 -7.70434201e-01
7.02999353e-01 -3.91652226e-01 1.20726919e+00 -8.02624747e-02
-1.40620112e-01 -1.30196467e-01 -3.58821511e-01 -8.99572819e-02
-8.71036649e-01 2.95887917e-01 1.43921092e-01 6.62634671e-02
1.61398992e-01 -7.37216234e-01 -3.85303527e-01 -5.38073897e-01
-1.26842082e+00 1.85577139e-01 7.74827898e-01 1.08458018e+00
6.54135585e-01 -2.66024321e-01 2.86439687e-01 -8.82347047e-01
8.87103558e-01 -3.92057717e-01 -3.16347897e-01 1.07198402e-01
-4.22988027e-01 6.40268087e-01 1.17575967e+00 -3.18384916e-01
-8.21741045e-01 8.95627290e-02 -4.60916370e-01 -4.79615748e-01
-3.87252957e-01 6.81241393e-01 -8.76979530e-02 9.01071191e-01
4.92371976e-01 9.61570799e-01 2.99401104e-01 -4.90223795e-01
4.93619531e-01 7.47250438e-01 6.96709335e-01 -4.88649130e-01
2.79967993e-01 2.77155526e-02 -4.02814686e-01 -1.00190687e+00
-5.80813050e-01 -2.27728546e-01 -7.16015339e-01 2.60441005e-01
8.91644955e-01 -7.42482483e-01 -4.39863116e-01 6.85348585e-02
-1.47437680e+00 -2.68545538e-01 -6.41568899e-01 4.88229036e-01
-9.43520904e-01 6.64355218e-01 -9.33139265e-01 -8.68098557e-01
-5.55931032e-01 -1.06809390e+00 9.53227043e-01 -1.24029135e-02
-3.92081589e-01 -9.49282408e-01 1.27522796e-01 3.38826738e-02
2.08301380e-01 -2.42635548e-01 1.18423927e+00 -7.87459075e-01
-3.92146558e-01 -2.70257711e-01 3.16293269e-01 9.10578012e-01
-1.51649550e-01 -2.70718783e-01 -6.90850079e-01 -4.58267391e-01
3.08135241e-01 -5.52546859e-01 1.17685974e+00 4.75225538e-01
1.41555357e+00 -8.61436486e-01 1.13367895e-02 4.74585861e-01
1.26895308e+00 7.85039738e-02 8.34234893e-01 3.96768562e-02
3.13094020e-01 6.27674401e-01 2.01993659e-01 7.19890893e-01
9.62829068e-02 2.61932671e-01 3.55375201e-01 2.30985016e-01
2.03874990e-01 -5.76985717e-01 5.64817429e-01 1.53341842e+00
2.18756586e-01 -4.61439639e-01 -5.79555035e-01 5.23214042e-01
-1.60864508e+00 -1.15413165e+00 3.29064339e-01 2.17973804e+00
1.12471974e+00 4.63216335e-01 -1.36178553e-01 1.24541737e-01
5.43374896e-01 4.51534927e-01 -4.39078569e-01 -7.15137005e-01
-5.97824939e-02 3.09366614e-01 1.91040412e-01 8.41534495e-01
-7.42836177e-01 7.75727212e-01 6.00712013e+00 8.83313656e-01
-9.30474102e-01 -1.52945280e-01 6.39695406e-01 -3.53804201e-01
-6.46224141e-01 1.54800620e-02 -6.63348377e-01 6.45141006e-01
1.34900105e+00 -7.07918331e-02 5.77560186e-01 6.34551585e-01
4.63639200e-01 -1.46609113e-01 -1.19527435e+00 7.38185525e-01
2.32050374e-01 -1.33600080e+00 3.22231680e-01 -1.51265860e-01
5.74186802e-01 -2.74521738e-01 -2.64522452e-02 3.46977383e-01
4.37298790e-02 -1.01451802e+00 8.69810224e-01 7.78614700e-01
8.77444625e-01 -8.58284235e-01 5.31294346e-01 9.93193388e-01
-7.06820726e-01 -7.51785263e-02 -6.31897271e-01 -1.47743762e-01
4.67538297e-01 6.59299374e-01 -8.17769170e-01 3.90085369e-01
2.16481909e-01 8.13605905e-01 -2.56245613e-01 4.48573828e-01
-1.96045667e-01 6.50598168e-01 -9.89662483e-02 -3.78753811e-01
1.50342271e-01 -5.15736222e-01 6.33103251e-01 1.57740533e+00
1.05966002e-01 -2.07719505e-02 2.62139827e-01 8.18901122e-01
-2.31442302e-01 1.14863381e-01 -5.99893332e-01 -4.30012554e-01
2.48745486e-01 7.48380184e-01 -4.28583801e-01 -6.42026067e-01
-1.24823615e-01 1.42353964e+00 5.37420452e-01 3.99672657e-01
-4.40143466e-01 -5.63664198e-01 2.09448174e-01 1.83861202e-03
4.77644473e-01 -3.53891015e-01 -3.50868940e-01 -1.32008028e+00
1.65815622e-01 -1.00523198e+00 1.77729353e-01 -4.63686377e-01
-8.89562547e-01 7.58121014e-01 -1.10237263e-02 -1.00244737e+00
-8.11749279e-01 -8.52466151e-02 -4.40690845e-01 1.04862773e+00
-1.36069071e+00 -5.79387605e-01 4.12733436e-01 2.93674469e-01
1.15478587e+00 -4.15811688e-01 9.92248952e-01 8.29125345e-02
-2.93169856e-01 5.89388311e-01 4.85265225e-01 4.30593640e-02
2.99390435e-01 -1.15175748e+00 5.72715700e-01 8.89473617e-01
2.34272689e-01 7.28023171e-01 1.01488554e+00 -7.26351678e-01
-1.25988507e+00 -8.27647567e-01 1.21884084e+00 7.12431818e-02
3.64661843e-01 -7.65371025e-01 -9.39402878e-01 4.92140532e-01
7.23555267e-01 -4.72760469e-01 5.74426651e-01 -1.76483989e-01
3.68145970e-03 1.02993533e-01 -7.10260093e-01 5.70167482e-01
6.00712895e-01 -6.94893241e-01 -7.22726047e-01 2.98544079e-01
1.04993856e+00 -2.88359642e-01 -4.52645868e-01 -1.35529175e-01
2.07210854e-01 -8.57039094e-01 7.20103621e-01 -6.78022563e-01
1.16723204e+00 1.21200942e-01 -2.06531137e-01 -1.39004922e+00
-2.07012758e-01 -4.38741297e-01 -5.43644547e-01 9.70001578e-01
4.19769406e-01 -8.00940245e-02 8.42555046e-01 2.64779896e-01
-2.14580148e-01 -8.72590244e-01 -7.71365285e-01 -6.88344359e-01
1.63361818e-01 -3.34727228e-01 3.37335497e-01 3.98564845e-01
7.33253136e-02 4.85329449e-01 -6.48868620e-01 -2.45684907e-01
5.41204810e-01 -8.95365775e-02 3.67668182e-01 -7.74979293e-01
-9.26580727e-01 -4.78889972e-01 1.95956349e-01 -1.58560848e+00
3.97785842e-01 -1.02955103e+00 3.87509316e-01 -1.34941125e+00
3.58349264e-01 1.64941698e-01 -3.13565391e-03 1.77028731e-01
-2.44918600e-01 -4.02598828e-01 3.67628753e-01 1.15325205e-01
-5.28061926e-01 1.26377308e+00 1.07592154e+00 -2.69113451e-01
-1.66654080e-01 -1.83971189e-02 -3.33234608e-01 5.29320478e-01
6.43243313e-01 -5.81451297e-01 -3.73601735e-01 -6.69649601e-01
1.07296661e-01 6.22335672e-01 -7.26380572e-02 -6.99677944e-01
3.90838623e-01 -1.70022279e-01 1.95394009e-01 -6.10706866e-01
4.54539299e-01 -6.53564572e-01 -2.21794456e-01 7.72502720e-01
-1.05967331e+00 1.58398062e-01 -1.43221825e-01 6.46608353e-01
-4.47123945e-01 -7.91588068e-01 8.36986244e-01 -4.12597656e-01
-7.11376071e-02 1.79262638e-01 -5.34915924e-01 1.50755560e-02
6.27144992e-01 3.51495035e-02 3.51907790e-01 -6.80193186e-01
-7.20360279e-01 8.87951180e-02 2.14299604e-01 2.73797303e-01
8.06072116e-01 -1.05315375e+00 -9.64429080e-01 3.52164179e-01
-3.15961748e-01 2.82472759e-01 2.55883753e-01 2.54723549e-01
-5.08333683e-01 5.80705583e-01 -1.32762626e-01 -5.67637026e-01
-1.02641940e+00 6.94405556e-01 -5.84512725e-02 -6.37000084e-01
-5.77516139e-01 8.23278666e-01 -2.14877248e-01 -2.96500087e-01
5.01993537e-01 -1.94112197e-01 -1.36040270e-01 1.71033051e-02
6.33666396e-01 1.70153216e-01 -1.07011490e-01 -2.23235831e-01
1.86830178e-01 4.01422344e-02 -5.90935171e-01 -5.79807401e-01
1.64632642e+00 -2.64223933e-01 -3.21443051e-01 6.06263518e-01
1.46841085e+00 -2.81731606e-01 -1.32579923e+00 -3.04749161e-01
1.03257671e-01 -2.56423414e-01 -2.56881267e-01 -2.54239738e-01
-5.41504562e-01 1.28545058e+00 1.27869159e-01 2.45784849e-01
1.13626087e+00 -2.52001941e-01 1.00953364e+00 5.34033000e-01
3.45369689e-02 -1.27049112e+00 2.84776181e-01 9.19064641e-01
8.37355733e-01 -7.59634018e-01 -1.70954004e-01 1.65400788e-01
-8.43819380e-01 1.13491631e+00 1.91531479e-01 -2.41308227e-01
1.21789366e-01 2.75284737e-01 -4.45430487e-01 3.33693206e-01
-1.11218739e+00 1.14333421e-01 -4.22807559e-02 3.40180159e-01
4.75322187e-01 -1.34991510e-02 -5.39873362e-01 4.64112669e-01
-3.83366436e-01 -3.88343222e-02 5.68575382e-01 8.14936042e-01
-6.78881228e-01 -1.40716970e+00 -2.53262490e-01 5.11212707e-01
-3.02443951e-01 -4.45245892e-01 5.64908274e-02 -4.12624627e-02
-1.59253463e-01 6.35072589e-01 3.63825291e-01 -8.12746063e-02
1.21234968e-01 3.99609149e-01 2.43633181e-01 -6.47871792e-01
-3.99066687e-01 2.92230994e-01 -2.10650489e-01 -1.48779422e-01
8.29637125e-02 -8.71240914e-01 -1.14019418e+00 -2.02439234e-01
-1.44887343e-01 4.10948813e-01 7.22148895e-01 1.05379963e+00
3.43133867e-01 5.89313328e-01 7.39433467e-01 -8.19861412e-01
-1.08249342e+00 -1.21287549e+00 -4.55358058e-01 5.31205475e-01
6.19438112e-01 1.90246284e-01 -9.79309529e-02 5.81301749e-01] | [12.151823997497559, 9.282356262207031] |
f8c7e8d4-da84-43c3-8e00-b10fe9292784 | few-shot-class-incremental-pill-recognition | 2304.11959 | null | https://arxiv.org/abs/2304.11959v1 | https://arxiv.org/pdf/2304.11959v1.pdf | Few-shot Class-incremental Pill Recognition | The automatic pill recognition system is of great significance in improving the efficiency of the hospital, helping people with visual impairment, and avoiding cross-infection. However, most existing pill recognition systems based on deep learning can merely perform pill classification on the learned pill categories with sufficient training data. In practice, the expensive cost of data annotation and the continuously increasing categories of new pills make it meaningful to develop a few-shot class-incremental pill recognition system. In this paper, we develop the first few-shot class-incremental pill recognition system, which adopts decoupled learning strategy of representations and classifiers. In learning representations, we propose the novel Center-Triplet loss function, which can promote intra-class compactness and inter-class separability. In learning classifiers, we propose a specialized pseudo pill image construction strategy to train the Graph Attention Network to obtain the adaptation model. Moreover, we construct two new pill image datasets for few-shot class-incremental learning. The experimental results show that our framework outperforms the state-of-the-art methods. | ['Dewen Hu', 'Kai Gao', 'Li Liu', 'Jinghua Zhang'] | 2023-04-24 | null | null | null | null | ['class-incremental-learning', 'few-shot-class-incremental-learning', 'incremental-learning'] | ['computer-vision', 'methodology', 'methodology'] | [ 4.03098583e-01 -8.81449506e-02 -6.06073380e-01 -5.23532271e-01
-6.48562193e-01 -8.15125629e-02 2.94288337e-01 2.25534141e-01
-1.76222133e-03 5.12643874e-01 -8.65187645e-02 -4.99405600e-02
-1.58644736e-01 -1.07415152e+00 -5.04059792e-01 -8.47264409e-01
3.81571978e-01 3.52251053e-01 -4.67209779e-02 7.56773278e-02
2.22580612e-01 1.77441493e-01 -1.26478124e+00 3.53601664e-01
1.14317727e+00 1.01518309e+00 3.24094176e-01 2.96904385e-01
-4.63719219e-01 7.35693097e-01 -5.05704105e-01 -4.40048009e-01
-5.68702221e-02 -7.46970415e-01 -2.14938059e-01 3.59976619e-01
2.99484193e-01 -4.45259601e-01 -3.61401767e-01 1.02833796e+00
8.27337146e-01 -6.77373856e-02 8.95160139e-01 -7.38167405e-01
-1.37296367e+00 4.19109285e-01 -6.06523514e-01 6.58641160e-02
1.80849478e-01 -1.53925762e-01 5.34790993e-01 -1.08805847e+00
4.00817573e-01 9.78036463e-01 7.26709902e-01 7.31551290e-01
-7.02981651e-01 -6.81673050e-01 2.12492213e-01 2.20989883e-01
-1.40265977e+00 -4.41466749e-01 8.96403968e-01 -3.72891784e-01
4.33530062e-01 2.31082588e-01 8.95249248e-01 9.34118807e-01
1.32113084e-01 1.20992076e+00 4.77718234e-01 -3.26803952e-01
3.16464394e-01 1.86811298e-01 3.36344779e-01 8.54189515e-01
5.37148535e-01 -6.07453763e-01 4.22079377e-02 1.30624086e-01
8.84909272e-01 9.95892167e-01 -2.06104010e-01 -5.68504214e-01
-5.94001055e-01 9.74152505e-01 6.30486369e-01 7.95049608e-01
-6.72097653e-02 -5.05090281e-02 3.81426752e-01 -1.46598741e-01
5.33218145e-01 7.98200667e-02 -2.02779740e-01 1.98581323e-01
-7.53787398e-01 -2.29545727e-01 4.68174130e-01 1.10845900e+00
4.89841968e-01 4.47809733e-02 -5.63144803e-01 9.72275794e-01
1.70207128e-01 5.12597442e-01 7.13425517e-01 -2.19602719e-01
2.73214459e-01 9.88175929e-01 -2.26405770e-01 -1.03322291e+00
-5.27917922e-01 -6.21520698e-01 -1.12383914e+00 -5.03010690e-01
-1.36766955e-01 6.32148609e-02 -1.30129099e+00 1.03569746e+00
1.43707931e-01 2.48455748e-01 1.12016238e-01 2.83237249e-01
1.25971842e+00 5.83665550e-01 -2.49673091e-02 -7.99018443e-01
1.21152079e+00 -1.15014577e+00 -1.13081551e+00 5.44760749e-02
8.80403399e-01 -5.17762601e-01 1.15708375e+00 1.61621243e-01
-8.41009736e-01 -6.96125746e-01 -1.17211545e+00 -5.57853729e-02
-5.93377471e-01 3.57618183e-01 9.98765528e-01 9.47760880e-01
-2.94363618e-01 5.26362479e-01 -7.50851631e-01 -3.19029152e-01
1.24044943e+00 4.73122567e-01 2.49700639e-02 -6.92381024e-01
-7.43987083e-01 5.45443892e-01 3.27682376e-01 3.03621083e-01
-6.48851275e-01 -5.90795517e-01 -1.06197476e+00 1.98673472e-01
3.65121335e-01 -3.48952919e-01 9.33601081e-01 -9.24852371e-01
-1.43088019e+00 7.31509805e-01 9.12786797e-02 -5.74601144e-02
2.15742454e-01 5.55865318e-02 -3.66070151e-01 1.86289325e-02
1.01555809e-01 2.96312273e-01 5.82113385e-01 -1.10601187e+00
-3.31772059e-01 -6.40020490e-01 -6.52725846e-02 2.53956765e-01
-7.07742929e-01 -3.62477571e-01 -7.22028613e-01 -5.98210156e-01
-8.33070949e-02 -6.34320617e-01 1.85892335e-03 -5.92946857e-02
-2.66154438e-01 -5.01320302e-01 7.69762337e-01 -4.64372575e-01
1.25251412e+00 -2.18887210e+00 1.93397012e-02 -1.21480696e-01
3.18154469e-02 4.83033538e-01 -1.25074521e-01 -1.11656308e-01
1.90114796e-01 1.92259485e-03 -4.32299435e-01 -4.77771789e-01
-4.15277183e-01 1.93861112e-01 7.03193769e-02 4.40880835e-01
7.36229122e-02 1.15325081e+00 -9.86675680e-01 -6.38086021e-01
3.80207688e-01 6.10825121e-01 -4.41303998e-01 1.63132325e-01
-1.86637729e-01 1.84747621e-01 -7.55810678e-01 1.19659376e+00
7.66232491e-01 -4.74612832e-01 1.00426346e-01 -3.06718975e-01
3.61154020e-01 -4.35788810e-01 -6.87018931e-01 2.07957625e+00
-1.71243802e-01 1.01959985e-02 -6.80402160e-01 -1.26005816e+00
9.99583185e-01 1.84242249e-01 4.65397626e-01 -7.61775315e-01
4.25606281e-01 3.66615027e-01 -1.02705255e-01 -7.62458086e-01
-3.20392102e-02 5.10905907e-02 1.30246237e-01 -1.82481200e-01
1.53762519e-01 3.16981860e-02 2.15039879e-01 -1.39928669e-01
6.04951799e-01 1.40238941e-01 4.23770696e-01 -6.05900474e-02
4.46852684e-01 -3.89571935e-01 4.73616481e-01 6.60192609e-01
-1.12716213e-01 6.64433420e-01 1.83208779e-01 -5.53752005e-01
-5.35187662e-01 -5.94607890e-01 -3.92488539e-01 8.18684876e-01
3.15048724e-01 -2.19034672e-01 -7.21703827e-01 -9.77847636e-01
-1.64567441e-01 4.49382335e-01 -8.23811829e-01 -5.27385771e-01
-4.98980314e-01 -1.14841592e+00 6.90248236e-02 6.99130595e-01
6.20398760e-01 -1.01122630e+00 -1.29714176e-01 1.76415130e-01
2.30742306e-01 -5.77958524e-01 -6.77303255e-01 -1.34648964e-01
-8.50856483e-01 -1.36013710e+00 -1.28442264e+00 -1.25999904e+00
8.59359741e-01 4.83253747e-01 5.99550545e-01 3.66060853e-01
-5.89531124e-01 1.92508951e-01 -3.49143088e-01 -4.48414057e-01
2.96395868e-01 1.89238042e-01 -1.65231049e-01 1.57753915e-01
6.03261054e-01 -3.76223713e-01 -8.71482015e-01 -8.15146044e-02
-7.02145934e-01 1.22715466e-01 6.18996680e-01 8.21319342e-01
8.46185803e-01 -1.69467777e-01 9.13693070e-01 -1.18380237e+00
5.00335336e-01 -4.48701710e-01 -2.55923212e-01 6.78258896e-01
-6.30529284e-01 -4.41048473e-01 6.40272141e-01 -6.20133758e-01
-9.14939940e-01 8.31418708e-02 -6.42176792e-02 -5.11806428e-01
3.05072725e-01 5.57703078e-01 -4.19675022e-01 -1.85707241e-01
3.51771772e-01 4.38329399e-01 -3.96641254e-01 -6.21536493e-01
4.02695686e-01 9.45677400e-01 2.71467060e-01 1.23538580e-02
1.06801502e-01 2.39710674e-01 -1.11847788e-01 -8.21761191e-01
-1.12409258e+00 -3.75424117e-01 -4.02923644e-01 8.84310305e-02
1.06576109e+00 -8.08698237e-01 -6.59081995e-01 5.42194903e-01
-9.33733404e-01 -7.16392472e-02 -2.27530569e-01 3.47276390e-01
-3.64148110e-01 2.84898788e-01 -4.47069973e-01 -8.67207468e-01
-6.94472730e-01 -1.01663935e+00 1.12760651e+00 7.11251140e-01
4.09698039e-01 -7.89610744e-01 6.31068423e-02 2.41202250e-01
3.08783859e-01 1.63147986e-01 9.93718445e-01 -8.46851468e-01
-4.57297176e-01 -5.43494999e-01 -4.02872503e-01 1.38428286e-01
7.57005274e-01 -3.07057977e-01 -8.56969178e-01 -3.72964412e-01
2.04795465e-01 -4.67085540e-01 1.10376716e+00 6.40288889e-01
1.71846139e+00 -1.23933487e-01 -5.69802642e-01 7.22844183e-01
1.35812390e+00 7.68320382e-01 8.78338575e-01 -3.21493357e-01
1.03124070e+00 9.53679830e-02 4.91307557e-01 1.76050484e-01
4.10695672e-01 5.40352881e-01 3.04372087e-02 -1.81368530e-01
-2.06753254e-01 -1.32883757e-01 -1.77983031e-01 1.02222109e+00
-1.21624038e-01 -7.51868114e-02 -5.84852636e-01 4.21826184e-01
-2.04564881e+00 -7.04869509e-01 3.31521213e-01 2.19527292e+00
8.94007564e-01 6.70415089e-02 -1.08160146e-01 -2.09847596e-02
5.26320875e-01 -1.56055346e-01 -6.29521012e-01 1.50282472e-01
1.54539838e-01 2.39193410e-01 1.14599675e-01 2.62787580e-01
-1.41916454e+00 8.43173683e-01 5.66495514e+00 1.35873771e+00
-8.99793923e-01 1.35987535e-01 8.89677346e-01 -3.83608155e-02
-9.26918983e-02 -5.15449405e-01 -9.19292808e-01 7.12999403e-01
4.15882379e-01 1.12605371e-01 7.94482380e-02 1.02060962e+00
-2.41245806e-01 2.43595675e-01 -1.27758658e+00 1.36961877e+00
7.66161919e-01 -1.60991848e+00 2.79208034e-01 -1.52127057e-01
7.83725142e-01 -4.51442897e-01 -1.77513018e-01 5.55183113e-01
-3.28551769e-01 -1.06555378e+00 -6.99868053e-02 8.87852430e-01
9.82051551e-01 -7.82520413e-01 7.73129344e-01 1.03905804e-01
-1.30527556e+00 -5.57556003e-02 -6.11789703e-01 2.85237521e-01
-3.19137871e-01 3.13482225e-01 -2.26417318e-01 5.75073183e-01
2.67432541e-01 1.39048016e+00 -7.05134511e-01 1.34463859e+00
1.21348664e-01 1.40060574e-01 4.82130237e-02 -3.19997847e-01
-4.33768593e-02 -3.78122747e-01 -2.08199710e-01 1.02507293e+00
4.42930102e-01 4.08688992e-01 5.33703089e-01 4.44397122e-01
-4.08090711e-01 4.18527752e-01 -7.15838969e-01 -3.41049671e-01
2.67016858e-01 1.05566514e+00 -9.51369643e-01 -6.00199401e-01
-6.03343844e-01 1.15467286e+00 1.15348667e-01 1.67414010e-01
-5.65572143e-01 -6.99820161e-01 -1.02727257e-01 -1.62296087e-01
4.08479840e-01 3.51978570e-01 -4.08792436e-01 -1.32132661e+00
-3.05046253e-02 -6.05741024e-01 3.76210898e-01 -5.52237630e-01
-1.22270095e+00 2.24516973e-01 -3.13299388e-01 -1.18393302e+00
2.27727413e-01 -4.72900093e-01 -6.61809504e-01 2.71630228e-01
-1.55662990e+00 -1.57773650e+00 -3.78294408e-01 6.40729189e-01
9.61856008e-01 -4.42609340e-01 1.23874080e+00 7.03286111e-01
-8.55248988e-01 7.93254495e-01 2.48607695e-01 2.08697453e-01
4.04800773e-01 -8.55780959e-01 -2.03656763e-01 2.34359890e-01
-1.99858919e-02 6.15522802e-01 6.99135959e-02 -7.76191711e-01
-1.54399419e+00 -1.19023347e+00 5.26871443e-01 -1.93601385e-01
1.72964782e-01 -3.61137986e-01 -9.56059217e-01 5.51491797e-01
-9.20698121e-02 1.84816986e-01 1.06324351e+00 -1.54078919e-02
-4.67067771e-02 -4.10907090e-01 -1.16090798e+00 5.01735330e-01
1.09890258e+00 -2.49573082e-01 -3.61236721e-01 9.75721955e-01
9.41897094e-01 -2.33762190e-01 -7.28695929e-01 6.16804838e-01
7.48553157e-01 -3.75255138e-01 9.56692100e-01 -4.25101399e-01
3.61065060e-01 5.27334958e-02 -7.47622699e-02 -7.85492539e-01
-2.90978223e-01 -1.23922341e-01 -5.76867163e-01 1.06085694e+00
1.66659534e-01 -3.66446942e-01 8.89413834e-01 3.09821784e-01
-2.43699998e-01 -1.26519477e+00 -7.81792998e-01 -5.83351016e-01
-2.31685378e-02 1.56838387e-01 4.48660791e-01 1.06850004e+00
1.27234161e-01 4.73692745e-01 -7.06495404e-01 -4.12946343e-01
5.04279017e-01 3.71267736e-01 4.86791581e-01 -1.21625745e+00
-4.41473663e-01 -3.08766633e-01 -4.98965889e-01 -8.82464945e-01
-3.61164749e-01 -8.78753722e-01 -1.67366281e-01 -2.01599550e+00
8.19785178e-01 -4.59560841e-01 -6.47281647e-01 7.13428080e-01
-3.11072022e-01 3.37056339e-01 9.36180055e-02 3.00613493e-01
-8.53693008e-01 7.80366898e-01 1.47776425e+00 -6.63495421e-01
-5.18223166e-01 1.67788655e-01 -9.00569797e-01 7.83499837e-01
6.36685014e-01 -4.15123433e-01 -6.77365065e-01 -4.16192025e-01
-1.05007563e-03 -1.71752200e-01 -2.88760185e-01 -8.48022878e-01
-4.40314878e-03 -1.99958667e-01 7.43731737e-01 -6.89544082e-01
4.44992423e-01 -7.59784579e-01 -2.40472868e-01 7.51312196e-01
-7.47071803e-02 -8.42148125e-01 -3.78512479e-02 7.81646132e-01
2.80466974e-01 -4.43254799e-01 6.41520381e-01 -3.07847470e-01
-3.57980579e-01 6.43509805e-01 1.03797048e-01 -2.96948552e-01
1.15134442e+00 -2.46465519e-01 -3.18967223e-01 1.09764665e-01
-6.43555105e-01 3.38788092e-01 1.36234760e-01 6.20775819e-01
8.60549033e-01 -1.66057527e+00 -3.87393415e-01 3.00131887e-01
3.64040434e-01 6.29318580e-02 2.61517048e-01 7.69834220e-01
-3.60738009e-01 2.98333377e-01 1.11563221e-01 -4.35495377e-01
-1.22360325e+00 8.59549582e-01 4.06643808e-01 -1.59132659e-01
-8.20748866e-01 8.72263670e-01 4.11813378e-01 -2.21457556e-01
5.55266738e-01 -2.59953856e-01 -6.79608881e-01 1.46675289e-01
8.99616182e-01 9.11858454e-02 -1.03186648e-02 -9.88350734e-02
-3.16989362e-01 8.60445321e-01 -3.99232805e-01 8.22840273e-01
1.41444528e+00 3.57378781e-01 -2.38227714e-02 6.43595815e-01
1.21089792e+00 -3.48538995e-01 -9.22060907e-01 -1.09737024e-01
-3.43276352e-01 -5.73768675e-01 -8.40877369e-02 -6.24628961e-01
-1.10821593e+00 9.08094406e-01 1.24927688e+00 8.10061395e-02
9.79929984e-01 -1.15700200e-01 7.91361272e-01 7.62803316e-01
2.71442920e-01 -1.01588476e+00 3.84943962e-01 2.23559840e-03
7.98357844e-01 -1.28373909e+00 1.72235593e-01 -5.23959696e-01
-3.17666501e-01 9.84838486e-01 5.02686501e-01 -1.85670450e-01
7.77789772e-01 -1.91632450e-01 -3.19552004e-01 -3.65550309e-01
3.77937108e-02 -5.96501045e-02 3.18073779e-01 6.09887421e-01
4.70351398e-01 1.42349666e-02 -6.28293335e-01 1.03868496e+00
4.75333154e-01 6.00500464e-01 2.75309861e-01 1.03457117e+00
-6.90063179e-01 -1.05449677e+00 2.64835984e-01 8.96710157e-01
-2.39473492e-01 -2.50744820e-01 -1.20451756e-01 4.86374557e-01
4.66821969e-01 7.23240554e-01 1.15686305e-01 -1.33133471e-01
3.96099389e-01 3.13114896e-02 9.04572368e-01 -9.93800104e-01
-3.80379111e-02 3.62725556e-01 -2.20393106e-01 -2.26462148e-02
-5.78479350e-01 -3.09123367e-01 -1.28279960e+00 2.37046897e-01
-7.39266455e-01 -7.43033811e-02 5.03667891e-01 1.12737834e+00
2.58313358e-01 7.62340426e-01 7.16722727e-01 -8.25804651e-01
-2.82546103e-01 -1.03264976e+00 -5.47433436e-01 2.23982751e-01
8.24327543e-02 -6.49815440e-01 2.35963076e-01 -2.38237418e-02] | [15.031317710876465, -2.5583224296569824] |
a12a06de-9f9a-4a98-a77e-03243a092f67 | explaining-legal-concepts-with-augmented | 2306.09525 | null | https://arxiv.org/abs/2306.09525v2 | https://arxiv.org/pdf/2306.09525v2.pdf | Explaining Legal Concepts with Augmented Large Language Models (GPT-4) | Interpreting the meaning of legal open-textured terms is a key task of legal professionals. An important source for this interpretation is how the term was applied in previous court cases. In this paper, we evaluate the performance of GPT-4 in generating factually accurate, clear and relevant explanations of terms in legislation. We compare the performance of a baseline setup, where GPT-4 is directly asked to explain a legal term, to an augmented approach, where a legal information retrieval module is used to provide relevant context to the model, in the form of sentences from case law. We found that the direct application of GPT-4 yields explanations that appear to be of very high quality on their surface. However, detailed analysis uncovered limitations in terms of the factual accuracy of the explanations. Further, we found that the augmentation leads to improved quality, and appears to eliminate the issue of hallucination, where models invent incorrect statements. These findings open the door to the building of systems that can autonomously retrieve relevant sentences from case law and condense them into a useful explanation for legal scholars, educators or practicing lawyers alike. | ['Huihui Xu', 'Hannes Westermann', 'Morgan A. Gray', 'Kevin D. Ashley', 'Jaromir Savelka'] | 2023-06-15 | null | null | null | null | ['information-retrieval'] | ['natural-language-processing'] | [ 4.57728535e-01 8.84617329e-01 -5.56923449e-03 -4.37184036e-01
-1.15340328e+00 -6.99789584e-01 7.69442499e-01 3.68134290e-01
-4.36665453e-02 7.66586781e-01 9.01154935e-01 -9.75642681e-01
-5.34636974e-01 -4.89938200e-01 -5.26321530e-01 -1.81102663e-01
4.80710953e-01 5.01932502e-01 1.07831098e-02 -6.04418337e-01
7.75818408e-01 3.00223976e-01 -1.44889343e+00 7.90429592e-01
1.21310282e+00 2.67094493e-01 -3.09483819e-02 4.19515878e-01
-4.60963786e-01 1.14046633e+00 -8.88376534e-01 -9.78937387e-01
7.29114283e-03 -4.45782959e-01 -1.08139765e+00 -3.22874069e-01
8.81553292e-01 -4.99239594e-01 -1.68038085e-01 8.57429028e-01
2.54219085e-01 -2.61690915e-01 6.44637942e-01 -6.19182110e-01
-1.16866565e+00 6.35522902e-01 -7.09866583e-02 3.94165486e-01
8.66058052e-01 2.07327539e-03 1.00813878e+00 -6.39278769e-01
1.14251196e+00 1.59088075e+00 3.52466613e-01 7.86032915e-01
-1.22448897e+00 -3.53588432e-01 3.23212981e-01 4.57971662e-01
-8.78397286e-01 -7.17466354e-01 3.03335339e-01 -6.39890671e-01
1.23701429e+00 4.81288970e-01 5.63109457e-01 1.21408594e+00
6.65719569e-01 3.22281301e-01 1.04452133e+00 -6.25012457e-01
3.00693929e-01 3.14999282e-01 5.12571037e-01 1.84042603e-01
7.22304702e-01 4.13193442e-02 -4.38080668e-01 -4.57735062e-01
4.21466857e-01 -3.55409920e-01 -2.45721146e-01 1.53454885e-01
-6.04433060e-01 8.63081753e-01 3.20437729e-01 6.43745780e-01
-5.50909340e-01 -3.75321582e-02 1.53431967e-01 2.97622412e-01
3.78297985e-01 8.17646027e-01 -1.32267103e-01 -3.29143465e-01
-6.72870040e-01 7.79169142e-01 8.20499837e-01 5.91237366e-01
2.16682196e-01 -3.59654337e-01 -5.32920957e-01 6.27103925e-01
4.15815175e-01 4.48061913e-01 2.53593028e-01 -1.28571260e+00
7.31098473e-01 8.15779090e-01 2.18251243e-01 -1.15967739e+00
1.05218142e-01 -4.68550324e-01 1.24655582e-01 3.33413243e-01
3.61908793e-01 4.68268096e-02 -8.42305481e-01 1.63137496e+00
1.08538635e-01 -4.91347283e-01 3.83712500e-01 8.82735550e-01
7.84416437e-01 3.25643539e-01 2.62399733e-01 1.16409138e-02
1.40608382e+00 -3.37433666e-01 -9.99836206e-01 -8.06435466e-01
6.55520916e-01 -8.42107177e-01 1.05927086e+00 1.94756255e-01
-1.13803589e+00 -1.24301061e-01 -8.97867858e-01 -5.55030644e-01
-2.78681099e-01 -3.22753906e-01 6.07790112e-01 4.92973685e-01
-9.34539855e-01 5.92476606e-01 -1.67262003e-01 -5.39727092e-01
7.25426435e-01 -1.78785861e-01 -3.00641507e-01 -3.54423881e-01
-1.13792217e+00 1.50956249e+00 1.22859918e-01 7.95746371e-02
-1.97102606e-01 -4.67168391e-01 -9.45212662e-01 2.29139581e-01
4.64685261e-01 -8.57380986e-01 1.38399291e+00 -6.30152643e-01
-7.30662882e-01 8.27448785e-01 -6.10799730e-01 -4.38362628e-01
4.02159780e-01 -3.03081691e-01 -4.31436539e-01 2.91105390e-01
8.14094365e-01 4.22681570e-01 3.39141130e-01 -1.27443612e+00
-3.06799978e-01 -4.90786165e-01 5.54555297e-01 -6.13865703e-02
8.58907327e-02 4.61585224e-02 -9.66570675e-02 -5.18625379e-01
3.07788253e-01 -6.43572092e-01 -2.46129513e-01 -1.33375064e-01
-3.86637568e-01 -1.40790224e-01 4.82967496e-01 -9.94104922e-01
1.35275364e+00 -2.02988482e+00 -1.17894262e-01 1.68275610e-01
2.87830591e-01 2.41201639e-01 3.28668281e-02 8.80089283e-01
-2.55149692e-01 6.58928931e-01 -1.78297669e-01 5.83611950e-02
1.66251570e-01 5.47571003e-01 -8.83770227e-01 -1.60591945e-01
2.62027621e-01 1.01156497e+00 -9.00110304e-01 -6.01850629e-01
2.66586505e-02 4.98404413e-01 -5.73325753e-01 -3.31070840e-01
-2.48464704e-01 2.69351095e-01 -6.27022386e-01 4.05179471e-01
3.05984735e-01 -1.81479454e-01 1.16973050e-01 2.73642600e-01
-7.75406417e-03 9.24061596e-01 -5.45062840e-01 1.65343785e+00
-2.98068762e-01 8.65480006e-01 -2.05405548e-01 -3.52355540e-01
6.72535121e-01 5.48092246e-01 -3.10403764e-01 -1.02442217e+00
4.32283571e-03 4.39728260e-01 2.51661777e-01 -1.11319911e+00
5.94692945e-01 -4.96664077e-01 6.79976195e-02 5.51005483e-01
-5.58122933e-01 -3.27768803e-01 3.63080710e-01 6.60459399e-01
1.18963003e+00 -3.33792120e-02 3.13018233e-01 -5.26371337e-02
2.26173714e-01 5.31765699e-01 2.39527702e-01 1.04538131e+00
2.45406870e-02 5.33680499e-01 7.43220747e-01 -5.48140049e-01
-7.19940722e-01 -6.77983224e-01 -2.92225689e-01 4.68192697e-01
-1.02426820e-01 -5.14673293e-01 -8.96349669e-01 -6.08871520e-01
-4.74486835e-02 1.72060061e+00 -7.11398780e-01 -2.23538801e-01
-4.19244409e-01 -4.37652394e-02 2.12271810e-01 3.48881751e-01
1.78592771e-01 -1.11960447e+00 -1.07828724e+00 2.92009830e-01
-5.35759330e-01 -1.03278828e+00 -1.49006516e-01 -3.25427979e-01
-9.18137074e-01 -1.11537218e+00 -2.13678807e-01 -2.45625541e-01
6.22332931e-01 1.32359013e-01 9.77555811e-01 6.44263208e-01
-1.76889822e-02 3.11573207e-01 -4.10133958e-01 -7.85853446e-01
-9.12233114e-01 -8.64958018e-02 -5.50471008e-01 -4.52338636e-01
6.26842082e-01 -5.24171889e-01 -3.32129657e-01 -3.13827783e-01
-1.31182110e+00 1.29601568e-01 6.34688318e-01 4.22784895e-01
-5.94981201e-02 -5.25230825e-01 6.95795536e-01 -1.17338443e+00
1.35447335e+00 -2.92008847e-01 -1.76136285e-01 3.88785541e-01
-7.48013377e-01 4.55071330e-01 2.31980771e-01 7.16922581e-02
-1.25152862e+00 -6.72306836e-01 -3.04215163e-01 1.58763126e-01
-2.89106399e-01 6.56895638e-01 2.11544156e-01 3.42163563e-01
1.22189784e+00 -1.60189182e-01 -7.99585730e-02 -3.13130111e-01
3.48521888e-01 7.17816949e-01 5.19582272e-01 -5.69798112e-01
7.83107817e-01 3.75846148e-01 -3.19457114e-01 -3.88702065e-01
-1.21741283e+00 -6.10744096e-02 -3.71847630e-01 -1.64297000e-01
7.56250322e-01 -4.22894746e-01 -2.43670762e-01 -5.89227080e-01
-1.64476252e+00 1.78874247e-02 -5.06004214e-01 3.16559076e-01
-2.44821072e-01 3.94233972e-01 -4.07290399e-01 -9.05627012e-01
-2.33099341e-01 -1.02443409e+00 1.04479408e+00 7.23279044e-02
-9.68188465e-01 -8.60135794e-01 -5.94703434e-03 9.93638754e-01
5.64467430e-01 3.67205292e-01 1.54228461e+00 -7.81073809e-01
-4.05581623e-01 -4.64168608e-01 -1.84778288e-01 -6.26377612e-02
3.27440538e-02 -1.59309074e-01 -1.13077605e+00 2.33028159e-01
1.50364071e-01 -9.26515758e-02 6.02244914e-01 4.06080671e-02
5.16682267e-01 -7.23102212e-01 -2.93321282e-01 -1.26336515e-01
1.29916668e+00 4.55310106e-01 9.92195010e-01 4.67208177e-01
8.86845365e-02 9.98520315e-01 4.95674998e-01 -6.52191415e-02
4.51821059e-01 6.72113359e-01 2.58246213e-01 2.53654364e-03
-2.00858861e-01 -3.24028194e-01 1.48379415e-01 1.52804703e-01
-1.28745288e-01 -5.20987362e-02 -1.04015934e+00 5.88313043e-01
-1.97118902e+00 -1.42669916e+00 -3.44931811e-01 1.92413354e+00
5.32325029e-01 2.22469538e-01 -5.46535432e-01 1.37607291e-01
4.27233666e-01 -1.21580444e-01 -3.74607205e-01 -1.03235519e+00
-3.22783589e-02 8.29737931e-02 -1.39112487e-01 9.36401486e-01
-3.47108424e-01 7.81261504e-01 6.75092125e+00 5.18491387e-01
-7.90920436e-01 4.84585613e-02 5.14825106e-01 1.66218337e-02
-9.37864304e-01 4.32029635e-01 -2.55562663e-01 2.16465041e-01
8.20221782e-01 -3.30988616e-01 1.53342903e-01 6.71616256e-01
5.82728028e-01 -2.98039943e-01 -1.18057251e+00 3.73521298e-01
1.90917253e-01 -1.42021465e+00 5.78571856e-01 3.24131519e-01
4.67034370e-01 -5.34632862e-01 -5.70107400e-02 2.57687837e-01
-8.89443904e-02 -1.14318526e+00 1.00135326e+00 6.67089105e-01
2.57992029e-01 -2.21803308e-01 9.34179008e-01 4.42180157e-01
-3.91567409e-01 -1.58677056e-01 -3.28298807e-01 -5.55585146e-01
4.22275037e-01 2.88839519e-01 -1.21474099e+00 5.73192418e-01
3.48916143e-01 3.52749467e-01 -8.10111880e-01 8.94998968e-01
-6.49816096e-01 3.23051810e-01 3.98454554e-02 1.78493243e-02
2.27664933e-01 -2.61405148e-02 6.21621847e-01 9.74493742e-01
2.30205819e-01 4.83025610e-01 -3.35337907e-01 1.24393654e+00
4.38801706e-01 1.44739896e-01 -1.04411113e+00 1.27504185e-01
4.08469170e-01 8.30537558e-01 -4.18463022e-01 -6.49944305e-01
-1.81921735e-01 7.69658744e-01 2.02112541e-01 4.99383450e-01
-5.44397175e-01 -1.86279550e-01 3.52939904e-01 4.49438632e-01
4.37085927e-02 3.34800482e-01 -5.26636660e-01 -1.04804981e+00
4.90758240e-01 -1.10944116e+00 3.31624955e-01 -1.47346425e+00
-9.80837226e-01 6.82301104e-01 3.51319760e-01 -8.25279951e-01
-5.26166677e-01 -4.01097715e-01 -6.75658107e-01 1.02032399e+00
-1.48021746e+00 -8.53865981e-01 -1.70762688e-02 6.82420144e-03
5.77110171e-01 2.51573086e-01 8.38672996e-01 7.00952262e-02
-1.86919808e-01 3.02046817e-03 -5.12488663e-01 -1.27599567e-01
4.53425139e-01 -1.17215216e+00 4.13929701e-01 9.46530700e-01
2.37129658e-01 1.32046866e+00 1.17244577e+00 -8.88607800e-01
-8.83393943e-01 -4.79635656e-01 1.65878212e+00 -8.65894556e-01
3.36416245e-01 2.26886086e-02 -1.24316061e+00 7.16968298e-01
5.97930908e-01 -7.60265887e-01 7.21824348e-01 1.59354150e-01
-4.89760816e-01 2.14513093e-01 -1.21725893e+00 9.89968717e-01
9.85119164e-01 -6.69069588e-01 -1.66517174e+00 5.76818943e-01
6.05616391e-01 -3.86141807e-01 -4.42753047e-01 -1.43934742e-01
7.14588583e-01 -1.12802422e+00 8.08484256e-01 -8.83787572e-01
7.96066761e-01 -1.30219862e-01 9.77224559e-02 -1.04396546e+00
-1.70094177e-01 -3.77399236e-01 5.14404237e-01 9.97038186e-01
1.03920770e+00 -8.58125985e-01 2.37941861e-01 1.44518709e+00
-2.98270881e-01 -7.01456666e-01 -1.15544629e+00 -3.76431435e-01
1.97280198e-01 -6.18814886e-01 5.90620816e-01 7.68888593e-01
5.57924330e-01 6.31128609e-01 1.50315806e-01 1.37061059e-01
2.31439501e-01 2.25954264e-01 5.64804971e-01 -1.29887366e+00
-7.93562904e-02 -2.92352200e-01 -2.17850491e-01 -6.67250097e-01
9.71064940e-02 -8.92154515e-01 -1.58793539e-01 -2.37920332e+00
2.28863776e-01 1.32491305e-01 3.59701574e-01 6.22618377e-01
-3.01205546e-01 -2.73627549e-01 4.37804729e-01 2.58481592e-01
-2.76810378e-01 8.65782350e-02 1.38004625e+00 7.72358030e-02
-1.70566127e-01 -3.97633940e-01 -1.63834786e+00 7.52585649e-01
6.54656351e-01 -6.62477195e-01 -5.37153423e-01 -9.21697497e-01
4.39719647e-01 1.12282448e-01 7.04546809e-01 -7.51289070e-01
3.40336680e-01 -2.12358668e-01 1.96151912e-01 -2.58851886e-01
3.98074657e-01 -7.53718197e-01 2.04550192e-01 4.50319052e-01
-7.10080445e-01 1.47290945e-01 4.44464386e-01 3.05126578e-01
-1.71332404e-01 -5.49045205e-01 2.02801414e-02 -2.91922718e-01
-1.88276798e-01 -6.37317300e-01 -6.41014397e-01 1.98611375e-02
6.96411729e-01 -6.16481543e-01 -8.46019864e-01 -8.13256681e-01
-6.37247503e-01 8.31348523e-02 3.99526566e-01 3.69366199e-01
7.28625298e-01 -1.00713360e+00 -7.87679315e-01 -2.03155279e-01
1.81343749e-01 -2.16292143e-01 2.07540944e-01 7.15450585e-01
-3.84101242e-01 9.68436897e-01 2.04309538e-01 -1.94862768e-01
-1.08632004e+00 2.80319780e-01 2.44655520e-01 -1.20827235e-01
-7.80482948e-01 2.59720802e-01 1.07235961e-01 -1.46279037e-01
-1.29969984e-01 -7.34703004e-01 -4.80402648e-01 1.48635602e-03
7.25446820e-01 2.06001848e-02 1.03785641e-01 -6.89566135e-01
-3.17741394e-01 6.07919812e-01 -3.21836591e-01 -5.22754669e-01
1.34534824e+00 -2.76197417e-04 -6.73001558e-02 3.13577801e-01
5.56595325e-01 2.96522021e-01 -5.77275753e-01 1.78408384e-01
2.42042422e-01 -7.79643178e-01 -2.24768803e-01 -1.27753663e+00
-4.48921621e-01 8.31804991e-01 8.13117474e-02 4.49886382e-01
6.31825030e-01 2.40048155e-01 5.57602882e-01 6.15730047e-01
2.32099205e-01 -7.81811535e-01 -2.71000475e-01 3.22950602e-01
1.49997711e+00 -9.77878034e-01 5.64449430e-02 -5.72403550e-01
-7.06442118e-01 1.26523912e+00 3.59599262e-01 3.22434723e-01
-1.09302431e-01 -1.83579668e-01 6.80487156e-01 -9.08272743e-01
-9.47702348e-01 -8.07035491e-02 3.19358408e-01 4.41644877e-01
6.79414988e-01 -3.57864708e-01 -7.24033833e-01 5.33017516e-01
-5.12696922e-01 7.41803274e-02 7.90902376e-01 1.03169107e+00
-4.19315994e-01 -1.14445436e+00 -7.18376517e-01 4.31033283e-01
-5.17521858e-01 -3.60362589e-01 -1.02713895e+00 8.76245081e-01
1.66183524e-02 1.44586480e+00 -3.26336831e-01 9.83970761e-02
7.57645726e-01 4.43768054e-01 1.34881079e-01 -9.60486114e-01
-9.30528343e-01 -2.57984698e-01 5.53833485e-01 -4.50437814e-01
-1.93197325e-01 -5.70897222e-01 -1.29528189e+00 -6.39149696e-02
-2.66277850e-01 4.99099731e-01 5.44102907e-01 1.51577044e+00
6.57469630e-01 5.48880339e-01 -2.99168974e-01 -2.00259089e-01
-5.49738884e-01 -9.55345929e-01 -6.28070682e-02 5.85344613e-01
3.68836999e-01 -3.48145872e-01 -3.58045280e-01 -5.44811860e-02] | [10.18641185760498, 7.95232629776001] |
f1a19372-11be-4446-a60a-efa9aee5c3e1 | reference-conditioned-super-resolution-by | 1804.03360 | null | http://arxiv.org/abs/1804.03360v1 | http://arxiv.org/pdf/1804.03360v1.pdf | Reference-Conditioned Super-Resolution by Neural Texture Transfer | With the recent advancement in deep learning, we have witnessed a great
progress in single image super-resolution. However, due to the significant
information loss of the image downscaling process, it has become extremely
challenging to further advance the state-of-the-art, especially for large
upscaling factors. This paper explores a new research direction in super
resolution, called reference-conditioned super-resolution, in which a reference
image containing desired high-resolution texture details is provided besides
the low-resolution image. We focus on transferring the high-resolution texture
from reference images to the super-resolution process without the constraint of
content similarity between reference and target images, which is a key
difference from previous example-based methods. Inspired by recent work on
image stylization, we address the problem via neural texture transfer. We
design an end-to-end trainable deep model which generates detail enriched
results by adaptively fusing the content from the low-resolution image with the
texture patterns from the reference image. We create a benchmark dataset for
the general research of reference-based super-resolution, which contains
reference images paired with low-resolution inputs with varying degrees of
similarity. Both objective and subjective evaluations demonstrate the great
potential of using reference images as well as the superiority of our results
over other state-of-the-art methods. | ['Zhifei Zhang', 'Zhaowen Wang', 'Zhe Lin', 'Hairong Qi'] | 2018-04-10 | null | null | null | null | ['image-stylization', 'reference-based-super-resolution'] | ['computer-vision', 'computer-vision'] | [ 7.63979673e-01 -1.16396144e-01 5.03537320e-02 -2.16434240e-01
-1.17872739e+00 -3.73278558e-02 5.56877553e-01 -7.79888988e-01
4.84277084e-02 9.22114611e-01 5.21177948e-01 4.73378211e-01
-8.07465389e-02 -9.66525733e-01 -8.06263447e-01 -9.44088399e-01
4.38531429e-01 2.19774153e-02 2.87039220e-01 -6.55932605e-01
2.99300104e-01 3.16376120e-01 -1.75953865e+00 7.35492468e-01
9.27411914e-01 9.34363663e-01 6.05429232e-01 4.59255695e-01
-8.27818140e-02 5.53329229e-01 -4.15302873e-01 -2.15853915e-01
2.94442564e-01 -6.87113702e-01 -7.57491112e-01 2.58661896e-01
8.88393760e-01 -5.54229140e-01 -3.60925019e-01 1.30728102e+00
5.89264691e-01 2.05797747e-01 3.18719685e-01 -4.66263771e-01
-1.48463285e+00 5.11451125e-01 -1.06424403e+00 4.94104326e-01
3.48883361e-01 -2.31611475e-01 5.05609214e-01 -9.14953589e-01
8.56155634e-01 1.47744310e+00 5.04576743e-01 7.90074944e-01
-1.45547020e+00 -7.06118941e-01 2.32692454e-02 9.06431451e-02
-1.29106617e+00 -4.24999744e-01 8.76594245e-01 -1.48367092e-01
4.55681294e-01 9.77329910e-02 6.94095716e-02 1.25719821e+00
2.98263788e-01 2.37008929e-01 1.57737803e+00 -2.83199489e-01
-1.25150934e-01 4.02369089e-02 -4.22204316e-01 3.73806477e-01
8.49118456e-02 5.87500513e-01 -5.59786797e-01 4.16126788e-01
1.74309540e+00 5.46279810e-02 -4.87849146e-01 -1.28888398e-01
-1.19285810e+00 4.38257784e-01 5.39522648e-01 4.75221217e-01
-3.85492265e-01 -1.73642188e-01 5.47518842e-02 2.37378895e-01
8.04198444e-01 2.50786364e-01 -1.02100149e-01 3.95153642e-01
-1.01334059e+00 1.65273860e-01 4.50511687e-02 9.46201980e-01
6.73266590e-01 4.14057344e-01 -3.08249235e-01 1.05954969e+00
-3.81785154e-01 3.10120225e-01 4.98646706e-01 -1.04454565e+00
4.05050278e-01 1.74660131e-01 4.96265411e-01 -8.52705240e-01
1.12127714e-01 -5.95513761e-01 -1.56050372e+00 4.38489348e-01
1.75900310e-01 2.36618489e-01 -8.97823036e-01 1.64723825e+00
3.14669997e-01 2.55670369e-01 2.39691541e-01 1.30074024e+00
9.81188118e-01 7.37991512e-01 -2.19354525e-01 -4.60289747e-01
1.29395187e+00 -9.77753460e-01 -8.92533898e-01 2.06045255e-01
-3.20781589e-01 -9.61467385e-01 1.33367896e+00 4.19200063e-01
-1.27886045e+00 -1.21115363e+00 -1.03806436e+00 -4.61199880e-01
-8.45740661e-02 -1.74524769e-01 3.62855881e-01 2.81637132e-01
-1.13197386e+00 7.22748816e-01 -2.89500594e-01 -7.01742172e-02
5.29927254e-01 -6.50010258e-02 -5.28161645e-01 -3.40472907e-01
-1.25593305e+00 7.96684563e-01 2.78636515e-01 4.31907251e-02
-8.00789714e-01 -9.64393198e-01 -7.16627955e-01 -4.98735607e-02
3.72793019e-01 -8.24017942e-01 8.94869208e-01 -1.21595991e+00
-1.60846269e+00 9.01918530e-01 -1.92173928e-01 -7.96062872e-02
4.73382860e-01 -2.73356050e-01 -6.24121606e-01 3.15283760e-02
2.44070902e-01 3.06694031e-01 1.23604751e+00 -1.64747608e+00
-9.20504093e-01 -2.32133582e-01 -3.86083052e-02 3.50881398e-01
2.22591441e-02 3.72460559e-02 -3.74124706e-01 -9.44274545e-01
5.13441712e-02 -3.62950712e-01 -1.87300265e-01 -1.97285593e-01
-1.16649501e-01 2.34378546e-01 9.35466468e-01 -7.20452130e-01
1.05201328e+00 -2.02699304e+00 4.30944830e-01 -5.33142149e-01
2.98775762e-01 1.20400362e-01 -2.98937649e-01 -9.46130529e-02
-3.26430351e-01 1.43007681e-01 -2.29222894e-01 -1.35376737e-01
-4.61622745e-01 -1.23610780e-01 -7.43742049e-01 1.52544320e-01
3.10442805e-01 8.00815344e-01 -9.96415734e-01 -4.47626919e-01
3.54896575e-01 1.04871082e+00 -2.49448106e-01 4.46595043e-01
4.97957356e-02 1.08913159e+00 -4.06042635e-01 4.78253037e-01
1.00609958e+00 -4.02817667e-01 -2.69253612e-01 -7.31263816e-01
-2.69836515e-01 -1.84085295e-01 -1.10319948e+00 1.72944641e+00
-6.91388309e-01 3.92023563e-01 1.59660518e-01 -7.38228798e-01
1.02727246e+00 1.83773085e-01 3.16608548e-01 -1.27625811e+00
-1.97572172e-01 7.00272769e-02 -4.30283278e-01 -4.03491527e-01
8.26230109e-01 -5.23583233e-01 1.06057696e-01 3.08765978e-01
-1.50982440e-01 -2.15237141e-01 -1.04938596e-01 -1.71616971e-01
4.01577264e-01 4.98939604e-01 2.07982600e-01 -1.25660673e-01
6.99688315e-01 -3.05685878e-01 3.72627825e-01 4.41520512e-01
1.52210757e-01 1.28310752e+00 -6.85829893e-02 -5.80732226e-01
-1.56736183e+00 -1.31054747e+00 -3.02658141e-01 1.07216275e+00
4.89321351e-01 1.84931085e-01 -8.54536593e-01 -1.83605149e-01
-4.83485878e-01 2.63031334e-01 -9.47916329e-01 1.47208884e-01
-8.14431846e-01 -7.63750017e-01 1.13338726e-02 4.62749809e-01
1.13501501e+00 -1.09808695e+00 -2.91670442e-01 1.43690944e-01
-5.36815047e-01 -1.26972258e+00 -7.18359530e-01 -3.16677421e-01
-8.58112097e-01 -6.71028733e-01 -1.18213868e+00 -8.35716009e-01
4.87095416e-01 6.47275686e-01 1.32180452e+00 -1.33413255e-01
-2.71059930e-01 -4.38003540e-02 -3.24675143e-01 1.97060242e-01
-3.34366500e-01 -2.06014454e-01 -1.08354136e-01 3.39617014e-01
-1.61197528e-01 -6.19410038e-01 -7.97143042e-01 4.20008749e-01
-1.22445607e+00 5.98009646e-01 8.42577457e-01 1.03245163e+00
1.26575518e+00 5.27816892e-01 4.93413359e-01 -9.04162645e-01
5.88989735e-01 -2.57791340e-01 -4.95245546e-01 2.03789935e-01
-3.84413809e-01 2.07230955e-01 8.15189481e-01 -5.86199760e-01
-1.71070135e+00 -3.57810557e-01 1.18397541e-01 -5.90173483e-01
-1.35425672e-01 5.77007532e-02 -2.97827840e-01 -1.45292759e-01
6.55940354e-01 5.97334027e-01 -2.10296109e-01 -6.90595448e-01
5.09943306e-01 4.93295610e-01 1.01494193e+00 -5.79020202e-01
9.30513561e-01 7.89247692e-01 2.87428554e-02 -7.54752040e-01
-1.33760989e+00 8.13522115e-02 -6.91354573e-01 8.11261535e-02
9.88641739e-01 -9.93852258e-01 -2.95144141e-01 3.30522805e-01
-8.70586157e-01 -2.68916816e-01 -3.44400257e-01 1.48396626e-01
-7.36315489e-01 3.76826584e-01 -6.54011667e-01 -3.68350983e-01
-5.58206260e-01 -1.20520782e+00 1.26467633e+00 5.49666882e-01
2.32485801e-01 -7.19240189e-01 -3.29415239e-02 5.21640480e-01
9.16522801e-01 4.22800183e-01 7.59126127e-01 3.24158013e-01
-9.11571801e-01 3.63457143e-01 -8.53100598e-01 4.35738087e-01
4.91311640e-01 -3.76335621e-01 -8.64410579e-01 -3.32974583e-01
3.11098009e-01 -1.78244397e-01 7.44739294e-01 4.11180943e-01
1.26443946e+00 -2.29335234e-01 7.64072016e-02 1.07030833e+00
1.80607915e+00 -7.97620043e-02 1.01498353e+00 5.37995875e-01
8.92573357e-01 5.07230103e-01 8.56087446e-01 1.18026778e-01
1.71703011e-01 9.85592723e-01 1.26319349e-01 -4.55392778e-01
-7.02849746e-01 -3.29451382e-01 -2.61133183e-02 4.69697148e-01
-6.14001989e-01 1.53148592e-01 -2.51188070e-01 3.26354682e-01
-1.49575996e+00 -1.31088579e+00 1.62162125e-01 2.26435781e+00
1.08341312e+00 -1.42213896e-01 -1.56745031e-01 -1.88956484e-01
1.01561809e+00 3.32494557e-01 -6.81479633e-01 -1.49047136e-01
-3.95040691e-01 3.68785858e-01 1.67443767e-01 5.86304188e-01
-1.09762645e+00 1.14206004e+00 6.01919079e+00 1.24797368e+00
-1.32688332e+00 1.51709676e-01 1.08644938e+00 -2.44921315e-02
-3.12592536e-01 -4.57822055e-01 -8.67035687e-01 5.68992972e-01
6.49385929e-01 -2.21080422e-01 7.92973876e-01 5.77746212e-01
2.22610593e-01 7.83943664e-03 -8.86823475e-01 1.25980127e+00
1.49180204e-01 -1.46251512e+00 5.24235785e-01 -1.35020167e-01
1.23224306e+00 -3.56571257e-01 5.56726635e-01 9.17940736e-02
7.61768147e-02 -1.43574691e+00 4.36597764e-01 7.36610353e-01
1.70678329e+00 -8.03641617e-01 5.78791142e-01 -4.20817174e-02
-1.42770064e+00 7.60255158e-02 -6.99790895e-01 2.32240766e-01
1.43366039e-01 4.49851841e-01 1.82773899e-02 9.85271573e-01
1.04123890e+00 7.06493914e-01 -2.51634896e-01 4.23267484e-01
-3.16815637e-02 8.31485167e-02 1.62447229e-01 8.46050024e-01
-1.36392310e-01 -4.05674338e-01 2.83455849e-01 9.87443447e-01
3.35806012e-01 5.03311992e-01 -1.37394533e-01 1.26888466e+00
-1.16638429e-01 -1.01803482e-01 -4.86312687e-01 3.64512235e-01
3.32703948e-01 1.36776650e+00 -5.85581601e-01 -4.31891859e-01
-4.71157908e-01 1.19924712e+00 2.64430404e-01 5.17463803e-01
-6.50273681e-01 -1.81978703e-01 4.35943991e-01 3.53390098e-01
4.35507387e-01 4.56514098e-02 -5.19359887e-01 -1.26517463e+00
3.27704027e-02 -1.07266176e+00 1.03654183e-01 -1.17754829e+00
-1.44751501e+00 1.11830127e+00 -7.15247989e-02 -1.53802574e+00
-7.35017583e-02 -1.31188363e-01 -3.14557016e-01 1.52784729e+00
-1.79793942e+00 -1.28305709e+00 -7.36283481e-01 7.34535575e-01
8.55462909e-01 -5.99332526e-02 5.33509374e-01 2.51880944e-01
-2.95828819e-01 4.83592868e-01 1.77671567e-01 -7.85013437e-02
8.11514020e-01 -8.99872541e-01 3.35975498e-01 8.58894706e-01
-3.23547632e-01 5.20908833e-01 7.99135268e-01 -7.00288236e-01
-1.18864870e+00 -1.17607617e+00 3.05413991e-01 -5.75814605e-01
3.55094165e-01 -1.44567773e-01 -1.31125367e+00 2.71253586e-01
1.40333593e-01 2.12912798e-01 1.15827873e-01 -2.63283163e-01
-4.30448145e-01 -2.57653177e-01 -1.26648450e+00 7.33471572e-01
1.13888836e+00 -5.79470992e-01 -6.81626379e-01 -1.84303224e-01
9.18928981e-01 -6.52561486e-01 -1.14233470e+00 6.06965601e-01
5.65816581e-01 -1.21233594e+00 1.34032011e+00 -2.14513451e-01
1.10368752e+00 -4.65857744e-01 -2.75190711e-01 -1.14203727e+00
-7.34704196e-01 -5.03046930e-01 1.03623562e-01 1.25794590e+00
-7.64992535e-02 -3.90132785e-01 6.51206732e-01 4.05357957e-01
1.38229743e-01 -6.26332402e-01 -7.45932937e-01 -4.44496751e-01
1.69785112e-01 3.42292935e-01 7.18924999e-01 1.00610781e+00
-7.24769056e-01 5.19575953e-01 -8.22427511e-01 2.36841723e-01
1.13327026e+00 6.25212908e-01 7.20683634e-01 -9.25264716e-01
-2.74275810e-01 -4.42498893e-01 1.11851968e-01 -9.56686378e-01
-1.89298734e-01 -4.26365733e-01 1.48136448e-02 -1.45965254e+00
5.83969176e-01 -1.48845285e-01 -4.25400048e-01 -7.62413442e-02
-4.07102346e-01 6.64448321e-01 -1.13442555e-01 2.95051247e-01
-4.52582031e-01 6.56373203e-01 2.13722873e+00 1.76893651e-01
-2.10091472e-01 -2.57279605e-01 -1.08358729e+00 5.18564522e-01
5.41549027e-01 1.37414911e-03 -4.61349845e-01 -5.46842992e-01
-1.07917003e-01 5.24775803e-01 3.49324226e-01 -7.13819385e-01
-1.53598726e-01 -3.09849024e-01 7.84367621e-01 -5.61862111e-01
3.29241574e-01 -6.24763966e-01 3.30316454e-01 -1.38708085e-01
-5.84218919e-01 -1.63852051e-01 8.92625824e-02 6.31124318e-01
-3.81597966e-01 3.08021516e-01 1.26019418e+00 -3.67107302e-01
-8.80342603e-01 6.51548624e-01 3.26010674e-01 1.23340875e-01
5.04224420e-01 -4.42426085e-01 -5.40189505e-01 -2.49640360e-01
-5.95218837e-01 -3.65572184e-01 7.45474815e-01 5.85800350e-01
8.97258580e-01 -1.32707500e+00 -1.08612609e+00 2.96470672e-01
-6.88654184e-02 2.10640684e-01 9.13589895e-01 5.16799152e-01
-1.34484604e-01 1.07663922e-01 -7.63709605e-01 -4.03189331e-01
-1.12276733e+00 1.02890682e+00 4.03810710e-01 -2.94638634e-01
-1.14339924e+00 5.95827460e-01 9.73859131e-01 5.77521250e-02
-8.55202079e-02 1.37885353e-02 -3.97673190e-01 -4.91854221e-01
1.09316838e+00 1.95286214e-01 -2.48356253e-01 -8.06898355e-01
2.80035257e-01 1.21480596e+00 -3.66811961e-01 -4.84962873e-02
1.46785343e+00 -5.45556366e-01 -1.08597390e-01 2.76647896e-01
9.59707320e-01 -8.26997310e-02 -1.57557094e+00 -6.88657880e-01
-4.16817993e-01 -9.28281546e-01 2.21793294e-01 -7.65038490e-01
-1.27251852e+00 8.35365117e-01 8.21485758e-01 -4.84999120e-02
1.54069221e+00 -1.31737918e-01 7.62188494e-01 -2.26113141e-01
6.86677516e-01 -8.80891502e-01 2.78696269e-01 2.48215497e-01
1.32424891e+00 -1.47253609e+00 1.07148401e-01 -4.53109562e-01
-5.97583771e-01 9.01220858e-01 7.21136749e-01 -4.37585622e-01
2.63374060e-01 3.36389095e-01 6.77385852e-02 7.03468472e-02
-6.76306188e-01 -2.49549717e-01 3.40829104e-01 9.33290362e-01
5.03257394e-01 -1.88303262e-01 -1.20405830e-01 6.63819969e-01
-1.33690774e-01 2.99364626e-01 6.47908390e-01 2.08519101e-01
-3.73414814e-01 -8.70350182e-01 -6.37240648e-01 1.68873832e-01
-6.57929480e-01 -3.28197867e-01 1.01921968e-01 6.37565851e-01
-9.81484423e-04 6.75401807e-01 3.17553000e-04 -3.56093138e-01
3.43246222e-01 -5.81755817e-01 6.69516146e-01 -4.72081453e-01
-8.92226249e-02 2.90124297e-01 -2.61655688e-01 -7.34074831e-01
-6.32481098e-01 -1.26195565e-01 -8.38808358e-01 -3.24836403e-01
1.06377602e-01 -8.05497766e-02 2.40051642e-01 6.10140979e-01
3.26338798e-01 8.23455215e-01 6.36244297e-01 -1.47797430e+00
-3.09622675e-01 -9.90252674e-01 -7.82450497e-01 6.47798717e-01
5.68823934e-01 -6.92188859e-01 -3.33172053e-01 2.95361280e-01] | [10.969175338745117, -2.084467649459839] |
d8916b6a-4bab-43f0-a82a-4353dac2f9f4 | towards-unsupervised-learning-of-temporal | 1401.6427 | null | http://arxiv.org/abs/1401.6427v1 | http://arxiv.org/pdf/1401.6427v1.pdf | Towards Unsupervised Learning of Temporal Relations between Events | Automatic extraction of temporal relations between event pairs is an
important task for several natural language processing applications such as
Question Answering, Information Extraction, and Summarization. Since most
existing methods are supervised and require large corpora, which for many
languages do not exist, we have concentrated our efforts to reduce the need for
annotated data as much as possible. This paper presents two different
algorithms towards this goal. The first algorithm is a weakly supervised
machine learning approach for classification of temporal relations between
events. In the first stage, the algorithm learns a general classifier from an
annotated corpus. Then, inspired by the hypothesis of "one type of temporal
relation per discourse, it extracts useful information from a cluster of
topically related documents. We show that by combining the global information
of such a cluster with local decisions of a general classifier, a bootstrapping
cross-document classifier can be built to extract temporal relations between
events. Our experiments show that without any additional annotated data, the
accuracy of the proposed algorithm is higher than that of several previous
successful systems. The second proposed method for temporal relation extraction
is based on the expectation maximization (EM) algorithm. Within EM, we used
different techniques such as a greedy best-first search and integer linear
programming for temporal inconsistency removal. We think that the experimental
results of our EM based algorithm, as a first step toward a fully unsupervised
temporal relation extraction method, is encouraging. | ['Gholamreza Ghassem-Sani', 'Seyed Abolghasem Mirroshandel'] | 2014-01-23 | null | null | null | null | ['temporal-relation-extraction'] | ['natural-language-processing'] | [ 3.09211642e-01 4.96230036e-01 -3.19050103e-01 -4.05483574e-01
-8.10460806e-01 -4.87183452e-01 8.84678960e-01 8.70910287e-01
-5.08085847e-01 9.43406880e-01 2.26742730e-01 -3.80532384e-01
-2.26347283e-01 -8.01106036e-01 -4.03976053e-01 -6.69048071e-01
-2.53146917e-01 5.60293972e-01 7.05763161e-01 -2.46725023e-01
5.04745066e-01 3.46354783e-01 -1.43101668e+00 5.13615072e-01
7.97731042e-01 7.77346194e-01 6.57358691e-02 4.37715679e-01
-4.22654629e-01 1.26874626e+00 -5.41162968e-01 -1.96020529e-01
-1.75470963e-01 -7.65349925e-01 -1.43937838e+00 1.72012419e-01
-4.09235984e-01 1.85120925e-01 1.12423591e-01 7.42611885e-01
1.42550826e-01 4.94420618e-01 7.04965711e-01 -1.01049542e+00
1.77411392e-01 8.53400648e-01 -4.22851890e-01 3.30605835e-01
6.11948252e-01 -4.28297132e-01 9.53174949e-01 -4.34627146e-01
9.33751583e-01 9.10226285e-01 4.77840692e-01 1.91837102e-01
-1.03248382e+00 -2.26310983e-01 3.48177999e-01 5.12131870e-01
-1.34378994e+00 -3.36525530e-01 1.04500389e+00 -5.65577447e-01
1.25778985e+00 4.40588593e-01 6.27148211e-01 7.66110182e-01
1.08596355e-01 7.36793339e-01 1.04745758e+00 -1.01389277e+00
3.95467460e-01 4.78277028e-01 5.42508781e-01 4.45179939e-01
-1.49960160e-01 -3.46356124e-01 -6.03341401e-01 -3.36098313e-01
2.65459478e-01 -5.56453049e-01 -1.94532201e-01 -7.13195130e-02
-1.04202104e+00 5.95943511e-01 -1.16174251e-01 1.09745324e+00
-4.65732306e-01 -2.80129492e-01 6.17474675e-01 2.80182749e-01
8.69657278e-01 2.16809422e-01 -4.68457043e-01 -1.21915437e-01
-1.12830007e+00 1.02276407e-01 1.16426229e+00 6.41116083e-01
4.42805201e-01 -6.38232172e-01 1.57228857e-01 7.89674401e-01
1.72733054e-01 -1.27007499e-01 6.63415730e-01 -6.48733377e-01
5.36410749e-01 8.46809864e-01 1.91990390e-01 -1.24110091e+00
-4.90624219e-01 -8.58572964e-03 -5.49219429e-01 -2.03362018e-01
4.62559551e-01 -9.31120738e-02 -3.07086408e-01 1.63383877e+00
5.72235107e-01 3.12296059e-02 2.53283679e-01 4.72531974e-01
6.23214245e-01 7.70768166e-01 4.34774999e-03 -1.24244738e+00
1.34033847e+00 -7.23506629e-01 -1.11778724e+00 1.13414533e-01
8.87365401e-01 -8.75420809e-01 3.28377098e-01 4.75478798e-01
-9.62244093e-01 -1.50015503e-01 -9.18708563e-01 1.64508149e-01
-5.26592374e-01 8.80863145e-03 6.90632343e-01 2.29321033e-01
-6.82088435e-01 6.51965439e-01 -9.50931311e-01 -9.13618803e-01
-1.85726762e-01 3.00514936e-01 -3.93979967e-01 5.37354231e-01
-1.08236766e+00 1.06266022e+00 9.21346128e-01 1.69426039e-01
-1.76642478e-01 1.15912169e-01 -6.60457850e-01 -1.82030752e-01
6.52922153e-01 -2.32848793e-01 1.28496253e+00 -1.00028348e+00
-1.32501030e+00 8.97051752e-01 -6.67780638e-01 -6.35670125e-01
2.82072186e-01 -9.42212194e-02 -4.99854416e-01 4.13820535e-01
1.55466467e-01 -1.50771085e-02 4.82950658e-01 -1.15496600e+00
-7.15262234e-01 -2.94703752e-01 -1.32774338e-01 6.03706427e-02
-2.97521651e-01 4.94700760e-01 -2.53354192e-01 -5.10201991e-01
4.13045704e-01 -8.37091446e-01 -2.36196011e-01 -6.84486568e-01
-2.99781144e-01 -7.37348616e-01 7.34680831e-01 -7.13603616e-01
1.65702045e+00 -1.98258853e+00 1.20634638e-01 2.21268162e-01
-2.16560587e-01 3.91067564e-02 4.14621860e-01 7.29227304e-01
-4.24190998e-01 -9.03907791e-03 -3.46222937e-01 -8.87179002e-02
-2.82933086e-01 2.32766226e-01 -2.97681421e-01 2.85185784e-01
1.25533670e-01 3.31556112e-01 -9.79768515e-01 -1.05241680e+00
9.84075144e-02 5.99460192e-02 -2.46318519e-01 2.33381435e-01
-3.86469394e-01 4.20678943e-01 -5.18224299e-01 1.68642104e-01
1.97203726e-01 2.38372516e-02 6.43831968e-01 -2.09633574e-01
-4.21393603e-01 6.49796844e-01 -1.44973528e+00 1.54512668e+00
-2.83203334e-01 6.42640889e-01 -3.65338027e-01 -1.63648927e+00
1.00213909e+00 9.48873281e-01 9.24358010e-01 -3.95825595e-01
2.44545281e-01 2.00127438e-01 -2.00179800e-01 -9.49294746e-01
5.20926654e-01 -1.95986331e-01 -7.78433233e-02 4.54822212e-01
1.26232296e-01 1.17399096e-01 6.68301880e-01 1.68789148e-01
1.05110013e+00 3.23532283e-01 9.41129684e-01 -2.31068477e-01
8.26480985e-01 3.76604289e-01 6.50291443e-01 3.39015931e-01
-4.75274883e-02 2.06551597e-01 7.08270311e-01 -5.28460681e-01
-7.15681374e-01 -5.59503436e-01 -1.19505145e-01 7.67674029e-01
-6.21601474e-03 -7.04796314e-01 -6.32390857e-01 -7.58031487e-01
-6.83195174e-01 7.80291617e-01 -3.82854849e-01 3.24039578e-01
-8.64777863e-01 -9.75777388e-01 2.34588534e-01 2.17048123e-01
4.67722058e-01 -1.13079202e+00 -6.54661357e-01 4.59778219e-01
-7.24421918e-01 -1.18120003e+00 1.51516944e-01 4.85721558e-01
-1.08483827e+00 -1.19554400e+00 -9.59100574e-02 -8.38955164e-01
4.25967544e-01 -1.10545851e-01 8.77504766e-01 -1.00830287e-01
1.91480175e-01 1.31005704e-01 -7.45154798e-01 -3.52441281e-01
-5.68980277e-01 6.73577785e-02 -7.33734518e-02 1.22276470e-01
5.92541099e-01 -7.71690905e-01 -1.01970024e-01 1.04691565e-01
-7.79999614e-01 -8.31920132e-02 3.75296175e-01 5.57780623e-01
3.26281965e-01 5.86947560e-01 5.95675528e-01 -8.69019210e-01
7.38456130e-01 -5.18513858e-01 -3.45962405e-01 3.15782458e-01
-4.47980255e-01 2.56079793e-01 4.81803209e-01 -5.44752181e-01
-1.51104987e+00 2.08050981e-01 -8.30372125e-02 2.97294706e-01
-3.73957664e-01 7.05627322e-01 5.52381538e-02 4.14086521e-01
7.92715609e-01 2.21515968e-01 -3.67265582e-01 -4.00394559e-01
1.37849838e-01 7.69941866e-01 3.97001743e-01 -6.59368694e-01
4.12185699e-01 4.04160589e-01 -6.77360147e-02 -1.05255187e+00
-8.99039447e-01 -8.41649532e-01 -8.55992377e-01 -4.44410414e-01
1.02599049e+00 -4.30824578e-01 -5.29569983e-01 1.34960845e-01
-1.49697340e+00 -9.76788104e-02 -2.33914569e-01 7.92833924e-01
-6.27425492e-01 6.35376155e-01 -4.70563680e-01 -1.25372207e+00
-1.25565246e-01 -5.80668330e-01 7.12558150e-01 1.64508417e-01
-7.20387876e-01 -1.07602406e+00 4.07025188e-01 5.00110745e-01
-2.09533051e-01 4.20926899e-01 7.80417621e-01 -1.10803056e+00
-3.35587442e-01 -2.81644911e-01 2.57750332e-01 8.55447352e-02
1.88174903e-01 -1.77299771e-02 -7.04862416e-01 2.09780961e-01
3.77922475e-01 -1.87444612e-01 5.54554582e-01 1.89491034e-01
4.70176786e-01 -5.03387034e-01 -6.26123011e-01 -1.97432056e-01
1.29224205e+00 6.42385781e-01 5.45333624e-01 5.30396044e-01
1.38544902e-01 9.39691603e-01 1.01402521e+00 4.15233850e-01
3.59513819e-01 7.61050820e-01 -4.53031063e-02 2.61284977e-01
3.38048369e-01 3.98673192e-02 2.74374694e-01 1.00835693e+00
-3.92625749e-01 -1.51751980e-01 -8.77200067e-01 9.06922519e-01
-2.35645986e+00 -1.33426797e+00 -4.28645849e-01 2.14457965e+00
1.05689204e+00 3.53754789e-01 2.43519843e-01 6.36766076e-01
7.35102475e-01 8.19821283e-02 1.80638850e-01 -6.24417186e-01
9.14331302e-02 7.95963332e-02 -2.12632809e-02 5.61731815e-01
-1.19652390e+00 8.79029095e-01 5.50307035e+00 6.58692479e-01
-7.40544021e-01 1.08384304e-01 2.42732093e-01 1.95551753e-01
4.27308343e-02 3.84858370e-01 -6.59633279e-01 2.57133543e-01
1.04438496e+00 -1.91898182e-01 -1.06043145e-01 4.90923911e-01
5.46497405e-01 -7.06681252e-01 -1.05804646e+00 6.02332592e-01
7.19392970e-02 -1.09423947e+00 -3.09669495e-01 -1.23823799e-01
5.76961398e-01 -5.27870536e-01 -8.42558503e-01 3.51984389e-02
7.90587887e-02 -5.59242964e-01 5.24133503e-01 5.41264415e-01
-1.42235950e-01 -7.47562945e-01 9.85522449e-01 8.68653297e-01
-1.44102001e+00 1.38130262e-01 1.26474038e-01 -3.57256204e-01
4.13865983e-01 7.88791299e-01 -9.19930756e-01 8.97566259e-01
5.79044580e-01 3.32625359e-01 -1.96143404e-01 9.27342534e-01
-2.47505933e-01 6.65079176e-01 -5.44777513e-01 -2.55697846e-01
2.87627727e-02 -2.08506241e-01 7.13946521e-01 1.41204417e+00
6.20858558e-02 6.07061803e-01 2.20417276e-01 4.75016087e-01
4.11705196e-01 5.75464547e-01 -7.66789079e-01 -3.61993425e-02
3.22200328e-01 1.12418211e+00 -1.21557665e+00 -5.05344331e-01
-3.37042749e-01 8.31008732e-01 2.74847746e-01 1.13459770e-02
-7.14818597e-01 -2.83688873e-01 -2.11484313e-01 4.73438352e-02
7.21573681e-02 -3.82854760e-01 -1.15429118e-01 -1.13607204e+00
3.10392410e-01 -6.36665583e-01 8.17459822e-01 -5.30002117e-01
-8.47033799e-01 7.95341849e-01 5.77603281e-01 -1.15695643e+00
-6.73811078e-01 -1.40555188e-01 -6.10310197e-01 4.83881503e-01
-8.85854900e-01 -9.76807773e-01 1.21611536e-01 5.72519779e-01
7.42563128e-01 6.04450367e-02 7.44753122e-01 1.53841019e-01
-4.87759739e-01 -8.21010210e-03 -3.67726326e-01 1.56728104e-01
7.30490625e-01 -1.12467480e+00 -5.07829905e-01 1.10564280e+00
3.05850267e-01 6.21961057e-01 1.06434810e+00 -7.55355477e-01
-7.48704851e-01 -6.28848672e-01 1.86677980e+00 -1.22424871e-01
8.89777362e-01 -1.46874031e-02 -1.00191963e+00 7.47569919e-01
5.04587650e-01 -4.85501111e-01 7.02073038e-01 3.32252353e-01
5.81082031e-02 -7.69643933e-02 -8.75228703e-01 4.20258343e-01
7.75109649e-01 -3.79031658e-01 -1.26414168e+00 6.61406338e-01
4.90555406e-01 -1.36343077e-01 -1.04020846e+00 4.25627708e-01
1.37277007e-01 -9.79934752e-01 5.97045243e-01 -4.92386460e-01
3.62047523e-01 -4.08931613e-01 2.67705917e-02 -9.38998938e-01
5.96219413e-02 -7.74113297e-01 -7.73704126e-02 1.72273278e+00
4.63082910e-01 -4.39996779e-01 5.48710704e-01 4.66437697e-01
1.38922542e-01 -4.37570691e-01 -8.24490428e-01 -7.20851600e-01
-2.83798397e-01 -6.61986172e-01 1.03772217e-02 1.33107102e+00
7.81168699e-01 7.27506101e-01 -2.39605457e-01 1.79349989e-01
4.48119432e-01 3.62930357e-01 5.13432741e-01 -1.25310040e+00
-3.95389557e-01 -2.47032464e-01 -3.52341861e-01 -6.36340559e-01
3.36112142e-01 -6.89852536e-01 6.11116588e-02 -1.50333428e+00
2.21041456e-01 -1.69748917e-01 -6.24420010e-02 4.23495382e-01
2.33550332e-02 -3.16179425e-01 -1.68136954e-01 1.85505211e-01
-6.37295127e-01 2.11677685e-01 6.87594950e-01 7.94365630e-02
-7.12039232e-01 1.35443002e-01 -3.24288934e-01 1.05373538e+00
7.80271530e-01 -6.27321839e-01 -4.16938424e-01 1.47629142e-01
2.04242691e-01 5.39808214e-01 2.46646237e-02 -7.70386755e-01
5.59633076e-01 -3.02043855e-01 -2.18407288e-01 -7.21922219e-01
7.48339817e-02 -9.02200639e-01 2.70271987e-01 3.64825726e-01
-3.12967598e-01 -2.09475815e-01 1.42471381e-02 3.77250969e-01
-6.54258013e-01 -6.11609697e-01 5.74792027e-01 -2.83992171e-01
-7.31047153e-01 -4.53815192e-01 -6.81381822e-01 -1.49628237e-01
1.17175329e+00 -6.18173964e-02 -4.56842184e-02 -2.71539241e-01
-1.03837037e+00 4.42783870e-02 -7.05212206e-02 1.36318356e-01
2.09911913e-01 -1.07127690e+00 -4.44143236e-01 -5.23157120e-01
-3.59491855e-02 1.14497900e-01 -8.32323506e-02 1.11463547e+00
-2.75642365e-01 5.52377582e-01 9.59150568e-02 -4.87444907e-01
-1.61169398e+00 8.02006841e-01 2.16694782e-03 -7.45343626e-01
-6.34588659e-01 4.22828466e-01 -2.66243219e-01 8.56799260e-02
1.20467111e-01 -2.99616873e-01 -6.47485137e-01 5.27473927e-01
3.58013570e-01 1.64573148e-01 1.20709881e-01 -6.20838106e-01
-3.71266633e-01 4.19821084e-01 -9.53603350e-03 -5.71742773e-01
1.44100726e+00 -2.83309489e-01 -5.19838631e-01 8.24240685e-01
9.22912478e-01 1.01319745e-01 -3.96764457e-01 -2.16444790e-01
8.49944234e-01 -2.11670958e-02 -2.45232046e-01 -2.93155909e-01
-3.63222808e-01 3.17047685e-01 2.68030595e-02 9.16205585e-01
1.50766015e+00 2.50778973e-01 3.90968353e-01 5.17568052e-01
3.84900510e-01 -1.34414947e+00 -1.43433973e-01 4.99319166e-01
7.40827203e-01 -1.19504666e+00 1.00048289e-01 -7.14741349e-01
-4.25551653e-01 1.22418880e+00 3.64753574e-01 7.63178766e-02
5.90139151e-01 3.64114434e-01 -2.18097225e-01 -1.86435267e-01
-1.03242612e+00 -4.38589215e-01 2.28243306e-01 2.58509964e-01
8.04566562e-01 -1.64350674e-01 -1.29949367e+00 5.62286615e-01
-6.77969530e-02 1.92580953e-01 4.21693057e-01 1.18806326e+00
-6.29789233e-01 -1.41177070e+00 -4.64899719e-01 7.78415278e-02
-5.82471490e-01 1.49628043e-01 -6.25181615e-01 9.18139219e-01
3.19754660e-01 1.30943251e+00 -4.51396815e-02 -2.29288280e-01
2.45754927e-01 5.04181445e-01 5.55363536e-01 -4.55657184e-01
-6.06055796e-01 3.15641373e-01 7.36331463e-01 -4.11847323e-01
-1.15818465e+00 -9.49437678e-01 -1.24860823e+00 1.19059242e-01
-5.10394156e-01 6.59447849e-01 5.26646674e-01 1.46216595e+00
-2.95501888e-01 3.98794442e-01 5.38817465e-01 -5.04159570e-01
-7.44677475e-03 -9.21836317e-01 -2.68791884e-01 4.60960239e-01
-1.75485894e-01 -4.67437625e-01 -2.85731435e-01 5.29275119e-01] | [9.094813346862793, 9.183455467224121] |
b93d616c-a15f-43de-8e18-f06c00c2c8d6 | shot-by-shot-movie-version-comparison | null | null | https://videoprocessing.ai/other/film-comparison.html | https://storage.videoprocessing.ml/compression/video-beta/video-diff/article.pdf | Shot-by-Shot Movie Version Comparison | Some movies are released in two or more versions. One of the frequent causes is shrinking the movie for theatrical showcase with subsequent release of both the theatrical (shrunk) cut and the director’s (original, extended) cut. Automatic editing difference detection is complicated by potential changes to color gamut, aspect ratio, reduced or increased length of individual scenes, object addition or removal. These difficulties make it impossible to directly apply existing approaches to editing map construction. We propose an algorithm for fully automatic construction of an editing map of two movie versions. | ['Dmitriy Vatolin', 'Ivan Molodetskikh'] | 2018-12-01 | null | null | null | world-of-technique-of-cinema-2018-12 | ['video-alignment'] | ['computer-vision'] | [ 4.69466150e-01 -2.91386098e-01 2.82267034e-01 -2.18823239e-01
-2.11352393e-01 -1.25759304e+00 5.35266697e-01 3.20836246e-01
-1.43295795e-01 6.52387381e-01 2.23416299e-01 -9.63345692e-02
-2.19716012e-01 -4.11411673e-01 -5.46505868e-01 -7.38589019e-02
-1.02017172e-01 1.30051866e-01 5.25816977e-01 -2.00504556e-01
5.52917480e-01 4.97164458e-01 -1.54799819e+00 4.88011479e-01
5.45128167e-01 8.97954404e-01 3.31382215e-01 9.96941924e-01
-2.42303953e-01 3.95152003e-01 -7.97917902e-01 -7.78743029e-01
7.54532516e-01 -6.50748014e-01 -2.90904969e-01 5.38294971e-01
8.62693191e-01 -4.90239173e-01 -4.64900397e-02 1.21358085e+00
-3.56105939e-02 2.58302480e-01 2.89805233e-01 -1.21002090e+00
-6.01404250e-01 7.24801481e-01 -9.96968687e-01 3.88019532e-02
6.87545002e-01 -3.61188352e-01 5.95587850e-01 -7.39070654e-01
1.22789872e+00 9.10151124e-01 6.37084424e-01 -1.74047370e-02
-1.32775092e+00 -3.11420709e-01 1.25478894e-01 -1.31206498e-01
-1.36816239e+00 -4.81138647e-01 9.69802737e-01 -6.22354984e-01
5.72785258e-01 8.57320011e-01 1.03228033e+00 5.54977417e-01
8.22117254e-02 1.72213167e-01 1.25851870e+00 -4.95383561e-01
1.86902419e-01 5.19621909e-01 -3.11457932e-01 3.63277644e-01
2.62207389e-01 -4.03574526e-01 -2.40697071e-01 -3.53911430e-01
1.00686920e+00 -5.52197977e-04 -2.77534246e-01 -6.74226165e-01
-1.18987942e+00 3.66089284e-01 -7.81021714e-02 2.17400283e-01
-2.12037861e-01 -8.95018354e-02 4.41460043e-01 6.56834900e-01
3.33579093e-01 7.94399202e-01 -2.97808826e-01 -4.09425199e-01
-1.33742285e+00 5.52937686e-01 4.71935868e-01 1.28333414e+00
5.74993491e-01 -2.53514946e-01 1.67339608e-01 8.65365803e-01
-4.10339952e-01 -7.43160471e-02 -1.34004299e-02 -1.17389393e+00
3.85236144e-01 4.18223917e-01 3.95385772e-01 -1.42774510e+00
-1.87681720e-01 -7.03725144e-02 -5.18535197e-01 6.37050867e-01
4.73964423e-01 -1.67021096e-01 -3.56276512e-01 1.12133157e+00
3.32095027e-01 -3.83237153e-01 -5.79055130e-01 6.96413934e-01
3.79859656e-01 4.31663394e-01 -4.48889494e-01 -5.61149061e-01
1.20312738e+00 -8.61519039e-01 -9.85951006e-01 1.93623289e-01
1.05123423e-01 -1.38582218e+00 9.54135120e-01 9.52396929e-01
-1.48109424e+00 -5.98731279e-01 -1.16438472e+00 -9.21242982e-02
-4.47467476e-01 1.57100484e-01 4.97866601e-01 4.27183092e-01
-9.09250736e-01 9.13724840e-01 -1.30493507e-01 -4.57492381e-01
6.75816536e-02 1.21853799e-01 -6.08901858e-01 7.07669184e-02
-6.25997007e-01 7.29323447e-01 7.65340179e-02 -7.35542178e-02
-7.85175040e-02 -6.45376265e-01 -5.17646015e-01 -2.07657963e-01
4.14257497e-01 -3.63997996e-01 9.70414400e-01 -1.43121517e+00
-1.56926072e+00 1.10011244e+00 3.13269854e-01 7.66171739e-02
9.21351969e-01 -2.41362512e-01 -5.99779546e-01 1.43648582e-02
-3.85885164e-02 5.29935360e-01 9.84491527e-01 -1.29039741e+00
-6.27238989e-01 -4.42674965e-01 2.94623286e-01 3.54931712e-01
-1.72358558e-01 4.40992683e-01 -6.34767711e-01 -1.18286538e+00
2.59999961e-01 -7.14640260e-01 -3.86647210e-02 3.16234678e-01
-6.07480168e-01 5.53451180e-01 8.06885421e-01 -1.04521942e+00
1.70962954e+00 -2.23439503e+00 3.60505104e-01 9.76850986e-02
1.35598242e-01 -2.45465577e-01 -7.35045504e-03 7.20559657e-01
-2.43753955e-01 7.24305138e-02 8.30876604e-02 -2.52438098e-01
-2.37842515e-01 -2.98985898e-01 -1.00433812e-01 4.19636458e-01
-4.92587388e-01 5.73637895e-02 -5.67155004e-01 -2.71071464e-01
-1.66237131e-02 1.71685562e-01 -5.09069204e-01 3.14946808e-02
8.04040432e-02 3.43858331e-01 7.54815787e-02 5.00632226e-01
1.08256316e+00 4.80954200e-01 1.49300724e-01 -3.78585458e-01
-7.82722056e-01 1.41845241e-01 -1.68270242e+00 2.00994849e+00
6.25091568e-02 8.88609529e-01 3.78441542e-01 -1.90124810e-01
9.48987484e-01 -3.92295932e-03 4.98056740e-01 -3.85303557e-01
-1.02979973e-01 -5.64795099e-02 -1.83120668e-01 -6.32617950e-01
1.28607678e+00 -2.36308888e-01 -8.82985368e-02 5.35326660e-01
-5.91000378e-01 -4.82025474e-01 4.64316785e-01 4.07044649e-01
8.48094046e-01 2.96495438e-01 3.31921160e-01 -2.25982100e-01
1.14257531e-02 1.23128779e-01 4.48773444e-01 4.27203804e-01
-1.98022649e-01 1.07233012e+00 7.62530029e-01 -5.89847028e-01
-1.56877220e+00 -9.20880616e-01 -1.22423597e-01 1.09878230e+00
2.50583857e-01 -7.53995836e-01 -9.55101907e-01 -4.42883402e-01
-7.05514625e-02 5.32425702e-01 -8.37426543e-01 1.74863458e-01
-5.71259439e-01 -1.22411519e-01 8.45004469e-02 1.57101825e-01
1.75288081e-01 -5.15087903e-01 -6.85140431e-01 2.16558024e-01
-2.55110878e-02 -7.93853879e-01 -9.43561375e-01 -1.46648794e-01
-8.78299236e-01 -9.58872437e-01 -6.21270418e-01 -6.70294404e-01
8.86820972e-01 4.06604171e-01 9.19334173e-01 -1.13451503e-01
-4.02492136e-01 2.42723495e-01 -5.49170971e-01 -8.37940872e-02
-3.59006435e-01 -4.93283451e-01 9.88294333e-02 5.72666787e-02
-3.44062060e-01 -6.32481992e-01 -4.28256661e-01 3.66494864e-01
-1.02223933e+00 3.68350387e-01 -7.99440667e-02 8.15355107e-02
5.12753665e-01 2.20956966e-01 1.71278849e-01 -9.73635435e-01
7.81425357e-01 -5.64467497e-02 -6.42842352e-01 2.93702424e-01
-8.61315429e-02 -2.71397412e-01 5.63086927e-01 -4.60884303e-01
-1.03878093e+00 1.06125049e-01 3.69436920e-01 -3.73671889e-01
2.40075514e-02 1.53343931e-01 -1.02821611e-01 6.20739087e-02
5.85635602e-01 -2.07318753e-01 -3.24799448e-01 -8.29895496e-01
3.41315240e-01 4.41934645e-01 7.45465219e-01 -3.39442380e-02
7.94412255e-01 4.56293881e-01 -3.12096387e-01 -6.48025274e-01
-1.08548246e-01 -4.47937883e-02 -1.15181530e+00 -5.44137359e-01
7.50427246e-01 -5.33844829e-01 -1.41725674e-01 4.53737587e-01
-1.18169641e+00 -7.91599974e-02 -3.88416380e-01 3.19981158e-01
-1.89476669e-01 4.66354907e-01 -4.69868153e-01 -3.56159508e-01
1.87928066e-01 -8.29296350e-01 5.49386084e-01 3.03218246e-01
-4.80264872e-01 -6.63921297e-01 4.42070991e-01 1.93613216e-01
4.75590944e-01 5.95781922e-01 8.09916258e-01 -1.37689084e-01
-2.29270339e-01 -5.22392392e-01 -2.23399792e-02 1.86084434e-01
3.99059743e-01 8.06557357e-01 -4.15164024e-01 -2.65564978e-01
-5.68715855e-02 1.88224256e-01 2.89854705e-01 4.34931189e-01
9.27686512e-01 -3.42726916e-01 -6.14628904e-02 4.92130041e-01
1.53367364e+00 5.51832616e-01 8.31104219e-01 4.78314936e-01
2.45494246e-01 6.46418691e-01 6.61264598e-01 9.33310270e-01
-1.38158932e-01 1.14457858e+00 2.57023931e-01 6.83984980e-02
-2.13246301e-01 -2.19749883e-01 4.21207428e-01 8.12729895e-01
-3.60493094e-01 -1.06409974e-01 -3.03842872e-01 2.37354696e-01
-1.40610886e+00 -1.04004979e+00 -5.61619222e-01 2.74916649e+00
4.97624040e-01 2.93418884e-01 1.54013127e-01 -1.42273426e-01
1.15205014e+00 1.28759414e-01 -1.59244001e-01 -8.91834199e-01
-3.68671060e-01 -3.15857708e-01 4.55596536e-01 5.16585469e-01
-1.00128293e+00 5.13498724e-01 7.74234867e+00 4.88925010e-01
-6.58771276e-01 5.45976609e-02 5.28357863e-01 -5.73083699e-01
-5.29472053e-01 3.14129889e-01 -4.69654351e-01 5.29365063e-01
1.29077286e-01 -1.79403812e-01 7.56158829e-01 6.03629410e-01
3.84998381e-01 -7.29553103e-01 -1.03828263e+00 1.06323409e+00
4.11196291e-01 -1.42163956e+00 -2.54127607e-02 -2.29962505e-02
1.07831419e+00 -7.34572828e-01 -1.88178226e-01 -1.59565002e-01
-2.16613203e-01 -3.88434112e-01 1.10145307e+00 4.76458460e-01
1.06623030e+00 -6.24814391e-01 1.83122769e-01 -2.50044078e-01
-1.17654049e+00 7.25259632e-02 -5.08186221e-01 -4.30566669e-01
4.05418068e-01 4.58058119e-01 -2.79791594e-01 2.85366386e-01
6.97882414e-01 3.76539946e-01 -8.72636616e-01 1.22716069e+00
1.84966996e-01 -9.27567780e-02 -2.54759073e-01 3.38331163e-01
-1.72765270e-01 -7.11954117e-01 1.03559518e+00 1.04266655e+00
5.44479370e-01 -5.17404862e-02 -2.62760147e-02 7.71908224e-01
-1.43908113e-01 3.02700341e-01 -6.53047383e-01 8.70898291e-02
4.27169681e-01 1.38778555e+00 -1.15009761e+00 -5.92078567e-01
-3.74181151e-01 1.50176072e+00 -7.96832070e-02 1.85368866e-01
-1.02268171e+00 -6.70403719e-01 4.07874107e-01 7.19855487e-01
3.35297018e-01 -2.07234055e-01 -3.11538368e-01 -9.90951419e-01
1.83534607e-01 -7.21123397e-01 2.47898832e-01 -1.09278703e+00
-8.48802030e-01 7.29596615e-01 2.03245729e-01 -1.48787141e+00
2.59302169e-01 -1.70016155e-01 -8.85783315e-01 4.00185853e-01
-4.08863008e-01 -7.94634640e-01 -3.08769673e-01 4.43011820e-01
7.48570800e-01 -1.58543080e-01 4.54775900e-01 4.55364347e-01
-4.04717207e-01 4.63936120e-01 4.64344352e-01 -2.68977255e-01
1.13172412e+00 -1.20273769e+00 1.61060747e-02 9.02742326e-01
-5.61981760e-02 3.17063421e-01 1.22610378e+00 -7.60013402e-01
-1.19599771e+00 -3.78233314e-01 6.96979642e-01 -3.62648904e-01
6.66797221e-01 -6.30062759e-01 -6.52453721e-01 6.36823893e-01
4.62470949e-01 -5.57532668e-01 6.79426074e-01 -1.29072666e-01
-2.95721143e-01 -2.82708079e-01 -1.12322414e+00 7.02255905e-01
9.22342122e-01 -1.20759517e-01 -4.34856713e-01 3.04149747e-01
4.24822837e-01 -3.82840335e-01 -8.79428804e-01 -1.18247457e-01
1.02817547e+00 -1.36720693e+00 5.67644179e-01 -4.62608516e-01
7.33757615e-01 -3.07063729e-01 -1.05027430e-01 -1.18550897e+00
-4.94697332e-01 -1.05978405e+00 3.82631958e-01 1.32978845e+00
3.13998997e-01 -6.31298423e-02 5.27698040e-01 8.76991451e-01
-2.83733636e-01 -3.43282968e-01 -7.12795794e-01 -6.07163250e-01
-3.30388397e-01 -1.85929269e-01 4.01661336e-01 1.21805441e+00
3.98457617e-01 9.91464704e-02 -7.19554067e-01 -2.91109502e-01
3.39259893e-01 3.17786723e-01 1.00720310e+00 -1.21464396e+00
-5.37589848e-01 -7.08619475e-01 -2.79953003e-01 -7.92239368e-01
-9.50154424e-01 -2.75810838e-01 -3.23326737e-01 -1.42606854e+00
4.22058165e-01 -1.51893079e-01 8.73103887e-02 4.59776707e-02
1.49578422e-01 3.92172635e-01 5.66466570e-01 1.91655129e-01
-5.83059847e-01 -1.63829643e-02 1.33478522e+00 3.31078500e-01
-6.45137966e-01 -2.08312094e-01 -5.71977735e-01 8.26838195e-01
6.35973215e-01 -5.86328030e-01 -3.23393226e-01 -3.83554518e-01
8.54165018e-01 4.60826829e-02 -5.12707382e-02 -1.01084149e+00
1.82019368e-01 -2.39984468e-01 3.68574858e-01 -6.64870620e-01
1.42024606e-01 -8.21854651e-01 9.34260070e-01 1.44578174e-01
-5.06613016e-01 7.67981708e-01 3.56098533e-01 3.63191515e-01
-1.18672084e-02 -4.48888481e-01 8.46525908e-01 -2.57112384e-01
-4.27605987e-01 -1.91291168e-01 -6.22340381e-01 -3.89008820e-01
1.25281322e+00 -9.29563344e-01 -8.06881338e-02 -5.01090467e-01
-9.93409097e-01 -3.84810567e-01 1.10291862e+00 6.00111425e-01
3.96634668e-01 -1.40571773e+00 -5.82850754e-01 1.55767471e-01
-1.14687636e-01 -4.79127496e-01 5.06194890e-01 1.01369715e+00
-1.17353427e+00 -1.52211443e-01 -7.90018916e-01 9.48285758e-02
-1.63500142e+00 7.58969128e-01 -7.38392621e-02 -7.80511424e-02
-5.64117551e-01 7.57792711e-01 2.03012407e-01 -6.16632635e-04
1.45093769e-01 -3.50197375e-01 -3.25005911e-02 3.67167175e-01
8.00876081e-01 7.71818101e-01 -1.38675636e-02 -5.87622285e-01
-1.24958130e-02 7.24808455e-01 -2.11555019e-01 -2.01367468e-01
1.31060529e+00 -5.44633806e-01 -3.09261501e-01 5.99902689e-01
1.06214988e+00 6.99509978e-01 -1.28589356e+00 2.48453021e-01
-6.54666066e-01 -9.68988359e-01 -2.64703214e-01 -9.28576350e-01
-9.30651486e-01 4.69326288e-01 4.32375163e-01 5.88858008e-01
1.23472846e+00 -1.78064615e-01 6.63714945e-01 -2.16295384e-03
3.07186991e-01 -1.52204418e+00 -9.47922021e-02 3.83567326e-02
1.42826617e+00 -7.70281017e-01 6.42530799e-01 -4.58257049e-01
-7.09494948e-01 1.29867542e+00 4.04286742e-01 -1.86771780e-01
4.06837672e-01 3.24350059e-01 -1.96911559e-01 -3.34456861e-01
-6.34927273e-01 5.42091250e-01 1.90767348e-01 3.87166440e-01
4.39746201e-01 5.74955493e-02 -6.11893773e-01 5.06748319e-01
-2.89936006e-01 -2.95739472e-01 1.15610242e+00 1.06423891e+00
-5.55305839e-01 -1.04046679e+00 -7.00237453e-01 4.11470294e-01
-2.76604831e-01 1.38248904e-02 -8.86003733e-01 8.52387905e-01
5.80629706e-01 5.44441283e-01 3.86811674e-01 -4.82713580e-01
5.42050302e-01 -1.52311131e-01 6.31264925e-01 -3.75506967e-01
-7.99130917e-01 5.11680067e-01 1.65368825e-01 -4.05650020e-01
-2.25966111e-01 -9.11117315e-01 -8.30535710e-01 -5.99761844e-01
-2.69774526e-01 -7.36751854e-02 7.76468754e-01 2.43487820e-01
5.14596462e-01 1.65961638e-01 6.41035318e-01 -8.97591412e-01
8.76131579e-02 -8.38968039e-01 -1.16524112e+00 8.54304850e-01
-1.10509969e-01 -4.69064176e-01 -1.86356187e-01 6.56648695e-01] | [11.040456771850586, -0.9231205582618713] |
603aa6fc-6d2f-4397-ac45-ca9757154282 | emotion-controllable-generalized-talking-face | 2205.01155 | null | https://arxiv.org/abs/2205.01155v1 | https://arxiv.org/pdf/2205.01155v1.pdf | Emotion-Controllable Generalized Talking Face Generation | Despite the significant progress in recent years, very few of the AI-based talking face generation methods attempt to render natural emotions. Moreover, the scope of the methods is majorly limited to the characteristics of the training dataset, hence they fail to generalize to arbitrary unseen faces. In this paper, we propose a one-shot facial geometry-aware emotional talking face generation method that can generalize to arbitrary faces. We propose a graph convolutional neural network that uses speech content feature, along with an independent emotion input to generate emotion and speech-induced motion on facial geometry-aware landmark representation. This representation is further used in our optical flow-guided texture generation network for producing the texture. We propose a two-branch texture generation network, with motion and texture branches designed to consider the motion and texture content independently. Compared to the previous emotion talking face methods, our method can adapt to arbitrary faces captured in-the-wild by fine-tuning with only a single image of the target identity in neutral emotion. | ['Brojeshwar Bhowmick', 'Ravindra Yadav', 'Sandika Biswas', 'Sanjana Sinha'] | 2022-05-02 | null | null | null | null | ['texture-synthesis', 'talking-face-generation'] | ['computer-vision', 'computer-vision'] | [ 1.99742347e-01 4.36865628e-01 2.71178633e-01 -5.69929004e-01
-2.48128489e-01 -4.05058384e-01 6.57716870e-01 -7.75722802e-01
1.88097030e-01 6.42452538e-01 1.78066209e-01 2.16712609e-01
1.90960079e-01 -9.61603701e-01 -5.36582947e-01 -8.24217558e-01
1.93823636e-01 4.06677783e-01 -7.17230067e-02 -5.61520159e-01
-3.03254239e-02 7.95899034e-01 -1.84428954e+00 3.46367121e-01
4.64794219e-01 1.23776114e+00 -1.61577046e-01 6.67532980e-01
-2.94589549e-01 8.61308932e-01 -5.03359973e-01 -4.94214147e-01
2.03797996e-01 -9.15879965e-01 -8.60310495e-01 4.92962807e-01
6.22227132e-01 -2.97866911e-01 -2.14362875e-01 7.05532610e-01
7.36208022e-01 3.16579700e-01 7.04277396e-01 -1.46459138e+00
-8.25203419e-01 -1.87910367e-02 -4.32149440e-01 -3.14284235e-01
4.66001898e-01 7.86587521e-02 4.75680828e-01 -9.61682379e-01
1.24556375e+00 1.58442497e+00 4.40632313e-01 1.34836817e+00
-1.05614614e+00 -5.46956956e-01 2.13409573e-01 1.02504544e-01
-1.27368677e+00 -8.80496800e-01 1.42026746e+00 -3.12398970e-01
5.68011880e-01 9.42610204e-03 8.29343259e-01 1.32980335e+00
-6.39097951e-03 3.89317304e-01 1.06055522e+00 -3.53315353e-01
4.18321729e-01 -5.44528663e-02 -9.51028705e-01 1.07587302e+00
-4.43078011e-01 1.59908429e-01 -7.03754663e-01 7.45521933e-02
9.59642231e-01 -3.88497502e-01 -2.35875830e-01 -5.54983437e-01
-8.61062706e-01 9.31716144e-01 3.66246492e-01 2.21268237e-01
-1.98708370e-01 3.51302713e-01 2.33229116e-01 3.04220557e-01
9.06250060e-01 2.57646233e-01 -1.87658578e-01 -2.73631196e-02
-1.05119646e+00 1.06615230e-01 9.00074720e-01 7.92836368e-01
9.71585691e-01 5.63862801e-01 -2.40201592e-01 7.22301424e-01
1.41852051e-01 5.11947811e-01 3.33719462e-01 -1.18439043e+00
-2.72161901e-01 4.50467765e-01 -1.29124880e-01 -1.32348371e+00
-4.27262366e-01 1.28460050e-01 -8.51532638e-01 5.09924233e-01
1.79205090e-01 -4.78746563e-01 -9.96891081e-01 2.03884482e+00
7.20884562e-01 4.18487430e-01 -1.22756343e-02 1.03359652e+00
1.09992361e+00 6.07996643e-01 -7.93177411e-02 -3.40690732e-01
1.05212748e+00 -8.83071959e-01 -9.26505864e-01 1.04928546e-01
4.92458999e-01 -6.97686017e-01 9.00048494e-01 3.32877010e-01
-9.89632964e-01 -5.56948483e-01 -7.17895210e-01 1.35768228e-03
-4.18591499e-01 1.37477368e-03 7.72953272e-01 8.07339370e-01
-1.53063965e+00 4.00555819e-01 -3.81441176e-01 -6.06956422e-01
5.86377144e-01 4.15851086e-01 -5.73901296e-01 9.89912078e-02
-1.04842401e+00 6.84039414e-01 -1.27174020e-01 2.17817768e-01
-9.59236085e-01 -4.19244766e-01 -9.59231079e-01 -1.45793512e-01
2.80722558e-01 -9.65178847e-01 8.84041429e-01 -1.65998805e+00
-2.35643435e+00 9.35844302e-01 -3.91564518e-01 2.21943572e-01
5.04394531e-01 3.38705420e-01 -2.89323390e-01 5.45290232e-01
-8.53649229e-02 1.27428377e+00 1.27545011e+00 -1.43105412e+00
-6.80304170e-02 -3.53997618e-01 -3.00545599e-02 1.51383266e-01
-2.94522256e-01 -5.68564199e-02 -3.51017326e-01 -8.16027641e-01
-2.32351005e-01 -8.90593410e-01 -1.79342732e-01 5.00873864e-01
-2.50026107e-01 -1.84202846e-03 1.32079792e+00 -1.47028133e-01
6.52016282e-01 -2.00581789e+00 4.35443848e-01 7.29698874e-03
9.47422385e-02 1.92553494e-02 -6.25546873e-01 2.92851508e-01
-3.44922133e-02 8.65542218e-02 -1.55913144e-01 -5.17224908e-01
-1.77207306e-01 1.52703673e-01 -2.49038890e-01 4.13386077e-01
6.27411127e-01 1.07739389e+00 -9.91661251e-01 -6.45709455e-01
2.96737701e-01 9.76094961e-01 -7.77734578e-01 2.94233859e-01
-4.67951149e-01 9.25222039e-01 -5.18207967e-01 5.62447548e-01
8.19310904e-01 1.97136849e-01 1.65520668e-01 -2.20775098e-01
7.89173320e-02 -4.60765034e-01 -8.52425039e-01 1.91269696e+00
-6.98776901e-01 5.53044081e-01 2.84733564e-01 -9.33663070e-01
1.22100818e+00 5.44101715e-01 6.68770075e-01 -5.42023718e-01
4.15320247e-01 8.89317170e-02 -3.78067195e-01 -6.79627538e-01
2.24450126e-01 -5.31954110e-01 1.50896057e-01 3.76114964e-01
3.61069858e-01 -6.14720106e-01 -9.61223915e-02 -4.37998958e-03
7.58430243e-01 5.09108067e-01 -2.11387724e-01 -1.39733300e-01
7.65429378e-01 -3.73069465e-01 5.17081261e-01 4.64439392e-02
-1.84605226e-01 9.37136531e-01 4.94647950e-01 -7.11619675e-01
-8.71621966e-01 -7.54707456e-01 1.04059041e-01 1.07599020e+00
2.29935482e-01 -1.42636746e-01 -1.01863742e+00 -7.91068673e-01
-3.33276629e-01 3.15131009e-01 -9.68370736e-01 -2.74128616e-01
-4.86031413e-01 -3.09457511e-01 5.00539839e-01 1.78612098e-01
6.41570032e-01 -1.42239547e+00 -5.42304039e-01 1.38251677e-01
-3.39653999e-01 -1.34775007e+00 -4.16236907e-01 -4.46460754e-01
-4.51976836e-01 -8.77638757e-01 -9.33071017e-01 -8.10453594e-01
6.62688017e-01 -1.76621880e-02 8.95997405e-01 2.05520853e-01
-5.78111231e-01 7.31584251e-01 -4.66070801e-01 -2.33858645e-01
-2.51885593e-01 -2.82918721e-01 3.93634662e-03 9.17421103e-01
-3.12934965e-02 -7.26814985e-01 -7.00578630e-01 2.41428539e-01
-1.04130328e+00 7.99733922e-02 1.43128157e-01 8.67080510e-01
2.25874633e-01 -2.51410067e-01 7.92986989e-01 -5.77873945e-01
4.54503894e-01 -2.66973436e-01 -1.59470513e-01 1.19889371e-01
3.87150852e-04 8.90250802e-02 7.56621420e-01 -6.59510970e-01
-1.37559760e+00 3.65281910e-01 -2.18667582e-01 -6.51232660e-01
-4.59811926e-01 -9.45415944e-02 -4.17499989e-01 -5.92836082e-01
4.64533389e-01 9.16715041e-02 1.95968136e-01 2.90766302e-02
7.29398668e-01 3.30362588e-01 2.87860751e-01 -6.53549492e-01
6.78108811e-01 9.80755389e-01 3.81323367e-01 -1.12572443e+00
-6.05906665e-01 6.28108531e-02 -8.38873088e-01 -7.86966741e-01
1.07929647e+00 -4.57277328e-01 -9.40622687e-01 5.32610536e-01
-1.24725771e+00 -5.00254929e-01 -3.18610281e-01 4.83516306e-02
-1.18248558e+00 2.72002757e-01 -3.70280892e-01 -1.01779377e+00
-3.61146569e-01 -1.15907264e+00 1.53870392e+00 1.14501894e-01
-8.77339169e-02 -1.00870049e+00 -3.18255909e-02 1.24143250e-01
6.86307073e-01 7.10380912e-01 5.92719853e-01 9.91385952e-02
-4.75049496e-01 1.78952605e-01 -7.98288137e-02 8.69248062e-02
2.60502458e-01 2.48188153e-01 -1.19847417e+00 -7.41542652e-02
-1.09042913e-01 -6.86829805e-01 6.13083959e-01 4.05443758e-01
1.08244693e+00 -3.63132149e-01 -1.26719266e-01 8.78527284e-01
1.17186451e+00 1.60113603e-01 6.60378218e-01 -2.18613088e-01
7.38952160e-01 1.03648460e+00 2.75564551e-01 5.13614535e-01
1.12106174e-01 9.74005580e-01 4.82137531e-01 -3.48574668e-01
-4.26995486e-01 -1.63923562e-01 3.81787956e-01 3.59449714e-01
-3.93230945e-01 -3.22317451e-01 -4.20072198e-01 4.91109312e-01
-1.67213416e+00 -1.14764690e+00 3.56187612e-01 1.71891463e+00
4.93151605e-01 -5.79352736e-01 1.62708703e-02 -8.83626640e-02
5.88926971e-01 2.77273953e-01 -4.87212032e-01 -6.78154051e-01
-2.82041341e-01 5.47372341e-01 -1.78462967e-01 5.65728545e-01
-8.38289261e-01 1.47844946e+00 6.06299782e+00 7.35708654e-01
-1.65199733e+00 -3.73073705e-02 9.04955685e-01 -1.40887097e-01
-3.76139015e-01 -1.41172856e-01 -4.11875367e-01 7.44828209e-02
7.04308987e-01 -7.06426799e-02 5.26889563e-01 6.99988842e-01
3.17647696e-01 1.59971669e-01 -8.38523686e-01 1.12313855e+00
6.09201968e-01 -1.26184595e+00 5.03022254e-01 -1.23131216e-01
8.73342335e-01 -6.22045398e-01 2.83141017e-01 -8.73617828e-02
1.72698125e-02 -1.23945773e+00 6.95895970e-01 7.57642448e-01
1.27714610e+00 -8.73959124e-01 3.43322873e-01 -2.59221662e-02
-1.30405641e+00 7.26757348e-02 -2.55997241e-01 1.12843849e-01
3.22830647e-01 3.24497342e-01 -6.26990378e-01 5.02920389e-01
5.84491014e-01 6.17340744e-01 -2.83743620e-01 3.58576536e-01
-9.38278064e-03 2.35541046e-01 -1.58926859e-01 -2.18599699e-02
2.24061728e-01 -2.53509402e-01 4.39714313e-01 1.00941253e+00
5.63204050e-01 1.60504088e-01 1.36692837e-01 9.65901256e-01
-1.72985017e-01 3.67421836e-01 -9.50121045e-01 -1.03669241e-02
1.61813218e-02 1.57834864e+00 -7.77422547e-01 -2.59587318e-01
-2.38386810e-01 1.41903567e+00 3.14606965e-01 6.11199498e-01
-7.36254811e-01 -3.31849575e-01 7.63861358e-01 1.15145192e-01
2.79223710e-01 -4.98360768e-02 1.54449895e-01 -1.17844260e+00
-2.49774739e-01 -7.29138732e-01 -1.62946135e-01 -9.85087931e-01
-1.23503625e+00 1.11646962e+00 -3.63972515e-01 -9.40377474e-01
-3.77689272e-01 -6.48443699e-01 -7.29094863e-01 7.49917567e-01
-1.48670805e+00 -1.60949981e+00 -5.93231261e-01 1.04310477e+00
5.28532922e-01 -1.92114457e-01 9.18764055e-01 1.06208496e-01
-2.92362332e-01 5.78890383e-01 -5.98457575e-01 1.78517491e-01
8.59199464e-01 -7.38558650e-01 1.62632912e-01 4.19353396e-01
-5.76449446e-02 2.34945804e-01 4.99703616e-01 -4.82832223e-01
-1.45786023e+00 -1.23460984e+00 6.97716951e-01 -2.42822662e-01
3.38414013e-01 -5.74034870e-01 -6.25985384e-01 4.14320499e-01
3.02811623e-01 3.57928991e-01 4.03857321e-01 -3.83404851e-01
-3.14752132e-01 -2.22573429e-01 -1.43524337e+00 7.65199721e-01
1.40253019e+00 -4.56892639e-01 -7.73544610e-02 1.41182244e-01
4.58895683e-01 -1.69057667e-01 -6.24418139e-01 6.28505707e-01
5.61062753e-01 -1.06716526e+00 7.90062428e-01 -6.80239677e-01
5.39300501e-01 -1.81193516e-01 8.78584012e-02 -1.36431932e+00
-1.40703633e-01 -1.01969290e+00 1.13463618e-01 1.18390965e+00
1.07474901e-01 -4.98880655e-01 8.94145429e-01 2.29820788e-01
6.16554264e-03 -6.13764226e-01 -1.03897166e+00 -5.04699290e-01
5.73947616e-02 -9.52555165e-02 8.20683718e-01 1.06836879e+00
-1.05490446e-01 3.94194961e-01 -6.55226648e-01 -2.72975028e-01
4.46616203e-01 1.76255509e-01 9.76770520e-01 -1.28288591e+00
4.92171012e-03 -5.04217863e-01 -5.98522663e-01 -8.09497058e-01
6.23145103e-01 -8.27115476e-01 1.59137070e-01 -1.26232100e+00
-7.30134472e-02 -2.68496424e-01 4.04866725e-01 4.38047200e-01
1.53140709e-01 6.07904911e-01 1.50127277e-01 -3.68091077e-01
-2.72588193e-01 1.13655555e+00 1.76758385e+00 4.51628724e-03
-1.38825282e-01 -4.46906507e-01 -5.39489269e-01 6.71615958e-01
6.64222002e-01 -4.07111607e-02 -6.93478048e-01 -1.32137522e-01
9.59366113e-02 2.91796088e-01 4.94174480e-01 -1.08080590e+00
3.56199741e-02 -4.20378387e-01 4.61606771e-01 -1.55890482e-02
7.04987884e-01 -7.04903543e-01 2.09396347e-01 2.51499154e-02
-2.03370646e-01 -1.81250125e-01 1.48464143e-01 5.28541923e-01
-1.86053872e-01 2.69865304e-01 1.00696659e+00 -1.50370523e-01
-5.76812148e-01 8.36489677e-01 -4.18084860e-01 1.70295127e-02
1.15732694e+00 -3.77381206e-01 -1.35832727e-01 -8.88626635e-01
-9.54893947e-01 -2.32394278e-01 5.76851666e-01 6.27974927e-01
8.15317333e-01 -1.42895412e+00 -6.85226083e-01 4.37793791e-01
-2.58449800e-02 -1.88379273e-01 4.50809449e-01 5.12492299e-01
-4.51566845e-01 8.47309083e-02 -4.86849785e-01 -4.81639802e-01
-1.08464396e+00 6.66481137e-01 6.72255278e-01 2.29097128e-01
-5.03314734e-01 8.95210087e-01 5.01032233e-01 -5.74681997e-01
-5.57554178e-02 2.15624839e-01 -1.91534236e-01 -5.90477660e-02
4.69870687e-01 1.01746218e-02 -1.60369575e-01 -1.12964368e+00
-1.59108579e-01 1.20254254e+00 5.66787660e-01 -4.44090366e-01
1.16469216e+00 -1.50172651e-01 -2.35329181e-01 3.05194616e-01
1.26603079e+00 2.27100179e-02 -1.32397354e+00 2.13096902e-01
-4.15049493e-01 -3.83807063e-01 -1.11865260e-01 -5.11879146e-01
-1.70194244e+00 1.08087754e+00 4.24146980e-01 -1.74531460e-01
1.44478059e+00 -9.69981775e-02 7.30052710e-01 4.62361798e-02
4.86104012e-01 -9.60011661e-01 5.68471611e-01 4.58524168e-01
1.14483702e+00 -9.90321517e-01 -4.76719707e-01 -6.60530984e-01
-7.32060969e-01 1.39137721e+00 8.21884751e-01 -3.01511101e-02
8.12373757e-01 2.71103710e-01 3.25479090e-01 -2.93743491e-01
-8.41780782e-01 -2.44388491e-01 2.85814136e-01 9.55706596e-01
4.89586949e-01 -2.59244055e-01 -7.57420436e-02 -1.23037724e-02
-3.11352074e-01 3.47102731e-01 4.57661480e-01 6.05582714e-01
-2.60336250e-01 -1.00064898e+00 -2.29705960e-01 -4.68144380e-02
-2.54778385e-01 1.09626450e-01 -7.24187970e-01 5.41283846e-01
2.65413344e-01 1.13014853e+00 1.77010745e-01 -3.72189999e-01
1.58112973e-01 2.26449490e-01 6.78627908e-01 -4.51502621e-01
-2.51039147e-01 -1.05087440e-02 5.08767590e-02 -8.90376627e-01
-6.90400541e-01 -1.67315200e-01 -1.25378358e+00 -4.67629611e-01
4.50298786e-02 -1.05933808e-02 5.14369130e-01 7.15964675e-01
7.11613953e-01 3.08591187e-01 8.86133552e-01 -1.36769736e+00
2.94139624e-01 -8.30072641e-01 -5.93031704e-01 7.06375718e-01
3.88852328e-01 -9.84028995e-01 -5.14737308e-01 3.39653820e-01] | [12.82950496673584, -0.22038288414478302] |
e980e299-c213-4497-a788-cc72aab71009 | learning-spatial-and-spatio-temporal-pixel | 2101.10760 | null | https://arxiv.org/abs/2101.10760v1 | https://arxiv.org/pdf/2101.10760v1.pdf | Learning Spatial and Spatio-Temporal Pixel Aggregations for Image and Video Denoising | Existing denoising methods typically restore clear results by aggregating pixels from the noisy input. Instead of relying on hand-crafted aggregation schemes, we propose to explicitly learn this process with deep neural networks. We present a spatial pixel aggregation network and learn the pixel sampling and averaging strategies for image denoising. The proposed model naturally adapts to image structures and can effectively improve the denoised results. Furthermore, we develop a spatio-temporal pixel aggregation network for video denoising to efficiently sample pixels across the spatio-temporal space. Our method is able to solve the misalignment issues caused by large motion in dynamic scenes. In addition, we introduce a new regularization term for effectively training the proposed video denoising model. We present extensive analysis of the proposed method and demonstrate that our model performs favorably against the state-of-the-art image and video denoising approaches on both synthetic and real-world data. | ['Ming-Hsuan Yang', 'Wenxiu Sun', 'Muchen Li', 'Xiangyu Xu'] | 2021-01-26 | null | null | null | null | ['video-denoising'] | ['computer-vision'] | [ 5.14974773e-01 -4.89549786e-01 4.37415212e-01 -4.06864703e-01
-1.07812512e+00 -3.15659493e-01 5.42444825e-01 -3.40823650e-01
-5.51327467e-01 6.38232887e-01 2.23714262e-01 5.17077409e-02
-7.43944272e-02 -7.98597693e-01 -9.32360709e-01 -1.09172523e+00
-6.43267436e-03 -3.23188156e-01 8.57802704e-02 -1.68388456e-01
1.78188354e-01 3.16997737e-01 -1.49344337e+00 4.45757270e-01
1.10096788e+00 1.12670279e+00 3.48868102e-01 8.53385270e-01
1.31114498e-01 7.48791516e-01 -5.15954018e-01 -1.42805710e-01
5.74240088e-01 -6.01818740e-01 -2.59808362e-01 5.77799022e-01
7.82285213e-01 -6.34693861e-01 -6.16672635e-01 1.20538974e+00
4.25079763e-01 3.18415016e-01 2.85570115e-01 -7.02085972e-01
-6.30098999e-01 2.30027229e-01 -6.70389891e-01 3.02582920e-01
5.52105680e-02 3.76515567e-01 6.37572885e-01 -7.93417335e-01
5.42482555e-01 1.31452882e+00 1.02143276e+00 4.18412060e-01
-1.37797582e+00 -3.66814554e-01 2.70377785e-01 1.87125221e-01
-1.12234437e+00 -5.47503233e-01 8.86940122e-01 -1.09512396e-01
5.40619850e-01 2.65896916e-01 5.25770128e-01 1.02246320e+00
1.81899682e-01 7.54521549e-01 1.14433825e+00 -3.23390156e-01
2.43077904e-01 -6.69995964e-01 -1.29475430e-01 2.99924076e-01
8.61075968e-02 2.11784497e-01 -4.84432042e-01 1.71309952e-02
1.10306966e+00 1.50896519e-01 -4.82558101e-01 -2.80345738e-01
-1.03040063e+00 4.23856676e-01 4.12860036e-01 2.34901682e-01
-7.41412282e-01 7.08140016e-01 3.89436632e-01 2.62346774e-01
8.15805435e-01 6.66587576e-02 -2.71178514e-01 7.20799118e-02
-1.52802670e+00 3.60854805e-01 3.32297534e-01 5.34826934e-01
7.47530103e-01 4.43574786e-01 -3.70734006e-01 1.00502551e+00
1.29956543e-01 4.32669431e-01 9.83298346e-02 -1.72732306e+00
3.78533036e-01 -1.32288948e-01 3.14690351e-01 -9.48312700e-01
6.77659437e-02 -3.83229315e-01 -1.25308263e+00 5.01709223e-01
4.32130992e-01 -5.50874770e-02 -1.14577508e+00 1.60076582e+00
1.59619719e-01 6.61577642e-01 -6.42213672e-02 9.45891798e-01
4.51707751e-01 8.38523030e-01 2.23424416e-02 -5.05037844e-01
9.05702412e-01 -1.10062051e+00 -1.10198307e+00 -1.77732557e-01
3.14628668e-02 -7.06802547e-01 8.41169715e-01 7.21409798e-01
-1.45920706e+00 -7.44937301e-01 -1.01138699e+00 -2.05227494e-01
1.50653437e-01 7.96484351e-02 3.72231394e-01 4.80650276e-01
-1.26433682e+00 1.00149512e+00 -1.16533792e+00 -1.57149389e-01
5.56242049e-01 2.13503361e-01 -2.78532773e-01 -4.80864376e-01
-8.77102911e-01 5.16419828e-01 1.59203187e-01 6.92503154e-01
-1.10830545e+00 -6.36746228e-01 -8.53482127e-01 -1.65370274e-02
3.79213214e-01 -8.71634185e-01 1.03864157e+00 -1.30248892e+00
-1.44799805e+00 5.37138760e-01 -4.85831052e-01 -5.95606446e-01
7.02489257e-01 -6.03314102e-01 -3.18229020e-01 3.77354503e-01
-3.50791402e-02 5.68916678e-01 1.39250028e+00 -1.68527412e+00
-5.08996546e-01 -1.20592915e-01 -9.57082361e-02 9.29765329e-02
-3.13254088e-01 -2.49731570e-01 -6.98347449e-01 -1.26655233e+00
3.08497667e-01 -4.55618739e-01 -5.67970991e-01 2.54582107e-01
-1.26206622e-01 4.98534828e-01 9.66634393e-01 -1.02990389e+00
1.24180472e+00 -2.18413281e+00 2.82111824e-01 5.19454032e-02
1.87216714e-01 3.38057488e-01 -3.79736453e-01 2.63656765e-01
-2.03085350e-04 -1.33873731e-01 -6.49660528e-01 -7.83768237e-01
-2.72624731e-01 6.19917274e-01 -2.31197208e-01 5.97462595e-01
7.21857324e-02 6.25520349e-01 -9.15767074e-01 -1.15786418e-01
5.74398935e-01 9.42034185e-01 -5.64940989e-01 2.43452206e-01
-9.83129889e-02 7.10182071e-01 -2.30737925e-01 4.75804418e-01
1.09733701e+00 1.66904658e-01 1.31438300e-01 -4.95764524e-01
-5.97691350e-02 -6.23431290e-04 -1.33753169e+00 1.80780435e+00
-5.02045751e-01 6.21118188e-01 7.90514886e-01 -1.12835693e+00
6.42598569e-01 3.00289303e-01 4.79785949e-01 -7.05083013e-01
-2.02344544e-02 2.74814427e-01 -4.99696642e-01 -5.61624289e-01
5.56871533e-01 -9.83702391e-02 4.53009099e-01 8.07051957e-02
3.21604051e-02 -1.02358438e-01 2.95354396e-01 1.90345570e-02
1.16701174e+00 3.08927387e-01 -4.08884622e-02 -2.62218118e-01
7.36015618e-01 -3.65646958e-01 8.14035594e-01 9.30824935e-01
-1.93528369e-01 1.04944515e+00 2.98771471e-01 -5.63866138e-01
-1.19270420e+00 -1.04404747e+00 2.70535294e-02 8.81490767e-01
1.61339119e-01 -2.26570413e-01 -9.68798220e-01 -2.88171172e-01
-2.06848443e-01 4.95414346e-01 -5.40511966e-01 1.21782891e-01
-8.15933466e-01 -8.81750166e-01 2.54462183e-01 5.05226612e-01
8.55450571e-01 -8.83096516e-01 -3.33964556e-01 3.88181299e-01
-4.17059869e-01 -1.09853792e+00 -5.96809328e-01 1.19979687e-01
-1.22475290e+00 -8.44405353e-01 -6.99530303e-01 -6.78429663e-01
6.67008221e-01 6.01812124e-01 1.13624632e+00 2.52825141e-01
-1.32806987e-01 4.72468704e-01 -2.33794957e-01 1.80305094e-01
-3.76611799e-01 -3.53118062e-01 -2.22043768e-01 2.92159617e-01
-9.05534253e-03 -8.79096508e-01 -1.01963270e+00 8.20158869e-02
-1.47454917e+00 -3.37116383e-02 4.27040517e-01 1.06451833e+00
7.11357653e-01 3.77006501e-01 2.68357396e-01 -7.67532110e-01
6.63882911e-01 -1.37807041e-01 -5.85002720e-01 6.64347857e-02
-2.66433179e-01 6.13353439e-02 5.52999854e-01 -2.96351224e-01
-1.28547502e+00 1.57811716e-01 -2.92520523e-01 -5.90562344e-01
-1.90289930e-01 2.58691341e-01 -1.42096892e-01 -2.38572255e-01
4.51096594e-01 2.41072521e-01 1.24058165e-01 -7.42528200e-01
4.61544812e-01 2.54963338e-01 1.01270390e+00 -5.25427520e-01
8.17679644e-01 1.03218782e+00 -8.23504031e-02 -8.62892687e-01
-6.09566152e-01 -3.23554575e-01 -6.53552413e-01 -2.24459693e-01
9.27638829e-01 -1.18321776e+00 -4.29186910e-01 9.18730319e-01
-1.23741543e+00 -5.79702199e-01 -2.18803197e-01 1.97401866e-01
-6.45688117e-01 8.04002285e-01 -1.01376688e+00 -7.73793340e-01
-2.60037929e-01 -1.18148482e+00 1.23405278e+00 -6.12972341e-02
1.42165288e-01 -1.11416125e+00 -1.04421221e-01 2.08668515e-01
5.39420187e-01 3.84456694e-01 4.59464401e-01 3.27799469e-01
-8.98990571e-01 -8.90094507e-03 -2.54988402e-01 8.68270934e-01
1.81819975e-01 -1.08675715e-02 -1.01984191e+00 -3.42015773e-01
3.59752476e-01 2.02996004e-02 1.49614084e+00 9.81979370e-01
1.37104189e+00 -2.51761407e-01 4.54778261e-02 9.99359906e-01
1.64812005e+00 -1.40718296e-01 1.05594695e+00 4.85077828e-01
7.80589700e-01 4.60164934e-01 4.07756478e-01 4.13497955e-01
5.91630228e-02 5.39535463e-01 6.65777266e-01 -2.76730001e-01
-3.21197271e-01 1.07319228e-01 3.56133968e-01 5.26117265e-01
-3.42758894e-01 -2.69546300e-01 -4.85859036e-01 6.11507893e-01
-2.01153541e+00 -1.07005692e+00 -2.49124989e-01 2.01505232e+00
8.14354241e-01 -4.51474413e-02 -3.40019614e-01 1.58028767e-01
6.20792150e-01 5.84843278e-01 -2.69613713e-01 -1.04428224e-01
-4.70613718e-01 3.41988176e-01 6.70738339e-01 8.23594511e-01
-1.42982948e+00 7.47810245e-01 7.00309944e+00 9.23031926e-01
-9.33993518e-01 1.15530342e-01 8.16636264e-01 -1.75376132e-01
-3.08468044e-01 -2.65438557e-01 -2.09447443e-01 3.66851211e-01
6.36637747e-01 2.89634526e-01 6.54576004e-01 3.65154415e-01
8.12869787e-01 -2.22674236e-01 -8.47265244e-01 1.08177674e+00
-9.02707428e-02 -1.44424188e+00 2.01479942e-02 -1.32070914e-01
1.03088880e+00 -2.46781692e-01 9.16640833e-02 -1.91776589e-01
1.15329869e-01 -1.07424891e+00 7.04907894e-01 7.46823132e-01
4.23765302e-01 -5.72904468e-01 7.25560963e-01 8.65702927e-02
-1.07042670e+00 -1.07203275e-01 -2.78526157e-01 -1.83804885e-01
5.92637539e-01 9.98815954e-01 1.58178195e-01 5.13700485e-01
9.73437488e-01 1.13852036e+00 -4.43326116e-01 1.02429771e+00
-3.11168551e-01 5.80001116e-01 -2.43268311e-01 8.43587637e-01
3.43302220e-01 -6.56155527e-01 5.60308099e-01 1.18343711e+00
5.17533898e-01 1.42348200e-01 -5.10597080e-02 5.88350117e-01
-8.27869624e-02 -4.85750645e-01 -3.55216593e-01 2.41827831e-01
4.18857746e-02 1.13932598e+00 -4.81881857e-01 -4.64933336e-01
-4.57754135e-01 1.30910671e+00 -5.03582172e-02 9.71200466e-01
-6.91286027e-01 4.67481976e-03 9.27652359e-01 1.30113378e-01
7.35475659e-01 -4.55032349e-01 -6.10749185e-01 -1.15278447e+00
2.29678869e-01 -9.39782023e-01 1.35308355e-01 -9.04821277e-01
-1.40509832e+00 5.76116621e-01 -2.47673050e-01 -1.19422269e+00
-5.45795402e-03 -4.64384764e-01 -7.38424063e-01 8.24869394e-01
-1.65687811e+00 -9.94300187e-01 -5.31484067e-01 6.12545311e-01
6.41882718e-01 2.36690536e-01 3.73820901e-01 5.88702083e-01
-3.86457384e-01 8.20661187e-02 5.65404832e-01 -9.88932997e-02
8.46043587e-01 -1.04561651e+00 5.45017183e-01 1.40900624e+00
-1.45080954e-01 4.94273603e-01 1.09181917e+00 -4.85964119e-01
-1.21242189e+00 -1.15613270e+00 3.50676566e-01 9.92084853e-03
4.82329339e-01 -1.47450015e-01 -1.20273602e+00 6.06481791e-01
6.56451344e-01 3.81480157e-01 1.04483560e-01 -3.54539633e-01
-2.37476334e-01 -4.95898277e-01 -1.11798799e+00 6.01612449e-01
1.03121746e+00 -3.79818588e-01 -2.90263116e-01 1.52797684e-01
5.12990952e-01 -4.14892316e-01 -5.59193492e-01 4.01137739e-01
3.68520975e-01 -1.36564684e+00 1.40895486e+00 -2.18966827e-01
6.78595066e-01 -5.33930659e-01 -1.81919709e-01 -1.33587110e+00
-3.49032849e-01 -6.58539176e-01 -2.36289173e-01 1.02074313e+00
-5.00587970e-02 -3.01009059e-01 7.32205868e-01 5.15122473e-01
-2.50996351e-01 -4.19797152e-01 -9.62705493e-01 -5.23214817e-01
-1.70426562e-01 -5.36088407e-01 6.89990819e-02 6.29092872e-01
-8.48988116e-01 -2.54853606e-01 -7.81704843e-01 3.98532599e-01
1.29349303e+00 -2.36876801e-01 6.61812663e-01 -7.56140172e-01
-3.42147321e-01 -1.87154442e-01 -2.12928131e-01 -1.36195159e+00
-6.67577088e-02 -1.04767591e-01 4.00733382e-01 -1.67344165e+00
-4.13705967e-02 4.38640304e-02 -2.97409415e-01 1.85352042e-01
-4.12532657e-01 7.28127599e-01 7.62664452e-02 1.45956531e-01
-5.31303346e-01 6.06618047e-01 1.38105202e+00 -1.73377156e-01
-1.05000407e-01 -1.68949142e-01 -4.15958673e-01 7.31982887e-01
5.93389094e-01 -2.54506141e-01 -3.22508305e-01 -9.13281977e-01
-5.16371317e-02 -1.07493564e-01 7.68669248e-01 -1.08399153e+00
1.71813786e-01 -1.05902553e-01 5.28302729e-01 -5.35111785e-01
4.11294550e-01 -9.96697962e-01 2.67988443e-01 2.36036122e-01
-1.75126001e-01 -2.32478485e-01 1.82856232e-01 8.66043150e-01
-5.70866346e-01 -3.60854045e-02 1.06924188e+00 -2.71945357e-01
-8.76412690e-01 1.79291070e-01 -5.07883012e-01 -2.44230181e-01
7.57286370e-01 -4.11219090e-01 -1.24999449e-01 -6.36578619e-01
-9.05927122e-01 9.92388576e-02 5.45289636e-01 6.01749159e-02
6.20052636e-01 -1.29688597e+00 -6.26851499e-01 2.37803325e-01
-4.48673606e-01 2.03562644e-03 6.58025265e-01 9.15710449e-01
-9.50861514e-01 -1.81487009e-01 -1.29538015e-01 -6.86635256e-01
-1.21320248e+00 4.42876250e-01 4.22237933e-01 -2.59671926e-01
-7.66620159e-01 7.59584367e-01 2.41429657e-01 -1.46363288e-01
2.61400133e-01 -4.51628119e-01 2.25238711e-01 -1.75318763e-01
6.47232950e-01 4.74377692e-01 4.31743152e-02 -4.95720118e-01
3.10336817e-02 6.29850745e-01 2.06920326e-01 -1.56082481e-01
1.70805335e+00 -5.97781718e-01 -3.73756826e-01 1.07160054e-01
1.15302503e+00 -1.49627775e-01 -1.99722755e+00 -2.07715601e-01
-3.02117914e-01 -7.39400327e-01 3.77867937e-01 -4.48021412e-01
-1.49411285e+00 8.16076040e-01 6.09134853e-01 1.14477314e-01
1.70932889e+00 -6.38027251e-01 1.05623031e+00 1.15762517e-01
9.45795700e-02 -1.24572551e+00 1.44377261e-01 3.52043837e-01
7.75157213e-01 -1.23376620e+00 2.48272389e-01 -6.52184844e-01
-3.46488059e-01 1.20000386e+00 2.86026180e-01 -4.62989300e-01
5.59680581e-01 4.13084179e-01 2.96267807e-01 1.24118313e-01
-5.39563239e-01 -7.59030804e-02 1.06048755e-01 6.50478244e-01
3.89625162e-01 -3.99960697e-01 -3.49292785e-01 1.32619843e-01
4.65453476e-01 1.53483436e-01 4.40480918e-01 9.19258356e-01
-5.13705969e-01 -1.22807956e+00 -7.87701309e-01 1.88651234e-01
-5.57316601e-01 -2.30854779e-01 2.28336439e-01 4.62788880e-01
1.79940656e-01 9.09924924e-01 2.00008750e-02 4.55441177e-02
3.23991179e-01 -1.56474531e-01 4.79418457e-01 -9.06945020e-02
-5.47519982e-01 6.50671005e-01 -1.08664207e-01 -8.86648536e-01
-9.79267299e-01 -6.11054003e-01 -7.45610058e-01 -3.35643411e-01
1.05975732e-01 -2.19017237e-01 2.95805961e-01 7.76267648e-01
2.71628618e-01 8.72819245e-01 4.10667121e-01 -1.46003652e+00
-3.22705030e-01 -8.26927662e-01 -7.41567373e-01 7.02048421e-01
7.16254652e-01 -3.24747086e-01 -5.53372025e-01 5.76381981e-01] | [11.386716842651367, -2.2272017002105713] |
1c80b0ac-5111-41b1-a0bd-7db8f44d50be | knowledge-distillation-for-small-footprint | 1608.00892 | null | http://arxiv.org/abs/1608.00892v3 | http://arxiv.org/pdf/1608.00892v3.pdf | Knowledge Distillation for Small-footprint Highway Networks | Deep learning has significantly advanced state-of-the-art of speech
recognition in the past few years. However, compared to conventional Gaussian
mixture acoustic models, neural network models are usually much larger, and are
therefore not very deployable in embedded devices. Previously, we investigated
a compact highway deep neural network (HDNN) for acoustic modelling, which is a
type of depth-gated feedforward neural network. We have shown that HDNN-based
acoustic models can achieve comparable recognition accuracy with much smaller
number of model parameters compared to plain deep neural network (DNN) acoustic
models. In this paper, we push the boundary further by leveraging on the
knowledge distillation technique that is also known as {\it teacher-student}
training, i.e., we train the compact HDNN model with the supervision of a high
accuracy cumbersome model. Furthermore, we also investigate sequence training
and adaptation in the context of teacher-student training. Our experiments were
performed on the AMI meeting speech recognition corpus. With this technique, we
significantly improved the recognition accuracy of the HDNN acoustic model with
less than 0.8 million parameters, and narrowed the gap between this model and
the plain DNN with 30 million parameters. | ['Steve Renals', 'Liang Lu', 'Michelle Guo'] | 2016-08-02 | null | null | null | null | ['acoustic-modelling'] | ['speech'] | [ 2.29465559e-01 2.35042065e-01 2.23604605e-01 -4.86815035e-01
-7.21460402e-01 -2.46614695e-01 3.08532357e-01 -4.60961968e-01
-7.98440754e-01 3.49058032e-01 -1.02402873e-01 -9.40168977e-01
1.64681524e-01 -4.46134001e-01 -7.73979545e-01 -7.77729273e-01
2.83353508e-01 5.38255334e-01 1.76740587e-01 1.09432034e-01
-3.14264864e-01 4.32966709e-01 -1.49712563e+00 -4.31128889e-02
5.85849285e-01 1.00589764e+00 4.46603894e-01 1.03426230e+00
-1.50620192e-01 6.85769498e-01 -1.02878666e+00 -3.43004286e-01
-1.99174896e-01 -1.91374183e-01 -4.71459597e-01 -1.68647170e-01
3.67252499e-01 -4.58978176e-01 -7.54839361e-01 8.70883703e-01
8.91549945e-01 4.58770633e-01 5.01161933e-01 -9.96250272e-01
-6.34935856e-01 1.01486611e+00 -1.47806481e-01 2.17492402e-01
-3.66807878e-01 -1.48861080e-01 5.66305578e-01 -1.02759111e+00
-2.84457922e-01 1.26290238e+00 7.15722084e-01 1.02339125e+00
-9.18074429e-01 -7.86628425e-01 1.04993053e-01 3.51888537e-01
-1.41448629e+00 -9.97471213e-01 5.57781398e-01 -1.99576989e-01
1.41872025e+00 1.88226625e-01 5.21968961e-01 1.40489733e+00
-4.37120140e-01 1.06253004e+00 5.39823830e-01 -8.01279068e-01
3.75426173e-01 6.05860241e-02 2.53424853e-01 3.97608340e-01
-1.18251726e-01 1.78677380e-01 -4.09257829e-01 -1.38442218e-01
7.25780308e-01 -5.46570867e-02 -7.28109553e-02 4.43836376e-02
-7.84055829e-01 8.14869165e-01 -1.11192971e-01 4.22621340e-01
-1.49878129e-01 4.10907269e-01 3.59715939e-01 5.30454814e-02
4.76391375e-01 -1.11946896e-01 -6.36263609e-01 -7.67868817e-01
-9.38035131e-01 -3.19128811e-01 1.10544086e+00 7.40713537e-01
5.25201917e-01 8.54927361e-01 4.02649373e-01 1.48115253e+00
3.31504673e-01 5.60457289e-01 9.58214223e-01 -7.53208339e-01
3.47520292e-01 -5.66889346e-02 -5.05230010e-01 -3.10358316e-01
-2.00954914e-01 -6.14692569e-01 -1.11636567e+00 -1.10880949e-01
1.73542902e-01 -4.59026515e-01 -1.35896730e+00 1.69342339e+00
1.75425336e-01 5.94283640e-01 3.22198421e-01 4.80906010e-01
8.56571555e-01 9.82530475e-01 -2.11638398e-02 -5.71536720e-02
9.44575071e-01 -1.26029265e+00 -8.58617783e-01 -3.49208862e-01
9.26159859e-01 -6.53282642e-01 9.77469146e-01 6.78085148e-01
-1.05015206e+00 -9.54183578e-01 -1.14506352e+00 1.71725154e-01
-3.72499228e-01 2.15204179e-01 5.20798802e-01 1.24607551e+00
-1.34185398e+00 2.93934613e-01 -1.12392175e+00 -1.18913695e-01
1.98277175e-01 6.24517143e-01 -9.04219598e-02 6.83975965e-02
-1.13356376e+00 8.07009339e-01 5.42884886e-01 3.64851534e-01
-1.07393241e+00 -6.03959382e-01 -8.37806106e-01 2.74283767e-01
2.96406806e-01 -3.76330674e-01 1.80785811e+00 -8.27026904e-01
-2.17556930e+00 3.43659818e-01 -8.89616087e-02 -7.23988116e-01
-8.80185142e-02 -3.47742856e-01 -7.08450139e-01 -2.29753569e-01
-7.36243844e-01 7.34943092e-01 6.80726826e-01 -9.81660545e-01
-5.91956258e-01 -1.18838139e-01 -2.70833433e-01 -3.57872508e-02
-7.17469692e-01 1.61296502e-02 -4.53759581e-01 -6.96263015e-01
7.47103766e-02 -9.48924124e-01 -3.55816692e-01 -7.24006414e-01
-1.55078486e-01 -5.36021948e-01 9.58217263e-01 -6.25925899e-01
1.42621768e+00 -2.34041643e+00 -1.05833739e-01 2.61778146e-01
-1.37056336e-02 1.00950909e+00 -2.12457120e-01 2.52719373e-01
4.15402325e-03 5.97137399e-02 -1.09988838e-01 -6.99219465e-01
1.85587406e-01 7.08600998e-01 -2.83645511e-01 6.27176687e-02
-1.95665270e-01 7.25494683e-01 -4.80206043e-01 -1.69796363e-01
1.76638275e-01 7.70865142e-01 -6.14610434e-01 4.65461463e-01
-1.32099286e-01 3.65952104e-02 -2.32805451e-03 4.19905007e-01
5.36912203e-01 1.49782777e-01 1.49096534e-01 8.67153481e-02
1.28657266e-01 5.58728158e-01 -9.85780180e-01 1.72947240e+00
-7.07306921e-01 7.55011201e-01 2.35645920e-01 -1.46336806e+00
1.12334347e+00 6.87399983e-01 1.40597820e-01 -3.75251800e-01
1.51048511e-01 4.12163436e-01 3.64696056e-01 -2.18382508e-01
3.23976755e-01 -1.47052661e-01 1.50964841e-01 4.07002926e-01
5.11304021e-01 -1.46351099e-01 -2.73954421e-01 -2.23245218e-01
8.23715031e-01 -3.29196155e-01 -1.63156644e-01 1.79737046e-01
3.32168400e-01 -5.96029162e-01 4.84209418e-01 1.00894594e+00
-1.32051378e-01 4.43959951e-01 -8.92357975e-02 -3.57586533e-01
-9.57084715e-01 -9.86613631e-01 -3.56512927e-02 1.26101947e+00
-4.91741210e-01 -4.16644961e-01 -1.13373518e+00 -3.70019257e-01
-3.49336714e-01 8.08576107e-01 -1.55816644e-01 -3.84019136e-01
-7.87987411e-01 -8.39122236e-01 1.28849244e+00 7.96990931e-01
4.55184221e-01 -1.01217699e+00 -1.67637110e-01 4.41305876e-01
1.94572106e-01 -1.37329030e+00 -2.85614371e-01 6.15628004e-01
-8.02456141e-01 -3.23468179e-01 -8.32163036e-01 -1.08454585e+00
2.35225976e-01 -2.15835515e-02 9.53625083e-01 -1.36967123e-01
1.12562016e-01 3.67515475e-01 -3.56268942e-01 -6.85527325e-01
-8.27570617e-01 2.49680310e-01 6.85193062e-01 -2.37019379e-02
6.53838217e-01 -9.02034521e-01 -1.94322720e-01 1.20138071e-01
-9.70381379e-01 -2.55710203e-02 9.06514168e-01 1.08015847e+00
4.41625148e-01 1.62480831e-01 8.56012762e-01 -5.74413240e-01
5.42727828e-01 -1.37639120e-01 -3.82103443e-01 1.74254164e-01
-5.27097404e-01 -1.53439745e-01 4.99606222e-01 -9.54109371e-01
-9.59415495e-01 1.87917321e-03 -1.01313937e+00 -7.18365550e-01
-4.03223872e-01 6.99764729e-01 -3.50773424e-01 4.77963947e-02
2.74472535e-01 4.61641133e-01 1.29317477e-01 -8.84155154e-01
3.10680628e-01 1.25573981e+00 6.19725764e-01 -5.54758489e-01
2.77270734e-01 -1.40584722e-01 -5.83604634e-01 -1.26160979e+00
-5.24213433e-01 -3.13432515e-01 -4.99736577e-01 1.61024779e-01
5.74611425e-01 -9.23777282e-01 -7.56346524e-01 8.09767425e-01
-1.18709910e+00 -7.73826957e-01 -2.19635263e-01 9.47735488e-01
-3.78185302e-01 4.88383263e-01 -9.49021161e-01 -1.13278699e+00
-2.83350110e-01 -1.20282006e+00 5.27515709e-01 1.27373859e-01
-6.00517020e-02 -1.04017699e+00 2.58860383e-02 3.52710992e-01
7.34203935e-01 -7.29697347e-01 8.93954277e-01 -1.32149363e+00
-1.00234941e-01 -2.70558774e-01 2.12571084e-01 1.19495189e+00
-2.93205399e-02 -5.22492230e-02 -1.47683096e+00 -3.06937844e-01
4.00506198e-01 -2.50463933e-01 8.09328020e-01 5.34639120e-01
1.60229719e+00 -2.87149519e-01 -2.25635573e-01 5.43082416e-01
7.37165093e-01 6.60217285e-01 6.59978569e-01 1.00809470e-01
9.17724013e-01 2.59417236e-01 -4.29442041e-02 1.13780275e-02
1.57100722e-01 7.02716231e-01 7.23547041e-02 -1.96928635e-01
-1.61857069e-01 -1.77610382e-01 6.17526591e-01 1.92033827e+00
1.45797893e-01 -4.83050704e-01 -1.12107205e+00 4.85007226e-01
-1.62551057e+00 -5.69886863e-01 3.67860764e-01 1.98711765e+00
9.35379326e-01 2.91791677e-01 -3.52797024e-02 2.02590495e-01
5.91699302e-01 -5.56108318e-02 -3.17764014e-01 -6.40787065e-01
-1.06828630e-01 5.96038938e-01 2.39442587e-01 5.46492338e-01
-9.83368158e-01 9.60101485e-01 6.51607752e+00 1.60366082e+00
-1.08621871e+00 3.54171991e-01 6.98710203e-01 -1.04282610e-01
4.30666544e-02 -3.21743786e-01 -1.27444696e+00 4.30294722e-01
1.96160889e+00 3.98686111e-01 2.41147310e-01 1.02764773e+00
-4.71037142e-02 2.93409467e-01 -1.16990530e+00 1.18960285e+00
1.36432678e-01 -1.09002376e+00 1.46782577e-01 2.43277475e-01
4.63348836e-01 1.83543965e-01 2.28247911e-01 8.80321026e-01
3.88546139e-01 -1.32998466e+00 4.07777429e-01 1.05230764e-01
7.11222351e-01 -9.13238525e-01 9.60356414e-01 6.79065406e-01
-7.84237683e-01 -3.07177082e-02 -5.77785373e-01 -2.33037118e-02
2.32146338e-01 6.11213148e-01 -1.06464720e+00 1.26683325e-01
6.55502260e-01 1.69302538e-01 -1.21194787e-01 8.42567623e-01
1.28125846e-01 1.36624873e+00 -6.49126112e-01 -1.53852269e-01
3.85228813e-01 -5.42224236e-02 2.13202655e-01 1.44229341e+00
4.96659815e-01 3.15021947e-02 5.99811636e-02 4.57366675e-01
-1.08438961e-01 -2.43813053e-01 -3.40694934e-01 -2.38651976e-01
6.81783080e-01 8.19890678e-01 -1.86914548e-01 -5.87768674e-01
-4.33239877e-01 6.95109546e-01 4.12363201e-01 5.01561999e-01
-7.13150203e-01 -3.40551466e-01 6.57298505e-01 -3.57212454e-01
5.87832391e-01 -4.10491973e-01 3.61310057e-02 -7.91140616e-01
-2.27489382e-01 -9.70261753e-01 -1.04202129e-01 -5.59735835e-01
-1.11593330e+00 8.70838642e-01 -2.21066594e-01 -7.31357872e-01
-6.04763508e-01 -8.37720633e-01 -5.30875325e-01 9.35243905e-01
-1.20727265e+00 -9.46349978e-01 2.93155998e-01 3.19663167e-01
7.43932545e-01 -5.17234504e-01 1.20044816e+00 6.22800052e-01
-9.08954859e-01 1.05926907e+00 3.85922849e-01 3.83391529e-01
2.76456207e-01 -9.53150213e-01 6.71233714e-01 7.04769969e-01
4.44084883e-01 6.25772893e-01 3.30615729e-01 -1.98212847e-01
-1.26000059e+00 -8.95264268e-01 7.81944871e-01 -2.17961952e-01
6.00309908e-01 -6.60300612e-01 -1.17044926e+00 7.61148930e-01
1.27537608e-01 -1.49305567e-01 1.09068120e+00 1.73095301e-01
-2.04924986e-01 -1.24559969e-01 -5.78143299e-01 5.32997251e-01
7.91788697e-01 -8.33965421e-01 -5.75943172e-01 -6.66910857e-02
1.24897838e+00 -3.49440694e-01 -7.49853730e-01 6.29695654e-01
5.10729909e-01 -6.19123757e-01 1.07355607e+00 -4.46900785e-01
-2.12668926e-01 1.01930656e-01 -5.36027312e-01 -1.35541975e+00
-4.90285195e-02 -6.37470186e-01 -7.17996776e-01 1.36476672e+00
6.05731249e-01 -6.16455495e-01 9.43388641e-01 5.01720011e-01
-8.09487820e-01 -9.57067430e-01 -1.28019011e+00 -9.77821589e-01
2.82312810e-01 -9.05403435e-01 5.71292341e-01 6.46395862e-01
-2.19294161e-01 3.30751330e-01 -2.95314163e-01 1.61900640e-01
1.51011735e-01 -6.29606009e-01 6.86998188e-01 -9.13636863e-01
-6.28119588e-01 -5.59711397e-01 -4.40003008e-01 -1.74173379e+00
4.00454313e-01 -5.39326906e-01 2.69671977e-01 -1.07578051e+00
-1.84671789e-01 -6.86426759e-01 -4.72709179e-01 4.92878109e-01
-3.39281484e-02 1.96256135e-02 -1.64048657e-01 -7.18009546e-02
-2.76526183e-01 6.81522489e-01 6.83195531e-01 -1.42970383e-01
-1.79328531e-01 2.65338838e-01 -2.52927035e-01 7.72873163e-01
8.68773401e-01 -3.50161523e-01 -5.11407375e-01 -6.44920170e-01
-3.79827887e-01 -7.91336745e-02 5.42064384e-02 -1.12632465e+00
5.96938908e-01 3.32740575e-01 1.68012142e-01 -6.95508897e-01
8.85139227e-01 -7.08122075e-01 1.09901667e-01 3.35217625e-01
-2.50827193e-01 -2.58665949e-01 5.04281640e-01 4.22376096e-01
-3.78015250e-01 -4.26444143e-01 4.81633782e-01 8.47332850e-02
-7.60017455e-01 1.77609384e-01 -7.80586898e-01 -2.90098399e-01
3.99337023e-01 -2.05785349e-01 -6.39184117e-02 -4.16964322e-01
-9.42885756e-01 -3.23406339e-01 -2.02359363e-01 3.14708948e-01
7.51754105e-01 -1.17573559e+00 -6.11292183e-01 5.09623766e-01
-3.62902015e-01 2.87948430e-01 3.86177659e-01 6.50088072e-01
-1.72468767e-01 6.29344642e-01 4.52216774e-01 -7.36312747e-01
-1.21339941e+00 4.38273400e-01 5.84649265e-01 -6.34357408e-02
-4.99354124e-01 1.24028254e+00 3.60881299e-01 -9.31667984e-01
9.46783006e-01 -4.22820598e-01 -1.26469642e-01 -3.49572957e-01
5.75741649e-01 2.76365787e-01 3.58385593e-01 -4.54576433e-01
-1.45993695e-01 1.94244072e-01 -2.01079980e-01 -3.54485452e-01
1.24058163e+00 -1.28852893e-02 1.93500906e-01 6.67607486e-01
1.22351635e+00 -2.42464513e-01 -1.02627063e+00 -3.13551396e-01
-8.74720141e-02 9.99957547e-02 3.71770263e-01 -5.98258257e-01
-1.00605989e+00 1.30694830e+00 6.69423699e-01 3.24258953e-01
9.11667168e-01 -5.81965074e-02 1.11433780e+00 5.51179826e-01
3.20410170e-02 -1.13267028e+00 -6.94749951e-02 8.35692286e-01
3.85424852e-01 -1.00272858e+00 -6.04645967e-01 -1.58053547e-01
-2.74645627e-01 1.11447823e+00 7.85141528e-01 3.00567985e-01
1.06011367e+00 6.42659605e-01 2.79194057e-01 9.82111692e-02
-8.42868090e-01 1.54158622e-01 2.03538612e-01 6.43500447e-01
4.16114032e-01 3.70979518e-01 5.26835144e-01 1.03759670e+00
-4.78815794e-01 -9.94028151e-02 3.60581487e-01 8.07894886e-01
-6.59173489e-01 -1.03243065e+00 -2.91503549e-01 2.58675784e-01
-4.85073656e-01 -5.01242340e-01 2.03519650e-02 4.02258635e-01
-1.87205046e-01 1.16997743e+00 2.98338860e-01 -8.07709336e-01
2.95784742e-01 4.07736003e-01 2.96910286e-01 -6.79364800e-01
-4.29221213e-01 2.90309459e-01 1.67367354e-01 -4.64834198e-02
-2.76038051e-01 -1.26942813e-01 -9.98812258e-01 -3.51361096e-01
-6.10726833e-01 5.07916868e-01 1.01628292e+00 1.07047570e+00
3.50325137e-01 7.82464921e-01 2.96373487e-01 -9.76864219e-01
-8.91155660e-01 -1.44550109e+00 -5.99953294e-01 -4.04403508e-01
4.37482536e-01 -3.00455838e-01 -5.88023067e-01 -7.71345990e-03] | [14.425528526306152, 6.42282772064209] |
22c6a55c-1620-45ec-a862-6a025f754dd5 | what-s-in-a-name-beyond-class-indices-for | 2304.02364 | null | https://arxiv.org/abs/2304.02364v1 | https://arxiv.org/pdf/2304.02364v1.pdf | What's in a Name? Beyond Class Indices for Image Recognition | Existing machine learning models demonstrate excellent performance in image object recognition after training on a large-scale dataset under full supervision. However, these models only learn to map an image to a predefined class index, without revealing the actual semantic meaning of the object in the image. In contrast, vision-language models like CLIP are able to assign semantic class names to unseen objects in a `zero-shot' manner, although they still rely on a predefined set of candidate names at test time. In this paper, we reconsider the recognition problem and task a vision-language model to assign class names to images given only a large and essentially unconstrained vocabulary of categories as prior information. We use non-parametric methods to establish relationships between images which allow the model to automatically narrow down the set of possible candidate names. Specifically, we propose iteratively clustering the data and voting on class names within them, showing that this enables a roughly 50\% improvement over the baseline on ImageNet. Furthermore, we tackle this problem both in unsupervised and partially supervised settings, as well as with a coarse-grained and fine-grained search space as the unconstrained dictionary. | ['Xuhui Jia', 'Jie Li', 'Sagar Vaze', 'Yandong Li', 'Kai Han'] | 2023-04-05 | null | null | null | null | ['object-recognition'] | ['computer-vision'] | [ 4.28088248e-01 -7.45988190e-02 -5.08916140e-01 -6.19700313e-01
-6.75161839e-01 -8.36911798e-01 1.01337147e+00 -1.19545072e-01
-7.57022917e-01 3.41870993e-01 2.09957850e-03 -6.11479245e-02
-6.47118688e-02 -7.08031893e-01 -6.87396705e-01 -5.63289404e-01
3.03185970e-01 9.44679976e-01 3.36572826e-01 3.40575397e-01
3.05172831e-01 3.96891326e-01 -1.93895888e+00 3.41594666e-01
2.99858928e-01 9.52269137e-01 3.97423148e-01 4.50935453e-01
-2.68324703e-01 7.46520340e-01 -4.29260641e-01 -3.08278382e-01
3.29281479e-01 -2.79852778e-01 -1.04184330e+00 6.75323248e-01
1.08223259e+00 -1.63512155e-01 -3.90538871e-01 1.33297753e+00
-5.19557633e-02 1.88850835e-01 7.55597055e-01 -1.04013395e+00
-7.98777461e-01 3.19030732e-01 -2.86884367e-01 1.55652940e-01
7.67319500e-02 4.14116532e-02 1.41221583e+00 -1.08210695e+00
8.20737302e-01 1.01884282e+00 4.03754830e-01 6.30646110e-01
-1.57187247e+00 -5.08329690e-01 3.25423330e-01 5.20220160e-01
-1.78309834e+00 -5.97436547e-01 4.27416861e-01 -7.03602552e-01
9.16490018e-01 3.17363948e-01 3.05472970e-01 7.84313023e-01
-3.91953260e-01 5.48214614e-01 1.18053448e+00 -5.78227103e-01
3.09998572e-01 3.17429066e-01 4.63995129e-01 7.13535011e-01
1.20263547e-01 3.74601819e-02 -4.93588865e-01 7.66265541e-02
6.82095945e-01 4.93100509e-02 -1.53188154e-01 -8.05321336e-01
-1.56474459e+00 9.54378724e-01 5.90974212e-01 4.56779301e-01
-3.20892602e-01 2.41810724e-01 8.35584626e-02 1.83508813e-01
6.59183934e-02 6.77434266e-01 -5.95766306e-01 4.01667029e-01
-1.05611789e+00 -1.21759042e-01 8.54428232e-01 1.04395211e+00
1.20421004e+00 -2.54388660e-01 -8.38937685e-02 9.21006203e-01
1.67394564e-01 5.50426066e-01 5.58862627e-01 -1.02830994e+00
-6.72502071e-02 3.35605025e-01 5.43804951e-02 -8.97510350e-01
-9.32586342e-02 -3.45772833e-01 -5.68496287e-01 1.91784516e-01
5.68610251e-01 3.73798490e-01 -1.43931162e+00 1.89877737e+00
7.19032586e-02 4.34098214e-01 -1.32372573e-01 8.85640621e-01
8.48165870e-01 3.54184687e-01 1.02373339e-01 -1.67141650e-02
1.41105723e+00 -1.00637841e+00 -3.96005154e-01 -5.89606524e-01
3.22952449e-01 -4.95782316e-01 1.11933577e+00 3.07225853e-01
-5.45878768e-01 -6.68844044e-01 -7.22799778e-01 -3.87920439e-02
-5.75838864e-01 2.11673543e-01 5.24380326e-01 4.41748559e-01
-1.28311086e+00 3.95624250e-01 -6.79521203e-01 -6.83821082e-01
5.73999882e-01 6.27727926e-01 -4.80663240e-01 -2.53202826e-01
-7.32120216e-01 9.08409178e-01 5.75329483e-01 -2.84934551e-01
-1.28152585e+00 -4.15544033e-01 -9.04726923e-01 -1.00033544e-01
5.69238544e-01 -6.11452222e-01 1.15899396e+00 -1.31565368e+00
-9.69873667e-01 1.53240716e+00 -3.71072739e-01 -5.20489037e-01
-2.70429766e-03 4.10282202e-02 -1.85434654e-01 2.48083264e-01
4.76110905e-01 1.29939997e+00 9.52434719e-01 -1.46317959e+00
-8.70608509e-01 -2.26697534e-01 3.82043511e-01 1.00616254e-01
-3.47778618e-01 3.86037864e-02 -9.09073591e-01 -5.66935182e-01
1.89438134e-01 -1.06400204e+00 -1.93390310e-01 1.18319884e-01
-3.81565154e-01 -3.06571335e-01 5.37977040e-01 -2.56406516e-01
7.75200963e-01 -2.08849645e+00 2.11251572e-01 1.25843346e-01
2.46512473e-01 1.41925395e-01 -3.16799343e-01 -4.72861193e-02
-7.80764362e-03 3.86863872e-02 -2.53771693e-01 -1.36456296e-01
1.05580986e-01 7.29138732e-01 -4.47943211e-01 5.28405964e-01
1.88477352e-01 8.86036634e-01 -8.73154402e-01 -4.68329817e-01
3.74381721e-01 1.21625461e-01 -5.63339591e-01 1.61853209e-01
-3.06645960e-01 3.79186243e-01 -4.14302707e-01 4.70376283e-01
2.87140429e-01 -5.98461807e-01 3.96832302e-02 -2.78590560e-01
8.84367526e-02 -5.31653427e-02 -1.25181878e+00 1.63888586e+00
-3.67783219e-01 6.10589683e-01 -1.08065329e-01 -1.35628855e+00
7.70998538e-01 7.14377686e-02 2.35529006e-01 -4.07449067e-01
2.85388585e-02 1.15534835e-01 -8.03472772e-02 -5.04105330e-01
1.13407999e-01 -4.72330987e-01 -6.20016977e-02 3.23348403e-01
4.05662328e-01 -1.52792424e-01 2.35360414e-01 7.15029379e-03
9.84846771e-01 -7.18183890e-02 3.83447796e-01 -2.72258431e-01
5.89982033e-01 2.12978050e-01 3.04353803e-01 1.32221580e+00
-2.15850204e-01 7.69092798e-01 -5.94657250e-02 -5.55030763e-01
-9.39967573e-01 -1.14071739e+00 -4.16761577e-01 1.26235127e+00
3.92320067e-01 -1.64000019e-01 -4.91102397e-01 -7.98611820e-01
-2.45369468e-02 9.22237039e-01 -7.67853260e-01 -1.62983760e-02
-2.68457174e-01 -4.78450477e-01 2.11285487e-01 4.03324455e-01
3.74929935e-01 -1.30990446e+00 -4.13949817e-01 -9.26470160e-02
-3.64299789e-02 -1.24144685e+00 -5.41797400e-01 4.44474399e-01
-4.66933131e-01 -1.18943393e+00 -3.12823236e-01 -1.20066524e+00
9.17000115e-01 2.00174391e-01 1.29338372e+00 -3.54431979e-02
-3.70412171e-01 8.09841096e-01 -2.64855355e-01 -1.09455265e-01
-1.71768099e-01 -2.46844620e-01 3.76531817e-02 3.41318667e-01
7.51744449e-01 -2.88879782e-01 -4.23223078e-01 3.66762400e-01
-8.06692719e-01 -6.16668798e-02 7.40249395e-01 9.42403734e-01
7.77298868e-01 -1.90776676e-01 3.10753822e-01 -9.19165730e-01
-5.09683378e-02 -4.29624885e-01 -5.75294971e-01 4.18701679e-01
-5.30144751e-01 2.01233149e-01 4.62487489e-01 -6.25037074e-01
-6.73827529e-01 4.45291311e-01 1.33313626e-01 -6.40047431e-01
-6.57724977e-01 3.88963461e-01 -2.02607140e-01 -2.21701115e-01
6.58555925e-01 5.17761886e-01 -2.44326234e-01 -4.54191476e-01
8.08469415e-01 5.09824455e-01 7.92175949e-01 -4.04120892e-01
9.79106605e-01 6.55645251e-01 -2.85937101e-01 -8.48060727e-01
-1.38025784e+00 -1.00307024e+00 -1.18096220e+00 1.32005155e-01
1.14375246e+00 -9.73423123e-01 -2.91975141e-01 7.16420934e-02
-9.97633278e-01 -2.82657117e-01 -4.99051571e-01 5.35258055e-01
-7.34645665e-01 2.58431673e-01 -2.49638468e-01 -2.60440469e-01
1.81770995e-01 -1.01186585e+00 1.10635054e+00 6.72268346e-02
-3.40419143e-01 -9.86734569e-01 -2.24725246e-01 3.88880700e-01
1.44168407e-01 -1.76884353e-01 8.13313723e-01 -1.12653804e+00
-7.58352101e-01 -2.61054039e-01 -4.65817541e-01 4.06258255e-01
2.26315588e-01 -5.00373900e-01 -9.91798162e-01 -3.01144600e-01
-2.33105402e-02 -5.54364383e-01 1.18787301e+00 2.82262474e-01
1.27582598e+00 -2.46475786e-01 -4.86281276e-01 7.26833045e-01
1.58420002e+00 7.61382058e-02 2.07871988e-01 3.55484575e-01
8.55592310e-01 5.05687714e-01 3.64310354e-01 1.50978759e-01
2.51735300e-01 7.38661110e-01 3.85371119e-01 1.50538804e-02
-2.80643374e-01 -6.37802780e-02 1.17122475e-02 5.46401501e-01
2.16904566e-01 -1.54149850e-04 -9.52247798e-01 8.82195354e-01
-1.63316143e+00 -8.83461297e-01 2.26515263e-01 2.44655442e+00
1.05763328e+00 9.08573568e-02 -1.82850257e-01 -3.20560753e-01
9.39124882e-01 9.69550088e-02 -6.64123297e-01 1.60073429e-01
-8.66796970e-02 3.27741861e-01 6.31887913e-01 5.27589798e-01
-1.50779414e+00 1.17425311e+00 7.00781250e+00 8.09303343e-01
-1.02625346e+00 9.51679125e-02 4.81496125e-01 8.27521160e-02
-9.38470662e-02 1.97467327e-01 -8.61545146e-01 1.54014021e-01
8.64224851e-01 -2.24498242e-01 4.74458307e-01 8.28821599e-01
-2.17129439e-01 -1.19534209e-02 -1.53296757e+00 1.23890471e+00
5.70932209e-01 -1.36602139e+00 3.45920801e-01 7.63465539e-02
7.62966037e-01 3.57568324e-01 -1.65159732e-01 2.43630618e-01
5.24985552e-01 -1.12761247e+00 8.18677366e-01 5.10512531e-01
9.01562154e-01 -1.63559556e-01 4.38888252e-01 3.68667722e-01
-9.36335981e-01 -1.65894449e-01 -4.35302824e-01 -7.95893669e-02
-1.87191457e-01 4.07008976e-02 -7.83419549e-01 -1.30062597e-02
6.58823609e-01 8.26616764e-01 -8.58518362e-01 1.12400079e+00
-3.90488029e-01 6.37015820e-01 -2.88830489e-01 2.17958093e-01
3.18966717e-01 -6.84082657e-02 2.12315485e-01 1.12940562e+00
4.64696437e-02 3.06999892e-01 7.40240157e-01 8.96289170e-01
-1.80148661e-01 -1.76272690e-01 -5.00827968e-01 -4.65881526e-02
4.53942776e-01 1.16400909e+00 -9.30043697e-01 -5.85831583e-01
-7.13766456e-01 1.07201552e+00 4.24961209e-01 5.04322648e-01
-5.34383833e-01 -2.38370001e-01 4.55910355e-01 1.64916851e-02
7.41576076e-01 -2.74955273e-01 -1.02763064e-01 -1.35452032e+00
-2.42605925e-01 -6.67234182e-01 4.26430911e-01 -6.95622981e-01
-1.37538791e+00 4.21084046e-01 4.05349620e-02 -1.06700540e+00
-1.67775810e-01 -9.63861704e-01 -2.86193043e-01 6.10765100e-01
-1.59022641e+00 -1.27927351e+00 -2.22260892e-01 7.93943703e-01
7.46108949e-01 -2.24741742e-01 1.06597853e+00 7.55983591e-02
-7.10008889e-02 3.79272789e-01 2.51532912e-01 5.11998475e-01
7.84928083e-01 -1.39371538e+00 8.91980603e-02 6.91771924e-01
9.50577021e-01 7.10121870e-01 7.39762306e-01 -3.77602994e-01
-1.18382609e+00 -1.22067082e+00 9.61995244e-01 -8.10108244e-01
8.59424233e-01 -4.05858010e-01 -9.34477687e-01 9.48589742e-01
6.11137040e-02 4.98142004e-01 6.97471023e-01 1.69495955e-01
-8.33007932e-01 -1.06258225e-02 -8.48763227e-01 2.24594027e-01
1.04266834e+00 -8.16512585e-01 -1.06416118e+00 6.38658166e-01
6.42643809e-01 -3.59481480e-03 -6.38034523e-01 1.99688107e-01
3.50261748e-01 -6.83836043e-01 1.13216078e+00 -7.19432831e-01
1.59743428e-01 -4.93417859e-01 -4.30611283e-01 -9.02691424e-01
-4.64177430e-01 -2.25286763e-02 3.74027371e-01 1.04116869e+00
3.76438379e-01 -3.18903983e-01 7.90890694e-01 5.26728749e-01
1.03832953e-01 -2.70535171e-01 -8.76545191e-01 -7.69070268e-01
-8.90164450e-02 -5.04798830e-01 -1.41994758e-02 1.02623928e+00
-4.34015036e-01 5.12968183e-01 -2.91440129e-01 3.09517801e-01
7.51050413e-01 4.43744570e-01 5.78696251e-01 -1.26801848e+00
-3.14978242e-01 -3.50950152e-01 -7.60367453e-01 -1.06035793e+00
5.16633451e-01 -1.12298000e+00 4.57730025e-01 -1.42886949e+00
6.86053872e-01 -5.63958406e-01 -5.98840296e-01 7.94839799e-01
-6.11612797e-02 6.67734742e-01 2.21258342e-01 6.53140068e-01
-1.08479440e+00 2.10895494e-01 8.63539279e-01 -4.64541972e-01
1.90555274e-01 -1.65668011e-01 -6.07353389e-01 7.96752691e-01
3.83857220e-01 -5.05595684e-01 -4.46783155e-01 -6.02836370e-01
-1.91390842e-01 -3.90295297e-01 7.62993574e-01 -8.25451791e-01
4.88619536e-01 -4.82429445e-01 2.33003154e-01 -2.13855624e-01
2.78326333e-01 -7.95078397e-01 1.84202623e-02 2.50567794e-01
-6.54440105e-01 -5.66575468e-01 3.52352075e-02 6.87314749e-01
-3.25203866e-01 -4.10795450e-01 9.38808799e-01 -3.17495555e-01
-1.46153128e+00 4.01375592e-01 -2.07763970e-01 1.79870665e-01
9.30934191e-01 -3.89389694e-01 -1.28993867e-02 -1.58657059e-01
-1.22471762e+00 7.49111176e-02 6.91576004e-01 5.14485121e-01
5.48435688e-01 -1.24703562e+00 -5.45811057e-01 2.84736335e-01
6.23089910e-01 -1.88185990e-01 -2.23992784e-02 3.74025226e-01
-1.39436916e-01 6.48238361e-01 -1.71805605e-01 -9.36740398e-01
-1.10915303e+00 8.91286373e-01 3.20272714e-01 2.40499243e-01
-5.75576603e-01 1.06817937e+00 7.12712288e-01 -5.87363958e-01
2.52394170e-01 -1.08409837e-01 -4.10443395e-01 1.07457392e-01
4.38322395e-01 -3.41778338e-01 -1.58009246e-01 -1.11958826e+00
-5.10450363e-01 1.03211582e+00 -1.33667573e-01 4.63427423e-05
1.11657536e+00 -2.97011077e-01 -2.34135389e-01 6.85402572e-01
1.27902317e+00 -2.11190462e-01 -1.23600078e+00 -8.49020123e-01
4.50534701e-01 -5.99574804e-01 4.58501875e-02 -8.27001572e-01
-8.95015895e-01 6.93232358e-01 6.27963185e-01 -1.69011980e-01
1.01030934e+00 6.22349083e-01 2.74915397e-01 7.88781524e-01
4.27675128e-01 -7.55662799e-01 1.64260060e-01 5.45381486e-01
6.28979266e-01 -1.54543972e+00 -1.49680153e-01 -3.39553833e-01
-5.61882317e-01 1.01112342e+00 4.62448239e-01 -2.53581911e-01
6.87924027e-01 -1.30873501e-01 1.43265292e-01 -2.96184480e-01
-8.06712210e-01 -6.41596019e-01 5.76664805e-01 6.96623147e-01
4.36141640e-02 4.88483235e-02 7.85921365e-02 3.21534932e-01
-9.08908769e-02 -9.53610316e-02 4.71412152e-01 6.33865833e-01
-7.99122155e-01 -1.01279294e+00 -2.25581139e-01 6.19988978e-01
-3.00466150e-01 -2.64718711e-01 -2.51545161e-01 7.10527182e-01
4.47672546e-01 8.63319039e-01 3.18874478e-01 -2.33948439e-01
1.80268601e-01 1.17399484e-01 4.01077837e-01 -1.26088881e+00
-4.54728268e-02 8.14939514e-02 -1.43703237e-01 -4.46791351e-01
-6.97692335e-01 -7.34024942e-01 -1.11752462e+00 3.38381201e-01
-3.65464985e-01 1.10842600e-01 4.47366297e-01 1.23498666e+00
1.01965450e-01 8.15714896e-02 4.54055429e-01 -6.48588598e-01
-6.16370201e-01 -6.32270098e-01 -6.10356450e-01 8.87511551e-01
3.49362850e-01 -7.18490481e-01 -4.78614122e-01 6.46343708e-01] | [9.922109603881836, 1.9385303258895874] |
9a6db18b-1432-43cc-aeba-a9d51809f27c | sjtu-nict-s-supervised-and-unsupervised | 2010.05122 | null | https://arxiv.org/abs/2010.05122v1 | https://arxiv.org/pdf/2010.05122v1.pdf | SJTU-NICT's Supervised and Unsupervised Neural Machine Translation Systems for the WMT20 News Translation Task | In this paper, we introduced our joint team SJTU-NICT 's participation in the WMT 2020 machine translation shared task. In this shared task, we participated in four translation directions of three language pairs: English-Chinese, English-Polish on supervised machine translation track, German-Upper Sorbian on low-resource and unsupervised machine translation tracks. Based on different conditions of language pairs, we have experimented with diverse neural machine translation (NMT) techniques: document-enhanced NMT, XLM pre-trained language model enhanced NMT, bidirectional translation as a pre-training, reference language based UNMT, data-dependent gaussian prior objective, and BT-BLEU collaborative filtering self-training. We also used the TF-IDF algorithm to filter the training set to obtain a domain more similar set with the test set for finetuning. In our submissions, the primary systems won the first place on English to Chinese, Polish to English, and German to Upper Sorbian translation directions. | ['Eiichiro Sumita', 'Masao Utiyama', 'Kehai Chen', 'Rui Wang', 'Hai Zhao', 'Zuchao Li'] | 2020-10-11 | null | null | null | null | ['unsupervised-machine-translation'] | ['natural-language-processing'] | [ 4.28322643e-01 -7.61061907e-02 -5.00126123e-01 -4.81320769e-01
-1.41777658e+00 -6.33336484e-01 8.60600591e-01 -4.23834115e-01
-7.89848328e-01 1.27360523e+00 5.28490782e-01 -9.74864244e-01
4.96752374e-02 -2.40283191e-01 -4.79447961e-01 -2.92431593e-01
7.27022171e-01 1.55821335e+00 -3.22941452e-01 -5.39137244e-01
1.32699192e-01 4.91999201e-02 -5.54382741e-01 8.47495675e-01
1.21963978e+00 2.70673037e-01 4.13696885e-01 6.75267696e-01
-4.90868948e-02 1.28643945e-01 -1.85102776e-01 -8.98607135e-01
4.37982917e-01 -7.14567006e-01 -1.10608280e+00 -4.97965693e-01
2.43123099e-01 1.12497643e-01 4.98674549e-02 8.39190423e-01
9.07699585e-01 8.11423212e-02 8.52289438e-01 -7.26850390e-01
-1.04966068e+00 1.00413072e+00 -3.37604105e-01 1.52654201e-01
4.84223142e-02 1.85657088e-02 7.46363521e-01 -1.57551348e+00
1.11416078e+00 1.34138656e+00 6.20149016e-01 9.00548398e-01
-1.34091330e+00 -4.84079599e-01 -3.02066803e-01 1.07900381e-01
-9.99142170e-01 -5.89845955e-01 1.75995037e-01 -1.48408204e-01
1.63783860e+00 4.03014332e-01 1.68137565e-01 1.60706604e+00
3.78055066e-01 7.91643023e-01 1.49007118e+00 -9.89616752e-01
-1.31427377e-01 5.77595770e-01 -2.45639622e-01 7.65622780e-02
-1.43210083e-01 3.32748622e-01 -7.20457733e-01 -1.52218342e-01
3.46759945e-01 -6.06097102e-01 5.46359234e-02 3.09179485e-01
-1.79666770e+00 5.50871193e-01 -2.76805431e-01 5.50358772e-01
-3.80107999e-01 -2.48240098e-01 5.28392434e-01 1.02293098e+00
8.17647338e-01 5.99109292e-01 -1.26397991e+00 -5.07981002e-01
-1.07104063e+00 -5.94131425e-02 8.08883786e-01 1.28364062e+00
6.09908521e-01 -3.13878655e-01 -5.79738617e-01 1.38910031e+00
9.20843780e-02 9.92865860e-01 7.74593592e-01 -4.80943203e-01
1.12811565e+00 1.19217277e-01 1.08131491e-01 -5.31504229e-02
-9.44693200e-03 -5.82355499e-01 -8.64972115e-01 -3.66067797e-01
2.54705757e-01 -6.10464990e-01 -9.76347983e-01 1.57621741e+00
-1.31444395e-01 -4.84183520e-01 3.88458818e-01 6.71973765e-01
2.99893051e-01 7.38817453e-01 -2.30635151e-01 -4.51988071e-01
1.12979984e+00 -1.55598879e+00 -7.26534009e-01 -2.39096373e-01
9.02592123e-01 -1.67011404e+00 1.10615015e+00 2.10848108e-01
-1.31867111e+00 -7.14982986e-01 -6.49762630e-01 -1.53769672e-01
-4.59678054e-01 6.07354164e-01 2.76861876e-01 5.72315753e-01
-1.17076206e+00 9.17462766e-01 -7.13558614e-01 -8.04098129e-01
-1.20982811e-01 5.18583417e-01 -4.56992656e-01 -1.74301669e-01
-1.51808703e+00 1.58028638e+00 3.51863891e-01 -6.38093278e-02
-6.19885802e-01 -6.40990794e-01 -3.12078118e-01 -5.06017268e-01
-3.17587405e-01 -1.10286236e+00 1.26824808e+00 -1.23260605e+00
-2.04397583e+00 1.16874087e+00 -4.04604167e-01 -4.40543562e-01
9.08048332e-01 -3.49366188e-01 -6.94998503e-01 -6.06442034e-01
2.08579734e-01 6.49408758e-01 4.53329653e-01 -6.40675664e-01
-5.72580874e-01 -3.23173463e-01 -8.21442842e-01 5.97352087e-01
-1.92089781e-01 6.83950484e-01 -3.26214999e-01 -7.79457390e-01
2.94814520e-02 -1.07972741e+00 -2.75222838e-01 -9.20670390e-01
-2.76516438e-01 -4.23142873e-02 4.10417348e-01 -1.15193844e+00
9.32137489e-01 -1.81961799e+00 5.82019567e-01 -1.18875206e-02
-5.64776897e-01 1.85336068e-01 -6.15754426e-01 8.65010083e-01
1.14571340e-01 1.00809447e-01 -2.27674656e-02 -7.73106337e-01
1.67312160e-01 2.43883446e-01 -2.09102109e-01 8.02168623e-02
3.02504301e-01 1.16499138e+00 -8.03869903e-01 -4.65182394e-01
-1.41122073e-01 1.71735734e-01 -4.79575358e-02 4.31948416e-02
-1.64029568e-01 8.98521781e-01 5.49216419e-02 5.09565771e-01
5.86534619e-01 2.00294897e-01 1.28412321e-01 5.35309799e-02
-1.66103199e-01 8.49469543e-01 -4.79780257e-01 2.02761722e+00
-7.03449428e-01 6.18908048e-01 -9.62423831e-02 -4.73809958e-01
1.06554604e+00 5.66015303e-01 1.50310636e-01 -8.66031528e-01
1.01756819e-01 1.09736383e+00 3.16267818e-01 -1.64217100e-01
7.01717317e-01 -1.73971012e-01 2.14440912e-01 6.24858379e-01
4.44982737e-01 -1.94144905e-01 2.38447204e-01 -2.44633079e-01
6.59039021e-01 6.63905323e-01 -1.31556720e-01 -5.90599060e-01
3.37557107e-01 2.76165098e-01 6.31187499e-01 5.82094073e-01
-7.34104738e-02 7.78062403e-01 -2.67041683e-01 -4.01143283e-01
-1.56543946e+00 -9.46544111e-01 -1.77835330e-01 1.29689956e+00
-5.15572548e-01 -3.84545624e-01 -7.77477443e-01 -6.00850046e-01
-5.43220401e-01 9.78738129e-01 -2.93482572e-01 -1.27630100e-01
-1.04539835e+00 -9.48570490e-01 7.68434703e-01 9.15524364e-02
4.01787668e-01 -9.89008188e-01 4.20233756e-01 3.88611436e-01
-6.57516837e-01 -9.75300908e-01 -9.10021245e-01 5.10258496e-01
-1.11000943e+00 -1.47659078e-01 -1.30358016e+00 -1.16008925e+00
3.15193951e-01 -2.92457640e-01 1.42976177e+00 -6.80495322e-01
3.80026996e-01 -2.03239664e-01 -3.32068920e-01 -4.79042798e-01
-9.68915045e-01 6.12781584e-01 4.03432310e-01 -3.33161682e-01
7.32534468e-01 -4.89593357e-01 -1.72752529e-01 6.30876124e-01
-1.92039788e-01 5.35188794e-01 8.53077888e-01 1.15871584e+00
5.71071923e-01 -8.33379269e-01 4.54703391e-01 -6.92473233e-01
9.40310597e-01 -1.56641617e-01 -1.86940566e-01 7.18980372e-01
-8.79909396e-01 6.06948882e-02 4.66765434e-01 -7.21995831e-01
-1.09327340e+00 -2.24500418e-01 -1.78374156e-01 -1.55928165e-01
1.67552352e-01 5.08859634e-01 -2.68872499e-01 2.38556236e-01
9.28315639e-01 6.06626034e-01 -1.36251837e-01 -6.44535005e-01
4.58362073e-01 1.06156027e+00 3.20819706e-01 -7.93084383e-01
6.03150129e-01 -3.62636536e-01 -4.59865302e-01 -2.28876591e-01
-4.84405458e-02 -2.33450145e-01 -1.01669574e+00 3.73623312e-01
7.63696015e-01 -9.15033162e-01 2.67483175e-01 4.64745760e-01
-1.49223769e+00 -5.61346650e-01 -8.09303224e-02 1.07170355e+00
-7.41461873e-01 4.50495929e-02 -9.68733549e-01 -4.76937830e-01
-1.06670761e+00 -1.39029527e+00 9.87418413e-01 -3.06951046e-01
-5.52271068e-01 -1.05011082e+00 5.65591753e-01 4.66273189e-01
7.59382546e-01 -4.89719540e-01 1.00888956e+00 -9.10988212e-01
-3.82180288e-02 1.59969702e-01 -1.08773552e-01 5.61987102e-01
1.55780092e-01 -1.28375173e-01 -4.61083263e-01 -4.33458209e-01
-2.12982208e-01 -1.86400250e-01 6.41511858e-01 2.52863467e-01
8.44993666e-02 -1.29870316e-02 -3.22085559e-01 5.68293869e-01
1.03299582e+00 1.38806224e-01 6.04543090e-01 4.72515613e-01
3.34336668e-01 3.90899450e-01 7.06946194e-01 -2.38728732e-01
2.05864176e-01 8.47642481e-01 -4.64120448e-01 -2.88302243e-01
-2.81264901e-01 -3.03143740e-01 8.19755793e-01 1.44594860e+00
-3.42796534e-01 -2.99863100e-01 -9.88195896e-01 5.59413850e-01
-1.94652903e+00 -5.50386667e-01 -2.31213659e-01 2.27846360e+00
1.21817935e+00 4.72250916e-02 2.03404911e-02 -5.24815619e-01
8.19425106e-01 -6.64754212e-01 -1.00309514e-01 -1.02786374e+00
-4.50037867e-01 4.89371508e-01 8.08987498e-01 6.85129404e-01
-6.60500824e-01 1.49768066e+00 5.91323996e+00 1.27430117e+00
-1.03574777e+00 7.82208204e-01 6.53315842e-01 -5.95827121e-04
-5.11071444e-01 2.26738274e-01 -9.04542565e-01 4.96578693e-01
1.52844918e+00 -2.50094742e-01 8.11605275e-01 3.42874020e-01
3.06985527e-01 3.43873113e-01 -1.20020652e+00 8.21367264e-01
-2.32093215e-01 -1.31750786e+00 7.36527741e-02 1.30205482e-01
1.25551891e+00 8.92793477e-01 -1.71502121e-02 6.84274018e-01
5.27594566e-01 -8.90328109e-01 5.57319343e-01 4.81085181e-01
1.17306781e+00 -5.39772332e-01 9.01888013e-01 5.77683032e-01
-5.32532096e-01 4.48484987e-01 -5.63258469e-01 1.05669819e-01
9.90653187e-02 4.76120055e-01 -1.03057528e+00 9.97271776e-01
2.57178664e-01 5.18548608e-01 -2.48082742e-01 4.71659750e-01
-1.02404125e-01 7.17495382e-01 -2.86545575e-01 -1.61509335e-01
3.08941752e-01 -4.68136162e-01 6.47831440e-01 1.51081622e+00
5.69913626e-01 -5.62008679e-01 -1.42631715e-03 4.95045006e-01
-1.54114366e-01 5.84478438e-01 -3.33544523e-01 -1.61248624e-01
2.36687228e-01 9.47803378e-01 -4.19163823e-01 -4.89039332e-01
-2.33493194e-01 1.63773668e+00 2.13415980e-01 4.14350778e-01
-4.84158754e-01 -1.49278849e-01 5.64195633e-01 -2.13103801e-01
-2.39236400e-01 -2.82049388e-01 -5.55594504e-01 -1.36802483e+00
-8.47977549e-02 -1.32024479e+00 -1.21286355e-01 -5.48382223e-01
-1.36897230e+00 1.23019397e+00 -3.17557216e-01 -1.29884624e+00
-5.08459091e-01 -6.83434784e-01 -3.93003225e-01 1.85149109e+00
-9.25312400e-01 -1.63737750e+00 8.38131309e-01 4.62384045e-01
8.76548886e-01 -8.46344590e-01 1.15794885e+00 6.88751936e-01
-2.22978830e-01 1.02897894e+00 8.77444863e-01 5.23240678e-02
1.49444079e+00 -9.41543579e-01 1.04887211e+00 6.98272407e-01
2.60408968e-01 9.50757325e-01 5.75451195e-01 -8.97648335e-01
-1.28807759e+00 -1.13951492e+00 1.97102225e+00 -8.04182172e-01
5.86860657e-01 -5.09445786e-01 -3.44608307e-01 6.81951582e-01
7.81773448e-01 -6.38041794e-01 6.77391946e-01 3.95142704e-01
-7.45581761e-02 1.79302841e-01 -9.43659723e-01 6.54836774e-01
1.01967204e+00 -6.62808895e-01 -5.97143948e-01 8.75784993e-01
8.17928672e-01 -4.30567920e-01 -8.26993167e-01 4.37557250e-01
5.89941680e-01 -1.08017936e-01 5.75148880e-01 -1.08335769e+00
5.87527752e-01 -3.62328440e-02 -4.76911426e-01 -1.89585221e+00
-4.03641999e-01 -9.78619874e-01 4.96714622e-01 1.24295580e+00
1.24068069e+00 -7.54841328e-01 4.65480864e-01 7.07269162e-02
-4.20849591e-01 -6.86720788e-01 -1.15268445e+00 -9.95249867e-01
8.50587845e-01 -2.22151920e-01 5.72936773e-01 9.07132924e-01
5.11501506e-02 7.39539385e-01 -6.90213978e-01 -4.00560319e-01
1.56527951e-01 -2.22229451e-01 8.60300601e-01 -7.53875792e-01
-5.94561875e-01 -5.43740034e-01 2.30883807e-01 -9.08288479e-01
1.23598576e-02 -1.35769892e+00 -9.68551636e-02 -1.57916081e+00
3.99273247e-01 -1.55748010e-01 -2.80386955e-01 3.22695106e-01
-1.61701098e-01 2.66370803e-01 3.93536910e-02 5.00406563e-01
-9.99841318e-02 3.72081637e-01 1.46474659e+00 -2.59937227e-01
-2.57034749e-01 3.78394812e-01 -2.85654932e-01 -5.78647368e-02
8.63952637e-01 -6.10900640e-01 -1.20765962e-01 -9.66480613e-01
2.00686485e-01 5.15871979e-02 -3.44801664e-01 -5.54840446e-01
2.82455655e-03 -2.57281184e-01 3.42701972e-01 -7.42126584e-01
1.93357751e-01 -4.48296756e-01 3.49662960e-01 4.93523896e-01
-5.11796117e-01 5.88790119e-01 2.45684251e-01 -2.16982976e-01
-8.26233029e-02 -7.46743083e-02 6.37067795e-01 -2.50248581e-01
-6.31152764e-02 2.20511794e-01 -4.02624160e-01 -1.03186540e-01
2.79575527e-01 -9.94791016e-02 -1.51835844e-01 -1.26409784e-01
-7.78894126e-01 1.26841173e-01 3.28455061e-01 7.68886507e-01
1.82527155e-01 -1.45563698e+00 -1.61119366e+00 1.19554147e-01
1.00413844e-01 -8.50447714e-01 -9.97523218e-02 1.28058660e+00
-2.42783800e-01 6.31094277e-01 -3.81583571e-01 -4.82083440e-01
-1.18413997e+00 2.46570691e-01 2.95466632e-01 -7.92984545e-01
-1.08426996e-01 8.92629504e-01 -4.40406352e-01 -1.47272480e+00
-1.67211086e-01 7.61697292e-02 3.90828907e-01 -3.11927497e-01
2.93501914e-01 2.94767022e-01 6.30913436e-01 -4.97143120e-01
-3.53384137e-01 3.65045309e-01 -3.34415019e-01 -9.57840562e-01
9.50768948e-01 -2.80086070e-01 -4.35906649e-01 3.90670210e-01
1.16447866e+00 2.14957595e-01 -4.22514349e-01 -5.14722049e-01
2.76198149e-01 -1.02325514e-01 -1.99843228e-01 -1.53561080e+00
-2.84380704e-01 8.25561821e-01 6.68609738e-01 -7.40549266e-01
8.66308212e-01 -4.50393379e-01 8.68875742e-01 4.50095981e-01
6.37350261e-01 -1.43180919e+00 -5.98399997e-01 1.21758890e+00
9.43511546e-01 -1.35649037e+00 -4.73308593e-01 3.91313396e-02
-5.93402505e-01 1.21703541e+00 3.10298622e-01 2.83162862e-01
2.69609481e-01 1.64159790e-01 4.73340154e-01 5.79171002e-01
-9.99825776e-01 -1.37821184e-02 5.34031272e-01 7.26719856e-01
8.36230397e-01 3.68924409e-01 -9.75613177e-01 5.03010333e-01
-4.19544101e-01 9.68065336e-02 -1.80244759e-01 5.13482451e-01
-8.75114184e-03 -1.99275780e+00 -2.97898769e-01 2.90262073e-01
-4.52788889e-01 -8.00538719e-01 -7.95130610e-01 5.97878933e-01
9.87650156e-02 9.87157285e-01 -3.76704246e-01 -7.05032766e-01
1.45591632e-01 4.11223024e-01 7.51070917e-01 -6.72620893e-01
-1.12007713e+00 3.42269570e-01 4.50337082e-01 -5.21827154e-02
1.44867539e-01 -7.52710223e-01 -6.13643289e-01 -3.38404894e-01
-2.87333965e-01 5.44310987e-01 1.01943040e+00 1.08710980e+00
4.91386771e-01 1.90521568e-01 4.78805989e-01 -6.25008941e-01
-8.44773829e-01 -1.87540472e+00 1.65908173e-01 1.66851684e-01
-2.12360546e-01 6.03436977e-02 -1.64507609e-02 -7.15957135e-02] | [11.604384422302246, 10.426523208618164] |
5300c967-8278-484c-9dc7-0c967eb683e4 | situational-perception-guided-image-matting | 2204.09276 | null | https://arxiv.org/abs/2204.09276v3 | https://arxiv.org/pdf/2204.09276v3.pdf | Situational Perception Guided Image Matting | Most automatic matting methods try to separate the salient foreground from the background. However, the insufficient quantity and subjective bias of the current existing matting datasets make it difficult to fully explore the semantic association between object-to-object and object-to-environment in a given image. In this paper, we propose a Situational Perception Guided Image Matting (SPG-IM) method that mitigates subjective bias of matting annotations and captures sufficient situational perception information for better global saliency distilled from the visual-to-textual task. SPG-IM can better associate inter-objects and object-to-environment saliency, and compensate the subjective nature of image matting and its expensive annotation. We also introduce a textual Semantic Transformation (TST) module that can effectively transform and integrate the semantic feature stream to guide the visual representations. In addition, an Adaptive Focal Transformation (AFT) Refinement Network is proposed to adaptively switch multi-scale receptive fields and focal points to enhance both global and local details. Extensive experiments demonstrate the effectiveness of situational perception guidance from the visual-to-textual tasks on image matting, and our model outperforms the state-of-the-art methods. We also analyze the significance of different components in our model. The code will be released soon. | ['Yong Tang', 'Yandong Guo', 'Cheng Lu', 'Ziwen Li', 'Han Huang', 'Jiake Xie', 'Bo Xu'] | 2022-04-20 | null | null | null | null | ['image-matting'] | ['computer-vision'] | [ 4.69857097e-01 -3.93142328e-02 -1.13812454e-01 -5.75518847e-01
-4.51963276e-01 -1.91540092e-01 5.12128234e-01 -1.58639029e-01
-1.19937167e-01 3.79074097e-01 2.40467086e-01 -1.40099600e-03
1.15322366e-01 -6.04106128e-01 -9.71694291e-01 -6.49353445e-01
5.91688335e-01 2.71273069e-02 8.30379605e-01 -1.99907467e-01
4.29623157e-01 1.57925427e-01 -1.71258783e+00 7.45914519e-01
1.25376797e+00 1.01398420e+00 1.27162552e+00 3.69999468e-01
-4.45595086e-01 8.65313768e-01 -4.30330038e-01 -1.70006812e-01
3.30341399e-01 -5.17179489e-01 -4.46531415e-01 5.14005899e-01
9.48947370e-01 -3.63816708e-01 -2.01100424e-01 1.29811215e+00
3.02161872e-01 1.27321050e-01 3.88473928e-01 -1.45585740e+00
-9.15217698e-01 6.43145084e-01 -8.28760982e-01 4.20507699e-01
1.71881430e-02 3.39299470e-01 6.74569786e-01 -1.23501039e+00
5.76165080e-01 1.44060469e+00 2.38607928e-01 3.13960969e-01
-1.13772750e+00 -5.96158326e-01 6.98344350e-01 3.51407111e-01
-1.02658820e+00 -4.45004582e-01 9.68227267e-01 -4.13218319e-01
4.74628061e-01 3.76450449e-01 8.91569495e-01 8.37715805e-01
1.39699042e-01 1.11187482e+00 1.18829226e+00 -2.05465540e-01
8.40840787e-02 2.47332990e-01 -1.63134769e-01 6.15185559e-01
2.05141217e-01 -2.22270682e-01 -8.60110283e-01 3.50590289e-01
1.13533735e+00 3.34670216e-01 -1.66301280e-01 -5.09710729e-01
-1.44852841e+00 4.09144461e-01 6.80294275e-01 2.80236721e-01
-5.12833893e-01 2.14316487e-01 4.80713397e-02 -1.43782586e-01
6.56939030e-01 4.09331203e-01 -3.39781463e-01 2.83979833e-01
-1.23281777e+00 2.23574236e-01 -1.07822426e-01 1.22379899e+00
1.04801226e+00 3.59582543e-01 -6.33363008e-01 7.45597720e-01
2.90120572e-01 7.23232210e-01 3.19226444e-01 -1.10738635e+00
5.42095065e-01 7.90415764e-01 1.85720921e-01 -1.19992757e+00
-2.95580141e-02 -5.63479066e-01 -5.67405939e-01 8.11467022e-02
1.59601659e-01 3.84392738e-01 -1.19918156e+00 1.70345211e+00
5.04683495e-01 2.65968502e-01 -2.60285079e-01 1.28370988e+00
1.13386011e+00 6.55231655e-01 2.56926417e-01 -1.93463176e-01
1.48068821e+00 -1.43868184e+00 -9.63422775e-01 -9.48705256e-01
6.47973269e-02 -9.04034913e-01 1.40416074e+00 -1.02466889e-01
-1.39886940e+00 -9.30032372e-01 -8.63658547e-01 -2.97620952e-01
-3.57047558e-01 3.24844122e-01 6.25229359e-01 1.39150321e-01
-1.07783115e+00 1.79563642e-01 -6.04720652e-01 -2.82755286e-01
5.81367850e-01 1.02754347e-02 -4.06449400e-02 -1.11983307e-01
-1.14362645e+00 9.44249272e-01 6.02216959e-01 1.14137061e-01
-1.26401329e+00 -7.46066689e-01 -9.57750797e-01 4.08570189e-03
6.85276747e-01 -9.13019478e-01 1.01027119e+00 -1.52812338e+00
-9.28504229e-01 8.19540560e-01 -6.01393163e-01 -1.90109193e-01
2.04413831e-01 -2.55678356e-01 -1.92035839e-01 2.90288597e-01
6.03241205e-01 1.10066772e+00 1.25911605e+00 -1.75692642e+00
-8.44184875e-01 -2.33547598e-01 1.24994069e-02 6.84815884e-01
-3.05049717e-01 -7.61417067e-03 -7.37816334e-01 -1.05182195e+00
4.30183291e-01 -5.32825112e-01 -1.71464309e-01 7.08821341e-02
-2.82282352e-01 3.20082456e-02 9.65072989e-01 -8.74202967e-01
1.11106563e+00 -2.20531797e+00 3.96954238e-01 -3.85542870e-01
3.35316241e-01 1.31810099e-01 -3.46547782e-01 9.10376757e-02
1.94025710e-02 -2.65353411e-01 -1.40832022e-01 -4.60109293e-01
-2.68253803e-01 1.99628353e-01 -4.92991388e-01 1.66166931e-01
2.01380193e-01 1.33594179e+00 -1.19358552e+00 -8.51170778e-01
4.83638644e-01 1.53615892e-01 -3.22775066e-01 4.59761620e-01
-4.87906128e-01 3.62192035e-01 -4.83290017e-01 8.42608273e-01
8.32963467e-01 -3.66307735e-01 -2.33007133e-01 -6.33594871e-01
-1.36308432e-01 2.16816664e-01 -9.33504879e-01 1.89218354e+00
-1.27744541e-01 6.19305730e-01 2.00167060e-01 -7.64491022e-01
7.02028334e-01 -7.23109022e-02 2.75837749e-01 -7.60402083e-01
1.38487723e-02 1.22069895e-01 -1.59448624e-01 -5.13950348e-01
8.35773706e-01 6.86612129e-02 2.88107097e-01 8.35132599e-02
6.49017841e-02 -2.86388814e-01 -2.78021991e-02 5.52716851e-01
4.50313270e-01 2.74095029e-01 1.72292352e-01 -3.69760752e-01
2.86705375e-01 1.01551339e-02 6.58417940e-01 6.92832470e-01
-3.82834136e-01 8.03358555e-01 -6.11165911e-03 -2.31115460e-01
-9.25922990e-01 -1.11978507e+00 2.30964467e-01 1.45516407e+00
9.55639422e-01 -3.65072638e-01 -9.12303388e-01 -6.81547225e-01
-1.19947031e-01 9.20452893e-01 -8.32550406e-01 -2.73900509e-01
-3.62604618e-01 -3.61157268e-01 -5.17894030e-02 6.97092652e-01
7.66383946e-01 -1.34925866e+00 -5.49577832e-01 9.20625683e-03
-7.21820593e-01 -1.14341092e+00 -8.91182542e-01 2.55845450e-02
-8.19917440e-01 -6.56817675e-01 -6.66285634e-01 -9.71430600e-01
8.62667024e-01 1.05903101e+00 1.07987356e+00 1.09918453e-01
-1.11679003e-01 1.78196862e-01 -3.30596983e-01 -5.33362150e-01
-1.44795448e-01 -9.60851163e-02 -1.63152277e-01 2.40408629e-01
1.45265907e-01 -2.26906583e-01 -8.49591434e-01 6.33779943e-01
-1.12522721e+00 1.00081396e+00 7.39781022e-01 5.96262097e-01
6.91222131e-01 -2.46197414e-02 4.69060659e-01 -5.82744837e-01
1.17842257e-01 -3.46090317e-01 -1.92812204e-01 3.42800736e-01
-4.16428238e-01 -2.12507233e-01 8.59853178e-02 -6.10813379e-01
-1.29198265e+00 2.52036601e-02 3.68334413e-01 -7.40906715e-01
-8.48314464e-02 1.78678647e-01 -5.95013261e-01 -6.15472384e-02
3.30837041e-01 6.23687863e-01 -1.72323048e-01 -2.40457892e-01
5.93290925e-01 3.21850181e-01 6.76095545e-01 -5.01753867e-01
9.23951030e-01 7.25670993e-01 -3.31662655e-01 -4.94835377e-01
-1.52930498e+00 -4.69909996e-01 -5.87598503e-01 -5.04721403e-01
1.19546688e+00 -1.09923172e+00 -6.15890399e-02 5.50967276e-01
-1.12925935e+00 -5.35815001e-01 -3.10746044e-01 8.15864205e-02
-5.87303996e-01 3.75866085e-01 -4.60934490e-01 -6.55819118e-01
-2.20146298e-01 -1.27618122e+00 1.33933234e+00 4.42540228e-01
1.22365624e-01 -7.56707251e-01 -6.39377177e-01 6.93596542e-01
5.51312447e-01 -9.34779793e-02 4.68409985e-01 -6.84685558e-02
-1.00563145e+00 5.05660236e-01 -6.94530964e-01 2.17898145e-01
3.35501939e-01 -2.61083275e-01 -1.05023265e+00 -1.82223499e-01
2.30918705e-01 -1.27464041e-01 1.11957872e+00 7.41960287e-01
1.41087544e+00 -2.25829825e-01 -3.34440023e-01 6.50295734e-01
1.11630297e+00 1.41141087e-01 7.43217468e-01 5.76882601e-01
1.18193293e+00 7.12003410e-01 1.36234689e+00 1.05448812e-01
5.40245533e-01 7.32500553e-01 7.27384508e-01 -6.65325999e-01
-5.90119839e-01 -5.01264036e-01 3.65744174e-01 5.54444432e-01
4.02765721e-02 -7.33377934e-02 -3.58732700e-01 6.88346922e-01
-2.11954427e+00 -1.01474595e+00 -1.45513535e-01 1.77198744e+00
1.04099774e+00 1.30131096e-01 -6.72357455e-02 -2.59519398e-01
9.29082751e-01 3.06340933e-01 -6.54379964e-01 4.15207446e-02
-4.63521212e-01 -4.38487262e-01 4.19467270e-01 4.68791425e-01
-1.15482962e+00 1.46256006e+00 5.67257738e+00 1.22191024e+00
-1.06430030e+00 3.06132436e-01 8.57154548e-01 5.83794825e-02
-5.42755663e-01 9.46132541e-02 -7.98282385e-01 6.19645655e-01
1.01259351e-01 3.03623118e-02 5.67174315e-01 8.58697593e-01
3.17898631e-01 -3.30378592e-01 -9.46106553e-01 1.02208185e+00
3.46719742e-01 -1.33046794e+00 4.61033404e-01 -1.92810625e-01
9.00888801e-01 -3.45119655e-01 3.53828788e-01 3.56991962e-03
-1.88775510e-01 -8.28198314e-01 1.44554794e+00 5.90279996e-01
7.01344967e-01 -3.07800502e-01 4.45451498e-01 1.81857124e-01
-1.50707006e+00 -8.58137012e-02 -4.63499606e-01 9.46924984e-02
2.60294080e-01 6.24644518e-01 -8.47422361e-01 4.95760977e-01
8.37985218e-01 8.70046258e-01 -1.01736593e+00 9.02830720e-01
-2.33048841e-01 4.09807920e-01 1.51800901e-01 1.27665222e-01
2.64432877e-01 -1.34240776e-01 6.93211317e-01 1.09855103e+00
2.12313995e-01 -5.30088618e-02 3.80229115e-01 1.24008405e+00
3.51808369e-01 -5.87397162e-03 -2.49790817e-01 4.36379537e-02
4.99148220e-01 1.36406207e+00 -1.01394939e+00 -8.11161160e-01
-1.82390481e-01 1.06854272e+00 2.24075496e-01 6.03243530e-01
-1.03545487e+00 3.94727513e-02 5.13172090e-01 4.80520666e-01
4.49081540e-01 -1.27073854e-01 -6.71144426e-01 -1.20584989e+00
3.55668925e-02 -9.14515734e-01 1.91779286e-01 -1.47404909e+00
-1.16282892e+00 3.60621840e-01 2.65371680e-01 -1.14478946e+00
4.15483475e-01 -2.65917182e-01 -5.33683717e-01 8.71700287e-01
-1.33870864e+00 -1.55202961e+00 -7.03586280e-01 6.05340540e-01
1.08706248e+00 9.34678502e-03 -1.40856206e-03 1.94410101e-01
-3.31072360e-01 2.30494440e-01 -5.05202591e-01 -1.93828642e-01
9.44344997e-01 -1.14364004e+00 3.03063631e-01 1.28361952e+00
-5.64058237e-02 7.28902817e-01 7.89557815e-01 -1.06046736e+00
-1.18912458e+00 -1.42094624e+00 5.62047482e-01 -5.62588751e-01
4.11510110e-01 -4.53649014e-01 -1.00999808e+00 5.90289235e-01
3.83565843e-01 -1.97057575e-01 -6.25161976e-02 -3.35834801e-01
-2.38704279e-01 -2.98108697e-01 -8.36127162e-01 8.89771998e-01
1.14903736e+00 -3.93376201e-01 -7.09138632e-01 1.20879054e-01
1.22244823e+00 -4.62199032e-01 -2.19784126e-01 4.98290300e-01
2.21228033e-01 -9.66040850e-01 1.24828780e+00 -1.35153621e-01
5.90383053e-01 -9.53337669e-01 -2.74279952e-01 -1.21334946e+00
-6.78404212e-01 -3.33767712e-01 -6.62570298e-02 1.40990400e+00
6.73212707e-02 -3.78288388e-01 3.88209760e-01 3.01838040e-01
-4.93072152e-01 -5.12807488e-01 -5.86531401e-01 -4.12035108e-01
-5.58066010e-01 -3.65392268e-01 5.75537205e-01 1.05657148e+00
-3.78189445e-01 3.16413462e-01 -6.12838089e-01 2.57250249e-01
7.80993104e-01 4.49872375e-01 6.61859334e-01 -7.33348727e-01
-1.57276466e-01 -4.46176469e-01 -2.68251121e-01 -1.21423864e+00
-1.58964872e-01 -7.53245473e-01 2.95131624e-01 -1.79766488e+00
7.44482398e-01 -2.59073853e-01 -4.39092577e-01 5.73865175e-01
-8.70939553e-01 4.46292877e-01 3.20357352e-01 1.24980703e-01
-1.00444293e+00 7.22168565e-01 1.98385811e+00 -2.85842657e-01
-5.94854914e-02 -3.95700932e-01 -9.48940814e-01 6.20005608e-01
6.60346985e-01 -2.33326793e-01 -6.70811474e-01 -4.87054169e-01
5.21533005e-03 -9.85541791e-02 6.39612436e-01 -7.23861694e-01
1.27937496e-01 -7.07604289e-01 6.96977019e-01 -9.72578824e-01
4.26306367e-01 -5.87643921e-01 3.23306173e-02 1.99711561e-01
-3.44262183e-01 -1.33697540e-01 1.80138111e-01 6.56635582e-01
-9.69074517e-02 -4.88673337e-03 8.36790085e-01 -2.64563620e-01
-1.00223541e+00 3.66751909e-01 -2.29843333e-01 -4.26722020e-02
1.10134172e+00 -3.35469335e-01 -4.26044345e-01 -3.46884966e-01
-4.44299519e-01 3.38315666e-01 8.04214239e-01 6.49855018e-01
9.91076946e-01 -1.36262918e+00 -5.55915475e-01 2.88302213e-01
2.03673065e-01 1.47729918e-01 5.21125853e-01 9.76516724e-01
-2.44541407e-01 1.35262430e-01 -3.72635365e-01 -8.55242193e-01
-1.38465154e+00 8.24806452e-01 2.51358062e-01 3.09826195e-01
-5.84596395e-01 9.19869781e-01 1.24222016e+00 6.55881464e-02
2.20882580e-01 -4.53654826e-01 -4.11069468e-02 -1.43332958e-01
6.84453785e-01 1.64219886e-01 -2.69737065e-01 -6.56098962e-01
-2.06890360e-01 4.15103793e-01 -8.07465464e-02 -1.43726647e-01
9.92239237e-01 -5.84256828e-01 -4.12310898e-01 5.89143157e-01
4.55248207e-01 -1.63289756e-01 -1.62995911e+00 -3.18477035e-01
-3.89697254e-01 -9.58545983e-01 7.62010738e-02 -7.64075518e-01
-1.15788865e+00 9.63593841e-01 5.92171907e-01 2.33735740e-02
1.30221879e+00 2.46063232e-01 7.90243268e-01 -2.04283237e-01
2.65902013e-01 -9.93462563e-01 7.23896623e-01 1.90876335e-01
1.07435703e+00 -1.23270488e+00 2.86778957e-02 -8.06797147e-01
-1.00417304e+00 5.94411016e-01 1.24144244e+00 -2.86880601e-02
1.79596424e-01 1.08818159e-01 1.41900122e-01 -1.84085459e-01
-6.14560723e-01 -3.32294941e-01 6.95406556e-01 6.79027915e-01
1.42502710e-01 3.25660482e-02 1.66940112e-02 5.73250234e-01
1.18582837e-01 -3.08847129e-01 2.35181466e-01 6.96461320e-01
-8.14333916e-01 -6.60950661e-01 -5.27295291e-01 4.17720437e-01
-6.24349751e-02 -4.16946054e-01 -4.00105774e-01 2.52245635e-01
4.41608965e-01 1.00482333e+00 2.57085580e-02 -2.58402348e-01
-5.57905585e-02 -3.12345296e-01 4.47783172e-01 -7.67508566e-01
-4.07451093e-01 6.30053043e-01 -2.10085481e-01 -7.10948825e-01
-7.01049328e-01 -5.25350749e-01 -1.20157182e+00 1.41478777e-01
-3.91959220e-01 -1.68158710e-01 4.45371687e-01 9.43040490e-01
3.52766305e-01 8.67527723e-01 3.46476704e-01 -1.28879786e+00
5.31321205e-02 -1.08531952e+00 -5.93613029e-01 5.51123857e-01
2.58715957e-01 -9.53133821e-01 -2.18221396e-01 3.71689647e-01] | [10.429675102233887, -0.7281911373138428] |
44819952-4bc5-46dd-8dc2-bebdb6376c9e | a-case-study-on-designing-evaluations-of-ml | 2302.07444 | null | https://arxiv.org/abs/2302.07444v2 | https://arxiv.org/pdf/2302.07444v2.pdf | A Case Study on Designing Evaluations of ML Explanations with Simulated User Studies | When conducting user studies to ascertain the usefulness of model explanations in aiding human decision-making, it is important to use real-world use cases, data, and users. However, this process can be resource-intensive, allowing only a limited number of explanation methods to be evaluated. Simulated user evaluations (SimEvals), which use machine learning models as a proxy for human users, have been proposed as an intermediate step to select promising explanation methods. In this work, we conduct the first SimEvals on a real-world use case to evaluate whether explanations can better support ML-assisted decision-making in e-commerce fraud detection. We study whether SimEvals can corroborate findings from a user study conducted in this fraud detection context. In particular, we find that SimEvals suggest that all considered explainers are equally performant, and none beat a baseline without explanations -- this matches the conclusions of the original user study. Such correspondences between our results and the original user study provide initial evidence in favor of using SimEvals before running user studies. We also explore the use of SimEvals as a cheap proxy to explore an alternative user study set-up. We hope that this work motivates further study of when and how SimEvals should be used to aid in the design of real-world evaluations. | ['Pedro Saleiro', 'Sérgio Jesus', 'Valerie Chen', 'Ada Martin'] | 2023-02-15 | null | null | null | null | ['fraud-detection'] | ['miscellaneous'] | [-2.33611967e-02 3.57869595e-01 -5.79149842e-01 -6.05843604e-01
-4.17222917e-01 -4.55959916e-01 6.82202101e-01 5.67411542e-01
-6.01168036e-01 5.84926128e-01 1.88951269e-01 -8.49504471e-01
4.03693877e-02 -6.93032503e-01 -5.10831118e-01 1.14650868e-01
1.06959462e-01 6.01978183e-01 1.89391244e-02 -2.31080428e-01
5.97347021e-01 2.17887253e-01 -1.59364724e+00 7.67155766e-01
1.23453474e+00 3.48042250e-01 -7.36156851e-02 4.29257810e-01
2.69870572e-02 4.27665859e-01 -5.22980928e-01 -9.50456381e-01
2.98813820e-01 -6.01294339e-01 -6.49735868e-01 4.89299148e-02
3.61180186e-01 -5.92912436e-01 3.52190405e-01 6.87601626e-01
2.80748755e-01 8.36475864e-02 6.28528357e-01 -1.54255509e+00
-6.67207181e-01 5.62402189e-01 -2.51615375e-01 1.92978308e-01
8.67423356e-01 3.62953782e-01 1.25113273e+00 -6.83311284e-01
6.85964823e-01 1.14045000e+00 7.20129907e-01 4.17726725e-01
-1.41440010e+00 -7.80289292e-01 2.03717634e-01 7.11955875e-02
-8.07902515e-01 -2.79327989e-01 2.88184285e-01 -5.23804724e-01
1.02556086e+00 5.65922678e-01 6.58751786e-01 1.08527970e+00
-2.97893912e-01 1.01698768e+00 1.12958920e+00 -8.03597033e-01
3.23953271e-01 9.30358529e-01 7.34195888e-01 8.32640469e-01
1.03258717e+00 1.85445622e-01 -4.12734807e-01 -7.55927682e-01
7.23242760e-01 2.81212237e-02 -2.96846926e-01 -3.51224035e-01
-1.03183460e+00 1.21161652e+00 1.98173642e-01 3.69376004e-01
-5.98597527e-01 -1.81570932e-01 2.26083949e-01 5.85559845e-01
2.96029687e-01 1.01334560e+00 -5.82489789e-01 -2.86324263e-01
-9.17682528e-01 6.48542345e-01 1.18582380e+00 4.96220231e-01
4.60304588e-01 -4.35211271e-01 -2.17788264e-01 7.05579102e-01
4.52639192e-01 -7.50972331e-02 7.37628996e-01 -9.25244689e-01
3.90979499e-01 8.54437172e-01 6.30496085e-01 -7.43712485e-01
-3.10985327e-01 -3.02374274e-01 -9.53580886e-02 3.67828816e-01
7.23326266e-01 -1.08109646e-01 -4.54079837e-01 1.33262765e+00
-1.29894271e-01 -1.36636898e-01 -2.34808281e-01 1.02836609e+00
2.75696695e-01 6.16788417e-02 2.63442039e-01 -3.00448447e-01
1.38581240e+00 -5.33494532e-01 -3.06718856e-01 -4.96786870e-02
1.20693159e+00 -7.12963223e-01 1.65790915e+00 6.68932915e-01
-9.86613452e-01 -3.66597295e-01 -9.88770664e-01 3.28724295e-01
-1.57377243e-01 2.73886230e-02 1.08913851e+00 1.09692967e+00
-5.53770065e-01 8.27455759e-01 -9.17601168e-01 -7.06243753e-01
3.87933254e-02 3.62916648e-01 -1.22367777e-01 9.23554674e-02
-1.05311394e+00 1.05149710e+00 7.35785365e-02 -3.82523090e-01
-1.40006080e-01 -4.99791592e-01 -7.27381647e-01 1.02153368e-01
3.58483881e-01 -7.44954824e-01 1.65392840e+00 -1.18997180e+00
-7.52802968e-01 6.10005438e-01 -1.85547799e-01 -6.26605630e-01
7.72560239e-01 -3.24119478e-01 -3.02435905e-01 -4.39721286e-01
2.55946636e-01 2.50272661e-01 9.43115130e-02 -1.29267013e+00
-7.25219905e-01 -1.65892184e-01 3.80705833e-01 -8.32174122e-02
-1.11147560e-01 8.08256641e-02 -1.74754292e-01 -5.57308733e-01
-1.08861111e-01 -9.00701046e-01 -3.28761995e-01 -3.32345784e-01
-2.14330003e-01 -1.15868859e-01 2.12963551e-01 -5.05063832e-01
1.48906231e+00 -1.72487152e+00 -5.78091383e-01 5.42364061e-01
3.36159281e-02 2.38640264e-01 -5.70613332e-02 4.55719382e-01
-1.44585788e-01 7.62381613e-01 -1.11697666e-01 -3.44367623e-01
2.34557196e-01 4.79376987e-02 -1.57631516e-01 1.80886403e-01
7.99146742e-02 6.82462394e-01 -7.94316769e-01 -4.01198834e-01
1.58797681e-01 1.10306002e-01 -9.70063984e-01 1.40073806e-01
-2.10661560e-01 -1.86846741e-02 -4.54292923e-01 5.88995218e-01
1.62032738e-01 -2.11705178e-01 1.79855734e-01 -1.40169365e-02
-5.11367954e-02 4.78316873e-01 -1.18152559e+00 9.46008265e-01
-5.94039142e-01 6.44535780e-01 -5.10422945e-01 -5.77384293e-01
5.72298646e-01 1.15751229e-01 1.25931412e-01 -8.03726256e-01
5.36021888e-02 4.75171298e-01 4.56045330e-01 -6.82953477e-01
3.58912319e-01 -3.00041020e-01 1.29651487e-01 9.47439432e-01
-2.87952572e-01 1.01286709e-01 4.13399488e-01 1.15956008e-01
1.00716984e+00 8.68913233e-02 6.22601032e-01 -1.58581868e-01
2.05510005e-01 2.50988692e-01 3.88121963e-01 1.10129762e+00
-1.45371050e-01 6.73248768e-01 5.90279818e-01 -4.82799083e-01
-8.66912842e-01 -5.29282689e-01 -2.16182724e-01 9.84586716e-01
-7.01177642e-02 -5.83570063e-01 -8.02601397e-01 -1.18619585e+00
2.99600035e-01 1.42794406e+00 -6.08972430e-01 2.25481447e-02
-1.69991925e-01 -9.33457613e-01 3.68436426e-01 3.89607549e-01
3.00351739e-01 -9.41334426e-01 -1.11768997e+00 2.25669444e-01
-2.07415119e-01 -6.71262503e-01 -2.82840252e-01 3.83009203e-02
-1.05010033e+00 -1.42819977e+00 -4.25639719e-01 -9.50530916e-03
5.28000474e-01 2.22497299e-01 1.37047958e+00 8.58866274e-01
1.44837677e-01 2.86101907e-01 -5.84900796e-01 -5.66544712e-01
-5.77544093e-01 -1.25936884e-02 1.39659777e-01 -2.74381548e-01
8.23763013e-01 -4.15181458e-01 -6.59397840e-01 6.81414843e-01
-7.26627707e-01 -1.11759324e-02 5.67120433e-01 7.51708984e-01
-5.15912077e-04 -2.76534259e-01 4.58879292e-01 -1.62887716e+00
1.25625503e+00 -6.53096259e-01 -3.91404778e-01 3.44545662e-01
-1.41894853e+00 3.94682527e-01 3.51714790e-01 -5.38369238e-01
-8.53905737e-01 -3.45966727e-01 6.33537536e-03 2.70199001e-01
-2.42117733e-01 8.51192534e-01 2.13902723e-02 2.03286067e-01
1.14863896e+00 -5.60312808e-01 -8.96895304e-02 -3.91727716e-01
5.31980060e-02 8.02602172e-01 7.53690675e-02 -5.05462646e-01
4.65768516e-01 -9.62539017e-02 -6.51264608e-01 -4.96135503e-01
-4.57028270e-01 -4.15871471e-01 -3.37386191e-01 1.14133351e-01
3.11480820e-01 -5.50773561e-01 -7.91373014e-01 -4.48567390e-01
-1.04751194e+00 -4.27044094e-01 -3.93834077e-02 6.14192486e-01
-4.74912405e-01 4.24228072e-01 -3.46575886e-01 -1.16874897e+00
7.77012706e-02 -1.03072941e+00 7.22371876e-01 -2.33424939e-02
-1.30279660e+00 -9.53370929e-01 1.04766592e-01 6.92820609e-01
1.61437124e-01 1.09851480e-01 9.86785889e-01 -1.37101221e+00
-3.17008555e-01 -5.19031584e-01 -2.11808607e-01 -5.28298020e-02
-6.55061379e-02 1.30057439e-01 -9.03495014e-01 -1.19455829e-01
-2.35940471e-01 -1.37990475e-01 6.90407574e-01 1.40994102e-01
8.77795339e-01 -4.76170182e-01 -3.02283764e-01 6.54920761e-04
1.17593050e+00 2.25036293e-01 5.54858685e-01 7.41514087e-01
1.65442944e-01 7.34837472e-01 9.41588998e-01 5.57298541e-01
3.05392116e-01 6.70109808e-01 1.93922356e-01 -3.44768882e-01
4.24186051e-01 -3.11192513e-01 3.48524690e-01 -1.89779401e-02
-3.87740970e-01 -4.10417795e-01 -8.74379337e-01 3.12640190e-01
-2.08786511e+00 -1.01480055e+00 -4.54826266e-01 2.70571709e+00
3.86569381e-01 5.95167398e-01 6.90511763e-01 2.54605472e-01
2.90693939e-01 -4.47798610e-01 -2.40835458e-01 -6.17048502e-01
4.25946504e-01 7.17184693e-02 3.62723649e-01 5.04060984e-01
-6.46648765e-01 2.74259269e-01 6.48119640e+00 -3.27624753e-02
-8.71207178e-01 4.70578223e-02 7.48215497e-01 -8.10642540e-02
-6.24236703e-01 2.00531155e-01 -3.28855574e-01 5.85562229e-01
1.15519881e+00 -2.58491151e-02 1.90875679e-01 9.27495539e-01
6.48322344e-01 -3.39781404e-01 -1.55812895e+00 6.83830082e-01
-2.17569619e-01 -1.04704881e+00 -6.09606765e-02 2.85957873e-01
3.72647136e-01 -2.20584139e-01 -1.49635836e-01 6.74791038e-01
4.83054638e-01 -9.13052499e-01 6.19192362e-01 2.96176076e-01
2.73222804e-01 -5.33300042e-01 1.27666187e+00 5.82081139e-01
-4.15179908e-01 -1.27362162e-01 -9.70784202e-02 -3.63113552e-01
6.11641034e-02 4.71543819e-01 -1.56815994e+00 2.31879354e-01
2.95099407e-01 1.17642969e-01 -6.70609355e-01 1.18390846e+00
-6.14779741e-02 1.02513695e+00 -1.45569041e-01 -1.44788414e-01
5.18532246e-02 -4.38687429e-02 1.03396133e-01 1.56068730e+00
4.38577652e-01 3.07903647e-01 1.46487743e-01 1.02194619e+00
2.01129213e-01 3.23062092e-01 -6.13707721e-01 -1.33882284e-01
3.19670439e-01 8.49552512e-01 -8.07016730e-01 -4.18037564e-01
-5.93216240e-01 1.04913688e+00 8.33802894e-02 3.33767444e-01
-7.69612372e-01 -4.56630811e-02 5.68839490e-01 8.38127136e-01
-2.27300614e-01 7.67288879e-02 -6.75290227e-01 -1.27784359e+00
-1.49473790e-02 -1.29405487e+00 5.74573815e-01 -7.44947433e-01
-1.05913377e+00 4.52796459e-01 2.17475727e-01 -1.14446473e+00
-5.97526610e-01 -4.40915704e-01 -9.01598513e-01 9.60147202e-01
-1.05607367e+00 -5.57129800e-01 -2.92606920e-01 5.10510355e-02
5.75844109e-01 4.28688377e-02 7.03482866e-01 8.08240324e-02
-5.10167122e-01 6.56110168e-01 -3.62284541e-01 -2.27912799e-01
7.10986495e-01 -1.37118387e+00 5.49423218e-01 6.85656607e-01
5.09895742e-01 1.17462647e+00 1.25199211e+00 -9.99848187e-01
-7.57205069e-01 -5.56998312e-01 1.04361045e+00 -8.20975959e-01
2.22496822e-01 -4.62939367e-02 -1.13772774e+00 6.09393716e-01
3.82317714e-02 -6.26101136e-01 9.78652656e-01 8.68482709e-01
-4.70457077e-02 5.41947842e-01 -1.42426467e+00 6.04597330e-01
9.04752612e-01 -3.57854337e-01 -7.15476215e-01 1.39892220e-01
3.99408638e-01 -4.17805426e-02 -6.35443985e-01 1.75978914e-01
9.35057461e-01 -1.39108694e+00 6.01929188e-01 -8.09610546e-01
4.25302893e-01 1.94166005e-02 3.36167999e-02 -1.26053417e+00
-1.42498702e-01 -3.40814590e-01 2.70892292e-01 1.24145198e+00
1.04452384e+00 -5.45702994e-01 8.97777677e-01 1.52549624e+00
2.42015064e-01 -7.48457074e-01 -2.53845509e-02 -5.27911961e-01
-3.25397253e-01 -7.97332823e-01 5.95075428e-01 1.29161286e+00
3.53688627e-01 1.86206609e-01 -6.42613247e-02 -3.70789841e-02
3.37724090e-01 -1.22135490e-01 9.91898537e-01 -1.42463136e+00
-6.54631913e-01 -4.51324522e-01 -1.18693918e-01 -5.45565724e-01
-7.82852992e-02 -7.35032022e-01 -3.17695528e-01 -1.56084931e+00
2.93877095e-01 -4.81437474e-01 -2.15531006e-01 5.87333918e-01
-3.58149111e-01 -8.56226087e-02 3.11849654e-01 2.07006916e-01
-2.07236439e-01 -6.71212971e-02 6.69522762e-01 2.64017910e-01
-7.07719982e-01 4.78775710e-01 -1.10711503e+00 8.34333241e-01
7.06169128e-01 -4.67305601e-01 -5.04652679e-01 2.95712538e-02
3.56183916e-01 9.45172459e-03 4.28505957e-01 -6.01296425e-01
-2.35294819e-01 -4.64559495e-01 2.93638766e-01 3.31876837e-02
-1.20164908e-01 -7.90802598e-01 1.11673720e-01 5.46118736e-01
-4.94020492e-01 4.17902946e-01 1.02347836e-01 3.24120849e-01
2.29210127e-03 -6.77153289e-01 1.50864929e-01 -3.04808885e-01
-5.69465101e-01 -1.58132955e-01 -4.59248513e-01 -3.06713998e-01
6.54376924e-01 -5.79444706e-01 8.49973708e-02 -6.23726785e-01
-5.40675044e-01 2.28541464e-01 7.23064423e-01 1.51884079e-01
3.35446119e-01 -1.06116188e+00 -6.61642849e-01 1.79705068e-01
3.60062331e-01 -7.91552603e-01 -5.21245658e-01 6.72730148e-01
-3.85773957e-01 3.53716046e-01 -1.03739373e-01 -1.84969202e-01
-1.40577328e+00 3.93812835e-01 1.89509153e-01 -2.14222968e-01
-2.85501063e-01 4.54321086e-01 -2.58531589e-02 -4.40919608e-01
2.11184263e-01 -6.64150417e-01 -6.26770481e-02 4.14104797e-02
3.18410933e-01 3.87092799e-01 1.71981379e-01 1.40957922e-01
-1.23159021e-01 -2.75437444e-01 -1.79146558e-01 -6.02544248e-01
1.26942301e+00 5.41942641e-02 4.15985227e-01 5.26728392e-01
7.60664225e-01 2.33130127e-01 -7.52335906e-01 1.96859166e-01
5.27072132e-01 -9.03570473e-01 -1.86569050e-01 -1.13721526e+00
-5.44342458e-01 5.48132122e-01 4.59718525e-01 6.04537964e-01
8.75538826e-01 -2.94677973e-01 3.35121185e-01 5.84140956e-01
4.65793163e-01 -8.60434890e-01 -2.46202156e-01 -1.86039358e-01
9.72007453e-01 -1.67323136e+00 1.46683469e-01 -3.53441864e-01
-1.02605796e+00 9.47345793e-01 6.92571878e-01 4.09468338e-02
1.81231916e-01 -2.64292330e-01 3.11976466e-02 -2.98386633e-01
-8.90697181e-01 -1.47927895e-01 3.61796945e-01 2.54514635e-01
1.27823424e+00 2.22533733e-01 -7.73456037e-01 1.03630733e+00
-3.70028734e-01 3.83492678e-01 7.18086123e-01 7.85106421e-01
-2.76730448e-01 -1.51704967e+00 -4.42298114e-01 9.03915823e-01
-3.87538701e-01 -5.76593652e-02 -7.95587301e-01 1.35592639e+00
2.08936617e-01 1.06955159e+00 -2.51034945e-01 -5.13226569e-01
8.65476310e-01 2.92997301e-01 2.90711313e-01 -7.87145615e-01
-9.83278394e-01 -3.03856820e-01 6.82766020e-01 -6.47988558e-01
-1.17832176e-01 -9.28401649e-01 -1.13365173e+00 -4.53693092e-01
-6.85195327e-01 4.57584053e-01 4.28097636e-01 9.81909633e-01
2.50885665e-01 4.15240079e-02 2.37763211e-01 -3.33910823e-01
-6.37342811e-01 -9.37783301e-01 -2.70367324e-01 9.03796434e-01
1.56098083e-01 -7.17199504e-01 -3.88246268e-01 -1.27764121e-01] | [8.828702926635742, 5.888394355773926] |
1d6e8f8d-52ab-43d5-8cf7-0c2f8dae122f | towards-reliable-image-outpainting-learning | 2204.05543 | null | https://arxiv.org/abs/2204.05543v2 | https://arxiv.org/pdf/2204.05543v2.pdf | Towards Reliable Image Outpainting: Learning Structure-Aware Multimodal Fusion with Depth Guidance | Image outpainting technology generates visually plausible content regardless of authenticity, making it unreliable to be applied in practice. Thus, we propose a reliable image outpainting task, introducing the sparse depth from LiDARs to extrapolate authentic RGB scenes. The large field view of LiDARs allows it to serve for data enhancement and further multimodal tasks. Concretely, we propose a Depth-Guided Outpainting Network to model different feature representations of two modalities and learn the structure-aware cross-modal fusion. And two components are designed: 1) The Multimodal Learning Module produces unique depth and RGB feature representations from the perspectives of different modal characteristics. 2) The Depth Guidance Fusion Module leverages the complete depth modality to guide the establishment of RGB contents by progressive multimodal feature fusion. Furthermore, we specially design an additional constraint strategy consisting of Cross-modal Loss and Edge Loss to enhance ambiguous contours and expedite reliable content generation. Extensive experiments on KITTI and Waymo datasets demonstrate our superiority over the state-of-the-art method, quantitatively and qualitatively. | ['Yao Zhao', 'Chunyu Lin', 'Kang Liao', 'Lei Zhang'] | 2022-04-12 | null | null | null | null | ['image-outpainting'] | ['computer-vision'] | [ 4.12956417e-01 7.49244317e-02 -2.88553387e-01 -3.29087108e-01
-9.91040111e-01 -3.86475265e-01 4.93130624e-01 -2.70419002e-01
-1.23236805e-01 6.96114302e-01 2.35646829e-01 2.31486876e-02
-3.99285443e-02 -7.68083632e-01 -7.92290270e-01 -7.39971638e-01
3.68625879e-01 7.11451620e-02 8.19295421e-02 -2.14659229e-01
2.14698225e-01 5.23951530e-01 -1.67734063e+00 3.19174170e-01
1.20584130e+00 1.25843155e+00 3.55742067e-01 3.58316034e-01
-3.15398455e-01 6.76261961e-01 -2.08402872e-02 -5.37023962e-01
6.71602368e-01 -4.15222108e-01 -5.45525253e-01 5.89725792e-01
7.08425522e-01 -7.91623652e-01 -5.78962922e-01 1.05658340e+00
4.38675642e-01 1.64495796e-01 4.85610485e-01 -1.42014778e+00
-7.38723159e-01 1.45545810e-01 -1.14078069e+00 -3.84860605e-01
5.47815084e-01 3.50177735e-01 8.84710014e-01 -9.27354157e-01
6.62428916e-01 1.17199457e+00 4.90803182e-01 5.06758332e-01
-1.15195000e+00 -4.78592038e-01 1.38932213e-01 9.00092274e-02
-1.22303808e+00 -4.14153844e-01 1.46003914e+00 -1.39757901e-01
2.12105170e-01 1.66337956e-02 6.44679666e-01 9.88587320e-01
-6.92314729e-02 8.95556927e-01 1.10854840e+00 -3.56411695e-01
1.05099194e-01 6.47885352e-02 -5.34523249e-01 1.05889809e+00
-2.73487493e-02 3.34026396e-01 -9.45994198e-01 -1.14119938e-02
1.00292754e+00 1.95617005e-01 -5.92467070e-01 -5.97058773e-01
-1.23355913e+00 7.70170510e-01 5.90214849e-01 -1.67378932e-01
-3.63065422e-01 2.41393715e-01 1.12373948e-01 1.04555473e-01
3.52149099e-01 9.63182524e-02 -2.32391775e-01 2.07016841e-01
-9.28853810e-01 1.37680769e-01 4.45968695e-02 1.08268380e+00
1.11367834e+00 1.64880574e-01 1.00370318e-01 7.91707695e-01
4.95617896e-01 5.77604294e-01 1.55683592e-01 -1.64139092e+00
6.29834235e-01 5.83915830e-01 5.40235220e-03 -9.38401937e-01
-8.13151374e-02 -3.70953195e-02 -8.35314691e-01 4.85111266e-01
3.09912324e-01 3.20191421e-02 -9.49414015e-01 1.62554741e+00
4.71979707e-01 2.48965263e-01 4.83098179e-02 1.03031659e+00
7.71178246e-01 6.03454351e-01 -8.59842449e-02 -1.59069404e-01
1.25791693e+00 -7.92876065e-01 -7.36597598e-01 -2.00882494e-01
1.89574927e-01 -7.74801075e-01 1.18415368e+00 5.55971920e-01
-1.44851148e+00 -5.07538378e-01 -1.07099795e+00 -5.73979974e-01
-1.25832617e-01 2.45634168e-02 8.82494092e-01 3.64766121e-01
-9.99078393e-01 3.89400095e-01 -7.03844547e-01 1.24878017e-02
5.31390488e-01 5.61599359e-02 -5.72325349e-01 -5.70956409e-01
-9.75055814e-01 5.05041242e-01 4.72484052e-01 5.25460988e-02
-8.10073137e-01 -8.42510223e-01 -1.24153650e+00 -2.36685246e-01
2.02322394e-01 -1.14449406e+00 7.68500149e-01 -7.88490176e-01
-1.50711310e+00 8.60495210e-01 -9.84989628e-02 1.30462172e-02
6.33149922e-01 -7.32117146e-02 -3.48031111e-02 8.04918826e-01
1.22611478e-01 1.24673033e+00 1.11208296e+00 -1.95798540e+00
-8.69994700e-01 -3.49055350e-01 3.90419625e-02 4.84464824e-01
-3.78883392e-01 -6.01377487e-01 -7.44118869e-01 -7.91054726e-01
4.93969530e-01 -5.21526873e-01 -7.94975758e-02 7.35481024e-01
-5.04468739e-01 2.95994222e-01 8.88724685e-01 -7.27665782e-01
7.49135911e-01 -2.14196515e+00 3.18333626e-01 1.54604405e-01
2.70336300e-01 -3.23894411e-01 -2.63424486e-01 1.97084993e-01
1.21151790e-01 -1.61791623e-01 -7.05986261e-01 -8.09038937e-01
-1.27115428e-01 3.82669091e-01 -4.27001387e-01 4.50806469e-01
4.91187215e-01 9.27411675e-01 -1.01883996e+00 -7.26291895e-01
5.54796100e-01 7.88272321e-01 -6.00122929e-01 3.00407380e-01
-2.56106287e-01 6.58338070e-01 -4.44403768e-01 1.25082386e+00
1.00808656e+00 3.66640016e-02 -3.52979243e-01 -7.12284803e-01
7.37783441e-04 -2.66377747e-01 -9.71084237e-01 2.48979735e+00
-5.03228843e-01 4.02848125e-01 2.61340529e-01 -4.45413381e-01
7.84663796e-01 -3.09423842e-02 6.06785357e-01 -7.59548247e-01
7.61250779e-02 2.72556275e-01 -6.13924146e-01 -4.83092099e-01
7.20364392e-01 -3.35636735e-01 3.67259197e-02 2.59067625e-01
-2.50919741e-02 -6.32333040e-01 -1.11919262e-01 3.51680607e-01
6.64046764e-01 3.87159079e-01 -2.81530887e-01 2.87128091e-01
5.40416062e-01 -2.76831567e-01 4.88354325e-01 2.69347191e-01
-1.18231960e-01 1.20380402e+00 1.70356721e-01 -6.34379610e-02
-9.73128319e-01 -1.37320423e+00 -1.07879557e-01 7.00647354e-01
7.32839167e-01 -1.08745648e-02 -3.87626976e-01 -6.14845335e-01
2.99326684e-02 5.81995547e-01 -6.37076378e-01 -2.34238178e-01
-3.95511001e-01 -2.32525557e-01 2.22394571e-01 5.26408255e-01
8.07353377e-01 -7.57879615e-01 -4.74440575e-01 -1.22735053e-01
-5.46311140e-01 -1.30519545e+00 -6.57201350e-01 2.06390209e-02
-9.40057158e-01 -9.44713235e-01 -8.47062945e-01 -6.21327937e-01
6.24450088e-01 6.11709237e-01 9.08097446e-01 1.17728226e-01
-1.82320774e-01 6.26557887e-01 -3.24882030e-01 1.58072725e-01
-2.18748584e-01 -1.75918609e-01 -2.20520303e-01 2.44434923e-01
-3.05208325e-01 -9.32771146e-01 -9.69738424e-01 1.18374033e-03
-1.49066424e+00 2.87707895e-01 8.38585675e-01 8.30373049e-01
6.52260125e-01 1.43216853e-03 3.66293222e-01 -5.29427767e-01
2.19296381e-01 -3.59660834e-01 -3.38925838e-01 2.91398257e-01
-4.19343024e-01 1.07691780e-01 3.20995629e-01 -2.56747335e-01
-1.39981902e+00 3.14652860e-01 -1.63482100e-01 -8.78796518e-01
2.70497054e-03 1.15441352e-01 -5.35912037e-01 -3.24690759e-01
2.18212470e-01 4.42024559e-01 1.59818083e-01 -3.26594919e-01
8.46149743e-01 3.67342323e-01 9.74751055e-01 -7.80840337e-01
1.02822280e+00 1.01042390e+00 6.56852573e-02 -6.85764194e-01
-6.74401104e-01 -3.31139296e-01 -4.31279391e-01 -3.33612591e-01
8.95966887e-01 -1.09640789e+00 -5.73074281e-01 5.11805356e-01
-1.18279219e+00 -1.65997937e-01 -3.92486215e-01 1.55058831e-01
-7.50199616e-01 7.60701835e-01 -6.91395819e-01 -6.99343145e-01
-2.74899244e-01 -1.04221463e+00 1.66875970e+00 4.09776032e-01
3.09101403e-01 -9.18266237e-01 -2.85404921e-01 6.63673282e-01
1.73856765e-01 5.38252354e-01 8.10119927e-01 4.32822853e-01
-9.28341687e-01 2.02825755e-01 -5.59007943e-01 3.99513215e-01
1.53019831e-01 2.60227220e-03 -1.25264466e+00 -1.52956739e-01
-1.72411025e-01 -4.90087062e-01 1.01340818e+00 1.71428114e-01
1.15001118e+00 3.74767557e-02 -9.05980989e-02 1.10183620e+00
1.45717478e+00 -1.26578078e-01 8.49338770e-01 2.37115726e-01
1.03392208e+00 8.44493687e-01 7.18317091e-01 5.31853437e-01
5.33581436e-01 4.92042005e-01 7.21795738e-01 -4.04873759e-01
-4.47761923e-01 -6.19244158e-01 3.17683518e-01 5.18190086e-01
1.40380666e-01 -5.02804294e-02 -4.85858977e-01 4.88947779e-01
-1.69475424e+00 -7.18153238e-01 2.88488064e-02 2.05839062e+00
1.01286101e+00 -4.75715883e-02 -1.17870674e-01 1.32606223e-01
4.58439171e-01 1.73261330e-01 -5.42629242e-01 2.02211455e-01
-5.00408173e-01 -3.45255882e-02 3.26799005e-01 6.20601058e-01
-8.22252870e-01 7.91012883e-01 5.33264732e+00 1.04187608e+00
-1.05114424e+00 3.27287018e-02 8.41814101e-01 -3.73164628e-04
-9.88871217e-01 -1.41150015e-03 -4.00752217e-01 4.33192819e-01
1.08335756e-01 1.99618369e-01 2.79490769e-01 4.93752420e-01
1.01166785e-01 -4.26640689e-01 -8.37934732e-01 1.27805948e+00
1.81443676e-01 -1.31175423e+00 3.40534568e-01 9.66333076e-02
9.38911974e-01 -2.82972515e-01 3.35894585e-01 -1.66519016e-01
1.11519555e-02 -6.00216269e-01 1.02750659e+00 7.71321237e-01
9.85407770e-01 -9.73404288e-01 3.34939808e-01 4.25974801e-02
-1.18011796e+00 -1.51457697e-01 -1.97585076e-01 4.72565383e-01
5.84299803e-01 7.00984120e-01 -1.43800810e-01 8.85615945e-01
6.24168634e-01 8.77734721e-01 -5.48760056e-01 9.02878761e-01
-3.12107027e-01 -3.39285620e-02 -3.59294266e-01 8.64844561e-01
1.27535641e-01 -3.43279451e-01 4.00899738e-01 7.60337710e-01
3.89758497e-01 1.70336261e-01 8.92960578e-02 1.06550992e+00
-1.59037963e-01 -1.75753698e-01 -6.01842165e-01 2.46977434e-01
5.49561024e-01 1.39112771e+00 -5.66779733e-01 -3.18398289e-02
-4.01829749e-01 1.38781726e+00 8.27244669e-02 5.66457748e-01
-7.29378223e-01 -2.33967423e-01 6.14175081e-01 2.19328493e-01
2.24279270e-01 -1.43857047e-01 -6.38660729e-01 -1.28341544e+00
1.12014957e-01 -6.17902160e-01 2.04506099e-01 -1.16752696e+00
-1.59065831e+00 5.17196417e-01 -3.59603502e-02 -1.55161405e+00
-1.39412716e-01 -4.08511698e-01 -5.80331087e-01 7.85170436e-01
-2.13260388e+00 -1.73226190e+00 -6.50802016e-01 9.24854875e-01
3.99543375e-01 1.83378756e-01 3.21447372e-01 3.97076815e-01
-3.38967055e-01 5.28385699e-01 -1.90260366e-01 -1.93930566e-01
7.91837156e-01 -9.92610276e-01 -1.38395742e-01 9.31506991e-01
-1.11607142e-01 4.19321358e-01 5.24586916e-01 -5.83807826e-01
-1.57626724e+00 -1.07016385e+00 2.11177126e-01 -1.93930387e-01
3.17592382e-01 -5.05118743e-02 -8.36967528e-01 3.16869497e-01
2.18380168e-01 1.44530088e-01 4.23271418e-01 -6.61909223e-01
-4.60164279e-01 -2.22586036e-01 -1.35856402e+00 5.31429231e-01
1.03318393e+00 -6.86278999e-01 -2.29364187e-01 -2.94589601e-03
9.17260408e-01 -4.37983066e-01 -8.26805651e-01 4.39484149e-01
5.16067505e-01 -1.28439271e+00 1.25224316e+00 6.43433928e-02
8.31794202e-01 -6.60728574e-01 -4.07409519e-01 -9.67410922e-01
1.59395009e-01 -7.49250948e-01 -3.80481601e-01 1.48120379e+00
4.02759425e-02 -3.16444874e-01 8.77884030e-01 7.09414005e-01
-3.82566333e-01 -7.39676654e-01 -9.37213480e-01 -3.67496878e-01
-6.94882572e-02 -5.79374909e-01 4.34768975e-01 9.40289557e-01
-2.53894955e-01 6.49544299e-02 -6.04736447e-01 1.22496322e-01
9.97161090e-01 3.81353736e-01 7.07813263e-01 -8.13780606e-01
-3.60714227e-01 -3.52781951e-01 -2.88054883e-01 -1.39010465e+00
1.94928050e-02 -5.67405581e-01 1.42001569e-01 -1.40538263e+00
3.53600420e-02 -4.50031728e-01 3.95915359e-02 4.38404977e-01
-2.56808728e-01 6.27206385e-01 2.36141786e-01 1.90815285e-01
-4.04805541e-01 1.07700789e+00 1.61382258e+00 -1.99945971e-01
-1.57875314e-01 -3.76685798e-01 -9.34204400e-01 6.84419513e-01
4.64189708e-01 -1.76736847e-01 -6.83019698e-01 -4.96589273e-01
2.26478472e-01 2.52672374e-01 5.76354861e-01 -8.76994491e-01
1.54728457e-01 -2.64802963e-01 6.31163120e-01 -7.59102941e-01
7.81265616e-01 -9.58690464e-01 -4.87192236e-02 -8.79462285e-04
-8.64443481e-02 -1.98711663e-01 1.26275033e-01 8.48543227e-01
-3.30941439e-01 5.91246970e-02 8.71390045e-01 2.13299394e-02
-7.66204417e-01 6.14124596e-01 2.00222030e-01 1.72904995e-03
9.51877832e-01 -4.56990421e-01 -1.45113036e-01 -5.69850087e-01
-4.29098845e-01 2.59255707e-01 9.39839959e-01 2.24330470e-01
1.23305476e+00 -1.54778218e+00 -5.72167397e-01 5.30448735e-01
3.13062191e-01 3.60555321e-01 5.32246292e-01 8.38827550e-01
-6.52444839e-01 -2.14699119e-01 -2.51380563e-01 -7.50194848e-01
-7.54549503e-01 3.44991028e-01 2.33221009e-01 1.08210593e-01
-6.35019362e-01 8.97324979e-01 4.68508631e-01 -2.60057360e-01
2.30643138e-01 -2.62982965e-01 2.43560448e-01 -4.98772226e-02
5.07570028e-01 1.90204814e-01 -1.19585842e-01 -6.01214647e-01
-7.16116428e-02 7.72406459e-01 -4.26020510e-02 -4.54701215e-01
1.22278214e+00 -6.94669902e-01 -2.80392859e-02 1.59773365e-01
1.25324845e+00 7.50510693e-02 -1.87580848e+00 -2.17855126e-01
-4.39272046e-01 -9.43045259e-01 2.80038625e-01 -6.37475789e-01
-1.50403559e+00 1.00190699e+00 5.41761458e-01 -1.35562673e-01
1.63678861e+00 -5.52214868e-02 1.24156415e+00 -4.61273901e-02
3.50406170e-01 -8.70110869e-01 4.12499219e-01 3.83087397e-02
8.87145996e-01 -1.43357050e+00 9.17046517e-02 -6.78940594e-01
-7.27431118e-01 1.12846613e+00 5.44326663e-01 6.58652633e-02
3.90503258e-01 1.34685114e-01 9.62195918e-02 -1.31695285e-01
-4.00687039e-01 -2.05388620e-01 3.75429124e-01 9.25927281e-01
7.14684129e-02 -4.03842568e-01 1.04816847e-01 3.41851264e-01
1.32383823e-01 -1.44634828e-01 5.14556468e-01 8.89617801e-01
-2.56425649e-01 -1.01014757e+00 -4.77620780e-01 7.38213360e-02
5.15470989e-02 -1.96907103e-01 -2.52727300e-01 7.44185686e-01
3.63933772e-01 8.18596423e-01 -1.84565172e-01 -3.86966586e-01
1.87915921e-01 -2.29837969e-01 6.86744153e-01 -2.43948519e-01
-1.20990969e-01 1.84788704e-01 -3.30388576e-01 -6.94715261e-01
-5.02654076e-01 -5.60471952e-01 -1.39665103e+00 -3.90696377e-01
-1.68196335e-01 -3.18614900e-01 6.46470249e-01 7.41761386e-01
3.38192195e-01 2.54664332e-01 8.19433391e-01 -1.38670087e+00
-1.48204640e-01 -5.11612058e-01 -7.33116567e-01 5.34636021e-01
5.04352033e-01 -8.19235265e-01 -5.49121380e-01 3.24401140e-01] | [11.188953399658203, -1.2971534729003906] |
c039a525-13df-4a43-95c9-c249f939cc63 | how-the-move-acceptance-hyper-heuristic-copes | 2304.10414 | null | https://arxiv.org/abs/2304.10414v1 | https://arxiv.org/pdf/2304.10414v1.pdf | How the Move Acceptance Hyper-Heuristic Copes With Local Optima: Drastic Differences Between Jumps and Cliffs | In recent work, Lissovoi, Oliveto, and Warwicker (Artificial Intelligence (2023)) proved that the Move Acceptance Hyper-Heuristic (MAHH) leaves the local optimum of the multimodal cliff benchmark with remarkable efficiency. With its $O(n^3)$ runtime, for almost all cliff widths $d,$ the MAHH massively outperforms the $\Theta(n^d)$ runtime of simple elitist evolutionary algorithms (EAs). For the most prominent multimodal benchmark, the jump functions, the given runtime estimates of $O(n^{2m} m^{-\Theta(m)})$ and $\Omega(2^{\Omega(m)})$, for gap size $m \ge 2$, are far apart and the real performance of MAHH is still an open question. In this work, we resolve this question. We prove that for any choice of the MAHH selection parameter~$p$, the expected runtime of the MAHH on a jump function with gap size $m = o(n^{1/2})$ is at least $\Omega(n^{2m-1} / (2m-1)!)$. This renders the MAHH much slower than simple elitist evolutionary algorithms with their typical $O(n^m)$ runtime. We also show that the MAHH with the global bit-wise mutation operator instead of the local one-bit operator optimizes jump functions in time $O(\min\{m n^m,\frac{n^{2m-1}}{m!\Omega(m)^{m-2}}\})$, essentially the minimum of the optimization times of the $(1+1)$ EA and the MAHH. This suggests that combining several ways to cope with local optima can be a fruitful approach. | ['Aurélien Stumpf', 'Johannes Lutzeyer', 'Arthur Dremaux', 'Benjamin Doerr'] | 2023-04-20 | null | null | null | null | ['open-question'] | ['natural-language-processing'] | [ 3.27818394e-01 1.70695037e-01 5.63087650e-02 5.61162494e-02
-9.74468350e-01 -7.14659512e-01 -7.30175525e-02 2.81101465e-01
-9.46415365e-01 9.32968616e-01 -8.64060223e-01 -7.02104867e-01
-6.28268540e-01 -9.37697768e-01 -7.76075363e-01 -1.24663734e+00
-6.89186215e-01 6.46868050e-01 1.12769440e-01 -5.23709416e-01
6.97432220e-01 2.04250768e-01 -1.96844053e+00 -4.41052169e-01
1.05693138e+00 1.22693956e+00 2.04296067e-01 1.04955268e+00
-2.10588485e-01 -1.54477715e-01 -7.09365785e-01 -7.78710961e-01
3.72328401e-01 -8.34763765e-01 -8.75676095e-01 -4.51714724e-01
7.39314184e-02 3.77508640e-01 1.74430907e-01 1.41377819e+00
5.99408209e-01 3.36056679e-01 4.70573187e-01 -1.14594519e+00
3.07851955e-02 8.33061039e-01 -9.49314892e-01 8.97792429e-02
1.84630882e-02 1.13720365e-01 8.46288025e-01 -4.23030496e-01
5.90273201e-01 1.00219333e+00 6.55707300e-01 4.23373014e-01
-1.01867032e+00 -7.06718981e-01 2.35167686e-02 -6.15752451e-02
-1.70040798e+00 2.25791745e-02 3.57604831e-01 2.26725340e-02
1.02822697e+00 9.83552933e-01 6.32552385e-01 1.72788426e-01
2.30774701e-01 4.98125136e-01 1.05438197e+00 -8.38984251e-01
4.35622036e-01 -1.39637291e-01 -2.40571484e-01 9.39447165e-01
4.87676382e-01 -2.73557790e-02 -1.10987432e-01 -2.09956050e-01
3.13899100e-01 -4.07318443e-01 -3.11754979e-02 1.03939451e-01
-7.18247771e-01 1.06471372e+00 1.00379795e-01 4.38496619e-01
-3.54098715e-02 5.30336201e-01 2.07739137e-02 5.83601415e-01
1.44198969e-01 7.21605778e-01 -6.40393913e-01 -6.08135462e-01
-8.57838511e-01 3.07711750e-01 8.10290933e-01 8.85354102e-01
6.85119152e-01 -5.79504371e-02 4.78782564e-01 7.46847689e-01
9.85884294e-02 8.65670919e-01 2.32733220e-01 -1.09156084e+00
5.24144530e-01 6.03044033e-01 2.07653493e-01 -7.45105088e-01
-5.99355817e-01 -5.69620013e-01 -9.02480006e-01 4.81982976e-01
7.10307598e-01 -3.79710048e-01 -3.74655068e-01 1.91134036e+00
2.89330125e-01 -7.24030137e-01 -3.03303510e-01 4.15313631e-01
-5.78481741e-02 6.42517447e-01 -2.19880715e-01 -8.17290187e-01
1.25052965e+00 -5.61643541e-01 -3.11624825e-01 -9.69775915e-02
8.53031576e-01 -8.88945520e-01 1.00244844e+00 5.44697285e-01
-1.88155723e+00 -1.30731896e-01 -1.10429072e+00 5.85836112e-01
-3.68287951e-01 -3.28270823e-01 7.08406568e-01 1.07451761e+00
-1.22141349e+00 7.04043686e-01 -6.36835694e-01 -1.17797785e-01
7.38848895e-02 9.04087842e-01 -7.34746233e-02 -2.31862795e-02
-7.74768591e-01 6.78672612e-01 1.82454377e-01 1.80703178e-01
-3.46324921e-01 -1.91056848e-01 -4.97991592e-01 -1.20685093e-01
6.08836830e-01 -4.00219321e-01 6.64195538e-01 -6.82946384e-01
-1.37915599e+00 8.51303697e-01 -2.16049597e-01 -4.25748020e-01
4.61135149e-01 3.89518350e-01 -4.07416187e-02 -2.50644255e-02
-2.18819067e-01 5.46680152e-01 4.66720462e-01 -9.89532471e-01
-8.83600652e-01 -6.99028373e-01 6.03021085e-02 8.72826055e-02
-4.54504155e-02 2.42998570e-01 -1.22203164e-01 -3.02155137e-01
3.53331417e-01 -1.15541327e+00 -4.12822485e-01 -4.14471775e-01
-2.00304776e-01 -8.78681317e-02 8.04202184e-02 -2.80278176e-01
1.53173220e+00 -2.03219771e+00 5.81587732e-01 7.16280937e-01
6.94330484e-02 1.81701705e-01 9.39973220e-02 5.21728992e-01
2.10377470e-01 3.88430804e-01 -6.97927952e-01 -1.36539027e-01
2.56451845e-01 1.09632015e-01 3.36087584e-01 5.30050099e-01
-5.32192051e-01 5.56222677e-01 -6.83779001e-01 -2.74104387e-01
-3.10771883e-01 2.66556293e-01 -6.14739954e-01 -4.09781694e-01
-1.17092654e-01 -1.36855105e-02 -1.97032347e-01 6.20375276e-01
9.12084997e-01 -1.42508507e-01 3.95358324e-01 3.33707899e-01
-4.88724053e-01 -2.89652258e-01 -1.55751979e+00 1.54125214e+00
-2.57884145e-01 3.90347809e-01 4.39829558e-01 -1.17997968e+00
7.72296250e-01 -3.37916426e-02 5.39966702e-01 -8.42728436e-01
4.32629824e-01 7.85275877e-01 2.39723906e-01 -2.28226766e-01
4.67948586e-01 -4.18198913e-01 -3.25068861e-01 5.67061901e-01
-2.77882278e-01 -1.20268069e-01 3.85116905e-01 -3.33506912e-01
1.42031896e+00 -2.91797727e-01 -5.80305606e-02 -4.32097554e-01
4.71588671e-01 -4.87076938e-02 5.70569336e-01 9.07816172e-01
-2.33356729e-01 2.04086572e-01 9.05382156e-01 -3.26098144e-01
-9.88924623e-01 -7.64900386e-01 -2.56300390e-01 1.27655160e+00
5.49652696e-01 -2.41921425e-01 -1.16051912e+00 -3.40386152e-01
-2.66201496e-01 8.85527670e-01 -7.66387284e-01 -9.97476131e-02
-8.99608076e-01 -1.51619637e+00 6.57905698e-01 1.97837502e-01
5.85853636e-01 -1.05831230e+00 -1.22091019e+00 1.76902398e-01
3.36641222e-02 -2.60828465e-01 -2.99591392e-01 7.56426632e-01
-9.01864171e-01 -8.03983033e-01 -5.97932935e-01 -7.37780690e-01
8.09473813e-01 -2.04603240e-01 7.85003901e-01 1.45039067e-01
-6.69231772e-01 1.45144947e-02 -2.98631430e-01 -2.26719826e-01
-3.05816680e-01 7.46895596e-02 -2.26299122e-01 -5.36440015e-01
2.21395805e-01 -6.15948558e-01 -8.21871400e-01 4.39246893e-01
-8.49531353e-01 -3.87366354e-01 5.17440140e-01 9.55382824e-01
7.24591434e-01 5.51357627e-01 9.29035172e-02 -4.01754826e-01
3.46532434e-01 -9.87480804e-02 -1.00116146e+00 2.38095716e-01
-8.28828335e-01 2.73750335e-01 5.79126716e-01 -4.10402536e-01
-5.22334039e-01 -2.82929182e-01 -3.64044189e-01 6.45998260e-03
2.88651615e-01 3.47323507e-01 -1.07707754e-02 -3.73349518e-01
5.77691376e-01 3.09031457e-01 -4.64092568e-02 -4.78768408e-01
1.03576072e-01 3.01849812e-01 3.76889378e-01 -6.82411075e-01
5.18230021e-01 3.43492091e-01 4.43780422e-01 -5.56483328e-01
-1.96276456e-01 2.43016362e-01 -3.35118920e-02 2.04841346e-01
8.02181900e-01 9.59824491e-03 -1.44947016e+00 1.11681662e-01
-5.98181486e-01 -1.95229962e-01 -3.92470449e-01 2.42561176e-01
-6.16332471e-01 2.53745854e-01 -3.08351278e-01 -1.51697803e+00
-4.00855660e-01 -1.25415170e+00 6.73452675e-01 3.36317718e-01
-9.68016461e-02 -6.06483757e-01 -2.05015615e-01 2.14858368e-01
5.39873242e-01 2.87985146e-01 1.45341229e+00 -4.33328927e-01
-3.34499627e-01 -3.71092021e-01 8.91864449e-02 5.65970056e-02
-4.36917335e-01 -1.04792528e-01 -1.95933193e-01 -4.85309303e-01
1.97599098e-01 1.05048619e-01 5.02824903e-01 4.80807155e-01
9.64708984e-01 -4.11025673e-01 -3.57394546e-01 6.90546930e-01
1.63239467e+00 9.29338872e-01 7.72158921e-01 6.24393463e-01
-5.48285581e-02 3.51426005e-01 5.49141109e-01 6.81127667e-01
8.85204598e-02 7.06749499e-01 7.91630805e-01 1.84283659e-01
4.75676000e-01 4.05737638e-01 3.06730539e-01 6.14317417e-01
-3.64710838e-01 -4.55684453e-01 -7.92490721e-01 5.05471945e-01
-1.55670369e+00 -8.82952988e-01 -2.27088686e-02 2.69637394e+00
7.78632164e-01 3.68746281e-01 1.59318626e-01 5.13517261e-01
8.41629684e-01 -3.50426614e-01 -3.26770633e-01 -1.08296120e+00
-4.19606566e-01 1.03853178e+00 7.49249995e-01 6.02775693e-01
-7.25724041e-01 5.24559975e-01 5.30443716e+00 1.25306952e+00
-7.30269730e-01 -6.27411855e-03 6.48910284e-01 -5.76508582e-01
-2.99367815e-01 2.19614267e-01 -7.92302966e-01 1.00185907e+00
9.14615631e-01 -1.29002128e-02 8.03523242e-01 6.01777136e-01
-3.56782556e-01 -6.15894258e-01 -7.48548627e-01 1.00973022e+00
-5.25049716e-02 -1.00933409e+00 -5.09002090e-01 4.22846496e-01
7.29931235e-01 -3.45404297e-01 3.44226778e-01 1.19858727e-01
2.74383724e-01 -1.28221357e+00 5.35752833e-01 -1.15550622e-01
6.76396310e-01 -1.36541545e+00 8.77502739e-01 5.51860213e-01
-1.19676626e+00 -4.27877873e-01 -3.09243351e-01 2.41269320e-01
2.42658257e-01 3.36590111e-01 -9.45774838e-02 5.77448785e-01
9.88357782e-01 -6.87219203e-01 -8.82863626e-02 1.03193176e+00
3.75691146e-01 2.85980683e-02 -8.45762193e-01 -3.38026851e-01
5.44397175e-01 -4.64147449e-01 6.66147530e-01 8.13320577e-01
7.98926413e-01 3.01810443e-01 -2.32760906e-01 4.32616025e-01
1.12915844e-01 1.98801532e-01 -7.85568655e-02 1.21457148e-02
4.36670691e-01 8.19593847e-01 -1.09523201e+00 7.54508674e-02
2.38622829e-01 8.87745798e-01 3.29536684e-02 -1.02277054e-02
-1.01369548e+00 -1.06530523e+00 5.13849378e-01 2.54654661e-02
6.79766238e-01 -2.02345550e-01 -4.78329569e-01 -5.24037242e-01
1.17012382e-01 -7.46770918e-01 6.00838959e-01 -1.98668867e-01
-7.24535167e-01 6.97603524e-01 2.08053757e-02 -8.63144040e-01
-2.34839156e-01 -6.95485532e-01 -2.45944604e-01 7.48383701e-01
-9.51991260e-01 -3.84842902e-01 2.26057500e-01 3.79175782e-01
4.74695377e-02 -1.11126691e-01 8.96944463e-01 2.63628364e-01
-4.58841026e-01 1.12297618e+00 4.79365587e-01 -4.62421358e-01
5.64957596e-02 -1.10857379e+00 -7.56933987e-02 4.81894374e-01
-1.80861861e-01 5.98619103e-01 1.09676695e+00 -2.87905514e-01
-1.80172384e+00 -2.46541366e-01 9.98657763e-01 -8.56850743e-02
3.44984025e-01 -3.98112684e-02 -4.96592999e-01 1.62581712e-01
2.01173931e-01 -3.75429541e-01 7.23251104e-01 -1.92866400e-01
9.06332433e-02 -8.34794566e-02 -1.50477803e+00 8.31365824e-01
1.32184100e+00 5.50533049e-02 3.83772701e-02 1.82570085e-01
2.94826686e-01 -4.26707745e-01 -9.30018961e-01 5.84806144e-01
6.15332305e-01 -1.31267595e+00 8.32593918e-01 -1.68726116e-01
-5.22938594e-02 -1.54356584e-01 -4.77369010e-01 -7.31578529e-01
9.96363685e-02 -1.17025173e+00 1.03125788e-01 6.52860224e-01
6.75346434e-01 -8.28148484e-01 7.16916680e-01 3.27894568e-01
6.26105815e-02 -1.09802628e+00 -1.61815023e+00 -7.42372096e-01
6.23450696e-01 -2.76210338e-01 5.72481692e-01 5.98318875e-01
-7.34143564e-03 -2.16938421e-01 -1.39539868e-01 -2.05689728e-01
5.35091341e-01 1.76562801e-01 3.45963836e-01 -1.00797772e+00
-9.53505516e-01 -1.03404903e+00 -2.85747021e-01 -7.68263161e-01
-4.48404551e-01 -6.65741980e-01 -5.77715747e-02 -8.64285827e-01
1.50737405e-01 -8.41842532e-01 -3.04654032e-01 3.37307423e-01
9.58086103e-02 1.25015050e-01 7.59206191e-02 -2.72374958e-01
-5.08027911e-01 7.01806024e-02 9.64888871e-01 5.13178632e-02
-2.38842875e-01 -2.42791660e-02 -5.50477207e-01 7.25915432e-01
6.02098882e-01 -6.68667018e-01 -2.50197381e-01 -5.05156279e-01
9.57777739e-01 4.93300736e-01 -1.42255753e-01 -7.24055052e-01
1.59325078e-01 -1.99436128e-01 -1.51325352e-02 -4.37949926e-01
3.90906483e-01 -5.32997012e-01 4.87870425e-01 8.21375370e-01
-1.90464735e-01 6.87548101e-01 3.07313293e-01 2.82571822e-01
1.59100831e-01 -7.85247743e-01 8.15342724e-01 -3.33757848e-01
-5.68907298e-02 -8.28872435e-03 -2.89064795e-01 4.58067656e-03
1.20304215e+00 -4.16247576e-01 -4.46405977e-01 -3.12176108e-01
-6.64088905e-01 1.18007973e-01 5.12765884e-01 -2.72193372e-01
1.04492247e-01 -8.71599674e-01 -4.19188708e-01 1.68303803e-01
-3.59206408e-01 3.07535678e-01 4.01510686e-01 1.23399162e+00
-8.30033123e-01 1.86745971e-02 1.51104331e-02 -5.88416636e-01
-1.14436221e+00 5.13929725e-01 3.81392479e-01 -4.50114131e-01
-1.57252133e-01 1.51203752e+00 -3.91447842e-01 -7.60218874e-02
2.75691092e-01 1.04124345e-01 5.30495882e-01 -2.05251444e-02
2.94514418e-01 8.45612407e-01 3.10233999e-02 -2.55933821e-01
-4.62617934e-01 7.98060715e-01 4.29192223e-02 -6.09486818e-01
1.25333154e+00 -1.98915303e-02 -5.45683563e-01 1.36847481e-01
1.32521200e+00 1.24094449e-02 -7.41959393e-01 4.70917016e-01
-5.54787181e-02 -3.14131945e-01 -4.20956999e-01 -7.30747581e-01
-9.85131025e-01 8.26359451e-01 8.30897689e-01 4.86888468e-01
1.42928004e+00 -1.13061093e-01 7.23287106e-01 2.30684817e-01
8.38583529e-01 -1.44076753e+00 -2.60673404e-01 3.21586013e-01
4.56682086e-01 -7.07983553e-01 -5.13936169e-02 -1.28564209e-01
-3.42414320e-01 1.03130078e+00 3.41343880e-01 -4.03038301e-02
4.95145470e-01 4.38544273e-01 -3.87465179e-01 -6.16447218e-02
-5.47502458e-01 -1.99435465e-02 -1.39590427e-01 -8.65291059e-02
1.76644936e-01 -4.73085511e-03 -1.02134204e+00 6.28045619e-01
-5.46828389e-01 -3.91558081e-01 2.42562536e-02 1.44034004e+00
-7.92767644e-01 -1.54392970e+00 -3.78927976e-01 2.05124095e-01
-5.29896200e-01 -3.93526033e-02 -1.42720327e-01 8.81018996e-01
5.05497634e-01 9.72926438e-01 3.34296264e-02 -3.35652709e-01
2.24789791e-02 4.51537013e-01 9.46125746e-01 1.99455470e-01
-8.77546370e-01 2.96911657e-01 -7.73900002e-02 -4.37058896e-01
-9.99205094e-03 -6.55555665e-01 -1.48878312e+00 -8.86657953e-01
-3.33600521e-01 5.64779043e-01 9.62066293e-01 8.40423048e-01
2.23032564e-01 -6.16380759e-03 5.99889576e-01 -4.85606015e-01
-5.64251721e-01 -5.25009811e-01 -7.58711278e-01 -8.41779113e-02
-2.31976002e-01 -5.84255099e-01 -6.76861107e-01 -4.81208503e-01] | [6.338179111480713, 4.543881893157959] |
940f5aa1-5798-45ab-b9b8-f09ec5602a00 | semantic-positive-pairs-for-enhancing | 2306.16122 | null | https://arxiv.org/abs/2306.16122v1 | https://arxiv.org/pdf/2306.16122v1.pdf | Semantic Positive Pairs for Enhancing Contrastive Instance Discrimination | Self-supervised learning algorithms based on instance discrimination effectively prevent representation collapse and produce promising results in representation learning. However, the process of attracting positive pairs (i.e., two views of the same instance) in the embedding space and repelling all other instances (i.e., negative pairs) irrespective of their categories could result in discarding important features. To address this issue, we propose an approach to identifying those images with similar semantic content and treating them as positive instances, named semantic positive pairs set (SPPS), thereby reducing the risk of discarding important features during representation learning. Our approach could work with any contrastive instance discrimination framework such as SimCLR or MOCO. We conduct experiments on three datasets: ImageNet, STL-10 and CIFAR-10 to evaluate our approach. The experimental results show that our approach consistently outperforms the baseline method vanilla SimCLR across all three datasets; for example, our approach improves upon vanilla SimCLR under linear evaluation protocol by 4.18% on ImageNet with a batch size 1024 and 800 epochs. | ['Mingjun Zhong', 'Georgios Leontidis', 'Mohammad Alkhalefi'] | 2023-06-28 | null | null | null | null | ['self-supervised-learning'] | ['computer-vision'] | [ 4.61191446e-01 5.10709360e-02 -3.54282439e-01 -3.31490099e-01
-5.91776311e-01 -5.53825796e-01 7.38137484e-01 2.12062627e-01
-6.56412482e-01 7.60616958e-01 2.01604329e-02 -6.97210878e-02
-1.98758971e-02 -8.04849088e-01 -7.09195197e-01 -8.20534706e-01
-1.90884054e-01 2.94482946e-01 2.52792150e-01 -1.06350714e-02
4.42752659e-01 2.56364763e-01 -1.73843586e+00 7.15642452e-01
7.22054303e-01 1.00962400e+00 4.42099087e-02 1.33758813e-01
-9.90728438e-02 1.14761436e+00 -7.06022084e-01 -3.60400230e-01
5.09982228e-01 -2.94878572e-01 -8.78109336e-01 -9.14633125e-02
7.57954240e-01 -6.84347227e-02 -3.60504657e-01 1.14954698e+00
5.12597203e-01 2.43459523e-01 6.82852626e-01 -1.37860262e+00
-9.08503890e-01 5.92601717e-01 -1.03004599e+00 4.25281793e-01
5.26995622e-02 -9.76287052e-02 1.05878508e+00 -1.12610483e+00
7.19040751e-01 1.40287662e+00 3.80930513e-01 8.26820552e-01
-1.46484470e+00 -1.02709258e+00 4.33677226e-01 2.29883149e-01
-1.48197234e+00 -5.49554408e-01 7.81602621e-01 -2.80776560e-01
7.52282798e-01 3.08334708e-01 3.95144135e-01 1.09364057e+00
-2.40036905e-01 1.08152044e+00 1.34191322e+00 -3.18782717e-01
3.63254458e-01 4.41137284e-01 5.99184334e-01 3.15449953e-01
4.77199197e-01 5.93406986e-03 -7.28249788e-01 -3.72457176e-01
5.67804754e-01 1.21745169e-01 -2.04158410e-01 -5.12912929e-01
-1.08208525e+00 9.30705965e-01 8.05087805e-01 2.15678990e-01
-2.98462868e-01 6.50295019e-02 4.76352006e-01 4.57738250e-01
5.46761751e-01 4.14918512e-01 -1.84168771e-01 3.94233018e-01
-6.23576283e-01 1.99483752e-01 3.39179307e-01 6.89367592e-01
8.56464446e-01 1.62947662e-02 -2.47820497e-01 1.16186857e+00
1.13043010e-01 3.93542588e-01 7.70947993e-01 -7.78973937e-01
5.48827291e-01 8.26175272e-01 -1.94830909e-01 -1.05425346e+00
7.86715597e-02 -5.14766753e-01 -9.36159849e-01 1.93367690e-01
8.28607976e-02 3.85769010e-02 -1.10330570e+00 1.78796232e+00
-5.67512773e-02 3.67233455e-01 4.46686655e-01 9.78471935e-01
1.10690868e+00 6.64904118e-01 4.14436072e-01 4.53102514e-02
1.28276408e+00 -1.03477407e+00 -4.58927840e-01 -4.64481413e-01
6.67604864e-01 -3.43830615e-01 8.19963336e-01 2.29622334e-01
-8.34882855e-01 -4.46012586e-01 -9.80217516e-01 3.31366330e-01
-4.58475530e-01 1.13357224e-01 7.23467052e-01 2.63998210e-01
-9.23811078e-01 5.65742552e-01 -4.59955007e-01 -9.58072096e-02
9.81847525e-01 3.60360831e-01 -7.24195778e-01 -2.07371011e-01
-1.11367095e+00 5.97510993e-01 4.45504099e-01 -2.38017619e-01
-8.40859652e-01 -6.31617069e-01 -7.39708364e-01 1.89657912e-01
-6.63309032e-03 -3.39254975e-01 7.34362423e-01 -1.61247897e+00
-6.73483908e-01 1.08400524e+00 -3.32775921e-01 -7.43528605e-01
3.32034767e-01 -3.67160857e-01 -3.62709880e-01 2.13894099e-01
2.90832043e-01 1.10321820e+00 9.35954094e-01 -1.62301362e+00
-4.73904580e-01 -5.57242572e-01 9.83477477e-03 4.30757135e-01
-6.67542398e-01 -3.43050696e-02 -5.50273433e-02 -6.44652843e-01
3.95575881e-01 -9.64088261e-01 -2.89963156e-01 -9.52318907e-02
-2.43394941e-01 -3.06262076e-01 8.51015389e-01 -1.91896200e-01
1.00440013e+00 -2.46050978e+00 -2.07698375e-01 2.34217539e-01
3.40125561e-01 5.29747069e-01 -4.62676793e-01 1.38167813e-01
-5.16060174e-01 3.35245281e-01 -2.40760356e-01 -5.51771112e-02
-4.44927126e-01 5.12640066e-02 -7.40253210e-01 3.67127955e-01
3.90754789e-01 8.53781879e-01 -1.25100851e+00 -4.66839552e-01
2.20634282e-01 5.01143634e-01 -5.41413844e-01 -1.59389023e-02
2.50198722e-01 1.80793107e-01 -4.07132715e-01 5.30289292e-01
8.46703589e-01 -3.51108104e-01 2.22890764e-01 -1.49776772e-01
3.58339608e-01 2.65577674e-01 -1.04159141e+00 1.15955937e+00
-3.03930610e-01 6.35895848e-01 -5.65480828e-01 -1.24682701e+00
9.76915359e-01 1.23009928e-01 2.82477349e-01 -1.01767075e+00
-2.00452000e-01 1.79454014e-01 -3.23705748e-02 -1.39607638e-01
4.50058788e-01 -7.77990222e-02 1.56900510e-01 5.09547412e-01
2.32512116e-01 5.82574248e-01 -2.87578087e-02 5.17016113e-01
1.04219854e+00 -1.50217801e-01 2.33397126e-01 -1.77035362e-01
4.46999341e-01 -2.06393629e-01 5.01196325e-01 9.96918678e-01
-4.29705918e-01 6.51379287e-01 3.49531144e-01 -5.03348768e-01
-6.15172505e-01 -1.27228355e+00 -2.17682093e-01 1.14489329e+00
5.08055329e-01 -4.29651171e-01 -2.68550903e-01 -1.07925045e+00
7.43280277e-02 7.54128933e-01 -7.34803736e-01 -4.99063164e-01
-3.81679982e-01 -8.27000678e-01 4.17266697e-01 4.95368958e-01
6.40352845e-01 -1.15774286e+00 -3.47717285e-01 -1.33989463e-02
-2.18658224e-01 -7.67134666e-01 6.56424612e-02 3.40917706e-01
-8.51762950e-01 -1.02171469e+00 -7.36334920e-01 -8.20735097e-01
1.03011310e+00 8.34008873e-01 1.23018837e+00 6.54913485e-02
-1.35907814e-01 1.72931403e-01 -4.34716433e-01 -3.21974307e-01
-6.61334768e-02 -1.26684606e-01 -3.82569619e-02 1.72589615e-01
6.80703938e-01 -4.42798525e-01 -8.98794770e-01 2.82559574e-01
-8.46214771e-01 -8.61036107e-02 4.94146138e-01 1.02546358e+00
6.44487560e-01 -3.63781266e-02 9.43191230e-01 -1.20875657e+00
5.40297091e-01 -7.10146129e-01 -1.29551619e-01 3.09578359e-01
-5.08991420e-01 -2.75813658e-02 6.87153637e-01 -6.22566164e-01
-8.48342121e-01 1.01797178e-01 3.65669161e-01 -7.37874091e-01
-9.06732082e-02 1.09947190e-01 1.06710315e-01 1.97513979e-02
8.77728164e-01 4.40792829e-01 -2.29724228e-01 -3.28341156e-01
2.44598493e-01 7.28546560e-01 3.18831652e-01 -2.66123414e-01
7.39513516e-01 7.49742091e-01 -4.93820488e-01 -7.14890897e-01
-9.21751022e-01 -6.33967817e-01 -2.64612943e-01 -1.07309684e-01
4.40489471e-01 -1.22155571e+00 -2.60477483e-01 1.82662919e-01
-9.52602506e-01 1.82861805e-01 -4.16795939e-01 5.03876925e-01
-1.98899880e-01 1.97732881e-01 -2.48641104e-01 -9.17433321e-01
-2.94597328e-01 -9.61133301e-01 9.49963927e-01 1.85716197e-01
-3.52902621e-01 -8.05737615e-01 4.82420288e-02 2.64285415e-01
2.76503801e-01 2.98743378e-02 7.10202813e-01 -9.99744833e-01
-4.42356914e-01 -2.23695785e-01 -4.26178843e-01 4.85066265e-01
5.61127439e-02 -3.82502884e-01 -1.42338085e+00 -5.17112613e-01
-1.51689276e-01 -7.22244740e-01 1.60729599e+00 2.84498055e-02
1.42499113e+00 -5.06895781e-01 -5.52260578e-01 3.87865692e-01
1.55888569e+00 7.41111115e-02 8.09171796e-01 4.39319968e-01
5.94322979e-01 6.47134602e-01 6.00294352e-01 2.27012515e-01
2.62655094e-02 4.12008196e-01 5.52409828e-01 -2.56862134e-01
-1.99406013e-01 -3.40770423e-01 2.66008139e-01 4.49742347e-01
-7.29558021e-02 -2.55459636e-01 -6.69510305e-01 7.13776469e-01
-1.85813451e+00 -1.18436623e+00 8.54022801e-02 2.33683324e+00
7.70841658e-01 2.24761851e-02 -1.57316431e-01 1.36820018e-01
9.40671325e-01 3.77973467e-01 -6.04469478e-01 -2.61974037e-01
-3.07573736e-01 3.27551305e-01 5.99481285e-01 8.45636055e-02
-1.33054745e+00 1.04142904e+00 5.76921797e+00 9.15167868e-01
-1.18750381e+00 8.97508860e-02 9.93110240e-01 -2.49646351e-01
-3.09730768e-01 -1.02369897e-02 -6.63413107e-01 4.72124010e-01
5.54133773e-01 -1.04140684e-01 3.55366588e-01 9.95056152e-01
-2.96249598e-01 2.30246894e-02 -9.72622752e-01 1.16078687e+00
2.45305642e-01 -1.27340758e+00 6.42869711e-01 -1.58075690e-01
8.26320231e-01 2.69899368e-01 3.31881762e-01 5.18088758e-01
4.80991274e-01 -1.24143767e+00 6.07879281e-01 5.86919971e-02
6.39388621e-01 -6.49722040e-01 6.51824117e-01 1.53852791e-01
-8.04314613e-01 -1.64334267e-01 -8.04334641e-01 1.11959666e-01
-4.40995634e-01 4.95239854e-01 -4.95304912e-01 3.57177705e-01
8.52578819e-01 1.05297768e+00 -7.88011134e-01 9.01009083e-01
-4.80624139e-02 7.67738521e-01 -6.21408001e-02 2.74276584e-01
2.41043463e-01 1.19432762e-01 4.48175102e-01 9.74807262e-01
-4.21274528e-02 -2.62495670e-02 -4.53787185e-02 7.39223182e-01
-3.60625952e-01 1.59065008e-01 -9.38311040e-01 1.65204644e-01
6.11114860e-01 1.10742080e+00 -7.23231554e-01 -5.66015124e-01
-3.44898671e-01 8.68988156e-01 6.52371407e-01 4.86992270e-01
-6.80078447e-01 -3.01603913e-01 4.44295496e-01 1.29305765e-01
2.35344842e-01 3.66122335e-01 -1.75110966e-01 -1.33769107e+00
3.30293253e-02 -8.33250046e-01 5.76248288e-01 -5.21429598e-01
-1.60237658e+00 7.80514002e-01 -2.34211326e-01 -1.55521476e+00
7.76944533e-02 -4.57118124e-01 -5.99301875e-01 6.04435146e-01
-1.73955226e+00 -8.25487852e-01 -1.86216503e-01 5.98199606e-01
4.38213766e-01 -3.65015239e-01 7.82378614e-01 2.04866499e-01
-3.45332384e-01 8.49626720e-01 1.22609571e-01 2.24707574e-01
7.51835644e-01 -1.00534868e+00 1.00549601e-01 4.07981575e-01
5.21305501e-01 8.41003418e-01 2.91442722e-01 -4.42964673e-01
-9.40002322e-01 -1.42343712e+00 7.42264628e-01 -3.55417281e-01
3.55087250e-01 -3.43474030e-01 -1.13725173e+00 6.40132189e-01
1.12251416e-01 3.17519724e-01 7.73265779e-01 7.03526363e-02
-8.51956069e-01 -6.75406978e-02 -1.21307755e+00 6.89991415e-01
1.20719814e+00 -4.48028833e-01 -8.17419708e-01 3.26479226e-01
4.42869931e-01 1.54127494e-01 -3.76175821e-01 5.47640741e-01
4.86217469e-01 -1.04048455e+00 1.21822822e+00 -6.93522096e-01
5.82567573e-01 -2.20380425e-01 -1.98540330e-01 -1.32176292e+00
-4.47259098e-01 -1.09383307e-01 -1.33617833e-01 1.12820959e+00
3.04105878e-01 -8.92312407e-01 6.70023799e-01 1.34029701e-01
4.24282223e-01 -7.23251760e-01 -9.22028959e-01 -9.76828694e-01
1.74859330e-01 2.15572026e-02 3.46397430e-01 1.24906278e+00
-2.09548905e-01 4.86943901e-01 -1.70573771e-01 -1.43104503e-02
7.33976901e-01 1.55678377e-01 5.46394289e-01 -1.13409936e+00
4.96612489e-02 -2.73185879e-01 -6.18893802e-01 -7.97536910e-01
4.32771146e-01 -1.33232057e+00 -3.06149185e-01 -1.26693010e+00
7.17716813e-01 -6.91703081e-01 -1.03710341e+00 7.33773470e-01
-2.85590619e-01 6.02552652e-01 3.07131201e-01 6.33376956e-01
-9.53616858e-01 5.22916734e-01 9.92156863e-01 -4.24643189e-01
-5.47113083e-02 -3.87798250e-01 -9.02616739e-01 7.57942379e-01
9.18469548e-01 -7.58239448e-01 -5.77051640e-01 -3.34958553e-01
-1.51656076e-01 -5.12584805e-01 7.24880338e-01 -9.55983222e-01
-3.55040394e-02 5.07407226e-02 7.58375168e-01 -3.93684179e-01
2.05725729e-01 -7.68019557e-01 -1.78600729e-01 5.13759673e-01
-8.95682395e-01 -2.21995875e-01 1.69340566e-01 6.97782636e-01
-4.04136151e-01 -2.30456203e-01 8.57840538e-01 -2.36833662e-01
-8.63860786e-01 1.69143155e-01 -1.63336828e-01 3.07518750e-01
1.20579374e+00 -4.21884596e-01 -7.05238819e-01 -1.17884055e-01
-5.64110994e-01 1.96066156e-01 3.68260473e-01 7.65802860e-01
8.84435833e-01 -1.48734128e+00 -7.72071540e-01 4.39779013e-01
5.64720035e-01 -1.96158499e-01 1.91482618e-01 3.04425895e-01
-1.36652485e-01 3.35644811e-01 -2.13560060e-01 -6.42937899e-01
-1.33270729e+00 4.52067137e-01 2.09691212e-01 -2.97451675e-01
-5.20705283e-01 9.24926281e-01 9.87539470e-01 -3.70687664e-01
1.72137499e-01 4.08089519e-01 -5.26117563e-01 1.28320158e-01
6.70961559e-01 1.24786325e-01 -9.01118889e-02 -7.12661326e-01
-6.60933435e-01 1.91964433e-01 -6.38336301e-01 1.18252032e-01
1.47606373e+00 1.50469020e-01 -5.67524880e-02 3.49220514e-01
1.50110006e+00 -1.42516837e-01 -1.04485822e+00 -3.36792439e-01
3.19077261e-02 -7.51967788e-01 4.94843908e-02 -7.53948629e-01
-1.32161236e+00 8.20561409e-01 8.35573673e-01 -6.57760426e-02
8.84217560e-01 1.51849121e-01 5.33410072e-01 3.99924308e-01
4.93063807e-01 -8.50100279e-01 3.81744534e-01 3.43230456e-01
8.32723796e-01 -1.35998440e+00 -1.21725544e-01 -2.77656704e-01
-9.05159593e-01 6.30029440e-01 9.46895421e-01 -6.14917934e-01
3.93653631e-01 -2.44004220e-01 7.86956549e-02 -1.91111311e-01
-1.00926280e+00 -6.59351274e-02 1.49671480e-01 6.94631636e-01
4.60076272e-01 2.42667068e-02 -2.74565727e-01 3.90579581e-01
1.42067492e-01 -1.32643640e-01 2.85027623e-01 1.05509555e+00
-3.48056138e-01 -7.58320689e-01 -2.22573146e-01 7.43448436e-01
-5.59715629e-01 -3.73861700e-01 -5.87328374e-01 5.84893763e-01
-3.09794471e-02 6.97348654e-01 3.20830464e-01 -4.09391195e-01
1.13253079e-01 -2.08004974e-02 4.18781668e-01 -7.25528181e-01
-5.98664641e-01 -3.82653415e-01 -1.45048603e-01 -3.70226175e-01
-4.66784179e-01 -3.90940338e-01 -1.12823951e+00 5.53305086e-04
-4.63921338e-01 1.49786964e-01 2.40062580e-01 4.40848798e-01
4.19597924e-01 3.12421918e-01 9.13652778e-01 -5.23596108e-01
-6.68200135e-01 -6.13637626e-01 -4.69563246e-01 9.35430706e-01
3.44139040e-01 -8.49247277e-01 -6.59509897e-01 -4.37446795e-02] | [9.6100435256958, 2.7150485515594482] |
6e1cdf32-9b3e-4203-9aff-aa62e2c92bad | prediction-of-amino-acid-side-chain | 1707.08381 | null | http://arxiv.org/abs/1707.08381v1 | http://arxiv.org/pdf/1707.08381v1.pdf | Prediction of amino acid side chain conformation using a deep neural network | A deep neural network based architecture was constructed to predict amino
acid side chain conformation with unprecedented accuracy. Amino acid side chain
conformation prediction is essential for protein homology modeling and protein
design. Current widely-adopted methods use physics-based energy functions to
evaluate side chain conformation. Here, using a deep neural network
architecture without physics-based assumptions, we have demonstrated that side
chain conformation prediction accuracy can be improved by more than 25%,
especially for aromatic residues compared with current standard methods. More
strikingly, the prediction method presented here is robust enough to identify
individual conformational outliers from high resolution structures in a protein
data bank without providing its structural factors. We envisage that our amino
acid side chain predictor could be used as a quality check step for future
protein structure model validation and many other potential applications such
as side chain assignment in Cryo-electron microscopy, crystallography model
auto-building, protein folding and small molecule ligand docking. | ['Günter Blobel', 'Shengwen Peng', 'Zhenyu Zhou', 'Suocheng Tan', 'Junqiu Wu', 'Qilin Dong', 'Jun Ma', 'Jie Fan', 'Xiangyan Sun', 'Ke Liu'] | 2017-07-26 | null | null | null | null | ['protein-design'] | ['medical'] | [ 2.79712439e-01 6.11002408e-02 -5.51555902e-02 -5.43675482e-01
-7.03070998e-01 -4.64392185e-01 1.72324385e-02 5.03529370e-01
-4.94571239e-01 1.33332312e+00 -4.47965302e-02 -9.02976155e-01
1.50150597e-01 -4.61468637e-01 -1.26400054e+00 -1.11438119e+00
-1.78662106e-01 8.67999375e-01 -2.16270193e-01 -1.72582820e-01
4.14104253e-01 1.08755326e+00 -1.04738545e+00 4.14994389e-01
9.50444043e-01 7.64504075e-01 2.35841483e-01 3.48907351e-01
-1.70559697e-02 3.69612902e-01 -1.15999304e-01 -1.91261560e-01
-2.73655038e-02 -4.94900644e-01 -8.00147116e-01 -4.46010619e-01
2.47653961e-01 -8.01159665e-02 2.44449869e-01 7.28313208e-01
8.00712526e-01 6.01141602e-02 6.65018499e-01 -8.87721255e-02
-5.50656319e-01 -6.10300004e-02 -2.27091312e-01 -1.19935967e-01
4.88430321e-01 7.56466031e-01 8.65488768e-01 -1.02127624e+00
9.94618535e-01 8.03077698e-01 1.00043845e+00 5.12761235e-01
-1.73264945e+00 -6.09675229e-01 -3.32488298e-01 6.37391150e-01
-9.76657987e-01 -2.06711087e-02 6.09588981e-01 -4.22470331e-01
1.80952132e+00 8.34038705e-02 8.95103514e-01 1.34676850e+00
8.41979206e-01 1.43860638e-01 1.05845523e+00 -4.40081835e-01
4.53742772e-01 -4.66437608e-01 2.62002349e-01 7.70616651e-01
9.44513157e-02 2.50992775e-01 -3.94345492e-01 -7.53709972e-01
3.22497338e-01 1.75551072e-01 -2.51624316e-01 -7.50702798e-01
-1.01784289e+00 8.90435338e-01 5.47429442e-01 -1.05122522e-01
-8.19609702e-01 4.18900177e-02 3.77635330e-01 2.93972820e-01
1.67939320e-01 9.24964845e-01 -1.19472206e+00 -2.09565252e-01
-7.22004831e-01 5.35790503e-01 6.43255532e-01 2.77456582e-01
8.12796950e-01 -1.78907186e-01 5.17152905e-01 3.07914525e-01
4.03942496e-01 2.32073694e-01 3.86012077e-01 -7.53657579e-01
-2.57783920e-01 4.39118952e-01 4.23106104e-01 -4.09935415e-01
-8.26898873e-01 9.04269069e-02 -6.79159999e-01 6.35615528e-01
3.82032782e-01 5.07883653e-02 -1.12687910e+00 1.49596381e+00
3.08610380e-01 -4.27162468e-01 -5.97782992e-03 8.47922266e-01
6.70185268e-01 5.30114532e-01 3.55664432e-01 -6.65706873e-01
1.41380525e+00 -2.49640122e-01 -2.38471866e-01 5.42162061e-01
8.16964805e-01 -6.70974076e-01 7.18945324e-01 7.12370574e-01
-1.05845737e+00 -2.91211635e-01 -1.22582853e+00 -1.04151934e-01
-2.64889210e-01 -1.53453618e-01 8.46505582e-01 3.71942073e-01
-6.01814687e-01 1.27896404e+00 -9.46946681e-01 -2.12431982e-01
2.33445555e-01 8.07809353e-01 -7.91474819e-01 5.48663139e-01
-1.14875221e+00 1.25163245e+00 7.76825309e-01 -2.52483189e-01
-6.59183860e-01 -9.21805739e-01 -3.12207848e-01 -1.02504790e-02
-7.29354694e-02 -7.93999970e-01 1.07930267e+00 -1.00533164e+00
-1.68139899e+00 9.62729216e-01 -5.12783110e-01 -8.57606351e-01
-1.34586636e-02 -1.00814670e-01 -3.49689692e-01 -8.60671550e-02
-5.54531932e-01 6.31013393e-01 3.70473593e-01 -7.58668780e-01
3.69821070e-03 -6.33001268e-01 -5.40522337e-01 1.64729401e-01
6.89507902e-01 4.50335667e-02 6.55862570e-01 -2.48271897e-01
2.91485041e-01 -8.79370809e-01 -5.88762701e-01 -1.29922748e-01
-3.27679515e-01 -1.29840940e-01 4.41243768e-01 -5.61026335e-01
4.56749409e-01 -1.16998243e+00 5.10491490e-01 5.15453696e-01
4.90549624e-01 6.40353143e-01 3.23528498e-01 9.12394404e-01
-7.14180350e-01 -1.48030862e-01 -3.38370837e-02 5.26942551e-01
-4.84248787e-01 -2.20637053e-01 -1.90897390e-01 6.96315527e-01
9.14952904e-02 1.02739477e+00 -3.36000204e-01 1.66478902e-01
3.20723861e-01 7.19978988e-01 -5.38162708e-01 2.14725718e-01
-5.65821707e-01 7.47035265e-01 -2.84164250e-01 5.84848642e-01
8.08940649e-01 -4.58633244e-01 6.49823308e-01 -4.36176807e-01
-7.89335072e-02 5.58619082e-01 -3.01585972e-01 1.58210170e+00
-4.71334346e-02 3.57183725e-01 -3.57570738e-01 -6.30105495e-01
1.01213944e+00 1.60646424e-01 6.16677701e-01 -7.31676459e-01
5.74069917e-02 2.11358353e-01 7.69399285e-01 -4.02783900e-01
1.06549315e-01 -6.31018519e-01 5.28034747e-01 3.85680944e-01
4.98574339e-02 6.81991160e-01 -3.28810334e-01 -2.30862081e-01
8.03641856e-01 6.48619711e-01 4.91653860e-01 -2.90707052e-01
7.09368050e-01 3.78356695e-01 3.75756443e-01 1.33247226e-01
-1.46188736e-01 6.33617818e-01 4.22161013e-01 -1.31182861e+00
-1.81797671e+00 -7.36820757e-01 -3.72959673e-01 1.14740038e+00
-2.96363145e-01 -4.74123538e-01 -7.72853315e-01 -4.72505212e-01
6.62138388e-02 2.68402100e-01 -4.90370959e-01 -2.85716206e-01
-9.28786337e-01 -1.17165029e+00 2.76559383e-01 2.82239944e-01
-2.57091135e-01 -1.56890368e+00 -5.67202508e-01 8.00570786e-01
3.05672437e-01 -3.49235833e-01 5.82385845e-02 1.08413720e+00
-9.00245726e-01 -1.35321450e+00 -8.26018095e-01 -5.21157324e-01
3.53347540e-01 -3.64233285e-01 1.05820370e+00 3.38736586e-02
-4.18732792e-01 -6.91967666e-01 1.90658066e-02 -3.73947948e-01
-8.87751579e-01 1.06870838e-01 4.00617838e-01 -4.92083281e-01
1.17718542e+00 -9.37293589e-01 -1.20108521e+00 8.40814114e-02
-3.37044388e-01 1.58901870e-01 7.36803472e-01 8.77672553e-01
9.72240567e-01 -7.60332406e-01 5.74275136e-01 -1.00733113e+00
5.46343684e-01 -1.49283633e-01 -6.43224299e-01 1.70284331e-01
-1.01896000e+00 6.61200285e-01 9.90236878e-01 -2.86072120e-02
-7.95184314e-01 4.55111533e-01 -8.24663281e-01 -4.95719872e-02
-6.11194134e-01 3.91242802e-01 -3.00558031e-01 -4.52804476e-01
7.44774461e-01 6.00034237e-01 4.74070787e-01 -8.65151525e-01
-1.35428250e-01 8.25110078e-02 1.91833273e-01 -4.84590083e-01
1.14797205e-01 1.96526542e-01 5.23432314e-01 -6.27710819e-01
-2.81956941e-01 -3.20294112e-01 -7.30183840e-01 3.69314343e-01
8.45657706e-01 -6.13861918e-01 -1.67601097e+00 1.07295342e-01
-1.25306129e+00 7.76641257e-03 3.69343817e-01 5.49780428e-01
-8.86187315e-01 8.60593796e-01 -5.05460799e-01 -2.47542113e-01
-8.17543030e-01 -1.56490684e+00 9.62044001e-01 -9.75809619e-02
-5.97165227e-01 -8.06008875e-01 4.55847353e-01 2.46469349e-01
1.36071756e-01 5.13128340e-01 1.48091447e+00 -7.94279397e-01
-3.59007627e-01 5.59290051e-02 5.80693446e-02 -8.78772582e-04
-1.86095610e-01 5.12402598e-03 -7.35666275e-01 -2.65483558e-01
-3.40914547e-01 -5.20136476e-01 9.16984141e-01 5.99816501e-01
1.14458477e+00 -1.30758226e-01 -3.24351221e-01 8.96873057e-01
1.45408094e+00 5.95759213e-01 7.21618652e-01 6.40131414e-01
8.55653763e-01 3.98008078e-01 3.32980067e-01 2.45043367e-01
-4.82200325e-01 6.78171158e-01 5.14307857e-01 -1.73008516e-02
3.65879178e-01 -1.17782153e-01 2.22067237e-01 1.20789714e-01
-4.93025720e-01 -1.53878182e-01 -1.02875829e+00 -2.46125191e-01
-1.68058789e+00 -1.01402807e+00 -3.38962317e-01 2.14212275e+00
1.06256413e+00 5.18368632e-02 1.96991175e-01 -2.98023045e-01
2.69228578e-01 -3.99737269e-01 -1.19376385e+00 -7.89830446e-01
-1.02406971e-01 5.19028544e-01 6.69858217e-01 3.24432880e-01
-7.51935840e-01 8.02756369e-01 7.34521198e+00 4.80342746e-01
-1.22666430e+00 -2.65680522e-01 5.11330903e-01 -5.90878390e-02
-2.55323529e-01 2.26229504e-01 -7.73701906e-01 5.57940662e-01
1.26894903e+00 1.44239053e-01 1.95198610e-01 8.24034810e-01
5.15158534e-01 1.04912706e-02 -1.12351251e+00 7.33228385e-01
-7.22845554e-01 -2.09659290e+00 1.88556775e-01 5.17873466e-01
3.43404561e-01 2.61474311e-01 -9.27382037e-02 -2.83064693e-01
3.88178676e-02 -1.45984602e+00 1.21555692e-02 5.27208269e-01
7.37544239e-01 -1.25317049e+00 5.92675090e-01 2.53800154e-01
-4.06752586e-01 5.94592154e-01 -6.02547586e-01 1.34608918e-03
5.79682961e-02 2.77851135e-01 -9.12857234e-01 1.58344746e-01
2.81076193e-01 3.19680512e-01 -7.00427517e-02 7.08563089e-01
2.91071326e-01 4.85284209e-01 -2.36125395e-01 -1.00282274e-01
3.57915998e-01 -5.44443071e-01 2.15238169e-01 8.88873935e-01
-1.96110889e-01 1.57768935e-01 9.92968008e-02 8.72946501e-01
-9.31991190e-02 2.40191653e-01 -2.45782942e-01 8.73235837e-02
2.26899058e-01 8.80815685e-01 -3.90844584e-01 -1.66552842e-01
-2.04734385e-01 8.58861327e-01 4.51014936e-01 3.59906703e-01
-8.58722568e-01 -4.95675176e-01 1.15103388e+00 2.60376871e-01
2.96061814e-01 -3.06450613e-02 2.55045682e-01 -8.78663123e-01
-1.07929230e-01 -9.47301865e-01 -2.17477337e-01 -6.35974109e-01
-8.19433093e-01 2.84070432e-01 -3.96795660e-01 -6.23523951e-01
-3.69311690e-01 -1.19463193e+00 -5.19355536e-01 1.22618258e+00
-1.27856767e+00 -9.35232043e-01 3.96044999e-01 -4.99919951e-02
2.03320026e-01 -4.97906148e-01 1.18000042e+00 -1.78827196e-01
-3.96916151e-01 5.40457308e-01 9.47194040e-01 -6.28474951e-01
7.22168326e-01 -1.42688656e+00 5.39044917e-01 1.43547982e-01
-2.86422640e-01 9.95595396e-01 1.42435265e+00 -8.56012523e-01
-1.51421845e+00 -8.72980952e-01 6.98486269e-01 -4.51469541e-01
3.57275993e-01 -1.83410898e-01 -1.43377507e+00 4.38490480e-01
2.35716894e-01 -2.23464713e-01 1.27023113e+00 1.06933318e-01
-1.71534672e-01 4.75599736e-01 -1.23778665e+00 2.06696987e-01
6.90900624e-01 -1.91497341e-01 -5.14197528e-01 5.93486428e-01
4.23553765e-01 -3.87180239e-01 -1.16367412e+00 3.08926761e-01
9.37724352e-01 -1.30273902e+00 1.13415539e+00 -1.38606358e+00
1.54656395e-01 -1.39113277e-01 1.86307803e-02 -9.68059421e-01
-6.20986462e-01 -7.48119652e-01 -4.13169652e-01 2.20656097e-01
5.92381477e-01 -6.21247828e-01 1.41880047e+00 4.50448751e-01
-2.18119815e-01 -1.06055522e+00 -9.94762719e-01 -7.04146683e-01
5.69483221e-01 -3.51207666e-02 6.08233392e-01 4.92764235e-01
-1.98390596e-02 3.35684419e-01 -4.06791657e-01 -2.09252968e-01
3.93571824e-01 3.39421690e-01 5.74921131e-01 -1.35644281e+00
-3.89122099e-01 -2.20960140e-01 -2.70857334e-01 -7.92587101e-01
4.29389238e-01 -8.66683960e-01 -4.07548964e-01 -1.19108093e+00
4.74203438e-01 1.12808734e-01 -3.66634756e-01 2.70609856e-01
2.43654802e-01 1.16376594e-01 -4.13257778e-01 3.49736989e-01
-3.01764637e-01 6.48628891e-01 1.03722739e+00 1.05509058e-01
-2.10463136e-01 -1.40941667e-03 -3.80309552e-01 7.87424028e-01
8.63755345e-01 -5.52673697e-01 3.16346884e-02 5.08938849e-01
5.21653831e-01 -2.65861630e-01 3.15255105e-01 -8.05289805e-01
-3.84223491e-01 -3.62493575e-01 9.11746681e-01 -8.97978067e-01
2.97871858e-01 -9.05072868e-01 4.87433225e-01 7.40740120e-01
-3.02703470e-01 1.84751421e-01 8.71019661e-02 5.39453447e-01
1.47167116e-01 8.71581137e-02 1.06014633e+00 -3.59177321e-01
-5.60413122e-01 3.28257889e-01 -8.52921426e-01 -5.79097986e-01
1.00966597e+00 -6.07690156e-01 -2.05141213e-02 7.94161260e-02
-1.18432891e+00 -3.12385052e-01 9.94149327e-01 -1.07977420e-01
7.43649721e-01 -1.02647519e+00 -3.40194374e-01 3.37065756e-01
3.66048366e-01 -4.55085933e-01 2.45440081e-01 7.62469709e-01
-1.34571314e+00 8.15059006e-01 -4.37339455e-01 -6.37019455e-01
-1.57409179e+00 8.14083040e-01 6.07885063e-01 -1.96469888e-01
-7.33040094e-01 5.06630421e-01 2.73012836e-02 -3.95950615e-01
-1.81917902e-02 -1.85311645e-01 2.70629674e-01 -3.94484431e-01
4.89864230e-01 9.23048006e-04 4.17757064e-01 -7.26936579e-01
-5.72487235e-01 4.57957685e-01 -5.37890613e-01 7.74750173e-01
1.58978879e+00 2.08254412e-01 4.28942814e-02 4.98222299e-02
1.45002329e+00 -5.44475317e-01 -1.26353991e+00 7.28583038e-02
8.61296058e-02 1.93472013e-01 -1.69210285e-01 -1.23660004e+00
-2.22682431e-01 9.15184379e-01 9.97791529e-01 -4.61976945e-01
7.58312702e-01 -8.47641528e-02 9.20256376e-01 1.33879125e+00
3.20206493e-01 -9.24927413e-01 -5.54840386e-01 5.98682821e-01
7.86033154e-01 -1.66572452e+00 2.43931442e-01 3.29509586e-01
-3.44308078e-01 1.59577692e+00 2.70227879e-01 -1.11954913e-01
1.71985433e-01 -3.94280106e-02 2.93332320e-02 -5.32437861e-01
-9.04720545e-01 1.36359438e-01 2.44244635e-01 8.26877534e-01
1.01642168e+00 -7.49443024e-02 -3.57880443e-01 1.16743661e-01
2.25157011e-02 -5.04523441e-02 2.30414525e-01 8.28933537e-01
-8.41015816e-01 -1.66394365e+00 -2.95895010e-01 2.99571514e-01
-7.41885364e-01 -3.03193480e-01 -9.40743327e-01 6.73393250e-01
-9.99571607e-02 7.20909983e-02 -2.43664652e-01 -6.12489693e-02
-3.90545605e-03 6.71322644e-01 8.63178790e-01 -2.46629968e-01
-8.22887897e-01 -4.84404415e-02 -7.01762214e-02 -6.80042326e-01
-3.04704994e-01 -1.89994365e-01 -1.59902716e+00 -6.35912478e-01
-4.08717811e-01 3.25102597e-01 7.66149402e-01 8.53096604e-01
1.02610350e+00 1.73672825e-01 2.91477650e-01 -9.64039207e-01
-4.11989897e-01 -8.13587129e-01 -4.06360596e-01 3.26502889e-01
4.73492801e-01 -5.12490094e-01 1.43724322e-01 -1.05536275e-01] | [4.759714603424072, 5.583127975463867] |
5ce2b220-c648-4926-89a6-f3cf69c551ad | supervised-machine-learning-models-for | null | null | https://link.springer.com/article/10.1007/s42979-020-00394-7 | https://rdcu.be/cSbRX | Supervised Machine Learning Models for Prediction of COVID-19 Infection using Epidemiology Dataset | COVID-19 or 2019-nCoV is no longer pandemic but rather endemic, with more than 651,247 people around world having lost their lives after contracting the disease. Currently, there is no specific treatment or cure for COVID-19, and thus living with the disease and its symptoms is inevitable. This reality has placed a massive burden on limited healthcare systems worldwide especially in the developing nations. Although neither an effective, clinically proven antiviral agents' strategy nor an approved vaccine exist to eradicate the COVID-19 pandemic, there are alternatives that may reduce the huge burden on not only limited healthcare systems but also the economic sector; the most promising include harnessing non-clinical techniques such as machine learning, data mining, deep learning and other artificial intelligence. These alternatives would facilitate diagnosis and prognosis for 2019-nCoV pandemic patients. Supervised machine learning models for COVID-19 infection were developed in this work with learning algorithms which include logistic regression, decision tree, support vector machine, naive Bayes, and artificial neutral network using epidemiology labeled dataset for positive and negative COVID-19 cases of Mexico. The correlation coefficient analysis between various dependent and independent features was carried out to determine a strength relationship between each dependent feature and independent feature of the dataset prior to developing the models. The 80% of the training dataset were used for training the models while the remaining 20% were used for testing the models. The result of the performance evaluation of the models showed that decision tree model has the highest accuracy of 94.99% while the Support Vector Machine Model has the highest sensitivity of 93.34% and Naïve Bayes Model has the highest specificity of 94.30%. | ['Chinmay Chakraborty & I. A. Mohammed', 'Abdulkadir Ahmad', 'Sani Sharif Usman', 'Ebrahem A. Algehyne', 'L. J. Muhammad'] | 2020-11-27 | null | null | null | springer-nature-singapore-pte-ltd-2020-2020 | ['epidemiology'] | ['medical'] | [-2.35935241e-01 -2.69089013e-01 -1.59069136e-01 -7.96783641e-02
2.50933707e-01 -5.00917733e-01 3.64942700e-01 4.52214003e-01
-6.37089849e-01 1.05085433e+00 5.02057513e-03 -5.25694609e-01
-1.92623168e-01 -7.41196930e-01 -1.44283503e-01 -7.30963767e-01
-2.58823514e-01 1.04072380e+00 1.35987597e-02 -2.17040613e-01
1.52778327e-01 4.77104008e-01 -1.19385552e+00 -2.69823950e-02
1.14634001e+00 6.87431693e-01 2.08855033e-01 5.85147202e-01
3.85395028e-02 4.60812628e-01 -5.83191633e-01 -6.19906466e-03
3.54041249e-01 -4.22296971e-01 -2.06990853e-01 -8.66670310e-01
-7.83275783e-01 -5.08294225e-01 3.70898426e-01 6.75904930e-01
5.03136337e-01 -3.43975842e-01 1.21715271e+00 -1.26763391e+00
-5.49968719e-01 1.17542177e-01 -6.28210604e-01 3.08206826e-01
2.05050647e-01 2.14414477e-01 2.68717349e-01 -7.00484157e-01
7.62689471e-01 9.90677059e-01 9.62068677e-01 4.95643258e-01
-7.91752577e-01 -9.44764435e-01 -5.93051016e-01 6.84723035e-02
-1.32440519e+00 1.65979385e-01 1.67081088e-01 -1.06412172e+00
1.27130103e+00 -1.01221018e-01 1.06227148e+00 7.21713960e-01
7.37576008e-01 -1.93458602e-01 1.21122718e+00 -1.76657796e-01
1.66763961e-01 5.51292896e-01 1.99014932e-01 5.80128729e-01
8.65072548e-01 3.28214973e-01 1.57071948e-01 -5.67454815e-01
4.79418695e-01 5.43316364e-01 2.49086674e-02 -7.81394169e-02
-7.16393709e-01 1.33326411e+00 6.29758984e-02 4.84646469e-01
-7.01285779e-01 -6.37857735e-01 4.92324829e-01 2.86569834e-01
3.57886642e-01 3.13477159e-01 -8.92813027e-01 1.88573137e-01
-6.62616014e-01 1.87433496e-01 4.97874707e-01 1.19984999e-01
4.96339858e-01 2.05628440e-01 1.71362147e-01 5.25331140e-01
5.92632592e-01 1.03628862e+00 4.42935765e-01 -1.78015471e-01
-5.17160669e-02 8.09256792e-01 1.72827065e-01 -1.30350316e+00
-8.29768121e-01 -2.58697093e-01 -1.01020050e+00 -4.57808822e-02
1.94863886e-01 -7.44202018e-01 -1.05751562e+00 1.67638803e+00
3.02844107e-01 1.14368394e-01 3.14296573e-01 7.07948267e-01
5.63579857e-01 1.12412620e+00 2.38789216e-01 -5.37089050e-01
1.44930267e+00 -2.68404543e-01 -6.17327392e-01 1.83953255e-01
5.67381799e-01 -4.76638436e-01 3.74193996e-01 2.09504083e-01
-4.37941551e-01 -1.59281790e-01 -1.04025292e+00 6.57547116e-01
-6.27651155e-01 -3.08201313e-01 4.88713205e-01 9.38809335e-01
-7.74033487e-01 9.56084952e-02 -7.47344673e-01 -8.74127150e-01
2.44674712e-01 5.56546271e-01 -3.54359984e-01 -7.08931759e-02
-1.38518083e+00 1.20839441e+00 5.73301494e-01 -1.15988187e-01
-7.35013843e-01 -6.26642406e-01 -4.06416208e-01 -2.02090874e-01
-3.43186885e-01 -8.24729800e-01 3.24265331e-01 -6.24459386e-01
-8.56346250e-01 8.33583832e-01 -8.84201899e-02 -4.12795305e-01
2.58057863e-01 7.21395239e-02 -6.58391297e-01 -6.34145364e-02
9.23686996e-02 5.17445624e-01 3.53751421e-01 -9.86160278e-01
-7.50935674e-01 -7.87459433e-01 -4.64403242e-01 -9.96302217e-02
-1.43110588e-01 5.04662216e-01 3.55199337e-01 -2.56701708e-01
-3.41563344e-01 -1.15748882e+00 -2.02628061e-01 -6.94601595e-01
-6.70378953e-02 -2.85426944e-01 1.01060486e+00 -8.35676074e-01
8.76815796e-01 -1.91941476e+00 -3.34016174e-01 3.67390573e-01
1.13586962e-01 7.51881301e-01 1.34073392e-01 6.56777084e-01
1.36188507e-01 1.41066164e-01 -2.38558024e-01 5.91606498e-01
-4.84372586e-01 2.51798689e-01 -1.70407891e-01 5.11844516e-01
1.95174947e-01 5.03948927e-01 -8.58641505e-01 -2.57780790e-01
2.57312149e-01 8.82436872e-01 -5.74002028e-01 4.14159238e-01
-8.32381472e-02 6.97509706e-01 -4.93731827e-01 4.63951737e-01
8.64808738e-01 -2.90216953e-01 4.59689409e-01 3.92225266e-01
-7.62141645e-02 -3.14915121e-01 -6.16403878e-01 7.19825447e-01
7.01039732e-02 2.91648000e-01 -2.05763504e-01 -9.97222006e-01
1.06966364e+00 6.06785178e-01 5.26449919e-01 -5.10016024e-01
3.30270380e-01 3.55832636e-01 3.13806236e-01 -9.21501160e-01
6.02855571e-02 -4.58240926e-01 2.01872781e-01 2.35559687e-01
-3.20666939e-01 4.58707362e-01 -7.25605041e-02 -3.28013152e-02
8.05072308e-01 -2.37819552e-01 5.62774718e-01 -3.42342615e-01
4.54718649e-01 4.98515904e-01 8.36550593e-01 3.24366450e-01
-2.79780716e-01 5.50969914e-02 3.87700856e-01 -6.00783646e-01
-8.58233154e-01 -1.03255761e+00 -4.39187586e-01 7.63496757e-01
-1.31472409e-01 2.23412067e-01 -6.83644176e-01 -1.78158879e-01
-2.47881748e-03 5.69162965e-01 -4.05059457e-01 -3.49866748e-02
-2.39183545e-01 -1.21794927e+00 6.08045280e-01 -8.16511922e-03
4.58676487e-01 -9.97741342e-01 -8.81951809e-01 7.94448704e-03
7.57258907e-02 -5.26414573e-01 3.31328750e-01 1.63439587e-01
-9.18412149e-01 -1.30786133e+00 -5.62044263e-01 -8.82692695e-01
5.72527111e-01 -1.09207720e-01 5.29850662e-01 2.10832775e-01
-2.99905539e-01 -2.24479362e-01 -2.08244532e-01 -8.58616769e-01
-5.20901978e-01 -2.17354253e-01 6.02260947e-01 -4.03524160e-01
7.70841897e-01 -2.91941494e-01 -7.75982797e-01 -8.70360155e-03
-5.95101535e-01 -3.13313723e-01 5.89092135e-01 6.22654796e-01
2.46795043e-01 1.18680611e-01 1.18136668e+00 -8.43549609e-01
5.52883983e-01 -1.14874506e+00 -4.68290061e-01 -2.83551440e-02
-9.86103058e-01 -5.07371604e-01 6.25235081e-01 -2.11565420e-01
-8.75007272e-01 -2.48180732e-01 -2.76738703e-01 1.83924474e-02
-1.59106717e-01 7.09149420e-01 3.32853705e-01 5.89344144e-01
5.32326341e-01 -4.54384647e-02 1.81985646e-01 -2.71697998e-01
-2.02210873e-01 1.02558565e+00 1.56512007e-01 1.62124664e-01
6.10360980e-01 3.60986650e-01 -1.21970244e-01 -1.07515609e+00
-2.22633839e-01 -4.73418772e-01 -2.62899131e-01 -1.86239153e-01
1.43307853e+00 -9.34081733e-01 -7.97936141e-01 6.72297955e-01
-1.18530142e+00 2.50387996e-01 4.02520865e-01 8.69256973e-01
6.97516054e-02 -2.01658651e-01 -4.87537384e-01 -9.88365173e-01
-6.68026447e-01 -9.87011075e-01 2.02494666e-01 3.62547725e-01
-4.96080905e-01 -1.05060065e+00 7.02053368e-01 4.14035469e-01
7.99606025e-01 4.84891802e-01 1.45309448e+00 -1.25045812e+00
-6.95838556e-02 -1.90347657e-01 -3.05703670e-01 1.80147111e-01
4.00136232e-01 1.26369774e-01 -7.26518929e-01 -4.55459684e-01
1.54521391e-01 -3.23210984e-01 5.03340125e-01 7.32100368e-01
1.04903981e-01 -3.31266105e-01 -5.70760548e-01 3.73149067e-01
1.59771907e+00 1.19227147e+00 4.51376379e-01 1.65642440e-01
3.54287773e-01 6.78152204e-01 4.89902139e-01 4.41869050e-01
3.83645386e-01 2.50363886e-01 3.33762407e-01 -6.29796684e-02
6.33700132e-01 -1.02039166e-01 1.70424253e-01 8.50176752e-01
-4.68600780e-01 -2.32011288e-01 -1.34719372e+00 3.82112265e-01
-1.41656518e+00 -1.15864432e+00 -2.96442032e-01 2.16692781e+00
5.67088783e-01 -1.36314541e-01 1.72160670e-01 -5.05321240e-03
6.95881248e-01 -3.69576365e-01 -5.65489709e-01 -7.53427863e-01
-7.73099065e-02 8.12208354e-02 2.22456336e-01 1.50525317e-01
-9.35572207e-01 4.80127871e-01 6.39892483e+00 2.63562560e-01
-1.41620564e+00 4.13265824e-03 6.35983825e-01 3.53087872e-01
1.93685926e-02 -1.06171109e-01 -5.05524337e-01 4.93532032e-01
1.18146658e+00 -4.52622324e-02 2.91071802e-01 6.20687664e-01
5.56680560e-01 -1.30039334e-01 -3.85903418e-01 6.61921799e-01
3.64984646e-02 -1.25484753e+00 -2.83640120e-02 1.62519783e-01
7.60077000e-01 3.62858862e-01 -2.20819339e-01 1.84342667e-01
2.72012234e-01 -1.02751410e+00 -6.62694797e-02 6.11184001e-01
6.88270807e-01 -9.08980608e-01 1.31266642e+00 6.18536711e-01
-7.08745778e-01 -7.29899406e-02 -1.43735424e-01 -2.90776193e-01
1.42650723e-01 4.20648128e-01 -1.22237146e+00 1.31354153e-01
7.59803712e-01 2.17019394e-01 -1.92924455e-01 7.20494688e-01
1.90222815e-01 7.08782852e-01 -2.28145733e-01 -2.51504362e-01
1.94838960e-02 -4.74379271e-01 4.04060900e-01 1.17368150e+00
4.38078195e-01 1.64427832e-01 -2.90653110e-02 4.33471680e-01
5.60177565e-01 3.95199060e-01 -1.00163674e+00 -2.49836475e-01
4.50455904e-01 7.92688668e-01 -8.95553708e-01 -2.51050055e-01
-3.76999795e-01 4.48265612e-01 -1.23931184e-01 2.04042077e-01
-8.27590942e-01 -3.55007172e-01 6.33433282e-01 2.94737130e-01
1.50998026e-01 1.71004087e-01 -1.71211079e-01 -6.55713499e-01
-7.02337146e-01 -9.42098439e-01 5.63702226e-01 -5.46770573e-01
-1.31361175e+00 6.59473062e-01 8.76891017e-02 -8.05419743e-01
-6.21523023e-01 -6.03467822e-01 -5.78617454e-01 1.01878142e+00
-1.06146789e+00 -9.87531483e-01 7.14067370e-02 4.49890584e-01
-9.71294567e-02 -5.52166164e-01 1.17848563e+00 1.85225993e-01
-4.92734998e-01 1.28517583e-01 4.70416665e-01 3.51728722e-02
4.38507497e-01 -6.46491706e-01 -4.04320717e-01 4.54649329e-01
-8.38521481e-01 9.88596678e-01 6.65220141e-01 -1.12025428e+00
-1.13980377e+00 -9.51430738e-01 1.21344662e+00 -9.53079388e-02
3.38025302e-01 -1.35532722e-01 -7.35624433e-01 5.75289249e-01
4.19362664e-01 -5.72398067e-01 1.05304921e+00 -3.51762146e-01
-2.16979161e-01 1.31489202e-01 -1.73068380e+00 3.19764763e-01
2.36787185e-01 -9.92295891e-02 -7.02889085e-01 3.05231005e-01
5.84819794e-01 2.43134573e-01 -8.42043579e-01 6.07070684e-01
8.66102934e-01 -1.02303493e+00 8.30348134e-01 -9.29056168e-01
1.19598284e-01 -3.54717106e-01 -1.85100242e-01 -1.06274474e+00
-1.59023300e-01 -1.14254653e-01 3.94242287e-01 1.00906515e+00
6.69958889e-01 -1.21590781e+00 5.08008420e-01 2.59960115e-01
3.22399408e-01 -8.67489576e-01 -7.26154804e-01 -5.45138776e-01
2.41162181e-01 -9.29063484e-02 5.58996081e-01 1.23474407e+00
-1.36803672e-01 4.47396994e-01 -4.82151657e-01 9.12708789e-02
2.93258965e-01 -4.36663590e-02 5.67173839e-01 -1.43290353e+00
2.65481565e-02 -1.56883039e-02 -5.37350178e-01 5.03338873e-02
-2.27270961e-01 -7.26122737e-01 -4.89835232e-01 -1.68707967e+00
6.32348657e-01 -7.40346968e-01 -3.24278206e-01 4.91142601e-01
1.24672651e-01 1.39160529e-01 -6.93015084e-02 3.26753974e-01
2.35912621e-01 7.92524368e-02 8.07859719e-01 8.00514594e-02
-5.16649425e-01 5.34464754e-02 -5.31286240e-01 7.64639020e-01
1.18285930e+00 -6.78228378e-01 -7.34257817e-01 3.35860322e-03
3.80868167e-01 1.77281633e-01 -9.72230174e-03 -6.12028956e-01
-1.58301279e-01 -4.59719419e-01 6.83998585e-01 -8.46633375e-01
1.49419218e-01 -8.50790858e-01 7.63124585e-01 1.32061672e+00
3.73352200e-01 4.16722655e-01 1.47614881e-01 3.38737190e-01
1.46005511e-01 -6.60007447e-02 9.51718628e-01 1.99565157e-01
-2.46324807e-01 6.75514340e-02 -8.76932442e-01 1.06965147e-01
1.49075115e+00 -2.58496791e-01 -4.55037713e-01 -1.40019909e-01
-3.85152668e-01 1.85744673e-01 4.13936257e-01 3.54696035e-01
5.47586560e-01 -7.68397748e-01 -8.45223367e-01 4.28853929e-01
-1.03323504e-01 -5.19134879e-01 2.59099782e-01 8.77379835e-01
-1.02465415e+00 8.11014473e-01 -7.77579248e-01 -5.05451739e-01
-1.07097983e+00 8.80174339e-01 1.24180250e-01 -2.78946102e-01
-2.45858759e-01 2.50535548e-01 -2.56541967e-02 -5.36440790e-01
-1.52603447e-01 1.03476651e-01 -7.58188188e-01 1.97165966e-01
5.04711270e-01 5.60435891e-01 -2.10378945e-01 -9.55115438e-01
-8.60443771e-01 5.58091223e-01 6.62562177e-02 1.62812442e-01
1.44696605e+00 2.23513052e-01 -5.10625362e-01 3.56419325e-01
9.13910627e-01 -3.22694592e-02 -2.99314946e-01 4.25471395e-01
-7.26946443e-02 -1.61622893e-02 -3.19068283e-01 -1.09270883e+00
-7.41134346e-01 5.31376362e-01 1.01663971e+00 1.26383334e-01
1.07089972e+00 -2.56168216e-01 6.77753627e-01 1.41530439e-01
3.69222701e-01 -6.17444634e-01 -5.80471992e-01 4.69840288e-01
4.59906071e-01 -1.23681188e+00 -1.60570398e-01 8.39081332e-02
-5.01812637e-01 5.53227067e-01 2.36649483e-01 -1.11746468e-01
9.81582105e-01 3.88017297e-01 3.78690392e-01 -2.95229435e-01
-7.73631036e-01 2.71623343e-01 -2.66936451e-01 1.04207349e+00
4.92938757e-01 3.70153248e-01 -8.72731686e-01 7.68400311e-01
-1.06598504e-01 2.88063645e-01 2.10258797e-01 7.48567104e-01
-4.84094799e-01 -8.24719310e-01 -4.91231382e-01 8.69405270e-01
-7.19861150e-01 -6.59793988e-02 -2.95270741e-01 9.69952464e-01
6.06545091e-01 1.14214301e+00 2.06890285e-01 -4.42807138e-01
-1.24753699e-01 2.72354126e-01 9.43900459e-03 -2.94500083e-01
-5.57616889e-01 -2.24063784e-01 -3.31757478e-02 2.04736188e-01
-3.99590433e-01 -6.64924562e-01 -1.75499403e+00 -6.93556249e-01
-1.77879468e-01 3.62922758e-01 8.82648647e-01 9.27598834e-01
4.03156638e-01 -9.95203853e-02 6.40398145e-01 -4.11414474e-01
-3.26754034e-01 -1.06139314e+00 -6.13321126e-01 1.54955149e-01
2.55749464e-01 -6.09115481e-01 -3.05669248e-01 -1.01889312e-01] | [5.535154342651367, 4.955984115600586] |
4fca2fe9-9113-44e3-823a-c0d10d58e62c | faspell-a-fast-adaptable-simple-powerful | null | null | https://aclanthology.org/D19-5522 | https://aclanthology.org/D19-5522.pdf | FASPell: A Fast, Adaptable, Simple, Powerful Chinese Spell Checker Based On DAE-Decoder Paradigm | We propose a Chinese spell checker {--} FASPell based on a new paradigm which consists of a denoising autoencoder (DAE) and a decoder. In comparison with previous state-of-the-art models, the new paradigm allows our spell checker to be Faster in computation, readily Adaptable to both simplified and traditional Chinese texts produced by either humans or machines, and to require much Simpler structure to be as much Powerful in both error detection and correction. These four achievements are made possible because the new paradigm circumvents two bottlenecks. First, the DAE curtails the amount of Chinese spell checking data needed for supervised learning (to {\textless}10k sentences) by leveraging the power of unsupervisedly pre-trained masked language model as in BERT, XLNet, MASS etc. Second, the decoder helps to eliminate the use of confusion set that is deficient in flexibility and sufficiency of utilizing the salient feature of Chinese character similarity. | ['Xianguo Yu', 'Junhui Liu', 'Yuzhong Hong', 'Nan Liu', 'Neng He'] | 2019-11-01 | null | null | null | ws-2019-11 | ['chinese-spell-checking'] | ['natural-language-processing'] | [ 1.93943352e-01 -1.38871655e-01 5.25090754e-01 -1.02234662e-01
-6.33539319e-01 -4.55296606e-01 3.77500534e-01 2.45222241e-01
-7.50290513e-01 7.76699066e-01 1.78938568e-01 -5.48096180e-01
2.44234696e-01 -7.58946896e-01 -5.75820446e-01 -5.19799173e-01
2.82367408e-01 2.56201625e-01 3.65074694e-01 -5.56877971e-01
4.18194950e-01 1.79173514e-01 -1.43755865e+00 4.13638353e-01
1.08112574e+00 7.66827345e-01 6.73319459e-01 6.23320520e-01
-2.41383672e-01 9.12494063e-01 -7.08485663e-01 -5.89652538e-01
1.30384117e-01 -4.94446248e-01 -5.46427667e-01 -1.45926401e-01
1.20991744e-01 -2.50950724e-01 -3.07540268e-01 1.37187850e+00
4.14874434e-01 -7.54399672e-02 5.34415007e-01 -3.69485408e-01
-1.09113276e+00 7.77557194e-01 -3.21609646e-01 3.52979712e-02
3.43821615e-01 -4.19182442e-02 6.74544215e-01 -7.07259834e-01
4.56377029e-01 8.03847075e-01 8.44135106e-01 7.88828373e-01
-7.12585509e-01 -5.94362438e-01 -5.40128648e-02 5.68542667e-02
-1.24937022e+00 -4.72693175e-01 4.06240165e-01 -2.55355388e-01
1.24618268e+00 2.14517638e-01 4.46405262e-01 1.08006501e+00
2.60773212e-01 8.58312368e-01 8.20373833e-01 -1.00779879e+00
1.93918422e-01 2.87621886e-01 1.69974416e-01 9.18206155e-01
3.60740274e-01 -5.90027720e-02 -3.51743698e-01 1.37894094e-01
5.97390771e-01 3.53277922e-02 -3.80746454e-01 2.41075933e-01
-9.69876230e-01 6.64987803e-01 -8.88175741e-02 8.48274291e-01
-1.32662416e-01 -8.04727003e-02 4.16849285e-01 5.62604666e-01
2.28160664e-01 6.00870192e-01 -5.70677817e-01 -4.13123488e-01
-1.29131627e+00 -1.63446754e-01 7.41129279e-01 1.16086626e+00
5.52658796e-01 4.46136922e-01 2.38344371e-01 5.36716700e-01
2.24375740e-01 5.38049877e-01 9.68296051e-01 -4.14710760e-01
4.92048413e-01 7.25225806e-01 3.81162413e-03 -8.37422729e-01
-2.05232427e-01 -5.69743454e-01 -9.69754994e-01 7.91054890e-02
2.94565946e-01 -3.26119006e-01 -8.73444438e-01 1.37773442e+00
-3.91273886e-01 -3.21883619e-01 1.32654682e-02 4.39492404e-01
3.41468245e-01 6.50763988e-01 -1.75990149e-01 -7.28393951e-03
1.19484603e+00 -9.31000888e-01 -8.35358024e-01 -1.83269903e-01
8.89374614e-01 -8.95917118e-01 1.20424163e+00 6.99479342e-01
-1.08851898e+00 -5.81118345e-01 -1.32495880e+00 -1.98278040e-01
-5.88450730e-01 4.68827844e-01 4.86028761e-01 9.81645346e-01
-1.08054125e+00 6.49472952e-01 -6.51006162e-01 -1.67197987e-01
9.23884660e-02 2.72821069e-01 -4.16355908e-01 8.23410302e-02
-1.15807188e+00 1.02351868e+00 6.02591276e-01 2.33823046e-01
-6.99159086e-01 -3.03181827e-01 -8.27924132e-01 3.50436717e-01
1.89221263e-01 -2.91621238e-01 1.21261001e+00 -1.34801304e+00
-1.64630675e+00 6.27553046e-01 -2.45921418e-01 -5.78405082e-01
5.58000207e-01 -5.15825212e-01 -8.82215858e-01 1.36023685e-01
-8.93244147e-02 1.67842999e-01 9.53761816e-01 -7.83175349e-01
-6.41827404e-01 -7.53760636e-02 -5.20255446e-01 -2.09710240e-01
-5.81107199e-01 3.44652086e-01 -5.32258630e-01 -9.38980699e-01
-1.21248782e-01 -6.22591197e-01 -1.99637488e-01 -3.85726482e-01
-2.85824955e-01 4.51510437e-02 8.40280294e-01 -9.49385405e-01
1.75586915e+00 -2.28701496e+00 -2.17297748e-01 1.32853627e-01
2.70347707e-02 8.86872232e-01 -5.35502806e-02 7.36811697e-01
6.26638830e-02 1.98932499e-01 -3.76040697e-01 -4.65522915e-01
-1.16526470e-01 1.07744388e-01 -1.74781099e-01 3.86165768e-01
2.56904989e-01 5.92902243e-01 -7.71023095e-01 -3.19008738e-01
-7.77311577e-03 1.57716170e-01 -5.99256217e-01 1.33286893e-01
-2.77382404e-01 7.68356100e-02 -4.33996230e-01 5.98033130e-01
6.53921664e-01 -1.72424436e-01 1.60668388e-01 4.34418797e-01
-3.69206935e-01 4.26649481e-01 -1.32151949e+00 1.70023715e+00
-2.40525231e-01 6.04891419e-01 2.17854362e-02 -8.57938170e-01
1.01129735e+00 4.64604199e-01 -1.94387227e-01 -6.64159477e-01
3.17937136e-01 3.77951443e-01 -1.61372840e-01 -4.08097655e-01
7.18490422e-01 4.96644229e-02 -1.26567662e-01 5.25158882e-01
9.28686708e-02 4.58625369e-02 2.58883834e-01 2.21405834e-01
9.71459627e-01 7.11060166e-02 2.36058235e-01 -4.93465334e-01
7.30057359e-01 -5.43723293e-02 6.85975313e-01 1.06715548e+00
-6.56861737e-02 4.72321093e-01 3.53177369e-01 -4.63141948e-01
-1.15523255e+00 -7.59619594e-01 1.18569136e-01 9.23567832e-01
-2.58364141e-01 -4.18602854e-01 -1.05658162e+00 -4.00527507e-01
-4.02907878e-01 7.26551235e-01 -3.72661740e-01 -5.22872843e-02
-7.69137621e-01 -5.41911900e-01 1.02525377e+00 4.97252375e-01
6.97983325e-01 -1.04995823e+00 -4.17922258e-01 3.73275310e-01
1.03887349e-01 -1.00493741e+00 -4.67128575e-01 5.33231020e-01
-8.62665653e-01 -7.46708095e-01 -5.90116680e-01 -1.21990788e+00
5.76277018e-01 -6.90281466e-02 9.64506507e-01 4.92118657e-01
-9.42856967e-02 -1.32716432e-01 -7.50547767e-01 -5.96266150e-01
-8.37065756e-01 7.53891170e-02 1.83173269e-01 -1.28637969e-01
7.38553464e-01 -4.26850498e-01 -1.60135955e-01 -1.78128630e-01
-1.01516283e+00 -2.48782113e-02 7.14633167e-01 9.27876055e-01
1.30466819e-01 1.80061281e-01 5.02233446e-01 -9.71082628e-01
6.89714074e-01 -9.75635722e-02 -8.36405814e-01 3.95329088e-01
-7.35843658e-01 2.40365937e-01 9.35989916e-01 -1.91340491e-01
-1.19545257e+00 2.17519235e-02 -3.38166684e-01 2.26398557e-01
-8.76946747e-02 2.91417569e-01 -7.62843788e-02 -1.10948250e-01
6.83483720e-01 6.77515566e-01 -9.49435905e-02 -8.89594495e-01
2.41748467e-01 1.07453072e+00 6.29394650e-01 -2.87736714e-01
5.98786116e-01 1.25525042e-01 -4.97335732e-01 -8.87518942e-01
-3.65705103e-01 -2.28382245e-01 -7.78111875e-01 1.80757448e-01
8.45264137e-01 -9.55681443e-01 -6.70708418e-01 6.76988542e-01
-1.13730001e+00 -1.18349316e-02 5.84785752e-02 4.63777363e-01
-1.54028684e-01 8.94557059e-01 -1.12516963e+00 -6.43940449e-01
-5.06973982e-01 -1.04239118e+00 6.66492820e-01 2.60038555e-01
2.75579211e-03 -7.17028379e-01 -1.22226760e-01 1.42802432e-01
5.49367011e-01 -2.46060252e-01 1.05751085e+00 -7.08487570e-01
-6.02274179e-01 -5.78310251e-01 -1.59432307e-01 9.82316732e-01
2.15332173e-02 -6.24505756e-03 -8.61621082e-01 -2.48674005e-01
2.96377540e-01 -2.63402373e-01 8.64305854e-01 -1.25810623e-01
1.08147776e+00 -2.55797118e-01 -2.82031531e-03 4.76327926e-01
1.56747210e+00 4.88705099e-01 7.00852990e-01 3.08667749e-01
3.88419181e-01 7.33929127e-02 6.52583614e-02 3.94006580e-01
1.29860654e-01 9.76356715e-02 6.52477369e-02 5.99612892e-02
-1.75007936e-02 -3.13473105e-01 4.52326387e-01 1.49229157e+00
-8.03702138e-03 -3.72593373e-01 -9.13620830e-01 5.74934542e-01
-1.64359021e+00 -9.17491138e-01 -1.99594289e-01 1.96318769e+00
1.04088545e+00 1.50398672e-01 -3.63847613e-01 4.03581381e-01
6.16465032e-01 -1.30406022e-01 1.84158936e-01 -7.56243289e-01
-3.08426321e-01 4.48424786e-01 4.49476629e-01 3.75852436e-01
-1.05556822e+00 1.16088402e+00 6.84614277e+00 9.20785904e-01
-9.27771628e-01 -2.78562587e-02 2.93963075e-01 3.27178419e-01
-2.48222217e-01 -5.77023514e-02 -8.52684200e-01 6.65338278e-01
9.95877087e-01 4.22756612e-01 5.11196911e-01 8.09399247e-01
2.44371220e-02 -3.08762699e-01 -9.22349215e-01 6.91582024e-01
2.65865058e-01 -1.52034795e+00 6.46549910e-02 -3.86806577e-01
6.53953075e-01 1.08727455e-01 -8.89368802e-02 4.12208200e-01
3.83437514e-01 -9.15768385e-01 7.83963799e-01 5.53516030e-01
6.39184237e-01 -6.64845705e-01 9.42132115e-01 6.60969615e-01
-6.59151852e-01 -1.22881986e-01 -4.52048957e-01 -3.38254809e-01
-1.61897838e-02 6.00247443e-01 -4.50900733e-01 4.42449808e-01
6.33840621e-01 2.45091915e-01 -5.96312046e-01 7.95579851e-01
-3.19306135e-01 8.16804349e-01 -1.04247063e-01 -5.48590243e-01
5.57060063e-01 -8.55038464e-02 1.94836676e-01 1.72118115e+00
3.64737958e-01 -5.34998141e-02 -1.53164744e-01 5.43126702e-01
9.20498446e-02 1.07330922e-02 -4.21470821e-01 -2.72773623e-01
4.42476183e-01 6.81735456e-01 -5.05785644e-01 -4.21862483e-01
-6.09244168e-01 1.29022431e+00 3.56070548e-01 1.58857405e-01
-4.55787569e-01 -8.57181370e-01 3.63251008e-02 -1.71528131e-01
6.61860526e-01 -4.06209350e-01 -3.86520982e-01 -1.48776197e+00
3.77377607e-02 -1.31652105e+00 -2.24348679e-02 -5.44690609e-01
-1.09845209e+00 9.53484058e-01 -7.02012479e-01 -1.00391042e+00
-3.07070285e-01 -9.81254578e-01 -4.00168300e-01 9.18537617e-01
-1.38454890e+00 -9.11243618e-01 9.01126266e-02 6.80950761e-01
4.97839123e-01 -5.51294804e-01 1.14616847e+00 6.65282428e-01
-5.46073675e-01 8.35778654e-01 5.25356233e-01 6.26214981e-01
6.25326991e-01 -1.14971936e+00 4.37293857e-01 1.28593171e+00
2.79551148e-01 7.49679625e-01 5.03157437e-01 -5.91503739e-01
-1.35244930e+00 -6.47852957e-01 1.40145659e+00 -4.50083971e-01
6.62722349e-01 -6.04584217e-01 -7.05364823e-01 5.63799143e-01
1.58968583e-01 -4.49415565e-01 6.09386206e-01 3.74159776e-02
-3.42891783e-01 -1.52574152e-01 -8.59556198e-01 6.88160658e-01
6.12557888e-01 -7.19725192e-01 -8.44593108e-01 6.07040413e-02
6.46859825e-01 -1.16765879e-01 -3.45397294e-01 -1.17567768e-02
2.59025007e-01 -1.08338368e+00 3.67314041e-01 -5.78291714e-01
3.70560884e-01 -3.57643723e-01 -1.98626265e-01 -1.00147009e+00
-4.95178640e-01 -1.03757405e+00 1.59620360e-01 1.10622573e+00
6.29451752e-01 -3.55526268e-01 5.90797305e-01 2.20290676e-01
-3.69779736e-01 -3.12120914e-01 -8.75034630e-01 -6.70521975e-01
5.80794178e-02 -4.28429037e-01 6.69594407e-01 9.04798567e-01
1.73556447e-01 2.76773907e-02 -6.73617184e-01 -1.83616467e-02
2.77285695e-01 -1.23675272e-01 3.01292807e-01 -1.01691508e+00
-6.54343963e-01 -3.47596765e-01 -1.66214898e-01 -1.05366373e+00
-1.83186188e-01 -7.05794811e-01 1.85677454e-01 -9.20062125e-01
-5.65400608e-02 -3.79163861e-01 -3.29524428e-01 4.73318875e-01
-1.11503810e-01 -3.00656795e-03 3.36587608e-01 -1.10971659e-01
-4.49896634e-01 2.32700467e-01 9.50415194e-01 1.48909494e-01
5.17201275e-02 -8.76164809e-02 -7.54375219e-01 9.34562981e-01
6.44895732e-01 -6.32445514e-01 -5.59947453e-02 -1.07764125e+00
3.69025260e-01 -7.08319917e-02 -1.62652895e-01 -1.12694466e+00
4.40970540e-01 3.17808658e-01 4.28394496e-01 -4.04379278e-01
-1.25051767e-01 -7.77802825e-01 -3.27156216e-01 5.64698279e-01
-1.14703558e-01 5.23108959e-01 1.24578111e-01 5.18707275e-01
-3.98901761e-01 -7.66956151e-01 7.34481633e-01 -6.10602796e-01
-8.31954122e-01 -2.46203572e-01 -8.81482720e-01 -2.90897731e-02
5.67353070e-01 -3.76537204e-01 -8.29425454e-02 -2.18311802e-01
-3.61436486e-01 -1.48232818e-01 3.82423788e-01 2.75573641e-01
3.28043342e-01 -7.22975314e-01 -5.16779959e-01 6.21275842e-01
-1.39303774e-01 -2.40834177e-01 1.89994887e-01 4.09558654e-01
-1.17672038e+00 5.26308119e-01 -2.16503203e-01 -1.50790752e-03
-9.23211336e-01 6.31342292e-01 5.28285205e-02 -3.88147354e-01
-7.18609095e-01 1.01853597e+00 -3.98485422e-01 -1.41869128e-01
5.42932868e-01 -3.66993904e-01 1.10001847e-01 -3.23207319e-01
8.40093076e-01 2.02070430e-01 3.77932698e-01 -1.60748914e-01
-2.74501085e-01 2.48851493e-01 -3.00059289e-01 4.49269302e-02
1.44529235e+00 -4.23072884e-03 -2.78371692e-01 1.36507049e-01
8.12564671e-01 5.04350483e-01 -8.76533508e-01 -1.15136236e-01
3.50104511e-01 -7.23929629e-02 1.04684867e-01 -1.01353943e+00
-5.73694170e-01 9.72301006e-01 4.03871834e-01 1.32157290e-02
1.27850926e+00 -6.52125716e-01 1.05621386e+00 9.30991411e-01
3.24682146e-01 -1.57637465e+00 -1.13500386e-01 9.34627712e-01
3.54467303e-01 -1.03647935e+00 -2.42923871e-01 6.08418882e-02
-7.78075695e-01 1.29030025e+00 2.76621640e-01 -2.98813879e-01
6.80225670e-01 7.35915601e-01 1.94048896e-01 1.78936869e-01
-5.51067829e-01 -8.04408640e-03 3.95870432e-02 5.02129972e-01
4.81010616e-01 1.86884906e-02 -3.60551745e-01 9.10424650e-01
-1.82247013e-01 7.66103193e-02 7.04326928e-01 1.01136827e+00
-7.31777191e-01 -1.32737672e+00 -1.49507344e-01 1.26356095e-01
-7.90290475e-01 -5.61030209e-01 -1.45954311e-01 7.64290333e-01
3.94341171e-01 8.46651435e-01 3.19595672e-02 -3.59530330e-01
-7.09250197e-02 1.85409471e-01 2.33527407e-01 -4.92647260e-01
-1.02845049e+00 1.34779751e-01 1.36453249e-02 -2.33868420e-01
-4.80680214e-03 -3.92414242e-01 -1.00003099e+00 -3.41476679e-01
-4.95216608e-01 1.66530997e-01 5.71388245e-01 1.02564478e+00
3.67287368e-01 2.20653385e-01 2.47961462e-01 -2.65190244e-01
-8.00740957e-01 -1.00552416e+00 -8.07283342e-01 9.57462713e-02
3.84636998e-01 3.30776945e-02 -1.03262909e-01 8.29431936e-02] | [10.883604049682617, 10.676987648010254] |
fd1229ea-64e8-49c4-b41a-9599bb3b7fff | language-independent-approach-for | null | null | https://aclanthology.org/2022.coling-1.470 | https://aclanthology.org/2022.coling-1.470.pdf | Language-Independent Approach for Morphological Disambiguation | This paper presents a language-independent approach for morphological disambiguation which has been regarded as an extension of POS tagging, jointly predicting complex morphological tags. In the proposed approach, all words, roots, POS and morpheme tags are embedded into vectors, and contexts representations from surface word and morphological contexts are calculated. Then the inner products between analyses and the context’s representations are computed to perform the disambiguation. The underlying hypothesis is that the correct morphological analysis should be closer to the context in a vector space. Experimental results show that the proposed approach outperforms the existing models on seven different language datasets. Concretely, compared with the baselines of MarMot and a sophisticated neural model (Seq2Seq), the proposed approach achieves around 6% improvement in average accuracy for all languages while running about 6 and 33 times faster than MarMot and Seq2Seq, respectively. | ['Rustam Mussabayev', 'Gulmira Tolegen', 'Alymzhan Toleu'] | null | null | null | null | coling-2022-10 | ['morphological-analysis', 'morphological-disambiguation'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.55884994e-02 -3.78395170e-01 -8.74814689e-02 -2.96285540e-01
-5.63895583e-01 -9.82888579e-01 4.28990066e-01 5.95530570e-01
-1.10998237e+00 5.43063164e-01 4.51452404e-01 -5.51582932e-01
2.56217241e-01 -6.24612033e-01 -2.08757341e-01 -5.59455216e-01
-2.92888343e-01 1.99547321e-01 2.86112159e-01 -2.27647364e-01
4.78161871e-01 3.34876388e-01 -1.15149271e+00 1.36032745e-01
9.46229517e-01 7.50006378e-01 5.29637992e-01 4.00740802e-01
-7.20978200e-01 3.39873224e-01 -5.32591462e-01 -6.87178314e-01
8.24797079e-02 -2.26158742e-02 -7.76646435e-01 -3.94285053e-01
4.49349642e-01 4.51409012e-01 -5.42987091e-03 1.43888593e+00
4.19874787e-01 4.22314048e-01 5.96687496e-01 -2.96096951e-01
-1.03160048e+00 1.22256100e+00 -5.05191565e-01 6.67130291e-01
2.71620512e-01 -1.36626303e-01 1.60722935e+00 -1.36835384e+00
5.28326809e-01 1.26134920e+00 6.14372969e-01 3.74909222e-01
-9.54015255e-01 -4.58989054e-01 4.60761487e-01 2.82153547e-01
-1.39910007e+00 -5.18420748e-02 5.37374854e-01 -2.61099488e-01
1.41856265e+00 1.77875802e-01 4.68185961e-01 7.80805230e-01
3.10310364e-01 7.52280951e-01 1.14543760e+00 -5.88277996e-01
2.77131587e-01 -1.18071444e-01 6.29930317e-01 6.28348887e-01
3.23843122e-01 -1.13263160e-01 -4.79108751e-01 -1.61735967e-01
4.21907544e-01 -3.70844960e-01 5.91533370e-02 1.23442508e-01
-1.28684473e+00 8.32223535e-01 3.15471858e-01 7.49420524e-01
-5.55377662e-01 -2.17534602e-01 5.40509641e-01 -2.54063960e-02
3.16050291e-01 6.52849555e-01 -7.33438790e-01 5.24655357e-02
-7.95133829e-01 6.98836446e-02 6.60290003e-01 8.61011267e-01
5.54023683e-01 4.50937390e-01 -2.67069221e-01 1.00343144e+00
4.17995751e-01 5.50595641e-01 9.89810646e-01 -1.74254850e-01
6.00938857e-01 6.68380499e-01 -1.39226228e-01 -8.24976981e-01
-4.84534860e-01 -2.16325775e-01 -3.61673474e-01 -3.09010953e-01
3.56019258e-01 -2.80590028e-01 -8.53300273e-01 1.64192212e+00
3.60719532e-01 3.42663825e-01 2.69276530e-01 7.11287320e-01
9.55448091e-01 9.63451684e-01 6.54197097e-01 -1.98452428e-01
1.83447075e+00 -8.13519955e-01 -8.26627016e-01 -5.34300804e-01
4.93575960e-01 -1.03544462e+00 1.20213008e+00 4.35017735e-01
-1.06089628e+00 -6.68583512e-01 -1.20220363e+00 -6.01697750e-02
-7.35070050e-01 2.55690485e-01 7.76928186e-01 5.19584715e-01
-9.22245324e-01 6.03305697e-01 -6.50281370e-01 -4.99762118e-01
-1.42971501e-01 2.83532649e-01 -4.36677247e-01 4.12045091e-01
-1.26193249e+00 1.03770363e+00 1.18443048e+00 9.86706614e-02
-3.04200917e-01 -3.17955643e-01 -1.14218950e+00 -3.00016571e-02
1.30233169e-01 -1.69016615e-01 1.06645155e+00 -6.92277253e-01
-1.29920805e+00 8.19311023e-01 -2.24540979e-01 -5.36604941e-01
-3.88200790e-01 -4.74773020e-01 -6.80962622e-01 -2.53689736e-01
3.27639692e-02 4.91812348e-01 2.48370290e-01 -8.36144805e-01
-1.06071150e+00 -2.79468745e-01 -2.72098511e-01 3.06591868e-01
-2.97617495e-01 6.00900769e-01 -4.65227276e-01 -8.40717912e-01
2.39235282e-01 -6.59199715e-01 -4.31281269e-01 -8.39519978e-01
-2.00601295e-01 -7.51529396e-01 2.15332031e-01 -1.09079087e+00
1.51025999e+00 -2.15172529e+00 2.14903161e-01 -1.40505852e-02
-2.72159547e-01 6.80592954e-01 -3.66864473e-01 4.41752911e-01
9.99928415e-02 2.46278033e-01 -2.90436983e-01 -1.33100346e-01
1.48528367e-01 4.68102723e-01 -3.00606668e-01 3.03782940e-01
3.88133377e-01 8.72235537e-01 -1.06613171e+00 -6.59878016e-01
1.14834733e-01 3.41078609e-01 -4.08442885e-01 7.91493133e-02
-4.58103493e-02 -5.02654426e-02 -1.05815008e-01 7.75263667e-01
6.88933432e-01 5.68368852e-01 9.86322165e-01 1.53807372e-01
-3.92723739e-01 8.50787222e-01 -1.25428581e+00 1.56385684e+00
-5.68670094e-01 1.57918036e-01 -2.90554017e-01 -7.91662455e-01
1.21232867e+00 1.96781203e-01 -1.03423141e-01 -6.07197404e-01
2.36499682e-01 3.37709546e-01 2.05043197e-01 -3.32993835e-01
8.58837783e-01 -2.36564219e-01 -5.77480137e-01 -4.53727394e-02
4.95244861e-01 9.82144549e-02 3.59070987e-01 -2.53368169e-01
8.24870050e-01 5.46898425e-01 1.05582201e+00 -4.90790695e-01
7.10318089e-01 -2.08168581e-01 1.14159977e+00 3.36135179e-01
-1.88606769e-01 -1.29930064e-01 1.99571505e-01 -4.95850056e-01
-6.26030028e-01 -1.37332821e+00 -1.18111819e-01 1.39680767e+00
5.68364225e-02 -5.73297143e-01 -7.45418489e-01 -1.00433505e+00
-2.83172615e-02 8.85164618e-01 -4.84651417e-01 4.50588614e-01
-1.21420681e+00 -9.30420995e-01 7.17222393e-01 8.97700489e-01
1.95161402e-01 -1.47788322e+00 -2.26049572e-01 5.67714393e-01
1.21261496e-02 -1.10417891e+00 -4.85478014e-01 4.41668600e-01
-9.09240484e-01 -7.78535366e-01 -5.20983823e-02 -1.45181262e+00
2.28509858e-01 1.70121938e-02 9.83045578e-01 -1.56467257e-03
3.61422710e-02 -4.59049791e-01 -6.26150131e-01 -3.96042228e-01
-1.75075799e-01 -3.70460451e-02 1.97193578e-01 -5.58766052e-02
7.53944337e-01 -5.64879119e-01 -2.16600552e-01 -2.13052332e-01
-8.24644923e-01 -5.86594164e-01 7.74475515e-01 7.90617228e-01
9.16065514e-01 -2.18070984e-01 4.19037044e-01 -8.14129472e-01
5.86849332e-01 -5.74139535e-01 -6.85564697e-01 2.54510581e-01
-5.07546961e-01 3.12555522e-01 7.76566625e-01 -5.52926183e-01
-1.00265610e+00 2.72822440e-01 -3.11152697e-01 1.86579913e-01
-2.45220408e-01 7.73372591e-01 -5.17730951e-01 2.79280484e-01
4.08463091e-01 3.53788495e-01 -6.76108420e-01 -8.92898977e-01
6.57401145e-01 5.41524708e-01 5.86026669e-01 -6.04547322e-01
6.12986624e-01 -3.79521362e-02 -3.09546649e-01 -7.97735989e-01
-6.54323161e-01 -7.51468718e-01 -1.18608665e+00 3.08463693e-01
9.92332697e-01 -5.86218059e-01 -3.43111068e-01 2.05443949e-01
-1.33504605e+00 1.05561204e-01 -5.66468574e-03 6.60952687e-01
-9.73852500e-02 7.55586267e-01 -8.22870374e-01 -8.28979135e-01
-5.49827576e-01 -8.11115146e-01 8.57690632e-01 3.22014362e-01
-4.04337645e-01 -1.19351923e+00 2.53676742e-01 -4.99927439e-02
6.47025779e-02 -1.27545998e-01 1.38668048e+00 -1.38322270e+00
2.07893085e-02 7.28414953e-02 -1.78832598e-02 2.91139305e-01
9.07799825e-02 -5.16115949e-02 -7.46483982e-01 2.94719394e-02
-2.28003234e-01 1.43112615e-01 8.68200898e-01 1.43485563e-02
6.29125595e-01 -3.42665195e-01 -2.29393616e-01 5.52792013e-01
1.60668468e+00 6.26424849e-01 4.18212801e-01 2.64961779e-01
5.29403329e-01 5.15890181e-01 7.90863037e-01 7.25826696e-02
3.90764683e-01 5.24170518e-01 1.56137049e-01 4.07361269e-01
-1.23365164e-01 -4.48663205e-01 6.98068380e-01 1.62282443e+00
2.59097572e-02 -1.57620370e-01 -9.77220356e-01 8.72440279e-01
-1.54302168e+00 -7.53166139e-01 -2.60218441e-01 2.25799561e+00
9.25753772e-01 -1.07912114e-02 3.55574600e-02 -2.36435216e-02
7.73261249e-01 3.01617503e-01 -1.42507493e-01 -1.00435138e+00
-3.54307026e-01 7.84357429e-01 3.95656288e-01 6.51356578e-01
-1.45073950e+00 1.78838301e+00 6.62068892e+00 8.94301474e-01
-1.12437856e+00 3.11126351e-01 -1.57314651e-02 2.17194960e-01
-1.74123749e-01 9.68137905e-02 -1.09610689e+00 3.41550767e-01
9.40814674e-01 5.68059273e-03 3.05714279e-01 7.25578487e-01
-2.28727102e-01 7.29752108e-02 -8.73257101e-01 6.66289568e-01
3.83805990e-01 -8.46636534e-01 2.94064820e-01 -2.38703676e-02
4.35871482e-01 1.27056316e-01 -1.35947302e-01 5.56817532e-01
5.96723557e-01 -8.51796091e-01 7.52434671e-01 1.76045254e-01
4.18873310e-01 -9.48811054e-01 9.37481940e-01 1.61616102e-01
-1.58674741e+00 -6.56221732e-02 -7.42142975e-01 -1.03700548e-01
1.48666903e-01 2.65412509e-01 -1.14249229e+00 6.36022210e-01
3.86737466e-01 4.07656789e-01 -7.39946187e-01 8.90531778e-01
-8.23354542e-01 1.00192869e+00 5.70195615e-02 -5.67570329e-01
5.15275478e-01 -2.43294075e-01 6.30964458e-01 1.91707766e+00
3.29805166e-01 6.25226721e-02 3.76987368e-01 4.03115541e-01
8.56847316e-02 1.00297844e+00 -3.83192927e-01 -1.19461067e-01
8.69591355e-01 1.48680782e+00 -8.37836444e-01 -4.44603592e-01
-4.70492482e-01 9.53186274e-01 8.82655084e-01 -1.47160403e-02
-6.23435020e-01 -6.22459590e-01 7.99454331e-01 -5.80339968e-01
7.35948384e-01 -4.77505863e-01 -4.29890424e-01 -1.06047356e+00
9.26688388e-02 -6.79680645e-01 7.65987515e-01 -4.37013388e-01
-1.39192164e+00 8.32375824e-01 -2.80370831e-01 -9.52994347e-01
8.02095011e-02 -1.09902310e+00 -6.79124832e-01 9.43489611e-01
-1.51354086e+00 -1.11034262e+00 6.28515422e-01 1.63872197e-01
7.41397440e-01 -2.76502937e-01 1.04548335e+00 -4.33387700e-03
-5.62311292e-01 6.47537827e-01 -7.88455829e-02 5.73218942e-01
4.82924104e-01 -1.62442875e+00 8.08210850e-01 1.35256147e+00
8.67053688e-01 8.47234011e-01 3.31072301e-01 -9.26487744e-01
-1.33778369e+00 -1.13497806e+00 1.80812681e+00 -5.01148999e-01
1.07196975e+00 -2.74243057e-01 -1.06372190e+00 6.21114314e-01
6.31741464e-01 -1.29976958e-01 9.76676524e-01 3.35863143e-01
-6.82636023e-01 1.23801857e-01 -8.76716375e-01 8.40351939e-01
1.20446658e+00 -4.27817345e-01 -1.51847589e+00 -6.24102019e-02
7.11289763e-01 -8.68061557e-02 -7.80292392e-01 1.43316045e-01
2.76297510e-01 -3.48618686e-01 7.23371029e-01 -9.89081383e-01
-1.12275279e-03 -4.24234062e-01 -8.00957143e-01 -1.56667936e+00
-6.90053165e-01 -4.11640048e-01 2.14187682e-01 1.44516766e+00
7.83918858e-01 -3.76733899e-01 1.32793814e-01 -3.76174182e-01
-3.28597069e-01 -6.57699823e-01 -1.13289523e+00 -9.28014934e-01
4.12121207e-01 -5.39955556e-01 8.24338555e-01 1.10912311e+00
3.70070398e-01 6.44772232e-01 -2.90144067e-02 4.89901215e-01
1.25585452e-01 1.15114816e-01 -1.16836742e-01 -1.20004773e+00
-2.32189462e-01 -5.84470630e-01 -8.08915615e-01 -8.45689416e-01
5.68453908e-01 -1.44947469e+00 1.89800948e-01 -1.45415533e+00
-2.30055507e-02 -2.07557544e-01 -8.87310088e-01 7.74827957e-01
-6.85798049e-01 2.44367227e-01 2.50110447e-01 -2.94096559e-01
-3.55857134e-01 1.35942727e-01 6.47436261e-01 -6.24612831e-02
-1.78147614e-01 -5.02245367e-01 -7.76573420e-01 1.03339565e+00
8.33503723e-01 -4.11906987e-01 1.32676065e-01 -6.05953217e-01
3.40858668e-01 -2.00674698e-01 -1.82013988e-01 -9.21337605e-01
6.78353012e-02 -3.52692127e-01 3.16235065e-01 -6.62371814e-01
2.25746810e-01 -4.60745901e-01 -4.30237740e-01 3.89034659e-01
-1.95470050e-01 7.71541119e-01 3.45465004e-01 3.31392586e-01
-1.63727030e-01 -4.83641058e-01 5.53963840e-01 -1.36591181e-01
-1.20472074e+00 5.96438013e-02 -6.70369148e-01 1.70340180e-01
8.42635870e-01 -3.93144030e-04 -9.96560976e-02 5.02232909e-01
-7.57782221e-01 -1.10044613e-01 4.91367141e-03 6.79523706e-01
6.79146767e-01 -1.47047353e+00 -7.74302542e-01 1.90195590e-01
3.75087820e-02 -4.95553821e-01 -1.93032846e-01 4.43665087e-01
-3.29331398e-01 3.70097280e-01 -1.56136706e-01 -4.85260338e-02
-1.45050931e+00 8.74345422e-01 -8.59996751e-02 -3.29196870e-01
-4.00089055e-01 9.82512712e-01 4.09086002e-03 -7.57010877e-01
3.19306627e-02 -6.60616994e-01 -8.40802252e-01 3.62440437e-01
6.61856055e-01 1.44661471e-01 1.93952143e-01 -1.08823597e+00
-5.63965976e-01 5.22797406e-01 -6.99845329e-02 -2.97208369e-01
1.23843086e+00 1.69775233e-01 -3.20351422e-01 5.50286233e-01
6.45791590e-01 6.91586494e-01 -5.31985998e-01 -4.65389222e-01
8.66775334e-01 -8.48678797e-02 -2.75324404e-01 -9.21912432e-01
-7.42390692e-01 1.04164088e+00 3.98041397e-01 -9.21155140e-02
8.66158783e-01 8.48181769e-02 9.16640162e-01 4.30031449e-01
4.12484139e-01 -1.26634598e+00 -4.12341863e-01 1.05119836e+00
6.48524761e-01 -6.34096920e-01 -3.19094151e-01 -5.64835846e-01
-9.33161080e-01 9.64935362e-01 6.77227616e-01 -3.07567328e-01
4.80725855e-01 3.24033856e-01 3.28444511e-01 1.46742254e-01
-7.25018978e-01 -5.33643782e-01 4.54667151e-01 6.72527075e-01
8.05388689e-01 5.59817314e-01 -1.02325881e+00 1.10512781e+00
-4.96304333e-01 -7.39822388e-01 1.11158088e-01 6.86165452e-01
-7.56428242e-01 -1.37060416e+00 -2.38636777e-01 8.09403434e-02
-8.49065483e-01 -6.68208480e-01 -6.28524005e-01 8.56350064e-01
5.24943888e-01 8.87344420e-01 1.76977247e-01 -4.40274388e-01
4.75645095e-01 6.75925434e-01 3.79953414e-01 -9.71005857e-01
-1.03728437e+00 1.42748132e-01 3.37052852e-01 -4.39287037e-01
-9.87902880e-02 -8.80377769e-01 -1.57002628e+00 3.23662013e-01
-2.47650295e-01 3.54993045e-01 6.34592652e-01 1.03492403e+00
1.36992738e-01 2.09523514e-01 5.01667261e-01 -4.69662189e-01
-7.03463018e-01 -1.18608367e+00 -6.59699023e-01 2.72023439e-01
-3.61555070e-01 -4.87285227e-01 -2.24188417e-02 -1.21940918e-01] | [10.311624526977539, 10.169166564941406] |
a781205a-0441-4863-b3d6-7e7684938633 | passnet-learning-pass-probability-surfaces | null | null | https://openreview.net/forum?id=r1xxKJBKvr | https://openreview.net/pdf?id=r1xxKJBKvr | PassNet: Learning pass probability surfaces from single-location labels. An architecture for visually-interpretable soccer analytics | We propose a fully convolutional network architecture that is able to estimate a full surface of pass probabilities from single-location labels derived from high frequency spatio-temporal data of professional soccer matches. The network is able to perform remarkably well from low-level inputs by learning a feature hierarchy that produces predictions at different sampling levels that are merged together to preserve both coarse and fine detail. Our approach presents an extreme case of weakly supervised learning where there is just a single pixel correspondence between ground-truth outcomes and the predicted probability map. By providing not just an accurate evaluation of observed events but also a visual interpretation of the results of other potential actions, our approach opens the door for spatio-temporal decision-making analysis, an as-yet little-explored area in sports. Our proposed deep learning architecture can be easily adapted to solve many other related problems in sports analytics; we demonstrate this by extending the network to learn to estimate pass-selection likelihood. | ['Luke Bornn', 'Javier Fernández'] | 2019-09-25 | null | null | null | null | ['sports-analytics'] | ['computer-vision'] | [ 1.11406036e-01 1.30205050e-01 -1.80334777e-01 -5.85932851e-01
-9.68041420e-01 -4.23051804e-01 6.40763521e-01 7.39888430e-01
-6.03881121e-01 5.72482646e-01 5.32979965e-01 4.85332608e-02
-3.09102267e-01 -1.17870390e+00 -1.10620058e+00 -4.45422113e-01
-4.74131972e-01 6.45984173e-01 6.71596944e-01 -2.83965617e-01
1.96494162e-01 6.21476352e-01 -2.01510906e+00 9.50259089e-01
2.29253545e-01 9.37660933e-01 -5.21701761e-02 6.53423131e-01
2.31938183e-01 9.24353063e-01 -5.41405380e-01 -4.61729288e-01
1.66100204e-01 -2.40415916e-01 -7.47876108e-01 -1.36197358e-01
6.19087815e-01 -8.19156691e-02 -4.05038297e-01 5.63079238e-01
2.58958519e-01 1.62661955e-01 5.83750725e-01 -9.41859722e-01
-1.95373669e-02 7.92260051e-01 -4.38827425e-01 4.91193175e-01
3.44222873e-01 4.89139795e-01 9.85481322e-01 -6.37375951e-01
6.45502150e-01 1.24595511e+00 1.00904632e+00 -2.35421449e-01
-1.26496148e+00 -3.14628929e-01 1.97427228e-01 5.14645100e-01
-1.29249156e+00 8.75454023e-02 3.53663981e-01 -6.29179537e-01
7.92479157e-01 1.91722974e-01 1.02945399e+00 8.45043242e-01
3.09905857e-01 8.77942264e-01 1.12840819e+00 -1.73308074e-01
2.99558997e-01 -3.92533720e-01 5.65433828e-03 6.57163918e-01
-5.03673032e-02 3.98163319e-01 -1.25372684e+00 8.52336884e-02
8.48238587e-01 1.12212442e-01 2.56839842e-01 -3.39149445e-01
-1.47172511e+00 6.28134966e-01 7.18657613e-01 -1.11815743e-01
-4.41350400e-01 3.47238004e-01 5.25207877e-01 6.17930926e-02
3.63881707e-01 3.13704193e-01 -3.33453923e-01 -4.11938131e-01
-1.27213407e+00 6.83495879e-01 6.04066312e-01 2.96772927e-01
8.85718226e-01 -1.82490721e-01 -5.23240507e-01 2.82298923e-01
-1.25406787e-01 -1.40440334e-02 2.40949597e-02 -1.02884209e+00
3.40722024e-01 6.94035769e-01 2.83030301e-01 -9.23518419e-01
-6.70663238e-01 -8.11820328e-01 -4.00399387e-01 8.10006499e-01
9.10850525e-01 6.74852058e-02 -9.68167245e-01 1.42968225e+00
3.64576668e-01 4.78808343e-01 -4.40671951e-01 1.19770551e+00
6.04732990e-01 4.43075627e-01 2.33708218e-01 5.63256681e-01
1.31585956e+00 -5.55296898e-01 -2.51906097e-01 -3.54038566e-01
4.17117029e-01 -4.15478975e-01 1.18189323e+00 5.43650866e-01
-1.12515700e+00 -7.70273268e-01 -1.06667376e+00 -2.05489591e-01
-4.97947037e-01 2.38902152e-01 7.39794970e-01 1.91958651e-01
-9.26023602e-01 9.72712576e-01 -1.13669574e+00 -2.80163884e-01
6.17285967e-01 1.87185511e-01 -4.04015064e-01 1.12697378e-01
-1.20501745e+00 8.28930497e-01 4.76618469e-01 2.18719497e-01
-1.08908308e+00 -8.60183239e-01 -9.48220789e-01 3.23079258e-01
4.05781239e-01 -3.69862407e-01 1.04952538e+00 -6.32466018e-01
-1.15111661e+00 9.47849989e-01 7.72106797e-02 -6.65121436e-01
7.91414142e-01 -4.08177972e-01 -1.00242801e-01 -4.35000993e-02
3.93436432e-01 5.89732945e-01 2.29611903e-01 -9.72806394e-01
-1.12869120e+00 -3.90638947e-01 8.95019472e-02 2.49599695e-01
8.80993158e-02 -8.57678503e-02 -3.87978733e-01 -6.70706689e-01
2.62790233e-01 -5.49232423e-01 -4.31194514e-01 2.55256951e-01
-3.87062103e-01 -1.55106246e-01 5.51589765e-02 -5.00473499e-01
1.03615129e+00 -2.02176452e+00 -9.24369879e-03 3.20354491e-01
9.10935923e-02 -1.41111150e-01 1.07242331e-01 6.76271737e-01
6.62404485e-03 -2.08792835e-01 -1.68474048e-01 -1.26808912e-01
4.48719300e-02 6.27911612e-02 -3.48572761e-01 5.11989295e-01
4.24610615e-01 9.96194243e-01 -1.15001845e+00 -4.85167563e-01
5.57701051e-01 3.44511896e-01 -3.86118144e-01 -1.09268434e-01
-1.97874248e-01 2.79306144e-01 -9.30144563e-02 5.94455659e-01
2.78789192e-01 -1.19465932e-01 3.95750478e-02 -7.49676153e-02
-4.04750675e-01 4.09160197e-01 -1.59356654e+00 2.00583911e+00
-1.48038715e-01 8.66220057e-01 -3.52145672e-01 -1.11186671e+00
7.84024477e-01 -1.79460600e-01 5.57634950e-01 -5.61926723e-01
-1.44455329e-01 -1.02037534e-01 -2.53174007e-01 -4.73984212e-01
6.15128398e-01 -2.93765306e-01 -4.73838329e-01 3.93080622e-01
2.25382864e-01 3.29342991e-01 2.26710379e-01 1.61198407e-01
1.23862159e+00 7.04625249e-01 6.70175776e-02 -1.57693565e-01
-6.58868402e-02 6.13483846e-01 6.25933230e-01 1.27439535e+00
6.24838518e-03 8.61841261e-01 7.18978286e-01 -1.09148216e+00
-9.68731642e-01 -1.59550476e+00 -1.52585045e-01 1.21576047e+00
2.25938633e-01 -5.62468767e-01 -4.60671037e-01 -5.47213018e-01
2.01355353e-01 3.39565665e-01 -8.70721221e-01 -6.65768906e-02
-7.80644655e-01 -6.13854945e-01 5.96032321e-01 9.03558791e-01
2.37618461e-01 -1.12906027e+00 -8.92619491e-01 4.35924202e-01
-1.61381826e-01 -8.47976387e-01 1.22957617e-01 4.34488982e-01
-6.37760878e-01 -1.07368791e+00 -4.65380639e-01 -5.11087120e-01
9.68717709e-02 -3.58245760e-01 1.57788050e+00 -8.27603266e-02
-7.28280067e-01 -1.26047336e-04 -1.47349536e-01 -6.09721780e-01
-1.66508391e-01 1.35219201e-01 -1.18172891e-01 1.51732400e-01
6.98129237e-01 -5.44633031e-01 -7.44251013e-01 2.71074951e-01
-7.10126042e-01 8.41037631e-02 4.59914267e-01 6.21572971e-01
8.75700772e-01 -9.16399881e-02 2.76315212e-01 -5.88838041e-01
1.83285445e-01 -4.85165805e-01 -4.62311715e-01 1.48618191e-01
5.27776703e-02 1.30442798e-01 2.82913655e-01 -1.13111190e-01
-9.70917463e-01 4.51980501e-01 -2.02365696e-01 -9.73828062e-02
-7.23611891e-01 6.02876961e-01 2.34629244e-01 4.55425799e-01
1.04759347e+00 -3.31179239e-02 -2.40379572e-01 -5.96184671e-01
5.21786153e-01 2.34041736e-01 1.01452816e+00 -5.02170146e-01
5.00229776e-01 9.65055287e-01 2.10003138e-01 -4.35242921e-01
-1.16342080e+00 -5.05836487e-01 -1.00208330e+00 -6.02085531e-01
9.62512732e-01 -1.06019056e+00 -1.09537184e+00 2.68234819e-01
-7.35334098e-01 -5.69045842e-01 -7.59996712e-01 6.62055790e-01
-7.20299840e-01 1.90532617e-02 -7.23940074e-01 -6.41947985e-01
4.48158026e-01 -7.17345893e-01 1.43242335e+00 1.53129250e-01
-4.26489174e-01 -7.18295217e-01 3.12502831e-01 1.47506341e-01
-2.03816861e-01 6.29916847e-01 5.31982660e-01 -3.62187773e-01
-7.43206680e-01 -2.02409297e-01 1.68451443e-02 -1.25944465e-01
-3.38677287e-01 -5.07800765e-02 -9.40332711e-01 1.33194089e-01
-6.68574452e-01 -4.23328429e-01 1.44419539e+00 6.91671014e-01
1.26051116e+00 2.47352377e-01 -3.86396199e-01 6.32632792e-01
1.26561177e+00 -6.18488312e-01 6.34380281e-01 6.93579733e-01
5.45171022e-01 8.32667828e-01 9.59862053e-01 4.90830451e-01
4.35645700e-01 7.87397802e-01 5.75914145e-01 -5.38027227e-01
-3.60816151e-01 -5.67375124e-01 -2.23333538e-02 -2.45551661e-01
-2.88607001e-01 -3.34091708e-02 -1.12039506e+00 8.39763701e-01
-2.22889733e+00 -1.35317373e+00 -4.32783425e-01 2.36081123e+00
6.81268752e-01 5.31308115e-01 5.34100592e-01 4.09728408e-01
5.73877037e-01 1.27351612e-01 -4.66918826e-01 -2.98218250e-01
-2.52342429e-02 3.71037573e-01 6.14355803e-01 2.68312901e-01
-1.41861093e+00 7.69151866e-01 7.15312624e+00 8.52353334e-01
-9.17832792e-01 -1.62259064e-04 6.66514218e-01 -5.68999946e-01
-1.06700130e-01 -9.51718132e-04 -6.17301524e-01 3.55435669e-01
7.90511787e-01 4.16362315e-01 -1.08892389e-01 6.22134209e-01
3.59293282e-01 -5.23688197e-01 -1.17569339e+00 7.40463316e-01
-2.39498705e-01 -1.93788767e+00 -2.62792498e-01 9.37663615e-02
4.51093376e-01 2.21259624e-01 -1.15726449e-01 2.00754687e-01
7.14453280e-01 -1.23709166e+00 1.11876154e+00 8.68931413e-01
6.74669087e-01 -8.13230336e-01 5.70422053e-01 4.62349236e-01
-1.20065069e+00 -2.74636686e-01 -2.57217199e-01 -6.65483057e-01
4.91112739e-01 6.71314180e-01 -7.36734688e-01 3.28753948e-01
1.19000280e+00 8.73747945e-01 -5.33258140e-01 1.30553842e+00
-1.69602126e-01 7.00092375e-01 -5.32394409e-01 2.58533835e-01
1.99602783e-01 2.53123254e-01 3.67407292e-01 1.40963745e+00
1.97958887e-01 -1.96179122e-01 3.52399975e-01 8.32519591e-01
3.15452665e-01 -1.43067360e-01 -4.97421384e-01 5.17206013e-01
5.99757247e-02 1.10952008e+00 -1.07323360e+00 -3.73416811e-01
-1.18983641e-01 5.89770973e-01 7.68145859e-01 2.11565703e-01
-8.07469308e-01 -8.03071558e-02 8.48074257e-01 6.27072573e-01
3.63626063e-01 -3.37148875e-01 -6.92010939e-01 -1.08114100e+00
2.84245480e-02 -4.19569880e-01 7.45290339e-01 -8.75714183e-01
-1.14846206e+00 2.71114260e-01 5.80453277e-02 -1.44288898e+00
-2.63146609e-01 -6.01205111e-01 -7.42581606e-01 8.40565145e-01
-1.34231150e+00 -1.03390586e+00 -2.04252973e-01 3.68673533e-01
4.91088450e-01 1.37270065e-02 6.13096833e-01 1.24382466e-01
-1.61336511e-01 1.66760340e-01 -1.14539981e-01 2.14004308e-01
5.51951945e-01 -1.49468672e+00 4.99272645e-01 9.08789515e-01
6.49037540e-01 -1.77242886e-02 8.41870189e-01 -6.01859629e-01
-5.96668005e-01 -8.82825017e-01 9.75760341e-01 -7.09592462e-01
5.13181746e-01 -4.63968635e-01 -7.71719038e-01 6.27455771e-01
-2.15317667e-01 9.51391011e-02 6.54364705e-01 4.85512525e-01
-1.30649135e-01 -3.11914414e-01 -6.51133060e-01 3.98783326e-01
1.12516797e+00 -3.59963596e-01 -6.02565527e-01 2.95244187e-01
6.55923486e-02 -7.54277468e-01 -8.34028065e-01 2.90819764e-01
6.29174411e-01 -1.41985595e+00 1.19868958e+00 -8.24388683e-01
8.26705039e-01 -4.11031276e-01 2.87365824e-01 -1.03746510e+00
-3.79739970e-01 -1.83769330e-01 1.52656928e-01 7.53240883e-01
3.76549363e-01 8.76735151e-02 1.15215874e+00 3.82284820e-01
-2.64291883e-01 -7.60104239e-01 -9.92477357e-01 -3.89842570e-01
-6.09370545e-02 -9.18717265e-01 6.34588599e-01 6.06475830e-01
1.47959702e-02 -1.29757002e-01 -3.73096049e-01 3.25875014e-01
7.55066514e-01 3.07748020e-01 6.85019970e-01 -1.30486917e+00
-3.53992611e-01 -2.39703506e-01 -8.02862585e-01 -9.57683325e-01
-1.68480098e-01 -7.77945817e-01 2.28667036e-01 -1.71274567e+00
3.31524424e-02 -3.05918604e-01 -4.99750674e-01 5.61925292e-01
-6.07067794e-02 8.03590953e-01 1.24856107e-01 8.40778574e-02
-6.58205330e-01 1.56407449e-02 9.47218716e-01 9.03110728e-02
-8.29305425e-02 1.13159761e-01 -3.24033767e-01 9.02044654e-01
6.37209475e-01 -6.91668510e-01 6.36750013e-02 -3.99936646e-01
6.10293627e-01 1.04469702e-01 1.05656350e+00 -1.43894923e+00
8.36761221e-02 -3.45745653e-01 7.83758342e-01 -8.37035000e-01
3.08914393e-01 -4.00940239e-01 1.12732321e-01 3.35427195e-01
-8.72536898e-01 -1.03383541e-01 1.66132918e-03 8.36459696e-01
-3.80090147e-01 1.39683679e-01 3.00885022e-01 -4.30708140e-01
-1.08685112e+00 1.52874649e-01 -6.04216039e-01 -8.17556009e-02
9.14909184e-01 -5.22585511e-01 -1.52715221e-01 -3.83547992e-01
-1.33510268e+00 2.43951470e-01 3.76779318e-01 3.67167979e-01
3.46915156e-01 -1.24791956e+00 -9.36493635e-01 1.37686849e-01
2.13674560e-01 -9.44954250e-03 4.18752223e-01 5.92213392e-01
-7.62600124e-01 -8.66906531e-03 -4.84710157e-01 -1.09288466e+00
-8.26268971e-01 2.41856515e-01 3.47712576e-01 -4.83784378e-01
-9.58553553e-01 7.76520491e-01 -1.14524513e-02 -2.63067573e-01
2.84709722e-01 -4.62512076e-01 -3.00560266e-01 -2.49879658e-02
5.87107122e-01 3.29787761e-01 1.36960179e-01 -6.24176323e-01
-3.50522548e-01 3.02742660e-01 4.98753905e-01 -2.87104696e-01
1.61690116e+00 8.05760622e-02 3.81723195e-01 6.43880069e-01
6.29151523e-01 -2.23049015e-01 -1.93496704e+00 -2.72550613e-01
1.38353348e-01 -6.32668734e-01 8.57053176e-02 -9.21651006e-01
-8.66903663e-01 1.17318857e+00 6.58570826e-01 1.48129582e-01
8.07926118e-01 3.38724732e-01 4.85895455e-01 1.80198461e-01
2.81155854e-01 -1.30904198e+00 1.18017077e-01 2.51401871e-01
5.93068361e-01 -1.30556357e+00 1.13588043e-01 -1.47921860e-01
-7.20705867e-01 1.19282007e+00 5.81405044e-01 -4.28105205e-01
5.26657164e-01 5.77610016e-01 1.32873610e-01 -5.29306889e-01
-6.10579729e-01 -6.27554357e-01 5.07180572e-01 6.34474993e-01
3.79057288e-01 1.20793112e-01 3.09227016e-02 4.81157213e-01
-3.61128151e-01 2.00261220e-01 4.60648239e-01 8.33079875e-01
-5.45230925e-01 -8.25308621e-01 -3.86525065e-01 5.51160634e-01
-4.42454278e-01 1.69637725e-01 -3.84521633e-02 8.65495741e-01
4.97885287e-01 5.07100999e-01 6.53814077e-01 -3.21734339e-01
5.40528655e-01 1.77990664e-02 2.22487465e-01 -6.85657680e-01
-8.06631684e-01 -7.71902278e-02 1.80661663e-01 -9.35458243e-01
-4.77473468e-01 -8.64519715e-01 -1.16660702e+00 -1.51451185e-01
3.24514121e-01 -1.55852482e-01 4.58188266e-01 9.95056808e-01
2.85611331e-01 7.40050137e-01 9.49110463e-02 -1.18288565e+00
-1.28237516e-01 -6.49525702e-01 -7.50565886e-01 7.23341584e-01
2.99393564e-01 -7.88423479e-01 1.02740593e-01 1.65191069e-01] | [7.950044631958008, 0.19564521312713623] |
b8138870-3637-4c76-8211-985815120658 | evaluating-the-adversarial-robustness-of-a | null | null | https://openreview.net/forum?id=HyhSFQ1hOgV | https://openreview.net/pdf?id=HyhSFQ1hOgV | Evaluating the Adversarial Robustness of a Foveated Texture Transform Module in a CNN | Biologically inspired mechanisms such as foveation and multiple fixation points have previously been shown to help alleviate adversarial examples (Reddy et al., 2020). By mimicking the effects of visual crowding present in human vision, foveated, texture-based computations may provide another route for increasing adversarial robustness. Previous statistical models of texture rendering (Portilla & Simoncelli, 2000; Gatys et al., 2015) paved the way for the development of a Foveated Texture Transform (FTT) module which utilizes localized texture synthesis in foveated receptive fields (Deza et al., 2017). The FTT module was added to a VGG-11 CNN architecture and ten random initializations were trained on 20-class subsets of the Places and EcoSet datasets for scene and object classification respectively. The trained networks were attacked using Projected Gradient Descent (PGD) and the adversarial accuracies were calculated at multiple epochs to evaluate changes in robustness as the networks trained. The results indicate that the FTT module significantly improved adversarial robustness for scene classification, especially when the validation loss was at a minimum. However, the FTT module did not provide a statistically significant increase in adversarial robustness for object classification. Furthermore, we do not find a trade-off between accuracy and robustness (Tsipras et al., 2018) for the FTT module suggesting a benefit of using foveated, texture-based distortions in the latent space during learning compared to non-perturbed latent space representations. Finally, we investigate the nature of latent space distortions with additional controls that probe other directions in the latent space that are not texture-based. | ['Arturo Deza', 'Andrzej Banburski', 'Jonathan M Gant'] | 2021-10-12 | null | null | null | neurips-workshop-svrhm-2021-12 | ['texture-synthesis', 'foveation'] | ['computer-vision', 'computer-vision'] | [ 2.54168600e-01 2.16525659e-01 5.58036268e-01 2.96105556e-02
-2.61933744e-01 -7.22744286e-01 9.53346908e-01 -2.01399878e-01
-5.18005431e-01 6.51768029e-01 8.36524516e-02 -3.16690505e-01
-8.57566148e-02 -9.10391390e-01 -1.11902058e+00 -8.79879117e-01
-1.39186069e-01 -4.02422905e-01 1.82239652e-01 -2.12488726e-01
3.21999788e-01 8.56459260e-01 -1.70676005e+00 4.67566937e-01
6.28954053e-01 8.86509418e-01 1.77533645e-02 7.70916700e-01
3.75705838e-01 6.32298052e-01 -8.84867847e-01 -4.93993640e-01
6.68604732e-01 -1.07889064e-01 -4.55447704e-01 -1.87418908e-01
7.30857968e-01 -3.16943496e-01 -4.58405107e-01 8.45034301e-01
7.16191471e-01 1.57802865e-01 6.44064486e-01 -1.16583097e+00
-7.69026399e-01 1.58027768e-01 -1.96978539e-01 2.49107644e-01
1.33519188e-01 7.26154625e-01 2.93887645e-01 -7.11660922e-01
5.08602560e-01 1.45589960e+00 6.41101360e-01 6.25138760e-01
-1.45423305e+00 -5.57681739e-01 8.62626880e-02 -9.07348692e-02
-1.13761020e+00 -4.53641951e-01 6.06821179e-01 -5.47036231e-01
1.20107806e+00 3.73284042e-01 6.79633796e-01 1.42874563e+00
5.45402825e-01 2.06343323e-01 1.60611141e+00 -3.88924092e-01
2.43683368e-01 1.06337972e-01 -5.10559142e-01 4.30032015e-01
3.56515229e-01 5.12430489e-01 -4.35839504e-01 -1.79399893e-01
9.98993814e-01 -5.45748413e-01 -1.53726503e-01 -1.55485466e-01
-1.05124986e+00 7.05453336e-01 6.97072029e-01 1.98924601e-01
-3.58719289e-01 3.94137472e-01 5.01225531e-01 3.06867927e-01
6.45067632e-01 9.41362858e-01 -2.62924522e-01 1.23723269e-01
-6.62305176e-01 3.25618327e-01 3.50332320e-01 3.41328323e-01
6.05559826e-01 4.75908518e-01 -3.69163215e-01 5.60631096e-01
9.38377976e-02 4.30115342e-01 3.15388411e-01 -1.06229019e+00
3.45287889e-01 2.77123421e-01 3.76221836e-02 -1.20668328e+00
-2.63092905e-01 -3.75875026e-01 -4.95341510e-01 8.56869996e-01
8.46322119e-01 -1.58854678e-01 -1.00539923e+00 1.90065312e+00
2.69761682e-01 1.70418814e-01 9.44937468e-02 8.80462527e-01
3.01151216e-01 3.02081048e-01 2.78578073e-01 2.32477859e-01
1.02236032e+00 -5.63970208e-01 -3.19852859e-01 -1.73976138e-01
4.88288939e-01 -8.54676425e-01 1.44656551e+00 2.80093670e-01
-9.28425968e-01 -5.44256985e-01 -1.05652666e+00 1.65393949e-01
-6.70687199e-01 -1.66721761e-01 6.77566409e-01 1.06296384e+00
-1.09138894e+00 8.01737487e-01 -8.08562398e-01 -4.91329700e-01
7.11622119e-01 2.43058860e-01 -4.27598417e-01 1.96858972e-01
-1.26849532e+00 1.25736248e+00 1.98974639e-01 1.70803100e-01
-1.18227613e+00 -5.51477015e-01 -6.98836029e-01 -2.72171944e-01
-3.14909220e-02 -6.90141261e-01 3.30624193e-01 -1.36409366e+00
-1.70187962e+00 8.84689689e-01 2.59340852e-01 -6.80309176e-01
7.29600906e-01 -2.11512089e-01 -1.46692440e-01 1.22090675e-01
-1.14451624e-01 1.05804431e+00 1.14371622e+00 -1.11407173e+00
1.36100933e-01 -2.52054065e-01 4.00142521e-01 2.96376318e-01
-1.98248640e-01 9.81228277e-02 4.06928360e-01 -9.69769239e-01
-2.43788570e-01 -1.03896677e+00 1.09957889e-01 3.07388842e-01
-3.29892576e-01 2.85129637e-01 7.24724770e-01 -7.14710295e-01
4.92554367e-01 -2.16835904e+00 -7.59781748e-02 1.70786291e-01
1.15350507e-01 4.40137386e-01 -1.87755659e-01 1.20837063e-01
-4.50301379e-01 5.14411449e-01 -1.81382522e-01 -1.57181427e-01
-1.22957826e-01 1.19618632e-01 -3.98148149e-01 8.57304513e-01
5.04864872e-01 9.49617147e-01 -5.92162371e-01 5.81520945e-02
3.66447002e-01 5.81023633e-01 -7.23518968e-01 -1.76674187e-01
-9.49827656e-02 6.79031014e-01 1.06733583e-01 4.47170585e-01
7.45462418e-01 3.96817774e-01 -3.73748451e-01 7.03678578e-02
-2.70663619e-01 1.76800460e-01 -7.77870834e-01 1.33152664e+00
-3.54204476e-01 1.02331960e+00 -2.72328079e-01 -6.32018209e-01
7.88238525e-01 2.66079634e-01 5.15165403e-02 -8.72946322e-01
2.69401819e-01 1.25326112e-01 2.82611370e-01 -2.91449994e-01
2.60329574e-01 -1.45333111e-01 1.07747376e-01 1.75778061e-01
-1.04233064e-01 -2.04066604e-01 -2.62097448e-01 -1.91070989e-01
1.06773674e+00 5.09804785e-01 -2.66770422e-01 -4.99559045e-01
2.91979730e-01 -4.32987288e-02 2.20542029e-01 7.70447016e-01
-3.78113508e-01 7.33151913e-01 6.39336526e-01 -4.25221413e-01
-1.32618046e+00 -1.23064709e+00 -3.02871794e-01 8.46656680e-01
-2.57273585e-01 1.45304993e-01 -1.12901342e+00 -4.53071237e-01
3.87512073e-02 8.30620110e-01 -1.11839318e+00 -6.50983572e-01
-2.82039285e-01 -5.32624125e-01 1.16829348e+00 1.48660436e-01
7.51825988e-01 -1.06575799e+00 -9.18923438e-01 -2.42497802e-01
3.15770768e-02 -7.25492954e-01 -2.50304472e-02 1.46132320e-01
-7.10046768e-01 -8.76761913e-01 -7.44354784e-01 -1.24145888e-01
7.51092315e-01 5.93998581e-02 6.20210767e-01 -1.20248377e-01
-4.47871208e-01 4.22939807e-01 -8.54196697e-02 -7.35914946e-01
-3.49284172e-01 -3.58496279e-01 2.78602660e-01 4.99861641e-03
-8.50285739e-02 -4.88085359e-01 -5.58567166e-01 4.31580335e-01
-9.91731584e-01 -1.04184002e-01 1.96252704e-01 8.14955354e-01
3.34806204e-01 -1.18664280e-01 3.41228217e-01 -4.68544930e-01
4.81296957e-01 -2.73353487e-01 -5.85390389e-01 -4.61824015e-02
-1.61365509e-01 9.75391641e-02 6.15409017e-01 -8.50376725e-01
-9.12552595e-01 -3.61354083e-01 1.02643430e-01 -7.70569563e-01
-3.52270246e-01 1.36492282e-01 -2.42003247e-01 -5.75436175e-01
1.24739861e+00 -4.65586968e-02 4.79076803e-02 -1.90763623e-02
1.47268817e-01 3.99954058e-02 1.25919506e-01 -7.31570780e-01
1.07143223e+00 5.92583299e-01 1.13130152e-01 -6.02604389e-01
-4.38672781e-01 2.92668164e-01 -5.67009747e-01 -5.14894187e-01
1.01874197e+00 -6.76546633e-01 -6.29337907e-01 8.45555425e-01
-1.00396979e+00 -8.42832386e-01 -3.63940179e-01 4.89775181e-01
-7.52202868e-01 3.23032849e-02 -1.43479615e-01 -8.09889853e-01
1.07239507e-01 -1.13804996e+00 6.78885698e-01 1.32693917e-01
-1.80617213e-01 -9.04145420e-01 2.69320654e-03 1.79236785e-01
5.60853720e-01 8.93895268e-01 9.37065125e-01 -3.75785649e-01
-5.01197934e-01 2.09096018e-02 -5.13107106e-02 4.28273678e-01
1.13624677e-01 1.13499783e-01 -1.53005445e+00 -3.95446032e-01
1.27420291e-01 -5.27604461e-01 8.01538467e-01 2.70103097e-01
1.08444560e+00 -5.02201378e-01 3.11849535e-01 9.16595817e-01
1.22999501e+00 8.36714655e-02 1.15168941e+00 8.75601888e-01
5.51680028e-01 7.90467203e-01 3.77339602e-01 7.54440874e-02
-2.72065878e-01 5.78444839e-01 5.73934853e-01 -2.83441752e-01
-3.55184942e-01 -1.51375338e-01 5.60363829e-01 -1.90283775e-01
-2.84840912e-01 -9.02899727e-02 -8.42483222e-01 3.08797389e-01
-1.28588402e+00 -1.13371801e+00 7.27109909e-02 2.47078395e+00
5.61577737e-01 4.38806921e-01 6.26925156e-02 1.91918179e-01
7.84362495e-01 -5.37213758e-02 -6.26306474e-01 -7.64992535e-01
-5.44756174e-01 1.90213367e-01 7.50151873e-01 3.55653375e-01
-1.08551812e+00 1.11862135e+00 6.24025869e+00 5.55450737e-01
-1.52605391e+00 -7.31480569e-02 9.62296247e-01 -2.32009113e-01
-2.27738649e-01 -3.56218778e-02 -2.23762542e-01 4.40607101e-01
1.00597787e+00 1.66295707e-01 8.56787145e-01 5.20863950e-01
2.88692564e-01 -3.15587372e-01 -5.64948142e-01 4.20727372e-01
-1.15014948e-01 -1.03739607e+00 -8.84238258e-03 2.49537587e-01
6.83971584e-01 -1.29262414e-02 6.74176514e-01 1.03255562e-01
2.13673990e-02 -1.15481341e+00 1.04318142e+00 8.03327262e-01
9.85974610e-01 -8.92483830e-01 4.36581105e-01 6.77425414e-02
-5.81572354e-01 -7.87250698e-02 -5.87283671e-01 -1.78581357e-01
-4.37225521e-01 1.22284926e-01 -7.16418743e-01 1.23018853e-01
9.59151864e-01 2.34862026e-02 -9.87904310e-01 7.61362672e-01
-1.79289892e-01 6.09493732e-01 -3.32926929e-01 2.82740682e-01
1.63760006e-01 1.08557738e-01 8.57186556e-01 8.05190206e-01
5.53010330e-02 -3.26599956e-01 -6.07289612e-01 1.11154389e+00
2.96534121e-01 -2.47338116e-01 -9.89624202e-01 -7.94800296e-02
3.13301355e-01 9.95535314e-01 -9.66414869e-01 1.27389997e-01
-1.88872695e-01 9.33491707e-01 3.35202754e-01 6.81136429e-01
-1.09101450e+00 -4.26233828e-01 8.36161852e-01 2.96564847e-01
1.36555612e-01 -2.62782812e-01 -6.79577768e-01 -9.09015059e-01
-1.11628406e-01 -9.10403490e-01 -1.98066801e-01 -1.04171884e+00
-1.03068876e+00 4.57900345e-01 1.04241498e-01 -1.00371492e+00
7.81535208e-02 -7.48394728e-01 -6.56796932e-01 1.32092547e+00
-1.16419816e+00 -1.18032801e+00 -3.13853413e-01 8.96492660e-01
7.22222477e-02 -1.85507506e-01 8.57220769e-01 -8.68180171e-02
-5.40378332e-01 9.19608772e-01 -2.35698074e-02 1.24329276e-01
7.84610987e-01 -1.02423835e+00 5.99392474e-01 1.16662288e+00
-2.72169411e-01 1.03702259e+00 7.15952933e-01 -8.35039258e-01
-1.02187777e+00 -1.27103424e+00 2.30546057e-01 -7.16284096e-01
4.30064917e-01 -5.90685010e-01 -7.89061844e-01 5.26117921e-01
1.04563147e-01 -1.05612297e-02 3.82469326e-01 -3.94231856e-01
-5.39780796e-01 1.70359984e-01 -1.50089359e+00 9.65783179e-01
9.33668315e-01 -7.20300615e-01 -2.90271759e-01 9.00854915e-02
7.53495932e-01 -3.64463896e-01 -7.00566292e-01 3.91552806e-01
5.18060386e-01 -1.03733683e+00 1.08393490e+00 -5.00796199e-01
4.01800036e-01 -2.46429414e-01 -1.88545853e-01 -1.18646407e+00
-2.74123877e-01 -5.06621897e-01 3.85355800e-01 1.12817562e+00
3.34105998e-01 -1.06028616e+00 5.73464811e-01 7.02412248e-01
-2.84307659e-01 -5.18588066e-01 -1.10926235e+00 -8.11108708e-01
4.44247007e-01 -4.24571544e-01 2.97260612e-01 8.59671712e-01
-5.01453459e-01 -4.99005049e-01 -1.38482913e-01 2.07129925e-01
4.12493378e-01 -7.84188330e-01 7.58365870e-01 -7.28602886e-01
-2.46001616e-01 -5.16203225e-01 -6.38424814e-01 5.19727841e-02
1.28460824e-01 -8.49476337e-01 -1.66399732e-01 -1.22491229e+00
-3.60825896e-01 -3.88575405e-01 -3.36736500e-01 6.46262407e-01
-1.75332472e-01 4.18762803e-01 3.92746627e-01 1.77690119e-01
1.35188684e-01 4.04348075e-01 1.20779216e+00 6.55885711e-02
8.73208046e-04 -2.53683329e-01 -6.19465351e-01 5.71814179e-01
1.00297821e+00 -4.03012186e-01 -4.78303969e-01 -4.44017708e-01
6.03719056e-02 -5.44981241e-01 9.48642969e-01 -1.24765956e+00
-1.14377208e-01 -2.53242940e-01 8.41969967e-01 1.57186747e-01
5.05111158e-01 -6.00676656e-01 1.29118562e-01 6.16732895e-01
-1.95478112e-01 -7.04986677e-02 8.71231616e-01 3.92827243e-01
3.03095311e-01 6.20002970e-02 1.07863188e+00 -7.17630908e-02
-1.84504837e-01 -9.27078649e-02 -8.00432622e-01 -1.31119877e-01
1.08154035e+00 -6.47315085e-01 -5.85172534e-01 -1.13611557e-01
-5.71878254e-01 -4.44033831e-01 8.27750862e-01 4.07439262e-01
5.34830093e-01 -1.25668144e+00 -5.70697367e-01 2.87797898e-01
-1.64989546e-01 -2.58595139e-01 2.70806581e-01 7.27090240e-01
-7.47837186e-01 3.07460129e-01 -8.02876651e-01 -4.06970739e-01
-9.62868631e-01 6.83140874e-01 7.16787934e-01 1.55981436e-01
-2.14144826e-01 1.04083061e+00 3.86984348e-01 -3.57504785e-01
5.92394695e-02 -7.73293227e-02 -5.04917577e-02 -8.74595717e-02
2.22329542e-01 3.67943376e-01 3.97649035e-02 -6.22529805e-01
-2.74918020e-01 3.67174625e-01 2.40859479e-01 -3.23704511e-01
1.10318696e+00 2.06721798e-02 2.25043613e-02 5.18288136e-01
9.29094732e-01 4.70759608e-02 -1.68466330e+00 2.46533826e-01
-3.16793352e-01 -6.34211361e-01 -7.04565691e-03 -9.37112987e-01
-8.87779772e-01 9.69160140e-01 9.02902782e-01 -5.71476705e-02
1.20756626e+00 -4.50730562e-01 2.57550254e-02 2.49652699e-01
4.17208701e-01 -9.79649007e-01 2.47299656e-01 5.70515990e-01
1.21173155e+00 -7.74159968e-01 -1.89445019e-01 -4.66967048e-03
-6.47516251e-01 9.32497621e-01 8.34477603e-01 -3.28919500e-01
3.30545664e-01 1.76074326e-01 5.97010478e-02 -4.66671772e-02
-6.12701118e-01 -1.02023363e-01 2.77314395e-01 9.13741946e-01
1.68281749e-01 -1.01072729e-01 1.70454904e-02 9.98802856e-02
-5.05754650e-01 -4.73555058e-01 3.78251284e-01 8.99167359e-01
-1.61265746e-01 -7.48830438e-01 -8.58501554e-01 3.33170831e-01
-4.49564606e-01 -1.92677900e-01 -5.75095296e-01 8.90897989e-01
4.63678896e-01 8.28349113e-01 1.71119228e-01 -6.06848121e-01
2.85577208e-01 2.88058788e-01 6.49220407e-01 -4.27092761e-01
-8.96029651e-01 -3.02193630e-02 -3.92738990e-02 -6.02500081e-01
-2.45546438e-02 -7.52032101e-01 -7.64291942e-01 -5.96004546e-01
-3.71725589e-01 -3.41104299e-01 7.85872281e-01 8.55218530e-01
2.32266873e-01 7.62314975e-01 5.16029835e-01 -1.06051838e+00
-4.22482938e-02 -1.05148935e+00 -2.47188151e-01 3.36728185e-01
1.89819083e-01 -9.25101519e-01 -6.53401136e-01 8.71269181e-02] | [5.5259599685668945, 7.896059989929199] |
91220fe5-a82e-49e4-bbe9-32d63902739c | face-anti-spoofing-with-human-material | 2007.02157 | null | https://arxiv.org/abs/2007.02157v1 | https://arxiv.org/pdf/2007.02157v1.pdf | Face Anti-Spoofing with Human Material Perception | Face anti-spoofing (FAS) plays a vital role in securing the face recognition systems from presentation attacks. Most existing FAS methods capture various cues (e.g., texture, depth and reflection) to distinguish the live faces from the spoofing faces. All these cues are based on the discrepancy among physical materials (e.g., skin, glass, paper and silicone). In this paper we rephrase face anti-spoofing as a material recognition problem and combine it with classical human material perception [1], intending to extract discriminative and robust features for FAS. To this end, we propose the Bilateral Convolutional Networks (BCN), which is able to capture intrinsic material-based patterns via aggregating multi-level bilateral macro- and micro- information. Furthermore, Multi-level Feature Refinement Module (MFRM) and multi-head supervision are utilized to learn more robust features. Comprehensive experiments are performed on six benchmark datasets, and the proposed method achieves superior performance on both intra- and cross-dataset testings. One highlight is that we achieve overall 11.3$\pm$9.5\% EER for cross-type testing in SiW-M dataset, which significantly outperforms previous results. We hope this work will facilitate future cooperation between FAS and material communities. | ['Xiaobai Li', 'Jingang Shi', 'Xuesong Niu', 'Zitong Yu', 'Guoying Zhao'] | 2020-07-04 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/318_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123520545.pdf | eccv-2020-8 | ['material-recognition'] | ['computer-vision'] | [ 3.85467857e-01 -3.61215115e-01 -5.40276542e-02 -1.73987731e-01
-3.56213123e-01 -3.38777155e-01 5.01708329e-01 -2.92198300e-01
1.43512383e-01 3.39985937e-01 6.82094023e-02 6.29406795e-02
-5.14671132e-02 -8.54636312e-01 -6.28264010e-01 -1.07580888e+00
-1.06414281e-01 -3.45057398e-01 2.33034611e-01 -2.78603166e-01
2.85647959e-01 8.93020391e-01 -1.69879425e+00 6.07106924e-01
4.64820206e-01 1.51345706e+00 -3.07267368e-01 2.15775669e-01
1.79006770e-01 3.98451149e-01 -4.68216568e-01 -6.57451093e-01
1.11200325e-01 -6.89678937e-02 -2.74317771e-01 5.95822409e-02
6.62831962e-01 -4.09412622e-01 -3.98959070e-01 1.11432767e+00
4.63068873e-01 -3.88180166e-01 8.07729721e-01 -1.35681057e+00
-6.38683975e-01 7.80833587e-02 -9.27464485e-01 -4.59277034e-02
5.81786335e-01 1.02276370e-01 4.20525610e-01 -1.13270020e+00
2.76061058e-01 1.71522951e+00 6.38098300e-01 6.62877202e-01
-8.06167245e-01 -1.37769830e+00 -1.65231228e-01 4.65917319e-01
-1.42087960e+00 -8.92639160e-01 1.36124659e+00 -3.47370476e-01
4.45252657e-01 3.42968941e-01 3.48147720e-01 1.41775143e+00
4.31710541e-01 6.08787477e-01 1.25306070e+00 -2.14364663e-01
-1.86023161e-01 2.40715250e-01 -2.17394069e-01 7.49702275e-01
5.56772709e-01 2.82928944e-01 -8.27072561e-01 -1.51619136e-01
6.52625382e-01 8.69871974e-02 -4.29157913e-01 -3.96211036e-02
-7.17081070e-01 4.39058512e-01 4.75301474e-01 2.38983214e-01
-4.71947603e-02 -2.34281763e-01 2.46374443e-01 1.48072287e-01
3.45252037e-01 -1.03540152e-01 -1.33322999e-01 4.04891521e-01
-7.78198540e-01 -1.42890317e-02 5.08623421e-01 7.04152882e-01
6.27618790e-01 1.98473886e-01 -2.49256343e-02 9.33428764e-01
9.41601574e-01 1.14950383e+00 1.94019035e-01 -4.34831768e-01
5.71634710e-01 3.75829875e-01 -1.10956535e-01 -1.78059351e+00
-2.60129958e-01 -2.35424832e-01 -1.11890268e+00 1.57213435e-01
5.84001057e-02 1.50751621e-01 -7.50270605e-01 1.45669365e+00
5.09076834e-01 5.04694223e-01 -7.38504305e-02 6.87736988e-01
9.47033048e-01 3.57343823e-01 -1.25884831e-01 -4.90644604e-01
1.48294199e+00 -5.75283885e-01 -7.69830823e-01 1.90991443e-02
6.83707446e-02 -1.09110832e+00 5.79062164e-01 6.93016648e-01
-7.29912937e-01 -8.30361009e-01 -1.30949223e+00 2.86758393e-01
-3.27744275e-01 2.02636220e-04 2.78482139e-01 1.07018769e+00
-6.86833084e-01 4.68764395e-01 -6.08083069e-01 1.02801416e-02
8.75667989e-01 5.76499283e-01 -6.76135540e-01 -2.57706642e-01
-1.09704697e+00 2.88077444e-01 1.88195072e-02 4.71381098e-01
-9.90358829e-01 -4.62955981e-01 -8.10288966e-01 -3.45221221e-01
1.17303342e-01 -1.92942191e-02 4.29372609e-01 -6.35342002e-01
-1.55482781e+00 8.15430105e-01 -1.03147253e-01 4.27465886e-02
2.87430733e-01 -1.49202719e-01 -9.81039464e-01 4.96424675e-01
-1.67719051e-01 1.87396258e-01 1.46833253e+00 -1.62954807e+00
-2.00424612e-01 -8.26175988e-01 -3.38560671e-01 -4.49027777e-01
-1.00360191e+00 3.45010787e-01 -3.28300059e-01 -8.33769858e-01
2.79473752e-01 -8.16477656e-01 6.51037455e-01 1.46622762e-01
-5.51227570e-01 -1.08263247e-01 1.35909379e+00 -8.36231887e-01
1.03840220e+00 -2.31590152e+00 -1.49354622e-01 2.48711795e-01
3.28219414e-01 7.13364244e-01 -2.86721826e-01 2.37635657e-01
-1.13698751e-01 7.85438046e-02 -1.96921192e-02 -4.44492042e-01
-1.20699771e-01 -3.19163054e-01 -1.15151152e-01 9.94183898e-01
4.69141275e-01 5.45612097e-01 -4.02898878e-01 -7.05005944e-01
1.89419270e-01 9.91998672e-01 -3.40915799e-01 1.46271840e-01
2.45685354e-01 5.70340157e-01 -4.83466774e-01 1.21828353e+00
1.61869657e+00 1.53502345e-01 -1.01708867e-01 -6.31678939e-01
2.51508445e-01 -3.11208908e-02 -1.22630453e+00 1.00998080e+00
-3.53311628e-01 4.06829178e-01 6.31528735e-01 -7.98322201e-01
1.09540761e+00 4.43077773e-01 5.16358852e-01 -8.54378700e-01
3.12079489e-01 2.60798931e-01 -2.10007951e-01 -7.39321709e-01
-2.91669481e-02 -6.98432996e-05 3.12207371e-01 1.79301992e-01
4.55280505e-02 1.53118297e-01 -3.97258341e-01 -2.15208843e-01
7.77260602e-01 -1.86899692e-01 -2.63481975e-01 -4.05156881e-01
9.61008430e-01 -9.14181769e-01 7.49322772e-01 1.47890598e-01
-3.34732860e-01 3.58777970e-01 1.37144685e-01 -2.22374335e-01
-5.08050323e-01 -9.77352321e-01 -5.08344769e-01 7.46089756e-01
5.23947537e-01 -1.79777443e-01 -7.66121626e-01 -7.36668646e-01
2.68152952e-01 -6.52296618e-02 -6.93718314e-01 -2.65147209e-01
-7.28251278e-01 -6.09008431e-01 7.17830718e-01 3.08155149e-01
8.48682046e-01 -9.95222509e-01 -1.73141271e-01 -5.53305969e-02
-2.56451312e-02 -1.22647536e+00 -1.68995723e-01 -4.45045173e-01
-4.78545815e-01 -1.26419997e+00 -4.70690668e-01 -8.47754896e-01
4.96941626e-01 5.74497163e-01 6.13118649e-01 4.84778225e-01
-4.11157936e-01 1.23220682e-01 -3.00880671e-01 -4.85462248e-01
1.08075538e-03 -4.56020027e-01 4.47478175e-01 7.85917580e-01
3.64004374e-01 -6.26910508e-01 -7.08986938e-01 7.49073327e-01
-8.91524136e-01 -9.14836079e-02 5.75656295e-01 7.33262718e-01
1.73017979e-01 3.48374844e-01 3.61948907e-01 -5.79077482e-01
2.49662027e-01 -4.64338481e-01 -2.68110901e-01 2.01226220e-01
-3.60109001e-01 -4.28070515e-01 4.16720331e-01 -4.39730942e-01
-1.05327094e+00 -3.37475389e-01 -2.07860291e-01 -4.81852531e-01
-3.38977426e-01 -2.77711879e-02 -9.77935016e-01 -6.18991017e-01
3.39071631e-01 2.29457349e-01 -6.31720945e-02 -4.81304467e-01
-2.41973445e-01 9.71100986e-01 3.36613387e-01 -5.23808181e-01
1.25920761e+00 8.32949638e-01 2.77477145e-01 -1.17664146e+00
-2.60236084e-01 -3.18145484e-01 -3.71024966e-01 -4.31206435e-01
5.30790389e-01 -8.58170688e-01 -1.25125611e+00 1.22597325e+00
-1.29909992e+00 9.38918963e-02 6.02376342e-01 3.42469603e-01
-7.79040740e-04 6.39932573e-01 -8.20384026e-01 -1.04688478e+00
-2.76621521e-01 -1.29254401e+00 1.28920507e+00 3.52214187e-01
4.38957095e-01 -7.37665117e-01 -4.19360310e-01 6.74139440e-01
5.48064113e-01 4.31068361e-01 4.24624920e-01 -3.18303972e-01
-4.58706588e-01 -2.33555362e-01 -5.60319126e-01 4.96388614e-01
5.11471033e-01 1.22929402e-01 -1.44869268e+00 -6.56993449e-01
3.31861347e-01 -3.24837685e-01 1.04678738e+00 -7.49565214e-02
1.24205482e+00 -3.96141201e-01 -5.24889708e-01 8.19646895e-01
1.19841218e+00 2.99818456e-01 8.03117931e-01 -9.31087509e-03
1.06262636e+00 9.50340986e-01 5.45055270e-01 4.63291794e-01
-5.18082529e-02 7.60450006e-01 6.82628751e-01 -7.57275075e-02
-2.78155535e-01 -3.24321873e-02 6.53517187e-01 6.76085949e-01
-3.31413224e-02 -3.93058956e-01 -6.46690249e-01 1.93938866e-01
-1.15316188e+00 -8.13940227e-01 -7.98820779e-02 1.85584342e+00
4.77165401e-01 1.36424899e-01 -1.17769442e-01 6.15973830e-01
8.24081004e-01 3.38550508e-01 -4.28134054e-01 6.04486838e-02
-3.07809889e-01 3.56854677e-01 3.56695741e-01 3.40280265e-01
-1.29547489e+00 6.59866810e-01 4.88182497e+00 1.24778152e+00
-1.45289969e+00 -7.32971951e-02 6.26543760e-01 1.99570760e-01
-2.91152358e-01 -5.55286646e-01 -1.01614904e+00 7.66603887e-01
4.96726662e-01 4.84990239e-01 3.74682128e-01 4.48899508e-01
3.90320756e-02 3.10396314e-01 -7.68146694e-01 1.25513101e+00
5.10470152e-01 -9.13295925e-01 -8.47097337e-02 3.17047983e-01
4.86535132e-01 -4.58218127e-01 4.68914717e-01 -3.12607408e-01
-2.42557123e-01 -1.25529361e+00 5.64339280e-01 3.95787656e-01
9.89264786e-01 -6.96143746e-01 7.25303292e-01 -1.41870096e-01
-1.60132992e+00 -2.31395617e-01 -1.81936845e-01 4.71189648e-01
-2.14840680e-01 6.96259856e-01 -4.14086014e-01 8.64595830e-01
8.62553954e-01 8.26940596e-01 -4.34225708e-01 7.08634794e-01
-1.58985078e-01 5.69140196e-01 -2.75897801e-01 2.66872168e-01
-3.76163781e-01 2.35533357e-01 5.18425345e-01 1.31425011e+00
2.63572216e-01 -1.12920232e-01 7.24336132e-04 5.78069985e-01
-2.18264848e-01 6.27074093e-02 -5.68976879e-01 -3.14094052e-02
4.42414135e-01 1.31787264e+00 -5.83269060e-01 4.74385656e-02
-3.57917339e-01 8.02703619e-01 -1.20062105e-01 1.81966871e-01
-6.93869531e-01 -5.02411127e-01 7.41825223e-01 2.46715456e-01
2.98828661e-01 -2.68212646e-01 -1.14563864e-03 -1.04893816e+00
3.01149040e-01 -9.71134841e-01 1.91627815e-01 -3.63583297e-01
-1.46278203e+00 5.29302537e-01 -3.06086868e-01 -1.13655758e+00
4.30626869e-01 -1.06671166e+00 -5.24789512e-01 7.76673496e-01
-1.84327149e+00 -1.53769743e+00 -6.58991933e-01 7.82369614e-01
2.12121665e-01 -3.54206890e-01 5.87025583e-01 6.39634907e-01
-8.12218070e-01 9.97084677e-01 -1.68648228e-01 4.64372963e-01
8.14568579e-01 -4.34181303e-01 7.49393106e-02 7.49016464e-01
-3.11278515e-02 7.01069832e-01 3.85195404e-01 -7.20066965e-01
-1.91581070e+00 -1.00472188e+00 4.09842640e-01 -1.76085889e-01
4.62049037e-01 -4.90713090e-01 -8.92807245e-01 5.57655245e-02
6.59709498e-02 3.91685545e-01 7.01880932e-01 -3.26779723e-01
-9.39371467e-01 -5.12911141e-01 -1.55708420e+00 1.15025483e-01
1.12833643e+00 -8.12931478e-01 -1.31609309e-02 1.44929186e-01
4.62956935e-01 -1.52181625e-01 -9.30068254e-01 7.69456208e-01
9.54628229e-01 -1.25961685e+00 1.07137311e+00 -2.03930408e-01
3.17043692e-01 -2.69462138e-01 -4.81177211e-01 -6.92838967e-01
-2.52049685e-01 -6.99797809e-01 -3.16042542e-01 1.69449425e+00
-1.27240390e-01 -7.26862252e-01 9.38319623e-01 -1.46706536e-01
7.90063813e-02 -7.51499474e-01 -8.91744614e-01 -8.39660943e-01
-2.63734341e-01 -4.07512277e-01 1.01598418e+00 8.86905372e-01
-2.58439422e-01 -2.29170322e-01 -5.26171803e-01 6.21029079e-01
1.03714108e+00 -9.69344005e-02 5.39830327e-01 -1.36001086e+00
-9.60537642e-02 -3.81691158e-01 -5.69263101e-01 -8.16746891e-01
3.54114830e-01 -5.43754458e-01 -1.17424414e-01 -6.88784242e-01
1.01624727e-01 -5.71971059e-01 -5.80480993e-01 2.76641756e-01
7.59270489e-02 8.03061366e-01 7.91026875e-02 1.10474899e-01
-2.69871086e-01 6.38200760e-01 1.29836905e+00 -5.65205932e-01
1.78611755e-01 -5.31218909e-02 -5.72867453e-01 6.23861074e-01
7.79269636e-01 -2.15648577e-01 -9.86442566e-02 -3.58657748e-01
-1.68652043e-01 -2.72887223e-03 6.22090816e-01 -1.23949707e+00
1.08166695e-01 -3.62553447e-01 7.62291193e-01 -4.44991201e-01
5.84158063e-01 -9.98804569e-01 -2.82340255e-02 4.65172797e-01
1.35778338e-01 -3.52337003e-01 4.00254995e-01 5.74947059e-01
-2.56176740e-01 1.98958591e-01 1.12566435e+00 2.90346175e-01
-2.90625662e-01 5.66626728e-01 1.10475026e-01 -2.81438410e-01
8.62352967e-01 -3.70897055e-01 -6.14364207e-01 1.36399046e-01
-2.44564801e-01 -2.66201019e-01 4.36236203e-01 5.10998726e-01
1.03703761e+00 -1.54906583e+00 -7.62858689e-01 8.49934578e-01
1.91395596e-01 -3.13390881e-01 4.51670676e-01 7.68940270e-01
-3.93100709e-01 3.83920074e-01 -4.57075089e-01 -5.80265701e-01
-1.56350648e+00 4.58746374e-01 1.76440254e-01 1.37775883e-01
-1.20059870e-01 1.14827073e+00 1.37511492e-01 7.82021508e-02
2.04651251e-01 2.34756306e-01 -3.69286895e-01 9.84359980e-02
9.28514838e-01 4.76290852e-01 2.64072657e-01 -1.18150353e+00
-6.63715065e-01 9.32758868e-01 -9.63953957e-02 1.97879046e-01
1.30823100e+00 -1.71761960e-02 -4.46039349e-01 -9.78009105e-02
1.58950579e+00 4.84466225e-01 -1.10039949e+00 -1.85064092e-01
-3.40853065e-01 -9.15831327e-01 -8.19175541e-02 -2.66117871e-01
-1.51263344e+00 1.12018955e+00 9.32533920e-01 2.82202899e-01
1.23043001e+00 -1.67327985e-01 1.00831473e+00 -7.31342435e-02
4.60402489e-01 -8.64848256e-01 6.10791922e-01 -3.79794315e-02
1.09666407e+00 -1.16859746e+00 5.44891134e-02 -8.71536911e-01
4.21051271e-02 1.16626477e+00 7.75518835e-01 9.55187231e-02
1.11294830e+00 3.09957206e-01 -8.48894939e-02 -3.02579552e-01
-2.72455454e-01 2.46174604e-01 4.52921152e-01 7.56589770e-01
2.99061686e-01 1.51296869e-01 2.52795458e-01 4.84575987e-01
-5.34938537e-02 -2.46244177e-01 -1.66736409e-01 8.55876386e-01
-2.84859776e-01 -1.18097925e+00 -6.99247599e-01 3.59441787e-01
-6.28516018e-01 2.64844805e-01 -3.29146117e-01 3.39118630e-01
7.10458457e-01 1.35025251e+00 -1.70500547e-01 -1.03428769e+00
5.56723960e-02 -3.90534967e-01 5.52881360e-01 -2.54196137e-01
-2.43137136e-01 1.45602643e-01 -2.22322509e-01 -6.45651340e-01
-4.84547973e-01 -6.31216705e-01 -6.55180991e-01 -5.84992349e-01
-4.63431239e-01 -2.04971462e-01 6.58612490e-01 7.55794704e-01
2.89962381e-01 2.05262735e-01 1.25070095e+00 -1.00645626e+00
-1.39854476e-01 -8.99503410e-01 -6.91776097e-01 4.79083627e-01
7.26052403e-01 -1.14429939e+00 -5.26657462e-01 -2.13066321e-02] | [12.983641624450684, 1.1238634586334229] |
81dcfd7d-32ba-4367-8227-ad37ebaabd9e | neustip-a-novel-neuro-symbolic-model-for-link | 2305.11301 | null | https://arxiv.org/abs/2305.11301v1 | https://arxiv.org/pdf/2305.11301v1.pdf | NeuSTIP: A Novel Neuro-Symbolic Model for Link and Time Prediction in Temporal Knowledge Graphs | While Knowledge Graph Completion (KGC) on static facts is a matured field, Temporal Knowledge Graph Completion (TKGC), that incorporates validity time into static facts is still in its nascent stage. The KGC methods fall into multiple categories including embedding-based, rule-based, GNN-based, pretrained Language Model based approaches. However, such dimensions have not been explored in TKG. To that end, we propose a novel temporal neuro-symbolic model, NeuSTIP, that performs link prediction and time interval prediction in a TKG. NeuSTIP learns temporal rules in the presence of the Allen predicates that ensure the temporal consistency between neighboring predicates in a given rule. We further design a unique scoring function that evaluates the confidence of the candidate answers while performing link prediction and time interval prediction by utilizing the learned rules. Our empirical evaluation on two time interval based TKGC datasets suggests that our model outperforms state-of-the-art models for both link prediction and the time interval prediction task. | ['Mausam', 'Garima Gaur', 'Navdeep Kaur', 'Ishaan Singh'] | 2023-05-15 | null | null | null | null | ['link-prediction', 'knowledge-graph-completion', 'temporal-knowledge-graph-completion'] | ['graphs', 'knowledge-base', 'knowledge-base'] | [-4.13956881e-01 4.28108513e-01 -9.48586106e-01 -3.47137064e-01
-1.39619589e-01 -4.05910254e-01 6.82421744e-01 5.74460924e-01
-1.12660415e-01 7.70923078e-01 3.03436041e-01 -5.32577753e-01
-8.39542747e-01 -1.18150175e+00 -6.74678326e-01 -1.68815404e-02
-6.85754299e-01 7.22241819e-01 7.84188688e-01 -3.08897853e-01
-5.91694843e-03 6.58353791e-02 -1.44825625e+00 1.14053503e-01
1.02487242e+00 1.11586130e+00 -5.02758622e-01 2.67932355e-01
-2.34978244e-01 1.15028965e+00 -2.66689599e-01 -7.26602793e-01
-8.59716609e-02 1.16423555e-02 -9.03763771e-01 -8.25583220e-01
-1.48801818e-01 -7.41513446e-02 -8.29200149e-01 7.25490510e-01
-3.25798988e-03 5.03267348e-01 4.78467613e-01 -1.66874301e+00
-9.56464112e-01 1.10015774e+00 3.16329226e-02 3.51459354e-01
6.83147609e-01 -2.88513988e-01 9.92996216e-01 -5.39986312e-01
9.51156139e-01 1.00735438e+00 8.93812120e-01 2.26372883e-01
-8.22377682e-01 -4.83506680e-01 4.65165019e-01 1.08960569e+00
-1.39260530e+00 9.79608744e-02 9.12288129e-01 -2.66685665e-01
1.43716729e+00 -1.31389558e-01 1.06917620e+00 1.04462314e+00
4.31303144e-01 4.82974768e-01 7.98643053e-01 -4.16447282e-01
3.89856517e-01 -2.51123548e-01 3.71134579e-01 1.01846707e+00
1.77596211e-01 5.00554979e-01 -1.26842487e+00 -8.25325251e-02
5.45787275e-01 -1.78026393e-01 -1.53109685e-01 -1.62926689e-01
-1.26117241e+00 6.64344311e-01 1.93250284e-01 2.02265531e-01
-3.19400787e-01 3.23168278e-01 5.42063475e-01 4.88033921e-01
5.84019184e-01 3.18100274e-01 -8.56028855e-01 -2.88027406e-01
-8.67417872e-01 3.69979471e-01 9.91017640e-01 1.16653347e+00
4.53578860e-01 4.16857656e-03 -4.09511954e-01 2.63688564e-01
1.40700072e-01 2.92474870e-03 3.85495126e-01 -7.57098794e-01
4.76663560e-01 1.03272915e+00 8.45204145e-02 -1.32732284e+00
-4.99779940e-01 -3.22236091e-01 -3.49048227e-01 -2.83201456e-01
5.78621589e-02 1.58330053e-01 -8.05527329e-01 1.82591712e+00
4.89385426e-01 9.49435651e-01 3.13604087e-01 4.65933621e-01
8.08388531e-01 6.75532937e-01 1.42852724e-01 -3.76701742e-01
9.90830243e-01 -7.21110940e-01 -9.55426931e-01 1.31774515e-01
6.71752751e-01 -6.10967800e-02 5.84210992e-01 3.80447954e-01
-6.91114783e-01 -3.66192400e-01 -1.17325187e+00 -1.04529336e-01
-1.03091860e+00 -3.00134748e-01 1.18794382e+00 2.35099494e-01
-9.52331722e-01 8.52639735e-01 -9.88573730e-01 -3.77489924e-01
-5.39488122e-02 2.78102547e-01 -2.78211564e-01 -1.36155769e-01
-2.06422186e+00 1.30893481e+00 1.20383883e+00 -8.18128809e-02
-7.05821335e-01 -9.09343779e-01 -1.08527184e+00 -3.49708907e-02
1.01491201e+00 -5.47355711e-01 8.53501141e-01 1.02332123e-01
-1.23240817e+00 4.19867039e-01 2.48450274e-03 -1.00718057e+00
1.71925530e-01 -1.86097726e-01 -1.52075708e+00 -6.93615898e-02
2.27794275e-01 3.63856673e-01 7.05924213e-01 -8.19472492e-01
-7.13181615e-01 -2.62194067e-01 2.16709644e-01 -7.27121010e-02
-2.03996927e-01 -4.84079331e-01 -8.69342804e-01 -5.90744853e-01
1.35105014e-01 -6.90546989e-01 4.22386192e-02 -2.37064138e-01
-3.98482531e-01 -8.08790743e-01 8.59384894e-01 -7.34163165e-01
1.82662272e+00 -1.86207533e+00 8.64023715e-02 3.93965930e-01
-1.67011060e-02 1.17596433e-01 8.68454352e-02 7.34873056e-01
2.97655966e-02 -4.56540808e-02 2.99349159e-01 4.95386496e-02
1.60748944e-01 5.71127117e-01 -6.00074708e-01 -4.19228687e-04
-8.53170007e-02 1.29691148e+00 -1.19106078e+00 -8.66444886e-01
6.62259832e-02 1.55665889e-01 -3.45714688e-01 -2.27002904e-01
-7.89874554e-01 -1.35273397e-01 -4.72273886e-01 8.76853764e-01
1.84120890e-02 -2.45341837e-01 4.94667143e-01 -4.51531023e-01
-6.98540509e-02 2.15340003e-01 -8.83095741e-01 1.95330930e+00
-5.24544157e-02 1.96696579e-01 -9.78745878e-01 -8.01663280e-01
9.23286140e-01 5.05974352e-01 5.21611810e-01 -7.13232517e-01
-1.49715737e-01 1.04343295e-01 -2.67451376e-01 -6.70627356e-01
9.16608274e-01 1.51872516e-01 -2.63452947e-01 9.29337144e-02
3.06135446e-01 1.41969129e-01 3.80023122e-01 4.53669220e-01
1.31447935e+00 6.40040100e-01 5.07213771e-01 5.34125790e-02
2.59466797e-01 3.43178689e-01 9.48640049e-01 4.39653873e-01
-3.30555409e-01 -2.52156377e-01 5.62157512e-01 -5.15466869e-01
-5.12913048e-01 -1.02880764e+00 2.24215508e-01 8.72392833e-01
2.97946364e-01 -9.79808569e-01 -1.90297253e-02 -1.01239002e+00
5.52311838e-01 1.08991659e+00 -9.37473476e-01 -4.79588717e-01
-2.46766075e-01 4.51429784e-02 8.02340925e-01 9.13182855e-01
3.71283740e-01 -8.84971023e-01 -2.76915520e-01 5.03179729e-01
-2.12326914e-01 -1.11940777e+00 -2.58361697e-01 2.31004193e-01
-8.96387875e-01 -1.35547590e+00 2.41572216e-01 -6.72087491e-01
2.11264312e-01 -3.71504694e-01 9.37025905e-01 -2.13172555e-01
1.41226128e-01 7.08596170e-01 -6.02395177e-01 -4.59476769e-01
2.50000119e-01 -4.53844182e-02 3.06464702e-01 -2.80813962e-01
5.76682150e-01 -9.24619317e-01 -3.77737135e-01 2.49081224e-01
-6.30132914e-01 4.24766429e-02 1.97538957e-01 4.92720664e-01
9.61446643e-01 6.48532331e-01 8.75490069e-01 -8.44752073e-01
6.38027489e-01 -7.39671886e-01 -6.41660035e-01 8.28931451e-01
-1.32453060e+00 3.43165070e-01 5.10963261e-01 -5.96076727e-01
-1.01488733e+00 -4.40442175e-01 5.13019323e-01 -7.31445789e-01
3.37852925e-01 1.46919215e+00 1.88451007e-01 2.75612861e-01
5.31008780e-01 3.39456737e-01 -3.31533551e-01 -1.88877106e-01
7.50579655e-01 -2.27069095e-01 1.05459321e+00 -9.81649518e-01
8.18141699e-01 3.60240310e-01 2.36168057e-01 -3.29707973e-02
-1.08062053e+00 -3.61318767e-01 -4.53711092e-01 -4.60944563e-01
6.70866668e-01 -7.06454575e-01 -8.47029746e-01 -1.59710750e-01
-1.14106703e+00 -3.38359356e-01 -3.98273557e-01 6.49357200e-01
-4.79083925e-01 2.56999254e-01 -3.92183036e-01 -7.65392005e-01
-2.69140214e-01 -3.35267127e-01 5.01273453e-01 2.95580532e-02
-3.63802105e-01 -1.37305045e+00 1.71334535e-01 7.40079954e-02
2.13413909e-01 7.37455547e-01 1.37630045e+00 -8.98932219e-01
-6.50047123e-01 -4.10920978e-01 1.38809666e-01 -2.27757886e-01
1.70428492e-02 7.82326236e-03 -2.76715070e-01 1.92627966e-01
-4.15507674e-01 -2.49609604e-01 7.03917384e-01 9.68387797e-02
1.10683322e+00 -3.41857761e-01 -7.38228977e-01 4.62152064e-01
1.46501756e+00 5.78664720e-01 6.11617684e-01 4.19371724e-01
4.66629475e-01 3.31769019e-01 1.14463103e+00 3.49452794e-01
8.21569562e-01 5.43797135e-01 3.57715875e-01 8.18294704e-01
-3.72584127e-02 -8.86976957e-01 1.82479620e-01 8.83985996e-01
-5.06428242e-01 -3.00398290e-01 -9.64373708e-01 1.01658225e+00
-2.41760898e+00 -1.10295057e+00 -1.32896435e-02 1.83857667e+00
1.17479813e+00 5.25485039e-01 -2.37085789e-01 3.00320417e-01
5.95615745e-01 2.10107058e-01 -6.97867215e-01 -1.21046148e-01
-2.45634671e-02 1.79643556e-01 3.91628295e-01 4.10431623e-01
-9.09771323e-01 1.22645807e+00 5.78536081e+00 9.62634683e-01
-7.24032521e-01 8.26133117e-02 -2.83069760e-02 1.08795196e-01
-4.10281986e-01 4.45307046e-01 -8.18426311e-01 3.40401709e-01
8.56578588e-01 -6.48980439e-01 4.05252874e-01 9.37865019e-01
-7.43084475e-02 -6.86964020e-02 -1.28115010e+00 8.93111348e-01
1.76429339e-02 -1.51907575e+00 8.54543298e-02 -2.36751795e-01
8.23942721e-01 -4.80861962e-01 -5.23953997e-02 1.00072944e+00
4.41573054e-01 -7.99606144e-01 6.74623609e-01 1.07823169e+00
8.01138401e-01 -7.64329374e-01 4.67512250e-01 7.85568953e-02
-1.47380579e+00 -3.92666319e-03 -6.96473196e-02 6.30538724e-03
3.67165357e-01 7.66941726e-01 -9.92381990e-01 1.45700848e+00
5.74346960e-01 1.04267883e+00 -6.22118950e-01 7.15465724e-01
-6.15805149e-01 8.40530694e-01 -1.61966518e-01 6.09730072e-02
-4.86068008e-03 7.82517269e-02 4.16975290e-01 5.95552266e-01
3.87407869e-01 2.64609814e-01 1.61618397e-01 7.33946621e-01
6.13908358e-02 -1.85184240e-01 -6.92551851e-01 -1.64946556e-01
9.31451380e-01 7.08188295e-01 -5.00209272e-01 -3.27764124e-01
-3.54036659e-01 6.31170392e-01 5.93211353e-01 4.96602774e-01
-1.18368459e+00 -3.68281722e-01 2.62200564e-01 -1.32841796e-01
4.90731418e-01 -3.79736930e-01 -3.07496846e-01 -1.00184619e+00
2.03438193e-01 -2.98596203e-01 1.02199399e+00 -8.46645713e-01
-1.04978085e+00 3.52364182e-01 4.91380662e-01 -1.24326897e+00
-4.01364654e-01 -4.13316041e-01 -6.25480175e-01 4.31333870e-01
-1.51267612e+00 -1.33995390e+00 -9.58437845e-02 9.43085909e-01
1.05431713e-01 -2.91317821e-01 8.55332971e-01 9.15825963e-02
-4.06284869e-01 5.21340013e-01 -3.60179752e-01 -2.06381716e-02
5.04150629e-01 -1.15218782e+00 1.32660732e-01 9.13087785e-01
9.04667284e-03 9.10973489e-01 6.25059485e-01 -1.40173388e+00
-1.42434883e+00 -1.54500270e+00 1.26345611e+00 -5.08405149e-01
1.01014256e+00 2.53811866e-01 -1.00151956e+00 1.18033755e+00
-1.38968587e-01 1.25046939e-01 6.46983743e-01 6.54781520e-01
-7.53821492e-01 -2.65773147e-01 -7.93735147e-01 5.62701643e-01
1.38697851e+00 -6.67548835e-01 -1.03567505e+00 1.83766276e-01
1.26484978e+00 -4.11042124e-01 -1.41841996e+00 6.33100986e-01
4.38916892e-01 -4.73842174e-01 9.72010612e-01 -9.23406959e-01
2.88271129e-01 -5.56315720e-01 -7.34030157e-02 -1.04805708e+00
-3.24705511e-01 -6.65059865e-01 -1.25811315e+00 1.27596998e+00
6.19624794e-01 -5.97564459e-01 7.87553966e-01 6.15838826e-01
-3.83033663e-01 -1.01770234e+00 -1.08163440e+00 -1.37980294e+00
-4.52962458e-01 -9.29792821e-01 9.09588516e-01 1.25181234e+00
5.66144586e-01 1.13230221e-01 -4.20608222e-01 3.12156349e-01
3.79758656e-01 3.15392077e-01 1.89888448e-01 -1.49841642e+00
-1.71105817e-01 -5.23200259e-03 -5.67068517e-01 -4.38160330e-01
4.34031457e-01 -1.07792377e+00 -3.76675993e-01 -1.90424454e+00
-3.86740744e-01 -3.19859624e-01 -3.55002761e-01 1.04007888e+00
1.14043718e-02 -6.00118697e-01 -4.03827071e-01 4.71906103e-02
-9.41169560e-01 7.78385997e-01 9.75843668e-01 -2.29382172e-01
-2.60886431e-01 -4.09841359e-01 -3.71092737e-01 3.76364797e-01
7.50547528e-01 -3.42639923e-01 -1.06515968e+00 -1.17388420e-01
7.85669684e-01 5.75394571e-01 4.02273059e-01 -1.13491642e+00
9.02121127e-01 -4.01480436e-01 8.72246772e-02 -9.65344369e-01
4.38353509e-01 -7.14721441e-01 4.11299378e-01 3.04966658e-01
-2.03651950e-01 2.29927972e-01 3.27822208e-01 1.27708101e+00
-4.15800422e-01 2.36966968e-01 -1.08151153e-01 2.58413792e-01
-1.52772427e+00 5.50146580e-01 1.50321871e-01 1.23460786e-02
1.08085191e+00 -2.38025099e-01 -6.57278776e-01 -2.35902742e-01
-8.54526520e-01 5.87313056e-01 -1.70481473e-01 5.49791276e-01
9.65098441e-01 -1.66122282e+00 -9.25195888e-02 -4.24757719e-01
5.59954524e-01 -1.13291889e-01 1.43195465e-01 8.36431921e-01
-7.00251758e-02 7.13769138e-01 1.11567087e-01 4.40627262e-02
-6.76886618e-01 1.12936091e+00 -5.97408526e-02 -6.69128239e-01
-5.88382423e-01 5.61084211e-01 -7.83622265e-01 -2.61855334e-01
4.24415469e-01 -6.44684792e-01 -4.34224904e-01 -7.03977197e-02
5.37809655e-02 5.44010282e-01 6.39977902e-02 3.70584056e-02
-4.80112702e-01 2.21871942e-01 -6.64727911e-02 -2.12686181e-01
1.16016030e+00 3.10232311e-01 -6.51268736e-02 6.45548522e-01
6.55998290e-01 -2.87534088e-01 -6.79383099e-01 -5.35493970e-01
4.08451378e-01 -1.03337705e-01 -5.75384200e-02 -1.16535950e+00
-6.59793556e-01 2.97304373e-02 3.64505351e-02 1.78424418e-01
9.88847613e-01 -1.35848299e-02 8.37551117e-01 5.51236153e-01
7.59753108e-01 -1.41552925e+00 3.76875214e-02 9.10813451e-01
7.83178687e-01 -8.99082541e-01 1.57243997e-01 -6.63178146e-01
-5.89104116e-01 9.70931411e-01 9.27761316e-01 2.00112939e-01
9.94754732e-01 -2.30809823e-01 -5.34321666e-01 -5.77000558e-01
-1.22725701e+00 -3.70988607e-01 5.56767285e-01 8.56946886e-01
9.01375040e-02 1.95386350e-01 -5.84416509e-01 9.78851020e-01
-4.65326339e-01 4.20954645e-01 5.09329960e-02 1.02245545e+00
-1.90905288e-01 -9.97473300e-01 2.54631620e-02 5.58752418e-01
-9.72627401e-02 -4.50373366e-02 -2.12458670e-01 1.06468749e+00
2.79777527e-01 9.09265935e-01 -3.47043276e-01 -8.92610431e-01
4.91280943e-01 5.52811444e-01 4.27761227e-01 -5.31240404e-01
-1.93338603e-01 -6.36547863e-01 4.60571200e-01 -7.28517056e-01
-3.84453416e-01 -4.47553813e-01 -1.52495265e+00 -3.33902985e-01
-2.29965225e-01 2.77651757e-01 2.15688780e-01 9.50783670e-01
4.29383516e-01 6.25692785e-01 -7.47422725e-02 4.03792970e-02
-6.17823489e-02 -6.41802669e-01 -6.34215713e-01 1.88737527e-01
-1.98821828e-01 -1.15191877e+00 5.36853224e-02 9.33055580e-03] | [8.552845001220703, 7.916233539581299] |
941db61d-76c9-48c3-8237-057173deac77 | gmlight-lighting-estimation-via-geometric | 2102.10244 | null | https://arxiv.org/abs/2102.10244v2 | https://arxiv.org/pdf/2102.10244v2.pdf | GMLight: Lighting Estimation via Geometric Distribution Approximation | Inferring the scene illumination from a single image is an essential yet challenging task in computer vision and computer graphics. Existing works estimate lighting by regressing representative illumination parameters or generating illumination maps directly. However, these methods often suffer from poor accuracy and generalization. This paper presents Geometric Mover's Light (GMLight), a lighting estimation framework that employs a regression network and a generative projector for effective illumination estimation. We parameterize illumination scenes in terms of the geometric light distribution, light intensity, ambient term, and auxiliary depth, which can be estimated by a regression network. Inspired by the earth mover's distance, we design a novel geometric mover's loss to guide the accurate regression of light distribution parameters. With the estimated light parameters, the generative projector synthesizes panoramic illumination maps with realistic appearance and high-frequency details. Extensive experiments show that GMLight achieves accurate illumination estimation and superior fidelity in relighting for 3D object insertion. The codes are available at \href{https://github.com/fnzhan/Illumination-Estimation}{https://github.com/fnzhan/Illumination-Estimation}. | ['WenBo Hu', 'Xuansong Xie', 'Feiying Ma', 'Ling Shao', 'Shijian Lu', 'Changgong Zhang', 'Rongliang Wu', 'Yingchen Yu', 'Fangneng Zhan'] | 2021-02-20 | null | null | null | null | ['lighting-estimation'] | ['computer-vision'] | [ 2.11773500e-01 -4.65847313e-01 2.42646247e-01 -6.30376041e-01
-5.47538996e-01 -4.27476853e-01 3.25965106e-01 -7.34229863e-01
-8.68873578e-03 5.44069767e-01 1.05357051e-01 -1.97560459e-01
2.03714713e-01 -8.02963495e-01 -7.32713103e-01 -1.02356899e+00
6.26437426e-01 1.01860892e-03 -1.80416107e-01 -9.64020640e-02
2.49209434e-01 3.97920668e-01 -1.53096473e+00 -1.74229726e-01
9.18671906e-01 9.10329223e-01 3.62873226e-01 8.55597138e-01
-9.08654928e-02 6.56927049e-01 -4.67945039e-01 -2.33034834e-01
5.51921487e-01 -6.19470894e-01 -9.66599509e-02 6.64726198e-02
6.58947766e-01 -6.37873590e-01 -4.77614909e-01 1.18736124e+00
6.35483921e-01 2.00105175e-01 6.31960988e-01 -1.33681893e+00
-8.94069552e-01 -1.58662245e-01 -7.51221836e-01 -2.19237417e-01
4.12591159e-01 3.78896147e-01 4.39812899e-01 -8.80300939e-01
4.14910197e-01 1.22249556e+00 5.30151844e-01 5.28088808e-01
-1.15787113e+00 -8.66694927e-01 -3.12714800e-02 7.06881210e-02
-1.56021607e+00 -3.82468820e-01 1.20849872e+00 -1.36232302e-01
2.63922483e-01 4.15342033e-01 6.54462576e-01 9.43770587e-01
2.79091120e-01 5.31793952e-01 1.21747458e+00 -6.17284834e-01
2.74070241e-02 2.61540294e-01 -3.59635949e-01 7.95521617e-01
3.64787392e-02 2.93340653e-01 -4.53631818e-01 1.31886572e-01
1.28432012e+00 1.29099652e-01 -7.98841357e-01 -1.49155512e-01
-9.39901173e-01 5.74212551e-01 7.17663109e-01 -1.97594821e-01
-1.97914034e-01 4.09268588e-01 -2.82109320e-01 1.63811427e-02
6.51241839e-01 1.57894194e-01 -1.17625624e-01 1.00316852e-01
-5.90191364e-01 -3.74245383e-02 6.60106242e-01 1.18426764e+00
1.11140156e+00 2.47941643e-01 -8.42980444e-02 1.00476956e+00
5.05727410e-01 1.15170515e+00 5.95848225e-02 -1.29788363e+00
2.34063074e-01 2.77174324e-01 2.65934199e-01 -1.05923772e+00
-1.62733749e-01 -9.96591225e-02 -9.29305911e-01 5.15957057e-01
3.33323509e-01 -2.92636245e-01 -9.10091043e-01 1.59440124e+00
6.68414474e-01 2.83561379e-01 -2.49413863e-01 1.20360243e+00
9.04444516e-01 8.86704743e-01 -3.34233940e-01 -2.07260609e-01
1.06002891e+00 -9.76150632e-01 -7.25341558e-01 -3.08039133e-02
8.15862566e-02 -1.11718702e+00 1.22498643e+00 5.29583395e-01
-1.12883413e+00 -5.90680838e-01 -8.36882770e-01 -4.91268218e-01
2.02198941e-02 2.23787129e-01 5.44913530e-01 5.94554126e-01
-1.07771993e+00 3.50775450e-01 -6.05016530e-01 -9.71783176e-02
3.08970362e-01 2.11801939e-02 2.49741077e-01 -4.23096955e-01
-8.43120873e-01 6.47511840e-01 -1.14618547e-01 1.52392194e-01
-8.20321858e-01 -7.99149156e-01 -8.39194834e-01 -2.93106943e-01
9.85698402e-02 -8.30903888e-01 9.55924273e-01 -9.59636748e-01
-1.91618109e+00 6.44847035e-01 -3.50366890e-01 2.65082359e-01
5.19199550e-01 -1.81309506e-01 -2.04966903e-01 4.69362959e-02
-2.27387443e-01 6.12371981e-01 9.70324457e-01 -1.61279047e+00
-4.48961765e-01 -1.80891901e-01 -1.59266040e-01 5.22222638e-01
-1.75092369e-01 -2.66165644e-01 -7.55973876e-01 -4.02610779e-01
1.13665394e-01 -7.73731589e-01 -7.26625770e-02 5.37426293e-01
-6.36592686e-01 1.74993083e-01 7.78719187e-01 -6.59523487e-01
8.97619843e-01 -2.03050709e+00 -6.65622577e-02 1.20692968e-01
5.73968738e-02 -1.30685493e-01 -1.56996056e-01 9.35167894e-02
2.30061356e-02 -4.04350311e-01 -2.41797611e-01 -5.03089488e-01
-3.37082893e-03 5.82261607e-02 -3.15181762e-01 7.23269880e-01
-2.99977899e-01 8.56974900e-01 -9.11133289e-01 -3.55026931e-01
7.44148016e-01 1.34466326e+00 -3.78913313e-01 4.80921209e-01
-1.49557978e-01 7.80759513e-01 -2.93120712e-01 7.25001335e-01
8.58540535e-01 -5.67342378e-02 -2.57567137e-01 -5.72454095e-01
-2.25774124e-01 -1.59323514e-01 -1.18588781e+00 1.74230158e+00
-7.67508149e-01 8.25518847e-01 -1.51646193e-02 -1.96552753e-01
1.20877266e+00 -3.08211651e-02 2.83357233e-01 -8.41180444e-01
4.62363958e-01 9.58832800e-02 -6.61541402e-01 -4.83584940e-01
4.10055727e-01 -3.51368934e-02 4.41657513e-01 4.13780719e-01
-4.54547018e-01 -9.09384549e-01 -1.90756574e-01 -1.53483093e-01
5.62396884e-01 6.07925177e-01 3.98357250e-02 3.09023093e-02
5.10734439e-01 -4.65487361e-01 6.13143623e-01 3.32573503e-01
4.34047952e-02 9.84081686e-01 -7.88372159e-02 -5.49415410e-01
-1.19135153e+00 -1.10291255e+00 -2.89177954e-01 8.52458119e-01
5.30970573e-01 2.71143243e-02 -8.15371454e-01 -1.27282396e-01
-2.79373765e-01 1.02168858e+00 -4.87249583e-01 1.01359533e-02
-5.27190149e-01 -5.62214136e-01 4.86040721e-03 5.41035011e-02
6.73087418e-01 -8.03335905e-01 -4.49398220e-01 -3.07919264e-01
-5.05602181e-01 -8.56872678e-01 -7.05666304e-01 -2.47319400e-01
-4.97006625e-01 -9.32109177e-01 -9.30773139e-01 -5.26664317e-01
1.14276993e+00 6.30975246e-01 1.21040845e+00 8.82833675e-02
-6.95783973e-01 5.30548513e-01 -8.96609053e-02 -5.60112596e-01
-2.97744989e-01 -4.28975582e-01 -1.32596731e-01 9.25542191e-02
1.58765525e-01 -8.00806105e-01 -1.26428175e+00 6.66258097e-01
-7.99419284e-01 5.97406566e-01 2.95841634e-01 3.90319377e-01
7.48158097e-01 4.82948795e-02 -3.11471730e-01 -6.45797849e-01
1.32738546e-01 -2.17010647e-01 -1.08593905e+00 1.15821756e-01
-4.84414786e-01 -2.25916788e-01 5.82218349e-01 -4.56824034e-01
-1.44633794e+00 -1.14914790e-01 9.97718144e-03 -7.77003527e-01
-2.18696713e-01 -4.71144408e-01 -2.59767234e-01 -3.33713382e-01
7.34301090e-01 3.66834551e-01 -2.34623969e-01 -3.43457252e-01
4.48743701e-01 5.37118196e-01 5.22204578e-01 -5.85491419e-01
1.16331494e+00 8.24478209e-01 1.84655353e-01 -8.55649233e-01
-9.34318721e-01 -2.33970582e-01 -3.29441458e-01 -5.08203089e-01
7.64867127e-01 -9.20362711e-01 -8.80414844e-01 5.86821616e-01
-9.96942282e-01 -8.42122376e-01 -2.30854824e-01 4.25995111e-01
-5.50077200e-01 2.83977807e-01 -5.02915978e-01 -7.53927052e-01
-3.97472858e-01 -1.06339085e+00 1.14039671e+00 6.25944138e-01
2.47107521e-01 -1.06311965e+00 2.02112451e-01 3.81252438e-01
5.41678429e-01 3.54571223e-01 5.86261213e-01 8.09146941e-01
-9.98589516e-01 5.58166020e-02 -4.18000579e-01 4.35627401e-01
4.01354969e-01 3.04769784e-01 -1.20099127e+00 -2.50805050e-01
1.20570593e-01 5.93033694e-02 5.16164303e-01 7.65916824e-01
1.40683103e+00 -2.37070352e-01 -4.99867462e-02 1.46217394e+00
1.94978976e+00 2.55016297e-01 8.63935351e-01 1.93015680e-01
1.06419861e+00 4.02354151e-01 4.25413549e-01 5.80168009e-01
4.54261780e-01 6.30018830e-01 6.51426017e-01 -3.46725106e-01
-5.61176121e-01 -3.25964063e-01 1.94424972e-01 7.91541874e-01
-3.61982048e-01 -3.75462234e-01 -5.77756226e-01 1.20624863e-01
-1.38631022e+00 -7.23193586e-01 -2.12946162e-01 2.36279058e+00
9.53684151e-01 -3.67924392e-01 -2.43294567e-01 -1.93928838e-01
7.24811018e-01 8.26124400e-02 -7.16096103e-01 -2.83543557e-01
-1.57172337e-01 -1.79478139e-01 7.10880220e-01 9.49917972e-01
-5.70779443e-01 7.43848264e-01 5.01068687e+00 6.84896231e-01
-1.26155603e+00 3.92715037e-02 7.85226107e-01 -2.95331389e-01
-7.74835408e-01 -3.25165913e-02 -8.80491912e-01 7.06898153e-01
3.54016006e-01 -2.01313160e-02 9.60693359e-01 5.96533000e-01
3.64670008e-01 -2.98081189e-01 -7.11931527e-01 1.26929212e+00
3.67609322e-01 -7.66828418e-01 -7.75167197e-02 -4.33209017e-02
1.02388918e+00 -7.01658726e-02 3.95296276e-01 -1.98874190e-01
4.53009903e-01 -7.55750775e-01 4.15368229e-01 9.42965567e-01
1.11851668e+00 -5.82704008e-01 1.61249474e-01 5.42480610e-02
-1.05083156e+00 2.17247605e-01 -5.94339788e-01 2.42709950e-01
3.84298205e-01 7.42270231e-01 -4.21786696e-01 2.87030369e-01
6.71726763e-01 7.22880900e-01 -2.72550672e-01 9.50678527e-01
-7.83474863e-01 3.93654436e-01 -3.91754717e-01 4.21484597e-02
-2.38220289e-01 -7.89123833e-01 4.90811259e-01 9.07502353e-01
5.73264122e-01 2.59654224e-01 -4.44394425e-02 1.10309052e+00
-1.27464786e-01 6.14225306e-02 -4.54725564e-01 7.47992575e-01
4.33660656e-01 1.45353758e+00 -5.58465958e-01 -4.68913764e-02
-3.07941467e-01 1.09739697e+00 3.78102288e-02 8.65485489e-01
-1.15417325e+00 -5.23068726e-01 5.99360943e-01 1.25151485e-01
-8.41094032e-02 -4.31482866e-02 -2.37367421e-01 -1.14185953e+00
-3.79179441e-03 -4.23922390e-01 -2.17285186e-01 -1.57252383e+00
-1.04729688e+00 4.85551357e-01 -8.75327513e-02 -1.33073807e+00
2.95746952e-01 -5.32821596e-01 -7.54369318e-01 1.02100801e+00
-1.86958253e+00 -1.23539114e+00 -9.78719592e-01 8.04745615e-01
6.24140203e-01 2.79549986e-01 5.95630527e-01 3.23252648e-01
-6.34328663e-01 3.96789432e-01 2.96450317e-01 -1.48387402e-01
9.94997084e-01 -1.11828661e+00 1.76857919e-01 7.93277085e-01
-1.45439684e-01 3.14889878e-01 8.76444280e-01 -3.08104813e-01
-1.48939860e+00 -1.12649059e+00 2.66468555e-01 -4.21470881e-01
2.11064860e-01 -3.29229534e-01 -6.98560834e-01 6.97582066e-01
3.17457169e-01 6.28601983e-02 3.29751790e-01 -3.74376088e-01
-3.27720553e-01 -4.32116240e-01 -1.15041173e+00 8.39586496e-01
9.82776701e-01 -3.77618670e-01 1.48921296e-01 4.69874561e-01
6.67152405e-01 -6.50386751e-01 -6.85013890e-01 -4.53338632e-03
6.41675711e-01 -1.14436758e+00 1.06704521e+00 3.63178104e-01
6.27257228e-01 -5.48313856e-01 -2.74974227e-01 -1.34326553e+00
-2.10876524e-01 -8.50868583e-01 -2.72399094e-02 1.26295602e+00
4.30078320e-02 -8.24614584e-01 5.91089904e-01 7.85151124e-01
-7.17164129e-02 -6.53959692e-01 -4.44844335e-01 -4.50963497e-01
-2.48160928e-01 -3.91723990e-01 7.84408331e-01 9.46675718e-01
-7.61915624e-01 1.30469218e-01 -6.49317682e-01 3.54390413e-01
1.26333177e+00 2.13921458e-01 1.12834787e+00 -8.14978957e-01
-2.10496634e-01 -2.45116577e-01 9.10701416e-03 -1.22995782e+00
-1.44983232e-01 -4.83937234e-01 2.90009230e-01 -1.66045761e+00
2.04955637e-01 -5.68234801e-01 -5.13256192e-02 2.30629936e-01
-1.86581135e-01 5.48469484e-01 9.32906121e-02 2.39905715e-01
-2.93977469e-01 6.34317696e-01 1.62124324e+00 6.50563091e-02
-3.26137364e-01 -5.56262545e-02 -5.38506925e-01 8.06089401e-01
9.73219037e-01 -2.31498599e-01 -5.06381094e-01 -8.71558309e-01
4.26910162e-01 -1.30883321e-01 4.29289609e-01 -8.69953990e-01
2.39221141e-01 -3.79498243e-01 6.42045796e-01 -5.39106488e-01
6.03775501e-01 -7.69935548e-01 4.45007026e-01 1.01024739e-01
-1.08883627e-01 -2.28318051e-01 1.55726084e-02 3.29040766e-01
2.05718488e-01 1.94681231e-02 1.06110251e+00 -1.59314021e-01
-4.43597943e-01 6.33132517e-01 2.38514334e-01 -3.08921151e-02
7.91020751e-01 -4.70325500e-01 -5.63358724e-01 -5.07231414e-01
-9.27507207e-02 3.47304740e-03 8.07058513e-01 1.61464810e-01
8.77916455e-01 -1.48250973e+00 -6.94252074e-01 3.61030251e-01
-4.26551104e-02 3.00940394e-01 4.29424375e-01 7.72508025e-01
-9.15271282e-01 -3.61368023e-02 -7.81700108e-03 -6.72862649e-01
-1.24248683e+00 3.33095163e-01 4.85012531e-01 4.31404263e-01
-7.10170865e-01 1.04000461e+00 6.83946192e-01 -3.61920118e-01
1.50292203e-01 -1.68217734e-01 2.16196850e-01 -4.81504083e-01
5.85687697e-01 5.29886425e-01 -3.78657371e-01 -5.70411980e-01
1.05481811e-01 1.10166287e+00 3.04098696e-01 1.80237636e-01
1.31743073e+00 -6.91203475e-01 -7.22863525e-02 4.46596652e-01
1.39254737e+00 -6.82438724e-03 -1.62423646e+00 -3.35074067e-01
-1.11584640e+00 -1.06747913e+00 3.49645048e-01 -6.69275284e-01
-1.32482016e+00 9.06966090e-01 7.44936109e-01 -2.04873547e-01
1.54333949e+00 -3.37076306e-01 8.05331588e-01 6.90770596e-02
3.82338494e-01 -9.26348627e-01 5.75258136e-02 1.96688473e-01
9.29838955e-01 -1.29333651e+00 2.25907221e-01 -3.85002464e-01
-3.68097097e-01 1.06588292e+00 6.00443304e-01 -8.57774019e-02
5.76775789e-01 2.71244347e-01 4.88704979e-01 -1.08131222e-01
-3.43740404e-01 8.90640989e-02 2.88354009e-01 5.52130938e-01
2.91519105e-01 -9.43301432e-03 2.51270831e-01 -3.63935679e-01
-2.12304249e-01 -1.66340321e-01 4.60305095e-01 4.63993043e-01
-3.81824672e-01 -6.89873040e-01 -5.94879806e-01 8.07668641e-02
-6.56533986e-02 -1.86008483e-01 2.01596543e-01 4.06778455e-01
1.52461946e-01 7.36476660e-01 1.73745472e-02 1.26145482e-01
2.54585207e-01 -4.38488305e-01 6.91679895e-01 -2.69874722e-01
9.55013037e-02 3.05677086e-01 -3.92721951e-01 -7.82352626e-01
-3.94492984e-01 -3.05558562e-01 -7.93682635e-01 -5.12088895e-01
-3.75171870e-01 -2.62354910e-01 1.10567284e+00 3.16752523e-01
1.18040964e-02 4.60160434e-01 1.15190530e+00 -1.19027698e+00
3.43821459e-02 -6.18191719e-01 -7.77745903e-01 3.35239917e-01
3.45894039e-01 -5.24340808e-01 -7.46850312e-01 3.49314868e-01] | [9.837608337402344, -2.876744508743286] |
f9507ac8-d2b9-43f8-8685-e7e4083109ab | zero-shot-temporal-relation-extraction-with | 2304.05454 | null | https://arxiv.org/abs/2304.05454v1 | https://arxiv.org/pdf/2304.05454v1.pdf | Zero-shot Temporal Relation Extraction with ChatGPT | The goal of temporal relation extraction is to infer the temporal relation between two events in the document. Supervised models are dominant in this task. In this work, we investigate ChatGPT's ability on zero-shot temporal relation extraction. We designed three different prompt techniques to break down the task and evaluate ChatGPT. Our experiments show that ChatGPT's performance has a large gap with that of supervised methods and can heavily rely on the design of prompts. We further demonstrate that ChatGPT can infer more small relation classes correctly than supervised methods. The current shortcomings of ChatGPT on temporal relation extraction are also discussed in this paper. We found that ChatGPT cannot keep consistency during temporal inference and it fails in actively long-dependency temporal inference. | ['Sophia Ananiadou', 'Qianqian Xie', 'Chenhan Yuan'] | 2023-04-11 | null | null | null | null | ['temporal-relation-extraction'] | ['natural-language-processing'] | [-5.41763008e-02 4.38038200e-01 -5.61912954e-01 -4.58903342e-01
-6.74871743e-01 -6.48117006e-01 9.54453409e-01 2.76948392e-01
-1.22349732e-01 8.05884480e-01 1.62626296e-01 -6.42645538e-01
-3.83087784e-01 -5.46819866e-01 -5.86732812e-02 -2.47594729e-01
-5.04227936e-01 6.86186731e-01 8.94492447e-01 -3.21349114e-01
1.95208251e-01 2.31122583e-01 -1.21495461e+00 7.04211891e-01
1.82518989e-01 5.60240149e-01 -2.33367205e-01 9.17910099e-01
-4.59657103e-01 1.68984830e+00 -7.49588192e-01 -1.78002700e-01
-1.90929681e-01 -6.68343782e-01 -1.67914999e+00 -4.05798048e-01
-2.90835679e-01 -1.74697101e-01 -5.27956963e-01 3.32295775e-01
2.65086982e-02 3.20285439e-01 6.56779706e-01 -1.61453629e+00
2.95538679e-02 1.06881034e+00 -4.43328947e-01 9.63330746e-01
8.12914431e-01 -9.23232436e-02 1.14205456e+00 -4.86734897e-01
1.00715446e+00 1.20061886e+00 8.90820622e-01 3.78957897e-01
-1.19941247e+00 -5.14019132e-01 3.83965135e-01 4.66401041e-01
-1.33804822e+00 -6.37915134e-01 4.30112779e-01 -4.19281006e-01
1.78655064e+00 4.20262247e-01 4.61155713e-01 1.13308561e+00
4.05646086e-01 5.42120218e-01 1.07925177e+00 -6.99449301e-01
-2.15088204e-02 -1.53638810e-01 7.53327906e-01 4.48714316e-01
-3.49232405e-01 1.80600688e-01 -1.11560810e+00 -2.58354962e-01
5.05901873e-01 -5.61619222e-01 5.39404787e-02 5.10307372e-01
-9.94811237e-01 5.34808815e-01 -3.07835728e-01 8.78832817e-01
2.03016028e-01 8.21776986e-02 5.90216041e-01 5.03579915e-01
7.31060326e-01 4.34577554e-01 -6.16853356e-01 -8.63996923e-01
-7.78048754e-01 2.77395636e-01 1.29785216e+00 1.40100408e+00
1.29571512e-01 -6.16336942e-01 -2.89570481e-01 4.32595938e-01
1.91873051e-02 -1.49353102e-01 8.06278065e-02 -1.04338098e+00
5.65902114e-01 3.92581135e-01 1.85468510e-01 -9.23263252e-01
-5.69239259e-01 3.72780085e-01 -1.59883812e-01 -2.74503320e-01
7.97422767e-01 -2.24984005e-01 -4.17555869e-01 1.53467739e+00
2.09233016e-01 -3.51664238e-02 8.43993649e-02 4.45838183e-01
8.66914630e-01 8.38386178e-01 2.57869661e-01 -1.00118661e+00
1.36774445e+00 -6.64404869e-01 -1.29393589e+00 -2.90779442e-01
1.18226039e+00 -7.80443072e-01 5.22347033e-01 1.58031926e-01
-9.80505586e-01 7.02062482e-03 -8.12643230e-01 -1.63145527e-01
-3.17523599e-01 -3.64992738e-01 1.17985666e+00 1.37428924e-01
-6.89551950e-01 7.93563128e-01 -1.10538661e+00 -8.33356678e-01
-8.35098252e-02 2.05390885e-01 -8.15518871e-02 3.21246445e-01
-1.45129228e+00 1.36536276e+00 5.67923546e-01 1.08379070e-02
-3.89991999e-01 -3.02179992e-01 -7.66250134e-01 -2.08441362e-01
8.55307102e-01 -4.55834158e-02 2.16676068e+00 -2.01634720e-01
-1.19652140e+00 6.85667872e-01 -5.75794458e-01 -5.25269628e-01
2.97370404e-01 -1.53460056e-01 -7.60867834e-01 2.38755032e-01
3.39010060e-01 1.49433598e-01 2.52772659e-01 -7.13606894e-01
-7.35096216e-01 3.67545597e-02 9.35299043e-03 -7.09667355e-02
-1.03508914e-02 8.13681304e-01 -4.64504510e-01 -3.80165726e-01
7.84210265e-02 -8.71907115e-01 -3.70362736e-02 -4.91007566e-01
-4.02630806e-01 -8.46863985e-01 1.12729371e+00 -4.11434740e-01
1.80100143e+00 -1.99884152e+00 -2.50311136e-01 -1.00817449e-01
1.54908106e-01 3.76700647e-02 2.53792703e-01 1.12529421e+00
-4.20288175e-01 3.26988846e-01 1.38411939e-01 -2.21956849e-01
-1.04033537e-01 6.95655346e-01 -4.31187034e-01 1.58983544e-01
2.58180439e-01 1.00576997e+00 -1.13381648e+00 -1.23754907e+00
3.97810787e-02 -4.45257127e-02 1.14032671e-01 2.53573477e-01
-4.88633394e-01 2.72820115e-01 -5.15062273e-01 4.20585871e-01
-1.30502522e-01 -4.47334498e-01 7.61259913e-01 1.37392685e-01
-4.68337387e-01 1.22901285e+00 -5.90762794e-01 1.51220608e+00
-2.86759902e-02 1.03022695e+00 -5.22600710e-01 -7.21263409e-01
7.16324449e-01 8.83089304e-01 5.08819222e-01 -5.78526676e-01
1.53082147e-01 -1.10391013e-01 9.53469649e-02 -8.88119996e-01
3.57051194e-01 -2.95056045e-01 -2.60178685e-01 1.03687131e+00
2.11354882e-01 -2.27115732e-02 6.10770464e-01 5.69394529e-01
1.46900260e+00 4.61156040e-01 7.36721575e-01 -2.97431588e-01
-8.78321845e-03 5.02053022e-01 6.63128734e-01 7.09699690e-01
-1.72802731e-01 1.14681110e-01 1.19406581e+00 -4.94086474e-01
-5.69694161e-01 -6.18784964e-01 7.54197268e-03 1.08268428e+00
2.98837572e-03 -1.23759651e+00 -3.14101070e-01 -1.17940605e+00
-4.97330785e-01 9.58504736e-01 -6.29445493e-01 9.09458771e-02
-9.14624512e-01 -7.34810710e-01 7.69398034e-01 7.83041000e-01
1.68242157e-01 -9.63321388e-01 -6.65414751e-01 2.75559306e-01
-7.41965413e-01 -1.52673709e+00 -2.52736270e-01 6.93266511e-01
-8.12594414e-01 -1.31684053e+00 3.27110887e-01 -6.51620686e-01
1.88530788e-01 -1.41206477e-02 1.20316112e+00 -1.42623559e-01
7.20767826e-02 2.78769106e-01 -7.93447196e-01 -4.54821467e-01
-6.04381263e-01 2.63383776e-01 -1.96955577e-01 -6.73931777e-01
8.54551733e-01 -9.05579209e-01 7.11390451e-02 5.47169685e-01
-3.66757244e-01 2.23933995e-01 -3.38795744e-02 4.01094705e-01
-1.30983457e-01 5.31352460e-01 3.57929647e-01 -1.02605820e+00
6.76308393e-01 -3.47387165e-01 -2.89012522e-01 2.74104148e-01
-8.19036067e-01 1.39871642e-01 2.14980021e-01 -6.14292383e-01
-1.32272041e+00 -2.20364869e-01 2.48698577e-01 6.27004504e-02
-1.10034674e-01 7.21596479e-01 3.31946552e-01 3.82713407e-01
7.89234102e-01 -2.36863464e-01 -3.08661342e-01 -4.01686549e-01
8.81176591e-02 2.15179831e-01 5.19563258e-01 -9.31733668e-01
5.81939340e-01 9.85339358e-02 -6.41540289e-02 -4.66792136e-01
-1.20513368e+00 -5.45107543e-01 -7.96533823e-01 -4.37416315e-01
7.26965308e-01 -4.99460936e-01 -7.65013456e-01 -2.42505083e-03
-1.59404171e+00 -8.74659300e-01 -3.83891851e-01 4.36566353e-01
-3.76836568e-01 2.67314523e-01 -1.28959060e+00 -1.17878354e+00
-2.34807823e-02 -2.80409038e-01 7.11698592e-01 -2.26759240e-02
-1.27216959e+00 -1.15451324e+00 2.95001209e-01 4.08776999e-02
-1.66551948e-01 1.54040337e-01 6.21524692e-01 -7.79708982e-01
-3.82029444e-01 -1.31906837e-01 -1.87685415e-01 -4.88186508e-01
2.47846216e-01 3.48561883e-01 -8.33082855e-01 2.60532051e-01
6.97932541e-02 -2.00052723e-01 2.48804346e-01 -1.77029036e-02
4.05463636e-01 -4.37010527e-01 -7.84363747e-01 -8.96500200e-02
9.90078807e-01 7.81025529e-01 5.48286319e-01 2.28348315e-01
3.74869078e-01 1.03532875e+00 1.23333061e+00 2.03564852e-01
3.17027539e-01 8.67314875e-01 -3.71247619e-01 3.75417411e-01
-3.30458172e-02 -2.47466251e-01 2.95119286e-01 7.73407221e-01
-3.37615460e-01 -3.93010736e-01 -1.28766143e+00 6.86651587e-01
-2.43279862e+00 -1.15631366e+00 -7.37042606e-01 1.36823499e+00
1.31446517e+00 5.59142053e-01 1.97772965e-01 4.14555699e-01
4.30211306e-01 9.75365713e-02 2.33906582e-01 -6.62849367e-01
1.33561984e-01 3.71932723e-02 5.88189699e-02 6.25378728e-01
-1.06015551e+00 1.04001534e+00 7.70350313e+00 5.43693244e-01
-5.21650314e-01 2.22287327e-01 1.92389548e-01 6.39126077e-02
2.94155180e-01 5.77341199e-01 -9.81137574e-01 2.12883934e-01
1.34427643e+00 -3.88797969e-01 -1.03953481e-02 3.29046458e-01
4.50055093e-01 -5.26837826e-01 -1.59965134e+00 6.90805674e-01
-2.21325338e-01 -1.10310018e+00 -5.75365901e-01 -2.25810394e-01
3.17359209e-01 -4.15331632e-01 -7.73704886e-01 4.58364666e-01
5.88981092e-01 -8.72878432e-01 5.16722381e-01 3.62423122e-01
4.31929022e-01 -5.27615666e-01 7.11280704e-01 4.86061245e-01
-1.49196661e+00 2.39831761e-01 5.58061123e-01 -8.15786481e-01
5.41605175e-01 5.17555475e-01 -1.13800979e+00 6.93229318e-01
8.18215191e-01 9.31647301e-01 -4.02919561e-01 4.80720669e-01
-7.20251322e-01 8.71036351e-01 -4.14504707e-01 -4.27683182e-02
-4.69202362e-02 2.03613251e-01 4.85809863e-01 1.56092608e+00
-1.09058075e-01 8.11167002e-01 2.70201892e-01 5.50552666e-01
6.43567383e-01 -2.95129985e-01 -7.12555707e-01 -3.14486980e-01
4.98291105e-01 8.21373403e-01 -1.09425557e+00 -5.93668044e-01
-3.94261211e-01 5.81920803e-01 1.79845810e-01 3.57561946e-01
-9.66824055e-01 -1.88452333e-01 1.56354442e-01 1.25927508e-01
8.03739130e-02 -4.76716161e-01 -2.78964162e-01 -1.00055587e+00
1.46781400e-01 -5.77847421e-01 9.03041720e-01 -1.06552231e+00
-1.21720469e+00 4.95968342e-01 8.34768474e-01 -8.79156291e-01
-6.84366703e-01 -1.60919219e-01 -6.62440717e-01 5.37138939e-01
-6.63475990e-01 -9.56251800e-01 9.09489319e-02 4.67740744e-01
4.29988801e-01 6.10210299e-01 9.56271172e-01 2.48451546e-01
-6.38646960e-01 1.26828641e-01 -1.10528123e+00 4.40304369e-01
8.07969987e-01 -1.29985917e+00 5.25237024e-01 1.14469814e+00
1.42917901e-01 7.99018443e-01 1.20006239e+00 -9.22852337e-01
-9.19374406e-01 -7.17536867e-01 1.88257349e+00 -8.16688299e-01
1.17698681e+00 -2.15689927e-01 -9.58263099e-01 1.36434400e+00
5.16933501e-01 -1.60809696e-01 5.70180297e-01 7.48929262e-01
-5.54959238e-01 1.61962345e-01 -4.75104451e-01 6.51113272e-01
1.25221431e+00 -9.29981828e-01 -1.13744640e+00 4.79113758e-01
8.46918344e-01 -5.24283707e-01 -1.14970660e+00 3.38328660e-01
3.50924760e-01 -7.45758355e-01 5.79214752e-01 -7.23739803e-01
3.43293369e-01 -2.72976190e-01 2.83450305e-01 -6.09786749e-01
-9.83822495e-02 -1.21243536e+00 -5.17889977e-01 1.69028139e+00
6.23779297e-01 -5.34860015e-01 5.83514214e-01 9.60627794e-01
2.79373359e-02 -5.02304852e-01 -7.17347443e-01 -1.10931265e+00
-3.96755427e-01 -8.07510436e-01 1.41647356e-02 1.39001167e+00
1.09333372e+00 1.09517395e+00 -2.99941361e-01 -4.09419537e-02
2.23228559e-01 1.90520287e-01 5.18632054e-01 -1.29769373e+00
-3.03361207e-01 -5.71745113e-02 -1.75750572e-02 -7.94928551e-01
5.04924133e-02 -3.96265954e-01 2.92510390e-01 -1.52459335e+00
1.68980896e-01 -2.88061410e-01 2.38490760e-01 1.13752496e+00
1.81031287e-01 -2.16153741e-01 -2.42456868e-01 2.86840349e-01
-8.32291543e-01 -1.20590925e-01 9.68511641e-01 -6.69233128e-02
-5.07218063e-01 -5.09569123e-02 -3.93860757e-01 5.72675347e-01
9.23080564e-01 -8.59831393e-01 -7.58101165e-01 -3.58897299e-02
4.12569642e-01 7.05323815e-01 2.22812429e-01 -6.10677540e-01
6.19016588e-01 -4.74315614e-01 -2.14708880e-01 -7.38327861e-01
1.50349513e-01 -5.75063407e-01 3.23221385e-01 1.99563280e-02
-5.08479536e-01 9.02866647e-02 4.62112427e-01 3.91338855e-01
-3.19364786e-01 2.75364164e-02 3.00223142e-01 -8.40683877e-02
-7.26890266e-01 -1.05413549e-01 -9.42156792e-01 1.94609120e-01
9.86491919e-01 1.58744082e-02 -5.02674341e-01 -3.71614188e-01
-1.07157516e+00 2.43471652e-01 1.86138116e-02 3.28924716e-01
2.46047586e-01 -1.02369344e+00 -2.22026750e-01 -6.11350298e-01
1.00996017e-01 -1.03068948e-02 -3.14572453e-01 1.31258154e+00
-1.33063450e-01 6.53002024e-01 3.35469961e-01 -5.48057675e-01
-1.71362793e+00 8.26847017e-01 7.46856555e-02 -6.07219696e-01
-1.02930415e+00 6.25388741e-01 -3.04955512e-01 2.03629524e-01
1.99364871e-01 -5.23902237e-01 -2.99061030e-01 3.36438477e-01
5.26845932e-01 -1.91482399e-02 3.70034911e-02 -1.03250332e-01
-7.15417624e-01 1.74472034e-01 -3.18490803e-01 -6.77467167e-01
1.23435330e+00 -1.88366726e-01 -4.23455715e-01 1.17588508e+00
9.03226614e-01 -2.00782299e-01 -6.92961931e-01 -2.21806169e-01
7.96965182e-01 -1.22710370e-01 -2.65246183e-01 -6.61135793e-01
-2.12344810e-01 1.72937602e-01 -2.81215817e-01 8.68298948e-01
9.71776962e-01 4.16205287e-01 7.23210037e-01 3.68803173e-01
7.19494283e-01 -1.06184006e+00 -2.84299143e-02 6.76984012e-01
5.31408191e-01 -1.00239432e+00 3.80005032e-01 -1.04106581e+00
-3.99699062e-01 1.14227688e+00 7.43641496e-01 4.47338909e-01
5.78172207e-01 8.74815524e-01 4.36485559e-02 -7.12772310e-01
-1.52197468e+00 -2.64019728e-01 -2.14134413e-03 4.04844671e-01
8.24366450e-01 -1.63641006e-01 -6.20297194e-01 3.86309892e-01
-1.33713365e-01 2.00367302e-01 3.93264383e-01 1.64068937e+00
-6.80939015e-03 -1.31996191e+00 -3.57913822e-01 7.29537532e-02
-4.25611079e-01 -7.11437315e-02 -8.52897584e-01 1.13074887e+00
-1.10721320e-01 1.54131997e+00 -8.19164291e-02 -6.42498672e-01
1.91682428e-01 4.13771749e-01 6.69218123e-01 -8.87896538e-01
-5.88780940e-01 2.10246012e-01 1.20763338e+00 -8.23696971e-01
-8.16589892e-01 -1.11869860e+00 -1.30245125e+00 -4.03453618e-01
-2.56943047e-01 7.70916462e-01 -1.00542955e-01 1.42521489e+00
-1.49074525e-01 6.87335193e-01 1.39975816e-01 -2.58999735e-01
2.08933428e-01 -1.01037574e+00 -1.89791620e-01 -3.77130136e-02
8.56894329e-02 -6.26657784e-01 -4.15772766e-01 3.91317934e-01] | [9.05298900604248, 9.229907035827637] |
7ba2647b-36ef-44b9-bfbd-e4ae5efd9a19 | state-of-the-art-of-audio-and-video-based | 2207.01487 | null | https://arxiv.org/abs/2207.01487v2 | https://arxiv.org/pdf/2207.01487v2.pdf | State of the Art of Audio- and Video-Based Solutions for AAL | The report illustrates the state of the art of the most successful AAL applications and functions based on audio and video data, namely (i) lifelogging and self-monitoring, (ii) remote monitoring of vital signs, (iii) emotional state recognition, (iv) food intake monitoring, activity and behaviour recognition, (v) activity and personal assistance, (vi) gesture recognition, (vii) fall detection and prevention, (viii) mobility assessment and frailty recognition, and (ix) cognitive and motor rehabilitation. For these application scenarios, the report illustrates the state of play in terms of scientific advances, available products and research project. The open challenges are also highlighted. | ['Andrej Zgank', 'Hilda Tellioglu', 'Lacramioara Stoicu-Tivadar', 'Anna Sigridur Islind', 'Maria Jose Santofimia', 'Albert Ali Salah', 'Susanna Spinsante', 'Mara Pudane', 'Angelica Poli', 'Peter Pocta', 'Sintija Petrovica', 'Galidiya Petrova', 'Rodrigo Rodriguez Peerez', 'Zada Pajalic', 'Sophie Noiret', 'Wiktor Mucha', 'Jennifer Lumetzberger', 'Lambros Lambrinos', 'Andrzej Klimczuk', 'William Kearns', 'Martin Kampel', 'Mladjan Jovanovic', 'Ivo Iliev', 'Murat Emirzeoglu', 'Steinunn Gróa Sigurðardóttir', 'Nicole Grech', 'Danila Germanese', 'Francisco Florez-Revuelta', 'Ekrem Erakin', 'Hazim Kemal Ekenel', 'Vladimir Despotovic', 'Stefania Cristina', 'Sara Colantonio', 'Pau Climent-Peerez', 'Anto Cartolovni', 'Kenneth Camilleri', 'Jean Calleja Agius', 'Michael Atanasov', 'Slavisa Aleksic'] | 2022-06-26 | null | null | null | null | ['gesture-recognition'] | ['computer-vision'] | [ 1.66280791e-01 -2.43873104e-01 -3.83884877e-01 -8.19737241e-02
-4.91592199e-01 -2.51205117e-01 1.95781365e-01 4.15913492e-01
-5.94410837e-01 1.15299737e+00 6.25280201e-01 -3.19250166e-01
-4.84578907e-01 -3.80544037e-01 -6.26721308e-02 -6.80228055e-01
-6.90309167e-01 1.43841967e-01 2.04545617e-01 -2.20755383e-01
1.95423722e-01 6.72690332e-01 -2.35888958e+00 1.42070070e-01
7.99101651e-01 1.38879752e+00 -3.00069988e-01 8.27393472e-01
1.20123528e-01 8.20157647e-01 -2.58343846e-01 2.83821225e-01
-6.04207456e-01 -2.73854852e-01 -6.90506816e-01 -1.11751929e-01
-3.06546122e-01 -6.74353778e-01 -2.24609315e-01 5.64425111e-01
6.92445040e-01 3.72548878e-01 3.89124334e-01 -1.36180651e+00
-1.93411976e-01 -1.65357769e-01 -4.33668755e-02 3.42290819e-01
1.18055296e+00 3.38614970e-01 -6.98550716e-02 -6.87034726e-01
2.62222379e-01 8.96796525e-01 4.82304186e-01 6.75388813e-01
-5.66863239e-01 -4.77301776e-01 5.08238710e-02 9.07809377e-01
-1.13810682e+00 -5.92823386e-01 2.41149709e-01 -4.71354336e-01
1.28058636e+00 7.10279465e-01 1.15960944e+00 1.14349389e+00
1.64672166e-01 6.08484328e-01 1.01071632e+00 -4.92487818e-01
6.14466965e-01 4.03225534e-02 5.41694045e-01 5.18715322e-01
3.24323803e-01 3.30041587e-01 -1.04817116e+00 -2.55117863e-01
3.87262523e-01 -3.21990736e-02 -2.40967408e-01 1.06194347e-01
-9.12890553e-01 3.49566281e-01 -3.00968945e-01 3.82881731e-01
-7.52621114e-01 -1.96311086e-01 5.56618869e-01 4.94803011e-01
4.49036732e-02 -4.05699939e-01 -1.56848878e-01 -8.97676528e-01
-8.16748798e-01 1.37889326e-01 6.73723221e-01 4.76122856e-01
1.16061501e-01 9.90711823e-02 -4.52518433e-01 5.45601547e-01
9.27125752e-01 8.34888399e-01 9.99184787e-01 -1.04434812e+00
1.87956661e-01 7.57148266e-01 4.18544263e-01 -4.05857086e-01
-8.52031231e-01 3.40946987e-02 -2.02181742e-01 8.98834348e-01
2.46596515e-01 -1.86828822e-01 -6.37997806e-01 1.25142360e+00
4.12454039e-01 -9.00807381e-02 -3.58934253e-02 6.84541464e-01
1.17744958e+00 -4.90247011e-02 4.66635436e-01 -6.96657360e-01
1.63439608e+00 -4.05617565e-01 -9.96091068e-01 -3.30174655e-01
5.26665628e-01 -3.60883415e-01 7.92294800e-01 5.90545475e-01
-1.05934608e+00 -1.59013227e-01 -1.03211904e+00 3.38400245e-01
-4.01106596e-01 2.15909407e-01 7.10532725e-01 1.42603421e+00
-9.65427220e-01 5.00492811e-01 -1.31994522e+00 -1.00056970e+00
3.04577768e-01 7.36423552e-01 -4.07808810e-01 4.99891229e-02
-1.03161478e+00 8.39718044e-01 2.58279275e-02 -2.15658844e-02
-4.70175862e-01 -3.01901013e-01 -6.96873069e-01 -3.38975042e-01
-1.80396870e-01 -4.88199979e-01 7.30335414e-01 -4.77961659e-01
-1.83735013e+00 1.26917231e+00 1.61269086e-03 -3.10215056e-01
6.45689607e-01 -7.66956449e-01 -8.92173350e-01 4.45184261e-01
-4.91356142e-02 -1.10805184e-02 1.47550181e-01 -2.36667827e-01
-9.08856452e-01 -9.34973359e-01 -3.64374369e-01 3.36862952e-01
-1.78178146e-01 6.28963828e-01 2.36143082e-01 -1.58675104e-01
-1.66661814e-02 -3.65439147e-01 3.33346754e-01 4.28587973e-01
-4.10518721e-02 -1.24012299e-01 8.53824258e-01 -1.17109275e+00
1.49970198e+00 -1.90999973e+00 -3.03310066e-01 1.04322128e-01
-8.67501423e-02 7.53648221e-01 6.75256729e-01 6.35534823e-01
-7.81412795e-02 -4.31552589e-01 3.99693221e-01 2.29060680e-01
6.85722977e-02 2.94958919e-01 5.02310693e-01 7.70555139e-01
-4.67126101e-01 7.42448866e-01 -7.40452528e-01 -2.93435097e-01
8.50901723e-01 5.87867558e-01 -1.11925200e-01 6.45728456e-03
6.06880367e-01 5.62953234e-01 -4.96624768e-01 1.29874885e+00
4.54882942e-02 4.68004972e-01 -1.29015937e-01 -1.14732064e-01
-4.56958771e-01 4.05874342e-01 -1.42802656e+00 1.33444428e+00
1.78817689e-01 3.40262383e-01 2.98762769e-01 -1.02029955e+00
9.79587376e-01 9.90497231e-01 8.44522953e-01 -8.33738804e-01
3.71647149e-01 3.42685282e-01 -4.44222718e-01 -1.62104881e+00
-4.79156703e-01 -1.85217232e-01 4.99203444e-01 4.75415230e-01
-7.92837739e-02 1.15759015e+00 4.47928198e-02 -3.20796460e-01
1.28852236e+00 2.88151354e-01 8.34176540e-01 3.62532027e-03
7.70893455e-01 -4.54053134e-01 1.03053726e-01 2.47258663e-01
-1.06479788e+00 5.85854240e-02 -1.04404837e-01 -1.33674353e-01
-2.78996397e-02 -1.16377544e+00 2.81721950e-02 1.14790869e+00
-7.51066878e-02 2.61774138e-02 -7.78000236e-01 -8.38136226e-02
1.58383965e-01 5.85400879e-01 -4.71726894e-01 -5.90250194e-01
-3.12089115e-01 -5.55330336e-01 4.82115537e-01 8.74208510e-01
5.47009110e-01 -1.47585344e+00 -1.29753447e+00 3.83346289e-01
-4.49662596e-01 -4.13242340e-01 1.29785016e-01 4.97234017e-01
-1.27154613e+00 -1.00608230e+00 -6.28222346e-01 -7.50039399e-01
9.53023583e-02 -1.30008310e-01 5.52421808e-01 4.20953520e-02
-4.63186026e-01 1.22933888e+00 -2.97610909e-01 -2.64349878e-01
5.48549555e-02 -5.66765785e-01 5.43722928e-01 3.85347492e-04
5.90196669e-01 -1.29172289e+00 -9.99924362e-01 4.31172192e-01
-3.67724210e-01 -6.73711777e-01 5.05854607e-01 1.50202245e-01
3.79250884e-01 -5.69714427e-01 5.53916693e-01 -1.51075646e-01
6.22781098e-01 -4.01752681e-01 1.29554316e-01 4.63706821e-01
-8.17673266e-01 -4.03749496e-01 -4.22898561e-01 -4.89385843e-01
-9.40798581e-01 -9.56282020e-03 -5.27400911e-01 2.32000247e-01
-9.01824653e-01 3.92942667e-01 -3.93289775e-01 -5.83237074e-02
9.60268080e-01 2.97216713e-01 3.88808757e-01 -7.09813535e-01
-1.15434892e-01 1.21123576e+00 1.30472291e+00 -2.88172960e-01
6.69175163e-02 5.21317184e-01 -1.56463414e-01 -1.07782984e+00
-1.28578544e-01 -7.66127765e-01 -7.58870006e-01 -1.05630875e+00
9.23213482e-01 -7.42920935e-01 -1.23006296e+00 6.41641796e-01
-4.99912083e-01 -2.26904273e-01 -4.38202083e-01 9.73342776e-01
-6.89511478e-01 3.59015703e-01 -4.95233625e-01 -1.62448156e+00
-8.39941442e-01 -2.67392844e-01 7.13263333e-01 4.92080897e-01
-9.79180574e-01 -7.69326389e-01 1.35618508e-01 6.16409302e-01
4.01453793e-01 9.52754796e-01 6.27817452e-01 -1.15145542e-01
1.71901062e-01 -8.48132312e-01 2.22230539e-01 5.38665764e-02
4.34695750e-01 -2.89253026e-01 -7.50383019e-01 -5.99735081e-02
5.68426177e-02 -4.33106631e-01 4.93883900e-03 6.30964875e-01
2.49390811e-01 3.58527154e-02 -5.10047853e-01 5.20334959e-01
8.46908033e-01 8.19652438e-01 1.43189967e+00 5.01416802e-01
9.50983837e-02 4.93923008e-01 8.54695857e-01 6.25700474e-01
1.95281617e-02 5.70877314e-01 4.81057703e-01 1.41649827e-01
-2.74595141e-01 2.15815082e-01 8.55937660e-01 2.31746584e-01
-9.18683290e-01 2.09969744e-01 -5.88595629e-01 2.12668344e-01
-2.00746608e+00 -1.25103009e+00 -4.18153763e-01 2.40648937e+00
2.56982625e-01 -3.17587378e-03 8.82948756e-01 8.91187072e-01
6.03026152e-01 -6.29096508e-01 -8.23791623e-01 -2.72099048e-01
9.34481248e-02 3.09249282e-01 2.29344606e-01 1.36659026e-01
-9.53873098e-01 1.93886742e-01 7.48462009e+00 1.31200299e-01
-1.02443719e+00 4.53883350e-01 -2.16873642e-02 -3.50418597e-01
4.85298753e-01 -5.78417182e-01 -4.37312812e-01 5.76680362e-01
1.05293417e+00 2.02365264e-01 3.86192381e-01 6.17257476e-01
5.25803566e-01 -5.25527179e-01 -6.80335999e-01 1.07388234e+00
9.06306058e-02 -6.17246866e-01 -5.09522617e-01 1.09747322e-02
-3.31133097e-01 -1.06852941e-01 -5.37393928e-01 9.53889638e-02
-4.32934314e-01 -7.67778456e-01 8.15655351e-01 9.95746791e-01
9.61710751e-01 -6.92268610e-01 4.75346774e-01 5.37934490e-02
-1.02974343e+00 -3.85022372e-01 3.88955176e-01 -4.71847981e-01
5.41538060e-01 1.72699183e-01 1.86709240e-01 9.96932462e-02
1.12054110e+00 4.22459006e-01 -3.55486386e-02 1.39191329e+00
-5.15481591e-01 8.05933297e-01 -6.11873269e-01 -2.27887332e-01
-2.89161116e-01 -7.53554925e-02 6.94275260e-01 9.93578374e-01
4.73641992e-01 4.87790048e-01 -2.40810722e-01 4.10156786e-01
9.57973957e-01 4.33465652e-02 -2.38514170e-01 1.27146210e-04
-7.98141211e-02 7.55597413e-01 -9.57085431e-01 9.67972949e-02
-5.42702138e-01 8.96834254e-01 -4.18951690e-01 3.77122194e-01
-4.29592013e-01 -5.04974604e-01 1.18526530e+00 5.51725805e-01
-1.08590126e-01 -6.39094338e-02 -1.32491335e-01 -5.64742684e-01
4.42805797e-01 -3.36173147e-01 9.86469984e-01 -7.70879865e-01
-5.09868145e-01 -2.37675514e-02 -6.78509101e-02 -1.44344270e+00
-5.02901375e-01 -5.82899451e-01 -7.13500679e-01 5.96653759e-01
-7.51937509e-01 -6.96219742e-01 -5.51677048e-01 9.46273267e-01
1.44557744e-01 -2.20926300e-01 1.13196158e+00 8.19958329e-01
-5.06719053e-01 4.19532508e-01 -5.41191734e-02 -4.28188026e-01
2.64535427e-01 -7.67802477e-01 -5.74176848e-01 3.02636772e-01
-7.52129972e-01 1.24245547e-01 7.17962682e-01 -9.00045455e-01
-1.35375941e+00 -6.17585361e-01 9.10680473e-01 -2.63127297e-01
1.96977049e-01 1.20692350e-01 -6.56156003e-01 4.18408811e-01
-5.84283292e-01 -2.88972110e-01 9.80734587e-01 4.98827174e-02
5.12073815e-01 -8.47495571e-02 -1.56670761e+00 2.56142884e-01
1.19072211e+00 -3.57061535e-01 -3.51925611e-01 5.49155831e-01
-2.59136140e-01 -4.17586237e-01 -8.35973859e-01 2.67472435e-02
1.44601393e+00 -9.48413134e-01 1.29329014e+00 -8.49106312e-01
-7.19140247e-02 -1.27712160e-01 -1.41872421e-01 -4.08957064e-01
-4.67603058e-01 -5.49143493e-01 -7.09923387e-01 1.41574872e+00
2.30210065e-03 -7.37365961e-01 5.20410657e-01 1.17959332e+00
-1.27364025e-01 -5.39196312e-01 -1.30144846e+00 -8.00477505e-01
-6.85215175e-01 -5.78145444e-01 4.48293984e-02 3.02719593e-01
9.59445238e-01 -9.72215012e-02 -4.26833361e-01 -1.38670191e-01
2.19005406e-01 -4.30961281e-01 3.17230299e-02 -1.48064017e+00
-1.14646114e-01 -3.07340801e-01 -1.12108397e+00 -1.97010502e-01
-6.64915085e-01 -4.62652504e-01 -3.99684042e-01 -2.10237861e+00
-2.39548892e-01 4.33201849e-01 -4.89478409e-01 5.71457088e-01
1.58960477e-01 1.99260816e-01 -4.68981534e-01 -7.59555325e-02
-7.09833324e-01 1.27943233e-01 4.88096923e-01 2.10928574e-01
-5.45745373e-01 5.01047969e-01 -7.25534499e-01 5.33769608e-01
5.73110938e-01 -2.31425971e-01 -3.14942062e-01 4.20204997e-01
2.33829916e-01 2.27921635e-01 5.62942922e-01 -1.45923960e+00
1.55891865e-01 -1.71633899e-01 6.08351111e-01 -5.97491145e-01
4.72361177e-01 -6.94817245e-01 2.30599001e-01 1.26310766e+00
2.37648010e-01 -3.12923312e-01 2.52895849e-03 3.80365610e-01
1.96646541e-01 -2.72642374e-01 6.27437294e-01 1.11447282e-01
-7.85964131e-01 -3.04131538e-01 -1.25671363e+00 -4.15106952e-01
1.09391272e+00 -9.60951209e-01 -3.15381467e-01 -6.35901272e-01
-1.76854479e+00 2.81201363e-01 -1.08248115e-01 2.65080303e-01
5.92390537e-01 -1.55334377e+00 -3.28719914e-01 3.11348915e-01
4.82560247e-02 -8.46642375e-01 4.32838112e-01 1.57585835e+00
-3.44831347e-01 2.74895757e-01 -6.71818376e-01 -3.60578686e-01
-1.98983479e+00 2.78779566e-01 3.27997833e-01 -3.92821897e-03
-9.58377361e-01 2.29131013e-01 -6.49031043e-01 4.08750921e-01
9.48555231e-01 -1.12258539e-01 -8.17223728e-01 1.56792432e-01
1.28396559e+00 1.32523179e+00 3.30480397e-01 -5.89232743e-01
-8.68337452e-01 4.84918416e-01 4.66773421e-01 -3.21089894e-01
1.27282226e+00 -6.46903336e-01 2.81378269e-01 6.74605966e-01
5.18366098e-01 -8.20121527e-01 -8.66806805e-01 9.55708325e-02
6.97779208e-02 -1.95356049e-02 1.28974125e-01 -1.31565964e+00
-8.49693418e-01 7.34712839e-01 1.69944382e+00 -3.83249559e-02
1.45558834e+00 -1.20865077e-01 8.18369031e-01 1.59490496e-01
5.53652763e-01 -1.56788862e+00 -2.72697091e-01 -2.59515680e-02
8.33167136e-01 -6.60943687e-01 3.74941193e-02 -2.42564708e-01
-4.31384861e-01 1.09046388e+00 -2.90352944e-02 1.35500059e-01
1.20404029e+00 4.29683894e-01 3.80764127e-01 -3.54836941e-01
-4.91564095e-01 -5.40279567e-01 2.21085519e-01 1.40464795e+00
6.27671599e-01 3.36028427e-01 -7.44238019e-01 1.20679748e+00
-2.70460919e-02 7.24872768e-01 -2.11960256e-01 1.48791432e+00
-7.54923701e-01 -5.53834438e-01 -8.46591473e-01 8.47968519e-01
-4.24387544e-01 7.33135939e-01 -3.83609802e-01 5.42265177e-01
1.85761556e-01 1.29343033e+00 -5.32539368e-01 -2.68104047e-01
7.47912169e-01 7.78405070e-01 5.83543956e-01 -1.34028494e-01
-6.35053098e-01 -1.37430290e-02 5.49758792e-01 -1.01132464e+00
-5.18612087e-01 -1.28774631e+00 -1.45961010e+00 2.34210920e-02
3.90269279e-01 -2.92618513e-01 8.24449837e-01 1.43856740e+00
7.61252344e-02 5.87305129e-01 -1.54637516e-01 -8.37954998e-01
1.39319628e-01 -9.94698346e-01 -1.29637218e+00 2.21813411e-01
9.90111604e-02 -1.04266644e+00 -1.67218909e-01 7.05975741e-02] | [7.15816593170166, 0.47466179728507996] |
124e0f1a-f131-45ee-9b34-e68b50c6d753 | lpat-learning-to-predict-adaptive-threshold | 1910.11285 | null | https://arxiv.org/abs/1910.11285v3 | https://arxiv.org/pdf/1910.11285v3.pdf | Towards Train-Test Consistency for Semi-supervised Temporal Action Localization | Recently, Weakly-supervised Temporal Action Localization (WTAL) has been densely studied but there is still a large gap between weakly-supervised models and fully-supervised models. It is practical and intuitive to annotate temporal boundaries of a few examples and utilize them to help WTAL models better detect actions. However, the train-test discrepancy of action localization strategy prevents WTAL models from leveraging semi-supervision for further improvement. At training time, attention or multiple instance learning is used to aggregate predictions of each snippet for video-level classification; at test time, they first obtain action score sequences over time, then truncate segments of scores higher than a fixed threshold, and post-process action segments. The inconsistent strategy makes it hard to explicitly supervise the action localization model with temporal boundary annotations at training time. In this paper, we propose a Train-Test Consistent framework, TTC-Loc. In both training and testing time, our TTC-Loc localizes actions by comparing scores of action classes and predicted threshold, which enables it to be trained with semi-supervision. By fixing the train-test discrepancy, our TTC-Loc significantly outperforms the state-of-the-art performance on THUMOS'14, ActivityNet 1.2 and 1.3 when only video-level labels are provided for training. With full annotations of only one video per class and video-level labels for the other videos, our TTC-Loc further boosts the performance and achieves 33.4\% mAP (IoU threshold 0.5) on THUMOS's 14. | ['Shih-Fu Chang', 'Xudong Lin', 'Zheng Shou'] | 2019-10-24 | null | null | null | null | ['weakly-supervised-temporal-action'] | ['computer-vision'] | [ 4.15259272e-01 -2.25992817e-02 -1.01281428e+00 -4.50161815e-01
-9.51075077e-01 -4.55393851e-01 3.40739816e-01 -2.14738891e-01
-4.95801806e-01 5.95669270e-01 2.31793091e-01 -7.41630271e-02
2.81602740e-01 -1.80798769e-01 -8.15827191e-01 -6.17459357e-01
-3.64292055e-01 1.67119011e-01 9.76744652e-01 3.79319310e-01
-7.51604736e-02 -9.52065960e-02 -1.34495437e+00 9.41181898e-01
6.89976752e-01 1.37479460e+00 2.56149352e-01 5.90931535e-01
1.71463102e-01 1.29648697e+00 -4.13521707e-01 9.10848007e-03
2.80057400e-01 -7.08071411e-01 -8.95837128e-01 3.31472754e-01
5.93997061e-01 -4.70882326e-01 -3.06806296e-01 8.02420735e-01
1.69656888e-01 4.93918173e-02 4.79821473e-01 -1.49555147e+00
-2.35563964e-01 5.65097213e-01 -6.20963395e-01 3.87985498e-01
3.40535730e-01 3.80208790e-01 1.13612282e+00 -8.40896428e-01
5.47996402e-01 9.26559448e-01 6.66300833e-01 6.60257101e-01
-1.01453876e+00 -5.91974974e-01 6.17450297e-01 4.60431874e-01
-1.14329576e+00 -2.65737593e-01 4.83833402e-01 -4.79824871e-01
1.04278982e+00 2.70154746e-03 6.56663477e-01 1.28509855e+00
1.87432524e-02 1.39751923e+00 8.45714211e-01 -1.28427923e-01
3.71748835e-01 -2.99765497e-01 -1.08291581e-01 7.32964218e-01
-3.26094508e-01 -1.25127718e-01 -6.84412241e-01 2.19348148e-01
8.33996058e-01 1.61932245e-01 -9.63490829e-02 -3.62493724e-01
-1.41064358e+00 3.70029390e-01 3.92514169e-01 4.65957820e-01
-3.10542375e-01 5.06740868e-01 6.68472648e-01 1.59576416e-01
5.08740664e-01 3.14180665e-02 -7.43735313e-01 -6.24154329e-01
-1.12716544e+00 -1.79737046e-01 2.78733701e-01 9.34072435e-01
5.49549520e-01 -2.61786301e-02 -4.70132291e-01 8.52363884e-01
1.48122255e-02 1.21140763e-01 5.21993041e-01 -1.22556520e+00
7.41811752e-01 7.17198610e-01 6.19575232e-02 -3.25875133e-01
-1.02209941e-01 -3.16621602e-01 -4.15254056e-01 8.70206878e-02
5.29220283e-01 -4.09367159e-02 -1.14245665e+00 1.70776176e+00
1.28742218e-01 7.51893401e-01 -1.35336399e-01 8.92976284e-01
2.87146628e-01 4.80091423e-01 3.65072608e-01 -3.41447204e-01
1.07808101e+00 -1.51839018e+00 -5.25291741e-01 -4.12791014e-01
1.24625897e+00 -3.04136336e-01 1.24270797e+00 3.59085828e-01
-9.74569440e-01 -8.15898180e-01 -9.14331019e-01 2.42078468e-01
5.22314943e-02 5.00357568e-01 5.13897061e-01 1.41628712e-01
-8.70554209e-01 7.47871280e-01 -1.43428671e+00 -3.48260105e-01
8.45946968e-01 1.88983381e-01 -4.35580552e-01 -1.61212847e-01
-9.82960463e-01 5.93876123e-01 4.71813709e-01 -3.53994071e-02
-1.30590439e+00 -5.43974996e-01 -9.08543408e-01 -1.25308961e-01
6.71316385e-01 -8.69251862e-02 1.47437036e+00 -1.36765218e+00
-1.20420516e+00 6.69530988e-01 -2.66227126e-01 -7.71981120e-01
6.98423028e-01 -4.69346136e-01 -5.02822161e-01 3.49173784e-01
5.48907459e-01 9.95058715e-01 7.75879443e-01 -7.43386626e-01
-1.00584257e+00 -4.12425585e-02 1.94217101e-01 1.89086512e-01
-2.93696195e-01 -2.71482710e-02 -9.00371790e-01 -5.32634497e-01
1.03550948e-01 -9.58759427e-01 -1.27075210e-01 2.01267377e-01
-1.39108971e-01 -5.12792110e-01 9.52132821e-01 -4.69803423e-01
1.40057564e+00 -2.41539502e+00 -1.98816463e-01 -2.04966635e-01
-1.51039258e-01 2.86849976e-01 -3.23314935e-01 9.87888947e-02
-2.51977563e-01 2.94226152e-03 -2.23416463e-01 -5.31543136e-01
-1.87223554e-01 4.36398059e-01 -1.64679661e-02 4.18309242e-01
3.96165073e-01 9.10374403e-01 -1.08651304e+00 -7.74739683e-01
2.45472431e-01 7.48159587e-02 -7.83038914e-01 4.54108380e-02
-4.48029041e-01 5.60926676e-01 -5.12645304e-01 9.35497880e-01
8.66362974e-02 -3.87840539e-01 1.21908568e-01 -2.48092160e-01
-3.97944003e-02 3.79113078e-01 -1.06275070e+00 1.93646085e+00
-2.65610904e-01 6.32649124e-01 -4.72989500e-01 -1.17107832e+00
4.29886281e-01 4.93108213e-01 8.88123572e-01 -6.90585077e-01
-2.23542392e-01 1.37052104e-01 8.67270455e-02 -5.95192075e-01
-1.59034327e-01 1.49740264e-01 -8.68525431e-02 3.01149547e-01
2.33720496e-01 4.64698315e-01 5.61816275e-01 1.55408412e-01
1.56679821e+00 8.70884717e-01 1.04100712e-01 1.50155127e-01
5.44368804e-01 9.16954875e-03 1.02653313e+00 6.47789180e-01
-6.33220375e-01 6.22736812e-01 6.78003669e-01 -4.76647466e-01
-7.33746171e-01 -8.94357920e-01 -4.02956232e-02 1.59534895e+00
5.72072715e-02 -6.44977808e-01 -7.74968445e-01 -1.42401528e+00
-1.54066414e-01 5.00941575e-01 -7.07560956e-01 -1.82216078e-01
-6.52526140e-01 -2.73269385e-01 4.63493973e-01 1.11081088e+00
7.91537464e-01 -1.21666694e+00 -5.67927361e-01 3.58656734e-01
-3.50636512e-01 -1.37394190e+00 -7.77170718e-01 3.71584862e-01
-1.03685892e+00 -1.11353469e+00 -6.45054996e-01 -7.53548324e-01
7.09160268e-01 1.86441630e-01 9.26817238e-01 4.90552420e-03
3.20239365e-02 2.78047979e-01 -6.81262076e-01 5.84457070e-02
-1.57524928e-01 -3.99174951e-02 3.53099555e-02 1.53368920e-01
4.21939611e-01 -5.46185076e-01 -7.79695392e-01 8.59624624e-01
-4.97118056e-01 9.06063765e-02 6.77587450e-01 6.44638956e-01
8.17176163e-01 -9.90462825e-02 4.83972192e-01 -6.30676270e-01
-2.89870858e-01 -2.91583836e-01 -2.72860020e-01 2.73017645e-01
-4.29680198e-01 -1.44688979e-01 5.03320932e-01 -7.81915069e-01
-8.27698290e-01 3.62688482e-01 1.00266963e-01 -8.13203156e-01
-2.09289461e-01 3.72815132e-01 -1.06226735e-01 3.60351443e-01
6.45782113e-01 1.46920636e-01 -2.45422885e-01 -4.64629799e-01
-4.35302481e-02 4.18161660e-01 5.95954359e-01 -3.82598341e-01
4.31568772e-01 3.88338834e-01 -4.21542853e-01 -2.80500621e-01
-1.34994805e+00 -6.58679962e-01 -9.11193311e-01 -4.93260056e-01
1.16764307e+00 -1.17228675e+00 -3.63479376e-01 3.58324617e-01
-7.94026136e-01 -9.52832520e-01 -3.53089839e-01 8.11065495e-01
-6.86370432e-01 1.77740499e-01 -6.67697668e-01 -6.16143584e-01
1.48653165e-01 -1.10619760e+00 1.25364530e+00 -1.26216024e-01
-3.29078287e-01 -8.22107553e-01 -5.02099581e-02 6.74812317e-01
-7.63340592e-02 1.06608152e-01 3.61054033e-01 -6.94244385e-01
-5.06352067e-01 -3.32992464e-01 -5.18269353e-02 7.64396667e-01
1.88463524e-01 -1.84278727e-01 -9.17696476e-01 -3.29630584e-01
-3.43543202e-01 -7.32075632e-01 1.01611936e+00 3.87940943e-01
1.56302023e+00 -1.05433404e-01 -3.69366586e-01 3.80272061e-01
9.49680686e-01 3.17852646e-01 6.82255507e-01 2.52276361e-01
6.95785284e-01 2.31973454e-01 1.12196386e+00 3.08016330e-01
2.11460158e-01 7.93992221e-01 4.30890769e-01 -6.03029467e-02
-1.52743280e-01 -4.17668194e-01 1.00221741e+00 4.74535674e-01
-3.45232546e-01 -1.13079034e-01 -6.88707709e-01 5.67082644e-01
-2.09846878e+00 -1.16628218e+00 1.48003479e-03 2.12487602e+00
9.66661215e-01 6.46652877e-01 3.11568558e-01 1.41847640e-01
6.52100623e-01 3.52465063e-01 -7.49983191e-01 5.09804375e-02
2.01967642e-01 -5.22600040e-02 4.83632028e-01 2.65788198e-01
-1.50592983e+00 1.08289135e+00 5.50039959e+00 9.63389635e-01
-9.03787613e-01 2.81956106e-01 7.36171365e-01 -4.10833567e-01
3.96479517e-01 1.31341442e-01 -7.93854535e-01 6.47839427e-01
8.96633565e-01 4.00027186e-01 3.41193862e-02 1.08187711e+00
5.64765751e-01 -4.26069945e-01 -1.56934464e+00 8.74311447e-01
6.45257309e-02 -1.04877710e+00 -2.55425096e-01 -6.16385266e-02
8.63850355e-01 1.96080878e-01 -2.08935231e-01 7.89858162e-01
1.22402035e-01 -7.63924778e-01 8.24012935e-01 1.55518979e-01
7.94476926e-01 -3.50105852e-01 6.86507881e-01 4.73009467e-01
-1.51867926e+00 -3.13323706e-01 -1.63472846e-01 -1.57012016e-01
2.56112486e-01 2.28457466e-01 -5.72262824e-01 2.71002889e-01
7.66341329e-01 1.34520161e+00 -5.96257150e-01 9.39525485e-01
-5.13881743e-01 1.09693670e+00 -1.49237499e-01 3.23821276e-01
6.68323815e-01 8.76124427e-02 7.57874325e-02 1.10981095e+00
2.44115829e-01 2.60361005e-02 8.47859859e-01 3.68748337e-01
2.91821808e-02 -2.10294187e-01 -2.22722247e-01 -3.34954597e-02
2.18920067e-01 9.36882138e-01 -7.52990186e-01 -5.97676873e-01
-6.61032856e-01 1.17204893e+00 2.31678173e-01 4.26359624e-01
-1.34413838e+00 1.29538670e-01 4.83030379e-01 2.98139364e-01
4.27001297e-01 -7.18235970e-02 -5.64068817e-02 -1.20002699e+00
2.94159204e-01 -6.34391725e-01 7.17504621e-01 -7.92843103e-01
-8.92265856e-01 2.77850717e-01 8.40286687e-02 -1.83379257e+00
-2.48988986e-01 -3.86037111e-01 -7.27828503e-01 2.42451608e-01
-1.23115063e+00 -1.06547904e+00 -1.95425138e-01 6.72575116e-01
1.05982840e+00 2.13534907e-02 4.67450470e-01 4.11978304e-01
-7.85706580e-01 6.22720122e-01 -3.88499737e-01 4.48513597e-01
8.04440320e-01 -1.22661483e+00 1.10595353e-01 9.73201990e-01
3.82376015e-01 1.23055615e-01 4.10869688e-01 -7.01086283e-01
-7.83012033e-01 -1.33323991e+00 7.26337552e-01 -5.79881191e-01
8.46407413e-01 -2.83254802e-01 -9.25326407e-01 8.89937520e-01
-3.70434821e-02 6.53488517e-01 5.97565830e-01 -9.69195925e-03
-3.53463769e-01 -2.49445930e-01 -6.99698091e-01 5.13083935e-01
1.56614232e+00 -6.72809720e-01 -4.95624214e-01 5.67901075e-01
5.65203428e-01 -2.37848356e-01 -8.67579162e-01 5.20233035e-01
4.33020175e-01 -9.44591582e-01 6.30733967e-01 -5.71691751e-01
5.14825165e-01 -4.41179246e-01 -3.85546386e-02 -8.49606156e-01
-1.85507283e-01 -4.02475536e-01 -2.45332226e-01 1.03781664e+00
6.08604550e-01 -1.40724704e-01 1.07362211e+00 3.55095655e-01
-4.90052968e-01 -9.90029633e-01 -9.19368684e-01 -1.09993339e+00
-3.67431015e-01 -8.69666278e-01 -3.07471510e-02 8.07155252e-01
2.30900422e-01 1.33084178e-01 -4.67247754e-01 9.68710557e-02
3.49154472e-01 -3.77757214e-02 6.46498322e-01 -7.77558327e-01
-4.51901197e-01 -1.63988471e-01 -5.04319191e-01 -1.49228621e+00
2.15538502e-01 -7.69500792e-01 2.32678726e-01 -1.43700159e+00
2.44343042e-01 -2.20209733e-01 -6.69649661e-01 1.21717715e+00
-1.52405962e-01 4.85714942e-01 1.03388213e-01 4.05835658e-01
-1.35021937e+00 4.69155669e-01 1.28231072e+00 -6.72824085e-02
-2.69642770e-01 1.01642817e-01 -1.91369355e-02 9.22694623e-01
6.57103360e-01 -4.97557223e-01 -6.75082147e-01 -2.28300542e-01
-2.50265270e-01 1.05235100e-01 5.00504196e-01 -1.32914805e+00
1.58253133e-01 -2.52868325e-01 4.69949782e-01 -7.68450499e-01
2.28459045e-01 -7.70240545e-01 -2.02364251e-01 4.66825128e-01
-6.12168133e-01 -3.03156853e-01 -1.09963700e-01 7.35288143e-01
-4.30265069e-01 -1.08797707e-01 7.77589619e-01 -6.66412264e-02
-1.16050327e+00 5.61182678e-01 -4.02442217e-01 1.91623047e-01
1.29122436e+00 -5.59175253e-01 -1.10940628e-01 -2.21163198e-01
-1.11075389e+00 5.26155651e-01 2.87359536e-01 3.71123791e-01
3.96645814e-01 -1.56354189e+00 -4.21424925e-01 4.04371172e-02
4.54041928e-01 -4.07194570e-02 2.74437994e-01 1.33535445e+00
-8.09310898e-02 2.67610192e-01 -9.24062207e-02 -9.86954093e-01
-1.26000822e+00 3.50237906e-01 2.60117084e-01 -4.12055314e-01
-6.32188022e-01 8.30511570e-01 2.89825201e-01 8.18334892e-02
5.34208298e-01 -6.33558273e-01 -8.87017325e-02 -2.01068688e-02
5.75921893e-01 1.82706997e-01 -1.88598767e-01 -4.87720162e-01
-5.33064485e-01 3.62589836e-01 -9.33571905e-02 -5.86512834e-02
1.17130637e+00 2.03183487e-01 4.03322101e-01 5.63111603e-01
1.23239410e+00 -4.31210756e-01 -2.17597747e+00 -2.33832553e-01
3.09805423e-01 -3.52382988e-01 -1.28353849e-01 -8.57810318e-01
-1.13898420e+00 7.92344749e-01 5.84494591e-01 -1.24337234e-01
1.09206903e+00 2.82908440e-01 7.23428428e-01 2.75829405e-01
4.70722854e-01 -1.41306317e+00 7.74530470e-01 5.08215129e-01
6.51144803e-01 -1.31933510e+00 -1.71277970e-01 -3.12348127e-01
-1.06019950e+00 6.86863005e-01 1.05034530e+00 -7.59717897e-02
3.05710852e-01 1.81212336e-01 -1.12696528e-01 1.38722360e-01
-1.01319528e+00 -3.34865749e-01 3.60601813e-01 4.19884682e-01
4.60151911e-01 -1.14389822e-01 -9.84958708e-02 4.88059521e-01
4.50489849e-01 2.75199831e-01 7.03410134e-02 1.14727342e+00
-5.79168797e-01 -9.61653233e-01 9.42430049e-02 4.77145523e-01
-4.80118304e-01 6.71820790e-02 -1.68978259e-01 6.49389088e-01
4.31544840e-01 8.40026855e-01 1.19809680e-01 -5.57159424e-01
3.15846294e-01 3.07431728e-01 2.59357870e-01 -7.76893079e-01
-3.12332928e-01 4.70816940e-01 3.36212337e-01 -1.17825317e+00
-8.46087515e-01 -8.81786942e-01 -1.56428778e+00 2.64423996e-01
-2.36310095e-01 1.29042923e-01 1.96278654e-02 1.12432742e+00
2.74066418e-01 5.31613350e-01 5.87089181e-01 -7.85334766e-01
-3.81946623e-01 -1.08895934e+00 -5.15403688e-01 4.22128081e-01
8.56865793e-02 -8.45111728e-01 -3.57384235e-01 5.06968260e-01] | [8.495060920715332, 0.6340492367744446] |
96d3e9ce-06fc-442e-80ec-be1948ae3c80 | safe-exploration-for-interactive-machine | 1910.13726 | null | https://arxiv.org/abs/1910.13726v1 | https://arxiv.org/pdf/1910.13726v1.pdf | Safe Exploration for Interactive Machine Learning | In Interactive Machine Learning (IML), we iteratively make decisions and obtain noisy observations of an unknown function. While IML methods, e.g., Bayesian optimization and active learning, have been successful in applications, on real-world systems they must provably avoid unsafe decisions. To this end, safe IML algorithms must carefully learn about a priori unknown constraints without making unsafe decisions. Existing algorithms for this problem learn about the safety of all decisions to ensure convergence. This is sample-inefficient, as it explores decisions that are not relevant for the original IML objective. In this paper, we introduce a novel framework that renders any existing unsafe IML algorithm safe. Our method works as an add-on that takes suggested decisions as input and exploits regularity assumptions in terms of a Gaussian process prior in order to efficiently learn about their safety. As a result, we only explore the safe set when necessary for the IML problem. We apply our framework to safe Bayesian optimization and to safe exploration in deterministic Markov Decision Processes (MDP), which have been analyzed separately before. Our method outperforms other algorithms empirically. | ['Andreas Krause', 'Matteo Turchetta', 'Felix Berkenkamp'] | 2019-10-30 | safe-exploration-for-interactive-machine-1 | http://papers.nips.cc/paper/8555-safe-exploration-for-interactive-machine-learning | http://papers.nips.cc/paper/8555-safe-exploration-for-interactive-machine-learning.pdf | neurips-2019-12 | ['safe-exploration'] | ['robots'] | [ 1.74959928e-01 5.65760851e-01 -3.15002680e-01 -2.57545501e-01
-1.02096844e+00 -7.12729216e-01 6.46166205e-01 4.89296675e-01
-6.46607101e-01 7.93297350e-01 -2.00476438e-01 -7.44174600e-01
-3.63347590e-01 -8.14708829e-01 -8.89270544e-01 -1.02235258e+00
-9.91776213e-02 7.68093348e-01 2.17923924e-01 2.04836294e-01
4.63749826e-01 3.52288634e-01 -1.19539344e+00 -2.26902038e-01
1.03771639e+00 8.62196326e-01 -7.11154416e-02 7.12988555e-01
8.17601904e-02 7.03651845e-01 -1.10844024e-01 -2.10588053e-01
3.58889014e-01 -2.89856326e-02 -8.67528379e-01 1.32861391e-01
7.46834185e-03 -4.05036062e-01 1.52467176e-01 1.10410571e+00
2.03096181e-01 6.37675405e-01 8.22214782e-01 -1.28776860e+00
-5.25255278e-02 7.55211294e-01 -3.83299083e-01 -1.19149595e-01
1.38545319e-01 3.70611459e-01 9.06260073e-01 -8.43644857e-01
2.62168348e-01 1.55266559e+00 2.51561165e-01 5.05157232e-01
-1.58728027e+00 -4.41484898e-01 7.21218050e-01 1.51369780e-01
-1.23356462e+00 -6.49581492e-01 5.53530276e-01 -5.47431171e-01
6.01837158e-01 2.58759022e-01 5.62719464e-01 1.10732138e+00
3.08791161e-01 1.09215820e+00 1.08274829e+00 -5.49974144e-01
9.95719850e-01 2.60825515e-01 2.05748126e-01 5.42121708e-01
4.40314174e-01 4.18760687e-01 -5.78006923e-01 -5.51710367e-01
4.91764337e-01 -7.62713701e-02 -2.52096146e-01 -6.34799302e-01
-1.03295934e+00 1.01480603e+00 -2.54571497e-01 -3.29541683e-01
-1.52953207e-01 3.08146477e-01 -2.59216968e-02 3.08616519e-01
3.98103863e-01 5.82310796e-01 -3.83789986e-01 -1.63811862e-01
-7.43633211e-01 5.00858903e-01 1.25327551e+00 8.64239514e-01
7.78514504e-01 -2.36638248e-01 -1.18478455e-01 3.77549320e-01
7.85476446e-01 3.55205715e-01 -3.02037120e-01 -1.11256421e+00
3.69297922e-01 8.84896144e-02 7.70583928e-01 -9.00029898e-01
-1.38003692e-01 -2.18460813e-01 -5.05946279e-01 5.65866292e-01
5.99375665e-01 -2.73977011e-01 -9.43957627e-01 1.56150544e+00
5.76512337e-01 3.13057080e-02 -1.20955385e-01 6.73152566e-01
-1.00398012e-01 7.43862808e-01 -6.65605441e-02 -6.42310381e-01
5.42972207e-01 -6.40491247e-01 -8.45332325e-01 -3.30252618e-01
5.19558191e-01 -4.12983865e-01 1.06635094e+00 1.01310289e+00
-1.15456617e+00 -1.92254305e-01 -9.94575262e-01 3.07828963e-01
9.15061682e-02 -4.91130710e-01 5.27802825e-01 7.45968103e-01
-8.37250829e-01 6.51329756e-01 -1.35448825e+00 1.19466372e-01
6.52280569e-01 5.14699876e-01 1.44856095e-01 7.74117038e-02
-7.87283897e-01 9.41363215e-01 6.33275449e-01 2.72946864e-01
-1.64902425e+00 -4.96093571e-01 -8.36940348e-01 -8.62123445e-03
1.52842271e+00 -2.99342513e-01 1.59910035e+00 -5.90230346e-01
-1.67104971e+00 1.84621781e-01 -1.14720665e-01 -7.25962758e-01
9.19661939e-01 -6.81670904e-01 1.47135854e-01 1.22713475e-02
-3.05963576e-01 3.88056934e-01 1.10619330e+00 -1.34857094e+00
-6.33907199e-01 -8.89905542e-02 3.63414437e-01 1.54603928e-01
-3.89598720e-02 -1.93807557e-01 -4.07567143e-01 -3.64231080e-01
1.65245295e-01 -1.02598798e+00 -8.56216848e-01 2.38615215e-01
-7.22703099e-01 -2.73439020e-01 7.05392659e-01 -2.31056914e-01
1.20802736e+00 -1.95371890e+00 2.11635619e-01 6.04051530e-01
2.10320845e-01 -8.92865211e-02 3.52178901e-01 3.32330018e-01
4.92280424e-01 5.13995647e-01 -6.58122957e-01 -5.69040477e-01
2.09540144e-01 5.33947349e-01 -4.99671221e-01 8.88049066e-01
2.36781448e-01 5.05147934e-01 -1.02954137e+00 -6.26403332e-01
4.01950598e-01 1.53229525e-02 -7.41508961e-01 2.99071223e-01
-6.89543724e-01 7.00775146e-01 -6.04392946e-01 5.44737697e-01
4.33366418e-01 -1.05593748e-01 8.31480473e-02 3.93425971e-01
-1.91030174e-01 1.55857682e-01 -1.45399821e+00 1.36734080e+00
-5.23687363e-01 2.93260008e-01 1.78764835e-01 -1.13215256e+00
6.54694021e-01 2.82400370e-01 3.70710731e-01 6.20833449e-02
8.79568979e-02 1.09374568e-01 -1.14177369e-01 -3.21871459e-01
1.06801674e-01 -3.65942679e-02 -7.96302781e-02 5.99786580e-01
-1.54945165e-01 -2.70978957e-01 2.78705787e-02 1.15513079e-01
8.72992635e-01 2.69923270e-01 6.77103937e-01 -5.48359573e-01
3.14088434e-01 -2.29567200e-01 9.44633245e-01 1.47056651e+00
-1.28782734e-01 3.19921225e-01 8.07834208e-01 -2.70508409e-01
-6.51855588e-01 -1.04025638e+00 -1.63641319e-01 8.83335829e-01
3.20765555e-01 -3.43087167e-01 -5.28005719e-01 -1.11825275e+00
-1.02911294e-01 1.28274035e+00 -5.60296774e-01 -2.89263688e-02
-4.53549117e-01 -7.05118060e-01 -5.99194728e-02 2.62774140e-01
2.47845605e-01 -5.04364729e-01 -7.24618793e-01 3.99362981e-01
2.72485554e-01 -7.28049636e-01 -1.70485437e-01 4.60290074e-01
-9.04834628e-01 -1.00826514e+00 -1.92989051e-01 -1.08095944e-01
8.42553258e-01 -2.41622001e-01 9.74217951e-01 -2.88693547e-01
3.66723239e-02 4.49256659e-01 -1.98455483e-01 -8.13528419e-01
-4.72292364e-01 -1.97555870e-01 1.56183988e-01 7.74360821e-02
2.15294510e-01 -4.48915899e-01 -3.18814963e-01 3.71793181e-01
-7.52993464e-01 3.77421826e-02 4.00871515e-01 7.19110250e-01
8.77058268e-01 3.07099968e-01 5.40258661e-02 -1.11613441e+00
4.40846920e-01 -5.28937280e-01 -1.21394145e+00 3.35343868e-01
-8.06133389e-01 4.57845867e-01 4.45339233e-01 -5.70413589e-01
-1.16885245e+00 1.97429791e-01 1.74943432e-01 -4.01427299e-01
-2.22176731e-01 3.98714989e-01 -5.00078857e-01 7.84654021e-02
7.08586931e-01 -2.13341981e-01 -2.30221719e-01 -4.61856455e-01
2.85041779e-01 4.28062797e-01 1.30532548e-01 -1.07236254e+00
8.48626494e-01 5.98196208e-01 1.79203704e-01 -6.07492447e-01
-1.03931653e+00 -2.08573356e-01 -5.66805780e-01 -3.15134823e-01
4.12868589e-01 -3.76826674e-01 -9.31584299e-01 6.33601006e-03
-8.75985205e-01 -6.43584669e-01 -3.94386411e-01 4.66303229e-01
-8.32878590e-01 3.22269350e-01 -3.01092826e-02 -1.65999126e+00
1.12785902e-02 -1.37199080e+00 7.85263777e-01 1.05349019e-01
-5.34240305e-01 -1.08016217e+00 5.23852243e-04 2.61312313e-02
2.00098351e-01 2.79403508e-01 6.66759908e-01 -7.62140155e-01
-9.80102956e-01 3.69885564e-02 4.98039365e-01 2.41573811e-01
5.96821904e-02 1.55964911e-01 -8.09201300e-01 -4.32725161e-01
4.76423234e-01 -2.81684875e-01 1.04455805e+00 4.32387948e-01
1.53745294e+00 -8.86769950e-01 -3.92631471e-01 4.68312681e-01
1.16204202e+00 5.34360111e-01 2.00096235e-01 2.65134484e-01
3.93532515e-01 7.53032327e-01 8.99366617e-01 7.43803322e-01
6.20761281e-03 4.73004818e-01 6.70395434e-01 1.81612834e-01
7.05692589e-01 -5.16440161e-02 4.59304869e-01 2.19470948e-01
4.02085185e-02 -4.25952673e-01 -1.23628616e+00 4.30175394e-01
-2.32076502e+00 -6.22976243e-01 2.62451410e-01 2.68268299e+00
1.03375840e+00 6.09779298e-01 -6.65544048e-02 1.69313818e-01
4.73199457e-01 -9.09529999e-02 -9.15825725e-01 -3.95087153e-01
2.42273882e-01 -7.38907000e-03 6.44642830e-01 1.01472867e+00
-1.38719106e+00 5.74910164e-01 5.89400578e+00 8.37276697e-01
-6.90824747e-01 1.88336581e-01 7.80775249e-01 -4.05213475e-01
-4.80904877e-01 4.68565464e-01 -1.07935929e+00 4.14023787e-01
7.84293711e-01 -1.13800295e-01 4.44280952e-01 1.08285546e+00
4.68567550e-01 -5.49767256e-01 -1.55390310e+00 7.65177310e-01
-4.06643867e-01 -1.06715167e+00 -2.04141214e-01 1.17002092e-01
7.72515059e-01 -4.04396564e-01 3.30725387e-02 1.83828741e-01
9.39177513e-01 -1.15616858e+00 9.49557781e-01 7.11954176e-01
1.57928661e-01 -1.13495910e+00 6.20815217e-01 7.77437270e-01
-6.27109289e-01 -3.49327892e-01 -1.50194332e-01 -7.61249959e-02
3.37242752e-01 8.89005482e-01 -8.91041815e-01 1.74911365e-01
4.66238528e-01 5.74401736e-01 -1.43370852e-01 1.13904607e+00
-7.57865906e-01 1.05984998e+00 -7.18078077e-01 -3.69388536e-02
3.29718709e-01 -2.61700183e-01 9.39583361e-01 9.16924357e-01
-9.75795835e-03 -2.41051503e-02 6.34647012e-01 9.44699347e-01
5.06633043e-01 -1.07879989e-01 -5.34138858e-01 -6.33536875e-02
5.91392040e-01 7.11369693e-01 -7.68601477e-01 -2.05084771e-01
-2.01781243e-01 4.39595491e-01 1.52703419e-01 6.09149158e-01
-6.27380133e-01 -1.52686089e-02 4.89501894e-01 -7.41150901e-02
5.61040230e-02 -4.56593126e-01 -5.36791027e-01 -9.49617982e-01
-1.13688104e-01 -9.12951529e-01 4.12830710e-01 -1.02079228e-01
-1.22879589e+00 -3.56213115e-02 3.68324608e-01 -9.41227198e-01
-4.74561661e-01 -4.16637212e-01 -4.76253480e-01 6.94766819e-01
-1.12534440e+00 -5.42854965e-01 2.50082970e-01 3.81349713e-01
7.08277524e-01 1.94291830e-01 5.21992743e-01 -3.04467767e-01
-8.42443287e-01 2.47222632e-01 2.42200375e-01 -3.01519215e-01
5.13511539e-01 -1.42861831e+00 2.47966826e-01 1.17121685e+00
1.05104350e-01 7.99493015e-01 1.03865540e+00 -8.03350389e-01
-1.45987308e+00 -9.88354445e-01 2.22395092e-01 -4.01708961e-01
5.94721317e-01 -5.49068093e-01 -8.18878829e-01 8.56683731e-01
-2.33797327e-01 -1.47668645e-01 4.31339085e-01 3.76450986e-01
6.00269325e-02 -2.97697298e-02 -9.72084820e-01 8.95531058e-01
8.12081099e-01 -1.69183969e-01 -5.19693613e-01 4.07679111e-01
6.34003282e-01 -3.11497569e-01 -5.37914932e-01 4.57761556e-01
2.77274191e-01 -6.79195583e-01 8.42585862e-01 -6.16609931e-01
1.81034785e-02 -4.00024295e-01 -1.99327677e-01 -1.15591431e+00
-4.92775664e-02 -1.36760044e+00 -8.25509667e-01 9.07919049e-01
6.02399468e-01 -7.13585913e-01 6.28269613e-01 1.05423927e+00
-9.04090106e-02 -9.99367535e-01 -1.04601896e+00 -8.61587703e-01
2.37903818e-02 -9.04846847e-01 2.81967342e-01 5.37042618e-01
-1.22486562e-01 -1.54184431e-01 -5.27859867e-01 4.52746302e-01
1.15222132e+00 4.34460044e-02 6.43859208e-01 -1.17776537e+00
-6.99688077e-01 -3.00520867e-01 1.64069340e-01 -1.05791485e+00
1.05192818e-01 -4.16116267e-01 5.83833933e-01 -1.16549122e+00
1.09840535e-01 -8.01627517e-01 -2.40445897e-01 5.29950678e-01
-2.73449987e-01 -6.14227831e-01 1.19967408e-01 2.45393500e-01
-7.93645442e-01 5.86396992e-01 1.04044485e+00 -1.48002625e-01
-5.67034245e-01 5.00905693e-01 -6.56519830e-01 1.06717050e+00
9.12782967e-01 -6.84336245e-01 -6.25141978e-01 -7.01541901e-02
6.05928898e-01 7.88914971e-03 3.13743711e-01 -5.78147948e-01
3.42012286e-01 -7.40047932e-01 4.70390730e-02 -6.82387531e-01
3.01389873e-01 -7.89543331e-01 2.19213944e-02 4.47484434e-01
-6.47395790e-01 -7.62632012e-01 9.01504047e-03 1.01705253e+00
8.93801972e-02 -5.18592179e-01 7.45594919e-01 -6.65697977e-02
-3.92498791e-01 3.98486316e-01 -7.39046037e-01 1.25401780e-01
1.12744236e+00 2.92881969e-02 1.29626453e-01 -5.64437628e-01
-1.02051842e+00 5.94169676e-01 4.67059463e-01 -5.10985404e-02
4.77866530e-01 -6.76948905e-01 -5.40979385e-01 -6.06943779e-02
-1.21438287e-01 7.68227756e-01 -1.63000822e-01 9.64690983e-01
-1.94616497e-01 2.71450728e-01 4.61734354e-01 -7.53931820e-01
-1.00052345e+00 6.68109238e-01 2.14630589e-01 -2.76062071e-01
-5.53324819e-01 9.07596707e-01 2.95529515e-01 -1.57674193e-01
6.47149384e-01 -3.90213698e-01 -8.00372213e-02 4.34458163e-03
5.21041811e-01 5.23561180e-01 -8.79023820e-02 2.90284306e-01
-8.73831734e-02 1.91101506e-01 -3.17699671e-01 -5.55285275e-01
1.13273573e+00 -3.74075115e-01 1.29224017e-01 8.47251058e-01
5.92945278e-01 1.63220018e-02 -1.76949787e+00 -3.40419412e-01
4.49081957e-01 -6.44886494e-01 3.67415845e-01 -6.08585656e-01
-5.04201472e-01 8.77221584e-01 2.56919026e-01 1.30707473e-01
9.47788894e-01 -6.17342032e-02 2.22783625e-01 9.76868391e-01
6.13519490e-01 -1.28384793e+00 -1.35011151e-01 5.02083480e-01
7.75578201e-01 -1.36221099e+00 1.49275005e-01 -4.90150750e-01
-4.20044541e-01 1.06252694e+00 4.00977582e-01 1.09942727e-01
7.12702155e-01 4.89279449e-01 -4.19204950e-01 1.78218678e-01
-1.08189428e+00 -7.97288045e-02 8.02338943e-02 3.16204160e-01
-2.64926821e-01 1.62738394e-02 -1.53400734e-01 4.29242611e-01
1.60288766e-01 -9.21626911e-02 4.53405470e-01 1.40999699e+00
-6.53947771e-01 -9.50213075e-01 -7.74120808e-01 3.93994093e-01
-5.20559072e-01 -1.83300872e-03 -1.72740504e-01 6.58317387e-01
-6.66580498e-02 1.23375905e+00 -1.54481843e-01 -3.30807231e-02
-4.60205898e-02 -5.37366606e-02 4.02959883e-01 -7.87417948e-01
-1.69926912e-01 2.77743429e-01 3.51005495e-01 -8.17457199e-01
-9.51394141e-02 -8.65394711e-01 -9.89263654e-01 -1.26521349e-01
-4.46582437e-01 2.03718573e-01 4.73762542e-01 1.25479209e+00
-2.03669500e-02 2.06049412e-01 5.81137061e-01 -7.62685895e-01
-1.10149539e+00 -6.44343615e-01 -4.47635680e-01 -1.50727883e-01
4.58399832e-01 -8.05146277e-01 -6.19937122e-01 -1.35276150e-02] | [4.717435836791992, 2.7731192111968994] |
4b772ac4-e6d4-41f7-8e15-a3e2ee9b6e00 | self-supervised-metric-learning-in-multi-view | 2106.07138 | null | https://arxiv.org/abs/2106.07138v4 | https://arxiv.org/pdf/2106.07138v4.pdf | Self-Supervised Metric Learning in Multi-View Data: A Downstream Task Perspective | Self-supervised metric learning has been a successful approach for learning a distance from an unlabeled dataset. The resulting distance is broadly useful for improving various distance-based downstream tasks, even when no information from downstream tasks is utilized in the metric learning stage. To gain insights into this approach, we develop a statistical framework to theoretically study how self-supervised metric learning can benefit downstream tasks in the context of multi-view data. Under this framework, we show that the target distance of metric learning satisfies several desired properties for the downstream tasks. On the other hand, our investigation suggests the target distance can be further improved by moderating each direction's weights. In addition, our analysis precisely characterizes the improvement by self-supervised metric learning on four commonly used downstream tasks: sample identification, two-sample testing, $k$-means clustering, and $k$-nearest neighbor classification. When the distance is estimated from an unlabeled dataset, we establish the upper bound on distance estimation's accuracy and the number of samples sufficient for downstream task improvement. Finally, numerical experiments are presented to support the theoretical results in the paper. | ['Shulei Wang'] | 2021-06-14 | null | null | null | null | ['hypothesis-testing', 'hypothesis-testing'] | ['methodology', 'miscellaneous'] | [ 5.94001040e-02 -5.30033186e-03 -5.11057496e-01 -9.37445521e-01
-1.19131160e+00 -5.83764374e-01 2.22397432e-01 3.17318499e-01
-5.24209917e-01 6.61713243e-01 1.46252483e-01 -3.17224801e-01
-7.56062508e-01 -6.82743430e-01 -2.92841017e-01 -9.10330415e-01
1.00839220e-01 1.69701815e-01 -9.12353620e-02 1.10288919e-03
8.15786779e-01 4.65645939e-01 -1.41333485e+00 -1.15662977e-01
1.02479386e+00 9.84255493e-01 1.01145200e-01 4.75795239e-01
1.46205738e-01 3.35325241e-01 -3.18237633e-01 -4.72495079e-01
6.70857728e-01 -5.00129759e-01 -8.69607389e-01 -1.24141850e-01
2.59854734e-01 -2.42298812e-01 -7.42407367e-02 1.11474240e+00
5.27266204e-01 2.24738762e-01 1.08498442e+00 -1.61034644e+00
-6.92146957e-01 8.20460141e-01 -6.22369349e-01 2.96509057e-01
9.01505630e-03 -1.81558371e-01 1.21999681e+00 -1.07078445e+00
3.78937930e-01 9.04586256e-01 9.21747267e-01 3.69907528e-01
-9.66212392e-01 -5.89544594e-01 4.55782227e-02 4.23983842e-01
-1.58439624e+00 -3.45338404e-01 8.99833381e-01 -6.03195488e-01
5.55461705e-01 4.15344499e-02 1.11472802e-02 5.00917077e-01
8.35246295e-02 5.65167010e-01 1.04974306e+00 -5.10864556e-01
3.42632711e-01 3.28469574e-01 4.61832076e-01 6.52943969e-01
2.54281521e-01 3.62632692e-01 -6.30996823e-01 -1.32691830e-01
4.15462255e-01 -3.55051383e-02 -7.31184483e-02 -6.68716788e-01
-1.15961134e+00 9.53658640e-01 3.40349942e-01 4.51795369e-01
3.02463591e-01 -8.44663009e-02 3.92629683e-01 7.21974790e-01
7.83672154e-01 6.38861835e-01 -6.72389984e-01 -3.47080290e-01
-7.59648442e-01 -2.46434122e-01 5.61674297e-01 1.12039864e+00
9.26640511e-01 -3.77674043e-01 -3.79668362e-02 7.93194115e-01
2.10474759e-01 3.81187171e-01 5.68187118e-01 -1.14844573e+00
8.17358315e-01 5.91706455e-01 8.58988538e-02 -8.51264894e-01
-3.89357984e-01 -2.69045502e-01 -6.14149392e-01 9.21475515e-02
9.19000208e-01 -2.32810646e-01 -2.95367420e-01 1.89213097e+00
4.13269043e-01 -5.41035610e-04 8.34354199e-03 6.21052444e-01
5.33031300e-02 2.48777643e-01 -3.15468162e-01 -4.88378048e-01
6.98061943e-01 -8.98183823e-01 -3.60064447e-01 1.85048178e-01
1.58352983e+00 -5.58936894e-01 1.14659536e+00 2.52872884e-01
-5.96670032e-01 -6.65656567e-01 -1.18946898e+00 -6.26556063e-03
-5.38132071e-01 1.29102012e-02 4.34797943e-01 8.53712201e-01
-1.07053542e+00 9.04051363e-01 -5.50227106e-01 -4.36917812e-01
2.33223781e-01 3.58814925e-01 -3.41700941e-01 -3.51383835e-01
-9.36388433e-01 8.58663857e-01 2.82374442e-01 -9.43586007e-02
-6.35606647e-01 -8.82181525e-01 -8.29074502e-01 -7.03362375e-02
3.16255055e-02 -3.70396137e-01 1.04144835e+00 -3.48583192e-01
-1.11511219e+00 7.97214150e-01 -1.83062673e-01 -2.09746689e-01
5.23655593e-01 1.43277258e-01 -1.96610302e-01 -1.41694797e-02
3.85359228e-01 4.01448727e-01 7.10923851e-01 -8.98264110e-01
-1.11367226e+00 -9.28747237e-01 1.24570742e-01 3.58508170e-01
-8.08241546e-01 -3.63221854e-01 8.52029845e-02 -5.04280746e-01
1.52391300e-01 -8.64740074e-01 -3.80650371e-01 5.50608598e-02
-2.62984216e-01 -4.66219366e-01 9.73047554e-01 -1.82375878e-01
1.30028331e+00 -2.41755843e+00 7.60149434e-02 5.01743436e-01
2.97920287e-01 -1.88029572e-01 -2.83890337e-01 3.12428802e-01
-1.26249149e-01 2.41423979e-01 -3.04026812e-01 -1.70149326e-01
1.48608582e-02 -1.79248914e-01 1.53503492e-01 8.19861710e-01
-2.33320370e-01 7.30455279e-01 -1.02097785e+00 -4.53016549e-01
1.23485275e-01 -7.64332190e-02 -4.13562715e-01 9.61508006e-02
4.94454116e-01 2.59787917e-01 -4.48942333e-01 5.55084586e-01
4.21131849e-01 -1.77999422e-01 -7.42422491e-02 -1.62659332e-01
1.08602747e-01 -3.11490446e-01 -1.17057538e+00 1.76913428e+00
-5.14434278e-01 6.46327972e-01 -1.87504560e-01 -1.58384609e+00
1.12717271e+00 -4.82694879e-02 7.70814955e-01 -5.78519821e-01
-5.79044875e-03 3.54015827e-01 1.40530080e-01 -4.79729444e-01
7.62811229e-02 -3.42776716e-01 -3.02677482e-01 8.80317569e-01
-2.12944746e-01 -5.34963124e-02 2.33787641e-01 6.20111525e-02
1.10215569e+00 -1.15214720e-01 4.25217420e-01 -6.04506612e-01
6.68651998e-01 1.48959503e-01 5.02186954e-01 6.57415450e-01
-7.10446894e-01 2.82034338e-01 3.02939951e-01 -1.74296312e-02
-1.03116453e+00 -1.12653267e+00 -3.73843610e-01 1.37608314e+00
3.12916130e-01 -9.94813815e-02 -8.23204160e-01 -1.09434021e+00
2.84962893e-01 5.44134080e-01 -8.26091230e-01 -6.74225688e-01
-2.57584423e-01 -8.26763809e-01 4.62418377e-01 9.02135491e-01
3.21388632e-01 -4.30968702e-01 -6.76757246e-02 -8.07948709e-02
-1.11815609e-01 -6.33297384e-01 -9.09798503e-01 4.30636853e-01
-1.18529725e+00 -1.56374049e+00 -6.44062579e-01 -9.54482019e-01
8.24882627e-01 7.65210152e-01 5.73423684e-01 -3.34771425e-01
-1.99334137e-02 5.36656916e-01 -4.34027463e-01 -1.38946190e-01
-2.51675695e-01 2.56457895e-01 3.79557282e-01 -1.22860387e-01
8.30779135e-01 -5.38260937e-01 -5.60480118e-01 7.21523285e-01
-5.89154661e-01 -5.88626266e-01 4.16986674e-01 8.84456575e-01
4.83924121e-01 1.82490915e-01 1.10731757e+00 -9.00097966e-01
6.71604574e-01 -6.02075696e-01 -3.76591057e-01 3.29163879e-01
-1.29912221e+00 3.37277561e-01 7.88858354e-01 -2.98218459e-01
-8.93385112e-01 -8.73345584e-02 1.86055899e-01 -6.89015910e-02
3.77457775e-02 5.11462092e-01 -4.02060360e-01 -1.49979502e-01
1.04417574e+00 7.95686804e-03 2.34206870e-01 -3.25336158e-01
6.32160962e-01 9.87415314e-01 6.01618513e-02 -5.63013315e-01
9.11993921e-01 6.01547301e-01 1.33217409e-01 -5.28399169e-01
-1.29955578e+00 -8.96765351e-01 -1.15903282e+00 -8.17106850e-03
6.04194939e-01 -5.77874303e-01 -6.52008593e-01 3.00735891e-01
-4.86268789e-01 -4.80979502e-01 -3.97361994e-01 6.07541084e-01
-9.28642809e-01 4.25296992e-01 -4.57894504e-01 -5.93956828e-01
2.06490550e-02 -9.92509961e-01 8.01605523e-01 -6.51904419e-02
-1.41683847e-01 -1.36573040e+00 2.35346835e-02 4.96753216e-01
1.20078363e-01 -5.97777888e-02 1.06334603e+00 -1.00893998e+00
4.66283560e-02 -8.48061591e-02 -3.30496371e-01 5.59252381e-01
5.84470332e-01 -3.45492750e-01 -8.35051239e-01 -3.92480582e-01
2.62337834e-01 -2.40547970e-01 5.98188698e-01 5.03575146e-01
1.31905615e+00 4.35616337e-02 -4.85248715e-01 6.38448298e-01
1.32095206e+00 3.35295141e-01 8.29394385e-02 2.73155630e-01
7.63342679e-01 9.89666283e-01 1.33747482e+00 5.08957505e-01
3.10011834e-01 3.77801001e-01 1.56761065e-01 1.98557347e-01
2.13160306e-01 -1.38254434e-01 1.19460180e-01 8.25733840e-01
3.01954269e-01 1.21177211e-01 -8.09608400e-01 5.37468493e-01
-1.85161042e+00 -9.32806790e-01 -2.17105597e-01 2.31442308e+00
7.51582444e-01 -5.94425090e-02 2.40350276e-01 6.27157032e-01
1.01102531e+00 -1.61829352e-01 -8.88457656e-01 -3.47837895e-01
2.58008987e-01 -1.61753401e-01 5.66357195e-01 6.20720327e-01
-1.01211095e+00 6.31886482e-01 6.77841568e+00 8.88476074e-01
-6.67331159e-01 3.26075517e-02 8.81536782e-01 -3.12392619e-02
-3.82059842e-01 -2.57407904e-01 -9.68585789e-01 4.00432825e-01
9.67766702e-01 -6.50967896e-01 1.98983222e-01 1.18358803e+00
5.39857090e-01 4.62695211e-03 -1.75507104e+00 1.13203430e+00
-1.89644992e-02 -1.07083666e+00 -2.27397680e-01 3.03711832e-01
7.69647777e-01 -2.54977316e-01 2.33799666e-01 4.80659753e-01
2.72352725e-01 -7.88856149e-01 2.15128794e-01 8.51579234e-02
1.10591757e+00 -1.23808408e+00 8.33365858e-01 6.05331481e-01
-1.15090811e+00 -4.25081730e-01 -4.37210530e-01 -1.25259265e-01
-1.06807575e-01 8.96228135e-01 -8.48497331e-01 3.08796257e-01
4.47232902e-01 1.04982245e+00 -6.28467202e-01 7.80926168e-01
2.21623689e-01 2.65406936e-01 4.06106636e-02 2.04896517e-02
2.57642478e-01 -5.75146317e-01 1.12256870e-01 9.52377975e-01
6.15440965e-01 1.22004643e-01 -1.23288557e-01 3.89649153e-01
4.61172573e-02 2.93815672e-01 -1.11482024e+00 1.97921857e-01
7.88679957e-01 9.12330210e-01 -4.93022919e-01 -1.22326791e-01
-3.14745337e-01 9.08603132e-01 5.00590146e-01 2.65466511e-01
-8.03483367e-01 -7.42760837e-01 9.53009725e-01 -1.02474585e-01
5.49834110e-02 -1.57024100e-01 -6.61605716e-01 -8.84843826e-01
5.53524494e-02 -4.32430744e-01 6.58614576e-01 -2.22040415e-01
-1.76545012e+00 -6.49295375e-02 -3.02490164e-02 -1.53019249e+00
-3.76042634e-01 -6.43452108e-01 -2.94680595e-01 5.44371247e-01
-1.33165002e+00 -5.97385526e-01 -9.74700823e-02 7.85605729e-01
4.67742413e-01 -1.58576369e-01 6.00009441e-01 2.85619915e-01
-3.87094498e-01 1.00743842e+00 6.66059196e-01 1.98128790e-01
1.09429371e+00 -1.50189507e+00 -1.28191933e-01 8.00126612e-01
4.89158928e-02 4.12075698e-01 2.09799215e-01 -3.73069078e-01
-1.20724523e+00 -1.21951091e+00 9.30908501e-01 -6.16851389e-01
7.89833724e-01 -8.21306556e-02 -4.77439106e-01 5.09878993e-01
-6.16425514e-01 1.92281008e-01 1.35441768e+00 3.41386795e-01
-6.02090716e-01 -5.50626397e-01 -1.48892999e+00 5.08503914e-01
1.58055937e+00 -5.98230243e-01 -4.26119238e-01 4.09965664e-01
7.37494171e-01 9.74802449e-02 -1.27349746e+00 3.86131734e-01
4.04056519e-01 -1.08154607e+00 8.68398368e-01 -6.43900990e-01
4.10543054e-01 -1.43273503e-01 -5.62430739e-01 -1.53494370e+00
-2.90642470e-01 -1.69197530e-01 -2.07609218e-02 1.32616174e+00
6.01127088e-01 -6.02370441e-01 1.03299260e+00 7.00716138e-01
-2.29874477e-01 -7.36087322e-01 -1.01431096e+00 -1.13572669e+00
4.70380783e-01 -5.46872020e-01 3.20942253e-01 1.12588334e+00
5.33820033e-01 3.21238905e-01 -7.14099407e-02 -7.91161060e-02
8.60313773e-01 1.38939172e-01 6.19663715e-01 -1.24319756e+00
-1.27397031e-01 -6.14158273e-01 -4.77807730e-01 -1.15874922e+00
3.68159026e-01 -1.15959823e+00 1.75240412e-01 -1.24554312e+00
2.38349780e-01 -8.24786842e-01 -4.81049061e-01 -7.74582149e-03
-3.30686837e-01 4.44937870e-02 -6.62568510e-02 1.11098699e-01
-5.42640626e-01 4.36352491e-01 1.25753081e+00 -1.31054044e-01
-1.87442064e-01 3.10447752e-01 -1.26308477e+00 5.62566221e-01
9.40412879e-01 -5.12291014e-01 -9.60418582e-01 -3.03060055e-01
-1.17547929e-01 3.72675844e-02 -3.04690629e-01 -5.81533968e-01
1.86454698e-01 -3.63977641e-01 1.76271915e-01 -3.60997051e-01
1.43005894e-02 -9.82271791e-01 -5.91614604e-01 5.11271358e-01
-6.13562286e-01 4.25713416e-03 -4.47651565e-01 9.10892367e-01
-1.66740224e-01 -4.06256914e-01 9.62653100e-01 2.11139947e-01
-6.02711380e-01 5.26390493e-01 -1.51226446e-01 3.76177251e-01
1.39281297e+00 -4.69025552e-01 -1.86077461e-01 -2.74409115e-01
-7.57175446e-01 3.11147213e-01 3.37246090e-01 1.60096034e-01
6.25366986e-01 -1.60670984e+00 -6.40110135e-01 3.04012336e-02
5.30423045e-01 -1.55022338e-01 -7.91070238e-02 9.20253813e-01
-1.40400797e-01 4.03871238e-01 1.01437077e-01 -7.33777463e-01
-1.03814220e+00 8.20865214e-01 1.98100239e-01 -1.86767593e-01
-2.32667327e-01 8.52555156e-01 1.31150663e-01 -5.08818686e-01
3.45245302e-01 -3.41203272e-01 -3.63157578e-02 2.40009785e-01
5.44938445e-01 9.88065243e-01 5.60110137e-02 -2.59039670e-01
-3.51035148e-01 6.68093801e-01 4.66411300e-02 -2.48359233e-01
1.16436636e+00 -6.36035800e-01 1.41429201e-01 6.97097600e-01
1.88930511e+00 -1.89844444e-01 -1.25143802e+00 -3.01302165e-01
5.57391524e-01 -6.79098248e-01 -2.74082631e-01 -1.78919315e-01
-9.09136236e-01 7.97860444e-01 4.18379635e-01 1.25371441e-01
9.89360511e-01 2.77169812e-02 5.82741141e-01 6.33181691e-01
5.97706139e-01 -1.36019695e+00 3.71089041e-01 4.29087430e-01
4.47600365e-01 -1.66646111e+00 -1.87578306e-01 -4.04507309e-01
-4.22135144e-01 1.00138211e+00 7.87783146e-01 -3.81328650e-02
1.13675630e+00 1.53217047e-01 9.38576981e-02 3.56822796e-02
-3.69735062e-01 -2.64484845e-02 1.29514679e-01 1.01840794e+00
4.24819708e-01 -3.86960320e-02 -5.11932671e-01 3.47623765e-01
-2.48414725e-01 -4.37619686e-01 4.02692527e-01 7.65519261e-01
-5.20928621e-01 -1.10225904e+00 -1.13376379e-01 6.57143533e-01
-6.28398508e-02 -8.21539089e-02 -3.99974406e-01 7.47821033e-01
-1.28009632e-01 1.37833154e+00 4.09820583e-03 -5.38239062e-01
3.31255682e-02 6.75750449e-02 4.54298973e-01 -8.83590281e-01
-1.73602372e-01 -4.56365168e-01 -1.15279198e-01 -5.91159701e-01
-4.01823729e-01 -7.31835902e-01 -1.10070586e+00 -4.13746983e-01
-4.56883997e-01 4.51089859e-01 4.63437885e-01 1.13545644e+00
2.32124612e-01 -7.84742013e-02 1.35785282e+00 -3.69641095e-01
-1.10242462e+00 -9.36067700e-01 -1.01147676e+00 3.93920690e-01
1.12499721e-01 -5.94582975e-01 -7.65717506e-01 1.03469484e-01] | [9.343716621398926, 3.36238956451416] |
74a6e75e-5e46-46e7-852e-dfdd30d247b1 | what-s-in-a-name-answer-equivalence-for-open | 2109.05289 | null | https://arxiv.org/abs/2109.05289v1 | https://arxiv.org/pdf/2109.05289v1.pdf | What's in a Name? Answer Equivalence For Open-Domain Question Answering | A flaw in QA evaluation is that annotations often only provide one gold answer. Thus, model predictions semantically equivalent to the answer but superficially different are considered incorrect. This work explores mining alias entities from knowledge bases and using them as additional gold answers (i.e., equivalent answers). We incorporate answers for two settings: evaluation with additional answers and model training with equivalent answers. We analyse three QA benchmarks: Natural Questions, TriviaQA, and SQuAD. Answer expansion increases the exact match score on all datasets for evaluation, while incorporating it helps model training over real-world datasets. We ensure the additional answers are valid through a human post hoc evaluation. | ['Jordan Boyd-Graber', 'Chen Zhao', 'Chenglei Si'] | 2021-09-11 | null | null | null | null | ['triviaqa'] | ['miscellaneous'] | [-3.46905701e-02 7.93131948e-01 -2.15584233e-01 -5.83120763e-01
-1.29710531e+00 -9.57583368e-01 6.28557801e-01 6.81313097e-01
-6.92767024e-01 1.08788371e+00 5.77692986e-01 -5.86110890e-01
-4.39106897e-02 -1.01665783e+00 -8.43515813e-01 2.25368723e-01
4.30508673e-01 9.62983668e-01 9.21763778e-01 -6.01195753e-01
6.63152933e-02 -2.58328259e-01 -1.26140547e+00 8.03029537e-01
1.50505495e+00 8.59911084e-01 -4.64689255e-01 6.66934073e-01
-2.05401480e-01 1.07312715e+00 -8.01530957e-01 -1.67962801e+00
1.49552613e-01 -3.28975409e-01 -1.63015842e+00 -8.41597557e-01
8.08114946e-01 -2.92956661e-02 2.21446723e-01 7.91490555e-01
4.82181460e-01 7.27060363e-02 6.61635399e-01 -1.08093107e+00
-8.42099965e-01 6.06794894e-01 2.12134242e-01 1.60502329e-01
1.11536682e+00 1.54042140e-01 1.38736618e+00 -9.76211905e-01
7.90549695e-01 1.19917715e+00 1.01157677e+00 7.23513782e-01
-1.23027229e+00 -1.23327747e-01 -4.91216809e-01 5.91347337e-01
-1.09205413e+00 -3.96563083e-01 6.93855882e-02 -2.55899787e-01
1.37661147e+00 6.39391124e-01 8.24646130e-02 6.38937891e-01
6.77680736e-03 5.62113225e-01 9.52530265e-01 -5.89823306e-01
1.56090125e-01 4.06216502e-01 6.18682325e-01 4.48139459e-01
2.32508779e-01 -2.08463892e-01 -1.79786399e-01 -5.05484104e-01
-2.42558688e-01 -6.40741825e-01 -4.72623527e-01 1.72496866e-02
-8.11371088e-01 9.19082165e-01 4.65749443e-01 -1.51968688e-01
-5.33106685e-01 -1.87829331e-01 2.03472182e-01 6.76600456e-01
-5.92092536e-02 1.29424453e+00 -7.89582133e-01 7.05656037e-02
-5.97344041e-01 9.57412839e-01 1.33904111e+00 6.97206974e-01
1.03323185e+00 -6.48987830e-01 -5.33851802e-01 5.45088768e-01
1.28035873e-01 4.91031736e-01 5.85066259e-01 -1.37326956e+00
7.57419527e-01 9.83378291e-01 6.98004365e-01 -1.04538941e+00
-3.55383307e-01 -3.57606739e-01 -9.20822769e-02 -3.14689904e-01
6.58558726e-01 -2.96377718e-01 -7.25141108e-01 1.74673200e+00
4.90738600e-01 -1.44317403e-01 5.62223256e-01 6.69346511e-01
1.42744994e+00 3.60714346e-01 5.33155262e-01 1.20749958e-01
1.44925916e+00 -1.06769025e+00 -6.91298664e-01 -3.76437426e-01
1.10266900e+00 -8.28002334e-01 1.54457510e+00 9.20129269e-02
-1.12535381e+00 -5.07665396e-01 -7.50043809e-01 -3.00545752e-01
-3.39586943e-01 -1.93929840e-02 3.53606731e-01 6.72966957e-01
-8.39555562e-01 4.65193301e-01 -3.22490901e-01 -1.14350595e-01
-8.96691084e-02 2.28671014e-01 -3.51741523e-01 -1.72627449e-01
-1.90012801e+00 1.64248621e+00 5.70603311e-01 -4.61822391e-01
-3.74662220e-01 -9.19590056e-01 -1.02197886e+00 7.34020993e-02
4.09831047e-01 -1.09576142e+00 1.78218746e+00 -6.76095307e-01
-8.66355360e-01 1.04283249e+00 -4.51540530e-01 -9.26109016e-01
3.85321915e-01 -4.80372101e-01 -7.56058991e-01 5.09706102e-02
4.93393004e-01 7.43032455e-01 1.72705039e-01 -1.12490821e+00
-5.19666314e-01 -2.12416083e-01 8.48538637e-01 2.21326858e-01
-1.11937076e-01 1.83318809e-01 -5.96276447e-02 -7.22853467e-02
-1.00608669e-01 -5.95185339e-01 -1.60334632e-01 -5.42879462e-01
-3.09651852e-01 -6.25268161e-01 7.75252283e-02 -1.02234089e+00
1.63092244e+00 -1.35318780e+00 -2.79180229e-01 7.59003088e-02
2.70105839e-01 2.78825492e-01 -9.94077027e-02 4.68722314e-01
-1.31318551e-02 3.70223552e-01 -9.79851708e-02 -4.20303689e-03
2.57846564e-01 2.24560484e-01 -5.02314448e-01 -4.51608717e-01
3.48296046e-01 1.31711793e+00 -8.50210011e-01 -7.82414556e-01
-5.53975523e-01 -2.34835342e-01 -8.25746417e-01 1.45398438e-01
-7.57164180e-01 4.93051521e-02 -1.77396536e-01 5.23856163e-01
4.40434039e-01 -4.65181649e-01 -1.68961897e-01 -5.33858120e-01
5.31014562e-01 8.33468318e-01 -1.00695825e+00 1.22260964e+00
-4.71099436e-01 2.28941500e-01 -6.41074836e-01 -2.15259358e-01
6.96118772e-01 4.89868909e-01 -3.81828666e-01 -1.06976020e+00
-2.85637826e-01 5.64129055e-01 1.27548143e-01 -7.88438320e-01
9.53271508e-01 -3.00998539e-01 -1.17236093e-01 3.96465927e-01
2.71226078e-01 -3.54831845e-01 5.36117554e-01 4.41609383e-01
1.27001917e+00 4.93776575e-02 3.16164404e-01 -3.67472954e-02
6.45907044e-01 5.69360375e-01 5.66113055e-01 8.84124398e-01
-4.33697626e-02 3.96487027e-01 2.99938291e-01 -4.38781470e-01
-9.35262501e-01 -1.18266082e+00 -4.13963795e-02 9.07605529e-01
1.91085890e-01 -9.48338807e-01 -7.41397083e-01 -1.26621294e+00
-9.78964567e-02 1.37601411e+00 -6.06261432e-01 -2.83372670e-01
-6.97758973e-01 -3.46730888e-01 9.28388953e-01 6.75393581e-01
3.63663197e-01 -9.67309475e-01 -7.20163524e-01 1.04417130e-01
-9.19121504e-01 -9.70393300e-01 -2.44145077e-02 -6.59533292e-02
-4.41833764e-01 -1.36594927e+00 -3.79286677e-01 -4.67242092e-01
1.72922611e-01 -3.14519018e-01 2.01713276e+00 4.03515875e-01
4.15458888e-01 4.69353735e-01 -4.14886802e-01 -5.43202579e-01
-6.60494685e-01 5.17683513e-02 -2.56566077e-01 -6.34885311e-01
8.95214558e-01 -1.33462653e-01 -5.39315879e-01 5.36500871e-01
-7.05254495e-01 -2.69742727e-01 1.01092413e-01 7.46341109e-01
3.83202463e-01 -3.50641936e-01 9.51822937e-01 -1.22783887e+00
1.00111532e+00 -5.67710280e-01 -9.41523239e-02 8.86350274e-01
-7.45275855e-01 3.74042809e-01 5.57406723e-01 2.83891652e-02
-1.37553918e+00 -4.86719102e-01 -5.83898008e-01 2.06151292e-01
-6.84862360e-02 8.54965806e-01 -1.32268474e-01 1.18216135e-01
1.56967592e+00 -5.23661852e-01 -6.02275670e-01 -2.50918388e-01
6.84886456e-01 3.08237642e-01 6.38153434e-01 -8.04158092e-01
6.62333369e-01 4.17913869e-02 -3.69652361e-01 -1.20562010e-01
-1.43756580e+00 -4.77422714e-01 -3.40178072e-01 8.99029076e-02
7.22279608e-01 -5.23815691e-01 -5.67830980e-01 -3.59266996e-01
-1.39647698e+00 -1.63428962e-01 -5.90415120e-01 3.91967475e-01
-4.82443124e-01 5.25612295e-01 -4.75371689e-01 -4.63269055e-01
-5.70688307e-01 -7.22992420e-01 6.75492227e-01 3.89522702e-01
-1.03010881e+00 -1.08365512e+00 5.63216746e-01 6.73087478e-01
3.62088799e-01 9.52316225e-02 1.18421817e+00 -1.17451787e+00
-3.22739105e-03 -3.52509618e-01 6.96406793e-03 1.84958011e-01
-2.47843429e-01 -1.33859143e-01 -1.05461001e+00 2.69661039e-01
-8.08097199e-02 -9.35901999e-01 7.73348033e-01 -2.49152452e-01
3.59891772e-01 -6.06761813e-01 -3.26465331e-02 -2.80413091e-01
1.26193559e+00 -1.39202088e-01 9.23885703e-01 4.91658926e-01
1.75066605e-01 8.85345280e-01 7.95306742e-01 -2.91044235e-01
8.62085104e-01 6.43688917e-01 3.46344650e-01 2.87723541e-01
-1.66493505e-01 -4.77700680e-01 -9.27013755e-02 4.27792460e-01
3.42739038e-02 -3.56778055e-01 -1.28941977e+00 9.16649401e-01
-1.71548593e+00 -1.00976610e+00 -5.81500769e-01 2.21122456e+00
1.39247358e+00 3.05399418e-01 -5.38040623e-02 -6.01449199e-02
2.81581432e-01 -5.37380517e-01 -3.19973201e-01 -3.07106048e-01
-3.34025621e-01 8.01391363e-01 1.65206213e-02 7.25885093e-01
-9.02529001e-01 8.37211311e-01 6.27301407e+00 7.00135648e-01
-3.18485290e-01 2.75027096e-01 2.58994222e-01 4.53777283e-01
-9.62588847e-01 4.20374840e-01 -8.47782612e-01 2.34986514e-01
1.04312634e+00 -3.18212241e-01 -2.69792169e-01 7.07423270e-01
-4.35150415e-01 -1.63694352e-01 -1.17487228e+00 3.44978094e-01
2.48806830e-02 -1.50246704e+00 3.26692432e-01 -6.34530723e-01
7.60617018e-01 -1.49583623e-01 -2.85301328e-01 8.00289690e-01
7.76089191e-01 -1.17647254e+00 5.91495097e-01 7.53314972e-01
5.38770318e-01 -4.95744824e-01 1.41255021e+00 4.07533437e-01
-7.90671766e-01 -8.46238434e-03 -2.97434151e-01 -9.52792689e-02
4.43339020e-01 2.59027958e-01 -9.07184601e-01 9.11342144e-01
6.67901635e-01 7.73968101e-02 -1.26885223e+00 1.21988678e+00
-7.54620135e-01 7.81951487e-01 -3.04015845e-01 -9.16722789e-03
1.85758188e-01 3.89267087e-01 2.54187912e-01 1.09738004e+00
5.41754626e-02 2.05159843e-01 -5.93655184e-02 8.39436412e-01
-1.70391217e-01 6.12245612e-02 -2.17385665e-01 3.05261225e-01
6.83841527e-01 9.40640450e-01 2.60394942e-02 -4.89318341e-01
-4.61914331e-01 6.59065783e-01 4.17141378e-01 2.43061483e-01
-6.43349946e-01 -4.51514512e-01 1.20916918e-01 1.46786287e-01
-2.86092758e-01 4.74175960e-01 -2.63784200e-01 -1.18646622e+00
3.14952791e-01 -1.09598732e+00 1.02187276e+00 -9.60962713e-01
-1.50572431e+00 7.04499185e-01 -9.47360322e-02 -9.80596483e-01
-6.58989966e-01 -3.73456091e-01 -6.13726199e-01 1.08526075e+00
-1.48719752e+00 -1.00563979e+00 -3.86518300e-01 4.53155458e-01
-8.16650409e-03 3.56851131e-01 1.20061922e+00 5.55852711e-01
6.97088614e-02 8.77938509e-01 -6.43285215e-01 2.12521136e-01
1.07182598e+00 -1.56359541e+00 4.95042294e-01 6.44075155e-01
1.94806039e-01 1.00871515e+00 1.09723711e+00 -8.26153696e-01
-5.81555068e-01 -7.74706483e-01 1.51973557e+00 -1.26196325e+00
6.16418719e-01 4.90902305e-01 -1.31476808e+00 8.52222085e-01
5.43991745e-01 -5.69347382e-01 1.04199398e+00 2.94018447e-01
-5.90162635e-01 3.22703838e-01 -1.28318298e+00 5.81282079e-01
5.37773669e-01 -4.76008147e-01 -1.42921209e+00 2.68745840e-01
1.18501818e+00 -4.53531146e-01 -1.02268493e+00 8.07340562e-01
4.01544720e-01 -7.34578013e-01 8.93906176e-01 -1.40028596e+00
6.88232362e-01 -5.51774740e-01 -2.92482525e-01 -1.34936190e+00
1.45645559e-01 -2.33611822e-01 -4.05126840e-01 1.10599887e+00
1.23541176e+00 -4.26550806e-01 8.41289043e-01 1.24859869e+00
-1.62280470e-01 -7.90323555e-01 -7.96022415e-01 -6.28849208e-01
3.54165465e-01 -2.89440930e-01 8.70419085e-01 1.09722626e+00
1.42918825e-01 8.58508468e-01 1.58110410e-01 1.00319631e-01
3.16693395e-01 1.05507702e-01 8.87711287e-01 -1.15700257e+00
-3.97147954e-01 -9.97712314e-02 -2.33538508e-01 -1.10612178e+00
3.67770195e-02 -8.28392327e-01 -2.60950983e-01 -1.90013039e+00
1.60937384e-01 -4.69019979e-01 3.22181821e-01 5.51944554e-01
-7.01835811e-01 2.28386626e-01 -6.67030737e-02 3.98095325e-02
-1.00714743e+00 4.62225497e-01 9.28409278e-01 -5.93424179e-02
1.38463646e-01 2.11878628e-01 -7.05498040e-01 8.71656716e-01
6.65844738e-01 -4.34583724e-01 -3.94174874e-01 -6.05923474e-01
1.19340754e+00 2.14625262e-02 5.66085279e-01 -8.41928065e-01
4.84945446e-01 -2.49114316e-02 4.55371663e-02 -4.48556513e-01
2.19837010e-01 -6.81902766e-01 -9.46587175e-02 2.68979073e-01
-7.16078579e-01 3.65172684e-01 2.04694688e-01 2.87817538e-01
-5.16376555e-01 -8.90008450e-01 3.87268960e-01 -3.21704030e-01
-6.88091278e-01 -2.97885716e-01 2.21089959e-01 8.82253766e-01
5.67433298e-01 -2.81151623e-01 -8.02698672e-01 -6.56577826e-01
-8.45371068e-01 6.72930241e-01 2.66925395e-01 1.21290676e-01
5.53147495e-01 -1.33620584e+00 -1.02626324e+00 -3.96693885e-01
5.00477791e-01 -1.15893200e-01 2.34197065e-01 6.93694949e-01
-5.27085960e-01 5.04307151e-01 1.04294248e-01 -3.74182582e-01
-1.22660446e+00 2.63280392e-01 6.70549393e-01 -6.93060517e-01
6.68879747e-02 9.88201201e-01 -1.52873054e-01 -1.24458396e+00
2.50521898e-02 -1.84198290e-01 -6.23775423e-01 -1.23678125e-01
6.85897589e-01 3.88174862e-01 5.05218983e-01 -3.81916791e-01
-1.06838100e-01 2.23397031e-01 -6.89127818e-02 -2.79871374e-01
7.83570349e-01 -2.95827929e-02 -1.71621084e-01 1.47946417e-01
6.89013124e-01 4.47994083e-01 -4.15695935e-01 -6.18291438e-01
5.42917311e-01 -4.85412151e-01 -7.41842151e-01 -1.40134192e+00
-2.07083240e-01 8.21043253e-01 1.65806159e-01 3.14394504e-01
8.60511959e-01 2.38573566e-01 8.99148285e-01 8.61422718e-01
4.12564605e-01 -8.61592829e-01 1.81192905e-01 7.84909487e-01
8.93593192e-01 -1.17293346e+00 -1.91839188e-01 -3.66397262e-01
-8.85757625e-01 8.35069656e-01 1.10649312e+00 3.41401845e-01
1.16666779e-01 -2.00796142e-01 1.29775614e-01 -4.93079305e-01
-8.54589045e-01 -1.03045374e-01 6.71594203e-01 4.63491440e-01
5.27971923e-01 -2.98444722e-02 -5.95561743e-01 1.12705123e+00
-7.06938148e-01 -1.25064448e-01 4.47010130e-01 7.67874360e-01
-4.96453136e-01 -9.42703724e-01 -2.11614206e-01 6.34649456e-01
-6.37246251e-01 -4.14882749e-01 -7.94025481e-01 6.76493347e-01
-9.99937207e-03 1.25158310e+00 -3.80552232e-01 -4.50281322e-01
7.70108819e-01 3.70647758e-01 4.96893376e-01 -8.22515786e-01
-1.00559580e+00 -1.07372892e+00 8.44988346e-01 -3.60138953e-01
-3.23585331e-01 -1.03729323e-01 -1.40334761e+00 -3.59333307e-01
-7.02763200e-01 9.28429723e-01 7.22260848e-02 1.18915391e+00
3.76230717e-01 9.94751528e-02 -7.07707927e-02 3.99690032e-01
-8.75687838e-01 -1.01772058e+00 3.45595419e-01 7.56801426e-01
9.96331964e-03 -4.49759364e-01 -2.50500321e-01 6.35534674e-02] | [11.009427070617676, 7.946611404418945] |
a9965368-3196-4200-8c27-c4b450feb06a | deepmedix-a-deep-learning-driven-resource | 2307.00324 | null | https://arxiv.org/abs/2307.00324v1 | https://arxiv.org/pdf/2307.00324v1.pdf | DeepMediX: A Deep Learning-Driven Resource-Efficient Medical Diagnosis Across the Spectrum | In the rapidly evolving landscape of medical imaging diagnostics, achieving high accuracy while preserving computational efficiency remains a formidable challenge. This work presents \texttt{DeepMediX}, a groundbreaking, resource-efficient model that significantly addresses this challenge. Built on top of the MobileNetV2 architecture, DeepMediX excels in classifying brain MRI scans and skin cancer images, with superior performance demonstrated on both binary and multiclass skin cancer datasets. It provides a solution to labor-intensive manual processes, the need for large datasets, and complexities related to image properties. DeepMediX's design also includes the concept of Federated Learning, enabling a collaborative learning approach without compromising data privacy. This approach allows diverse healthcare institutions to benefit from shared learning experiences without the necessity of direct data access, enhancing the model's predictive power while preserving the privacy and integrity of sensitive patient data. Its low computational footprint makes DeepMediX suitable for deployment on handheld devices, offering potential for real-time diagnostic support. Through rigorous testing on standard datasets, including the ISIC2018 for dermatological research, DeepMediX demonstrates exceptional diagnostic capabilities, matching the performance of existing models on almost all tasks and even outperforming them in some cases. The findings of this study underline significant implications for the development and deployment of AI-based tools in medical imaging and their integration into point-of-care settings. The source code and models generated would be released at https://github.com/kishorebabun/DeepMediX. | ['Balasubramanian Raman', 'Uppala Vivek Narayan', 'Pradeep Singh', 'Kishore Babu Nampalle'] | 2023-07-01 | null | null | null | null | ['medical-diagnosis'] | ['medical'] | [ 3.22493136e-01 2.13735804e-01 -4.63947505e-01 -5.23265481e-01
-7.74360538e-01 -5.64223766e-01 2.37959310e-01 4.33356225e-01
-6.20926321e-01 5.32013535e-01 1.41688600e-01 -7.20394135e-01
-5.64626336e-01 -4.53940421e-01 -3.22552860e-01 -6.36286557e-01
-2.52337813e-01 4.67840105e-01 -2.22992837e-01 3.68219078e-01
-1.41259000e-01 4.08202350e-01 -1.22029257e+00 4.07904714e-01
7.13279545e-01 1.23643637e+00 -6.68296590e-02 5.08502364e-01
2.28056505e-01 9.42975461e-01 -3.02325994e-01 -7.06642687e-01
3.85273844e-01 7.79779032e-02 -8.64658058e-01 -1.90647230e-01
4.01890934e-01 -6.66668594e-01 -4.06020910e-01 7.63633788e-01
8.79183173e-01 -3.40549499e-01 1.35421321e-01 -1.22722363e+00
-4.16987181e-01 2.93309957e-01 -1.42473236e-01 2.22630128e-01
2.89667517e-01 4.03317243e-01 7.65152097e-01 -2.32666746e-01
7.67341733e-01 3.07820320e-01 9.80727851e-01 6.74170613e-01
-1.04162133e+00 -7.18493879e-01 -4.58502859e-01 1.04664892e-01
-1.16979051e+00 -5.92397630e-01 5.96589036e-03 -3.98216099e-01
7.33804226e-01 8.44958067e-01 6.64233923e-01 1.39189398e+00
4.66886729e-01 7.95835972e-01 1.24195433e+00 -8.67314562e-02
3.14008832e-01 2.36390904e-01 1.17233716e-01 7.52378047e-01
4.03431803e-01 3.03700622e-02 -4.83911067e-01 -5.97863853e-01
4.47110206e-01 5.56049824e-01 -5.18967748e-01 -4.68582451e-01
-1.21669018e+00 6.33660257e-01 2.81197608e-01 3.28606606e-01
-5.46055198e-01 3.43339778e-02 5.58917761e-01 1.13027222e-01
4.06951666e-01 2.94557840e-01 -5.12217224e-01 -2.89517581e-01
-9.26429093e-01 8.69637355e-03 7.86274135e-01 6.79215372e-01
-3.14163081e-02 -5.08555710e-01 -1.84244528e-01 5.31290114e-01
5.67754880e-02 4.01680648e-01 5.57594955e-01 -9.40421522e-01
2.52856553e-01 4.56490576e-01 -9.61723626e-02 -8.92390549e-01
-6.17756248e-01 -6.98956132e-01 -9.04254377e-01 3.87411527e-02
2.73710012e-01 -3.13594878e-01 -6.85191929e-01 1.40210211e+00
4.01076406e-01 1.68381914e-01 -2.69539040e-02 6.74353719e-01
7.21606672e-01 -6.79859370e-02 5.04631519e-01 4.65774424e-02
1.49156368e+00 -4.97990996e-01 -8.05581868e-01 -2.81202071e-03
7.75345325e-01 -4.55551058e-01 8.56664062e-01 5.37445724e-01
-1.01918542e+00 1.22940756e-01 -8.54075253e-01 2.01184303e-02
-4.00652945e-01 -1.19581513e-01 1.16347277e+00 1.07588053e+00
-1.08769715e+00 6.10231936e-01 -1.15780485e+00 -4.06427860e-01
1.31196690e+00 6.78234696e-01 -5.99719584e-01 -4.96880293e-01
-8.68125498e-01 7.30025828e-01 2.26924673e-01 -2.88134247e-01
-7.84636199e-01 -1.25160933e+00 -5.76956153e-01 5.32175414e-02
3.63027900e-01 -1.01862097e+00 1.22980547e+00 -6.94135725e-01
-9.03998613e-01 9.68681395e-01 3.66626561e-01 -7.43164003e-01
8.57642114e-01 4.55983728e-03 -6.09142542e-01 4.44800884e-01
6.36374727e-02 5.52031577e-01 5.03909826e-01 -6.20471001e-01
-6.20147407e-01 -5.98634362e-01 -2.95606911e-01 -8.71862005e-03
-6.79153323e-01 -1.80806175e-01 -3.56181145e-01 -4.23539758e-01
-3.68847698e-01 -9.15076852e-01 -4.07948196e-01 3.37963492e-01
-2.58580238e-01 3.12760949e-01 7.80637205e-01 -8.59108508e-01
1.01008809e+00 -2.21090555e+00 -4.34273899e-01 2.52278984e-01
7.25738883e-01 5.82790017e-01 1.44600540e-01 1.96886167e-01
1.65344477e-02 2.37074614e-01 -2.49210238e-01 -2.32769623e-01
-1.92570388e-01 -5.41498326e-02 2.95210421e-01 5.39645493e-01
-1.54764071e-01 1.23315334e+00 -7.64151335e-01 -4.68264550e-01
3.87982905e-01 6.48948610e-01 -4.47723597e-01 -1.03718579e-01
2.28607565e-01 5.62235296e-01 -3.60153586e-01 9.26203072e-01
4.71394718e-01 -6.27539158e-01 4.21262860e-01 -9.31592211e-02
2.62821853e-01 -1.67392984e-01 -6.97742522e-01 1.82709944e+00
-3.62193853e-01 2.67570347e-01 3.63802165e-01 -6.01200104e-01
3.64419460e-01 5.65562308e-01 1.13702881e+00 -8.26117754e-01
4.14019883e-01 1.94705203e-01 -3.76923792e-02 -8.13186467e-01
1.01567730e-02 -4.62961057e-03 1.43494412e-01 4.07494009e-01
3.53491195e-02 2.49448866e-01 -4.02545601e-01 2.53491253e-01
1.62446773e+00 -2.80968815e-01 3.05987865e-01 -2.62837201e-01
1.38779968e-01 -2.99911052e-02 3.70571554e-01 7.57185161e-01
-4.98958379e-01 3.64073455e-01 1.27396081e-02 -5.12808144e-01
-5.91284513e-01 -1.00747705e+00 -5.02072275e-01 7.43893445e-01
-3.02579075e-01 -4.42925423e-01 -6.72861934e-01 -8.18949521e-01
2.23074555e-01 5.92603385e-01 -6.70096159e-01 -1.87611356e-01
1.17326146e-02 -8.13567996e-01 7.22518325e-01 3.31339508e-01
4.84212488e-01 -8.34799409e-01 -1.21870339e+00 8.61306861e-02
-2.81166852e-01 -9.31093812e-01 -2.89126396e-01 8.76003653e-02
-8.29398870e-01 -1.34660292e+00 -6.76203370e-01 -3.03032130e-01
4.13242757e-01 -4.31453846e-02 8.99755597e-01 6.20688163e-02
-1.08149302e+00 8.00792158e-01 -1.60550907e-01 -8.20681155e-01
-1.98634416e-01 1.77453905e-01 -2.53778696e-01 -1.91734478e-01
4.46942955e-01 -4.43705857e-01 -9.92414236e-01 1.81323364e-02
-1.03605580e+00 -1.38543665e-01 5.23860157e-01 7.20805407e-01
5.65942287e-01 -9.47467908e-02 6.20533347e-01 -1.28926671e+00
5.25290251e-01 -8.92318547e-01 -1.91244751e-01 2.79709339e-01
-1.04939628e+00 -5.04525542e-01 2.64625758e-01 -5.37190288e-02
-8.98120940e-01 1.00918569e-01 -1.38149723e-01 -3.72499764e-01
-1.83095604e-01 5.45697808e-01 1.42225489e-01 -2.95733988e-01
6.31531417e-01 -4.83424999e-02 6.25728726e-01 -2.96448559e-01
-7.73685202e-02 9.13058460e-01 5.28645217e-01 -4.60682064e-02
1.73037529e-01 6.30903363e-01 -2.38453201e-03 -5.79933941e-01
-5.31591356e-01 -3.97541642e-01 -1.28107637e-01 -1.03487320e-01
8.81513357e-01 -8.69811237e-01 -9.71543431e-01 2.68475711e-01
-3.20964634e-01 -1.81327537e-01 -4.54137921e-01 4.63275790e-01
-2.08832413e-01 2.30459809e-01 -5.30102789e-01 -5.27762473e-01
-8.21062803e-01 -1.01060808e+00 7.55546033e-01 8.48187410e-05
-5.41477442e-01 -1.14739382e+00 -1.67835265e-01 9.17251766e-01
7.04604089e-01 5.13168454e-01 9.01499927e-01 -1.03608680e+00
-5.19419491e-01 -6.85106516e-01 3.62201966e-02 2.87144989e-01
2.44542599e-01 -3.94599587e-01 -9.89619672e-01 -4.94940549e-01
3.72724570e-02 -3.83722842e-01 3.90605956e-01 3.77041608e-01
1.62844884e+00 -2.13705927e-01 -4.61476356e-01 9.35004652e-01
1.29807961e+00 2.06990242e-01 5.67641616e-01 3.33652854e-01
5.20674884e-01 3.93411964e-01 4.76454139e-01 5.58451295e-01
4.69549567e-01 1.28619850e-01 6.87458038e-01 -3.62353474e-01
2.95489114e-02 7.32420087e-02 -2.95372903e-01 4.29130614e-01
2.39394918e-01 -1.23079777e-01 -1.38908470e+00 5.43385386e-01
-1.61060727e+00 -8.14052582e-01 9.62307770e-03 2.34105158e+00
5.48634827e-01 -3.04094851e-01 -6.14925064e-02 -3.59192975e-02
2.90537000e-01 -1.96290329e-01 -8.38643670e-01 -2.73883462e-01
1.06574863e-01 5.20664096e-01 8.51721287e-01 -5.00575779e-03
-1.03738272e+00 3.73528481e-01 6.32061815e+00 6.53660655e-01
-1.33210325e+00 4.67287630e-01 1.07801700e+00 -5.33863604e-01
-2.60983944e-01 -5.85537016e-01 -1.08176820e-01 5.50306082e-01
1.15523124e+00 -4.88293841e-02 3.48756015e-01 8.91639590e-01
3.23325582e-02 -1.35800362e-01 -9.73504424e-01 1.19577003e+00
-3.06012277e-02 -1.74525630e+00 -3.22877973e-01 4.32777733e-01
4.06039357e-01 4.79283512e-01 4.02200043e-01 -1.27547439e-02
1.37252495e-01 -1.29080176e+00 2.28162140e-01 4.18816328e-01
1.28317904e+00 -6.56644225e-01 9.78262663e-01 1.87199101e-01
-3.78597826e-01 -3.67552727e-01 2.56959647e-01 2.45428309e-01
-1.95326224e-01 4.95620489e-01 -1.11515558e+00 7.67830908e-01
9.48183656e-01 5.72242260e-01 -6.36262715e-01 9.66709673e-01
5.63311160e-01 6.99084699e-01 -1.60266146e-01 2.64556170e-01
-2.95888307e-03 9.39185545e-02 2.22659558e-01 1.14813161e+00
2.59327829e-01 1.95807263e-01 2.54025590e-02 2.04330012e-01
-1.03445224e-01 3.30124229e-01 -8.08561683e-01 -1.99432299e-01
5.46664298e-01 1.52352703e+00 -6.33944094e-01 -1.25590071e-01
-3.03816527e-01 8.46138537e-01 -3.06190532e-02 -1.13016360e-01
-7.95657575e-01 -1.09376989e-01 7.17600584e-01 3.75607401e-01
-8.58238526e-03 3.13839406e-01 -6.37492359e-01 -9.81911242e-01
-3.69339623e-02 -1.29100645e+00 7.36761272e-01 -3.75849813e-01
-1.13852084e+00 6.11761928e-01 -3.43478292e-01 -1.09812164e+00
-2.06708476e-01 -4.87257451e-01 -1.50657833e-01 5.45590997e-01
-1.36911488e+00 -1.32698715e+00 -3.94895524e-01 1.01827717e+00
-2.58756995e-01 -3.79143894e-01 1.32924390e+00 6.15746439e-01
-4.56511974e-01 8.89480472e-01 2.05648676e-01 -2.10567433e-02
6.60308421e-01 -8.62405598e-01 -9.22653526e-02 4.36492234e-01
-1.72634006e-01 7.43727207e-01 1.93519890e-01 -5.54022849e-01
-1.57128298e+00 -1.16554785e+00 5.68897247e-01 -6.58325434e-01
4.00425732e-01 -2.25323021e-01 -5.76918781e-01 8.19181502e-01
1.21843278e-01 1.97691679e-01 1.51351881e+00 1.68881088e-01
-1.78758293e-01 -2.59852767e-01 -1.84538245e+00 4.21559244e-01
9.76173520e-01 -5.87188482e-01 -8.73564705e-02 6.22677922e-01
4.40453976e-01 -4.08600569e-01 -1.24753213e+00 3.76150668e-01
5.95046818e-01 -9.79395568e-01 9.05097723e-01 -6.54056132e-01
3.28441471e-01 3.40480775e-01 -8.02548751e-02 -8.77904415e-01
-2.86197871e-01 -7.56503046e-01 -1.24357395e-01 1.05181074e+00
2.36397326e-01 -9.53005195e-01 1.00206435e+00 1.21313632e+00
1.17829971e-01 -1.15966225e+00 -1.12967038e+00 -4.63398546e-01
-2.21809223e-01 -6.23145759e-01 6.89050019e-01 1.27748132e+00
-5.66819534e-02 -2.72000939e-01 -2.07094669e-01 1.17441460e-01
7.29827464e-01 -3.81829202e-01 4.16131377e-01 -1.13457596e+00
-3.40528995e-01 -4.34031412e-02 -6.34657264e-01 1.02747694e-01
-1.43223614e-01 -1.29788125e+00 -6.15751445e-01 -1.39134049e+00
2.85669714e-01 -7.87163675e-01 -5.28658032e-01 8.97022247e-01
8.97951424e-02 5.91438174e-01 2.85136968e-01 2.51797467e-01
-5.12822628e-01 -6.71515167e-02 1.06141305e+00 -1.07056014e-01
1.36151254e-01 1.08457811e-01 -1.08268917e+00 4.93490607e-01
9.68411028e-01 -4.14487720e-01 -6.64473712e-01 -4.62672293e-01
-1.93551108e-01 1.55630440e-01 5.97266734e-01 -1.24914575e+00
3.23063165e-01 1.09824881e-01 4.69347417e-01 -9.57340524e-02
2.76495904e-01 -1.35976517e+00 5.94476342e-01 8.31387460e-01
-2.35805973e-01 -7.06211478e-02 1.53083444e-01 3.88604254e-01
1.25406086e-01 1.56120986e-01 5.84684134e-01 -1.32634521e-01
-4.01754498e-01 6.74383342e-01 -2.75217116e-01 -4.77251299e-02
1.58017731e+00 -2.03893140e-01 -3.47061872e-01 -1.79954156e-01
-7.44232595e-01 2.79430181e-01 4.32531863e-01 2.74781168e-01
4.63506609e-01 -8.41710091e-01 -4.90863681e-01 2.59873748e-01
1.47460833e-01 -1.08462535e-01 7.48898745e-01 1.01634276e+00
-6.00924909e-01 4.15809393e-01 -3.20863694e-01 -5.79447746e-01
-1.56205237e+00 4.86765712e-01 2.82532036e-01 -2.25670829e-01
-7.20016479e-01 7.23124206e-01 -2.88338125e-01 -4.81349349e-01
4.16431785e-01 -7.14489967e-02 3.23924899e-01 -2.05765497e-02
7.59749413e-01 3.79141837e-01 4.49450850e-01 -8.51838887e-02
-5.55093765e-01 -2.70063668e-01 -4.47535098e-01 1.10967755e-01
1.62199235e+00 -3.54249626e-02 -5.61258942e-02 2.93952916e-02
1.27429616e+00 -9.98374075e-02 -9.80354071e-01 7.56214410e-02
-1.02118887e-01 -4.60177004e-01 2.30890259e-01 -1.54771411e+00
-1.28260541e+00 5.09805202e-01 1.13333654e+00 -2.94323623e-01
1.30918968e+00 -1.71937719e-01 7.72602618e-01 1.38002992e-01
5.45206547e-01 -7.44734228e-01 -2.68434852e-01 -1.25186458e-01
2.44072303e-01 -1.28343272e+00 1.71287388e-01 -2.02118337e-01
-8.52474749e-01 7.27544308e-01 2.61689782e-01 4.67617691e-01
8.43047142e-01 5.16218841e-01 3.77080947e-01 -4.09343660e-01
-6.35647058e-01 4.52566266e-01 -9.69683453e-02 8.58207047e-01
2.44302154e-01 3.23084205e-01 -1.96649805e-01 7.26506889e-01
1.42535632e-02 6.22536361e-01 1.93919182e-01 1.27451062e+00
3.10023278e-01 -1.18891180e+00 -1.14462569e-01 9.77629185e-01
-1.01002097e+00 -1.64018720e-01 -1.61961168e-01 6.70665979e-01
3.86643082e-01 7.06784844e-01 -1.63952559e-01 -3.29167068e-01
1.70994371e-01 1.93105489e-01 1.72877178e-01 -3.12532991e-01
-1.12683189e+00 -3.33895504e-01 1.50609901e-02 -8.78856003e-01
-9.69990343e-02 -8.80246520e-01 -1.02621102e+00 -5.29231966e-01
1.10863701e-01 -1.23424120e-01 1.04598641e+00 5.61543584e-01
1.07888210e+00 5.91661334e-01 3.02282333e-01 -1.63386419e-01
-6.40451610e-01 -3.48950952e-01 -4.82843786e-01 4.76382881e-01
3.45310062e-01 -1.10462822e-01 1.44762948e-01 -2.15031132e-01] | [6.185377597808838, 6.501511096954346] |
ad477963-1145-4bbe-9273-8147afdf03b6 | ariadne-s-thread-using-text-prompts-to | 2307.03942 | null | https://arxiv.org/abs/2307.03942v1 | https://arxiv.org/pdf/2307.03942v1.pdf | Ariadne's Thread:Using Text Prompts to Improve Segmentation of Infected Areas from Chest X-ray images | Segmentation of the infected areas of the lung is essential for quantifying the severity of lung disease like pulmonary infections. Existing medical image segmentation methods are almost uni-modal methods based on image. However, these image-only methods tend to produce inaccurate results unless trained with large amounts of annotated data. To overcome this challenge, we propose a language-driven segmentation method that uses text prompt to improve to the segmentation result. Experiments on the QaTa-COV19 dataset indicate that our method improves the Dice score by 6.09% at least compared to the uni-modal methods. Besides, our extended study reveals the flexibility of multi-modal methods in terms of the information granularity of text and demonstrates that multi-modal methods have a significant advantage over image-only methods in terms of the size of training data required. | ['Ming Wu', 'Kaixin Chen', 'Kongming Liang', 'Mengqiu Xu', 'Yi Zhong'] | 2023-07-08 | null | null | null | null | ['semantic-segmentation', 'image-segmentation', 'medical-image-segmentation'] | ['computer-vision', 'computer-vision', 'medical'] | [ 2.71175861e-01 -1.45587862e-01 -2.05933884e-01 -3.89986746e-02
-1.06467950e+00 -6.13503098e-01 4.12305772e-01 5.42078959e-03
-4.32313144e-01 4.32823777e-01 1.12175524e-01 -4.72700238e-01
-6.76812753e-02 -6.26607776e-01 -2.48205990e-01 -6.50025249e-01
3.88471365e-01 9.49536085e-01 5.99851191e-01 3.52474838e-01
6.08285479e-02 4.74177182e-01 -8.44526529e-01 3.44932973e-01
8.30219269e-01 5.59681773e-01 3.48144323e-01 8.90312135e-01
-4.14648086e-01 5.07616162e-01 -3.75568867e-01 -2.18464389e-01
3.25688213e-01 -5.00819266e-01 -9.15016770e-01 3.76528472e-01
4.99409497e-01 -6.36189342e-01 -1.42101809e-01 7.20365942e-01
5.71112812e-01 -3.40075165e-01 9.12007511e-01 -8.36469591e-01
-2.14903057e-01 1.18204840e-01 -7.28092790e-01 3.57366562e-01
-4.48762625e-03 2.58964926e-01 7.81372428e-01 -7.95808196e-01
6.99733615e-01 1.02860832e+00 8.70993972e-01 5.18868029e-01
-1.07529902e+00 -2.91792125e-01 -4.08977717e-01 -1.56737000e-01
-1.24930465e+00 1.18687555e-01 4.59830731e-01 -8.13054919e-01
6.07788622e-01 3.97756368e-01 4.99610633e-01 4.13385957e-01
8.35480988e-02 7.70118177e-01 1.25586498e+00 -3.16997111e-01
-1.69489697e-01 1.47489803e-02 -1.67800054e-01 1.13804734e+00
4.25403327e-01 -2.94499636e-01 -2.40598582e-02 -2.69843578e-01
8.66681218e-01 3.19214523e-01 -1.23184249e-01 -2.63290733e-01
-1.67734790e+00 8.32389772e-01 3.11873704e-01 6.14120603e-01
-4.10785884e-01 8.83079469e-02 5.26687443e-01 -3.80707860e-01
6.31384075e-01 3.08596283e-01 -2.51053035e-01 9.58371982e-02
-1.61760986e+00 -2.24194527e-01 4.97863173e-01 5.83150029e-01
4.71050978e-01 -3.66193354e-01 -5.88840187e-01 8.87254775e-01
3.46871793e-01 7.96328843e-01 4.70942199e-01 -1.10250175e+00
4.77772802e-01 7.07816184e-01 -2.48922497e-01 -7.74595857e-01
-4.39727783e-01 -8.01328048e-02 -9.24499154e-01 -1.41609520e-01
6.33326113e-01 -1.61791563e-01 -1.34574044e+00 1.13629019e+00
3.94086212e-01 -2.25238297e-02 -3.18182826e-01 7.02237070e-01
8.33069861e-01 5.02965748e-01 2.54049987e-01 -2.22796410e-01
1.48228383e+00 -1.19013131e+00 -7.29228854e-01 -2.93775052e-02
7.11461902e-01 -8.64366710e-01 1.09945166e+00 4.43619937e-02
-1.00538135e+00 -2.84101814e-01 -4.52532619e-01 1.77055657e-01
-4.45310384e-01 2.42469043e-01 2.32084200e-01 7.91353106e-01
-1.17803884e+00 2.59489805e-01 -7.94844568e-01 -6.12332702e-01
6.57909632e-01 3.32891226e-01 -1.78253382e-01 -1.12901375e-01
-8.46625209e-01 6.24643564e-01 4.47362512e-01 -3.36744219e-01
-5.39252698e-01 -8.85286510e-01 -4.10379142e-01 -2.92097569e-01
4.43975538e-01 -9.47467923e-01 1.09377313e+00 -5.02575099e-01
-8.92018676e-01 1.09360373e+00 -1.94013789e-01 -2.47110233e-01
9.44475234e-01 1.15731329e-01 2.78455675e-01 9.89234984e-01
3.19251984e-01 1.05878639e+00 6.13725066e-01 -1.37792492e+00
-6.60085976e-01 -2.17408538e-01 -4.02025700e-01 2.22095907e-01
-2.39661649e-01 -6.11228719e-02 -1.13959849e+00 -7.63832867e-01
-9.16711763e-02 -1.13327968e+00 -4.91704553e-01 2.79891014e-01
-5.37948906e-01 -1.51355699e-01 1.09022057e+00 -9.59492385e-01
1.27747631e+00 -1.93053567e+00 -3.75672311e-01 2.04407021e-01
6.46118701e-01 4.47114736e-01 7.51062259e-02 1.06926247e-01
3.10298651e-01 6.32032514e-01 -7.63751626e-01 -2.17697427e-01
-2.40626827e-01 3.14672559e-01 3.53064686e-01 3.98325592e-01
-7.44988397e-02 1.07232034e+00 -6.40246451e-01 -1.55064392e+00
5.63813031e-01 5.41221321e-01 -5.27078152e-01 2.56832510e-01
-2.49779493e-01 6.75855279e-01 -6.40460968e-01 5.91836572e-01
5.69957793e-01 -6.55269682e-01 1.78612903e-01 3.46631296e-02
4.48631085e-02 -2.15870634e-01 -6.01389706e-01 1.36647642e+00
-3.88760537e-01 3.41273159e-01 -6.51585013e-02 -7.28787184e-01
1.13838479e-01 7.18621910e-01 1.18523586e+00 -2.38073215e-01
1.04268499e-01 1.92323372e-01 -6.35347590e-02 -6.76714301e-01
1.49042591e-01 -2.92530656e-01 3.29615772e-01 6.33858919e-01
-4.53597367e-01 -6.65229142e-01 4.14890975e-01 3.21820557e-01
1.04067016e+00 -3.45674366e-01 1.85285375e-01 -1.16010487e-01
5.92527568e-01 3.77831340e-01 2.38007173e-01 8.06473792e-01
-5.55003822e-01 1.11897492e+00 4.21825707e-01 -4.66527604e-02
-1.19528484e+00 -9.71955121e-01 -2.42018864e-01 5.95191836e-01
-1.38759732e-01 -2.14950114e-01 -1.21603918e+00 -1.25273705e+00
-1.36526749e-01 2.66985059e-01 -5.27404964e-01 5.57098746e-01
-5.73515892e-01 -9.32694495e-01 6.73474193e-01 5.91688156e-01
5.00719070e-01 -9.78314519e-01 -6.92315340e-01 2.93610748e-02
-4.95975941e-01 -1.45188534e+00 -7.24982977e-01 -1.68335333e-01
-1.23269570e+00 -1.21262896e+00 -1.40917385e+00 -7.81356514e-01
1.02946842e+00 3.47934425e-01 1.03299212e+00 4.86873329e-01
-6.16295576e-01 7.28686213e-01 -1.22736491e-01 -4.87541892e-02
-4.26308155e-01 1.27938077e-01 -6.29644275e-01 -4.52919602e-01
1.07312277e-01 8.46462473e-02 -9.80543435e-01 3.33746731e-01
-1.23379719e+00 1.91146493e-01 7.27073789e-01 7.90751994e-01
9.76780176e-01 2.14139640e-01 4.30112839e-01 -1.41596007e+00
5.95513582e-01 -5.14675200e-01 -1.26049787e-01 3.87685657e-01
-8.23998928e-01 -2.12455750e-01 4.14256096e-01 -1.67599007e-01
-1.11812007e+00 2.21492633e-01 -8.13607313e-03 -3.08967501e-01
-2.28316396e-01 2.75231987e-01 4.59908038e-01 -3.03888004e-02
3.24956447e-01 1.73051029e-01 9.41635296e-02 -3.64784658e-01
3.99050623e-01 7.21907794e-01 4.01164651e-01 -2.91285038e-01
7.21471786e-01 7.69699097e-01 1.56021535e-01 -8.04968357e-01
-7.32535422e-01 -9.89645720e-01 -1.12333298e+00 -3.50623935e-01
1.33480310e+00 -6.31437600e-01 -1.02083817e-01 4.03964311e-01
-8.93621743e-01 -3.65016013e-01 -2.15334982e-01 5.38688183e-01
-4.79859561e-01 8.52563322e-01 -7.08398759e-01 -3.74482900e-01
-5.82247615e-01 -1.36582530e+00 1.22225106e+00 -9.34258476e-02
-1.56200200e-01 -1.38342249e+00 1.15880601e-01 1.04740584e+00
3.84285122e-01 2.48299167e-01 1.06778955e+00 -6.87408805e-01
-5.14657021e-01 -2.67722696e-01 -6.41457498e-01 2.88954109e-01
3.46468598e-01 1.27318650e-01 -5.66456974e-01 -5.89091703e-02
-1.39446959e-01 -7.26204440e-02 8.36211979e-01 7.12209105e-01
1.40098059e+00 -3.96486185e-02 -4.49227929e-01 1.50497690e-01
1.43210590e+00 1.65549442e-01 3.88424963e-01 1.50310742e-02
8.44605029e-01 5.09208620e-01 6.10236585e-01 2.97935218e-01
3.42767179e-01 3.62199247e-01 2.06103206e-01 -6.06410742e-01
-5.02928734e-01 7.35842735e-02 -1.30333185e-01 1.05552304e+00
1.75720334e-01 -3.90547246e-01 -1.30406964e+00 9.33668375e-01
-1.42739892e+00 -6.13783836e-01 -4.98446941e-01 1.84772885e+00
9.46235359e-01 -1.08329132e-01 2.92391479e-01 -1.13068327e-01
8.22219014e-01 1.17789589e-01 -2.63907462e-01 -1.60820082e-01
2.26847872e-01 1.39071062e-01 6.17498159e-01 3.68244767e-01
-1.11647403e+00 8.91179025e-01 7.42848778e+00 1.09067762e+00
-1.05657327e+00 3.37507576e-01 8.97701800e-01 2.02054963e-01
-2.32488737e-01 -3.38206053e-01 -4.15548921e-01 4.67929870e-01
7.48019516e-01 2.45545372e-01 6.46214858e-02 4.85737801e-01
4.09254253e-01 -5.56777418e-01 -6.63230836e-01 7.78024018e-01
1.61128208e-01 -1.14151001e+00 1.19578123e-01 1.85907498e-01
9.83226955e-01 1.54277936e-01 -6.80135377e-03 -2.92670131e-01
-8.29959288e-02 -9.77692783e-01 6.66515529e-02 4.37338233e-01
1.13270664e+00 -3.38089913e-01 8.68862808e-01 3.80232275e-01
-1.16073000e+00 5.96710205e-01 -4.53168787e-02 7.12297022e-01
2.54254460e-01 7.38996923e-01 -1.41891158e+00 3.06240469e-01
5.10826230e-01 3.54428262e-01 -9.27913904e-01 1.09634745e+00
-1.80684011e-02 9.51721549e-01 -4.61088866e-01 2.51614392e-01
3.33240539e-01 -2.97865063e-01 3.25052410e-01 1.56012154e+00
2.29290545e-01 7.32951164e-02 3.09964418e-01 6.90442443e-01
-1.71843544e-01 4.01357591e-01 -6.57380641e-01 -3.25227320e-01
4.04434279e-02 1.32338297e+00 -1.31207013e+00 -7.88750231e-01
-6.21983945e-01 9.72132087e-01 -2.81737864e-01 1.27399474e-01
-8.37606966e-01 5.05413339e-02 -2.90499032e-01 4.19492662e-01
3.18644226e-01 -1.64885774e-01 -6.62429035e-01 -6.27312303e-01
-2.25751072e-01 -7.04604328e-01 5.82911789e-01 -7.11886227e-01
-1.10559189e+00 4.76273328e-01 3.45230587e-02 -1.02672672e+00
-1.10977106e-01 -5.52324891e-01 -3.94480050e-01 5.31608105e-01
-1.41660821e+00 -1.21639287e+00 -3.63980144e-01 4.92063940e-01
7.18556523e-01 2.08239406e-01 5.49766839e-01 4.82245475e-01
-5.29972374e-01 2.91949958e-01 2.97161162e-01 3.59860986e-01
7.01691628e-01 -1.42661810e+00 -5.94106317e-02 7.17494845e-01
-8.18066299e-02 4.08465803e-01 2.06660688e-01 -8.64650965e-01
-8.35559785e-01 -1.18138289e+00 8.92615199e-01 -6.23965979e-01
4.42487419e-01 3.23517412e-01 -7.81579792e-01 4.59740520e-01
1.79126784e-01 9.15751830e-02 8.70186090e-01 -6.29473269e-01
1.12356946e-01 3.10911059e-01 -1.39236283e+00 4.60811585e-01
5.23148596e-01 -5.46033263e-01 -5.73130012e-01 5.70551097e-01
4.40453947e-01 -2.07860753e-01 -1.12102461e+00 5.39135695e-01
3.43394876e-01 -8.69256318e-01 1.12168181e+00 -1.11245580e-01
4.38207090e-01 -3.61135751e-01 1.34827718e-01 -6.45183265e-01
-5.70172966e-02 1.94537975e-02 2.91396797e-01 1.04144454e+00
4.99478847e-01 -4.85817850e-01 9.58213270e-01 4.73466367e-01
1.18285991e-01 -8.95638585e-01 -7.90778816e-01 -5.05272448e-01
3.16075683e-01 -3.26645970e-01 1.79025158e-02 8.58408272e-01
-4.61242408e-01 -8.96583274e-02 -2.01730989e-02 -2.67592788e-01
5.75351238e-01 9.54525471e-02 5.61338007e-01 -1.05310392e+00
1.21329196e-01 -5.00374436e-01 4.09652963e-02 -8.95083547e-01
1.86143238e-02 -8.53210330e-01 2.36902371e-01 -2.06497836e+00
9.11107361e-01 -5.06251037e-01 -9.84887928e-02 4.14591938e-01
-4.40148890e-01 8.02182078e-01 5.88731654e-02 4.73144531e-01
-7.01403022e-01 -1.01572447e-01 1.79072022e+00 -1.87749848e-01
6.49159104e-02 -5.84816188e-03 -3.66460621e-01 8.91430736e-01
9.53510284e-01 -5.47754586e-01 -4.06177640e-01 -2.44508117e-01
-2.65956432e-01 1.81632251e-01 2.72316694e-01 -7.37234473e-01
7.02126101e-02 -2.69939572e-01 2.57855445e-01 -1.05163479e+00
-5.46451882e-02 -9.65831101e-01 1.15486778e-01 9.52676713e-01
-1.37927920e-01 -6.79261461e-02 4.19357792e-02 6.09470725e-01
-2.35037968e-01 -4.05164152e-01 1.04134643e+00 -4.20628935e-01
-3.24436039e-01 4.17903841e-01 -6.99240744e-01 5.15100420e-01
1.04404235e+00 -3.46387774e-01 -6.59129769e-02 -2.33176008e-01
-4.87475336e-01 1.91645309e-01 6.00982308e-01 -1.61272645e-01
3.47432375e-01 -8.35688889e-01 -6.42367125e-01 -2.37342551e-01
-1.00701101e-01 3.30002666e-01 1.91238850e-01 1.44174540e+00
-1.21629882e+00 8.92574191e-01 2.27708995e-01 -9.92201567e-01
-1.58826661e+00 4.44882870e-01 4.53640431e-01 -8.09445322e-01
-6.79284990e-01 5.14101267e-01 5.32376170e-01 -3.31579447e-01
7.73320943e-02 -3.39209259e-01 -2.69531533e-02 1.61643848e-01
-4.49782237e-02 4.99899745e-01 -3.24937925e-02 -7.43832767e-01
-4.49908018e-01 8.94136190e-01 4.19322439e-02 -1.96046904e-01
9.19013083e-01 -3.24352205e-01 -2.81861097e-01 1.64850786e-01
1.12210250e+00 1.44175008e-01 -9.33439314e-01 -2.09471345e-01
9.43577066e-02 -4.50538367e-01 1.38498902e-01 -9.28263187e-01
-1.35534108e+00 1.04346359e+00 7.34808445e-01 1.90654978e-01
1.18755412e+00 1.01575337e-01 1.13113678e+00 -6.05142787e-02
-2.16631085e-01 -1.02064502e+00 3.70504826e-01 1.92479536e-01
4.32844579e-01 -1.52947450e+00 2.73812056e-01 -5.70199907e-01
-8.06938589e-01 1.00220799e+00 3.15348238e-01 3.00299764e-01
7.16460824e-01 1.84438765e-01 3.28249186e-01 -3.47818494e-01
-3.06697071e-01 -3.79988134e-01 7.15028584e-01 5.26703358e-01
5.66125989e-01 8.89468193e-02 -7.01745331e-01 -8.69256561e-04
2.97382891e-01 -3.27587947e-02 8.85389298e-02 8.26851428e-01
-4.93044436e-01 -9.55020845e-01 -5.78345954e-01 7.23345876e-01
-1.06793535e+00 -2.58222878e-01 -5.87325215e-01 8.77570033e-01
1.90320656e-01 8.57775331e-01 -1.64763525e-01 1.69968724e-01
-1.10599421e-01 9.37199071e-02 2.87756801e-01 -6.72057807e-01
-6.09389961e-01 4.43517119e-01 -2.70505816e-01 -2.28128418e-01
-7.66982079e-01 -5.41877866e-01 -1.66731763e+00 -2.32600763e-01
-4.20751840e-01 -1.58933480e-03 5.25336027e-01 9.68681037e-01
8.49270821e-02 5.68097711e-01 5.43731570e-01 -3.84287030e-01
-1.18251495e-01 -7.21613288e-01 -3.45600724e-01 5.52098215e-01
1.63973093e-01 -3.13475072e-01 -5.00468314e-01 4.77362335e-01] | [14.928387641906738, -2.1022121906280518] |
d3453e1c-7a72-4a1c-83b3-bdc4c4ed3c35 | general-board-game-playing-for-education-and | 1907.06508 | null | https://arxiv.org/abs/1907.06508v1 | https://arxiv.org/pdf/1907.06508v1.pdf | General Board Game Playing for Education and Research in Generic AI Game Learning | We present a new general board game (GBG) playing and learning framework. GBG defines the common interfaces for board games, game states and their AI agents. It allows one to run competitions of different agents on different games. It standardizes those parts of board game playing and learning that otherwise would be tedious and repetitive parts in coding. GBG is suitable for arbitrary 1-, 2-, ..., N-player board games. It makes a generic TD($\lambda$)-n-tuple agent for the first time available to arbitrary games. On various games, TD($\lambda$)-n-tuple is found to be superior to other generic agents like MCTS. GBG aims at the educational perspective, where it helps students to start faster in the area of game learning. GBG aims as well at the research perspective by collecting a growing set of games and AI agents to assess their strengths and generalization capabilities in meaningful competitions. Initial successful educational and research results are reported. | ['Wolfgang Konen'] | 2019-07-11 | null | null | null | null | ['board-games'] | ['playing-games'] | [-4.87195700e-01 2.15518758e-01 1.44231811e-01 1.74052790e-01
-6.04700327e-01 -8.01361322e-01 3.70460242e-01 1.43377051e-01
-3.72330874e-01 8.33744884e-01 -3.13777417e-01 -6.25220656e-01
-4.69327271e-01 -1.15456414e+00 -3.50914687e-01 -4.86807853e-01
-4.68867987e-01 8.42639625e-01 7.77558148e-01 -1.07018173e+00
1.82152018e-01 1.42811611e-01 -2.08392477e+00 1.03775278e-01
7.89721906e-01 5.93928635e-01 3.63472819e-01 8.54993761e-01
-6.19367771e-02 1.10852861e+00 -7.31039405e-01 -6.03862286e-01
5.50287366e-01 -4.73042309e-01 -9.04148102e-01 -3.57740194e-01
9.85343754e-02 -1.48602858e-01 5.09311818e-02 1.03042710e+00
5.62285662e-01 4.46994245e-01 2.54603088e-01 -1.40459156e+00
-1.39705101e-02 7.35798001e-01 -2.83484217e-02 2.22855777e-01
6.78805768e-01 2.59487212e-01 1.02218068e+00 2.68656258e-02
5.06533682e-01 9.38825250e-01 2.31959909e-01 7.35840380e-01
-5.18320560e-01 -8.57088029e-01 1.46210477e-01 2.63253897e-01
-1.14881611e+00 2.59796172e-01 3.81858617e-01 -4.01660830e-01
1.02087724e+00 6.11409426e-01 1.44532979e+00 5.65550268e-01
1.95839986e-01 6.77089870e-01 1.14808118e+00 -4.70020503e-01
5.73627830e-01 -2.65999794e-01 -3.07111200e-02 6.14568949e-01
2.69041300e-01 2.21574992e-01 -6.73835278e-01 1.43773392e-01
1.05536771e+00 -6.17445588e-01 1.64634600e-01 -3.81522804e-01
-4.80568558e-01 1.05765200e+00 -5.89051507e-02 2.75223374e-01
-1.65678173e-01 2.52714872e-01 2.76120245e-01 8.92525911e-01
-1.18684759e-02 9.33584869e-01 -3.12640697e-01 -1.10214996e+00
-4.70962733e-01 8.42019260e-01 8.80238116e-01 8.87958884e-01
5.46639562e-01 1.34906918e-01 1.61586776e-01 6.86456561e-01
2.48146355e-01 -4.00282480e-02 5.25622189e-01 -1.30582845e+00
1.51902795e-01 9.07489955e-01 -3.64572518e-02 -5.31482816e-01
-7.06580579e-01 -3.85999322e-01 -2.91451097e-01 1.09997487e+00
5.70763528e-01 -1.34317592e-01 -4.43027616e-01 1.74932921e+00
3.66226822e-01 1.43003345e-01 1.17070429e-01 5.80452442e-01
1.29530180e+00 4.08249259e-01 3.34248580e-02 -1.71075203e-03
1.63984299e+00 -9.32951152e-01 -1.45595536e-01 -2.72681087e-01
8.86665285e-01 -5.00263929e-01 1.15805984e+00 9.36749101e-01
-1.67929852e+00 -4.03176278e-01 -1.06140423e+00 4.22052860e-01
-5.85097611e-01 -5.90986431e-01 9.35594618e-01 1.49364567e+00
-1.43748379e+00 3.74484122e-01 -7.38728285e-01 -1.84753254e-01
-6.01465814e-02 9.41899598e-01 -2.11304665e-01 -7.26632476e-02
-1.35068023e+00 8.99739325e-01 5.77616811e-01 -6.17022693e-01
-8.95450473e-01 -5.82058966e-01 -7.83430398e-01 4.94637638e-02
7.02963412e-01 -6.92765653e-01 1.60686219e+00 -5.23678005e-01
-2.05537987e+00 1.05667257e+00 6.79760516e-01 -3.60449493e-01
5.52868903e-01 2.56155103e-01 -7.46741444e-02 -2.19097853e-01
2.45670423e-01 6.38508379e-01 5.18978573e-02 -7.01234102e-01
-1.02294087e+00 -1.66657329e-01 1.15377414e+00 1.01209295e+00
2.36110196e-01 1.53497577e-01 -4.26125348e-01 -3.15207660e-01
2.65053157e-02 -8.76087964e-01 -3.34003061e-01 -7.74053395e-01
2.79441625e-01 -5.40784061e-01 2.44403869e-01 1.44031197e-01
1.24651861e+00 -1.84505820e+00 7.87939504e-02 1.37135416e-01
4.95714456e-01 3.33598256e-01 -5.26360385e-02 7.32818067e-01
-7.46631101e-02 1.68597978e-02 6.21257424e-01 9.55974981e-02
3.97562712e-01 3.15309376e-01 4.43936348e-01 -7.31931701e-02
-6.23515308e-01 7.37707675e-01 -9.92234647e-01 -1.57599583e-01
1.93970278e-01 -2.73790210e-01 -9.69384134e-01 2.15528384e-01
-3.78595233e-01 1.90118372e-01 -4.58224505e-01 4.17742252e-01
3.18709373e-01 -4.07907851e-02 8.49137157e-02 8.12612116e-01
-3.14692527e-01 5.28989017e-01 -1.59814274e+00 1.64475536e+00
-2.63083458e-01 3.64575386e-01 4.18959297e-02 -8.34608614e-01
7.48317719e-01 4.06224877e-01 2.73714274e-01 -9.21183765e-01
1.16376549e-01 4.18924004e-01 6.15838170e-01 -2.90602684e-01
7.85961688e-01 -2.05817342e-01 -5.53069055e-01 7.38152623e-01
1.45750627e-01 -6.38407290e-01 9.13484871e-01 2.59557486e-01
1.19462013e+00 7.44647719e-03 4.54454333e-01 -5.04357815e-01
4.85533237e-01 -2.96434537e-02 2.64419913e-01 9.57165897e-01
-1.16963096e-01 9.25224125e-02 6.96951389e-01 -6.07259095e-01
-4.83779192e-01 -9.45860922e-01 4.40786541e-01 1.81691360e+00
3.94095540e-01 -8.61565113e-01 -8.21448565e-01 -1.02662675e-01
-6.18687451e-01 4.34658915e-01 -5.09849608e-01 -8.59846100e-02
-2.50367284e-01 -3.74381989e-01 5.11465907e-01 2.56496072e-01
7.34022617e-01 -1.42076778e+00 -1.08713770e+00 4.98267442e-01
6.52396381e-02 -5.83423316e-01 -3.85918953e-02 4.21903908e-01
-4.73841250e-01 -1.10125387e+00 -1.57509476e-01 -8.38654757e-01
-3.99852134e-02 1.71862423e-01 1.34610128e+00 4.48797554e-01
-1.74898282e-02 4.33396816e-01 -6.84399545e-01 -7.79284894e-01
-3.47596556e-01 2.87996143e-01 -2.68020518e-02 -9.63620424e-01
3.24983597e-01 -6.15550637e-01 -3.98351580e-01 2.74127901e-01
-7.64447570e-01 2.62015432e-01 1.18411668e-01 5.65156281e-01
3.42395194e-02 3.25640410e-01 1.91570252e-01 -9.28677499e-01
1.21160257e+00 1.20475991e-02 -8.04960966e-01 -1.82302259e-02
-2.98155934e-01 -2.26226479e-01 5.50161362e-01 -2.38800526e-01
-4.04979616e-01 -6.18732750e-01 -2.63633519e-01 3.44366342e-01
-1.43238708e-01 6.07929349e-01 -2.35537007e-01 -4.23655093e-01
7.53502011e-01 2.20481470e-01 -1.80357993e-01 6.67342618e-02
8.74709338e-02 2.34960198e-01 2.33859599e-01 -9.85897541e-01
4.66566235e-01 -3.06102425e-01 -1.16757803e-01 -6.24351382e-01
-2.11669743e-01 -3.83410305e-01 -2.62632333e-02 -5.36011815e-01
6.74776793e-01 -8.18111300e-01 -1.34827971e+00 7.69099116e-01
-5.21562278e-01 -8.19270134e-01 -5.65833569e-01 4.53518927e-01
-8.54450703e-01 -1.42519459e-01 -6.83629990e-01 -6.67000175e-01
9.26691294e-02 -1.49228823e+00 4.25544053e-01 8.29288185e-01
-3.03448230e-01 -9.91912961e-01 4.19746697e-01 5.15501678e-01
3.10749590e-01 -1.48801625e-01 8.53700161e-01 -8.03493381e-01
-5.03355265e-01 9.54274274e-03 4.36194688e-01 -2.07467511e-01
-2.59585381e-01 -1.58822417e-01 -3.67642283e-01 -2.81034708e-01
-2.24520177e-01 -5.26833475e-01 3.25320363e-02 4.21118110e-01
4.34662074e-01 -3.55630368e-02 1.48643360e-01 4.07055855e-01
1.34287798e+00 1.02836657e+00 8.57905030e-01 1.14195597e+00
9.28261951e-02 4.26378399e-01 4.90454435e-01 4.09320444e-01
6.64170027e-01 7.59312510e-01 7.11012065e-01 1.96538493e-01
1.90261245e-01 -1.21374568e-02 2.57766694e-01 9.10815358e-01
-5.70214570e-01 -2.65805155e-01 -1.01403844e+00 3.23223680e-01
-1.73512185e+00 -1.07268178e+00 1.59054026e-02 1.95663071e+00
6.57157421e-01 5.95154703e-01 7.96965539e-01 2.25015551e-01
3.79271060e-01 -5.45306653e-02 -1.07308462e-01 -1.03693223e+00
1.16901524e-01 1.05045354e+00 9.54441279e-02 7.14153528e-01
-6.64500415e-01 1.32167518e+00 6.46877003e+00 1.22229052e+00
-7.93357730e-01 -4.93300594e-02 3.71322304e-01 -1.67717814e-01
-2.22458467e-01 -7.96163902e-02 -7.04375923e-01 3.68363410e-01
7.35796034e-01 -6.75818622e-01 6.43778622e-01 6.49870932e-01
-3.83144915e-02 -3.26748133e-01 -7.22908378e-01 9.66267943e-01
-4.21730429e-01 -1.41361165e+00 -1.03562869e-01 2.55076498e-01
7.15687931e-01 -1.19347930e-01 1.03386045e-01 8.13198864e-01
1.27390444e+00 -1.39916885e+00 8.25516403e-01 -1.90997496e-01
4.71661508e-01 -1.10936606e+00 8.74800861e-01 5.77721715e-01
-1.32242560e+00 -4.04073261e-02 -2.12311283e-01 -1.04845083e+00
-4.68791246e-01 -5.27431905e-01 -4.44917619e-01 6.29631221e-01
6.65023267e-01 -2.73276065e-02 -4.77308750e-01 1.39104664e+00
-2.33161613e-01 3.42779815e-01 -3.62389535e-01 -5.23565352e-01
7.68319309e-01 -6.28459513e-01 2.78518975e-01 4.89840716e-01
3.41725260e-01 7.73052990e-01 3.29099506e-01 2.73839623e-01
4.18534130e-01 3.87252808e-01 -4.79720742e-01 2.36401170e-01
5.93134999e-01 9.94275808e-01 -1.09211814e+00 -5.44264242e-02
-1.63131237e-01 4.65012252e-01 1.57601312e-01 -3.18607651e-02
-7.03878880e-01 -3.16217363e-01 9.79974926e-01 2.17961594e-01
1.73283696e-01 -2.25169569e-01 -1.51194885e-01 -7.45532334e-01
-4.75883365e-01 -1.41229510e+00 6.22170210e-01 -7.86375046e-01
-6.80149615e-01 6.94078803e-01 2.04952389e-01 -1.17212212e+00
-6.52792215e-01 -6.58537149e-01 -7.48153925e-01 6.73669100e-01
-6.80664480e-01 -7.45247424e-01 -2.38760859e-01 9.75811124e-01
3.94618481e-01 -6.41349316e-01 1.03567469e+00 -1.93883106e-01
-3.19603026e-01 5.33703566e-01 -1.50612295e-01 -1.08056583e-01
1.90495122e-02 -1.39291334e+00 2.77228683e-01 5.21747172e-01
1.47966594e-01 3.78940314e-01 7.52138138e-01 -2.14435712e-01
-9.63265657e-01 -2.13583291e-01 3.16509783e-01 -6.05893284e-02
5.19495428e-01 -1.44516230e-01 -2.63981551e-01 3.98648858e-01
3.13586891e-01 -6.03334367e-01 9.91867423e-01 2.09233254e-01
4.14390527e-02 3.87584269e-02 -9.43382740e-01 1.03903186e+00
9.95751917e-01 -3.68696153e-01 -4.55789596e-01 3.60257700e-02
3.18322301e-01 -1.12692010e+00 -6.10947669e-01 -1.44222155e-01
5.38303137e-01 -1.56677377e+00 7.50248015e-01 -5.71766973e-01
3.83180141e-01 -1.79787949e-01 -2.84058172e-02 -1.40470183e+00
-4.01170433e-01 -1.07340109e+00 4.62940931e-01 5.43456197e-01
3.33937585e-01 -6.78619444e-01 1.37498355e+00 6.73272252e-01
-3.10119599e-01 -6.11478627e-01 -1.11660743e+00 -9.41068053e-01
5.29083788e-01 -8.94169569e-01 8.31005931e-01 7.52103508e-01
4.79989082e-01 1.36319399e-01 -9.69226584e-02 -2.56434023e-01
2.88857281e-01 -2.54790708e-02 9.33882713e-01 -1.49638939e+00
-8.22288275e-01 -1.02006435e+00 -9.33517218e-01 -9.37246561e-01
-3.42321903e-01 -6.25694871e-01 -2.91447073e-01 -1.53997481e+00
-1.80551812e-01 -5.01790226e-01 7.40166008e-02 3.62310976e-01
-9.72668640e-03 4.11269426e-01 4.60134029e-01 -2.95384049e-01
-9.86925066e-01 -1.23712897e-01 1.50051308e+00 1.35723680e-01
-3.67694557e-01 2.77798057e-01 -9.90800202e-01 7.13725626e-01
1.00608814e+00 -2.43484631e-01 -9.00206745e-01 -1.69228360e-01
8.00642371e-01 2.90669113e-01 -1.26442924e-01 -1.44351912e+00
5.40951729e-01 -5.47241092e-01 -3.54149997e-01 9.44717228e-03
3.37945342e-01 -4.43942308e-01 5.18416464e-01 7.17100799e-01
-4.53844219e-02 3.18826735e-01 4.05094326e-01 -2.29968682e-01
-2.65802085e-01 -7.82385528e-01 4.53856915e-01 -7.44732141e-01
-9.35854375e-01 1.07507512e-01 -9.48696077e-01 4.62993115e-01
1.54346621e+00 -9.92008030e-01 -2.51139104e-01 -9.27136898e-01
-9.14057434e-01 2.73919016e-01 3.61023217e-01 1.06122479e-01
1.75669298e-01 -1.06128454e+00 -5.74627280e-01 2.28247762e-01
6.59856154e-03 1.50287360e-01 4.10521269e-01 2.08019823e-01
-1.16309214e+00 2.91973889e-01 -8.56385171e-01 -1.26344219e-01
-1.60778964e+00 5.87509237e-02 5.44999719e-01 -8.88172209e-01
-3.96070868e-01 1.36301196e+00 1.96593091e-01 -5.04973292e-01
1.51575491e-01 -7.72354081e-02 -4.21376556e-01 -4.97558340e-03
5.84787846e-01 3.26702058e-01 2.61370450e-01 -2.27283746e-01
-9.40238461e-02 3.35462004e-01 1.64655061e-03 -3.78849328e-01
1.34989583e+00 2.28485778e-01 1.04831666e-01 3.90295982e-01
2.56336659e-01 -2.14848183e-02 -8.57057810e-01 3.69811088e-01
-2.29597956e-01 -2.38942116e-01 -1.97683752e-01 -7.57837594e-01
-9.50525045e-01 4.59218889e-01 4.03962612e-01 8.10223281e-01
1.06294286e+00 -9.93637592e-02 2.08253011e-01 3.72192748e-02
1.22618520e+00 -9.89336669e-01 9.88509804e-02 9.84800637e-01
4.36802983e-01 -5.45805812e-01 -1.68227360e-01 -3.47874224e-01
-5.81477344e-01 1.20662498e+00 9.99839544e-01 -1.55989051e-01
3.85348976e-01 3.93402457e-01 1.37950763e-01 -5.79600573e-01
-1.08170319e+00 -5.43420732e-01 -6.51263371e-02 1.03716326e+00
4.43794459e-01 3.36799651e-01 -6.14772797e-01 7.54894495e-01
-1.23020244e+00 9.14544389e-02 1.00013459e+00 1.03748977e+00
-5.79271317e-01 -1.71629059e+00 -3.60687256e-01 2.88125932e-01
-4.93296564e-01 -1.63556919e-01 -3.55241030e-01 1.22457683e+00
4.52089816e-01 9.43354130e-01 -1.10605443e-02 -5.92879474e-01
3.02269042e-01 -1.94851518e-01 8.78276944e-01 -8.34011257e-01
-1.29405332e+00 -2.65792429e-01 1.50008336e-01 -4.62172449e-01
-2.64553100e-01 -4.55661744e-01 -1.06322205e+00 -9.45966661e-01
-8.16409215e-02 7.39341617e-01 2.12602496e-01 8.10514808e-01
-2.38000557e-01 4.83181208e-01 -6.09246641e-02 -5.22704542e-01
-2.09566414e-01 -6.72746956e-01 -1.18451166e+00 4.90965731e-02
-4.78236377e-01 -8.01838696e-01 4.07832451e-02 -5.63165307e-01] | [3.4020113945007324, 1.4697850942611694] |
8022ecc5-dba4-42a7-9124-5fabba89eb95 | uit-e10dot3-at-semeval-2021-task-5-toxic | 2104.07376 | null | https://arxiv.org/abs/2104.07376v1 | https://arxiv.org/pdf/2104.07376v1.pdf | UIT-E10dot3 at SemEval-2021 Task 5: Toxic Spans Detection with Named Entity Recognition and Question-Answering Approaches | The increment of toxic comments on online space is causing tremendous effects on other vulnerable users. For this reason, considerable efforts are made to deal with this, and SemEval-2021 Task 5: Toxic Spans Detection is one of those. This task asks competitors to extract spans that have toxicity from the given texts, and we have done several analyses to understand its structure before doing experiments. We solve this task by two approaches, Named Entity Recognition with spaCy library and Question-Answering with RoBERTa combining with ToxicBERT, and the former gains the highest F1-score of 66.99%. | ['Kiet Van Nguyen', 'Luan Thanh Nguyen', 'Phu Gia Hoang'] | 2021-04-15 | null | https://aclanthology.org/2021.semeval-1.125 | https://aclanthology.org/2021.semeval-1.125.pdf | semeval-2021 | ['toxic-spans-detection'] | ['natural-language-processing'] | [-6.34856284e-01 -6.98440000e-02 -1.29597768e-01 -5.64130656e-02
-9.41307306e-01 -7.27333307e-01 2.37971857e-01 9.08077136e-02
-5.10155916e-01 1.01449525e+00 6.57257557e-01 -3.65992844e-01
1.64699644e-01 -7.51157522e-01 -3.40416610e-01 -1.27036711e-02
-1.16470516e-01 1.25355303e-01 7.52096713e-01 -4.37969387e-01
5.27806044e-01 2.49095231e-01 -8.29621196e-01 5.86608768e-01
1.29221952e+00 7.40977466e-01 -4.24534410e-01 5.68352938e-01
-5.53469181e-01 9.21174347e-01 -1.11155820e+00 -1.20824146e+00
6.47698641e-02 -2.50378966e-01 -1.18656194e+00 -4.95972097e-01
8.06256458e-02 -1.45989940e-01 -6.24469817e-01 1.12733889e+00
8.27091992e-01 3.74663770e-02 5.84926069e-01 -1.18913746e+00
-8.59600902e-01 1.06272840e+00 -5.92817366e-01 7.19078362e-01
6.44897699e-01 2.54236292e-02 1.17970204e+00 -8.72767627e-01
5.37208200e-01 1.31188107e+00 8.14037621e-01 3.46634328e-01
-7.12573469e-01 -6.36609435e-01 -1.22382596e-01 4.66507554e-01
-1.61235356e+00 -1.63484886e-01 4.05380875e-01 -3.40862066e-01
9.99078572e-01 6.61702693e-01 2.72990257e-01 1.42971337e+00
-1.32043123e-01 8.70331228e-01 9.00636137e-01 -8.78673105e-04
1.50276288e-01 2.35011771e-01 6.61333501e-01 2.21578315e-01
3.75181019e-01 -4.27746236e-01 -5.46181440e-01 -5.24682939e-01
-1.76329195e-01 -1.17778629e-01 -2.91199535e-01 7.72945821e-01
-6.01645291e-01 8.17459643e-01 5.29643476e-01 4.93182659e-01
-1.12778664e-01 -1.90368131e-01 6.85605288e-01 1.27896547e-01
6.88673437e-01 8.10660064e-01 -6.24495566e-01 -3.40495318e-01
-9.19964015e-01 3.60405892e-01 1.27155769e+00 1.02977622e+00
2.70519644e-01 -5.25249600e-01 -5.53709626e-01 7.67584980e-01
2.35664591e-01 4.18877900e-01 4.56838638e-01 -3.76210302e-01
1.00802195e+00 8.92492175e-01 1.83122352e-01 -9.23336864e-01
-5.48848391e-01 -4.83094454e-01 -4.17662024e-01 -4.90283161e-01
5.27691007e-01 -6.89325094e-01 -5.14869452e-01 1.36898923e+00
2.84010679e-01 -1.20352544e-02 -3.74356300e-01 5.50985694e-01
7.94586957e-01 7.32861340e-01 4.43580240e-01 -9.49439555e-02
1.55838954e+00 -7.81857848e-01 -8.99118006e-01 6.74896538e-02
9.14833128e-01 -9.22578633e-01 1.20034468e+00 3.58647317e-01
-6.63465321e-01 -7.15041682e-02 -8.75355601e-01 -1.54138237e-01
-7.45437026e-01 -7.15631992e-02 2.31499448e-01 1.20216417e+00
-4.87251520e-01 6.88866496e-01 -2.33913675e-01 -4.16276157e-01
5.14802158e-01 -1.95067540e-01 7.66452588e-03 2.58441567e-01
-1.98832357e+00 1.14752114e+00 2.49041408e-01 -2.31084079e-01
-4.53025311e-01 -7.97162235e-01 -3.41001272e-01 8.49944130e-02
3.67998868e-01 -2.52522200e-01 1.13336897e+00 -2.64980763e-01
-7.65207410e-01 6.55864656e-01 6.90836012e-02 -2.41718978e-01
6.02408767e-01 -7.74909735e-01 -8.27747762e-01 -2.42074922e-01
2.52296686e-01 -1.83069199e-01 3.58258337e-01 -5.97019136e-01
-2.55992383e-01 -7.00472355e-01 1.57161742e-01 1.27613232e-01
-8.64464819e-01 9.32559133e-01 -2.79839486e-01 -5.66341758e-01
-3.81473035e-01 -7.71684885e-01 -1.31442502e-01 -5.03324449e-01
-1.11105597e+00 -8.30160201e-01 5.66317797e-01 -1.19666302e+00
2.16950417e+00 -1.83480942e+00 -4.12300885e-01 1.00367673e-01
4.49350476e-01 4.74648029e-01 -6.64304569e-02 8.65246415e-01
2.74237730e-02 8.81847978e-01 1.03458367e-01 -1.96774006e-02
1.27575353e-01 -5.16443789e-01 -4.86702561e-01 2.83523798e-01
-2.77292877e-02 8.21934462e-01 -5.75249314e-01 -7.07401633e-01
-7.90544212e-01 8.66084322e-02 -2.74237216e-01 2.73279548e-01
-1.55957550e-01 -3.83970477e-02 -8.98946941e-01 8.32220495e-01
7.93581724e-01 -5.25408089e-02 -2.00963616e-01 2.22906992e-01
-3.19489181e-01 6.51753545e-01 -9.97606039e-01 1.18169594e+00
-9.31666940e-02 4.71433043e-01 -2.34157786e-01 -1.70490310e-01
7.92956471e-01 2.61462718e-01 2.14349300e-01 -6.54771626e-01
1.77155823e-01 -3.17075439e-02 -1.51429757e-01 -1.27054131e+00
6.60111487e-01 1.07278347e-01 -5.16579866e-01 2.89500177e-01
-3.07068586e-01 3.38199794e-01 2.88676649e-01 6.56041384e-01
1.68034542e+00 -2.49242529e-01 1.19266436e-01 -1.90997541e-01
4.76962119e-01 -4.57246751e-02 6.82101071e-01 3.66710067e-01
-6.78459167e-01 4.51434165e-01 8.60529125e-01 -2.61422932e-01
-9.68901098e-01 -8.96250069e-01 -4.86693718e-02 1.29246604e+00
2.77629308e-02 -8.67899597e-01 -1.01373994e+00 -1.24021411e+00
-5.75722791e-02 1.05375016e+00 -5.26995897e-01 -3.11222356e-02
-5.03766894e-01 -6.98716819e-01 1.10353959e+00 4.28891540e-01
5.57824671e-01 -1.17402673e+00 -7.91705251e-02 1.48200840e-01
-6.13626599e-01 -9.21059430e-01 -8.00256073e-01 -2.08184853e-01
-4.46072698e-01 -1.11598563e+00 -7.17519045e-01 -2.96912014e-01
-1.71537660e-02 5.77429608e-02 1.19256747e+00 2.54598171e-01
-3.93853873e-01 -2.81475931e-01 -6.37059808e-01 -6.37549758e-01
-9.72828176e-03 6.75758541e-01 -1.35146841e-01 -2.89019406e-01
6.47181928e-01 -5.78996599e-01 -5.62596679e-01 3.65877509e-01
-6.73030436e-01 -6.41299844e-01 4.01366115e-01 1.91656262e-01
-3.36766541e-01 -1.95299581e-01 1.07190204e+00 -1.37119448e+00
1.04509759e+00 -9.83462751e-01 -8.79468918e-02 4.06427920e-01
-3.41852725e-01 -2.10089937e-01 6.97387636e-01 -3.30155343e-01
-1.11213291e+00 -1.14820622e-01 -6.35298550e-01 1.46951750e-01
-8.91200751e-02 4.77152407e-01 -6.81263804e-01 4.62058365e-01
1.29925251e+00 -2.23307922e-01 -7.56275475e-01 -8.19984734e-01
4.82694477e-01 9.91614580e-01 7.02461451e-02 -8.82337689e-02
8.23479831e-01 -1.87771603e-01 -7.83628643e-01 -7.91811049e-01
-1.28892469e+00 -9.21685815e-01 -4.45792347e-01 -1.51949525e-01
9.40188408e-01 -6.38355672e-01 -6.65145099e-01 4.64161605e-01
-1.49138534e+00 1.74411476e-01 4.52926345e-02 1.66503713e-01
1.40278667e-01 6.96839213e-01 -7.41607845e-01 -9.78799224e-01
-6.85885429e-01 -4.41055804e-01 2.89576858e-01 3.35414737e-01
-4.26702291e-01 -5.60437441e-01 4.99178469e-01 5.71281254e-01
5.58427513e-01 2.05993205e-01 6.86370730e-01 -1.31739128e+00
-9.85897556e-02 -2.80278623e-01 -4.40726340e-01 5.10803945e-02
-4.64916289e-01 1.29154120e-02 -9.91413057e-01 1.42239362e-01
-6.34807795e-02 -3.49419951e-01 8.43494117e-01 -2.79168725e-01
1.32897735e+00 -6.15027487e-01 -3.84894878e-01 -9.34418514e-02
1.22535264e+00 -1.54487425e-04 1.15210891e+00 3.93007994e-01
4.20226365e-01 7.43143022e-01 4.98502761e-01 8.58126104e-01
3.92819464e-01 4.82748121e-01 5.59717000e-01 3.57711881e-01
-2.56181397e-02 -5.16343594e-01 5.37622988e-01 7.33743608e-01
1.49787784e-01 -4.79201347e-01 -1.00785124e+00 5.89592457e-01
-1.72546530e+00 -1.22544014e+00 -8.55210721e-01 1.74176574e+00
1.07686079e+00 3.08933288e-01 4.54924166e-01 2.91517880e-02
9.36949372e-01 3.28420043e-01 -5.45885026e-01 -2.86412597e-01
-7.76792467e-02 -1.29917905e-01 4.90230680e-01 -6.20924458e-02
-1.11314940e+00 8.15477312e-01 6.69449091e+00 1.20287716e+00
-4.08104718e-01 3.11788201e-01 7.34768689e-01 1.27177924e-01
-3.29095095e-01 2.74404194e-02 -1.13388014e+00 9.26343501e-01
1.40265572e+00 -2.80354798e-01 1.56167388e-01 9.26828086e-01
1.82099253e-01 3.56358923e-02 -4.71736521e-01 7.66165555e-01
1.63161889e-01 -8.86550069e-01 -3.09500724e-01 -7.55490288e-02
5.47524333e-01 4.19206545e-02 -1.77949026e-01 8.29736054e-01
3.40235472e-01 -9.70832825e-01 5.55905819e-01 7.28764713e-01
3.67322683e-01 -7.61180580e-01 6.82260394e-01 8.32389414e-01
-8.80716443e-01 -2.46167570e-01 -6.66069150e-01 -2.97580864e-02
3.19785833e-01 1.08844686e+00 -4.87435013e-01 2.42015794e-01
8.07716548e-01 2.07300305e-01 -1.06256557e+00 1.69588447e+00
-4.76338416e-01 1.19064534e+00 -2.53198504e-01 -6.49556279e-01
-1.40121104e-02 2.11508691e-01 6.27088070e-01 1.38262331e+00
2.35667333e-01 1.43602252e-01 -9.57737565e-02 8.29285264e-01
-7.23530710e-01 5.88856578e-01 -5.02384484e-01 -4.35955971e-02
7.12837458e-01 1.50350749e+00 -3.88538599e-01 -3.80572289e-01
-4.24202651e-01 9.35411394e-01 5.45771956e-01 -4.39834818e-02
-1.24771464e+00 -8.45324636e-01 2.30409607e-01 3.39801848e-01
-1.29373595e-01 -4.89794910e-02 -6.54643059e-01 -1.22480559e+00
-1.04528196e-01 -6.58683896e-01 7.37305522e-01 -6.24951661e-01
-1.69562066e+00 7.40499139e-01 -4.57411021e-01 -9.87429142e-01
5.12370229e-01 -2.18962058e-01 -1.06796980e+00 8.33896577e-01
-1.16362178e+00 -7.76313066e-01 -2.07791284e-01 4.93624508e-01
6.07220232e-01 1.16069384e-01 5.15887856e-01 7.18875408e-01
-9.78698850e-01 9.08895433e-01 -6.52365312e-02 3.35096151e-01
9.43844557e-01 -1.33349895e+00 6.09582543e-01 8.09712172e-01
-2.82013901e-02 5.56075811e-01 6.71523452e-01 -9.17970777e-01
-7.83820391e-01 -1.27824795e+00 1.65226603e+00 -1.21899462e+00
1.00425315e+00 -3.34664822e-01 -9.46263194e-01 3.05800021e-01
4.50411618e-01 -3.20041835e-01 9.86294985e-01 4.05115753e-01
-6.81548715e-01 3.30638498e-01 -9.98291492e-01 5.01687109e-01
1.29622233e+00 -5.29423475e-01 -7.92044699e-01 7.08575666e-01
1.05888355e+00 7.36273080e-02 -9.40212965e-01 -1.23607665e-01
2.16350984e-02 -9.30284739e-01 5.80316782e-01 -1.09802520e+00
4.53045636e-01 -5.55646084e-02 1.35792241e-01 -1.13571668e+00
-2.69522130e-01 -6.73364460e-01 -5.46900667e-02 1.71339190e+00
1.11779237e+00 -3.61026734e-01 7.98891008e-01 1.06968868e+00
-1.70034856e-01 -6.52810693e-01 -7.15631425e-01 -8.30121517e-01
3.06603909e-01 -3.10934782e-01 5.84490001e-01 9.79563475e-01
4.45419520e-01 8.20588529e-01 -4.37689245e-01 1.11775905e-01
1.80488497e-01 -2.50716358e-01 5.00690222e-01 -1.08042932e+00
1.58425882e-01 -2.11905167e-01 2.33558398e-02 -8.17870021e-01
-1.31196946e-01 -7.91078269e-01 -2.13046759e-01 -1.26464021e+00
6.22659028e-01 -1.51924506e-01 -4.91301008e-02 5.31115055e-01
-4.93788630e-01 2.01281011e-01 1.53052390e-01 3.02900702e-01
-1.21961629e+00 6.34802341e-01 8.31818461e-01 3.01599856e-02
-1.24654830e-01 4.56989855e-01 -9.95097578e-01 6.73601389e-01
1.09104633e+00 -7.52291799e-01 1.19137131e-01 8.16072077e-02
8.61660957e-01 -1.82352141e-02 -1.35646805e-01 -8.59260917e-01
2.32015580e-01 -1.63695380e-01 2.41607428e-01 -9.05887663e-01
-1.48893207e-01 -4.57407206e-01 -2.61051685e-01 3.94826829e-01
-4.79910791e-01 2.88707554e-01 -1.53561488e-01 4.88371313e-01
4.80738655e-02 -7.34010875e-01 5.87145686e-01 -7.20448941e-02
-4.70335007e-01 3.38058472e-01 -5.96455097e-01 6.91650152e-01
9.34561312e-01 3.07368398e-01 -7.88692415e-01 -3.25083673e-01
-8.33443463e-01 2.31084630e-01 -1.64135173e-01 4.71176088e-01
3.32732886e-01 -1.22504818e+00 -1.03503847e+00 -5.44871271e-01
2.47305974e-01 -7.17007101e-01 3.31387639e-01 8.44902277e-01
-4.74280655e-01 2.09268034e-01 2.24499516e-02 4.00914490e-01
-1.18819964e+00 7.28402078e-01 1.71878323e-01 -7.35955358e-01
-2.06595749e-01 9.03202951e-01 -2.79762030e-01 -5.83524764e-01
2.50388175e-01 2.18828201e-01 -6.99939549e-01 3.82449061e-01
9.04110909e-01 1.15167212e+00 1.31629542e-01 -4.06325936e-01
-4.13462043e-01 -1.19948328e-01 -1.79753259e-01 1.86126992e-01
1.09361589e+00 -1.43553585e-01 -2.51464039e-01 1.88851744e-01
1.40877867e+00 5.57906032e-01 -4.93881911e-01 -1.91749230e-01
6.68863416e-01 -3.79364878e-01 -4.28923965e-01 -1.14873469e+00
-6.70063198e-01 7.81953216e-01 2.04141051e-01 7.24178493e-01
7.23735631e-01 1.03179105e-02 1.27357399e+00 2.39165127e-01
3.14982571e-02 -1.35145795e+00 2.77175128e-01 9.15112972e-01
1.15482390e+00 -9.45058167e-01 -4.15604003e-02 -6.21978283e-01
-6.75984502e-01 8.50586057e-01 8.90757442e-01 1.06386319e-02
6.97543800e-01 -7.32821673e-02 -9.35067162e-02 -2.90298581e-01
-7.25676715e-01 -2.85254776e-01 3.09808463e-01 2.83526748e-01
4.81436461e-01 4.43166420e-02 -9.62319434e-01 1.28077877e+00
-3.29721183e-01 -5.23786366e-01 5.65478981e-01 5.75214624e-01
-8.02030444e-01 -8.72593999e-01 -3.41228932e-01 6.94598436e-01
-1.08972824e+00 -3.67666334e-01 -8.83642375e-01 5.98710477e-01
5.04734693e-03 1.07609534e+00 -5.68228543e-01 -6.63392901e-01
6.43678248e-01 2.75396854e-01 -4.45525020e-01 -6.44056022e-01
-1.23418772e+00 -2.27151364e-01 6.26793504e-01 -6.43119693e-01
2.93637007e-01 -6.63579524e-01 -1.26265359e+00 -6.58164263e-01
-8.35088670e-01 5.59584379e-01 5.89818716e-01 5.82020104e-01
4.48163092e-01 1.62110314e-01 9.63362515e-01 1.01759508e-01
-1.08227336e+00 -1.35656321e+00 -7.47411907e-01 5.56310833e-01
-4.00111318e-01 -1.62527084e-01 -4.56626266e-01 -4.16332245e-01] | [8.972380638122559, 10.623212814331055] |
1918c7b4-b1a9-4c7f-bc81-87f643464bb0 | assamese-word-sense-disambiguation-using | null | null | https://aclanthology.org/2020.icon-main.40 | https://aclanthology.org/2020.icon-main.40.pdf | Assamese Word Sense Disambiguation using Genetic Algorithm | Word sense disambiguation (WSD) is a problem to determine a word according to a context in which it occurs. There are plenty amount of works done in WSD for some languages such as English, but research work on Assamese WSD remains limited. It is a more exigent task because Assamese has an intrinsic complexity in its writing structure and ambiguity, such as syntactic, semantic, and anaphoric ambiguity levels.A novel unsupervised genetic word sense disambiguation algorithm is proposed in this paper. The algorithm first uses WordNet to extract all possible senses for a given ambiguous word, then a genetic algorithm is used taking Wu-Palmer’s similarity measure as the fitness function and calculating the similarity measure for all extracted senses. The winner sense which will have the highest score declared as he winner sense. | ['Shikhar Kr. Sarma', 'Nomi Baruah', 'Arjun Gogoi'] | null | null | null | null | icon-2020-12 | ['word-sense-disambiguation'] | ['natural-language-processing'] | [ 3.47364873e-01 -2.84285188e-01 -3.63279819e-01 -6.61504790e-02
6.20648116e-02 -6.57960653e-01 4.31326032e-01 6.31105721e-01
-7.95393169e-01 9.96846020e-01 5.04443824e-01 -2.54210293e-01
-6.02499902e-01 -9.01968241e-01 3.60427409e-01 -4.10170585e-01
3.55689317e-01 4.94122624e-01 4.72984999e-01 -8.96989286e-01
8.13247383e-01 3.27772588e-01 -1.40175986e+00 -2.67937034e-01
1.10294747e+00 2.44138181e-01 8.06350052e-01 1.45179242e-01
-9.53672826e-01 2.35138163e-01 -1.06092477e+00 -1.48555920e-01
4.14875075e-02 -6.22115016e-01 -1.28470302e+00 -4.36916053e-02
-3.47849846e-01 6.43632770e-01 2.18446970e-01 1.33002234e+00
5.48445404e-01 3.38635504e-01 4.46770906e-01 -8.92186582e-01
-5.56508780e-01 7.72847235e-01 -4.38542545e-01 5.38301826e-01
6.27764165e-01 -5.37127852e-01 1.14813721e+00 -4.83227789e-01
8.13495576e-01 1.23153818e+00 7.01554120e-02 4.35138255e-01
-6.91450775e-01 -5.49953938e-01 5.23217879e-02 2.44591758e-01
-1.39513671e+00 2.88776010e-01 7.05161393e-01 -1.38019219e-01
1.16869438e+00 1.66091010e-01 8.29846501e-01 6.15253389e-01
3.55860740e-01 3.61231834e-01 1.22155297e+00 -9.43190157e-01
3.23014528e-01 -7.83185810e-02 5.34640014e-01 2.33505160e-01
6.35476649e-01 -3.74602586e-01 -5.12083888e-01 -9.79615375e-02
2.21099079e-01 -2.30458021e-01 -1.53607354e-02 1.69358492e-01
-9.59685206e-01 6.91702843e-01 -3.68044525e-02 1.18463922e+00
-2.85568923e-01 -5.86697042e-01 4.17124569e-01 3.79473716e-01
-2.23671459e-02 1.13222039e+00 -5.79020500e-01 -2.17918530e-01
-6.92362487e-01 5.60833514e-01 7.39704907e-01 5.64456046e-01
6.35829329e-01 -3.84990305e-01 8.50719884e-02 1.06008887e+00
3.29286456e-01 4.58113343e-01 1.15780175e+00 -4.08227444e-01
1.99065059e-01 1.10796881e+00 -7.95641020e-02 -1.35974967e+00
-4.09994721e-01 -1.66614726e-01 -3.53073001e-01 3.05323433e-02
-9.79037583e-02 -1.29496127e-01 -8.06757927e-01 1.63282347e+00
3.41996372e-01 4.82149795e-02 5.07529438e-01 8.86220813e-01
8.70890379e-01 5.07903874e-01 3.24301988e-01 -4.99291241e-01
1.68071663e+00 -1.98188096e-01 -9.22924280e-01 -6.68748438e-01
1.37272656e-01 -1.35419190e+00 9.11998570e-01 1.83839619e-01
-6.42186940e-01 -1.16839379e-01 -1.44186294e+00 1.16880715e-01
-9.05388355e-01 -3.33022237e-01 5.03763735e-01 6.40861869e-01
-6.52541935e-01 2.46242598e-01 -3.26565802e-01 -9.31319654e-01
-3.08568984e-01 1.66057959e-01 -2.64220744e-01 5.62928021e-02
-1.76072574e+00 1.27320492e+00 1.14780390e+00 -4.77565646e-01
2.17392385e-01 4.82636876e-02 -7.56269634e-01 -1.67039797e-01
7.61337101e-01 -5.91198087e-01 8.68774414e-01 -8.07889819e-01
-9.55924332e-01 1.09466302e+00 -1.84758291e-01 -3.19372028e-01
-2.67210305e-01 1.75064534e-01 -9.64684188e-01 -1.62944615e-01
5.12459576e-01 1.65747821e-01 3.46726328e-01 -9.37587440e-01
-8.39688897e-01 -6.37272716e-01 2.08157316e-01 6.37114942e-01
-2.47054875e-01 4.39985365e-01 -2.00838536e-01 -1.02775347e+00
6.37310028e-01 -7.68429160e-01 -8.43150541e-02 -7.97042966e-01
-1.30720675e-01 -7.90514946e-01 7.32780576e-01 -4.57393795e-01
1.92145467e+00 -1.93766153e+00 -8.90880451e-02 4.46960896e-01
6.24762550e-02 6.57907963e-01 9.42357928e-02 6.47892594e-01
-3.39988954e-02 2.56170958e-01 -3.31330627e-01 6.10016584e-01
-2.08901218e-03 6.90263569e-01 5.96421286e-02 -5.15540540e-01
3.13963480e-02 9.85509679e-02 -1.18796742e+00 -5.78271329e-01
-4.46982961e-03 -6.55809641e-02 -4.46726419e-02 -5.75907826e-02
-2.46102318e-01 -3.13807696e-01 -8.06337893e-01 5.51330507e-01
4.51051623e-01 1.69100538e-01 6.49273694e-01 -5.87349175e-04
-3.08066100e-01 5.34865260e-01 -1.58072376e+00 1.48423338e+00
-3.73184443e-01 1.83525309e-01 -3.57617348e-01 -8.67968142e-01
1.22847843e+00 3.11230451e-01 3.15416157e-01 -5.63433707e-01
1.76643655e-01 5.78108907e-01 1.29872218e-01 -8.19819629e-01
1.07953942e+00 -4.35492516e-01 -4.26572233e-01 3.02745670e-01
5.72836660e-02 -3.19348782e-01 8.13032985e-01 3.10017735e-01
8.83514047e-01 -1.32445395e-01 1.15139318e+00 -6.62617981e-01
5.57673812e-01 2.56074518e-01 8.55075419e-01 4.14238155e-01
4.77460101e-02 1.03044838e-01 2.81049162e-01 -2.55433559e-01
-7.38038957e-01 -9.75523531e-01 -6.50174692e-02 7.61676967e-01
5.16541362e-01 -7.58499980e-01 -6.34789288e-01 -4.37816173e-01
-1.49248675e-01 8.54617178e-01 -1.44422039e-01 -4.29047346e-02
-4.32610154e-01 -6.92678094e-01 4.32611197e-01 -4.68582008e-03
8.20516050e-01 -1.35187423e+00 -6.41203523e-01 5.89156985e-01
-1.02533869e-01 -6.65022433e-01 -2.37884551e-01 4.85709235e-02
-3.94532025e-01 -1.06948936e+00 -1.04804263e-01 -9.15560603e-01
3.06816250e-01 1.60533771e-01 9.34023082e-01 1.58439368e-01
-4.00912344e-01 -2.27834731e-01 -7.38634050e-01 -9.17118847e-01
-3.14293325e-01 3.21726352e-01 1.23891525e-01 -6.39600039e-01
1.06197119e+00 -4.33809578e-01 -1.13023229e-01 -2.97784172e-02
-1.17700648e+00 -3.97640288e-01 4.24839497e-01 7.17648983e-01
5.21996319e-01 5.82416832e-01 5.42711556e-01 -7.98824131e-01
1.33851779e+00 -5.13540387e-01 -3.61623287e-01 4.71397817e-01
-9.44247961e-01 3.28443915e-01 8.50178823e-02 -2.02240404e-02
-1.04781806e+00 -2.44530410e-01 -2.15752378e-01 4.91341770e-01
-1.28628731e-01 8.47652495e-01 -3.01680088e-01 2.34727055e-01
6.91287994e-01 1.03788391e-01 -2.18197823e-01 -2.44620115e-01
-9.25427303e-03 8.36310863e-01 4.92042601e-01 -7.75391102e-01
4.93446797e-01 -2.23704860e-01 2.00221296e-02 -8.82112622e-01
-9.88756657e-01 -6.82467163e-01 -3.47322434e-01 -1.68662868e-03
1.09786499e+00 -1.63745657e-01 -3.14328343e-01 1.84590071e-01
-1.06082797e+00 6.90198898e-01 1.31950248e-02 6.63783610e-01
6.95043355e-02 4.51399922e-01 9.50488895e-02 -7.92761505e-01
-4.86363113e-01 -8.36662591e-01 5.23756742e-01 6.88472569e-01
-9.17184591e-01 -8.41584325e-01 2.05979586e-01 1.31460816e-01
1.69569671e-01 2.18265697e-01 1.23154092e+00 -1.00215209e+00
1.49531975e-01 1.48975374e-02 1.02495231e-01 1.57560796e-01
6.22397184e-01 -1.78367645e-01 -2.22246960e-01 1.09372839e-01
5.22388928e-02 7.35958889e-02 5.59873641e-01 8.69768392e-03
5.13435662e-01 -2.91119576e-01 -3.70383024e-01 -1.60667658e-01
1.70177162e+00 7.78614700e-01 5.97725332e-01 7.61163831e-01
3.06451201e-01 3.34697455e-01 1.08828211e+00 2.95882016e-01
1.71361595e-01 5.16912758e-01 -5.42354137e-02 3.76975507e-01
1.79879013e-02 -1.62704624e-02 -6.61182124e-03 8.71105254e-01
2.75321323e-02 -5.14037013e-01 -1.20955265e+00 6.44525588e-01
-1.69982946e+00 -9.81154203e-01 2.24289715e-01 1.98551476e+00
1.01476538e+00 3.77081811e-01 -1.31429210e-01 6.24805808e-01
8.64075065e-01 3.05680543e-01 -7.53266886e-02 -6.71850562e-01
-5.13231814e-01 6.70678198e-01 2.09287763e-01 7.07665622e-01
-6.44785941e-01 1.40736437e+00 5.69122696e+00 9.90451992e-01
-9.78700399e-01 -2.08621219e-01 -6.50097504e-02 3.34271044e-01
-3.93897414e-01 2.25545913e-01 -7.34133184e-01 5.00436783e-01
1.16382018e-01 -7.57714272e-01 2.68378466e-01 3.18541527e-01
3.11198011e-02 -4.33032423e-01 -1.23935319e-01 8.12229812e-01
2.84396052e-01 -9.03044581e-01 3.04619074e-01 -3.23078007e-01
5.50492704e-01 -2.90150404e-01 -4.85844374e-01 -1.09834179e-01
4.35002714e-01 -7.08025634e-01 2.26489589e-01 3.55816782e-01
2.73146302e-01 -9.94755983e-01 9.61920202e-01 1.35767356e-01
-1.05265760e+00 1.47118673e-01 -4.52385008e-01 -3.05864006e-01
-4.49556783e-02 6.74256206e-01 -1.03844285e+00 7.24252284e-01
4.31303471e-01 3.51665646e-01 -4.20903593e-01 9.15674150e-01
-5.79262912e-01 4.97167706e-01 -4.01015908e-01 -6.84557021e-01
3.11316013e-01 -3.36218715e-01 1.02936792e+00 1.12661242e+00
6.79161787e-01 6.05801582e-01 3.73017967e-01 2.62394875e-01
2.72648454e-01 7.48685718e-01 -4.62741137e-01 -2.93999463e-01
1.07229352e+00 8.59662831e-01 -1.05145979e+00 -3.53510410e-01
-2.88967062e-02 9.51101422e-01 -1.02382749e-01 -8.98295082e-03
-1.02123350e-01 -7.60189414e-01 8.31932187e-01 5.19674271e-03
-1.68065205e-01 -3.22537214e-01 -3.62579584e-01 -6.72271907e-01
2.40211766e-02 -7.46256411e-01 1.06541574e+00 -7.43213415e-01
-1.28484249e+00 6.13379538e-01 5.75633124e-02 -1.01898158e+00
-2.21047267e-01 -6.69205427e-01 -5.73372781e-01 9.25594866e-01
-1.14336038e+00 -5.31717360e-01 5.74167781e-02 3.22194844e-01
6.86761975e-01 -4.06813294e-01 1.02070057e+00 -1.31811947e-01
-1.60231888e-01 2.72977144e-01 -1.90102383e-01 -2.53234431e-02
7.27634072e-01 -1.27020788e+00 1.14037454e-01 9.03461576e-01
5.89167476e-02 8.84936869e-01 9.75417793e-01 -1.03722179e+00
-7.56651998e-01 -3.97408485e-01 1.83028007e+00 1.26733452e-01
7.65065312e-01 3.79376441e-01 -6.38648331e-01 1.37864426e-01
3.99817228e-01 -6.78803146e-01 8.22719574e-01 2.83858851e-02
-7.47760981e-02 -5.38781732e-02 -1.38105428e+00 8.65427375e-01
1.14730191e+00 -1.97184876e-01 -1.46060979e+00 6.97810501e-02
7.82063723e-01 -3.46074164e-01 -7.44430482e-01 5.89983575e-02
2.34093651e-01 -4.97581899e-01 8.39762449e-01 -5.04780769e-01
1.37146220e-01 -8.36559653e-01 -2.02274278e-01 -1.54932988e+00
-3.24332416e-01 -5.31248152e-01 7.36772299e-01 1.46326995e+00
5.66776752e-01 -6.22756660e-01 1.07870035e-01 3.23034316e-01
1.17637096e-02 -4.89363879e-01 -8.58626604e-01 -7.67641902e-01
-3.06796461e-01 -1.62896961e-01 1.11397314e+00 1.52719808e+00
4.77626860e-01 7.79680789e-01 1.22466765e-01 -5.28274663e-03
2.96888411e-01 2.05635577e-01 -1.31976217e-01 -1.49580956e+00
5.07157706e-02 -4.91375297e-01 -7.73815215e-01 -1.98628768e-01
1.73509270e-01 -9.45518017e-01 -2.06659630e-01 -1.61900461e+00
-1.06332071e-01 -3.88823926e-01 -5.94057441e-01 6.13573015e-01
-5.26917338e-01 -1.76876321e-01 1.69234082e-01 -2.16704130e-01
-3.26643795e-01 5.41636022e-03 9.53913093e-01 -3.46228480e-02
-3.29235613e-01 -2.84694374e-01 -8.70611668e-01 6.79222167e-01
9.67886567e-01 -6.61636174e-01 -5.64521194e-01 -3.68673891e-01
5.82973003e-01 -1.28758624e-01 -3.69891495e-01 -9.46309328e-01
2.14453965e-01 -7.36245990e-01 1.04766779e-01 -1.85568526e-01
-1.68399513e-01 -6.82334244e-01 9.96621996e-02 4.64076251e-01
-1.96612418e-01 5.80936849e-01 9.35434774e-02 8.26014131e-02
-3.44751567e-01 -8.77023280e-01 5.06054461e-01 -4.13263977e-01
-1.43749726e+00 -1.14785314e-01 -4.95923847e-01 2.98474520e-01
8.74849975e-01 -4.49409932e-01 1.15837678e-01 1.36352763e-01
-6.35965943e-01 1.25563815e-01 4.29100722e-01 8.86715829e-01
5.09141564e-01 -1.31510592e+00 -6.83319807e-01 -1.95919454e-01
3.97987306e-01 -2.99010903e-01 -1.56625867e-01 -2.01569065e-01
-3.62743050e-01 1.20593190e-01 -3.98289204e-01 1.72418967e-01
-1.70092499e+00 4.07350630e-01 -9.01811197e-02 -3.35099727e-01
-6.58334494e-02 6.12446308e-01 -6.58474505e-01 -2.22286537e-01
-3.88220221e-01 1.85572673e-02 -9.11310315e-01 2.38423258e-01
4.48274910e-01 2.92930871e-01 1.46565646e-01 -7.90479541e-01
-6.49942756e-01 7.10125744e-01 9.53223035e-02 -2.90036112e-01
1.04641795e+00 -1.34380773e-01 -6.28849983e-01 3.09891284e-01
7.36244917e-01 3.02988201e-01 2.22288202e-02 -3.86093408e-01
6.16237044e-01 -5.97829580e-01 -5.80873601e-02 -1.08559287e+00
-4.32527453e-01 1.94018707e-01 5.80145717e-01 3.20878625e-01
1.18590355e+00 -3.70179377e-02 9.26318169e-01 5.47747791e-01
4.35085744e-01 -1.69318032e+00 -3.48245919e-01 9.70422864e-01
6.37727201e-01 -7.08583415e-01 1.59714371e-02 -5.76935768e-01
-6.28812850e-01 9.98163462e-01 5.77518642e-01 7.40962774e-02
5.13521910e-01 3.39366674e-01 1.91514462e-01 -3.59663635e-01
-2.13793039e-01 -4.38797086e-01 8.65212679e-02 6.09084070e-01
7.16788232e-01 1.36921689e-01 -1.76846182e+00 7.50362635e-01
-7.32279778e-01 -1.37228459e-01 4.11528707e-01 1.10561931e+00
-9.17985022e-01 -1.59731209e+00 -3.86420876e-01 3.34818035e-01
-5.05506694e-01 -2.61202753e-01 -6.82306945e-01 4.95749652e-01
5.57812095e-01 1.14479458e+00 -1.44053474e-01 -3.10760140e-01
2.86020249e-01 1.79575160e-01 3.74547273e-01 -7.63384879e-01
-4.83529270e-01 -2.41080552e-01 2.48553753e-01 -1.79378930e-02
-6.65243208e-01 -5.29497802e-01 -1.94099510e+00 -2.62127846e-01
-2.19898149e-01 6.31573141e-01 6.01353347e-01 1.34500325e+00
-9.36782956e-02 2.43780971e-01 4.23693269e-01 2.10346013e-01
-1.17652461e-01 -7.07047760e-01 -7.13277698e-01 5.80162466e-01
-3.42817664e-01 -7.53871262e-01 4.16331813e-02 -3.31316978e-01] | [10.142845153808594, 9.16417407989502] |
f5272e1b-742c-4883-91f0-1f5b0612a82e | shallow-discourse-parsing-for-open | null | null | https://aclanthology.org/2022.codi-1.9 | https://aclanthology.org/2022.codi-1.9.pdf | Shallow Discourse Parsing for Open Information Extraction and Text Simplification | We present a discourse-aware text simplification (TS) approach that recursively splits and rephrases complex English sentences into a semantic hierarchy of simplified sentences. Using a set of linguistically principled transformation patterns, sentences are converted into a hierarchical representation in the form of core sentences and accompanying contexts that are linked via rhetorical relations. As opposed to previously proposed sentence splitting approaches, which commonly do not take into account discourse-level aspects, our TS approach preserves the semantic relationship of the decomposed constituents in the output. A comparative analysis with the annotations contained in RST-DT shows that we capture the contextual hierarchy between the split sentences with a precision of 89% and reach an average precision of 69% for the classification of the rhetorical relations that hold between them. Moreover, an integration into state-of-the-art Open Information Extraction (IE) systems reveals that when applying our TS approach as a pre-processing step, the generated relational tuples are enriched with additional meta information, resulting in a novel lightweight semantic representation for the task of Open IE. | ['Siegfried Handschuh', 'André Freitas', 'Christina Niklaus'] | null | null | null | null | coling-codi-crac-2022-10 | ['discourse-parsing', 'open-information-extraction'] | ['natural-language-processing', 'natural-language-processing'] | [ 5.18728375e-01 1.28146553e+00 -1.85712427e-02 -2.83808410e-01
-7.76679337e-01 -6.58323467e-01 9.12731051e-01 9.39399004e-01
-2.05448225e-01 7.84715116e-01 9.26761568e-01 -2.80252784e-01
-2.28636712e-01 -1.01463330e+00 -4.48613673e-01 -2.26516952e-03
4.75362420e-01 5.24100959e-01 5.64778566e-01 -6.49485826e-01
2.43692413e-01 2.58305985e-02 -1.85131180e+00 9.19845998e-01
1.09705341e+00 7.99629211e-01 -5.41942641e-02 4.13277984e-01
-8.20213675e-01 1.31559396e+00 -8.98817658e-01 -8.68804395e-01
-8.49207714e-02 -4.76396173e-01 -1.50041735e+00 -1.07276335e-01
1.62218615e-01 3.71400356e-01 -4.02598046e-02 9.63827133e-01
-3.15174721e-02 2.82870252e-02 4.28841233e-01 -6.31851077e-01
-1.58987135e-01 1.18340719e+00 -9.94650088e-03 -1.98095381e-01
9.23256814e-01 -6.41569734e-01 1.36875987e+00 -6.47106469e-01
1.42501831e+00 1.43149853e+00 7.53299117e-01 2.88302600e-01
-1.36867833e+00 9.08548087e-02 8.41204152e-02 2.43757829e-01
-9.60318983e-01 -5.61685026e-01 6.31809711e-01 -3.29813063e-01
1.42858171e+00 7.68589854e-01 5.56378245e-01 6.91857994e-01
-6.04513995e-02 3.47330838e-01 8.72990787e-01 -9.55750942e-01
1.49795040e-01 4.76079732e-01 5.80365777e-01 3.65813375e-01
4.92638737e-01 -6.90354288e-01 -5.66319704e-01 -7.80174956e-02
-1.55900940e-01 -4.74811018e-01 2.83614919e-03 -1.90914735e-01
-9.65416610e-01 8.25491548e-01 2.12130651e-01 7.54021585e-01
-3.60996932e-01 -5.95681310e-01 1.06407511e+00 2.17909873e-01
6.60460711e-01 6.15937173e-01 -2.51801640e-01 -2.58676931e-02
-8.87924075e-01 6.75751507e-01 1.43311858e+00 1.11313665e+00
7.17058182e-01 -5.16248882e-01 -3.01506281e-01 7.06531227e-01
-3.75825018e-02 5.45352846e-02 3.23439211e-01 -1.32030606e+00
7.11655855e-01 1.29218698e+00 -2.09590588e-02 -1.09034348e+00
-1.59336343e-01 -2.38201454e-01 -3.29636723e-01 -2.22978398e-01
2.20102258e-02 1.67433813e-01 -1.99810639e-01 1.45930791e+00
6.05507314e-01 -6.01538181e-01 5.72027087e-01 2.16322511e-01
1.24884188e+00 4.26772535e-01 1.82326913e-01 -5.20088196e-01
1.68784738e+00 -7.86493778e-01 -1.05683267e+00 -7.80015893e-04
9.78003502e-01 -6.00964785e-01 9.56105113e-01 2.26245806e-01
-1.25642323e+00 -2.22400591e-01 -1.25456190e+00 -7.99455941e-01
-5.60066521e-01 -2.30252147e-01 3.76349747e-01 4.25212681e-01
-7.27651417e-01 7.94356406e-01 -2.83828706e-01 -4.07113105e-01
2.60142833e-01 -8.00853670e-02 -4.29372162e-01 2.90105373e-01
-1.28931391e+00 1.16753340e+00 7.12771952e-01 -3.85501981e-01
-2.32405066e-02 -8.82119894e-01 -1.14339554e+00 2.67163455e-01
9.23100770e-01 -8.56685221e-01 1.34381711e+00 -6.90606713e-01
-1.50972378e+00 1.13402832e+00 -5.90244234e-01 -8.43428373e-01
4.51296389e-01 -3.57369930e-01 -2.38391325e-01 3.55025232e-01
5.34967959e-01 2.85366178e-01 1.78576082e-01 -1.28918576e+00
-9.00882185e-01 -4.25414175e-01 6.30973935e-01 2.86774635e-01
-2.65670806e-01 5.39137602e-01 -1.84971929e-01 -4.15337116e-01
5.27463369e-02 -5.74967146e-01 3.03225107e-02 -6.28236890e-01
-6.70353353e-01 -5.03691256e-01 6.46081984e-01 -9.78073239e-01
1.98421204e+00 -1.87323987e+00 5.69922507e-01 4.19591274e-03
4.04542774e-01 2.11132929e-01 4.74131852e-01 8.28609943e-01
-7.81748295e-02 3.50654155e-01 -5.57117999e-01 -5.71344912e-01
3.53832632e-01 3.85741234e-01 -6.05789304e-01 -2.91060716e-01
-9.66784880e-02 7.93012738e-01 -9.80379820e-01 -9.19215918e-01
3.16247791e-01 1.86084181e-01 -5.61863244e-01 -1.47591271e-02
-6.15311325e-01 -3.92897725e-02 -2.48756379e-01 -5.87351760e-03
1.46873012e-01 2.60063767e-01 6.09049797e-01 -3.18746775e-01
-2.22403720e-01 1.10353279e+00 -9.74016011e-01 1.67679751e+00
-5.18769920e-01 5.91729939e-01 -8.62836391e-02 -8.27616215e-01
1.01105177e+00 3.98209214e-01 3.17933232e-01 -5.80171883e-01
9.07597840e-02 9.20783058e-02 -5.54037273e-01 -4.98148978e-01
1.02902675e+00 -2.16864184e-01 -5.47324419e-01 2.79729962e-01
1.07621156e-01 -4.41860974e-01 8.28625798e-01 6.23597503e-01
1.03979015e+00 3.54609579e-01 7.81490266e-01 -2.88137823e-01
1.02825296e+00 5.37460864e-01 5.75415492e-01 4.01819587e-01
3.50464165e-01 4.73185718e-01 1.06492829e+00 -4.81005013e-01
-1.17757356e+00 -6.22810304e-01 -2.01141283e-01 7.89807737e-01
-1.05724260e-01 -1.30655682e+00 -1.32588995e+00 -7.25624144e-01
-2.31916860e-01 1.35248995e+00 -6.15031660e-01 1.14520751e-01
-8.07720363e-01 -2.35574648e-01 5.93987226e-01 6.59008836e-03
1.74305871e-01 -1.05063081e+00 -9.12589967e-01 1.82704046e-01
-6.58635676e-01 -1.50522554e+00 3.15896720e-01 -5.22500696e-03
-5.22573590e-01 -1.25548410e+00 1.88777164e-01 -5.16415894e-01
3.36428761e-01 -2.37379372e-01 1.34120882e+00 -1.16907284e-01
1.05337344e-01 -1.21037982e-01 -6.54071331e-01 -5.75159073e-01
-1.12180150e+00 3.30020428e-01 -3.44083816e-01 -2.65939713e-01
4.09183323e-01 -4.98155624e-01 6.77341074e-02 -4.52014625e-01
-1.02769494e+00 3.46762598e-01 -5.50956689e-02 4.40779328e-01
4.44680631e-01 4.97282818e-02 3.07396442e-01 -1.68330657e+00
9.08617437e-01 -3.15042257e-01 -6.77966326e-02 5.89171350e-01
-5.66465020e-01 3.20491076e-01 6.82828546e-01 2.10800499e-01
-1.60619187e+00 -3.47262889e-01 -1.45385236e-01 3.00480902e-01
-1.70164868e-01 6.95684910e-01 -2.80852914e-01 6.01705670e-01
8.53812218e-01 -3.05298835e-01 -1.42993063e-01 -5.02990663e-01
7.42970347e-01 6.89989626e-01 6.81486249e-01 -6.72182381e-01
4.40546244e-01 4.77873236e-01 1.77099362e-01 -8.97741616e-01
-1.37038124e+00 -2.68106312e-01 -8.55632365e-01 1.00085087e-01
8.00446451e-01 -7.27038443e-01 -4.59668308e-01 -3.51358801e-01
-1.58489180e+00 1.56605586e-01 -1.05723655e+00 -1.48040965e-01
-5.67103326e-01 6.60221636e-01 -4.56513464e-01 -6.02007270e-01
-5.00448465e-01 -7.20624506e-01 1.04742694e+00 -1.13268279e-01
-1.27105951e+00 -9.06515300e-01 9.02332142e-02 6.68102324e-01
-6.97750747e-02 5.10533869e-01 1.45158577e+00 -1.00343788e+00
-3.98534611e-02 -5.05261794e-02 -1.46779209e-01 1.91766232e-01
8.37469175e-02 7.44297951e-02 -7.40537643e-01 4.77556288e-01
1.07540913e-01 2.78415671e-03 4.92486894e-01 -4.00693603e-02
3.61436099e-01 -6.81060672e-01 -2.47496799e-01 1.26435801e-01
1.19779015e+00 -6.55886233e-02 7.24497378e-01 6.39784515e-01
5.31064510e-01 1.19962418e+00 5.70615232e-01 2.51859695e-01
6.27924144e-01 7.37912953e-01 4.25938889e-03 3.62104952e-01
-4.32360113e-01 -5.33190131e-01 2.06692293e-01 8.19926679e-01
2.49619782e-02 1.80756181e-01 -7.55079150e-01 5.16984105e-01
-1.96929181e+00 -1.00490892e+00 -4.51177478e-01 1.93026543e+00
1.22657788e+00 3.41457158e-01 -5.10935555e-04 5.57955921e-01
4.87993509e-01 2.77175248e-01 1.40115753e-01 -8.67286026e-01
-3.41397762e-01 4.04958636e-01 7.10861012e-02 9.59567547e-01
-8.08643937e-01 1.10396266e+00 5.91624165e+00 7.89085925e-01
-5.20448387e-01 1.74481541e-01 1.99352372e-02 1.17120035e-02
-7.39646852e-01 5.51377535e-01 -7.19364882e-01 3.86311747e-02
1.14190459e+00 -5.63695729e-01 5.21760024e-02 7.47237742e-01
1.04079865e-01 -2.99565822e-01 -1.19878018e+00 4.33163851e-01
1.44721687e-01 -1.75213647e+00 4.38080013e-01 -2.68519640e-01
4.82749581e-01 -5.11904895e-01 -7.44283319e-01 2.37213030e-01
2.79156238e-01 -7.44140625e-01 1.17404819e+00 4.60370213e-01
4.40748185e-01 -6.74084425e-01 7.82443225e-01 3.06674749e-01
-9.76251364e-01 -7.56055117e-02 -1.52602628e-01 -3.97817343e-01
2.80465841e-01 6.93207502e-01 -6.55147851e-01 1.29220438e+00
5.14762342e-01 5.83605766e-01 -5.64629734e-01 1.32518440e-01
-6.28251195e-01 2.12254047e-01 -1.49498791e-01 -4.05499600e-02
-1.99862629e-01 -3.31932753e-01 9.24064040e-01 1.50481474e+00
-1.38901249e-01 1.21739261e-01 -1.33891866e-01 8.15965176e-01
-4.64090705e-02 3.53396237e-01 -5.98688662e-01 1.31794214e-01
6.42119646e-01 1.01653707e+00 -7.03302741e-01 -8.04924667e-01
-2.26375148e-01 6.80090606e-01 4.71142590e-01 3.15126888e-02
-2.81130254e-01 -7.06754982e-01 2.73528367e-01 1.27331868e-01
2.48274848e-01 1.49553776e-01 -7.27790236e-01 -1.17683029e+00
5.18069923e-01 -9.33540285e-01 4.40421373e-01 -6.73240006e-01
-6.52678132e-01 9.29898739e-01 3.78707826e-01 -6.72177315e-01
-5.14125586e-01 -8.69672894e-02 -3.71268392e-01 6.83422804e-01
-1.13362575e+00 -1.27029884e+00 -6.97822496e-03 8.74215551e-03
5.66359460e-01 2.72328198e-01 1.21628964e+00 4.02427930e-03
-4.75557894e-01 1.97151992e-02 -3.08861047e-01 -1.02807228e-02
2.98549980e-01 -1.33446717e+00 3.38703930e-01 1.05066669e+00
6.85800463e-02 7.64383078e-01 1.25284553e+00 -8.38131547e-01
-7.97100544e-01 -8.72374415e-01 2.01352119e+00 -5.38466632e-01
8.06555629e-01 -4.21429038e-01 -1.05363643e+00 6.90838337e-01
6.54853284e-01 -6.96593702e-01 5.84251821e-01 2.41390988e-01
-5.91435254e-01 -1.40775412e-01 -1.01830721e+00 6.02424204e-01
1.16902983e+00 -7.65967786e-01 -1.65344381e+00 2.25917071e-01
1.27336895e+00 -4.72632915e-01 -1.04198372e+00 4.33847040e-01
2.58388489e-01 -1.09141684e+00 7.77625144e-01 -7.26241767e-01
8.69069397e-01 -2.03812376e-01 -1.48926795e-01 -9.10272658e-01
1.61229953e-01 -7.96068609e-01 -1.97136283e-01 1.59875679e+00
5.50670683e-01 -3.08555752e-01 4.36293542e-01 7.53316522e-01
-3.54992688e-01 -5.04168570e-01 -9.87717986e-01 -3.73349726e-01
1.60888489e-02 -5.36285698e-01 5.80553353e-01 6.82013452e-01
8.45062494e-01 9.17819023e-01 2.29805499e-01 -3.77172500e-01
4.50715452e-01 3.54624927e-01 5.98046899e-01 -1.53683007e+00
3.57134640e-02 -4.15426552e-01 -2.15279967e-01 -4.07219946e-01
5.01312792e-01 -9.76346433e-01 -2.55639344e-01 -1.90577388e+00
1.31218135e-01 -1.27015799e-01 3.66333812e-01 2.42633805e-01
1.80240273e-01 -1.95777386e-01 3.19312543e-01 9.31939408e-02
-7.41232753e-01 5.08369982e-01 6.34904563e-01 4.81304973e-02
-4.27145630e-01 -4.23043728e-01 -1.10428143e+00 9.21316743e-01
3.70208293e-01 -5.35604239e-01 -2.65219986e-01 -7.52623528e-02
6.09800339e-01 -5.57026602e-02 8.68879631e-03 -7.60873199e-01
1.93126351e-01 -2.97178719e-02 -4.62201864e-01 -4.50468272e-01
1.26993448e-01 -6.54145479e-01 3.37658316e-01 3.33418995e-01
-6.43311560e-01 -3.15736026e-01 3.27199847e-02 6.86077178e-02
-5.22965372e-01 -5.30837297e-01 4.19032007e-01 -1.75815031e-01
-3.37971628e-01 -6.72852635e-01 -3.82846683e-01 3.51999760e-01
9.83219683e-01 -3.00756752e-01 -4.19349462e-01 -2.95443237e-02
-7.97362864e-01 5.53746894e-03 4.30587322e-01 3.26507211e-01
2.46347696e-01 -8.42650831e-01 -6.53787434e-01 -3.19484830e-01
7.25358650e-02 3.69414389e-01 4.34503742e-02 5.51656723e-01
-6.38489366e-01 8.01271856e-01 1.00606352e-01 -7.81293437e-02
-1.51672232e+00 5.56227744e-01 5.80159687e-02 -7.36719787e-01
-8.58240008e-01 7.03511313e-02 -7.09696636e-02 -3.84309560e-01
1.31460768e-03 -6.22220576e-01 -5.82166493e-01 4.97246444e-01
4.54051942e-01 4.92530435e-01 4.90735650e-01 -9.04541612e-01
-2.32536346e-01 2.76223689e-01 1.03855282e-01 -2.08224237e-01
1.57496202e+00 -5.28784633e-01 -7.62428522e-01 5.14040828e-01
8.88841927e-01 4.54738140e-01 -4.02307898e-01 -4.48417962e-01
8.50114822e-01 6.98193461e-02 -3.33260685e-01 -5.85849226e-01
-3.41894962e-02 4.75162476e-01 -5.43814957e-01 6.90034389e-01
8.64439011e-01 4.36978817e-01 7.07493663e-01 4.85982627e-01
6.25586137e-02 -1.25420952e+00 -5.06785035e-01 8.19212258e-01
1.09804380e+00 -5.49755514e-01 2.73562819e-01 -1.24455106e+00
-6.27353072e-01 1.14351714e+00 1.03415780e-01 6.59052283e-02
2.28904143e-01 4.00400430e-01 -9.24977437e-02 -4.15303111e-01
-7.40177095e-01 -2.66750872e-01 8.32925886e-02 4.80467737e-01
4.68070388e-01 9.69301388e-02 -8.31776619e-01 1.02380562e+00
-8.62925231e-01 -1.28677025e-01 7.54084826e-01 9.68495965e-01
-5.88960648e-01 -1.30469418e+00 -1.26370445e-01 5.87496981e-02
-4.82252270e-01 -1.74859077e-01 -8.28668952e-01 7.53720939e-01
1.94925919e-01 1.02211940e+00 1.26611637e-02 -1.77403107e-01
8.50312591e-01 4.50011075e-01 3.94064963e-01 -1.09787619e+00
-9.92463470e-01 -3.01288962e-01 9.50685918e-01 -5.50104260e-01
-7.91538179e-01 -7.43233085e-01 -1.57987559e+00 -2.78220832e-01
3.88021744e-03 4.80997473e-01 4.71169889e-01 1.47056639e+00
5.10022104e-01 7.70659447e-01 2.57984102e-01 -4.37283546e-01
-2.98713863e-01 -8.64013374e-01 -9.21273902e-02 7.52867699e-01
1.23200029e-01 -3.24702173e-01 -1.55654311e-01 3.41493040e-01] | [10.310153007507324, 9.180363655090332] |
81e56231-18da-4ad9-835b-3730af880309 | sipos-a-benchmark-dataset-for-sindhi-part-of | null | null | https://aclanthology.org/2021.ranlp-srw.4 | https://aclanthology.org/2021.ranlp-srw.4.pdf | SiPOS: A Benchmark Dataset for Sindhi Part-of-Speech Tagging | In this paper, we introduce the SiPOS dataset for part-of-speech tagging in the low-resource Sindhi language with quality baselines. The dataset consists of more than 293K tokens annotated with sixteen universal part-of-speech categories. Two experienced native annotators annotated the SiPOS using the Doccano text annotation tool with an inter-annotation agreement of 0.872. We exploit the conditional random field, the popular bidirectional long-short-term memory neural model, and self-attention mechanism with various settings to evaluate the proposed dataset. Besides pre-trained GloVe and fastText representation, the character-level representations are incorporated to extract character-level information using the bidirectional long-short-term memory encoder. The high accuracy of 96.25% is achieved with the task-specific joint word-level and character-level representations. The SiPOS dataset is likely to be a significant resource for the low-resource Sindhi language. | ['Jay Kumar', 'Zenglin Xu', 'Wazir Ali'] | null | null | null | null | ranlp-2021-9 | ['text-annotation'] | ['natural-language-processing'] | [-8.45661163e-02 1.35305479e-01 -3.84271055e-01 -4.79557484e-01
-1.38815939e+00 -6.26143277e-01 2.90366113e-01 6.28666952e-02
-9.66632307e-01 9.16889369e-01 5.65230906e-01 -3.74984980e-01
4.75119054e-01 -5.52535176e-01 -5.58032215e-01 -6.14115834e-01
3.38704996e-02 6.81413531e-01 1.02950998e-01 -3.29875685e-02
1.15290888e-01 2.64717221e-01 -7.46373117e-01 6.34259462e-01
6.83526337e-01 8.45684826e-01 5.07219374e-01 9.36995506e-01
-4.13952202e-01 7.03290820e-01 -9.43326175e-01 -8.55370283e-01
-1.34239405e-01 8.47892389e-02 -1.21939087e+00 -2.52324760e-01
1.98693350e-01 -1.44944206e-01 -4.98387903e-01 9.96158004e-01
7.83202946e-01 8.98320526e-02 4.13398206e-01 -6.95960045e-01
-9.28214788e-01 1.13897717e+00 -2.01325253e-01 3.68117899e-01
-2.31450666e-02 6.82339296e-02 1.40782869e+00 -1.06110168e+00
4.29304779e-01 1.20381641e+00 5.61585784e-01 5.47546387e-01
-6.81873381e-01 -8.59440148e-01 -5.79067543e-02 1.13728099e-01
-1.56395829e+00 -3.89534235e-01 4.25268710e-01 -3.96983832e-01
1.57096720e+00 -1.26908034e-01 2.36274704e-01 1.15396404e+00
2.45701507e-01 9.70807791e-01 9.93458390e-01 -5.92742920e-01
-1.16915576e-01 1.64854124e-01 2.77896613e-01 6.10986590e-01
9.51068103e-02 -3.01478684e-01 -7.43177772e-01 -9.91216525e-02
4.98088121e-01 -6.08272739e-02 1.93683580e-01 5.72152078e-01
-1.27450669e+00 7.65460312e-01 1.32036090e-01 6.72147691e-01
-2.61545569e-01 1.87704697e-01 7.61013567e-01 -1.33554265e-01
6.72335744e-01 2.60415018e-01 -1.07442200e+00 -3.70010823e-01
-9.20987129e-01 -4.66583580e-01 7.05007493e-01 1.38610005e+00
5.86817384e-01 2.11655766e-01 -4.79570836e-01 1.16422939e+00
5.06356359e-01 6.43559575e-01 8.54227781e-01 -4.67168331e-01
1.03636765e+00 1.10767066e-01 -7.86562487e-02 -2.49747604e-01
-1.87569082e-01 -4.50305670e-01 -6.98192000e-01 -5.41882634e-01
8.35461989e-02 -4.07714754e-01 -1.07730949e+00 1.58977985e+00
-6.00613914e-02 -1.07004836e-01 3.75485122e-01 5.24258733e-01
8.03291082e-01 9.79117811e-01 6.98972106e-01 8.59375820e-02
1.77219367e+00 -1.25872052e+00 -1.09803414e+00 -4.61028278e-01
9.56172347e-01 -1.10441732e+00 9.86463428e-01 -1.24604605e-01
-1.00391889e+00 -5.90753734e-01 -8.91850293e-01 -2.32713699e-01
-7.80590415e-01 6.46664798e-01 4.42231327e-01 7.99152970e-01
-8.73201847e-01 2.81253308e-01 -7.82281220e-01 -2.88343340e-01
2.55375713e-01 1.25351936e-01 -5.21430135e-01 6.92198351e-02
-1.76631498e+00 7.98051178e-01 8.76556873e-01 2.10714161e-01
-7.68852293e-01 -4.43810135e-01 -1.01512551e+00 1.58832610e-01
-2.09891307e-03 -3.86285521e-02 1.32553744e+00 -6.56672359e-01
-1.51177585e+00 1.15091527e+00 -1.68941990e-01 -4.50955510e-01
7.02582002e-02 -4.19697315e-01 -7.16625571e-01 7.66373873e-02
2.50836730e-01 7.12605655e-01 3.61526340e-01 -5.52475274e-01
-5.42802095e-01 -1.47930756e-01 -4.13193524e-01 2.21135050e-01
-5.32535195e-01 5.15254319e-01 -5.17849743e-01 -9.43013012e-01
-2.13605553e-01 -8.67267489e-01 -8.02343339e-02 -5.59497535e-01
-4.88947511e-01 -3.29601377e-01 4.17538762e-01 -1.37017620e+00
1.39475214e+00 -2.10887218e+00 -6.28805578e-01 -1.97574690e-01
-3.51897538e-01 6.69560254e-01 -3.12044203e-01 7.98340380e-01
4.26866971e-02 4.30178106e-01 1.08861420e-02 -4.73566592e-01
9.82548520e-02 2.32253760e-01 -2.56822616e-01 2.90408671e-01
4.37397450e-01 1.02641726e+00 -1.06233716e+00 -6.67551756e-01
2.71698255e-02 6.83752775e-01 -1.02060862e-01 4.87340778e-01
6.83676824e-02 2.07112536e-01 -3.25583637e-01 5.73944807e-01
5.32010436e-01 2.85660550e-02 2.51383871e-01 9.85780880e-02
-4.36993204e-02 1.00948751e+00 -6.92333102e-01 1.80172765e+00
-8.10256004e-01 5.37112772e-01 -9.53703746e-02 -6.58899188e-01
1.17721379e+00 8.00206900e-01 1.40938535e-01 -5.82867086e-01
-2.23280378e-02 4.40841585e-01 -1.12470776e-01 -4.90246892e-01
9.08477604e-01 -1.36930630e-01 -4.30323243e-01 6.74733222e-02
6.54095471e-01 1.88214332e-01 9.85365435e-02 2.17926025e-01
1.00639677e+00 1.64734796e-01 4.13525969e-01 -2.29038313e-01
5.03970563e-01 -1.35258436e-01 7.86821663e-01 5.83461702e-01
-2.76557386e-01 5.91963172e-01 2.59641796e-01 -1.78573295e-01
-1.25822711e+00 -8.46950948e-01 -2.04324946e-01 1.44303095e+00
-5.68231881e-01 -4.36354369e-01 -7.65851438e-01 -8.68659556e-01
-4.03158367e-01 6.51681781e-01 -3.47837627e-01 1.04178451e-01
-6.52452588e-01 -4.95652735e-01 1.41470277e+00 9.27124500e-01
7.15656936e-01 -1.45792317e+00 7.78568387e-02 5.17401457e-01
-4.28923965e-01 -1.39635658e+00 -7.92892218e-01 6.11076891e-01
-5.38493574e-01 -5.87239265e-01 -1.13472271e+00 -1.09920979e+00
2.48687908e-01 -9.49715078e-02 1.09177876e+00 -1.98405594e-01
1.35295436e-01 -7.38482997e-02 -6.27129138e-01 -1.98976994e-02
-4.00771499e-01 5.11098742e-01 -1.85832307e-01 -1.77798480e-01
7.99697399e-01 1.30484119e-01 -2.25679934e-01 1.74006864e-01
-6.64233863e-01 -1.70551732e-01 7.89795220e-01 9.41588283e-01
5.37558019e-01 -1.80736542e-01 5.87805808e-01 -9.58591342e-01
1.98942751e-01 -4.67598677e-01 -4.10811275e-01 3.02755207e-01
-2.18765080e-01 8.22823420e-02 5.91612518e-01 -3.81295025e-01
-1.43777561e+00 2.10402012e-01 -7.70241499e-01 2.58994941e-02
-2.20618770e-01 5.93366981e-01 -6.41773582e-01 6.37162983e-01
1.79634273e-01 2.45783329e-01 -8.26560557e-01 -8.20893586e-01
4.60626274e-01 1.52924061e+00 6.45673811e-01 -6.40026152e-01
3.78628492e-01 -7.56066144e-02 -6.70140743e-01 -1.00674891e+00
-1.15689182e+00 -9.11302924e-01 -1.11489773e+00 2.93050021e-01
1.25083852e+00 -1.29825878e+00 -2.86066413e-01 6.03320062e-01
-1.45796466e+00 -5.05939901e-01 -8.88077263e-03 4.99759704e-01
-2.33570263e-01 3.76704931e-01 -1.10372984e+00 -9.68564212e-01
-7.41896391e-01 -9.84236717e-01 1.22368884e+00 1.97235290e-02
-1.94033366e-02 -1.08119857e+00 1.02807172e-01 6.07490480e-01
1.62299305e-01 -5.54268062e-01 6.02704048e-01 -1.30545521e+00
2.63178628e-02 -3.42583477e-01 -2.27304831e-01 6.63208485e-01
-9.80845019e-02 -1.67934909e-01 -1.25819981e+00 -2.69972682e-01
-5.44614494e-01 -5.08723915e-01 8.47776353e-01 6.72279596e-02
1.06011367e+00 -4.51332092e-01 -7.64481872e-02 3.10132086e-01
1.18905246e+00 2.95212626e-01 7.01573491e-01 2.16136858e-01
7.14406967e-01 5.56294918e-01 6.56967759e-01 2.90703446e-01
4.96586949e-01 5.87079465e-01 -2.54354089e-01 1.78692594e-01
-3.35260302e-01 -6.14700735e-01 5.30425072e-01 1.53049791e+00
4.60647970e-01 -6.14022195e-01 -1.26933050e+00 8.69970679e-01
-1.51316881e+00 -7.87425399e-01 -1.94236040e-01 1.93050146e+00
1.14112377e+00 4.03655358e-02 -2.79606581e-01 -1.17786052e-02
1.03520024e+00 3.43449414e-01 1.07402980e-01 -5.74608266e-01
-2.44600162e-01 2.45735660e-01 8.50862026e-01 5.81876934e-01
-1.26100338e+00 1.53531408e+00 6.14393520e+00 1.17491388e+00
-8.93511117e-01 4.86779839e-01 5.66067398e-01 5.40419102e-01
8.70515592e-03 -6.16379082e-02 -1.62680948e+00 7.02855766e-01
1.74854112e+00 4.33103383e-01 -4.69878549e-03 9.60951686e-01
-1.88340917e-01 2.14159340e-01 -4.27278817e-01 7.03578591e-01
1.28745347e-01 -1.09154916e+00 -2.64498711e-01 1.94105078e-02
6.94471955e-01 4.71151054e-01 -2.16181219e-01 5.65464914e-01
6.30371690e-01 -8.79119098e-01 8.02694201e-01 2.48527363e-01
1.21517241e+00 -7.30340242e-01 1.25857651e+00 1.85949072e-01
-1.41874039e+00 2.95329034e-01 -6.35114074e-01 2.47803405e-01
2.75542498e-01 6.10224962e-01 -1.13046992e+00 3.33124846e-01
5.64121187e-01 5.25243163e-01 -3.31815988e-01 8.27952266e-01
-6.27017260e-01 1.22988808e+00 -1.57386675e-01 -3.05449188e-01
4.98603910e-01 2.52321839e-01 8.52175653e-02 2.16337085e+00
1.72082141e-01 -1.86228409e-01 1.28334329e-01 2.38725767e-01
-4.58026171e-01 3.19886506e-01 -2.55799413e-01 -6.55351698e-01
7.53096581e-01 1.30982673e+00 -6.43517315e-01 -6.58910990e-01
-5.39224565e-01 1.19818068e+00 5.61315417e-01 3.26330096e-01
-5.64845204e-01 -8.09677660e-01 5.60204983e-01 -3.35724741e-01
5.49747527e-01 -4.76639867e-01 -1.93023369e-01 -1.14347398e+00
-2.63786167e-01 -6.19083345e-01 4.81805742e-01 -8.46900940e-01
-1.47519529e+00 8.32646191e-01 -2.53365099e-01 -7.95290589e-01
-1.05552174e-01 -7.83896506e-01 -4.11663175e-01 1.25079155e+00
-1.69959879e+00 -1.48218918e+00 1.53557390e-01 2.28278130e-01
8.09576035e-01 -3.39478016e-01 1.05288613e+00 5.53994119e-01
-5.92811823e-01 6.41691566e-01 2.24347204e-01 1.05767727e+00
7.78926015e-01 -1.28844368e+00 7.49315083e-01 7.23229527e-01
2.65704125e-01 6.32501960e-01 1.83343664e-01 -9.71598387e-01
-8.23766589e-01 -1.35996997e+00 1.87361729e+00 -3.39372694e-01
9.41401541e-01 -8.02140236e-01 -9.84900415e-01 8.91122460e-01
3.41282010e-01 2.33915865e-01 9.47163939e-01 9.83041078e-02
-5.90886295e-01 1.14606947e-01 -9.30455148e-01 1.56518579e-01
7.55815268e-01 -1.16915405e+00 -7.10843980e-01 2.84038246e-01
1.04290426e+00 -1.76575497e-01 -1.11839056e+00 -1.51830018e-01
4.60069060e-01 -1.60117418e-01 7.01305091e-01 -7.34512508e-01
6.47894442e-02 1.31889805e-02 -5.36048234e-01 -1.26193810e+00
-3.90342057e-01 -5.31670511e-01 3.01649749e-01 1.97069144e+00
6.65701747e-01 -2.86966860e-01 5.73333502e-01 1.68740779e-01
-4.17963445e-01 -2.76053846e-01 -1.12875998e+00 -9.36996877e-01
3.35921258e-01 -5.77833593e-01 5.96113920e-01 8.56231272e-01
1.11479662e-01 6.36336803e-01 -5.70378780e-01 1.77866906e-01
1.33098841e-01 -5.92120826e-01 3.04273158e-01 -8.86572003e-01
-1.21570274e-01 5.62904775e-02 -3.79449815e-01 -1.15157187e+00
7.70619750e-01 -9.80284393e-01 4.64668661e-01 -1.41697991e+00
1.91555262e-01 -4.97886896e-01 -5.42760968e-01 8.47484887e-01
-2.01524049e-01 2.79972643e-01 -8.04598927e-02 -1.44524910e-02
-6.33766949e-01 6.01942182e-01 7.60124505e-01 -1.82121992e-01
1.32518113e-01 -1.38801083e-01 -1.81385338e-01 4.66623217e-01
1.00063264e+00 -8.11589122e-01 1.14006571e-01 -6.24269783e-01
2.25972444e-01 -6.42752275e-02 -1.94074526e-01 -7.40565836e-01
8.60840827e-02 -1.41546770e-03 4.32862103e-01 -7.78237581e-01
4.31548119e-01 -6.69951737e-01 -2.54185259e-01 3.70158315e-01
-6.27173066e-01 1.03865437e-01 1.25127867e-01 2.41810903e-01
-2.36062258e-01 -6.09974325e-01 7.54114032e-01 -2.53490239e-01
-7.82137692e-01 2.05514535e-01 -7.82255352e-01 3.49973112e-01
4.65490311e-01 1.79551780e-01 -4.13108289e-01 -1.91925615e-01
-4.88733679e-01 -7.20122829e-02 4.92843892e-03 5.33225298e-01
2.34147757e-01 -1.32985628e+00 -7.16195345e-01 2.65203029e-01
3.40157062e-01 -4.58128899e-01 1.77586943e-01 3.67825300e-01
-5.34418464e-01 9.80439663e-01 -8.57902393e-02 -1.41691774e-01
-9.81887400e-01 3.86727631e-01 6.25953302e-02 -4.93578166e-01
-2.77434409e-01 9.69498456e-01 -1.77321911e-01 -7.24221945e-01
1.61006987e-01 -9.95884165e-02 -2.37652734e-01 1.15950637e-01
5.55578947e-01 2.75063276e-01 2.61331350e-01 -1.13819456e+00
-6.22989953e-01 1.83201596e-01 -7.02225417e-02 -4.54761297e-01
1.20711386e+00 -2.57650822e-01 1.00647904e-01 4.83978540e-01
1.39088428e+00 1.98070094e-01 -9.77262855e-01 -4.10003901e-01
3.07409257e-01 -2.17667252e-01 8.39643553e-02 -8.67498875e-01
-7.75535047e-01 1.42602694e+00 3.08056593e-01 -2.34918147e-01
4.90800768e-01 3.92737985e-02 1.08913600e+00 4.84195203e-01
2.17741191e-01 -1.33492613e+00 4.33898112e-03 1.21964931e+00
4.69083577e-01 -1.20410717e+00 -4.35379118e-01 -1.18342116e-01
-9.03422534e-01 8.51557136e-01 5.11763752e-01 1.83895841e-01
5.42047262e-01 4.54974234e-01 3.95197362e-01 3.01867068e-01
-7.10540116e-01 -3.05894703e-01 3.42029542e-01 8.72500241e-01
1.08642006e+00 2.09861457e-01 -2.39240214e-01 7.93734372e-01
-3.31109822e-01 -4.29066449e-01 2.53514230e-01 6.48796737e-01
-6.10696316e-01 -1.25289285e+00 -1.21024750e-01 1.40734777e-01
-1.04174805e+00 -6.79757059e-01 -1.68923080e-01 3.89475167e-01
-1.99833050e-01 1.10070741e+00 2.26959229e-01 -1.86645344e-01
-7.92583302e-02 4.83333141e-01 -6.87506497e-02 -9.05043125e-01
-9.36120868e-01 3.70979220e-01 4.96224284e-01 -2.05107495e-01
-6.36638403e-02 -3.29455465e-01 -1.39022028e+00 -9.94168222e-03
-4.43452895e-01 5.04410446e-01 8.96751463e-01 8.69890690e-01
2.20029175e-01 2.57272571e-01 3.46394598e-01 -4.92527366e-01
-3.59684974e-01 -1.56527710e+00 -6.26803696e-01 2.20086977e-01
1.08780406e-01 -2.79028416e-01 -3.11595444e-02 1.31936699e-01] | [9.96670150756836, 9.818185806274414] |
55951237-3f8f-4735-85fe-be25aa192810 | encyclopedic-vqa-visual-questions-about | 2306.09224 | null | https://arxiv.org/abs/2306.09224v1 | https://arxiv.org/pdf/2306.09224v1.pdf | Encyclopedic VQA: Visual questions about detailed properties of fine-grained categories | We propose Encyclopedic-VQA, a large scale visual question answering (VQA) dataset featuring visual questions about detailed properties of fine-grained categories and instances. It contains 221k unique question+answer pairs each matched with (up to) 5 images, resulting in a total of 1M VQA samples. Moreover, our dataset comes with a controlled knowledge base derived from Wikipedia, marking the evidence to support each answer. Empirically, we show that our dataset poses a hard challenge for large vision+language models as they perform poorly on our dataset: PaLI [14] is state-of-the-art on OK-VQA [37], yet it only achieves 13.0% accuracy on our dataset. Moreover, we experimentally show that progress on answering our encyclopedic questions can be achieved by augmenting large models with a mechanism that retrieves relevant information from the knowledge base. An oracle experiment with perfect retrieval achieves 87.0% accuracy on the single-hop portion of our dataset, and an automatic retrieval-augmented prototype yields 48.8%. We believe that our dataset enables future research on retrieval-augmented vision+language models. | ['Vittorio Ferrari', 'André Araujo', 'Fei Sha', 'Howard Zhou', 'Felipe Cadar', 'Arushi Goel', 'Lluis Castrejon', 'Jasper Uijlings', 'Thomas Mensink'] | 2023-06-15 | null | null | null | null | ['visual-question-answering-1', 'question-answering'] | ['computer-vision', 'natural-language-processing'] | [-2.93750733e-01 1.19734459e-01 -3.68251681e-01 -6.60000741e-02
-1.49221063e+00 -1.09811664e+00 6.50077403e-01 2.72973403e-02
-4.78569537e-01 4.06520635e-01 3.59156519e-01 -4.28641856e-01
1.34531707e-01 -7.02318192e-01 -1.16503286e+00 -1.61963299e-01
2.24050358e-01 8.21591496e-01 6.28733635e-01 -4.00602728e-01
3.05119246e-01 8.44236240e-02 -1.62141037e+00 7.70677805e-01
5.48633516e-01 9.90440249e-01 2.69679993e-01 1.09944975e+00
-2.82744974e-01 1.31111097e+00 -4.81067300e-01 -8.23611081e-01
2.32651621e-01 -1.00187182e-01 -1.36993229e+00 -2.12878048e-01
1.49407089e+00 -5.95102727e-01 -7.69101441e-01 8.40175986e-01
1.16138890e-01 -1.73943296e-01 6.97909534e-01 -1.63758934e+00
-1.62457442e+00 2.03690991e-01 -3.64287078e-01 3.83342624e-01
4.91270065e-01 6.73881173e-01 1.48669386e+00 -1.19781685e+00
1.24474204e+00 1.22174537e+00 2.63750523e-01 5.80719113e-01
-8.96092236e-01 -3.19433808e-01 8.39622542e-02 7.96653688e-01
-1.33098722e+00 -4.44856584e-01 4.28917915e-01 -4.74229246e-01
1.24510407e+00 2.52595961e-01 7.05234706e-01 8.23521197e-01
-3.06187242e-01 1.08763707e+00 1.00895369e+00 -3.53423059e-01
-6.43793344e-02 1.83397532e-01 4.99975622e-01 1.07768703e+00
2.11931750e-01 -5.75017110e-02 -5.62569380e-01 -1.07387170e-01
6.27944112e-01 -4.57158357e-01 -3.88299346e-01 -7.51687407e-01
-1.31903207e+00 8.78873646e-01 9.73170936e-01 -1.39814392e-01
-7.92088285e-02 6.11413777e-01 4.53207880e-01 6.14190876e-01
-1.58101231e-01 7.84460306e-01 -4.72780466e-01 1.85201660e-01
-5.85808158e-01 5.29814720e-01 7.09442377e-01 1.20025277e+00
9.78508174e-01 -2.67607391e-01 -4.86089021e-01 6.25582278e-01
2.59449363e-01 9.80166137e-01 8.33541825e-02 -1.64707732e+00
5.82313299e-01 6.43069327e-01 2.10382879e-01 -9.33625996e-01
1.26948291e-02 -7.19798077e-03 -2.51908869e-01 -1.93178803e-02
7.95102835e-01 4.56711620e-01 -1.15156388e+00 1.67571533e+00
2.11281776e-01 -2.02141136e-01 2.07583949e-01 1.05774748e+00
1.49892688e+00 7.97101915e-01 2.41539091e-01 1.84745342e-01
1.71634889e+00 -1.36405480e+00 -3.63877684e-01 -4.37667370e-01
4.70501184e-01 -6.97697461e-01 1.52474070e+00 1.60217896e-01
-1.01945543e+00 -4.99935120e-01 -7.28246868e-01 -7.36148655e-01
-5.87415159e-01 6.42222911e-02 6.18575454e-01 2.79617935e-01
-1.52127600e+00 -3.77643913e-01 -1.41385570e-01 -5.30336142e-01
4.71905828e-01 -4.19730581e-02 -5.25544286e-01 -7.27645755e-01
-9.13751960e-01 1.15684438e+00 4.39501926e-02 -1.93835303e-01
-1.44996715e+00 -7.57756591e-01 -9.14214313e-01 -1.79365978e-01
7.03541517e-01 -1.03133106e+00 1.36182690e+00 -6.83349311e-01
-5.06111085e-01 1.29134512e+00 -3.95530164e-01 -4.72961873e-01
2.10269436e-01 2.17386186e-02 -2.64453620e-01 8.41000736e-01
3.33927363e-01 1.18843985e+00 8.58840048e-01 -1.63770103e+00
-5.57976425e-01 -4.45735782e-01 6.67206466e-01 1.21589996e-01
-1.06055737e-02 -5.31258956e-02 -1.17515075e+00 -2.48490691e-01
-2.97555715e-01 -7.84835875e-01 6.70174211e-02 5.43200016e-01
-1.68556526e-01 -3.76837760e-01 7.08361268e-01 -9.12905395e-01
7.61028230e-01 -1.81792474e+00 6.80934731e-03 -3.94911729e-02
5.93085706e-01 2.13912368e-01 -7.09773362e-01 4.46693271e-01
5.30735195e-01 1.24572307e-01 7.47393817e-02 3.71666625e-02
8.66078809e-02 4.14882272e-01 -6.12094402e-01 2.11748078e-01
2.39487499e-01 1.73401821e+00 -1.02844894e+00 -9.08304632e-01
1.52511727e-02 2.06420451e-01 -5.60235500e-01 7.67546296e-02
-6.10468626e-01 -2.57856339e-01 -2.44122654e-01 1.01603925e+00
3.77634406e-01 -7.35730171e-01 -6.82395548e-02 -3.48448873e-01
2.37372935e-01 -9.63295475e-02 -5.20271122e-01 1.55931926e+00
-2.50561059e-01 1.05154705e+00 1.10529490e-01 -6.43647373e-01
4.86684352e-01 -5.65959886e-02 -2.64387637e-01 -1.03574741e+00
-1.34241238e-01 2.49681380e-02 -2.74217844e-01 -7.38785505e-01
8.97174120e-01 4.35869724e-01 -9.01875868e-02 2.42500126e-01
3.47603053e-01 -4.88281310e-01 2.97802746e-01 1.02336884e+00
1.14951575e+00 -1.85827315e-01 4.64522950e-02 9.71206184e-03
2.51424968e-01 9.19326186e-01 -1.23087205e-01 1.32837462e+00
-5.97910106e-01 5.85099757e-01 4.03520346e-01 -4.00370181e-01
-1.35430372e+00 -1.29504144e+00 2.22935185e-01 1.22569680e+00
3.62897009e-01 -4.83186662e-01 -4.94023025e-01 -7.76538432e-01
2.06967354e-01 4.50149983e-01 -8.76837671e-01 9.64466780e-02
-3.52247864e-01 2.21374296e-02 7.21732795e-01 7.49032617e-01
3.64445448e-01 -1.29632437e+00 -2.98769146e-01 -3.00221413e-01
-5.11087060e-01 -1.26696372e+00 -5.32603204e-01 -2.85865873e-01
-4.64223057e-01 -1.49684083e+00 -7.02983379e-01 -1.19943917e+00
4.58761036e-01 6.76268458e-01 1.82214940e+00 3.68061900e-01
-4.08820033e-01 1.21114707e+00 -4.33886468e-01 -2.00462386e-01
-4.54653442e-01 -1.69123709e-01 -4.68438178e-01 -5.75855434e-01
4.37410533e-01 1.66062623e-01 -6.92043483e-01 1.79608226e-01
-6.16367280e-01 -1.67330697e-01 4.94306475e-01 7.48669088e-01
8.26360226e-01 -8.39832008e-01 6.03110075e-01 -4.97368515e-01
4.03452456e-01 -3.87243629e-01 -6.72903001e-01 8.44401479e-01
-3.92800689e-01 2.78043393e-02 2.49540344e-01 -3.39760572e-01
-6.72339916e-01 -1.45503730e-01 5.74540719e-02 -7.40830362e-01
7.91125894e-02 2.53106117e-01 1.30584702e-01 -2.60780960e-01
9.10716057e-01 3.98300529e-01 -4.18032147e-02 -2.02742621e-01
1.20885110e+00 4.16190952e-01 1.14972508e+00 -5.18292248e-01
8.41994226e-01 5.46281040e-01 -4.06028718e-01 -7.83860803e-01
-1.03524661e+00 -8.69930625e-01 -3.03950161e-01 -2.44766816e-01
1.02831340e+00 -1.21758783e+00 -9.60474551e-01 2.02784240e-01
-1.22068548e+00 -3.90025765e-01 -3.00115168e-01 -1.42881662e-01
-6.24762118e-01 4.01188493e-01 -8.20425808e-01 -5.68575442e-01
-5.45121193e-01 -9.57099140e-01 1.11715019e+00 4.28708047e-02
7.42898211e-02 -6.18919253e-01 1.21365897e-01 1.04104197e+00
2.84373462e-01 -2.18626663e-01 1.06086874e+00 -3.25581878e-01
-1.20737350e+00 9.04227793e-02 -8.66897643e-01 1.84809193e-01
-5.62599838e-01 -4.71314974e-02 -9.76214647e-01 -3.19606632e-01
-7.92104781e-01 -1.15674245e+00 1.21544874e+00 2.64350493e-02
1.08518791e+00 -2.18972176e-01 -1.76140144e-01 2.51242310e-01
1.61069381e+00 -2.41740316e-01 6.00460887e-01 3.68214130e-01
1.02383792e+00 5.26724637e-01 7.89316237e-01 -6.92819878e-02
1.08036828e+00 4.71668392e-01 8.61925066e-01 1.27367884e-01
-7.25153089e-01 -5.19495726e-01 -5.73237659e-03 6.51373982e-01
2.09031045e-01 -2.79138267e-01 -1.14713275e+00 1.24538147e+00
-1.80088067e+00 -1.05388916e+00 -3.55953783e-01 1.82585049e+00
7.58814514e-01 -2.33280331e-01 8.05944800e-02 -4.06026244e-01
3.19992959e-01 2.09609777e-01 -6.99149251e-01 -1.82713911e-01
-3.83624464e-01 1.71898380e-01 4.89686221e-01 6.63764536e-01
-9.59766269e-01 1.34752643e+00 7.33131838e+00 7.13213742e-01
-6.02006614e-01 1.06202357e-01 1.76205173e-01 -7.20523521e-02
-7.38126874e-01 5.41097000e-02 -6.76200211e-01 -5.24779931e-02
5.64608514e-01 -1.12567537e-01 5.64359188e-01 9.50253069e-01
-5.74492216e-01 -1.69716433e-01 -8.74534309e-01 1.23844743e+00
5.27059793e-01 -1.84950995e+00 6.22948050e-01 -1.03603661e-01
7.52088904e-01 5.05425811e-01 1.91836879e-01 6.35485888e-01
5.62378526e-01 -1.10413778e+00 8.66262257e-01 5.07746339e-01
1.05018067e+00 -4.26241428e-01 5.59870243e-01 1.60520449e-01
-1.11363900e+00 -1.42392501e-01 -5.53460598e-01 9.89446044e-02
1.02185830e-02 -1.10714272e-01 -9.46856678e-01 2.40002856e-01
9.89329576e-01 3.50748420e-01 -1.19806743e+00 1.00670230e+00
-3.09908003e-01 5.33102870e-01 -1.20356508e-01 -2.35711500e-01
3.58385265e-01 3.63816142e-01 2.67421275e-01 9.25064862e-01
-1.07524090e-01 2.44127303e-01 1.33107826e-01 5.91088116e-01
-5.42996943e-01 2.80668382e-02 -7.68283725e-01 -6.06925674e-02
6.48914695e-01 9.84653771e-01 -3.19013178e-01 -8.00847411e-01
-5.72147667e-01 8.91754031e-01 9.18862820e-01 5.57331443e-01
-5.54788589e-01 -1.55727148e-01 5.62199116e-01 -1.29634276e-01
6.25559807e-01 -7.99933076e-02 9.08424184e-02 -1.43766713e+00
-1.30793564e-02 -1.11127448e+00 7.93167293e-01 -1.59907997e+00
-1.56347835e+00 5.10532260e-01 -1.76798061e-01 -6.84550285e-01
-2.49244690e-01 -8.81151736e-01 -6.94729611e-02 5.30527592e-01
-1.68370891e+00 -1.59680605e+00 -6.87091351e-01 9.87355888e-01
4.44576621e-01 3.21869627e-02 7.33124137e-01 4.66467859e-03
1.14651941e-01 5.58230221e-01 -1.01439403e-02 3.28479916e-01
9.39470768e-01 -1.34472668e+00 5.89799821e-01 5.89457870e-01
6.54018939e-01 4.87900287e-01 6.58005476e-01 -4.64038223e-01
-1.84927666e+00 -9.95386243e-01 8.40629876e-01 -1.42618012e+00
8.51685286e-01 -2.91925907e-01 -1.03423512e+00 9.54114735e-01
4.53415930e-01 3.32263619e-01 3.68648320e-01 4.93481234e-02
-1.14482844e+00 5.67786880e-02 -9.78559852e-01 6.53803587e-01
1.19595277e+00 -1.17451227e+00 -1.02585220e+00 3.65491807e-01
1.08995104e+00 -3.57489437e-01 -7.41659403e-01 2.05312625e-01
6.21128082e-01 -5.76256514e-01 1.25378752e+00 -9.39350188e-01
5.68284154e-01 -5.65613508e-01 -5.51486731e-01 -9.32030141e-01
-3.67564142e-01 1.77271724e-01 -4.40680861e-01 8.28029156e-01
4.89451140e-01 -1.65927112e-01 7.43943989e-01 4.63645667e-01
9.99807492e-02 -5.43047130e-01 -6.67424977e-01 -7.17232823e-01
1.64761961e-01 -6.16095066e-01 3.40381950e-01 9.39038455e-01
-3.15730125e-01 3.82717609e-01 -2.97342211e-01 2.07493573e-01
6.40928686e-01 3.98208797e-01 1.15563178e+00 -7.94156253e-01
-2.56966919e-01 -1.97501481e-01 -3.52297544e-01 -1.27165902e+00
3.22285034e-02 -9.04722691e-01 -1.12017252e-01 -2.14287663e+00
6.68141901e-01 -1.31561577e-01 -1.09781101e-01 6.27147853e-01
-2.15262070e-01 8.34219515e-01 5.04380107e-01 3.80396605e-01
-1.29472649e+00 3.97123337e-01 1.52052712e+00 -6.23589277e-01
3.15742791e-01 -9.03498411e-01 -8.85893464e-01 4.41213429e-01
3.93410474e-01 -5.68002984e-02 -5.50114155e-01 -8.27221215e-01
3.79342794e-01 1.25478432e-01 9.24870551e-01 -4.61003006e-01
3.33220154e-01 -6.07944950e-02 3.35752845e-01 -8.16431522e-01
5.74812889e-01 -4.34462905e-01 -3.30485076e-01 2.31791794e-01
-3.88388872e-01 2.26862028e-01 2.11608380e-01 6.91151738e-01
-3.00146818e-01 -1.77388135e-02 4.92292136e-01 -4.07721311e-01
-1.52922916e+00 3.65043104e-01 -5.54094911e-02 7.95280099e-01
8.73452187e-01 1.19020194e-01 -1.27523589e+00 -6.18537605e-01
-5.56825280e-01 7.34124541e-01 7.65299678e-01 6.00431025e-01
8.24254334e-01 -1.28984821e+00 -6.63981378e-01 -4.52848643e-01
1.00508773e+00 -3.48615348e-01 4.91077453e-01 2.67927676e-01
-6.26302779e-01 6.46626472e-01 -1.43725380e-01 -6.10747755e-01
-1.31409299e+00 1.17915559e+00 3.94409001e-02 -2.99137123e-02
-3.87033761e-01 1.03804672e+00 2.77554750e-01 -5.13033807e-01
1.96744546e-01 -1.25851303e-01 -2.09862292e-01 6.73318580e-02
6.64735436e-01 1.85963258e-01 -2.04654500e-01 -7.16050804e-01
-6.67082071e-01 7.59313762e-01 -2.19745621e-01 -2.21092090e-01
8.16415727e-01 -8.89666751e-02 -9.43474993e-02 1.31420746e-01
1.32543623e+00 -3.15977670e-02 -9.39308703e-01 -4.35883939e-01
-2.90025920e-01 -4.84791905e-01 -1.78324506e-01 -1.14105344e+00
-7.68969774e-01 8.17495227e-01 3.72151792e-01 -8.58308524e-02
8.33415031e-01 9.26890790e-01 6.96163058e-01 1.02381563e+00
5.15076697e-01 -7.51621723e-01 4.80814219e-01 5.53769827e-01
1.13463950e+00 -1.64725518e+00 -2.30506837e-01 -2.74187952e-01
-8.22472870e-01 5.44293582e-01 9.11030352e-01 -1.61229178e-01
2.41547927e-01 -3.60955596e-01 4.86106932e-01 -6.69669628e-01
-1.16306424e+00 -5.69524765e-01 5.66114247e-01 8.59554768e-01
-2.01022476e-01 2.48752460e-02 1.65158004e-01 1.50258303e-01
-2.52348781e-01 -1.22304551e-01 5.10537148e-01 7.11324990e-01
-6.84660196e-01 -4.64179784e-01 -3.04445922e-01 4.77213681e-01
-5.74847646e-02 -3.79711241e-01 -6.25751555e-01 1.15077424e+00
-2.60444999e-01 1.10614574e+00 4.59770709e-02 -1.09912761e-01
4.02348250e-01 -2.71140207e-02 7.53360629e-01 -4.21037823e-01
-3.07626575e-01 -6.18667424e-01 2.97889650e-01 -7.76370883e-01
-1.07883438e-01 -2.75934696e-01 -1.10419428e+00 -2.94479400e-01
5.70209417e-03 1.81189962e-02 4.92235363e-01 7.28883862e-01
4.25097018e-01 -6.62647933e-03 -7.52824619e-02 -1.96675509e-01
-4.16695237e-01 -6.01781607e-01 -1.90275848e-01 5.68081737e-01
6.19386733e-01 -4.26759988e-01 -2.87174761e-01 6.49259016e-02] | [10.90605640411377, 1.6614762544631958] |
57337739-ad7c-4302-9eff-7ac7f3373638 | diagonal-rnns-in-symbolic-music-modeling | 1704.05420 | null | http://arxiv.org/abs/1704.05420v2 | http://arxiv.org/pdf/1704.05420v2.pdf | Diagonal RNNs in Symbolic Music Modeling | In this paper, we propose a new Recurrent Neural Network (RNN) architecture.
The novelty is simple: We use diagonal recurrent matrices instead of full. This
results in better test likelihood and faster convergence compared to regular
full RNNs in most of our experiments. We show the benefits of using diagonal
recurrent matrices with popularly used LSTM and GRU architectures as well as
with the vanilla RNN architecture, on four standard symbolic music datasets. | ['Y. Cem Subakan', 'Paris Smaragdis'] | 2017-04-18 | null | null | null | null | ['music-modeling'] | ['music'] | [ 1.73200876e-01 -2.13792861e-01 7.25933462e-02 1.17948808e-01
-5.28256536e-01 -3.41358691e-01 5.41878939e-01 -7.84576833e-01
-4.43416357e-01 9.77261901e-01 3.97630513e-01 -5.15840530e-01
-1.86323151e-01 -5.25123954e-01 -7.11428642e-01 -5.46105385e-01
-1.97116122e-01 2.13972971e-01 -1.97023496e-01 -4.56527710e-01
2.03552544e-01 3.39482546e-01 -1.46632314e+00 2.41549000e-01
3.73773247e-01 7.83794105e-01 -2.37367172e-02 7.61681437e-01
2.72662461e-01 1.46915209e+00 -7.27051556e-01 -1.92220315e-01
3.23521197e-01 -5.32115996e-01 -6.38699949e-01 -5.66701114e-01
4.24819663e-02 -2.38218918e-01 -5.56447029e-01 7.47161388e-01
8.01912904e-01 8.06272745e-01 4.11237866e-01 -6.58459604e-01
-6.36161923e-01 1.47898650e+00 -2.85502076e-01 8.85308236e-02
2.46307254e-01 -4.02890891e-01 1.29007089e+00 -8.74426544e-01
5.23876667e-01 1.27635908e+00 1.00124586e+00 4.95077521e-01
-1.42193329e+00 -9.38999712e-01 -9.99963805e-02 3.65600348e-01
-1.56277990e+00 -6.32732928e-01 7.41431773e-01 3.16562168e-02
1.49795496e+00 2.51952499e-01 5.41547298e-01 1.77948630e+00
1.01411797e-01 9.16380167e-01 8.22048247e-01 -5.15493453e-01
-1.61841772e-02 -4.28274721e-01 2.89896816e-01 4.62676466e-01
-6.31774515e-02 2.37751365e-01 -3.88450205e-01 -2.09248409e-01
9.88926172e-01 1.46015853e-01 -2.65898347e-01 -2.23711371e-01
-1.37035239e+00 8.34569216e-01 4.45780188e-01 5.69631577e-01
-4.00031745e-01 8.67576241e-01 6.44413948e-01 6.33847356e-01
1.65587857e-01 5.76001108e-01 -3.57375354e-01 -4.84944701e-01
-1.17205679e+00 7.16814213e-03 8.99914086e-01 6.41884744e-01
3.87289524e-01 9.93800581e-01 -1.78198889e-01 1.35122621e+00
2.54608691e-02 4.77757961e-01 1.18572545e+00 -1.21229112e+00
2.20029980e-01 7.65950978e-02 -3.17362070e-01 -9.06990528e-01
-5.14525890e-01 -9.94952798e-01 -1.09910893e+00 3.46304588e-02
6.85423091e-02 -2.89764524e-01 -1.02615023e+00 1.77023184e+00
-6.33826196e-01 8.96975517e-01 3.49820614e-01 5.75601816e-01
7.84418643e-01 5.79797864e-01 -6.35984361e-01 -8.64302889e-02
6.17754698e-01 -1.19038010e+00 -7.80667961e-01 -1.49807185e-01
4.73144501e-01 -7.58794904e-01 9.22843695e-01 6.78041458e-01
-9.55187261e-01 -6.64809823e-01 -1.32188511e+00 -1.77099705e-01
-4.39786792e-01 4.88475353e-01 7.60231793e-01 4.79749709e-01
-1.44961548e+00 1.14692414e+00 -9.36595976e-01 -2.19517022e-01
-3.56231153e-01 5.46896160e-01 1.12206936e-02 3.64627719e-01
-1.37871242e+00 8.08028698e-01 6.13313138e-01 5.73984265e-01
-7.00634360e-01 -2.17706308e-01 -9.04738784e-01 1.10454515e-01
2.79755741e-01 -4.18761671e-01 1.58492875e+00 -5.54552615e-01
-2.05253744e+00 1.94470331e-01 -9.03811827e-02 -1.05339038e+00
2.27034435e-01 -6.71992362e-01 -4.19105709e-01 -3.38694364e-01
-4.28325832e-01 5.24567127e-01 4.73493129e-01 -6.03790224e-01
5.60063533e-02 1.62616432e-01 -3.07459563e-01 -3.79221514e-03
-3.78132649e-02 -3.19038004e-01 -2.84435809e-01 -1.15436614e+00
3.23616624e-01 -1.26515579e+00 -3.84156942e-01 -1.14073133e+00
-5.61134517e-01 -4.09806706e-03 6.88982785e-01 -5.70152342e-01
1.41587949e+00 -2.14192533e+00 3.33113611e-01 4.59842861e-01
-2.59262770e-01 3.25126708e-01 -3.25386226e-01 2.62881219e-01
-5.10376573e-01 -7.04157213e-03 6.57900646e-02 -3.27682376e-01
-4.12990898e-02 6.45860434e-01 -9.84261811e-01 4.21640538e-02
-8.84431452e-02 1.14708018e+00 -5.63868821e-01 1.76127464e-01
2.14957058e-01 7.18702972e-01 -4.66802210e-01 -2.30946973e-01
-9.14469510e-02 6.87703863e-02 -1.56589344e-01 5.29326141e-01
-6.73869699e-02 -1.93679050e-01 2.00640857e-01 9.25856233e-02
-5.76770604e-02 8.02229166e-01 -1.27835941e+00 1.86650264e+00
-4.99513894e-01 9.03311253e-01 -5.05094171e-01 -7.04691648e-01
1.01451087e+00 5.68270981e-01 1.00821003e-01 -7.36967504e-01
2.62474626e-01 4.53309745e-01 2.34996751e-01 2.89465725e-01
6.74109995e-01 1.24500126e-01 1.69779584e-01 6.49345756e-01
2.64179736e-01 3.30322146e-01 2.91622400e-01 -1.05779670e-01
1.14463806e+00 4.23377901e-01 3.21646094e-01 3.89168747e-02
2.03243777e-01 -6.30674720e-01 6.70618773e-01 1.12662029e+00
4.48068053e-01 6.88705683e-01 5.21812856e-01 -2.22853437e-01
-8.92291903e-01 -8.43595326e-01 2.10930556e-02 1.22423053e+00
-6.88858926e-01 -6.70077860e-01 -1.43783018e-01 -3.74244750e-02
-1.72366232e-01 1.05271924e+00 -7.46258080e-01 -3.03360764e-02
-9.52657163e-01 -5.22904217e-01 1.17403507e+00 6.04866266e-01
4.78923172e-01 -1.39073586e+00 -5.08757234e-01 4.42109734e-01
-1.47649258e-01 -1.06878006e+00 -9.83971581e-02 6.56329274e-01
-1.12610447e+00 -6.15166962e-01 -8.47244442e-01 -3.79391730e-01
-2.66674340e-01 2.44710058e-01 1.03096962e+00 -1.86649442e-01
2.25565769e-02 -1.19067496e-02 -6.05814695e-01 -4.05218713e-02
-2.80115455e-01 3.47310364e-01 9.13435146e-02 -3.61891031e-01
1.57009326e-02 -1.07406282e+00 -1.01780817e-01 1.89039782e-01
-6.40781105e-01 -8.00314322e-02 7.85410345e-01 1.03955889e+00
3.89540792e-01 -5.21811128e-01 3.86208326e-01 -8.70519221e-01
8.89756262e-01 -1.38051108e-01 -4.27697569e-01 1.34211689e-01
-7.20235527e-01 5.40704072e-01 5.23630857e-01 -4.48543012e-01
-5.06844640e-01 -2.96049565e-01 -3.44153076e-01 -9.64355946e-01
4.57254261e-01 7.10250795e-01 7.42255926e-01 6.14080951e-02
6.38644814e-01 2.74646670e-01 -2.03887939e-01 -5.48321843e-01
2.92814910e-01 3.67871135e-01 5.48755109e-01 -4.28588837e-01
3.27366829e-01 -1.14668310e-01 2.46815634e-05 -6.19351685e-01
-6.25065148e-01 -1.62417442e-01 -4.09781694e-01 3.84296291e-02
3.24157029e-01 -7.82500386e-01 -8.77800226e-01 2.30734184e-01
-1.09923089e+00 -6.22064590e-01 -4.75505441e-01 9.28076267e-01
-6.68113649e-01 2.47134581e-01 -1.16647696e+00 -9.56433296e-01
-5.48187912e-01 -1.18327332e+00 7.31213808e-01 -1.63513556e-01
-2.91228324e-01 -7.71694064e-01 4.97285873e-01 -3.99981499e-01
7.09562182e-01 1.05932444e-01 5.41872323e-01 -7.67443240e-01
-2.75911570e-01 -1.96357947e-02 -4.00249399e-02 5.87399423e-01
-1.15441367e-01 1.79444790e-01 -1.15333486e+00 -1.72657192e-01
-5.46473190e-02 -4.63880152e-01 1.53888941e+00 5.77517092e-01
7.84632444e-01 -7.48977438e-02 1.35455325e-01 7.59502888e-01
1.03357911e+00 3.51985961e-01 8.75732601e-01 3.08782190e-01
5.95780909e-01 -4.17606756e-02 1.03674121e-01 3.75644177e-01
-1.92280114e-01 5.89654088e-01 -8.54098145e-03 9.41947177e-02
-1.59723580e-01 -2.10408762e-01 8.59462798e-01 1.50088739e+00
-4.13052142e-01 -2.10066494e-02 -8.84417117e-01 1.50020540e-01
-2.17488885e+00 -1.19727671e+00 6.91775307e-02 1.99326336e+00
5.28228283e-01 1.42065525e-01 -3.77269951e-03 4.98814434e-01
4.55502510e-01 3.09139341e-01 -4.13382232e-01 -7.46490419e-01
-4.53469574e-01 7.18214154e-01 5.02587378e-01 3.45634878e-01
-9.32733893e-01 1.12405801e+00 7.90285397e+00 8.23338091e-01
-1.27971029e+00 1.69335425e-01 1.70489192e-01 -5.91195464e-01
6.70946017e-02 -1.95368156e-01 -6.51351929e-01 -3.21207531e-02
1.52099478e+00 2.28574812e-01 7.80640721e-01 8.51937115e-01
-5.69006540e-02 3.62933218e-01 -1.01365685e+00 1.25914896e+00
1.69377521e-01 -1.29079866e+00 1.21551558e-01 -4.88342121e-02
5.81525028e-01 7.42325842e-01 2.38716006e-01 8.57539594e-01
8.47842216e-01 -1.27686191e+00 6.50882721e-01 5.36819220e-01
7.04517424e-01 -1.02102447e+00 8.69684517e-01 -1.64791360e-01
-1.19485819e+00 -1.83374107e-01 -5.44000387e-01 -2.98673481e-01
-1.24062195e-01 4.49271947e-01 -8.52571905e-01 5.24788558e-01
4.99708325e-01 1.13208139e+00 -5.44338942e-01 8.44289780e-01
-2.96797454e-01 8.09427261e-01 -4.19732064e-01 1.20989466e-02
5.20090699e-01 -3.09137464e-01 5.99800110e-01 1.47101641e+00
4.38212812e-01 -4.33348268e-01 -1.70004725e-01 7.58964717e-01
-2.15300620e-01 -5.65522164e-02 -9.42874670e-01 -9.62162763e-02
2.45926246e-01 1.15498841e+00 -6.15827143e-01 -2.45135948e-01
-7.86272287e-02 9.37652111e-01 3.54262978e-01 7.71344304e-01
-7.79977143e-01 -6.36453986e-01 5.08084595e-01 -6.58910334e-01
5.14911950e-01 -4.48035270e-01 5.91202937e-02 -1.34060800e+00
-1.55449972e-01 -1.06104600e+00 4.35363591e-01 -9.84677732e-01
-8.46632063e-01 9.06830847e-01 -3.30398560e-01 -1.12714136e+00
-1.19188166e+00 -7.01785147e-01 -5.39561152e-01 6.38666451e-01
-1.29375243e+00 -7.90129125e-01 3.81689459e-01 6.69217229e-01
3.78269494e-01 -4.87259239e-01 1.20472431e+00 4.11991805e-01
-7.97810137e-01 7.96564102e-01 3.78221959e-01 3.86964977e-01
5.51138043e-01 -1.11203325e+00 7.78021336e-01 7.00432599e-01
8.29686761e-01 1.08645141e+00 6.48977816e-01 -3.95308286e-01
-1.25258541e+00 -8.35192025e-01 4.51062709e-01 -2.34356942e-03
1.05460513e+00 -3.21043402e-01 -7.04714119e-01 1.18520224e+00
2.72796720e-01 -3.92634213e-01 6.66987062e-01 7.77437508e-01
-5.99408209e-01 1.10918671e-01 -3.98624212e-01 9.59210575e-01
9.59246635e-01 -9.07159686e-01 -6.57264531e-01 -1.00814067e-01
9.70112979e-01 -4.17433590e-01 -5.53292930e-01 7.67853618e-01
8.97101760e-01 -1.00837874e+00 1.06715691e+00 -3.91457975e-01
-1.83140114e-02 -8.17605704e-02 -5.29042423e-01 -1.31585431e+00
-4.73191023e-01 -8.81467879e-01 -4.45706367e-01 7.33117640e-01
7.28220522e-01 -7.25212872e-01 6.88530385e-01 -5.01711480e-03
-3.02273482e-01 -5.75203538e-01 -1.08659852e+00 -8.09725106e-01
-2.41338179e-01 -9.25705314e-01 1.18535720e-01 9.06918108e-01
-3.62138338e-02 5.94219446e-01 -8.35674584e-01 -3.61185640e-01
9.16798785e-02 1.24506623e-01 6.60755813e-01 -1.16739166e+00
-6.59940720e-01 -7.02657521e-01 -4.80424315e-01 -9.34287012e-01
3.03981453e-01 -1.08968842e+00 -2.26602092e-01 -1.23660147e+00
-8.85954872e-02 1.42793238e-01 -9.43689704e-01 7.61117041e-01
4.38034058e-01 5.85349679e-01 1.92089841e-01 1.16118126e-01
-5.38399756e-01 5.34740865e-01 7.40279973e-01 -4.76149209e-02
-4.54659104e-01 -1.46464650e-02 -3.54367614e-01 7.38708496e-01
8.61542225e-01 -3.37642968e-01 -2.95667320e-01 -2.59623319e-01
5.58777034e-01 1.12108149e-01 3.68188471e-01 -1.44374621e+00
2.58522481e-01 5.91245472e-01 5.03049850e-01 -9.88857210e-01
7.72180259e-01 -2.86043495e-01 3.57058555e-01 7.10383654e-01
-4.31194484e-01 4.97969419e-01 4.76966113e-01 3.59366417e-01
-3.04599404e-01 -1.95755556e-01 4.40773934e-01 3.97454500e-02
-3.31345320e-01 -4.48110312e-01 -3.18372071e-01 -2.85101622e-01
2.75252908e-01 -1.90281600e-01 -8.74178112e-02 -6.14679635e-01
-1.05402219e+00 -5.84861226e-02 -6.08370006e-02 4.80100483e-01
7.82166183e-01 -1.44350386e+00 -7.26875901e-01 3.53792638e-01
-4.56401289e-01 -7.66948462e-01 -1.56854317e-01 9.42286372e-01
-2.83342630e-01 8.94007504e-01 -2.02487275e-01 -3.54078770e-01
-1.10440218e+00 2.30227917e-01 3.39938074e-01 -5.92918396e-01
-9.27668810e-01 7.44378924e-01 -3.85570228e-01 -6.98592782e-01
6.55239820e-01 -7.60877848e-01 -3.84489387e-01 2.25030035e-01
3.76770705e-01 5.78143537e-01 1.89935669e-01 -3.07189196e-01
-2.09418796e-02 5.43762982e-01 -1.78826973e-01 -4.86095309e-01
1.44551420e+00 4.67627764e-01 -2.46979877e-01 1.24288726e+00
9.01107371e-01 3.34400237e-02 -3.67044419e-01 -3.09504718e-01
1.82865024e-01 9.76640135e-02 1.04111113e-01 -6.31028891e-01
-1.12645483e+00 7.48732626e-01 4.45168644e-01 1.17516316e-01
8.48880470e-01 -5.73446751e-01 7.53924191e-01 1.28831017e+00
3.58693212e-01 -1.06954372e+00 -1.85749352e-01 1.24553335e+00
9.14572597e-01 -4.15221006e-01 -6.54251948e-02 6.27736211e-01
-4.93806452e-01 1.23427963e+00 3.22271101e-02 -6.09854579e-01
5.33647299e-01 1.96723312e-01 -5.18670343e-02 -2.78474316e-02
-1.11974931e+00 -3.21700871e-01 3.43116730e-01 -1.09291896e-01
8.30952883e-01 -3.90800536e-02 -8.94148201e-02 2.93985963e-01
-6.77931190e-01 8.54275078e-02 6.82083130e-01 6.05921328e-01
-2.35625967e-01 -1.11025226e+00 -3.50486636e-01 5.50912023e-01
-5.24851799e-01 -4.49374139e-01 -4.36316609e-01 9.01157022e-01
-4.12834167e-01 8.58098626e-01 2.59052720e-02 -8.65959585e-01
2.74298280e-01 4.30483222e-01 3.81163388e-01 -4.45001721e-01
-8.80634546e-01 4.03199822e-01 1.53001100e-01 -8.50989401e-01
-3.26991707e-01 -5.98508835e-01 -1.27873731e+00 -2.82921940e-01
-4.84458685e-01 3.18299919e-01 6.88009679e-01 6.59880877e-01
4.47478294e-01 1.13013422e+00 3.26859266e-01 -1.04493916e+00
-4.78873163e-01 -1.41464698e+00 -6.43798470e-01 -1.47351176e-01
3.31914127e-01 -6.17767930e-01 -2.58807003e-01 -5.98052740e-01] | [10.89930534362793, 6.365229606628418] |
ad6b51e7-e490-4572-af69-0d3c8e548df3 | unsupervised-feature-learning-via-non-1 | null | null | http://openaccess.thecvf.com/content_cvpr_2018/html/Wu_Unsupervised_Feature_Learning_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Wu_Unsupervised_Feature_Learning_CVPR_2018_paper.pdf | Unsupervised Feature Learning via Non-Parametric Instance Discrimination | Neural net classifiers trained on data with annotated class labels can also capture apparent visual similarity among categories without being directed to do so. We study whether this observation can be extended beyond the conventional domain of supervised learning: Can we learn a good feature representation that captures apparent similarity among instances, instead of classes, by merely asking the feature to be discriminative of individual instances? We formulate this intuition as a non-parametric classification problem at the instance-level, and use noise-contrastive estimation to tackle the computational challenges imposed by the large number of instance classes. Our experimental results demonstrate that, under unsu- pervised learning settings, our method surpasses the state-of-the-art on ImageNet classification by a large margin. Our method is also remarkable for consistently improving test performance with more training data and better network architectures. By fine-tuning the learned feature, we further obtain competitive results for semi-supervised learning and object detection tasks. Our non-parametric model is highly compact: With 128 features per image, our method requires only 600MB storage for a million images, enabling fast nearest neighbour retrieval at the run time. | ['Zhirong Wu', 'Yuanjun Xiong', 'Stella X. Yu', 'Dahua Lin'] | 2018-06-01 | null | null | null | cvpr-2018-6 | ['self-supervised-image-classification'] | ['computer-vision'] | [ 3.89905959e-01 -9.87294540e-02 -3.79210174e-01 -7.51062453e-01
-8.94310892e-01 -6.91935182e-01 8.88562202e-01 1.36386886e-01
-6.96517467e-01 4.43651110e-01 -6.02381788e-02 -1.50817066e-01
-3.52189064e-01 -6.61251724e-01 -9.44071949e-01 -6.72760010e-01
-9.67580676e-02 5.12059093e-01 3.85597646e-01 2.30234355e-01
2.96701580e-01 5.69810748e-01 -1.89830518e+00 4.92468923e-01
3.67758483e-01 1.49328470e+00 3.26403715e-02 4.66082811e-01
3.22476476e-02 6.49495661e-01 -3.74336392e-01 -3.17047447e-01
4.14271504e-01 -5.99868111e-02 -9.60036576e-01 2.80651569e-01
1.20628071e+00 -2.17026427e-01 -3.88600886e-01 9.74313200e-01
3.67948771e-01 7.96473846e-02 9.36521471e-01 -1.33185327e+00
-7.55452812e-01 2.43534312e-01 -3.64172161e-01 3.55405807e-01
-9.81213227e-02 1.50030285e-01 1.23531294e+00 -9.41711009e-01
3.94707471e-01 1.19760931e+00 8.17098141e-01 4.37647969e-01
-1.49115872e+00 -6.44305587e-01 2.04317465e-01 5.57194531e-01
-1.41583610e+00 -5.47834337e-01 5.01548648e-01 -4.40039605e-01
9.10316348e-01 2.44346961e-01 3.52815121e-01 7.76367605e-01
-1.95342794e-01 7.99522102e-01 1.25285995e+00 -4.60322887e-01
2.65337944e-01 2.18222126e-01 3.01978916e-01 7.25137591e-01
2.53584534e-01 2.08150923e-01 -5.07254243e-01 -6.76201880e-02
6.61734045e-01 1.85921684e-01 -1.61995053e-01 -7.84955680e-01
-1.26577735e+00 8.84683073e-01 7.04283178e-01 1.30384818e-01
-5.09238653e-02 1.53167561e-01 5.94195962e-01 5.35817087e-01
4.79769409e-01 5.06358743e-01 -6.15177751e-01 7.34740421e-02
-1.00099170e+00 -2.01778691e-02 7.87906468e-01 9.00050223e-01
9.70215261e-01 -1.63142294e-01 -8.06285515e-02 8.46314311e-01
-7.02122450e-02 3.53232503e-01 6.80806279e-01 -1.05538619e+00
3.83877493e-02 5.71011186e-01 -2.66568452e-01 -6.69101775e-01
-3.67871851e-01 -6.01148844e-01 -7.92534590e-01 2.97969580e-01
6.81751966e-01 2.41230175e-01 -1.02461457e+00 1.66345406e+00
1.90158874e-01 2.70875692e-01 -7.97358602e-02 7.77666688e-01
7.66334116e-01 3.70792717e-01 1.29488438e-01 -5.29807769e-02
1.37905121e+00 -9.86582637e-01 -1.06846765e-01 -5.08592725e-01
7.39182413e-01 -5.39368093e-01 1.11017573e+00 3.53736430e-01
-6.65785491e-01 -6.30825043e-01 -1.06255198e+00 -3.35794277e-02
-5.26170492e-01 1.69430941e-01 8.17396939e-01 4.71392751e-01
-1.11158407e+00 6.81133807e-01 -6.50664628e-01 -4.12217289e-01
8.45304489e-01 4.57439929e-01 -7.11389899e-01 -3.25077683e-01
-7.44030595e-01 9.03300941e-01 4.44021493e-01 -1.01557948e-01
-7.60899901e-01 -7.40748465e-01 -8.29325914e-01 1.11747451e-01
3.63959968e-01 -3.89477521e-01 1.11753607e+00 -1.28031754e+00
-1.00170434e+00 1.16986704e+00 -1.33164346e-01 -5.13423324e-01
3.50335687e-01 9.32311341e-02 -2.03586012e-01 1.79748937e-01
1.49696153e-02 9.52762544e-01 9.81965482e-01 -1.17562318e+00
-7.00122356e-01 -4.18398291e-01 -7.44106844e-02 1.63580477e-01
-5.68665922e-01 -1.67955771e-01 -2.26578042e-01 -4.63937014e-01
1.74561754e-01 -9.51753497e-01 -2.11839061e-02 5.37718058e-01
-1.03867285e-01 -5.89432955e-01 8.40463400e-01 1.63828768e-02
5.73000789e-01 -2.31903338e+00 -2.99081445e-01 2.13713288e-01
4.04045880e-01 4.18284088e-01 -4.07697409e-01 3.83686312e-02
-2.06950516e-01 -6.82640150e-02 -1.82281628e-01 -1.15067609e-01
1.17628023e-01 2.75839210e-01 -4.31412309e-01 6.30310059e-01
3.21436673e-01 1.13844252e+00 -7.54585087e-01 -4.29669946e-01
1.14667766e-01 4.12942559e-01 -4.17910904e-01 5.28479740e-02
7.38034993e-02 -1.02709889e-01 -2.11957812e-01 4.56690460e-01
5.64304054e-01 -5.90933621e-01 1.14535548e-01 -2.61814177e-01
3.22438717e-01 2.43383840e-01 -1.29707468e+00 1.49104643e+00
-4.27721858e-01 9.22480643e-01 -2.27683306e-01 -1.59055364e+00
7.70464003e-01 -7.04922453e-02 1.48692027e-01 -6.75426066e-01
-1.92400604e-01 2.95323521e-01 1.39282867e-01 -3.29423428e-01
1.03340317e-02 -4.22112569e-02 1.21140905e-01 4.54430461e-01
3.26249689e-01 5.95567189e-02 6.80956841e-02 -3.95448990e-02
1.03222144e+00 -2.35192746e-01 2.60234594e-01 -4.01365936e-01
2.17075318e-01 -1.40831629e-02 1.87871918e-01 1.09077656e+00
-2.54234791e-01 3.38836074e-01 3.64314675e-01 -7.16408610e-01
-9.64202702e-01 -1.21413767e+00 -5.63886642e-01 1.52610695e+00
1.19442627e-01 -9.19884518e-02 -5.02332449e-01 -1.05095506e+00
2.53316551e-01 2.98099279e-01 -7.07748830e-01 -1.94377214e-01
-2.99872369e-01 -7.98894048e-01 4.60917860e-01 8.53954136e-01
5.20660877e-01 -1.08571255e+00 -4.19273615e-01 -5.19807301e-02
2.81257719e-01 -1.18014193e+00 -2.16331840e-01 5.94176710e-01
-8.56883049e-01 -1.13742244e+00 -5.78577459e-01 -1.26192510e+00
8.11641634e-01 3.20692718e-01 1.28758073e+00 1.60243228e-01
-5.04040539e-01 4.16538149e-01 -8.90762731e-02 -2.08617464e-01
-1.00182772e-01 1.94509439e-02 7.73095489e-02 -5.95027953e-02
8.16947341e-01 -6.15521133e-01 -7.50890791e-01 5.59759676e-01
-8.93151999e-01 -2.57160276e-01 9.72473204e-01 1.13672578e+00
6.67852581e-01 5.39575033e-02 6.30443513e-01 -8.70505452e-01
2.00004146e-01 -2.43010208e-01 -4.27411795e-01 2.98389494e-01
-6.97017074e-01 3.03416103e-01 4.93439585e-01 -8.22827458e-01
-6.38162136e-01 3.85229707e-01 1.42382279e-01 -2.69672185e-01
-5.78533769e-01 1.52437583e-01 3.35464068e-02 -3.81976515e-01
8.25900376e-01 4.10743654e-01 6.86282739e-02 -4.18076813e-01
5.48804104e-01 7.27886558e-01 6.24963045e-01 -5.20402789e-01
8.84344041e-01 6.20150149e-01 2.27809567e-02 -7.22672224e-01
-1.14264965e+00 -6.92456126e-01 -8.83762717e-01 5.78874610e-02
5.93451142e-01 -9.72694695e-01 -5.99077463e-01 2.47366205e-01
-8.76879990e-01 -5.17984092e-01 -4.23022747e-01 4.40789312e-01
-5.97399175e-01 2.13261217e-01 -5.19461095e-01 -3.57489794e-01
-5.03680930e-02 -1.01868331e+00 1.04100931e+00 -9.14598480e-02
5.41931987e-02 -8.78489316e-01 -1.45148844e-01 1.94705471e-01
4.62763846e-01 -1.31627321e-01 1.00661099e+00 -1.11783814e+00
-6.06163144e-01 -4.63550597e-01 -7.78621674e-01 5.96524537e-01
2.59714182e-02 -4.23469990e-01 -1.12790489e+00 -4.45525587e-01
-2.21451581e-01 -9.39238548e-01 1.25603700e+00 2.20886007e-01
1.52866018e+00 -4.06026095e-01 -3.51534426e-01 6.87771857e-01
1.38413835e+00 -3.70691806e-01 4.22619164e-01 2.84219742e-01
6.74449623e-01 7.19549239e-01 3.50788862e-01 9.41540822e-02
1.43168136e-01 6.63976312e-01 4.46279764e-01 -2.18824938e-01
-2.89644778e-01 -2.74144765e-02 -9.71920267e-02 2.89161921e-01
1.02750868e-01 4.64107507e-05 -8.01209688e-01 6.49173677e-01
-1.69791746e+00 -9.19807196e-01 4.65910025e-02 2.32481146e+00
9.05905008e-01 2.61721849e-01 1.58650339e-01 2.51704276e-01
6.71521425e-01 5.70190437e-02 -6.21569633e-01 -9.55930799e-02
-6.04345649e-02 2.87160516e-01 6.43208921e-01 1.56165645e-01
-1.39989400e+00 7.68588841e-01 6.95176458e+00 1.16458678e+00
-1.24520862e+00 -6.62984850e-04 7.95680642e-01 -8.98737535e-02
8.09281319e-02 -2.72132754e-01 -7.86717832e-01 2.88972259e-01
8.46506059e-01 2.01702952e-01 4.04754013e-01 1.12139392e+00
-4.61053759e-01 1.67372629e-01 -1.65804040e+00 1.04456186e+00
2.95841098e-01 -1.42072499e+00 1.17504597e-01 1.62095040e-01
7.26838648e-01 3.54405612e-01 2.33333901e-01 3.44230831e-01
2.08517507e-01 -1.32130873e+00 4.93051738e-01 2.70110399e-01
8.89791608e-01 -5.84189415e-01 6.22378588e-01 3.68589550e-01
-9.54668462e-01 -2.69312948e-01 -6.94992244e-01 -1.80502772e-01
-4.17569816e-01 4.65691835e-01 -7.88940549e-01 -9.63425860e-02
7.64625072e-01 6.93126082e-01 -1.03738391e+00 1.30224693e+00
1.46746906e-02 6.78796291e-01 -3.90280575e-01 -9.96690169e-02
3.19829971e-01 4.02075410e-01 1.50191978e-01 1.34605467e+00
7.06508011e-02 -1.65440083e-01 3.04833978e-01 5.36320448e-01
-2.54152387e-01 -1.18290866e-02 -5.77354133e-01 1.70720637e-01
5.40290713e-01 1.29894686e+00 -8.09681773e-01 -4.82625693e-01
-4.07556564e-01 9.61696386e-01 7.59729922e-01 2.40772873e-01
-3.79916012e-01 -4.52523887e-01 3.97487015e-01 2.10536242e-01
6.11040711e-01 -3.37088737e-03 -2.16115609e-01 -9.52865243e-01
1.53233573e-01 -7.03852892e-01 4.26036894e-01 -5.63147902e-01
-1.73115253e+00 5.44465899e-01 -6.77365735e-02 -1.23547816e+00
-3.41735482e-01 -1.20715714e+00 -3.91697139e-01 3.70490313e-01
-1.65603387e+00 -1.21998227e+00 -2.86730111e-01 6.28768086e-01
4.13011640e-01 -1.21573567e-01 9.66146171e-01 1.68616250e-01
-1.85885340e-01 8.86576533e-01 4.50600833e-01 3.71373713e-01
8.93852711e-01 -1.41299987e+00 1.98065639e-01 4.63049322e-01
7.39369512e-01 3.58932555e-01 4.28814024e-01 -3.64223346e-02
-1.10621309e+00 -1.13260245e+00 8.05807531e-01 -7.01713443e-01
7.71495223e-01 -5.56154490e-01 -1.09844351e+00 6.74481571e-01
-1.86120525e-01 8.12018037e-01 6.52310789e-01 3.16006571e-01
-1.08563066e+00 -3.17231983e-01 -1.04928315e+00 1.78112894e-01
1.17352736e+00 -7.84493506e-01 -6.70924902e-01 6.27779007e-01
5.59193194e-01 -8.48413408e-02 -7.25674510e-01 4.23267812e-01
5.80956221e-01 -7.62084603e-01 1.19850612e+00 -8.58205914e-01
3.94241840e-01 -3.20355967e-02 -3.15773815e-01 -1.08677602e+00
-4.71072018e-01 -1.31798744e-01 4.71412092e-02 9.71483350e-01
4.16949928e-01 -6.53209746e-01 8.17781568e-01 4.61822271e-01
1.44225821e-01 -8.64958167e-01 -1.02615583e+00 -1.08697701e+00
2.27599189e-01 -4.76062179e-01 3.11604381e-01 9.54816937e-01
-1.56467453e-01 3.20932299e-01 -6.27270192e-02 1.50581628e-01
7.05784440e-01 2.57895082e-01 5.09738445e-01 -1.48873937e+00
-4.50884759e-01 -5.66356301e-01 -9.98323500e-01 -1.14562535e+00
5.23486853e-01 -1.03466773e+00 2.56228000e-01 -1.09477127e+00
6.15451455e-01 -5.66165745e-01 -5.11282206e-01 8.18174064e-01
-1.39626548e-01 9.50182915e-01 7.00759292e-02 4.34618175e-01
-1.02562594e+00 1.67608082e-01 9.43685651e-01 -2.90369123e-01
2.31712312e-01 -1.32537454e-01 -6.92437470e-01 7.72201836e-01
4.64874029e-01 -4.91091967e-01 -3.14645529e-01 -4.73089010e-01
-1.69249550e-01 -4.33400154e-01 8.23809505e-01 -1.10575736e+00
3.56921464e-01 9.35652256e-02 7.26210296e-01 -2.67450511e-01
3.09781730e-01 -7.72343159e-01 -4.02994633e-01 3.02336663e-01
-9.08564448e-01 -1.82934508e-01 1.53560489e-01 8.21455240e-01
-1.45267189e-01 -2.10899323e-01 8.48509252e-01 1.80688854e-02
-1.01948786e+00 3.64358306e-01 -1.42290354e-01 2.92435557e-01
8.21018815e-01 -2.97401249e-01 -5.27181625e-01 -2.07257152e-01
-6.28818572e-01 -8.18798726e-04 4.23130602e-01 2.59210229e-01
4.68775779e-01 -1.36406124e+00 -6.47395074e-01 2.76211590e-01
4.89949822e-01 -1.36965573e-01 -1.46507164e-02 5.23856223e-01
-4.17956710e-02 4.69591409e-01 -1.51519254e-01 -9.93144989e-01
-1.24020255e+00 5.99434495e-01 3.33125174e-01 -6.77803755e-02
-5.19616902e-01 9.95359063e-01 5.39899826e-01 -5.38705051e-01
4.82722610e-01 4.88389134e-02 3.49664427e-02 -3.98516543e-02
6.33051395e-01 8.98947343e-02 2.99826443e-01 -6.04255319e-01
-4.21044022e-01 5.26997685e-01 -3.75042081e-01 2.27275699e-01
1.38065612e+00 1.25534192e-01 6.70980886e-02 3.51769239e-01
1.63673353e+00 -4.60880429e-01 -1.59370589e+00 -5.97711384e-01
1.06779411e-01 -4.16090190e-01 2.26107761e-01 -9.98842001e-01
-1.00715613e+00 9.11063850e-01 9.32154596e-01 1.69915870e-01
9.72138822e-01 4.26440060e-01 4.29546058e-01 9.72954452e-01
2.32759356e-01 -9.26365077e-01 2.18471140e-01 3.91627550e-01
6.25866711e-01 -1.66199136e+00 -2.09491104e-02 -2.38998204e-01
-4.45888996e-01 1.10809207e+00 5.83876550e-01 -5.19397736e-01
7.27022588e-01 1.05519503e-01 7.23796710e-02 -2.18402177e-01
-7.93078601e-01 -2.84540594e-01 8.18956137e-01 7.41768479e-01
2.45941296e-01 -1.37327462e-02 1.85477704e-01 2.64874667e-01
-1.94509979e-02 -2.43297875e-01 7.51315951e-02 6.82298720e-01
-7.28559852e-01 -8.70231211e-01 -5.01280650e-03 8.97871137e-01
-3.30248386e-01 -2.16537684e-01 -3.48139614e-01 8.49262655e-01
4.61545587e-02 7.61852622e-01 5.33734500e-01 -2.47142807e-01
1.05981164e-01 1.35884419e-01 5.04831433e-01 -6.35229766e-01
-1.17685691e-01 -2.89521933e-01 -9.64486524e-02 -6.43526077e-01
-5.01688421e-01 -4.49329317e-01 -8.58618259e-01 -9.39983502e-02
-4.19995040e-01 1.81776769e-02 5.33680260e-01 1.16239119e+00
2.83536643e-01 1.25276580e-01 6.31037533e-01 -9.50460732e-01
-9.99658585e-01 -7.68241405e-01 -4.22802269e-01 5.54784775e-01
4.14491087e-01 -6.89185083e-01 -6.44580066e-01 -7.33476691e-03] | [9.54239273071289, 2.563886880874634] |
e782d971-f359-4bb8-a570-e8ae8c309e07 | mixed-modality-representation-learning-and | 2210.05197 | null | https://arxiv.org/abs/2210.05197v1 | https://arxiv.org/pdf/2210.05197v1.pdf | Mixed-modality Representation Learning and Pre-training for Joint Table-and-Text Retrieval in OpenQA | Retrieving evidences from tabular and textual resources is essential for open-domain question answering (OpenQA), which provides more comprehensive information. However, training an effective dense table-text retriever is difficult due to the challenges of table-text discrepancy and data sparsity problem. To address the above challenges, we introduce an optimized OpenQA Table-Text Retriever (OTTeR) to jointly retrieve tabular and textual evidences. Firstly, we propose to enhance mixed-modality representation learning via two mechanisms: modality-enhanced representation and mixed-modality negative sampling strategy. Secondly, to alleviate data sparsity problem and enhance the general retrieval ability, we conduct retrieval-centric mixed-modality synthetic pre-training. Experimental results demonstrate that OTTeR substantially improves the performance of table-and-text retrieval on the OTT-QA dataset. Comprehensive analyses examine the effectiveness of all the proposed mechanisms. Besides, equipped with OTTeR, our OpenQA system achieves the state-of-the-art result on the downstream QA task, with 10.1\% absolute improvement in terms of the exact match over the previous best system. \footnote{All the code and data are available at \url{https://github.com/Jun-jie-Huang/OTTeR}.} | ['Nan Duan', 'Daxin Jiang', 'Ming Gong', 'Qian Liu', 'Wanjun Zhong', 'JunJie Huang'] | 2022-10-11 | null | null | null | null | ['open-domain-question-answering'] | ['natural-language-processing'] | [-9.42301750e-02 -1.32895872e-01 -2.66201079e-01 -2.40509152e-01
-2.09758282e+00 -5.18965602e-01 4.22974676e-01 1.15639992e-01
-3.16538215e-01 8.47158968e-01 4.36524630e-01 -2.17221990e-01
-3.08278084e-01 -6.12868071e-01 -8.15806210e-01 -6.15454912e-01
4.91503388e-01 7.89548695e-01 1.50370941e-01 -7.49266148e-01
1.10254250e-01 -1.56556666e-01 -1.45635986e+00 6.51465654e-01
1.32997608e+00 1.31893194e+00 1.56861275e-01 2.04914972e-01
-3.40493023e-01 7.56628036e-01 -4.04827952e-01 -7.73288429e-01
1.25672296e-01 -2.31794059e-01 -7.80231535e-01 -3.18288147e-01
5.71367919e-01 -2.77497619e-01 -6.34514451e-01 1.00885344e+00
7.66924500e-01 1.45801127e-01 6.44944072e-01 -9.86741424e-01
-8.80114317e-01 5.39597154e-01 -6.53404474e-01 3.43465626e-01
5.36624432e-01 -6.68527856e-02 1.48200214e+00 -1.29228914e+00
5.86341202e-01 1.04912305e+00 1.59809813e-01 5.02170026e-01
-5.84967792e-01 -6.91602945e-01 -1.71554565e-01 4.75115746e-01
-1.54720390e+00 -5.90184629e-01 6.68512344e-01 1.35542616e-01
5.34517467e-01 4.68123794e-01 -2.38454174e-02 9.28530872e-01
-2.60896921e-01 1.20927596e+00 9.73020554e-01 -4.80117321e-01
-2.43549898e-01 1.09967977e-01 1.87286198e-01 7.62499869e-01
1.21181078e-01 -3.44442427e-01 -7.77141035e-01 -2.41288498e-01
3.32111746e-01 -5.04141226e-02 -4.51709390e-01 -1.15656309e-01
-1.28724647e+00 6.21577203e-01 4.97911245e-01 2.03602508e-01
-3.82882267e-01 -6.59327134e-02 5.66937029e-01 4.79511619e-01
4.03147668e-01 1.98610157e-01 -5.68393409e-01 2.45789606e-02
-8.22770298e-01 2.75551051e-01 5.61087608e-01 9.93516922e-01
7.23820150e-01 -2.43848398e-01 -4.81729865e-01 1.31334674e+00
3.87608290e-01 1.08843362e+00 5.83373010e-01 -7.45597720e-01
1.22349405e+00 5.71266770e-01 9.23836231e-02 -7.75206089e-01
5.63577488e-02 -4.60149318e-01 -8.53231430e-01 -8.51325512e-01
3.61408025e-01 -1.28227491e-02 -9.66123283e-01 1.35168695e+00
3.63775760e-01 -3.72895330e-01 3.53759587e-01 1.25655198e+00
1.48777449e+00 1.04105198e+00 3.08223609e-02 -2.20188051e-01
1.74266863e+00 -1.09003770e+00 -9.28173244e-01 -2.16679260e-01
6.68107152e-01 -8.90365481e-01 1.40719676e+00 1.72065973e-01
-1.09057641e+00 -2.21045822e-01 -7.71026850e-01 -5.23008883e-01
-4.35598046e-01 4.99244094e-01 4.04758215e-01 4.18852776e-01
-6.02754891e-01 -2.63774008e-01 -3.42994690e-01 4.36386792e-03
4.23550665e-01 7.29981661e-02 -3.22108328e-01 -7.69327819e-01
-1.62349272e+00 6.17788613e-01 2.15142623e-01 1.86445087e-01
-8.01499903e-01 -7.60217726e-01 -8.20309103e-01 2.09597990e-01
8.55540216e-01 -8.01068604e-01 1.28774953e+00 -4.44183141e-01
-1.05116832e+00 6.80453777e-01 -4.23073709e-01 -1.69549063e-01
1.32860199e-01 -4.39784825e-01 -5.63540518e-01 4.62886751e-01
2.31080398e-01 5.12591362e-01 7.24714220e-01 -1.29051650e+00
-3.15425843e-01 -6.56731009e-01 1.69108901e-02 4.51489389e-01
-5.42780697e-01 -1.22053422e-01 -1.10380757e+00 -7.20489383e-01
2.69930452e-01 -7.19979525e-01 1.01907529e-01 -2.63390392e-01
-4.36580241e-01 -5.79847574e-01 3.77723068e-01 -9.82087612e-01
1.35121357e+00 -2.07213593e+00 -1.94378290e-02 2.30126232e-01
4.84829023e-02 1.96669027e-01 -2.37737179e-01 7.88564324e-01
3.13937932e-01 -1.74945012e-01 -2.49360129e-01 -2.32442155e-01
2.25336194e-01 7.14953691e-02 -6.93045199e-01 1.47342205e-01
6.58669621e-02 1.08810616e+00 -6.81461990e-01 -8.81785929e-01
-1.86207876e-01 3.54063958e-01 -2.26113528e-01 1.10594936e-01
-5.39553463e-01 2.97963053e-01 -9.27468181e-01 1.16456234e+00
5.99043965e-01 -5.27343869e-01 -3.91530979e-04 -3.92770380e-01
4.07456964e-01 3.76081735e-01 -9.47786808e-01 2.00060129e+00
-5.05748510e-01 2.55051583e-01 -1.87390819e-02 -8.01261783e-01
8.25755775e-01 4.34780180e-01 2.74964780e-01 -1.39861524e+00
2.52311707e-01 6.30116701e-01 -4.59396660e-01 -5.76830804e-01
8.01370621e-01 -3.89659368e-02 -3.06789905e-01 3.89891118e-01
1.12515993e-01 -5.04662581e-02 4.70803767e-01 7.20489144e-01
8.52683425e-01 -2.38459669e-02 -1.37884229e-01 1.21594608e-01
8.46793354e-01 1.89859852e-01 2.67555356e-01 6.11425102e-01
1.49146274e-01 5.29217303e-01 3.01386327e-01 1.89973935e-01
-7.05953956e-01 -9.72953916e-01 -1.61187500e-01 1.28601778e+00
3.54039013e-01 -4.06746626e-01 -3.38815898e-01 -6.92022443e-01
-7.68493861e-02 5.40556550e-01 -3.91101956e-01 -7.78319165e-02
-4.77582157e-01 -4.91605788e-01 6.61231637e-01 3.24558944e-01
5.74030936e-01 -8.74272645e-01 8.04975331e-02 -2.13752806e-01
-9.95233417e-01 -1.06166387e+00 -5.98591149e-01 -2.64827132e-01
-9.01473463e-01 -9.62396920e-01 -1.10258603e+00 -8.60265970e-01
5.54185569e-01 4.78065044e-01 1.24928057e+00 2.02659115e-01
-4.30616587e-02 5.13573885e-01 -6.25250280e-01 -1.18336365e-01
-1.99243985e-02 1.79897010e-01 -3.40921372e-01 -2.95741912e-02
3.31989825e-01 -2.21230671e-01 -8.47414553e-01 3.58432770e-01
-1.16402304e+00 -2.86325693e-01 9.07282054e-01 1.10670328e+00
8.57846797e-01 -2.10623920e-01 9.65071201e-01 -6.44536793e-01
8.68216634e-01 -7.50787497e-01 -4.77201223e-01 7.41410851e-01
-6.64212942e-01 1.40134618e-01 4.30616230e-01 -1.47562727e-01
-1.36303926e+00 -4.42828149e-01 -2.78158814e-01 -5.15826046e-01
2.76489884e-01 9.31631565e-01 -1.97193474e-01 2.24067569e-01
5.30985773e-01 4.40974832e-01 -2.57712722e-01 -6.12420797e-01
6.53623104e-01 9.23291743e-01 5.64792156e-01 -9.21272874e-01
7.99731374e-01 4.36945170e-01 -2.80703306e-01 -3.38395119e-01
-1.22585964e+00 -9.11607265e-01 -3.45401987e-02 -2.51564354e-01
4.74105597e-01 -1.30982888e+00 -6.14356637e-01 1.25474975e-01
-7.66799092e-01 2.92491000e-02 -1.26028329e-01 3.49131495e-01
-2.00461671e-01 5.91692388e-01 -6.99084818e-01 -7.54552543e-01
-9.01530683e-01 -8.41190934e-01 1.38962352e+00 2.08252549e-01
4.90208507e-01 -5.96862018e-01 2.03441847e-02 1.30538273e+00
2.19070256e-01 -4.17851895e-01 8.68928075e-01 -7.62462795e-01
-7.36469805e-01 -3.44021440e-01 -5.29258490e-01 2.15362757e-01
-2.78578609e-01 -5.21296859e-01 -9.11866844e-01 -2.38190696e-01
-3.07596952e-01 -1.08804190e+00 1.09707904e+00 -4.58245836e-02
1.18393600e+00 -1.94030970e-01 -5.23436777e-02 -1.62609033e-02
1.35536337e+00 -1.83109313e-01 7.11215734e-01 2.81005353e-01
5.95257819e-01 5.83020627e-01 1.30130208e+00 5.35148859e-01
7.04683840e-01 5.01339674e-01 2.62820244e-01 1.63749069e-01
-1.25191763e-01 -3.47158760e-01 2.11708620e-01 1.07785952e+00
2.11088434e-01 -3.96278501e-01 -7.86474168e-01 8.19053710e-01
-1.72730899e+00 -7.39036202e-01 -1.07425503e-01 2.09182787e+00
1.23449814e+00 -7.68301114e-02 -1.77997306e-01 5.07294014e-03
3.58082861e-01 1.99914753e-01 -4.93031949e-01 2.55634218e-01
-3.32957327e-01 2.41631702e-01 2.10025385e-01 1.90038562e-01
-6.72787011e-01 9.01637852e-01 4.57393503e+00 1.51731491e+00
-7.45265663e-01 3.22403044e-01 5.26445746e-01 -1.07182950e-01
-6.65768743e-01 -1.31208435e-01 -8.67034018e-01 2.80887485e-01
8.31083000e-01 -1.60343871e-01 3.81879061e-01 4.98347014e-01
-2.60631084e-01 -1.87101290e-01 -7.45291770e-01 9.99311030e-01
4.25281882e-01 -1.21091557e+00 3.90545368e-01 -2.72307128e-01
6.04024827e-01 6.05921373e-02 1.82488933e-01 8.67188811e-01
-9.70718116e-02 -7.43028581e-01 5.30208647e-01 5.60847223e-01
8.76450181e-01 -6.06865704e-01 8.84721935e-01 3.87152284e-01
-1.08097792e+00 3.48784551e-02 -2.91514814e-01 5.65103531e-01
2.22483933e-01 6.67696297e-01 -4.59971011e-01 1.09300041e+00
8.59248459e-01 4.17291909e-01 -5.76514661e-01 8.02606821e-01
-1.29816994e-01 6.12627685e-01 -3.84287864e-01 -9.35376287e-02
1.81465074e-01 -1.62328199e-01 4.17438149e-01 7.90431440e-01
1.66084379e-01 3.69301736e-01 -1.56351507e-01 5.79267144e-01
-6.18244469e-01 4.71896410e-01 -2.27063373e-01 -1.26203358e-01
6.38721287e-01 1.12966561e+00 -1.41754895e-01 -3.99495274e-01
-4.50162917e-01 7.08669126e-01 3.79821420e-01 4.25426543e-01
-7.61910498e-01 -4.48819906e-01 3.43107656e-02 -2.93883324e-01
4.58764017e-01 4.51583974e-02 5.93910925e-02 -1.38699317e+00
7.23674238e-01 -1.20928597e+00 1.01707530e+00 -1.12469208e+00
-1.53917611e+00 5.52646041e-01 2.76828855e-02 -1.37025714e+00
-3.16434093e-02 -2.48024985e-01 1.46863818e-01 7.45666027e-01
-1.87380397e+00 -1.17420542e+00 -2.40184456e-01 1.03925335e+00
6.48337543e-01 -1.65861428e-01 5.44492185e-01 8.33167791e-01
-4.35965896e-01 8.48738670e-01 3.52623641e-01 1.30109951e-01
1.04765511e+00 -8.42579186e-01 -2.72309870e-01 5.29034734e-01
3.54567677e-01 8.11870217e-01 3.67137313e-01 -6.16624355e-01
-2.00356603e+00 -7.72324026e-01 8.40957522e-01 -4.78600621e-01
7.47905195e-01 -9.36098099e-02 -1.01430750e+00 4.55786496e-01
3.69437426e-01 -2.51343064e-02 6.54524386e-01 1.81913421e-01
-6.05852425e-01 -4.45653886e-01 -9.41016555e-01 5.13662219e-01
5.49346924e-01 -7.82670259e-01 -8.66254032e-01 5.63498437e-01
9.14684355e-01 -6.60580635e-01 -1.09856343e+00 5.98541498e-01
4.23463434e-01 -4.84215558e-01 1.13370717e+00 -4.42095190e-01
5.25830805e-01 -2.40168020e-01 -4.58963424e-01 -8.47489595e-01
2.72312313e-01 -2.38054916e-01 -2.97848731e-01 1.32226944e+00
6.70321405e-01 -5.67938864e-01 5.88164032e-01 3.47291201e-01
-1.52162999e-01 -9.69280958e-01 -1.19195902e+00 -5.26420057e-01
7.55180642e-02 -3.83508593e-01 5.43720484e-01 8.54098380e-01
9.57384631e-02 6.54907823e-01 -3.81700963e-01 2.10833043e-01
4.86555934e-01 3.51977736e-01 6.36973798e-01 -6.92025423e-01
-8.96613747e-02 2.54952107e-02 3.08892816e-01 -1.57619536e+00
5.29942065e-02 -8.41618657e-01 4.85258251e-02 -1.71873498e+00
3.77967805e-01 -6.21733785e-01 -4.09018368e-01 3.52210552e-01
-3.04390311e-01 3.25503349e-01 1.17332771e-01 5.00473201e-01
-1.18429101e+00 1.15839517e+00 1.54621911e+00 -3.09360504e-01
2.24064603e-01 -2.91883200e-01 -7.84146190e-01 -4.97335345e-02
5.50346017e-01 -4.53383505e-01 -4.93806928e-01 -5.90590179e-01
5.38751185e-01 6.31587982e-01 3.63529533e-01 -6.56647623e-01
3.24910551e-01 5.75519614e-02 2.26320270e-02 -1.03038478e+00
7.67574787e-01 -7.89875507e-01 -4.43671048e-01 -6.98158005e-03
-4.35503960e-01 -3.03302868e-03 2.24245399e-01 7.69730687e-01
-6.44682646e-01 -2.22598538e-01 1.58062547e-01 -3.66015285e-02
-5.21427453e-01 2.97501653e-01 -9.58505645e-02 5.92603087e-01
3.66368175e-01 4.40927476e-01 -8.53355289e-01 -6.49725258e-01
-3.28066885e-01 7.27748930e-01 -1.03414416e-01 4.56124783e-01
7.95349836e-01 -1.30364025e+00 -8.79794240e-01 -2.39993155e-01
6.02956712e-01 7.87321925e-02 6.74046814e-01 1.11665201e+00
-1.14133231e-01 8.41021478e-01 2.38010794e-01 -4.15086329e-01
-1.13936031e+00 3.60659212e-01 -4.51859739e-03 -6.36089623e-01
-2.60682225e-01 6.72490597e-01 -5.05072214e-02 -6.60754621e-01
3.02843839e-01 4.70648929e-02 -3.22852135e-01 2.74494719e-02
4.33716893e-01 2.44462937e-01 4.27677333e-01 -4.81125534e-01
-1.60569042e-01 3.00913185e-01 -3.19430590e-01 -2.45436594e-01
1.04717004e+00 -3.21604997e-01 -2.22155571e-01 2.09950283e-01
1.06582665e+00 2.05236807e-01 -4.23133016e-01 -6.75104678e-01
-2.43277311e-01 -5.49047530e-01 1.29273072e-01 -1.06943834e+00
-1.13305283e+00 8.40898931e-01 5.21820128e-01 -1.56998694e-01
1.26135969e+00 2.98354775e-01 1.38872504e+00 8.80744219e-01
3.08450133e-01 -8.10375452e-01 2.81041533e-01 4.73789692e-01
1.05298877e+00 -1.53297973e+00 1.08449131e-01 -4.22244042e-01
-8.60007524e-01 6.82551265e-01 6.26999855e-01 3.39073122e-01
4.49082494e-01 -5.47680974e-01 2.49513239e-01 -7.09245801e-01
-8.64554644e-01 -4.20069546e-01 6.96070254e-01 3.96513268e-02
2.91442424e-01 -2.36023992e-01 -4.71466035e-01 6.56285524e-01
7.79210776e-02 -5.48991337e-02 6.07768223e-02 9.40063596e-01
-2.72810280e-01 -9.84662116e-01 -5.48679769e-01 6.94581270e-01
-4.88460064e-01 -5.26148140e-01 -1.81348160e-01 6.40075922e-01
-5.02404034e-01 1.14492607e+00 -3.99353445e-01 -4.96345721e-02
3.58478099e-01 1.08896889e-01 4.32783753e-01 -2.69123763e-01
-5.01720965e-01 7.62035549e-02 4.85941678e-01 -3.22073251e-01
-3.87640744e-01 -5.70372105e-01 -1.23863924e+00 -1.24400802e-01
-6.51302934e-01 5.97369850e-01 5.38953602e-01 9.21190858e-01
4.73148495e-01 2.82658637e-01 5.25409758e-01 -7.22152144e-02
-8.01661789e-01 -1.10287309e+00 -4.46715266e-01 3.44148278e-01
2.02970326e-01 -6.05557203e-01 -2.99119860e-01 -1.78958446e-01] | [11.345511436462402, 7.745286464691162] |
856a367a-a558-4283-9986-87d9886a9775 | optimistic-bounds-for-multi-output-learning | null | null | https://proceedings.icml.cc/static/paper_files/icml/2020/5808-Paper.pdf | https://proceedings.icml.cc/static/paper_files/icml/2020/5808-Paper.pdf | Optimistic bounds for multi-output learning | We investigate the challenge of multi-output learning, where the goal is to learn a vector-valued function based on a supervised data set. This includes a range of important problems in Machine Learning including multi-target regression, multi-class classification and multi-label classification. We begin our analysis by introducing the self-bounding Lipschitz condition for multi-output loss functions, which interpolates continuously between a classical Lipschitz condition and a multi-dimensional analogue of a smoothness condition. We then show that the self-bounding Lipschitz condition gives rise to optimistic bounds for multi-output learning, which are minimax optimal up to logarithmic factors. The proof exploits local Rademacher complexity combined with a powerful minoration inequality due to Srebro, Sridharan and Tewari. As an application we derive a state-of-the-art generalization bound for multi-class gradient boosting. | ['Ata Kaban', 'Henry Reeve'] | null | null | https://proceedings.icml.cc/static/paper_files/icml/2020/5808-Paper.pdf | https://proceedings.icml.cc/static/paper_files/icml/2020/5808-Paper.pdf | icml-2020-1 | ['multi-target-regression'] | ['miscellaneous'] | [ 2.10648358e-01 -2.66562682e-02 -5.17844260e-01 -7.00091362e-01
-1.45243394e+00 -8.11015368e-01 1.38340980e-01 6.24776661e-01
-5.48511326e-01 8.49714518e-01 -2.01669350e-01 -3.75134379e-01
-2.42577448e-01 -5.06524742e-01 -8.68975699e-01 -1.01167774e+00
-8.35537240e-02 2.47426882e-01 1.17053322e-01 -2.03059435e-01
1.70903862e-01 5.79924822e-01 -1.28606045e+00 3.79077882e-01
5.57412863e-01 1.41799080e+00 -5.23005962e-01 1.18476093e+00
1.55884802e-01 4.89347935e-01 -3.36768061e-01 -4.76826102e-01
3.42984796e-01 -3.17049921e-01 -1.03893101e+00 -2.54796147e-01
7.67497301e-01 7.25699065e-04 4.47404116e-01 1.17337930e+00
5.79863310e-01 1.91527545e-01 9.23690975e-01 -1.70968390e+00
-3.51863325e-01 2.15465635e-01 -8.10370266e-01 2.79867768e-01
-1.46742105e-01 -6.19353712e-01 1.17788279e+00 -7.78764486e-01
1.65314957e-01 1.21674907e+00 1.11617112e+00 5.56000888e-01
-1.23147571e+00 -5.80800831e-01 1.55916616e-01 2.47032326e-02
-9.94456351e-01 -6.48251697e-02 4.80774015e-01 -5.03045321e-01
5.36998332e-01 4.57832038e-01 -7.42864162e-02 6.97644949e-01
3.39433223e-01 7.79882252e-01 1.36632276e+00 -7.42711782e-01
1.01242669e-01 4.07480747e-01 4.53499585e-01 1.08889604e+00
4.66296673e-02 2.45196875e-02 -3.41587335e-01 -6.17057264e-01
5.05580127e-01 1.83229689e-02 -8.10499638e-02 -3.31248254e-01
-6.46030188e-01 1.48372984e+00 3.48188609e-01 2.45999351e-01
2.70008612e-02 3.15918505e-01 6.31502688e-01 5.80656111e-01
8.86468112e-01 4.07729640e-05 -7.87498653e-01 1.75640628e-01
-6.24288023e-01 2.97695249e-01 8.90972316e-01 7.09447205e-01
6.15771294e-01 -1.97345823e-01 -4.18755785e-02 9.55402374e-01
2.29042143e-01 5.55181682e-01 3.59034777e-01 -8.83512318e-01
5.27880728e-01 5.61091788e-02 1.19589113e-01 -5.18449783e-01
-6.11988664e-01 -5.00098407e-01 -7.23473728e-01 5.31847894e-01
9.74108160e-01 -1.68492943e-01 -3.97909611e-01 2.01632476e+00
5.84034145e-01 9.35836956e-02 -2.01481670e-01 6.99451327e-01
1.14859268e-01 4.80868906e-01 -4.79698181e-02 -4.31797951e-01
1.15544021e+00 -1.15437603e+00 -2.42855981e-01 4.79889624e-02
9.98462319e-01 -4.15759683e-01 1.02226901e+00 6.61140621e-01
-8.93682480e-01 -2.11097777e-01 -1.19431162e+00 -2.06499144e-01
-4.20918971e-01 -3.69274735e-01 6.84417963e-01 9.35969114e-01
-7.97838569e-01 6.43391848e-01 -5.49652278e-01 3.93014215e-02
3.30424070e-01 3.87406170e-01 -5.57540894e-01 3.32733095e-02
-9.52913642e-01 8.72815728e-01 2.06478253e-01 -1.15319639e-01
-6.72869146e-01 -8.00926566e-01 -7.30890632e-01 -1.75581470e-01
8.67266431e-02 -4.70198125e-01 1.49688792e+00 -1.18355107e+00
-1.40155590e+00 1.03292215e+00 -5.12457341e-02 -4.18525785e-01
9.34800565e-01 -1.90269053e-01 -9.86488312e-02 7.11511970e-02
-7.34670386e-02 -1.00999638e-01 8.31040442e-01 -9.68169272e-01
-1.01194394e+00 -8.88696253e-01 1.82072356e-01 1.45685315e-01
-4.37274486e-01 1.46926120e-01 2.09787458e-01 -6.35932028e-01
-2.25431189e-01 -6.97019815e-01 -3.93924028e-01 8.16649646e-02
3.41150053e-02 -3.09178233e-01 6.58563673e-01 -5.84606767e-01
9.23222780e-01 -1.88130534e+00 8.10474455e-02 2.55390793e-01
2.63516366e-01 -7.59665519e-02 -4.19620909e-02 1.27952009e-01
-1.27286747e-01 1.77951485e-01 -2.85137147e-01 -4.57108855e-01
1.06896445e-01 -1.16211630e-01 -3.91479522e-01 1.06087089e+00
-7.78285041e-02 6.71549380e-01 -9.26998496e-01 -3.61970127e-01
-1.40170440e-01 3.64159882e-01 -3.62751454e-01 -2.22951367e-01
-1.16005607e-01 3.29173118e-01 -3.38780999e-01 5.09754360e-01
7.36406565e-01 -4.40557986e-01 -1.47891626e-01 1.45594388e-01
1.27979577e-01 -4.22543883e-02 -1.30331278e+00 1.47059965e+00
-9.78315115e-01 2.69488215e-01 4.04929519e-01 -1.32316923e+00
5.08410633e-01 1.94313049e-01 4.31013554e-01 -2.07896188e-01
1.86567649e-01 5.98130584e-01 -6.84172094e-01 -8.40795599e-03
2.59604994e-02 -7.75645494e-01 -3.59221756e-01 4.13225859e-01
1.84255317e-02 1.12675905e-01 -5.16031012e-02 -1.30720019e-01
9.28069830e-01 -4.41830009e-02 4.97289181e-01 -6.17526472e-01
6.80853605e-01 -4.02345479e-01 4.40649241e-01 8.63046408e-01
-2.36874402e-01 2.86640048e-01 5.78143656e-01 -2.05071539e-01
-8.07947397e-01 -9.57366765e-01 -4.77230847e-01 1.82871234e+00
-1.02871306e-01 -8.68863538e-02 -4.86493111e-01 -1.17105782e+00
3.74058813e-01 3.55362117e-01 -9.54334855e-01 -1.19617134e-01
-4.62219507e-01 -9.68359888e-01 4.80511487e-01 6.11399829e-01
1.75480142e-01 -3.26950848e-01 -4.73810703e-01 9.78426263e-02
1.58475757e-01 -6.66512787e-01 -6.52230799e-01 6.29618108e-01
-9.46221352e-01 -1.05192482e+00 -9.95226443e-01 -8.74050856e-01
3.94550979e-01 -8.29009190e-02 1.03475237e+00 -1.27436817e-01
-1.46852568e-01 2.62919635e-01 -3.49786691e-02 -5.13995826e-01
-4.56585407e-01 1.98292598e-01 4.78719808e-02 8.20624828e-03
1.20840266e-01 -3.61761004e-01 -2.43114173e-01 3.46427441e-01
-7.50472844e-01 -9.30395573e-02 3.79343152e-01 1.09657300e+00
7.70584643e-01 -1.57354087e-01 9.41592753e-01 -1.08152616e+00
4.70098853e-01 -5.86138487e-01 -9.89988923e-01 6.85282469e-01
-7.64530182e-01 4.41338271e-01 9.77362394e-01 -6.66412532e-01
-6.82775736e-01 8.05693716e-02 -1.52004376e-01 -9.96513292e-02
2.43246302e-01 1.70237169e-01 -4.17652205e-02 -5.31873286e-01
7.63531446e-01 -2.07759947e-01 -2.84146994e-01 -5.82167447e-01
6.82264507e-01 9.71719503e-01 4.52642530e-01 -8.19346845e-01
3.87065858e-01 5.11736512e-01 4.42961425e-01 -6.04455888e-01
-1.37207592e+00 -7.81900048e-01 -5.31004012e-01 -4.73774821e-02
6.11524343e-01 -6.01183176e-01 -1.03510678e+00 4.15986449e-01
-1.00053072e+00 -5.72179437e-01 -1.50637060e-01 4.43838865e-01
-7.90329695e-01 2.83707500e-01 -9.16564167e-01 -1.15172529e+00
-4.55363274e-01 -8.22027147e-01 1.00128627e+00 4.64054570e-02
3.68517011e-01 -1.43321621e+00 2.29770243e-01 4.09760982e-01
2.55286455e-01 5.33009887e-01 1.00430524e+00 -7.71473110e-01
1.55295655e-01 -1.52001753e-01 -3.36322576e-01 6.31568372e-01
-2.26948395e-01 -6.72992527e-01 -9.83392537e-01 -5.31128526e-01
3.32565516e-01 -7.02521086e-01 1.08262050e+00 3.39992434e-01
1.20204735e+00 -5.31651437e-01 -2.06413835e-01 7.49926507e-01
1.65179050e+00 -3.80485922e-01 -1.56075060e-01 3.33524108e-01
6.52955234e-01 4.82989043e-01 5.99783421e-01 5.56637704e-01
2.41320789e-01 5.46769381e-01 3.84113401e-01 -1.62471756e-01
2.87616313e-01 2.54084449e-02 1.64779320e-01 5.94521046e-01
2.34578282e-01 1.43599257e-01 -6.62054300e-01 3.73108089e-01
-1.85654914e+00 -6.61405504e-01 -2.92042762e-01 2.59034371e+00
1.15253615e+00 1.66115388e-01 6.05918527e-01 3.42212558e-01
8.09076071e-01 -1.91398397e-01 -7.53510416e-01 -8.05517793e-01
-1.21183760e-01 4.66590196e-01 1.25033176e+00 1.03445518e+00
-1.35831189e+00 5.64964116e-01 6.67688179e+00 9.75956142e-01
-1.04068613e+00 8.21526289e-01 7.04398394e-01 -5.96930832e-02
3.72377783e-02 -2.35351905e-01 -1.14096344e+00 3.16955686e-01
1.21715164e+00 -2.81475872e-01 3.92323971e-01 1.15393591e+00
-3.51974279e-01 2.68615596e-02 -1.12379599e+00 9.69236195e-01
2.91789398e-02 -9.73833621e-01 -6.19405746e-01 1.19456939e-01
7.72470653e-01 -1.42947910e-02 1.23326689e-01 3.28766555e-01
3.27661633e-01 -1.09325397e+00 6.98177814e-01 -1.55630484e-01
9.58859682e-01 -1.09720671e+00 5.55319607e-01 3.62326562e-01
-1.26720226e+00 -3.93790901e-01 -1.39902696e-01 -1.08824223e-01
-4.62066568e-02 7.18250215e-01 -2.21238717e-01 5.86493075e-01
2.48137951e-01 3.34056407e-01 -3.05297762e-01 9.36836183e-01
-4.14426140e-02 4.34973598e-01 -6.28733218e-01 -7.33045116e-02
2.81791121e-01 -1.49555817e-01 3.24042477e-02 1.60360789e+00
-4.67559062e-02 -1.71254501e-01 2.47880980e-01 1.39899313e-01
-3.56934994e-01 4.05986309e-01 -3.27797532e-01 3.87879848e-01
8.09307843e-02 1.19517553e+00 -5.95025480e-01 -3.02772433e-01
-6.22467995e-01 1.01988018e+00 6.32607698e-01 2.24041805e-01
-8.88341486e-01 -6.07417226e-01 4.59876567e-01 -2.76457340e-01
2.97689885e-01 -7.26442738e-03 -6.38599932e-01 -9.93881464e-01
2.59690970e-01 -6.66674852e-01 9.09080863e-01 1.04585104e-01
-1.49916029e+00 1.78154573e-01 1.47420257e-01 -7.38534212e-01
-3.08976203e-01 -9.42232490e-01 -2.91789800e-01 1.02628982e+00
-1.72458422e+00 -1.31108737e+00 1.29617780e-01 5.40895522e-01
1.97398528e-01 1.62201434e-01 9.57341611e-01 5.76791048e-01
-2.36934602e-01 9.51701939e-01 5.76825559e-01 -1.37439817e-01
6.73087060e-01 -1.54189229e+00 -8.78587663e-02 6.86567068e-01
-1.98680349e-02 1.65943932e-02 7.03751147e-01 -2.01433226e-01
-1.03139102e+00 -9.60841715e-01 8.13481092e-01 -3.73484790e-01
8.40415061e-01 -4.10886973e-01 -8.02057981e-01 8.72434795e-01
-5.39937913e-01 5.36480963e-01 8.50434244e-01 2.83982277e-01
-7.34687388e-01 -2.54721910e-01 -1.36344099e+00 -5.71186543e-02
7.05420732e-01 -6.41469419e-01 -1.64457843e-01 7.80364990e-01
3.76970857e-01 -3.54673952e-01 -1.08709526e+00 2.28762716e-01
8.35103035e-01 -6.74459100e-01 9.34880257e-01 -1.15593779e+00
2.11441994e-01 2.02076897e-01 -4.48221296e-01 -1.21857035e+00
-1.07927084e-01 -6.82752013e-01 -3.71401966e-01 9.25081253e-01
5.08179903e-01 -8.71561110e-01 5.17446935e-01 4.78757769e-01
8.28789100e-02 -1.24859321e+00 -1.13106620e+00 -1.08893895e+00
8.12227786e-01 -3.64974797e-01 4.22083773e-02 9.54983175e-01
8.95758942e-02 3.67861867e-01 -5.94574869e-01 7.44836554e-02
9.19442773e-01 3.63196507e-02 3.03029180e-01 -1.35012007e+00
-8.19845736e-01 -6.80294096e-01 -3.55115443e-01 -9.98743951e-01
3.94029260e-01 -1.18612182e+00 -1.04891889e-01 -1.17353296e+00
1.81447804e-01 -7.75447488e-01 -7.92979658e-01 5.25132120e-01
-9.94441286e-02 3.31004113e-01 1.21595770e-01 -6.69444278e-02
-5.91344059e-01 2.21613690e-01 8.40022862e-01 7.01925484e-03
-9.59641039e-02 4.60671246e-01 -7.61559725e-01 6.97654903e-01
8.01531732e-01 -8.73926520e-01 -1.63286313e-01 -2.08682775e-01
3.75037789e-01 1.16016313e-01 2.43627936e-01 -4.86293137e-01
-9.40050632e-02 -2.81525314e-01 1.58838511e-01 -2.76460480e-02
4.22975749e-01 -6.06270432e-01 -3.18370283e-01 6.72793329e-01
-6.97888732e-01 5.00953458e-02 1.15804859e-02 5.95126927e-01
2.08721146e-01 -5.42870283e-01 1.41674244e+00 1.64993897e-01
-1.69856846e-01 3.23423982e-01 1.88409895e-01 5.58230162e-01
9.51544821e-01 3.21441293e-01 -3.07583541e-01 -2.94838309e-01
-5.32942176e-01 1.91758856e-01 1.34757176e-01 2.61534303e-01
2.04557121e-01 -1.45556080e+00 -9.41376746e-01 8.94530490e-02
-3.87343355e-02 -3.75815272e-01 -8.28155801e-02 1.06981003e+00
-2.22049624e-01 2.66676068e-01 2.43777316e-02 -3.75654906e-01
-1.44276226e+00 5.45967042e-01 3.62843603e-01 -5.65757930e-01
-3.34683239e-01 1.04705918e+00 6.24845289e-02 -1.87436581e-01
4.60394830e-01 -3.53391618e-01 3.80892545e-01 -1.41643390e-01
8.00034761e-01 7.17527628e-01 3.74238014e-01 -4.06455547e-01
-5.34034789e-01 6.31272554e-01 -6.74449280e-02 -4.56981272e-01
1.18367147e+00 1.17690161e-01 1.53874069e-01 9.74773228e-01
2.02099323e+00 -3.40610631e-02 -1.04803479e+00 -3.11133981e-01
2.22457364e-01 -2.77026176e-01 5.21668196e-02 -8.68281305e-01
-6.57769263e-01 9.39980924e-01 9.02005613e-01 2.55989790e-01
1.04875374e+00 1.59570631e-02 7.87811637e-01 4.17267978e-01
3.40814948e-01 -1.37274599e+00 -5.87783679e-02 3.19006950e-01
8.15047026e-01 -1.42261589e+00 -3.01287547e-02 -1.76552862e-01
-2.81471699e-01 1.16608489e+00 1.85139179e-01 -6.51500523e-02
1.00558639e+00 3.34077418e-01 -4.75962423e-02 4.50757384e-01
-5.15580237e-01 -1.56630147e-02 3.03719312e-01 2.30089232e-01
4.23048019e-01 1.53802097e-01 -4.55986202e-01 5.75676382e-01
1.12087131e-02 -1.19861491e-01 2.21752725e-03 7.56844580e-01
-5.13378918e-01 -1.18900847e+00 -3.88059348e-01 4.71683502e-01
-9.02882755e-01 1.24547079e-01 2.11649723e-02 5.96869051e-01
-1.09923951e-01 1.00389814e+00 -3.47152323e-01 -7.28477985e-02
-1.04420295e-03 2.41506830e-01 8.97860467e-01 -3.05849910e-01
-5.58425307e-01 -1.19803458e-01 -1.55645773e-01 -2.50823200e-01
-3.42588931e-01 -3.98096472e-01 -1.22930849e+00 -3.59294146e-01
-7.00854897e-01 1.64866403e-01 1.17725646e+00 1.06191874e+00
-1.10205799e-01 5.64848669e-02 8.47037911e-01 -5.65230787e-01
-1.43251169e+00 -9.68016803e-01 -6.83302224e-01 3.79705697e-01
8.50410283e-01 -6.21086419e-01 -8.09340537e-01 -1.34458482e-01] | [8.267651557922363, 4.124420642852783] |
cc578dfb-bc35-4d88-8e90-ce8678a0a172 | fsid-fully-synthetic-image-denoising-via | 2212.03961 | null | https://arxiv.org/abs/2212.03961v1 | https://arxiv.org/pdf/2212.03961v1.pdf | FSID: Fully Synthetic Image Denoising via Procedural Scene Generation | For low-level computer vision and image processing ML tasks, training on large datasets is critical for generalization. However, the standard practice of relying on real-world images primarily from the Internet comes with image quality, scalability, and privacy issues, especially in commercial contexts. To address this, we have developed a procedural synthetic data generation pipeline and dataset tailored to low-level vision tasks. Our Unreal engine-based synthetic data pipeline populates large scenes algorithmically with a combination of random 3D objects, materials, and geometric transformations. Then, we calibrate the camera noise profiles to synthesize the noisy images. From this pipeline, we generated a fully synthetic image denoising dataset (FSID) which consists of 175,000 noisy/clean image pairs. We then trained and validated a CNN-based denoising model, and demonstrated that the model trained on this synthetic data alone can achieve competitive denoising results when evaluated on real-world noisy images captured with smartphone cameras. | ['Rakesh Ranjan', 'Bo Zhu', 'Xiaoyu Xiang', 'Seonghyeon Nam', 'Beibei Du', 'Gyeongmin Choe'] | 2022-12-07 | null | null | null | null | ['scene-generation', 'synthetic-data-generation', 'synthetic-data-generation'] | ['computer-vision', 'medical', 'miscellaneous'] | [ 5.36712885e-01 -2.35772222e-01 6.67842388e-01 -4.48642582e-01
-1.25322783e+00 -5.92011571e-01 6.48266673e-01 -1.39998853e-01
-6.98536336e-01 3.91082793e-01 4.57232818e-02 4.11254317e-02
4.37754035e-01 -6.69639409e-01 -9.94872570e-01 -5.02405286e-01
3.82508278e-01 6.09747358e-02 6.62031621e-02 -1.51019618e-01
2.42244914e-01 3.60579699e-01 -1.71883440e+00 3.38037193e-01
1.02599227e+00 1.19383895e+00 3.39112610e-01 7.71133900e-01
3.10549468e-01 5.31457603e-01 -8.76583695e-01 -6.90287352e-01
7.85402060e-01 -1.94128111e-01 -2.20056400e-01 3.97174895e-01
9.51034129e-01 -7.76980639e-01 -3.70710611e-01 1.34736192e+00
6.26869857e-01 8.22305307e-02 1.84580773e-01 -1.12065089e+00
-6.62446976e-01 2.54400045e-01 -3.05143297e-01 -2.15731710e-01
2.14512885e-01 8.18947911e-01 4.66216773e-01 -8.79546463e-01
7.73473263e-01 1.19279599e+00 1.04348886e+00 4.96713251e-01
-1.71051443e+00 -6.20751321e-01 -2.81406432e-01 -1.54767215e-01
-1.14286852e+00 -5.84993601e-01 5.73731542e-01 -2.43824184e-01
5.73965907e-01 -9.44180936e-02 6.09939516e-01 1.77777708e+00
-1.17494976e-02 3.93419266e-01 1.28490078e+00 -1.05271943e-01
4.82272178e-01 -4.61562648e-02 -1.48297831e-01 1.81490481e-01
3.82111460e-01 2.18451172e-01 -4.58942473e-01 9.51810703e-02
6.28120005e-01 -2.95881271e-01 -4.19231147e-01 -3.01686764e-01
-1.18126237e+00 5.30179024e-01 4.20317411e-01 -2.99036860e-01
-3.00986141e-01 2.13914096e-01 4.16872799e-01 3.98261696e-01
4.19968039e-01 5.50022244e-01 -2.98091263e-01 5.03005199e-02
-9.80431497e-01 4.33982104e-01 7.46313214e-01 9.85305071e-01
6.53270125e-01 2.04903409e-01 9.20846313e-02 9.44980025e-01
-1.60161674e-01 7.30239689e-01 2.13635236e-01 -1.50701964e+00
3.81809205e-01 1.97683379e-01 1.47219926e-01 -1.07821584e+00
-8.01148936e-02 -3.77083749e-01 -1.06643677e+00 4.25510108e-01
3.89561504e-01 -1.17377095e-01 -1.10518956e+00 1.73129189e+00
-7.82982539e-03 1.73624426e-01 1.28973752e-01 8.76777291e-01
8.12462211e-01 3.79900783e-01 6.27882197e-04 5.84134050e-02
7.66120136e-01 -7.18605578e-01 -5.25855899e-01 -4.31325376e-01
1.64972171e-01 -8.49367142e-01 1.18004179e+00 7.88637638e-01
-1.07190192e+00 -8.48005176e-01 -1.04894567e+00 -4.01402891e-01
-3.29251796e-01 1.46783903e-01 1.84320658e-01 6.54686749e-01
-1.29496920e+00 7.01982498e-01 -7.79341757e-01 -3.62966329e-01
7.35385299e-01 1.04112014e-01 -6.20072424e-01 -5.11138916e-01
-8.17236960e-01 7.04481840e-01 1.56072110e-01 2.93021113e-01
-1.39891911e+00 -7.07480609e-01 -1.09293830e+00 -1.51091397e-01
2.83679307e-01 -8.46552730e-01 1.14364898e+00 -9.34444726e-01
-1.19501305e+00 9.61112082e-01 2.55589724e-01 -6.43470168e-01
9.27670121e-01 -3.67412239e-01 -2.31330186e-01 1.30746916e-01
2.68158048e-01 8.76619577e-01 1.41166627e+00 -1.70410407e+00
-2.94293076e-01 -3.23206276e-01 -1.33655086e-01 -6.61754832e-02
-8.43913183e-02 -2.74884939e-01 -7.21768916e-01 -7.67602146e-01
-4.31840494e-02 -7.48693228e-01 -5.53206861e-01 3.13634068e-01
-3.80414397e-01 8.45918238e-01 8.57405365e-01 -8.58548701e-01
3.40526015e-01 -2.57332563e+00 -1.00606397e-01 1.25433341e-01
1.35435835e-01 3.67063165e-01 -5.31679273e-01 -1.25654077e-03
-1.49822578e-01 1.86490878e-01 -3.02340627e-01 -8.79381835e-01
-3.16538095e-01 1.40346095e-01 -3.26045603e-01 3.36248219e-01
4.12961423e-01 6.80313051e-01 -9.32479441e-01 -1.34709284e-01
5.26042104e-01 6.35447979e-01 -7.66670406e-01 3.48482549e-01
-3.04468691e-01 6.50487185e-01 -8.08113813e-03 6.75691962e-01
1.06439769e+00 1.43974081e-01 -1.06888555e-01 -6.21050954e-01
2.70511240e-01 -1.91362604e-01 -1.11122429e+00 1.96078956e+00
-5.27389348e-01 6.43534243e-01 5.41619480e-01 -6.33452892e-01
8.54519129e-01 -6.46274984e-02 2.48047307e-01 -6.24988437e-01
3.30911785e-01 1.34960353e-01 -5.20608246e-01 -5.64084172e-01
5.32494605e-01 2.31500879e-01 2.67973598e-02 -3.78767232e-04
2.28148937e-01 -1.10785818e+00 1.32737949e-01 2.24569649e-01
1.40015543e+00 2.18143404e-01 -2.37692907e-01 -3.33772264e-02
1.89279303e-01 2.13758379e-01 4.19177026e-01 9.92932618e-01
-1.96307361e-01 1.42732370e+00 2.84759104e-01 -4.66383964e-01
-1.28857517e+00 -1.31318855e+00 1.32575154e-01 4.04030293e-01
1.21184856e-01 -4.28697526e-01 -1.13955367e+00 -3.65615129e-01
-2.94431150e-01 7.37568855e-01 -5.04192412e-01 -2.31889859e-01
-2.84868956e-01 -7.44644582e-01 4.85941470e-01 7.67388046e-02
9.01988864e-01 -8.95643294e-01 -3.25454533e-01 1.01013750e-01
-2.08158329e-01 -1.54748952e+00 -4.59500104e-01 2.04011887e-01
-5.33438504e-01 -1.11385322e+00 -1.87482864e-01 -5.59904277e-01
6.43597662e-01 4.50702369e-01 1.32326019e+00 1.97246328e-01
-6.21762812e-01 3.73195589e-01 -2.68001527e-01 -3.33034009e-01
-6.00536227e-01 -2.46662021e-01 1.74969193e-02 1.00367600e-02
-1.20132804e-01 -7.51093388e-01 -8.17203283e-01 1.35239691e-01
-1.40144002e+00 -3.24137807e-02 5.02850771e-01 9.57337141e-01
7.26496220e-01 2.01647788e-01 2.56601423e-01 -6.08802557e-01
7.28389263e-01 -7.67343640e-02 -8.75833809e-01 -6.63582683e-02
-1.73677251e-01 -1.23721026e-01 8.10387731e-01 -4.10133243e-01
-1.23865545e+00 4.24975455e-01 -4.34883237e-01 -5.15899837e-01
-3.56125832e-01 1.22594222e-01 -4.64002848e-01 -2.21122146e-01
9.49095964e-01 1.86909676e-01 1.50272861e-01 -3.45703483e-01
6.47001743e-01 7.60064244e-01 1.21263242e+00 -4.81103778e-01
8.19260061e-01 6.80283725e-01 -2.01139599e-01 -9.76852834e-01
-9.32149053e-01 1.24866236e-03 -4.38872367e-01 -1.25020906e-01
8.92598331e-01 -1.29804265e+00 -4.87275124e-01 1.07529056e+00
-1.10270774e+00 -5.43882251e-01 -3.13444793e-01 3.13219786e-01
-5.03040433e-01 3.54623854e-01 -6.44120634e-01 -4.10659701e-01
-2.50801235e-01 -1.57805884e+00 1.39878368e+00 9.07727610e-03
4.37531509e-02 -4.32272941e-01 -2.75758684e-01 6.00714445e-01
4.51408178e-01 4.55934525e-01 4.83735979e-01 -6.51587397e-02
-8.59656632e-01 -2.39451677e-01 -2.92038441e-01 1.07159328e+00
3.13612521e-02 4.87854406e-02 -1.18753135e+00 -2.24533275e-01
3.88339847e-01 -7.53258526e-01 9.37757254e-01 3.01165432e-01
1.49796629e+00 1.06263578e-01 1.71758503e-01 1.12627244e+00
1.51990795e+00 -3.23150516e-01 8.04930210e-01 2.69463509e-01
6.96885407e-01 3.80010307e-01 3.15939665e-01 3.05903614e-01
1.41921699e-01 2.10227609e-01 7.75532126e-01 -1.14594884e-01
-2.90055871e-01 -2.79137164e-01 3.46872747e-01 3.74138504e-01
2.69902945e-01 -2.54981846e-01 -5.84111273e-01 5.58978081e-01
-1.31736147e+00 -7.28016615e-01 -6.57720044e-02 2.09241605e+00
8.17311704e-01 2.85529554e-01 -4.79191333e-01 -3.58547568e-02
5.68717062e-01 6.13157405e-03 -6.20169580e-01 -4.64363582e-02
-5.09674549e-01 2.87313402e-01 6.56800032e-01 2.39509761e-01
-1.18051755e+00 9.44735765e-01 6.49457836e+00 7.11533964e-01
-1.20312703e+00 -1.32820308e-01 8.92638147e-01 -1.66870311e-01
-2.26397246e-01 -2.99055278e-01 -3.80166978e-01 4.04476076e-01
8.04905653e-01 2.14165136e-01 7.77833164e-01 9.18730259e-01
4.35350239e-01 -1.85150668e-01 -9.60308909e-01 1.26926839e+00
1.32277757e-01 -1.39043617e+00 2.43082941e-01 -9.77360010e-02
9.19056594e-01 3.73346090e-01 1.88485429e-01 5.49230836e-02
5.77066064e-01 -1.05816674e+00 8.31014931e-01 4.56801802e-01
8.02090943e-01 -4.49439555e-01 7.26661026e-01 3.06622297e-01
-5.44509768e-01 6.88302442e-02 -3.12537462e-01 1.08130582e-01
1.09225638e-01 6.79406703e-01 -5.09169757e-01 4.70336914e-01
1.18104959e+00 7.35029101e-01 -9.53969836e-01 9.30256009e-01
-3.04445267e-01 4.84230280e-01 -4.40371484e-01 7.04983115e-01
-1.91356912e-01 -5.10977149e-01 2.55983651e-01 8.56238008e-01
4.25550282e-01 -1.08145051e-01 -1.23002619e-01 8.27750027e-01
-4.97038603e-01 -4.61360902e-01 -7.68788218e-01 1.25519693e-01
3.22901279e-01 1.41092908e+00 -6.69877112e-01 -6.47864044e-02
-3.93475085e-01 1.22578037e+00 7.73071200e-02 5.00584424e-01
-7.23380685e-01 -9.93411988e-02 1.01104820e+00 2.97605321e-02
2.99681991e-01 -3.88883412e-01 -6.83081329e-01 -1.36075926e+00
1.77217349e-01 -1.41528141e+00 6.54604984e-03 -1.21318674e+00
-1.48161209e+00 8.67904425e-01 -3.47641319e-01 -1.07259440e+00
3.30417380e-02 -5.92841685e-01 -4.30740446e-01 7.40638435e-01
-1.11720300e+00 -1.30012333e+00 -9.82241631e-01 6.31212711e-01
5.12681901e-01 1.05875477e-01 8.29279423e-01 2.72303879e-01
-4.28282380e-01 3.86583507e-01 2.03114331e-01 1.74447283e-01
8.53228986e-01 -9.89481807e-01 9.27554846e-01 1.21068978e+00
-2.35094000e-02 5.46328664e-01 8.19105208e-01 -5.58709264e-01
-1.44657695e+00 -1.41663158e+00 1.02625839e-01 -5.12317896e-01
4.91597593e-01 -7.43030667e-01 -8.61891031e-01 5.14456093e-01
2.74819553e-01 2.88911134e-01 6.10966012e-02 -5.64084291e-01
-4.62383658e-01 -3.56486887e-01 -1.44370151e+00 8.11803520e-01
1.09813845e+00 -5.72076678e-01 -2.61973470e-01 2.32554764e-01
8.52258146e-01 -5.63225865e-01 -7.82500088e-01 3.68837476e-01
2.50839800e-01 -1.22424412e+00 1.25830865e+00 -8.81822780e-02
6.57626390e-01 -5.05984247e-01 -2.45000988e-01 -1.65412748e+00
3.41183335e-01 -7.22949266e-01 5.42886734e-01 1.14145219e+00
2.63344675e-01 -2.20709473e-01 6.42146945e-01 6.71784937e-01
-1.39554247e-01 -1.26057804e-01 -5.57008386e-01 -5.40746093e-01
-2.10140422e-01 -9.20696795e-01 4.17172015e-01 4.61612552e-01
-1.00151026e+00 6.62474856e-02 -4.65130270e-01 1.98272109e-01
9.52849209e-01 -2.40183771e-01 1.26671159e+00 -8.06343138e-01
-1.36361003e-01 7.55411899e-03 -4.37148809e-01 -7.65271068e-01
1.00805774e-01 -3.71306181e-01 3.41431797e-01 -1.28856647e+00
-8.58809352e-02 -2.12009311e-01 2.01983452e-01 1.74724713e-01
-1.50653318e-01 7.20102906e-01 4.84411232e-03 -5.30747809e-02
-3.56711507e-01 6.61688507e-01 1.16849148e+00 -2.30639055e-01
2.12938841e-02 -2.97956616e-01 -6.94557726e-01 7.43783891e-01
6.68286800e-01 -3.40541184e-01 -4.03926581e-01 -8.39028716e-01
9.96061787e-02 -2.93905973e-01 7.99887240e-01 -1.25345421e+00
8.65728557e-02 1.35607600e-01 5.53312421e-01 -3.77684593e-01
6.77905560e-01 -9.11997080e-01 3.38601798e-01 2.61328727e-01
-2.77207792e-01 -2.63041258e-01 1.72536522e-01 6.24695420e-01
-3.74171466e-01 6.89124763e-02 9.95089471e-01 -2.30709776e-01
-7.31960118e-01 2.12344706e-01 -1.18061572e-01 2.16201186e-01
8.03472042e-01 -1.17147855e-01 -3.48507702e-01 -6.58005297e-01
-5.78903258e-01 -7.40420669e-02 9.29733515e-01 4.05977726e-01
7.18093157e-01 -9.56216753e-01 -6.51183546e-01 5.73703885e-01
1.22765899e-01 4.82833475e-01 2.33940303e-01 2.61165380e-01
-9.20780838e-01 -3.92048389e-01 -3.50530624e-01 -7.01644182e-01
-1.09381628e+00 6.55009151e-01 2.91955173e-01 2.48197183e-01
-5.60867786e-01 8.53757381e-01 9.47122648e-03 -5.33155739e-01
4.05034184e-01 -6.58303320e-01 5.17128050e-01 -2.55985916e-01
4.63033140e-01 8.76167268e-02 3.91532779e-01 -4.99009311e-01
2.89805289e-02 3.71554226e-01 1.71569988e-01 -3.08302164e-01
1.60462356e+00 -1.54104978e-01 -1.23066790e-01 -1.17989898e-01
1.22583091e+00 -5.77237904e-02 -1.72405720e+00 -7.39226788e-02
-2.06396654e-01 -7.07835793e-01 9.90389660e-02 -5.98698258e-01
-1.26147544e+00 5.42478621e-01 6.80748582e-01 -1.13548459e-02
1.45563924e+00 -4.29169416e-01 9.10108984e-01 4.67608392e-01
5.45667827e-01 -1.01230156e+00 1.50918588e-01 3.71969819e-01
9.78642941e-01 -1.54500496e+00 -1.89653739e-01 -5.12189507e-01
-5.61026752e-01 8.68126690e-01 6.10606670e-01 -2.35149860e-01
7.16516972e-01 5.32492816e-01 5.67145228e-01 7.42716491e-02
-4.94234532e-01 1.34345153e-02 -2.22505108e-01 9.53537524e-01
-1.74623892e-01 -3.84830743e-01 2.70517915e-01 4.53563064e-01
-4.89011288e-01 1.07009448e-01 7.23676682e-01 8.13915372e-01
3.00452244e-02 -1.15832424e+00 -4.87759650e-01 4.04151767e-01
-4.04772490e-01 -2.35573694e-01 -4.02788401e-01 5.01253963e-01
4.99346435e-01 1.29527318e+00 -1.11659065e-01 -5.52946270e-01
4.08913225e-01 -4.87182617e-01 2.82928735e-01 -4.81615752e-01
-6.97691262e-01 -7.01364055e-02 -3.21186660e-03 -9.39252257e-01
-2.77004153e-01 -6.26988053e-01 -5.83546340e-01 -2.32169345e-01
3.12956143e-03 -3.44318479e-01 1.01112866e+00 6.46399021e-01
3.69664222e-01 3.96055520e-01 4.73355085e-01 -1.18390584e+00
-7.09377766e-01 -8.98493767e-01 -3.53128403e-01 9.55302715e-01
3.85403186e-01 -2.09496722e-01 -6.03354752e-01 5.55468917e-01] | [10.959970474243164, -2.3829073905944824] |
9c9a1b7d-3c43-46ae-b407-15c7c8876aa9 | transfer-learning-for-music-classification | 1703.09179 | null | http://arxiv.org/abs/1703.09179v4 | http://arxiv.org/pdf/1703.09179v4.pdf | Transfer learning for music classification and regression tasks | In this paper, we present a transfer learning approach for music
classification and regression tasks. We propose to use a pre-trained convnet
feature, a concatenated feature vector using the activations of feature maps of
multiple layers in a trained convolutional network. We show how this convnet
feature can serve as general-purpose music representation. In the experiments,
a convnet is trained for music tagging and then transferred to other
music-related classification and regression tasks. The convnet feature
outperforms the baseline MFCC feature in all the considered tasks and several
previous approaches that are aggregating MFCCs as well as low- and high-level
music features. | ['György Fazekas', 'Kyunghyun Cho', 'Keunwoo Choi', 'Mark Sandler'] | 2017-03-27 | null | null | null | null | ['music-classification'] | ['music'] | [ 2.01751381e-01 -4.23172742e-01 -9.10537541e-02 -2.67224282e-01
-6.82114959e-01 -5.16095757e-01 6.44652843e-01 -4.49478254e-03
-5.69373310e-01 5.50752401e-01 4.54483032e-01 3.31941843e-01
-3.85760248e-01 -5.89688003e-01 -4.39003140e-01 -5.41775167e-01
-3.47342551e-01 4.04444262e-02 -7.66596571e-02 -9.86414850e-02
4.01151061e-01 3.99985731e-01 -1.73931527e+00 6.92286253e-01
2.45573506e-01 1.35043550e+00 1.76721185e-01 7.98036218e-01
1.61081374e-01 6.74274564e-01 -9.45300221e-01 -1.08908378e-01
4.43681180e-02 -6.87263608e-01 -7.59223998e-01 -4.68368292e-01
5.32609522e-01 2.68866241e-01 -2.59149134e-01 8.83137882e-01
6.97148263e-01 4.61998522e-01 6.37944818e-01 -1.02813935e+00
-4.84947205e-01 1.18466997e+00 1.68537181e-02 1.95309252e-01
2.45782137e-01 -2.35291675e-01 1.22297788e+00 -7.48714268e-01
3.43064785e-01 8.58348310e-01 1.10538387e+00 3.33785772e-01
-9.70004320e-01 -9.22155321e-01 -2.86696672e-01 4.72128212e-01
-1.38159323e+00 -3.16598088e-01 1.07945800e+00 -4.44996536e-01
1.28685415e+00 1.91261798e-01 9.62769687e-01 1.13653612e+00
1.21968631e-02 7.05602765e-01 8.00891042e-01 -5.91092646e-01
-2.53172498e-02 -2.87581801e-01 -1.80638116e-02 3.38290423e-01
-1.86410084e-01 9.36922207e-02 -1.00765896e+00 -2.34175056e-01
9.56553102e-01 -1.83801085e-01 -1.28079325e-01 2.12184370e-01
-1.60278392e+00 7.96528578e-01 7.27644324e-01 1.07045209e+00
-4.42813993e-01 8.51347029e-01 6.71650767e-01 5.12165308e-01
3.16515297e-01 8.00728917e-01 -7.23420858e-01 -4.60074008e-01
-1.19217944e+00 1.71777710e-01 6.81526661e-01 5.00369191e-01
5.56181192e-01 3.76332343e-01 -5.17189622e-01 1.01529968e+00
-1.15857124e-01 1.87411666e-01 7.99024343e-01 -9.40597057e-01
1.46900058e-01 2.44063988e-01 -4.16917533e-01 -6.86229348e-01
-5.21218896e-01 -9.78836119e-01 -8.70895028e-01 9.01286155e-02
7.83228651e-02 -1.32163242e-01 -6.75363064e-01 1.74115741e+00
-4.90591764e-01 7.68844843e-01 -3.02167386e-02 9.72941220e-01
9.83879805e-01 3.93701136e-01 1.23789534e-01 9.18421894e-02
1.06214249e+00 -1.00818801e+00 -7.24427640e-01 3.80432904e-01
2.51277238e-01 -1.18337214e+00 7.79551446e-01 6.62389934e-01
-9.54787970e-01 -1.31822908e+00 -1.14026642e+00 -4.73747542e-03
-4.90483284e-01 8.40204477e-01 9.23606515e-01 3.58046591e-01
-1.05412257e+00 1.31481361e+00 -5.24959683e-01 -1.85986459e-01
3.40728045e-01 7.68093944e-01 -2.47909844e-01 6.32519186e-01
-1.12822628e+00 7.48290539e-01 7.00498819e-01 -9.71985683e-02
-5.69053888e-01 -4.65422153e-01 -6.75241828e-01 4.40415144e-01
-3.35089862e-01 -7.57287323e-01 1.32554865e+00 -1.31958246e+00
-1.76128173e+00 6.39106095e-01 2.43900269e-01 -5.04887521e-01
-1.45654514e-01 -3.37436080e-01 -7.35719740e-01 -8.54328424e-02
-9.33938324e-02 8.18441510e-01 9.84100223e-01 -7.61372924e-01
-6.23793602e-01 -1.23465303e-02 -6.59217611e-02 8.59524086e-02
-3.97848308e-01 9.11453590e-02 -2.82383673e-02 -1.18035483e+00
-2.45183669e-02 -9.09192979e-01 -7.97237754e-02 -5.93087196e-01
-3.85671973e-01 -5.10891676e-01 6.21051192e-01 -2.74892360e-01
1.27709866e+00 -2.41447496e+00 2.48871118e-01 2.43710473e-01
-3.04235697e-01 2.56533235e-01 -5.65366328e-01 4.01942104e-01
-3.31699729e-01 -6.11784607e-02 -1.19311266e-01 -2.75238186e-01
7.52597898e-02 -2.59531196e-02 -2.50027657e-01 1.87914371e-01
2.91997701e-01 9.76287127e-01 -7.02810466e-01 -1.59345627e-01
4.10954118e-01 7.02760339e-01 -6.85625315e-01 -6.88747093e-02
4.88951579e-02 8.02548885e-01 -6.75517470e-02 2.80686617e-01
3.89493167e-01 2.79541135e-01 -4.18286026e-02 -3.18773508e-01
-5.83544001e-02 8.83849680e-01 -1.02212620e+00 2.59053254e+00
-7.54892051e-01 1.05505311e+00 -4.63430077e-01 -1.07675600e+00
1.20288825e+00 5.74094534e-01 6.55383229e-01 -2.14454383e-01
4.66195047e-01 3.63607943e-01 4.05173123e-01 -1.06239498e-01
3.12558770e-01 -2.23403126e-01 -3.57529759e-01 4.04071569e-01
9.54162300e-01 -2.52917320e-01 -2.80361623e-02 -4.64239597e-01
1.08172369e+00 3.35689932e-01 2.48192638e-01 -1.77875653e-01
7.36800730e-01 -3.13453674e-01 3.61411989e-01 6.55757546e-01
1.82548791e-01 7.51895189e-01 1.35716185e-01 -5.37380278e-01
-8.06383491e-01 -7.29724348e-01 -2.84702480e-01 1.51988995e+00
-4.14403766e-01 -9.70386088e-01 -5.75033426e-01 -5.39747000e-01
9.67040956e-02 2.87890881e-01 -5.49308896e-01 -2.83283472e-01
-6.27057314e-01 -3.23798895e-01 9.96979952e-01 7.30599523e-01
4.03487980e-01 -1.77513742e+00 -3.46459925e-01 4.02980655e-01
-6.73529655e-02 -6.97122097e-01 -3.56536597e-01 9.68216896e-01
-9.38702106e-01 -9.46217179e-01 -5.87009311e-01 -1.27290273e+00
-2.01374829e-01 -2.92063832e-01 1.07776427e+00 2.87725180e-02
-2.10274249e-01 9.76628363e-02 -5.32287121e-01 -6.41348124e-01
-1.95875406e-01 4.78513241e-01 -5.01918197e-02 1.06602907e-01
5.94237924e-01 -1.12725723e+00 -4.98709142e-01 -1.79787576e-01
-5.24092555e-01 -3.19867909e-01 4.92941439e-01 9.63212252e-01
5.00237525e-01 -2.11220846e-01 8.66768956e-01 -3.88677418e-01
7.99443364e-01 -1.07597545e-01 -1.57149583e-02 -2.12483823e-01
-1.14101894e-01 3.22813317e-02 5.42837739e-01 -7.11071134e-01
-3.58116865e-01 5.24365902e-01 -4.75379884e-01 -6.35537326e-01
-4.32185352e-01 6.51573360e-01 2.51002878e-01 -1.16793543e-01
7.04881489e-01 -4.26823422e-02 -5.46419978e-01 -9.12808180e-01
4.07437295e-01 9.11510825e-01 7.79075384e-01 -3.34809810e-01
6.02772236e-01 -1.17638469e-01 1.68176055e-01 -5.97191453e-01
-9.74003255e-01 -6.79736257e-01 -1.07969689e+00 -1.95118636e-01
8.51804435e-01 -9.21531022e-01 -8.70083570e-01 2.82385409e-01
-1.20875525e+00 -1.67881519e-01 -8.01352561e-01 1.11989737e+00
-1.16423309e+00 -1.40821815e-01 -8.02022696e-01 -6.28395081e-01
-4.03456807e-01 -6.49544358e-01 1.14706337e+00 7.73959160e-02
-4.76556361e-01 -7.61901200e-01 5.15709162e-01 -2.88743675e-01
4.98575032e-01 6.94486350e-02 8.36272717e-01 -8.97998810e-01
8.07467476e-03 -1.73706532e-01 1.09308705e-01 5.94754338e-01
1.11659050e-01 -1.44089535e-01 -1.47413957e+00 -2.90097073e-02
-1.45141780e-01 -5.39971590e-01 1.53978240e+00 5.80386937e-01
1.55715823e+00 1.19219787e-01 1.25856668e-01 8.64009738e-01
1.26652086e+00 1.56142026e-01 5.87384760e-01 1.94313228e-01
5.13996780e-01 -5.50421178e-02 3.04873079e-01 4.47991103e-01
-3.04952383e-01 8.30829442e-01 2.30933264e-01 -2.58783221e-01
-4.57642078e-01 -2.15400755e-01 2.58604050e-01 1.19178939e+00
-7.95783401e-01 2.12748408e-01 -4.09840822e-01 3.03577423e-01
-2.03581977e+00 -1.10971212e+00 -1.51442349e-01 1.95726776e+00
6.37406707e-01 -1.41288817e-01 2.94611990e-01 7.65734792e-01
7.24303365e-01 5.83782047e-02 -2.17983708e-01 -6.23907924e-01
-2.52584755e-01 1.29741073e+00 5.60141876e-02 -5.89234531e-02
-1.73252130e+00 1.10502291e+00 7.24435949e+00 1.00570011e+00
-1.37461257e+00 4.06327844e-01 -2.71595120e-01 -4.31661494e-04
2.81754911e-01 -2.02995241e-01 -6.26322627e-02 -3.31901759e-02
1.07205582e+00 2.85955489e-01 6.52817011e-01 8.19831491e-01
-2.68731773e-01 2.81864762e-01 -1.16728020e+00 1.44469547e+00
2.08047777e-01 -1.25368404e+00 6.44954816e-02 -1.53065771e-01
7.52202630e-01 1.91340417e-01 1.44988090e-01 6.13623023e-01
-4.42772657e-02 -1.23258674e+00 7.42309868e-01 6.13795638e-01
9.88187671e-01 -1.05117536e+00 8.66132379e-01 -9.78595465e-02
-1.61507130e+00 -2.86341012e-01 -4.79310095e-01 -4.50916290e-01
-2.65699089e-01 1.48834541e-01 -4.75929379e-01 5.85009277e-01
6.89058304e-01 1.32485819e+00 -4.15009379e-01 1.41397369e+00
-2.57721245e-01 7.95927882e-01 -1.97443627e-02 1.13664024e-01
3.81821394e-01 8.86465050e-03 3.36106300e-01 1.71952832e+00
7.22534180e-01 -4.20985997e-01 5.25507480e-02 8.28237116e-01
-1.96076289e-01 3.47969949e-01 -5.32071888e-01 -1.43882036e-01
-4.47954237e-03 1.53994632e+00 -6.16207719e-01 -4.06504422e-01
-2.25177258e-01 1.11185288e+00 1.86716288e-01 2.59628534e-01
-7.06036925e-01 -8.19072723e-01 7.94847012e-01 -4.06001598e-01
6.73664689e-01 -6.81522489e-02 -2.67421901e-01 -1.19905436e+00
-4.24146175e-01 -4.13150877e-01 3.36305290e-01 -8.62275362e-01
-1.36502528e+00 7.70275712e-01 -5.82581162e-01 -1.68557549e+00
-4.71333802e-01 -7.61033714e-01 -1.00533795e+00 9.55148101e-01
-1.36674142e+00 -1.28694177e+00 -2.17615068e-02 1.01865733e+00
4.00791466e-01 -7.22892523e-01 1.50857282e+00 4.74370539e-01
8.92773122e-02 5.07816911e-01 -9.36839208e-02 4.68981624e-01
8.81477296e-01 -1.38550556e+00 2.12417036e-01 4.29609418e-02
1.10770702e+00 4.80667263e-01 8.32720473e-02 -1.10112511e-01
-7.45178998e-01 -1.13143587e+00 1.05640709e+00 -8.03689808e-02
5.25306821e-01 -2.45625660e-01 -5.54998279e-01 5.32941937e-01
4.36492324e-01 3.00600976e-02 1.02984715e+00 6.07550859e-01
-3.69557381e-01 -1.58608362e-01 -6.71404004e-01 -3.28355655e-02
1.06830823e+00 -9.66528416e-01 -8.42730105e-01 6.11924529e-02
3.86958122e-01 -5.68867102e-02 -1.02222121e+00 4.97684181e-01
8.07745576e-01 -7.37975419e-01 9.72169816e-01 -8.89648318e-01
2.34582499e-01 -2.67745376e-01 -3.01983595e-01 -1.51621056e+00
-6.77974403e-01 -5.86346686e-01 -5.91940805e-03 9.58209455e-01
4.39673871e-01 7.93666914e-02 4.38021183e-01 -7.19523370e-01
-4.75501776e-01 -1.96152225e-01 -1.12522745e+00 -9.31616545e-01
1.28742754e-01 -7.37105608e-01 5.41507006e-01 1.08915377e+00
4.17517483e-01 5.81583083e-01 -3.93972397e-01 -3.96298409e-01
-6.94873556e-03 3.24365914e-01 5.19844711e-01 -1.80678296e+00
-3.15122485e-01 -7.58207560e-01 -1.03610110e+00 -5.62509418e-01
4.05415058e-01 -1.63032234e+00 3.01189199e-02 -1.32013440e+00
2.17609808e-01 -3.17418098e-01 -1.30805707e+00 6.37980759e-01
3.29990357e-01 9.83875155e-01 5.06776512e-01 7.28844851e-02
-6.10001683e-01 3.63368124e-01 1.12266254e+00 -3.01565796e-01
-3.83343339e-01 2.85471022e-01 -1.61731109e-01 8.19776058e-01
1.06547832e+00 -6.84978426e-01 9.01608318e-02 -3.17269973e-02
7.06589743e-02 -1.24151468e-01 4.52662498e-01 -1.65179789e+00
1.01408757e-01 3.47699851e-01 8.40322912e-01 -6.92810595e-01
7.47704387e-01 -5.76969147e-01 1.37486041e-01 4.91065174e-01
-5.64987719e-01 -3.06965411e-01 1.25672311e-01 1.63025916e-01
-8.52322280e-01 -3.98849756e-01 3.25389266e-01 -2.76346598e-02
-5.01826942e-01 -7.23547637e-02 -4.30583000e-01 -5.22488296e-01
3.33592087e-01 9.88875404e-02 1.03504002e-01 -3.93934518e-01
-1.35602474e+00 -6.92726195e-01 -1.80897713e-01 6.73226714e-01
4.62203860e-01 -1.93501103e+00 -7.27626204e-01 2.95386463e-01
1.71777397e-01 -8.31271291e-01 -6.90283701e-02 9.11465168e-01
-1.03872478e-01 5.61478853e-01 -7.67825603e-01 -5.74732006e-01
-1.16847038e+00 4.51027483e-01 3.73856008e-01 -3.33510190e-02
-4.26491380e-01 6.41970158e-01 -2.94127733e-01 -4.27530199e-01
4.88802165e-01 -6.55522883e-01 -6.02023244e-01 2.34501481e-01
3.24069202e-01 7.24820346e-02 1.68852091e-01 -8.34723115e-01
-2.41063118e-01 9.88683820e-01 6.82954371e-01 -3.34885985e-01
1.62350059e+00 4.17732388e-01 -9.05563012e-02 9.45853949e-01
1.12429082e+00 -2.31089089e-02 -6.40099645e-01 -9.57999676e-02
1.94904253e-01 -3.49784642e-02 1.65532187e-01 -9.02641654e-01
-1.26400495e+00 1.10475016e+00 7.07076788e-01 1.59100056e-01
1.32068729e+00 1.11826472e-02 5.13079762e-01 6.29665017e-01
2.14995444e-01 -1.11138260e+00 7.17593431e-02 7.49158978e-01
1.14943135e+00 -7.96208322e-01 -3.38105559e-01 3.30719799e-02
-5.31357825e-01 1.72484457e+00 1.52621478e-01 -7.96504259e-01
1.06460369e+00 2.32586861e-01 8.41401964e-02 7.91713689e-03
-5.14191687e-01 -9.35891867e-01 9.41177487e-01 6.60417199e-01
9.62023437e-01 2.18466967e-01 -5.15593290e-01 1.24433112e+00
-7.40585804e-01 2.54707545e-01 1.42231673e-01 6.65318251e-01
-3.84544253e-01 -1.34669757e+00 -1.53048500e-01 4.16085005e-01
-7.35954523e-01 -1.93113506e-01 -6.70203567e-01 6.41530693e-01
8.89346898e-01 9.77040648e-01 3.79225999e-01 -9.66678977e-01
2.60323852e-01 3.76430452e-01 7.60582626e-01 -7.52056360e-01
-1.49171257e+00 5.13591707e-01 5.14348485e-02 -4.38967288e-01
-9.64150786e-01 -2.54265666e-01 -1.04127109e+00 1.61111742e-01
-5.29718399e-01 1.58786446e-01 9.94333982e-01 7.25447178e-01
6.07799031e-02 9.19823825e-01 6.15643620e-01 -1.30134308e+00
-1.34049326e-01 -1.60345805e+00 -8.60057533e-01 4.96900558e-01
3.05540323e-01 -5.99446833e-01 2.27302909e-02 1.43972307e-01] | [15.752266883850098, 5.223875522613525] |
13f31827-197f-4d13-8137-16e3f7ac8654 | look-more-but-care-less-in-video-recognition | 2211.09992 | null | https://arxiv.org/abs/2211.09992v1 | https://arxiv.org/pdf/2211.09992v1.pdf | Look More but Care Less in Video Recognition | Existing action recognition methods typically sample a few frames to represent each video to avoid the enormous computation, which often limits the recognition performance. To tackle this problem, we propose Ample and Focal Network (AFNet), which is composed of two branches to utilize more frames but with less computation. Specifically, the Ample Branch takes all input frames to obtain abundant information with condensed computation and provides the guidance for Focal Branch by the proposed Navigation Module; the Focal Branch squeezes the temporal size to only focus on the salient frames at each convolution block; in the end, the results of two branches are adaptively fused to prevent the loss of information. With this design, we can introduce more frames to the network but cost less computation. Besides, we demonstrate AFNet can utilize fewer frames while achieving higher accuracy as the dynamic selection in intermediate features enforces implicit temporal modeling. Further, we show that our method can be extended to reduce spatial redundancy with even less cost. Extensive experiments on five datasets demonstrate the effectiveness and efficiency of our method. | ['Yun Fu', 'Yi Xu', 'Huan Wang', 'Yue Bai', 'Yitian Zhang'] | 2022-11-18 | null | null | null | null | ['video-recognition'] | ['computer-vision'] | [ 1.61774054e-01 -2.42890596e-01 -3.03657979e-01 -3.39261293e-01
-5.30695543e-02 -1.28465295e-01 2.44095191e-01 -3.65982890e-01
-4.61692929e-01 4.86501098e-01 1.28253937e-01 6.92892745e-02
-1.01038702e-01 -8.17674279e-01 -3.47174168e-01 -8.13338161e-01
9.17224884e-02 -3.49013060e-01 7.46089339e-01 1.14202216e-01
1.77869275e-01 4.62320060e-01 -1.36431789e+00 3.31830114e-01
7.81529903e-01 1.33310449e+00 4.99823630e-01 5.67508414e-02
-9.08978954e-02 1.02455771e+00 -4.53655154e-01 1.15976706e-01
4.81200337e-01 -3.35214287e-01 -4.26606983e-01 4.37987238e-01
1.68894708e-01 -9.78843987e-01 -8.78993750e-01 1.01931417e+00
3.28472853e-01 4.03092027e-01 -4.74490747e-02 -1.12941229e+00
-4.08365071e-01 4.66068417e-01 -7.91183650e-01 4.39860582e-01
-3.33381630e-02 4.45957422e-01 6.32950783e-01 -8.84549975e-01
3.92827272e-01 1.26788425e+00 2.57509708e-01 5.49247086e-01
-8.53042066e-01 -8.53312612e-01 6.75291479e-01 3.66950125e-01
-1.27118790e+00 -5.34796000e-01 8.04907978e-01 -2.99147535e-02
5.50086498e-01 1.20834611e-01 8.63140702e-01 6.81819558e-01
-2.79728621e-01 1.03400135e+00 6.84340537e-01 6.77662389e-03
8.25701132e-02 -4.59977657e-01 2.52927244e-01 8.70325446e-01
1.02577493e-01 1.11615555e-02 -5.50883472e-01 2.90947855e-01
1.21739769e+00 6.86602414e-01 -6.94531560e-01 -1.86987266e-01
-1.35550022e+00 6.32243812e-01 5.03818989e-01 1.82387993e-01
-4.57504123e-01 2.31368124e-01 5.38672090e-01 1.41979516e-01
1.51642874e-01 -7.99486935e-02 -2.73179710e-01 -1.28796697e-01
-9.67126608e-01 5.03140017e-02 2.52850920e-01 8.94437909e-01
8.82195950e-01 2.05164850e-01 -3.26569498e-01 8.39710176e-01
9.27129835e-02 1.58579290e-01 5.14092267e-01 -1.20990825e+00
6.38412476e-01 8.99830341e-01 -1.21888906e-01 -1.20522416e+00
-3.24321508e-01 -3.05320710e-01 -1.01808822e+00 8.09677169e-02
3.51249009e-01 -4.96234931e-02 -8.74547005e-01 1.68275058e+00
2.70797729e-01 4.70744520e-01 -1.84125602e-01 1.19884396e+00
5.47578096e-01 6.66861296e-01 5.41362949e-02 -4.49170947e-01
1.24641645e+00 -1.25544202e+00 -6.53518379e-01 -1.55269772e-01
5.14297724e-01 -4.82141972e-01 9.89689112e-01 2.73484051e-01
-1.01254165e+00 -7.46788621e-01 -1.12166297e+00 -1.69106890e-02
1.14171758e-01 6.02631629e-01 9.09189045e-01 1.30043045e-01
-6.35815322e-01 6.29340053e-01 -1.20313323e+00 -5.39648496e-02
6.55698717e-01 3.74774665e-01 -1.48882911e-01 -2.01312646e-01
-1.04672933e+00 2.92696923e-01 5.47204912e-01 4.35212761e-01
-7.25108743e-01 -5.42370915e-01 -8.35125208e-01 2.20452264e-01
6.67137444e-01 -4.01975811e-01 1.02906048e+00 -1.10435319e+00
-1.50322652e+00 1.48108527e-01 -2.69230723e-01 -4.62239295e-01
5.67392886e-01 -3.71128201e-01 -3.13931972e-01 5.26098609e-01
5.16136130e-03 8.58165681e-01 7.60255575e-01 -6.50328696e-01
-8.73551786e-01 -3.78613204e-01 4.07770127e-01 1.88657567e-01
-7.21311331e-01 -1.44930303e-01 -8.96215081e-01 -9.36762691e-01
4.33368236e-01 -6.62120163e-01 -3.18269700e-01 4.41482693e-01
-1.01075977e-01 -1.75852537e-01 1.16309714e+00 -5.28866589e-01
1.47407436e+00 -2.41544628e+00 9.85300243e-02 -1.37714921e-02
4.20138776e-01 4.54709738e-01 -2.03647882e-01 -3.41068283e-02
-1.19545925e-02 -5.25422096e-02 -7.75019154e-02 -5.19132838e-02
-4.01949674e-01 3.82247627e-01 -2.96142131e-01 3.47651690e-01
2.12833896e-01 7.86955714e-01 -7.52756655e-01 -5.27194560e-01
3.69754523e-01 3.03669006e-01 -6.57145321e-01 1.43465458e-03
-7.12839589e-02 2.44025469e-01 -9.42756116e-01 5.70383489e-01
8.07715178e-01 -3.57202411e-01 8.22654068e-02 -5.57636917e-01
-2.46045023e-01 2.36173749e-01 -1.42756844e+00 1.72618556e+00
-1.27740264e-01 4.28654581e-01 2.31275395e-01 -1.13696778e+00
8.18429708e-01 2.35729129e-03 5.77021122e-01 -7.88044333e-01
1.54179215e-01 -2.27931570e-02 -2.64773034e-02 -5.49081385e-01
2.88733810e-01 3.38795096e-01 2.93537259e-01 3.66879672e-01
-3.70491117e-01 6.67404830e-01 2.46128827e-01 1.13796607e-01
9.88969147e-01 2.30280116e-01 2.18725763e-02 3.37934420e-02
7.39483118e-01 -2.81135798e-01 1.26690781e+00 3.80980700e-01
-4.81018305e-01 3.29115540e-01 2.86295176e-01 -7.56209254e-01
-6.47346199e-01 -7.31899559e-01 2.88540483e-01 7.68936932e-01
6.23755813e-01 -5.02889276e-01 -7.37184405e-01 -7.44819343e-01
-1.81063890e-01 1.70717672e-01 -4.21209872e-01 -3.03161353e-01
-1.04293191e+00 -4.30189103e-01 2.52295971e-01 7.73694456e-01
1.13874185e+00 -1.03079379e+00 -9.60861146e-01 1.95212632e-01
-2.02902287e-01 -1.01314211e+00 -8.19602966e-01 -2.87111312e-01
-1.01552379e+00 -1.01639521e+00 -6.47933960e-01 -7.46756911e-01
7.69552886e-01 7.42231250e-01 4.61976975e-01 3.75006557e-01
-1.10378735e-01 -1.75488085e-01 -5.57483733e-01 1.55564860e-01
3.74693364e-01 -8.81336033e-02 -6.40588403e-02 2.64419675e-01
4.12798971e-01 -8.59609425e-01 -1.01407838e+00 5.83765268e-01
-9.78252172e-01 4.02968526e-01 6.52417779e-01 9.32671368e-01
5.04587114e-01 1.20160408e-01 4.27518666e-01 -3.65438044e-01
1.82966948e-01 -1.95965528e-01 -5.40805459e-01 5.46339080e-02
-2.83289403e-01 -2.81893648e-02 8.78028870e-01 -6.46773458e-01
-9.68860865e-01 1.76412210e-01 2.91217476e-01 -8.55852187e-01
1.00724593e-01 2.77921170e-01 -2.47526959e-01 -8.28592628e-02
1.48712084e-01 6.03989244e-01 1.57338277e-01 -5.20842135e-01
1.22507617e-01 5.07157207e-01 4.07100379e-01 -3.62093717e-01
5.10343134e-01 6.29869580e-01 -9.07706991e-02 -5.18346906e-01
-6.81524813e-01 -2.07382202e-01 -5.45922756e-01 -3.22228551e-01
6.39165699e-01 -9.28462029e-01 -9.25338387e-01 4.30207640e-01
-1.07270133e+00 -1.99670911e-01 -7.78415129e-02 6.31140590e-01
-2.54695356e-01 6.50125384e-01 -7.59436429e-01 -8.09446216e-01
-2.99540341e-01 -1.20820129e+00 8.54614079e-01 5.21192074e-01
1.48710012e-01 -4.10601228e-01 -6.12765908e-01 1.56005904e-01
2.37077296e-01 -4.10572924e-02 6.08778954e-01 -1.21347032e-01
-1.03475320e+00 2.53655940e-01 -5.46541095e-01 2.81056970e-01
3.66550475e-01 -7.25300536e-02 -5.77696025e-01 -3.04538041e-01
4.92839329e-02 -1.70825601e-01 1.15614569e+00 2.34974250e-01
1.62417305e+00 -3.02408904e-01 -3.79455805e-01 7.38064945e-01
1.22422254e+00 6.29391253e-01 7.26869941e-01 3.89654994e-01
7.47268379e-01 4.80256915e-01 8.75921190e-01 4.81210083e-01
6.32492676e-02 6.45298839e-01 3.69084597e-01 -1.77229494e-01
-1.93710417e-01 -1.42205983e-01 3.92855465e-01 7.18835771e-01
-1.31980300e-01 8.85071158e-02 -4.40972209e-01 4.07289028e-01
-2.17655325e+00 -1.11080217e+00 3.26280862e-01 2.12221456e+00
7.00204730e-01 8.39225948e-02 1.47484049e-01 1.15730241e-01
7.53969848e-01 5.91583729e-01 -7.33020425e-01 3.48827064e-01
5.42747676e-02 -4.90592967e-04 3.28856051e-01 1.41695037e-01
-9.54579651e-01 9.25709903e-01 5.84127378e+00 1.05695999e+00
-1.24200511e+00 -1.31597832e-01 8.19012642e-01 -4.97933179e-01
-5.81488665e-03 1.42082676e-01 -7.54100263e-01 7.08208144e-01
3.85274559e-01 2.03611180e-02 4.46728587e-01 9.51239288e-01
6.05813503e-01 -1.22436389e-01 -8.08582067e-01 1.18486738e+00
-1.86547115e-01 -1.48381352e+00 3.45211118e-01 -6.74974471e-02
3.93595189e-01 -3.19966584e-01 -2.35608459e-01 1.64555326e-01
-1.98295116e-02 -6.43587768e-01 6.67433798e-01 5.22636950e-01
6.03877842e-01 -9.30622280e-01 3.51907343e-01 4.39832419e-01
-1.58753097e+00 -3.93870115e-01 -7.05619514e-01 -3.45692277e-01
2.83222556e-01 6.46692753e-01 -2.59237904e-02 5.82174361e-01
6.58874989e-01 1.03726435e+00 -3.97762060e-01 9.41373885e-01
8.24126005e-02 3.71869028e-01 -2.84615278e-01 -8.30727220e-02
4.56241369e-01 -4.98318791e-01 4.26544100e-01 8.82023513e-01
3.79860193e-01 5.87577522e-01 7.26465046e-01 7.26681232e-01
1.10848486e-01 -1.63773447e-01 -3.07752728e-01 1.45260081e-01
7.08926201e-01 1.17369330e+00 -6.41992927e-01 -4.54694062e-01
-6.53403938e-01 1.14752710e+00 4.27147895e-01 4.70455170e-01
-9.09821451e-01 -5.57352126e-01 6.21056616e-01 4.98801209e-02
4.82545614e-01 -2.96820462e-01 -1.54705748e-01 -1.40855920e+00
3.44179600e-01 -8.10668588e-01 3.69998693e-01 -6.66481256e-01
-8.17501366e-01 4.73036230e-01 -2.85341352e-01 -1.49953651e+00
6.17277212e-02 -4.21473831e-01 -4.99451846e-01 7.64962018e-01
-1.44028604e+00 -9.43483591e-01 -6.13501549e-01 8.14982414e-01
8.94418180e-01 -1.62941478e-02 2.42442593e-01 5.46771169e-01
-1.05996037e+00 4.50140446e-01 -2.46812701e-01 4.27680492e-01
4.64985967e-01 -6.12759888e-01 1.04269661e-01 1.05532658e+00
-1.94487706e-01 8.52203667e-01 1.12810388e-01 -5.58496594e-01
-1.24339175e+00 -1.18769670e+00 3.31915230e-01 2.98341811e-01
4.04095531e-01 -1.17536811e-02 -9.20713961e-01 4.47675824e-01
-1.58592522e-01 2.35271156e-01 3.50866348e-01 -3.29166681e-01
-3.70178849e-01 -4.79858637e-01 -8.07448804e-01 8.63453269e-01
1.32565653e+00 -2.22521618e-01 -4.66482341e-01 7.66315404e-03
8.36588562e-01 -3.16814959e-01 -5.83125174e-01 3.49758595e-01
5.95491827e-01 -1.07462239e+00 7.52832055e-01 -3.54807615e-01
5.07435918e-01 -7.90489733e-01 -6.10084198e-02 -8.04321766e-01
-5.88745356e-01 -5.94128728e-01 -3.06139499e-01 1.20736182e+00
-2.30378527e-02 -5.68802059e-01 8.70400608e-01 6.07196748e-01
-9.65648890e-02 -1.14951134e+00 -8.95264804e-01 -8.16874027e-01
-5.09328425e-01 -3.44097525e-01 6.73043191e-01 7.82351315e-01
-7.19867349e-02 1.09250769e-01 -5.59055686e-01 7.61525929e-02
4.36024308e-01 4.38606083e-01 6.82851374e-01 -8.79171610e-01
-3.62546444e-01 -4.09175992e-01 -4.06402498e-01 -1.90317500e+00
-1.03106402e-01 -3.64273965e-01 -2.24711835e-01 -1.20399594e+00
3.58229429e-01 -4.29577202e-01 -4.79276240e-01 7.62797356e-01
-3.02085102e-01 1.75610036e-01 4.04752344e-01 5.40031970e-01
-6.65853858e-01 7.58377373e-01 1.59557045e+00 -8.37761164e-02
-2.47558594e-01 -1.77332893e-01 -5.67143559e-01 8.62147748e-01
5.73475540e-01 -1.56018943e-01 -6.89463139e-01 -5.23347318e-01
-5.01051009e-01 1.25953466e-01 4.24825698e-01 -1.19854641e+00
2.59807348e-01 -3.60157907e-01 6.94117606e-01 -5.70269763e-01
2.14354724e-01 -8.43814492e-01 5.55470772e-03 5.73044300e-01
-2.32305184e-01 -5.78465573e-02 9.08613577e-02 6.54982090e-01
-4.33626175e-01 4.27785031e-02 9.20108199e-01 -8.63379091e-02
-9.56640959e-01 7.76938915e-01 -1.54529866e-02 -4.87610191e-01
1.18177581e+00 -4.71806318e-01 -1.63889498e-01 -1.64718837e-01
-6.03654325e-01 5.36837995e-01 3.81127536e-01 3.57746273e-01
7.39300609e-01 -1.61536181e+00 -1.99638933e-01 3.09018254e-01
-3.02574128e-01 1.38247445e-01 5.97925365e-01 1.02595961e+00
-4.94559824e-01 3.70335907e-01 -2.59036094e-01 -5.23709357e-01
-1.27623594e+00 6.59328818e-01 3.67736816e-01 -2.06477925e-01
-9.16418135e-01 5.93208671e-01 5.73176682e-01 2.11867392e-01
2.55106539e-01 -2.03871608e-01 -3.18165392e-01 -5.34774736e-02
9.58195269e-01 4.45365787e-01 -4.63946313e-01 -3.94783348e-01
-2.71118194e-01 5.61844528e-01 -4.52611774e-01 3.73311946e-03
1.30760133e+00 -2.56741792e-01 -1.24111660e-01 1.19367959e-02
1.14689600e+00 -1.78575352e-01 -1.89153564e+00 -2.61258006e-01
-4.18487161e-01 -8.67592037e-01 1.45139113e-01 -3.89945388e-01
-1.71528804e+00 8.32466066e-01 5.16953886e-01 -2.48253997e-02
1.71774387e+00 -5.57651997e-01 1.13559055e+00 3.67567927e-01
4.40555811e-01 -1.09003460e+00 1.41985878e-01 3.47559601e-01
5.86933672e-01 -8.11162412e-01 -6.48244768e-02 -6.71753228e-01
-5.57553709e-01 1.35292566e+00 1.02213621e+00 -1.87406123e-01
4.16342795e-01 9.87862200e-02 -2.22169235e-01 -1.44986421e-01
-5.93261898e-01 -6.31311908e-02 -7.52679855e-02 3.74379426e-01
8.12951252e-02 -3.68449777e-01 -6.36622548e-01 7.77711034e-01
4.03232336e-01 3.12086433e-01 2.69994974e-01 9.37421322e-01
-5.25675058e-01 -9.79721308e-01 -4.86640306e-03 3.96469802e-01
-3.21973532e-01 -1.04314141e-01 -3.71946730e-02 6.98481858e-01
2.79857993e-01 7.94728756e-01 1.87354028e-01 -4.72927630e-01
3.35192502e-01 -2.74268240e-01 9.06076431e-02 -3.99283528e-01
-2.09297344e-01 3.78821790e-01 -1.08842567e-01 -1.05859768e+00
-6.31164730e-01 -5.29913127e-01 -1.38738084e+00 -2.85224229e-01
-3.41700613e-01 7.43042752e-02 -1.06082469e-01 8.72367680e-01
6.79055631e-01 6.70859814e-01 6.97043598e-01 -7.60061681e-01
-4.52227414e-01 -9.21338260e-01 -5.99482417e-01 3.04463655e-01
2.72410721e-01 -7.51967549e-01 -1.96787670e-01 2.54067391e-01] | [9.207528114318848, -0.0316602848470211] |
f567ea52-c1a7-4909-a3da-f5a27461029e | that-s-bad-blind-anomaly-detection-by | 2307.03243 | null | https://arxiv.org/abs/2307.03243v1 | https://arxiv.org/pdf/2307.03243v1.pdf | That's BAD: Blind Anomaly Detection by Implicit Local Feature Clustering | Recent studies on visual anomaly detection (AD) of industrial objects/textures have achieved quite good performance. They consider an unsupervised setting, specifically the one-class setting, in which we assume the availability of a set of normal (\textit{i.e.}, anomaly-free) images for training. In this paper, we consider a more challenging scenario of unsupervised AD, in which we detect anomalies in a given set of images that might contain both normal and anomalous samples. The setting does not assume the availability of known normal data and thus is completely free from human annotation, which differs from the standard AD considered in recent studies. For clarity, we call the setting blind anomaly detection (BAD). We show that BAD can be converted into a local outlier detection problem and propose a novel method named PatchCluster that can accurately detect image- and pixel-level anomalies. Experimental results show that PatchCluster shows a promising performance without the knowledge of normal data, even comparable to the SOTA methods applied in the one-class setting needing it. | ['Takayuki Okatani', 'Masanori Suganuma', 'Jie Zhang'] | 2023-07-06 | null | null | null | null | ['anomaly-detection', 'clustering', 'outlier-detection'] | ['methodology', 'methodology', 'methodology'] | [ 5.56323588e-01 -4.78509739e-02 3.61305952e-01 -8.15689191e-02
-4.54590410e-01 -4.73456532e-01 6.27139270e-01 4.47223663e-01
-1.06496304e-01 1.82505190e-01 -5.08584380e-01 -4.19399261e-01
-2.90053338e-02 -5.97889185e-01 -9.87787068e-01 -1.08297884e+00
2.72856522e-02 2.83182979e-01 3.42284828e-01 1.01757124e-01
3.56392711e-01 5.31808376e-01 -1.86759198e+00 8.92499313e-02
7.95335829e-01 1.07358217e+00 -2.54256248e-01 9.25707698e-01
-5.10111563e-02 6.07985437e-01 -7.64715850e-01 -7.49477372e-02
5.11189640e-01 -3.69160682e-01 -6.05978251e-01 9.80158687e-01
8.10051739e-01 -2.42362365e-01 2.60364059e-02 1.39345133e+00
2.03790694e-01 1.13348708e-01 6.32142127e-01 -1.72103620e+00
-4.36125487e-01 -1.03379756e-01 -8.78396809e-01 4.36170608e-01
2.72790343e-01 2.28829145e-01 6.94479585e-01 -8.62859547e-01
4.45391566e-01 9.04908001e-01 3.57435405e-01 3.85988027e-01
-1.59279978e+00 -2.86381692e-01 5.02432644e-01 2.89758682e-01
-1.15031159e+00 -3.48521978e-01 7.95517266e-01 -8.34029436e-01
6.19164526e-01 3.90670121e-01 2.75578558e-01 1.26478481e+00
-2.25156583e-02 8.04612339e-01 1.02050531e+00 -7.44423330e-01
6.99815869e-01 -1.57349706e-01 3.12791646e-01 2.84703404e-01
6.50801957e-01 -7.52061009e-02 -2.50832260e-01 -4.77512330e-01
5.61198235e-01 3.55481863e-01 -2.17938244e-01 -8.11703146e-01
-1.17762256e+00 7.19093680e-01 9.08141136e-02 2.68841088e-01
-4.72418904e-01 -2.77772427e-01 5.10264337e-01 7.08000064e-01
5.59144020e-01 2.74669588e-01 -2.25512609e-01 1.96333095e-01
-7.17189133e-01 8.06982666e-02 5.08797944e-01 8.87858689e-01
6.30906761e-01 3.01413387e-01 9.09591541e-02 6.27926886e-01
1.05813928e-01 4.82250780e-01 5.97762883e-01 -7.51905620e-01
3.56214195e-01 5.99088788e-01 2.22124249e-01 -1.02952766e+00
-9.84056219e-02 -2.02531055e-01 -9.56565857e-01 7.27923691e-01
8.69025588e-01 2.49013618e-01 -1.22073007e+00 1.37403953e+00
3.75786901e-01 5.12503624e-01 3.99445593e-02 6.54452562e-01
5.54264598e-02 3.34898233e-01 -3.97411346e-01 -3.89474243e-01
9.40531969e-01 -7.65140593e-01 -7.97410131e-01 -2.30997086e-01
5.35105348e-01 -7.39871323e-01 1.28127801e+00 1.14242196e+00
-6.54868364e-01 -3.08204085e-01 -1.22408223e+00 5.37885547e-01
-5.80728948e-01 -9.99253020e-02 1.52024686e-01 6.63566530e-01
-8.15566063e-01 5.35117388e-01 -9.94599104e-01 -7.77244329e-01
3.48116040e-01 3.89213637e-02 -4.84395593e-01 -3.41860175e-01
-3.40956390e-01 2.22972497e-01 3.24864805e-01 1.53700083e-01
-1.10564399e+00 -1.06979944e-01 -8.24734867e-01 -3.43955457e-01
8.80659044e-01 -1.60439283e-01 1.16545749e+00 -1.56555998e+00
-9.86737669e-01 9.43535686e-01 -1.81079939e-01 -4.55121279e-01
7.22505867e-01 -3.62455785e-01 -5.94731748e-01 1.00421242e-01
1.62676975e-01 -2.11904794e-01 1.36960554e+00 -1.49156725e+00
-5.99109709e-01 -5.98518848e-01 -1.75774947e-01 -3.93200547e-01
-2.82807291e-01 -6.38165623e-02 -3.92236233e-01 -8.20180357e-01
4.10249382e-01 -8.71321619e-01 -2.82559633e-01 1.29291996e-01
-8.89490962e-01 9.56138149e-02 1.25271738e+00 -3.85412097e-01
1.04591417e+00 -2.44051051e+00 -3.62801969e-01 7.66386151e-01
2.40070239e-01 2.45747656e-01 3.90831428e-03 2.56965220e-01
-5.40907681e-01 -3.29162292e-02 -7.64393330e-01 -3.31717223e-01
-7.94629529e-02 6.73150539e-01 -4.43265915e-01 8.11887920e-01
3.33757520e-01 2.43846387e-01 -8.32833052e-01 -2.70295799e-01
3.15267891e-01 -2.58960217e-01 -4.16390300e-01 2.15223640e-01
-2.25229368e-01 6.40618742e-01 -4.52659309e-01 8.91763389e-01
8.68948758e-01 -1.62360176e-01 -8.91376436e-02 3.95834208e-01
6.75102696e-02 -5.11204541e-01 -1.57461083e+00 1.40934873e+00
6.42362633e-04 6.53779209e-01 6.01855144e-02 -1.35353065e+00
6.98739707e-01 2.32021004e-01 5.04858077e-01 -5.98368883e-01
-2.23233283e-01 1.90704584e-01 1.11565679e-01 -5.84325194e-01
2.09049121e-01 2.41528168e-01 1.13752067e-01 4.13585633e-01
-4.88910936e-02 3.43612552e-01 1.68752417e-01 1.29632354e-01
1.51059103e+00 -1.19946375e-01 4.22540486e-01 -1.90247208e-01
5.25504231e-01 -9.45439264e-02 6.40979290e-01 1.18238807e+00
-3.73566180e-01 9.25348341e-01 8.87426138e-01 -2.10130528e-01
-1.20211971e+00 -1.38015628e+00 -1.62818402e-01 7.59464622e-01
8.92939493e-02 -2.85165578e-01 -6.43524766e-01 -8.67848098e-01
5.32755926e-02 4.48717743e-01 -6.46628857e-01 -3.62887792e-02
-3.82551312e-01 -9.19106543e-01 2.13941589e-01 3.22670639e-01
2.28879213e-01 -8.66398275e-01 -4.53412980e-01 -2.32291427e-02
9.59289372e-02 -1.09919107e+00 -3.36153805e-02 2.07245663e-01
-8.91322374e-01 -1.38168550e+00 -5.13373077e-01 -5.39014578e-01
1.02354479e+00 1.75168160e-02 1.11379516e+00 9.26907584e-02
-6.16072357e-01 7.22745478e-01 -4.11363065e-01 -5.60056984e-01
-4.03764069e-01 -5.60540140e-01 2.18146250e-01 5.91153324e-01
3.95651937e-01 -5.94825089e-01 -4.71793890e-01 4.27352041e-01
-1.47642505e+00 -6.70535386e-01 5.29327512e-01 7.39336908e-01
9.20302749e-01 2.31796592e-01 3.44878674e-01 -1.06821144e+00
1.73333883e-01 -7.32955515e-01 -6.13365471e-01 -4.73348089e-02
-7.58338451e-01 -2.35236764e-01 7.01980472e-01 -3.66469979e-01
-6.53117120e-01 2.09947944e-01 1.29433319e-01 -6.95431530e-01
-8.44559789e-01 5.64228334e-02 -4.75223482e-01 4.08291630e-02
6.56106710e-01 1.53184861e-01 -7.31825903e-02 -7.08732367e-01
-3.78113687e-02 6.49977326e-01 8.54406655e-01 -4.11746830e-01
1.00679898e+00 7.22551823e-01 4.37630713e-02 -1.05849063e+00
-4.71206278e-01 -7.88984358e-01 -8.13720465e-01 -1.24857418e-01
6.85233951e-01 -6.68113112e-01 -2.44676873e-01 7.86369324e-01
-8.09544086e-01 -2.42549315e-01 -6.76939666e-01 2.15373725e-01
-6.53922200e-01 9.52668071e-01 -2.68621236e-01 -1.04455066e+00
2.72382766e-01 -9.20953274e-01 1.04700184e+00 -2.50118017e-01
7.04160556e-02 -8.70634675e-01 -9.72836316e-02 -6.19298555e-02
1.11160524e-01 8.30290973e-01 8.36552441e-01 -1.27456677e+00
-5.34505248e-01 -3.35981071e-01 6.78464025e-02 8.41164768e-01
3.48541051e-01 1.66221157e-01 -1.26158094e+00 -3.59257162e-01
1.48290098e-01 1.31982490e-01 7.19314992e-01 2.54000276e-01
1.53236330e+00 -4.11131054e-01 -2.15071797e-01 2.29831144e-01
1.32930708e+00 1.31787822e-01 7.29713380e-01 6.10578239e-01
7.65247822e-01 3.71406704e-01 8.34344268e-01 6.24442041e-01
-2.75899202e-01 7.39005327e-01 9.81924713e-01 -3.75545442e-01
3.81351709e-01 3.54678541e-01 4.52779621e-01 2.55533457e-01
9.22960788e-02 -4.06713158e-01 -1.01057768e+00 7.76187956e-01
-2.13681841e+00 -9.07154381e-01 -4.35775995e-01 2.65851712e+00
2.77172595e-01 2.25864395e-01 3.86388242e-01 7.23000228e-01
8.90641570e-01 9.46803205e-03 -6.81111217e-01 -4.52166080e-01
-1.35822400e-01 -6.84412643e-02 3.83505970e-01 1.30672812e-01
-1.53490877e+00 1.77731782e-01 5.86935711e+00 5.08811176e-01
-7.73571968e-01 -3.36061567e-02 5.94341338e-01 1.03832297e-01
2.23914221e-01 -1.96710438e-01 -1.93260342e-01 6.57330990e-01
9.23754990e-01 1.29019529e-01 1.55030852e-02 1.04549241e+00
1.35230854e-01 -2.99042672e-01 -1.29976332e+00 8.86584580e-01
3.07984978e-01 -4.06214952e-01 1.92168858e-02 2.91343808e-01
7.67627716e-01 -1.77148059e-01 1.12053178e-01 -1.03900179e-01
1.35189697e-01 -7.58133054e-01 5.06517112e-01 6.93339944e-01
3.93615574e-01 -6.45302534e-01 8.59249771e-01 3.52183133e-01
-7.81715512e-01 -2.82999992e-01 -1.96548536e-01 2.95231380e-02
-2.26337075e-01 9.52148080e-01 -5.71899772e-01 7.11338282e-01
1.00000393e+00 8.82661581e-01 -8.78367305e-01 1.36984682e+00
1.90647274e-01 8.13260198e-01 -4.37922329e-01 8.39066982e-01
1.78191528e-01 -3.02143306e-01 9.32045281e-01 9.53483641e-01
4.75573301e-01 -5.01816511e-01 5.64359546e-01 5.78739226e-01
2.93674260e-01 1.20620385e-01 -1.06136250e+00 2.50538319e-01
-6.87777624e-02 8.62191379e-01 -9.41887736e-01 -3.19556266e-01
-6.72398210e-01 1.32579803e+00 -2.36517236e-01 4.70099032e-01
-4.35219675e-01 -3.16060364e-01 7.90738463e-01 1.84775978e-01
5.86675406e-01 -7.87200183e-02 -1.14098102e-01 -1.19969893e+00
5.33410788e-01 -9.29353237e-01 8.23332489e-01 -5.70509017e-01
-1.73896694e+00 2.04452142e-01 8.13963562e-02 -1.70215964e+00
-3.59908134e-01 -9.24544930e-01 -8.80054832e-01 4.04432178e-01
-1.33902609e+00 -7.58925974e-01 -4.75547612e-01 9.21016395e-01
5.81991792e-01 -1.40028670e-01 6.03663862e-01 1.89917475e-01
-8.20388079e-01 4.06480581e-01 5.25389493e-01 1.57149374e-01
7.48904228e-01 -1.84658051e+00 2.54690737e-01 1.68444359e+00
1.90584198e-01 2.18368426e-01 8.94300520e-01 -5.00005066e-01
-1.01579034e+00 -1.43689799e+00 3.49669755e-01 -6.37050211e-01
7.53468335e-01 -4.48971957e-01 -1.20448411e+00 1.00849450e+00
-1.55586913e-01 6.85470164e-01 4.88065958e-01 -2.38441050e-01
-2.21835330e-01 -9.32969302e-02 -1.17271292e+00 4.50050414e-01
8.78830612e-01 -2.62436718e-01 -4.48778778e-01 5.20554900e-01
3.80006552e-01 -2.57291406e-01 -5.03816485e-01 2.36651361e-01
-5.44607118e-02 -1.10545778e+00 8.58966708e-01 -6.27511680e-01
1.32327499e-02 -6.91909194e-01 -2.86060125e-01 -1.08743274e+00
1.18005373e-01 -4.63723779e-01 -3.95360142e-01 1.26442814e+00
9.72579345e-02 -8.84677291e-01 4.98614341e-01 2.37913311e-01
-2.40762606e-01 -2.92082012e-01 -1.06668103e+00 -1.21735227e+00
-1.27971441e-01 -6.79705560e-01 4.69164759e-01 9.46928620e-01
-4.56327826e-01 -2.51551151e-01 -2.48057380e-01 6.77708328e-01
7.08508015e-01 -2.31994633e-02 9.86815333e-01 -1.51543009e+00
-2.25557074e-01 -1.86640635e-01 -1.03311789e+00 -5.55712879e-01
2.11614240e-02 -4.65005994e-01 2.98047572e-01 -1.03504920e+00
-8.72917287e-03 -5.09628356e-02 -5.35379708e-01 4.35198814e-01
-1.10708222e-01 4.73805815e-01 -3.08014482e-01 3.42997491e-01
-7.05037773e-01 2.86792994e-01 6.63503408e-01 -1.28209382e-01
-3.92166004e-02 2.73477316e-01 -3.17155421e-01 1.01734805e+00
6.72067404e-01 -5.33152282e-01 -2.90613592e-01 7.20672533e-02
5.84842218e-03 -5.55080652e-01 8.60567331e-01 -1.13198924e+00
-4.02111635e-02 -1.56840384e-01 3.55351925e-01 -4.74995941e-01
-1.89025953e-01 -1.24061263e+00 -2.37225518e-01 1.58923924e-01
-1.09210592e-02 4.32953566e-01 9.80016589e-02 1.21853697e+00
-3.27012479e-01 -2.52056658e-01 6.03099704e-01 -6.23950874e-03
-9.40739274e-01 2.03667119e-01 -6.67392254e-01 -1.26086503e-01
1.40723121e+00 -5.04773557e-01 -2.99412459e-01 -3.33122909e-01
-1.13874650e+00 -6.96760267e-02 8.14926267e-01 2.90863872e-01
4.98795927e-01 -1.24548566e+00 -5.08379400e-01 5.42764544e-01
7.68760264e-01 1.79758698e-01 1.35349855e-01 1.14866579e+00
-2.89295346e-01 -1.27486378e-01 -9.42365974e-02 -1.11529922e+00
-1.20981395e+00 9.23229575e-01 2.07086101e-01 -9.31883752e-02
-1.07079685e+00 1.10027045e-01 4.31027800e-01 -1.35177955e-01
2.54903585e-01 -3.98444474e-01 -1.78481005e-02 -1.68843001e-01
4.38234508e-01 3.98438185e-01 5.15358686e-01 -6.54936671e-01
-2.59070843e-01 2.72956312e-01 -1.10411651e-01 1.66536584e-01
9.68275845e-01 -2.04191282e-01 -1.53293163e-01 9.22811270e-01
7.99236238e-01 1.78080246e-01 -1.20582831e+00 -3.64836067e-01
5.34447670e-01 -7.42672622e-01 -2.37989530e-01 -4.23799723e-01
-1.13166559e+00 6.74077511e-01 1.08241403e+00 5.89843869e-01
1.41992676e+00 2.14008789e-04 2.94090837e-01 5.26683509e-01
-3.03593893e-02 -1.10582256e+00 3.75484616e-01 3.52153718e-01
6.34870946e-01 -1.40868747e+00 -1.60502464e-01 -2.55508631e-01
-4.35360700e-01 1.12926674e+00 7.41665661e-01 -2.99455822e-01
6.41973257e-01 1.01675503e-01 1.79965004e-01 -2.13268772e-01
-4.47805166e-01 -3.85130048e-01 2.82707065e-01 8.98971677e-01
-1.16311073e-01 -2.04151988e-01 3.47820342e-01 2.93586969e-01
3.05601150e-01 -3.71662855e-01 8.42718542e-01 1.27009237e+00
-3.10393065e-01 -8.96490514e-01 -7.68419802e-01 6.31771445e-01
-5.54600894e-01 2.84577906e-01 -3.82226914e-01 9.17153716e-01
2.43652776e-01 1.08269763e+00 4.65251863e-01 -2.19758060e-02
6.25257313e-01 4.14914280e-01 3.07781063e-02 -6.05324626e-01
-4.71107587e-02 -3.59764397e-02 -2.87716895e-01 -9.76369560e-01
-4.75815386e-01 -8.30544829e-01 -7.46600747e-01 4.87117395e-02
-3.41979176e-01 -7.13679194e-02 3.69604737e-01 1.04127824e+00
2.31473505e-01 4.70716715e-01 7.91032255e-01 -6.30512238e-01
-3.81365329e-01 -7.31405139e-01 -1.01354063e+00 8.28080297e-01
9.17201579e-01 -6.21946394e-01 -7.95878291e-01 3.26218992e-01] | [7.6765546798706055, 2.2155163288116455] |
f6e00d50-ab17-44ff-bfae-1163c93e5d32 | rethinking-diversified-and-discriminative | 1805.03508 | null | http://arxiv.org/abs/1805.03508v1 | http://arxiv.org/pdf/1805.03508v1.pdf | Rethinking Diversified and Discriminative Proposal Generation for Visual Grounding | Visual grounding aims to localize an object in an image referred to by a
textual query phrase. Various visual grounding approaches have been proposed,
and the problem can be modularized into a general framework: proposal
generation, multi-modal feature representation, and proposal ranking. Of these
three modules, most existing approaches focus on the latter two, with the
importance of proposal generation generally neglected. In this paper, we
rethink the problem of what properties make a good proposal generator. We
introduce the diversity and discrimination simultaneously when generating
proposals, and in doing so propose Diversified and Discriminative Proposal
Networks model (DDPN). Based on the proposals generated by DDPN, we propose a
high performance baseline model for visual grounding and evaluate it on four
benchmark datasets. Experimental results demonstrate that our model delivers
significant improvements on all the tested data-sets (e.g., 18.8\% improvement
on ReferItGame and 8.2\% improvement on Flickr30k Entities over the existing
state-of-the-arts respectively) | ['DaCheng Tao', 'Zhou Zhao', 'Qi Tian', 'Zhou Yu', 'Jun Yu', 'Chenchao Xiang'] | 2018-05-09 | null | null | null | null | ['phrase-grounding'] | ['natural-language-processing'] | [ 2.02859472e-02 3.40318948e-01 -4.52929974e-01 -1.50982201e-01
-1.03340542e+00 -5.26399255e-01 8.00585210e-01 -9.39478725e-03
-2.29393080e-01 4.50409412e-01 4.03212070e-01 3.25054526e-02
1.21719152e-01 -6.18643820e-01 -6.84171557e-01 -4.86470163e-01
2.03574479e-01 3.61394167e-01 7.09815443e-01 -3.44810486e-01
3.65108103e-01 2.12820947e-01 -1.54495788e+00 4.51999277e-01
6.43114448e-01 1.02298045e+00 3.79466116e-01 1.46657646e-01
-2.07124308e-01 6.32583201e-01 -3.95598501e-01 -5.55120111e-01
2.22793356e-01 -2.16180667e-01 -8.38909388e-01 2.04322800e-01
8.62729907e-01 -3.39636773e-01 -5.03486931e-01 9.75115120e-01
5.60773194e-01 2.08415389e-01 7.22194374e-01 -1.51917183e+00
-7.96298683e-01 6.88563883e-01 -7.46314645e-01 2.22812653e-01
4.15892452e-01 -7.20012039e-02 1.45029342e+00 -1.49419236e+00
1.01105702e+00 1.29445243e+00 3.11353773e-01 5.36871850e-01
-1.09568226e+00 -7.00691283e-01 5.27149498e-01 1.82014436e-01
-1.82693374e+00 -3.59678417e-01 7.28727937e-01 -3.96311760e-01
6.39976263e-01 3.07374477e-01 4.86690879e-01 8.44737470e-01
-1.50270924e-01 1.24076331e+00 8.63168895e-01 -3.41157854e-01
-2.74660066e-03 7.89207667e-02 8.04763064e-02 8.96427572e-01
3.27985793e-01 -4.99933511e-02 -7.63421774e-01 -1.55462012e-01
7.84220874e-01 -1.50735036e-03 -2.93028116e-01 -7.08716691e-01
-1.52177846e+00 7.25975156e-01 8.85038674e-01 1.65364072e-01
-1.80773139e-01 2.83480316e-01 1.98508576e-01 -6.15767613e-02
2.78376997e-01 3.59661877e-01 1.07423421e-02 4.91439968e-01
-1.08938003e+00 6.62074566e-01 3.36687654e-01 1.23671937e+00
8.60620856e-01 -1.48139387e-01 -8.85985136e-01 6.87377632e-01
7.94805765e-01 4.44657177e-01 2.89920151e-01 -6.47352040e-01
8.70308876e-01 8.02712262e-01 4.14440453e-01 -1.43854201e+00
-3.12235296e-01 -5.17061353e-01 -6.47172213e-01 -2.48764366e-01
-3.93397696e-02 1.75748065e-01 -1.13943923e+00 1.50452769e+00
4.35755640e-01 2.14011759e-01 -6.12212718e-02 1.06955802e+00
1.77578843e+00 7.90415585e-01 2.30746567e-01 2.43591014e-02
1.34802091e+00 -1.40857518e+00 -5.98099172e-01 -2.02227533e-01
4.01888371e-01 -8.50117147e-01 9.89611685e-01 1.03415705e-01
-1.07955086e+00 -7.76874304e-01 -1.03719687e+00 -1.98747829e-01
-1.90054432e-01 6.43881202e-01 5.92223346e-01 2.47965783e-01
-1.23266685e+00 2.34858587e-01 -4.35026973e-01 -4.99471247e-01
3.20688725e-01 3.69348645e-01 -3.84950250e-01 1.89394560e-02
-1.04741728e+00 6.23792171e-01 6.63247585e-01 1.12218857e-01
-1.08537543e+00 -4.99084771e-01 -7.12780416e-01 1.71276871e-02
4.91774350e-01 -8.49715173e-01 1.21244156e+00 -5.85393131e-01
-1.11916912e+00 1.09362578e+00 -1.13832131e-02 -3.72698724e-01
4.81294543e-01 -2.30148345e-01 -4.00835156e-01 1.98748931e-01
4.33922261e-01 1.34594858e+00 7.69352853e-01 -1.55154943e+00
-1.11352396e+00 -1.20959273e-02 3.91431600e-01 4.66083258e-01
-2.67972082e-01 6.77668303e-02 -1.22524619e+00 -7.74991989e-01
3.65737438e-01 -1.10701215e+00 -2.53568530e-01 -8.03567469e-03
-6.87286675e-01 -5.62525094e-01 5.25841355e-01 -4.38816756e-01
1.44153941e+00 -2.10469532e+00 2.88091302e-01 9.26004946e-02
3.43440384e-01 2.04657018e-01 -3.43225151e-02 5.98493457e-01
2.28833824e-01 1.57661587e-01 2.83563882e-01 -4.36458975e-01
3.06521177e-01 -1.04640834e-01 -5.57074487e-01 2.75527447e-01
1.12716645e-01 1.24929559e+00 -9.04380441e-01 -8.77120852e-01
4.95109297e-02 2.69979447e-01 -4.38243330e-01 2.00674579e-01
-1.47177577e-01 1.29491746e-01 -7.05343902e-01 8.23791444e-01
4.84276086e-01 -5.89150906e-01 5.80150560e-02 -5.65851450e-01
-4.47782688e-03 -7.99022838e-02 -1.44674623e+00 2.08015656e+00
2.04379886e-01 3.31490427e-01 -3.37317973e-01 -5.90386748e-01
1.06789076e+00 1.67687595e-01 5.04942201e-02 -6.39074147e-01
-1.70440525e-01 2.06539467e-01 -2.19992846e-01 -2.96475023e-01
1.13831937e+00 1.93449259e-01 -2.97879398e-01 2.31770381e-01
2.07663223e-01 1.35337994e-01 2.79047817e-01 6.68653488e-01
7.30985641e-01 2.40857735e-01 2.80120492e-01 -1.25096291e-01
3.31098646e-01 1.61813557e-01 4.55238998e-01 9.05494690e-01
-1.98344246e-01 8.47136855e-01 4.20351505e-01 -4.64151084e-01
-8.27848911e-01 -1.03330088e+00 2.59617180e-01 1.30592835e+00
8.09259892e-01 -6.98344409e-01 -4.18765038e-01 -1.06599379e+00
-6.52406886e-02 3.57186019e-01 -7.78211951e-01 7.00197369e-02
-5.80447257e-01 -5.44177413e-01 2.70355493e-01 6.65335894e-01
6.13280296e-01 -1.17543554e+00 -3.14535737e-01 -1.14027314e-01
-4.14255828e-01 -1.02302086e+00 -5.51463127e-01 -3.52802008e-01
-6.91051245e-01 -7.89042354e-01 -8.56736183e-01 -9.10821617e-01
7.52479196e-01 7.58363008e-01 1.34357512e+00 1.88950613e-01
1.72170792e-02 4.01422411e-01 -4.50347096e-01 -1.78525582e-01
4.74500693e-02 3.20099235e-01 -2.27740332e-01 1.36933938e-01
2.53341403e-02 -3.12616602e-02 -9.34434950e-01 5.58851719e-01
-7.97931612e-01 2.78809965e-01 8.95386338e-01 6.12776518e-01
9.79935706e-01 -4.08886462e-01 5.14426351e-01 -6.53300285e-01
3.57506275e-01 -3.43594491e-01 -4.41320568e-01 6.33094490e-01
-4.93143201e-01 9.45456699e-02 -6.36934862e-02 -7.55472183e-02
-9.66375470e-01 1.03576943e-01 1.93752974e-01 -4.67106491e-01
3.31466906e-02 4.79148924e-01 -1.91936269e-01 -1.60066873e-01
5.72055399e-01 1.07866608e-01 -5.03259003e-01 -2.90969551e-01
7.45377183e-01 4.09919083e-01 4.20217127e-01 -5.33922911e-01
9.49927568e-01 6.31706178e-01 -1.27682984e-01 -3.38294536e-01
-9.08594608e-01 -7.25564957e-01 -4.60885227e-01 -3.63714516e-01
7.37169385e-01 -1.15161848e+00 -2.98438758e-01 -1.38981223e-01
-1.04223692e+00 9.31727663e-02 -2.16846704e-01 2.15143800e-01
-3.48768830e-01 2.73495376e-01 -1.33441284e-01 -5.30333936e-01
-5.29830813e-01 -1.17553890e+00 1.36964571e+00 3.55086774e-01
1.00316033e-02 -5.17639756e-01 5.91846928e-02 3.30274642e-01
2.19017997e-01 2.23296732e-01 5.48542619e-01 -5.17810583e-01
-1.03244543e+00 -1.52950481e-01 -6.30085409e-01 -1.42224222e-01
-3.12734067e-01 -4.78299931e-02 -8.34838927e-01 -4.06417996e-01
-7.79538035e-01 -5.15128136e-01 1.19917679e+00 2.11156487e-01
1.01455069e+00 -1.07150286e-01 -8.67632270e-01 4.46943045e-01
1.43342078e+00 -1.09158941e-01 6.21061444e-01 3.90863538e-01
8.20780993e-01 5.42365193e-01 1.09653378e+00 2.27831572e-01
6.36926174e-01 1.19048953e+00 7.79424250e-01 -2.46487677e-01
-2.80976564e-01 -6.59612775e-01 4.46608253e-02 5.36279380e-01
-3.09377193e-01 -4.30812716e-01 -8.60582054e-01 7.63707578e-01
-2.23604417e+00 -9.04996872e-01 -4.51363064e-02 1.85753000e+00
4.66913700e-01 7.39203766e-02 3.57729375e-01 -2.88691282e-01
8.59518170e-01 4.32090640e-01 -2.31358007e-01 2.88384318e-01
-7.09685087e-02 -2.35508442e-01 2.06679612e-01 1.35854751e-01
-1.43369794e+00 1.16400421e+00 5.99333429e+00 1.12398612e+00
-9.76265728e-01 1.23168796e-01 5.69871187e-01 1.02869995e-01
-4.18757796e-01 -8.95494130e-03 -1.29817748e+00 6.30083755e-02
2.57621080e-01 -1.94542944e-01 -2.95497775e-01 9.57725883e-01
-4.80993055e-02 1.95172966e-01 -9.94997859e-01 1.16526520e+00
3.66004080e-01 -1.53793538e+00 4.25123304e-01 5.19502275e-02
9.12432492e-01 -2.82649696e-02 2.80784875e-01 4.91697341e-01
2.05050230e-01 -9.08649325e-01 1.09524500e+00 4.98001367e-01
6.18120670e-01 -5.89363635e-01 7.39300728e-01 2.56472919e-02
-1.55954957e+00 4.15235013e-02 -4.73775387e-01 3.81924510e-01
1.96783990e-01 1.05746634e-01 -9.53379154e-01 9.68829632e-01
6.41205132e-01 7.51577556e-01 -8.74193311e-01 1.40964830e+00
-3.93922567e-01 5.07720053e-01 -4.15472090e-02 -1.31863683e-01
4.93250459e-01 2.53990918e-01 4.22142178e-01 1.02213812e+00
3.72672558e-01 -8.16743597e-02 6.12413228e-01 7.77234912e-01
-2.57532269e-01 4.62661803e-01 -3.16033363e-01 1.64311796e-01
4.76103783e-01 1.59220004e+00 -1.04938209e+00 -3.69492352e-01
-3.55571240e-01 8.23003709e-01 4.81211126e-01 3.99034262e-01
-1.03858376e+00 -1.69516370e-01 8.64529703e-03 1.81239903e-01
6.35757208e-01 7.11008906e-02 2.94759840e-01 -1.21373510e+00
1.28304079e-01 -7.11607158e-01 5.53315639e-01 -8.96980047e-01
-1.15852010e+00 6.98537290e-01 2.43570015e-01 -1.55417919e+00
9.51786041e-02 -4.16215628e-01 -4.52821821e-01 5.88757932e-01
-1.51913834e+00 -1.66625822e+00 -6.08290851e-01 3.42501700e-01
5.02743065e-01 -9.73737389e-02 5.55532277e-01 2.84393102e-01
-4.03339118e-01 5.97476065e-01 -2.47083187e-01 9.03586969e-02
7.82662511e-01 -1.25135124e+00 4.92288321e-01 8.95085335e-01
6.65112734e-01 6.94120586e-01 4.62661147e-01 -5.94163954e-01
-1.09316039e+00 -1.38059616e+00 1.03053260e+00 -4.91806239e-01
4.82306927e-01 -3.05266857e-01 -5.68322003e-01 7.03961313e-01
3.75588924e-01 3.76591176e-01 5.49361587e-01 -9.12622362e-02
-3.49418789e-01 -7.75282159e-02 -7.68752694e-01 7.46672630e-01
1.22827089e+00 -1.26070961e-01 -7.04159021e-01 3.28242213e-01
8.67717803e-01 -6.15594625e-01 -5.10466337e-01 5.52833438e-01
6.31591916e-01 -8.54327202e-01 1.23735225e+00 -6.06044054e-01
3.26092213e-01 -8.12506557e-01 -2.98709422e-01 -8.14903140e-01
-6.51589990e-01 -5.38393021e-01 -2.45883137e-01 1.38969707e+00
4.07419443e-01 2.54090466e-02 9.71304178e-01 2.27571040e-01
-3.00645202e-01 -8.27116132e-01 -7.61093378e-01 -5.70395410e-01
-4.02968824e-01 -2.15957254e-01 6.15339816e-01 6.88256979e-01
-1.53192580e-01 4.38590407e-01 -5.74104369e-01 2.08925843e-01
4.65091169e-01 4.45771873e-01 1.02536976e+00 -1.03802884e+00
-1.58337072e-01 -2.39288688e-01 -6.21832550e-01 -1.43826568e+00
1.76522229e-02 -1.05529106e+00 1.46904781e-01 -1.98889101e+00
6.57224119e-01 -3.74098748e-01 -5.04634857e-01 5.08257210e-01
-4.63243455e-01 5.64524353e-01 4.51044708e-01 4.97394949e-01
-1.45399654e+00 6.50290370e-01 1.02423751e+00 -4.17967647e-01
-1.96266785e-01 -1.75864950e-01 -7.73169518e-01 7.08086550e-01
3.54564309e-01 -3.56334358e-01 -5.43793082e-01 -4.63703811e-01
4.09493655e-01 -1.85336098e-01 5.77052116e-01 -8.85379732e-01
1.70403913e-01 -1.89972684e-01 2.14488730e-01 -1.13985860e+00
2.80864865e-01 -5.52506864e-01 1.03187025e-01 2.59738296e-01
-4.15553212e-01 1.27682865e-01 5.51186167e-02 9.21483517e-01
-3.93003434e-01 -3.13532092e-02 3.60222548e-01 -1.28798693e-01
-1.30379295e+00 4.23247695e-01 2.00492308e-01 1.49100170e-01
1.12471688e+00 -8.02652389e-02 -5.09235442e-01 -1.99768007e-01
-7.20293820e-01 3.36007267e-01 2.82752454e-01 5.69777727e-01
8.46328795e-01 -1.77788198e+00 -7.22491384e-01 -3.05681616e-01
6.68533266e-01 -1.55390605e-01 2.66059816e-01 8.02863061e-01
-4.16308135e-01 6.79267526e-01 -4.00840938e-02 -6.47824109e-01
-1.36118543e+00 5.18083274e-01 6.21017739e-02 -4.77445066e-01
-4.19880897e-01 1.21696234e+00 6.26344085e-01 -9.84316692e-02
3.65687102e-01 -3.65810186e-01 -6.13938570e-01 1.38786659e-01
4.21508670e-01 1.20559849e-01 -1.02186918e-01 -1.04935467e+00
-4.54950541e-01 7.13601708e-01 -3.49520177e-01 -8.11224878e-02
1.05187750e+00 -2.12872744e-01 1.59724057e-01 2.56597430e-01
7.67581224e-01 -8.71919841e-02 -9.24678087e-01 -4.10729557e-01
-1.03830487e-01 -4.84581500e-01 -6.18142709e-02 -5.41153312e-01
-1.06395292e+00 7.41211236e-01 6.36215329e-01 2.63493538e-01
7.91171491e-01 4.52761620e-01 4.08705324e-01 4.42200840e-01
5.68400443e-01 -8.23126793e-01 5.01357615e-01 1.24957532e-01
1.17893922e+00 -1.20610905e+00 2.10008144e-01 -5.59253335e-01
-7.83814728e-01 7.65190184e-01 9.19181287e-01 -1.42777845e-01
3.51508498e-01 -6.05325401e-01 -1.99777365e-01 -5.20321548e-01
-5.68954706e-01 -3.96270484e-01 8.83938611e-01 3.21287096e-01
3.85382950e-01 -5.46857305e-02 -4.12499815e-01 5.65755069e-01
1.06405385e-01 -1.24053583e-01 1.44102871e-01 7.14067161e-01
-6.16490841e-01 -1.03476298e+00 -3.03038090e-01 3.11048299e-01
-3.04683566e-01 -1.10853329e-01 -5.03990352e-01 9.26241398e-01
2.81468421e-01 7.85490334e-01 -1.99081153e-01 -3.98946792e-01
5.25399685e-01 -1.88098297e-01 2.00538635e-01 -8.73087585e-01
-4.57830369e-01 1.18116014e-01 1.94000974e-01 -6.70397818e-01
-8.72204959e-01 -5.62956035e-01 -9.45323348e-01 1.88877136e-01
-7.15812147e-01 2.26337358e-01 2.36797333e-01 7.54749715e-01
5.34568310e-01 4.24574792e-01 2.30585366e-01 -9.68308687e-01
-3.01728636e-01 -9.10290897e-01 -4.62873757e-01 4.07127440e-01
-7.67427906e-02 -9.02344108e-01 -1.08090162e-01 1.18532747e-01] | [10.404012680053711, 1.4689068794250488] |
5c9b0fcd-2b68-4834-a5f8-9dd8c3f8b79e | adic-anomaly-detection-integrated-circuit-in | 2008.09442 | null | https://arxiv.org/abs/2008.09442v1 | https://arxiv.org/pdf/2008.09442v1.pdf | ADIC: Anomaly Detection Integrated Circuit in 65nm CMOS utilizing Approximate Computing | In this paper, we present a low-power anomaly detection integrated circuit (ADIC) based on a one-class classifier (OCC) neural network. The ADIC achieves low-power operation through a combination of (a) careful choice of algorithm for online learning and (b) approximate computing techniques to lower average energy. In particular, online pseudoinverse update method (OPIUM) is used to train a randomized neural network for quick and resource efficient learning. An additional 42% energy saving can be achieved when a lighter version of OPIUM method is used for training with the same number of data samples lead to no significant compromise on the quality of inference. Instead of a single classifier with large number of neurons, an ensemble of K base learner approach is chosen to reduce learning memory by a factor of K. This also enables approximate computing by dynamically varying the neural network size based on anomaly detection. Fabricated in 65nm CMOS, the ADIC has K = 7 Base Learners (BL) with 32 neurons in each BL and dissipates 11.87pJ/OP and 3.35pJ/OP during learning and inference respectively at Vdd = 0.75V when all 7 BLs are enabled. Further, evaluated on the NASA bearing dataset, approximately 80% of the chip can be shut down for 99% of the lifetime leading to an energy efficiency of 0.48pJ/OP, an 18.5 times reduction over full-precision computing running at Vdd = 1.2V throughout the lifetime. | ['Sumon Kumar Bose', 'Pradeep Kumar Gopalakrishnan', 'Mohendra Roy', 'Bapi Kar', 'Arindam Basu'] | 2020-08-21 | null | null | null | null | ['one-class-classifier'] | ['methodology'] | [ 1.13638110e-01 4.76514809e-02 -3.45487356e-01 -4.30556059e-01
-2.00277314e-01 -3.60615551e-01 1.16263792e-01 4.05718744e-01
-8.51628840e-01 8.38195324e-01 -7.77833939e-01 -4.23886806e-01
1.25864679e-02 -7.55513847e-01 -7.95131743e-01 -7.50529051e-01
9.56909657e-02 8.84445682e-02 4.15324211e-01 1.79283589e-01
3.27231050e-01 7.87655354e-01 -1.67150915e+00 1.27185345e-01
6.98513210e-01 1.54523611e+00 -3.37099671e-01 6.17784142e-01
-8.51380154e-02 6.29396021e-01 -7.64784575e-01 -1.98288620e-01
3.92081201e-01 -1.46157697e-01 -2.07788959e-01 -7.58335769e-01
4.88349617e-01 -6.97567046e-01 -4.80775326e-01 9.85907078e-01
5.34125209e-01 -1.07972018e-01 7.65260994e-01 -1.42506206e+00
1.84135407e-01 6.56197309e-01 -6.03952527e-01 3.51860791e-01
-1.41420990e-01 1.10596873e-01 4.98362094e-01 -5.80233574e-01
6.79873005e-02 7.57167995e-01 6.31292462e-01 5.32530546e-01
-1.04487753e+00 -1.16800559e+00 -2.95746088e-01 2.41470292e-01
-1.77735901e+00 -5.08370638e-01 6.77680194e-01 3.54576617e-01
1.62148821e+00 2.17341900e-01 9.54540372e-01 7.14134812e-01
1.06945384e+00 2.63715953e-01 1.04614878e+00 -4.74114656e-01
8.70711684e-01 1.81390792e-01 4.21008825e-01 7.69944966e-01
1.00182545e+00 -5.58746569e-02 -7.22828269e-01 -4.17393446e-01
4.32430178e-01 3.42447847e-01 2.21315280e-01 1.73792690e-01
-3.21621656e-01 3.33357662e-01 4.88781989e-01 -2.94286609e-02
-2.60008425e-01 8.25619936e-01 7.42937505e-01 2.02877089e-01
-2.60133684e-01 2.64132351e-01 -4.39431250e-01 -2.35162333e-01
-1.05887508e+00 2.11699858e-01 7.69195557e-01 6.72815621e-01
6.65996730e-01 6.75798535e-01 1.52669623e-01 3.59307230e-01
2.70375550e-01 7.40166903e-01 6.41293287e-01 -9.20572698e-01
-2.58061849e-02 1.01651406e+00 -1.83453441e-01 -4.08761889e-01
-6.06352329e-01 -4.74448353e-01 -9.02946889e-01 5.87554812e-01
2.14764595e-01 -1.18705764e-01 -1.11991334e+00 1.44305384e+00
1.33948192e-01 -6.55584596e-03 3.24831307e-02 2.69169718e-01
8.35119262e-02 7.22926855e-01 -3.48231159e-02 -2.85625070e-01
1.18334424e+00 -4.84230429e-01 -7.32700467e-01 -1.65097490e-01
3.75295132e-01 -1.60537809e-01 6.00607753e-01 6.72174275e-01
-8.53615999e-01 -4.48852986e-01 -2.02810860e+00 -1.96759012e-02
-5.36354721e-01 2.22044975e-01 5.20341277e-01 9.73753870e-01
-8.83639514e-01 6.83457315e-01 -1.33181083e+00 -1.79830894e-01
4.94950652e-01 1.05638945e+00 2.46943831e-01 1.58021420e-01
-6.85700178e-01 5.07685661e-01 3.91086370e-01 8.93414244e-02
-7.70175934e-01 -7.32312322e-01 -5.27797043e-01 6.37664422e-02
-9.40925702e-02 -3.40969265e-01 8.75654101e-01 -1.01238894e+00
-1.52748394e+00 3.68418932e-01 -9.86597613e-02 -1.02507842e+00
-1.45823769e-02 -2.86903441e-01 -6.24366522e-01 1.18133642e-01
-4.34552699e-01 4.74960327e-01 8.83669496e-01 -6.38881087e-01
-5.18515110e-01 -5.44056118e-01 -4.18178529e-01 -2.15744644e-01
-8.33140910e-01 -3.51644278e-01 8.62745643e-02 -2.46287540e-01
3.23765039e-01 -1.06628573e+00 4.65005776e-03 3.42901587e-01
-4.09108698e-02 1.24336869e-01 1.24598086e+00 -3.20885599e-01
1.35992289e+00 -1.97278690e+00 -5.63613474e-01 4.82159376e-01
-2.95507349e-02 2.93579757e-01 6.81548059e-01 -3.22749987e-02
3.12678635e-01 -1.47264868e-01 -4.91321802e-01 3.98549475e-02
-2.89402723e-01 3.78255606e-01 -2.17587501e-01 6.51813388e-01
4.74189781e-02 6.28676832e-01 -4.24042642e-01 -6.34455830e-02
-4.84955013e-02 3.34180683e-01 -4.50197816e-01 -2.61152893e-01
2.20992997e-01 -3.74751508e-01 -3.91281873e-01 1.09298003e+00
5.84375560e-01 7.30821639e-02 4.69311565e-01 -4.86671805e-01
-9.31899771e-02 6.92046434e-02 -1.17172003e+00 1.46604824e+00
-4.64119047e-01 7.04538643e-01 1.42962962e-01 -5.88552058e-01
1.12872028e+00 1.08520631e-02 2.22610176e-01 -8.37151349e-01
4.15919155e-01 6.37284636e-01 1.71433464e-01 2.69202620e-01
4.84801292e-01 1.28901526e-01 -3.58438700e-01 5.48479855e-01
1.57479882e-01 4.05325592e-01 -3.00856203e-01 9.19958502e-02
1.63374937e+00 -1.54710218e-01 2.56769568e-01 -7.84428060e-01
5.03655612e-01 4.95651215e-02 8.60580027e-01 7.45921910e-01
-4.56240892e-01 -2.19448641e-01 1.65246740e-01 -5.05832911e-01
-8.54697049e-01 -1.28755510e+00 -5.16873538e-01 6.87725663e-01
1.19201809e-01 -2.09092334e-01 -7.30611980e-01 -3.72542679e-01
1.12688996e-01 9.17682528e-01 -7.13329315e-02 -6.17001057e-01
-6.49029970e-01 -9.70248997e-01 1.15703595e+00 8.50955844e-01
7.89817274e-01 -5.60888469e-01 -1.42618382e+00 8.91372263e-02
8.34667087e-01 -8.27931464e-01 1.78193927e-01 9.87534165e-01
-1.50137019e+00 -5.00297189e-01 2.24196956e-01 -3.41446787e-01
8.75587881e-01 -5.00143826e-01 7.77410507e-01 1.78572908e-01
-5.59425652e-01 1.40347794e-01 1.01662837e-01 -5.95811188e-01
-1.47521466e-01 2.48946827e-02 6.66724980e-01 -3.44331831e-01
5.68628311e-01 -9.17015672e-01 -7.04414546e-01 -7.37624690e-02
-5.65357983e-01 -3.75444561e-01 1.00591469e+00 8.57619464e-01
6.57201469e-01 3.47751617e-01 7.26149797e-01 -7.18973935e-01
-2.58563310e-02 -3.09054762e-01 -7.24397540e-01 -1.37632251e-01
-1.25500524e+00 3.54905427e-01 1.00834370e+00 -6.84056044e-01
-9.29538548e-01 3.71311098e-01 6.13120534e-02 -3.99373323e-01
3.21957976e-01 -2.33138680e-01 1.67491492e-02 -5.65096557e-01
8.66672099e-01 1.27046034e-01 -7.32354745e-02 3.78534198e-02
-2.19959840e-01 7.89864302e-01 4.86369789e-01 -4.77586746e-01
5.41901410e-01 3.91493231e-01 5.48726737e-01 -8.12570870e-01
9.23943967e-02 3.30842882e-01 -4.08701211e-01 -1.91965833e-01
4.50405508e-01 -1.10389912e+00 -1.10445464e+00 4.77651238e-01
-4.56607610e-01 -4.09439653e-01 5.60042150e-02 2.14991942e-01
1.85284600e-01 -2.94884980e-01 -7.55122125e-01 -1.39392388e+00
-1.13751447e+00 -8.64488423e-01 4.91839498e-01 6.74912035e-01
-3.40028554e-01 -4.17036086e-01 -7.16651201e-01 -3.17307860e-01
6.81636691e-01 2.21757710e-01 1.02084064e+00 -6.78727269e-01
-6.22561574e-01 -5.43661237e-01 -4.10460718e-02 3.85660887e-01
-7.98645020e-02 -1.11952655e-01 -1.17876875e+00 -5.85760534e-01
1.10277228e-01 -3.33029389e-01 5.44235170e-01 2.52512842e-01
1.19540966e+00 -1.33673608e-01 -5.94502687e-01 5.93523443e-01
1.97912884e+00 6.32233620e-01 4.65478390e-01 -1.18990041e-01
5.56754887e-01 -3.08028907e-01 5.13280392e-01 4.54967707e-01
-1.95452776e-02 2.66795397e-01 4.99635607e-01 5.07120311e-01
2.13788942e-01 -1.56311065e-01 5.33308029e-01 7.33729541e-01
4.30916727e-01 6.99004233e-02 -9.81925786e-01 2.63798892e-01
-1.24279904e+00 -4.16785657e-01 5.07942140e-02 2.76420617e+00
8.67958844e-01 7.64555871e-01 -2.11995408e-01 3.63521725e-01
2.75131464e-01 -2.49924853e-01 -1.34410596e+00 -9.01670337e-01
3.05179298e-01 8.05847049e-01 1.20822823e+00 2.46056810e-01
-4.56652701e-01 3.26595336e-01 4.77850485e+00 9.75374341e-01
-1.52510512e+00 -9.82522443e-02 7.35106051e-01 -6.78174555e-01
1.80272877e-01 7.86138028e-02 -1.35811460e+00 7.19042361e-01
1.73806262e+00 5.72604761e-02 2.76385903e-01 1.07301271e+00
-3.97313923e-01 -7.57511973e-01 -1.18621540e+00 9.64731216e-01
-4.82596532e-02 -1.10685205e+00 -1.00479893e-01 5.93701899e-02
4.86268282e-01 -1.51023060e-01 -2.17356130e-01 5.97611248e-01
-4.09837663e-01 -8.79512787e-01 5.63385665e-01 3.70964617e-01
1.15990472e+00 -1.43464398e+00 7.43702710e-01 3.27933252e-01
-1.13290203e+00 -6.52820766e-01 -3.48080665e-01 -2.57924139e-01
-3.15538436e-01 6.84157491e-01 -4.91563290e-01 1.49974301e-01
1.14342678e+00 -3.75360176e-02 -5.79168797e-01 4.57246691e-01
2.94398397e-01 7.62043595e-01 -9.10673618e-01 -6.48747921e-01
-1.55820549e-01 -2.64567044e-02 2.55124331e-01 9.25911725e-01
3.12388361e-01 1.52237758e-01 -2.42062464e-01 7.09923923e-01
-2.53892183e-01 -4.80564505e-01 -2.99724877e-01 3.11715186e-01
1.27201021e+00 1.46704543e+00 -8.71647358e-01 -1.85927913e-01
7.13573471e-02 1.15686977e+00 1.22152440e-01 -9.39093456e-02
-7.59312868e-01 -9.72369254e-01 6.55668437e-01 2.87812471e-01
-1.84050333e-02 -2.00923637e-01 -7.02914298e-01 -1.70587346e-01
1.36036202e-01 -3.75192493e-01 1.52578026e-01 -2.67013311e-01
-6.40451729e-01 2.74939775e-01 -1.33821433e-02 -7.21035004e-01
-1.89934611e-01 -7.72895455e-01 -8.06958973e-01 6.19305730e-01
-1.10488415e+00 -6.12920880e-01 -4.81808603e-01 6.91088513e-02
2.08467245e-01 -9.14335847e-02 8.49601865e-01 2.23856524e-01
-7.60348678e-01 1.24796605e+00 2.13775188e-01 -1.06585354e-01
4.23512965e-01 -1.09177101e+00 2.17867672e-01 7.35226929e-01
-5.51868379e-01 7.16511667e-01 5.34441471e-01 -7.32365787e-01
-2.22055578e+00 -1.05001426e+00 2.24543065e-01 -1.86980551e-03
2.65782058e-01 -6.79485321e-01 -1.08757305e+00 3.17121178e-01
2.42637359e-02 2.65409082e-01 5.62558711e-01 -3.35567862e-01
4.98895384e-02 -9.15858448e-01 -1.79381645e+00 5.91378987e-01
7.59029508e-01 -4.00907129e-01 1.85277149e-01 -1.67250782e-01
3.84291261e-01 -4.98546243e-01 -9.19820487e-01 7.07442522e-01
9.23855603e-01 -7.30575979e-01 6.09319746e-01 3.32927167e-01
-2.22814366e-01 -3.84368867e-01 -2.81072378e-01 -3.05080354e-01
2.20832571e-01 -5.31650603e-01 -9.27516937e-01 1.15244341e+00
3.67001683e-01 -8.68291080e-01 1.13348711e+00 6.26245081e-01
-1.91659890e-02 -1.42993248e+00 -1.19433415e+00 -7.25799739e-01
-1.40240073e-01 -2.54686803e-01 7.07090124e-02 1.76144183e-01
-3.21146578e-01 3.18071663e-01 1.16765358e-01 2.91649669e-01
7.74534523e-01 -4.06044394e-01 1.76431000e-01 -1.11532128e+00
-2.82289892e-01 -1.36533320e-01 -6.47781491e-01 -5.24496853e-01
3.57260220e-02 -5.95900357e-01 -2.61195093e-01 -6.91680968e-01
-6.66704029e-02 -6.27586186e-01 -5.29342294e-01 6.93588972e-01
2.75558561e-01 2.66857624e-01 -4.15949374e-01 -2.53250450e-01
-1.16865888e-01 2.50940561e-01 7.34071508e-02 1.38105005e-01
-3.14276576e-01 -3.13455999e-01 -1.56657264e-01 5.70702732e-01
9.41258669e-01 -6.36597157e-01 -6.16575778e-01 -1.79951936e-02
2.29601383e-01 -7.77531639e-02 8.68472382e-02 -1.64860713e+00
6.64754450e-01 3.70348334e-01 1.17285764e+00 -7.08186746e-01
6.00290358e-01 -9.72153306e-01 1.24771409e-01 1.04614639e+00
2.98093945e-01 3.35485309e-01 9.20357347e-01 6.90548062e-01
3.92890871e-01 -4.83705163e-01 9.00758624e-01 2.83930868e-01
-6.07131124e-01 -1.26056850e-01 -5.93738198e-01 -4.40764785e-01
1.01943994e+00 -4.35960978e-01 -2.28988543e-01 1.80902675e-01
-9.04377922e-02 7.53462464e-02 5.73113263e-01 -6.80291131e-02
6.28879964e-01 -1.33396721e+00 2.16716472e-02 4.49944139e-01
-2.11371720e-01 -9.80890766e-02 3.26479167e-01 5.41357517e-01
-6.95910096e-01 2.05915421e-01 -4.07570750e-01 -5.71417570e-01
-1.12980115e+00 6.76547224e-03 5.14629722e-01 -2.42059976e-02
-5.20285726e-01 7.67948568e-01 -8.89527380e-01 1.70290947e-01
2.39837542e-01 -2.85061806e-01 3.75128895e-01 -3.13952833e-01
5.39738595e-01 7.82794833e-01 3.71519089e-01 2.15333283e-01
-7.09027529e-01 2.75054067e-01 -4.28111732e-01 -1.08674981e-01
8.93786967e-01 2.66967714e-01 -1.76157072e-01 7.50837624e-01
1.18220901e+00 -1.67665347e-01 -1.19913542e+00 -4.89221625e-02
9.48925763e-02 1.54813007e-01 6.23963177e-01 -9.08640921e-01
-7.94349790e-01 4.61116195e-01 1.36078072e+00 -3.74808401e-01
1.71308017e+00 -6.66078568e-01 8.64316225e-01 8.96527350e-01
5.50222814e-01 -1.54013407e+00 -3.52946781e-02 2.96178401e-01
1.36761397e-01 -6.95077717e-01 6.19396508e-01 3.05489987e-01
-2.91077401e-02 1.28869975e+00 9.87405002e-01 -4.56020266e-01
6.59831583e-01 1.00670469e+00 -4.29457307e-01 -2.12874692e-02
-7.96962380e-01 6.83763087e-01 -7.81606287e-02 1.88812852e-01
4.47251312e-02 -1.00502230e-01 -1.91032529e-01 7.07645893e-01
1.11727387e-01 -1.61501557e-01 4.09518600e-01 1.49650013e+00
-6.52439833e-01 -7.50217736e-01 -5.94475903e-02 1.12902892e+00
-5.32094240e-01 1.77407339e-01 3.31704132e-02 6.45322740e-01
1.34971604e-01 5.81622303e-01 5.74093342e-01 -5.23162961e-01
-3.66639122e-02 4.12920862e-01 5.38354814e-01 -1.16437562e-01
-6.30813658e-01 -4.98911917e-01 -2.01925024e-01 -8.66485596e-01
2.83370525e-01 -3.87167841e-01 -1.92273915e+00 -3.28238040e-01
-2.84912765e-01 -2.12608889e-01 1.18437898e+00 8.00076962e-01
3.73186916e-01 7.55326748e-01 4.65078354e-01 -6.44321203e-01
-5.66692948e-01 -6.56571567e-01 -5.86515725e-01 -7.54921556e-01
1.10331699e-01 -6.15561545e-01 -4.41696554e-01 -5.71676254e-01] | [8.294916152954102, 2.5768611431121826] |
6dc464b9-38a1-442e-8f48-64f455d65ec9 | query2box-reasoning-over-knowledge-graphs-in-1 | 2002.05969 | null | https://arxiv.org/abs/2002.05969v2 | https://arxiv.org/pdf/2002.05969v2.pdf | Query2box: Reasoning over Knowledge Graphs in Vector Space using Box Embeddings | Answering complex logical queries on large-scale incomplete knowledge graphs (KGs) is a fundamental yet challenging task. Recently, a promising approach to this problem has been to embed KG entities as well as the query into a vector space such that entities that answer the query are embedded close to the query. However, prior work models queries as single points in the vector space, which is problematic because a complex query represents a potentially large set of its answer entities, but it is unclear how such a set can be represented as a single point. Furthermore, prior work can only handle queries that use conjunctions ($\wedge$) and existential quantifiers ($\exists$). Handling queries with logical disjunctions ($\vee$) remains an open problem. Here we propose query2box, an embedding-based framework for reasoning over arbitrary queries with $\wedge$, $\vee$, and $\exists$ operators in massive and incomplete KGs. Our main insight is that queries can be embedded as boxes (i.e., hyper-rectangles), where a set of points inside the box corresponds to a set of answer entities of the query. We show that conjunctions can be naturally represented as intersections of boxes and also prove a negative result that handling disjunctions would require embedding with dimension proportional to the number of KG entities. However, we show that by transforming queries into a Disjunctive Normal Form, query2box is capable of handling arbitrary logical queries with $\wedge$, $\vee$, $\exists$ in a scalable manner. We demonstrate the effectiveness of query2box on three large KGs and show that query2box achieves up to 25% relative improvement over the state of the art. | ['Jure Leskovec', 'Weihua Hu', 'Hongyu Ren'] | 2020-02-14 | null | https://openreview.net/forum?id=BJgr4kSFDS | https://openreview.net/pdf?id=BJgr4kSFDS | iclr-2020-1 | ['complex-query-answering'] | ['knowledge-base'] | [-1.33245662e-01 5.01561165e-01 -2.23889142e-01 -2.33123481e-01
-5.53918064e-01 -8.74810338e-01 -1.00200661e-01 5.66128910e-01
-1.80908963e-01 6.06620491e-01 -1.35006681e-01 -5.93684316e-01
-3.79727125e-01 -1.69018590e+00 -1.19136989e+00 -7.58801401e-02
-5.49800634e-01 8.88801873e-01 6.38813853e-01 -5.08125007e-01
-3.30666333e-01 3.31320465e-01 -1.65773284e+00 2.85486192e-01
6.36740863e-01 1.01640952e+00 -4.08626825e-01 3.78442526e-01
-3.45309734e-01 4.69206452e-01 -6.91906393e-01 -7.16788590e-01
3.89962614e-01 1.86408892e-01 -1.05487502e+00 -5.07503927e-01
4.84740615e-01 -4.06552225e-01 -2.59672731e-01 1.14888918e+00
3.47458124e-02 -9.63775143e-02 2.90244818e-01 -1.81026638e+00
-4.61994380e-01 8.10615778e-01 -2.28525758e-01 -2.44829610e-01
6.28736675e-01 -2.89410353e-01 1.66078150e+00 -7.68918395e-01
7.76506305e-01 1.21945739e+00 4.45647240e-01 1.55105457e-01
-1.46015930e+00 -5.66053510e-01 -4.54651825e-02 2.55940467e-01
-1.87970424e+00 -1.47762835e-01 3.59762102e-01 -2.26158589e-01
1.34750521e+00 6.96337223e-01 4.90783811e-01 -1.42100435e-02
-2.08616197e-01 6.41718864e-01 4.79818493e-01 -4.09280628e-01
4.10919219e-01 2.60625273e-01 5.91018319e-01 9.65952694e-01
7.95198202e-01 -2.44296029e-01 -4.10182089e-01 -6.01428092e-01
4.19991195e-01 -2.02885941e-01 -2.17813805e-01 -7.68280983e-01
-1.05453408e+00 1.19397581e+00 5.00875175e-01 4.19326387e-02
6.41470775e-02 4.86464024e-01 3.04969519e-01 5.04028320e-01
-2.19476327e-01 5.02699137e-01 -7.22105622e-01 3.12275231e-01
-5.49819291e-01 6.44919157e-01 1.28331876e+00 1.33435893e+00
1.17482245e+00 -2.83342540e-01 7.50622600e-02 3.56415927e-01
1.60113975e-01 6.54327691e-01 -4.79904056e-01 -9.41545427e-01
7.06888974e-01 1.29948676e+00 3.07546198e-01 -1.05327332e+00
-3.47279549e-01 -1.02239043e-01 -4.21953142e-01 1.38155758e-01
3.87440115e-01 3.47523428e-02 -3.03919852e-01 1.98514199e+00
6.40882432e-01 -4.21222925e-01 3.34702730e-01 5.83726764e-01
7.47230232e-01 5.82955360e-01 -1.52834699e-01 -4.74485457e-02
1.81925428e+00 -4.92110044e-01 -4.78335977e-01 -1.24486677e-01
1.21623540e+00 -1.43680155e-01 1.04021680e+00 2.70311326e-01
-1.22364962e+00 2.54955348e-02 -1.22068036e+00 -3.53657395e-01
-8.39768529e-01 -1.96936131e-01 9.31893051e-01 6.33290827e-01
-1.06839931e+00 -1.92839839e-02 -6.19305968e-01 -9.26314816e-02
1.86196908e-01 7.81418324e-01 -7.55039811e-01 -4.25543278e-01
-1.58474529e+00 6.35590732e-01 6.48905575e-01 -1.84454754e-01
-1.77834228e-01 -8.41401815e-01 -1.32763827e+00 3.63744110e-01
1.23099685e+00 -8.92997563e-01 9.28328276e-01 9.62096602e-02
-4.78457183e-01 6.07805550e-01 -4.39762115e-01 -5.49802721e-01
8.31847917e-03 1.41768292e-01 -4.87965614e-01 1.85962364e-01
1.46143794e-01 3.82034779e-01 9.92558524e-02 -1.09285188e+00
-6.08345032e-01 -8.49978805e-01 1.11708486e+00 2.60626879e-02
-2.67464072e-01 3.77322286e-02 -5.95058560e-01 -7.86126032e-02
4.68838573e-01 -7.30360329e-01 -9.20085236e-03 7.72984996e-02
-5.98027050e-01 -5.45883894e-01 7.22486913e-01 -1.31102383e-01
1.60444307e+00 -2.05039215e+00 7.57891461e-02 5.64676404e-01
6.49197996e-01 2.32835989e-02 3.23466092e-01 7.66560078e-01
9.40605178e-02 6.21869385e-01 -2.42202312e-01 1.48707122e-01
7.22913861e-01 5.97167015e-01 -5.12875140e-01 1.24290250e-01
1.04372472e-01 1.02043784e+00 -7.42654145e-01 -5.51413536e-01
-8.08379203e-02 -8.09619874e-02 -9.97513831e-01 -1.11712404e-01
-8.61286879e-01 -8.62078369e-01 -4.69783783e-01 8.87785375e-01
8.85580897e-01 -4.03170228e-01 3.74716043e-01 -5.13981283e-01
2.02059582e-01 1.04448549e-01 -1.85922492e+00 1.27554178e+00
-1.67998582e-01 6.84836954e-02 3.26866835e-01 -8.39670002e-01
6.28761053e-01 9.08912644e-02 3.11291158e-01 -4.11920309e-01
-3.21699917e-01 2.65237987e-01 -3.83147210e-01 -3.09379786e-01
7.98874795e-01 -3.32407594e-01 -6.11666679e-01 4.35499936e-01
-2.37776428e-01 -3.40489298e-01 4.90570098e-01 7.99922705e-01
1.53337443e+00 -4.83056784e-01 1.48410439e-01 -2.43836362e-02
4.36503947e-01 2.91855693e-01 6.35884047e-01 8.83975983e-01
1.27401620e-01 -1.80723127e-02 1.11862099e+00 -3.64276320e-01
-6.92352831e-01 -1.19358790e+00 -1.14775203e-01 9.06840444e-01
4.15327907e-01 -1.19358242e+00 -2.53521979e-01 -6.54643357e-01
5.57780325e-01 6.88136756e-01 -3.38327229e-01 -1.60424665e-01
-4.10369873e-01 -4.25310522e-01 8.18400502e-01 8.59884381e-01
3.25798690e-01 -6.10838294e-01 -5.87440193e-01 2.04252005e-01
-1.70240462e-01 -1.23696268e+00 -6.25294372e-02 3.45944375e-01
-6.68566406e-01 -1.39986002e+00 2.43553147e-01 -5.61035097e-01
7.62009978e-01 7.39279091e-02 1.33218455e+00 2.08872423e-01
-2.54993975e-01 4.85141754e-01 -1.73495919e-01 -3.67800444e-01
7.92911202e-02 -2.46320605e-01 -1.36679485e-01 -4.98982459e-01
5.99388838e-01 -2.02672467e-01 -4.17230368e-01 4.26538825e-01
-1.46062624e+00 -3.74502331e-01 9.49652791e-02 7.65316963e-01
8.86647701e-01 6.98881388e-01 2.26717293e-01 -1.25490308e+00
3.76740545e-01 -5.77125013e-01 -1.07790709e+00 5.97424567e-01
-5.57595253e-01 3.94235194e-01 6.36970222e-01 2.01738277e-03
-3.66125226e-01 -1.28047898e-01 1.61453471e-01 -3.59369189e-01
1.50515720e-01 1.00187171e+00 -4.91263509e-01 1.23802178e-01
5.56865573e-01 -1.86625108e-01 -4.14036751e-01 -4.12231982e-02
5.99162936e-01 2.41757676e-01 4.66640651e-01 -1.07551157e+00
1.01829314e+00 6.89784348e-01 3.96351665e-01 -4.02212590e-01
-4.16603744e-01 -4.03501093e-01 -2.29881763e-01 4.24954802e-01
3.24940830e-01 -8.62890840e-01 -1.29267192e+00 -2.18509734e-01
-1.05178666e+00 -7.23403320e-02 -5.68557978e-01 7.69694224e-02
-3.56312901e-01 3.37431490e-01 -2.88198292e-01 -7.54939616e-01
-1.72859177e-01 -1.09924746e+00 1.11956966e+00 -1.45877898e-01
-3.06074411e-01 -7.08748877e-01 -1.84440047e-01 3.93293768e-01
2.45143011e-01 2.87770838e-01 1.61328614e+00 -7.86559880e-01
-9.54759479e-01 -5.23653865e-01 -3.68207723e-01 6.35787770e-02
-2.85901546e-01 -1.10021263e-01 -3.79513621e-01 -2.44821995e-01
-3.90837014e-01 -4.47584897e-01 4.02962089e-01 -2.90334284e-01
8.48477721e-01 -4.84356254e-01 -5.18521309e-01 3.76204759e-01
1.77819836e+00 -1.45689324e-01 7.24308074e-01 1.40426576e-01
2.94016838e-01 3.99575591e-01 5.10482073e-01 5.22998095e-01
7.97448754e-01 6.21678770e-01 6.16719782e-01 2.72645205e-01
4.23858374e-01 -3.55960727e-01 -7.24937394e-02 -2.60219094e-03
4.26714659e-01 -2.12792590e-01 -1.02040195e+00 6.74727499e-01
-1.65776980e+00 -8.47260296e-01 -9.67068225e-02 2.44627476e+00
7.77160823e-01 1.99413434e-01 -6.87417313e-02 4.23865199e-01
3.55571419e-01 -3.14142779e-02 -5.09708583e-01 -5.05529404e-01
-3.26963753e-01 5.61290622e-01 5.45467079e-01 5.54346681e-01
-8.37494135e-01 9.89942670e-01 5.97788286e+00 4.90332484e-01
-7.14428484e-01 -1.66593954e-01 -1.41331017e-01 -1.40198424e-01
-8.02413702e-01 4.36767280e-01 -8.46984923e-01 1.09083094e-01
6.05314374e-01 -4.19492722e-01 4.87481147e-01 8.34869206e-01
-5.21572113e-01 -3.37365508e-01 -1.55305445e+00 8.14871371e-01
-2.20994726e-01 -1.22943103e+00 -4.27323431e-02 2.07790986e-01
5.08970976e-01 -2.97521502e-01 -3.71944830e-02 7.43499398e-01
5.36276937e-01 -9.97752786e-01 5.58979034e-01 -4.20033140e-03
9.38642204e-01 -7.26783812e-01 6.92430019e-01 2.92431831e-01
-1.57419658e+00 -2.40699530e-01 -4.40625191e-01 1.13703050e-01
-8.33161473e-02 4.97076362e-01 -8.01202297e-01 9.99995410e-01
6.63797379e-01 -8.27045217e-02 -2.81386971e-01 6.98849082e-01
-3.10340971e-02 1.47345975e-01 -9.22117412e-01 1.17358059e-01
-7.32280090e-02 2.68496815e-02 3.36641431e-01 9.42019045e-01
2.24176705e-01 4.64465529e-01 6.95217699e-02 1.08765221e+00
-3.95755231e-01 9.13089141e-03 -7.43598223e-01 -1.82431877e-01
7.06177294e-01 8.10287058e-01 -3.27597469e-01 -5.74424565e-01
-6.48605168e-01 3.88448656e-01 5.01932442e-01 5.02021432e-01
-8.35944593e-01 -6.98347390e-01 8.50034118e-01 2.92357177e-01
5.37786007e-01 -9.08378586e-02 -1.48586616e-01 -1.24599814e+00
6.63930357e-01 -8.60865116e-01 1.02299058e+00 -8.38481784e-01
-9.13375318e-01 1.83396861e-01 3.36747468e-01 -6.81168020e-01
-7.10038096e-02 -6.60587788e-01 -2.36653924e-01 8.65814447e-01
-1.28277791e+00 -8.70126188e-01 -1.73115104e-01 9.72825706e-01
-3.75752985e-01 5.31959951e-01 1.11646008e+00 2.07188964e-01
-3.31525683e-01 7.35459864e-01 -1.98166460e-01 1.10488839e-01
2.67733485e-01 -1.49970174e+00 -8.57541114e-02 5.29738069e-01
1.36626378e-01 1.09753263e+00 6.37008250e-01 -5.37170768e-01
-2.18094563e+00 -8.97517323e-01 1.10140860e+00 -3.77120942e-01
7.02841818e-01 -4.82375264e-01 -8.85666311e-01 1.10113800e+00
-3.17193776e-01 5.64665020e-01 7.44862258e-01 4.59936380e-01
-1.02613580e+00 -4.40545976e-01 -1.38383043e+00 5.99798560e-01
1.04640365e+00 -7.81476080e-01 -7.33789206e-01 3.29891294e-01
9.17543769e-01 -4.62245435e-01 -1.22620559e+00 7.02415407e-01
4.27216023e-01 -8.47513974e-01 1.08998692e+00 -8.75857532e-01
-1.16286367e-01 -7.35987782e-01 -6.53543890e-01 -7.84176648e-01
1.47995129e-01 -3.41567874e-01 -6.51796520e-01 7.45317400e-01
5.28602242e-01 -9.38008845e-01 8.73652220e-01 1.20603836e+00
1.55170470e-01 -9.92983699e-01 -1.04029274e+00 -7.59727299e-01
4.38844673e-02 -7.16441691e-01 1.13360584e+00 6.84632480e-01
5.70026278e-01 1.44750282e-01 2.51581460e-01 7.15265155e-01
5.12818158e-01 5.99055231e-01 8.79652858e-01 -1.27457619e+00
-3.97992998e-01 -2.01639891e-01 -7.38507569e-01 -9.15982783e-01
2.43061576e-02 -1.01410985e+00 -4.99698281e-01 -1.81379294e+00
7.54132271e-02 -7.25091100e-01 -2.95472853e-02 9.26500201e-01
1.94470808e-01 -6.81599379e-02 1.77665427e-01 -1.03249460e-01
-7.93023407e-01 2.00094387e-01 8.78441632e-01 -4.97116566e-01
3.00378613e-02 -3.53071660e-01 -8.55119646e-01 5.14749110e-01
3.56461704e-01 -3.18853915e-01 -5.22229552e-01 -4.05222952e-01
1.17255270e+00 4.70427692e-01 5.37851572e-01 -7.30463743e-01
5.90346217e-01 -4.05487120e-02 -2.18036786e-01 -7.57159889e-01
6.72386646e-01 -9.97353017e-01 5.13078749e-01 2.63272345e-01
-6.06970005e-02 1.40189111e-01 2.79038578e-01 5.95173955e-01
-5.03449678e-01 -4.47679386e-02 1.08920835e-01 -5.87579645e-02
-6.44566536e-01 1.95152998e-01 3.22092324e-01 3.08014095e-01
1.21124530e+00 -1.67348802e-01 -8.02370787e-01 -1.68175071e-01
-8.22257519e-01 7.14229882e-01 4.96486634e-01 -8.91972482e-02
7.90967047e-01 -1.17903006e+00 -3.02833050e-01 1.54525280e-01
5.17152846e-01 6.62910759e-01 2.14602157e-01 5.64523220e-01
-4.96792018e-01 6.32713854e-01 3.72746050e-01 -2.75011361e-01
-1.24494541e+00 9.58614647e-01 2.31101230e-01 -4.17015344e-01
-1.52082369e-01 9.83732164e-01 2.59764552e-01 -5.30954599e-01
4.08901304e-01 -7.45034695e-01 3.05515379e-01 -9.59935412e-02
3.16048652e-01 3.62927407e-01 1.53973922e-01 -3.90951097e-01
-6.63513005e-01 4.69825685e-01 -5.99846989e-02 1.26367975e-02
1.00807166e+00 3.02975476e-01 -6.49440527e-01 2.90820580e-02
1.17917264e+00 2.92325526e-01 -1.65841699e-01 -4.69211727e-01
-6.69348538e-02 -4.75570560e-01 -4.69559461e-01 -6.68611586e-01
-8.85961533e-01 4.78581190e-01 -1.35792851e-01 4.24089700e-01
1.05125177e+00 6.14959478e-01 6.93995893e-01 9.09258783e-01
9.90994275e-01 -8.35820377e-01 -3.66362154e-01 3.20092678e-01
6.83853030e-01 -9.64012861e-01 5.14077954e-02 -9.36216712e-01
-2.31264502e-01 9.41399515e-01 6.53731048e-01 2.59254724e-01
5.71886361e-01 4.26610678e-01 -4.38992679e-01 -6.05726182e-01
-9.49574053e-01 -2.48043507e-01 1.14606516e-02 1.28618106e-01
-2.05635786e-01 2.87237793e-01 -1.78498864e-01 9.82199192e-01
-3.05623323e-01 6.07689060e-02 4.36785161e-01 1.27832770e+00
-3.35146606e-01 -1.28875995e+00 -5.34762919e-01 4.63700205e-01
-2.64844388e-01 -8.27710405e-02 -2.21895024e-01 1.09520495e+00
3.42336416e-01 9.94138002e-01 -1.55282274e-01 -4.61217910e-01
6.36686862e-01 2.45242134e-01 4.65096593e-01 -7.61021972e-01
-2.97977358e-01 -4.62647706e-01 2.26093084e-01 -5.48362315e-01
1.41358063e-01 -2.17466012e-01 -1.90811288e+00 -5.41003525e-01
-5.84816277e-01 4.46216404e-01 4.00372416e-01 5.62453210e-01
4.59964901e-01 1.15348287e-01 1.72996998e-01 4.82628107e-01
-8.10415149e-01 -3.00126970e-01 -8.21304083e-01 3.58285367e-01
1.62877604e-01 -6.12814724e-01 -3.30944687e-01 -2.59430677e-01] | [9.031747817993164, 7.578350067138672] |
21f624f9-8cea-4a8a-be75-f1e91aca3dbe | stock-price-prediction-using-cnn-and-lstm | 2010.13891 | null | https://arxiv.org/abs/2010.13891v1 | https://arxiv.org/pdf/2010.13891v1.pdf | Stock Price Prediction Using CNN and LSTM-Based Deep Learning Models | Designing robust and accurate predictive models for stock price prediction has been an active area of research for a long time. While on one side, the supporters of the efficient market hypothesis claim that it is impossible to forecast stock prices accurately, many researchers believe otherwise. There exist propositions in the literature that have demonstrated that if properly designed and optimized, predictive models can very accurately and reliably predict future values of stock prices. This paper presents a suite of deep learning based models for stock price prediction. We use the historical records of the NIFTY 50 index listed in the National Stock Exchange of India, during the period from December 29, 2008 to July 31, 2020, for training and testing the models. Our proposition includes two regression models built on convolutional neural networks and three long and short term memory network based predictive models. To forecast the open values of the NIFTY 50 index records, we adopted a multi step prediction technique with walk forward validation. In this approach, the open values of the NIFTY 50 index are predicted on a time horizon of one week, and once a week is over, the actual index values are included in the training set before the model is trained again, and the forecasts for the next week are made. We present detailed results on the forecasting accuracies for all our proposed models. The results show that while all the models are very accurate in forecasting the NIFTY 50 open values, the univariate encoder decoder convolutional LSTM with the previous two weeks data as the input is the most accurate model. On the other hand, a univariate CNN model with previous one week data as the input is found to be the fastest model in terms of its execution speed. | ['Jaydip Sen', 'Sidra Mehtab'] | 2020-10-22 | null | null | null | null | ['stock-price-prediction'] | ['time-series'] | [-4.71927196e-01 -1.78800151e-01 -2.34702811e-01 -3.71952713e-01
-3.39037240e-01 -5.44756949e-01 6.53764129e-01 -1.52881863e-02
-2.84995437e-01 8.59168172e-01 1.53862119e-01 -7.52714932e-01
-1.32537514e-01 -1.17754745e+00 -6.87412739e-01 -6.06381595e-01
-3.94055367e-01 4.77214545e-01 3.48388143e-02 -4.39108461e-01
7.46193588e-01 4.16599035e-01 -1.41361451e+00 -1.78703275e-02
2.86351353e-01 1.82050753e+00 -3.84668559e-02 4.42188203e-01
-1.19322166e-01 1.21973574e+00 -4.25363898e-01 -4.34914470e-01
8.64421964e-01 -6.73949942e-02 -4.17523205e-01 -5.30579627e-01
-1.30700186e-01 -7.92533755e-01 -2.36858338e-01 7.57802308e-01
2.19269603e-01 -1.57850608e-01 1.84915975e-01 -1.14991820e+00
-7.25055754e-01 8.54641438e-01 -2.45110109e-01 5.12315333e-01
-3.14489037e-01 1.47953495e-01 1.19571435e+00 -8.00283670e-01
1.93285167e-01 6.37337208e-01 8.70615005e-01 -8.36175755e-02
-7.94920385e-01 -1.05277944e+00 2.65488457e-02 5.97050339e-02
-1.01535213e+00 -2.84686863e-01 6.65484309e-01 -6.78191006e-01
1.28309536e+00 8.22610185e-02 9.67448056e-01 4.58024412e-01
7.78161228e-01 3.07227015e-01 1.05649829e+00 -1.30512670e-01
2.23292962e-01 1.96519732e-01 -2.88736094e-02 1.07877545e-01
3.62983048e-01 6.65527344e-01 -3.31049800e-01 -1.16019309e-01
5.91917872e-01 1.91111922e-01 2.51084268e-02 4.15294886e-01
-1.13905752e+00 9.93510604e-01 4.13122833e-01 4.34379578e-01
-8.48798573e-01 1.81082070e-01 4.09719169e-01 7.45884657e-01
8.14763010e-01 5.68269908e-01 -1.00585568e+00 -2.85342723e-01
-1.52424371e+00 4.65012610e-01 1.04370975e+00 2.78516561e-01
4.34783846e-01 6.65147066e-01 1.22277513e-01 2.23602176e-01
2.21110672e-01 4.80406880e-01 8.63685250e-01 -4.59918469e-01
4.07669306e-01 4.40527499e-01 4.39313173e-01 -9.93677735e-01
-4.16163057e-01 -8.47755432e-01 -6.39997721e-01 4.08152759e-01
2.49181673e-01 -4.03797239e-01 -7.36922920e-01 1.25984526e+00
-3.87304097e-01 4.85093266e-01 3.76440525e-01 6.05005443e-01
2.73834288e-01 1.15980601e+00 -1.23300344e-01 -3.91096175e-01
8.02035034e-01 -7.44583845e-01 -4.11434412e-01 -1.38799310e-01
4.59684402e-01 -6.08672917e-01 2.63157070e-01 2.28601575e-01
-9.77126241e-01 -5.19116759e-01 -1.12179589e+00 4.65821147e-01
-6.91982567e-01 -1.31942466e-01 5.68863511e-01 1.74375325e-01
-1.01165676e+00 1.08941507e+00 -6.11640811e-01 4.14388359e-01
5.66850901e-02 3.44915688e-01 1.33716479e-01 6.28334224e-01
-1.62078118e+00 1.26547658e+00 6.65307820e-01 3.93949956e-01
-5.10492325e-01 -7.68829823e-01 -3.81112337e-01 4.44532275e-01
-2.83280343e-01 -7.90521652e-02 1.46577787e+00 -1.19083977e+00
-1.31099319e+00 2.06052557e-01 3.61042410e-01 -1.22913265e+00
4.75126058e-01 -1.51181442e-03 -7.16809630e-01 -4.85620022e-01
-1.12189353e-01 2.94684410e-01 6.03974700e-01 -5.93467236e-01
-8.59968185e-01 -6.78546876e-02 -8.73434395e-02 -1.50895625e-01
-2.33371519e-02 6.89920550e-03 3.29582661e-01 -7.75831282e-01
1.11625947e-01 -8.02222729e-01 -9.23492238e-02 -5.53921461e-01
-3.42274271e-02 -8.62055868e-02 6.18676245e-01 -1.10391116e+00
1.36525619e+00 -1.86125898e+00 -5.29039800e-01 4.58409458e-01
-2.35028595e-01 1.57219455e-01 2.65226305e-01 5.71774602e-01
-4.10980642e-01 1.46150798e-01 -1.79197863e-01 6.76958263e-02
1.05887644e-01 1.47900954e-01 -1.13740480e+00 2.26509824e-01
1.07085675e-01 1.06999242e+00 -3.67788255e-01 2.66712964e-01
9.52221602e-02 2.50358999e-01 -3.10742073e-02 1.60118043e-01
-4.65458006e-01 1.47688985e-01 -2.39207074e-01 4.90579247e-01
6.09222293e-01 -3.76156420e-01 -1.59831911e-01 1.66821867e-01
-8.03279638e-01 5.23515284e-01 -8.88079166e-01 6.39712572e-01
-2.45456517e-01 6.84924603e-01 -5.79482198e-01 -9.33045924e-01
1.26302433e+00 5.75728655e-01 5.52873909e-01 -1.09891641e+00
5.25598526e-02 8.64314437e-01 3.02166134e-01 -1.13492884e-01
5.58389664e-01 -6.06634736e-01 1.32896945e-01 5.99618495e-01
-4.54622120e-01 3.90360951e-02 1.49076939e-01 -5.06792843e-01
6.19928956e-01 1.84042275e-01 2.22402066e-01 -8.86156633e-02
3.82987618e-01 1.26815423e-01 7.02807963e-01 2.53918648e-01
6.10866658e-02 3.35441172e-01 5.74964523e-01 -1.19172001e+00
-1.30132461e+00 -4.92918134e-01 -2.34860986e-01 8.04466724e-01
-4.91942793e-01 2.30641901e-01 -1.21170983e-01 -8.31815526e-02
3.52022886e-01 1.12491632e+00 -6.52089655e-01 2.21530870e-01
-4.48380619e-01 -9.19239283e-01 3.22047174e-01 5.53158700e-01
7.15309024e-01 -1.35027909e+00 -1.04184747e+00 4.86494273e-01
3.25817555e-01 -5.84030688e-01 1.30092695e-01 4.27580833e-01
-1.02423561e+00 -8.41090918e-01 -9.74304318e-01 -4.94128168e-01
8.64975676e-02 -3.48047763e-01 1.09100604e+00 4.08314727e-02
4.60539311e-01 -3.47865641e-01 -1.00761630e-01 -9.17990148e-01
-2.39051774e-01 5.62477559e-02 -1.30572498e-01 -1.16812680e-02
4.21846867e-01 -4.75374430e-01 -5.61956823e-01 -1.01391226e-01
-6.37904406e-01 1.03123330e-01 6.10454977e-01 5.54635763e-01
5.24873555e-01 1.71340197e-01 9.71315861e-01 -5.53785563e-01
5.29742599e-01 -7.96208858e-01 -1.22508299e+00 8.52228776e-02
-1.31328011e+00 2.67103612e-02 7.08371699e-01 1.00942976e-04
-8.22080851e-01 -2.32488424e-01 -3.66533905e-01 -3.29639673e-01
4.78857011e-01 1.13447845e+00 7.81808853e-01 1.20201275e-01
3.70674506e-02 6.64530575e-01 3.41155343e-02 -5.27245462e-01
-2.95472324e-01 4.93801981e-01 4.47165936e-01 8.73281956e-02
8.43575418e-01 1.34974584e-01 -6.93439022e-02 -3.82818043e-01
-8.57494116e-01 -9.34763327e-02 -5.23943365e-01 -2.23114982e-01
5.11970997e-01 -1.10611403e+00 -4.62213218e-01 7.46737659e-01
-7.13253975e-01 -2.75504500e-01 -2.86011457e-01 5.81099987e-01
-3.49162817e-01 -3.17043930e-01 -6.86988711e-01 -1.03502178e+00
-7.52763391e-01 -8.62625003e-01 3.78746957e-01 2.11382493e-01
-2.05584794e-01 -1.21840501e+00 2.73175806e-01 -2.85624504e-01
1.06688809e+00 5.35069406e-01 1.00096273e+00 -1.20020938e+00
-5.17642558e-01 -7.34758019e-01 -1.77633330e-01 5.69144249e-01
-4.70382646e-02 1.08850569e-01 -7.45668232e-01 -1.89328045e-01
2.35128701e-01 -9.79910493e-02 7.75179744e-01 5.85634768e-01
8.01846027e-01 -5.72834492e-01 1.94246635e-01 5.37743747e-01
1.72291982e+00 7.28805482e-01 6.25054598e-01 9.65361536e-01
1.06105141e-01 2.15953901e-01 3.60092729e-01 6.09670043e-01
4.05550957e-01 1.28546432e-01 4.82675523e-01 1.10210352e-01
5.52973092e-01 -2.78667003e-01 5.00080466e-01 9.41480696e-01
-2.84514874e-01 1.25540555e-01 -9.30823803e-01 4.28050429e-01
-1.51868176e+00 -1.22426140e+00 6.14692867e-02 2.29615188e+00
4.94643062e-01 6.02893353e-01 3.33905779e-02 3.66535485e-01
1.31495506e-01 4.36181575e-01 -7.19883025e-01 -4.28183675e-01
-2.29999825e-01 1.94720626e-01 9.51387525e-01 2.36573189e-01
-1.04793239e+00 5.31481743e-01 6.59553528e+00 7.35637397e-02
-1.70123196e+00 -3.14605981e-01 1.06166852e+00 -1.11478694e-01
-2.84861058e-01 1.04638055e-01 -9.09844279e-01 9.07448113e-01
1.68658864e+00 -6.23402894e-01 1.90939307e-01 1.02161586e+00
4.42744344e-01 -7.22636357e-02 -7.65975952e-01 5.39382577e-01
-2.56063938e-01 -1.80380833e+00 -7.34417811e-02 1.29342750e-01
7.89265990e-01 3.19190234e-01 3.19600701e-01 4.88898486e-01
5.33090495e-02 -1.04108047e+00 1.07706189e+00 1.05929542e+00
3.62817675e-01 -9.61875618e-01 1.32548654e+00 6.11009538e-01
-9.93745029e-01 -5.26325047e-01 -4.15632784e-01 -6.79065645e-01
1.16132371e-01 6.26286864e-01 -4.99481767e-01 2.70373583e-01
6.60598636e-01 7.42148459e-01 -3.68430316e-01 8.83542836e-01
1.95404276e-01 7.56664395e-01 -2.86457568e-01 -3.93482968e-02
6.93724394e-01 -3.14981908e-01 6.48655789e-03 7.49610543e-01
8.03355038e-01 3.12248274e-04 -1.84959024e-01 6.83149874e-01
7.97066987e-02 3.67364511e-02 -5.25114715e-01 -1.66584492e-01
1.95852950e-01 6.45588577e-01 -4.51051086e-01 -4.67630535e-01
-7.52398729e-01 2.14628085e-01 -6.39062896e-02 1.93952978e-01
-7.30620980e-01 -1.89298749e-01 3.77714962e-01 1.85824677e-01
4.99407679e-01 -1.11742683e-01 -7.17552006e-01 -9.74350512e-01
-1.13719180e-01 -6.04065657e-01 2.09202111e-01 -8.06807458e-01
-1.15322697e+00 6.18705094e-01 -1.61229834e-01 -1.30366540e+00
-7.51473010e-01 -7.29624987e-01 -9.00832772e-01 1.28595781e+00
-1.95917726e+00 -6.25194848e-01 3.29555869e-01 9.38629508e-02
5.04698992e-01 -6.82800472e-01 7.82566190e-01 -5.52510172e-02
-4.15622294e-01 -9.06735733e-02 6.77022755e-01 4.69631523e-01
6.20896788e-03 -9.51484382e-01 8.46482694e-01 7.25246727e-01
-6.71318499e-03 3.82744849e-01 7.60738194e-01 -8.92122924e-01
-1.00074708e+00 -1.06207192e+00 1.27008843e+00 2.02055454e-01
1.03045654e+00 3.29473138e-01 -1.15030611e+00 1.01217186e+00
1.72202408e-01 -2.14399591e-01 4.32656884e-01 -4.69326377e-01
-1.51960582e-01 -2.26421773e-01 -9.24421549e-01 3.50487903e-02
-2.01819062e-01 -1.98782608e-01 -7.03079343e-01 1.76996261e-01
3.60980898e-01 -2.77408242e-01 -1.06754601e+00 3.02468002e-01
8.68304908e-01 -1.16374028e+00 6.03956759e-01 -3.19117904e-01
4.79319006e-01 -7.02094957e-02 -2.90520966e-01 -1.37379980e+00
-2.64149070e-01 -3.32293808e-01 -2.01537624e-01 7.72022009e-01
8.58213484e-01 -1.09336269e+00 7.39190280e-01 9.38174725e-01
7.33906701e-02 -1.06754732e+00 -9.58601058e-01 -7.62030661e-01
4.98928487e-01 -4.60216939e-01 1.09934628e+00 7.91506827e-01
-2.15209275e-01 -1.25935674e-01 -5.81535339e-01 8.09760168e-02
2.97629714e-01 6.55116320e-01 3.08417767e-01 -1.39292490e+00
-1.75642744e-01 -6.07544065e-01 -1.54771417e-01 -5.80484509e-01
7.16799423e-02 -5.41006267e-01 -3.18337411e-01 -1.48301017e+00
-2.87830502e-01 -4.81729865e-01 -6.42955780e-01 4.67941016e-01
2.54650623e-01 -8.05850402e-02 3.47099960e-01 6.57720089e-01
4.32138294e-01 2.69657254e-01 7.47959375e-01 -5.06707840e-02
-5.37036397e-02 4.91888762e-01 -3.82602632e-01 6.78135455e-01
1.23132563e+00 -1.18860371e-01 -1.03500456e-01 -2.12729558e-01
5.95899284e-01 2.77030021e-01 2.56804109e-01 -1.00780344e+00
9.66359079e-02 -2.76755631e-01 8.56846571e-01 -9.95952308e-01
1.64420962e-01 -7.37763941e-01 5.12836695e-01 8.72986674e-01
-2.94354349e-01 7.01648295e-01 2.71206081e-01 2.03165531e-01
-5.31206787e-01 -2.16800556e-01 5.87995470e-01 -3.71826738e-01
-8.92756462e-01 4.77010936e-01 -2.08984762e-01 -3.80179435e-01
1.00723362e+00 -2.10137397e-01 -4.23770696e-02 -6.66022539e-01
-5.08546412e-01 2.85773247e-01 5.73950633e-02 4.29935277e-01
5.18294156e-01 -1.17792022e+00 -9.08138573e-01 3.84120405e-01
-4.21099752e-01 -4.72836196e-01 -9.82078761e-02 6.02080464e-01
-7.96685100e-01 9.26477849e-01 -2.63363332e-01 1.49172647e-02
-4.00141925e-01 5.31707048e-01 6.66635454e-01 -4.62850422e-01
-7.31457889e-01 6.22131526e-01 -2.91236639e-01 1.47156879e-01
8.88371617e-02 -7.02804863e-01 -4.05349076e-01 4.26165760e-01
7.82677412e-01 2.27557197e-01 2.20547229e-01 -7.61486590e-01
5.98021597e-02 4.92582977e-01 1.24815241e-01 -7.53450543e-02
1.96721196e+00 1.43976703e-01 -7.01374784e-02 9.71751213e-01
1.05911613e+00 -5.39577186e-01 -1.21709120e+00 -5.59792109e-02
4.76707458e-01 -1.55782372e-01 2.36301884e-01 -9.93110836e-01
-1.39289570e+00 7.00935900e-01 4.90157068e-01 5.78106642e-01
1.04373145e+00 -6.12569511e-01 1.07667649e+00 2.57483989e-01
1.66553319e-01 -1.07920837e+00 -7.87023783e-01 8.08410287e-01
1.01431382e+00 -1.09418952e+00 -4.04264927e-02 5.05392373e-01
-6.77031696e-01 1.42651939e+00 7.26301223e-02 -5.05920649e-01
1.08166575e+00 3.27249080e-01 1.96876526e-01 -4.84069735e-02
-1.19095790e+00 2.07026973e-01 1.15162849e-01 -3.64616849e-02
5.28895438e-01 7.08299801e-02 1.60868224e-02 6.12822115e-01
-7.69534767e-01 2.98021287e-01 5.04640520e-01 7.47390568e-01
-6.04296207e-01 -5.38259685e-01 -4.20493335e-01 9.65100408e-01
-8.04487407e-01 -2.74617881e-01 3.08417864e-02 8.55112731e-01
-2.14712486e-01 5.37732482e-01 5.49079120e-01 -2.80689687e-01
2.52882808e-01 6.35924995e-01 -3.21992308e-01 -4.82249670e-02
-8.89095426e-01 -1.28574163e-01 -3.79246399e-02 -1.60790160e-01
-1.17503829e-01 -6.96723402e-01 -1.15764630e+00 -6.96397364e-01
2.20987387e-02 2.26913914e-01 6.80173934e-01 1.13779008e+00
2.09553204e-02 2.56852835e-01 1.01390302e+00 -9.38798308e-01
-1.07233477e+00 -1.14115155e+00 -8.23421061e-01 -4.92229685e-02
6.67794943e-01 -3.01659882e-01 -5.49933136e-01 -1.26339905e-02] | [4.471744060516357, 4.2193217277526855] |
61f51e01-5dc4-4be9-bb58-aaf94abc1a73 | synonym-replacement-based-on-a-study-of-basic | null | null | https://aclanthology.org/2021.nodalida-main.26 | https://aclanthology.org/2021.nodalida-main.26.pdf | Synonym Replacement based on a Study of Basic-level Nouns in Swedish Texts of Different Complexity | Basic-level terms have been described as the most important to human categorisation. They are the earliest emerging words in children’s language acquisition, and seem to be more frequently occurring in language in general. In this article, we explored the use of basic-level nouns in texts of different complexity, and hypothesise that hypernyms with characteristics of basic-level words could be useful for the task of lexical simplification. We conducted two corpus studies using four different corpora, two corpora of standard Swedish and two corpora of simple Swedish, and explored whether corpora of simple texts contain a higher proportion of basic-level nouns than corpora of standard Swedish. Based on insights from the corpus studies, we developed a novel algorithm for choosing the best synonym by rewarding high relative frequencies and monolexemity, and restricting the climb in the word hierarchy not to suggest synonyms of a too high level of inclusiveness. | ['Arne Jönsson', 'Evelina Rennes'] | null | null | null | null | nodalida-2021-5 | ['lexical-simplification'] | ['natural-language-processing'] | [ 1.03868254e-01 3.41649562e-01 -1.45135015e-01 -3.58670861e-01
1.14204451e-01 -3.87340546e-01 5.89296162e-01 9.28845763e-01
-1.04933810e+00 5.11742949e-01 5.89899838e-01 -5.28019369e-01
-5.97837865e-01 -5.34990191e-01 -1.45083264e-01 -2.46385619e-01
3.47326666e-01 6.07222497e-01 3.86822462e-01 -5.10052621e-01
3.71256411e-01 2.72764713e-01 -1.92128789e+00 8.64396319e-02
1.15542769e+00 -1.31586641e-01 5.12124658e-01 -3.07194833e-02
-3.58320147e-01 2.59360105e-01 -8.55912268e-01 -6.54208243e-01
-8.42996836e-02 -2.60268688e-01 -7.28050709e-01 -7.05829784e-02
3.13758314e-01 1.90826416e-01 8.38117301e-02 1.03546047e+00
4.77615863e-01 2.04793394e-01 8.20314169e-01 -5.90208709e-01
-3.22642237e-01 1.53061974e+00 -1.99519530e-01 4.15829897e-01
4.51866716e-01 -2.62431234e-01 1.08325982e+00 -5.08465409e-01
8.34806204e-01 1.50912821e+00 4.95073825e-01 6.63553178e-01
-1.32066393e+00 -6.98750556e-01 3.39053243e-01 -9.49824229e-03
-1.37539983e+00 -3.44240844e-01 3.49061549e-01 -5.42102814e-01
1.17699301e+00 2.81984121e-01 1.14670146e+00 7.31128514e-01
-1.81566283e-01 2.27838978e-01 1.03408384e+00 -1.12653232e+00
-6.42604604e-02 3.13212395e-01 5.01596451e-01 6.51156232e-02
1.06940544e+00 -1.96850300e-01 -4.90129650e-01 6.95300922e-02
3.13715130e-01 -4.14155781e-01 -1.18907258e-01 -6.95382580e-02
-1.16106808e+00 1.00060356e+00 -1.74328059e-01 1.14772463e+00
-9.59959999e-02 -1.01324536e-01 7.05938578e-01 2.67118543e-01
2.02638879e-01 9.20821965e-01 -4.62285489e-01 -2.35060155e-01
-7.00941920e-01 4.25681531e-01 7.19661117e-01 9.19913352e-01
2.26805434e-01 -9.65824425e-02 3.04584414e-01 1.31122112e+00
2.12661475e-01 3.28540295e-01 8.43414128e-01 -6.00575089e-01
2.35714853e-01 8.23732197e-01 -4.41004425e-01 -5.61671436e-01
-5.83330154e-01 -2.65269905e-01 -4.92836982e-01 -7.42322952e-02
4.43873912e-01 1.17855012e-01 -6.89342022e-01 2.01314712e+00
3.14685106e-01 -6.95733666e-01 1.37321174e-01 1.68863997e-01
8.66149068e-01 4.16208684e-01 6.22917950e-01 -6.17130756e-01
1.57459438e+00 -3.06285501e-01 -6.93400979e-01 -1.54005721e-01
1.01806486e+00 -8.67080510e-01 1.29274249e+00 3.68028224e-01
-1.07316601e+00 -3.84894669e-01 -1.10307300e+00 -1.73894048e-01
-7.03503251e-01 -2.55101591e-01 8.06313217e-01 1.02404475e+00
-7.70421088e-01 5.73903322e-01 -3.86947304e-01 -5.54786444e-01
-2.66703218e-01 2.80576497e-01 -4.39410001e-01 2.72626549e-01
-1.36454964e+00 1.30742884e+00 9.51090932e-01 -5.11397898e-01
2.05699459e-01 -6.22928798e-01 -8.43622625e-01 -6.16860799e-02
2.35173434e-01 -2.96073377e-01 1.14834309e+00 -8.97061229e-01
-8.98796082e-01 1.38794839e+00 4.10572141e-01 -2.35517278e-01
4.83043306e-03 -1.41659200e-01 -5.22263765e-01 -8.83784220e-02
3.56940478e-01 3.63388807e-01 4.85356778e-01 -8.26025248e-01
-8.71478677e-01 -1.55848294e-01 5.82729615e-02 3.53261858e-01
-2.27784619e-01 5.33233166e-01 1.68264508e-01 -9.92327392e-01
3.94033104e-01 -6.88276947e-01 -1.06277578e-01 -6.98260128e-01
6.69107363e-02 -8.17063510e-01 -1.74902231e-01 -5.57534695e-01
1.84268773e+00 -2.19081283e+00 8.03274661e-02 3.66306096e-01
2.29588524e-01 5.24113536e-01 1.80688530e-01 4.83760625e-01
-2.25236207e-01 4.79623139e-01 7.57506117e-02 1.89402536e-01
1.65593058e-01 4.03628021e-01 2.37081364e-01 1.72156632e-01
-1.14870355e-01 5.54936886e-01 -1.14900005e+00 -5.90752244e-01
2.48741105e-01 1.01921663e-01 -4.31545049e-01 -3.38557214e-01
-2.92049386e-02 -1.81216836e-01 -5.41178472e-02 2.72947848e-01
1.71246812e-01 5.29398978e-01 3.64451170e-01 1.70800477e-01
-4.61957186e-01 1.00112545e+00 -1.10273433e+00 1.09900379e+00
-4.70209360e-01 4.81737852e-01 -3.07433784e-01 -6.50223374e-01
5.41432440e-01 3.44370097e-01 -3.68968956e-02 -6.97074711e-01
4.38214004e-01 5.87014258e-01 1.02917194e+00 -4.02524292e-01
5.87591946e-01 -2.89766639e-01 -3.08778524e-01 3.72803509e-01
1.14132121e-01 -7.57518411e-01 9.29898977e-01 1.13925114e-01
5.72912633e-01 -1.75943732e-01 1.07630467e+00 -8.63544643e-01
5.46967030e-01 -2.24829718e-01 5.10137618e-01 5.31735778e-01
2.39028767e-01 7.56236240e-02 3.84341121e-01 -2.50631839e-01
-1.07956660e+00 -8.64832520e-01 -5.97791612e-01 1.33353853e+00
-3.27038765e-01 -8.72868061e-01 -8.37063491e-01 -3.60820852e-02
-2.12410226e-01 1.35930371e+00 -4.23361003e-01 -1.04999602e-01
-8.38269413e-01 -5.18550396e-01 5.64738452e-01 -3.51839215e-02
-2.40213290e-01 -1.21704769e+00 -1.01819777e+00 3.49689364e-01
2.13830322e-01 -8.31208706e-01 -6.53819516e-02 4.24362481e-01
-4.62197006e-01 -8.98170471e-01 -6.16935551e-01 -1.01576924e+00
4.95337695e-01 1.84802432e-02 1.21837604e+00 6.76065266e-01
4.14862856e-02 6.77491575e-02 -8.45893860e-01 -1.00354016e+00
-9.18698013e-01 4.06920612e-01 3.07149887e-01 -8.59767973e-01
7.74891555e-01 -6.28989577e-01 1.44182250e-01 -1.85857132e-01
-1.05532444e+00 -1.01808816e-01 2.17115596e-01 7.10148871e-01
7.78381005e-02 -7.14094192e-02 3.68992060e-01 -1.14670753e+00
6.51194990e-01 -1.89565822e-01 -3.19094837e-01 6.86540678e-02
-6.12646699e-01 1.97416022e-01 3.41142625e-01 -8.88827503e-01
-7.05448985e-01 -5.13518751e-01 -4.74484950e-01 4.30516154e-01
-3.39079231e-01 4.78630781e-01 -2.25454539e-01 6.73220903e-02
7.17585027e-01 -1.25019193e-01 -1.95937485e-01 -4.47594643e-01
4.01931226e-01 6.13840878e-01 1.78811237e-01 -8.85602534e-01
6.28795981e-01 -2.64077216e-01 6.18689461e-03 -1.41327035e+00
-5.98042309e-01 -5.19951344e-01 -8.32106888e-01 -6.94458373e-03
6.77599430e-01 -6.50734127e-01 -3.23995143e-01 2.75166839e-01
-1.07623434e+00 -1.80839315e-01 -5.67150116e-01 6.78897381e-01
-2.20782489e-01 4.26203340e-01 -1.16416566e-01 -5.73957443e-01
-1.65652797e-01 -8.51355374e-01 5.39039612e-01 2.14270785e-01
-1.05289614e+00 -9.64379609e-01 1.05221428e-01 -2.32463509e-01
1.13675736e-01 1.87178627e-01 1.67480075e+00 -1.33604360e+00
2.43998438e-01 1.80750355e-01 3.35157603e-01 1.68714598e-01
2.50483006e-01 2.59606630e-01 -4.72239941e-01 -2.06857901e-02
1.06758669e-01 -4.53778356e-02 4.81048018e-01 7.57846609e-02
3.28066319e-01 -2.99983323e-01 -1.55616671e-01 3.01470220e-01
1.29168999e+00 3.69121104e-01 3.72864097e-01 5.61361551e-01
5.42403281e-01 1.17535961e+00 5.22267520e-01 6.11251546e-03
2.88560897e-01 6.80676520e-01 -2.96869010e-01 2.49573246e-01
-2.84693539e-01 -7.82019496e-02 1.52652472e-01 1.54061007e+00
7.31765181e-02 -6.75404519e-02 -1.15629864e+00 1.03720760e+00
-1.25799632e+00 -6.87462091e-01 -4.55161959e-01 2.23943329e+00
1.16607130e+00 5.65834939e-01 3.69887561e-01 5.33406794e-01
8.41585100e-01 3.47528234e-02 3.26099396e-01 -9.74434733e-01
-2.61427850e-01 6.25522792e-01 2.27538645e-01 5.04910171e-01
-5.09088278e-01 1.24344897e+00 6.37761736e+00 1.12280238e+00
-8.13077331e-01 -1.52451964e-03 9.61432755e-02 -1.17152110e-01
-5.36286414e-01 5.31044267e-02 -8.51902664e-01 5.34496069e-01
9.78810847e-01 -5.38157403e-01 3.64194289e-02 3.99704307e-01
-2.13201437e-02 -3.20300967e-01 -1.21341145e+00 6.67583704e-01
-8.41197744e-03 -7.04210520e-01 3.13925356e-01 -1.98884398e-01
5.31877339e-01 -3.50958854e-01 -4.20742989e-01 2.39481136e-01
2.17250705e-01 -1.02099109e+00 1.03087354e+00 -8.44214018e-03
7.48497486e-01 -7.13796020e-01 5.96423626e-01 3.62836331e-01
-8.90669763e-01 2.32117280e-01 -4.42868948e-01 -4.49680150e-01
1.88120008e-01 2.22840145e-01 -4.39272702e-01 1.41923875e-01
4.27948654e-01 1.67729124e-01 -6.88099504e-01 8.90365183e-01
-5.73674917e-01 6.89515173e-01 -6.74614429e-01 -6.45814002e-01
1.76040113e-01 -1.67151541e-01 7.13902295e-01 1.44732165e+00
7.18954727e-02 2.13667274e-01 -2.45493189e-01 1.78190485e-01
4.06396985e-01 8.77823472e-01 -5.81793487e-01 -2.06886604e-02
8.58443201e-01 9.04518306e-01 -1.13677692e+00 -3.47296655e-01
-5.06817341e-01 3.41315657e-01 2.23721221e-01 -1.63562387e-01
-3.41789097e-01 -5.24903238e-01 6.06403768e-01 5.05792618e-01
-6.23683631e-02 -3.15175764e-02 -4.12124425e-01 -8.07506382e-01
-1.43829703e-01 -9.93952215e-01 2.65593439e-01 -2.40902528e-01
-1.18087125e+00 5.76262355e-01 7.33728647e-01 -6.82086527e-01
-3.67125213e-01 -7.01102555e-01 -4.01790440e-01 8.38603914e-01
-8.32076430e-01 -9.17513669e-01 3.65181953e-01 2.55557835e-01
5.91486335e-01 4.78176139e-02 9.27983999e-01 1.28222749e-01
-2.65550315e-01 6.49286687e-01 4.45498377e-02 -4.33984548e-02
3.74776334e-01 -1.31308198e+00 5.72183430e-01 7.48515725e-01
1.64803773e-01 9.96922016e-01 1.15868545e+00 -7.98827112e-01
-2.96652019e-01 -3.32148373e-01 1.53984916e+00 -8.89429599e-02
9.24285889e-01 -4.75044787e-01 -1.04433513e+00 3.58894110e-01
3.08937818e-01 -1.09290469e+00 8.21518302e-01 2.29934201e-01
-7.78792128e-02 2.06855714e-01 -1.04711211e+00 9.46847498e-01
1.20729423e+00 -1.80104420e-01 -1.45614409e+00 2.15843111e-01
1.01057887e+00 -2.01857612e-01 -8.30451131e-01 3.33772182e-01
5.46145499e-01 -6.44130886e-01 7.56977081e-01 -2.58174777e-01
-2.73826830e-02 7.58193135e-02 1.88703045e-01 -1.61798322e+00
-5.29009521e-01 -7.39066720e-01 5.89249671e-01 1.56859767e+00
4.77854699e-01 -7.18943059e-01 1.76833048e-01 4.37084615e-01
-2.88143456e-01 -4.97736305e-01 -1.06243825e+00 -7.48881638e-01
6.32821441e-01 -3.12719703e-01 5.88502228e-01 8.52006495e-01
4.54495311e-01 5.81025183e-01 2.44929627e-01 -4.93148685e-01
3.57507378e-01 -4.67952698e-01 3.10793161e-01 -1.84085083e+00
8.59939978e-02 -8.62913787e-01 -6.72077715e-01 -3.13327849e-01
2.67159224e-01 -9.15371597e-01 -1.12823382e-01 -1.35665512e+00
-1.75899670e-01 -4.24416751e-01 1.84082732e-01 2.12946028e-01
-8.94155577e-02 -3.73320691e-02 3.30293834e-01 -1.63260981e-01
2.25280263e-02 6.89041838e-02 7.70554245e-01 3.21982950e-01
-3.86557162e-01 -1.05773270e-01 -8.22607279e-01 1.25109732e+00
7.15481818e-01 -7.93783247e-01 -4.56012726e-01 -4.20605212e-01
6.45006478e-01 -6.65168345e-01 -3.78285408e-01 -8.38137686e-01
-1.02498597e-02 -4.51437235e-01 -3.97457592e-02 -3.09485883e-01
-1.41840473e-01 -8.34863603e-01 2.48197600e-01 6.65824234e-01
-3.01141441e-01 2.56210864e-01 2.68004745e-01 -4.03288543e-01
9.61885974e-02 -1.14908862e+00 9.12943959e-01 -2.04188138e-01
-7.35438228e-01 -4.62358236e-01 -7.82224417e-01 4.63600576e-01
1.11770439e+00 -7.25061834e-01 1.27130181e-01 1.99725777e-01
-5.63938081e-01 -4.74954881e-02 6.94858074e-01 5.70279717e-01
1.56209201e-01 -9.59851980e-01 -7.30137885e-01 1.66638494e-01
2.67097294e-01 -1.61795229e-01 -2.04606742e-01 4.11008626e-01
-9.31997359e-01 5.19833326e-01 -3.19768786e-01 -4.41493690e-02
-1.43372679e+00 5.07301688e-01 4.85920087e-02 6.24721684e-02
-6.02252066e-01 8.71420383e-01 1.13774903e-01 -4.10208315e-01
3.96202743e-01 -6.85930490e-01 -7.27991462e-01 5.42218685e-01
3.89635742e-01 5.68658054e-01 1.16837397e-01 -1.20305252e+00
-4.11425829e-01 5.85833192e-01 -2.09467366e-01 -2.66368270e-01
1.18462157e+00 -2.15896234e-01 -4.09170985e-01 8.70418549e-01
5.77091455e-01 5.62272787e-01 -1.07993744e-01 -3.23331803e-02
5.45077443e-01 -2.82757133e-01 -4.11634684e-01 -4.93945897e-01
-1.66897520e-01 5.28953731e-01 -4.33887132e-02 4.59886879e-01
9.55815196e-01 1.70684680e-01 3.99691254e-01 1.63049310e-01
4.11490321e-01 -1.36838686e+00 -5.41520894e-01 8.16488445e-01
6.79004848e-01 -4.04620469e-01 3.31201136e-01 -7.70708442e-01
-4.02486682e-01 9.33470607e-01 5.08224249e-01 -5.48160896e-02
5.67303598e-01 3.29924300e-02 -6.21243231e-02 -2.62937248e-01
-6.43345058e-01 -5.79935253e-01 3.16332102e-01 6.87456191e-01
7.96642542e-01 1.96638361e-01 -1.77931678e+00 7.46031642e-01
-9.89194036e-01 -6.44292474e-01 8.78683984e-01 7.20402002e-01
-6.02421165e-01 -1.51425505e+00 -3.87677163e-01 7.12525427e-01
-9.27756250e-01 -5.84665596e-01 -5.28196275e-01 1.29780650e+00
7.91674793e-01 8.37793171e-01 2.28876635e-01 -5.61578386e-02
3.82874280e-01 2.08556712e-01 6.50856614e-01 -1.19931793e+00
-9.61860240e-01 7.64500648e-02 3.52714211e-01 1.88697353e-01
-4.75037366e-01 -9.23669338e-01 -1.14792168e+00 -3.23124677e-01
-6.48229539e-01 4.20280159e-01 5.79265773e-01 1.16597295e+00
-6.19895697e-01 4.03247803e-01 -1.45804107e-01 -5.30870914e-01
-2.77800232e-01 -1.21381366e+00 -8.08515787e-01 3.66259158e-01
-2.15786457e-01 -6.73702657e-01 -4.05330896e-01 -2.49014974e-01] | [10.68879222869873, 10.22141170501709] |
d29914da-e121-4b54-aab2-8f252fcb92e2 | wider-and-higher-intensive-integration-and | 2210.06919 | null | https://arxiv.org/abs/2210.06919v1 | https://arxiv.org/pdf/2210.06919v1.pdf | Wider and Higher: Intensive Integration and Global Foreground Perception for Image Matting | This paper reviews recent deep-learning-based matting research and conceives our wider and higher motivation for image matting. Many approaches achieve alpha mattes with complex encoders to extract robust semantics, then resort to the U-net-like decoder to concatenate or fuse encoder features. However, image matting is essentially a pixel-wise regression, and the ideal situation is to perceive the maximum opacity correspondence from the input image. In this paper, we argue that the high-resolution feature representation, perception and communication are more crucial for matting accuracy. Therefore, we propose an Intensive Integration and Global Foreground Perception network (I2GFP) to integrate wider and higher feature streams. Wider means we combine intensive features in each decoder stage, while higher suggests we retain high-resolution intermediate features and perceive large-scale foreground appearance. Our motivation sacrifices model depth for a significant performance promotion. We perform extensive experiments to prove the proposed I2GFP model, and state-of-the-art results can be achieved on different public datasets. | ['Xin Yang', 'Qiang Zhang', 'Dongsheng Zhou', 'Yuxin Wang', 'Yuhao Liu', 'Ziqi Wei', 'Yu Qiao'] | 2022-10-13 | null | null | null | null | ['image-matting'] | ['computer-vision'] | [ 6.58974946e-01 8.14330727e-02 -2.07348049e-01 -3.91057193e-01
-5.78087151e-01 2.67067924e-02 4.66766894e-01 -2.80258417e-01
-1.47557318e-01 5.46580672e-01 2.94222653e-01 -2.52824724e-01
3.68946135e-01 -1.02559316e+00 -1.28939795e+00 -7.07500994e-01
2.57514626e-01 -2.44850875e-03 4.22887444e-01 -1.05963156e-01
5.56926169e-02 1.62187740e-01 -1.45914233e+00 9.52063918e-01
9.64935660e-01 1.22649026e+00 4.72404331e-01 7.95139015e-01
-4.09263313e-01 1.07986176e+00 -4.62849945e-01 -7.79791713e-01
4.45297211e-01 -5.06066263e-01 -6.18923664e-01 4.73869473e-01
7.08250344e-01 -8.38209629e-01 -4.81354713e-01 1.07266819e+00
9.54385847e-02 -3.04557770e-01 3.89149219e-01 -1.35811710e+00
-8.82436574e-01 7.93093383e-01 -8.51227224e-01 3.05469446e-02
1.21007137e-01 4.62183923e-01 9.13952410e-01 -8.63838911e-01
4.30762649e-01 1.33540428e+00 4.18763965e-01 3.06970447e-01
-1.00346065e+00 -4.03111488e-01 4.36322749e-01 1.76943600e-01
-9.93861437e-01 -3.45423013e-01 7.80725121e-01 -1.02294847e-01
5.54485023e-01 4.33026314e-01 8.85871828e-01 1.13603795e+00
4.26642716e-01 1.06044650e+00 1.15266776e+00 -2.10355192e-01
-1.37578294e-01 3.50453928e-02 -2.22332031e-01 9.63294923e-01
3.25321972e-01 -1.88454702e-01 -8.98028433e-01 3.43759179e-01
1.22212505e+00 3.99871498e-01 -3.61098111e-01 -1.26518518e-01
-1.61099648e+00 4.80589598e-01 8.90970409e-01 8.37697014e-02
-3.19651008e-01 5.43660164e-01 2.23080553e-02 4.62675184e-01
4.71684992e-01 2.00053938e-02 -1.51903898e-01 -9.23025757e-02
-1.17913175e+00 3.93044092e-02 3.09299737e-01 9.73065317e-01
1.10813987e+00 1.82801783e-01 -2.48061955e-01 5.70748150e-01
4.26022351e-01 6.72682106e-01 1.56511605e-01 -1.37251878e+00
5.37787080e-01 6.16051793e-01 -2.87318360e-02 -1.01966381e+00
2.25456595e-01 -3.71185929e-01 -1.08652437e+00 5.30198872e-01
2.76689142e-01 -1.20893214e-02 -1.20942235e+00 1.51201260e+00
7.65252039e-02 4.96654898e-01 6.08307403e-03 1.14648926e+00
7.78403163e-01 9.44232464e-01 -1.67121857e-01 -1.16807431e-01
1.40570247e+00 -1.46278369e+00 -7.66504645e-01 -3.87854487e-01
1.12379074e-01 -9.81454611e-01 1.05096781e+00 3.36768419e-01
-1.53131711e+00 -8.93134058e-01 -1.26333129e+00 -5.07709444e-01
-1.81977257e-01 -1.16663709e-01 8.03931534e-01 3.87976378e-01
-1.07956719e+00 6.70131445e-01 -8.23697746e-01 -1.64299414e-01
6.63425148e-01 2.17706025e-01 -1.88402995e-01 -1.91310227e-01
-9.21343386e-01 6.31812036e-01 2.45467126e-01 1.81969374e-01
-1.08129478e+00 -5.06404757e-01 -7.69685030e-01 -3.51218097e-02
2.28571907e-01 -1.21571445e+00 9.68645573e-01 -1.30145216e+00
-1.44940186e+00 8.26298296e-01 -3.63160729e-01 -5.75782061e-01
7.09983051e-01 -4.44905460e-01 -1.17701463e-01 2.97833592e-01
-8.31827223e-02 1.06527853e+00 1.18903375e+00 -1.79655218e+00
-6.90150321e-01 -2.16367953e-02 3.23058575e-01 2.45671973e-01
-3.36602747e-01 -2.05383137e-01 -5.33314228e-01 -8.30002546e-01
3.03549081e-01 -4.36529040e-01 -1.26187831e-01 4.36534375e-01
-3.17464918e-01 3.72297674e-01 8.48001122e-01 -7.82451391e-01
1.01513076e+00 -2.13054037e+00 2.42138267e-01 -9.77549702e-02
7.84310043e-01 1.37624785e-01 -2.13292152e-01 1.85002089e-01
2.16733992e-01 -6.97480887e-02 -2.97695547e-01 -5.97058654e-01
-2.79049147e-02 4.02223647e-01 -6.11753464e-01 2.19254345e-01
4.75359499e-01 1.33151424e+00 -7.25492835e-01 -6.31917238e-01
4.62330729e-01 5.50291896e-01 -6.07237041e-01 5.11643887e-01
-4.67498988e-01 1.83150023e-01 -3.47351074e-01 7.69374728e-01
9.68750238e-01 -4.03206378e-01 -1.67040437e-01 -6.69716775e-01
-2.42290888e-02 5.23453318e-02 -9.72317159e-01 1.99472594e+00
-3.08811456e-01 6.29791796e-01 2.73023546e-01 -7.02037871e-01
7.20108449e-01 -4.10167053e-02 4.72807229e-01 -6.75936103e-01
2.35163376e-01 1.32566556e-01 -1.38688117e-01 -2.33502328e-01
5.72463632e-01 1.58772945e-01 3.62930894e-01 3.25211942e-01
-6.03708103e-02 -9.23257601e-03 -2.52392411e-01 4.94876832e-01
9.91627872e-01 4.23880875e-01 -9.97041464e-02 5.11425454e-03
2.59391099e-01 -2.78702438e-01 3.86285305e-01 7.25275755e-01
-1.30121052e-01 9.36584890e-01 4.16738153e-01 -4.36411977e-01
-1.08854163e+00 -1.44983065e+00 2.29853451e-01 1.02811062e+00
7.04895020e-01 -4.43623930e-01 -9.25754786e-01 -5.56580901e-01
-9.97033268e-02 5.23767471e-01 -7.09161043e-01 -1.46234751e-01
-6.25044465e-01 -6.25983477e-01 2.89197028e-01 5.65751314e-01
1.07626653e+00 -9.93773758e-01 -5.52580416e-01 8.23046714e-02
-4.13429379e-01 -1.07620764e+00 -4.04694378e-01 1.29367530e-01
-9.06489193e-01 -6.09784365e-01 -8.21023285e-01 -7.10026503e-01
5.87738693e-01 7.56556213e-01 1.22258270e+00 4.51820076e-01
-1.34835187e-02 8.75651836e-02 -3.12835246e-01 -1.50674582e-01
-3.21157396e-01 -1.50404468e-01 -4.44077551e-01 2.51060605e-01
2.96310112e-02 -8.15758288e-01 -8.79034042e-01 4.69335765e-02
-1.13367629e+00 7.89959550e-01 1.12204862e+00 6.91140473e-01
5.10046363e-01 -2.16043994e-01 3.14900100e-01 -8.93806577e-01
1.30578995e-01 -2.05295473e-01 -1.01962209e-01 3.50183398e-01
-3.80584985e-01 -5.97374961e-02 3.38658690e-01 -1.64177954e-01
-1.24499798e+00 -1.44743487e-01 -2.65692860e-01 -6.45071685e-01
-1.01608932e-01 1.64541915e-01 -5.17358661e-01 -1.69804379e-01
3.31272006e-01 4.40534562e-01 -1.80352684e-02 -3.74834508e-01
8.76041830e-01 4.90432709e-01 7.42923439e-01 -6.58849955e-01
1.03911889e+00 8.12872112e-01 -1.60158649e-01 -4.70408767e-01
-1.19547820e+00 -6.78130573e-06 -6.21989071e-01 -3.37210238e-01
1.16307998e+00 -1.10638344e+00 -6.19567990e-01 4.89876390e-01
-1.26719213e+00 -4.05585051e-01 -2.72593975e-01 1.48138655e-02
-6.38622224e-01 4.33790505e-01 -9.20097589e-01 -5.61586678e-01
-4.21215951e-01 -1.32002115e+00 1.35941482e+00 1.95375964e-01
3.33608001e-01 -6.67834461e-01 -5.03201365e-01 6.03994548e-01
6.01460278e-01 2.12507546e-01 6.35286450e-01 2.90917456e-01
-1.50882339e+00 4.44500238e-01 -8.86421204e-01 4.01492536e-01
1.66595489e-01 6.37182891e-02 -1.31468546e+00 -1.47776842e-01
1.96188707e-02 -2.28210509e-01 1.46060824e+00 2.38418594e-01
1.51687706e+00 -4.19714063e-01 -1.04351386e-01 1.05735719e+00
1.47284114e+00 -2.08440617e-01 1.10516787e+00 2.62171179e-01
1.24113357e+00 2.77706593e-01 5.50340831e-01 2.82436371e-01
5.23861051e-01 3.43081385e-01 8.03450823e-01 -7.17639506e-01
-6.20091319e-01 -2.98176825e-01 7.17834890e-01 1.03104615e+00
-1.46882087e-01 -2.72148401e-01 -3.09167594e-01 1.94420278e-01
-2.05297470e+00 -8.98892581e-01 -1.27109393e-01 1.80043745e+00
1.02282512e+00 4.26910937e-01 -3.60559486e-02 -1.49308681e-01
5.21684289e-01 4.59913790e-01 -4.59138781e-01 -2.45401725e-01
-4.20746654e-01 1.59060240e-01 5.09383380e-01 5.66217721e-01
-1.01465392e+00 1.06532264e+00 6.21988153e+00 1.00111187e+00
-1.17676795e+00 8.14639255e-02 1.05783796e+00 -4.79061864e-02
-6.43442035e-01 2.72204801e-02 -3.73168349e-01 5.14726579e-01
6.60384834e-01 1.18762083e-01 5.54824591e-01 5.41392624e-01
-7.83816166e-03 -1.40522882e-01 -1.04942894e+00 1.02007568e+00
1.82972565e-01 -1.70072758e+00 5.40089071e-01 8.79891738e-02
8.43045175e-01 -1.60121337e-01 3.25903058e-01 4.61627059e-02
1.89560711e-01 -1.10327923e+00 9.15529013e-01 6.85280800e-01
7.30955124e-01 -5.63089252e-01 5.92436254e-01 -1.14256022e-02
-1.29315042e+00 1.37432292e-01 -6.05152071e-01 -1.11898459e-01
3.00419211e-01 8.49038839e-01 -3.80933017e-01 7.69576430e-01
5.85722446e-01 9.08675253e-01 -7.43848264e-01 8.13083529e-01
-5.93854487e-03 6.03236914e-01 -1.36496365e-01 3.74065667e-01
2.54477650e-01 -2.36761004e-01 2.02507183e-01 1.20829356e+00
3.14601511e-01 -1.66114658e-01 2.50648350e-01 1.04535520e+00
-3.30108106e-01 -1.42237410e-01 -3.19253981e-01 -3.55684198e-04
1.50344923e-01 1.27943885e+00 -7.19031274e-01 -7.15855420e-01
-5.88219166e-01 1.48449433e+00 2.96597213e-01 4.02292818e-01
-1.22834098e+00 -6.59454241e-02 8.60240757e-01 2.16018677e-01
3.33712697e-01 -2.65480012e-01 -7.31610537e-01 -1.36912608e+00
1.55934421e-02 -9.46482420e-01 1.01959899e-01 -1.04899180e+00
-1.18400836e+00 3.88863742e-01 -2.90886521e-01 -1.19594157e+00
3.59646678e-01 -6.75852239e-01 -6.59126222e-01 6.47643507e-01
-1.73110354e+00 -1.71299124e+00 -6.94828451e-01 5.93332112e-01
6.38235748e-01 2.21824393e-01 3.42752159e-01 4.57145035e-01
-4.36861306e-01 4.66619939e-01 -1.84885472e-01 1.13716997e-01
8.27640355e-01 -1.32921243e+00 4.77940828e-01 1.04070318e+00
3.26458812e-01 5.65728307e-01 5.01616836e-01 -6.50188029e-01
-1.66435134e+00 -1.26781559e+00 4.45359945e-01 -4.40055370e-01
5.66207886e-01 -4.12342191e-01 -8.54188025e-01 7.10294366e-01
7.84997582e-01 -8.15924481e-02 2.44123653e-01 -3.49761903e-01
-4.00023699e-01 -3.63826156e-01 -9.08578515e-01 8.40298295e-01
1.28184295e+00 -3.76114488e-01 -4.22972649e-01 1.10747345e-01
1.26075411e+00 -3.47805351e-01 -8.02390516e-01 4.07328427e-01
5.38382709e-01 -1.35863256e+00 1.16022575e+00 -1.76005572e-01
9.65995848e-01 -7.26379037e-01 -4.38053071e-01 -9.30595636e-01
-4.08749670e-01 -6.32934868e-01 -3.34702909e-01 1.37095439e+00
1.24628909e-01 -3.06123793e-01 7.97523737e-01 1.41440451e-01
-2.94078022e-01 -8.64923120e-01 -6.14396334e-01 -2.27199376e-01
-1.58645019e-01 -4.13440824e-01 6.57967389e-01 7.27996409e-01
-4.91665393e-01 2.55198449e-01 -8.12627256e-01 2.09889635e-01
9.03739035e-01 5.13490319e-01 9.06503201e-01 -6.49729192e-01
-4.31606323e-01 -3.90153915e-01 -3.77213180e-01 -1.59042418e+00
-2.92345375e-01 -6.80614412e-01 1.14221700e-01 -1.66501236e+00
5.15697360e-01 -2.12033376e-01 -4.17581767e-01 3.26376110e-01
-5.15706241e-01 5.80842137e-01 4.78545189e-01 1.83002904e-01
-8.66288185e-01 7.27846682e-01 1.76443422e+00 -3.53635013e-01
3.28450382e-01 -3.52792859e-01 -7.95968175e-01 6.79615259e-01
6.63978338e-01 8.76543298e-03 -3.64806622e-01 -8.89330626e-01
9.34680924e-02 -2.24019513e-01 5.94547927e-01 -1.03304946e+00
6.00290895e-02 -4.78945911e-01 9.36159670e-01 -6.37734354e-01
6.73531830e-01 -8.31801772e-01 2.82458011e-02 3.47230166e-01
-2.04894751e-01 -2.23348942e-02 -2.00873479e-01 5.83249569e-01
-2.22816363e-01 1.24122746e-01 7.86847234e-01 -3.57265204e-01
-8.29294086e-01 4.69727933e-01 -1.38953343e-01 -1.99718654e-01
8.30443442e-01 -3.21832508e-01 -5.89302421e-01 -4.61744100e-01
-4.97333050e-01 -4.63311560e-02 8.74778867e-01 2.78677523e-01
9.62806106e-01 -1.40688419e+00 -7.01580226e-01 3.49226147e-01
-7.67274052e-02 1.57629400e-01 1.80204332e-01 7.52856612e-01
-8.09022784e-01 -1.80385575e-01 -4.49812561e-01 -8.00355017e-01
-9.88562286e-01 5.12686431e-01 1.53228521e-01 4.58697230e-02
-7.96067595e-01 1.14002562e+00 7.28534400e-01 1.64759785e-01
2.59593576e-01 -5.89929461e-01 4.03183669e-01 -2.10446745e-01
5.97083092e-01 8.37252215e-02 -3.14711064e-01 -4.61515635e-01
8.31186678e-03 4.57985312e-01 -8.86050984e-02 -1.93565097e-02
1.16411126e+00 -5.12218952e-01 -3.81024629e-01 4.03771609e-01
8.85529876e-01 -6.90476298e-02 -1.68900585e+00 -3.06091189e-01
-4.76148039e-01 -9.03993368e-01 2.79993881e-02 -5.14793754e-01
-1.37461197e+00 1.17290020e+00 5.42709410e-01 8.33370388e-02
1.39545059e+00 -7.48948455e-02 1.26920688e+00 1.54882118e-01
2.67625093e-01 -7.90673077e-01 5.11338115e-01 1.48941353e-01
7.73794413e-01 -1.28766131e+00 1.19575568e-01 -7.02579618e-01
-7.12832630e-01 1.02742350e+00 1.01100039e+00 -1.74372539e-01
3.94118160e-01 5.39092600e-01 3.08300406e-01 6.58452958e-02
-8.80919337e-01 -3.28748643e-01 2.10474119e-01 5.12764692e-01
4.00372386e-01 5.52811660e-02 1.39920741e-01 3.09508055e-01
-1.84099257e-01 -2.01823503e-01 4.04854029e-01 7.58695781e-01
-8.79488766e-01 -1.08760858e+00 -3.46602350e-01 5.45554221e-01
-2.09948257e-01 -2.49268413e-01 -2.90893376e-01 2.88949579e-01
4.72268671e-01 8.12784493e-01 9.20086429e-02 -4.88062501e-01
-2.00395450e-01 -1.97354361e-01 7.15571821e-01 -3.70489657e-01
-3.97910863e-01 2.66853780e-01 -1.51006594e-01 -9.05281425e-01
-5.36010385e-01 -2.00358361e-01 -1.17067432e+00 -7.81301856e-01
-7.38082752e-02 -3.39587778e-01 4.08108443e-01 8.66940618e-01
2.69459784e-01 9.00370002e-01 5.24717629e-01 -9.27714169e-01
-7.77299255e-02 -7.58719742e-01 -4.07826066e-01 5.04720688e-01
4.39552158e-01 -5.19582808e-01 -1.15569368e-01 4.98763740e-01] | [10.63437557220459, -0.9048005938529968] |
3b8a5c60-61c6-4aa0-9967-c43fd8eb49e8 | bridging-images-and-videos-a-simple-learning | 2212.10147 | null | https://arxiv.org/abs/2212.10147v1 | https://arxiv.org/pdf/2212.10147v1.pdf | Bridging Images and Videos: A Simple Learning Framework for Large Vocabulary Video Object Detection | Scaling object taxonomies is one of the important steps toward a robust real-world deployment of recognition systems. We have faced remarkable progress in images since the introduction of the LVIS benchmark. To continue this success in videos, a new video benchmark, TAO, was recently presented. Given the recent encouraging results from both detection and tracking communities, we are interested in marrying those two advances and building a strong large vocabulary video tracker. However, supervisions in LVIS and TAO are inherently sparse or even missing, posing two new challenges for training the large vocabulary trackers. First, no tracking supervisions are in LVIS, which leads to inconsistent learning of detection (with LVIS and TAO) and tracking (only with TAO). Second, the detection supervisions in TAO are partial, which results in catastrophic forgetting of absent LVIS categories during video fine-tuning. To resolve these challenges, we present a simple but effective learning framework that takes full advantage of all available training data to learn detection and tracking while not losing any LVIS categories to recognize. With this new learning scheme, we show that consistent improvements of various large vocabulary trackers are capable, setting strong baseline results on the challenging TAO benchmarks. | ['Joon-Young Lee', 'In So Kweon', 'Seoung Wug Oh', 'KwanYong Park', 'Sanghyun Woo'] | 2022-12-20 | null | null | null | null | ['video-object-detection'] | ['computer-vision'] | [ 3.78023013e-02 -4.32323098e-01 -4.16088760e-01 -8.05231705e-02
-7.14932561e-01 -7.14622200e-01 6.09316051e-01 -4.45143968e-01
-3.76803070e-01 5.20727873e-01 7.52189308e-02 5.46677224e-02
2.14829966e-01 -1.06208839e-01 -1.02223396e+00 -7.35655487e-01
-2.17320487e-01 4.38659102e-01 7.82310545e-01 -4.33518104e-02
-2.14537784e-01 3.70022297e-01 -1.91416764e+00 1.45167753e-01
3.48637879e-01 1.10268748e+00 8.56359452e-02 5.76349795e-01
1.12111598e-01 8.44986141e-01 -6.85089588e-01 -3.54600430e-01
7.12513268e-01 -1.27854506e-02 -1.18444785e-01 1.22639716e-01
1.59276927e+00 -4.83303368e-01 -8.12994897e-01 9.87750411e-01
3.23112011e-01 -1.10148020e-01 4.46622103e-01 -1.55779481e+00
-5.44807732e-01 5.28450131e-01 -6.41218901e-01 4.35614496e-01
-9.50827450e-02 4.24257368e-01 1.02615118e+00 -1.18296492e+00
6.25706017e-01 1.24011910e+00 1.01361895e+00 6.27979934e-01
-8.99890780e-01 -1.00955725e+00 4.85967338e-01 2.34716147e-01
-1.73909128e+00 -7.20797539e-01 1.66187391e-01 -6.78568006e-01
7.90946543e-01 2.53981978e-01 6.84727788e-01 1.22336709e+00
8.32222849e-02 9.92352247e-01 6.53298378e-01 -1.64387092e-01
-1.00671552e-01 1.81506395e-01 2.67146647e-01 5.99061787e-01
7.30241954e-01 5.11402607e-01 -8.23851347e-01 4.47467621e-03
6.74812496e-01 3.38162668e-02 -7.62491347e-03 -7.98380613e-01
-1.48909855e+00 4.68593270e-01 2.92054564e-01 1.39683366e-01
9.21737999e-02 4.97060061e-01 5.90491712e-01 4.82750654e-01
2.11298287e-01 2.00963825e-01 -3.74898762e-01 -5.87603338e-02
-1.16433191e+00 2.54265696e-01 4.37185913e-01 1.38165867e+00
4.08954740e-01 4.16878700e-01 -5.83085716e-01 4.90518183e-01
2.55469680e-01 1.12665796e+00 2.36452118e-01 -8.36655796e-01
4.04501200e-01 1.94597349e-01 1.53470755e-01 -7.97200859e-01
-2.53611863e-01 -6.94843948e-01 -5.08318901e-01 2.37304568e-01
5.87023616e-01 1.47331906e-02 -1.02180195e+00 1.78518391e+00
3.60127956e-01 6.28872752e-01 -3.04512054e-01 7.36087501e-01
8.04740608e-01 2.67432630e-01 1.21243112e-01 -1.75461248e-01
1.36165583e+00 -1.01426685e+00 -6.94041371e-01 -2.60603875e-01
5.86732805e-01 -7.43013978e-01 8.09847772e-01 4.35157478e-01
-5.93783140e-01 -7.63334453e-01 -1.25527334e+00 2.03170866e-01
-2.38085210e-01 2.76735216e-01 7.02861488e-01 8.34269643e-01
-1.07712436e+00 1.27413064e-01 -9.02364492e-01 -6.57454312e-01
6.72278166e-01 4.84409899e-01 -3.29448342e-01 -1.33691117e-01
-8.72807980e-01 7.89005160e-01 2.31471255e-01 -1.03407316e-01
-1.38892519e+00 -8.36462915e-01 -4.98100102e-01 2.90390272e-02
8.93258870e-01 -3.23844165e-01 1.09987628e+00 -6.85880065e-01
-8.14242780e-01 6.54765844e-01 -2.07747091e-02 -5.62542439e-01
6.41963065e-01 -2.98219562e-01 -6.68307841e-01 -1.63446128e-01
2.76651293e-01 8.71727169e-01 1.30128419e+00 -1.05129695e+00
-1.03746569e+00 -2.28584930e-01 -1.96559817e-01 -7.90004153e-03
-7.36391008e-01 2.22334296e-01 -7.75281072e-01 -7.77222037e-01
-2.82696396e-01 -1.11159325e+00 2.29199752e-01 4.12522882e-01
-3.82556133e-02 -3.81998867e-01 1.17877614e+00 -3.41369748e-01
1.07868862e+00 -2.17956591e+00 -6.17183596e-02 -1.98749110e-01
6.32785082e-01 6.77673697e-01 -3.32966805e-01 -3.70546099e-04
1.68674186e-01 -1.36856079e-01 4.78933811e-01 -4.32459623e-01
-3.68969850e-02 3.68720829e-01 -5.90352356e-01 5.65972924e-01
1.26496375e-01 9.13981438e-01 -8.79999697e-01 -5.93337715e-01
1.75060496e-01 4.06946331e-01 -5.59684873e-01 -1.60484817e-02
-2.50448853e-01 4.91286010e-01 -2.41711929e-01 9.80458319e-01
6.23338163e-01 -3.77148211e-01 4.80908295e-03 -3.90486419e-01
-1.83370605e-01 1.26193827e-02 -1.32549095e+00 1.43078613e+00
1.95823178e-01 8.76277626e-01 1.02313356e-02 -8.81447256e-01
5.61185539e-01 3.30984563e-01 7.39299953e-01 -5.23986161e-01
3.04036750e-03 1.86075136e-01 -1.73486501e-01 -2.91090816e-01
4.46723074e-01 1.66057289e-01 1.71482921e-01 1.43971860e-01
3.81417423e-01 2.90656775e-01 4.75037605e-01 4.63916838e-01
1.16923237e+00 2.20090419e-01 -1.36307970e-01 -1.62431106e-01
2.70809621e-01 2.01677665e-01 8.42512012e-01 1.24842739e+00
-5.77037215e-01 5.16098261e-01 7.57581964e-02 -6.59362555e-01
-1.09368050e+00 -1.16924775e+00 -2.03995198e-01 1.35226727e+00
1.81543559e-01 -7.52938867e-01 -3.72536182e-01 -7.69879460e-01
2.98448592e-01 -2.98342928e-02 -5.89079440e-01 -9.75507423e-02
-7.22119391e-01 -5.41684687e-01 8.65242064e-01 7.65488625e-01
2.17365384e-01 -6.11685276e-01 -4.65628028e-01 -5.27211349e-04
-1.09198913e-01 -1.46896911e+00 -8.88363421e-01 1.11959964e-01
-6.24050021e-01 -1.11930788e+00 -6.87661707e-01 -6.25792980e-01
2.82770276e-01 7.80743599e-01 1.08791888e+00 2.61228621e-01
-5.45042098e-01 5.76745868e-01 -2.24119395e-01 -5.85003018e-01
-2.50909597e-01 5.59418835e-02 6.95062757e-01 2.13426538e-02
4.85950649e-01 -3.12035214e-02 -2.92938262e-01 6.89859807e-01
-5.43832183e-01 -3.58271778e-01 5.45698106e-01 7.40546525e-01
4.68410969e-01 -2.69003361e-01 3.84992242e-01 -4.32654351e-01
-2.02197805e-01 -2.48553649e-01 -1.05069327e+00 4.60019976e-01
-5.49017668e-01 -9.03307050e-02 1.51777416e-01 -8.57505739e-01
-6.64506257e-01 1.63483933e-01 4.61380109e-02 -9.50184762e-01
2.32491031e-01 -2.78843474e-02 2.93774493e-02 -6.60160363e-01
7.19577312e-01 2.11125076e-01 -2.88840123e-02 -5.69992542e-01
2.26115033e-01 5.09229600e-01 6.58778310e-01 -4.11288321e-01
1.21652853e+00 5.96720934e-01 -2.42252633e-01 -8.88336420e-01
-1.11761248e+00 -7.62257874e-01 -6.70601904e-01 -2.77003765e-01
7.01757789e-01 -1.48409677e+00 -6.48209870e-01 4.47484821e-01
-7.57312357e-01 -3.72859016e-02 -5.38199127e-01 4.82369810e-01
-3.29631209e-01 5.58546126e-01 -3.33699703e-01 -4.26320910e-01
-6.38842359e-02 -1.20901299e+00 1.16908944e+00 1.24110118e-01
1.09733388e-01 -5.03634810e-01 5.93049973e-02 3.72641683e-01
6.18331969e-01 -2.51521528e-01 3.77802327e-02 -5.24902582e-01
-1.11185133e+00 -3.70624781e-01 -3.94431025e-01 4.23667222e-01
6.73202425e-02 6.60918802e-02 -9.65529025e-01 -9.17028248e-01
-2.85779536e-01 -5.11816859e-01 9.90475893e-01 1.94736600e-01
6.55892611e-01 3.44429165e-02 -7.20258415e-01 9.05234218e-01
1.16050887e+00 -6.39425823e-03 2.20323801e-01 2.65241653e-01
1.04633963e+00 -9.26167369e-02 8.46492112e-01 3.29497576e-01
4.17550862e-01 1.17784548e+00 3.12587053e-01 3.08783185e-02
-8.38410258e-01 -2.43387640e-01 8.09757113e-01 6.71375394e-01
-2.92127617e-02 -1.53093645e-02 -7.53016889e-01 5.32318711e-01
-1.85573423e+00 -1.12791502e+00 1.80403180e-02 2.46819186e+00
6.14967763e-01 1.39062524e-01 3.15899372e-01 -3.58223766e-01
5.60822666e-01 2.09232822e-01 -6.99855149e-01 5.56511998e-01
-3.45879048e-01 -1.88761145e-01 9.20408845e-01 1.87197924e-01
-1.48675287e+00 1.22501647e+00 7.21242237e+00 8.68617356e-01
-1.27170122e+00 4.41070616e-01 -9.91877466e-02 -4.45391148e-01
2.79224217e-01 -3.74806561e-02 -1.67746484e+00 4.01725650e-01
7.69981444e-01 -8.08191001e-02 2.70526618e-01 9.48461533e-01
-1.93226144e-01 3.51330221e-01 -1.12443829e+00 1.23085797e+00
3.41368824e-01 -1.45361531e+00 1.28450513e-01 -3.08625437e-02
7.17575550e-01 7.38220572e-01 8.80856588e-02 7.47046351e-01
3.36754888e-01 -6.40215099e-01 1.01550305e+00 5.61839268e-02
1.07754421e+00 -3.61426659e-02 3.87015849e-01 1.78725332e-01
-1.55331790e+00 -3.08176160e-01 -7.53462017e-01 1.73022285e-01
1.05235323e-01 1.82962254e-01 -7.02251911e-01 2.54024148e-01
1.06085289e+00 9.91711140e-01 -8.88514519e-01 1.45218766e+00
3.38821560e-02 6.76993370e-01 -6.58297539e-01 3.67708355e-01
-3.73686813e-02 3.08314949e-01 7.31257200e-01 1.04733622e+00
1.82118222e-01 -3.28611434e-01 7.17937529e-01 3.84797126e-01
-3.15848649e-01 -1.96643963e-01 -6.09024286e-01 -5.75318150e-02
5.16702414e-01 1.18284845e+00 -5.93878031e-01 -7.24825621e-01
-7.60327816e-01 4.15375412e-01 3.08738470e-01 2.08977848e-01
-1.21694422e+00 3.10103327e-01 1.03410327e+00 2.68144071e-01
7.71661520e-01 -4.29836690e-01 7.62561411e-02 -1.78459275e+00
6.46730140e-02 -1.09531224e+00 5.45993388e-01 -3.07543516e-01
-1.09154630e+00 4.35165226e-01 3.81465927e-02 -1.52588356e+00
1.77038550e-01 -5.85330725e-01 -2.89039612e-01 2.01032534e-01
-1.57419538e+00 -1.44518840e+00 -5.05628467e-01 7.51113892e-01
6.81161821e-01 -4.94739413e-01 2.39750370e-01 9.52571034e-01
-4.81295645e-01 1.08423889e+00 2.42013916e-01 1.44132778e-01
1.30774176e+00 -8.18749547e-01 3.84266227e-01 1.03797066e+00
6.87577903e-01 5.12036264e-01 7.55661130e-01 -8.14622045e-01
-1.66607368e+00 -1.19159532e+00 6.73637688e-01 -1.04835927e+00
9.56525505e-01 -7.91582704e-01 -8.41640711e-01 1.15434849e+00
-3.24272998e-02 5.50597906e-01 3.49591881e-01 2.20408231e-01
-7.22179532e-01 -2.94735491e-01 -6.19905412e-01 3.47644985e-01
1.43527472e+00 -2.54022568e-01 -4.66861457e-01 6.86681092e-01
6.84377491e-01 -4.43427205e-01 -7.96070039e-01 5.44236064e-01
6.95431113e-01 -6.40816391e-01 1.35928762e+00 -6.15881443e-01
-4.99227881e-01 -7.70282209e-01 -1.50286332e-01 -7.37339854e-01
-4.23785836e-01 -4.97246027e-01 -4.74124998e-01 1.23405135e+00
1.06855370e-02 -4.06954646e-01 1.03037214e+00 2.22147629e-01
-1.53548270e-01 -6.43240139e-02 -1.06950867e+00 -1.44002247e+00
-7.60409608e-02 -2.90860116e-01 3.13591808e-01 7.85664678e-01
-4.72287178e-01 1.98383987e-01 -1.10101092e+00 2.54454911e-01
1.04068112e+00 -1.11185238e-01 1.12321305e+00 -1.32871187e+00
-2.88669735e-01 -1.71986431e-01 -6.61458075e-01 -1.44250870e+00
-3.42798620e-01 -7.47696340e-01 1.33177251e-01 -8.55099022e-01
3.88207644e-01 -5.77347457e-01 -4.76957530e-01 6.78238273e-01
-1.96016684e-01 6.72665358e-01 6.67779922e-01 6.88867331e-01
-1.29516304e+00 5.52792192e-01 9.51710999e-01 -5.47487676e-01
2.11597785e-01 -3.99213731e-02 -6.19620860e-01 4.88907337e-01
1.85743734e-01 -7.12284565e-01 -2.71284074e-01 -5.93482792e-01
-4.41818796e-02 -4.40551013e-01 4.07009363e-01 -1.48801398e+00
5.96694708e-01 1.41896069e-01 3.74148369e-01 -8.21832657e-01
3.25826555e-01 -7.96988308e-01 1.17304496e-01 3.97360444e-01
-2.68923379e-02 -7.14811385e-02 2.43127272e-01 6.41813934e-01
-1.55469909e-01 1.99100554e-01 8.26897740e-01 7.24377260e-02
-1.19489026e+00 7.18011975e-01 -2.13074535e-01 2.82535881e-01
1.23627925e+00 -2.77730376e-01 -4.49340373e-01 -3.44547890e-02
-6.76970422e-01 5.88049948e-01 5.24775922e-01 9.98183966e-01
2.07723513e-01 -1.44358110e+00 -7.28746593e-01 3.27994436e-01
4.72861737e-01 -4.51295048e-01 9.00844783e-02 1.01863658e+00
-1.26213521e-01 6.96369827e-01 -2.76619464e-01 -1.03279114e+00
-1.58479941e+00 9.62950826e-01 2.13791549e-01 -1.76891550e-01
-7.89632022e-01 8.75182688e-01 4.74719465e-01 -1.40679121e-01
7.18036771e-01 -9.49277058e-02 -9.83357653e-02 8.07891935e-02
8.08600426e-01 1.64012864e-01 2.44094599e-02 -7.72674143e-01
-5.40734887e-01 7.77852356e-01 -3.68961751e-01 4.11341101e-01
1.15006006e+00 -2.03967646e-01 4.44644898e-01 1.49316281e-01
7.25333095e-01 4.74276654e-02 -1.61636579e+00 -5.12421370e-01
1.43771350e-01 -7.50304818e-01 -1.00725338e-01 -5.15514910e-01
-1.15177214e+00 7.41273820e-01 9.11354363e-01 -1.21234782e-01
6.63380206e-01 7.52761960e-02 7.17167437e-01 5.39308488e-01
6.37332678e-01 -1.04210949e+00 1.34977818e-01 6.65781796e-01
5.46334863e-01 -1.52309942e+00 8.99836645e-02 -3.74766886e-01
-4.64106411e-01 7.09980547e-01 8.42820346e-01 5.48801050e-02
5.10969996e-01 5.26681006e-01 2.51735866e-01 -2.86074668e-01
-9.43503141e-01 -2.75328934e-01 3.32917273e-01 6.71983063e-01
2.08741263e-01 -2.24289909e-01 1.01558775e-01 5.95618226e-02
3.57655734e-01 1.78920388e-01 2.90177643e-01 8.39837849e-01
-6.40110195e-01 -1.13717091e+00 -5.41440666e-01 5.34363270e-01
-2.82516479e-01 -3.40392813e-02 -1.04698181e-01 9.23317790e-01
1.56384036e-01 6.46449387e-01 -1.17945045e-01 -3.30278844e-01
3.56136262e-01 -1.08975030e-01 5.23872912e-01 -5.70725858e-01
-4.32205975e-01 1.27795801e-01 -1.04517043e-01 -7.92167902e-01
-2.87871689e-01 -9.11437869e-01 -6.99625552e-01 -2.17641175e-01
-6.33975625e-01 2.15635821e-02 3.92020077e-01 8.52526546e-01
3.98108125e-01 3.88990879e-01 1.52212158e-01 -9.86587822e-01
-1.05499637e+00 -8.27135682e-01 -5.21507502e-01 4.10319269e-01
5.50118029e-01 -1.29194009e+00 -2.27512017e-01 1.86016083e-01] | [6.320186138153076, -2.0639238357543945] |
259bb9ff-170a-43a2-a0d8-c78dc01b640d | slicematch-geometry-guided-aggregation-for | 2211.14651 | null | https://arxiv.org/abs/2211.14651v3 | https://arxiv.org/pdf/2211.14651v3.pdf | SliceMatch: Geometry-guided Aggregation for Cross-View Pose Estimation | This work addresses cross-view camera pose estimation, i.e., determining the 3-Degrees-of-Freedom camera pose of a given ground-level image w.r.t. an aerial image of the local area. We propose SliceMatch, which consists of ground and aerial feature extractors, feature aggregators, and a pose predictor. The feature extractors extract dense features from the ground and aerial images. Given a set of candidate camera poses, the feature aggregators construct a single ground descriptor and a set of pose-dependent aerial descriptors. Notably, our novel aerial feature aggregator has a cross-view attention module for ground-view guided aerial feature selection and utilizes the geometric projection of the ground camera's viewing frustum on the aerial image to pool features. The efficient construction of aerial descriptors is achieved using precomputed masks. SliceMatch is trained using contrastive learning and pose estimation is formulated as a similarity comparison between the ground descriptor and the aerial descriptors. Compared to the state-of-the-art, SliceMatch achieves a 19% lower median localization error on the VIGOR benchmark using the same VGG16 backbone at 150 frames per second, and a 50% lower error when using a ResNet50 backbone. | ['Ted Lentsch', 'Julian F. P. Kooij', 'Holger Caesar', 'Zimin Xia'] | 2022-11-26 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Lentsch_SliceMatch_Geometry-Guided_Aggregation_for_Cross-View_Pose_Estimation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Lentsch_SliceMatch_Geometry-Guided_Aggregation_for_Cross-View_Pose_Estimation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['visual-localization'] | ['computer-vision'] | [ 1.73376486e-01 -2.58589447e-01 -1.10081926e-01 -8.00467432e-02
-1.00153327e+00 -1.12279987e+00 5.06707907e-01 -3.27212140e-02
-4.61860001e-01 -7.73672312e-02 3.71709391e-02 2.71754801e-01
-1.14295147e-01 -8.12781096e-01 -7.20200777e-01 -6.69363141e-01
-9.27639976e-02 1.94004849e-01 3.13305616e-01 -4.95890528e-02
3.06645066e-01 7.95379996e-01 -1.52771318e+00 -1.57156602e-01
3.49502563e-01 1.63126612e+00 3.17578077e-01 8.27535629e-01
6.85136735e-01 3.40078771e-01 -3.93807679e-01 -2.51117140e-01
8.38861406e-01 1.28873959e-01 -6.34476960e-01 6.75795496e-01
1.40583944e+00 -5.44807076e-01 -5.94009459e-01 7.81643331e-01
4.58212703e-01 1.58547655e-01 3.99199039e-01 -1.10773647e+00
-2.07861677e-01 -2.29730397e-01 -5.60211837e-01 3.50297540e-01
6.09435022e-01 2.24232420e-01 1.22529411e+00 -1.19560707e+00
7.64578462e-01 7.51488209e-01 6.48727477e-01 -4.07521054e-02
-1.10957348e+00 -4.66337740e-01 3.85344982e-01 -1.38877869e-01
-1.96674192e+00 -2.29043022e-01 6.48571610e-01 -5.17125905e-01
1.00768161e+00 2.15365857e-01 7.78787732e-01 5.12672186e-01
2.83850133e-01 4.29318696e-01 5.22959709e-01 -1.77344400e-02
3.11160721e-02 -2.77932882e-01 -3.65454316e-01 1.19168162e+00
6.54979516e-03 7.75786340e-02 -5.69646299e-01 -3.05029362e-01
8.48757803e-01 3.22767228e-01 -5.73430181e-01 -7.43193090e-01
-1.24944973e+00 7.15616763e-01 8.41447592e-01 -2.65309066e-01
-4.44515437e-01 2.20796794e-01 -6.04226142e-02 -2.29095854e-02
3.53361577e-01 6.82780504e-01 -6.47045016e-01 3.52404147e-01
-1.24385035e+00 1.58261523e-01 3.52506727e-01 1.26714480e+00
1.02824855e+00 -4.48815301e-02 7.49589130e-03 2.05099195e-01
2.18118355e-01 9.93776798e-01 -4.05452214e-02 -7.98210680e-01
7.44426608e-01 8.95832062e-01 8.65933299e-02 -1.33471966e+00
-3.67041796e-01 -5.98861456e-01 -4.23089504e-01 -1.71243530e-02
2.53673494e-01 -8.73533040e-02 -8.09858978e-01 1.34841526e+00
4.65552866e-01 -1.35530047e-02 -2.31063098e-01 1.10365760e+00
5.97219348e-01 5.92970192e-01 -3.28068912e-01 2.62756765e-01
1.28182244e+00 -9.54516590e-01 1.91286772e-01 -5.70717752e-01
5.06664455e-01 -8.55798542e-01 5.76406240e-01 4.13418800e-01
-1.03586423e+00 -5.91838896e-01 -1.12850118e+00 -7.54382014e-02
-3.52871805e-01 9.02188241e-01 2.78926730e-01 2.96997190e-01
-9.54530597e-01 4.91169512e-01 -7.02103019e-01 -3.13794583e-01
1.77754849e-01 7.19828606e-01 -7.42199957e-01 -2.86468565e-01
-7.10534632e-01 6.53022885e-01 1.07389390e-01 9.95587744e-03
-9.85154569e-01 -5.68246484e-01 -1.09618473e+00 5.07842973e-02
5.42642295e-01 -8.44532907e-01 7.28245735e-01 -8.41735363e-01
-1.01627445e+00 9.90154624e-01 1.67210340e-01 -2.18043566e-01
2.68847257e-01 -3.47797066e-01 -1.35045379e-01 5.60565770e-01
3.45773965e-01 5.29943883e-01 1.07909739e+00 -9.07454371e-01
-9.55878437e-01 -7.17592835e-01 2.48773366e-01 5.20854235e-01
-4.66801040e-02 -2.54022509e-01 -9.93055761e-01 -4.31003630e-01
4.78366405e-01 -1.36602950e+00 -1.64120674e-01 1.99910462e-01
-4.62908089e-01 2.73428500e-01 7.79456675e-01 -6.41259134e-01
1.13086677e+00 -2.20501351e+00 4.17580456e-01 4.03419197e-01
3.31144065e-01 -5.42957932e-02 -2.46805474e-01 5.21914005e-01
-8.07958003e-03 -2.05986261e-01 1.55636325e-01 -2.19263867e-01
-3.40337038e-01 -4.72407043e-01 -2.70967782e-01 1.04711628e+00
2.31187671e-01 5.73152423e-01 -9.06971097e-01 -3.08078319e-01
6.32715285e-01 4.49085653e-01 -7.39833415e-01 2.84005582e-01
1.41885057e-01 3.74637432e-02 -6.27246857e-01 9.36247945e-01
8.66624773e-01 -2.41145834e-01 -1.48608282e-01 -6.35405421e-01
-1.90761223e-01 -8.41218382e-02 -1.16523659e+00 1.90835607e+00
-5.53907871e-01 5.17339528e-01 -5.93292490e-02 -9.05048102e-02
9.68704879e-01 -1.42973617e-01 5.76857626e-01 -2.29052901e-02
3.17478836e-01 6.11737296e-02 -5.73120594e-01 -7.81662241e-02
7.23940492e-01 5.40514112e-01 -4.09678698e-01 -2.14007758e-02
3.98740292e-01 -4.56711888e-01 -3.45186070e-02 1.29586235e-01
1.35619318e+00 1.97116047e-01 4.77696717e-01 -1.17130853e-01
6.22370064e-01 1.33947330e-02 2.51694977e-01 2.35549629e-01
1.85679436e-01 1.03938687e+00 1.20248705e-01 -8.17334116e-01
-7.33871281e-01 -1.11403942e+00 2.33501047e-01 8.53398025e-01
5.62502623e-01 -8.30156565e-01 -7.38356411e-01 -9.45679545e-01
6.25458499e-03 9.56981704e-02 -6.64956808e-01 -1.40315786e-01
-3.36624265e-01 -1.02031827e-01 1.73374310e-01 4.95507836e-01
7.19291270e-01 -2.50382930e-01 -1.14075935e+00 -2.27785408e-01
-1.43715322e-01 -1.51159596e+00 -1.00582957e+00 2.75614679e-01
-6.72695518e-01 -1.41193080e+00 -3.41914237e-01 -5.52408338e-01
1.02985907e+00 9.03938532e-01 1.07220948e+00 5.10167144e-02
-2.56768942e-01 6.35696113e-01 -3.92746985e-01 1.63084701e-01
4.76754099e-01 4.63029772e-01 -1.91102666e-03 1.68779090e-01
1.11578271e-01 -2.57867008e-01 -9.26284730e-01 5.88364780e-01
-7.39735305e-01 -6.86892048e-02 7.45585442e-01 7.27112651e-01
1.01051557e+00 -1.88068524e-01 -5.64971566e-01 -1.44374952e-01
-1.78742245e-01 -3.46990168e-01 -1.11721587e+00 1.60031319e-01
-3.15123767e-01 -1.18404165e-01 7.62681365e-01 -1.28238365e-01
-1.93740353e-01 9.68174279e-01 3.65806967e-01 -1.04079545e+00
2.10605487e-02 1.99461743e-01 -3.36237282e-01 -5.26895761e-01
4.99859184e-01 2.89994359e-01 -3.61330032e-01 6.56218035e-03
2.38368243e-01 3.28106433e-01 4.99792486e-01 -2.21822262e-01
1.20219624e+00 6.34971380e-01 1.84659034e-01 -9.95034039e-01
-1.07240081e+00 -9.59710717e-01 -1.13799024e+00 -3.57967615e-01
1.12999713e+00 -1.38253915e+00 -5.52000999e-01 1.61948085e-01
-1.14496791e+00 -4.98796348e-03 -1.41486645e-01 4.93761867e-01
-6.41751468e-01 3.51289064e-02 8.58984217e-02 -3.14837575e-01
-3.41005623e-01 -1.32447922e+00 2.01593423e+00 1.37449875e-02
7.02010319e-02 -6.15651488e-01 -7.60491714e-02 4.16774631e-01
-1.15962863e-01 5.48257768e-01 1.77430347e-01 -2.95011550e-01
-1.21553767e+00 -7.18056738e-01 -1.24371499e-01 2.89621741e-01
-1.11108653e-01 7.92348087e-02 -1.00105643e+00 -5.98287344e-01
-3.26116443e-01 -1.48577169e-01 8.01459134e-01 4.06906754e-01
8.33224237e-01 -9.14907902e-02 -3.70991528e-01 1.34121978e+00
1.76261973e+00 -2.43513230e-02 3.58362108e-01 2.75336832e-01
9.11226988e-01 1.74642712e-01 9.51439202e-01 6.55715883e-01
2.87868440e-01 1.01148736e+00 9.94609535e-01 -2.55297631e-01
1.25487775e-01 -4.32626128e-01 4.54284877e-01 2.07482874e-01
-1.52421400e-01 -2.57713199e-01 -8.94941747e-01 4.72859621e-01
-1.42101812e+00 -7.47613668e-01 9.71663520e-02 2.37463951e+00
-2.96294168e-02 -2.18930826e-01 -3.80848311e-02 -2.30020508e-01
5.26425362e-01 4.93612170e-01 -5.89007378e-01 3.85701627e-01
-2.73755342e-02 -4.57068533e-02 1.22881627e+00 3.98443460e-01
-1.61663616e+00 1.09468126e+00 5.18465185e+00 5.66264808e-01
-1.13886023e+00 -3.05585235e-01 3.75749022e-01 -2.55815178e-01
3.76552731e-01 1.10479645e-01 -1.15664053e+00 2.34438524e-01
5.46363592e-01 1.63485348e-01 4.31903481e-01 1.19712341e+00
-1.53698012e-01 -2.17358723e-01 -1.01859307e+00 1.20686376e+00
6.08629644e-01 -1.40185261e+00 6.42434359e-02 5.16232193e-01
7.65976608e-01 4.43902373e-01 1.51623085e-01 -1.75870076e-01
-3.91502529e-02 -7.74450600e-01 1.01625228e+00 3.46292555e-01
9.23105657e-01 -9.63386595e-01 7.53309011e-01 9.17345732e-02
-1.76659310e+00 -3.48420024e-01 -4.30452764e-01 2.45162711e-01
-1.88349217e-01 2.44046152e-01 -6.85673475e-01 7.52026200e-01
6.65393114e-01 9.06441987e-01 -9.89737332e-01 9.33952749e-01
-2.06346110e-01 1.11369647e-01 -4.19270277e-01 4.48075563e-01
5.85101187e-01 -1.25438824e-01 5.23583829e-01 8.43201876e-01
6.10622823e-01 -3.49351056e-02 7.16582358e-01 5.21420717e-01
-6.24421425e-02 5.06534874e-02 -8.69567037e-01 3.19655716e-01
5.06184220e-01 1.67505288e+00 -8.60900998e-01 5.22927083e-02
-4.25264329e-01 1.24894917e+00 9.54808369e-02 5.96806258e-02
-9.34237540e-01 -2.34177068e-01 7.78570473e-01 5.63498378e-01
7.36694336e-01 -4.11424607e-01 3.70613188e-01 -1.29679036e+00
3.29249451e-04 -6.92833126e-01 1.82349980e-01 -9.99547005e-01
-7.47797906e-01 1.04498291e+00 9.11247656e-02 -1.82859111e+00
-1.42894447e-01 -7.25048661e-01 -4.11911458e-01 7.63424218e-01
-1.17595696e+00 -1.59648025e+00 -1.02656841e+00 6.91163778e-01
6.03147805e-01 -2.80352235e-01 7.92850971e-01 5.71117736e-02
-3.16009223e-01 3.75774890e-01 -1.75630555e-01 3.17410380e-01
4.65782136e-01 -1.07318163e+00 6.99789762e-01 1.06845796e+00
3.47876877e-01 3.66619051e-01 2.65663058e-01 -5.60319483e-01
-1.88420904e+00 -1.65317988e+00 4.57974434e-01 -8.16316962e-01
5.23997724e-01 -4.94914263e-01 -1.13849327e-01 7.14897990e-01
-6.76339343e-02 5.98280609e-01 4.44865346e-01 -3.28463554e-01
-3.01461697e-01 -2.05790788e-01 -8.48938048e-01 3.27188075e-01
9.74666715e-01 -5.78855753e-01 -2.51013469e-02 3.40374887e-01
4.23386067e-01 -7.71814406e-01 -9.20689225e-01 6.55138493e-02
5.12737989e-01 -8.93662512e-01 1.24304414e+00 -4.46064025e-02
2.52053350e-01 -7.28827775e-01 -6.98394537e-01 -1.23667300e+00
-5.12687385e-01 -5.35219610e-01 1.61228836e-01 8.36038828e-01
1.26289725e-01 -1.15875334e-01 8.53949249e-01 7.31286332e-02
-1.56401902e-01 -9.24372971e-01 -7.79827237e-01 -9.17474091e-01
-5.99328101e-01 -1.25842169e-01 4.47842747e-01 4.91393209e-01
-5.38200438e-01 5.21046460e-01 -2.35526755e-01 8.07864130e-01
5.23717225e-01 3.02655965e-01 1.01870310e+00 -1.19247007e+00
-1.51313156e-01 -1.46435080e-02 -9.81705487e-01 -1.22103393e+00
6.31165802e-02 -8.65861297e-01 -6.20335490e-02 -1.34330761e+00
-4.01827991e-02 1.08241523e-02 1.21733621e-01 2.31011853e-01
1.16018847e-01 4.72669870e-01 4.74032134e-01 1.18132450e-01
-8.21336150e-01 4.65884596e-01 9.44057047e-01 -1.81269586e-01
-1.32167011e-01 -3.89871304e-03 -4.30562407e-01 9.12962556e-01
5.79600275e-01 -4.05581743e-01 -4.44255590e-01 -5.24910092e-01
1.88075066e-01 7.36526549e-02 6.26072943e-01 -1.41756582e+00
3.29144269e-01 -4.41707596e-02 7.09521651e-01 -8.76560748e-01
8.34216475e-01 -1.03336787e+00 1.45183280e-01 2.61827350e-01
-1.66441966e-02 6.34738982e-01 6.73867688e-02 6.58572853e-01
-1.38568506e-01 7.48855099e-02 7.52749443e-01 -1.92035899e-01
-7.85578489e-01 7.59530902e-01 5.64464144e-02 2.81721186e-02
1.21286249e+00 -4.66687858e-01 -1.51539713e-01 -1.52425408e-01
-3.85567009e-01 5.48364818e-02 9.82228577e-01 2.24043652e-01
9.49405253e-01 -1.16008210e+00 -4.63299721e-01 5.24738193e-01
4.63667810e-01 2.28277728e-01 1.66262865e-01 8.35136712e-01
-6.92066669e-01 4.73724127e-01 -8.43138024e-02 -6.80834353e-01
-1.57811999e+00 4.19038653e-01 5.01425028e-01 -2.91885018e-01
-3.64402235e-01 9.01497126e-01 4.32663113e-01 -1.86526269e-01
1.48605183e-02 -6.17090523e-01 1.18385680e-01 1.54887393e-01
4.42025065e-01 2.92494893e-01 2.93704838e-01 -1.23303854e+00
-5.99189460e-01 1.21766710e+00 1.43778324e-01 8.72603878e-02
1.22092831e+00 -1.63555071e-01 2.32779041e-01 -9.16461796e-02
1.44362175e+00 2.28437364e-01 -1.62566686e+00 -1.35399953e-01
-4.04066414e-01 -8.79973173e-01 5.72309792e-01 -5.42431355e-01
-1.31627548e+00 6.30438209e-01 7.16879845e-01 -1.46212026e-01
1.28910983e+00 3.25288810e-02 5.57098925e-01 4.66751695e-01
6.86015069e-01 -8.36788356e-01 1.27463803e-01 4.82073963e-01
1.00467324e+00 -1.21674502e+00 4.79732573e-01 -4.40798253e-01
-5.70817947e-01 1.12663901e+00 6.40193045e-01 -5.74488103e-01
4.86569703e-01 1.36784418e-02 -2.02295139e-01 -6.01970077e-01
-5.35195708e-01 -4.59846467e-01 7.17636228e-01 4.97922003e-01
-2.39131358e-02 -1.13063507e-01 6.25313878e-01 2.58337140e-01
-3.19679022e-01 -4.50998515e-01 2.89140403e-01 7.12625563e-01
-3.15453053e-01 -5.24765432e-01 -4.21101749e-01 3.47344607e-01
-1.38574079e-01 -2.12339342e-01 -6.45019531e-01 9.09107566e-01
4.15319741e-01 7.41121113e-01 7.50808045e-02 -1.00360215e+00
6.46938920e-01 -4.15470183e-01 4.74333823e-01 -8.08423102e-01
-7.64743865e-01 3.05241030e-02 -1.31342739e-01 -1.07281792e+00
-2.03369811e-01 -7.50815034e-01 -7.20064521e-01 -5.64585365e-02
-6.19192302e-01 -1.77800238e-01 6.41748726e-01 5.74427187e-01
5.59594572e-01 1.44346610e-01 1.02879369e+00 -1.43328822e+00
-3.30103964e-01 -4.27453399e-01 -6.75705075e-01 -5.02755307e-02
4.29554701e-01 -7.99709976e-01 -3.95999819e-01 1.60300180e-01] | [7.634697914123535, -2.249546527862549] |
71855a5e-9711-42c4-9982-3ecf81c3cba3 | conaclip-exploring-distillation-of-fully | 2305.17652 | null | https://arxiv.org/abs/2305.17652v1 | https://arxiv.org/pdf/2305.17652v1.pdf | ConaCLIP: Exploring Distillation of Fully-Connected Knowledge Interaction Graph for Lightweight Text-Image Retrieval | Large-scale pre-trained text-image models with dual-encoder architectures (such as CLIP) are typically adopted for various vision-language applications, including text-image retrieval. However,these models are still less practical on edge devices or for real-time situations, due to the substantial indexing and inference time and the large consumption of computational resources. Although knowledge distillation techniques have been widely utilized for uni-modal model compression, how to expand them to the situation when the numbers of modalities and teachers/students are doubled has been rarely studied. In this paper, we conduct comprehensive experiments on this topic and propose the fully-Connected knowledge interaction graph (Cona) technique for cross-modal pre-training distillation. Based on our findings, the resulting ConaCLIP achieves SOTA performances on the widely-used Flickr30K and MSCOCO benchmarks under the lightweight setting. An industry application of our method on an e-commercial platform further demonstrates the significant effectiveness of ConaCLIP. | ['Lianwen Jin', 'Jun Huang', 'Xiaodan Wang', 'Chengyu Wang', 'Jiapeng Wang'] | 2023-05-28 | null | null | null | null | ['model-compression'] | ['methodology'] | [ 3.68473530e-01 -1.63188860e-01 -5.61067760e-01 -7.66711310e-02
-7.51528800e-01 -1.44739211e-01 7.95077801e-01 -1.71744198e-01
-5.34083247e-01 3.79871488e-01 6.89182356e-02 -5.31752527e-01
-1.44615084e-01 -4.70240891e-01 -7.68825412e-01 -5.88992834e-01
3.12565714e-01 5.37832558e-01 1.06274724e-01 1.17970824e-01
-1.99341729e-01 5.11350259e-02 -1.84658396e+00 3.41256171e-01
7.34526396e-01 8.60994220e-01 3.11575979e-01 7.37288237e-01
2.67908666e-02 1.00779152e+00 -1.50334939e-01 -8.55992436e-01
1.37169763e-01 -2.01312572e-01 -8.83994579e-01 2.56066173e-01
9.23286080e-01 -6.96250558e-01 -8.98371041e-01 9.81323898e-01
7.13995695e-01 1.55289248e-01 4.92520124e-01 -1.17919230e+00
-6.36445880e-01 7.65192926e-01 -8.78507316e-01 7.73898363e-02
1.03736324e-02 -1.54016837e-01 9.36445773e-01 -8.48619282e-01
6.32160783e-01 1.19156837e+00 2.50819981e-01 3.82194370e-01
-9.46791828e-01 -7.83479989e-01 4.65258546e-02 1.01168382e+00
-1.70390868e+00 -5.07618666e-01 5.87701917e-01 -1.88772734e-02
9.67664182e-01 1.77819625e-01 6.29981637e-01 1.18228734e+00
-1.92945883e-01 1.24078465e+00 9.60529447e-01 -7.74361610e-01
-1.05451904e-01 2.72947252e-01 -4.54654917e-02 8.19696724e-01
2.17104897e-01 -3.78364116e-01 -7.72359133e-01 1.96658328e-01
4.80825126e-01 1.88460816e-02 -2.79741943e-01 -1.68162152e-01
-1.10669780e+00 6.39953375e-01 1.32000804e-01 3.36316198e-01
-9.66792181e-02 3.17107916e-01 6.35861516e-01 2.27783203e-01
4.11030143e-01 -1.71215564e-01 -1.54200390e-01 -3.74176860e-01
-1.24846458e+00 -5.38186841e-02 7.37419486e-01 1.27659559e+00
5.10887682e-01 -1.13107495e-01 -6.92002401e-02 1.01035738e+00
2.62062907e-01 6.87674046e-01 4.63679940e-01 -8.67368519e-01
4.43848252e-01 5.01190484e-01 -5.08246720e-01 -8.84696960e-01
4.39461134e-02 -3.64902318e-01 -1.19156241e+00 -4.43674058e-01
7.10432157e-02 1.95855603e-01 -1.07695889e+00 1.40373218e+00
3.63676459e-01 6.36644483e-01 4.47721817e-02 7.36378372e-01
1.02742982e+00 5.32655418e-01 6.96170554e-02 -9.02190953e-02
1.55713522e+00 -1.28341115e+00 -8.63718629e-01 -2.10689887e-01
8.19509327e-01 -9.23890233e-01 1.11539328e+00 5.08354127e-01
-1.04609489e+00 -5.22789180e-01 -7.90169001e-01 -5.21806538e-01
-5.01006007e-01 3.87452722e-01 8.05878341e-01 4.05490488e-01
-1.07993042e+00 1.09603062e-01 -7.77041197e-01 -4.47201639e-01
6.00064397e-01 1.98902965e-01 -3.22866648e-01 -7.17901766e-01
-1.14954770e+00 7.38539875e-01 5.10174155e-01 5.61019033e-02
-8.12750459e-01 -6.53868735e-01 -6.12886846e-01 2.32861966e-01
6.03431940e-01 -7.47783244e-01 1.13924658e+00 -6.23149693e-01
-1.52047694e+00 9.52947795e-01 1.89896356e-02 -5.82971275e-01
4.26606119e-01 -4.51021105e-01 -4.01210517e-01 4.13169146e-01
-3.77106607e-01 9.99323010e-01 9.75239217e-01 -9.72577512e-01
-5.73488235e-01 -3.28717977e-01 4.82930779e-01 5.77321827e-01
-1.01566184e+00 -2.03010216e-01 -1.34351325e+00 -5.36014557e-01
-3.20181578e-01 -1.09915972e+00 -7.35302120e-02 -6.13939688e-02
-4.50808257e-01 -3.09464246e-01 1.32486486e+00 -6.05439544e-01
1.52210712e+00 -2.22443008e+00 1.36137143e-01 -5.35224266e-02
2.41030887e-01 5.11373758e-01 -1.36928782e-01 4.12078321e-01
1.45289958e-01 -1.59874618e-01 1.71001107e-01 -6.00692570e-01
-2.47399621e-02 3.31433475e-01 -3.31331789e-01 2.18969479e-01
-4.00944501e-01 1.15806723e+00 -6.24402881e-01 -1.06040180e+00
5.48423231e-01 8.88650596e-01 -6.33369029e-01 1.63709223e-02
-2.82113194e-01 -8.24872181e-02 -2.60764003e-01 6.43340170e-01
4.97658223e-01 -6.86759412e-01 1.95779741e-01 -5.11159360e-01
2.40759060e-01 -1.39706001e-01 -1.06264770e+00 2.17479157e+00
-6.00314915e-01 8.98558378e-01 -2.26320419e-02 -1.01997554e+00
1.03162028e-01 3.00476462e-01 4.55830276e-01 -8.31596434e-01
1.48850486e-01 -6.12738431e-02 -2.85813212e-01 -4.94643062e-01
8.00273299e-01 2.95009941e-01 3.57450962e-01 3.38590294e-01
2.22565919e-01 -1.51935056e-01 5.14642358e-01 6.95642531e-01
8.72643828e-01 8.32541510e-02 -6.01034276e-02 5.92377260e-02
4.23429102e-01 -1.44699574e-01 -1.45482585e-01 7.04907238e-01
1.61784645e-02 2.59267956e-01 5.76349124e-02 -8.89840275e-02
-9.07918632e-01 -5.24841130e-01 -2.70375103e-01 1.20910084e+00
1.66119561e-01 -8.05455506e-01 -7.95965731e-01 -7.27063477e-01
-2.91643620e-01 6.88765407e-01 -1.90439343e-01 -2.49726847e-01
-2.12952226e-01 -5.59882522e-01 7.35773742e-01 3.78218323e-01
8.47147167e-01 -6.20680273e-01 -4.54618335e-01 -5.28336093e-02
-3.53642315e-01 -1.76679432e+00 -4.65850353e-01 -1.95844859e-01
-7.08712935e-01 -1.02566218e+00 -7.82234848e-01 -7.94323921e-01
5.19525647e-01 7.72364855e-01 1.03859496e+00 8.94021317e-02
-4.71962363e-01 9.07140911e-01 -3.83291125e-01 -3.33844185e-01
-5.57110682e-02 4.74145263e-02 -1.53394699e-01 -6.01180363e-03
5.91753006e-01 -4.56545979e-01 -8.44615757e-01 9.45504084e-02
-1.41484261e+00 5.85773587e-01 7.30332911e-01 8.74488652e-01
7.58204699e-01 4.17437255e-01 2.54517615e-01 -8.38132203e-01
3.10201287e-01 -1.72209635e-01 -5.18727303e-01 6.62479818e-01
-7.88228869e-01 -8.12294260e-02 3.30043674e-01 -6.65005207e-01
-1.07707286e+00 -9.01986659e-02 4.71114740e-02 -8.58735740e-01
-9.13703721e-03 8.50926936e-01 7.23093078e-02 -1.71620920e-01
9.79953036e-02 5.27894974e-01 -2.30236471e-01 -2.84331053e-01
7.86903203e-01 9.68113720e-01 6.23079896e-01 -5.06782949e-01
5.90125799e-01 6.24553561e-01 2.00086430e-01 -1.15387118e+00
-9.43620563e-01 -4.97597039e-01 -3.84237111e-01 -1.89480096e-01
8.15913916e-01 -1.48293376e+00 -6.90470159e-01 3.84217978e-01
-8.86187255e-01 -3.52509439e-01 -2.09737778e-01 6.39183700e-01
-4.51056957e-01 8.34113002e-01 -5.56921601e-01 -4.70781773e-01
-5.17350554e-01 -1.30522335e+00 1.29296064e+00 2.83360362e-01
4.35838610e-01 -1.15160668e+00 -1.79563805e-01 9.82128084e-01
4.00686592e-01 -3.77571195e-01 9.03040648e-01 -4.05163080e-01
-7.46466279e-01 -1.99787825e-01 -4.66877550e-01 3.32100064e-01
-2.87502885e-01 -6.66235164e-02 -1.09278417e+00 -4.72350180e-01
-5.53578079e-01 -9.80223835e-01 9.39578354e-01 2.66735286e-01
1.77720535e+00 -9.62875858e-02 -3.55145633e-01 5.59180558e-01
1.36816740e+00 -2.59567589e-01 6.25750005e-01 -4.75185700e-02
9.26823437e-01 2.93929100e-01 4.01380718e-01 3.22828770e-01
7.41954684e-01 8.24692845e-01 3.89720470e-01 -6.28190786e-02
-5.55345893e-01 -3.66680026e-01 2.89188057e-01 1.52153850e+00
-1.12057477e-01 -5.36367297e-01 -8.81102026e-01 7.45097101e-01
-1.79636502e+00 -8.50911677e-01 3.82759538e-03 1.98666036e+00
1.05054855e+00 -4.34727147e-02 -4.04418975e-01 -5.34029270e-04
4.18410540e-01 8.30754191e-02 -5.16452730e-01 5.10898344e-02
-2.00334832e-01 1.91682458e-01 5.22936821e-01 1.28408253e-01
-1.01011503e+00 9.22182143e-01 5.28834629e+00 1.68133163e+00
-9.99856651e-01 3.03495497e-01 6.66840136e-01 -2.98489839e-01
-1.13131270e-01 -9.57540572e-02 -9.73353028e-01 1.83575436e-01
1.13498497e+00 -1.92898020e-01 5.31022549e-01 7.37089872e-01
-2.91269422e-01 -2.61598915e-01 -1.17870653e+00 1.58024704e+00
3.42429847e-01 -1.31014562e+00 3.95465463e-01 1.14049077e-01
9.16739106e-01 1.83039203e-01 2.55152166e-01 6.67229533e-01
1.02063321e-01 -8.98564696e-01 2.22977296e-01 1.63033634e-01
1.31995511e+00 -6.54872179e-01 5.66598058e-01 3.50187331e-01
-1.23315501e+00 1.56834379e-01 -3.32406372e-01 4.58172023e-01
9.37177241e-02 6.39555693e-01 -6.35097086e-01 7.81945109e-01
7.76780307e-01 7.17343986e-01 -7.32967734e-01 6.85796142e-01
1.70359425e-02 8.64334106e-01 -6.27206624e-01 3.52976441e-01
1.82496533e-01 1.85482483e-02 1.14944525e-01 1.27652264e+00
2.10149139e-01 -4.47506197e-02 7.72196352e-02 2.51889795e-01
-4.79521573e-01 2.07145780e-01 -6.79005623e-01 -3.72928888e-01
3.69990706e-01 1.28144848e+00 -6.14717364e-01 -5.79964578e-01
-9.80044365e-01 1.01037014e+00 2.36424550e-01 5.22852778e-01
-1.07762933e+00 -1.94298968e-01 4.03182566e-01 4.33868095e-02
4.04863536e-01 -2.31474504e-01 1.48347422e-01 -1.48530686e+00
-5.90238795e-02 -1.02468228e+00 5.50303459e-01 -8.84908855e-01
-9.15326655e-01 3.59197646e-01 3.21468532e-01 -1.01745522e+00
-6.92716837e-02 -4.67354387e-01 -1.03443405e-02 2.80929387e-01
-1.91863286e+00 -1.60879552e+00 -3.46155524e-01 9.50074315e-01
6.84565902e-01 -7.28834718e-02 8.73461366e-01 8.46255779e-01
-7.28788555e-01 9.50420320e-01 3.39375496e-01 4.90676053e-02
7.02277601e-01 -8.50714505e-01 -1.11920781e-01 7.67467320e-01
6.16778672e-01 3.60012382e-01 3.17140043e-01 -3.34588766e-01
-1.95499218e+00 -1.06393623e+00 6.53513312e-01 -1.96520403e-01
6.97028637e-01 -3.29984605e-01 -8.10863435e-01 7.67761886e-01
6.99684501e-01 1.58473670e-01 7.48551011e-01 -3.06926761e-02
-3.83044899e-01 -1.03832453e-01 -6.37305379e-01 8.97642374e-01
1.02982998e+00 -7.76417792e-01 -1.75672531e-01 6.49819195e-01
7.81910300e-01 -5.15765667e-01 -1.06343818e+00 5.27657032e-01
4.73526388e-01 -6.53510392e-01 9.74509597e-01 -3.67269129e-01
7.04693615e-01 2.30794419e-02 -2.08348155e-01 -9.27227974e-01
-2.66861916e-02 -4.40892547e-01 -7.50200987e-01 1.17247307e+00
-3.17479111e-02 -7.11017847e-02 8.44896674e-01 4.73498404e-01
1.16334461e-01 -9.19957519e-01 -1.12467766e+00 -4.60263401e-01
-2.65670478e-01 -6.59609020e-01 2.64768839e-01 9.87788141e-01
-1.10374741e-01 5.56657970e-01 -5.98345757e-01 -9.17420238e-02
7.89453626e-01 1.74158528e-01 8.50044489e-01 -8.70891511e-01
-4.55900699e-01 -2.70662487e-01 -3.30285996e-01 -1.51328456e+00
2.49132380e-01 -9.05027866e-01 -3.47948372e-01 -1.41909468e+00
6.82476997e-01 -2.83713341e-01 -1.05123580e-01 5.48223913e-01
-8.83901343e-02 4.14435863e-01 2.20951870e-01 1.10683218e-01
-1.15873730e+00 7.32351303e-01 1.33910620e+00 -4.49213833e-01
2.20387563e-01 -4.65860724e-01 -3.56968045e-01 6.07683480e-01
3.55161577e-01 -7.94619396e-02 -9.01763916e-01 -7.59720325e-01
4.06973213e-01 2.75523197e-02 3.13851058e-01 -9.04746771e-01
5.60311854e-01 -1.01200899e-03 -5.34034185e-02 -6.83939874e-01
6.29768670e-01 -1.30682075e+00 9.73100588e-02 1.76798195e-01
-2.16272637e-01 -1.85224816e-01 3.42424482e-01 6.40312791e-01
-5.09702682e-01 -6.49590418e-02 6.07338786e-01 2.09895477e-01
-8.30948174e-01 4.63617504e-01 6.98414519e-02 1.26331031e-01
1.16260839e+00 -7.13508129e-02 -5.93539119e-01 -6.64407134e-01
-4.09890115e-01 3.32612962e-01 1.34848684e-01 3.99584144e-01
7.83059597e-01 -1.33020651e+00 -6.89747572e-01 2.72574127e-01
3.17611396e-01 1.34793729e-01 7.29962289e-01 1.10920918e+00
-3.93742889e-01 7.93374717e-01 1.48342267e-01 -7.60436416e-01
-1.65183556e+00 5.71693897e-01 -7.55566210e-02 -5.13044894e-01
-7.37324953e-01 8.08032632e-01 1.83252424e-01 3.03143542e-02
5.72816133e-01 -1.00909621e-01 -1.10974126e-01 5.16122878e-02
6.24650538e-01 4.34200615e-01 1.81087494e-01 -6.54433310e-01
1.27190957e-02 5.01407266e-01 -3.83460283e-01 1.22784361e-01
1.15616310e+00 -4.28882480e-01 -4.84625669e-03 9.83263254e-02
1.14058018e+00 -3.02595586e-01 -9.66161668e-01 -6.82787299e-01
-3.73645902e-01 -3.75516087e-01 5.32422066e-01 -4.83117878e-01
-1.35825896e+00 1.13401163e+00 6.08060896e-01 -9.20404494e-02
1.47496414e+00 9.40792356e-03 9.83265102e-01 6.93507135e-01
4.76749390e-01 -1.21382010e+00 1.10368446e-01 5.38183093e-01
4.64357704e-01 -1.24102545e+00 2.32132733e-01 -2.64080644e-01
-5.89901447e-01 8.48171592e-01 6.10716760e-01 6.00924551e-01
8.07904780e-01 9.75559652e-02 -2.25249872e-01 -2.09039137e-01
-1.01045954e+00 -2.39446878e-01 4.40609396e-01 2.99380124e-01
3.67643803e-01 5.80358028e-04 -1.51890874e-01 7.94715881e-02
-1.43453060e-02 3.11890721e-01 4.26569045e-01 8.07536364e-01
5.66770993e-02 -9.00049627e-01 -6.03957027e-02 5.89408040e-01
-6.31815016e-01 -2.82208681e-01 3.40552852e-02 6.54635489e-01
1.55160166e-02 9.23112690e-01 -8.37599188e-02 -4.59437281e-01
1.24530964e-01 1.48740709e-01 6.77482665e-01 -4.09687370e-01
-2.34976739e-01 1.72063500e-01 -3.23778093e-02 -5.28577924e-01
-8.57320249e-01 -2.26231307e-01 -9.55022216e-01 -5.86651206e-01
-6.16119444e-01 -8.20615441e-02 8.37565601e-01 7.95198500e-01
6.19319975e-01 5.36347210e-01 9.60407630e-02 -6.10458910e-01
-4.42846686e-01 -1.02952504e+00 -4.40884948e-01 4.91894037e-01
-2.15434685e-01 -5.21498322e-01 -5.03124073e-02 3.82622659e-01] | [10.24763011932373, 0.9905239939689636] |
6df25ddc-407f-4cd8-b7fc-2c9243bee122 | generating-virtual-on-body-accelerometer-data | 2305.03187 | null | https://arxiv.org/abs/2305.03187v1 | https://arxiv.org/pdf/2305.03187v1.pdf | Generating Virtual On-body Accelerometer Data from Virtual Textual Descriptions for Human Activity Recognition | The development of robust, generalized models in human activity recognition (HAR) has been hindered by the scarcity of large-scale, labeled data sets. Recent work has shown that virtual IMU data extracted from videos using computer vision techniques can lead to substantial performance improvements when training HAR models combined with small portions of real IMU data. Inspired by recent advances in motion synthesis from textual descriptions and connecting Large Language Models (LLMs) to various AI models, we introduce an automated pipeline that first uses ChatGPT to generate diverse textual descriptions of activities. These textual descriptions are then used to generate 3D human motion sequences via a motion synthesis model, T2M-GPT, and later converted to streams of virtual IMU data. We benchmarked our approach on three HAR datasets (RealWorld, PAMAP2, and USC-HAD) and demonstrate that the use of virtual IMU training data generated using our new approach leads to significantly improved HAR model performance compared to only using real IMU data. Our approach contributes to the growing field of cross-modality transfer methods and illustrate how HAR models can be improved through the generation of virtual training data that do not require any manual effort. | ['Thomas Plötz', 'Hyeokhyen Kwon', 'Zikang Leng'] | 2023-05-04 | null | null | null | null | ['motion-synthesis', 'human-activity-recognition', 'human-activity-recognition'] | ['computer-vision', 'computer-vision', 'time-series'] | [ 3.89644861e-01 2.50441842e-02 -5.69823943e-02 2.05604844e-02
-1.00983191e+00 -5.32600999e-01 1.19045079e+00 -4.51445997e-01
-6.23983562e-01 8.42701674e-01 6.52633131e-01 7.71073550e-02
4.21396703e-01 -4.44553018e-01 -8.88354778e-01 -2.02225491e-01
1.27113327e-01 6.29534602e-01 3.53039652e-01 -1.83875978e-01
2.29683835e-02 2.07419485e-01 -1.49426377e+00 5.47117591e-01
2.06688628e-01 5.86061597e-01 2.95263320e-01 1.11209714e+00
-2.56259665e-02 1.34962296e+00 -6.28032565e-01 1.08283550e-01
1.63474441e-01 -9.24842656e-01 -8.74276876e-01 2.98035234e-01
5.69122493e-01 -5.07835686e-01 -6.12140357e-01 2.44325876e-01
6.73824012e-01 4.50649679e-01 8.38515937e-01 -1.47556257e+00
-5.12058020e-01 3.11553299e-01 -7.56074712e-02 1.59742206e-01
9.47421551e-01 4.46130544e-01 5.73387861e-01 -7.72595108e-01
1.43045175e+00 1.16781938e+00 6.79897010e-01 8.74977052e-01
-1.33869696e+00 -3.02606434e-01 -3.97352189e-01 3.83454412e-01
-1.20779908e+00 -5.25456131e-01 5.85761309e-01 -6.05140746e-01
1.43373680e+00 2.06208043e-02 9.20189083e-01 1.94824970e+00
-6.92270771e-02 1.22442245e+00 6.85855150e-01 -4.60512906e-01
3.57525945e-02 -6.40756935e-02 -3.23203385e-01 6.63503587e-01
-2.74255931e-01 4.31319792e-03 -9.51914787e-01 1.80928614e-02
9.59642112e-01 -4.27828521e-01 -3.56816351e-01 -4.27200556e-01
-2.03677034e+00 6.01571918e-01 -2.28189379e-02 3.59298259e-01
-4.45952922e-01 5.10943890e-01 4.53348756e-01 1.82850435e-01
1.17577516e-01 4.11129355e-01 -1.78168058e-01 -7.56662786e-01
-1.05814886e+00 5.54180622e-01 8.47602904e-01 1.06831539e+00
5.05086541e-01 2.93100864e-01 -2.90588260e-01 7.35189855e-01
5.98454997e-02 7.31692791e-01 8.58299613e-01 -1.37034309e+00
6.47784710e-01 1.08661942e-01 2.75420129e-01 -7.92931199e-01
-2.64315456e-01 1.87495500e-01 -2.14699820e-01 -1.44307464e-01
4.94751483e-01 -1.27832815e-01 -9.52663243e-01 1.69142342e+00
9.08658747e-03 4.56500620e-01 2.63365746e-01 8.33875775e-01
8.60138834e-01 8.06000590e-01 2.25777090e-01 1.43497869e-01
1.00515258e+00 -1.13868558e+00 -5.65326214e-01 -3.84233832e-01
9.70188558e-01 -5.30235827e-01 8.19285214e-01 1.58257082e-01
-1.17078781e+00 -7.54499674e-01 -9.60596621e-01 -7.74589702e-02
-2.38426492e-01 -5.09874746e-02 2.84397542e-01 2.70507097e-01
-1.14884603e+00 6.14826500e-01 -9.25176263e-01 -9.54133034e-01
3.55383933e-01 9.53341275e-02 -8.13485861e-01 -8.60989988e-02
-1.06793463e+00 9.90157306e-01 6.70290828e-01 -1.95293158e-01
-1.11276817e+00 -3.38281870e-01 -1.30598223e+00 -4.39388424e-01
2.84711659e-01 -1.08215952e+00 1.51325607e+00 -9.43000495e-01
-1.48108387e+00 6.23948574e-01 -6.53389543e-02 -5.30549049e-01
7.25374043e-01 -1.93436563e-01 -1.77133322e-01 4.66480255e-01
1.33867458e-01 1.24366522e+00 7.47304261e-01 -1.10379612e+00
-5.49003124e-01 1.55042782e-01 -3.03751349e-01 3.40567648e-01
-2.08915085e-01 2.40096878e-02 -5.90458393e-01 -6.78159416e-01
-4.63362485e-01 -1.22657228e+00 -7.10831508e-02 -3.27651411e-01
-4.93974648e-02 8.52227584e-02 8.21050048e-01 -9.72488463e-01
8.96949470e-01 -1.72008538e+00 6.82126880e-01 -4.87086847e-02
-1.49491489e-01 3.30800086e-01 -4.71723884e-01 6.81044698e-01
2.00946257e-01 -9.05463472e-02 -2.91795224e-01 -3.60656023e-01
1.82696357e-01 3.85500431e-01 -6.29642159e-02 9.12037715e-02
2.04505309e-01 1.33501506e+00 -1.15747511e+00 -7.09743738e-01
5.14006436e-01 6.25356972e-01 -5.29880047e-01 4.24307019e-01
-2.55666912e-01 6.78402901e-01 -1.33058414e-01 3.92393470e-01
-2.41494030e-01 -2.76100159e-01 6.95179030e-02 -3.25154699e-02
1.22260608e-01 -4.96586133e-03 -9.74866211e-01 2.27558708e+00
-3.12861383e-01 9.93880212e-01 -5.65857708e-01 -8.33831906e-01
5.48466682e-01 6.98259771e-01 6.48890913e-01 -5.97709835e-01
1.84017643e-01 4.29725684e-02 -4.08781230e-01 -7.23582327e-01
6.96355343e-01 2.58751330e-03 -2.78116971e-01 5.70673525e-01
6.24048650e-01 -3.56889606e-01 4.12584037e-01 2.02517942e-01
1.32499242e+00 1.06563985e+00 3.08202207e-01 4.76777405e-01
3.75710875e-01 5.65894842e-01 4.19206142e-01 8.30143094e-01
-3.56603533e-01 1.04161489e+00 7.03463405e-02 -1.77957103e-01
-1.50428379e+00 -1.04758751e+00 3.64085287e-01 8.78332078e-01
-3.93796206e-01 -5.66482544e-01 -9.53134716e-01 -6.60879433e-01
-2.34134525e-01 7.68044949e-01 -6.48555100e-01 4.88417000e-02
-8.10990393e-01 -4.38571513e-01 9.74411249e-01 8.95973325e-01
6.31692410e-01 -1.54794085e+00 -8.87494266e-01 3.64869565e-01
-6.54842198e-01 -1.63564527e+00 -5.39853275e-01 -4.52989012e-01
-7.75130570e-01 -8.48358750e-01 -1.22317004e+00 -6.94602847e-01
3.73367697e-01 1.60981610e-01 1.10335302e+00 -2.39059567e-01
-3.61210316e-01 1.05906260e+00 -6.52432442e-01 -2.86039740e-01
-8.13035429e-01 2.48501841e-02 2.26954445e-01 -1.76193014e-01
2.80984908e-01 -4.11663741e-01 -2.22910270e-01 2.05446765e-01
-8.89132857e-01 5.71101606e-01 5.33245146e-01 7.91572332e-01
4.07283604e-01 -8.91556799e-01 4.73650217e-01 -5.35017610e-01
3.11475277e-01 -6.96336687e-01 1.51159791e-02 4.63243760e-02
-4.45670635e-02 1.58877075e-01 3.16150635e-01 -7.57929385e-01
-1.22856414e+00 3.52084011e-01 6.04325086e-02 -8.01930130e-01
-4.26769227e-01 5.61866045e-01 2.58574814e-01 5.11827469e-02
9.13802564e-01 4.36542481e-01 3.79812568e-02 -2.62693524e-01
5.93942046e-01 6.67017996e-01 9.38849211e-01 -3.84165525e-01
7.53093362e-01 3.79809082e-01 -7.58167207e-02 -1.23475683e+00
-2.00709790e-01 -2.88895875e-01 -9.16325629e-01 -4.47224021e-01
1.24689162e+00 -8.81028831e-01 -1.33946776e-01 4.69111353e-01
-1.11549556e+00 -8.12358081e-01 -2.16156840e-01 6.86624646e-01
-1.27383316e+00 5.65692008e-01 -7.78874397e-01 -6.36400700e-01
-1.38362899e-01 -8.46453905e-01 1.17281365e+00 -4.05313745e-02
-8.22732389e-01 -1.06042278e+00 4.22580898e-01 7.43257403e-01
1.70250580e-01 5.07419050e-01 4.29098189e-01 -6.61169231e-01
-5.19579649e-01 -2.51273394e-01 2.57552505e-01 1.48104727e-01
-1.42099157e-01 -2.36872703e-01 -8.69468212e-01 9.63312685e-02
-5.76642871e-01 -7.53234088e-01 4.62829411e-01 1.34911388e-01
4.97631431e-01 1.12379631e-02 -2.54743069e-01 4.58548635e-01
9.13721144e-01 3.21474075e-02 8.12507808e-01 6.04834020e-01
9.68481541e-01 4.96014446e-01 7.09213138e-01 3.03505808e-01
5.30359089e-01 8.66082966e-01 -1.47564411e-01 1.63095683e-01
-4.16433007e-01 -5.99480271e-01 6.97492599e-01 1.00072026e+00
-4.83532250e-01 -1.13192052e-01 -1.04974639e+00 7.94224322e-01
-2.24025106e+00 -1.38588369e+00 4.83848676e-02 1.93113732e+00
4.68066245e-01 -1.88399538e-01 2.89691925e-01 -1.73415110e-01
3.20582032e-01 1.29643947e-01 -1.35519996e-01 -1.36340857e-01
-8.22449625e-02 1.34922322e-02 3.50641161e-01 2.79687524e-01
-1.20411634e+00 1.09128940e+00 6.80232906e+00 5.34891903e-01
-7.38373458e-01 1.66655630e-01 1.55631844e-02 -3.07585955e-01
-5.37238345e-02 -1.52296722e-01 -3.95886004e-01 2.86148429e-01
1.49660563e+00 -9.16806012e-02 3.74879748e-01 5.80180585e-01
3.92189115e-01 -1.18606634e-01 -1.34191513e+00 1.30723155e+00
5.04298866e-01 -1.56092334e+00 2.68625617e-01 6.26827478e-02
7.26998448e-01 1.47148371e-01 -4.97745395e-01 5.99263549e-01
2.40007430e-01 -8.70188475e-01 9.79727805e-01 7.37248540e-01
6.16259038e-01 -5.18882215e-01 6.43551826e-01 5.57857692e-01
-1.25380838e+00 1.79156065e-01 -2.45491192e-02 8.44125524e-02
5.57186067e-01 -5.56005001e-01 -1.23583233e+00 7.01057196e-01
4.76703256e-01 9.38843787e-01 -5.58959067e-01 6.61217928e-01
-1.23873562e-01 4.80069518e-01 -8.64880756e-02 -1.78522430e-02
2.60542423e-01 -6.62647262e-02 6.38145208e-01 1.49573278e+00
5.17538011e-01 -1.25799969e-01 1.55022398e-01 7.24395752e-01
1.20489620e-01 1.03534006e-01 -1.14598966e+00 -5.22955239e-01
1.11950137e-01 9.35965717e-01 -4.42636043e-01 -6.18383110e-01
-6.52145445e-01 1.52251017e+00 2.18290418e-01 4.28406596e-01
-8.34138393e-01 -7.50421062e-02 2.65909255e-01 4.52305265e-02
1.75430804e-01 -6.17554903e-01 3.43474239e-01 -1.42158520e+00
-3.30845267e-01 -1.05175149e+00 3.42666596e-01 -1.31417525e+00
-8.25365722e-01 5.70094764e-01 4.52665567e-01 -1.52006054e+00
-1.24869955e+00 -4.86042887e-01 -2.11528704e-01 6.97145224e-01
-7.99141645e-01 -1.46736026e+00 -5.23447812e-01 5.81242263e-01
9.02764976e-01 -3.71417701e-01 9.21400070e-01 2.11004242e-01
-2.70260721e-01 1.64568961e-01 -4.18642700e-01 4.13939208e-01
8.49285722e-01 -1.01794636e+00 8.52586508e-01 7.37619221e-01
4.66288179e-01 2.88950235e-01 7.35063672e-01 -7.78143346e-01
-1.43260038e+00 -9.90499914e-01 7.72734880e-01 -1.21981454e+00
6.76228821e-01 -2.41160244e-01 -8.26189697e-01 1.15776324e+00
2.87794441e-01 1.52485892e-02 7.64306426e-01 -7.48464763e-01
-2.25369129e-02 6.78104043e-01 -7.94679165e-01 8.04791987e-01
1.41695511e+00 -6.50297165e-01 -1.04401362e+00 1.39790764e-02
3.37870687e-01 -3.14002067e-01 -9.77433443e-01 2.49097884e-01
8.02549064e-01 -5.05625546e-01 1.09523284e+00 -9.16470170e-01
5.70347905e-01 -2.35059977e-01 -2.78829932e-01 -1.46097350e+00
-5.59411421e-02 -5.45393705e-01 -3.71420026e-01 8.88353348e-01
3.54179829e-01 -3.39742638e-02 7.37893641e-01 5.37075996e-01
-1.22393005e-01 4.54015732e-02 -6.62022710e-01 -7.39409387e-01
-2.32575655e-01 -7.24522591e-01 -7.03078462e-03 1.02205348e+00
1.35721818e-01 3.43805939e-01 -7.86114156e-01 -3.83593589e-01
5.57978570e-01 -3.83354068e-01 1.42357910e+00 -9.11399305e-01
-5.00237644e-01 -1.60040960e-01 -8.34549069e-01 -1.01226807e+00
3.42903405e-01 -9.72743213e-01 2.26749778e-01 -1.86350381e+00
3.01043153e-01 2.38569826e-01 1.25392482e-01 5.10439754e-01
-1.30298454e-02 5.80578566e-01 6.01013780e-01 4.30234641e-01
-6.37385547e-01 5.73065519e-01 1.16520405e+00 -4.24541049e-02
-2.32928813e-01 -5.37045538e-01 6.45071268e-02 7.41498768e-01
4.28878307e-01 -2.40838751e-01 -4.27041322e-01 -2.88559109e-01
-9.79878232e-02 3.80522281e-01 4.68696505e-01 -1.49206591e+00
1.19262114e-02 -1.64585978e-01 5.57309687e-01 -3.84526849e-01
5.95205367e-01 -4.60022718e-01 3.39385897e-01 3.77301514e-01
-2.09988624e-01 1.66068878e-02 3.66373450e-01 6.14506185e-01
-9.51368883e-02 -1.11295236e-02 2.38144815e-01 -4.49348837e-01
-1.38026381e+00 1.40265143e-02 -9.01433885e-01 -3.61402817e-02
1.09088886e+00 -4.87089723e-01 -2.06246063e-01 -8.02004158e-01
-9.34716702e-01 6.87777176e-02 5.53632557e-01 7.76816607e-01
5.91159701e-01 -1.61619556e+00 -7.68682778e-01 -4.35910113e-02
3.95038515e-01 -3.17229211e-01 1.27419740e-01 8.32079530e-01
-8.14097106e-01 6.25758827e-01 -6.86039150e-01 -7.35896707e-01
-1.28353822e+00 3.82625669e-01 2.43968129e-01 5.11359274e-02
-8.80715072e-01 3.31938744e-01 -1.46051422e-01 -5.44763744e-01
-2.28188913e-02 -5.33966795e-02 -2.37008989e-01 -1.66477952e-02
6.57044828e-01 4.52359706e-01 -3.09785455e-01 -1.16969252e+00
-1.88129067e-01 3.60181689e-01 4.72717285e-01 -9.42962289e-01
1.19485545e+00 3.32956687e-02 6.63758337e-01 7.30878890e-01
1.02662551e+00 -3.53296310e-01 -1.49905956e+00 -1.15000471e-01
1.12724505e-01 -3.70862484e-01 -3.23336750e-01 -4.19768304e-01
-4.71233487e-01 7.58342385e-01 4.78868306e-01 -3.51477921e-01
6.67079449e-01 3.50436270e-02 1.07975197e+00 7.46239305e-01
7.43056536e-01 -1.23250198e+00 5.07687807e-01 6.35569036e-01
8.51841569e-01 -1.23078167e+00 -3.87678385e-01 9.04725045e-02
-1.14062190e+00 9.61218297e-01 5.39367437e-01 1.14710279e-01
-1.39309457e-02 3.98916490e-02 1.72516584e-01 2.00192571e-01
-7.33687758e-01 -3.85580182e-01 3.26753646e-01 9.15789485e-01
3.66165280e-01 -1.92459524e-01 -6.43615499e-02 3.54810059e-01
-8.31162930e-02 6.02092624e-01 7.47110367e-01 1.42756593e+00
-2.89951235e-01 -1.09878278e+00 -4.88906413e-01 1.12240158e-01
-2.15576336e-01 1.78648651e-01 -4.50963438e-01 1.07116210e+00
-1.54987425e-01 6.80821240e-01 9.61466134e-02 -5.73004901e-01
3.34977031e-01 6.31006837e-01 9.12451982e-01 -7.15858519e-01
-2.82362163e-01 4.20818180e-02 6.72663867e-01 -7.92200208e-01
-6.77154660e-01 -9.54995573e-01 -1.33950162e+00 -1.07006609e-01
4.74930495e-01 -5.57949208e-02 5.43695807e-01 1.15332544e+00
3.25649351e-01 5.25076568e-01 -9.05117765e-02 -1.61357510e+00
-1.61719471e-01 -1.18852329e+00 -1.66990563e-01 8.73012483e-01
5.20196296e-02 -6.67053819e-01 -1.56066820e-01 7.24779844e-01] | [7.385823726654053, -0.06998332589864731] |
659e1308-1ec5-4495-a565-9d5428c64d01 | hyperspectral-remote-sensing-image | 2106.14804 | null | https://arxiv.org/abs/2106.14804v1 | https://arxiv.org/pdf/2106.14804v1.pdf | Hyperspectral Remote Sensing Image Classification Based on Multi-scale Cross Graphic Convolution | The mining and utilization of features directly affect the classification performance of models used in the classification and recognition of hyperspectral remote sensing images. Traditional models usually conduct feature mining from a single perspective, with the features mined being limited and the internal relationships between them being ignored. Consequently, useful features are lost and classification results are unsatisfactory. To fully mine and utilize image features, a new multi-scale feature-mining learning algorithm (MGRNet) is proposed. The model uses principal component analysis to reduce the dimensionality of the original hyperspectral image (HSI) to retain 99.99% of its semantic information and extract dimensionality reduction features. Using a multi-scale convolution algorithm, the input dimensionality reduction features were mined to obtain shallow features, which then served as inputs into a multi-scale graph convolution algorithm to construct the internal relationships between eigenvalues at different scales. We then carried out cross fusion of multi-scale information obtained by graph convolution, before inputting the new information obtained into the residual network algorithm for deep feature mining. Finally, a flexible maximum transfer function classifier was used to predict the final features and complete the classification. Experiments on three common hyperspectral datasets showed the MGRNet algorithm proposed in this paper to be superior to traditional methods in recognition accuracy. | ['Guojin Liu', 'Tianchong Qiu', 'Zhihan Chen', 'Yin Li', 'Yunsong Zhao'] | 2021-06-28 | null | null | null | null | ['remote-sensing-image-classification'] | ['miscellaneous'] | [ 5.57988226e-01 -2.95850039e-01 2.16464221e-01 -4.07141566e-01
1.66707281e-02 -2.07651809e-01 4.80157882e-02 -2.44347826e-01
-1.76158398e-01 4.73596543e-01 -2.78482586e-01 -2.84809560e-01
-7.30612576e-01 -1.29246616e+00 -7.30928779e-02 -1.01450074e+00
-1.92683280e-01 -8.88425410e-02 -3.11936438e-01 -2.60887444e-01
2.40395918e-01 9.63209271e-01 -1.80844617e+00 4.07925278e-01
1.11510432e+00 1.24684966e+00 4.48194981e-01 3.14911574e-01
-2.34150350e-01 5.31648755e-01 -2.46851861e-01 2.48576030e-01
4.73311424e-01 -2.74049997e-01 -6.01673901e-01 4.97447640e-01
-1.36415675e-01 -4.44696993e-02 -5.43104172e-01 1.26775265e+00
4.02729362e-01 4.70803976e-01 4.82380420e-01 -1.10500741e+00
-7.62136281e-01 5.47531605e-01 -9.49336886e-01 2.08044961e-01
-1.12907603e-01 -2.26763189e-01 7.70602942e-01 -1.20196819e+00
8.96272659e-02 1.06800997e+00 5.16408145e-01 -2.15651952e-02
-9.91359890e-01 -7.97110021e-01 -1.07632145e-01 3.81156445e-01
-1.76108336e+00 -1.21867381e-01 1.09574258e+00 -3.28036219e-01
1.04413486e+00 4.71934348e-01 8.81105900e-01 1.00911520e-01
-6.27584681e-02 2.75190532e-01 1.06409943e+00 -4.46395814e-01
-2.05440506e-01 1.09414652e-01 3.49087656e-01 8.96023154e-01
3.09614927e-01 8.79989713e-02 -2.20989853e-01 -6.58614561e-02
5.32320499e-01 4.98920530e-01 -4.35863286e-01 1.24305278e-01
-8.48774016e-01 1.01517081e+00 9.11831975e-01 5.08780897e-01
-7.36822546e-01 -6.92228317e-01 2.40545301e-03 6.30793393e-01
4.70967025e-01 2.62279004e-01 -3.59293699e-01 5.21956325e-01
-8.02096128e-01 -4.92842793e-01 4.58137870e-01 3.57756853e-01
1.51874518e+00 2.46452227e-01 1.73852786e-01 9.97752190e-01
1.88975722e-01 5.70333302e-01 6.93747818e-01 -3.86791617e-01
3.09636384e-01 1.33841038e+00 -4.72803801e-01 -1.50312376e+00
-8.04879844e-01 -5.80383360e-01 -1.26225722e+00 3.07357669e-01
-2.07535535e-01 -3.38827521e-01 -8.63193691e-01 9.95010257e-01
2.68205464e-01 -1.20931249e-02 3.08146268e-01 1.24022961e+00
7.60572851e-01 9.80812848e-01 -2.44458243e-01 -1.82959795e-01
9.73779857e-01 -5.24196327e-01 -3.72800499e-01 -1.26968294e-01
5.02138197e-01 -7.19473779e-01 8.41403604e-01 3.58888656e-01
-4.34929281e-01 -7.63469934e-01 -1.11228657e+00 3.89819115e-01
-6.71419322e-01 5.21666527e-01 1.00554109e+00 3.60661238e-01
-8.35939527e-01 7.75358021e-01 -4.74857807e-01 -7.84838349e-02
6.08025372e-01 6.47315085e-01 -4.91426677e-01 -2.65233725e-01
-1.04739535e+00 6.68084562e-01 9.26954985e-01 6.02264225e-01
-2.20436111e-01 -4.25502360e-01 -7.85789907e-01 1.67805403e-01
1.88451782e-01 -2.56734908e-01 4.20693010e-01 -1.33672500e+00
-1.09889472e+00 3.75490695e-01 2.21618786e-02 1.51027292e-01
-2.76551843e-01 2.66837180e-01 -7.91511834e-01 3.08348894e-01
-1.02363564e-01 2.35107794e-01 1.04064929e+00 -8.09476137e-01
-6.62103772e-01 -7.06784844e-01 -2.35165849e-01 4.00865436e-01
-8.79777908e-01 -2.51136154e-01 2.13235378e-01 -3.33474159e-01
8.35230172e-01 -6.81618333e-01 -1.62523910e-01 -3.64952266e-01
-3.32897633e-01 1.53354099e-02 1.50535750e+00 -6.78936601e-01
1.00315619e+00 -2.41206956e+00 2.33830675e-01 8.52363169e-01
3.40800375e-01 3.28031361e-01 -3.86628807e-01 2.96513230e-01
-5.99707186e-01 -1.85360670e-01 -3.57483566e-01 4.39191401e-01
-5.21726072e-01 3.80907096e-02 -1.02884263e-01 6.00263536e-01
3.94432545e-01 7.40990281e-01 -5.27492464e-01 -1.65320456e-01
4.75078255e-01 5.95640182e-01 -2.94085182e-02 1.82742938e-01
3.02033663e-01 4.08982597e-02 -7.51946032e-01 9.28875089e-01
9.37196374e-01 -3.82207781e-01 -1.90844193e-01 -5.82683623e-01
-3.12220365e-01 -3.26815128e-01 -1.24055469e+00 1.27822506e+00
-2.91504920e-01 4.35747355e-01 1.19331554e-01 -1.42632234e+00
1.31298280e+00 5.48875425e-03 8.39578986e-01 -4.99705881e-01
1.40229166e-01 3.13371681e-02 1.59088790e-01 -6.76840365e-01
2.37425491e-01 -2.38710448e-01 4.86104220e-01 3.91408831e-01
-2.22481892e-01 -1.40946684e-03 -6.45078719e-03 -2.08687112e-01
5.45803130e-01 -2.77114898e-01 1.81421325e-01 -1.98374704e-01
7.24985898e-01 3.91017854e-01 2.96289742e-01 2.53198594e-01
2.56955057e-01 1.80919036e-01 -5.84541075e-02 -6.95188224e-01
-7.24611163e-01 -4.89070088e-01 -2.61993468e-01 8.99004102e-01
2.98314728e-02 2.07938217e-02 -2.80837238e-01 -5.81188679e-01
8.27192888e-02 3.42511773e-01 -5.24379551e-01 -4.31574613e-01
4.96722348e-02 -1.31552660e+00 2.34182939e-01 1.36323825e-01
8.14002991e-01 -1.08315170e+00 -4.59294468e-01 2.75727928e-01
1.89944282e-02 -6.00776017e-01 1.13616195e-02 3.55786294e-01
-1.10998809e+00 -1.17848623e+00 -2.79034287e-01 -8.14635158e-01
8.97567391e-01 9.52379107e-01 2.49527305e-01 1.98075429e-01
-7.32776523e-01 -1.45635143e-01 -4.83337462e-01 -5.15510321e-01
-1.33588277e-02 -6.77279234e-02 -2.19897740e-02 4.10671324e-01
6.59949839e-01 -8.52907479e-01 -4.20001894e-01 8.41891393e-02
-1.10850799e+00 1.39292002e-01 1.07469928e+00 1.07317901e+00
4.38346863e-01 9.36476827e-01 6.84434593e-01 -5.84653974e-01
6.72519326e-01 -4.62612063e-01 -6.54191494e-01 1.09204859e-01
-7.59221554e-01 -1.62409171e-01 8.26183200e-01 -4.81566608e-01
-1.03476858e+00 2.37223744e-01 3.04324239e-01 -4.85898405e-01
5.40012456e-02 1.17093635e+00 -2.49841884e-01 -5.76517999e-01
7.14720190e-01 6.72711492e-01 3.45884562e-01 -2.85772741e-01
7.24731237e-02 1.00669253e+00 -1.75126716e-02 1.23745792e-01
9.44271684e-01 3.28216583e-01 1.46988675e-01 -1.34126472e+00
-7.90761173e-01 -5.95674276e-01 -8.56082141e-01 -1.29664540e-01
6.41415596e-01 -8.69716883e-01 -5.36923409e-01 5.16507864e-01
-6.93297565e-01 2.46633604e-01 -2.22174913e-01 8.24178815e-01
7.83008635e-02 4.25798655e-01 -3.80169481e-01 -7.61254728e-01
-5.41852713e-01 -7.58485317e-01 6.18289769e-01 4.73546892e-01
4.88548964e-01 -7.97778606e-01 -3.65211368e-01 2.25662008e-01
4.80937868e-01 2.06703618e-02 1.17051780e+00 -5.08826733e-01
-1.03821389e-01 -3.86949003e-01 -6.94654703e-01 5.86826801e-01
5.34497261e-01 -8.30017254e-02 -9.51062679e-01 -3.37764174e-01
2.09428534e-01 -2.30875388e-01 9.84549940e-01 3.50488991e-01
1.48406553e+00 -1.83392331e-01 -2.87086904e-01 1.03700602e+00
1.68273389e+00 3.64763647e-01 5.24510682e-01 2.54779607e-01
9.35653746e-01 5.63641131e-01 4.95116889e-01 6.08461261e-01
-9.46472362e-02 -1.23874292e-01 4.42222357e-01 -6.34100676e-01
3.32885176e-01 2.02347472e-01 1.07619084e-01 8.82719874e-01
-4.13943261e-01 3.47230941e-01 -7.00633228e-01 8.66290703e-02
-1.63421834e+00 -1.07941294e+00 -1.17532685e-01 1.74050438e+00
4.36379671e-01 -2.60982424e-01 -2.74902880e-01 5.15862882e-01
9.22032297e-01 3.63004580e-02 -8.46472800e-01 -3.12071852e-02
-5.20114183e-01 3.26848537e-01 4.56667274e-01 1.33380905e-01
-1.19356728e+00 8.54657352e-01 5.48908281e+00 4.66590405e-01
-1.54053330e+00 -2.64479935e-01 3.63455564e-01 7.02571794e-02
-5.76779172e-02 -7.70276561e-02 -3.61569345e-01 -1.21585891e-01
5.80163181e-01 -1.60440370e-01 9.05743182e-01 7.06057072e-01
3.51256013e-01 9.62483734e-02 -4.14171159e-01 1.25872421e+00
-6.63726106e-02 -1.03802550e+00 3.83208275e-01 2.18489096e-01
5.99959493e-01 1.37250274e-01 -7.31393844e-02 1.50938496e-01
-3.35873966e-03 -1.13802147e+00 2.71810126e-03 7.81770527e-01
7.61197448e-01 -9.90606844e-01 6.59148693e-01 3.78798425e-01
-1.27043557e+00 -5.64297557e-01 -9.93299425e-01 -3.43646020e-01
-4.19245869e-01 9.86165583e-01 -6.96698368e-01 9.99983549e-01
3.95400524e-01 9.97744620e-01 -5.25407076e-01 8.67931962e-01
1.26189753e-01 4.05722111e-01 -3.27114344e-01 -1.65686421e-02
4.41150695e-01 -7.51409769e-01 3.54057997e-01 9.54394698e-01
6.52560472e-01 5.10026693e-01 4.21214670e-01 8.49142313e-01
1.35230377e-01 4.21393752e-01 -7.10029304e-01 -4.29864585e-01
1.98562846e-01 1.75450921e+00 -6.25057936e-01 -2.49580353e-01
-5.97936690e-01 9.14947212e-01 2.41084516e-01 4.61697549e-01
-4.72543091e-01 -8.80210757e-01 4.43912148e-01 -1.92065388e-01
3.04892600e-01 -2.16434285e-01 -1.80709139e-01 -1.23558688e+00
-2.29243226e-02 -6.78278029e-01 5.79406857e-01 -8.10213685e-01
-1.23754025e+00 7.71458566e-01 -3.34189296e-01 -1.21824706e+00
2.46027693e-01 -7.97544956e-01 -6.00652277e-01 1.48913884e+00
-1.92812192e+00 -1.18735969e+00 -7.07335651e-01 1.07097828e+00
2.28232101e-01 -5.58356822e-01 9.95609999e-01 1.51357889e-01
-7.18872070e-01 6.41948963e-03 1.58464953e-01 2.40503713e-01
-1.42510179e-02 -7.33942211e-01 -4.19051290e-01 8.82880747e-01
-9.87059847e-02 3.23688000e-01 -4.19237465e-02 -7.66557515e-01
-1.74333584e+00 -1.38612378e+00 3.35126191e-01 7.06741214e-01
6.67282104e-01 2.14773282e-01 -9.68670964e-01 2.91502148e-01
-3.51631880e-01 9.47611257e-02 1.03480816e+00 -1.53005645e-01
-9.84747633e-02 -3.64403188e-01 -1.06746352e+00 1.37991995e-01
7.61017323e-01 -4.96683687e-01 -4.69520450e-01 4.18996066e-01
4.88728732e-01 2.51910746e-01 -9.68581438e-01 4.08334255e-01
4.92445499e-01 -7.16577709e-01 7.69795001e-01 -7.10738420e-01
2.78345078e-01 -3.04752529e-01 -2.45424941e-01 -1.45926499e+00
-8.50475132e-01 -5.62955998e-03 2.94860452e-01 5.92182517e-01
6.24184191e-01 -8.64364445e-01 6.13216341e-01 1.98845297e-01
-2.25104213e-01 -5.15990853e-01 -5.92437387e-01 -4.04697001e-01
-4.94932353e-01 -2.37516463e-01 8.17054451e-01 1.25554538e+00
1.87799022e-01 4.73449886e-01 -3.97108048e-02 6.29453540e-01
6.73743904e-01 7.63414860e-01 1.33734733e-01 -1.62466025e+00
2.80258409e-03 -4.80761528e-01 -4.86251175e-01 -3.81215632e-01
2.78842419e-01 -1.39769256e+00 -4.40623075e-01 -1.59398842e+00
1.47593558e-01 -4.94743049e-01 -4.58269358e-01 8.73888373e-01
-1.05130248e-01 3.18145126e-01 -3.46306935e-02 5.24559140e-01
2.28533134e-01 6.08166993e-01 1.48824191e+00 -3.85025531e-01
-5.35212219e-01 1.56135917e-01 -9.01141286e-01 5.46372056e-01
9.30945218e-01 -2.78429300e-01 -4.68264490e-01 -3.32283288e-01
-1.18792020e-01 3.32059786e-02 1.78976879e-01 -1.05545402e+00
1.99266180e-01 -3.97505045e-01 1.05183804e+00 -5.49284101e-01
3.39413881e-01 -1.04323006e+00 2.11303443e-01 5.42411208e-01
1.21891022e-01 -4.30911213e-01 1.84065238e-01 1.20100781e-01
-3.79704386e-01 -1.80907473e-01 6.16690099e-01 -1.69174552e-01
-9.89208877e-01 6.35164678e-01 -1.71814665e-01 -7.76159465e-01
1.04417682e+00 -6.15190029e-01 2.09661629e-02 1.61931012e-02
-1.06758857e+00 1.41981087e-04 -4.61626202e-02 1.59695938e-01
1.21203470e+00 -1.20664394e+00 -9.61040139e-01 9.16480601e-01
5.56594543e-02 -9.71278250e-02 3.35997313e-01 6.81405485e-01
-4.60051715e-01 1.73600033e-01 -4.78543550e-01 -5.55158257e-01
-1.15864146e+00 4.79142308e-01 4.70552832e-01 1.80523187e-01
-6.78930223e-01 5.29654682e-01 -1.75225183e-01 -3.84763896e-01
-3.03330660e-01 1.11827264e-02 -6.66392148e-01 3.84560525e-01
7.33098447e-01 3.49421382e-01 3.44404399e-01 -7.60695100e-01
-1.81869939e-01 6.36346161e-01 1.95368916e-01 2.16795668e-01
1.92874241e+00 -1.20822601e-01 -6.99006081e-01 1.25712350e-01
1.33614945e+00 -3.66347432e-01 -8.55811775e-01 -4.53441083e-01
-2.96797216e-01 -5.63077271e-01 5.66981614e-01 -6.24337375e-01
-1.45171928e+00 7.79127359e-01 6.72502279e-01 4.13754433e-01
1.75733531e+00 -4.19466555e-01 3.85503441e-01 8.76402378e-01
2.37805601e-02 -9.79796529e-01 -3.63809705e-01 4.22763258e-01
7.01856732e-01 -1.18516135e+00 1.11042283e-01 -4.91864592e-01
-4.43870813e-01 1.68648195e+00 5.39017260e-01 -4.33292165e-02
9.56090212e-01 1.45074815e-01 -2.35204883e-02 -4.67120439e-01
-1.19510755e-01 -3.10332000e-01 4.83843893e-01 5.61457455e-01
2.37247333e-01 1.68844804e-01 -9.54029039e-02 4.69961524e-01
-2.87436336e-01 -1.93387028e-02 3.25612247e-01 8.40817213e-01
-9.14297819e-01 -6.57786131e-01 -5.25490582e-01 9.95880544e-01
-4.27880920e-02 -2.38077119e-01 -3.92060369e-01 4.33179468e-01
4.93692793e-02 1.04195619e+00 -9.18266848e-02 -8.25609982e-01
3.05966437e-01 8.39325637e-02 1.66714013e-01 -6.07339561e-01
-2.23566264e-01 6.15344830e-02 -2.49643981e-01 -3.03992271e-01
-4.54862654e-01 -3.80592585e-01 -1.46053040e+00 -1.64520532e-01
-6.34025931e-01 4.46741372e-01 8.41867030e-01 8.77200186e-01
3.03301781e-01 5.38711548e-01 1.26545751e+00 -9.34404016e-01
-7.42056549e-01 -1.16197562e+00 -1.22907853e+00 1.81582168e-01
2.44589493e-01 -3.76066029e-01 -3.97642851e-01 -2.75581386e-02] | [9.89140796661377, -1.631320834159851] |
90aa012f-1832-4bd6-9d81-c306b9f72846 | simulating-time-to-event-prediction-with | 2103.02583 | null | https://arxiv.org/abs/2103.02583v1 | https://arxiv.org/pdf/2103.02583v1.pdf | Simulating time to event prediction with spatiotemporal echocardiography deep learning | Integrating methods for time-to-event prediction with diagnostic imaging modalities is of considerable interest, as accurate estimates of survival requires accounting for censoring of individuals within the observation period. New methods for time-to-event prediction have been developed by extending the cox-proportional hazards model with neural networks. In this paper, to explore the feasibility of these methods when applied to deep learning with echocardiography videos, we utilize the Stanford EchoNet-Dynamic dataset with over 10,000 echocardiograms, and generate simulated survival datasets based on the expert annotated ejection fraction readings. By training on just the simulated survival outcomes, we show that spatiotemporal convolutional neural networks yield accurate survival estimates. | ['William Hiesinger', 'Curtis P. Langlotz', 'John P. Cunningham', 'Jeffrey Teuteberg', 'Michelle C. Li', 'Kate M. Callon', 'Cayley Bowles', 'Patpilai Kasinpila', 'Robyn Fong', 'Nicolas Quach', 'Rohan Shad'] | 2021-03-03 | null | null | null | null | ['time-to-event-prediction'] | ['time-series'] | [-8.98325220e-02 -1.51029825e-01 -1.91213623e-01 -6.37478173e-01
-8.72470319e-01 -4.17983681e-01 1.90281663e-02 -2.63186339e-02
-2.65970528e-01 1.04384005e+00 3.83757710e-01 -6.38289750e-01
-3.85216445e-01 -6.31487012e-01 -3.71798903e-01 -4.38443542e-01
-1.20549679e+00 7.18581259e-01 -3.01899850e-01 2.52830267e-01
-3.03670287e-01 5.79821587e-01 -5.40976405e-01 2.40068182e-01
5.77071548e-01 1.19953918e+00 -5.76547742e-01 1.23346651e+00
4.57497895e-01 1.17580903e+00 -5.75786471e-01 -9.22216922e-02
4.46427912e-02 -6.67084217e-01 -5.85435212e-01 -4.92484421e-01
4.91319984e-01 -1.03651452e+00 -8.84230435e-01 2.40676939e-01
7.52683163e-01 -2.96990294e-02 7.85861015e-01 -9.85739529e-01
-5.20587504e-01 6.31573200e-01 2.53251374e-01 6.98304474e-01
-2.48259619e-01 1.70967534e-01 3.96590441e-01 -6.89714789e-01
7.07210541e-01 6.29450619e-01 1.32130432e+00 5.69533885e-01
-1.09026635e+00 -4.06459361e-01 -4.98504519e-01 2.84334291e-02
-9.26056921e-01 -2.09618732e-01 3.90440136e-01 -6.19503677e-01
7.90762603e-01 1.97997227e-01 9.58863795e-01 1.29884470e+00
9.32231367e-01 4.76913214e-01 7.67713547e-01 3.52674455e-04
-3.09282281e-02 -6.04116440e-01 2.62283653e-01 7.18421698e-01
-1.74902901e-01 8.48006904e-01 -1.80215016e-01 -5.33942461e-01
1.18148386e+00 4.36154962e-01 -2.33132914e-01 -2.73780376e-01
-1.44266295e+00 6.78302169e-01 3.73715818e-01 4.96108085e-02
-4.85144377e-01 6.79893970e-01 8.65990818e-01 1.75978392e-01
5.99168718e-01 1.40142664e-01 -6.44575119e-01 -1.16356194e-01
-1.28891695e+00 3.06988835e-01 7.15741217e-01 3.70704889e-01
-3.22003275e-01 4.07366365e-01 -5.24703145e-01 4.24801826e-01
-1.68690667e-01 4.36504453e-01 3.83344889e-01 -1.38389182e+00
1.67090353e-02 7.01647848e-02 3.18291545e-01 -6.73273206e-01
-1.21127057e+00 -1.00125504e+00 -1.37532055e+00 3.17116439e-01
8.23029339e-01 -6.97291791e-01 -1.15504527e+00 1.68552804e+00
-4.27628979e-02 5.25477409e-01 1.80643365e-01 9.68610525e-01
6.70418918e-01 3.40810448e-01 4.07641023e-01 -3.01665753e-01
1.06543446e+00 -5.75109184e-01 -6.55879855e-01 2.51068324e-01
1.14066184e+00 1.48818776e-01 4.40010399e-01 1.75935879e-01
-1.07980335e+00 -3.40753883e-01 -6.83426797e-01 1.63743615e-01
-3.56250852e-02 4.89787668e-01 7.17316926e-01 3.77885193e-01
-9.93669093e-01 1.32402933e+00 -1.44731224e+00 -2.47902080e-01
7.88345456e-01 1.93380132e-01 -2.46578559e-01 2.36992821e-01
-1.62584412e+00 7.78422415e-01 1.04618080e-01 2.79818237e-01
-1.52087426e+00 -1.21008050e+00 -6.71612561e-01 3.58144552e-01
-7.99960792e-02 -1.15546227e+00 9.67632949e-01 -7.72069037e-01
-9.67241645e-01 7.01888800e-01 1.20249435e-01 -9.48904037e-01
7.24525630e-01 -1.30435184e-01 -4.93543833e-01 2.54674792e-01
-8.64184201e-02 3.66901845e-01 3.74557257e-01 -4.16024953e-01
-2.68568307e-01 -4.33648705e-01 -2.38904014e-01 -1.79930910e-01
-1.74114391e-01 1.99013159e-01 2.36024484e-01 -9.90573049e-01
-2.05860302e-01 -1.13484752e+00 -6.02590859e-01 4.20566052e-01
-1.62032872e-01 1.45268902e-01 4.15081292e-01 -1.28542912e+00
1.14877880e+00 -1.98898411e+00 -1.67795733e-01 -9.54675600e-02
8.03340375e-01 -4.03739437e-02 2.30321750e-01 2.40929872e-02
-3.74914736e-01 1.97389767e-01 -2.37524003e-01 -2.51220584e-01
-4.91001993e-01 -8.04554820e-02 -2.36537367e-01 4.91918892e-01
-5.30162491e-02 1.24070930e+00 -6.91543102e-01 -5.73042929e-01
1.75326765e-01 6.10163271e-01 -4.70152676e-01 2.06165031e-01
2.82826483e-01 9.00234461e-01 -2.92875618e-01 5.96150994e-01
1.91248685e-01 -8.03423285e-01 3.93451184e-01 7.49310106e-02
1.47365034e-01 -1.41192079e-01 -2.14754224e-01 1.89887476e+00
-1.50718212e-01 7.63793826e-01 -2.97668606e-01 -9.45292592e-01
4.57006395e-01 8.21027398e-01 9.24105883e-01 -3.23177636e-01
1.28667012e-01 1.36701092e-01 1.60269171e-01 -4.05409306e-01
-1.94076821e-01 -6.08192146e-01 -2.58972406e-01 3.46904844e-01
1.97134361e-01 7.06881106e-01 -2.38524109e-01 1.95811674e-01
1.49658132e+00 2.39351079e-01 1.25218257e-01 -1.34805992e-01
-9.58085954e-02 2.75276214e-01 8.99304748e-01 1.09891975e+00
-6.81937814e-01 9.33151841e-01 6.48954511e-01 -1.36921906e+00
-1.15980244e+00 -1.54699051e+00 -4.82504040e-01 5.76694191e-01
-4.17816401e-01 1.70145184e-01 -4.15548474e-01 -8.67670119e-01
-9.62371286e-03 4.16837275e-01 -9.64637637e-01 -2.85724252e-01
-9.37297165e-01 -1.05700970e+00 1.10121083e+00 1.29460835e+00
1.32036924e-01 -9.72497046e-01 -1.00803030e+00 6.15192473e-01
-2.53233910e-01 -6.96658671e-01 -2.85149157e-01 3.89824286e-02
-1.37188411e+00 -1.01193941e+00 -1.42965198e+00 -5.98948598e-01
1.74487874e-01 -8.01081181e-01 1.47085416e+00 2.51512080e-01
-6.10896289e-01 1.09804079e-01 1.14363477e-01 -1.71190575e-01
-5.74200332e-01 1.50654003e-01 6.04159720e-02 -4.18939620e-01
-2.36982927e-01 -7.21190929e-01 -1.27504790e+00 7.77783468e-02
-3.58745754e-01 1.15580484e-01 2.52258867e-01 1.09615242e+00
5.06564081e-01 -6.58990026e-01 1.10503399e+00 -8.68413508e-01
1.76833764e-01 -7.92913675e-01 -3.83382231e-01 2.58948594e-01
-7.15791404e-01 -2.90970296e-01 8.51070344e-01 -4.43967640e-01
-9.24101532e-01 4.51172143e-02 8.17153305e-02 -9.34765160e-01
-1.81323811e-01 8.83104146e-01 7.18544424e-01 2.48580053e-01
6.14864111e-01 2.30553135e-01 1.05494529e-01 -1.52877465e-01
8.25235397e-02 4.56725657e-02 8.13859642e-01 -2.32150644e-01
-6.43553212e-02 6.57963336e-01 6.48234487e-01 -1.20378293e-01
-6.75842047e-01 2.79297620e-01 -6.44437790e-01 -4.29356754e-01
9.45871949e-01 -8.36433530e-01 -7.78049052e-01 5.40527463e-01
-1.02913404e+00 -4.78439093e-01 -4.47992861e-01 9.10110652e-01
-8.41076672e-01 -1.23832552e-02 -1.10703862e+00 -5.07953227e-01
-5.29200673e-01 -7.98007965e-01 5.98964870e-01 -6.47589564e-02
-1.71116114e-01 -1.43666327e+00 1.87132120e-01 -2.43442595e-01
7.35199451e-01 8.70874882e-01 1.10630465e+00 -8.42289269e-01
-4.55177307e-01 -4.77193952e-01 -2.07580879e-01 2.22696334e-01
-7.40291132e-03 -3.28672171e-01 -4.54364061e-01 -3.09309185e-01
-3.24918270e-01 -2.36715287e-01 9.05843079e-01 1.38837624e+00
1.55136764e+00 -6.27892232e-03 -5.81389248e-01 1.11863744e+00
1.05979228e+00 4.09488976e-01 5.70532978e-01 -1.08326346e-01
5.78506827e-01 2.74281293e-01 2.54474074e-01 5.64917028e-01
4.94721711e-01 2.79943824e-01 2.16923609e-01 -5.02763331e-01
-2.58640528e-01 1.32760648e-02 -3.39338154e-01 4.58515495e-01
-4.60125297e-01 -3.19584966e-01 -1.22457933e+00 6.70367658e-01
-1.77566755e+00 -9.68314350e-01 -1.80135533e-01 2.14487433e+00
6.03632092e-01 -8.00487995e-02 7.40629956e-02 -4.98577416e-01
4.96104121e-01 1.38525981e-02 -7.91554272e-01 5.51455421e-03
-1.80623844e-01 1.52283192e-01 6.30424619e-01 1.61991388e-01
-1.22834480e+00 3.11495602e-01 7.84817362e+00 1.60252959e-01
-1.21862388e+00 3.34242374e-01 1.35445046e+00 -4.13475841e-01
2.53103405e-01 1.86058417e-01 -1.70985430e-01 3.50772560e-01
1.47093356e+00 -2.13042453e-01 1.70172863e-02 4.31103289e-01
4.50572640e-01 4.26778674e-01 -1.16818309e+00 5.23916423e-01
-2.35869840e-01 -1.72902524e+00 -3.85336578e-01 -1.49718702e-01
5.89385331e-01 1.13799073e-01 2.81418487e-02 3.39726180e-01
3.63765717e-01 -1.24716723e+00 2.29889661e-01 1.25632715e+00
1.42269421e+00 -4.08224821e-01 8.21328580e-01 2.32859254e-02
-8.62998307e-01 -1.25193000e-01 1.01123415e-01 3.90895866e-02
7.22862601e-01 5.37485123e-01 -7.13291883e-01 3.38422060e-01
7.38992274e-01 8.85742784e-01 -2.44111106e-01 9.94188547e-01
3.38383466e-01 1.13959515e+00 -1.89538136e-01 6.18601441e-01
-1.97798431e-01 1.89506665e-01 6.25661731e-01 8.68313015e-01
7.29275465e-01 2.08763376e-01 1.36611819e-01 1.07730544e+00
-3.50571759e-02 -1.97738424e-01 -5.10822952e-01 1.90367678e-03
7.66193196e-02 9.93018985e-01 -7.07873344e-01 -6.92955673e-01
-3.73874396e-01 7.80820549e-01 1.13034718e-01 5.63787162e-01
-1.03324294e+00 -2.52948374e-01 2.17104256e-01 4.79198277e-01
-1.06577031e-01 -1.93433061e-01 -6.10340178e-01 -1.15569139e+00
-3.70967329e-01 -3.22229147e-01 8.20617139e-01 -9.75971758e-01
-1.20733547e+00 5.78129113e-01 -1.77589998e-01 -1.17170846e+00
-5.33824503e-01 -4.19751227e-01 -6.76226556e-01 1.13494706e+00
-1.03088427e+00 -1.15772307e+00 -2.87434369e-01 3.08753639e-01
2.77378321e-01 -1.12409562e-01 9.34199214e-01 4.31753486e-01
-4.18792546e-01 5.29270232e-01 9.54249352e-02 6.29585624e-01
7.91397154e-01 -1.31970346e+00 6.22055173e-01 4.41089451e-01
-4.20979768e-01 4.65772033e-01 4.23711628e-01 -9.55546021e-01
-7.74433792e-01 -1.26276886e+00 8.69894207e-01 -8.23900044e-01
4.75397617e-01 2.35538647e-01 -1.00627768e+00 9.98372555e-01
-2.00002581e-01 6.93779230e-01 5.15084684e-01 5.26749492e-02
2.91713536e-01 5.68167493e-02 -8.82257164e-01 3.17783922e-01
1.15116847e+00 -3.34332347e-01 -3.05270702e-01 2.42613897e-01
4.50675994e-01 -6.69057250e-01 -1.51971877e+00 8.82811487e-01
1.00048757e+00 -6.63492978e-01 1.13320613e+00 -1.41846848e+00
7.77674198e-01 2.47311071e-01 3.37213099e-01 -1.30414546e+00
-2.53763020e-01 -3.06160927e-01 -2.43074507e-01 5.69066167e-01
3.96069169e-01 -6.15208983e-01 8.74867320e-01 6.63136661e-01
-2.66236365e-01 -9.59088922e-01 -1.26251948e+00 -6.63990855e-01
5.82093656e-01 -2.34095618e-01 3.71625811e-01 1.02703953e+00
-2.69516528e-01 -3.41092646e-01 -7.55904973e-01 2.04857871e-01
7.42603958e-01 1.75360501e-01 2.29222910e-03 -1.38614702e+00
-2.64579922e-01 -1.39590085e-01 -3.20052981e-01 -4.06562239e-01
1.48678616e-01 -8.80006611e-01 -2.02719077e-01 -1.43523264e+00
4.70555663e-01 -6.73777163e-01 -8.62968743e-01 3.97964805e-01
-2.18795463e-01 3.04621100e-01 -1.29267484e-01 3.23643535e-01
-3.61573935e-01 3.32322687e-01 1.21701932e+00 1.50692537e-01
-1.74062420e-02 7.08207116e-02 -2.70017609e-02 7.04865515e-01
8.44824612e-01 -7.25195408e-01 -1.25421837e-01 -1.64046705e-01
4.45555449e-02 1.42250383e+00 1.09959686e+00 -1.12344134e+00
-5.53714521e-02 5.48236519e-02 9.85577047e-01 -5.49081385e-01
1.98662981e-01 -4.14305270e-01 4.38593447e-01 7.54653931e-01
-7.24652052e-01 3.31055313e-01 1.23265825e-01 5.83156407e-01
-5.61796874e-02 4.05654728e-01 5.78585923e-01 -2.27183044e-01
-2.28310809e-01 7.41203308e-01 -5.97158194e-01 2.38461390e-01
8.15153778e-01 -5.18507920e-02 -2.88951874e-01 -5.05818069e-01
-1.69568777e+00 2.15499803e-01 1.10710468e-02 2.02613473e-01
6.48098528e-01 -1.49056947e+00 -1.02673411e+00 -2.33779270e-02
-1.49669170e-01 -4.98934507e-01 9.19625938e-01 1.28237212e+00
-8.00348103e-01 5.47637641e-01 -2.92795271e-01 -6.41949773e-01
-9.14800704e-01 4.53324169e-01 1.14189672e+00 -5.64052522e-01
-9.37485099e-01 3.33801717e-01 4.92033482e-01 -3.00477207e-01
2.69365846e-03 -3.80402863e-01 -8.06284100e-02 -3.86793077e-01
2.83178538e-01 5.69717109e-01 -2.16677442e-01 -1.49867281e-01
-3.45108986e-01 4.21982408e-02 1.15538150e-01 -5.04367165e-02
1.22099304e+00 -2.61879712e-02 2.76127934e-01 4.96882766e-01
1.04873967e+00 -7.01591790e-01 -1.47493505e+00 7.29938745e-02
-2.48888016e-01 -6.91576302e-03 2.98228830e-01 -1.30599582e+00
-1.34979725e+00 1.10065520e+00 1.19020522e+00 4.20422852e-03
1.08101118e+00 -1.90287426e-01 9.68569636e-01 -9.26313698e-02
1.53486401e-01 -5.62153935e-01 -3.93299103e-01 2.05122024e-01
6.88934803e-01 -1.10890758e+00 -2.31519416e-01 6.29335269e-02
-4.88689601e-01 1.37712657e+00 3.69193584e-01 -3.16721499e-01
9.82894123e-01 3.77813935e-01 3.92775983e-01 -2.19013363e-01
-9.17043328e-01 5.25146544e-01 1.46657810e-01 3.73018891e-01
5.59934020e-01 3.64817411e-01 -3.17725003e-01 9.44290698e-01
4.47176576e-01 8.57109368e-01 5.10534823e-01 6.96008682e-01
1.00186370e-01 -5.65335512e-01 -4.83465120e-02 9.75741506e-01
-1.05884099e+00 -3.65945846e-01 2.81933576e-01 5.88873744e-01
-2.02138841e-01 2.58697361e-01 1.27051115e-01 -2.87848059e-02
1.14958800e-01 2.86593139e-01 2.37503976e-01 -1.01181611e-01
-4.18785095e-01 -2.39878640e-01 1.38008624e-01 -5.85642457e-01
-6.48100004e-02 -7.84986496e-01 -1.44563568e+00 -1.69789612e-01
9.06169415e-02 -2.87109792e-01 3.11177701e-01 8.50276768e-01
6.58820450e-01 1.18125188e+00 4.35489982e-01 -6.74765944e-01
-7.42509663e-01 -9.15359139e-01 -7.21497953e-01 3.58206004e-01
6.02476776e-01 -4.63436604e-01 -4.67078149e-01 2.11289153e-01] | [7.936946392059326, 5.794567108154297] |
94f25e2c-f8f2-4991-b921-3b3db22c9575 | learning-to-generate-novel-scientific | 2305.14259 | null | https://arxiv.org/abs/2305.14259v1 | https://arxiv.org/pdf/2305.14259v1.pdf | Learning to Generate Novel Scientific Directions with Contextualized Literature-based Discovery | Literature-Based Discovery (LBD) aims to discover new scientific knowledge by mining papers and generating hypotheses. Standard LBD is limited to predicting pairwise relations between discrete concepts (e.g., drug-disease links). LBD also ignores critical contexts like experimental settings (e.g., a specific patient population where a drug is evaluated) and background knowledge and motivations that human scientists consider (e.g., to find a drug candidate without specific side effects). We address these limitations with a novel formulation of contextualized-LBD (C-LBD): generating scientific hypotheses in natural language, while grounding them in a context that controls the hypothesis search space. We present a new modeling framework using retrieval of ``inspirations'' from a heterogeneous network of citations and knowledge graph relations, and create a new dataset derived from papers. In automated and human evaluations, our models improve over baselines, including powerful large language models (LLMs), but also reveal challenges on the road to building machines that generate new scientific knowledge. | ['Tom Hope', 'Heng Ji', 'Doug Downey', 'Qingyun Wang'] | 2023-05-23 | null | null | null | null | ['contextualized-literature-based-discovery'] | ['natural-language-processing'] | [ 2.80033678e-01 5.31042516e-01 -1.02884817e+00 -7.54962564e-02
-6.34327829e-01 -8.55812967e-01 8.04814398e-01 7.85995901e-01
5.79018779e-02 1.42608333e+00 3.26646149e-01 -9.15085495e-01
-5.07940948e-01 -9.60862756e-01 -1.03394520e+00 -2.18086705e-01
-2.14908436e-01 7.15943992e-01 -1.26882106e-01 2.09677309e-01
4.19915736e-01 5.64231038e-01 -8.78035903e-01 2.84167200e-01
9.00860012e-01 1.92006588e-01 1.63495794e-01 1.63851127e-01
-2.97537982e-01 8.79789710e-01 -6.54358506e-01 -4.54208165e-01
-1.69137627e-01 -6.15292072e-01 -1.08061683e+00 -4.85514432e-01
3.72465372e-01 4.03831542e-01 -3.14291090e-01 8.01477611e-01
4.11645979e-01 -9.82342362e-02 6.36840940e-01 -1.38791203e+00
-9.56404567e-01 9.75887299e-01 -3.86442959e-01 1.99618846e-01
5.20860136e-01 6.25026822e-02 1.44610763e+00 -8.76488984e-01
1.51086485e+00 1.24701917e+00 3.95781040e-01 7.82846928e-01
-1.49626267e+00 -8.76538396e-01 3.47552925e-01 3.21075469e-01
-1.37752783e+00 -2.16531307e-01 5.46774566e-01 -6.44432604e-01
9.65604484e-01 3.67853194e-01 3.85344386e-01 1.51726317e+00
1.28520191e-01 4.80822444e-01 1.06481338e+00 -3.35163146e-01
5.96937954e-01 2.94992533e-02 2.01620355e-01 6.75185740e-01
1.02321923e+00 1.82101056e-02 -7.55430281e-01 -6.79034293e-01
6.17726326e-01 -1.61960587e-01 -4.60172623e-01 1.13539295e-02
-1.57250810e+00 6.68308258e-01 2.31573522e-01 2.85972685e-01
-3.63457173e-01 -2.35275850e-01 -2.99507137e-02 -2.22780295e-02
2.64759213e-01 1.34659934e+00 -7.97787964e-01 4.05947626e-01
-1.05838835e+00 4.63263899e-01 1.15004289e+00 1.30802739e+00
5.92481732e-01 -6.03448629e-01 -4.27763820e-01 3.62627685e-01
1.24101453e-01 -4.96896356e-02 1.26085043e-01 -7.83009827e-01
3.62354517e-01 6.48113191e-01 1.73882023e-01 -9.13240373e-01
-4.93550003e-01 -7.22720265e-01 -7.10844219e-01 -4.32245046e-01
1.64638460e-02 -2.72848159e-01 -7.62335002e-01 1.95310771e+00
1.22025587e-01 7.65049279e-01 9.87099037e-02 7.08400428e-01
1.45978403e+00 4.13807213e-01 5.42218864e-01 -5.84583461e-01
1.19367576e+00 -5.93004882e-01 -6.68779194e-01 -1.25318170e-01
7.28925049e-01 -4.58370090e-01 5.75749278e-01 3.57430309e-01
-1.02971911e+00 -1.13893405e-01 -7.80929804e-01 9.44896042e-02
-7.08596587e-01 -2.98303574e-01 8.92444015e-01 5.94700798e-02
-1.11221373e+00 6.71104252e-01 -1.75759956e-01 -2.98869193e-01
8.28241408e-01 4.21728879e-01 -2.87345678e-01 -2.03892231e-01
-1.60554230e+00 8.59527290e-01 3.85187596e-01 -2.82994062e-01
-8.62647176e-01 -1.41288328e+00 -4.46572244e-01 -1.29767537e-01
6.52202308e-01 -1.33357131e+00 6.77521050e-01 -3.97959948e-02
-9.15490270e-01 9.38536823e-01 -4.41534847e-01 -4.40172762e-01
1.43881068e-01 2.69934505e-01 -5.94301760e-01 -8.25530142e-02
5.76390743e-01 7.29072869e-01 2.59107854e-02 -1.22469711e+00
-3.27915847e-01 -1.84422940e-01 2.16799140e-01 -1.56280860e-01
1.85852852e-02 4.34134454e-02 -5.85178912e-01 -6.03678286e-01
-3.45840305e-01 -8.51192176e-01 -6.15449607e-01 -3.50826383e-01
-1.11793542e+00 -6.00292265e-01 5.28936684e-01 -1.72793880e-01
1.29241490e+00 -1.32900107e+00 4.15419549e-01 4.48265672e-01
9.48930323e-01 -1.39273688e-01 -1.88953474e-01 4.86428469e-01
-4.24149841e-01 9.44021523e-01 1.11258049e-02 2.23199740e-01
-2.49281242e-01 1.67326465e-01 -3.30703110e-01 5.60353249e-02
3.61678660e-01 1.31082869e+00 -1.38590753e+00 -7.51972437e-01
-4.17987347e-01 1.48302406e-01 -4.63899314e-01 -6.80528581e-02
-8.22622180e-01 5.08258283e-01 -9.59574401e-01 7.62863815e-01
2.28240401e-01 -9.57357705e-01 6.06623471e-01 -2.24211589e-01
-1.57129794e-01 5.49329996e-01 -6.97788119e-01 1.54365349e+00
-1.93859473e-01 3.69619101e-01 -2.59469807e-01 -9.85900283e-01
7.48816013e-01 2.94139922e-01 6.59497201e-01 -2.01096401e-01
-2.69588023e-01 2.09321246e-01 1.59920976e-01 -4.86031264e-01
-2.76258230e-01 -6.85316930e-03 3.41481328e-01 3.70274544e-01
-8.83773318e-04 9.50264707e-02 2.90652394e-01 6.45398259e-01
1.52269459e+00 -4.76682931e-02 3.32678974e-01 -4.67489213e-01
2.51387298e-01 4.18671817e-01 7.67829776e-01 1.01431847e+00
4.91229177e-01 1.88210592e-01 7.38993347e-01 -2.32893437e-01
-6.52864099e-01 -8.22846889e-01 -2.30121911e-01 4.56885099e-01
-3.47658880e-02 -8.93179655e-01 -2.83517301e-01 -9.08409178e-01
2.81623453e-01 8.74980330e-01 -8.57524812e-01 -4.20118570e-02
-2.02101544e-01 -9.84565616e-01 4.55962926e-01 1.57252029e-01
-3.12315673e-02 -6.83287859e-01 7.87002891e-02 2.63062656e-01
-5.34385517e-02 -1.21765113e+00 -3.13093930e-01 7.48457238e-02
-9.25108850e-01 -1.46051908e+00 -4.76792365e-01 -8.02557826e-01
7.62262523e-01 -4.54276465e-02 1.63765347e+00 -7.01311156e-02
-7.14526653e-01 3.39864008e-02 -8.31290334e-02 -7.48086572e-01
-2.40497813e-01 -5.16914278e-02 1.04947358e-01 -5.95645428e-01
6.31438732e-01 -5.54061115e-01 -6.47855103e-01 2.39568025e-01
-6.91564202e-01 1.89816535e-01 8.94804478e-01 7.45233893e-01
8.41568649e-01 -6.53586909e-03 9.71815228e-01 -1.29634011e+00
9.88997221e-01 -1.14813387e+00 -4.41424549e-01 4.81127918e-01
-9.56751466e-01 1.99935272e-01 4.68048841e-01 -4.01274085e-01
-5.03801048e-01 -3.34874719e-01 2.73891985e-01 -2.03660980e-01
-1.46629974e-01 1.09319973e+00 -2.85754561e-01 1.03641622e-01
7.52291381e-01 -2.44828075e-01 -2.21954316e-01 -5.74022949e-01
5.47735870e-01 2.26532608e-01 2.19240695e-01 -8.95981550e-01
6.73900545e-01 1.23422511e-01 6.66913748e-01 -5.69318891e-01
-1.02568245e+00 -2.22000152e-01 -3.80649239e-01 3.33005190e-01
5.56185126e-01 -8.32046032e-01 -1.06431925e+00 -5.56505263e-01
-1.44729412e+00 3.97959128e-02 -2.15217903e-01 6.54209316e-01
-6.24687113e-02 -6.20586388e-02 -3.90808642e-01 -2.33788356e-01
-3.29691410e-01 -8.59054089e-01 7.97900677e-01 7.34708384e-02
-6.68064713e-01 -1.34277487e+00 3.60290229e-01 2.25455314e-01
-5.73680550e-03 6.03114486e-01 1.53104031e+00 -9.84328985e-01
-6.31948352e-01 8.92389864e-02 -2.94629931e-01 -4.49246168e-01
3.65013927e-01 -9.33227688e-02 -5.50804079e-01 1.53792948e-01
-5.22565246e-01 -1.69990107e-01 6.62757456e-01 4.96290326e-01
1.35877442e+00 -7.08386660e-01 -1.21090305e+00 2.42903367e-01
1.15845096e+00 3.16751808e-01 3.17521125e-01 6.97036684e-02
6.83969200e-01 6.40145361e-01 1.38608636e-02 1.27197191e-01
3.43290716e-01 6.38212740e-01 -8.63996297e-02 -3.55340719e-01
-1.92342728e-01 -4.30798113e-01 -1.92894548e-01 3.86011392e-01
-3.07413876e-01 -4.36614752e-01 -1.10868943e+00 4.97270942e-01
-1.75209296e+00 -8.45250189e-01 -3.28707635e-01 1.94503570e+00
1.61930954e+00 2.27911845e-01 -7.65193254e-02 -6.12919867e-01
7.65694916e-01 -4.24705833e-01 -9.41409945e-01 -2.82216966e-02
-5.57460129e-01 7.08667517e-01 2.96335310e-01 6.42571330e-01
-5.25731027e-01 9.89950299e-01 6.61899281e+00 8.19838345e-01
-7.89407492e-01 -6.84758574e-02 7.63980865e-01 -2.34540105e-02
-8.21936965e-01 2.67549127e-01 -9.14969504e-01 1.81294814e-01
8.79471898e-01 -7.57833421e-01 3.73378620e-02 4.80800718e-01
4.01448309e-01 4.90809232e-01 -1.69519556e+00 8.15753698e-01
-2.38725934e-02 -2.20392895e+00 5.52412450e-01 2.43771493e-01
8.89382422e-01 -5.72513528e-02 -2.95316279e-01 -6.82750791e-02
8.29826176e-01 -1.45989645e+00 -1.16341310e-02 8.49557698e-01
7.44860828e-01 -1.82072774e-01 4.36515391e-01 1.16147652e-01
-5.34395933e-01 3.72640401e-01 -1.58017814e-01 8.48951638e-02
-3.67772162e-01 1.01967740e+00 -1.34608150e+00 9.90450978e-01
2.49803737e-01 1.11453831e+00 -5.47768950e-01 1.01347995e+00
-3.63808632e-01 8.79053473e-01 1.60665791e-02 -3.52929413e-01
1.66685700e-01 4.79892939e-02 5.97074687e-01 1.37307620e+00
1.40096307e-01 3.56756300e-01 2.16571450e-01 1.53789508e+00
-6.96915746e-01 2.05661312e-01 -9.34380651e-01 -5.01993179e-01
6.66585147e-01 1.09029555e+00 -4.35981601e-01 -2.76443630e-01
-2.87814409e-01 3.93582016e-01 2.52057277e-02 6.27751946e-01
-4.29713696e-01 -3.90144318e-01 4.64332998e-01 3.61299992e-01
-3.08551580e-01 3.66632566e-02 -3.09895635e-01 -1.03116453e+00
-3.61481488e-01 -8.01102281e-01 5.19587398e-01 -8.30712318e-01
-1.64031923e+00 4.24807042e-01 2.66096462e-02 -8.37016284e-01
-7.15799034e-02 -4.94518995e-01 -5.96748829e-01 1.08879483e+00
-1.39572537e+00 -9.40276682e-01 1.62894905e-01 4.20793295e-01
3.43479037e-01 -1.58838168e-01 8.48941088e-01 2.03502312e-01
-5.17422616e-01 3.69964212e-01 -2.43148506e-01 -2.20411047e-01
7.76931107e-01 -1.19749117e+00 4.47775245e-01 2.49807477e-01
1.65141195e-01 1.25600290e+00 7.14199603e-01 -1.11399603e+00
-1.61473930e+00 -1.21499872e+00 1.26471114e+00 -1.06314802e+00
9.98811066e-01 -4.18465823e-01 -8.48665178e-01 5.03741086e-01
-2.16748416e-02 -2.06168845e-01 1.19200468e+00 5.11856019e-01
-4.20005471e-01 3.64637733e-01 -8.07195187e-01 8.71091902e-01
1.62186086e+00 -2.17100456e-01 -6.65563107e-01 8.86228740e-01
9.14922297e-01 -1.79701112e-02 -9.63101268e-01 4.40440714e-01
1.83686152e-01 5.65997958e-02 1.09152722e+00 -1.63518834e+00
8.07429969e-01 -3.06278318e-01 3.43051970e-01 -1.29800320e+00
-5.81858993e-01 -9.82368529e-01 -2.50937283e-01 8.72697473e-01
1.21191704e+00 -2.58405387e-01 8.07756841e-01 7.78594255e-01
1.99630763e-02 -1.11824965e+00 -5.71521401e-01 -6.84393764e-01
2.85330474e-01 -4.88159098e-02 5.82596600e-01 1.67516792e+00
4.54458207e-01 8.52372587e-01 6.73608854e-02 1.15381978e-01
7.31625795e-01 3.39662373e-01 3.58289361e-01 -1.76303148e+00
-2.48116642e-01 -6.04975939e-01 -2.45434925e-01 -7.24025309e-01
5.11266112e-01 -1.35022342e+00 -5.05455971e-01 -2.03990006e+00
5.92635691e-01 -5.61527014e-01 -4.18552458e-01 8.12427342e-01
-3.60282153e-01 -2.36599132e-01 -4.76525813e-01 2.75459647e-01
-6.48578107e-01 8.24883953e-02 1.28616154e+00 -4.15940404e-01
-2.12202564e-01 -2.84495682e-01 -1.37023413e+00 5.54097891e-01
5.18637955e-01 -4.72522050e-01 -5.52473366e-01 -7.33508542e-02
6.43004954e-01 1.10770360e-01 4.49186951e-01 -1.60786942e-01
7.66083479e-01 -7.41643846e-01 3.56495410e-01 -9.84415561e-02
-3.43202770e-01 -2.36269012e-01 3.46609682e-01 3.62613529e-01
-8.55291009e-01 -1.57499507e-01 3.28843653e-01 7.30322242e-01
2.59939916e-02 3.12064290e-02 1.04308069e-01 -4.12149966e-01
-2.97577590e-01 3.96143079e-01 3.79873067e-02 4.36394840e-01
7.37116992e-01 1.50920242e-01 -9.44981515e-01 -1.93439886e-01
-8.06770384e-01 5.63904047e-01 9.56230536e-02 6.79048359e-01
6.81550682e-01 -1.00523973e+00 -1.07932067e+00 -5.71778476e-01
3.27683121e-01 -8.66343752e-02 -2.21273765e-01 5.98052144e-01
-4.68077026e-02 8.02493513e-01 3.95035684e-01 -3.99073586e-02
-9.46333945e-01 8.59357238e-01 -4.09687636e-03 -3.18386346e-01
-4.27060872e-01 9.60311234e-01 4.41740125e-01 -1.41871229e-01
2.44984105e-01 -1.22361377e-01 -2.42330104e-01 -1.44303784e-01
2.74261981e-01 6.40523434e-02 -1.87517405e-02 -2.18362920e-02
-7.12912858e-01 3.39924991e-01 -3.13525319e-01 2.05098212e-01
1.57501197e+00 5.74723661e-01 -6.40134513e-01 2.88188398e-01
9.20660555e-01 3.67429614e-01 -2.99004972e-01 -2.84577131e-01
3.52085352e-01 -5.40001132e-02 1.23566747e-01 -1.48803353e+00
-7.14367747e-01 2.76617438e-01 -2.29009122e-01 -1.24302749e-02
4.90457147e-01 4.89471614e-01 2.93120235e-01 6.59945250e-01
2.68806756e-01 -7.95574665e-01 -2.48224735e-01 1.48007393e-01
1.05549073e+00 -9.73039985e-01 4.01479453e-01 -6.85484111e-01
-2.87271410e-01 1.07617581e+00 4.99678016e-01 5.11999130e-01
8.45427871e-01 2.06715062e-01 -3.91165882e-01 -8.22700799e-01
-1.15018463e+00 5.38703501e-02 5.19358158e-01 3.61448646e-01
7.67588913e-01 3.97624820e-02 -6.61068797e-01 9.07517850e-01
-6.31620288e-02 4.00394291e-01 3.82207423e-01 7.55528212e-01
1.31098226e-01 -1.48106778e+00 -1.34535164e-01 8.56949449e-01
-5.03866136e-01 -7.13332772e-01 -1.17683637e+00 6.15195274e-01
4.58517730e-01 9.17170703e-01 -3.59883398e-01 -4.25762683e-01
2.55772054e-01 4.98261070e-03 3.07689995e-01 -1.15035093e+00
-2.62715757e-01 -1.19913831e-01 2.91856438e-01 -1.79995835e-01
-5.98169744e-01 -3.24246407e-01 -1.16042876e+00 -1.53636172e-01
-3.48565042e-01 8.94141197e-01 3.52207303e-01 9.14611220e-01
1.04690313e+00 6.40352488e-01 3.30974787e-01 1.05610073e-01
2.97531992e-01 -5.04150331e-01 -3.49455297e-01 1.95504010e-01
3.31281847e-03 -5.85457623e-01 -2.66538173e-01 2.95347959e-01] | [8.71737003326416, 8.334962844848633] |
9572e8bb-3e8c-4e4b-bef2-bdf153dbdf58 | spectral-algorithms-optimally-recover | 2203.11847 | null | https://arxiv.org/abs/2203.11847v2 | https://arxiv.org/pdf/2203.11847v2.pdf | Spectral Algorithms Optimally Recover Planted Sub-structures | Spectral algorithms are an important building block in machine learning and graph algorithms. We are interested in studying when such algorithms can be applied directly to provide optimal solutions to inference tasks. Previous works by Abbe, Fan, Wang and Zhong (2020) and by Dhara, Gaudio, Mossel and Sandon (2022) showed the optimality for community detection in the Stochastic Block Model (SBM), as well as in a censored variant of the SBM. Here we show that this optimality is somewhat universal as it carries over to other planted substructures such as the planted dense subgraph problem and submatrix localization problem, as well as to a censored version of the planted dense subgraph problem. | ['Colin Sandon', 'Elchanan Mossel', 'Julia Gaudio', 'Souvik Dhara'] | 2022-03-22 | null | null | null | null | ['stochastic-block-model'] | ['graphs'] | [ 3.99901807e-01 3.34820181e-01 -2.61363924e-01 1.41974598e-01
-5.64385533e-01 -6.63732648e-01 4.73521143e-01 6.52551427e-02
2.52570957e-01 1.07262731e+00 5.56047596e-02 -7.25842535e-01
-8.44845712e-01 -7.62866795e-01 -6.21757627e-01 -8.08828890e-01
-8.06801796e-01 5.94452739e-01 2.84295529e-01 -1.37880826e-02
1.06358483e-01 4.93054956e-01 -1.16953850e+00 -1.31768808e-01
7.84459412e-01 3.00546855e-01 1.81943059e-01 7.80197680e-01
1.06569931e-01 5.12068629e-01 -1.12662040e-01 -3.91162872e-01
2.21922100e-01 -4.53182489e-01 -9.78907049e-01 3.12685758e-01
4.19840544e-01 2.40971088e-01 -5.61201513e-01 1.18592203e+00
2.78103471e-01 -1.38286740e-01 7.92798579e-01 -1.61648560e+00
-3.86125892e-01 1.06230688e+00 -1.07410216e+00 2.41769895e-01
3.06583345e-01 -7.66413391e-01 1.10619056e+00 -7.99776316e-01
5.97288370e-01 1.04613543e+00 8.51789713e-01 2.43698746e-01
-1.67485821e+00 -2.70370364e-01 3.74316913e-03 4.81894493e-01
-1.55340576e+00 -2.05268130e-01 5.00970006e-01 -4.38838273e-01
4.29517299e-01 5.39209843e-01 5.00451207e-01 8.91507685e-01
-1.43072724e-01 9.97647524e-01 1.32107341e+00 -8.24192405e-01
2.48844206e-01 -3.44565690e-01 3.18867505e-01 8.10272455e-01
9.69047666e-01 1.33627355e-01 -2.82700419e-01 -5.81670284e-01
5.76545060e-01 -5.85332960e-02 -3.95387977e-01 -6.49195254e-01
-1.02555203e+00 1.17133701e+00 9.93358791e-02 3.17342311e-01
4.13045101e-02 1.30421802e-01 1.32119134e-01 3.83157074e-01
3.57457399e-01 -5.73302293e-03 -1.89247191e-01 1.94940194e-01
-1.38325965e+00 2.07470089e-01 1.18747866e+00 1.24156523e+00
6.94985509e-01 -1.32824987e-01 5.65511361e-02 5.76856077e-01
1.76658496e-01 3.56279224e-01 -2.58237928e-01 -1.03825438e+00
5.02968490e-01 8.20374042e-02 3.13724950e-02 -9.39123988e-01
-4.65034842e-01 -5.78004003e-01 -1.21434152e+00 -1.34793654e-01
7.08307445e-01 1.17420712e-02 -7.34952927e-01 1.74106979e+00
1.80174783e-01 4.53899711e-01 -2.60268539e-01 4.25518364e-01
3.10216725e-01 3.68299007e-01 -4.38622981e-01 -5.18231928e-01
9.88400161e-01 -7.39692926e-01 -4.27376032e-01 -1.21243149e-01
5.32298446e-01 -5.52123785e-01 3.81553441e-01 6.24321282e-01
-8.31578791e-01 -5.38242646e-02 -9.29100752e-01 3.70339006e-01
1.00375544e-02 -9.81109291e-02 8.87598336e-01 9.92656767e-01
-1.43460464e+00 6.05241776e-01 -7.04439223e-01 -1.04989147e+00
3.93950701e-01 3.24820936e-01 -2.29436606e-01 -5.63953578e-01
-8.78377736e-01 5.11349797e-01 2.56513923e-01 1.53632760e-01
-9.61398840e-01 -1.42462790e-01 -4.93186116e-01 8.55525434e-02
7.05532193e-01 -5.28212845e-01 6.39291227e-01 -6.66249633e-01
-7.19858229e-01 9.45115626e-01 -3.15145999e-01 -6.10356152e-01
2.77622253e-01 2.93528974e-01 -1.70545995e-01 1.07077144e-01
1.59748554e-01 -5.34162074e-02 4.53073263e-01 -9.77274001e-01
-3.62033218e-01 -5.03790975e-01 1.69825628e-01 -3.42043459e-01
-6.47002086e-02 3.51698287e-02 4.10531685e-02 -7.04637229e-01
5.11595547e-01 -1.11794972e+00 -5.03242254e-01 -3.90043795e-01
-6.67912006e-01 -1.05581470e-01 5.99480271e-01 -5.90189993e-01
1.15380859e+00 -1.79859483e+00 4.68742698e-01 6.14270151e-01
4.09500718e-01 -2.52922386e-01 -1.65165827e-01 1.15696168e+00
-4.06247288e-01 1.70984372e-01 -3.88033926e-01 2.61434540e-02
9.37052909e-03 3.55002195e-01 -2.07052037e-01 8.99420977e-01
-1.95061445e-01 8.01168978e-01 -8.80779088e-01 -4.75600511e-01
-2.57453412e-01 -3.15544546e-01 -4.17333663e-01 -3.29390258e-01
2.61952728e-03 1.78793713e-01 -2.28953421e-01 6.76871598e-01
7.27315068e-01 -7.36168802e-01 6.06983662e-01 5.66584527e-01
-2.39639282e-02 -2.29704604e-02 -1.50846767e+00 1.39085364e+00
-1.38800358e-02 7.28964150e-01 6.06146097e-01 -1.50832629e+00
6.15026295e-01 4.16188538e-01 4.58442897e-01 1.27923623e-01
-1.57562420e-01 1.33384675e-01 -3.17994148e-01 -3.77121985e-01
2.41671413e-01 -1.38799101e-01 -1.16067119e-01 5.97361088e-01
-3.15419525e-01 4.61902767e-01 3.09412777e-01 7.30130911e-01
1.53000259e+00 -3.21284354e-01 5.96324265e-01 -9.03610528e-01
2.95329928e-01 1.64186489e-02 3.29677820e-01 1.42709434e+00
6.36169910e-02 7.30210185e-01 7.74951756e-01 1.86345607e-01
-1.14492488e+00 -1.40031981e+00 -2.97413439e-01 8.52278948e-01
-4.46761772e-02 -4.10757750e-01 -6.47179008e-01 -4.83346403e-01
4.38375659e-02 2.70103544e-01 -6.54745519e-01 -1.28846288e-01
-2.08616972e-01 -9.03257787e-01 3.54850918e-01 4.08335865e-01
2.15428129e-01 -4.10168201e-01 1.51727200e-01 1.60630003e-01
-3.00463945e-01 -9.74606454e-01 -4.03120756e-01 4.09253150e-01
-1.19727790e+00 -1.29111314e+00 -7.70708263e-01 -8.37509274e-01
5.06965399e-01 4.05200094e-01 1.01052308e+00 1.31730661e-01
-3.62544030e-01 3.36553127e-01 -4.45864111e-01 3.10571808e-02
-3.06655556e-01 2.80709416e-01 1.05318144e-01 -1.76172610e-02
1.40658081e-01 -8.23801041e-01 -1.20328188e-01 2.56074727e-01
-7.44319320e-01 3.21062021e-02 3.81721646e-01 7.60076165e-01
8.33889097e-02 1.65597022e-01 5.11089265e-01 -1.11972976e+00
2.19291627e-01 -8.68839204e-01 -8.98366213e-01 3.22950184e-01
-7.91254461e-01 -7.47422827e-03 3.24167341e-01 -1.29377767e-01
-4.29894567e-01 1.67624354e-01 6.84840158e-02 -2.28188887e-01
1.32681444e-01 7.64962614e-01 -3.53838295e-01 -1.58860624e-01
4.07187343e-01 2.58217394e-01 -1.84114516e-01 -6.05692625e-01
2.81259418e-01 7.36760676e-01 2.36981928e-01 -5.71286142e-01
1.05500054e+00 6.58976614e-01 6.00072801e-01 -1.06766212e+00
-8.09670031e-01 -6.33733988e-01 -8.21859360e-01 -1.92248717e-01
5.60792804e-01 -7.33801663e-01 -3.56358558e-01 2.81680495e-01
-9.22451675e-01 -1.31066158e-01 1.82181969e-02 3.94143552e-01
-6.83059096e-01 8.10159266e-01 -5.77898860e-01 -1.16480172e+00
3.38627011e-01 -5.41126788e-01 1.02723777e+00 -2.51273185e-01
5.01881652e-02 -1.23778450e+00 2.18616232e-01 2.87324339e-01
4.05943654e-02 4.72637862e-01 1.19264281e+00 -6.07895255e-01
-8.13871682e-01 -1.60428867e-01 -2.72468776e-01 -6.18581437e-02
-1.57089710e-01 -1.41336575e-01 -5.34003377e-01 -6.40371680e-01
-1.62703380e-01 7.67848417e-02 1.02728009e+00 6.68674052e-01
7.95277953e-01 -4.35841024e-01 -8.39650869e-01 5.47366202e-01
1.49766254e+00 -2.54412800e-01 8.03313255e-01 1.01707600e-01
5.64242661e-01 4.88991618e-01 2.07560137e-01 1.62717059e-01
1.78043574e-01 6.55185699e-01 3.07828635e-01 4.98691648e-02
1.67450011e-01 -2.48788387e-01 7.22821616e-03 8.83273065e-01
-2.66460888e-02 -3.46185416e-01 -7.62249231e-01 8.44129145e-01
-2.14489317e+00 -1.37840700e+00 -8.75595927e-01 2.35154700e+00
6.17185056e-01 -1.47000581e-01 5.53829551e-01 4.33799475e-01
1.27514529e+00 9.61855352e-02 2.95981187e-02 -1.55354753e-01
-3.91924143e-01 4.41947579e-01 8.40789974e-01 5.12862325e-01
-9.52135146e-01 6.26627684e-01 7.72062302e+00 1.06110334e+00
-2.06470758e-01 3.12537432e-01 3.92178357e-01 1.11789308e-01
-2.48447135e-01 6.02695644e-01 -5.48226237e-01 2.25474507e-01
9.85627294e-01 -2.65025049e-01 6.34063065e-01 6.28687620e-01
1.58997089e-01 -3.25201899e-01 -9.95096087e-01 6.91683412e-01
8.76984373e-02 -1.19894195e+00 -5.61168313e-01 5.61350644e-01
1.11277723e+00 -3.56164835e-02 -4.50156748e-01 -1.63233563e-01
5.58659852e-01 -1.02557135e+00 1.88150480e-01 2.87028909e-01
6.10560715e-01 -6.37447417e-01 3.62006575e-01 5.86212456e-01
-1.10464287e+00 -1.43739609e-02 -4.48808342e-01 -3.38199288e-01
3.46784778e-02 9.90910769e-01 -5.96507668e-01 7.77510762e-01
3.20073128e-01 5.66339076e-01 -3.82978648e-01 1.28221989e+00
-4.74314280e-02 1.15231717e+00 -5.15368700e-01 -1.07257657e-01
1.23870850e-01 -4.82426316e-01 7.09076166e-01 8.77037764e-01
2.93305248e-01 6.90017715e-02 1.39483675e-01 8.04357231e-01
8.15086290e-02 -2.31118992e-01 -6.83806717e-01 -9.98597220e-02
4.35991228e-01 1.09573221e+00 -1.16020012e+00 1.88801512e-02
-5.48268855e-01 8.13085794e-01 2.35483557e-01 6.07206404e-01
-6.04659379e-01 -2.82701254e-01 3.73244464e-01 4.68365043e-01
8.90171826e-01 -5.11490166e-01 -1.16595879e-01 -1.18865502e+00
-2.36641869e-01 -7.17691839e-01 6.53437257e-01 -3.90001178e-01
-1.53885627e+00 5.33674657e-03 3.63531023e-01 -7.21803546e-01
-2.27747913e-02 -5.24211228e-01 -6.54718935e-01 8.28746617e-01
-8.05382490e-01 -1.03092420e+00 6.03157207e-02 6.68264389e-01
-8.09675083e-02 -4.17132229e-02 4.99806136e-01 3.32581282e-01
-5.71119070e-01 1.79994002e-01 8.37362289e-01 3.78638022e-02
4.88046110e-02 -1.43770838e+00 4.14113343e-01 1.19275200e+00
6.77527666e-01 4.18768257e-01 9.14732516e-01 -7.07374692e-01
-1.59634531e+00 -1.09158969e+00 8.72374177e-01 -3.67153376e-01
9.03130710e-01 -7.17497408e-01 -6.90606654e-01 8.17020655e-01
6.77688094e-03 -1.64014861e-01 4.54843074e-01 4.83999491e-01
-2.10890278e-01 2.33976692e-02 -8.81078660e-01 4.41550791e-01
1.58790064e+00 -4.35962647e-01 -3.20112318e-01 9.54753280e-01
3.06028545e-01 -7.16097234e-03 -6.56325996e-01 1.18478097e-01
4.38033909e-01 -1.04058564e+00 7.32172847e-01 -7.98413515e-01
1.63962007e-01 -8.21511969e-02 -2.33636573e-01 -7.86938787e-01
-6.97551072e-01 -8.98758888e-01 -1.43989041e-01 1.25348699e+00
3.73399615e-01 -5.93099117e-01 1.06277800e+00 -2.45573074e-01
2.12591976e-01 -5.22092164e-01 -1.32346463e+00 -1.19207656e+00
1.42469183e-01 -2.54891485e-01 3.32220942e-01 8.63835454e-01
7.25487173e-02 2.49511302e-01 -5.35228968e-01 3.43171984e-01
1.18255448e+00 4.47613865e-01 5.23067713e-01 -1.49499953e+00
-7.27666378e-01 -3.57355624e-01 -5.86064577e-01 -1.04792953e+00
2.36889064e-01 -1.38951242e+00 -1.85179174e-01 -1.78301394e+00
6.77525163e-01 -5.39558530e-01 -8.52675587e-02 1.69192821e-01
-8.37171227e-02 3.10296208e-01 -5.23638465e-02 3.12468797e-01
-4.97250706e-01 -1.14851065e-01 6.27120674e-01 -5.48655502e-02
2.19320476e-01 6.02469683e-01 -7.81672001e-01 3.07681262e-01
4.23569173e-01 -7.99606621e-01 -1.91094905e-01 -1.32381469e-01
4.51841265e-01 4.83342737e-01 6.15152597e-01 -7.58023202e-01
3.43783975e-01 -2.04663500e-01 5.17664291e-03 -5.99038541e-01
2.91337315e-02 -5.76088250e-01 6.47225559e-01 6.62806749e-01
9.44042671e-03 -3.34791914e-02 -2.09237039e-01 1.08999658e+00
2.31370896e-01 -8.42471004e-01 5.37797153e-01 -8.25828165e-02
-3.01872909e-01 3.80040348e-01 -4.82209444e-01 3.11113119e-01
1.09543252e+00 -4.41296309e-01 -2.74015307e-01 -7.24797726e-01
-9.67762470e-01 3.61844689e-01 4.61237460e-01 -1.20150737e-01
2.54476935e-01 -1.21862555e+00 -9.13047075e-01 -4.04445119e-02
-1.39020234e-01 -4.13754672e-01 3.25463410e-03 1.41832435e+00
-2.56753653e-01 5.14359117e-01 3.04425746e-01 -6.97006702e-01
-1.32476592e+00 7.92442918e-01 8.76533166e-02 -2.08747789e-01
-5.60275793e-01 8.53055179e-01 2.19821315e-02 1.89443201e-01
1.12377934e-01 1.21390358e-01 3.79774332e-01 -8.62395242e-02
1.69211045e-01 6.58887863e-01 -1.46646872e-01 -2.92869687e-01
-6.87591314e-01 4.05505240e-01 2.90026248e-01 8.39071348e-02
1.35342872e+00 -4.21102643e-01 -4.46676850e-01 5.31439602e-01
1.16822898e+00 2.00078577e-01 -4.97444272e-01 -2.34124377e-01
4.79919225e-01 -2.89063215e-01 -1.49892777e-01 -3.64385575e-01
-8.89619589e-01 6.18514180e-01 1.04631729e-01 9.41783547e-01
9.82575297e-01 5.79272449e-01 3.70348781e-01 1.26636803e-01
7.85195708e-01 -5.65035403e-01 -5.19467711e-01 3.75062972e-01
5.13683677e-01 -7.40462840e-01 2.89507896e-01 -7.19119787e-01
-8.09223130e-02 1.00597227e+00 -1.92496777e-01 -1.42694026e-01
8.52454007e-01 4.64070976e-01 -8.18311214e-01 -1.13332905e-01
-7.68388212e-01 -6.38012171e-01 6.54118657e-02 8.01158369e-01
6.72130287e-02 4.89488840e-02 -4.91057605e-01 3.03367764e-01
2.35222969e-02 -9.39708054e-02 8.68054867e-01 6.59894466e-01
-5.56433916e-01 -1.11770630e+00 -5.60517132e-01 9.16127682e-01
-1.09617010e-01 -2.54348397e-01 -6.87747478e-01 8.75611067e-01
-2.75719129e-02 9.67457592e-01 -2.98658699e-01 -1.30521119e-01
-9.61163566e-02 4.95435968e-02 8.98274481e-01 -8.62040401e-01
-7.57199526e-02 1.95256844e-01 2.91017205e-01 -1.20574035e-01
-5.89588821e-01 -1.08236635e+00 -5.44843495e-01 -6.09185636e-01
-8.30974817e-01 6.37806058e-01 1.63648844e-01 1.04639924e+00
-1.21083139e-02 7.07373917e-02 6.46548331e-01 -2.99745321e-01
-7.16512203e-01 -1.01653302e+00 -1.38452041e+00 -2.56318897e-01
2.48056442e-01 -5.80042303e-01 -7.77069747e-01 -1.17840990e-01] | [6.87750768661499, 5.143777370452881] |
95cf1dd5-a535-40d3-bf37-564a9ef37495 | predicting-gene-expression-from-network | 2005.03961 | null | https://arxiv.org/abs/2005.03961v2 | https://arxiv.org/pdf/2005.03961v2.pdf | A Graph Feature Auto-Encoder for the Prediction of Unobserved Node Features on Biological Networks | Motivation: Molecular interaction networks summarize complex biological processes as graphs, whose structure is informative of biological function at multiple scales. Simultaneously, omics technologies measure the variation or activity of genes, proteins, or metabolites across individuals or experimental conditions. Integrating the complementary viewpoints of biological networks and omics data is an important task in bioinformatics, but existing methods treat networks as discrete structures, which are intrinsically difficult to integrate with continuous node features or activity measures. Graph neural networks map graph nodes into a low-dimensional vector space representation, and can be trained to preserve both the local graph structure and the similarity between node features. Results: We studied the representation of transcriptional, protein-protein and genetic interaction networks in E. Coli and mouse using graph neural networks. We found that such representations explain a large proportion of variation in gene expression data, and that using gene expression data as node features improves the reconstruction of the graph from the embedding. We further proposed a new end-to-end graph feature auto-encoder which is trained on the feature reconstruction task, and showed that it performs better at predicting unobserved node features than auto-encoders that are trained on the graph reconstruction task before learning to predict node features. When applied to the problem of imputing missing data in single-cell RNAseq data, our graph feature auto-encoder outperformed a state-of-the-art imputation method that does not use protein interaction information, showing the benefit of integrating biological networks and omics data using graph representation learning. | ['Ramin Hasibi', 'Tom Michoel'] | 2020-05-08 | null | null | null | null | ['graph-reconstruction'] | ['graphs'] | [ 3.92306983e-01 2.67145634e-01 -3.60473841e-01 -2.50196844e-01
-2.43440289e-02 -6.02813303e-01 3.78685802e-01 7.37229645e-01
6.92137107e-02 7.33334124e-01 2.78112441e-01 -2.66608953e-01
-3.24930072e-01 -9.93843734e-01 -1.08526731e+00 -8.46740484e-01
-3.60118032e-01 5.85920632e-01 -3.94749463e-01 5.46440631e-02
-2.53711313e-01 5.91360807e-01 -1.08495438e+00 2.87412167e-01
5.17514348e-01 7.08541214e-01 -1.04371428e-01 8.26824605e-01
-3.77486199e-01 6.59593165e-01 -2.40656286e-01 -2.42583126e-01
-3.27461623e-02 -5.95847249e-01 -4.64721382e-01 -9.45204049e-02
2.32489854e-01 4.27218229e-01 -7.80124962e-01 7.98166990e-01
3.73135984e-01 -2.41767630e-01 7.37416148e-01 -1.53806782e+00
-9.71970797e-01 3.02506119e-01 -1.43287078e-01 -1.72615111e-01
3.03643256e-01 2.38124449e-02 1.39855218e+00 -6.79307520e-01
9.91058886e-01 1.14780200e+00 8.90520692e-01 3.46814781e-01
-2.03206038e+00 -2.54014313e-01 -4.96168099e-02 8.25976506e-02
-1.31237566e+00 -2.87450284e-01 6.71767533e-01 -6.52439296e-01
1.31901443e+00 2.30036050e-01 9.00629282e-01 1.08941638e+00
8.66663992e-01 3.05471241e-01 6.58546984e-01 -8.69651064e-02
2.17012435e-01 -4.84963804e-01 2.71949321e-01 1.21566880e+00
4.22802567e-01 -4.35635448e-02 -5.01910746e-01 -6.42614841e-01
6.48931086e-01 6.58845127e-01 -3.51049215e-01 -7.04003990e-01
-1.45625246e+00 8.64511132e-01 6.03519797e-01 3.67502421e-01
-4.21815842e-01 4.56435502e-01 4.45673853e-01 4.98572320e-01
6.97930694e-01 6.23879135e-01 -8.05403471e-01 4.33565795e-01
-4.44404423e-01 -3.42646509e-01 1.06065035e+00 6.39051557e-01
1.03444993e+00 -1.01356924e-01 1.74886912e-01 6.17479026e-01
2.29127765e-01 2.45664060e-01 4.51495737e-01 -7.33659804e-01
-5.60123809e-02 9.82182622e-01 -4.91612315e-01 -1.27395654e+00
-7.45163798e-01 -4.62805867e-01 -1.35789156e+00 -1.60862446e-01
3.82391095e-01 -6.77000433e-02 -9.56930280e-01 1.99240077e+00
2.56302536e-01 3.97955924e-01 -5.53810596e-02 4.04117405e-01
8.93878460e-01 6.24075413e-01 -4.78938296e-02 -2.68563122e-01
1.13078117e+00 -5.33285975e-01 -8.72273088e-01 2.83563733e-02
9.06974435e-01 -8.77019092e-02 4.74304289e-01 3.25768329e-02
-6.97653770e-01 -3.07544827e-01 -9.85693395e-01 -1.90245181e-01
-8.45299363e-01 -2.02835828e-01 1.01193011e+00 6.53653592e-02
-1.01031554e+00 1.12147510e+00 -1.02174199e+00 -3.72793078e-01
3.10160965e-01 6.22230828e-01 -1.05941272e+00 -1.08062387e-01
-9.86840069e-01 6.46214306e-01 1.22747004e-01 4.10212763e-02
-5.70165455e-01 -7.85641670e-01 -1.12784767e+00 4.00894344e-01
2.22323701e-01 -1.14047086e+00 1.87254578e-01 -6.79340065e-01
-1.30810463e+00 5.66532671e-01 -4.18087900e-01 -4.52423580e-02
-2.88554728e-01 3.99971962e-01 -3.30720216e-01 -4.57767807e-02
-5.87715358e-02 4.29253608e-01 4.93329614e-01 -7.57931232e-01
3.50738883e-01 -8.19135487e-01 -3.57455671e-01 -1.48242682e-01
-9.81573910e-02 -6.89320326e-01 -1.27535090e-02 -1.63865685e-01
5.02144516e-01 -9.56877112e-01 -3.53074878e-01 2.68182367e-01
-6.31565750e-01 2.30330927e-03 5.21945953e-01 -5.27432740e-01
6.47149026e-01 -1.90521550e+00 7.47224867e-01 2.93278664e-01
7.50021756e-01 -1.77149639e-01 -6.28155172e-01 8.70208859e-01
-6.40834630e-01 4.56223100e-01 -3.00866365e-01 -4.01219837e-02
-9.51266065e-02 3.73650551e-01 2.45687246e-01 8.46724570e-01
4.39062834e-01 1.25503576e+00 -9.23175097e-01 -5.79724275e-02
4.57798243e-02 8.33399415e-01 -4.20726597e-01 2.91465104e-01
-3.01212639e-01 2.94915259e-01 -2.65522480e-01 6.96222544e-01
3.05444717e-01 -8.88451219e-01 7.60225952e-01 -3.75934154e-01
5.10865271e-01 6.90857992e-02 -7.07690656e-01 1.82377911e+00
-2.75163114e-01 5.47575355e-01 2.04582978e-02 -1.38996100e+00
9.79982615e-01 3.38529825e-01 1.02165663e+00 -2.18561053e-01
-2.33423039e-02 -1.64172530e-01 2.96010494e-01 -3.01869631e-01
-3.14382255e-01 -2.67179646e-02 8.76573920e-02 2.05221146e-01
5.22747159e-01 1.37202412e-01 -6.44476414e-02 2.85168797e-01
1.87970769e+00 7.09409639e-03 5.16103983e-01 -3.67406271e-02
3.79892796e-01 -1.98860317e-01 7.25182652e-01 2.42965266e-01
1.61802977e-01 4.23150867e-01 1.11731231e+00 -5.43621659e-01
-9.50360060e-01 -1.00230384e+00 -7.32648745e-03 1.03951454e+00
-3.22315812e-01 -6.69331014e-01 -3.71932209e-01 -7.70714998e-01
4.80211049e-01 1.21300958e-01 -9.69508231e-01 -4.18139160e-01
-8.80392417e-02 -1.02485406e+00 4.59287912e-01 2.61692792e-01
-3.50958198e-01 -6.85539901e-01 4.43209141e-01 3.78270626e-01
4.35377173e-02 -9.27357554e-01 -4.04806942e-01 8.33201051e-01
-9.44800973e-01 -1.53861916e+00 -6.73075765e-02 -6.18192375e-01
1.06239831e+00 1.22842543e-01 1.10148919e+00 3.77013713e-01
-5.51196456e-01 2.32253209e-01 1.57297552e-02 -1.65498585e-01
-5.14320433e-01 -2.78892089e-03 8.74086469e-02 1.68360199e-03
2.90378630e-01 -9.63558137e-01 -3.33116680e-01 7.83437639e-02
-9.04520094e-01 5.35276346e-02 5.64919055e-01 1.43983245e+00
8.42427433e-01 -1.88475400e-01 7.84326911e-01 -1.17390442e+00
4.96572018e-01 -7.11532056e-01 -3.99949074e-01 3.64824682e-01
-6.94120586e-01 5.07210314e-01 9.70924795e-01 -2.22374294e-02
-2.06994817e-01 4.81203824e-01 -6.56224862e-02 -2.23148763e-01
2.14782376e-02 9.25338089e-01 -4.37879324e-01 -1.67147964e-01
4.28589880e-01 1.82010174e-01 5.96729994e-01 -1.95125103e-01
4.87975359e-01 3.42623681e-01 2.11963400e-01 -2.61838168e-01
2.42417753e-01 2.94249356e-01 9.36845779e-01 -8.33280027e-01
-4.57888484e-01 -5.08708000e-01 -1.02728271e+00 2.03932837e-01
7.83634782e-01 -9.07067120e-01 -1.07231343e+00 -3.76245342e-02
-1.09001517e+00 1.91303119e-02 -1.60079747e-01 4.26663756e-01
-6.96755826e-01 4.70139831e-01 -9.09261763e-01 -3.74614507e-01
-2.17496589e-01 -8.62187088e-01 1.16310060e+00 -4.26984221e-01
-3.74156609e-02 -1.39224291e+00 3.97313178e-01 -7.49733299e-02
7.53075480e-02 6.98818862e-01 1.36230528e+00 -8.94705772e-01
-4.73393887e-01 -3.93506706e-01 -1.66204676e-01 6.89631850e-02
5.32359481e-01 1.16797738e-01 -6.85698330e-01 -3.28894049e-01
-4.05962467e-01 -1.39634416e-01 1.03369617e+00 4.33005840e-01
1.20679915e+00 -4.38817561e-01 -6.63028359e-01 8.74721587e-01
1.56212568e+00 -2.99333662e-01 5.60365736e-01 -3.52919430e-01
9.99714196e-01 5.59411407e-01 -2.71824718e-01 4.04597260e-02
6.97934348e-03 5.82195640e-01 6.04757249e-01 -1.91762164e-01
-6.04109429e-02 -5.66856563e-01 3.86775166e-01 1.04167342e+00
3.99575941e-02 -5.32480180e-01 -6.49666071e-01 3.15489411e-01
-2.09002590e+00 -8.56050432e-01 -4.35193419e-01 2.00438952e+00
7.47288942e-01 -3.23665202e-01 -8.66035447e-02 -1.12349160e-01
6.95858359e-01 -1.81571916e-02 -7.48032570e-01 -3.88563514e-01
-3.57728571e-01 9.43671018e-02 5.18911183e-01 4.58448589e-01
-6.69261932e-01 5.55504680e-01 6.65206718e+00 1.13427050e-01
-8.90315413e-01 -8.38125572e-02 4.94143844e-01 2.76788116e-01
-4.49704081e-01 1.15846872e-01 -3.07849288e-01 3.53827745e-01
1.38881230e+00 -2.30994076e-01 5.18611729e-01 5.00988483e-01
4.15329747e-02 5.05746901e-01 -1.59665620e+00 7.26231277e-01
-2.62429826e-02 -1.70195925e+00 -5.53728938e-02 3.66777718e-01
5.44441462e-01 3.80235255e-01 -3.67554486e-01 -6.05617976e-03
6.17963731e-01 -1.34488642e+00 -3.92850667e-01 8.50331724e-01
6.60072863e-01 -4.62455839e-01 8.95905793e-01 3.65767926e-01
-9.23992932e-01 1.67281434e-01 -7.56346762e-01 -1.36591226e-01
-1.03429906e-01 1.02384102e+00 -1.20181322e+00 6.32519007e-01
4.70684916e-02 1.25595105e+00 -4.84431088e-01 5.42365789e-01
-8.84270370e-02 5.30998051e-01 -1.85848385e-01 -5.03493026e-02
-1.16932951e-01 -5.06585777e-01 2.43386313e-01 1.08512866e+00
2.86174268e-01 -1.90442745e-02 2.15207815e-01 9.29786325e-01
-4.02053565e-01 1.05522268e-01 -1.17906201e+00 -6.91527784e-01
9.25851837e-02 1.46801651e+00 -4.86356854e-01 -2.03106895e-01
-5.47521770e-01 1.00009727e+00 7.23307014e-01 4.31468248e-01
-4.51889902e-01 -3.05608392e-01 1.05693269e+00 1.91501066e-01
3.36395297e-03 -3.49250108e-01 -5.11356927e-02 -1.27956963e+00
-4.25135046e-01 -6.12256229e-01 2.71367997e-01 -7.39042640e-01
-1.70633686e+00 2.09462196e-01 -5.89635134e-01 -7.96901762e-01
-2.31041923e-01 -1.03941631e+00 -3.06420237e-01 7.51148880e-01
-1.27010775e+00 -1.18029785e+00 -1.17156804e-01 2.36594766e-01
-1.48819983e-01 -1.61336616e-01 1.39729762e+00 -1.46020232e-02
-7.33342588e-01 2.52139777e-01 6.27982438e-01 1.39234409e-01
4.85508502e-01 -1.36160207e+00 4.76178229e-01 1.97940171e-01
3.98414880e-01 8.58650208e-01 3.97547752e-01 -9.09181833e-01
-2.18068171e+00 -1.42497230e+00 1.06909537e+00 -4.70033735e-01
7.72775173e-01 -7.42538810e-01 -1.10784864e+00 9.83541131e-01
-6.95486516e-02 7.50263333e-01 1.18397164e+00 4.89057809e-01
-5.96928596e-01 9.12742913e-02 -1.01439166e+00 2.89919585e-01
1.29211020e+00 -7.90516138e-01 -1.79829791e-01 7.13609636e-01
7.59220719e-01 4.39636558e-02 -1.50956154e+00 1.90752134e-01
5.61669111e-01 -4.67446625e-01 1.13893843e+00 -1.34189606e+00
4.31507170e-01 -2.71703243e-01 -1.92860931e-01 -1.80355155e+00
-8.47944200e-01 -2.95322984e-01 -2.67104477e-01 6.77432239e-01
6.26342475e-01 -8.09296250e-01 8.64531338e-01 3.29861760e-01
-1.24283522e-01 -6.37086868e-01 -8.54830921e-01 -5.93881130e-01
-1.18820354e-01 1.88568339e-01 5.23092031e-01 1.22908521e+00
4.73135144e-01 8.09174359e-01 -2.88608521e-01 2.56474409e-02
5.24387598e-01 1.28689855e-01 6.57341301e-01 -1.63977218e+00
-4.78868574e-01 -1.05594814e-01 -1.16831255e+00 -5.99597633e-01
6.96803451e-01 -1.60518491e+00 -1.43125907e-01 -1.67748618e+00
3.68352532e-01 9.98946801e-02 -5.71539998e-01 8.07197332e-01
-1.51543528e-01 1.05069960e-02 -2.04982042e-01 -6.37756214e-02
-4.43451375e-01 6.28342032e-01 1.27133989e+00 -5.47615707e-01
1.11175828e-01 -3.68726403e-01 -7.02402353e-01 4.04050022e-01
5.60066462e-01 -5.85060000e-01 -2.46865824e-01 -1.39992699e-01
4.51295674e-01 4.39529687e-01 4.19457257e-01 -5.84350765e-01
3.71576160e-01 -9.46974382e-03 8.67900431e-01 -1.28018618e-01
4.88275021e-01 -9.75389957e-01 7.73742080e-01 6.16267025e-01
-5.18377066e-01 1.92169130e-01 -9.51114222e-02 1.24585497e+00
-1.28623560e-01 2.62759686e-01 2.74404615e-01 -2.36666366e-01
-7.62935206e-02 5.57895720e-01 -2.80646592e-01 -4.37139302e-01
6.75809681e-01 -4.35048044e-02 -5.04396260e-01 -4.03367072e-01
-1.12481582e+00 -8.83531868e-02 5.71215987e-01 1.19350061e-01
7.03291476e-01 -1.30272114e+00 -6.96276546e-01 4.25586551e-01
1.36453599e-01 -4.12566662e-01 -4.67489921e-02 7.46483564e-01
-1.93329617e-01 3.93827975e-01 -2.74371862e-01 -7.14771867e-01
-1.20486164e+00 7.71034181e-01 2.70505160e-01 -2.52763301e-01
-3.70144278e-01 4.29368049e-01 1.86938539e-01 -9.47453618e-01
-2.61545032e-01 -2.41602570e-01 -6.69433996e-02 -1.08137071e-01
1.66981510e-04 9.33733881e-02 1.49203360e-01 -5.20397365e-01
-3.67838740e-01 5.17396629e-01 2.40665451e-01 5.74529350e-01
1.73760355e+00 1.24117084e-01 -7.08559811e-01 6.40036285e-01
1.69658935e+00 -2.05641598e-01 -8.37556362e-01 -1.33656934e-01
-2.86290632e-03 -1.91179976e-01 3.67667351e-04 -7.68533647e-01
-1.02846122e+00 7.96365976e-01 8.35460797e-02 2.33020738e-01
7.12433279e-01 -1.43247703e-02 3.91748786e-01 7.50445724e-01
1.48997128e-01 -5.67425370e-01 -3.46271209e-02 4.68166888e-01
7.53112733e-01 -1.15137637e+00 2.55287707e-01 -7.79230416e-01
-3.18513550e-02 1.27461600e+00 2.48540893e-01 -1.04400203e-01
8.41595829e-01 2.59142518e-01 -3.87481332e-01 -6.31663918e-01
-1.21736538e+00 3.41532868e-03 3.61750990e-01 7.47673392e-01
8.74726176e-01 2.34657750e-01 -1.01507828e-01 5.04622102e-01
1.39379412e-01 2.95453854e-02 5.03284812e-01 5.45988262e-01
-2.04002470e-01 -1.25991666e+00 2.21451133e-01 8.77041280e-01
-3.58537495e-01 -1.19207807e-01 -9.36638057e-01 5.70215344e-01
-4.97420132e-02 7.44473517e-01 -5.08892573e-02 -4.95982289e-01
1.48510203e-01 4.61954176e-01 3.62246752e-01 -7.33625412e-01
-3.82600307e-01 -2.00080737e-01 2.24701598e-01 -7.24901676e-01
-2.49401405e-01 -5.09395182e-01 -1.27342963e+00 -3.57374430e-01
-4.23712671e-01 2.94999957e-01 9.14412975e-01 6.32116735e-01
1.00633681e+00 8.91311884e-01 6.02209926e-01 -4.96292681e-01
-2.99098253e-01 -7.59090006e-01 -8.61280143e-01 5.08582294e-01
3.28168601e-01 -4.80520993e-01 -4.29029584e-01 6.52218983e-02] | [6.058343410491943, 5.806270599365234] |
ec89ab90-1414-4458-ab90-6490820c92d9 | on-the-role-of-the-zero-conditional-mean | 2211.09502 | null | https://arxiv.org/abs/2211.09502v1 | https://arxiv.org/pdf/2211.09502v1.pdf | On the Role of the Zero Conditional Mean Assumption for Causal Inference in Linear Models | Many econometrics textbooks imply that under mean independence of the regressors and the error term, the OLS parameters have a causal interpretation. We show that even when this assumption is satisfied, OLS might identify a pseudo-parameter that does not have a causal interpretation. Even assuming that the linear model is "structural" creates some ambiguity in what the regression error represents and whether the OLS estimand is causal. This issue applies equally to linear IV and panel data models. To give these estimands a causal interpretation, one needs to impose assumptions on a "causal" model, e.g., using the potential outcome framework. This highlights that causal inference requires causal, and not just stochastic, assumptions. | ['Joeri Smits', 'Giovanni Mellace', 'Michael C. Knaus', 'Federico Crudu'] | 2022-11-17 | null | null | null | null | ['econometrics'] | ['miscellaneous'] | [-7.21060187e-02 2.74708539e-01 -9.02688861e-01 -4.51271802e-01
-3.48571032e-01 -6.53054297e-01 6.03956401e-01 -1.71479076e-01
-7.31172934e-02 9.61807549e-01 6.34507120e-01 -1.19825816e+00
-4.24131125e-01 -7.63829291e-01 -1.02906442e+00 -4.46306586e-01
5.38083613e-02 1.06555112e-01 -3.86877120e-01 3.43956798e-01
3.07993203e-01 9.56655890e-02 -7.13511050e-01 -2.45330706e-01
9.13960993e-01 3.22432488e-01 -1.55673608e-01 4.60733414e-01
-3.96792814e-02 1.15245652e+00 -5.55121183e-01 -4.16994035e-01
2.20804989e-01 -4.88439798e-01 -7.86971629e-01 -9.82689261e-02
1.35586157e-01 -7.05956280e-01 -4.48531300e-01 7.02550530e-01
-1.20669305e-01 -1.78033352e-01 1.17204475e+00 -1.29756176e+00
-1.10453486e+00 8.95872235e-01 -5.43888688e-01 3.14900726e-01
4.16050941e-01 1.70465976e-01 1.03568959e+00 -8.03792298e-01
5.42097628e-01 1.59810185e+00 6.70142472e-01 -3.27406302e-02
-1.71792734e+00 -6.48021400e-01 3.49745303e-01 -2.67332762e-01
-9.31632400e-01 -4.64079827e-01 3.64722937e-01 -8.77874255e-01
2.89949447e-01 3.03308278e-01 4.23634201e-01 1.51311457e+00
7.22701728e-01 2.57794708e-01 1.46448672e+00 -5.67977011e-01
-2.92177242e-03 1.14352003e-01 5.18647671e-01 1.03785902e-01
1.11946630e+00 4.93083060e-01 -1.76233113e-01 -4.10240382e-01
1.10416269e+00 -1.00105137e-01 -1.78352609e-01 -4.81189489e-02
-9.29124653e-01 1.16880310e+00 -7.45059848e-02 7.95892179e-02
-7.45548248e-01 7.60092139e-01 2.52480865e-01 3.32281291e-01
1.89528286e-01 2.63701379e-01 -4.91977245e-01 2.22902000e-01
-4.33925450e-01 2.39272311e-01 8.67854595e-01 3.72857988e-01
3.85284096e-01 2.70223945e-01 6.08002208e-02 3.11685920e-01
6.00059807e-01 9.51050580e-01 -5.05557209e-02 -1.18689013e+00
3.12448770e-01 2.55289942e-01 5.69002151e-01 -9.49959159e-01
-4.26329643e-01 -1.61363229e-01 -2.96094954e-01 1.62453607e-01
6.95852041e-01 -7.16498971e-01 -7.44296134e-01 1.94097531e+00
5.32703027e-02 8.45247880e-02 -4.30329442e-02 6.95070148e-01
5.12383401e-01 6.29017830e-01 4.98399884e-01 -5.34978092e-01
1.01258194e+00 2.01951172e-02 -8.77180398e-01 -2.06214711e-01
7.41385520e-01 -4.45259809e-01 9.40464437e-01 6.77786246e-02
-1.13878536e+00 -3.17922793e-02 -4.69215989e-01 1.28498912e-01
2.81089991e-01 -2.78991908e-01 9.92075145e-01 7.95233428e-01
-8.37594450e-01 1.74122289e-01 -5.72939336e-01 -3.49632114e-01
-3.43334645e-01 3.38230014e-01 -8.16350505e-02 2.96844155e-01
-1.05129957e+00 9.02857542e-01 -6.40123487e-02 -1.27866432e-01
-8.18127036e-01 -9.28839147e-01 -5.75740278e-01 4.13901836e-01
3.17148209e-01 -9.80982125e-01 1.22607601e+00 -1.38829350e+00
-8.54240894e-01 3.30334276e-01 -4.16928500e-01 -1.23074062e-01
4.43785429e-01 -9.78248008e-03 -3.75492781e-01 -2.31956750e-01
5.13355672e-01 -2.10384518e-01 4.50243533e-01 -1.60897195e+00
-6.63634896e-01 -3.89610559e-01 1.77261218e-01 -1.20712565e-02
3.05889368e-01 5.64193845e-01 2.63221532e-01 -3.69594157e-01
1.94957629e-01 -9.05246854e-01 -2.19100602e-02 -7.16776371e-01
-7.30446875e-01 -2.27803931e-01 2.49233648e-01 -6.25765085e-01
1.37154853e+00 -1.87196660e+00 -3.63198221e-01 4.99150276e-01
5.02809472e-02 -7.39561319e-01 3.12893212e-01 5.60508132e-01
-8.06143701e-01 7.59445488e-01 -5.07488735e-02 1.95321649e-01
6.87444434e-02 1.21865280e-01 -5.62376320e-01 8.86677444e-01
-5.33668734e-02 8.11865330e-01 -4.09235746e-01 -2.85521150e-01
1.17021143e-01 6.59861490e-02 -5.61039150e-01 -2.51862019e-01
4.78344321e-01 3.47110808e-01 -5.19853413e-01 2.72092491e-01
4.19391185e-01 -1.78607002e-01 6.83728635e-01 4.80665714e-01
-7.43677795e-01 7.21910715e-01 -1.28327489e+00 5.81947923e-01
-2.31736749e-01 7.73821771e-01 -1.30599225e-02 -1.32098997e+00
3.90625805e-01 5.74790061e-01 4.53224033e-01 -5.14485359e-01
1.18494481e-01 2.01263502e-01 3.64950746e-01 -6.15843356e-01
1.84756860e-01 -5.34327865e-01 -1.85857236e-01 4.45022166e-01
-2.50646353e-01 2.98726171e-01 -1.25016347e-01 1.73293442e-01
1.02519286e+00 -3.61171700e-02 3.31216961e-01 -7.42668509e-01
-1.50890008e-01 2.42515296e-01 9.06316519e-01 1.07151735e+00
3.85125697e-01 9.92362276e-02 1.06746078e+00 -1.35106489e-01
-1.16045296e+00 -1.15527105e+00 -4.94326383e-01 7.36538053e-01
-1.89638942e-01 2.97160268e-01 -3.72827798e-01 -3.32633048e-01
5.09120643e-01 1.24748695e+00 -1.00739539e+00 4.39221002e-02
-4.01283562e-01 -9.93675411e-01 4.26244199e-01 3.52713555e-01
-2.09822148e-01 -3.01630825e-01 -3.08317780e-01 1.15756392e-01
-5.77341914e-02 -3.82537574e-01 4.65045273e-02 6.80529401e-02
-8.67999434e-01 -1.05875897e+00 -1.89444333e-01 -2.84354780e-02
7.10678160e-01 3.72083843e-01 9.22955751e-01 3.07223890e-02
5.72327971e-01 5.16267359e-01 -3.47491801e-02 -8.24020922e-01
-4.76946145e-01 -4.02417451e-01 9.44011193e-03 -2.74195462e-01
3.50549906e-01 -5.09448647e-01 -5.41103840e-01 -7.50727281e-02
-4.51274246e-01 -3.13868970e-01 2.73728341e-01 6.77487612e-01
-9.81314927e-02 -2.01897502e-01 9.13465977e-01 -1.11639416e+00
4.03824598e-01 -9.73186076e-01 -7.34316766e-01 3.18800688e-01
-8.87357354e-01 -1.35440603e-01 3.17469448e-01 -4.11196947e-01
-1.51928818e+00 -5.94620883e-01 4.71352637e-01 1.06582887e-01
-1.95597872e-01 1.06563950e+00 -1.44966096e-01 3.55505735e-01
7.11838245e-01 -5.53435504e-01 -3.07370394e-01 -4.93078560e-01
2.98921345e-03 3.71551961e-01 1.11360937e-01 -7.08354473e-01
7.27756798e-01 5.35114884e-01 2.59882212e-01 -3.23970020e-01
-6.16378486e-01 -8.88054743e-02 -2.10042179e-01 8.17474201e-02
1.00636744e+00 -1.21698976e+00 -1.00961602e+00 -2.61697382e-01
-1.11671674e+00 -5.64150631e-01 -3.41637194e-01 1.32949054e+00
-5.45076907e-01 -2.89877534e-01 -5.66026986e-01 -1.42043293e+00
7.07893908e-01 -9.79932845e-01 2.47688070e-01 -2.16146559e-02
-6.73031628e-01 -1.45112276e+00 3.36496718e-03 1.54656783e-01
4.82616387e-02 3.05263698e-01 1.03542101e+00 -4.83348072e-01
-4.41853791e-01 -1.45973578e-01 -2.88991332e-01 -4.47683990e-01
1.69933215e-01 7.03636289e-01 -6.47552848e-01 2.04891205e-01
1.57348529e-01 1.82807565e-01 5.27603686e-01 1.26880765e+00
7.54912853e-01 -7.47657597e-01 -4.25320417e-01 3.81388932e-01
1.58692658e+00 7.96954989e-01 5.03071964e-01 2.34091923e-01
6.55234694e-01 8.78495038e-01 1.28954887e-01 5.12317002e-01
7.73938835e-01 3.52570295e-01 1.43490061e-01 -2.69681185e-01
4.49569643e-01 -6.49522781e-01 3.20104033e-01 3.13248158e-01
-1.12350218e-01 -3.75369012e-01 -1.03376436e+00 6.14779949e-01
-1.86671031e+00 -1.38378000e+00 -1.17061818e+00 2.37801790e+00
8.24776530e-01 -2.20817477e-02 3.71280760e-01 -3.29624236e-01
6.24835968e-01 -2.30033606e-01 -2.13602006e-01 -8.69464219e-01
-2.82196850e-01 -2.16962963e-01 1.19995165e+00 8.60763311e-01
-5.00176013e-01 3.53235483e-01 8.60922718e+00 5.27660437e-02
-8.30400586e-01 1.08031824e-01 9.07094777e-01 1.10886000e-01
-1.13437212e+00 8.23973417e-01 -5.95127404e-01 5.32467544e-01
1.27609336e+00 -6.09151721e-01 -1.44297451e-01 4.90696907e-01
1.15834200e+00 -5.77309251e-01 -1.03282821e+00 5.43836839e-02
-5.75291514e-01 -9.01422918e-01 -2.71693558e-01 7.47879565e-01
9.32435274e-01 -5.14757335e-01 1.45147249e-01 1.63290918e-01
1.33895183e+00 -1.29264569e+00 1.09296107e+00 5.16177297e-01
7.35033393e-01 -6.59915328e-01 6.13652885e-01 3.17704260e-01
-4.12452221e-01 -3.80509883e-01 -2.49738172e-01 -7.33680248e-01
2.86265224e-01 7.22443044e-01 -5.34663916e-01 3.52198243e-01
3.95509928e-01 1.04284905e-01 1.00196498e-02 7.69569039e-01
-4.12265837e-01 1.42462373e+00 -8.51630196e-02 4.80819702e-01
1.66317955e-01 -5.29383659e-01 3.49614024e-01 1.07819378e+00
5.36796272e-01 5.44896901e-01 -3.04040462e-01 9.73529816e-01
1.82288393e-01 3.65271531e-02 -9.28462267e-01 -3.48788537e-02
6.30171835e-01 4.94481027e-01 -3.52730662e-01 -3.63432676e-01
-1.04248691e+00 1.64847493e-01 -1.78469583e-01 8.34369302e-01
-8.07520151e-01 3.35216761e-01 5.82506418e-01 3.55357498e-01
-2.88112104e-01 -2.06295878e-01 -9.52277005e-01 -1.14519596e+00
-2.92797863e-01 -6.29473507e-01 2.71922916e-01 -7.06834733e-01
-8.78339052e-01 -9.52555418e-01 5.81537247e-01 -3.42259020e-01
-4.67209190e-01 -3.82912666e-01 -8.64598751e-01 1.25220418e+00
-1.04001546e+00 -8.10071826e-01 3.59445930e-01 2.06253856e-01
1.43562779e-01 5.76378763e-01 3.69850844e-01 -4.61332425e-02
-7.80329823e-01 1.79313764e-01 2.37729549e-01 -2.60899216e-02
7.43307829e-01 -1.43820512e+00 1.90493286e-01 9.19519544e-01
-3.22989106e-01 1.08065224e+00 1.20032239e+00 -9.97549951e-01
-1.38764918e+00 -5.51752627e-01 1.15061402e+00 -5.51062644e-01
9.57970917e-01 1.73792150e-02 -8.03789735e-01 1.43561530e+00
2.90751874e-01 -6.64052248e-01 6.32722855e-01 7.70455778e-01
-2.40912959e-01 1.34647101e-01 -8.59548509e-01 7.89716005e-01
8.98308754e-01 -3.27981502e-01 -6.19146526e-01 1.94594711e-01
7.83987820e-01 1.17745042e-01 -7.76798308e-01 2.32438356e-01
8.11038196e-01 -9.30646122e-01 8.12137663e-01 -9.00216937e-01
4.86176014e-01 1.24378532e-01 -1.11357749e-01 -1.34783244e+00
-7.79903412e-01 -3.74285012e-01 6.14709139e-01 1.45503867e+00
7.28887260e-01 -1.02096462e+00 3.12660605e-01 1.25032401e+00
2.64920723e-02 -6.60991520e-02 -6.55891895e-01 -8.95354152e-01
7.01156497e-01 -5.18012881e-01 5.95408797e-01 1.66365933e+00
1.98257163e-01 3.16711992e-01 -2.95333117e-01 2.21240267e-01
5.51908255e-01 -1.20269142e-01 7.93689311e-01 -1.61185336e+00
-2.16789022e-01 -2.28672758e-01 1.89192221e-01 -5.12978315e-01
3.91220897e-01 -2.71293819e-01 -2.76885331e-01 -1.57195246e+00
4.98582721e-01 -5.77697754e-01 5.62655739e-02 2.06047714e-01
-1.90000325e-01 -4.01527554e-01 -5.46797365e-02 3.42915952e-01
3.16140413e-01 -5.72174974e-02 7.97032595e-01 3.88259351e-01
-4.45884883e-01 -9.54805166e-02 -1.15449774e+00 1.01231694e+00
8.95755231e-01 -6.77234828e-01 -4.33531314e-01 -3.44910532e-01
6.76576197e-01 9.00042832e-01 9.51843798e-01 6.96849525e-02
-7.17880651e-02 -1.16712284e+00 3.87229949e-01 -6.38905540e-02
-1.03828736e-01 -7.00364590e-01 9.64927495e-01 4.51989084e-01
-3.99440289e-01 3.26466054e-01 6.58859983e-02 2.95645595e-01
2.13653862e-01 -4.15500402e-01 1.70721442e-01 -7.54389837e-02
8.59837979e-02 -1.95073858e-01 -7.27687657e-01 -5.96841052e-02
7.49969244e-01 -2.45532319e-01 -4.40485746e-01 -6.94375217e-01
-6.07046902e-01 4.25893553e-02 7.58387148e-01 -1.85385704e-01
1.44293085e-02 -1.26006067e+00 -8.97560894e-01 -4.76025522e-01
-3.02586585e-01 -6.93050325e-01 7.86312595e-02 1.32148302e+00
-5.24781533e-02 7.75135338e-01 3.45462054e-01 -3.54300812e-02
-8.84766996e-01 7.37734199e-01 2.48629957e-01 1.81744680e-01
-3.38386089e-01 3.50659251e-01 1.14609396e+00 1.82948187e-01
-4.85290200e-01 -1.15622498e-01 7.76214898e-02 1.61635593e-01
1.72632098e-01 6.07783258e-01 -7.05412567e-01 -5.96307874e-01
-3.47156018e-01 1.59673408e-01 4.93975580e-01 -7.25237012e-01
1.02207541e+00 -7.83331454e-01 -2.55085945e-01 1.14320040e+00
8.57635915e-01 5.62740147e-01 -1.11594057e+00 1.60002738e-01
2.58106083e-01 -4.78143364e-01 -9.32808369e-02 -5.81588626e-01
-5.64978242e-01 4.96854514e-01 -7.83296302e-02 5.67634344e-01
6.30353808e-01 -1.12151705e-01 -2.81357199e-01 -1.72732651e-01
1.55400485e-01 -8.52509618e-01 -6.26612604e-01 5.81318475e-02
9.21880603e-01 -1.25734401e+00 1.37265056e-01 -4.78854984e-01
-5.22571027e-01 7.20823228e-01 1.81956831e-02 -3.76598120e-01
5.48458695e-01 2.18207523e-01 -1.18825980e-01 -1.97606996e-01
-1.05434310e+00 -4.30579744e-02 9.69041288e-02 3.54215145e-01
9.40137267e-01 6.13579094e-01 -1.25150466e+00 4.28383499e-01
-3.25861573e-01 -1.10815160e-01 1.05927706e+00 3.61566246e-01
-3.01218241e-01 -8.63006830e-01 -9.31386769e-01 5.75774252e-01
-1.08630133e+00 -4.79242206e-02 -4.60987180e-01 1.46657932e+00
-1.43261537e-01 1.27792299e+00 4.55271363e-01 1.46296158e-01
2.95648515e-01 1.53019458e-01 4.12023887e-02 -5.16247809e-01
-3.55941236e-01 4.94919151e-01 4.85160112e-01 -2.90491462e-01
-4.61060554e-01 -1.13771033e+00 -1.13011181e+00 -1.04373932e+00
-5.10099411e-01 1.63656145e-01 5.29421568e-01 9.68829513e-01
-1.61283478e-01 2.60601789e-01 8.54445636e-01 -8.44240263e-02
-4.91353780e-01 -8.05516541e-01 -7.32883453e-01 2.55654454e-02
3.85484129e-01 -7.65130460e-01 -7.95569897e-01 2.78545823e-02] | [7.993809223175049, 5.247262477874756] |
3ba9eda1-83c7-4a7e-a1be-d715234afeab | speech-watermarking-a-solution-for | 2203.02275 | null | https://arxiv.org/abs/2203.02275v2 | https://arxiv.org/pdf/2203.02275v2.pdf | Speech watermarking: an approach for the forensic analysis of digital telephonic recordings | In this article, the authors discuss the problem of forensic authentication of digital audio recordings. Although forensic audio has been addressed in several articles, the existing approaches are focused on analog magnetic recordings, which are less prevalent because of the large amount of digital recorders available on the market (optical, solid state, hard disks, etc.). An approach based on digital signal processing that consists of spread spectrum techniques for speech watermarking is presented. This approach presents the advantage that the authentication is based on the signal itself rather than the recording format. Thus, it is valid for usual recording devices in police-controlled telephone intercepts. In addition, our proposal allows for the introduction of relevant information such as the recording date and time and all the relevant data (this is not always possible with classical systems). Our experimental results reveal that the speech watermarking procedure does not interfere in a significant way with the posterior forensic speaker identification. | ['Martin Hagmueller', 'Jose Juan Lucena-Molina', 'Marcos Faundez-Zanuy'] | 2022-02-23 | null | null | null | null | ['speaker-identification'] | ['speech'] | [ 4.31571573e-01 -7.37333670e-02 -3.58848572e-02 2.59724140e-01
-4.64213312e-01 -5.88135958e-01 2.86256760e-01 6.04589581e-01
-6.10082448e-01 8.23488593e-01 -3.23003978e-01 -4.25054580e-01
-2.04731867e-01 -6.41574919e-01 -3.19291025e-01 -7.02436447e-01
7.10593760e-02 -3.39403632e-03 4.28305775e-01 9.07663442e-03
4.59005147e-01 8.17205071e-01 -1.59647822e+00 -2.21059144e-01
5.56720078e-01 1.06637645e+00 1.82598121e-02 3.57391685e-01
-2.31521383e-01 4.95165557e-01 -1.00754964e+00 -3.98760110e-01
1.12682700e-01 -4.68867362e-01 -3.03960085e-01 8.14954340e-02
-1.32354811e-01 -5.28518200e-01 -3.30189139e-01 1.35688126e+00
4.29010361e-01 -9.10566598e-02 5.36931574e-01 -8.95122409e-01
-3.30432445e-01 7.02235758e-01 -7.78152943e-01 2.97824889e-01
6.13656402e-01 -1.36895761e-01 3.61591786e-01 -4.48233724e-01
4.85395640e-01 7.88497746e-01 4.12047595e-01 1.52511269e-01
-9.44143176e-01 -7.54431546e-01 -3.59388918e-01 6.18418872e-01
-1.51509035e+00 -4.29311365e-01 1.20783675e+00 -4.27759111e-01
2.11392716e-01 1.96995080e-01 5.41191578e-01 9.77563083e-01
1.37900993e-01 2.57311076e-01 1.20510149e+00 -8.58764887e-01
3.18483084e-01 5.16696513e-01 5.49867868e-01 2.79898167e-01
6.14566088e-01 -2.20494732e-01 -4.83044893e-01 -2.07792521e-01
4.49727088e-01 -2.17648596e-01 -6.79865420e-01 -2.33641900e-02
-7.87193120e-01 4.06956077e-01 -4.59663182e-01 9.37896907e-01
-5.23328662e-01 -2.73764610e-01 6.63337409e-01 2.38961309e-01
6.63307831e-02 1.30187690e-01 4.05419260e-01 -3.24320525e-01
-1.18968928e+00 -7.30471127e-03 7.05394626e-01 3.57091933e-01
3.44180107e-01 3.07409227e-01 4.14733648e-01 2.39052072e-01
3.98192078e-01 6.86882198e-01 5.59385836e-01 -4.82953131e-01
4.46158975e-01 6.56923056e-02 1.97042823e-01 -1.31761658e+00
-6.63833544e-02 -8.17688629e-02 -6.12361968e-01 2.97784060e-01
6.14682496e-01 -1.69827268e-01 -2.00762495e-01 1.25977826e+00
2.23586157e-01 2.75207758e-01 1.55193642e-01 5.93582630e-01
4.81413931e-01 6.17758930e-01 -2.32602015e-01 -5.60761154e-01
1.47791040e+00 -1.17663689e-01 -1.52383924e+00 2.35820994e-01
-7.42132291e-02 -1.06634033e+00 6.67442203e-01 8.89109433e-01
-8.16119730e-01 -4.37793076e-01 -1.30587280e+00 2.69573569e-01
-3.43021363e-01 7.22682551e-02 1.30513743e-01 1.35084105e+00
-7.01724529e-01 4.11122471e-01 -5.67495286e-01 -1.76161319e-01
-1.01230092e-01 3.57685238e-01 -3.41255307e-01 4.17822838e-01
-1.05678141e+00 8.19874823e-01 3.63672137e-01 2.15793386e-01
-2.89739400e-01 -5.18975556e-02 -5.11241376e-01 2.78314739e-01
2.13575795e-01 9.31805074e-02 1.04391658e+00 -6.61146879e-01
-1.76575780e+00 6.60268962e-01 1.27017647e-01 -6.67039573e-01
6.00423932e-01 -7.51474351e-02 -9.35682476e-01 9.30824578e-01
-2.80761302e-01 -3.05719018e-01 1.28000760e+00 -8.00563931e-01
-6.73014373e-02 -4.54135329e-01 -6.26865327e-02 -5.00611961e-01
-6.29249632e-01 3.22536498e-01 -7.29335193e-03 -5.85729063e-01
1.05075598e-01 -6.58590496e-01 2.68109798e-01 -9.30402279e-02
-3.61967146e-01 2.94506371e-01 1.08509076e+00 -9.36308980e-01
1.39200258e+00 -2.55365753e+00 -3.11185241e-01 5.75682282e-01
-1.64556012e-01 6.54074728e-01 5.19679606e-01 6.86293781e-01
-6.44609109e-02 -1.53331697e-01 -1.56273767e-01 -1.56348929e-01
-1.70460626e-01 -3.73869687e-01 -5.86218238e-01 8.66458774e-01
-2.44946375e-01 -1.29237529e-02 -3.86447996e-01 -5.75495064e-01
4.41860110e-01 6.50221944e-01 -4.20267582e-02 -3.29423130e-01
3.00469488e-01 5.61521351e-01 -5.56689620e-01 4.86100405e-01
8.68371487e-01 2.72098184e-01 3.63942713e-01 -1.65128782e-01
-4.21856314e-01 3.01933587e-01 -1.62039137e+00 1.23389184e+00
-1.29963070e-01 5.92415571e-01 5.08830190e-01 -9.68751013e-01
1.16968441e+00 8.81288648e-01 4.67775345e-01 -2.87613779e-01
3.91095549e-01 4.20212030e-01 -1.30482152e-01 -6.86974347e-01
6.72210038e-01 -6.87328428e-02 2.39041686e-01 6.14491642e-01
-2.72459656e-01 8.00648630e-02 2.46483058e-01 -1.45334601e-01
6.90554082e-01 -8.12991336e-02 3.79790604e-01 -1.06033787e-01
8.19978476e-01 -4.84797776e-01 2.38417134e-01 1.98932439e-01
-1.21511795e-01 2.66430676e-01 3.35506201e-01 3.02900255e-01
-7.27048457e-01 -6.52647913e-01 -2.60481715e-01 -1.21695437e-02
2.58490860e-01 -3.12150180e-01 -9.52322066e-01 -1.31729707e-01
-2.05415174e-01 3.14688981e-01 -1.09122708e-01 -1.06156670e-01
-4.98008847e-01 -2.08512694e-01 9.43796456e-01 8.58300552e-02
3.89402539e-01 -9.73670244e-01 -8.46196771e-01 3.51174355e-01
-1.42713189e-01 -1.17749834e+00 -8.98547024e-02 -5.87648079e-02
-9.52841699e-01 -1.08570099e+00 -7.03973472e-01 -4.51114297e-01
3.79264832e-01 2.75707096e-01 -2.59552500e-03 -1.18650801e-01
-1.62222117e-01 4.16643769e-01 -4.35714900e-01 -4.87746894e-01
-6.01298034e-01 -2.07147986e-01 1.19194105e-01 6.05205417e-01
4.75348145e-01 -7.30382681e-01 -1.49460465e-01 1.12414658e-01
-1.01335144e+00 -7.29809463e-01 2.95097888e-01 2.28622898e-01
1.12910554e-01 3.60583454e-01 7.89581537e-01 -5.84123969e-01
8.61002624e-01 -2.08486170e-01 -7.73794353e-01 1.19656809e-01
-4.22210932e-01 -2.61269093e-01 7.48704076e-01 -3.78748834e-01
-9.46212351e-01 -2.08594918e-01 -2.17079207e-01 -2.31221929e-01
-3.71914893e-01 3.77391309e-01 -3.45098972e-01 -2.52158195e-01
4.87483799e-01 3.89281064e-01 1.94798291e-01 -8.76977026e-01
-1.63236916e-01 1.15735030e+00 5.58164418e-01 -4.02543694e-01
7.91732192e-01 4.72894460e-01 7.81850666e-02 -1.39102769e+00
1.98796764e-01 -3.83786142e-01 -3.38818014e-01 -3.17201078e-01
7.77503192e-01 -3.48863453e-01 -1.01395810e+00 6.28517926e-01
-1.28241289e+00 5.73172331e-01 -2.51412898e-01 1.03819764e+00
-2.60957271e-01 1.22409654e+00 -6.44772828e-01 -1.38246202e+00
-2.85673261e-01 -1.21739662e+00 5.18938720e-01 1.13742806e-01
-1.94098890e-01 -7.30778456e-01 -2.52745658e-01 2.94102073e-01
2.52702594e-01 2.06747338e-01 7.79751301e-01 -4.94728476e-01
-5.23904145e-01 -6.43386841e-01 1.53349847e-01 4.21514213e-01
1.90882757e-01 2.89942443e-01 -1.07225657e+00 -1.62237406e-01
6.47486746e-01 1.50853127e-01 3.79064888e-01 3.92593384e-01
9.28299606e-01 2.10336279e-02 -2.00064346e-01 6.64695352e-02
1.17549253e+00 6.47531748e-01 8.87029409e-01 4.25536782e-01
3.69575806e-02 7.74537742e-01 5.27277231e-01 5.64731181e-01
-1.60469308e-01 8.66515458e-01 2.80961007e-01 3.07742357e-01
1.24578267e-01 9.61947627e-03 5.44116557e-01 8.22490573e-01
-1.69889465e-01 -2.81122774e-01 -5.94886184e-01 3.12498480e-01
-1.30725729e+00 -1.26171637e+00 -4.16192383e-01 2.64149022e+00
4.97227579e-01 1.15790695e-01 2.00641334e-01 1.16667926e+00
1.22635794e+00 5.89648709e-02 4.77332808e-02 -3.83677006e-01
-1.56342894e-01 4.76208836e-01 4.60538596e-01 3.90285641e-01
-9.00906920e-01 9.79142711e-02 5.66912031e+00 8.40322018e-01
-1.43968880e+00 1.59699202e-01 -2.44146228e-01 1.67521268e-01
-1.03653431e-01 -8.76190215e-02 -5.80893219e-01 8.79932404e-01
8.66181910e-01 -1.28336027e-01 6.97944313e-02 4.57851142e-01
4.10947293e-01 -3.65004122e-01 -5.91124356e-01 1.17391717e+00
6.24239519e-02 -8.01412463e-01 -1.81514010e-01 3.53286088e-01
-1.21005207e-01 -9.39948022e-01 2.25486338e-01 -3.72968405e-01
-9.45444763e-01 -4.29265082e-01 8.84292245e-01 3.40390265e-01
5.64862430e-01 -8.40059400e-01 8.06554914e-01 2.36168161e-01
-9.22756732e-01 1.81874055e-02 -1.87999949e-01 1.46202128e-02
5.70215523e-01 8.34576964e-01 -7.00625300e-01 7.80120194e-01
3.24264765e-01 1.55704752e-01 -1.26998171e-01 1.29400396e+00
-2.62447387e-01 8.37252617e-01 -4.71696526e-01 -3.56126577e-02
-1.21006593e-01 -3.93065810e-01 8.64203453e-01 9.78075147e-01
6.80063069e-01 -1.52030334e-01 -2.44370654e-01 5.50872028e-01
2.93755919e-01 5.38365960e-01 -7.89660156e-01 -2.06194356e-01
6.39741123e-01 8.13597322e-01 -8.82271826e-01 -9.66609344e-02
-3.85748118e-01 5.60451508e-01 -7.51734078e-01 8.09950158e-02
-6.23460293e-01 -8.14969659e-01 3.19816396e-02 5.41729808e-01
4.70681489e-01 -4.08977568e-01 1.36524085e-02 -8.82585049e-01
3.24733824e-01 -9.52126861e-01 4.38963622e-02 -2.54078776e-01
-8.52743030e-01 4.90591407e-01 9.02731344e-03 -1.62639368e+00
-3.27595979e-01 -4.23095942e-01 -2.78258443e-01 7.28537977e-01
-1.31314993e+00 -5.28292716e-01 1.55268729e-01 7.53055930e-01
-1.19624259e-02 -2.57513881e-01 8.82931173e-01 7.96268761e-01
-4.30241913e-01 4.35737222e-01 3.33350748e-01 7.64891431e-02
7.20645249e-01 -6.07514083e-01 -1.14768468e-01 9.70408857e-01
1.26335055e-01 8.82891715e-01 9.05100405e-01 -5.67092061e-01
-1.26242304e+00 9.09964442e-02 8.24515939e-01 5.15451550e-01
7.56745577e-01 -7.69875944e-02 -1.06419981e+00 3.03625643e-01
3.88464659e-01 -4.77515966e-01 9.48877573e-01 -3.39111298e-01
-1.49989411e-01 -2.31858179e-01 -1.36063123e+00 2.27193698e-01
1.39841452e-01 -8.49265039e-01 -8.48151624e-01 -1.08570576e-01
1.07367694e-01 2.36766879e-02 -6.42728269e-01 -1.21101178e-01
5.75338542e-01 -1.02482212e+00 6.68515742e-01 1.63943052e-01
-1.79508686e-01 -5.86614251e-01 5.83957545e-02 -6.35533631e-01
3.37784201e-01 -7.58793890e-01 -1.45983566e-02 1.60070670e+00
4.23951633e-02 -9.29537654e-01 5.55527449e-01 1.94161594e-01
2.96646953e-01 1.23206176e-01 -1.16712213e+00 -9.43893850e-01
-6.62468672e-01 -2.86205202e-01 5.13867378e-01 9.85408187e-01
3.93278450e-01 9.79894623e-02 -5.54914653e-01 2.52144068e-01
7.76544690e-01 1.42782986e-01 5.05355537e-01 -1.44595921e+00
-6.12061322e-01 -2.32463688e-01 -8.53122473e-01 -6.96191847e-01
8.71292055e-02 -4.59974945e-01 -4.39679086e-01 -9.36632097e-01
-4.11282510e-01 -1.89929694e-01 -2.29857162e-01 -7.58657232e-02
2.77369231e-01 1.53630242e-01 3.16831321e-01 1.73364997e-01
8.38092193e-02 2.56918877e-01 7.47449696e-01 -7.00012967e-02
-1.16899520e-01 2.72386700e-01 -1.28908277e-01 7.51799345e-01
7.85177410e-01 -5.52570939e-01 -4.93529111e-01 -6.22024834e-02
-2.43995771e-01 2.91890830e-01 2.05508158e-01 -1.38255382e+00
3.40948731e-01 2.47749761e-01 -1.85967758e-01 -6.01142347e-01
5.55897772e-01 -1.30755615e+00 3.77290606e-01 4.48112190e-01
9.68922675e-02 7.49283060e-02 8.94386694e-02 5.45397222e-01
-5.58929086e-01 -9.50537443e-01 7.04975009e-01 1.70888156e-01
-2.32078806e-01 -4.70332086e-01 -8.33313465e-01 -6.68170393e-01
1.02498436e+00 -7.36700535e-01 -2.89847314e-01 -4.87953156e-01
-6.94001317e-01 -6.23129010e-01 4.31468487e-01 -1.58837155e-01
4.41819876e-01 -6.88544631e-01 -1.33125275e-01 1.43583640e-01
-2.85700113e-01 -6.12852752e-01 4.38379765e-01 9.99359548e-01
-8.00200939e-01 4.08230960e-01 -5.19240081e-01 -2.44131699e-01
-1.42962992e+00 7.42498338e-01 -3.45875323e-02 1.47501780e-02
-6.71121538e-01 4.97771651e-02 -4.74634230e-01 5.37022889e-01
2.91208863e-01 -1.72023103e-02 -5.19888520e-01 4.61038560e-01
9.43278313e-01 6.90389276e-01 3.27775180e-01 -6.66275620e-01
-3.66234869e-01 6.71770751e-01 1.24339305e-01 -5.33549070e-01
1.11872983e+00 -1.83074325e-01 -1.97619364e-01 5.34622133e-01
9.01162386e-01 8.46260130e-01 -4.45660859e-01 -9.23978351e-03
4.00173850e-02 -4.97514576e-01 8.04917887e-02 -1.14215143e-01
-8.16727161e-01 1.09100568e+00 5.65818906e-01 6.89962566e-01
1.08993685e+00 -6.93510771e-01 7.96145916e-01 3.06597531e-01
7.13616014e-01 -1.10965562e+00 -3.09310257e-01 -7.10271597e-02
4.86691087e-01 -4.68330413e-01 7.77172670e-02 -5.45381546e-01
-1.85740232e-01 1.39026463e+00 -1.43526658e-01 1.20758519e-01
7.18331099e-01 4.70718622e-01 4.78276014e-02 1.14006728e-01
1.24575503e-01 -6.73308410e-03 -2.56149530e-01 6.91170752e-01
3.78635257e-01 -2.51688838e-01 -1.08831131e+00 6.32838428e-01
-8.04012492e-02 2.40129858e-01 9.84306872e-01 1.12381577e+00
-4.35860336e-01 -1.46319067e+00 -1.16322422e+00 -6.03941716e-02
-1.10077691e+00 2.38539234e-01 -2.13630378e-01 5.93836069e-01
-4.94342037e-02 1.18891871e+00 -2.10742280e-01 -8.52969736e-02
3.16165477e-01 3.33860815e-01 5.32487988e-01 -3.05172980e-01
-4.75036204e-01 1.82871997e-01 -7.62010664e-02 -3.92337851e-02
-6.22959733e-01 -8.70990455e-01 -1.31634915e+00 -4.31662649e-01
-5.24309278e-01 5.82324862e-01 1.17060316e+00 6.33629084e-01
5.75621575e-02 2.96011090e-01 4.60963190e-01 -6.30932629e-01
-5.95928609e-01 -8.02147925e-01 -1.19325578e+00 -2.27910578e-02
4.22981381e-01 -6.62804246e-01 -4.34912413e-01 7.10686594e-02] | [14.989420890808105, 5.558487892150879] |
316ceff0-eb24-436f-9f62-ca4ba80e542b | signal-reconstruction-from-quantized-noisy | 2201.03114 | null | https://arxiv.org/abs/2201.03114v1 | https://arxiv.org/pdf/2201.03114v1.pdf | Signal Reconstruction from Quantized Noisy Samples of the Discrete Fourier Transform | In this paper, we present two variations of an algorithm for signal reconstruction from one-bit or two-bit noisy observations of the discrete Fourier transform (DFT). The one-bit observations of the DFT correspond to the sign of its real part, whereas, the two-bit observations of the DFT correspond to the signs of both the real and imaginary parts of the DFT. We focus on images for analysis and simulations, thus using the sign of the 2D-DFT. This choice of the class of signals is inspired by previous works on this problem. For our algorithm, we show that the expected mean squared error (MSE) in signal reconstruction is asymptotically proportional to the inverse of the sampling rate. The samples are affected by additive zero-mean noise of known distribution. We solve this signal estimation problem by designing an algorithm that uses contraction mapping, based on the Banach fixed point theorem. Numerical tests with four benchmark images are provided to show the effectiveness of our algorithm. Various metrics for image reconstruction quality assessment such as PSNR, SSIM, ESSIM, and MS-SSIM are employed. On all four benchmark images, our algorithm outperforms the state-of-the-art in all of these metrics by a significant margin. | ['Animesh Kumar', 'Mohak Goyal'] | 2022-01-09 | null | null | null | null | ['ms-ssim'] | ['computer-vision'] | [ 7.65590787e-01 -5.11063710e-02 -4.40880917e-02 1.09517880e-01
-6.03260100e-01 -3.10547113e-01 2.06156462e-01 -1.13591716e-01
-5.16756356e-01 6.17682040e-01 -8.12165365e-02 -2.84557581e-01
-1.06086373e-01 -3.68088126e-01 -7.09633827e-01 -9.42634106e-01
-4.73495305e-01 -6.01771235e-01 -3.12707108e-03 -2.78338119e-02
3.64590973e-01 1.61869660e-01 -1.16110504e+00 -2.27807209e-01
4.81741667e-01 1.58544624e+00 1.57554597e-01 7.93761492e-01
4.40059781e-01 4.30336207e-01 -6.67183936e-01 -2.28654608e-01
6.53632820e-01 -7.88930416e-01 -4.14993912e-01 1.54428661e-01
-2.30342597e-01 -3.90484571e-01 -3.33205163e-01 1.56876934e+00
5.48085034e-01 -9.43793803e-02 4.78342861e-01 -8.84483874e-01
-3.64885390e-01 3.27549070e-01 -6.69262350e-01 3.47595990e-01
3.43145877e-01 -5.16558252e-02 5.62059999e-01 -9.26463187e-01
3.48010749e-01 8.03570688e-01 7.64628232e-01 6.30637184e-02
-1.06955159e+00 -4.37981546e-01 -5.65501928e-01 4.04502392e-01
-1.67074084e+00 -4.76144642e-01 7.34333873e-01 -2.93549985e-01
2.21515283e-01 3.52310926e-01 3.90258729e-01 5.34605026e-01
4.26298827e-01 3.48252952e-01 1.16078687e+00 -7.34960616e-01
2.70768285e-01 3.31621128e-03 -1.20283224e-01 5.41146100e-01
3.19777042e-01 2.17356563e-01 -3.88547271e-01 -2.65479386e-01
1.02006960e+00 -1.14371844e-01 -1.00079739e+00 -1.52970791e-01
-1.29743075e+00 5.51802278e-01 1.41141266e-01 4.46158469e-01
-5.81854582e-01 3.45337600e-01 1.83652177e-01 6.50799155e-01
2.17762396e-01 -1.19496599e-01 1.13320621e-02 -9.34055969e-02
-8.13589931e-01 -2.02965870e-01 7.31654942e-01 8.41529489e-01
3.74424517e-01 1.60600543e-01 -1.67847335e-01 5.26150584e-01
3.79924506e-01 6.44242764e-01 7.31880665e-01 -8.78273308e-01
2.44720176e-01 -2.26328522e-01 2.92386889e-01 -1.16178036e+00
2.11572144e-02 -6.94892764e-01 -1.13342464e+00 3.22097465e-02
4.19128269e-01 -1.14307292e-01 -4.03052777e-01 1.52584314e+00
1.24981917e-01 4.78271633e-01 2.42855772e-01 1.02971625e+00
2.94083238e-01 8.13260078e-01 -6.40502214e-01 -7.39414871e-01
1.22961032e+00 -4.85143930e-01 -9.90357935e-01 7.05110058e-02
-7.61318654e-02 -9.22938049e-01 7.04368174e-01 6.39717638e-01
-1.31568539e+00 -6.59954309e-01 -1.47893655e+00 4.80537266e-01
3.70932698e-01 4.32576954e-01 -2.52106100e-01 7.98388064e-01
-7.87879229e-01 6.36124432e-01 -5.92953801e-01 1.10492073e-01
3.84475999e-02 6.60392344e-02 -2.36269444e-01 -1.70639865e-02
-9.91331875e-01 5.16655564e-01 8.36577192e-02 2.60461241e-01
-5.28982580e-01 -5.36526561e-01 -5.32346368e-01 1.17048137e-01
1.00762121e-01 -2.71741241e-01 1.04607153e+00 -1.13594234e+00
-1.51184940e+00 5.56222022e-01 -1.09056786e-01 -6.66621506e-01
5.55276453e-01 8.01768899e-02 -5.33947468e-01 4.94621128e-01
-7.12061897e-02 -5.83369061e-02 1.30769944e+00 -8.66897047e-01
-2.71342278e-01 -1.92933977e-01 -3.72333914e-01 -1.23145148e-01
-1.86680973e-01 -2.50101924e-01 -2.88994551e-01 -9.10462499e-01
4.60807353e-01 -6.84526563e-01 -2.02621117e-01 2.43618369e-01
-1.91104606e-01 6.00736141e-01 7.10812271e-01 -8.22131872e-01
1.32876670e+00 -2.66692185e+00 -1.15201809e-01 3.12579662e-01
-2.66924471e-01 1.57810777e-01 1.35039300e-01 5.58089793e-01
-2.09754288e-01 -1.55834377e-01 -5.20752847e-01 -4.65572514e-02
-3.07654798e-01 -2.54108403e-02 -3.91734451e-01 1.15716541e+00
-3.81849557e-02 2.13591516e-01 -6.82238638e-01 -1.08731888e-01
-1.01836594e-02 4.62404013e-01 -2.50365436e-01 8.01662803e-02
4.28657413e-01 4.77577955e-01 -1.89133957e-01 1.19529895e-01
9.89332318e-01 -2.09239304e-01 2.42997140e-01 -4.38283563e-01
-1.81397200e-01 1.12484753e-01 -1.55041718e+00 1.41833591e+00
-5.23690641e-01 7.65806973e-01 2.71346688e-01 -1.01117754e+00
7.61877418e-01 6.73206091e-01 3.51591259e-01 -5.89413881e-01
3.13682079e-01 4.64959681e-01 1.03853635e-01 -4.79525834e-01
5.08271493e-02 -3.16951841e-01 2.93439001e-01 3.34301651e-01
-8.23504478e-02 3.66802886e-03 1.52592659e-01 -1.08321838e-01
9.24322784e-01 -4.13994312e-01 8.13496470e-01 -3.02724421e-01
8.70253026e-01 -6.93274498e-01 4.40593183e-01 5.39981544e-01
-1.80330604e-01 5.83012283e-01 5.40695846e-01 6.28459156e-02
-1.06078744e+00 -8.51111829e-01 -4.43446606e-01 1.74998075e-01
3.27297568e-01 -1.90039784e-01 -8.69498968e-01 -1.68166593e-01
-2.98561156e-01 4.31479454e-01 -2.27931648e-01 -6.24749884e-02
-3.33740354e-01 -6.33964360e-01 4.45276380e-01 -3.40857469e-02
8.62933099e-01 -5.60202301e-01 -9.10087764e-01 9.29840207e-02
-1.95301548e-01 -1.27311206e+00 -6.21037543e-01 -1.70930196e-03
-7.60198295e-01 -1.09044993e+00 -1.21900249e+00 -7.53668904e-01
8.58381867e-01 3.15860569e-01 5.31907558e-01 1.24983881e-02
-1.73978880e-01 4.54940230e-01 -3.97182822e-01 -6.23080358e-02
-4.57958788e-01 -7.54011512e-01 -1.66723683e-01 6.16833270e-01
-4.26610470e-01 -6.15841687e-01 -8.60562503e-01 5.91317594e-01
-1.19223523e+00 -2.01764733e-01 7.13004470e-01 9.15146112e-01
7.29888856e-01 4.62718189e-01 3.57794225e-01 -2.85222322e-01
7.44683743e-01 -3.12909395e-01 -7.71974266e-01 -1.13863155e-01
-4.13381875e-01 1.56214848e-01 9.18597698e-01 -5.09019852e-01
-4.57035959e-01 -1.26981456e-02 4.18503210e-02 -3.60905200e-01
3.25340003e-01 5.31160414e-01 2.44755261e-02 -3.61536622e-01
5.27794123e-01 5.72053194e-01 1.24539196e-01 -2.71820068e-01
5.77652007e-02 9.25767362e-01 8.99433613e-01 -2.00372547e-01
7.21841693e-01 5.66013515e-01 3.42502475e-01 -1.14309752e+00
-2.06954151e-01 -5.02518117e-01 5.37759662e-02 -1.35540500e-01
3.36198956e-01 -6.06260657e-01 -8.14132512e-01 6.95828915e-01
-1.18473005e+00 1.56670049e-01 -2.44702712e-01 7.75272608e-01
-7.36822724e-01 8.45066130e-01 -4.72506166e-01 -1.04961538e+00
-5.20190597e-01 -1.18518758e+00 5.92515767e-01 -1.32859632e-01
1.12960421e-01 -6.81432009e-01 -1.62015274e-01 -2.96172559e-01
4.57804680e-01 3.59611303e-01 6.41241431e-01 -1.64913908e-01
-3.66125941e-01 -4.53183830e-01 -1.06757857e-01 8.53328347e-01
9.45449099e-02 -4.08486515e-01 -6.39794827e-01 -3.89293462e-01
9.38503623e-01 1.67552352e-01 4.56198394e-01 5.10865748e-01
1.19864035e+00 -6.26143754e-01 8.48257616e-02 6.01556361e-01
1.71172607e+00 2.25492239e-01 8.40219378e-01 -2.40725856e-02
-2.46358216e-01 1.86986312e-01 6.07438505e-01 8.96489322e-01
-2.81273961e-01 7.61062205e-01 4.58507717e-01 1.75517440e-01
-6.61095902e-02 1.09332904e-01 4.43517447e-01 8.21541369e-01
6.95646182e-02 -2.19539225e-01 -4.94962841e-01 4.10054326e-01
-1.50093246e+00 -8.37196529e-01 -3.00525606e-01 2.73219299e+00
7.27457166e-01 1.38664052e-01 -9.00840163e-02 8.90864432e-01
7.27206469e-01 7.74955377e-02 -4.38295364e-01 -5.45605458e-02
-1.81197345e-01 3.01722348e-01 9.53130424e-01 4.27560419e-01
-7.64964581e-01 -1.74574733e-01 6.02535105e+00 9.64180410e-01
-1.33936405e+00 1.85461175e-02 6.57857358e-01 3.07405502e-01
3.46723199e-02 5.87731553e-03 -3.14162582e-01 8.41465890e-01
8.70706141e-01 -2.55718440e-01 6.72166109e-01 4.68867123e-01
2.83427984e-01 -2.16147929e-01 -7.87182510e-01 1.27390230e+00
8.13557878e-02 -1.00370586e+00 -3.82377654e-01 1.08092301e-01
6.33339107e-01 -4.03695732e-01 3.70946050e-01 -4.99297976e-01
-5.97594857e-01 -8.01889479e-01 9.79287624e-01 5.31041801e-01
1.02321827e+00 -7.32468724e-01 7.99793184e-01 4.26487833e-01
-1.05315804e+00 -1.42598033e-01 -2.68667489e-01 1.92548390e-02
1.15433857e-01 1.04310644e+00 -6.25673652e-01 5.38765371e-01
3.34944457e-01 4.95822847e-01 1.28510058e-01 1.24509144e+00
-1.65782616e-01 6.94041014e-01 -4.30186719e-01 -8.97036791e-02
6.79615363e-02 -3.79681557e-01 7.31895208e-01 8.89981091e-01
8.41232181e-01 3.02290559e-01 -2.39450902e-01 6.45576537e-01
-4.48536985e-02 1.29713759e-01 -2.86932707e-01 9.25281420e-02
4.61881906e-01 8.03026199e-01 -4.20333564e-01 -8.21123272e-02
-3.35148364e-01 1.01022422e+00 -7.04880416e-01 3.76498699e-01
-6.44079387e-01 -8.22734892e-01 3.24096859e-01 2.09363773e-01
6.10564411e-01 -3.85007858e-01 -2.95441657e-01 -6.65082753e-01
3.97337377e-01 -9.77897584e-01 2.43184157e-03 -5.39618492e-01
-7.91728139e-01 3.64336580e-01 -1.43411711e-01 -1.66409659e+00
-3.39084089e-01 -3.10124040e-01 -4.62442636e-01 9.01951134e-01
-1.48873425e+00 -9.98912230e-02 -1.21178724e-01 5.56240797e-01
1.77634761e-01 -3.35768759e-02 6.36807442e-01 4.34581310e-01
-1.48923695e-01 4.68976319e-01 5.16155422e-01 1.33383989e-01
2.38112882e-01 -6.08958602e-01 2.59950519e-01 1.00493062e+00
-8.24986994e-02 4.38618094e-01 1.03264260e+00 -2.16649234e-01
-1.61366427e+00 -5.20712435e-01 4.96798366e-01 5.99170744e-01
4.93916333e-01 2.05349810e-02 -4.88102347e-01 3.45536023e-01
1.60791010e-01 3.82130772e-01 4.50993776e-01 -7.86160231e-01
3.21022607e-02 -1.51689142e-01 -1.45035613e+00 3.71937543e-01
5.65406501e-01 -2.37320766e-01 -3.96528989e-02 9.36967880e-02
2.86564201e-01 -4.06168699e-01 -8.51127446e-01 4.15121466e-01
5.02114356e-01 -1.21555448e+00 9.39913630e-01 1.95345953e-01
4.18924183e-01 -5.28981149e-01 -6.31096482e-01 -1.19263327e+00
-5.06069772e-02 -9.69363093e-01 -1.45228654e-01 6.55968904e-01
1.21074304e-01 -6.87060773e-01 4.19339001e-01 -3.90072554e-01
1.64306507e-01 -7.78381348e-01 -1.24266064e+00 -9.14392948e-01
-5.49774349e-01 -3.44992369e-01 3.30182493e-01 5.17713368e-01
9.40917730e-02 -8.55825394e-02 -4.95132715e-01 3.88966322e-01
8.78123164e-01 -1.26288936e-01 4.33540016e-01 -6.87631488e-01
-7.96404004e-01 -2.06681073e-01 -8.90007794e-01 -1.54054618e+00
-3.94598961e-01 -4.84206051e-01 6.86743036e-02 -9.90828395e-01
-1.71523243e-01 -8.01878497e-02 -1.73017055e-01 -1.28695533e-01
2.48698115e-01 3.49709928e-01 1.04112506e-01 5.34942783e-02
5.59304953e-02 4.76312816e-01 1.24528039e+00 -3.93238179e-02
1.26529023e-01 3.00776899e-01 -3.23189557e-01 6.55448556e-01
4.52176064e-01 -4.90320891e-01 -4.43156987e-01 -8.50114077e-02
1.09509505e-01 5.51187456e-01 3.99492651e-01 -1.35430157e+00
1.27217382e-01 2.88154095e-01 1.39408588e-01 -3.77392143e-01
4.62461650e-01 -1.11635590e+00 3.10972810e-01 8.25233877e-01
-2.83677816e-01 -1.52987391e-01 -6.03951737e-02 7.35014856e-01
-4.09296691e-01 -7.29697168e-01 1.08384204e+00 1.59077108e-01
-8.79130661e-02 -1.29977018e-01 -3.19580019e-01 -9.68195423e-02
9.19503093e-01 -2.73183942e-01 6.25059083e-02 -8.20846081e-01
-4.42490637e-01 -5.01990020e-01 2.02541247e-01 -3.02731037e-01
7.86838830e-01 -1.34459341e+00 -7.73266375e-01 5.84764481e-01
-3.15981179e-01 -3.52949798e-01 4.63490821e-02 1.16362441e+00
-5.73909521e-01 3.28218460e-01 -2.03266200e-02 -5.52914560e-01
-1.27768981e+00 3.32522184e-01 3.55620116e-01 -1.43311813e-01
-3.60235423e-01 5.00484288e-01 -1.88123450e-01 3.52107793e-01
3.13587397e-01 -4.57701325e-01 2.03960344e-01 -3.41343820e-01
7.57318139e-01 5.51266193e-01 1.39839083e-01 -6.06619179e-01
-5.25084473e-02 7.29570270e-01 4.53233033e-01 -5.51480114e-01
1.04644275e+00 -1.79037407e-01 -1.30761445e-01 3.70686710e-01
1.67283177e+00 2.81586736e-01 -1.00579941e+00 -1.92894325e-01
-3.37753236e-01 -7.11801410e-01 1.64124981e-01 -4.55229849e-01
-1.03249633e+00 6.78775489e-01 9.38992381e-01 5.13050318e-01
1.60660863e+00 -5.89709759e-01 8.49505663e-01 8.55279043e-02
4.62088823e-01 -6.53016508e-01 8.43821764e-02 2.13201538e-01
8.42370689e-01 -6.86574697e-01 1.99060161e-02 -4.40235704e-01
-1.34361044e-01 1.22290730e+00 -4.45423692e-01 -3.02796662e-01
9.27047253e-01 1.61825016e-01 -1.99830741e-01 3.39253217e-01
-2.37625837e-01 2.38644749e-01 2.29026884e-01 2.68533945e-01
3.07546705e-01 -1.86706394e-01 -7.36309767e-01 4.45056498e-01
-1.58196524e-01 2.40474880e-01 7.03616381e-01 8.32300365e-01
-6.00264728e-01 -8.09049487e-01 -8.55297923e-01 5.21043949e-02
-7.93941796e-01 -1.13217890e-01 1.18517913e-01 3.24766487e-01
-2.59506583e-01 1.00342917e+00 -1.01130098e-01 -6.05810136e-02
2.95771778e-01 -3.73966098e-01 6.20392859e-01 9.37248245e-02
-2.39403676e-02 1.72255322e-01 -2.39764839e-01 -2.38027439e-01
-3.57217520e-01 -5.56504548e-01 -1.23632491e+00 -3.11784297e-01
-3.93370837e-01 1.93803668e-01 1.05226052e+00 5.75090528e-01
1.11879170e-01 2.73252398e-01 1.13824451e+00 -4.75659788e-01
-1.12965178e+00 -7.90504873e-01 -8.39885890e-01 1.62292257e-01
7.20087230e-01 -1.16284855e-01 -7.68818378e-01 1.69878617e-01] | [6.695490837097168, 1.4945605993270874] |
3fcd049c-9f76-44cb-b3a1-4334c3ecd2f7 | from-plane-crashes-to-algorithmic-harm | 2210.03535 | null | https://arxiv.org/abs/2210.03535v1 | https://arxiv.org/pdf/2210.03535v1.pdf | From plane crashes to algorithmic harm: applicability of safety engineering frameworks for responsible ML | Inappropriate design and deployment of machine learning (ML) systems leads to negative downstream social and ethical impact -- described here as social and ethical risks -- for users, society and the environment. Despite the growing need to regulate ML systems, current processes for assessing and mitigating risks are disjointed and inconsistent. We interviewed 30 industry practitioners on their current social and ethical risk management practices, and collected their first reactions on adapting safety engineering frameworks into their practice -- namely, System Theoretic Process Analysis (STPA) and Failure Mode and Effects Analysis (FMEA). Our findings suggest STPA/FMEA can provide appropriate structure toward social and ethical risk assessment and mitigation processes. However, we also find nontrivial challenges in integrating such frameworks in the fast-paced culture of the ML industry. We call on the ML research community to strengthen existing frameworks and assess their efficacy, ensuring that ML systems are safer for all people. | ['Negar Rostamzadeh', 'AJung Moon', 'Joshua Kroll', 'Edgar Jatho', 'Andrew Smart', 'Renee Shelby', 'Shalaleh Rismani'] | 2022-10-06 | null | null | null | null | ['culture'] | ['speech'] | [ 2.64042735e-01 5.85730433e-01 -3.58684547e-02 7.63720693e-03
-5.41604519e-01 -7.68020868e-01 3.08964789e-01 4.52255905e-01
-9.05914605e-02 1.43183336e-01 5.19641459e-01 -1.05834150e+00
-3.72515798e-01 -4.99262750e-01 -4.87294286e-01 -4.28112298e-01
5.42526007e-01 -3.70064408e-01 -3.42617840e-01 3.66871618e-02
6.56123281e-01 3.93142164e-01 -1.23497427e+00 1.07098721e-01
5.85707903e-01 6.41756237e-01 -5.41442394e-01 3.00777286e-01
2.69417495e-01 1.33710563e+00 -4.24479723e-01 -6.80442512e-01
4.51017052e-01 -2.87087202e-01 -7.61146843e-01 -2.80585676e-01
5.53000644e-02 -6.19238198e-01 1.29090145e-01 8.71356785e-01
3.59727800e-01 -2.64677584e-01 5.71294904e-01 -1.77150428e+00
-9.21529055e-01 5.61947227e-01 -3.70892346e-01 -3.50309640e-01
6.38807043e-02 8.33116531e-01 6.55806720e-01 -5.06231189e-01
3.88725579e-01 1.13657629e+00 9.27079320e-01 4.35891598e-01
-1.32313383e+00 -1.17865431e+00 8.02686960e-02 -8.74802023e-02
-1.14605713e+00 -5.30057967e-01 5.17631173e-01 -1.16538644e+00
1.12201881e+00 1.50486693e-01 7.57956445e-01 1.19367218e+00
9.53201890e-01 2.93145049e-03 1.01047623e+00 -4.58743423e-01
4.60721493e-01 5.71299613e-01 3.71632993e-01 -4.01756093e-02
1.07684553e+00 1.64360210e-01 -1.48203030e-01 -5.56190193e-01
5.24869919e-01 1.34011209e-01 3.85553271e-01 6.13254029e-03
-8.46774638e-01 5.53127170e-01 -3.09191167e-01 3.53582352e-01
-4.49402958e-01 5.99352360e-01 4.55386341e-01 4.23936784e-01
2.22364515e-01 7.57387221e-01 -3.26999158e-01 -4.85084981e-01
-1.16575480e-01 1.03317641e-01 8.76787603e-01 6.13219202e-01
4.75971937e-01 -1.85462102e-01 3.42559852e-02 3.99679065e-01
8.68697524e-01 2.46842086e-01 -6.21883392e-01 -1.53574026e+00
1.26903817e-01 5.02828717e-01 4.97457981e-01 -1.27028489e+00
4.97545442e-03 -1.93741471e-01 -6.98156580e-02 7.92548001e-01
2.45933980e-01 -5.20492733e-01 -2.59074807e-01 1.43871820e+00
-7.65630379e-02 -8.69572386e-02 -5.25861494e-02 3.88073534e-01
-7.07249418e-02 5.35592139e-01 9.38001990e-01 -4.10912544e-01
8.68249655e-01 -2.59667765e-02 -9.89919782e-01 -5.37230134e-01
8.68083060e-01 -5.81342280e-01 1.15455019e+00 7.44684279e-01
-9.86593306e-01 -7.46170580e-02 -1.05656910e+00 4.88497317e-01
1.69700131e-01 -7.04671323e-01 6.65428787e-02 1.39579916e+00
-8.14411044e-01 5.69477618e-01 -7.29872525e-01 -3.65616053e-01
6.33482456e-01 1.09411003e-02 -9.27446485e-02 5.88527173e-02
-8.92139733e-01 1.29347897e+00 -8.72140080e-02 3.76204491e-01
-8.34728241e-01 -8.99674535e-01 -5.03673017e-01 -3.20188403e-01
4.88182843e-01 -5.72423518e-01 1.51898956e+00 -8.29138339e-01
-8.86887729e-01 1.51136845e-01 7.61886001e-01 3.42625119e-02
4.23271060e-01 -7.61631727e-01 -4.85953748e-01 -2.58009046e-01
-4.50673066e-02 -3.92652638e-02 4.41977590e-01 -1.72634280e+00
-6.39124930e-01 -3.30923855e-01 -7.22855628e-02 -9.87528488e-02
-5.72089255e-01 6.33717537e-01 7.77766526e-01 -1.28573880e-01
-3.07495952e-01 -1.00344741e+00 -5.53692639e-01 -2.14826345e-01
-2.32967272e-01 -3.01541258e-02 7.79385209e-01 -7.35996544e-01
1.54753911e+00 -2.17857361e+00 -5.92785656e-01 2.48840168e-01
2.45228514e-01 1.39925912e-01 2.54099011e-01 1.14178073e+00
-7.62652084e-02 7.72407651e-01 1.13780729e-01 2.66429156e-01
2.94023335e-01 -2.82681078e-01 -4.61918801e-01 5.15834033e-01
3.10883105e-01 4.01948988e-01 -8.13907266e-01 -1.33795768e-01
4.56362307e-01 5.80641806e-01 -4.12097901e-01 5.06082876e-03
1.87242493e-01 3.42199624e-01 -4.46645260e-01 5.25408924e-01
4.84812915e-01 3.07788044e-01 3.66157740e-01 2.52771437e-01
-3.95318598e-01 8.44410658e-02 -9.63075638e-01 5.64565778e-01
-4.64794189e-01 4.94935572e-01 1.68273374e-01 -7.99327418e-02
7.24624395e-01 5.93575239e-01 5.76063752e-01 -4.78727013e-01
4.00722414e-01 1.51596973e-02 1.92534968e-01 -8.18996787e-01
-2.85402462e-02 -4.25091714e-01 -2.79223949e-01 8.64328802e-01
-4.29172486e-01 -4.55234479e-03 -6.71867847e-01 4.99565527e-02
1.35673237e+00 1.79697394e-01 1.14594311e-01 -3.86822224e-01
7.88236260e-02 -2.40212068e-01 7.02172518e-01 2.21678451e-01
-7.17977822e-01 2.94746235e-02 8.48207891e-01 -2.15933144e-01
-7.88584292e-01 -9.35254395e-01 2.11545989e-01 8.55169356e-01
-1.45792728e-02 -6.22364521e-01 -1.01903188e+00 -9.11044955e-01
1.81967869e-01 1.66394472e+00 -1.44759327e-01 -7.32762277e-01
-2.12128013e-01 -2.52320856e-01 5.01252294e-01 5.14610469e-01
-1.60137832e-01 -1.00063455e+00 -1.15825725e+00 4.40020412e-01
2.89685220e-01 -6.49194598e-01 -1.75133705e-01 -1.16161615e-01
-3.86108637e-01 -1.14063668e+00 2.63950735e-01 -1.71250939e-01
3.71233791e-01 -4.57066782e-02 6.41986549e-01 -2.76696105e-02
-3.06583703e-01 6.51027024e-01 3.72197479e-02 -1.04644227e+00
-1.05694056e+00 -7.92003572e-01 1.34651095e-01 -3.39292079e-01
4.86037135e-01 -6.16669178e-01 -7.21372724e-01 4.35521722e-01
-7.66903460e-01 -3.43871355e-01 7.01488376e-01 -2.23854512e-01
-1.18985102e-01 2.34872431e-01 1.26757431e+00 -7.16107070e-01
1.11367965e+00 -5.28880656e-01 -1.28479913e-01 2.99876273e-01
-1.59734249e+00 -5.68793535e-01 2.00375289e-01 -2.99792469e-01
-1.39175558e+00 -2.87767231e-01 -8.96795094e-02 5.26839793e-02
-5.70594490e-01 3.44138801e-01 -3.82695615e-01 -4.95493896e-02
9.52877700e-01 -7.91826427e-01 3.50319952e-01 1.03737451e-01
3.21509153e-01 7.80994833e-01 -1.23225138e-01 -4.19266313e-01
6.93321288e-01 -2.88413949e-02 -1.76672161e-01 -5.26843369e-01
-2.68168718e-01 -7.10472912e-02 -1.16213836e-01 -8.47149849e-01
8.56804669e-01 -8.10554981e-01 -8.86850715e-01 3.19455713e-01
-7.51714170e-01 -5.17538071e-01 -8.99557546e-02 4.76379365e-01
-3.27352077e-01 3.62382412e-01 -7.46577978e-01 -1.67325807e+00
-2.59812862e-01 -1.04966402e+00 6.48169577e-01 -9.44784209e-02
-9.57930803e-01 -7.56957650e-01 1.57649338e-01 8.00140977e-01
5.66672981e-01 7.57473350e-01 1.14586782e+00 -4.99998927e-01
-3.82465601e-01 -3.24413985e-01 8.49932581e-02 9.74865556e-01
1.33784384e-01 3.62356156e-01 -1.00967729e+00 4.99941148e-02
5.24769783e-01 -1.38529316e-01 -1.73085675e-01 1.17234968e-01
5.42149365e-01 -8.63787532e-01 -2.96806842e-02 -5.12365878e-01
1.56098616e+00 7.80712247e-01 6.66556418e-01 3.81127596e-01
7.19107985e-01 1.45442879e+00 9.93847311e-01 4.66275960e-01
2.78123111e-01 -1.38384596e-01 4.28515017e-01 3.33206803e-01
5.19177437e-01 -2.14150876e-01 1.00212550e+00 3.75934422e-01
1.58276752e-01 1.22349642e-01 -1.37515759e+00 2.82536924e-01
-1.96547818e+00 -9.15608823e-01 -5.58850527e-01 2.13878775e+00
5.12566209e-01 6.35316610e-01 2.47327611e-01 1.50593743e-01
5.87323189e-01 -2.94773668e-01 -2.49530911e-01 -7.69538403e-01
7.92798996e-01 -3.06141257e-01 5.95676064e-01 2.49996126e-01
-6.74490154e-01 3.94670397e-01 6.92569876e+00 2.14814886e-01
-5.65560997e-01 1.80159941e-01 9.33574378e-01 -1.17900692e-01
-7.65538037e-01 3.80317390e-01 -3.50539833e-01 6.69295937e-02
1.41863751e+00 -5.35512090e-01 -2.37336606e-02 9.53506827e-01
6.51320398e-01 9.40850303e-02 -1.20553875e+00 3.22389305e-01
-4.20821160e-01 -9.82811153e-01 -6.86356783e-01 4.33758765e-01
3.84497732e-01 -5.68531692e-01 1.76422328e-01 1.53441250e-01
4.96133596e-01 -1.05322063e+00 1.12732005e+00 3.87308151e-01
2.90633619e-01 -1.21574223e+00 7.01653302e-01 2.77939141e-01
-5.47298312e-01 -7.63678968e-01 1.82260826e-01 -6.65198922e-01
4.04197991e-01 4.56534386e-01 -7.10772872e-01 2.47753039e-02
7.86591470e-01 8.15610960e-02 -2.11031109e-01 4.20860827e-01
1.17246062e-01 9.22919750e-01 2.42263839e-01 1.63993523e-01
-8.89020320e-03 -4.66679782e-01 1.68909967e-01 1.10438526e+00
8.95346180e-02 -9.57778916e-02 -3.94340515e-01 9.26223993e-01
8.45193088e-01 -1.68065965e-01 -1.04895520e+00 -5.78578413e-01
7.22438812e-01 1.36659098e+00 -6.25036716e-01 4.12691921e-01
-5.43991625e-01 -7.21573159e-02 -4.35855508e-01 2.96252429e-01
-7.42430568e-01 -2.71879733e-01 1.07437897e+00 5.22402108e-01
-6.43243313e-01 -3.56745452e-01 -8.19509447e-01 -2.02754810e-02
-1.16017438e-01 -1.15451491e+00 1.74043700e-01 -3.58215898e-01
-1.52448666e+00 -4.53782734e-04 2.25245297e-01 -9.18760896e-01
-6.54572388e-04 -4.08191383e-01 -4.97379780e-01 5.37813604e-01
-7.03278005e-01 -1.40767300e+00 1.59451753e-01 -1.35958165e-01
-7.93755949e-02 1.62537202e-01 4.42240685e-01 6.56941459e-02
-6.86961949e-01 1.72167331e-01 -3.50848526e-01 -6.41360223e-01
9.00297761e-01 -8.33661139e-01 3.36856484e-01 7.81084776e-01
-7.13827610e-01 1.06843638e+00 8.60020339e-01 -1.17672002e+00
-1.67666209e+00 -1.09788096e+00 7.87013590e-01 -1.07461560e+00
8.87545943e-01 -4.53938425e-01 -7.98364282e-01 9.52437937e-01
5.23328364e-01 -9.30926621e-01 1.30925941e+00 5.14879264e-02
-1.26781523e-01 -1.25653014e-01 -1.46640301e+00 1.01517892e+00
1.06523383e+00 -8.39360833e-01 -4.48564261e-01 9.68916640e-02
1.05432141e+00 5.39795816e-01 -1.28678083e+00 1.85278565e-01
6.69365525e-01 -9.88273442e-01 6.86111510e-01 -2.71466821e-01
2.35176131e-01 -1.30854204e-01 1.28394216e-02 -7.02218890e-01
-6.53112888e-01 -1.04328239e+00 2.03060001e-01 1.64371300e+00
4.14041042e-01 -8.87733102e-01 4.08067763e-01 1.74736333e+00
-3.07049453e-01 -5.20710409e-01 -4.95494545e-01 -5.98782897e-01
3.17445040e-01 -1.19351315e+00 3.65160108e-01 1.19716465e+00
4.66472924e-01 1.11409046e-01 -1.43774673e-01 4.55412567e-01
7.54260480e-01 -9.41262424e-01 4.88572270e-01 -1.18369377e+00
4.57772017e-02 -3.93502563e-01 -1.37226149e-01 6.10363901e-01
-1.34918123e-01 -3.60103130e-01 1.70494676e-01 -1.71166170e+00
1.81611970e-01 -1.43173009e-01 -4.47025239e-01 5.63725591e-01
1.00523926e-01 -5.02544820e-01 4.53438789e-01 4.95453812e-02
-9.21698064e-02 1.97829217e-01 5.52541018e-01 5.22831559e-01
-2.47997120e-01 -2.90222913e-01 -1.64866745e+00 7.73352802e-01
8.34753096e-01 -5.61027408e-01 -8.06615293e-01 2.48787589e-02
6.21741533e-01 -2.18101107e-02 5.96355081e-01 -1.04712772e+00
1.09635845e-01 -9.23658550e-01 -1.27326414e-01 -1.28564686e-01
-3.24461937e-01 -1.47824013e+00 1.15555668e+00 9.02449310e-01
-4.37958270e-01 5.75904846e-02 2.51440793e-01 3.65124464e-01
4.71687645e-01 -1.59772724e-01 3.82101119e-01 3.14135581e-01
-7.98442066e-02 -2.97514796e-01 -1.09510422e+00 -4.61701810e-01
1.77987039e+00 -2.33195499e-01 -7.89192319e-01 -3.81089002e-01
-4.61717814e-01 2.60571837e-01 7.40714550e-01 5.47501266e-01
5.21750510e-01 -1.12803280e+00 -3.76925588e-01 -4.37319987e-02
1.50969043e-01 -4.00599390e-01 4.12706405e-01 8.15205991e-01
-5.50651908e-01 1.30687177e-01 -2.03514755e-01 1.06750011e-01
-9.34570789e-01 6.30179107e-01 2.13746354e-01 4.39672083e-01
-5.96311808e-01 4.47413683e-01 3.48705798e-01 -1.28409013e-01
2.61427552e-01 -1.34195551e-01 1.33776113e-01 5.25583848e-02
3.58779997e-01 9.72130537e-01 -2.98485845e-01 -2.77170986e-01
-4.35591072e-01 1.25813603e-01 2.19199061e-01 -5.82294226e-01
1.21189439e+00 -1.70179814e-01 -2.76401401e-01 6.28790379e-01
6.45698905e-01 -3.52226198e-02 -1.47928143e+00 7.23542571e-01
6.64396167e-01 -5.13114452e-01 5.73851131e-02 -8.09272707e-01
-4.57134962e-01 6.18005455e-01 6.16618752e-01 5.43352425e-01
8.60742688e-01 -1.68537691e-01 5.35214961e-01 2.28583261e-01
3.33870023e-01 -1.73607051e+00 2.51424581e-01 -2.80937850e-01
1.11323309e+00 -3.80189061e-01 -2.57944226e-01 -5.81018090e-01
-1.11713815e+00 7.55637825e-01 8.92963469e-01 1.47637039e-01
1.22346830e+00 6.13722086e-01 2.10528076e-01 -1.38266087e-01
-9.95211720e-01 7.17182219e-01 -4.60498303e-01 7.86116004e-01
5.67102015e-01 5.63129224e-02 -4.69663709e-01 1.09266412e+00
3.57020080e-01 3.53844643e-01 8.04201722e-01 1.32529747e+00
-3.43392730e-01 -1.18492162e+00 -5.24263263e-01 3.83065850e-01
-7.08302200e-01 3.03705990e-01 -8.18126440e-01 7.92959690e-01
2.83396780e-01 1.62688959e+00 -4.20784354e-01 -9.50323343e-01
6.20925188e-01 2.80825734e-01 -1.03201576e-01 -6.41595185e-01
-1.13022721e+00 2.67829567e-01 7.28737175e-01 -7.48194993e-01
1.97649628e-01 -1.19045627e+00 -1.14407122e+00 -2.92450577e-01
-2.30499804e-01 -2.56656498e-01 7.44782805e-01 8.63189876e-01
6.64392471e-01 4.72976685e-01 6.85348392e-01 -2.05562279e-01
-9.91517127e-01 -5.89729607e-01 -5.88611305e-01 1.55761793e-01
8.63143653e-02 -2.61799961e-01 -4.02095497e-01 -5.09842858e-02] | [9.109076499938965, 6.37045431137085] |
ae6b19fe-e1fb-4ed8-92a3-a8d3ac72f18a | the-impact-of-random-models-on-clustering | 1701.06508 | null | http://arxiv.org/abs/1701.06508v2 | http://arxiv.org/pdf/1701.06508v2.pdf | The Impact of Random Models on Clustering Similarity | Clustering is a central approach for unsupervised learning. After clustering
is applied, the most fundamental analysis is to quantitatively compare
clusterings. Such comparisons are crucial for the evaluation of clustering
methods as well as other tasks such as consensus clustering. It is often argued
that, in order to establish a baseline, clustering similarity should be
assessed in the context of a random ensemble of clusterings. The prevailing
assumption for the random clustering ensemble is the permutation model in which
the number and sizes of clusters are fixed. However, this assumption does not
necessarily hold in practice; for example, multiple runs of K-means clustering
returns clusterings with a fixed number of clusters, while the cluster size
distribution varies greatly. Here, we derive corrected variants of two
clustering similarity measures (the Rand index and Mutual Information) in the
context of two random clustering ensembles in which the number and sizes of
clusters vary. In addition, we study the impact of one-sided comparisons in the
scenario with a reference clustering. The consequences of different random
models are illustrated using synthetic examples, handwriting recognition, and
gene expression data. We demonstrate that the choice of random model can have a
drastic impact on the ranking of similar clustering pairs, and the evaluation
of a clustering method with respect to a random baseline; thus, the choice of
random clustering model should be carefully justified. | ['Alexander J. Gates', 'Yong-Yeol Ahn'] | 2017-01-23 | null | null | null | null | ['clustering-ensemble'] | ['graphs'] | [ 2.60063767e-01 -3.34577531e-01 1.77837178e-01 -4.25108254e-01
-5.51712990e-01 -7.70355701e-01 8.43595326e-01 6.31693184e-01
-6.63082361e-01 4.75497931e-01 1.00601554e-01 -1.86202034e-01
-6.07613504e-01 -6.93704069e-01 -3.14872652e-01 -1.30098021e+00
-4.89271693e-02 6.09007895e-01 1.95944563e-01 1.27456933e-01
4.62950438e-01 4.37768728e-01 -1.76356900e+00 6.19625300e-02
7.74427056e-01 5.12447298e-01 2.31137529e-01 5.09666920e-01
-2.88016610e-02 1.28018394e-01 -8.65419328e-01 -1.46420747e-01
2.29091495e-01 -6.73345089e-01 -7.70401299e-01 3.59164923e-01
-4.35170904e-02 5.24902105e-01 4.50771004e-01 1.21068764e+00
5.02689302e-01 4.05935764e-01 1.03560948e+00 -1.16491783e+00
1.08081527e-01 8.08400512e-01 -5.29441237e-01 -1.95771992e-01
1.37475028e-03 1.34192362e-01 1.17054415e+00 -1.88741580e-01
6.12786174e-01 9.81912613e-01 4.42879975e-01 3.20376791e-02
-1.68272662e+00 -2.20622614e-01 -1.03395954e-01 -5.30116819e-02
-1.64886177e+00 -3.68919700e-01 4.83114094e-01 -6.54014826e-01
1.53041542e-01 4.00054574e-01 2.98489571e-01 5.30986130e-01
1.07520574e-03 1.00645134e-02 1.27523530e+00 -6.93583071e-01
7.28414476e-01 1.05105545e-02 3.54264200e-01 6.60484135e-02
5.08606672e-01 -2.17964172e-01 -1.15288109e-01 -2.65079170e-01
1.12089835e-01 -2.75959671e-01 -2.45810077e-01 -8.32390606e-01
-1.40299070e+00 7.39602625e-01 1.29571706e-01 8.88880908e-01
-2.25823745e-01 -8.64680186e-02 3.97072971e-01 2.60891557e-01
1.05199039e-01 7.36904860e-01 -1.32650614e-01 -3.38413343e-02
-1.14598107e+00 1.54432595e-01 7.47701108e-01 3.51473421e-01
9.25780475e-01 -5.70137143e-01 -1.71433225e-01 7.78130412e-01
1.80073828e-02 1.04967549e-01 6.84132934e-01 -9.68653381e-01
-9.62673202e-02 5.39324284e-01 1.42256260e-01 -1.18857634e+00
-3.57878774e-01 -2.24686220e-01 -9.67836320e-01 2.10180521e-01
1.00750291e+00 -1.10672154e-01 -4.27837491e-01 1.79879224e+00
1.93512216e-01 -8.16826746e-02 -2.28031185e-02 7.82369256e-01
2.41307199e-01 1.75205410e-01 -1.14336595e-01 -5.76663971e-01
9.46387410e-01 -2.03743339e-01 -5.60022295e-01 3.06748897e-01
8.17137480e-01 -7.66217291e-01 1.03754401e+00 3.73352438e-01
-7.94659972e-01 -4.95036244e-01 -7.97446668e-01 4.99931753e-01
-4.51107204e-01 -9.48646106e-03 1.28525749e-01 7.45728552e-01
-1.07304990e+00 9.31222081e-01 -6.88596487e-01 -7.68161416e-01
-1.50080964e-01 4.54177856e-01 -3.18080097e-01 1.95040733e-01
-7.41632760e-01 6.33585036e-01 5.99085689e-01 9.68264639e-02
-1.82705477e-01 -1.36691704e-01 -2.79521853e-01 2.15701208e-01
2.02617079e-01 -3.88789266e-01 7.73879051e-01 -1.04109144e+00
-1.16751456e+00 8.50078285e-01 -8.88454095e-02 -2.29246289e-01
5.41435957e-01 4.24334556e-01 -1.84303716e-01 -1.77253798e-01
1.37863696e-01 4.08468276e-01 4.23541933e-01 -1.36893654e+00
-4.30246234e-01 -4.29299742e-01 -2.93388635e-01 7.05159903e-02
-4.60243449e-02 -6.27466142e-02 -4.11906451e-01 -2.72633553e-01
2.44525149e-01 -1.02714288e+00 -4.44254458e-01 -4.76243287e-01
-4.70978141e-01 -1.60167217e-01 2.58560240e-01 3.40844989e-02
1.38069153e+00 -2.38528657e+00 2.92030424e-01 8.29598784e-01
8.06566887e-03 -2.31361181e-01 -1.91701520e-02 5.73348224e-01
-2.82893717e-01 2.42326707e-01 -5.25413871e-01 -6.36894405e-02
4.06596996e-02 1.08423978e-01 9.05482620e-02 6.74400628e-01
1.28670281e-03 3.52635324e-01 -9.36973572e-01 -4.41695929e-01
2.98906595e-01 2.50430316e-01 -4.50496912e-01 1.18416354e-01
-1.54934162e-02 6.09408081e-01 -1.82500958e-01 -8.64525884e-02
3.88509452e-01 -4.17380214e-01 8.66579354e-01 -2.46887431e-01
-3.10293555e-01 2.49607116e-02 -1.56824458e+00 1.26792884e+00
3.49703804e-02 5.48827767e-01 -2.91432947e-01 -1.17550325e+00
9.95255828e-01 4.01957631e-02 6.45247221e-01 -1.82432130e-01
1.93035379e-01 1.82345897e-01 6.62169993e-01 -7.13109449e-02
1.97077081e-01 -1.26105204e-01 -7.38949180e-02 7.21858084e-01
-2.04587489e-01 -1.51936054e-01 6.03848100e-01 1.24335721e-01
1.24891543e+00 -2.48553231e-01 3.78351986e-01 -7.05229640e-01
4.90718901e-01 -4.04806547e-02 3.72030705e-01 8.03052008e-01
8.36245269e-02 8.95112574e-01 8.84318948e-01 2.31301151e-02
-9.67427731e-01 -1.09582150e+00 -3.38990092e-01 7.97786176e-01
1.19438067e-01 -4.60992903e-01 -9.40633237e-01 -5.53106427e-01
-1.86130598e-01 5.38224578e-01 -6.45960867e-01 -1.32055506e-01
-2.23903298e-01 -1.22755229e+00 3.73670042e-01 -1.07067682e-01
2.21456110e-01 -8.22616935e-01 -6.60675824e-01 -8.88418257e-02
-1.49046123e-01 -6.93018258e-01 -1.86442167e-01 4.76361305e-01
-7.51773059e-01 -1.48309255e+00 -4.62779850e-01 -5.31680584e-01
8.95708382e-01 1.56952009e-01 1.11736226e+00 2.52478153e-01
9.04801413e-02 3.23793679e-01 -5.41149318e-01 1.17100857e-01
-6.24430418e-01 1.23447925e-01 1.49731249e-01 1.83902606e-01
4.33575481e-01 -6.72683656e-01 -5.12068748e-01 5.08885324e-01
-1.17545247e+00 -3.42645854e-01 4.38135564e-01 7.63254166e-01
4.76825148e-01 5.97392201e-01 3.22535187e-01 -9.36958075e-01
8.01756084e-01 -3.22280318e-01 -5.04841506e-01 4.88595694e-01
-5.94792306e-01 1.82038829e-01 6.70036256e-01 -2.81085223e-01
-6.30516768e-01 7.51651525e-02 2.45753109e-01 -1.73263669e-01
-5.01124918e-01 6.22707129e-01 -4.11386222e-01 2.67945260e-01
7.17677951e-01 9.84092429e-02 8.49770159e-02 -2.88031220e-01
4.31190163e-01 6.97796345e-01 3.47812921e-01 -6.78409576e-01
5.84224582e-01 3.78680915e-01 1.34355351e-01 -1.10005414e+00
-2.57069856e-01 -6.96404815e-01 -1.20758450e+00 -2.92825073e-01
6.97141707e-01 -2.95256674e-01 -6.75472021e-01 3.22425842e-01
-7.80143559e-01 -3.64692986e-01 -2.19173178e-01 4.15926486e-01
-6.00179017e-01 6.51162982e-01 1.42653752e-02 -6.09450340e-01
2.99038768e-01 -1.40976882e+00 7.52351940e-01 -1.39767095e-01
-6.69526398e-01 -1.11820853e+00 2.77546197e-01 -9.62420739e-03
2.39464827e-02 4.42887813e-01 1.19942141e+00 -9.59001780e-01
-1.12966418e-01 6.57083094e-02 7.87079036e-02 1.77622288e-01
5.93164265e-01 4.46805179e-01 -7.28101313e-01 -2.74482757e-01
-1.82220340e-01 1.61744431e-01 7.68982768e-01 5.27319491e-01
1.02719355e+00 -9.21118706e-02 -4.06800836e-01 1.92271560e-01
1.30403972e+00 2.71763653e-01 5.69516540e-01 1.44746825e-01
4.00160998e-01 1.00423574e+00 3.74055922e-01 2.14626059e-01
-1.06302232e-01 7.90688574e-01 -3.89637202e-02 2.40561545e-01
3.15913469e-01 2.04013288e-01 1.10496648e-01 9.96145904e-01
-1.90205444e-02 -3.04369122e-01 -1.05773282e+00 3.89118940e-01
-1.74938226e+00 -9.95530486e-01 -1.41008556e-01 2.82308960e+00
6.43173635e-01 9.64642614e-02 3.31393272e-01 5.63081682e-01
1.11874998e+00 -5.30052260e-02 -1.59222439e-01 -2.47154087e-01
-1.08466469e-01 -6.50416762e-02 2.14205846e-01 4.29667056e-01
-1.02819359e+00 4.14563656e-01 6.45446062e+00 8.07834983e-01
-1.10050762e+00 -3.43579650e-01 8.05978000e-01 3.19333710e-02
-3.89163382e-02 1.39403686e-01 -3.40465844e-01 8.62505198e-01
8.20793509e-01 -2.81459600e-01 2.52651274e-01 3.31316799e-01
3.89962405e-01 -3.90526921e-01 -1.29466581e+00 6.93795979e-01
-3.90885353e-01 -9.50543523e-01 -1.34423807e-01 3.02066386e-01
7.53994524e-01 -1.73035577e-01 -1.68088779e-01 -1.72221199e-01
5.53940892e-01 -9.92817223e-01 2.60364175e-01 4.34617639e-01
4.52837378e-01 -9.22223151e-01 8.37881029e-01 2.87567437e-01
-8.98810923e-01 9.35772136e-02 -2.41010621e-01 5.28589301e-02
-1.10032827e-01 9.99318957e-01 -6.14161432e-01 5.48975885e-01
6.08394682e-01 4.03300077e-01 -5.57153642e-01 1.11070943e+00
5.17415591e-02 7.31666148e-01 -4.55699444e-01 -8.05557445e-02
1.07572399e-01 -6.79914474e-01 3.42301905e-01 1.17536592e+00
2.67458349e-01 -1.24448009e-01 3.52995880e-02 6.72455549e-01
1.71350986e-01 2.84835070e-01 -4.51399833e-01 -1.23748891e-02
8.13614964e-01 1.22088051e+00 -1.41775286e+00 -1.60853282e-01
-1.08195245e-01 6.44210041e-01 2.86713123e-01 3.07885468e-01
-4.61617172e-01 -1.63294584e-01 5.94476879e-01 1.83963194e-01
3.88288796e-01 -2.18327999e-01 -1.56831622e-01 -6.43006980e-01
-2.13440001e-01 -8.86096418e-01 3.95920902e-01 -3.77154946e-01
-1.53142309e+00 3.25394779e-01 1.83363602e-01 -1.21912384e+00
-2.73350388e-01 -3.45292360e-01 -8.47822905e-01 5.54437280e-01
-7.85215795e-01 -3.24689448e-01 -1.31123990e-01 2.54992276e-01
1.51789645e-02 1.01354934e-01 5.73780477e-01 1.92198937e-03
-6.34340107e-01 3.85440707e-01 7.11303353e-01 1.90431654e-01
9.89649117e-01 -1.39153790e+00 -1.67164177e-01 7.21072972e-01
3.74049127e-01 9.82479930e-01 9.20773029e-01 -4.19584662e-01
-8.75859141e-01 -1.08887959e+00 6.84154272e-01 -3.65121096e-01
5.35687506e-01 -2.16782629e-01 -1.02002406e+00 2.28976637e-01
-4.21967879e-02 -3.46897870e-01 9.54288900e-01 3.55976462e-01
9.65035148e-03 -1.00884072e-01 -9.65657949e-01 7.28880644e-01
6.98048830e-01 -2.23169148e-01 -3.55726361e-01 1.52995020e-01
1.42668620e-01 2.90319771e-01 -1.18090940e+00 3.32863271e-01
5.81878006e-01 -1.40052629e+00 6.13794267e-01 -3.24613929e-01
2.47572333e-01 -7.12509215e-01 -2.11200625e-01 -1.45084000e+00
-3.78388405e-01 -3.40531498e-01 6.58117473e-01 1.39281797e+00
2.84483820e-01 -4.94262516e-01 6.39535904e-01 5.66647947e-01
3.82631481e-01 -5.13196468e-01 -8.83185744e-01 -8.02874386e-01
1.81531399e-01 -1.41657382e-01 4.65234101e-01 1.16916549e+00
8.63124728e-02 3.88997495e-01 2.11759463e-01 8.83907005e-02
4.22574073e-01 1.79494321e-01 9.85746801e-01 -1.50542080e+00
-2.06745982e-01 -8.26098144e-01 -6.01122022e-01 -5.02368093e-01
1.38489649e-01 -7.14719117e-01 2.24193588e-01 -1.18304491e+00
2.76878893e-01 -6.04617059e-01 -3.87704402e-01 3.20105851e-02
-1.94171205e-01 7.90163875e-02 1.50219560e-01 4.43196774e-01
-6.21011078e-01 1.82685062e-01 6.42273188e-01 3.03168166e-02
-3.96555603e-01 -4.10790853e-02 -5.32132864e-01 7.48288572e-01
8.65041792e-01 -4.76679176e-01 -3.56504142e-01 1.97706744e-01
2.28166401e-01 -2.87565857e-01 1.12354174e-01 -1.03687537e+00
4.38522100e-01 -1.67298019e-01 1.86477259e-01 -3.19397867e-01
-1.68181479e-01 -7.75321960e-01 6.32464588e-01 2.61826158e-01
-6.13308609e-01 1.23859234e-01 -1.99097842e-01 4.33530629e-01
-2.72860765e-01 -3.92605811e-01 8.03705394e-01 -2.08132580e-01
-3.41400564e-01 -2.23619431e-01 -5.01129806e-01 -8.01796317e-02
1.05994666e+00 -5.21180809e-01 1.62633955e-01 -2.58323103e-01
-8.15714359e-01 1.01814799e-01 9.72301900e-01 1.15632392e-01
6.77080303e-02 -1.11108994e+00 -6.32640660e-01 9.54544451e-03
1.61349222e-01 4.61963937e-02 -9.73792151e-02 9.92969573e-01
-2.50818282e-01 2.63560236e-01 6.86593205e-02 -9.02086735e-01
-1.48020089e+00 6.18785858e-01 3.02509248e-01 -2.54787326e-01
-2.67390143e-02 1.92703590e-01 2.98630983e-01 -5.19926548e-01
8.24109763e-02 -7.75365606e-02 -3.14177305e-01 3.14256191e-01
1.98167369e-01 4.56876934e-01 1.76839963e-01 -5.74128211e-01
-2.66834587e-01 6.31151259e-01 2.00145110e-01 -1.31590277e-01
1.05351341e+00 -2.38098532e-01 -4.69636381e-01 8.48317146e-01
9.01630104e-01 3.00610494e-02 -7.33644843e-01 1.17357604e-01
4.70384210e-01 -3.85934293e-01 -2.98211336e-01 -4.77447063e-01
-9.00219500e-01 5.82612932e-01 4.28746670e-01 4.56453651e-01
1.09919298e+00 -4.25425544e-02 -2.19367057e-01 4.64378625e-01
3.34837824e-01 -9.59265828e-01 -1.47489756e-01 1.27398968e-01
4.47247446e-01 -1.03556430e+00 1.43893123e-01 -2.05086738e-01
-4.98894572e-01 9.53558922e-01 1.75022617e-01 -6.20604046e-02
7.04016328e-01 1.49486586e-01 9.69886929e-02 -2.08135828e-01
-6.07691467e-01 -3.32331777e-01 1.98759645e-01 4.40809906e-01
7.76758850e-01 9.06295404e-02 -6.31968439e-01 1.68052122e-01
-2.91070998e-01 -3.07089865e-01 4.74536240e-01 5.79962134e-01
-2.85259157e-01 -1.29641140e+00 -4.76565659e-01 4.74727809e-01
-2.96958715e-01 2.81450879e-02 -7.03870356e-01 8.34554970e-01
2.61534333e-01 1.08214331e+00 3.62226903e-01 -3.36857796e-01
4.12743688e-02 2.22008079e-01 3.52095604e-01 -6.09884322e-01
-6.19965374e-01 1.03050411e-01 -1.36088654e-01 -1.30981162e-01
-8.82300436e-01 -8.93728197e-01 -1.01377690e+00 -3.14812601e-01
-6.00023925e-01 6.61280215e-01 4.73085940e-01 1.00640857e+00
1.89370513e-01 2.48738825e-01 7.26505518e-01 -5.38962901e-01
-3.78200948e-01 -9.16037858e-01 -7.97562182e-01 7.28732467e-01
-1.26243502e-01 -5.88866174e-01 -6.87794030e-01 2.45801583e-01] | [7.631160736083984, 4.546947479248047] |
e8a37646-acfa-47a7-b384-6354c9d9acc9 | provably-efficient-primal-dual-reinforcement | 2201.11965 | null | https://arxiv.org/abs/2201.11965v4 | https://arxiv.org/pdf/2201.11965v4.pdf | Provably Efficient Primal-Dual Reinforcement Learning for CMDPs with Non-stationary Objectives and Constraints | We consider primal-dual-based reinforcement learning (RL) in episodic constrained Markov decision processes (CMDPs) with non-stationary objectives and constraints, which plays a central role in ensuring the safety of RL in time-varying environments. In this problem, the reward/utility functions and the state transition functions are both allowed to vary arbitrarily over time as long as their cumulative variations do not exceed certain known variation budgets. Designing safe RL algorithms in time-varying environments is particularly challenging because of the need to integrate the constraint violation reduction, safe exploration, and adaptation to the non-stationarity. To this end, we identify two alternative conditions on the time-varying constraints under which we can guarantee the safety in the long run. We also propose the \underline{P}eriodically \underline{R}estarted \underline{O}ptimistic \underline{P}rimal-\underline{D}ual \underline{P}roximal \underline{P}olicy \underline{O}ptimization (PROPD-PPO) algorithm that can coordinate with both two conditions. Furthermore, a dynamic regret bound and a constraint violation bound are established for the proposed algorithm in both the linear kernel CMDP function approximation setting and the tabular CMDP setting under two alternative conditions. This paper provides the first provably efficient algorithm for non-stationary CMDPs with safe exploration. | ['Javad Lavaei', 'Yuhao Ding'] | 2022-01-28 | null | null | null | null | ['safe-exploration'] | ['robots'] | [-9.14154053e-02 2.25331381e-01 -4.74435270e-01 1.11810505e-01
-7.24149764e-01 -7.86543906e-01 1.59942836e-01 2.02557534e-01
-7.93397605e-01 1.46916926e+00 -4.47229475e-01 -5.30400753e-01
-9.09663081e-01 -6.60365462e-01 -6.54781282e-01 -1.15452361e+00
-6.05604053e-01 7.43597209e-01 -1.45498291e-01 9.65057581e-04
1.49134308e-01 6.15645111e-01 -1.26249290e+00 -5.80831110e-01
1.07095504e+00 1.26082659e+00 2.84730077e-01 8.08043599e-01
2.74528146e-01 6.32098734e-01 -4.50410157e-01 2.37904340e-02
4.32660311e-01 -2.11427003e-01 -3.96688759e-01 1.94689497e-01
-6.45580649e-01 -4.05222327e-01 -2.46680930e-01 9.21734095e-01
5.23640513e-01 6.06386721e-01 4.70796019e-01 -1.81360769e+00
2.21074023e-03 4.93507802e-01 -9.70524311e-01 4.48961079e-01
7.89449587e-02 3.60156417e-01 8.56039762e-01 -1.82592705e-01
4.39449698e-01 1.16659427e+00 -5.68903610e-02 6.06522560e-01
-1.05360627e+00 -5.81488609e-01 8.85842323e-01 -3.17772105e-02
-1.19808686e+00 -1.10815475e-02 2.29685172e-01 -1.08260503e-02
7.00118780e-01 3.48689586e-01 6.89101636e-01 8.71031702e-01
3.34229380e-01 1.03137600e+00 1.09891224e+00 -2.31595129e-01
6.38471425e-01 -3.33997309e-02 -2.07251295e-01 3.55661064e-01
3.69664609e-01 3.79781663e-01 -1.98420405e-01 -2.34025910e-01
9.35143769e-01 1.57358572e-02 -5.01251556e-02 -3.96734804e-01
-9.14627373e-01 9.47763205e-01 -2.09338263e-01 -2.49189436e-01
-6.73977375e-01 2.98014551e-01 2.76900202e-01 4.05462146e-01
1.47864550e-01 3.22680384e-01 -5.26583135e-01 -4.09898072e-01
-6.98799789e-01 5.95753491e-01 7.74208963e-01 1.25903428e+00
1.37746394e-01 3.91373873e-01 -4.99393940e-01 3.90016377e-01
1.03806183e-02 6.72399998e-01 -6.60451278e-02 -1.22526956e+00
9.09049809e-01 -2.73680210e-01 1.06973171e+00 -4.18837905e-01
-5.15587449e-01 -6.98471010e-01 -5.86581528e-01 3.22287887e-01
5.68396449e-01 -8.21240664e-01 -5.09200752e-01 2.01191688e+00
4.37520087e-01 -2.17228793e-02 2.56404072e-01 8.54627609e-01
-4.70585555e-01 8.22311997e-01 -3.57531048e-02 -1.12990284e+00
9.48537230e-01 -5.50828636e-01 -8.08188856e-01 -7.34617114e-02
2.59701520e-01 -2.68758178e-01 9.98204768e-01 6.21422172e-01
-1.40355384e+00 1.30753860e-01 -9.43627656e-01 7.24103272e-01
2.33790562e-01 -1.97458744e-01 3.98050964e-01 6.29965127e-01
-5.90259910e-01 4.47110891e-01 -8.60289991e-01 1.44485369e-01
1.41911060e-01 5.18862069e-01 2.53965944e-01 1.55116245e-02
-1.10249519e+00 7.44364560e-01 6.91942990e-01 2.31342137e-01
-1.38882244e+00 -5.35122275e-01 -4.62501794e-01 1.68503985e-01
1.22077703e+00 -1.81878060e-01 1.37814951e+00 -5.60235739e-01
-1.49978065e+00 1.95516884e-01 3.35477442e-01 -6.15384936e-01
1.13659263e+00 -4.54773121e-02 -2.78765559e-01 1.99708596e-01
3.21504921e-02 1.17374495e-01 7.55337656e-01 -1.15426970e+00
-9.81406629e-01 -3.53847057e-01 3.98940712e-01 8.18488538e-01
-1.36538029e-01 -1.78883165e-01 -2.66466528e-01 -4.82302159e-01
-2.21658692e-01 -1.11483479e+00 -5.12805700e-01 -3.70415211e-01
-4.78427052e-01 -6.50496688e-03 6.37678742e-01 -3.21042210e-01
1.27249634e+00 -1.87086296e+00 1.83403671e-01 3.17854524e-01
-4.05324042e-01 -4.59219441e-02 7.62933865e-02 5.44343412e-01
4.37696218e-01 -4.02173772e-02 -1.95304856e-01 -1.28268942e-01
1.66223437e-01 7.50864029e-01 -4.55933899e-01 5.09308219e-01
-2.02632949e-01 3.15632552e-01 -9.93340194e-01 -3.36198270e-01
-1.99660379e-02 -1.89062551e-01 -3.58811527e-01 1.91597939e-01
-6.32157207e-01 6.25449538e-01 -8.26897562e-01 6.68248832e-01
6.60728812e-01 2.94098467e-01 3.58393729e-01 7.06048846e-01
-3.01956505e-01 -4.45341587e-01 -1.57823730e+00 1.02307522e+00
-6.50716841e-01 -4.87537868e-02 6.13738000e-01 -1.00972056e+00
5.36301374e-01 3.04309160e-01 8.29015195e-01 -6.93952739e-01
1.47309244e-01 4.20989171e-02 -1.89458296e-01 -2.61799783e-01
4.98836607e-01 -2.94536620e-01 -2.78338939e-01 5.84760129e-01
-4.70605731e-01 -7.67720565e-02 2.60522157e-01 -2.47651935e-02
8.61953855e-01 2.66291350e-01 2.44525939e-01 -3.80269140e-01
3.75454903e-01 -2.28079006e-01 1.04977334e+00 8.19687307e-01
-4.33287561e-01 -1.79969564e-01 1.20794022e+00 3.45103256e-02
-8.34881604e-01 -8.65280151e-01 -4.20275256e-02 1.02132702e+00
4.03101802e-01 4.57772791e-01 -3.18462402e-01 -5.78838885e-01
1.38289124e-01 1.11039734e+00 -6.37393594e-01 6.14359677e-02
-3.43595833e-01 -8.70201766e-01 3.06635320e-01 3.43107551e-01
3.70000094e-01 -9.00408030e-01 -1.10306895e+00 3.58980536e-01
1.11364566e-01 -1.07809281e+00 -5.18344939e-01 7.95371532e-01
-5.48640609e-01 -7.96895504e-01 -7.75548577e-01 -1.39074221e-01
7.25194752e-01 -9.22873095e-02 5.87819397e-01 -5.50412774e-01
1.28343880e-01 5.33749759e-01 -1.15227818e-01 -5.94402075e-01
4.59864587e-02 -9.18809995e-02 2.35013083e-01 -9.17014182e-02
-2.26656795e-01 -2.91487992e-01 -5.38843334e-01 4.89139736e-01
-9.26170826e-01 -2.21708208e-01 2.70882159e-01 8.36041629e-01
1.07567501e+00 3.80364656e-01 8.76731694e-01 -5.57257652e-01
8.08398783e-01 -6.00085080e-01 -1.18745553e+00 4.86983895e-01
-6.84407711e-01 2.41339505e-01 6.95892036e-01 -7.65288651e-01
-1.13662100e+00 -1.11324742e-01 2.71488547e-01 -6.20820403e-01
3.34591925e-01 4.64866489e-01 -4.03822422e-01 3.53861004e-01
2.06971258e-01 2.35976964e-01 -9.84896049e-02 -3.83860171e-02
1.37201576e-02 2.30051309e-01 2.72357970e-01 -1.26603723e+00
7.02791989e-01 3.35518569e-01 5.07791996e-01 -4.50446010e-01
-6.72961235e-01 6.08750544e-02 -1.87897012e-01 -4.31072563e-01
4.98358816e-01 -7.55535424e-01 -1.46530569e+00 1.41554675e-03
-4.66661811e-01 -7.69591570e-01 -4.83544260e-01 4.81549770e-01
-1.21730781e+00 3.97861719e-01 -2.37520158e-01 -1.69892406e+00
-2.46528946e-02 -9.07759905e-01 3.30813438e-01 4.68908638e-01
3.99849445e-01 -7.97737777e-01 -1.31039783e-01 -6.01682775e-02
2.24001687e-02 6.64364278e-01 6.98726892e-01 -3.61115962e-01
-5.83451867e-01 5.19426130e-02 7.50595853e-02 2.07611442e-01
-2.79555470e-01 -2.50634372e-01 -1.75486624e-01 -7.15093434e-01
3.72924693e-02 -3.94497037e-01 3.45131248e-01 4.65703964e-01
1.11044085e+00 -8.71836126e-01 -1.90014854e-01 3.90163600e-01
1.43742466e+00 8.52825046e-01 2.64918745e-01 6.24789476e-01
-6.06830344e-02 4.80008125e-01 1.46178746e+00 1.42623627e+00
2.93469042e-01 6.19255245e-01 1.02474797e+00 4.10410881e-01
1.00786328e+00 -1.60151467e-01 5.01964450e-01 -2.35008731e-01
-9.32721496e-02 -6.18353307e-01 -6.69518828e-01 5.84879220e-01
-2.09962368e+00 -9.75107908e-01 1.85759008e-01 2.84764099e+00
9.04080570e-01 4.04277384e-01 3.52055728e-01 3.72325927e-02
6.62349999e-01 -3.37851495e-02 -1.20017314e+00 -8.74020278e-01
-2.71844678e-02 -1.42356530e-01 1.08145142e+00 4.75457728e-01
-7.62465298e-01 6.02053761e-01 4.76291323e+00 9.55759764e-01
-8.69135797e-01 1.56900719e-01 7.25411475e-01 -6.49334729e-01
-6.91847429e-02 5.15473541e-03 -8.11998010e-01 7.90515542e-01
1.09791386e+00 -4.07135516e-01 6.75466418e-01 9.19657648e-01
6.42816424e-01 -6.88226581e-01 -9.20790374e-01 7.19384193e-01
-5.10491312e-01 -7.12682426e-01 -7.11850822e-01 3.04729491e-01
7.55890727e-01 -4.77690011e-01 2.72205919e-01 5.74407816e-01
6.51550114e-01 -7.60527551e-01 9.17616487e-01 5.38148105e-01
8.81278217e-01 -1.56205010e+00 3.61808568e-01 6.50408149e-01
-1.08116615e+00 -6.37866497e-01 -2.71213055e-01 8.18038881e-02
4.56344515e-01 3.89924943e-01 -3.02674860e-01 6.71121359e-01
3.12991738e-01 1.22110602e-02 5.43018341e-01 9.29218888e-01
-4.42902416e-01 1.40779138e-01 -4.25176591e-01 -9.43906382e-02
6.23508692e-01 -4.25519735e-01 6.63419962e-01 6.67902410e-01
3.85588199e-01 1.97347060e-01 6.35128736e-01 4.91597325e-01
4.06269759e-01 -1.80536717e-01 -2.30373248e-01 -3.83816063e-02
7.21715093e-01 8.85728061e-01 -8.56670022e-01 -1.04100974e-02
2.15120256e-01 8.07308733e-01 -2.06699278e-02 5.47574818e-01
-1.30447757e+00 -3.75200778e-01 7.53064036e-01 -2.17043117e-01
4.10784036e-01 -3.91338587e-01 -2.10132673e-01 -8.24660718e-01
1.17144465e-01 -4.74219710e-01 7.61830091e-01 -2.94336557e-01
-9.98234570e-01 1.20625734e-01 2.68751353e-01 -1.18759763e+00
-3.28279108e-01 -1.52259052e-01 -3.73891294e-01 8.29117417e-01
-1.75972712e+00 -5.96495807e-01 2.62560248e-01 8.26274395e-01
4.95227337e-01 4.88469228e-02 3.53189826e-01 -3.65054933e-03
-9.29361105e-01 3.73845726e-01 6.66625082e-01 -6.57264352e-01
1.81098238e-01 -1.34012723e+00 -6.43205285e-01 7.49451220e-01
-6.94301605e-01 8.20161998e-02 1.01875126e+00 -5.74256539e-01
-1.50250316e+00 -9.97385323e-01 4.68793772e-02 2.57562876e-01
7.44466901e-01 -1.48767173e-01 -5.70365310e-01 5.43090820e-01
-2.04435796e-01 -3.75258201e-03 2.85289884e-01 -1.67979240e-01
1.65062591e-01 -8.47452730e-02 -1.39733064e+00 6.58409655e-01
7.40668356e-01 -4.56536282e-03 3.44091691e-02 6.36544228e-01
5.74162006e-01 -6.12385273e-01 -8.19491208e-01 2.28760719e-01
3.35920513e-01 -4.53957468e-01 5.68757355e-01 -7.07427025e-01
-1.29882842e-01 -6.56266585e-02 -1.82181999e-01 -1.19138026e+00
1.50512084e-01 -1.25569487e+00 -5.47878087e-01 9.57841277e-01
3.68593395e-01 -6.58031285e-01 7.45528281e-01 8.93884003e-01
-1.40847163e-02 -9.49652851e-01 -1.54939377e+00 -1.46168077e+00
3.64432156e-01 -2.85144150e-01 3.72745246e-01 6.15412593e-01
2.87147760e-02 -3.81375760e-01 -8.74261200e-01 4.17022288e-01
7.98502266e-01 1.21292807e-01 3.21472585e-01 -7.21836925e-01
-6.72450542e-01 -8.13851580e-02 4.85290527e-01 -8.42033327e-01
1.98366761e-01 -3.28325957e-01 1.94709539e-01 -1.38681781e+00
4.58412133e-02 -9.50044990e-01 -5.31393647e-01 3.73213083e-01
5.10881692e-02 -7.06016719e-01 4.15158093e-01 1.92581624e-01
-8.53430629e-01 8.85603547e-01 1.29631519e+00 1.22651607e-01
-6.28140926e-01 6.69333875e-01 -4.04893905e-01 1.69231981e-01
8.49447906e-01 -4.61250752e-01 -9.14920866e-01 -7.56693333e-02
2.98098326e-01 1.13109577e+00 -1.31742284e-01 -6.26614571e-01
-1.02038637e-01 -1.14053249e+00 -2.43214697e-01 -7.90306211e-01
4.27578628e-01 -9.05398071e-01 1.73792720e-01 7.93458521e-01
-4.39081848e-01 1.91433817e-01 8.70607868e-02 1.01109922e+00
3.08855057e-01 -3.50845635e-01 9.33965683e-01 -1.46373227e-01
-4.02265757e-01 5.99769235e-01 -6.71982229e-01 1.91358060e-01
1.72474885e+00 -1.16636967e-02 -1.04218110e-01 -7.37748086e-01
-9.29562151e-01 1.15408826e+00 9.46216807e-02 1.61145568e-01
2.92293310e-01 -8.71185720e-01 -2.20184326e-01 -3.66986513e-01
-4.11243170e-01 -1.20693848e-01 5.58992803e-01 1.07724833e+00
-1.25705287e-01 5.85443676e-01 -3.18752944e-01 -1.47852898e-01
-8.44337642e-01 9.98508811e-01 4.18788821e-01 -7.30121553e-01
-2.65570432e-01 6.09011292e-01 -1.27952158e-01 6.52747750e-02
6.32149398e-01 -1.80622965e-01 1.52415097e-01 3.12771052e-01
2.36421183e-01 7.82417119e-01 -2.71618128e-01 3.88442986e-02
-6.50380254e-02 -1.64871037e-01 -3.88730019e-02 -5.33257663e-01
1.30516589e+00 -5.15652955e-01 3.18142563e-01 2.15517879e-01
3.72363716e-01 -3.48543704e-01 -1.96418345e+00 -1.93741456e-01
9.91205499e-02 -4.20862466e-01 8.17793682e-02 -8.92272949e-01
-9.15894151e-01 5.44766605e-01 5.38394928e-01 2.63470054e-01
1.05358315e+00 -5.55789590e-01 4.81847048e-01 3.41280639e-01
8.26163411e-01 -1.68787920e+00 9.62442905e-02 6.15525663e-01
7.79900074e-01 -7.22492814e-01 1.43179283e-01 7.03982834e-04
-1.22315657e+00 8.88729453e-01 5.85319221e-01 5.65903299e-02
4.19199497e-01 3.47737759e-01 -6.76365733e-01 3.15583706e-01
-9.68730569e-01 -4.52550352e-01 -5.87518573e-01 5.74397564e-01
-3.43308866e-01 4.12505180e-01 -8.21322262e-01 5.71499646e-01
3.44157696e-01 -8.48070160e-02 7.99431860e-01 1.37084377e+00
-5.64674616e-01 -1.02731729e+00 -6.15204155e-01 1.88761458e-01
-4.94118929e-01 3.14561218e-01 2.82420844e-01 9.41125453e-01
-1.58250466e-01 1.02096474e+00 -1.61453396e-01 4.27811116e-01
2.26819396e-01 -8.39452147e-02 3.52687001e-01 -2.73754150e-01
-2.60231346e-01 4.11416560e-01 1.76613122e-01 -4.20278043e-01
-8.99506286e-02 -8.49272013e-01 -1.55249453e+00 -3.15962970e-01
-1.21107779e-01 2.71452188e-01 6.18544698e-01 9.69463766e-01
1.17888175e-01 5.43226421e-01 1.10590601e+00 -3.96804631e-01
-1.19134557e+00 -4.82630163e-01 -1.13210094e+00 -3.85959387e-01
1.53379038e-01 -1.00167692e+00 -3.91186833e-01 -7.74058580e-01] | [4.390807628631592, 2.7648656368255615] |
9b648528-7d13-4536-b818-07ff2cd9d2ae | effective-learning-based-illuminant | null | null | http://openaccess.thecvf.com/content_cvpr_2015/html/Cheng_Effective_Learning-Based_Illuminant_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Cheng_Effective_Learning-Based_Illuminant_2015_CVPR_paper.pdf | Effective Learning-Based Illuminant Estimation Using Simple Features | Illumination estimation is the process of determining the chromaticity of the illumination in an imaged scene in order to remove undesirable color casts through white-balancing. While computational color constancy is a well-studied topic in computer vision, it remains challenging due to the ill-posed nature of the problem. One class of techniques relies on low-level statistical information in the image color distribution and works under various assumptions (e.g. Grey-World, White-Patch, etc). These methods have an advantage that they are simple and fast, but often do not perform well. More recent state-of-the-art methods employ learning-based techniques that produce better results, but often rely on complex features and have long evaluation and training times. In this paper, we present a learning-based method based on four simple color features and show how to use this with an ensemble of regression trees to estimate the illumination. We demonstrate that our approach is not only faster than existing learning-based methods in terms of both evaluation and training time, but also gives the best results reported to date on modern color constancy data sets. | ['Brian Price', 'Scott Cohen', 'Michael S. Brown', 'Dongliang Cheng'] | 2015-06-01 | null | null | null | cvpr-2015-6 | ['color-constancy'] | ['computer-vision'] | [ 4.39366579e-01 -7.85363913e-01 2.17191368e-01 -4.00304675e-01
-6.12176955e-01 -5.97762525e-01 4.69719619e-01 -1.19499527e-01
-3.73884588e-01 7.37233639e-01 -5.55161476e-01 -2.86361665e-01
1.69477481e-02 -4.16122854e-01 -3.67580682e-01 -1.09683156e+00
5.86861968e-02 2.80923605e-01 4.75422949e-01 -2.40310431e-01
5.33035696e-01 5.55827558e-01 -1.91994846e+00 2.64650118e-03
1.01600766e+00 1.13762689e+00 -3.74956131e-02 9.27975953e-01
-1.94269776e-01 7.86001444e-01 -5.39913237e-01 -2.45542035e-01
2.35940829e-01 -8.81714165e-01 -5.05498588e-01 3.13042194e-01
6.89049900e-01 2.36290991e-02 3.45509380e-01 1.23064482e+00
4.45781738e-01 1.26752451e-01 7.55180538e-01 -1.27211583e+00
-3.44656408e-01 -3.66517335e-01 -8.62049758e-01 -1.27386212e-01
2.24236265e-01 -7.17558041e-02 8.07667077e-01 -7.57509351e-01
3.78058434e-01 9.81098831e-01 7.41002440e-01 4.14228261e-01
-1.50349426e+00 -4.94444042e-01 -1.04319647e-01 4.96574938e-01
-1.07992685e+00 -5.78511417e-01 9.46545959e-01 -3.97987127e-01
4.05166119e-01 5.89766502e-01 6.56649053e-01 4.10134196e-01
1.48676127e-01 5.71735322e-01 2.17863512e+00 -1.06847918e+00
4.09521759e-01 2.78579384e-01 -1.60357580e-01 1.02242696e+00
6.99092001e-02 1.45238131e-01 -3.90202075e-01 -6.81819171e-02
7.07372963e-01 -2.19809875e-01 -3.11277181e-01 -6.49459243e-01
-8.26830029e-01 7.70382285e-01 3.94817382e-01 2.21294448e-01
-1.19880609e-01 4.91877943e-02 1.39794320e-01 2.69568384e-01
6.21002197e-01 4.82959956e-01 -3.78902584e-01 2.97504873e-03
-1.09232831e+00 -2.63729226e-02 9.11853731e-01 5.65292418e-01
1.18072236e+00 6.88604219e-03 1.25454411e-01 1.04502058e+00
1.48455814e-01 5.97211659e-01 1.78957567e-01 -1.00439298e+00
-1.27771959e-01 5.08636534e-01 2.42201716e-01 -8.91736805e-01
-4.69989479e-01 -9.07537416e-02 -7.49994934e-01 1.28911710e+00
7.55235136e-01 -6.73618913e-02 -1.29951060e+00 1.44219863e+00
3.90325874e-01 -1.73167691e-01 -3.19786578e-01 9.06379223e-01
3.13089043e-01 4.01780874e-01 -2.72006273e-01 -3.69155675e-01
1.14246583e+00 -8.69121373e-01 -6.82893455e-01 -1.43516168e-01
-3.01370397e-02 -1.20115244e+00 1.02735066e+00 9.86884713e-01
-1.02820253e+00 -5.23763657e-01 -1.08404660e+00 6.85818540e-03
-6.12658143e-01 1.61268964e-01 7.74244666e-01 1.20205748e+00
-1.23895407e+00 6.61107719e-01 -5.49770772e-01 -3.10850531e-01
2.06703320e-02 2.26861551e-01 -2.26461038e-01 -2.68071681e-01
-5.31077266e-01 1.22191894e+00 5.94053380e-02 1.81344628e-01
-3.16699147e-01 -2.90796191e-01 -6.90893233e-01 -2.98043072e-01
4.51429725e-01 -2.32924238e-01 1.10213733e+00 -1.66838706e+00
-1.83526683e+00 8.38103533e-01 -2.88940489e-01 1.43916145e-01
7.64269173e-01 -1.69164807e-01 -2.99468845e-01 1.59748241e-01
-4.12041396e-01 2.43236899e-01 1.30446076e+00 -1.78273380e+00
-6.76001787e-01 -3.56241018e-01 -1.50576696e-01 9.02417079e-02
-3.08610294e-02 1.17671438e-01 -7.04741895e-01 -3.36956799e-01
3.72633934e-01 -1.17909980e+00 -1.78596079e-01 3.65784287e-01
-4.12825316e-01 1.39508426e-01 7.32000887e-01 -6.07348025e-01
9.15596187e-01 -2.01538110e+00 -1.11530319e-01 4.30675000e-01
-3.29057947e-02 3.16082537e-01 3.51477489e-02 3.13557327e-01
-1.19987234e-01 -2.54817784e-01 -5.27910709e-01 -2.84560621e-01
-8.84876922e-02 2.03160360e-01 1.32042855e-01 8.14048171e-01
1.72454059e-01 3.11737418e-01 -7.59547234e-01 -7.11197555e-01
6.86212063e-01 6.22511327e-01 -3.13865215e-01 1.87608033e-01
-1.40224218e-01 3.53633910e-01 8.41039270e-02 7.62767911e-01
9.61411417e-01 1.28089607e-01 1.01567075e-01 -3.82721692e-01
-5.35980940e-01 -2.45640412e-01 -1.48029292e+00 1.34685254e+00
-5.21609426e-01 8.50707054e-01 1.73673078e-01 -9.17536855e-01
7.74381638e-01 -7.52354506e-03 5.10625839e-01 -7.87108183e-01
1.56958148e-01 5.10340214e-01 -1.94328781e-02 -4.74631578e-01
3.43183607e-01 -2.21295640e-01 4.22279567e-01 4.53852236e-01
-2.08232298e-01 -6.01903498e-01 4.41391855e-01 -3.75456065e-01
6.40967727e-01 5.90911746e-01 3.07641894e-01 -3.13837498e-01
6.95550919e-01 -2.48640906e-02 5.52863717e-01 6.13349438e-01
-1.91638216e-01 8.42077017e-01 4.03542280e-01 -3.39072585e-01
-9.45108652e-01 -7.20744491e-01 -1.61508352e-01 1.13779783e+00
3.48355800e-01 5.22967465e-02 -8.35194945e-01 -3.78368437e-01
-1.60206854e-01 6.69643581e-01 -6.14674985e-01 1.53450385e-01
-4.68820721e-01 -1.01929557e+00 -1.58573404e-01 1.97668701e-01
4.51409608e-01 -7.49459505e-01 -8.83879125e-01 8.04931950e-03
2.58737560e-02 -9.08881962e-01 -4.27025147e-02 4.43537831e-01
-8.32287192e-01 -1.35365641e+00 -7.93682396e-01 -5.36358178e-01
9.56435740e-01 4.08785671e-01 1.26182222e+00 2.27106348e-01
-6.96614563e-01 4.16049033e-01 -4.23118114e-01 -5.65838456e-01
-3.05137157e-01 -4.09546465e-01 -3.69583637e-01 2.59326577e-01
2.85175622e-01 -1.96070239e-01 -6.01838887e-01 4.99310315e-01
-8.49033833e-01 7.74537474e-02 8.07881534e-01 9.99050319e-01
6.87413990e-01 3.10558885e-01 -1.41187370e-01 -1.16751015e+00
2.19507739e-01 2.01169491e-01 -9.78917003e-01 5.22955179e-01
-8.94760191e-01 2.27728993e-01 5.20543635e-01 -2.36804724e-01
-1.17903149e+00 4.19035643e-01 1.15134634e-01 -1.96333423e-01
-1.35916889e-01 3.23668830e-02 1.33432060e-01 -7.12857127e-01
7.16390967e-01 1.11076608e-01 -2.25180015e-02 -3.99821669e-01
3.66050988e-01 3.45584452e-01 5.55150390e-01 -3.48996967e-01
1.03842795e+00 6.53046727e-01 4.39174920e-01 -9.17725146e-01
-7.74381518e-01 -6.79454446e-01 -8.44331205e-01 -6.01372182e-01
7.61930883e-01 -3.43206286e-01 -6.35869503e-01 6.40208244e-01
-7.38214672e-01 -3.30310374e-01 1.02734871e-01 3.38078529e-01
-5.56632280e-01 3.68040442e-01 -1.13212667e-01 -1.25326371e+00
-4.96750884e-02 -9.52388644e-01 9.06798482e-01 4.75264639e-01
4.12890643e-01 -1.11890066e+00 1.30024359e-01 7.51544461e-02
7.29242325e-01 5.40881991e-01 1.01631749e+00 2.89891601e-01
-4.81169462e-01 -2.08787024e-01 -5.26548386e-01 5.00528574e-01
4.06910598e-01 4.42887783e-01 -1.28545845e+00 -2.75861233e-01
-5.87034635e-02 -3.47860992e-01 1.03764987e+00 5.11391878e-01
1.18434072e+00 1.89566344e-01 8.56744871e-02 5.92956722e-01
2.07318258e+00 1.41617656e-01 7.08287954e-01 4.89359796e-01
5.29886007e-01 7.40805566e-01 6.92202032e-01 2.34616503e-01
1.20519064e-02 6.60362601e-01 6.97125554e-01 -8.16723168e-01
-3.31135482e-01 3.65276307e-01 1.30061492e-01 3.85985613e-01
-5.47909617e-01 1.55259579e-01 -6.06828392e-01 1.65441692e-01
-1.75504160e+00 -9.14987504e-01 -4.03578639e-01 2.54665136e+00
9.57822382e-01 3.07787079e-02 2.12809876e-01 5.63706338e-01
6.73147500e-01 -1.29819259e-01 -5.16912758e-01 -4.94182020e-01
-2.85280555e-01 2.54090488e-01 7.53459811e-01 5.74873745e-01
-1.20751202e+00 6.48966253e-01 6.59495544e+00 4.81809705e-01
-1.39341593e+00 -1.18940562e-01 6.68257177e-01 3.04435939e-01
-7.03281611e-02 3.40125784e-02 -2.86858946e-01 2.46509060e-01
4.34408933e-01 2.89276928e-01 7.36690164e-01 7.00822949e-01
9.98106226e-03 -8.68510962e-01 -8.91161621e-01 1.24978745e+00
6.07260346e-01 -6.00963473e-01 -6.43933952e-01 -2.81108558e-01
8.98233473e-01 -1.63710549e-01 2.24296942e-01 -1.10235162e-01
2.10167706e-01 -8.46036494e-01 7.68455982e-01 6.60438180e-01
8.08050215e-01 -6.32748902e-01 6.59594893e-01 -5.71233183e-02
-1.00954425e+00 2.63731703e-02 -4.30565298e-01 1.38722062e-01
-1.90153286e-01 8.43081594e-01 -3.67673934e-01 4.57352996e-01
7.08783925e-01 3.30748439e-01 -9.86089110e-01 1.64793968e+00
-4.36905593e-01 3.49897325e-01 -3.59208673e-01 -1.92318201e-01
1.07079849e-01 -5.27551651e-01 4.73347753e-02 1.30730104e+00
1.02828152e-01 -2.59101450e-01 1.34674549e-01 6.67691946e-01
3.13688427e-01 1.84893906e-01 -2.42640167e-01 3.29857528e-01
-1.44122556e-01 1.58730698e+00 -9.91593301e-01 -1.61617145e-01
-5.31766295e-01 1.13255477e+00 1.74233928e-01 5.53603292e-01
-5.83847702e-01 -5.77605665e-01 2.91639209e-01 -9.22759548e-02
2.19849885e-01 -3.69198620e-01 -3.25017214e-01 -9.43340182e-01
-4.02856320e-02 -8.50265205e-01 2.12779924e-01 -9.04564619e-01
-1.34124053e+00 5.65228164e-01 -2.30218023e-01 -1.39932537e+00
-2.07132161e-01 -1.11871040e+00 -5.45262039e-01 9.15064812e-01
-2.10895848e+00 -1.12734294e+00 -6.66805923e-01 7.80751407e-01
4.74943519e-01 7.44628459e-02 9.71645951e-01 1.36252731e-01
-3.70769918e-01 1.72311366e-01 5.39922118e-01 -2.11981922e-01
1.14811707e+00 -1.84492695e+00 -1.96409374e-01 9.77418542e-01
2.35566229e-01 2.59404212e-01 1.01044893e+00 -7.17841312e-02
-1.60185790e+00 -5.81808984e-01 5.51712453e-01 -2.04331800e-01
4.14692730e-01 -3.18192363e-01 -7.48258054e-01 4.65059690e-02
3.04094613e-01 1.33952215e-01 6.80606008e-01 2.53582329e-01
-4.68594849e-01 -3.72915328e-01 -1.04067504e+00 4.90377635e-01
3.38031054e-01 -3.28230590e-01 -1.87431410e-01 2.80070007e-01
-3.14701259e-01 -3.73084486e-01 -3.67470801e-01 1.65150046e-01
7.59852529e-01 -1.62019253e+00 7.04399824e-01 -1.69679940e-01
1.49318278e-01 -5.96670151e-01 8.49782396e-03 -1.33499253e+00
-2.99636930e-01 -5.66679835e-01 3.44991714e-01 1.00215018e+00
4.66480225e-01 -6.40728652e-01 6.83649361e-01 7.61807442e-01
1.61818519e-01 -4.25909698e-01 -7.94819117e-01 -7.62875795e-01
-2.53044605e-01 -1.89054534e-01 6.07566498e-02 7.16964185e-01
-4.93122607e-01 3.84190865e-02 -5.39576292e-01 -5.50795123e-02
9.83644009e-01 5.76387405e-01 8.23744655e-01 -1.43451202e+00
-1.83330879e-01 -5.73737800e-01 -2.87000626e-01 -2.53728569e-01
-9.12131220e-02 -2.30106741e-01 4.38496411e-01 -1.55245268e+00
1.23056814e-01 -7.46197343e-01 -4.37497556e-01 5.40582240e-01
-4.79470342e-01 7.53295064e-01 1.39277698e-02 6.20682463e-02
-5.28537393e-01 1.32616758e-01 1.11915398e+00 -5.63082658e-02
-1.91305622e-01 5.96136153e-02 -3.42024654e-01 8.25335324e-01
8.04725647e-01 -2.66605645e-01 -2.57033259e-01 -3.05451691e-01
2.22290844e-01 -5.28920889e-01 3.27291429e-01 -9.82745945e-01
2.50584215e-01 -3.07988077e-01 8.36523056e-01 -4.75950122e-01
5.38870752e-01 -9.93329227e-01 -9.46735963e-02 4.57332164e-01
7.79305920e-02 9.35652554e-02 4.84228656e-02 3.46951902e-01
-1.65540740e-01 -3.33662152e-01 1.32114708e+00 -2.61365741e-01
-9.78972554e-01 -4.35994416e-02 -1.79061040e-01 -1.90131366e-01
9.00424719e-01 -3.24733168e-01 -1.83353662e-01 -3.76751721e-01
-2.85364479e-01 -3.00381988e-01 8.24927807e-01 -6.15683477e-03
3.86877626e-01 -1.00327456e+00 -5.77609420e-01 2.32197210e-01
1.39572740e-01 -4.94507432e-01 3.22289183e-03 9.61753428e-01
-9.59323287e-01 -1.04122534e-01 -4.67565954e-01 -7.87783206e-01
-1.49010170e+00 4.78896320e-01 4.22743529e-01 1.05272330e-01
-3.53293419e-01 7.69801617e-01 -1.40259624e-01 -7.92165920e-02
1.50837690e-01 1.24263782e-02 -1.45665869e-01 2.69751512e-02
5.28774500e-01 4.62525219e-01 3.10632229e-01 -5.55664778e-01
-2.35623926e-01 1.09227180e+00 3.84295434e-01 -2.21820399e-01
1.21057153e+00 -1.38527423e-01 -5.12583494e-01 5.77578843e-01
1.00259531e+00 5.00139147e-02 -1.30507529e+00 -1.58164591e-01
7.31432587e-02 -1.08975637e+00 3.37871850e-01 -9.47109401e-01
-1.17684031e+00 9.59466338e-01 1.00711668e+00 4.83605444e-01
1.75063694e+00 -4.35621619e-01 1.40676066e-01 2.53375083e-01
2.13560998e-01 -1.58845794e+00 1.04148634e-01 7.01826215e-02
6.88473344e-01 -1.61327136e+00 5.23513615e-01 -5.82659423e-01
-5.87057829e-01 1.46798182e+00 4.81327623e-01 -6.83840439e-02
5.44197798e-01 1.54216930e-01 6.23996556e-01 4.39391769e-02
-4.02790576e-01 -5.96666813e-01 5.10582507e-01 6.70800745e-01
6.26277208e-01 -1.56988442e-01 -2.38004401e-01 -5.08194447e-01
-3.89379933e-02 -2.71481901e-01 4.71689850e-01 9.51313376e-01
-5.59477866e-01 -1.21056151e+00 -7.21372485e-01 2.79132187e-01
-3.83866519e-01 -1.14860879e-02 -4.66960549e-01 7.25594282e-01
2.10523009e-01 1.10691774e+00 -2.82330781e-01 3.14389430e-02
2.59389162e-01 -2.65413560e-02 9.22122896e-01 -2.52177119e-01
-3.28135997e-01 4.27252769e-01 -2.14610696e-02 -7.66887307e-01
-9.60409582e-01 -7.33992159e-01 -6.39252067e-01 -1.91935092e-01
-6.66230619e-01 -6.58127666e-02 1.27259457e+00 7.26022661e-01
-4.07500207e-01 3.02310407e-01 9.55741286e-01 -1.20965922e+00
-2.19811931e-01 -6.80861890e-01 -7.75766909e-01 5.04618227e-01
5.46712697e-01 -7.34637558e-01 -4.32853669e-01 3.91332865e-01] | [10.372610092163086, -2.5443942546844482] |
74ba9318-e01e-49c2-9d67-9bd363e4349b | multimodal-multi-user-surface-recognition | 2303.04930 | null | https://arxiv.org/abs/2303.04930v1 | https://arxiv.org/pdf/2303.04930v1.pdf | Multimodal Multi-User Surface Recognition with the Kernel Two-Sample Test | Machine learning and deep learning have been used extensively to classify physical surfaces through images and time-series contact data. However, these methods rely on human expertise and entail the time-consuming processes of data and parameter tuning. To overcome these challenges, we propose an easily implemented framework that can directly handle heterogeneous data sources for classification tasks. Our data-versus-data approach automatically quantifies distinctive differences in distributions in a high-dimensional space via kernel two-sample testing between two sets extracted from multimodal data (e.g., images, sounds, haptic signals). We demonstrate the effectiveness of our technique by benchmarking against expertly engineered classifiers for visual-audio-haptic surface recognition due to the industrial relevance, difficulty, and competitive baselines of this application; ablation studies confirm the utility of key components of our pipeline. As shown in our open-source code, we achieve 97.2% accuracy on a standard multi-user dataset with 108 surface classes, outperforming the state-of-the-art machine-learning algorithm by 6% on a more difficult version of the task. The fact that our classifier obtains this performance with minimal data processing in the standard algorithm setting reinforces the powerful nature of kernel methods for learning to recognize complex patterns. | ['Katherine J. Kuchenbecker', 'Sebastian Trimpe', 'Friedrich Solowjow', 'Behnam Khojasteh'] | 2023-03-08 | null | null | null | null | ['hypothesis-testing', 'hypothesis-testing'] | ['methodology', 'miscellaneous'] | [ 4.70985711e-01 -1.63441330e-01 2.18871593e-01 -2.14106664e-01
-1.29299974e+00 -6.00076556e-01 5.03168821e-01 2.95476198e-01
-3.93381834e-01 3.05099458e-01 -1.94181100e-01 -2.07003281e-01
-3.39451909e-01 -5.88928998e-01 -8.33759248e-01 -6.19491339e-01
-3.37539792e-01 2.04865292e-01 3.56060922e-01 -2.88345695e-01
2.61513293e-01 3.86613339e-01 -2.27698779e+00 8.14060450e-01
5.81738591e-01 1.54779661e+00 -9.24589038e-02 8.17295790e-01
3.19334008e-02 2.19325110e-01 -4.83732104e-01 5.45437112e-02
1.11929014e-01 3.37550556e-03 -5.68784475e-01 -3.91933471e-01
8.77108037e-01 -5.37396334e-02 2.66135335e-01 5.13882756e-01
8.91775012e-01 2.19725985e-02 1.02594388e+00 -1.23010838e+00
-4.08769757e-01 -1.86055601e-01 -2.39835352e-01 -1.78106457e-01
5.29027343e-01 1.76486105e-01 1.07115352e+00 -9.35548067e-01
3.53703052e-01 9.53024864e-01 1.10543454e+00 3.08688343e-01
-1.73751414e+00 -4.27993715e-01 -3.24291617e-01 -5.59413582e-02
-1.09059989e+00 -4.14187938e-01 5.66090286e-01 -8.65543306e-01
1.06345797e+00 4.56456691e-01 5.17584324e-01 1.39838326e+00
2.29772344e-01 8.80417109e-01 1.49416387e+00 -4.33558851e-01
4.81442839e-01 2.90872127e-01 -7.23680258e-02 4.71631616e-01
1.63304489e-02 2.90937632e-01 -6.11179829e-01 -5.34924686e-01
6.56258225e-01 -3.31896842e-01 -1.23083100e-01 -6.32349133e-01
-1.22241163e+00 4.83297586e-01 -1.57855928e-01 5.12959212e-02
-2.84727275e-01 1.27541035e-01 6.67920411e-01 6.26518965e-01
4.99741822e-01 6.74410880e-01 -9.93852675e-01 -6.46911442e-01
-5.90734839e-01 3.57161969e-01 1.05352640e+00 5.06737709e-01
5.31511486e-01 -2.04625860e-01 1.15353368e-01 1.03180730e+00
-1.39798224e-01 6.01155818e-01 2.61880130e-01 -6.41641796e-01
4.85845879e-02 3.60533863e-01 7.85194486e-02 -9.23081815e-01
-3.82769972e-01 -1.68572128e-01 -4.70873952e-01 8.72122943e-01
7.21871734e-01 -3.10632363e-02 -6.48003340e-01 1.35784233e+00
3.97117347e-01 6.56702295e-02 -1.57644600e-01 7.86426485e-01
5.70426226e-01 1.98427454e-01 -7.30149448e-02 3.55339319e-01
1.05052853e+00 -3.56240064e-01 -2.22610131e-01 1.02779448e-01
6.22148395e-01 -8.48720908e-01 1.83883059e+00 9.62969244e-01
-7.45802999e-01 -7.91695416e-01 -1.10210574e+00 1.94706947e-01
-5.32509983e-01 2.70013362e-01 7.70911157e-01 6.53512359e-01
-8.11062098e-01 1.00903201e+00 -7.92831719e-01 -3.15603793e-01
5.02941370e-01 5.84850729e-01 -6.39510512e-01 3.50735843e-01
-9.47590292e-01 6.52023315e-01 -9.15660188e-02 -3.18363458e-01
-4.31504071e-01 -1.14736402e+00 -6.24703884e-01 -3.17136556e-01
1.59568116e-01 -1.88913569e-01 1.19736362e+00 -9.58597541e-01
-1.74574184e+00 9.24896002e-01 4.57258016e-01 1.13230377e-01
5.97193718e-01 -7.35878646e-01 -3.92115474e-01 2.52552833e-02
-2.44707271e-01 3.77279848e-01 1.23288965e+00 -1.31271076e+00
-5.56879282e-01 -2.98994798e-02 -1.96209505e-01 -3.55753005e-01
-4.37063426e-01 -2.45420873e-01 -6.42001331e-02 -5.79664171e-01
-2.10547000e-01 -1.00838971e+00 1.57750830e-01 2.45616198e-01
-2.65803635e-01 -1.68924958e-01 6.64679646e-01 -6.08108819e-01
8.82540762e-01 -2.56805182e+00 -3.01366802e-02 4.18741345e-01
2.65691373e-02 1.35616720e-01 -2.78158337e-01 5.66683650e-01
-2.91008651e-01 -8.95520523e-02 -1.08685821e-01 -2.05606505e-01
3.16745907e-01 -2.80154079e-01 -3.12752575e-01 5.12265682e-01
4.58376378e-01 7.78541327e-01 -7.72778928e-01 -1.23107888e-01
3.39210451e-01 4.42969590e-01 -5.70376575e-01 3.55428368e-01
-8.35394561e-02 3.03516895e-01 3.11630638e-03 8.54018092e-01
3.25136960e-01 -2.13341981e-01 -1.68232769e-01 -6.51868463e-01
-8.17481545e-04 1.37193695e-01 -1.31195354e+00 1.90926933e+00
-7.31347561e-01 7.12110341e-01 2.03102410e-01 -9.08661127e-01
7.69423366e-01 9.58934501e-02 5.78575552e-01 -6.78769708e-01
5.78427576e-02 5.80525339e-01 7.30614811e-02 -8.24056983e-01
1.13123953e-01 -2.51897741e-02 -1.51562959e-01 4.92552698e-01
4.95867208e-02 -4.78909910e-01 -4.61708903e-01 -3.02988529e-01
1.27472532e+00 3.98952395e-01 -7.54204392e-02 -4.55287665e-01
5.02725728e-02 -1.39385700e-01 -1.20997593e-01 8.44258010e-01
-4.41378243e-02 8.78074944e-01 4.32254463e-01 -4.57206249e-01
-1.06626964e+00 -1.39544833e+00 -4.56083000e-01 1.38492894e+00
-1.33706078e-01 -3.53824049e-01 -5.69526970e-01 -4.63024169e-01
6.39105916e-01 3.20710152e-01 -9.37282443e-01 -2.40384117e-01
-1.74049318e-01 -4.53595638e-01 5.79201639e-01 6.69735789e-01
-7.03787357e-02 -9.41380739e-01 -7.59837389e-01 3.65538187e-02
4.61712956e-01 -1.00208688e+00 3.42715383e-02 3.04808170e-01
-6.05116248e-01 -1.13451958e+00 -3.58855218e-01 -5.83996594e-01
1.36629805e-01 -1.77042171e-01 1.23922026e+00 -3.00165951e-01
-8.86990845e-01 8.39212298e-01 -2.02012032e-01 -5.06096959e-01
-4.28548783e-01 -5.92521683e-04 2.72972703e-01 -4.13279794e-02
2.44123697e-01 -8.49563122e-01 -6.90008879e-01 3.29803586e-01
-7.83002436e-01 -9.48548540e-02 7.33740389e-01 1.05706203e+00
4.54415590e-01 -1.69169307e-01 6.94330335e-01 -6.66029453e-01
9.52347577e-01 -4.85976607e-01 -3.82856190e-01 8.88301060e-02
-3.73989135e-01 8.07263479e-02 3.49535018e-01 -9.31965351e-01
-5.12093902e-01 8.64444003e-02 -4.14480045e-02 -2.67179042e-01
-5.16276419e-01 4.94580090e-01 1.01159796e-01 -3.61179799e-01
1.07847881e+00 -9.07001272e-02 5.58763921e-01 -4.48167652e-01
3.44504505e-01 9.33855236e-01 4.32113856e-01 -1.03010428e+00
5.58289289e-01 4.78694588e-01 -5.69940656e-02 -9.90473032e-01
-4.52430308e-01 -3.34915459e-01 -5.97111046e-01 -2.07683131e-01
6.35111868e-01 -7.89384663e-01 -1.13269413e+00 7.51921773e-01
-7.83010185e-01 -6.81925297e-01 -4.00093198e-01 4.67775255e-01
-8.52677763e-01 1.67828783e-01 -5.16993761e-01 -9.11146104e-01
-2.08595574e-01 -7.94145286e-01 1.68726313e+00 -3.13762009e-01
-6.93191826e-01 -1.01978946e+00 2.49845713e-01 1.49304777e-01
6.92939460e-01 7.67190576e-01 1.05871344e+00 -7.36117721e-01
-7.47519135e-02 -4.76897091e-01 -1.78551957e-01 6.88372552e-01
4.81278747e-01 1.16315655e-01 -1.48406863e+00 -3.33338857e-01
-3.27640146e-01 -9.78571117e-01 6.28625870e-01 9.19745490e-02
1.34292018e+00 3.22988361e-01 -1.24400169e-01 3.80485505e-01
1.17997444e+00 -2.82375604e-01 5.00335395e-01 2.92440593e-01
6.55869365e-01 8.57142806e-01 5.51273406e-01 4.13669914e-01
-1.28381491e-01 6.75139546e-01 1.89821497e-01 -4.87427354e-01
1.78953022e-01 1.33565381e-01 2.91468054e-01 4.72496450e-01
8.10640380e-02 3.41519535e-01 -1.01379502e+00 3.55203927e-01
-1.71790814e+00 -6.38824224e-01 1.26870163e-02 2.29842877e+00
9.81658280e-01 3.54600638e-01 2.19568580e-01 4.48703170e-01
1.76452696e-01 -1.59653634e-01 -5.85242689e-01 -3.59216005e-01
3.70324329e-02 7.65419304e-01 1.59811407e-01 8.97265449e-02
-1.41508090e+00 5.69133997e-01 6.93918324e+00 1.00960290e+00
-1.44774210e+00 -2.25875363e-01 3.68396759e-01 -2.68778890e-01
-7.10455924e-02 -5.96540034e-01 -1.04470901e-01 3.85902286e-01
1.00624597e+00 4.51717675e-01 4.91560787e-01 8.65803480e-01
-5.90267368e-02 -2.26640895e-01 -1.56350541e+00 1.36378837e+00
-1.57746434e-01 -1.36730278e+00 -3.94694954e-01 4.72365227e-03
3.98308605e-01 1.06038861e-01 3.41183722e-01 3.54026049e-01
-4.45914827e-02 -1.27383995e+00 7.72144020e-01 5.70849836e-01
1.21334958e+00 -3.20377558e-01 4.49182749e-01 2.81291548e-02
-9.49138165e-01 2.31050178e-02 8.00287053e-02 -2.89060801e-01
-2.94967353e-01 7.52307832e-01 -6.44376934e-01 2.65191317e-01
8.98590207e-01 5.28822541e-01 -4.83030975e-01 7.39232004e-01
2.95864075e-01 7.15867937e-01 -5.64829051e-01 -1.43742308e-01
-2.87710875e-01 1.44649774e-01 1.50418371e-01 1.49382055e+00
1.25861049e-01 -2.53593415e-01 1.97786987e-01 8.00387442e-01
4.42780197e-01 1.82209849e-01 -7.94987917e-01 -1.78987145e-01
3.27394158e-01 1.27690768e+00 -4.07722384e-01 6.11672513e-02
-4.22870576e-01 9.07680809e-01 3.01201165e-01 2.97501653e-01
-6.07829809e-01 -7.43899703e-01 1.07439625e+00 2.08151251e-01
2.09671244e-01 -4.50382262e-01 -4.98403639e-01 -7.52397418e-01
3.15939337e-01 -1.12109876e+00 -3.20640542e-02 -6.64117575e-01
-1.80733144e+00 3.84705007e-01 -4.92370240e-02 -1.34978092e+00
-1.33772701e-01 -1.27836454e+00 -4.66926485e-01 9.72286999e-01
-1.02862954e+00 -1.08468473e+00 -4.53426391e-01 4.37762827e-01
7.14989603e-02 -1.89450786e-01 1.18584001e+00 2.51114398e-01
6.80629760e-02 7.22058475e-01 1.64787725e-01 -1.57262255e-02
9.00453687e-01 -1.46887755e+00 5.27865171e-01 -3.97830494e-02
1.52140325e-02 6.43905222e-01 4.84520316e-01 -3.73377502e-01
-1.74542749e+00 -5.32898843e-01 1.25062317e-02 -9.28255260e-01
9.77946341e-01 -8.77235591e-01 -1.09414530e+00 -2.21508071e-02
-1.18892051e-01 1.17279388e-01 1.26864767e+00 5.07631183e-01
-5.93709350e-01 -9.90395322e-02 -9.46643174e-01 4.51884955e-01
1.01892722e+00 -1.02372479e+00 -3.97679150e-01 2.34135643e-01
3.71090829e-01 -4.71413821e-01 -1.26056147e+00 6.63297236e-01
1.09001815e+00 -1.04678392e+00 9.65449572e-01 -7.27068126e-01
4.47564453e-01 -1.09125085e-01 -3.81749690e-01 -1.27383983e+00
-1.23153031e-01 -6.01334989e-01 4.41670902e-02 9.04043436e-01
4.53062385e-01 -5.28026521e-01 6.34735465e-01 6.46260500e-01
-6.49079084e-02 -1.08950579e+00 -9.03679967e-01 -7.72702396e-01
8.23330879e-03 -8.90765011e-01 1.90006256e-01 1.00381541e+00
1.64006799e-01 9.23252553e-02 -2.14930270e-02 -4.91205454e-02
4.69345838e-01 9.16545540e-02 1.12068236e+00 -1.54805636e+00
-6.18613899e-01 -3.89143437e-01 -6.68428361e-01 -4.43759441e-01
2.53960639e-01 -5.54886460e-01 9.76519138e-02 -1.14831781e+00
-1.84120908e-01 -7.81753182e-01 -4.53669488e-01 5.59757710e-01
-1.56473234e-01 1.55268058e-01 -3.29600096e-01 1.21461473e-01
-4.06259000e-01 4.86245722e-01 9.49545503e-01 -1.08852506e-01
-3.20654958e-01 -2.42426097e-01 -3.17986250e-01 4.93178934e-01
6.04669690e-01 -1.35919750e-02 -3.98939818e-01 -1.04031645e-01
2.93119639e-01 -4.78301913e-01 6.65955782e-01 -1.36102271e+00
-1.33635804e-01 -7.08074495e-02 4.75344062e-01 -8.26295987e-02
4.64731663e-01 -8.04836452e-01 9.14319307e-02 2.74138868e-01
-3.93440723e-01 -6.45071790e-02 7.76819289e-01 5.50765395e-01
-3.52724791e-02 3.72761041e-01 5.42515278e-01 3.58399719e-01
-6.72373712e-01 -7.62874931e-02 -2.17514127e-01 1.16468683e-01
8.47617447e-01 -4.55738932e-01 -1.94574520e-01 -1.43785119e-01
-6.60876691e-01 -3.89761299e-01 5.55659294e-01 7.29160249e-01
4.95368749e-01 -1.27147591e+00 -5.51285863e-01 5.56799889e-01
4.80904877e-01 -2.35458538e-01 1.68725103e-01 8.41545165e-01
-4.11163449e-01 -1.37245238e-01 -4.27994519e-01 -1.10715091e+00
-1.16743922e+00 1.62155420e-01 3.99187833e-01 2.29701698e-01
-5.28992474e-01 8.39900613e-01 -2.52409652e-02 -6.34775698e-01
1.88692912e-01 -6.60611808e-01 2.80948669e-01 5.26241101e-02
3.25519443e-01 3.55255663e-01 6.01528049e-01 6.39514551e-02
-3.53573382e-01 5.91761589e-01 9.76866931e-02 -2.59566784e-01
1.28775883e+00 5.89479506e-01 1.51101395e-01 1.05467463e+00
1.50453651e+00 1.95113331e-01 -1.51878405e+00 1.41337156e-01
3.93696055e-02 -3.67448241e-01 -1.62685618e-01 -9.61943567e-01
-3.94909799e-01 8.31906021e-01 1.06496012e+00 3.45427006e-01
8.99373293e-01 8.89587775e-03 6.43045425e-01 4.70318258e-01
5.14185786e-01 -1.39717972e+00 3.89045835e-01 3.13070267e-01
1.06339133e+00 -1.44836032e+00 -1.90135136e-01 -4.10644889e-01
-3.92896593e-01 1.08909893e+00 4.49404150e-01 -1.34713843e-01
8.74714017e-01 6.18388772e-01 3.54250371e-01 -2.84269780e-01
-7.09485471e-01 9.98261422e-02 5.33768773e-01 8.32501352e-01
5.77059686e-01 1.82046741e-02 2.81733453e-01 4.80620265e-01
-1.33637533e-01 -1.69329997e-02 2.20286399e-02 1.20540428e+00
-2.76333928e-01 -1.02982616e+00 -2.05773473e-01 6.20539904e-01
-2.25476250e-01 -2.11983714e-02 -3.24182868e-01 9.20328379e-01
-4.42364514e-02 7.77732253e-01 -3.33890282e-02 -8.95282388e-01
6.93761647e-01 1.19699843e-01 6.46028578e-01 -6.13035619e-01
-6.85623407e-01 -1.19988203e-01 2.45521903e-01 -9.74652767e-01
-2.32351005e-01 -7.74557471e-01 -9.57864523e-01 -2.60731615e-02
-1.54323831e-01 -1.62485152e-01 8.75145555e-01 7.79440403e-01
7.20242202e-01 5.12848198e-01 3.99003893e-01 -1.29290724e+00
-8.39922726e-01 -9.15309072e-01 -6.20964170e-01 1.02227700e+00
4.36803222e-01 -9.62588131e-01 -4.34713155e-01 1.09781325e-01] | [10.019116401672363, -0.043466828763484955] |
de2b2a10-9c25-4b97-94f3-647cd5afa7a5 | ho-3d-a-multi-user-multi-object-dataset-for | 1907.01481 | null | https://arxiv.org/abs/1907.01481v6 | https://arxiv.org/pdf/1907.01481v6.pdf | HOnnotate: A method for 3D Annotation of Hand and Object Poses | We propose a method for annotating images of a hand manipulating an object with the 3D poses of both the hand and the object, together with a dataset created using this method. Our motivation is the current lack of annotated real images for this problem, as estimating the 3D poses is challenging, mostly because of the mutual occlusions between the hand and the object. To tackle this challenge, we capture sequences with one or several RGB-D cameras and jointly optimize the 3D hand and object poses over all the frames simultaneously. This method allows us to automatically annotate each frame with accurate estimates of the poses, despite large mutual occlusions. With this method, we created HO-3D, the first markerless dataset of color images with 3D annotations for both the hand and object. This dataset is currently made of 77,558 frames, 68 sequences, 10 persons, and 10 objects. Using our dataset, we develop a single RGB image-based method to predict the hand pose when interacting with objects under severe occlusions and show it generalizes to objects not seen in the dataset. | ['Vincent Lepetit', 'Shreyas Hampali', 'Markus Oberweger', 'Mahdi Rad'] | 2019-07-02 | honnotate-a-method-for-3d-annotation-of-hand | http://openaccess.thecvf.com/content_CVPR_2020/html/Hampali_HOnnotate_A_Method_for_3D_Annotation_of_Hand_and_Object_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Hampali_HOnnotate_A_Method_for_3D_Annotation_of_Hand_and_Object_CVPR_2020_paper.pdf | cvpr-2020-6 | ['hand-object-pose'] | ['computer-vision'] | [ 1.09895386e-01 -9.20047015e-02 1.90455779e-01 -1.63291410e-01
-4.99729425e-01 -8.07064235e-01 2.74913698e-01 -5.39140940e-01
-3.45633626e-01 2.92231351e-01 1.64509520e-01 3.33504558e-01
1.72343850e-01 4.88882810e-02 -7.26106942e-01 -4.69010562e-01
2.80925259e-03 1.07758272e+00 4.14826095e-01 1.65311068e-01
1.02797583e-01 8.60246599e-01 -1.64262068e+00 1.82821184e-01
5.27823623e-03 9.50637758e-01 4.30448174e-01 1.07172811e+00
3.12171906e-01 7.91412055e-01 -5.99265397e-01 -2.34135836e-01
6.66931987e-01 -8.73286054e-02 -9.01797473e-01 7.34275997e-01
9.11736012e-01 -1.00705707e+00 -5.73695719e-01 5.36661148e-01
7.65935957e-01 1.53653711e-01 3.43126714e-01 -1.65462196e+00
8.49590376e-02 2.65516993e-02 -7.55369008e-01 -3.45532805e-01
1.05299926e+00 3.57036948e-01 8.35218251e-01 -7.83807218e-01
1.14506936e+00 1.57835448e+00 5.96498668e-01 6.74022853e-01
-1.14979732e+00 -2.81692088e-01 1.55132726e-01 1.31221756e-01
-1.54129672e+00 -4.60328400e-01 7.55863726e-01 -8.29053760e-01
8.34087968e-01 2.57330984e-01 1.10628724e+00 1.43782449e+00
-3.42155397e-01 9.63957429e-01 8.83018196e-01 -6.41235888e-01
3.93010601e-02 -4.46908772e-02 -1.09803885e-01 6.27414286e-01
1.62915383e-02 -4.86361384e-02 -6.91524148e-01 -1.81258246e-01
1.15874732e+00 2.58827031e-01 -3.54037195e-01 -8.40882540e-01
-1.43818164e+00 3.51010747e-02 1.46280229e-01 -8.93878341e-02
-4.22139823e-01 3.71598035e-01 2.91743856e-02 -3.89333457e-01
8.87736082e-02 1.06804483e-01 -6.42447889e-01 -4.60047334e-01
-5.74327707e-01 5.29692650e-01 8.81268203e-01 1.36852121e+00
4.64837819e-01 -4.48520243e-01 -6.39766827e-02 2.52912849e-01
3.25751036e-01 5.87229848e-01 -2.08842114e-01 -1.58159626e+00
4.77990597e-01 5.58351815e-01 6.09746039e-01 -9.03819442e-01
-5.17026305e-01 2.55082607e-01 -1.18592598e-01 5.83718598e-01
8.69604647e-01 3.36932726e-02 -8.82801592e-01 1.52446342e+00
7.37538755e-01 -1.41085893e-01 -3.48871499e-01 1.32909405e+00
5.01653671e-01 2.09763467e-01 -1.89420089e-01 4.73847538e-02
1.52412570e+00 -9.37983572e-01 -7.83098161e-01 -3.03150982e-01
1.61511377e-01 -8.70453835e-01 9.80741322e-01 6.27269626e-01
-1.08476114e+00 -6.48484111e-01 -6.38245225e-01 -2.76804924e-01
-1.23502530e-01 1.87444314e-01 6.37406826e-01 3.44271570e-01
-7.99673617e-01 3.31945866e-01 -1.11012948e+00 -4.24101889e-01
1.84970662e-01 4.68200415e-01 -6.86451554e-01 -1.95154384e-01
-4.34769988e-01 1.04417264e+00 4.26094562e-01 3.47620070e-01
-8.60814869e-01 -2.77056754e-01 -7.10833371e-01 -4.72218633e-01
7.05928087e-01 -4.52621371e-01 1.17937398e+00 -7.02865541e-01
-1.28538489e+00 1.08939373e+00 -9.52913463e-02 2.95587331e-01
9.18991446e-01 -7.14345276e-01 2.83250153e-01 3.35074127e-01
-8.38250965e-02 8.12500715e-01 8.10694516e-01 -1.63738763e+00
-3.74828726e-01 -8.65598738e-01 3.18304420e-01 2.59432048e-01
-4.45429236e-02 4.73544262e-02 -1.10135043e+00 -4.66888011e-01
2.18692496e-01 -1.34239328e+00 -2.17665397e-02 5.40047824e-01
-6.84650600e-01 -1.62596554e-01 1.17415941e+00 -8.93967867e-01
4.23692107e-01 -2.02680993e+00 6.95686817e-01 3.39244455e-02
1.47510856e-01 4.70574982e-02 -1.05327971e-01 9.13544223e-02
1.50554910e-01 -4.65764523e-01 7.94536546e-02 -8.51389468e-01
2.74817273e-02 5.38140178e-01 -4.79618423e-02 5.44211984e-01
-3.96610275e-02 6.96593225e-01 -8.71187210e-01 -6.74347341e-01
4.01835918e-01 9.50403988e-01 -4.25666004e-01 7.65460670e-01
-4.65003520e-01 8.67822230e-01 -4.39278066e-01 9.33790207e-01
5.47756016e-01 -2.18268797e-01 2.87385404e-01 -5.16461074e-01
1.37367353e-01 -1.48261085e-01 -1.72691739e+00 2.16452932e+00
-3.94318923e-02 4.54141110e-01 1.90869108e-01 -2.34085426e-01
3.63249987e-01 6.56931102e-01 8.54005039e-01 -4.79346374e-03
3.01033050e-01 1.20208882e-01 -3.11526567e-01 -8.82168174e-01
2.47173354e-01 2.42111787e-01 2.56631076e-01 7.07394898e-01
9.10313129e-02 -5.45222998e-01 8.52150172e-02 2.62822751e-02
9.79029477e-01 7.74654746e-01 -3.54971439e-02 3.80237013e-01
1.60770118e-02 -2.76048500e-02 1.51372999e-01 4.82066065e-01
-2.33786881e-01 1.06272042e+00 3.08427989e-01 -6.23457611e-01
-1.30995834e+00 -9.48328674e-01 1.27873644e-01 8.61809611e-01
2.28214175e-01 -3.30065042e-01 -7.64998198e-01 -4.49529737e-01
2.91121811e-01 3.27842869e-02 -6.57527208e-01 3.89101595e-01
-7.17519701e-01 -2.76185796e-02 2.27297857e-01 7.73046911e-01
1.81259528e-01 -9.82774734e-01 -1.08633173e+00 -5.20120151e-02
-4.11161453e-01 -1.61895728e+00 -6.70001090e-01 2.52671745e-02
-6.17100358e-01 -1.61102974e+00 -7.83939421e-01 -5.14825404e-01
7.64821589e-01 1.10107258e-01 1.02256203e+00 1.18937772e-02
-7.30586648e-01 1.00799119e+00 -3.39646041e-01 -1.76115066e-01
-1.52074277e-01 -2.51978546e-01 2.74287492e-01 -4.81805615e-02
4.82374765e-02 -4.00104553e-01 -5.77562630e-01 5.17160118e-01
-5.40818512e-01 1.47647426e-01 2.24775225e-01 2.81778902e-01
5.23559451e-01 -3.66350651e-01 -4.55198318e-01 -3.23632032e-01
-1.32793501e-01 2.20841974e-01 -5.84233403e-01 2.96357572e-01
3.29701841e-01 -2.03537956e-01 -1.06150649e-01 -7.58838773e-01
-9.37522233e-01 1.04970777e+00 1.80101335e-01 -8.32884550e-01
-3.03227484e-01 -3.29877764e-01 -4.09370124e-01 -1.36706308e-01
5.15192688e-01 -3.16180408e-01 4.62357439e-02 -6.65261447e-01
3.90570521e-01 6.97481811e-01 9.10389781e-01 -7.75633633e-01
5.94116628e-01 6.50375009e-01 1.10468499e-01 -6.61465883e-01
-8.92192006e-01 -5.60652733e-01 -1.47342467e+00 -5.86514771e-01
1.01214957e+00 -8.88993204e-01 -1.33734310e+00 6.71591282e-01
-1.66752517e+00 -4.51791942e-01 -2.85649270e-01 6.56138122e-01
-9.72042859e-01 3.67028803e-01 -4.20671433e-01 -1.09525561e+00
3.85434972e-03 -1.41146863e+00 1.72257864e+00 -2.61978418e-01
-4.95119125e-01 -5.50692677e-01 -1.63735971e-01 5.30842900e-01
-2.24233702e-01 7.09169984e-01 2.51898766e-01 -5.13731912e-02
-7.30905712e-01 -4.22771513e-01 -1.67772267e-02 2.57077754e-01
2.48397067e-01 -8.39860179e-03 -1.09221339e+00 -2.72986621e-01
-1.66565418e-01 -5.68512976e-01 1.62592217e-01 2.59795874e-01
1.14815736e+00 -1.04036681e-01 -2.81490803e-01 3.88901204e-01
9.88362908e-01 7.74475411e-02 4.40603256e-01 2.38650709e-01
1.22133470e+00 7.79722989e-01 5.51175833e-01 6.51806593e-01
3.15018356e-01 1.35342479e+00 6.93968296e-01 2.13113189e-01
-3.17530453e-01 -2.62644049e-02 2.61312395e-01 2.79343277e-01
-7.67620683e-01 -1.95446208e-01 -9.31271315e-01 2.24933773e-01
-1.68693900e+00 -6.41748965e-01 -1.91404194e-01 2.14030170e+00
8.81104946e-01 -2.72882015e-01 4.81891125e-01 4.08577830e-01
6.04293823e-01 -1.50230825e-01 -7.54844308e-01 5.57776928e-01
2.35248590e-03 3.90113853e-02 2.92873561e-01 5.78791738e-01
-9.77586627e-01 8.21546257e-01 6.89416456e+00 -3.09282565e-03
-7.25766838e-01 -1.05135784e-01 -1.07533544e-01 -3.58684063e-01
3.18368942e-01 2.21074410e-02 -6.55479074e-01 1.16127945e-01
1.70216665e-01 6.38679862e-01 6.38791442e-01 7.52548158e-01
-7.02285022e-03 -2.95517713e-01 -1.56741738e+00 1.19566190e+00
3.70783031e-01 -5.74943304e-01 -2.55796999e-01 9.59106311e-02
5.79465628e-01 -3.17837745e-01 -2.12935060e-01 -5.11244059e-01
6.94903135e-02 -8.10744286e-01 1.29945838e+00 6.48056328e-01
7.24017203e-01 -1.97905660e-01 3.68975431e-01 3.66621405e-01
-1.02018511e+00 6.00477196e-02 1.14132494e-01 -2.17457622e-01
3.94352347e-01 1.91285893e-01 -7.99253583e-01 1.19016552e-02
9.35597003e-01 6.24326408e-01 -4.64048684e-01 7.11522937e-01
-3.55248749e-01 -1.96231484e-01 -5.76026678e-01 4.61595356e-01
-3.83255035e-01 8.68008584e-02 5.93021750e-01 7.26745188e-01
1.45862877e-01 2.99750358e-01 4.78408247e-01 6.94719970e-01
1.93001151e-01 -5.00111997e-01 -4.47058588e-01 1.16204672e-01
3.19124132e-01 1.05193162e+00 -8.04694533e-01 -2.99725384e-01
-5.34125082e-02 1.47049546e+00 1.13253683e-01 4.13066566e-01
-6.08783841e-01 -5.37282676e-02 6.53301537e-01 2.21563354e-01
3.11434507e-01 -8.40093136e-01 1.03540225e-02 -1.23022473e+00
5.63784480e-01 -8.45478773e-01 1.10125065e-01 -1.39249706e+00
-9.27731335e-01 4.91525859e-01 1.96378782e-01 -1.14285171e+00
-3.83607954e-01 -9.69561160e-01 2.57979661e-01 7.50536740e-01
-8.57011080e-01 -1.38930070e+00 -9.65145469e-01 8.64211857e-01
4.68530983e-01 2.43799478e-01 8.94274414e-01 1.74350619e-01
-1.29185900e-01 -1.48164798e-02 -5.29950857e-01 2.03375280e-01
8.13051522e-01 -1.27100468e+00 1.94979787e-01 2.97227085e-01
3.30052584e-01 3.82003814e-01 7.76564181e-01 -4.63897347e-01
-2.00572920e+00 -4.45286691e-01 4.24928635e-01 -1.35388207e+00
1.76371232e-01 -8.51675272e-01 -4.85308677e-01 1.12532711e+00
-2.42790982e-01 4.11195129e-01 1.95790231e-01 -2.00521514e-01
-3.87634695e-01 8.49706531e-02 -1.05827212e+00 2.87854671e-01
1.46634185e+00 -6.40243888e-01 -5.29561818e-01 7.00441360e-01
4.81597275e-01 -1.22589552e+00 -9.28578377e-01 8.90128613e-02
1.19959748e+00 -9.87410605e-01 1.31011164e+00 -4.34632897e-01
1.55746460e-01 -5.96437633e-01 -4.13039953e-01 -8.52519572e-01
2.88778812e-01 -4.41379696e-01 -5.65319359e-01 9.30518031e-01
-3.72251749e-01 -1.85259222e-03 9.87635434e-01 1.08439386e+00
3.94659132e-01 -3.27036172e-01 -9.44506586e-01 -7.11650014e-01
-5.46385705e-01 -6.03646934e-01 4.03100610e-01 5.16765356e-01
-2.73071945e-01 -8.95423293e-02 -6.87122941e-01 2.23025337e-01
8.22233260e-01 1.20319739e-01 1.42117548e+00 -1.19409275e+00
-3.92417043e-01 3.24964821e-02 -7.18322635e-01 -1.17497373e+00
2.67160445e-01 -2.50623584e-01 3.27950120e-01 -1.26501048e+00
4.16379958e-01 -4.21961457e-01 4.71001565e-01 7.81365871e-01
-1.63612291e-01 5.01389027e-01 4.59871531e-01 4.53191012e-01
-5.27708352e-01 1.59700841e-01 1.43491030e+00 -4.55017425e-02
-1.41663924e-01 -8.32982510e-02 1.27134562e-01 8.43850553e-01
1.48043096e-01 -3.30904216e-01 -7.58013651e-02 -6.66224182e-01
-5.14786243e-02 2.43621245e-01 9.59304631e-01 -9.96527612e-01
1.24680541e-01 -1.35136202e-01 6.98516369e-01 -8.21818829e-01
9.33368266e-01 -1.39016271e+00 7.44731724e-01 4.51818973e-01
-2.79205710e-01 1.05398819e-01 3.57812718e-02 4.15032476e-01
2.65562832e-01 4.35546823e-02 4.16293472e-01 -4.02842790e-01
-7.06133604e-01 3.62628162e-01 2.05710500e-01 -1.28559485e-01
1.19395781e+00 -3.41148674e-01 6.73456639e-02 -4.20868397e-01
-1.15578067e+00 5.35023548e-02 7.04421401e-01 6.11493528e-01
5.15005529e-01 -1.31859577e+00 -4.88175124e-01 3.84834588e-01
2.65213456e-02 5.28737187e-01 1.06662728e-01 6.75409377e-01
-5.91679573e-01 1.15029126e-01 -3.39477688e-01 -1.10684299e+00
-1.49478245e+00 4.80720669e-01 2.91411430e-01 3.38253081e-01
-7.24348545e-01 7.07723379e-01 -1.07150994e-01 -4.40763891e-01
7.37150192e-01 -2.48626739e-01 2.64737874e-01 -4.95343748e-03
5.99347532e-01 5.97179651e-01 5.97811900e-02 -9.18609738e-01
-4.67418134e-01 1.10375857e+00 3.55858147e-01 -3.02833468e-01
1.35262132e+00 -1.37636527e-01 -1.96364358e-01 5.45103550e-01
1.25786078e+00 -1.74255162e-01 -1.87753904e+00 -2.42860198e-01
-3.68881077e-01 -1.07660651e+00 -3.60935986e-01 -6.71285927e-01
-1.12137449e+00 9.27402258e-01 6.05779409e-01 -1.39919475e-01
8.20044219e-01 3.61544728e-01 5.34460604e-01 5.31639934e-01
7.63103843e-01 -9.64414597e-01 6.12638295e-01 4.08765316e-01
1.24193406e+00 -1.18597949e+00 1.76918387e-01 -4.86351281e-01
-5.20263314e-01 1.21561205e+00 6.80355430e-01 1.33719653e-01
3.17525744e-01 3.73466045e-01 8.42979848e-02 -2.27775753e-01
-1.96538121e-01 -1.81058541e-01 2.35215470e-01 7.37607121e-01
1.76305741e-01 6.84305886e-03 5.07296145e-01 2.34345309e-02
-1.50122210e-01 8.64098892e-02 2.47326002e-01 1.42208278e+00
1.34817079e-01 -1.13354313e+00 -8.48517358e-01 -5.22473641e-02
-1.75403669e-01 5.33987164e-01 -5.45149565e-01 8.91843498e-01
3.06823492e-01 7.58731067e-01 5.87042235e-02 -2.97983170e-01
7.27328479e-01 1.60762928e-02 1.21599960e+00 -5.44331849e-01
-2.21567035e-01 1.10614397e-01 -1.88120052e-01 -8.29777658e-01
-9.93098497e-01 -8.93083274e-01 -1.09236717e+00 -1.06083333e-01
-3.45841467e-01 -3.72599751e-01 8.20017934e-01 8.83249283e-01
1.67857885e-01 1.76806748e-01 3.42982084e-01 -1.90266299e+00
-3.60141426e-01 -8.37025940e-01 -8.46403360e-01 7.83447266e-01
6.03237987e-01 -1.15014160e+00 -2.18520895e-01 5.14120638e-01] | [6.547422885894775, -1.0084484815597534] |
f1289b35-17cc-4bae-bfbc-16f31fdd3279 | convergence-of-first-order-algorithms-for | 2301.06806 | null | https://arxiv.org/abs/2301.06806v1 | https://arxiv.org/pdf/2301.06806v1.pdf | Convergence of First-Order Algorithms for Meta-Learning with Moreau Envelopes | In this work, we consider the problem of minimizing the sum of Moreau envelopes of given functions, which has previously appeared in the context of meta-learning and personalized federated learning. In contrast to the existing theory that requires running subsolvers until a certain precision is reached, we only assume that a finite number of gradient steps is taken at each iteration. As a special case, our theory allows us to show the convergence of First-Order Model-Agnostic Meta-Learning (FO-MAML) to the vicinity of a solution of Moreau objective. We also study a more general family of first-order algorithms that can be viewed as a generalization of FO-MAML. Our main theoretical achievement is a theoretical improvement upon the inexact SGD framework. In particular, our perturbed-iterate analysis allows for tighter guarantees that improve the dependency on the problem's conditioning. In contrast to the related work on meta-learning, ours does not require any assumptions on the Hessian smoothness, and can leverage smoothness and convexity of the reformulation based on Moreau envelopes. Furthermore, to fill the gaps in the comparison of FO-MAML to the Implicit MAML (iMAML), we show that the objective of iMAML is neither smooth nor convex, implying that it has no convergence guarantees based on the existing theory. | ['Peter Richtárik', 'Slavomír Hanzely', 'Konstantin Mishchenko'] | 2023-01-17 | null | null | null | null | ['personalized-federated-learning'] | ['methodology'] | [-5.91685921e-02 3.31418753e-01 -3.23330432e-01 -7.70170242e-02
-1.00597131e+00 -6.15200520e-01 3.67610753e-01 3.80377710e-01
-3.39995861e-01 6.88470960e-01 7.47649968e-02 -3.47398221e-01
-5.67510366e-01 -8.12806189e-01 -1.18912554e+00 -8.49529386e-01
-2.59194642e-01 3.02014619e-01 -1.66035458e-01 -5.38548052e-01
1.94858149e-01 2.99714774e-01 -1.32430542e+00 1.66763425e-01
1.04414082e+00 1.07641876e+00 -2.05312416e-01 7.08234966e-01
-1.08911417e-01 7.07918942e-01 -5.64795732e-02 -4.02439922e-01
6.11532867e-01 -4.56204295e-01 -1.21423876e+00 6.56726807e-02
6.97211742e-01 -6.94327056e-02 -2.12797970e-01 1.28468704e+00
2.74544805e-01 4.79542464e-01 3.30479801e-01 -1.45718241e+00
-1.97423011e-01 8.39495361e-01 -2.60983199e-01 -1.06696889e-01
-8.93766619e-03 -8.35118145e-02 1.23635757e+00 -7.14470327e-01
5.08089542e-01 1.02740586e+00 9.13274229e-01 4.93378878e-01
-1.23447549e+00 -1.32834941e-01 4.45427388e-01 7.54635930e-02
-1.06188536e+00 -3.17498863e-01 4.41809386e-01 -3.04538101e-01
5.96249700e-01 5.94498813e-01 5.06658614e-01 5.74533224e-01
-1.34011395e-02 9.94116604e-01 9.85309482e-01 -6.96032226e-01
4.72568512e-01 1.39963746e-01 3.87401372e-01 1.08400261e+00
2.92929262e-01 1.34327397e-01 -5.10452271e-01 -3.37377548e-01
4.78966653e-01 9.96869057e-02 -4.19229746e-01 -5.30963540e-01
-1.06123471e+00 9.33540702e-01 3.71898323e-01 4.99961168e-01
-2.05353290e-01 2.49676570e-01 3.52230757e-01 7.07573473e-01
6.83153629e-01 4.91090953e-01 -5.13256609e-01 -1.58202663e-01
-1.08083951e+00 2.85874099e-01 1.24485958e+00 9.10455227e-01
1.07602012e+00 2.94178016e-02 1.53679531e-02 3.15230876e-01
6.19765697e-03 1.77245736e-01 3.19703072e-01 -1.09513450e+00
4.48875368e-01 5.13554513e-01 2.38765880e-01 -7.40287244e-01
-2.65178084e-01 -7.33725011e-01 -5.87855279e-01 2.10464507e-01
7.30367780e-01 -1.71711579e-01 -3.70972931e-01 1.88050902e+00
5.34850597e-01 3.34258527e-01 -3.47674750e-02 8.72315228e-01
5.16009405e-02 5.50647259e-01 -4.75254774e-01 -4.96936023e-01
6.81009293e-01 -1.25244236e+00 -4.27464038e-01 4.04785573e-02
1.10175872e+00 -3.00673217e-01 1.03780258e+00 5.36104500e-01
-1.37495804e+00 -1.19701721e-01 -9.74752426e-01 -7.59432912e-02
-3.04484665e-01 -5.01766384e-01 9.78365421e-01 7.97883093e-01
-1.28477073e+00 1.02116251e+00 -9.31598902e-01 -1.48686424e-01
2.15001270e-01 2.16770664e-01 -2.07166553e-01 6.47920519e-02
-8.59745145e-01 6.57640159e-01 4.32684779e-01 1.90749526e-01
-9.16624308e-01 -1.17257595e+00 -6.65653586e-01 8.10853913e-02
8.21699321e-01 -7.32245505e-01 1.32646489e+00 -1.25469995e+00
-1.62096798e+00 7.13622034e-01 -2.41388768e-01 -6.27037883e-01
1.01776588e+00 -2.56809860e-01 2.08257362e-01 -9.78412852e-03
-2.24989831e-01 -2.07017109e-01 7.41757035e-01 -9.50768054e-01
-8.50504577e-01 -4.52005208e-01 7.17522919e-01 1.40327528e-01
-4.67215151e-01 -4.76750553e-01 -1.31654307e-01 -3.42453927e-01
-1.43286632e-03 -7.38357902e-01 -5.62597394e-01 -6.10141046e-02
-9.40739512e-02 -2.83328176e-01 5.60837686e-01 -1.17904916e-01
1.17897439e+00 -2.02467155e+00 6.38259649e-01 2.98584849e-01
2.14552000e-01 7.82394856e-02 -8.29055533e-02 7.57453978e-01
1.26799300e-01 2.95863807e-01 -3.91545326e-01 -5.84453225e-01
4.57390964e-01 1.44228041e-01 -3.83358747e-01 9.68724728e-01
-3.98290962e-01 7.75754571e-01 -1.03781021e+00 -2.63164252e-01
2.81816162e-02 1.40037417e-01 -9.49379206e-01 6.13255939e-03
-5.88857114e-01 1.69415161e-01 -5.69445550e-01 3.73999327e-01
5.74125767e-01 -4.71317410e-01 2.03812242e-01 -6.94645941e-02
-3.03031594e-01 2.70294815e-01 -1.22915375e+00 1.92115498e+00
-8.62694323e-01 9.36795250e-02 6.12123728e-01 -1.34517741e+00
1.74192458e-01 1.60050005e-01 7.54626036e-01 -1.48062170e-01
7.95024559e-02 4.72437173e-01 -4.05574679e-01 -2.07048804e-01
3.91821474e-01 -2.66755790e-01 1.57676861e-01 4.60067570e-01
-3.60814817e-02 2.98466414e-01 2.29198113e-01 2.54623681e-01
9.27912414e-01 1.06364809e-01 6.09255061e-02 -6.24673188e-01
6.72168016e-01 2.55065914e-02 4.97306705e-01 9.85329807e-01
1.83447242e-01 1.23631477e-01 4.63570476e-01 -2.92587459e-01
-9.29833353e-01 -8.52256596e-01 -1.55676425e-01 1.38595462e+00
-4.26890813e-02 -6.85293853e-01 -8.78393173e-01 -8.34155917e-01
1.29451141e-01 5.75203001e-01 -6.63295925e-01 3.50826792e-02
-5.14648736e-01 -7.35798597e-01 2.11421803e-01 1.56316921e-01
4.29786175e-01 -2.48636752e-01 -4.27418470e-01 1.46630749e-01
5.95474951e-02 -5.88103592e-01 -6.96652591e-01 1.43076077e-01
-1.04366016e+00 -1.20122504e+00 -5.28883934e-01 -5.06304741e-01
5.13205647e-01 6.60369545e-02 1.00800097e+00 1.02575481e-01
7.84877539e-02 8.96881640e-01 -2.19635099e-01 -1.60964847e-01
-3.91471475e-01 4.74962354e-01 -3.89088131e-02 3.07465792e-01
-1.12000704e-01 -6.27279758e-01 -5.15404880e-01 -1.28760591e-01
-9.72124636e-01 7.22446218e-02 2.57847071e-01 1.00185800e+00
5.93592763e-01 1.78530529e-01 5.01110494e-01 -1.08406711e+00
3.14048052e-01 -7.23768532e-01 -9.90956008e-01 3.66658330e-01
-9.33663070e-01 3.97550076e-01 9.89052355e-01 -2.03589350e-01
-7.95791626e-01 -1.10475525e-01 1.16103061e-01 -5.48794806e-01
3.98323298e-01 7.41582155e-01 -1.53690815e-01 -3.52080792e-01
4.71030056e-01 1.72886118e-01 1.25830919e-01 -4.81514215e-01
6.02142751e-01 3.68337780e-01 4.45932150e-01 -1.04968190e+00
8.49528134e-01 6.03647232e-01 4.17193651e-01 -7.33645916e-01
-1.20417917e+00 -4.03465211e-01 -3.31584811e-01 -9.42452103e-02
3.04129422e-01 -5.67522943e-01 -1.30239391e+00 9.23438519e-02
-6.92893147e-01 -6.15294755e-01 -5.61267316e-01 2.98385888e-01
-8.90962780e-01 5.34038186e-01 -5.94287455e-01 -1.20114207e+00
-3.88527155e-01 -8.58079255e-01 6.37292385e-01 -1.01271667e-01
2.26635039e-01 -1.59979963e+00 6.82248771e-02 1.78261951e-01
5.10955453e-01 4.08103615e-01 7.79145956e-01 -6.28156304e-01
-7.57488251e-01 -9.70493481e-02 3.39013278e-01 3.07219416e-01
-4.77494225e-02 -3.33695173e-01 -7.64303207e-01 -7.19682097e-01
5.65013111e-01 -3.52052450e-01 8.97012472e-01 2.08373889e-01
1.05331135e+00 -1.02212107e+00 -9.18018669e-02 1.03431427e+00
1.75188422e+00 -4.57398593e-01 1.04319826e-01 5.77738941e-01
5.69278359e-01 4.90590602e-01 5.73949575e-01 6.61176205e-01
2.98932463e-01 4.47532773e-01 6.57890499e-01 1.70098573e-01
4.64006215e-01 -2.35744596e-01 4.49673921e-01 6.03923678e-01
-5.30157462e-02 1.55995727e-01 -5.29353976e-01 5.54009855e-01
-2.26666832e+00 -9.82115269e-01 -6.29954487e-02 2.63253212e+00
9.99265134e-01 -2.83647895e-01 4.36669558e-01 7.34534413e-02
5.55073559e-01 1.80255026e-01 -6.54924035e-01 -4.94769156e-01
-2.89145354e-02 3.92209679e-01 9.36435938e-01 9.42323565e-01
-9.03448522e-01 6.44007862e-01 6.40590763e+00 7.82179415e-01
-1.07231295e+00 4.53009158e-01 3.59589398e-01 -4.49223459e-01
-4.86334413e-01 1.79303288e-01 -6.07558072e-01 3.87155026e-01
1.03422582e+00 -5.30545771e-01 1.03542304e+00 1.16416121e+00
4.99863401e-02 2.27168247e-01 -1.43592167e+00 6.87507629e-01
-2.07842022e-01 -1.43086421e+00 -3.81899834e-01 2.94002980e-01
1.04706049e+00 -7.92984813e-02 2.76259750e-01 2.15840474e-01
4.81502086e-01 -9.74130511e-01 7.04207361e-01 4.04216886e-01
4.15320754e-01 -1.10805261e+00 3.73095632e-01 5.19847214e-01
-9.67685044e-01 -2.87537396e-01 -3.88997316e-01 -2.20941022e-01
-1.98606342e-01 6.84046447e-01 -3.56289089e-01 1.06246400e+00
2.24290818e-01 6.22300744e-01 -6.09834902e-02 9.39247489e-01
2.39238963e-01 6.46454394e-01 -5.57405591e-01 9.64505002e-02
5.33625007e-01 -4.31898594e-01 6.37428164e-01 1.13428390e+00
1.42866865e-01 -8.45767781e-02 4.33839023e-01 7.78861880e-01
-2.44848862e-01 3.11899960e-01 -2.83283949e-01 -2.06931289e-02
3.05862546e-01 1.19430387e+00 -6.78134477e-03 -2.57480562e-01
-4.93768066e-01 9.09161091e-01 5.65438211e-01 2.36316457e-01
-5.55877328e-01 -1.64999053e-01 6.19746685e-01 -4.26669419e-02
3.52732033e-01 -1.51896596e-01 -1.68905914e-01 -1.39794254e+00
2.85982639e-01 -1.02985287e+00 7.43120492e-01 2.09739611e-01
-1.50885499e+00 2.07993627e-01 -5.35255112e-02 -7.61965811e-01
-3.05872589e-01 -5.21290779e-01 -5.29061377e-01 8.79824877e-01
-1.58178997e+00 -1.10045481e+00 2.12113917e-01 7.56737232e-01
3.50499779e-01 1.41472653e-01 6.43839657e-01 1.89363986e-01
-4.62293565e-01 8.06683362e-01 4.95274812e-01 -4.41073745e-01
3.62049848e-01 -1.48603749e+00 -9.95126069e-02 1.03085089e+00
-6.91221207e-02 6.48904681e-01 8.56056988e-01 -3.23415071e-01
-2.11650801e+00 -1.15194809e+00 6.20782614e-01 -3.25233698e-01
9.51867938e-01 -1.92990780e-01 -7.38125324e-01 1.00781703e+00
-1.01179257e-01 1.91131175e-01 3.53768796e-01 4.32871103e-01
-2.63702005e-01 -2.69757271e-01 -1.14682901e+00 4.84277010e-01
1.23495257e+00 -4.05280828e-01 -2.07782999e-01 6.71115637e-01
6.58682108e-01 -5.86251557e-01 -1.18410659e+00 2.00113431e-01
3.48718196e-01 -9.04747248e-01 7.28750288e-01 -1.03710830e+00
2.14044988e-01 -7.69483224e-02 -5.23316979e-01 -1.10207760e+00
-2.41553977e-01 -1.25932097e+00 -6.58133984e-01 1.04855752e+00
1.82717204e-01 -9.72011089e-01 9.60246146e-01 7.26481795e-01
-3.66783470e-01 -1.00445461e+00 -1.03142834e+00 -1.14132190e+00
6.04299784e-01 -4.74354237e-01 5.50196767e-01 1.01521397e+00
3.38403523e-01 -9.84450802e-02 -5.42368710e-01 4.32445407e-01
9.48556900e-01 2.54171133e-01 7.21674502e-01 -1.00453174e+00
-9.08986509e-01 -6.67911291e-01 -1.29108325e-01 -1.08200109e+00
2.82566875e-01 -1.44277036e+00 -1.70656264e-01 -1.21817374e+00
8.09096321e-02 -6.70613587e-01 -4.65517819e-01 3.73788327e-01
-4.92842644e-02 -3.22254479e-01 3.55842024e-01 4.47713621e-02
-7.06681907e-01 5.30323684e-01 1.09772003e+00 3.21739732e-04
-2.36465901e-01 2.54825711e-01 -7.14115560e-01 5.41216373e-01
6.53447449e-01 -2.62099266e-01 -4.37037587e-01 -2.05844879e-01
4.75691855e-01 1.88622475e-01 3.57853144e-01 -7.22692490e-01
4.75221962e-01 -3.80038649e-01 -3.81146878e-01 -3.17699760e-01
7.60466158e-02 -8.19627285e-01 4.71264422e-02 6.36709869e-01
-5.62531948e-01 -1.09626614e-01 -7.95972794e-02 5.84916949e-01
1.32796288e-01 -6.28394127e-01 5.87799549e-01 -2.64522940e-01
-1.22245215e-01 5.95786572e-01 -9.86064076e-02 3.87754202e-01
6.19513988e-01 1.37237802e-01 -1.58720076e-01 -3.56138676e-01
-5.41446209e-01 4.29653674e-01 7.29182959e-01 -2.44262502e-01
2.04980180e-01 -1.19268668e+00 -6.78239346e-01 -2.27116495e-01
-3.14202696e-01 6.36760592e-02 3.42861444e-01 1.22262776e+00
-1.68584540e-01 4.38570201e-01 4.02050257e-01 -3.84587735e-01
-8.52330744e-01 7.46271610e-01 6.79813027e-01 -4.72081125e-01
-6.63830757e-01 5.80216825e-01 7.46361390e-02 -3.21614653e-01
4.33153033e-01 -1.31263539e-01 6.72695637e-01 -3.11101288e-01
4.65220779e-01 8.05766940e-01 1.74943700e-01 -8.93498510e-02
-2.46838540e-01 1.85030371e-01 -1.66676164e-01 -2.55232990e-01
1.44652951e+00 -7.13766441e-02 -2.45737240e-01 6.07661724e-01
1.40280199e+00 3.11766922e-01 -1.32871485e+00 -3.93596888e-01
7.38294721e-02 -3.51828665e-01 7.08940998e-02 -5.98656178e-01
-1.13408685e+00 6.63901925e-01 2.98024684e-01 3.04215312e-01
1.02729893e+00 -1.85139835e-01 7.47925162e-01 7.25660801e-01
6.11947238e-01 -1.13313353e+00 -1.89500824e-01 4.55352277e-01
6.56977177e-01 -9.88048553e-01 -9.23788846e-02 -1.04697764e-01
-1.30885914e-01 1.28292644e+00 2.13312879e-01 -2.25577742e-01
4.43613231e-01 1.17705315e-01 -4.98695612e-01 3.92736867e-03
-9.28556979e-01 -1.40812308e-01 1.50322065e-01 1.78225011e-01
3.10331792e-01 -1.69073895e-01 -4.35169399e-01 4.54906464e-01
-4.00833547e-01 1.86020453e-02 4.59080696e-01 9.67162967e-01
-5.68769753e-01 -1.22656727e+00 -1.78420186e-01 1.97501346e-01
-7.69086063e-01 -5.20286784e-02 -5.25331683e-02 7.79886425e-01
-1.78147063e-01 9.48270261e-01 -2.06005603e-01 -1.43588847e-02
1.11739021e-02 1.67510644e-01 8.65491092e-01 -4.71704215e-01
-7.76011884e-01 -8.36904645e-02 -6.22916706e-02 -6.26759470e-01
-3.84249300e-01 -5.32992125e-01 -1.19088995e+00 -7.94080853e-01
-3.79119903e-01 4.91970748e-01 6.61089599e-01 1.07247674e+00
1.88158244e-01 1.73173919e-01 9.14798498e-01 -5.70524991e-01
-1.24879742e+00 -5.39706647e-01 -5.81170499e-01 2.88037986e-01
6.17219448e-01 -1.74245000e-01 -7.55420804e-01 -3.35467398e-01] | [6.343122482299805, 4.737835884094238] |
245b073e-13cd-4d16-8e9f-ca1c6667a1d9 | light-yolov5-a-lightweight-algorithm-for | 2208.13422 | null | https://arxiv.org/abs/2208.13422v3 | https://arxiv.org/pdf/2208.13422v3.pdf | Light-YOLOv5: A Lightweight Algorithm for Improved YOLOv5 in Complex Fire Scenarios | Fire-detection technology is of great importance for successful fire-prevention measures. Image-based fire detection is one effective method. At present, object-detection algorithms are deficient in performing detection speed and accuracy tasks when they are applied in complex fire scenarios. In this study, a lightweight fire-detection algorithm, Light-YOLOv5 (You Only Look Once version five), is presented. First, a separable vision transformer (SepViT) block is used to replace several C3 modules in the final layer of a backbone network to enhance both the contact of the backbone network to global in-formation and the extraction of flame and smoke features; second, a light bidirectional feature pyramid network (Light-BiFPN) is designed to lighten the model while improving the feature extraction and balancing speed and accuracy features during a fire-detection procedure; third, a global attention mechanism (GAM) is fused into the network to cause the model to focus more on the global dimensional features and further improve the detection accuracy of the model; and finally, the Mish activation function and SIoU loss are utilized to simultaneously increase the convergence speed and enhance the accuracy. The experimental results show that compared to the original algorithm, the mean average accuracy (mAP) of Light-YOLOv5 increases by 3.3%, the number of parameters decreases by 27.1%, and the floating point operations (FLOPs) decrease by 19.1%. The detection speed reaches 91.1 FPS, which can detect targets in complex fire scenarios in real time. | ['Fei Zhong', 'Bo Li', 'Hao Xu'] | 2022-08-29 | null | null | null | null | ['fire-detection'] | ['time-series'] | [ 1.74173653e-01 -8.64803433e-01 1.60973668e-01 2.66429543e-01
1.45038873e-01 -1.35350272e-01 3.14747214e-01 -3.34713012e-01
-6.11130774e-01 3.63422751e-01 -3.50982159e-01 -6.43182397e-02
-2.34094262e-02 -1.23375225e+00 -2.87380010e-01 -1.09204876e+00
1.31833121e-01 -3.14777732e-01 8.43661427e-01 -1.38181582e-01
1.18235022e-01 7.78913021e-01 -1.53001165e+00 1.59128487e-01
8.09241652e-01 1.24796367e+00 5.37065804e-01 4.44974840e-01
1.26167193e-01 6.02973044e-01 -3.12352777e-01 2.14254007e-01
3.88096362e-01 -1.97898671e-01 -1.57810703e-01 -4.18546759e-02
3.66043784e-02 -7.63202310e-01 -5.75226128e-01 1.09486079e+00
7.18324959e-01 1.15969323e-01 6.48683846e-01 -1.34439552e+00
-2.56407738e-01 1.62273601e-01 -9.70880687e-01 5.15920460e-01
-1.49280295e-01 7.07192719e-01 3.39231849e-01 -1.12201619e+00
7.80200809e-02 1.20494485e+00 6.91536367e-01 2.76649565e-01
-9.04204249e-01 -1.05275261e+00 7.83765912e-02 4.33855414e-01
-1.60145509e+00 -2.09050208e-01 6.91706479e-01 -2.86415547e-01
8.21245134e-01 3.09105664e-01 1.13580382e+00 6.51381850e-01
1.67242005e-01 4.18878585e-01 7.92865217e-01 -1.61401570e-01
-1.06872030e-01 -2.72549927e-01 -6.61192536e-02 9.67849612e-01
5.38275540e-01 5.56121349e-01 -1.43956289e-01 4.80328739e-01
1.23548150e+00 5.17905056e-01 -4.45327997e-01 3.04988056e-01
-9.47455049e-01 7.68492579e-01 1.11825478e+00 3.66335720e-01
-6.10720396e-01 4.55354825e-02 3.45262736e-01 -2.15051830e-01
2.38210008e-01 -3.37668024e-02 -6.16705641e-02 1.37566432e-01
-8.46451819e-01 1.70470446e-01 2.76982039e-01 6.88205838e-01
6.85263574e-01 3.79977822e-01 -4.44301158e-01 6.79659545e-01
3.40949684e-01 1.09622514e+00 5.57511253e-03 -6.47170424e-01
2.98566788e-01 1.06525242e+00 -2.68760938e-02 -1.21372151e+00
-6.00537360e-01 -4.85997826e-01 -1.27572429e+00 5.58232188e-01
2.11520776e-01 -1.92575723e-01 -8.84067297e-01 1.23431134e+00
3.17767143e-01 2.00116962e-01 -4.23600435e-01 1.20029521e+00
8.21159661e-01 9.33127344e-01 1.88282073e-01 -4.13855970e-01
1.42240191e+00 -7.59469390e-01 -3.51803064e-01 -2.59626418e-01
-9.64786708e-02 -6.92079127e-01 8.49296153e-01 1.48530424e-01
-8.34570646e-01 -8.79943132e-01 -9.68583226e-01 3.06647003e-01
-3.48008305e-01 2.40334690e-01 3.76770049e-01 2.49467373e-01
-5.88714004e-01 2.79545218e-01 -5.61257064e-01 -4.01844233e-01
6.33473456e-01 2.43946865e-01 2.23585844e-01 -2.44121000e-01
-1.12067890e+00 7.00624883e-01 6.02794945e-01 3.66868019e-01
-1.11044598e+00 -7.60983467e-01 -4.09162492e-01 2.86688358e-01
2.66028017e-01 -7.77258217e-01 7.59006619e-01 -5.55237830e-01
-1.11249709e+00 3.24079841e-01 2.88643893e-02 -8.03148821e-02
5.87527215e-01 -3.30607630e-02 -3.94326359e-01 3.62882555e-01
-1.47557169e-01 8.16554666e-01 9.47992086e-01 -1.12473381e+00
-1.06624806e+00 -3.19704622e-01 1.65593699e-01 3.47621709e-01
-2.95605898e-01 1.73188463e-01 -1.37493908e-01 -5.06620228e-01
6.26965612e-02 -6.04026258e-01 -1.74443468e-01 4.77190226e-01
-6.19961023e-02 -1.25110000e-01 1.39918041e+00 -4.96439159e-01
1.31451857e+00 -2.21098185e+00 -1.13677062e-01 1.58351108e-01
1.87130615e-01 8.14111471e-01 4.48258780e-02 5.00720926e-02
6.04380928e-02 -5.56124896e-02 -4.55907166e-01 3.05778652e-01
-5.35970032e-01 3.68557721e-02 7.93281272e-02 4.43733066e-01
2.03102350e-01 6.78490400e-01 -7.39277244e-01 -5.57496190e-01
6.78663790e-01 7.13335335e-01 -4.49184805e-01 6.91767633e-02
9.62784514e-02 1.38996288e-01 -4.59432930e-01 6.67158127e-01
9.93817329e-01 1.29230484e-01 -4.82303172e-01 -4.73437428e-01
-6.28701806e-01 -2.88190544e-01 -1.39180338e+00 1.07092488e+00
-3.94423008e-01 3.03754389e-01 3.83524716e-01 -5.09336174e-01
1.07230175e+00 2.10173070e-01 6.16240978e-01 -6.93802953e-01
5.43579876e-01 -7.67308548e-02 -7.17633590e-02 -6.98856890e-01
-7.63640702e-02 -1.30837739e-01 4.09778297e-01 1.17359586e-01
-4.35422629e-01 -9.34375450e-02 2.98433423e-01 -3.02660972e-01
8.27638388e-01 -2.51037806e-01 -2.36256067e-02 -2.88078368e-01
7.36073732e-01 8.31551626e-02 5.84532499e-01 4.84056115e-01
-2.92179197e-01 2.35755891e-01 -3.22497249e-01 -6.17612660e-01
-7.24009991e-01 -9.88955438e-01 -1.98238641e-01 7.23433435e-01
5.48345506e-01 -2.66544104e-01 -6.96656406e-01 -4.90166605e-01
-3.02975513e-02 5.51431358e-01 -1.09341338e-01 -5.86501539e-01
-4.88879114e-01 -8.71870637e-01 3.50949973e-01 7.15321302e-01
1.34662807e+00 -1.32886183e+00 -1.17140090e+00 2.76587546e-01
-3.19218159e-01 -7.74060905e-01 -2.21078798e-01 2.24593934e-03
-7.58396447e-01 -1.20637476e+00 -5.37818015e-01 -7.65808284e-01
5.07302403e-01 8.80037308e-01 3.97437990e-01 4.52228904e-01
-7.07617521e-01 -7.42349774e-02 -4.12438750e-01 -5.76403856e-01
1.20088317e-01 -1.76995978e-01 -5.25373593e-02 -7.83846825e-02
5.04570425e-01 -6.06760025e-01 -9.21417058e-01 3.44366193e-01
-9.53695536e-01 2.14487746e-01 6.24837101e-01 7.80278563e-01
8.42604786e-02 4.66320634e-01 9.59771685e-03 -3.62685807e-02
3.63595515e-01 -7.16919266e-03 -7.24334002e-01 3.16875242e-02
-6.36511266e-01 -5.74251771e-01 7.66964018e-01 -3.74535531e-01
-1.01939154e+00 -1.54500036e-02 1.23424627e-01 -5.62853932e-01
-1.12195522e-01 1.18766241e-01 -5.48023507e-02 -2.94898182e-01
5.74493825e-01 4.55337822e-01 3.09208836e-02 -2.09337458e-01
2.73548104e-02 6.84670806e-01 4.45178449e-01 -9.39956233e-02
1.12733459e+00 4.58493322e-01 3.21790963e-01 -9.52618003e-01
-5.09083807e-01 -2.89253891e-01 -6.07725441e-01 -8.06133389e-01
8.68589699e-01 -8.29041719e-01 -1.19773281e+00 8.47767413e-01
-1.21183658e+00 6.39155135e-02 -4.94791791e-02 5.97929120e-01
-1.07868515e-01 1.62519053e-01 -6.83127701e-01 -9.14957047e-01
-8.78167868e-01 -1.06057227e+00 7.04436839e-01 8.30664575e-01
4.95227456e-01 -3.16262096e-01 -4.03620839e-01 6.49258345e-02
6.38431609e-01 -4.32644556e-05 6.44967496e-01 3.33806574e-01
-7.63697743e-01 2.37442758e-02 -8.80921304e-01 3.74067664e-01
1.69786856e-01 2.66090751e-01 -8.91116977e-01 -1.92833513e-01
-7.29132295e-02 2.41233006e-01 1.17249525e+00 5.70743382e-01
1.07180870e+00 -7.42176473e-02 -4.97808665e-01 7.81606555e-01
1.64125609e+00 6.96534932e-01 5.80484450e-01 4.28239405e-01
7.96387613e-01 2.01996088e-01 5.17915368e-01 6.58643246e-01
4.83420901e-02 3.87275010e-01 9.50263262e-01 -5.14083266e-01
-4.44264293e-01 -1.64996952e-01 4.30037349e-01 4.86849189e-01
-6.22205198e-01 -3.38552631e-02 -5.97552776e-01 2.35334218e-01
-1.68580449e+00 -1.42890894e+00 -4.97998238e-01 1.92463386e+00
3.55774105e-01 1.45426378e-01 1.38075069e-01 3.72860581e-01
1.02034712e+00 1.70155093e-01 -6.27879441e-01 1.88833907e-01
4.85774688e-02 -1.58428699e-02 6.04316294e-01 2.89400041e-01
-1.31367075e+00 8.48996520e-01 5.79069853e+00 8.14752162e-01
-1.28327143e+00 -4.09574658e-02 1.69068128e-01 -8.24110135e-02
3.68331462e-01 -1.16631001e-01 -8.16129029e-01 5.48527598e-01
5.09104393e-02 1.26774907e-01 8.26070905e-01 6.77545965e-01
5.80886364e-01 -1.79840371e-01 -2.84854949e-01 1.21173549e+00
-1.52591690e-01 -8.94810855e-01 9.23060998e-02 -1.95622087e-01
2.32794657e-01 -1.29480392e-01 -2.24143565e-01 5.13522960e-02
6.68238178e-02 -6.94850683e-01 5.72681844e-01 6.20845079e-01
8.34990263e-01 -1.12449861e+00 7.06529379e-01 4.84852135e-01
-1.68756187e+00 -5.29392660e-01 -7.64220417e-01 -3.11185837e-01
3.78888577e-01 7.94070482e-01 -4.29494143e-01 5.42997658e-01
1.00734484e+00 7.02724040e-01 -4.68429267e-01 1.30173647e+00
-4.14771378e-01 4.63559687e-01 -7.40077794e-01 -9.80333611e-02
3.83613378e-01 -5.17014027e-01 7.32830524e-01 1.12759650e+00
2.90066242e-01 1.91445455e-01 5.57529807e-01 9.63861465e-01
3.28127623e-01 -2.04458728e-01 -5.24933755e-01 3.17714632e-01
4.75035816e-01 1.51819980e+00 -6.52366817e-01 -1.99248254e-01
-2.44471043e-01 7.35646009e-01 -1.12958893e-01 2.62861669e-01
-1.18065774e+00 -6.47275925e-01 5.13249993e-01 2.83739805e-01
3.21264297e-01 -1.77506462e-01 -1.69080690e-01 -7.53444552e-01
-5.05587086e-02 -4.29538965e-01 4.49983180e-01 -1.06287813e+00
-1.09567595e+00 5.18188059e-01 -1.64760634e-01 -1.15260887e+00
5.33332229e-01 -6.12344503e-01 -7.57542551e-01 8.70245516e-01
-1.58134222e+00 -1.29215443e+00 -9.98934686e-01 8.02477956e-01
6.34741426e-01 1.64736137e-02 4.90262121e-01 5.78018367e-01
-1.06253624e+00 9.05938223e-02 -2.93098152e-01 1.36414245e-01
3.09313774e-01 -4.37954605e-01 -1.81637868e-01 1.29484081e+00
-3.19243342e-01 2.85619080e-01 1.94115102e-01 -6.80328190e-01
-1.14415073e+00 -1.35863328e+00 5.09594381e-01 2.55539536e-01
1.80672541e-01 -1.52498364e-01 -6.18538380e-01 6.76670820e-02
-4.13594097e-02 2.07854901e-02 -1.45624757e-01 -5.18845916e-01
-8.87143165e-02 -6.18710399e-01 -1.26988840e+00 6.21315897e-01
1.22043002e+00 -2.58678436e-01 -2.48689473e-01 2.04363719e-01
4.63530988e-01 -9.41283256e-02 -5.55822492e-01 4.70509619e-01
6.35966778e-01 -9.95063901e-01 1.13122880e+00 -1.44731894e-01
2.28016749e-01 -8.98378849e-01 3.92817520e-02 -9.94336188e-01
-1.02024806e+00 -1.14625759e-01 5.06635308e-02 1.23252130e+00
-8.51059109e-02 -5.65093935e-01 3.25000554e-01 4.62149717e-02
-6.18728548e-02 -6.49810910e-01 -8.39199305e-01 -6.43210053e-01
-5.07339239e-01 -2.47889310e-01 4.61882651e-01 5.36846638e-01
-2.31587544e-01 3.70117188e-01 -3.08029745e-02 4.04468626e-01
5.49885750e-01 2.70806290e-02 2.87001759e-01 -1.33889997e+00
3.36516052e-01 -6.76368833e-01 -1.74440250e-01 -9.21685755e-01
-4.66618270e-01 -8.24319422e-01 -2.57256906e-02 -1.79829109e+00
3.09102327e-01 -3.38296831e-01 -3.40786636e-01 6.19979084e-01
-3.22233051e-01 2.78468639e-01 5.66681623e-01 2.12441728e-01
-7.97818080e-02 6.74722791e-01 1.47912574e+00 -4.03654665e-01
-5.65046258e-02 1.68067083e-01 -3.25152069e-01 8.43212008e-01
7.35479832e-01 -4.60237235e-01 -2.47618511e-01 -4.47023958e-01
-3.04096133e-01 -2.06806660e-01 8.03906798e-01 -1.44909179e+00
5.08255184e-01 -2.93932825e-01 9.07895446e-01 -7.52933502e-01
2.56112397e-01 -1.30653894e+00 3.03843729e-02 1.16371548e+00
4.79669929e-01 2.07850128e-01 2.62768149e-01 2.57588297e-01
6.99065402e-02 -2.75282562e-01 1.13874090e+00 -2.30707988e-01
-8.46848607e-01 6.20594680e-01 -5.44162691e-01 -2.67529577e-01
1.30227649e+00 -3.89016509e-01 -4.53885972e-01 5.97646832e-02
-2.75076568e-01 1.81827173e-01 2.00667620e-01 3.18918169e-01
8.90612304e-01 -1.41682053e+00 -7.44209588e-01 5.06620109e-01
-2.59931594e-01 2.17224196e-01 4.49683249e-01 9.24871087e-01
-8.39300752e-01 1.82208136e-01 -3.79591495e-01 -7.24711716e-01
-1.25308454e+00 6.88060760e-01 4.45428371e-01 3.26367244e-02
-8.48901451e-01 7.40449905e-01 2.21866652e-01 7.67701864e-02
6.21731393e-04 -2.35160947e-01 -3.21422666e-01 1.11802101e-01
7.94956625e-01 9.42005515e-01 -1.81557745e-01 -4.59680229e-01
-5.23598135e-01 8.71633351e-01 2.05648601e-01 1.78960830e-01
1.35076571e+00 8.86816531e-02 -1.71464220e-01 -1.37679547e-01
5.83671153e-01 -3.65645975e-01 -1.54712021e+00 4.80707362e-02
-7.81183839e-01 -6.70260370e-01 5.89186907e-01 -8.90828311e-01
-1.41504252e+00 1.10869527e+00 1.04066288e+00 2.14505479e-01
1.63376987e+00 -5.17706096e-01 7.76531458e-01 3.18627618e-02
2.15157688e-01 -8.52235258e-01 -1.20267116e-01 6.26481295e-01
9.07184184e-01 -6.70198143e-01 -1.34846857e-02 -5.56935608e-01
-3.34039360e-01 1.29674995e+00 9.86331105e-01 -1.99711859e-01
5.88229537e-01 5.86513102e-01 -1.22154631e-01 -1.81689098e-01
-2.18268752e-01 -5.11508584e-01 -4.08225544e-02 6.95960164e-01
-2.92901784e-01 -4.50623110e-02 -1.65744781e-01 4.40242618e-01
2.54260242e-01 8.73878002e-02 -3.07759997e-02 7.80707657e-01
-9.11335409e-01 -3.86033893e-01 -5.85047066e-01 4.55261320e-01
8.58040601e-02 -2.42085591e-01 -7.45147243e-02 7.24148750e-01
7.03234792e-01 1.23888445e+00 4.13439684e-02 -6.80049956e-01
5.87700486e-01 -2.65132517e-01 4.58864182e-01 -1.16118923e-01
-6.85407579e-01 1.60108835e-01 -3.45169008e-01 -3.42593551e-01
-4.15648282e-01 -3.18667263e-01 -1.16663194e+00 -5.29691100e-01
-6.43129826e-01 -2.99639970e-01 4.42742676e-01 8.12975109e-01
2.23795190e-01 8.87375772e-01 7.74882853e-01 -8.94562483e-01
-2.37425834e-01 -9.58606482e-01 -5.82644463e-01 1.06317095e-01
-4.42872681e-02 -1.00983119e+00 -4.47027534e-01 -1.82592124e-02] | [9.234251976013184, -1.024664044380188] |
f200b630-71fb-41e0-b6c4-022b0f9b625c | hunor-a-hungarianrussian-parallel-corpus | null | null | https://aclanthology.org/L12-1106 | https://aclanthology.org/L12-1106.pdf | HunOr: A Hungarian---Russian Parallel Corpus | In this paper, we present HunOr, the first multi-domain Hungarian―Russian parallel corpus. Some of the corpus texts have been manually aligned and split into sentences, besides, named entities also have been annotated while the other parts are automatically aligned at the sentence level and they are POS-tagged as well. The corpus contains texts from the domains literature, official language use and science, however, we would like to add texts from the news domain to the corpus. In the future, we are planning to carry out a syntactic annotation of the HunOr corpus, which will further enhance the usability of the corpus in various NLP fields such as transfer-based machine translation or cross lingual information retrieval. | ["Istv{\\'a}n Nagy T.", "Martina Katalin Szab{\\'o}", 'Veronika Vincze'] | 2012-05-01 | null | null | null | lrec-2012-5 | ['cross-lingual-information-retrieval'] | ['natural-language-processing'] | [-1.98922023e-01 3.66531521e-01 -3.20674390e-01 -4.18019086e-01
-6.05007470e-01 -7.25631177e-01 6.00540459e-01 4.93454397e-01
-7.44504809e-01 1.27176285e+00 6.08430982e-01 -5.52221179e-01
1.66960619e-02 -7.15657055e-01 -3.02659392e-01 -1.31292492e-01
4.94464755e-01 1.08438158e+00 4.08165812e-01 -6.08831286e-01
2.68235177e-01 1.56991571e-01 -6.59413815e-01 3.64928991e-01
1.09833527e+00 1.35374352e-01 5.28400719e-01 5.38928472e-02
-4.77043420e-01 5.96060157e-01 -5.07751763e-01 -1.01742995e+00
1.04857214e-01 -5.99884272e-01 -1.53517067e+00 -5.29122092e-02
-2.20085140e-02 1.16694868e-01 -1.43952265e-01 1.15936506e+00
4.62005287e-01 -1.26031071e-01 3.59518379e-01 -6.80398762e-01
-5.60738385e-01 1.14252174e+00 -3.72018158e-01 2.98135281e-01
4.92121518e-01 -3.51608336e-01 9.94612277e-01 -7.44933784e-01
1.45958579e+00 1.07000268e+00 2.62116224e-01 3.80055755e-01
-6.35376036e-01 -5.98096907e-01 -1.46981105e-01 5.13061523e-01
-1.19762743e+00 -3.25373590e-01 7.10756481e-01 -4.30560350e-01
1.27517211e+00 -1.26874641e-01 3.52833986e-01 1.12169743e+00
1.61181286e-01 6.74421728e-01 1.01687980e+00 -1.18816507e+00
-1.70664772e-01 3.86596411e-01 2.97210157e-01 3.35387051e-01
3.18725109e-01 -3.77495825e-01 -2.06355959e-01 -2.45475285e-02
5.46373487e-01 -5.27891397e-01 -2.57674605e-01 3.70377973e-02
-1.48484051e+00 7.12591112e-01 -2.48652145e-01 1.22055304e+00
-3.87899965e-01 -5.96156359e-01 8.45403612e-01 3.96564275e-01
5.28963983e-01 6.05081201e-01 -9.65098619e-01 -1.31850407e-01
-7.08640695e-01 2.79604524e-01 1.06795871e+00 1.36033225e+00
5.76159239e-01 -3.18235934e-01 1.68550372e-01 1.07160187e+00
1.39359698e-01 4.79199320e-01 1.00190008e+00 -5.76153696e-01
1.14363265e+00 6.60280287e-01 2.33854041e-01 -5.77408552e-01
-3.19003880e-01 9.36106220e-02 -4.10856932e-01 -5.42250454e-01
4.81496990e-01 -5.12118399e-01 -7.19622076e-01 1.42875600e+00
2.92115867e-01 -4.80746835e-01 8.12417209e-01 8.27094555e-01
1.12386751e+00 7.56277978e-01 9.35200825e-02 -5.10557592e-01
1.85610294e+00 -9.53284025e-01 -1.24634087e+00 -9.02273133e-02
7.72357345e-01 -1.41056514e+00 7.36481965e-01 1.39489248e-01
-1.19694448e+00 -3.31354499e-01 -9.26654458e-01 -1.33874446e-01
-5.97354650e-01 3.44299048e-01 1.81383535e-01 4.44548398e-01
-5.39207220e-01 4.07362223e-01 -6.27529740e-01 -8.53078306e-01
-2.75104702e-01 -2.18849126e-02 -6.07822478e-01 -6.40807077e-02
-1.84015501e+00 1.38516176e+00 1.10508275e+00 -2.06758559e-01
8.39875862e-02 -4.65464741e-02 -9.77891386e-01 -1.72941610e-01
2.26152748e-01 -4.11948085e-01 1.29553640e+00 -9.54012096e-01
-1.53987265e+00 1.22372389e+00 -3.10110152e-01 -4.13830996e-01
2.24548921e-01 -1.96424961e-01 -8.87670457e-01 -2.30641570e-02
5.08965552e-01 1.23771839e-01 -6.90454915e-02 -6.57748699e-01
-8.86987507e-01 -5.02123713e-01 -3.03690750e-02 2.79398412e-01
-1.62447050e-01 8.29420626e-01 -5.71902394e-01 -8.48795593e-01
2.28574425e-02 -9.17012930e-01 -3.34533691e-01 -1.09476745e+00
-1.03550940e-03 -2.99698204e-01 5.20675242e-01 -1.27456713e+00
1.29582822e+00 -2.02313137e+00 2.45488554e-01 -4.10184674e-02
-3.15729499e-01 2.09506586e-01 4.67583090e-02 9.86870408e-01
-7.31978007e-03 1.77006364e-01 1.38195166e-02 2.26689965e-01
-3.02336931e-01 5.03030121e-01 -3.68046820e-01 3.82654876e-01
-4.42642458e-02 7.73163915e-01 -9.09845591e-01 -7.76743174e-01
-2.28842534e-02 -2.64473647e-01 3.14084291e-02 -2.43873492e-01
-3.04510575e-02 3.67339522e-01 -6.87838912e-01 3.18509936e-01
4.55150217e-01 2.65393287e-01 4.16546315e-01 1.19604759e-01
-4.77105707e-01 8.80752981e-01 -8.56028795e-01 1.78717041e+00
-7.25330412e-01 4.73344624e-01 -2.04300925e-01 -8.86505961e-01
9.23045695e-01 9.17652428e-01 4.51402426e-01 -7.43481815e-01
2.22242922e-01 4.67435002e-01 3.59521389e-01 -6.97718680e-01
1.13178349e+00 -1.22958399e-01 -4.35386121e-01 1.11668579e-01
2.71210074e-01 -1.61623672e-01 7.28253782e-01 -1.13529734e-01
6.58091724e-01 4.49380279e-01 8.01521361e-01 -4.17561352e-01
8.31556201e-01 9.23365533e-01 8.12664688e-01 3.52236368e-02
-9.97859761e-02 3.87583226e-01 2.66111434e-01 -2.80779064e-01
-1.35201120e+00 -5.18595755e-01 -6.23335361e-01 7.75362849e-01
-3.55951488e-03 -5.75053811e-01 -7.24412143e-01 -8.29597414e-01
-6.82295084e-01 1.06685770e+00 -1.13377444e-01 6.15019798e-01
-9.65898156e-01 -5.45355976e-01 7.67560661e-01 2.30927959e-01
6.69893622e-01 -1.24855351e+00 -7.12172464e-02 7.18986154e-01
-6.19012892e-01 -1.63597262e+00 -1.83071554e-01 2.20447406e-01
-5.75903714e-01 -8.85960340e-01 -7.01903939e-01 -1.13494325e+00
2.41538942e-01 -2.43357792e-01 8.02269161e-01 -4.57339227e-01
3.48275065e-01 -2.95991421e-01 -9.25071955e-01 -6.16114974e-01
-7.24701226e-01 4.00952011e-01 -3.40131298e-02 -7.12758005e-01
6.45816565e-01 -1.84705064e-01 2.41603166e-01 3.96379232e-01
-5.51593304e-01 8.51354152e-02 4.70074356e-01 7.65255332e-01
4.36605006e-01 -2.03421768e-02 5.59595704e-01 -1.40311897e+00
4.65749323e-01 -4.07673210e-01 -5.47961831e-01 2.99082577e-01
-1.27772868e-01 -5.17168827e-02 7.80699193e-01 -3.12021911e-01
-1.46559489e+00 -1.22198068e-01 -6.36101902e-01 1.61325052e-01
-4.76868719e-01 8.50527406e-01 -4.45528954e-01 3.36198032e-01
6.74867451e-01 3.23321037e-02 -5.34219682e-01 -4.55982447e-01
2.22532213e-01 1.07267118e+00 4.87770230e-01 -6.30158186e-01
6.36641145e-01 -3.76024425e-01 -3.71914148e-01 -8.67034495e-01
-6.46398008e-01 -8.35211635e-01 -8.96640301e-01 -3.57061215e-02
1.00964940e+00 -8.18908155e-01 1.36378795e-01 -5.35504781e-02
-1.64853179e+00 -1.40449926e-01 -1.22192472e-01 9.62921381e-01
-3.42814058e-01 5.41606486e-01 -9.20038879e-01 -3.56768101e-01
-5.29136896e-01 -9.20255780e-01 6.66990161e-01 1.83130771e-01
-5.81506670e-01 -1.03364372e+00 4.68290299e-01 4.64647502e-01
-1.41927928e-01 -3.09191091e-04 1.01219571e+00 -1.27293873e+00
3.00435334e-01 -5.73227592e-02 1.26387715e-01 -5.64688779e-02
2.19824910e-01 -1.17045820e-01 -4.05334622e-01 3.00933584e-03
1.10739686e-01 -2.05883130e-01 1.47386640e-01 -1.02320224e-01
3.13133180e-01 -4.21425432e-01 -2.58478463e-01 -2.52595901e-01
1.29738498e+00 7.38300979e-01 8.25999379e-01 7.84410655e-01
2.90459573e-01 9.45086896e-01 8.91689122e-01 -2.94723473e-02
5.53814471e-01 8.40998769e-01 -4.85626578e-01 3.86657752e-02
-4.15907875e-02 -2.17421427e-01 5.81534684e-01 1.51462400e+00
-2.00985193e-01 -2.70493597e-01 -1.19625640e+00 7.38298714e-01
-1.99170935e+00 -1.00897741e+00 -7.61150360e-01 1.87658906e+00
1.02700758e+00 -1.00015670e-01 -1.52785480e-01 -1.63220674e-01
8.78831267e-01 -1.92403629e-01 2.73115546e-01 -6.43127680e-01
-2.81333983e-01 3.53523672e-01 6.27369642e-01 5.60817122e-01
-1.03083670e+00 1.61056626e+00 6.06710768e+00 8.68436873e-01
-9.17380810e-01 3.50300372e-01 1.49775147e-01 5.51559091e-01
-1.85532674e-01 3.37696522e-01 -9.41893816e-01 3.45601737e-01
1.26340246e+00 -5.01877546e-01 4.29121479e-02 8.99209201e-01
3.00207108e-01 -4.76436578e-02 -6.50442004e-01 5.43764114e-01
-3.15983817e-02 -1.19830978e+00 -2.54239589e-01 1.88308973e-02
6.27484739e-01 3.77573043e-01 -7.35062361e-01 5.39958954e-01
6.01769805e-01 -4.91370112e-01 6.69949234e-01 -1.04086936e-01
7.09492147e-01 -7.63958097e-01 1.48117602e+00 6.16157353e-01
-8.70703995e-01 6.60308838e-01 -6.96461558e-01 2.01123524e-02
4.67139870e-01 4.76491094e-01 -9.92233813e-01 1.40178597e+00
5.55814266e-01 4.73382682e-01 -4.69112515e-01 7.62354851e-01
-4.92476374e-01 7.04680145e-01 -2.15324033e-02 -3.00739914e-01
2.93238461e-01 -6.36718988e-01 6.22773349e-01 1.40807772e+00
2.84743726e-01 2.28183508e-01 3.96496058e-01 2.04019040e-01
-3.67965996e-02 1.08666360e+00 -6.59587443e-01 -1.99585989e-01
1.43138528e-01 1.15065336e+00 -7.19313145e-01 -4.81466770e-01
-8.49725246e-01 1.20079720e+00 2.78032392e-01 2.09103793e-01
-4.74344492e-01 -7.46050894e-01 1.68267637e-01 1.18163697e-01
1.01994105e-01 -3.20356339e-01 -2.45859951e-01 -1.38488436e+00
2.28822976e-03 -8.16540956e-01 4.48543727e-01 -6.35543585e-01
-1.43840337e+00 9.78368700e-01 1.06356844e-01 -1.08722901e+00
-5.56254447e-01 -6.22762442e-01 -2.37971708e-01 1.05158150e+00
-1.26431370e+00 -1.06558716e+00 5.00670969e-01 4.99962419e-01
7.18838036e-01 -4.32619989e-01 1.05570936e+00 5.07701933e-01
-4.81975466e-01 2.01077953e-01 1.22141145e-01 6.82814240e-01
1.07120883e+00 -9.96436059e-01 4.17235225e-01 8.79527628e-01
3.19963455e-01 6.37561142e-01 9.55923438e-01 -9.59761620e-01
-6.93434596e-01 -9.53588128e-01 1.85193813e+00 -1.67844623e-01
1.11470020e+00 -1.34463057e-01 -8.46453011e-01 8.84620845e-01
9.85047877e-01 -6.28271461e-01 8.78599942e-01 2.40434006e-01
1.02996297e-01 3.10555547e-01 -1.27076578e+00 5.76615870e-01
7.24017680e-01 -3.28922242e-01 -1.42839634e+00 5.49492300e-01
7.65579641e-01 -7.23668635e-01 -1.11134195e+00 2.81147569e-01
1.04023844e-01 -1.50411665e-01 1.59308553e-01 -6.74162149e-01
5.21742105e-01 -3.58793706e-01 -6.26717731e-02 -1.42702723e+00
-2.21035585e-01 -3.55418742e-01 7.37861574e-01 1.75437450e+00
8.72471452e-01 -6.02284908e-01 4.27604258e-01 3.64562809e-01
-3.85795683e-01 -1.51322141e-01 -1.06284320e+00 -7.01268375e-01
2.94259012e-01 -4.67491925e-01 5.05868614e-01 1.34350228e+00
9.83501375e-01 9.65995312e-01 -2.58193254e-01 -1.00245804e-01
-1.48268968e-01 3.26431915e-02 5.12443483e-01 -9.74473953e-01
-2.47506499e-01 -1.49685472e-01 -1.96202427e-01 -7.63158083e-01
5.09830475e-01 -1.03497446e+00 1.04534820e-01 -1.66542614e+00
6.26534969e-02 -1.78260788e-01 4.65166301e-01 6.21009290e-01
3.48326587e-03 -2.20304102e-01 9.97128990e-03 2.18156755e-01
-4.75928783e-02 4.49459195e-01 1.23740125e+00 1.77474037e-01
-3.80099922e-01 -5.17172515e-02 -4.04650688e-01 7.88041294e-01
8.90274584e-01 -5.45227826e-01 1.26099020e-01 -3.51422369e-01
-1.87628597e-01 2.79669255e-01 -6.73174739e-01 -4.39619958e-01
3.63970250e-02 -5.44712067e-01 3.08061391e-01 -6.23141766e-01
-1.62533522e-01 -9.42216754e-01 1.14181206e-01 3.13900143e-01
-9.98117626e-02 5.42777777e-01 3.42908621e-01 -1.69568256e-01
-6.16554916e-01 -7.97433972e-01 6.25279725e-01 -3.52988243e-01
-7.32117116e-01 4.22813144e-04 -7.67522156e-01 2.48481646e-01
1.10096502e+00 5.89552000e-02 -4.24935073e-01 4.38334160e-02
-4.38398123e-01 1.55115411e-01 4.72177267e-01 3.78683388e-01
1.22168615e-01 -1.20147800e+00 -8.14515054e-01 -3.34872633e-01
2.23705113e-01 -2.62387872e-01 -6.13308623e-02 5.97954273e-01
-8.11693728e-01 7.35343993e-01 -5.67067087e-01 3.54433730e-02
-1.17919087e+00 7.21459508e-01 -2.70126790e-01 -7.13973165e-01
-6.63429201e-01 -1.14140280e-01 -2.29266673e-01 -7.75909245e-01
-3.42617959e-01 5.57958782e-02 -7.78542221e-01 1.07428901e-01
2.74147868e-01 6.99644536e-02 1.83198005e-01 -1.25733888e+00
-3.55687678e-01 2.01281592e-01 -2.34181970e-01 -6.50113106e-01
1.23837578e+00 -3.21012974e-01 -4.16511327e-01 4.60792065e-01
8.01025808e-01 5.87210476e-01 1.22494273e-01 -2.06470266e-01
6.74713790e-01 -3.97606157e-02 -4.03738767e-01 -7.83896506e-01
-4.86577630e-01 3.73132735e-01 -5.29907309e-02 -8.11014771e-02
8.94463897e-01 1.37291625e-01 8.95393193e-01 5.17536581e-01
6.31808162e-01 -1.40961659e+00 -6.69559240e-01 1.22433448e+00
7.81011522e-01 -8.58260512e-01 -1.60651252e-01 -6.90128505e-01
-1.01678157e+00 1.23624957e+00 4.91924852e-01 1.11618459e-01
3.99574429e-01 4.53011036e-01 3.76032710e-01 5.96656129e-02
-3.73636574e-01 -5.24526238e-01 2.18277141e-01 5.02368033e-01
9.66031730e-01 -9.48123857e-02 -1.58507693e+00 8.39976668e-01
-6.72635198e-01 -2.90679606e-03 7.20647037e-01 5.87941110e-01
-1.27700105e-01 -1.96412170e+00 -3.24877769e-01 -1.66408084e-02
-1.02241695e+00 -3.34771305e-01 -5.47655523e-01 1.12193644e+00
1.90500066e-01 8.40127289e-01 -8.16522241e-02 -3.47954109e-02
5.23424268e-01 2.98610866e-01 5.51434755e-01 -7.89914370e-01
-8.12148631e-01 3.49681348e-01 8.73433828e-01 4.48505729e-02
-6.00161135e-01 -8.33396971e-01 -1.41863871e+00 -2.29636624e-01
-3.42919737e-01 7.43395209e-01 6.82176292e-01 1.32923985e+00
-1.69838786e-01 1.48403481e-01 8.14322159e-02 -3.23799133e-01
-3.99149880e-02 -1.34325647e+00 -4.28597927e-01 7.65906051e-02
-6.04224682e-01 -3.31097662e-01 2.74591893e-01 1.29131600e-01] | [10.42579174041748, 10.108892440795898] |
5c5071cb-4c39-4e2c-9fbc-d2aa1aa18bc9 | mvp-robust-multi-view-practice-for-driving | 2207.02042 | null | https://arxiv.org/abs/2207.02042v1 | https://arxiv.org/pdf/2207.02042v1.pdf | MVP: Robust Multi-View Practice for Driving Action Localization | Distracted driving causes thousands of deaths per year, and how to apply deep-learning methods to prevent these tragedies has become a crucial problem. In Track3 of the 6th AI City Challenge, researchers provide a high-quality video dataset with densely action annotations. Due to the small data scale and unclear action boundary, the dataset presents a unique challenge to precisely localize all the different actions and classify their categories. In this paper, we make good use of the multi-view synchronization among videos, and conduct robust Multi-View Practice (MVP) for driving action localization. To avoid overfitting, we fine-tune SlowFast with Kinetics-700 pre-training as the feature extractor. Then the features of different views are passed to ActionFormer to generate candidate action proposals. For precisely localizing all the actions, we design elaborate post-processing, including model voting, threshold filtering and duplication removal. The results show that our MVP is robust for driving action localization, which achieves 28.49% F1-score in the Track3 test set. | ['Yangguang Li', 'Haisheng Su', 'Kaibin Tian', 'Kunchang Li', 'Jingjie Shang'] | 2022-07-05 | null | null | null | null | ['action-localization'] | ['computer-vision'] | [-2.52831578e-02 -4.33002412e-01 -4.81732696e-01 -3.15596402e-01
-1.06517375e+00 -4.55591857e-01 5.60521603e-01 -5.73333442e-01
-5.31363845e-01 5.67005157e-01 6.57343268e-01 1.78477317e-01
1.95362553e-01 -5.00640810e-01 -7.25132406e-01 -8.21240366e-01
1.22186812e-02 8.28183964e-02 7.59060383e-01 -1.05881961e-02
3.13489884e-01 4.06292558e-01 -1.74705684e+00 5.96020103e-01
7.07395494e-01 8.51161003e-01 1.29414210e-02 7.14491248e-01
2.76398838e-01 1.05773222e+00 -5.66294670e-01 -1.56288326e-01
3.11618418e-01 -3.58245492e-01 -6.02773845e-01 4.04329970e-02
8.82168055e-01 -8.20941091e-01 -8.13990951e-01 8.22678268e-01
6.61805272e-01 2.18925789e-01 3.85498255e-01 -1.61473703e+00
-1.26898391e-02 2.18567193e-01 -5.96590817e-01 6.11668766e-01
3.23554665e-01 5.35208821e-01 6.97288930e-01 -9.89024162e-01
5.58674395e-01 1.19968247e+00 5.29375553e-01 7.59046555e-01
-6.80545270e-01 -8.81440639e-01 3.01932395e-01 8.47586811e-01
-1.42986321e+00 -8.68728042e-01 4.54471171e-01 -5.26351690e-01
1.03316391e+00 1.29897341e-01 8.48143518e-01 1.44195640e+00
4.36587870e-01 1.12143528e+00 6.94270849e-01 2.37196594e-01
1.21673726e-01 -5.30328929e-01 -1.64686903e-01 5.96595347e-01
2.61099879e-02 2.71965023e-02 -8.01987588e-01 1.67770699e-01
5.40845037e-01 2.85462111e-01 -4.83199954e-02 -1.35042757e-01
-1.51928329e+00 5.02922118e-01 3.67793232e-01 1.76656249e-04
-3.48553121e-01 5.78962743e-01 6.36219978e-01 -1.87226713e-01
3.44267190e-01 8.99969861e-02 -4.27958101e-01 -6.08325720e-01
-7.60348320e-01 3.57129008e-01 -3.50279622e-02 7.71641195e-01
5.17369986e-01 1.86240040e-02 -4.84914809e-01 5.85851789e-01
-1.72410160e-02 7.19084680e-01 2.85770446e-01 -1.50004816e+00
7.57656455e-01 6.75522923e-01 2.41276808e-02 -8.51716995e-01
-4.33408439e-01 6.31997734e-02 -6.87028170e-01 3.84333469e-02
3.91464084e-01 -2.22582087e-01 -1.04341805e+00 1.43775773e+00
4.20682907e-01 6.19028509e-01 -7.60439485e-02 9.21532452e-01
8.32408667e-01 4.53404486e-01 2.19843626e-01 8.73466283e-02
1.18317533e+00 -1.09051824e+00 -6.53074801e-01 -4.99227405e-01
9.15830433e-01 -3.18659604e-01 8.53058457e-01 3.12777698e-01
-7.67615736e-01 -7.55331516e-01 -8.75062585e-01 -1.23305671e-01
-1.30634934e-01 3.16605508e-01 4.56401676e-01 8.31726938e-02
-7.41231084e-01 2.69635051e-01 -1.14102983e+00 -3.51380199e-01
8.96842420e-01 8.83350894e-02 -6.46129310e-01 -4.81365055e-01
-1.21533787e+00 7.73885131e-01 1.63413107e-01 1.56338826e-01
-1.43609250e+00 -5.51250100e-01 -9.06189978e-01 -2.51904041e-01
5.09972870e-01 -5.14760911e-01 9.71782863e-01 -7.17033148e-01
-9.99281704e-01 7.13411152e-01 -4.13200974e-01 -3.72189820e-01
7.39335299e-01 -4.64402884e-01 -5.75108290e-01 1.41902849e-01
6.14266753e-01 9.78635013e-01 7.37927794e-01 -6.72828138e-01
-1.13516402e+00 -3.86561960e-01 8.87935981e-02 2.37613603e-01
-1.35676160e-01 1.13942489e-01 -8.55839074e-01 -2.86184430e-01
-2.53725052e-03 -1.04873085e+00 -2.52533168e-01 -7.11741149e-02
-2.32309029e-01 -4.35842007e-01 8.52021396e-01 -6.12680852e-01
1.32541931e+00 -2.41109681e+00 6.80122450e-02 -2.81187505e-01
3.41353863e-01 1.34427235e-01 -9.31782424e-02 1.93661362e-01
1.71334699e-01 -2.24699080e-01 7.69224018e-02 -1.91477776e-01
-2.81527042e-01 3.30189556e-01 -6.12174869e-02 8.28056991e-01
1.42887101e-01 9.23782825e-01 -1.02635550e+00 -5.85937440e-01
3.06060761e-01 2.96331197e-01 -5.64651906e-01 3.84745672e-02
2.59782039e-02 5.07100403e-01 -5.97582519e-01 8.18782926e-01
4.95921701e-01 -2.80343276e-03 -2.86456943e-01 -1.40048996e-01
-1.73926026e-01 1.59084871e-01 -1.21551931e+00 1.77326512e+00
-6.52219355e-02 8.50542247e-01 -1.30258992e-01 -8.67669344e-01
5.05268395e-01 1.14429921e-01 8.26547801e-01 -8.27967286e-01
2.41133459e-02 2.92499959e-02 -5.20193577e-02 -1.08287597e+00
3.50203484e-01 3.24123979e-01 -3.05809617e-01 6.87448829e-02
-1.61719322e-01 3.25650722e-01 4.30410415e-01 2.83059090e-01
1.49556911e+00 3.20287973e-01 3.35232280e-02 6.24665357e-02
6.69891000e-01 1.74051926e-01 1.08162332e+00 4.94019181e-01
-7.32675016e-01 8.16322982e-01 4.57818031e-01 -8.33073795e-01
-5.88777661e-01 -9.09123957e-01 1.89078838e-01 1.07030785e+00
2.33830154e-01 -4.48658019e-01 -7.78526127e-01 -9.55980301e-01
-1.02623692e-02 5.38340986e-01 -7.14972496e-01 -4.75595236e-01
-8.97537351e-01 -5.05325258e-01 6.43963218e-01 9.21530068e-01
8.73362780e-01 -1.03444660e+00 -7.37568438e-01 5.66490702e-02
-6.88740492e-01 -1.31562531e+00 -6.28411412e-01 -1.66790068e-01
-5.68881273e-01 -1.36858952e+00 -5.56847513e-01 -4.43616122e-01
5.01878440e-01 6.69998944e-01 9.03887808e-01 -7.31833465e-03
-2.34775886e-01 1.51426986e-01 -2.86226839e-01 -2.43235633e-01
-1.22577153e-01 -3.45529094e-02 2.27414668e-01 1.91780016e-01
7.78245091e-01 -2.12718129e-01 -8.57725918e-01 6.08659208e-01
-3.71155679e-01 -4.89581795e-03 5.58585942e-01 3.60606283e-01
6.86335444e-01 -8.27825889e-02 4.88231421e-01 -2.64500648e-01
-7.48653263e-02 -4.76167202e-01 -5.19414067e-01 -3.25775705e-02
1.20967835e-01 -3.14606696e-01 3.28085989e-01 -2.00824574e-01
-8.17422867e-01 4.94860470e-01 -1.50153935e-01 -5.87495446e-01
-4.23128843e-01 -1.27276599e-01 -4.51335251e-01 3.49865854e-01
5.77980340e-01 3.23506266e-01 -1.33197442e-01 -8.85964334e-02
2.63691843e-01 6.04491830e-01 6.12907946e-01 1.81084741e-02
5.74036658e-01 7.66467988e-01 -1.75915062e-01 -6.96411610e-01
-1.10035694e+00 -6.06202424e-01 -8.14599752e-01 -7.24857748e-01
1.49347603e+00 -1.52161157e+00 -6.18077755e-01 6.11456096e-01
-1.09809852e+00 -3.65320683e-01 1.33226067e-01 7.53601372e-01
-5.91004252e-01 4.72945087e-02 -2.89827913e-01 -4.16363299e-01
1.48910088e-02 -1.33739114e+00 1.25576651e+00 1.11134626e-01
-1.09787397e-01 -3.79115045e-01 1.83810726e-01 8.91747892e-01
1.37185976e-01 1.36336133e-01 1.38174385e-01 -3.18305284e-01
-7.58091569e-01 -1.84598327e-01 -1.45910963e-01 3.33060235e-01
4.38799933e-02 8.96841288e-02 -1.07225204e+00 -1.30581558e-01
-4.22639579e-01 -2.64742613e-01 1.29451096e+00 7.00848818e-01
1.30098164e+00 2.23135948e-01 -7.07110941e-01 7.12945461e-01
8.46438646e-01 2.00695485e-01 9.41970766e-01 4.28940952e-01
9.82037306e-01 4.08509612e-01 9.07077372e-01 3.81012589e-01
6.63625300e-01 7.22455382e-01 5.29265761e-01 1.18982203e-01
-1.77814707e-01 -1.77737430e-01 8.12982619e-01 4.17114854e-01
-2.16091141e-01 -2.30192155e-01 -8.99610102e-01 7.48641908e-01
-2.01826787e+00 -1.55939853e+00 -2.69593060e-01 1.92952812e+00
2.46641099e-01 4.01989400e-01 3.98770183e-01 -5.66900112e-02
5.60987175e-01 3.42240512e-01 -7.00717509e-01 1.68338150e-01
5.78702800e-02 -5.15610516e-01 7.62502432e-01 2.09552750e-01
-1.64347231e+00 9.86658216e-01 6.06136322e+00 8.34834874e-01
-9.98509347e-01 7.10528195e-02 6.16456091e-01 -5.09374261e-01
1.53622046e-01 -2.48462334e-01 -1.10470116e+00 5.78159630e-01
8.29508245e-01 1.72988072e-01 1.98707774e-01 8.99078369e-01
7.64040709e-01 -3.64239007e-01 -9.18625951e-01 1.09582639e+00
1.88934803e-01 -1.30868208e+00 -3.04225087e-01 4.72850576e-02
7.93145895e-01 5.91457248e-01 -3.32074583e-01 5.73919713e-01
2.46474907e-01 -9.38748240e-01 7.23779798e-01 5.94127297e-01
5.19720972e-01 -8.56516540e-01 5.39692342e-01 4.48312521e-01
-1.38045490e+00 -4.44911718e-01 -3.06737989e-01 -1.37204230e-01
3.22407365e-01 4.41617966e-01 -3.49123925e-01 2.87819117e-01
9.65763152e-01 1.33298767e+00 -6.05071723e-01 9.97392297e-01
-2.51308203e-01 6.57912731e-01 -1.64421663e-01 2.76758105e-01
2.05999225e-01 1.22880325e-01 3.35304767e-01 9.25974667e-01
3.30131292e-01 1.01473838e-01 4.01863426e-01 3.13952267e-01
2.98969075e-02 -4.38538641e-01 -7.56802022e-01 2.05723286e-01
4.22968596e-01 1.27097476e+00 -6.58690989e-01 -4.59807724e-01
-6.03769481e-01 9.60585952e-01 1.17215104e-01 1.33412987e-01
-1.40935373e+00 -8.17632973e-02 1.14475858e+00 2.56088495e-01
2.99177527e-01 -2.90409118e-01 -2.07955539e-02 -1.26959753e+00
1.39193013e-01 -8.36962402e-01 3.92929435e-01 -6.48680389e-01
-7.81385243e-01 2.65530169e-01 1.26668081e-01 -1.68423927e+00
-1.56439066e-01 -2.01280251e-01 -6.48031414e-01 2.59817034e-01
-1.20220721e+00 -1.07021677e+00 -5.23009360e-01 5.85973740e-01
9.47153211e-01 -1.34359151e-01 2.14072153e-01 7.61397839e-01
-9.71120596e-01 4.00484174e-01 -1.18116863e-01 3.26052785e-01
9.67187345e-01 -8.06833923e-01 5.33173084e-01 9.61130977e-01
4.18399163e-02 -2.73843762e-02 3.66684020e-01 -6.36409163e-01
-1.42986012e+00 -1.56555474e+00 8.62167776e-01 -9.15253282e-01
4.93900597e-01 -4.43281978e-01 -6.86289072e-01 8.08722079e-01
-6.55015483e-02 4.15345579e-01 2.95172036e-01 -3.73525500e-01
-2.33519487e-02 -3.50116819e-01 -6.63347960e-01 4.93774354e-01
1.46813357e+00 -2.30156124e-01 -3.67994040e-01 4.16224390e-01
5.24176896e-01 -3.45377922e-01 -4.82686996e-01 3.42550933e-01
5.14349461e-01 -1.08137500e+00 9.63201046e-01 -6.61165893e-01
6.46973908e-01 -5.93892157e-01 -1.40060291e-01 -1.06878197e+00
-4.55052108e-01 -2.88920850e-01 3.27201076e-02 1.12142229e+00
2.59839833e-01 -2.38817647e-01 8.68110359e-01 3.42170119e-01
-5.61336696e-01 -7.13816941e-01 -1.21367371e+00 -5.52597046e-01
-2.65231997e-01 -6.45558059e-01 2.68277854e-01 6.07472718e-01
-2.49772459e-01 1.51646495e-01 -6.39051020e-01 8.39941353e-02
4.02251899e-01 -7.47590140e-02 1.14241791e+00 -9.15902615e-01
-3.57978567e-02 -2.18912989e-01 -6.19615018e-01 -1.16942298e+00
2.00043082e-01 -6.02896452e-01 2.73335725e-01 -1.73357177e+00
2.74400085e-01 -4.42763492e-02 -2.88177818e-01 6.26111209e-01
-2.48742163e-01 3.67590904e-01 -3.37686501e-02 8.78360495e-02
-1.10070741e+00 6.77330256e-01 1.27433622e+00 -1.75035730e-01
-4.42474969e-02 -2.22181529e-02 -3.44074965e-01 9.50065434e-01
7.32880473e-01 -5.75366199e-01 -3.45593870e-01 -6.13526046e-01
-5.32282628e-02 -8.75246599e-02 5.84482551e-01 -1.38722277e+00
2.07333058e-01 -6.08724773e-01 6.65665329e-01 -9.72111404e-01
5.03878117e-01 -6.57874525e-01 -1.33765399e-01 5.72009623e-01
-1.85958579e-01 8.59629959e-02 1.13202706e-01 6.74128890e-01
-1.72378078e-01 4.12421465e-01 6.95667684e-01 -7.52919689e-02
-1.35624456e+00 4.11686331e-01 -6.64117992e-01 3.61132890e-01
1.45358253e+00 -2.43243963e-01 -5.60473204e-01 -2.56751776e-01
-6.13599896e-01 6.96953118e-01 2.85899520e-01 7.78283417e-01
5.78757226e-01 -1.51236916e+00 -8.75705004e-01 2.16876820e-01
3.36443990e-01 3.11662722e-02 6.99760258e-01 1.23579693e+00
-3.15339208e-01 2.62268186e-01 -2.47982740e-01 -7.50672519e-01
-1.33435833e+00 3.55632871e-01 4.32347804e-01 1.05335161e-01
-8.56037915e-01 8.75127792e-01 2.54567564e-01 -1.88572586e-01
2.03634888e-01 -1.62209556e-01 -3.14335942e-01 1.31919071e-01
8.49858284e-01 6.89605415e-01 -1.27768710e-01 -9.87985253e-01
-8.56373429e-01 6.00386500e-01 9.08245817e-02 1.53046519e-01
1.23464215e+00 -6.47360384e-02 2.91266114e-01 9.53081921e-02
1.20293176e+00 -2.11221740e-01 -1.82184947e+00 2.53737003e-01
-2.19934270e-01 -6.31158113e-01 6.08507656e-02 -4.16302234e-01
-1.36921608e+00 9.20170963e-01 6.17007971e-01 -3.91395390e-01
9.15246844e-01 1.44129485e-01 9.73880827e-01 2.98335910e-01
2.43114114e-01 -1.35743070e+00 1.84186235e-01 6.92599893e-01
7.78523982e-01 -1.37415504e+00 1.34205818e-02 -9.71396714e-02
-7.63027310e-01 9.21774089e-01 1.21265447e+00 -1.27475023e-01
3.94109935e-01 2.02677459e-01 2.42843285e-01 -2.54966050e-01
-8.73879313e-01 -1.99246630e-01 1.06013432e-01 5.14260530e-01
4.08044197e-02 -4.90243733e-02 -8.60135928e-02 2.82901049e-01
1.67125046e-01 1.63803160e-01 4.49432939e-01 8.88188422e-01
-6.90185487e-01 -6.14900947e-01 -2.44563431e-01 5.15901089e-01
-2.95696884e-01 3.40019047e-01 -3.64291281e-01 6.13573670e-01
5.18907249e-01 9.73370373e-01 2.20324188e-01 -7.94494331e-01
5.82163870e-01 -1.87757332e-02 1.19319618e-01 -2.40910903e-01
-2.43437082e-01 -1.45377335e-03 2.02431753e-01 -1.20329785e+00
-5.27777016e-01 -9.09955740e-01 -1.48636019e+00 -2.48274937e-01
2.77534463e-02 -3.92543107e-01 3.07971835e-01 1.13094532e+00
7.20294476e-01 6.10107839e-01 5.86743295e-01 -9.19123948e-01
-3.65309894e-01 -7.78232813e-01 -1.70702234e-01 3.97468269e-01
3.58217776e-01 -9.43558991e-01 -3.78701806e-01 1.03383176e-01] | [8.200764656066895, 0.35050448775291443] |
f8aa58e2-c502-443f-8b45-14515a772691 | natural-language-processing-of-clinical-notes | 1908.05780 | null | https://arxiv.org/abs/1908.05780v1 | https://arxiv.org/pdf/1908.05780v1.pdf | Natural Language Processing of Clinical Notes on Chronic Diseases: Systematic Review | Of the 2652 articles considered, 106 met the inclusion criteria. Review of the included papers resulted in identification of 43 chronic diseases, which were then further classified into 10 disease categories using ICD-10. The majority of studies focused on diseases of the circulatory system (n=38) while endocrine and metabolic diseases were fewest (n=14). This was due to the structure of clinical records related to metabolic diseases, which typically contain much more structured data, compared with medical records for diseases of the circulatory system, which focus more on unstructured data and consequently have seen a stronger focus of NLP. The review has shown that there is a significant increase in the use of machine learning methods compared to rule-based approaches; however, deep learning methods remain emergent (n=3). Consequently, the majority of works focus on classification of disease phenotype with only a handful of papers addressing extraction of comorbidities from the free text or integration of clinical notes with structured data. There is a notable use of relatively simple methods, such as shallow classifiers (or combination with rule-based methods), due to the interpretability of predictions, which still represents a significant issue for more complex methods. Finally, scarcity of publicly available data may also have contributed to insufficient development of more advanced methods, such as extraction of word embeddings from clinical notes. Further efforts are still required to improve (1) progression of clinical NLP methods from extraction toward understanding; (2) recognition of relations among entities rather than entities in isolation; (3) temporal extraction to understand past, current, and future clinical events; (4) exploitation of alternative sources of clinical knowledge; and (5) availability of large-scale, de-identified clinical corpora. | ['Joel T. Dudley', 'Venet Osmani', 'Alberto Lavelli', 'Fabio Rinaldi', 'Seyedmostafa Sheikhalishahi', 'Riccardo Miotto'] | 2019-08-15 | null | null | null | null | ['clinical-knowledge'] | ['miscellaneous'] | [-5.09817479e-03 2.16723368e-01 -6.35469675e-01 -1.39984041e-01
-6.37940109e-01 -5.11355042e-01 2.92342752e-01 1.03539252e+00
-5.29580772e-01 8.25559199e-01 7.27965474e-01 -5.51534772e-01
-4.45110559e-01 -6.77567661e-01 -7.80600533e-02 -5.70571780e-01
-3.34062368e-01 6.16655886e-01 -2.26433486e-01 1.79538906e-01
-8.91356170e-03 4.32687163e-01 -1.21816897e+00 2.56471962e-01
8.61063302e-01 7.49020636e-01 -8.03378131e-03 4.29308474e-01
-3.41941506e-01 5.88634431e-01 -4.30034071e-01 -3.72859538e-01
-1.43533275e-01 -5.81573904e-01 -6.93922520e-01 -1.85822159e-01
-4.71287817e-02 -1.54600188e-01 -1.40182748e-01 6.46000445e-01
7.51688480e-01 -2.28274852e-01 6.60892189e-01 -7.91084826e-01
-1.00183845e+00 3.80938679e-01 1.81707996e-03 1.86127380e-01
4.04369980e-01 1.21702090e-01 1.05064869e+00 -6.72371268e-01
1.02044463e+00 7.62327313e-01 6.63482130e-01 5.85125923e-01
-1.08982348e+00 -5.81723690e-01 7.73720094e-05 5.56019917e-02
-1.25015140e+00 -2.32431963e-01 2.30877951e-01 -6.77919805e-01
1.30905128e+00 2.42510438e-01 1.01891911e+00 1.08551383e+00
3.09774548e-01 1.41295731e-01 1.05744743e+00 -3.64332408e-01
1.60796419e-01 3.95454079e-01 1.50252551e-01 5.00843108e-01
7.23162651e-01 6.03542663e-02 -1.55550048e-01 -7.12771237e-01
3.88669521e-01 1.90308630e-01 -3.13520938e-01 -1.10890128e-01
-1.23582506e+00 8.70906055e-01 7.14915916e-02 5.33789933e-01
-5.79065502e-01 -6.04265630e-01 7.71569490e-01 1.43461585e-01
5.09117007e-01 6.14820778e-01 -8.50240886e-01 -1.52786314e-01
-1.07182193e+00 1.86455637e-01 1.02790070e+00 5.59690237e-01
4.46541570e-02 -7.19977468e-02 2.13653162e-01 9.26450133e-01
4.53347921e-01 2.91095763e-01 6.64543331e-01 -3.71580094e-01
3.71915251e-01 8.46430302e-01 -1.71674564e-01 -1.08887184e+00
-7.95312047e-01 -4.46814418e-01 -8.60486627e-01 -2.11885631e-01
3.83669019e-01 -3.53448212e-01 -8.16740811e-01 1.54312742e+00
1.86864972e-01 -3.32954973e-01 2.90716141e-01 6.96188509e-01
9.91342008e-01 1.66748628e-01 4.23265845e-01 -3.52884829e-01
1.69015586e+00 -4.38076258e-01 -8.45632553e-01 1.57110877e-02
9.16136444e-01 -6.17962778e-01 3.95913661e-01 4.17616248e-01
-9.02717471e-01 -4.11297269e-02 -8.05912435e-01 -5.67741022e-02
-1.03976178e+00 1.72022194e-01 7.25625277e-01 7.08233178e-01
-7.06764162e-01 4.18160647e-01 -9.60230529e-01 -8.87791276e-01
7.88594007e-01 4.40190047e-01 -5.53946853e-01 1.62876137e-02
-1.42362905e+00 1.36086583e+00 3.33740950e-01 -2.97583286e-02
-2.22024068e-01 -1.19347119e+00 -9.22719419e-01 -1.27570018e-01
2.36765936e-01 -1.06677210e+00 4.87136185e-01 -6.36873126e-01
-9.15162385e-01 1.17137873e+00 -9.15668756e-02 -4.03237075e-01
1.39723763e-01 -6.55688420e-02 -7.32139945e-01 1.62821084e-01
-5.39993495e-02 4.05785322e-01 -1.05847724e-01 -6.70371234e-01
-6.72911644e-01 -6.01827323e-01 -2.08987564e-01 3.61921787e-02
-4.27002132e-01 1.96657658e-01 -3.60339247e-02 -6.91966355e-01
-1.22110829e-01 -8.94046962e-01 -4.10820723e-01 9.64052379e-02
-3.56837839e-01 -2.89961427e-01 3.26680630e-01 -9.49531257e-01
1.68511152e+00 -1.87130380e+00 8.43370706e-02 -2.01925874e-01
5.65046072e-01 4.71383005e-01 3.12401026e-01 9.24820423e-01
-4.17660534e-01 6.36953056e-01 -4.81863096e-02 4.58327755e-02
-3.36681962e-01 2.56130278e-01 -1.30548840e-02 5.18671691e-01
5.11835694e-01 7.81872571e-01 -9.31596756e-01 -5.56884408e-01
3.91518503e-01 6.76515996e-01 -5.22020757e-01 -1.70765385e-01
8.19529369e-02 1.16355143e-01 -5.28827071e-01 8.28414619e-01
3.90701771e-01 -4.01127726e-01 3.92481804e-01 -5.35586551e-02
-2.92244673e-01 7.64826834e-01 -9.40732896e-01 1.37001419e+00
-3.06175351e-01 3.86983693e-01 -1.16885841e-01 -1.06131482e+00
5.92874110e-01 7.98706114e-01 9.73329067e-01 -2.10730508e-01
5.16002588e-02 3.28243732e-01 4.20535326e-01 -7.83791780e-01
-9.02003348e-02 -2.95436561e-01 9.01505575e-02 3.30794416e-02
-2.33126789e-01 3.07824194e-01 1.43687665e-01 2.93914527e-02
1.38308215e+00 4.02093045e-02 8.66434932e-01 -1.60152972e-01
3.50395441e-01 4.45747256e-01 9.36763525e-01 4.04516339e-01
-3.19315821e-01 6.08192027e-01 5.71335256e-01 -4.84987378e-01
-7.88833320e-01 -8.77783656e-01 -8.75398755e-01 2.15096146e-01
-5.67661941e-01 -6.77416921e-01 -1.08570009e-01 -4.44209546e-01
2.73462057e-01 4.24719691e-01 -6.68778479e-01 -3.99640240e-02
-3.56876761e-01 -1.31159043e+00 8.50094140e-01 4.46177244e-01
3.84959625e-04 -8.42247009e-01 -7.67206430e-01 5.61676860e-01
1.28079318e-02 -1.05961382e+00 1.22647636e-01 3.88593405e-01
-1.04850376e+00 -1.42589772e+00 -7.37457871e-01 -7.65931606e-01
5.41716993e-01 -5.77421427e-01 8.94398570e-01 6.80037588e-02
-5.90601563e-01 1.69419423e-01 -3.66386294e-01 -7.09954858e-01
-4.57282305e-01 1.67061254e-01 1.85386747e-01 -4.72077042e-01
8.77019107e-01 -4.36354250e-01 -7.11902499e-01 -3.18848401e-01
-8.72077405e-01 -2.07349956e-01 9.44276392e-01 9.01926398e-01
4.11333621e-01 -8.29662159e-02 9.24095809e-01 -1.12045789e+00
5.74163795e-01 -8.82055819e-01 -2.27673445e-02 5.36816344e-02
-1.14140785e+00 -3.41785908e-01 5.05340934e-01 -5.05998492e-01
-7.00234771e-01 -2.87237376e-01 -2.03736454e-01 4.20501940e-02
-6.31034136e-01 9.75833952e-01 2.84459203e-01 4.96595919e-01
5.01362681e-01 4.76885736e-02 2.36064970e-01 -4.86570597e-01
-1.56261042e-01 9.17633474e-01 -1.13951266e-01 -2.15950966e-01
3.57603073e-01 3.41734618e-01 -4.89207841e-02 -8.67051899e-01
-5.59529364e-01 -5.14741242e-01 -5.28728247e-01 4.88047808e-01
1.12090957e+00 -8.73597503e-01 -4.37151611e-01 6.17751703e-02
-8.19351315e-01 1.84100792e-02 -1.65426597e-01 9.33522522e-01
1.00275807e-01 2.07316130e-01 -7.31489301e-01 -6.20428622e-01
-4.50928003e-01 -1.07323313e+00 8.22931170e-01 -4.76263016e-02
-9.98660922e-01 -1.31914127e+00 4.08329278e-01 3.07003975e-01
2.94360995e-01 5.85373163e-01 1.67581415e+00 -1.14324331e+00
6.35404438e-02 -4.17423069e-01 -1.96298942e-01 1.50231749e-01
8.30369473e-01 1.22775137e-01 -6.08126938e-01 -1.50533644e-02
-2.23606095e-01 -6.12206385e-02 4.74660248e-01 4.30841416e-01
6.99004948e-01 -2.54416525e-01 -7.16326475e-01 3.72773588e-01
1.46220386e+00 6.24632359e-01 5.57815135e-01 2.72026777e-01
4.71176147e-01 8.74139667e-01 2.28994623e-01 3.51243734e-01
5.17073095e-01 5.72433710e-01 7.71453083e-02 -1.95358261e-01
-1.71800986e-01 1.91496521e-01 1.04229704e-01 9.15916979e-01
-1.08228877e-01 -1.02464139e-01 -1.13547862e+00 8.81437182e-01
-1.37039340e+00 -6.79264128e-01 -3.00206244e-01 2.05076146e+00
1.08866501e+00 4.58199764e-03 2.35376861e-02 -8.13038740e-03
4.05372322e-01 -9.62182060e-02 -3.95183206e-01 -5.91621876e-01
1.22056976e-02 4.22082752e-01 3.72597218e-01 -7.18576238e-02
-8.33109915e-01 3.54161918e-01 6.75728035e+00 2.45798811e-01
-1.24355304e+00 -6.54105321e-02 3.81113559e-01 -1.68190539e-01
-3.37777138e-02 3.24343815e-02 -8.04112732e-01 7.85561621e-01
1.12827373e+00 1.10296356e-02 -1.58952594e-01 3.89785558e-01
4.61378276e-01 1.75342374e-02 -1.10761440e+00 6.76505089e-01
-1.51243657e-01 -1.32555866e+00 -2.03876644e-01 4.26068395e-01
4.70668107e-01 1.85752958e-01 -2.05309615e-01 2.15570107e-01
-2.88420413e-02 -1.01735270e+00 5.83872460e-02 4.19951588e-01
9.19281542e-01 -1.38198048e-01 9.67545569e-01 -1.01724178e-01
-7.96247125e-01 7.96482190e-02 -2.12178946e-01 -4.05560695e-02
2.35588580e-01 9.22580540e-01 -9.46385145e-01 9.07339275e-01
6.75716937e-01 9.50296521e-01 -3.94677103e-01 1.04445624e+00
3.26562598e-02 9.18267727e-01 -2.71191746e-01 -4.29673195e-02
8.84388536e-02 -2.07199097e-01 4.28456277e-01 1.35009301e+00
9.97318178e-02 1.56916425e-01 1.05986133e-01 7.61616468e-01
1.12337060e-01 4.71316665e-01 -8.07266474e-01 -6.65913820e-01
3.48323911e-01 1.26655960e+00 -7.22465456e-01 -2.00113639e-01
-1.00436258e+00 4.82302636e-01 1.03397733e-02 2.43291304e-01
-5.86765647e-01 -3.23143572e-01 7.80910790e-01 3.93677205e-01
1.22532278e-01 7.24218637e-02 -3.42911780e-01 -9.87320900e-01
-5.64906783e-02 -1.10366046e+00 8.97144198e-01 -2.69173861e-01
-1.36240745e+00 4.88894880e-01 -1.75356507e-01 -1.02561319e+00
-1.73966512e-01 -6.27846301e-01 -9.45308730e-02 9.21789169e-01
-1.27561879e+00 -8.18678021e-01 2.13435784e-01 -9.47120599e-03
2.69149899e-01 -1.25245869e-01 1.50039101e+00 5.72205007e-01
-7.12023079e-01 3.45577121e-01 1.09974936e-01 2.44707152e-01
1.06039238e+00 -1.15565968e+00 -2.86113381e-01 1.61994830e-01
-3.07812780e-01 9.14653718e-01 3.29202533e-01 -9.98151243e-01
-1.13049459e+00 -9.79966819e-01 1.55501056e+00 -6.85580075e-01
7.62158751e-01 -1.43029720e-01 -9.23664331e-01 5.40159523e-01
1.36076838e-01 -1.90637156e-01 1.41613209e+00 3.48286927e-01
-1.03502221e-01 1.52727917e-01 -1.23605514e+00 6.62513733e-01
6.96127713e-01 -3.72059464e-01 -8.17711890e-01 2.36524969e-01
4.25196022e-01 -2.34544158e-01 -1.46811378e+00 4.63659137e-01
7.95826077e-01 -3.06292772e-01 8.86844516e-01 -1.00388575e+00
6.85268164e-01 4.23736833e-02 1.34943083e-01 -1.06974900e+00
-3.00745696e-01 -2.22070098e-01 -1.70238074e-02 1.18437517e+00
6.97666466e-01 -9.40695226e-01 6.52479470e-01 8.42542052e-01
3.18461703e-03 -1.38049304e+00 -6.62500739e-01 -3.14704746e-01
2.09085554e-01 -9.88712013e-02 2.19739661e-01 1.39272940e+00
2.34616622e-01 1.18688613e-01 4.96181771e-02 1.75336227e-02
-8.86143371e-02 -2.72497926e-02 2.38934420e-02 -1.40135360e+00
-7.27215633e-02 -4.11429048e-01 -5.68157613e-01 -1.75660059e-01
-8.09548199e-02 -1.04584551e+00 -5.20250022e-01 -2.09651661e+00
3.07437152e-01 -5.78339696e-01 -4.17956412e-01 5.85425973e-01
-1.42047942e-01 -1.94097124e-02 -3.68510991e-01 1.28945008e-01
4.44016680e-02 6.32979125e-02 9.14727807e-01 -2.59525049e-03
-4.55517173e-01 -2.00515702e-01 -1.14734268e+00 8.07877541e-01
7.13989317e-01 -8.64818096e-01 -2.96440005e-01 -1.38854235e-01
3.90041530e-01 3.82909901e-04 1.26271218e-01 -5.61295509e-01
8.42940956e-02 -1.10887952e-01 4.40681517e-01 -4.61716622e-01
2.02924386e-03 -8.61094952e-01 1.96104631e-01 8.10220301e-01
-3.18740815e-01 2.74642766e-01 3.92753214e-01 4.49813396e-01
-3.29634160e-01 -2.26819292e-01 2.97562450e-01 -1.28317684e-01
-2.71315038e-01 5.58102913e-02 -8.35283637e-01 2.22727597e-01
8.64952803e-01 -3.01189035e-01 -1.97146431e-01 -4.27525342e-02
-1.01613927e+00 1.88797459e-01 2.50201464e-01 3.29214752e-01
3.21832150e-01 -1.08537841e+00 -7.73549497e-01 -2.36601189e-01
5.59216589e-02 -4.21605147e-02 1.53047651e-01 1.50751317e+00
-5.66782534e-01 9.03321087e-01 1.03177510e-01 -1.79864109e-01
-1.29583848e+00 7.49539495e-01 1.42750815e-01 -5.90311944e-01
-9.60405529e-01 2.60584176e-01 1.00081034e-01 -3.38905215e-01
3.07287216e-01 -3.98501277e-01 -3.88660640e-01 5.58681250e-01
3.50897044e-01 3.95980358e-01 2.48145685e-01 -4.39288944e-01
-9.94877040e-01 3.44763488e-01 -2.80338764e-01 2.86789834e-01
1.66247487e+00 1.83186099e-01 -2.56356686e-01 4.13883060e-01
1.16270554e+00 1.59216106e-01 -1.96216136e-01 9.84838381e-02
1.03728399e-01 1.77848279e-01 -5.52615561e-02 -1.08606029e+00
-8.67065191e-01 6.16817117e-01 5.70991576e-01 1.50806494e-02
1.07646394e+00 -7.49388114e-02 6.36392951e-01 -1.36796152e-03
6.51547611e-02 -9.78765428e-01 -6.22488081e-01 9.02894512e-02
5.44057488e-01 -1.05969906e+00 2.44545266e-01 -4.15895760e-01
-5.18025279e-01 1.15163052e+00 1.01507634e-01 1.39430240e-01
9.57034945e-01 3.67805451e-01 8.25668573e-02 -4.52562600e-01
-9.22031760e-01 -1.74345002e-01 2.34779671e-01 5.42752385e-01
9.47263658e-01 -5.37934899e-02 -9.34712529e-01 1.11862087e+00
6.10036217e-02 3.11539501e-01 4.11193401e-01 9.45897400e-01
2.75579482e-01 -1.47543824e+00 -6.86259866e-02 1.05546582e+00
-1.14881384e+00 -4.20892417e-01 -5.42224228e-01 1.08583856e+00
6.55810118e-01 9.15244281e-01 -5.99764995e-02 6.38408735e-02
2.77085453e-01 5.39712191e-01 2.86562979e-01 -1.05205095e+00
-8.28311265e-01 2.45945137e-02 4.94939446e-01 -2.55756736e-01
-6.26265109e-01 -1.05955470e+00 -1.21869326e+00 1.09225251e-01
-3.04083884e-01 6.96751550e-02 4.43093419e-01 1.03086710e+00
6.74596488e-01 7.63623595e-01 -1.35934383e-01 1.66431427e-01
-3.06270778e-01 -8.14621568e-01 -5.02659917e-01 2.49442324e-01
1.17777526e-01 -4.88261044e-01 -3.58460188e-01 6.50119931e-02] | [8.403791427612305, 8.549577713012695] |
3acc3379-381f-4a39-bf35-8957bc3e051c | depth-map-generation-using-pixel-matching-in | 1902.03471 | null | https://arxiv.org/abs/1902.03471v3 | https://arxiv.org/pdf/1902.03471v3.pdf | Depth-Map Generation using Pixel Matching in Stereoscopic Pair of Images | Modern day multimedia content generation and dissemination is moving towards the presentation of more and more `realistic' scenarios. The switch from 2-dimensional (2D) to 3-dimensional (3D) has been a major driving force in that direction. Over the recent past, a large number of approaches have been proposed for creating 3D images/videos most of which are based on the generation of depth-maps. This paper presents a new algorithm for obtaining depth information pertaining to a depicted scene from a set of available pair of stereoscopic images. The proposed algorithm performs a pixel-to-pixel matching of the two images in the stereo pair for estimation of depth. It is shown that the obtained depth-maps show improvements over the reported counterparts. | ['Mohd. Samar Ansari', 'Asra Aslam'] | 2019-02-09 | null | null | null | null | ['3d-depth-estimation'] | ['computer-vision'] | [ 6.66876256e-01 -1.40202120e-01 1.12996943e-01 -4.44040358e-01
-8.11172605e-01 -4.16860580e-01 7.61550784e-01 1.80223510e-01
-3.38129371e-01 6.90541565e-01 2.80923307e-01 1.87508501e-02
-1.65614970e-02 -8.91207337e-01 -3.52882862e-01 -7.27356136e-01
-1.30745918e-01 1.68716431e-01 8.03995073e-01 -1.48706481e-01
7.09731162e-01 6.89462900e-01 -2.09810138e+00 4.79824156e-01
4.71284986e-01 1.17988575e+00 4.54484254e-01 1.06705701e+00
-2.94140011e-01 1.85175225e-01 -3.40894759e-01 -3.46298784e-01
2.90389270e-01 -4.52229619e-01 -4.77606267e-01 3.91118705e-01
5.28550684e-01 -5.99994481e-01 -1.72098085e-01 9.72776413e-01
6.32498920e-01 -1.34852692e-01 4.94628549e-01 -1.18519795e+00
1.53557688e-01 -2.93592632e-01 -8.26903701e-01 1.97504833e-01
9.25286710e-01 -2.30216950e-01 3.69654924e-01 -5.65141678e-01
8.86756718e-01 1.27967942e+00 1.40915930e-01 2.92796880e-01
-1.05326891e+00 -2.21785635e-01 -3.47352266e-01 4.06513751e-01
-1.32520425e+00 -3.65274310e-01 1.14801049e+00 -4.64704096e-01
7.44653940e-01 1.91099733e-01 6.67858481e-01 5.96991122e-01
2.22887829e-01 5.01757264e-01 1.34696233e+00 -9.36019003e-01
2.50488222e-01 3.93256724e-01 -4.01747167e-01 3.09751987e-01
2.68462986e-01 2.81528294e-01 -5.78596890e-01 4.24494334e-02
1.08134687e+00 -3.43686044e-01 -3.58837664e-01 -7.50923336e-01
-9.71314073e-01 3.96503121e-01 4.60907491e-03 4.25248355e-01
-3.91368985e-01 -1.20832384e-01 2.95889854e-01 6.46224394e-02
4.60921228e-01 1.10280596e-01 1.05462879e-01 -4.57578301e-01
-1.01426792e+00 2.92754412e-01 5.13532996e-01 9.08832908e-01
7.57838249e-01 -2.23611042e-01 5.80120921e-01 5.76315939e-01
2.66350776e-01 3.53857636e-01 3.38072687e-01 -1.14105809e+00
6.08431101e-01 3.89140606e-01 3.54576766e-01 -1.25244451e+00
-2.20430538e-01 1.59391522e-01 -5.52148998e-01 6.83127582e-01
4.34201270e-01 -3.47692780e-02 -5.44732630e-01 1.01371503e+00
6.49945199e-01 -1.90421849e-01 2.32561231e-01 9.56121683e-01
6.95365071e-01 8.04335356e-01 -3.41448098e-01 -2.01500490e-01
8.33799303e-01 -4.02448237e-01 -6.96121335e-01 -6.24141321e-02
1.24374211e-01 -1.07464767e+00 4.52454627e-01 8.05597246e-01
-1.56076097e+00 -5.77757001e-01 -1.18956220e+00 -5.59209287e-02
-2.50175714e-01 -3.99363786e-01 4.48649637e-02 8.21382225e-01
-1.21626854e+00 3.64012837e-01 -2.69274324e-01 -4.60091799e-01
-1.35546066e-02 1.17560029e-01 -6.12044871e-01 -2.04249144e-01
-9.89281178e-01 6.54151738e-01 4.13369447e-01 -1.46826968e-01
-4.71873492e-01 -2.23872706e-01 -6.17430747e-01 -2.34087050e-01
1.63739741e-01 -4.54526693e-01 1.08098233e+00 -9.16591406e-01
-1.43052626e+00 1.32408047e+00 -3.45289230e-01 -1.75032213e-01
7.90553212e-01 -4.60353419e-02 -3.44657481e-01 6.77478135e-01
-1.38976589e-01 6.84387684e-01 6.92215562e-01 -1.70492959e+00
-1.14043140e+00 -5.04319549e-01 2.52346605e-01 4.98865992e-01
-4.87142839e-02 2.88761277e-02 -6.12109125e-01 -3.06559533e-01
5.47520041e-01 -4.31225151e-01 5.44897430e-02 4.87349868e-01
-2.29549259e-01 3.78318965e-01 9.65041995e-01 -4.80764419e-01
9.26298022e-01 -2.00631785e+00 1.29960269e-01 4.78922166e-02
1.25560933e-03 1.73025891e-01 2.00649217e-01 7.49934077e-01
-7.71360621e-02 -1.94604635e-01 6.49419501e-02 -4.31411147e-01
-3.56239766e-01 7.54587678e-03 -9.44246277e-02 5.63473940e-01
-3.52226466e-01 1.49957851e-01 -8.29901576e-01 -7.99624860e-01
6.24198079e-01 6.80815518e-01 -4.45125997e-01 3.07114094e-01
7.53168575e-03 7.28374958e-01 -2.79224843e-01 3.39323580e-01
1.14750886e+00 2.62481838e-01 2.10665986e-01 2.29835603e-02
-5.33034623e-01 6.66365325e-02 -1.32972157e+00 1.63016868e+00
-3.96856636e-01 8.49747717e-01 1.16310224e-01 -5.94933391e-01
9.78184164e-01 5.45835257e-01 6.55585587e-01 -1.11416674e+00
9.82694477e-02 3.75728697e-01 -2.85713434e-01 -7.44633138e-01
6.82599425e-01 -2.90573895e-01 8.54871199e-02 2.89592057e-01
-3.96654308e-01 -5.26210189e-01 2.31932729e-01 -1.22595958e-01
5.75656712e-01 1.98539615e-01 4.09405649e-01 7.52875730e-02
9.48323727e-01 1.41230496e-02 1.60110906e-01 3.20215195e-01
-3.33980769e-01 9.14396763e-01 4.49322939e-01 -2.80421823e-01
-1.42823565e+00 -1.16016591e+00 -2.41958037e-01 1.08680345e-01
5.67141950e-01 -5.91436103e-02 -7.22707987e-01 -1.60738807e-02
-2.49697253e-01 4.29583013e-01 -2.84475356e-01 4.89552617e-01
-5.52923203e-01 -1.51917085e-01 6.54148161e-02 1.25448123e-01
9.60943997e-01 -5.29226065e-01 -1.20466542e+00 1.92895845e-01
-2.35109344e-01 -1.25041878e+00 1.06275696e-02 -2.31589809e-01
-1.21144402e+00 -1.04594600e+00 -1.13506448e+00 -7.34807849e-01
6.00730062e-01 7.97913969e-01 8.37432861e-01 -3.38026285e-01
-3.11215460e-01 5.15133500e-01 -5.24397075e-01 -2.80970097e-01
-4.51381654e-01 -5.05291462e-01 -2.14202657e-01 1.24321945e-01
6.16250001e-02 -8.26148212e-01 -9.76009369e-01 4.10006732e-01
-1.16844010e+00 5.04413188e-01 3.23186457e-01 1.58188239e-01
5.47455549e-01 3.98567945e-01 1.79160342e-01 -6.30539358e-01
2.26705343e-01 -2.68639028e-01 -8.26426387e-01 -1.98931053e-01
-2.19174370e-01 -2.62393624e-01 3.76506239e-01 9.96705741e-02
-1.37696540e+00 8.96356627e-02 -2.39729702e-01 -9.96059738e-03
-5.96556604e-01 4.06563692e-02 -2.90677816e-01 -8.85540470e-02
5.77096641e-01 2.66580254e-01 -9.72877517e-02 -4.46373940e-01
1.93108141e-01 9.96441960e-01 5.93446314e-01 -1.68242097e-01
7.83553779e-01 1.00573099e+00 3.31457973e-01 -1.10979724e+00
-2.42717370e-01 -7.05586493e-01 -9.41062093e-01 -7.51860857e-01
8.98119211e-01 -7.38624096e-01 -3.23034614e-01 6.60042822e-01
-1.29306626e+00 3.10758874e-03 -7.31442682e-03 4.85591561e-01
-8.37382913e-01 6.56391263e-01 -1.31790280e-01 -1.03009355e+00
7.74970353e-02 -1.12802482e+00 1.12945437e+00 3.51702958e-01
-5.04974090e-02 -1.14421809e+00 2.72345781e-01 2.81591982e-01
3.77097726e-01 5.04259169e-01 9.89344001e-01 3.12951714e-01
-7.66829133e-01 -5.00405967e-01 -3.58996779e-01 2.26826698e-01
3.22909892e-01 -8.29885900e-02 -1.21051788e+00 2.08695363e-02
1.93502560e-01 1.20897405e-01 7.22296163e-02 4.58529353e-01
6.70127213e-01 3.15925479e-02 -3.59890223e-01 3.66843939e-01
1.96551812e+00 6.48327231e-01 1.14523947e+00 6.48715794e-01
2.09644899e-01 8.69539976e-01 8.44005942e-01 5.72302222e-01
2.46066257e-01 1.12297618e+00 5.81376553e-01 -3.84487323e-02
-1.20897584e-01 -2.60884732e-01 2.29535010e-02 4.22116667e-01
-1.84642941e-01 -3.34720194e-01 -8.21042776e-01 5.96217453e-01
-1.37465668e+00 -1.11390424e+00 -3.54839236e-01 2.62926126e+00
3.43895406e-01 9.44884568e-02 1.06918171e-01 6.76974773e-01
9.49280262e-01 1.55243397e-01 -1.57310426e-01 -6.29691601e-01
-1.54353142e-01 -1.37823626e-01 3.43973428e-01 6.46145046e-01
-8.30635250e-01 3.09351891e-01 6.55836010e+00 4.16746497e-01
-1.32216656e+00 -2.08967596e-01 4.07425910e-01 1.27573922e-01
-4.52934116e-01 -1.26644224e-01 -5.60502589e-01 5.02023935e-01
7.77691841e-01 -5.56314349e-01 -2.10192934e-01 5.96462429e-01
6.98137105e-01 -1.13486612e+00 -8.31848502e-01 1.45488405e+00
1.87190861e-01 -1.18066311e+00 1.99193582e-01 2.23549873e-01
9.76238549e-01 -5.58698356e-01 9.96313170e-02 -6.73710227e-01
-4.25208598e-01 -4.84018683e-01 6.74472988e-01 4.69550163e-01
9.08476353e-01 -8.86040866e-01 4.76598173e-01 4.98984218e-01
-1.12641549e+00 2.09841818e-01 -1.05019495e-01 4.22032587e-02
6.56577229e-01 6.30309582e-01 -8.73898745e-01 6.43939674e-01
7.97691166e-01 4.77671742e-01 -3.25681090e-01 1.60741091e+00
8.81680846e-02 -1.38086677e-01 -7.14319721e-02 2.29609206e-01
2.13122115e-01 -3.18721533e-01 6.70453608e-01 9.33489442e-01
6.65362418e-01 -1.52266575e-02 -5.53644121e-01 3.40774059e-01
4.73682642e-01 5.81381284e-02 -9.66800034e-01 3.96799088e-01
1.29249781e-01 8.14861655e-01 -7.43164122e-01 -4.19549167e-01
-5.73776364e-01 1.19795012e+00 -4.05394524e-01 1.55554488e-01
-6.85930014e-01 -5.25928855e-01 3.73818815e-01 6.61324203e-01
3.13295066e-01 -3.72518837e-01 -1.81533039e-01 -7.18856812e-01
9.12037417e-02 -4.79242295e-01 1.83443710e-01 -1.20475292e+00
-7.94303358e-01 6.41797423e-01 4.66113150e-01 -1.64407718e+00
-5.56875050e-01 -5.39852202e-01 -3.42039585e-01 1.06701195e+00
-1.70365250e+00 -3.60162705e-01 -6.56125128e-01 4.96398628e-01
6.20802939e-01 2.37283111e-01 6.56988680e-01 5.67762375e-01
1.94843620e-01 -7.08585829e-02 5.18334210e-01 -3.97089571e-01
6.13473833e-01 -1.00044680e+00 2.71001667e-01 7.33809948e-01
-3.37485909e-01 1.43051585e-02 1.02390909e+00 -4.39164668e-01
-1.27680409e+00 -4.17749882e-01 9.72459018e-01 -6.07906803e-02
4.46878970e-02 -1.81538418e-01 -6.19634271e-01 1.39935747e-01
2.41458416e-01 -4.70903963e-01 3.60803783e-01 -8.12180758e-01
8.69466737e-02 -2.29495630e-01 -1.47709584e+00 4.99507129e-01
8.32457066e-01 -5.14577985e-01 -2.68283367e-01 -8.05669427e-02
9.17865720e-04 -4.57421064e-01 -7.01106250e-01 1.44957528e-01
6.66676521e-01 -1.90493631e+00 9.54194307e-01 2.90089607e-01
5.75666845e-01 -3.76765817e-01 -2.63891011e-01 -1.06465971e+00
3.91845793e-01 -6.41028702e-01 3.63384098e-01 8.62972736e-01
-6.07434250e-02 -4.37382877e-01 1.18206012e+00 4.63412255e-01
7.62395486e-02 -4.63837326e-01 -1.18487835e+00 -5.41785717e-01
-2.10822150e-01 -4.94050533e-01 2.98055530e-01 3.66179675e-01
8.75358060e-02 -2.78126486e-02 -1.39236555e-01 1.07110210e-01
7.44005620e-01 1.42350003e-01 9.44284379e-01 -1.27638376e+00
-1.17951550e-01 -3.31655145e-01 -1.10510767e+00 -1.39680803e+00
-6.04817688e-01 -3.65561843e-01 -2.00484246e-01 -1.82266176e+00
1.10153176e-01 -2.08005637e-01 4.80745822e-01 -7.01223016e-01
4.25799489e-01 3.61159205e-01 2.81089637e-02 8.39785933e-02
-2.68555045e-01 2.77939051e-01 1.39433658e+00 4.54320490e-01
-2.57312059e-02 7.79914334e-02 -2.07403675e-01 7.26907074e-01
7.14093745e-01 -1.08266138e-01 -4.93479848e-01 -4.64882612e-01
9.61304754e-02 4.78363842e-01 2.42119923e-01 -1.26486683e+00
1.83739766e-01 -3.63485850e-02 3.12797219e-01 -9.78819788e-01
9.05052543e-01 -1.01389968e+00 3.78079861e-01 2.52930760e-01
4.52419631e-02 2.03498736e-01 1.64980620e-01 6.06067240e-01
-5.79354227e-01 -4.29206401e-01 9.28794026e-01 -3.97325307e-01
-1.00776291e+00 -1.24254748e-01 -3.87033284e-01 -2.79474139e-01
1.25923800e+00 -1.01543295e+00 -1.64669469e-01 -6.78226471e-01
-5.41253209e-01 -3.14599156e-01 8.52103949e-01 1.78427860e-01
1.05507779e+00 -1.29284036e+00 -3.04095238e-01 1.30528063e-01
1.47206895e-03 6.07679365e-03 5.86760044e-01 5.86408317e-01
-1.16886389e+00 8.60977471e-01 -5.30385375e-01 -7.36432135e-01
-1.63702762e+00 2.97671854e-01 4.73167926e-01 -1.46759991e-02
-6.64609730e-01 3.33408535e-01 3.82732511e-01 3.28578562e-01
2.79329479e-01 2.06121922e-01 -3.30237418e-01 -6.16753101e-02
7.37592578e-01 6.66167080e-01 -3.18933874e-02 -1.03746653e+00
-7.02720359e-02 1.07126951e+00 1.56550810e-01 -7.63850510e-01
1.26547384e+00 -7.10535407e-01 1.65820792e-01 4.97409135e-01
1.42757428e+00 -3.72863151e-02 -1.45727837e+00 -2.60676127e-02
-1.33444041e-01 -1.20788157e+00 1.66241556e-01 -4.28232670e-01
-8.21339786e-01 1.09920537e+00 8.39044690e-01 2.74345875e-01
1.36340702e+00 -1.55930474e-01 6.83939040e-01 -2.23921716e-01
8.73283565e-01 -1.14410055e+00 2.72038244e-02 1.00623563e-01
7.74668574e-01 -1.11890864e+00 7.70595148e-02 -7.12780535e-01
-2.97500849e-01 1.38446236e+00 2.98668534e-01 1.82370439e-01
5.00928283e-01 3.01987659e-02 2.34660998e-01 -1.09513544e-01
-4.07860070e-01 -2.53745019e-01 5.20724244e-02 9.72052872e-01
4.62096721e-01 -4.54453915e-01 -2.84063280e-01 -4.56993669e-01
1.71303168e-01 3.62333134e-02 9.58800137e-01 1.19347560e+00
-6.59125447e-01 -1.32149041e+00 -8.06491077e-01 -2.20578685e-02
-1.91518441e-01 3.06494921e-01 -1.26607627e-01 9.14976180e-01
7.01365024e-02 1.03953183e+00 1.30964413e-01 4.28706296e-02
5.81666172e-01 -1.27420872e-01 8.30117524e-01 -4.03140038e-01
7.09628314e-02 1.72588065e-01 6.22838698e-02 -5.04277766e-01
-8.41812611e-01 -8.14468801e-01 -1.09170306e+00 -3.91811907e-01
1.26202554e-01 3.08071692e-02 1.09478676e+00 3.80326122e-01
1.59346566e-01 -3.36123668e-02 8.84062529e-01 -1.32678604e+00
2.11375937e-01 -3.46342742e-01 -8.68281603e-01 5.25353909e-01
3.16360623e-01 -5.44254184e-01 -3.15374583e-01 2.22010732e-01] | [9.219204902648926, -2.536027431488037] |
6c0c9002-1959-4086-915d-5c5ab702066f | parkinsons-disease-emg-data-augmentation-and | null | null | https://doi.org/10.3390/s20092605 | https://doi.org/10.3390/s20092605 | Parkinson’s Disease EMG Data Augmentation and Simulation with DCGANs and Style Transfer | This paper proposes two new data augmentation approaches based on Deep Convolutional Generative Adversarial Networks (DCGANs) and Style Transfer for augmenting Parkinson’s Disease (PD) electromyography (EMG) signals. The experimental results indicate that the proposed models can adapt to different frequencies and amplitudes of tremor, simulating each patient’s tremor patterns and extending them to different sets of movement protocols. Therefore, one could use these models for extending the existing patient dataset and generating tremor simulations for validating treatment approaches on different movement scenarios. | ['Esther Luna Colombini', 'Rafael Anicet Zanini'] | 2020-05-03 | null | null | null | null | ['electromyography-emg'] | ['medical'] | [ 2.78045624e-01 3.88631135e-01 -1.77574709e-01 -2.41512526e-02
-5.88639855e-01 -1.75143331e-01 3.33866626e-01 -1.12401271e+00
-3.28023165e-01 1.02417278e+00 5.54337680e-01 -1.95778102e-01
-8.63714069e-02 -5.98433375e-01 -4.55363542e-01 -4.98677462e-01
-5.08566022e-01 6.93968415e-01 -6.19749464e-02 -3.33674371e-01
-2.85599351e-01 4.68992442e-01 -9.06160355e-01 4.10628825e-01
7.74166822e-01 3.21583152e-01 2.40179688e-01 7.36404777e-01
5.25390506e-01 4.56671834e-01 -1.02904177e+00 2.01586142e-01
3.92335296e-01 -6.66970789e-01 -7.38897085e-01 1.52700290e-01
-1.33703694e-01 -8.24799955e-01 -8.87789965e-01 8.59446883e-01
9.48924005e-01 -1.26201794e-01 4.29143399e-01 -1.23303306e+00
-8.44195306e-01 8.49339545e-01 -1.50434643e-01 1.71333849e-01
4.05381709e-01 7.82861173e-01 1.85727894e-01 1.50427431e-01
7.97209203e-01 1.10075581e+00 1.00015187e+00 1.33298552e+00
-1.08241153e+00 -6.63937211e-01 -4.52514529e-01 2.28863150e-01
-4.34999019e-01 2.83229321e-01 8.26644599e-01 -2.55995929e-01
6.06158674e-01 2.43787661e-01 1.11031961e+00 2.26449180e+00
5.72600245e-01 8.54804575e-01 1.21790385e+00 1.24195449e-01
2.24984974e-01 -7.23580599e-01 -5.25536537e-02 1.16589196e-01
1.90807715e-01 4.57226098e-01 -9.43027064e-02 -2.55150586e-01
1.22371173e+00 -4.22490314e-02 -5.88132620e-01 2.40942851e-01
-1.56868041e+00 7.40359843e-01 5.02495468e-01 4.08610791e-01
-8.47353280e-01 5.70417345e-01 6.52293444e-01 5.25845647e-01
-1.50866397e-02 3.90305817e-01 -3.51641655e-01 -4.73823100e-01
-6.07880771e-01 7.93459058e-01 5.86840928e-01 9.84345794e-01
-3.37152511e-01 3.68884265e-01 -5.95786750e-01 4.68129516e-01
2.49726698e-01 6.35553241e-01 1.34055591e+00 -8.88574660e-01
5.51860213e-01 5.30620039e-01 -5.53659201e-02 -2.22751290e-01
-9.22885001e-01 -2.32827082e-01 -1.01514959e+00 4.04864192e-01
2.88406014e-01 -8.11071992e-01 -1.38337922e+00 1.85925555e+00
-1.15974657e-01 -3.79563868e-02 1.09644942e-01 9.18972313e-01
6.35675728e-01 -2.79179439e-02 2.53016263e-01 2.23053932e-01
1.22490811e+00 -5.71599066e-01 -1.13505316e+00 -4.06888038e-01
6.66052401e-01 -2.07141742e-01 1.29254889e+00 4.08826053e-01
-1.00131392e+00 -4.32914317e-01 -8.13452184e-01 2.79218972e-01
-3.94698307e-02 3.18351746e-01 8.16528440e-01 6.24502778e-01
-7.05297351e-01 1.10263407e+00 -1.46016574e+00 -3.33913863e-01
7.52990425e-01 7.43489325e-01 -2.35186800e-01 2.48374864e-01
-1.61126828e+00 8.99993896e-01 3.16597879e-01 1.51164919e-01
-9.50680614e-01 -5.37701905e-01 -4.23669666e-01 -4.61191535e-01
-4.89282578e-01 -1.13181412e+00 1.09324026e+00 -5.50533712e-01
-1.57075739e+00 5.20509660e-01 7.27564812e-01 -8.99029493e-01
9.23627794e-01 -4.90805119e-01 -5.27859449e-01 -2.33877432e-02
-3.21913362e-01 5.42908609e-01 6.68036520e-01 -6.37058854e-01
1.42667815e-01 -5.26305437e-01 -8.63441918e-03 3.60145532e-02
-7.27315769e-02 -1.15748696e-01 1.66026130e-01 -1.27052903e+00
-5.22878133e-02 -1.36060286e+00 -2.80108452e-01 2.50343680e-02
-9.06720221e-01 2.89370656e-01 1.21782374e+00 -9.51394498e-01
7.31384695e-01 -1.73104358e+00 7.13264465e-01 -8.12802389e-02
1.51991798e-02 4.72356915e-01 -2.97142327e-01 4.70348865e-01
-6.55567572e-02 -1.87037624e-02 -4.40944225e-01 -9.56262648e-02
5.97123355e-02 6.90671623e-01 9.05864239e-02 3.02480280e-01
2.35582128e-01 1.19774187e+00 -7.19618917e-01 1.75520316e-01
2.77082711e-01 5.08778632e-01 -6.81792557e-01 2.70885289e-01
-2.99696863e-01 1.37578058e+00 -5.89434266e-01 6.13727450e-01
3.08038414e-01 2.02077612e-01 2.22835645e-01 -4.00904596e-01
6.80459261e-01 1.77123565e-02 -7.66197085e-01 1.76376510e+00
-1.92035854e-01 4.12268281e-01 -9.10774097e-02 -7.45889544e-01
6.81249917e-01 4.98400152e-01 8.68747592e-01 -3.09429109e-01
6.51559889e-01 2.06797630e-01 6.60024226e-01 -8.82270932e-01
-3.25131476e-01 -1.84943825e-01 -1.45947844e-01 4.24723893e-01
7.86091536e-02 -1.76873788e-01 1.81901157e-02 -4.77887601e-01
1.68450022e+00 3.52364182e-01 -3.20078701e-01 6.01145960e-02
2.04263240e-01 2.41064103e-04 3.34455431e-01 3.44341785e-01
-4.80997562e-01 6.19176149e-01 2.54376292e-01 -2.11273998e-01
-1.02768600e+00 -1.19340777e+00 1.64595425e-01 2.15856805e-01
-3.89363647e-01 1.45064369e-01 -1.08264017e+00 -6.40383422e-01
5.72783276e-02 5.96338868e-01 -9.27671015e-01 -6.74853086e-01
-8.92641366e-01 -1.03789353e+00 1.19414163e+00 1.12933218e+00
6.71625018e-01 -1.65107512e+00 -6.83817029e-01 3.40837836e-01
-2.47752458e-01 -7.18964636e-01 -3.69627744e-01 2.23783806e-01
-1.16373134e+00 -1.48135483e+00 -1.16655147e+00 -7.87972689e-01
4.61593896e-01 -7.59617865e-01 4.69709545e-01 -3.49123597e-01
-2.88346469e-01 5.65631330e-01 -5.81680238e-01 -4.56388324e-01
-1.02559626e+00 -4.61862497e-02 4.25317347e-01 -6.75228953e-01
1.89989597e-01 -1.05338752e+00 -8.52327645e-01 1.75479233e-01
-1.14099765e+00 -6.12577051e-02 8.34187508e-01 1.02023947e+00
6.19433999e-01 -2.36337259e-01 8.38911414e-01 -7.04411626e-01
1.50653243e+00 -4.23133999e-01 1.94090992e-01 -1.83731079e-01
-3.45697194e-01 -3.65674943e-02 5.18576741e-01 -1.09187150e+00
-7.60551512e-01 7.35594928e-02 -5.89132190e-01 -6.93293989e-01
-2.81376421e-01 2.06500292e-01 -1.18781537e-01 5.27427569e-02
8.01143110e-01 4.00118977e-01 3.95940572e-01 -5.71484923e-01
4.31736678e-01 6.57323241e-01 1.02552152e+00 -1.70845762e-01
7.02869415e-01 2.47973919e-01 -8.23608711e-02 -3.54246140e-01
2.03159079e-02 3.85574520e-01 -6.91494048e-01 -1.82190295e-02
9.93878603e-01 -4.88237709e-01 -4.71558064e-01 1.02155352e+00
-8.66572917e-01 -8.87262106e-01 -6.75504625e-01 9.21714902e-01
-1.30389905e+00 3.35333228e-01 -1.18108940e+00 -1.51090562e-01
-8.78732324e-01 -1.38619244e+00 1.01799750e+00 -1.54074088e-01
-6.75574243e-01 -8.43789220e-01 5.07075787e-01 1.12401359e-01
6.54724240e-01 7.78522074e-01 8.26947868e-01 -6.49163246e-01
-1.09805278e-02 -5.55551231e-01 4.98792231e-01 6.28075600e-01
4.93479073e-01 -5.21254420e-01 -4.93978500e-01 -3.20838451e-01
3.94614518e-01 -4.67592299e-01 3.55982721e-01 6.24365747e-01
1.11917460e+00 -3.64602894e-01 -2.63976872e-01 6.80788100e-01
8.68952811e-01 3.42941940e-01 1.37701499e+00 6.15325868e-01
6.59171760e-01 -9.13971961e-02 2.19532520e-01 2.65114993e-01
-2.66339809e-01 5.99620521e-01 5.78221917e-01 -1.30858719e-01
-5.94830155e-01 8.88386834e-03 3.94282758e-01 5.19181132e-01
-5.89406073e-01 -1.75714388e-01 -5.93334913e-01 3.27249855e-01
-1.57255530e+00 -9.19453681e-01 -1.19197175e-01 1.72417128e+00
9.11651850e-01 -5.92498034e-02 4.56113368e-01 3.62105429e-01
6.52774215e-01 -2.71153241e-01 -1.09256387e+00 -5.77288605e-02
3.11057474e-02 6.61425531e-01 5.71585298e-01 -2.46018156e-01
-7.51599371e-01 5.03180265e-01 7.51980877e+00 1.73565641e-01
-1.04379249e+00 1.71741158e-01 -1.15756840e-01 -1.40188128e-01
-8.44297558e-02 -6.24295592e-01 -6.17293976e-02 8.68465006e-01
1.07031691e+00 -2.87132636e-02 6.42074049e-01 4.72306043e-01
3.60924423e-01 5.36367774e-01 -1.01269150e+00 6.23941720e-01
-4.19438541e-01 -9.78043914e-01 -3.66832665e-03 2.12949529e-01
4.48818713e-01 5.40049970e-01 -5.71399508e-03 3.07350576e-01
3.74703586e-01 -9.34624612e-01 1.33515358e-01 8.11288595e-01
6.89466476e-01 -5.44226885e-01 9.19149697e-01 1.21382184e-01
-6.10419154e-01 -1.17728181e-01 1.83792129e-01 4.23064649e-01
3.64451259e-01 -2.52581090e-01 -5.41208565e-01 3.26627284e-01
4.85667169e-01 5.00797451e-01 -2.66914815e-01 8.36532235e-01
-4.54283863e-01 8.31810594e-01 -2.84319907e-01 1.89613238e-01
1.66761622e-01 1.09204821e-01 8.34027052e-01 7.16149688e-01
5.32195210e-01 -1.45638615e-01 -3.21103305e-01 1.23070550e+00
3.86234769e-03 -4.59946454e-01 -4.60818440e-01 -4.53934968e-01
2.24098131e-01 7.54274011e-01 -1.69643492e-01 1.11940093e-02
-1.06409900e-01 1.42106009e+00 -3.34861100e-01 2.07322851e-01
-1.12471604e+00 -2.46844232e-01 6.55090988e-01 -3.55822295e-02
-1.57359064e-01 -1.06316447e-01 -2.00379238e-01 -8.54008734e-01
-1.81707479e-02 -1.09292865e+00 1.29907995e-01 -9.56539571e-01
-1.47298706e+00 7.12656140e-01 -9.75132808e-02 -1.71008945e+00
-5.53048134e-01 -6.11447334e-01 -9.30907011e-01 8.62563789e-01
-7.86217928e-01 -1.31680501e+00 -3.25044125e-01 8.52271020e-01
4.05882239e-01 -3.29383343e-01 1.06394660e+00 3.55572015e-01
-3.97464335e-01 5.07134914e-01 5.08601815e-02 1.65430382e-01
5.42294800e-01 -1.19903076e+00 6.34854734e-01 2.69005656e-01
-5.40190399e-01 3.86573046e-01 8.82606149e-01 -1.02075863e+00
-1.43636620e+00 -1.43227148e+00 -1.85689330e-02 -2.47164980e-01
8.26671481e-01 1.10353634e-01 -1.05356133e+00 1.05007315e+00
2.53754914e-01 -3.29936475e-01 4.80520189e-01 -6.40317917e-01
4.82788533e-01 5.00097394e-01 -1.76669681e+00 8.49172175e-01
1.31231129e+00 2.93283742e-02 -9.52850103e-01 8.48048568e-01
4.41389292e-01 -8.75067055e-01 -1.48770058e+00 6.75744891e-01
5.12980521e-01 -2.73530215e-01 9.80604768e-01 -1.00192571e+00
5.70778668e-01 1.14024490e-01 1.19591460e-01 -1.92103314e+00
-1.76273927e-01 -6.81687117e-01 -5.23272574e-01 5.90968013e-01
-1.50543340e-02 -7.45034873e-01 8.56478930e-01 3.72022003e-01
-4.71070945e-01 -7.91779876e-01 -9.05795336e-01 -9.80036676e-01
3.73319924e-01 -3.94684255e-01 8.69351149e-01 8.04711342e-01
1.70275837e-01 -1.09253652e-01 -7.76442587e-01 -1.86472107e-02
2.92430133e-01 -5.02319217e-01 6.81838334e-01 -1.01363540e+00
-3.78309578e-01 -2.94891417e-01 -8.54278147e-01 -2.02037364e-01
-1.09885052e-01 -9.76677835e-01 -9.63778570e-02 -1.73945785e+00
-2.70029962e-01 -1.14845939e-01 -1.44319460e-01 8.03788304e-01
3.35043818e-02 3.61388266e-01 -3.96554098e-02 3.16216171e-01
4.26476300e-01 6.96929336e-01 1.80095875e+00 -3.35727751e-01
-4.65767652e-01 3.69204789e-01 -2.49216780e-01 7.53181040e-01
1.21200311e+00 -4.52801526e-01 -4.20384109e-01 -2.07010195e-01
-3.87780994e-01 -2.70673558e-02 5.53361297e-01 -1.52744949e+00
-2.69259661e-01 1.03664085e-01 5.22894025e-01 -5.17016590e-01
1.04231156e-01 -8.09756339e-01 5.52883565e-01 1.25982785e+00
-2.71096617e-01 -9.03721228e-02 6.29510999e-01 6.37796760e-01
2.40803316e-01 1.31306991e-01 6.75788760e-01 -2.48119354e-01
-2.67743498e-01 3.33580822e-01 -7.44481027e-01 -3.62055972e-02
9.68734384e-01 -1.85233995e-01 -2.56506890e-01 -2.69611150e-01
-1.58141267e+00 9.41987783e-02 2.74114192e-01 5.42133868e-01
5.41572928e-01 -1.77266872e+00 -6.40083313e-01 2.72963166e-01
-1.26834258e-01 -3.96720290e-01 3.91226023e-01 9.56544757e-01
-5.58205247e-01 2.51458615e-01 -1.07302272e+00 -3.67868692e-01
-9.96684909e-01 1.79506555e-01 6.41454697e-01 -4.73393470e-01
-1.23420572e+00 2.44477004e-01 -5.90659678e-01 -5.82082152e-01
2.66515613e-01 -6.25892043e-01 -1.25459477e-01 -7.02867806e-01
2.40317196e-01 5.27757525e-01 -2.52574086e-02 -4.03454632e-01
-5.36429286e-02 2.15652794e-01 2.39327803e-01 1.45736948e-01
1.48955321e+00 3.83050948e-01 4.36879307e-01 3.31056342e-02
4.63467956e-01 -4.38003033e-01 -1.20352316e+00 2.64697045e-01
-3.70726287e-01 1.76790923e-01 -7.87280500e-02 -1.30487657e+00
-1.53868425e+00 4.16350991e-01 1.57508612e+00 -1.95746168e-01
1.36442685e+00 -1.57586411e-01 1.41982758e+00 4.74455416e-01
5.66499174e-01 -8.95747542e-01 -6.50228932e-02 2.85654031e-02
1.42699838e+00 -5.88769078e-01 -2.37424523e-01 3.83086456e-03
-6.83397174e-01 1.06277740e+00 4.26887155e-01 -6.64949298e-01
4.33017612e-01 5.30613661e-01 2.71613479e-01 -3.91777009e-01
-1.78100079e-01 -8.84659439e-02 2.54822224e-01 1.28666484e+00
2.63120532e-01 2.17349440e-01 -8.92285883e-01 9.49331224e-01
-2.15309694e-01 6.27746582e-01 5.09259582e-01 1.08001459e+00
8.99088383e-02 -1.54085159e+00 -3.18243027e-01 9.06362534e-01
-4.36921865e-01 2.59467065e-01 -6.31627619e-01 1.44169533e+00
2.17898190e-01 3.21809918e-01 -1.48222849e-01 -7.89981782e-01
7.02382445e-01 2.51812637e-01 9.28311944e-01 -5.76954484e-01
-8.15177858e-01 -5.70514984e-02 -6.20922493e-03 -6.76063955e-01
-6.31898463e-01 -7.20688879e-01 -1.63572812e+00 8.46939087e-02
-8.91512707e-02 -5.35880744e-01 4.52726662e-01 7.54107714e-01
3.89455289e-01 1.32761133e+00 1.83312789e-01 -1.07668042e+00
-9.60697591e-01 -1.83805084e+00 -8.22547197e-01 8.59177232e-01
-6.72873184e-02 -8.61689508e-01 -2.80540049e-01 -1.45227171e-03] | [6.911042213439941, 0.19230982661247253] |
cbb6d316-0d3e-48bc-95fc-030bed8a3ab7 | exploiting-category-names-for-few-shot | 2211.16594 | null | https://arxiv.org/abs/2211.16594v3 | https://arxiv.org/pdf/2211.16594v3.pdf | Exploiting Category Names for Few-Shot Classification with Vision-Language Models | Vision-language foundation models pretrained on large-scale data provide a powerful tool for many visual understanding tasks. Notably, many vision-language models build two encoders (visual and textual) that can map two modalities into the same embedding space. As a result, the learned representations achieve good zero-shot performance on tasks like image classification. However, when there are only a few examples per category, the potential of large vision-language models is often underperformed, mainly due to the gap between a large number of parameters and a relatively small amount of training data. This paper shows that we can significantly improve the performance of few-shot classification by using the category names to initialize the classification head. With the proposed category name initialization method, our model obtains the state-of-the-art performance on a number of few-shot image classification benchmarks (e.g., 87.37% on ImageNet and 96.08% on Stanford Cars, both using five-shot learning). | ['Ming-Hsuan Yang', 'Shengyang Dai', 'Jiahui Yu', 'Liangliang Cao', 'ZiRui Wang', 'Taihong Xiao'] | 2022-11-29 | null | null | null | null | ['few-shot-image-classification'] | ['computer-vision'] | [-7.83912838e-02 -7.81067312e-02 -5.40942073e-01 -4.64060843e-01
-6.96804106e-01 -2.78677404e-01 1.05473650e+00 -3.42872851e-02
-5.00607610e-01 3.55730712e-01 2.68905073e-01 -1.22913651e-01
4.65175301e-01 -7.50044405e-01 -7.94848442e-01 -4.83669400e-01
3.23854268e-01 2.29510918e-01 4.58241075e-01 -2.06135437e-01
1.20414287e-01 -8.80910456e-02 -1.99312377e+00 2.03949377e-01
4.41519648e-01 1.03935277e+00 2.79932767e-01 5.15303671e-01
-5.08517325e-01 1.09544837e+00 -2.14475587e-01 -5.10644376e-01
1.14119254e-01 -1.90065145e-01 -6.54417753e-01 2.30804071e-01
8.74083996e-01 -6.21667743e-01 -6.35759175e-01 1.18476117e+00
2.12778330e-01 2.63465196e-01 6.82368219e-01 -1.51009834e+00
-1.07651675e+00 5.15056193e-01 -6.08804882e-01 1.99905828e-01
3.74193233e-03 4.03292775e-01 1.19412625e+00 -1.25739861e+00
7.70332992e-01 1.22941744e+00 4.06809151e-01 7.40774155e-01
-1.14919960e+00 -6.50948882e-01 1.74864784e-01 5.65144837e-01
-1.38021576e+00 -6.28356278e-01 4.01865959e-01 -6.78028464e-01
1.13189030e+00 -2.74123758e-01 5.13606787e-01 1.06221306e+00
-5.47617041e-02 6.42258286e-01 8.60538840e-01 -3.45482528e-01
3.58277261e-01 3.36679369e-01 6.36824071e-01 7.72842586e-01
2.21761435e-01 9.66414344e-03 -5.53419232e-01 1.57776996e-01
3.96021843e-01 3.48420918e-01 6.56963736e-02 -5.88015318e-01
-1.09062827e+00 1.21074307e+00 7.96340346e-01 2.76303142e-01
2.06025206e-02 4.07669157e-01 5.40976405e-01 1.63223222e-01
3.09197456e-01 3.09767514e-01 1.68852471e-02 -3.60893011e-02
-7.01275647e-01 -7.37235323e-02 6.12581372e-01 1.11842597e+00
1.06100905e+00 1.55335262e-01 -3.17811668e-01 1.04121661e+00
2.26131275e-01 6.36209369e-01 5.13839245e-01 -8.85013819e-01
2.70346373e-01 3.87969583e-01 -1.18892245e-01 -6.80326581e-01
2.30579209e-02 -1.53137922e-01 -7.64510810e-01 2.53214121e-01
3.59561235e-01 1.27771080e-01 -1.48420215e+00 1.81080520e+00
-5.36011755e-02 4.71900284e-01 1.45588189e-01 8.58778536e-01
1.24646282e+00 7.94082999e-01 4.19469297e-01 1.56278834e-01
1.67024100e+00 -1.36281371e+00 -6.21801853e-01 -8.21375430e-01
5.82387567e-01 -4.36591983e-01 1.32883811e+00 -2.16411591e-01
-7.71617532e-01 -7.25520968e-01 -1.08876395e+00 -3.37122023e-01
-6.62468493e-01 -7.72522576e-03 7.39016891e-01 5.00161350e-01
-9.95193183e-01 2.03650028e-01 -7.36292422e-01 -8.38579834e-01
7.68299282e-01 -1.22191742e-01 -4.16790962e-01 -7.17170775e-01
-9.81269062e-01 9.91980553e-01 2.11402133e-01 -4.82974291e-01
-1.31026900e+00 -8.23235691e-01 -1.14762068e+00 1.42591313e-01
3.15815270e-01 -6.55231118e-01 1.21429837e+00 -6.78148389e-01
-1.08783555e+00 1.14317751e+00 -1.95747569e-01 -5.47945380e-01
1.94799095e-01 -1.84683248e-01 -2.54821897e-01 2.25161433e-01
2.99453288e-01 1.11726928e+00 9.09866154e-01 -1.19754350e+00
-6.01615787e-01 -2.62054741e-01 2.40075663e-01 9.62171629e-02
-6.95122123e-01 5.83963282e-02 -8.38126361e-01 -2.69576281e-01
-2.98340678e-01 -8.24169815e-01 -2.36218154e-01 3.80262852e-01
-1.78215608e-01 -4.28552032e-01 7.03866005e-01 -2.64440775e-01
8.17536235e-01 -2.25621939e+00 2.20491197e-02 -3.70981336e-01
3.37644905e-01 3.69565308e-01 -4.60775405e-01 1.27239943e-01
1.45415375e-02 1.09998725e-01 -3.28794532e-02 -4.78605062e-01
-2.73532588e-02 2.86683083e-01 -4.23253626e-01 3.65167081e-01
1.73779756e-01 1.16383660e+00 -9.22074139e-01 -3.76170516e-01
4.66162413e-01 3.82696658e-01 -4.21182752e-01 3.54016930e-01
-1.46497920e-01 -2.40908608e-01 -3.10085155e-02 5.66869378e-01
3.88597041e-01 -7.50826776e-01 -2.10788786e-01 -2.33628899e-01
4.47328202e-02 4.47071269e-02 -8.12173963e-01 1.92055690e+00
-4.69251812e-01 9.23109889e-01 -4.51877862e-01 -9.93524253e-01
6.60551548e-01 5.93808480e-02 1.79523513e-01 -5.85211515e-01
1.01962149e-01 -1.30322695e-01 1.04124297e-03 -3.42509061e-01
3.44354868e-01 -2.42164612e-01 -8.03853422e-02 4.34924692e-01
6.05051637e-01 -5.57936691e-02 3.72574598e-01 5.92074335e-01
1.11058760e+00 -3.37025493e-01 3.77314270e-01 5.59280766e-03
1.44618884e-01 7.81015530e-02 2.78361797e-01 9.90900695e-01
-3.87881666e-01 5.99613130e-01 3.78247023e-01 -4.38836306e-01
-1.25603116e+00 -1.09295750e+00 -8.99476930e-02 1.37642682e+00
2.66469657e-01 -5.58472157e-01 -4.68797147e-01 -5.05408645e-01
7.37398937e-02 8.98794949e-01 -8.11294019e-01 -6.06623411e-01
6.17736578e-02 -4.93370444e-01 4.54863608e-01 7.84377754e-01
4.50273961e-01 -9.33945060e-01 -4.92112458e-01 -3.18420529e-02
1.91716343e-01 -1.45628071e+00 -2.86908209e-01 1.04643844e-01
-5.37990451e-01 -9.69478786e-01 -7.70989001e-01 -9.17614937e-01
5.36419928e-01 9.24069107e-01 1.17433786e+00 7.17350245e-02
-5.68700373e-01 5.66116631e-01 -3.02122176e-01 -3.78284425e-01
-2.14512482e-01 -2.03004554e-01 -8.84744525e-02 -8.33151788e-02
8.45009685e-01 -1.99454397e-01 -4.71247047e-01 1.65351853e-01
-6.58699930e-01 1.77857831e-01 5.05757630e-01 1.02142942e+00
3.21626961e-01 -4.65331018e-01 5.28023362e-01 -6.76479697e-01
1.49399146e-01 -5.61871469e-01 -4.97714043e-01 3.64858180e-01
-6.00453138e-01 6.21524155e-02 5.87920129e-01 -6.39417648e-01
-9.17024672e-01 -3.23367938e-02 2.23574191e-01 -1.00521433e+00
-3.18632782e-01 3.92479688e-01 1.81414455e-01 -3.17233615e-02
7.02347338e-01 3.11117798e-01 1.50758326e-01 -3.44227076e-01
1.00987482e+00 6.98915660e-01 4.18162286e-01 -5.34246564e-02
8.39698851e-01 4.92445171e-01 -4.09204453e-01 -1.00582612e+00
-1.18882036e+00 -8.19110811e-01 -4.97307926e-01 -1.09061413e-02
1.13225079e+00 -1.36285460e+00 -3.10497195e-01 4.55728889e-01
-1.06937826e+00 -1.92035541e-01 -3.51403713e-01 4.28017706e-01
-4.75691438e-01 1.53618604e-01 -6.28235281e-01 -4.80959892e-01
-2.96161592e-01 -1.08136785e+00 8.98717701e-01 3.10610116e-01
1.39573276e-01 -7.68011689e-01 2.12464854e-02 3.14581931e-01
4.93220180e-01 -3.52756470e-01 9.41313922e-01 -6.02295578e-01
-6.47187233e-01 -1.54467538e-01 -7.28637218e-01 3.32427233e-01
2.80911047e-02 -2.03775004e-01 -1.24518192e+00 -3.09681386e-01
-3.86903137e-01 -9.76472735e-01 1.41604888e+00 2.68766284e-01
1.14646125e+00 1.32490888e-01 -3.34760606e-01 6.37550116e-01
1.68389666e+00 -1.73147526e-02 6.20081484e-01 1.25811398e-01
9.41231608e-01 2.29677841e-01 5.43829918e-01 3.48362535e-01
4.10219014e-01 5.69260418e-01 3.83096516e-01 1.77476015e-02
-5.11870027e-01 -3.57440352e-01 1.78816825e-01 6.72750354e-01
1.07303135e-01 6.59028674e-03 -1.06273866e+00 7.86721230e-01
-1.88714874e+00 -1.16316712e+00 3.33037794e-01 2.07460499e+00
7.98102379e-01 2.25590557e-01 8.34564567e-02 -5.34161687e-01
6.87396646e-01 4.53612328e-01 -7.34484255e-01 -9.20255482e-02
8.43395852e-03 2.15984527e-02 5.71031749e-01 2.66234130e-01
-1.29951823e+00 1.22280872e+00 6.64920759e+00 6.96229517e-01
-1.05763352e+00 2.76359409e-01 4.24405009e-01 -1.16797507e-01
-1.13563135e-01 6.88969940e-02 -1.14075065e+00 3.95323038e-01
1.09154963e+00 -5.01301587e-01 4.15442199e-01 1.27412140e+00
-2.95569122e-01 4.44721580e-02 -1.23641098e+00 1.36749542e+00
5.57418525e-01 -1.69805729e+00 2.57794559e-01 -2.41823960e-03
7.76501060e-01 6.06118977e-01 7.21261725e-02 6.45561278e-01
4.02482033e-01 -1.09696209e+00 5.80287397e-01 1.73892289e-01
1.09399807e+00 -3.67987156e-01 5.94695508e-01 2.50179470e-01
-1.14461339e+00 -2.18989462e-01 -9.24155831e-01 1.83748007e-02
1.48686752e-01 1.97956458e-01 -6.66936457e-01 -3.68800461e-02
6.70997977e-01 1.15023661e+00 -8.19872439e-01 1.09663093e+00
-1.42939046e-01 5.82957625e-01 -5.06630875e-02 -9.06400383e-02
3.72940361e-01 2.08747089e-02 1.31285965e-01 1.03036273e+00
7.99646750e-02 6.19548373e-02 4.04453129e-01 9.20945108e-01
-4.51098382e-01 -1.17225930e-01 -9.86547291e-01 -3.15983713e-01
4.87535924e-01 1.25716412e+00 -5.12196004e-01 -7.56522059e-01
-1.13477314e+00 7.26710677e-01 6.48264110e-01 3.95701677e-01
-8.88364375e-01 -4.61377501e-01 8.65923047e-01 -1.27067298e-01
5.04495978e-01 -5.67332022e-02 -9.23436731e-02 -1.42834568e+00
-3.81368399e-01 -7.43752003e-01 3.04040998e-01 -9.12821949e-01
-1.47548079e+00 4.12254602e-01 2.96411775e-02 -1.19772649e+00
-3.33214849e-01 -7.89480209e-01 -6.31812811e-01 5.03711641e-01
-1.60068464e+00 -1.33665168e+00 -5.56517899e-01 7.02625096e-01
8.95356715e-01 -4.65837151e-01 9.84186828e-01 2.82554746e-01
-4.64968830e-01 6.27422929e-01 1.36058807e-01 3.66623372e-01
7.95035541e-01 -1.14940608e+00 4.46081787e-01 8.46799493e-01
4.96644467e-01 5.38003147e-01 6.05839193e-01 -3.08286160e-01
-1.47579718e+00 -1.08824348e+00 6.83546007e-01 -5.73937297e-01
9.31953192e-01 -4.41105634e-01 -9.38252091e-01 8.56214523e-01
4.14876342e-01 5.05954981e-01 6.92784727e-01 2.97971666e-01
-9.35253024e-01 -8.52960795e-02 -7.50661135e-01 6.54022753e-01
1.04768395e+00 -9.21669006e-01 -8.98987114e-01 4.77898300e-01
8.36818039e-01 7.20019042e-02 -7.02872455e-01 -9.08202007e-02
4.80736911e-01 -6.30040169e-01 1.15095568e+00 -9.74454224e-01
8.08706880e-01 -1.57854885e-01 -4.28370565e-01 -1.17826033e+00
-5.14235139e-01 -6.34480128e-03 -2.37031415e-01 1.15121555e+00
2.02225000e-01 -3.03007573e-01 5.88381827e-01 7.06712246e-01
1.09931491e-01 -5.09954989e-01 -6.39599204e-01 -9.15311635e-01
2.47863494e-02 -2.62584060e-01 1.39023587e-01 8.63374531e-01
-1.75912082e-01 8.92950237e-01 -5.36588848e-01 -2.80978143e-01
9.79096949e-01 1.76095724e-01 8.85348022e-01 -1.08627367e+00
-1.14040114e-01 -2.80224442e-01 -7.78626680e-01 -8.03464174e-01
3.94160926e-01 -9.47793484e-01 1.89504117e-01 -1.60330474e+00
8.98918509e-01 -1.57126456e-01 -5.92345834e-01 8.61033082e-01
-2.29232103e-01 4.26170796e-01 5.65870583e-01 2.11468071e-01
-9.12350476e-01 6.57370985e-01 8.27540159e-01 -5.24619281e-01
1.91845030e-01 -4.24109876e-01 -7.46155918e-01 7.61048794e-01
5.54386377e-01 -3.39056909e-01 -5.85166991e-01 -5.39109111e-01
-3.02653164e-01 -3.74956489e-01 5.38697183e-01 -9.42858160e-01
3.30832541e-01 -2.95585304e-01 2.35222906e-01 -3.83830994e-01
6.09365761e-01 -5.43550491e-01 -2.45286360e-01 4.45597291e-01
-5.34378052e-01 -4.21252191e-01 4.83766161e-02 8.32770348e-01
-3.61220419e-01 -1.92828402e-01 1.13936031e+00 -2.26282805e-01
-1.65995216e+00 4.85807598e-01 -9.31159407e-02 2.68418044e-01
1.19106376e+00 -9.47888568e-02 -8.77824962e-01 -3.65959138e-01
-4.80883688e-01 2.19195217e-01 6.54124856e-01 7.79293537e-01
7.28413820e-01 -1.48176074e+00 -3.89825732e-01 1.45218804e-01
8.03508162e-01 -4.69891995e-01 2.64190853e-01 5.79171240e-01
-1.45702317e-01 4.72582668e-01 -4.85870898e-01 -7.82321572e-01
-1.22625744e+00 8.56602788e-01 6.48507774e-02 1.15229242e-01
-8.51155639e-01 9.82831120e-01 5.38922966e-01 -6.38949350e-02
4.01199996e-01 -3.01799253e-02 -2.67497540e-01 2.40336016e-01
1.00802004e+00 1.09633058e-01 -2.45989621e-01 -6.95504963e-01
-4.54459310e-01 7.59329617e-01 -4.04163092e-01 4.35186885e-02
1.33335888e+00 -8.13910887e-02 2.27321669e-01 6.69335723e-01
1.49145293e+00 -6.33107960e-01 -1.46097982e+00 -5.96259177e-01
-2.30094105e-01 -5.91977239e-01 2.91258454e-01 -5.32747090e-01
-9.94103432e-01 1.32489109e+00 7.79425800e-01 -1.29107982e-01
7.02034056e-01 4.66042250e-01 6.41224921e-01 6.58242881e-01
3.94517362e-01 -1.06471384e+00 4.04309839e-01 6.38182402e-01
5.72836041e-01 -1.82625699e+00 -4.90336455e-02 -1.85370073e-01
-1.03334808e+00 7.46957302e-01 8.59935582e-01 -2.33358428e-01
7.47360528e-01 -4.69587967e-02 6.24413155e-02 -1.53637141e-01
-1.14163721e+00 -6.76729739e-01 3.39790821e-01 6.40545964e-01
2.67886311e-01 -3.19093757e-04 1.33544326e-01 3.73097748e-01
2.59641618e-01 2.69127265e-02 6.07658207e-01 6.94726467e-01
-8.62693131e-01 -6.51839674e-01 1.62010238e-01 6.89905822e-01
-1.28453225e-01 -2.19073266e-01 -4.99865450e-02 7.04070508e-01
-2.34798595e-01 9.28695083e-01 2.84401953e-01 -3.72592539e-01
1.26931310e-01 1.98956236e-01 3.08170140e-01 -1.07338727e+00
1.08262263e-02 -2.32994691e-01 -3.78304190e-04 -8.19535673e-01
-4.15785074e-01 -3.22574884e-01 -1.07921576e+00 -2.93928176e-01
-1.93057284e-01 -3.04103822e-01 5.66538572e-01 9.47556198e-01
3.32800984e-01 2.99796313e-01 3.98577660e-01 -6.23526573e-01
-8.24858129e-01 -9.45042372e-01 -7.28597522e-01 8.24207962e-01
2.55038917e-01 -1.00158846e+00 -4.16067153e-01 2.07156837e-01] | [10.019664764404297, 2.221747875213623] |
9b7bc68f-8570-46dc-a0c5-464f6976a253 | sentence-extraction-based-machine-reading | 2105.09043 | null | https://arxiv.org/abs/2105.09043v2 | https://arxiv.org/pdf/2105.09043v2.pdf | Sentence Extraction-Based Machine Reading Comprehension for Vietnamese | The development of natural language processing (NLP) in general and machine reading comprehension in particular has attracted the great attention of the research community. In recent years, there are a few datasets for machine reading comprehension tasks in Vietnamese with large sizes, such as UIT-ViQuAD and UIT-ViNewsQA. However, the datasets are not diverse in answers to serve the research. In this paper, we introduce UIT-ViWikiQA, the first dataset for evaluating sentence extraction-based machine reading comprehension in the Vietnamese language. The UIT-ViWikiQA dataset is converted from the UIT-ViQuAD dataset, consisting of comprises 23.074 question-answers based on 5.109 passages of 174 Wikipedia Vietnamese articles. We propose a conversion algorithm to create the dataset for sentence extraction-based machine reading comprehension and three types of approaches for sentence extraction-based machine reading comprehension in Vietnamese. Our experiments show that the best machine model is XLM-R_Large, which achieves an exact match (EM) of 85.97% and an F1-score of 88.77% on our dataset. Besides, we analyze experimental results in terms of the question type in Vietnamese and the effect of context on the performance of the MRC models, thereby showing the challenges from the UIT-ViWikiQA dataset that we propose to the language processing community. | ['Ngan Luu-Thuy Nguyen', 'Anh Gia-Tuan Nguyen', 'Kiet Van Nguyen', 'Tin Van Huynh', 'Nhat Duy Nguyen', 'Phong Nguyen-Thuan Do'] | 2021-05-19 | null | null | null | null | ['vietnamese-datasets'] | ['natural-language-processing'] | [ 2.24584699e-01 2.86694676e-01 2.50813901e-01 -3.14925015e-01
-1.35486782e+00 -6.80198610e-01 3.05873603e-01 6.48034215e-01
-9.02115345e-01 7.89338887e-01 6.12250805e-01 -7.27202237e-01
4.74267304e-02 -9.91998911e-01 -7.34787643e-01 -2.07569197e-01
4.65045005e-01 5.60938895e-01 2.54175663e-01 -7.94412374e-01
5.53663731e-01 -4.50372100e-01 -1.23776984e+00 5.56682408e-01
1.81846726e+00 7.27472782e-01 5.65744698e-01 9.72221076e-01
-2.30995178e-01 7.33925939e-01 -9.23053086e-01 -8.10091913e-01
-8.54419768e-02 -5.84510803e-01 -1.39148474e+00 -4.99518991e-01
6.15207374e-01 -3.57439697e-01 -2.57159919e-01 7.76539803e-01
6.68576419e-01 1.04133785e-01 6.31311655e-01 -8.55307221e-01
-9.50581968e-01 8.30930650e-01 -3.07744175e-01 4.56210434e-01
1.13684380e+00 2.20543873e-02 1.12351716e+00 -8.27541828e-01
7.02115357e-01 1.31213880e+00 1.48237452e-01 4.98898268e-01
-5.38411438e-01 -2.63784409e-01 -6.41754121e-02 7.35035360e-01
-1.17067707e+00 -2.42372215e-01 3.44106793e-01 -9.41345319e-02
1.22433805e+00 3.66541684e-01 2.42394835e-01 1.06656682e+00
3.88165712e-01 1.07814133e+00 1.20463777e+00 -7.49436438e-01
5.50169460e-02 -1.93737388e-01 5.92279077e-01 5.96002162e-01
-2.94194277e-02 -5.71778953e-01 -7.26382732e-01 2.01855898e-01
1.33845523e-01 -5.89193404e-01 -5.00059545e-01 5.46294749e-01
-1.33840466e+00 1.02685285e+00 1.79913029e-01 1.67342946e-01
-1.14617288e-01 -6.84843361e-01 6.57022119e-01 7.18728602e-01
3.82889062e-01 6.42537832e-01 -7.84002602e-01 -4.02208120e-01
-3.74009281e-01 5.39064109e-01 1.23014092e+00 1.44569290e+00
3.21871310e-01 -4.81730580e-01 -4.72913235e-01 8.49824369e-01
-8.43874514e-02 9.49902534e-01 4.86714363e-01 -7.33816683e-01
1.40335107e+00 5.08295774e-01 -2.66944319e-02 -1.13432014e+00
-2.74631172e-01 -7.97516257e-02 -8.77547383e-01 -7.30316937e-01
5.58755100e-01 -3.29259336e-01 -5.58835030e-01 1.52632582e+00
2.37800293e-02 -9.13985074e-01 5.41739404e-01 7.01369643e-01
1.42563069e+00 1.17534816e+00 5.18405139e-02 -2.53311008e-01
1.65025640e+00 -1.20460689e+00 -9.00380135e-01 -2.93067962e-01
7.65186548e-01 -8.25733185e-01 1.57807529e+00 4.73868012e-01
-1.30757248e+00 -7.21415222e-01 -8.54971766e-01 -7.48432100e-01
-4.82693523e-01 3.68322828e-03 4.27895114e-02 3.70422751e-01
-7.61031330e-01 -1.52324840e-01 -2.94201761e-01 -6.34753823e-01
-4.50422568e-03 -2.41616771e-01 -1.91561192e-01 -5.65279365e-01
-1.59159029e+00 1.04734313e+00 5.99267185e-01 6.45705089e-02
-7.60362029e-01 -6.23606980e-01 -9.38868701e-01 1.84325010e-01
7.30838478e-01 -6.17129683e-01 1.48474276e+00 -6.38354838e-01
-1.25678062e+00 7.08253860e-01 -6.38778925e-01 -3.59757662e-01
3.75944853e-01 -5.29119134e-01 -4.08334255e-01 4.21328992e-01
4.20435816e-01 4.15888935e-01 4.44328904e-01 -8.04047465e-01
-7.59756863e-01 -3.03967774e-01 2.66600341e-01 4.15856361e-01
-1.17038824e-01 1.77726656e-01 -3.61023277e-01 -4.02368963e-01
-1.00403123e-01 -4.08856899e-01 1.15630567e-01 -8.18173170e-01
-3.92856300e-01 -7.96413481e-01 1.94632083e-01 -1.49168611e+00
1.62143576e+00 -1.65505195e+00 1.75789282e-01 -1.40829071e-01
2.86481529e-01 3.41948420e-01 -6.12417758e-01 1.10956800e+00
4.13788140e-01 3.73439699e-01 -3.30585837e-01 3.25206369e-02
3.67089175e-02 1.58352494e-01 -4.29015905e-01 -2.44930178e-01
3.28096896e-01 1.22500408e+00 -1.10559988e+00 -7.13695526e-01
-3.58186185e-01 -1.84788272e-01 -3.24819744e-01 7.86361635e-01
-4.25996214e-01 1.80126280e-01 -6.07318997e-01 6.50852799e-01
6.92050755e-01 1.66276649e-01 -3.00353672e-02 2.15976179e-01
-2.20167682e-01 6.90939367e-01 -2.61289239e-01 1.74129093e+00
-3.98480117e-01 6.14231646e-01 -1.13401368e-01 -5.01792669e-01
8.61838222e-01 2.70148844e-01 -2.88152903e-01 -1.33110499e+00
2.02364385e-01 2.61267424e-01 1.49710536e-01 -1.11144996e+00
9.34931397e-01 2.40780503e-01 -4.52366233e-01 3.14924061e-01
9.94130876e-03 -4.02327299e-01 8.34106684e-01 6.13005161e-01
1.17151892e+00 -2.22095877e-01 4.70451474e-01 -3.29670787e-01
8.25286210e-01 5.07871866e-01 2.86201566e-01 9.29132283e-01
-1.30435169e-01 6.24659896e-01 6.56746209e-01 1.38515577e-01
-9.75355923e-01 -1.05750823e+00 9.16196685e-03 1.32400072e+00
-7.04432279e-02 -6.38501227e-01 -1.24994802e+00 -6.45260811e-01
-4.36338186e-01 1.08233380e+00 -3.48391294e-01 5.24629243e-02
-9.61422980e-01 -3.80295724e-01 7.52466321e-01 4.00620043e-01
1.02142417e+00 -1.21716189e+00 -4.83219773e-01 2.27089435e-01
-1.05527639e+00 -1.31751001e+00 -5.38125575e-01 -2.57589757e-01
-5.04707873e-01 -1.14682484e+00 -6.71918869e-01 -1.31104112e+00
4.07131970e-01 2.19444409e-01 1.63442957e+00 2.19381198e-01
1.86515003e-01 3.82626504e-01 -1.14480031e+00 -7.68044651e-01
-5.52885532e-01 7.17647612e-01 -4.31763768e-01 -5.69979191e-01
7.55900264e-01 1.80724695e-01 -2.05119401e-01 1.53458685e-01
-9.72908318e-01 2.03879699e-01 6.14953637e-01 7.64746726e-01
2.42951617e-01 -1.69071004e-01 8.49432707e-01 -9.36732054e-01
1.30750072e+00 -7.21789300e-01 -2.92857498e-01 8.05763245e-01
-2.50139296e-01 -2.43873522e-01 9.18672681e-01 -1.38331950e-01
-1.32421613e+00 -7.80248046e-01 -6.83210373e-01 5.87730348e-01
-1.12424470e-01 1.17977512e+00 -4.58416522e-01 5.34321666e-01
4.90034133e-01 4.43680078e-01 -2.12910220e-01 -4.84757602e-01
3.07436168e-01 1.03963661e+00 7.77184010e-01 -5.55441201e-01
5.96440256e-01 -4.64884430e-01 -6.63484931e-01 -1.25794780e+00
-1.19413400e+00 -4.64573473e-01 -7.67683208e-01 3.20444107e-02
1.14032328e+00 -7.52972305e-01 -7.00813293e-01 7.42961943e-01
-1.45440412e+00 -1.54170692e-01 1.21569060e-01 1.43646076e-01
-2.63173908e-01 5.50693095e-01 -8.79397869e-01 -5.88081598e-01
-8.72279644e-01 -9.24043357e-01 9.33585703e-01 3.97795618e-01
-2.40756735e-01 -9.93881404e-01 -2.75511555e-02 1.05162895e+00
3.59949857e-01 -2.15911772e-02 1.55348670e+00 -7.61252880e-01
-2.84422845e-01 4.33677323e-02 -2.58405507e-01 4.29695219e-01
-3.77447188e-01 -2.50081241e-01 -5.33194125e-01 -3.17964733e-01
1.62554398e-01 -8.19537878e-01 6.80907369e-01 -1.62280351e-02
9.48401392e-01 -4.16590810e-01 4.38756227e-01 -6.71964288e-02
1.31845450e+00 2.19601974e-01 9.46575403e-01 4.04337406e-01
4.78239775e-01 9.09312248e-01 1.12652409e+00 9.11758840e-02
1.11104310e+00 8.82827118e-02 1.42230630e-01 1.53559253e-01
6.44699484e-02 -6.92405403e-01 5.17913818e-01 1.99822545e+00
2.36750580e-02 -6.68900549e-01 -1.17977750e+00 6.59908175e-01
-1.59426141e+00 -6.45440042e-01 -5.88829696e-01 1.84667969e+00
1.09352219e+00 -6.50851652e-02 -2.22918510e-01 -6.81044608e-02
1.83091104e-01 9.71625224e-02 -2.41692394e-01 -8.58222425e-01
-4.01931077e-01 4.97630447e-01 -1.08358671e-03 6.41627610e-01
-7.64266491e-01 1.13532233e+00 5.60709143e+00 9.36189592e-01
-3.92038912e-01 7.59337693e-02 3.68807644e-01 5.44574261e-01
-3.24382126e-01 -9.25680995e-02 -8.55892122e-01 2.42741093e-01
1.05663168e+00 -4.84656394e-01 1.99368805e-01 1.92090720e-01
1.59798816e-01 -5.70686877e-01 -8.74393523e-01 6.59505427e-01
5.63069880e-01 -8.06867957e-01 4.09588516e-01 -3.99599850e-01
5.41898549e-01 -1.11515231e-01 -2.37417579e-01 8.48384202e-01
-1.53206736e-01 -1.23952770e+00 6.24348640e-01 4.60488200e-01
5.26635766e-01 -8.05045962e-01 1.19422805e+00 9.82616544e-01
-8.97347450e-01 4.89410944e-02 -7.39657283e-01 -2.63924867e-01
3.06912512e-01 2.38508746e-01 -7.21325457e-01 8.30414236e-01
6.72345042e-01 3.69348854e-01 -9.71975327e-01 7.01371133e-01
-7.23964930e-01 1.07957947e+00 7.30469301e-02 -7.74628520e-01
2.51428336e-01 -3.61747175e-01 3.79785359e-01 1.14264631e+00
1.04261100e-01 5.26773572e-01 3.60956676e-02 6.13444984e-01
-1.88226521e-01 6.62354410e-01 -2.41757154e-01 -2.53052618e-02
4.72656041e-01 9.19021547e-01 -3.04444563e-02 -4.88751322e-01
-5.03693283e-01 9.00955975e-01 6.28565431e-01 4.05992538e-01
-5.63427508e-01 -9.39542234e-01 2.88155004e-02 -1.21545114e-01
-1.94259081e-02 -4.69564885e-01 -9.43974853e-02 -1.52587569e+00
3.65619987e-01 -1.62342751e+00 4.29336429e-01 -8.95012438e-01
-1.45214951e+00 7.82591164e-01 1.96151510e-01 -5.98954380e-01
-2.44758472e-01 -6.10856712e-01 -5.87590933e-01 1.07688141e+00
-1.48686588e+00 -1.10327148e+00 -4.72209781e-01 4.37629521e-01
1.15044737e+00 -1.47415297e-02 8.46754551e-01 8.63178372e-02
-6.07002556e-01 6.01529419e-01 2.16756780e-02 4.08385783e-01
6.21775091e-01 -1.33299756e+00 6.10323787e-01 9.92736161e-01
-2.07853451e-01 7.55372643e-01 6.14496410e-01 -6.74175262e-01
-1.63424206e+00 -8.53472650e-01 1.88123417e+00 -6.29105389e-01
6.20082676e-01 -4.85889375e-01 -1.21026087e+00 4.81680661e-01
9.83493924e-01 -1.27762520e+00 6.40177846e-01 2.91453246e-02
-1.09646313e-01 1.87184215e-01 -9.92727160e-01 7.09419131e-01
9.17907298e-01 -6.66754723e-01 -1.23334777e+00 3.46674383e-01
1.08213031e+00 -4.23077285e-01 -1.12349987e+00 3.07107717e-01
1.04221441e-01 -6.41191423e-01 5.25219917e-01 -6.96628094e-01
7.47200191e-01 5.83885983e-02 -1.42181158e-01 -1.33987010e+00
2.41927262e-02 -2.91545093e-01 2.23933533e-01 1.33964264e+00
7.65294015e-01 -6.17556512e-01 2.10766301e-01 6.63405061e-01
-1.77570283e-01 -7.97640622e-01 -9.07901883e-01 -3.69773865e-01
7.68702269e-01 -3.31737816e-01 6.01526856e-01 6.04706526e-01
1.01717487e-01 8.55286121e-01 3.70042063e-02 -7.45028481e-02
2.70684659e-01 8.97928774e-02 8.34963500e-01 -6.69494033e-01
-6.32168651e-02 6.29000068e-02 2.87147999e-01 -1.67906833e+00
1.04098722e-01 -7.51718998e-01 1.29472479e-01 -1.89352047e+00
3.56686652e-01 9.40951407e-02 4.18932110e-01 2.52547294e-01
-5.44162571e-01 -4.83794749e-01 1.89749941e-01 -1.29467780e-02
-6.71567798e-01 6.85607672e-01 1.65493798e+00 -2.00787708e-01
-1.41990678e-02 -2.00078189e-01 -7.24833071e-01 3.29747945e-01
9.57785487e-01 -9.89040583e-02 -4.21661586e-01 -1.09058237e+00
3.44584137e-01 2.96963245e-01 -4.66041304e-02 -5.98057330e-01
2.53986329e-01 -5.97878322e-02 7.95187354e-02 -8.24223578e-01
-1.03678927e-01 -1.99524626e-01 -7.03477025e-01 3.56151760e-01
-7.34052122e-01 7.10349739e-01 -1.14304781e-01 9.92366448e-02
-5.74343622e-01 -6.36397302e-01 1.68036312e-01 -2.43281290e-01
-8.15470159e-01 -3.36554870e-02 -7.48602629e-01 1.01319993e+00
5.72864711e-01 1.74952075e-01 -8.03094923e-01 -6.19203508e-01
-6.72516087e-03 8.01398098e-01 -4.03192304e-02 7.16634035e-01
9.02104855e-01 -7.18050480e-01 -1.35480833e+00 7.14487359e-02
3.95416588e-01 2.36446932e-01 1.65609896e-01 5.88083386e-01
-8.24811220e-01 8.48165870e-01 -2.10918427e-01 -1.79006398e-01
-1.26024544e+00 4.38300639e-01 -1.55224502e-01 -4.75310206e-01
-2.01993346e-01 6.32663071e-01 -1.89324364e-01 -7.05880225e-01
3.92365530e-02 -5.29305637e-01 -7.16284752e-01 4.16494999e-03
6.91184640e-01 6.07622325e-01 6.80865496e-02 -6.39548123e-01
1.24271274e-01 3.48282456e-01 -2.53499001e-01 -1.28412321e-01
8.34188044e-01 -4.73323047e-01 -5.39512813e-01 2.53414482e-01
1.22915733e+00 2.01705188e-01 -1.87900871e-01 -2.03821361e-01
3.00962269e-01 -3.52837384e-01 -4.63760853e-01 -1.18773365e+00
-3.14184695e-01 1.02791262e+00 -1.62365790e-02 1.86013907e-01
1.25790894e+00 4.76084314e-02 1.47534168e+00 8.03596556e-01
5.65851927e-01 -1.24135816e+00 -1.10237405e-01 1.26815963e+00
1.30154490e+00 -1.35012913e+00 -3.61096799e-01 -5.46099424e-01
-8.30105901e-01 1.03178906e+00 8.96220565e-01 2.97885656e-01
1.56338796e-01 -3.55409414e-01 5.14882624e-01 -5.17735630e-02
-7.80620396e-01 -1.86352298e-01 4.37311769e-01 5.66672981e-01
5.70455670e-01 1.01187475e-01 -9.26017225e-01 8.74272406e-01
-8.83471131e-01 -3.80136013e-01 5.90451956e-01 1.11081445e+00
-6.19515002e-01 -1.08007228e+00 -1.52025923e-01 5.52255511e-01
-3.62906575e-01 -4.65596557e-01 -5.80155134e-01 8.60260844e-01
-3.05791259e-01 1.70956635e+00 -2.32455626e-01 -2.89661974e-01
7.69643784e-01 -1.45433158e-01 3.89984190e-01 -6.11289203e-01
-9.16525841e-01 -4.48518872e-01 4.85584557e-01 -1.63444042e-01
-6.29883781e-02 -4.44632947e-01 -1.23701870e+00 -5.04836738e-01
-2.44480893e-01 5.33019245e-01 2.93471575e-01 9.81666207e-01
7.25244358e-02 1.39244989e-01 3.88882935e-01 6.55283481e-02
-7.77785599e-01 -1.39861107e+00 -3.78480673e-01 3.86477888e-01
1.25357404e-01 1.30840361e-01 -2.11490437e-01 -1.50691658e-01] | [11.390488624572754, 8.260286331176758] |
e6e46d45-0eb4-4177-8e7b-fd9eaf3639c0 | objects-matter-learning-object-relation-graph | 2205.13280 | null | https://arxiv.org/abs/2205.13280v1 | https://arxiv.org/pdf/2205.13280v1.pdf | Objects Matter: Learning Object Relation Graph for Robust Camera Relocalization | Visual relocalization aims to estimate the pose of a camera from one or more images. In recent years deep learning based pose regression methods have attracted many attentions. They feature predicting the absolute poses without relying on any prior built maps or stored images, making the relocalization very efficient. However, robust relocalization under environments with complex appearance changes and real dynamics remains very challenging. In this paper, we propose to enhance the distinctiveness of the image features by extracting the deep relationship among objects. In particular, we extract objects in the image and construct a deep object relation graph (ORG) to incorporate the semantic connections and relative spatial clues of the objects. We integrate our ORG module into several popular pose regression models. Extensive experiments on various public indoor and outdoor datasets demonstrate that our method improves the performance significantly and outperforms the previous approaches. | ['Xinglu Wang', 'Zhiyu Xiang', 'Chengyu Qiao'] | 2022-05-26 | null | null | null | null | ['camera-relocalization'] | ['computer-vision'] | [-2.94432282e-01 -3.04287493e-01 -1.93303823e-01 -5.61908007e-01
-2.88189530e-01 -4.15002495e-01 4.25374597e-01 -4.97911964e-03
-4.02826518e-01 5.42459369e-01 1.22507557e-01 3.50915641e-01
-2.79416114e-01 -5.90032697e-01 -9.43685234e-01 -4.85274941e-01
2.42017299e-01 3.70755315e-01 5.71841061e-01 -1.03289947e-01
1.98943555e-01 5.58827877e-01 -1.21481347e+00 -3.23899955e-01
5.96666217e-01 1.11278129e+00 4.38410163e-01 3.14834982e-01
3.18011165e-01 7.24002004e-01 -3.05693954e-01 -2.20558286e-01
2.28804797e-01 3.57508846e-02 -4.15501714e-01 2.84789056e-01
7.70068944e-01 -4.93137866e-01 -7.19040573e-01 1.04470885e+00
3.59537661e-01 7.21617043e-02 3.86705428e-01 -1.29065371e+00
-6.79958344e-01 2.32324496e-01 -9.56231296e-01 1.52211010e-01
5.24718583e-01 2.79604476e-02 8.98777902e-01 -1.08428037e+00
6.09100819e-01 1.26410198e+00 7.50550032e-01 3.45044397e-02
-1.13625443e+00 -7.37780333e-01 5.01613498e-01 5.22635818e-01
-1.81600988e+00 -1.72483400e-01 1.01580036e+00 -3.70245397e-01
7.11017787e-01 1.19506091e-01 8.17886353e-01 1.03470194e+00
1.31398693e-01 8.36310029e-01 7.93129086e-01 -7.91711658e-02
-1.89002380e-01 -1.02558010e-03 -5.67892008e-02 1.05319452e+00
3.77833247e-01 -3.26984763e-01 -6.78647816e-01 8.35669637e-02
9.17646050e-01 5.55586755e-01 -2.39499331e-01 -9.44394112e-01
-1.29250050e+00 5.27289987e-01 1.02752173e+00 -5.22618219e-02
-1.72889844e-01 4.38795328e-01 -5.97382709e-02 -2.17537582e-01
4.29490089e-01 2.96259731e-01 -3.46827328e-01 3.31369966e-01
-6.58304870e-01 1.10077061e-01 2.45239735e-01 1.17392802e+00
1.18537605e+00 -2.03388482e-01 1.17790557e-01 6.42060697e-01
4.80229020e-01 5.63425481e-01 1.46539614e-01 -8.87278199e-01
6.59708440e-01 7.86194384e-01 1.25366971e-01 -1.72568560e+00
-5.46396852e-01 -4.42144722e-01 -6.72626853e-01 -3.00884813e-01
1.13361119e-03 5.45145907e-02 -7.86439717e-01 1.50811303e+00
4.00325030e-01 3.64124686e-01 -5.13894916e-01 1.14524901e+00
7.59889483e-01 5.88285923e-01 -6.64548799e-02 4.39469032e-02
1.01078892e+00 -1.06686127e+00 -5.72030962e-01 -5.06136477e-01
9.47495624e-02 -6.65657997e-01 7.71708727e-01 1.95231110e-01
-7.32055962e-01 -7.33999968e-01 -1.16249287e+00 -1.84006393e-01
-2.35112339e-01 3.78924519e-01 8.61403406e-01 1.36069864e-01
-8.81877184e-01 3.02302241e-01 -1.00450766e+00 -4.25667226e-01
4.28841799e-01 6.31262422e-01 -5.73333859e-01 -2.10480049e-01
-7.22843051e-01 6.31348789e-01 3.55348408e-01 4.02972102e-01
-8.15551460e-01 -3.28116357e-01 -1.15836680e+00 -1.13381989e-01
5.36287129e-01 -6.84643269e-01 7.23529637e-01 -7.29247093e-01
-1.18799150e+00 6.80641830e-01 -1.28703997e-01 -9.61943418e-02
5.48867047e-01 -6.72563553e-01 -9.76610333e-02 1.11388601e-03
3.62999924e-02 8.00680459e-01 7.63466001e-01 -1.36054063e+00
-6.59209728e-01 -6.24827921e-01 2.12312356e-01 3.61968577e-01
-3.66159528e-01 -2.14945644e-01 -1.12200689e+00 -4.93591368e-01
6.68613672e-01 -1.14409733e+00 -3.64838213e-01 3.43952239e-01
-5.81574142e-01 -2.25435928e-01 1.10389841e+00 -5.53813159e-01
9.27986383e-01 -2.05963898e+00 4.67440635e-01 9.61993262e-02
4.69790846e-01 -4.52625230e-02 -1.06222212e-01 1.93320319e-01
1.57990277e-01 -1.66397706e-01 2.98794478e-01 -5.28130054e-01
-1.28747329e-01 3.13464463e-01 -8.88793394e-02 8.37482512e-01
1.68351799e-01 9.98341739e-01 -8.98468733e-01 -6.80344105e-01
4.99146134e-01 6.21214986e-01 -6.35423303e-01 3.30846697e-01
-1.01235338e-01 6.21781051e-01 -6.79307222e-01 6.44044101e-01
7.90664554e-01 -4.03574973e-01 -5.63930208e-03 -5.10384381e-01
9.08207893e-02 4.86361608e-02 -1.18840933e+00 2.10550356e+00
-2.32531667e-01 6.69620574e-01 -2.41501719e-01 -7.33382046e-01
9.90808964e-01 -2.61841148e-01 6.71054423e-01 -5.66054106e-01
6.72240555e-02 -1.42356604e-01 -3.54955435e-01 -3.58405888e-01
5.94412148e-01 4.00129408e-01 -8.39073509e-02 -7.55176470e-02
9.66806486e-02 1.07176118e-02 -1.20363519e-01 1.42100856e-01
7.69992352e-01 4.37124997e-01 3.75701964e-01 -2.74805375e-03
5.88491201e-01 -1.94534436e-01 8.14160407e-01 2.86894172e-01
-1.17102280e-01 8.27276349e-01 8.02787840e-02 -7.29552567e-01
-7.70509064e-01 -1.34919357e+00 3.67296152e-02 8.81235361e-01
1.03650331e+00 -6.40971959e-01 -3.85518849e-01 -8.09408605e-01
7.86372200e-02 -1.37253767e-02 -5.08291364e-01 -2.35981241e-01
-7.95990825e-01 -6.01626873e-01 -2.73329485e-02 7.61256218e-01
7.61026978e-01 -5.43885112e-01 -3.09667379e-01 -6.93484023e-02
-5.23540497e-01 -1.34498775e+00 -5.91205120e-01 -9.89361033e-02
-5.94760776e-01 -1.03368092e+00 -4.64355826e-01 -8.52701426e-01
9.35757041e-01 5.62673867e-01 8.02148104e-01 -5.97919077e-02
-3.49753320e-01 4.16021764e-01 -1.73266694e-01 -1.62211850e-01
5.14397264e-01 1.92017689e-01 3.44690233e-01 1.49250910e-01
2.62946576e-01 -5.74132085e-01 -8.52385223e-01 5.38443863e-01
-3.84230494e-01 1.80810496e-01 7.20201373e-01 4.51314092e-01
8.49087417e-01 -8.02764073e-02 2.90149208e-02 -4.67038095e-01
8.00537020e-02 -2.64027655e-01 -7.40002096e-01 2.76270390e-01
-4.16582108e-01 1.28733560e-01 2.53156394e-01 -3.77454102e-01
-8.64683449e-01 5.58928072e-01 2.46278927e-01 -6.44239068e-01
7.43898451e-02 2.84892559e-01 -4.25811529e-01 -2.75878817e-01
3.31448346e-01 1.89045727e-01 -4.09517199e-01 -5.64742088e-01
3.04299653e-01 2.79530764e-01 5.56564212e-01 -4.21322614e-01
1.14215171e+00 5.79064369e-01 1.64921820e-01 -6.30383074e-01
-1.11403751e+00 -5.09687364e-01 -1.11554015e+00 -3.92707407e-01
9.82581496e-01 -1.23389995e+00 -9.69921410e-01 2.30470151e-01
-1.22672772e+00 7.09595978e-02 1.50179192e-01 7.27699637e-01
-4.78385746e-01 3.61061633e-01 -4.73408729e-01 -4.14710730e-01
-1.73736997e-02 -1.18240857e+00 1.29965651e+00 3.56135875e-01
3.74579094e-02 -8.00900459e-01 -6.44265627e-03 3.82393301e-01
-1.61543395e-02 4.20102149e-01 3.52219075e-01 -2.78626364e-02
-1.12362957e+00 -2.28846014e-01 -4.54719692e-01 1.63424872e-02
2.75152415e-01 9.92092886e-04 -7.72503555e-01 -3.73479754e-01
-2.85515279e-01 -2.58518398e-01 9.06588316e-01 2.27320060e-01
1.46569335e+00 -1.90311641e-01 -6.44004524e-01 9.85946834e-01
1.42900634e+00 -1.58191681e-01 3.07501405e-01 3.09431195e-01
1.33753014e+00 4.71200615e-01 8.59477580e-01 4.90404487e-01
5.81746459e-01 8.88463557e-01 8.15952480e-01 -1.92458659e-01
-2.48955470e-02 -6.34816289e-01 2.06634745e-01 8.86365354e-01
-2.98086971e-01 -1.12421744e-01 -8.43926311e-01 2.56176740e-01
-2.13655901e+00 -5.51528931e-01 -1.24702886e-01 1.93380499e+00
5.86371779e-01 1.31866798e-01 -2.71456331e-01 -4.08650339e-01
7.63080239e-01 3.35042506e-01 -6.45563364e-01 4.92189169e-01
-1.45142619e-02 -3.23013574e-01 6.89007938e-01 4.53315109e-01
-1.36813641e+00 1.14668310e+00 5.96364594e+00 5.26818275e-01
-1.01603472e+00 -1.57412048e-02 4.33413565e-01 -6.83919266e-02
7.15692341e-02 -2.73398682e-03 -1.02113450e+00 2.56269306e-01
7.61284754e-02 1.15134314e-01 1.96340978e-01 1.09995306e+00
1.62183568e-02 -1.67899981e-01 -1.19636929e+00 1.35658765e+00
4.49864596e-01 -1.09739137e+00 4.28020805e-02 8.04888755e-02
7.52690613e-01 -2.97556119e-03 2.32141778e-01 1.39365286e-01
2.99796432e-01 -8.25018227e-01 8.88139367e-01 8.96552563e-01
4.93462682e-01 -8.54401529e-01 6.25895441e-01 3.17886353e-01
-1.50649154e+00 -8.69806185e-02 -5.79870105e-01 -1.27611682e-01
-4.72748792e-03 2.82971859e-01 -7.25381017e-01 5.31705081e-01
1.01480448e+00 1.21224093e+00 -1.11090994e+00 1.04288590e+00
-3.37455839e-01 2.27746159e-01 -3.16748649e-01 4.59607877e-02
1.09554254e-01 -1.86188459e-01 2.99050122e-01 7.65516281e-01
2.37774953e-01 -1.32402956e-01 5.68048954e-01 8.26767266e-01
-5.01341093e-03 -9.87679958e-02 -5.61674476e-01 3.36149126e-01
3.42957169e-01 1.36686504e+00 -8.09199572e-01 2.80344058e-02
-4.36727047e-01 1.18152428e+00 6.04027689e-01 3.72378916e-01
-1.12407053e+00 -8.91265199e-02 8.37385356e-01 2.95604199e-01
3.59198391e-01 -7.89411187e-01 1.24989405e-01 -1.25549364e+00
1.69011399e-01 -4.38405186e-01 1.73571751e-01 -1.07092655e+00
-1.07340789e+00 3.16815406e-01 1.36335110e-02 -1.35984433e+00
1.01168804e-01 -6.37024820e-01 -2.21985951e-01 4.85644221e-01
-1.32044256e+00 -1.56357801e+00 -8.50629210e-01 6.61790311e-01
5.26454389e-01 1.22236840e-01 3.10501009e-01 2.79711604e-01
-7.27116883e-01 4.26090121e-01 -2.00088412e-01 4.10143703e-01
7.90365636e-01 -1.12425983e+00 1.60278812e-01 7.89283156e-01
4.23611969e-01 8.10337245e-01 5.25690973e-01 -6.00360155e-01
-1.58529937e+00 -1.21023118e+00 5.11370540e-01 -7.29948401e-01
5.77259839e-01 -6.83583200e-01 -5.11066139e-01 9.46339965e-01
1.46909207e-02 3.99688214e-01 2.69593716e-01 3.20844240e-02
-3.60284299e-01 -4.42453235e-01 -6.03746295e-01 6.00400150e-01
1.46040297e+00 -4.21531528e-01 -2.94401497e-01 4.49569285e-01
8.63709748e-01 -6.72529876e-01 -5.56685507e-01 5.62521935e-01
4.15241748e-01 -8.63149405e-01 1.43464935e+00 -3.06322187e-01
1.26954347e-01 -6.38654113e-01 -2.43479058e-01 -8.97631943e-01
-5.33080161e-01 -2.18652263e-01 -1.45218700e-01 1.21394932e+00
1.99219715e-02 -2.98532516e-01 8.41937363e-01 4.25520509e-01
4.42105308e-02 -5.57416499e-01 -6.91021144e-01 -6.51160955e-01
-5.78031301e-01 -2.54937440e-01 4.56661612e-01 7.90257752e-01
-3.47503066e-01 5.29301107e-01 -7.99126804e-01 5.23618281e-01
6.09055102e-01 2.90816635e-01 1.12179029e+00 -1.28915799e+00
-1.75461784e-01 -4.58279774e-02 -1.01266944e+00 -1.41689980e+00
2.28718191e-01 -6.81466997e-01 3.76129225e-02 -1.61809278e+00
4.94407713e-01 -4.36163962e-01 -4.14042443e-01 5.98450899e-01
-2.77578622e-01 4.79499578e-01 1.34192899e-01 3.13577414e-01
-1.26768100e+00 8.61390352e-01 1.28777349e+00 -3.98182064e-01
5.85403666e-03 -4.12960388e-02 -4.24180061e-01 1.02893639e+00
5.55399954e-01 -4.55706716e-01 -4.96735603e-01 -6.90742791e-01
3.56783271e-01 -2.57651806e-01 6.58665776e-01 -1.30785441e+00
5.01388848e-01 -2.67872036e-01 9.07002926e-01 -9.13329184e-01
7.08441377e-01 -1.14689922e+00 9.90266204e-02 4.21949536e-01
-6.57660291e-02 3.21474165e-01 -6.67709038e-02 1.03514218e+00
-1.94864169e-01 2.13904679e-01 5.28913140e-01 -3.43597457e-02
-1.14030123e+00 7.13911772e-01 1.99108616e-01 -2.11571008e-01
1.25788796e+00 -9.88945290e-02 -1.17338903e-01 -4.08947051e-01
-6.43549144e-01 3.14850390e-01 5.59867024e-01 7.66959012e-01
8.80287647e-01 -1.60148227e+00 -2.56843328e-01 1.23661131e-01
3.93863827e-01 3.41534317e-01 1.20289326e-01 9.30778444e-01
-6.88712955e-01 4.55676556e-01 -1.58302844e-01 -9.66257334e-01
-1.41432273e+00 7.03117430e-01 3.06877136e-01 1.72405481e-01
-5.44128418e-01 8.23696673e-01 5.55803001e-01 -3.69824052e-01
2.98497528e-01 -3.23034436e-01 -2.12908044e-01 -1.25796974e-01
2.50824839e-01 2.32815042e-01 -2.84444898e-01 -1.01922667e+00
-5.86540282e-01 1.01635718e+00 -1.34009257e-01 2.00373143e-01
1.45988822e+00 -4.10825700e-01 -1.43830225e-01 4.13261831e-01
1.29531395e+00 3.74619961e-02 -1.66486967e+00 -5.60521066e-01
-3.18169631e-02 -7.95058191e-01 6.92596287e-02 -2.38771200e-01
-1.27029026e+00 7.76253581e-01 5.67841768e-01 -4.46530491e-01
9.15643096e-01 1.73515767e-01 6.13323331e-01 6.88655734e-01
7.14773953e-01 -1.02525079e+00 6.86901927e-01 3.66928607e-01
8.91281426e-01 -1.40678358e+00 3.97552848e-01 -6.50255859e-01
-3.50067735e-01 1.03287470e+00 1.08373427e+00 -2.74103850e-01
7.44969666e-01 8.31318274e-02 -1.47388875e-01 -1.75530136e-01
-1.76830858e-01 -2.68390954e-01 5.34584105e-01 5.66761196e-01
8.28907788e-02 -1.19736008e-01 -3.24404053e-02 3.83496255e-01
-1.26069829e-01 -4.32195395e-01 6.36617932e-03 7.75498092e-01
-4.63849574e-01 -8.86821151e-01 -3.21083903e-01 2.31328234e-02
-6.78649619e-02 1.11191630e-01 -6.21610343e-01 8.85237038e-01
3.60535651e-01 7.06694305e-01 1.38846353e-01 -6.88906193e-01
2.39767939e-01 -4.99013782e-01 5.69511652e-01 -5.75870037e-01
3.12493518e-02 1.01012416e-01 -3.74154568e-01 -8.95739675e-01
-5.16710401e-01 -5.86363137e-01 -1.16244471e+00 -1.47900984e-01
-4.97598588e-01 -3.17816585e-01 5.43911338e-01 8.78113508e-01
2.24432722e-01 4.42290217e-01 5.94551504e-01 -1.16119659e+00
-6.32540435e-02 -7.12472975e-01 -4.39691663e-01 3.69135082e-01
3.72845948e-01 -1.05187035e+00 2.33928135e-04 5.81173077e-02] | [7.788843154907227, -2.267019033432007] |
d568154c-f05c-4eb4-8783-56947a28942a | scalably-learning-quantum-many-body | 2209.14328 | null | https://arxiv.org/abs/2209.14328v1 | https://arxiv.org/pdf/2209.14328v1.pdf | Scalably learning quantum many-body Hamiltonians from dynamical data | The physics of a closed quantum mechanical system is governed by its Hamiltonian. However, in most practical situations, this Hamiltonian is not precisely known, and ultimately all there is are data obtained from measurements on the system. In this work, we introduce a highly scalable, data-driven approach to learning families of interacting many-body Hamiltonians from dynamical data, by bringing together techniques from gradient-based optimization from machine learning with efficient quantum state representations in terms of tensor networks. Our approach is highly practical, experimentally friendly, and intrinsically scalable to allow for system sizes of above 100 spins. In particular, we demonstrate on synthetic data that the algorithm works even if one is restricted to one simple initial state, a small number of single-qubit observables, and time evolution up to relatively short times. For the concrete example of the one-dimensional Heisenberg model our algorithm exhibits an error constant in the system size and scaling as the inverse square root of the size of the data set. | ['Jens Eisert', 'Ryan Sweke', 'Dominik Hangleiter', 'Ingo Roth', 'Augustine Kshetrimayum', 'Frederik Wilde'] | 2022-09-28 | null | null | null | null | ['tensor-networks'] | ['methodology'] | [ 5.44309877e-02 -3.15352887e-01 3.78130600e-02 -1.60464585e-01
-4.52360928e-01 -7.28531301e-01 6.49263501e-01 2.13890433e-01
-6.59910679e-01 9.16189075e-01 -8.76097158e-02 -3.99501055e-01
-2.94603825e-01 -8.13394845e-01 -4.87546414e-01 -1.02097797e+00
-2.82305062e-01 7.62392700e-01 -5.11104846e-03 -5.19412398e-01
3.55987698e-01 5.76294899e-01 -1.08469331e+00 -4.07797515e-01
4.51470703e-01 7.24863231e-01 -3.10705155e-01 1.00522304e+00
5.39326668e-01 5.96270919e-01 -6.68806657e-02 -5.56298420e-02
3.63707185e-01 -5.46729743e-01 -1.04366076e+00 -1.70329303e-01
7.25951791e-02 -4.06719446e-02 -8.38524401e-01 1.17817736e+00
3.66438210e-01 2.69835293e-01 5.48524320e-01 -8.15536916e-01
-3.35259110e-01 6.06277883e-01 -6.46087527e-02 2.62361199e-01
1.38127610e-01 5.29995203e-01 1.30422556e+00 -3.42268199e-01
8.08548927e-01 9.02254820e-01 4.46898013e-01 3.93730909e-01
-1.95714879e+00 -1.81403100e-01 -5.24843812e-01 2.17421725e-01
-1.42394400e+00 -3.75360280e-01 5.53086221e-01 -5.38728595e-01
1.07251465e+00 1.71846941e-01 6.58956587e-01 6.88612044e-01
5.68319380e-01 -1.00184180e-01 1.41703856e+00 -5.83829820e-01
5.39702237e-01 -1.03000030e-01 5.56255758e-01 8.29147816e-01
2.96548218e-01 5.04790843e-01 -3.71828467e-01 -4.05421048e-01
5.10890663e-01 -1.53357834e-01 -7.59509450e-04 -6.62466705e-01
-1.58231747e+00 8.33807826e-01 3.79050851e-01 3.57130259e-01
-1.99743658e-01 3.88223857e-01 3.44927818e-01 6.32724881e-01
-8.65095779e-02 7.57809222e-01 -4.09627885e-01 -2.44857982e-01
-7.36295342e-01 3.69306207e-01 1.13099360e+00 3.87814105e-01
1.00437379e+00 -3.24133307e-01 2.14668036e-01 -1.77813452e-02
-5.55820093e-02 8.14970374e-01 1.22564703e-01 -1.00360310e+00
6.56756610e-02 2.57552862e-01 3.64086688e-01 -6.23420894e-01
-6.98087752e-01 -1.22964136e-01 -1.01592040e+00 -1.70229655e-02
7.91028142e-01 -1.90381482e-01 -5.37584484e-01 1.75285554e+00
2.90642768e-01 -2.81972170e-01 -2.82091945e-02 9.64093089e-01
1.19668759e-01 6.37758195e-01 -4.81193006e-01 -5.13022304e-01
1.16215372e+00 -5.00345826e-01 -5.03990173e-01 1.31712984e-02
8.64966333e-01 -5.26266515e-01 7.04452217e-01 4.12385672e-01
-9.12469506e-01 -1.83487564e-01 -1.13595951e+00 -1.21125691e-01
-2.75378764e-01 -2.48662114e-01 9.11750257e-01 6.06443107e-01
-8.79779398e-01 1.17543983e+00 -1.06816697e+00 -2.82290876e-01
-2.82768846e-01 7.24490047e-01 -5.96396446e-01 2.90861330e-03
-1.11455083e+00 9.69418168e-01 5.79649150e-01 1.19518124e-01
-8.17162156e-01 -2.84331530e-01 -4.42246974e-01 -7.47877359e-02
4.09896255e-01 -7.14373291e-01 1.14463151e+00 -2.44832993e-01
-1.64279139e+00 5.17485976e-01 -1.20479226e-01 -3.96489054e-01
1.02319866e-01 3.09598565e-01 -2.28622749e-01 4.47529089e-03
-1.33061841e-01 1.08640408e-03 6.19338036e-01 -5.42802870e-01
2.58445382e-01 -5.09360313e-01 2.89127976e-01 -2.24885747e-01
-6.94135129e-02 -2.15251297e-01 3.05411875e-01 2.42910847e-01
3.66488278e-01 -1.58745480e+00 -4.51380432e-01 -3.18658531e-01
-5.53988814e-01 1.00942649e-01 2.87515014e-01 -1.90975815e-01
1.12430251e+00 -1.99848378e+00 6.77393675e-01 5.12111545e-01
3.42763662e-01 3.88839357e-02 6.47776853e-03 8.95471215e-01
-1.33150786e-01 -3.34563479e-02 -2.44613409e-01 1.09992377e-01
1.60428822e-01 2.23759666e-01 8.75465944e-03 7.70459056e-01
1.33590996e-01 8.07866931e-01 -9.53825474e-01 -1.67814940e-01
2.99484711e-02 1.81012914e-01 -6.19647264e-01 -1.22333514e-02
-1.02976389e-01 8.18686485e-01 -5.12701213e-01 1.48182243e-01
3.51172477e-01 -7.06301332e-01 5.34049034e-01 -2.21322954e-01
-3.30326557e-01 5.63287795e-01 -1.30642986e+00 1.56597793e+00
-2.13720277e-01 4.43240494e-01 -5.07843196e-02 -9.89024937e-01
5.25252521e-01 2.74603099e-01 6.03703141e-01 -5.39777637e-01
4.60461050e-01 4.43509638e-01 7.30567515e-01 -4.38519925e-01
3.36472183e-01 -5.55645168e-01 -2.84809351e-01 9.03954387e-01
2.10216060e-01 -4.51959729e-01 6.73410594e-01 3.27946901e-01
1.51638448e+00 -3.59318584e-01 5.02407968e-01 -3.71894449e-01
4.73243475e-01 -1.58870965e-02 2.44414076e-01 8.30216944e-01
-2.06688866e-01 1.35799810e-01 7.99527407e-01 -6.41413331e-01
-1.64655900e+00 -8.45726728e-01 -5.58627725e-01 5.39929211e-01
1.43658981e-01 -6.90047026e-01 -8.15288603e-01 -4.81249951e-03
7.99367949e-02 3.14871550e-01 -5.98746955e-01 -3.84181172e-01
-4.40094411e-01 -1.10460973e+00 2.28663966e-01 -7.69401565e-02
2.13635311e-01 -8.80308509e-01 -3.65069568e-01 3.30355972e-01
2.64884353e-01 -1.17425549e+00 -2.84097433e-01 3.85812759e-01
-9.21980381e-01 -1.01200116e+00 1.14364903e-02 -7.33821541e-02
4.91740465e-01 -1.42063856e-01 8.99798095e-01 -1.18311375e-01
-5.91159582e-01 -3.78459990e-02 1.38146415e-01 2.48117834e-01
-6.93618953e-01 1.71151042e-01 4.98447806e-01 -5.78485429e-02
1.85165480e-01 -6.62822545e-01 -4.44931000e-01 -1.37531534e-01
-8.09820652e-01 -2.18829036e-01 4.50520962e-01 1.13420141e+00
3.90895367e-01 2.23204732e-01 -1.61789656e-01 -7.29720473e-01
5.90722442e-01 -3.65106612e-01 -8.94135296e-01 1.03003658e-01
-5.73264837e-01 7.61454046e-01 8.51697981e-01 -2.63316810e-01
-4.98116523e-01 1.98093265e-01 2.99199849e-01 1.63717777e-01
-5.61692864e-02 5.39339662e-01 3.41649979e-01 -4.54197466e-01
8.30031753e-01 2.64443308e-01 3.35450545e-02 -3.29807937e-01
4.11634177e-01 5.91693938e-01 2.76275218e-01 -7.56569684e-01
8.71005297e-01 4.83456820e-01 7.89608836e-01 -8.86157691e-01
-6.89688385e-01 -1.76592022e-01 -1.17205346e+00 1.41775757e-01
5.06686211e-01 -4.84566987e-01 -1.38400984e+00 4.41240102e-01
-8.43080819e-01 -2.10210443e-01 -1.66331276e-01 8.34427118e-01
-5.85920513e-01 3.42885762e-01 -8.91474485e-01 -9.32606459e-01
-1.43949747e-01 -1.20483077e+00 8.16306591e-01 1.48392795e-02
-4.27501574e-02 -1.02969503e+00 6.15541041e-01 1.80205360e-01
4.55605745e-01 1.01318136e-01 1.18146873e+00 -3.21252853e-01
-8.53386998e-01 -5.24751067e-01 -1.73965399e-03 2.54267037e-01
-1.37983680e-01 1.39794871e-01 -5.31349897e-01 -6.53637469e-01
6.26406074e-02 -5.33508718e-01 6.75156891e-01 8.33257139e-02
5.64665556e-01 -9.80017260e-02 -6.36662394e-02 2.81313658e-01
1.51036727e+00 -1.39139965e-01 2.40560174e-01 -1.02663428e-01
7.24365771e-01 2.04119623e-01 2.57486366e-02 3.11249316e-01
3.70255619e-01 6.87986553e-01 7.42603615e-02 5.63234746e-01
4.38786179e-01 1.18775867e-01 3.74522448e-01 1.29829466e+00
-3.68434012e-01 1.94957539e-01 -8.81969392e-01 2.22146794e-01
-1.70665979e+00 -1.04078114e+00 -3.79566729e-01 2.60199142e+00
9.38768685e-01 2.18577981e-01 2.56954730e-01 1.43274710e-01
3.32073212e-01 2.25947089e-02 -7.55294204e-01 -4.38423276e-01
-6.50066584e-02 4.34872329e-01 6.61163867e-01 5.86214662e-01
-9.57526386e-01 6.78684652e-01 7.18626165e+00 2.63852626e-01
-1.32668960e+00 1.04569435e-01 1.33659273e-01 -2.59711474e-01
-1.32095935e-02 5.87380528e-01 -5.17244041e-01 4.51403558e-01
1.44144642e+00 -2.82157153e-01 1.12904537e+00 3.76985252e-01
2.27369189e-01 -2.86644459e-01 -1.30845094e+00 9.09790218e-01
-5.21619439e-01 -1.15812898e+00 -5.46030879e-01 3.91893297e-01
7.69786298e-01 4.19746250e-01 -2.06583455e-01 1.10970419e-02
2.38600552e-01 -9.67936635e-01 3.12593609e-01 6.18764699e-01
5.10142267e-01 -5.60365140e-01 6.15748167e-01 6.24018669e-01
-6.82832658e-01 -1.66412979e-01 -4.20567036e-01 -5.28464437e-01
1.30271122e-01 8.16268921e-01 -4.21769440e-01 1.84614539e-01
1.82418272e-01 4.40541625e-01 -3.18003625e-01 7.64568925e-01
7.34613985e-02 6.44292533e-01 -7.14187205e-01 -2.90322095e-01
3.73953670e-01 -6.91035390e-01 4.41314548e-01 6.38509154e-01
1.25974953e-01 3.56078506e-01 1.58373028e-01 7.66561210e-01
-1.75047740e-02 -8.13412443e-02 -7.28759944e-01 -6.77644014e-01
2.89983332e-01 1.40247023e+00 -5.58078825e-01 -1.90722004e-01
-3.07360202e-01 6.44036949e-01 6.13916874e-01 1.25599846e-01
-3.15798461e-01 -3.54415417e-01 3.93022418e-01 -1.02994129e-01
3.40466589e-01 -8.47008526e-01 1.58804819e-01 -1.57399559e+00
8.32658783e-02 -7.63563514e-01 -5.67199849e-02 -2.55708307e-01
-1.24452662e+00 2.98103303e-01 -1.88564733e-01 -5.95450401e-01
-4.59439367e-01 -8.91387403e-01 -3.72930110e-01 7.73022652e-01
-1.04374528e+00 -4.71710980e-01 3.44396502e-01 4.11892593e-01
-5.62835872e-01 2.88708303e-02 1.16049898e+00 1.72469392e-01
-8.11604261e-01 6.43364936e-02 8.14191699e-01 3.10150553e-02
3.15928370e-01 -1.33580673e+00 4.61368471e-01 5.29484272e-01
3.56134623e-01 9.86446559e-01 1.06195438e+00 -2.33541787e-01
-2.21965337e+00 -3.63415927e-01 8.00292015e-01 -3.79306316e-01
1.28312206e+00 -6.88463151e-01 -7.78229356e-01 5.79743743e-01
-4.13250178e-02 5.58004200e-01 6.61200225e-01 3.73314768e-01
-3.96399081e-01 -3.00136320e-02 -1.05173910e+00 3.36185426e-01
8.52730393e-01 -1.00707638e+00 -2.63013631e-01 8.79334688e-01
1.35706007e-01 -2.72884279e-01 -1.28611565e+00 4.98103648e-02
5.96983492e-01 -8.24681401e-01 6.03472054e-01 -1.01107109e+00
1.03376992e-01 -4.11524326e-01 -1.77917406e-01 -1.13304210e+00
-4.20156270e-01 -9.47283208e-01 -2.45598286e-01 2.93465048e-01
4.31894839e-01 -5.98792791e-01 5.74425340e-01 6.86013043e-01
4.87806857e-01 -4.85681802e-01 -1.00337255e+00 -7.16181874e-01
2.57436365e-01 -1.06453918e-01 2.89568782e-01 8.34666014e-01
4.06684846e-01 7.52138555e-01 -4.68027025e-01 1.33869693e-01
8.54109168e-01 4.58570093e-01 7.17297912e-01 -1.10708761e+00
-7.50320613e-01 -1.48720622e-01 -7.40739763e-01 -7.67543793e-01
1.50715724e-01 -9.22638893e-01 -1.41714171e-01 -6.20670736e-01
6.69463336e-01 -3.58024985e-01 -3.15381616e-01 1.69328481e-01
1.41688362e-01 1.55814543e-01 1.21681932e-02 2.93266058e-01
-6.68787897e-01 3.73713136e-01 1.21974325e+00 -6.50570355e-03
8.06867033e-02 -1.83828026e-01 -1.58467606e-01 3.22206467e-01
5.57424128e-01 -6.81462407e-01 -7.88215995e-02 -1.19045347e-01
5.86522937e-01 3.41144800e-01 5.74962854e-01 -8.73589814e-01
3.11990857e-01 -2.31058940e-01 -1.50875613e-01 -1.73358768e-01
4.66110528e-01 -3.57325345e-01 3.13402921e-01 6.01812840e-01
-3.58681470e-01 -1.13044254e-01 -2.16286764e-01 3.93083066e-01
-8.25294703e-02 -2.71495044e-01 8.28666687e-01 -2.23542109e-01
-2.05347747e-01 2.54472405e-01 -2.39530206e-01 5.78955710e-02
7.11771727e-01 4.66369629e-01 -3.06214213e-01 -2.75291920e-01
-9.16235030e-01 -5.41978031e-02 6.33959889e-01 -1.25251144e-01
-6.47554025e-02 -1.23877144e+00 -4.82841402e-01 2.89080858e-01
1.13768391e-02 -5.08521378e-01 2.91200340e-01 1.08225465e+00
-5.42433619e-01 8.37652385e-01 -2.19630882e-01 -5.25057554e-01
-9.15916443e-01 6.47881329e-01 4.25976187e-01 -3.11469734e-01
-3.42026919e-01 4.25225198e-01 -3.12092036e-01 -6.95195019e-01
-5.35602510e-01 -3.41376096e-01 5.05608320e-01 -4.41039324e-01
3.62527966e-01 3.41855913e-01 1.88080385e-01 -9.05524135e-01
-2.26948217e-01 6.03878319e-01 -1.93540588e-01 -1.74537793e-01
1.28223205e+00 1.81973279e-01 -6.98507190e-01 7.79727876e-01
1.37384892e+00 -2.12468743e-01 -9.13030565e-01 -4.92553681e-01
-2.85512432e-02 -2.13460535e-01 1.50325179e-01 -3.83461922e-01
-6.27826273e-01 9.36715424e-01 4.50191408e-01 7.34627843e-01
5.36850274e-01 8.13778490e-03 7.02871978e-01 1.30317378e+00
7.20504761e-01 -9.53465223e-01 -2.28485256e-01 7.42127538e-01
2.06140876e-01 -1.41492760e+00 2.05589339e-01 8.09708387e-02
-1.68877821e-02 1.25504398e+00 -7.32114539e-02 -3.65541250e-01
6.51328564e-01 5.99709153e-02 -4.12853837e-01 -4.26742077e-01
-9.99776423e-01 9.06006470e-02 2.31305018e-01 -1.26453731e-02
4.38642710e-01 3.36535126e-01 -2.77097076e-01 -1.32570267e-01
-5.52042663e-01 -1.57019883e-01 8.06257904e-01 9.05722558e-01
-4.34800208e-01 -1.49731433e+00 -2.26222396e-01 4.71327305e-01
-3.17744225e-01 -1.21959560e-01 -3.21662694e-01 7.11882710e-01
-2.57718980e-01 7.21202791e-01 -1.94261596e-01 -2.11500823e-01
5.35317250e-02 1.71673596e-01 9.48559821e-01 -6.77179694e-01
-9.02793035e-02 -4.05233115e-01 -7.06771165e-02 -6.89924896e-01
-4.15222168e-01 -9.48197484e-01 -1.17344165e+00 -9.00053620e-01
-5.00216901e-01 2.82132536e-01 8.22413862e-01 1.34837747e+00
2.61027247e-01 6.65260404e-02 8.44151795e-01 -8.54226172e-01
-1.13153040e+00 -8.11012566e-01 -9.84339058e-01 3.65708888e-01
6.23997509e-01 -5.30547559e-01 -5.14592826e-01 -3.01780075e-01] | [5.645634651184082, 4.899710655212402] |
75d949f4-fc7b-4c1a-a9cb-5a4cbb452f15 | casia-face-africa-a-large-scale-african-face | 2105.03632 | null | https://arxiv.org/abs/2105.03632v2 | https://arxiv.org/pdf/2105.03632v2.pdf | CASIA-Face-Africa: A Large-scale African Face Image Database | Face recognition is a popular and well-studied area with wide applications in our society. However, racial bias had been proven to be inherent in most State Of The Art (SOTA) face recognition systems. Many investigative studies on face recognition algorithms have reported higher false positive rates of African subjects cohorts than the other cohorts. Lack of large-scale African face image databases in public domain is one of the main restrictions in studying the racial bias problem of face recognition. To this end, we collect a face image database namely CASIA-Face-Africa which contains 38,546 images of 1,183 African subjects. Multi-spectral cameras are utilized to capture the face images under various illumination settings. Demographic attributes and facial expressions of the subjects are also carefully recorded. For landmark detection, each face image in the database is manually labeled with 68 facial keypoints. A group of evaluation protocols are constructed according to different applications, tasks, partitions and scenarios. The performances of SOTA face recognition algorithms without re-training are reported as baselines. The proposed database along with its face landmark annotations, evaluation protocols and preliminary results form a good benchmark to study the essential aspects of face biometrics for African subjects, especially face image preprocessing, face feature analysis and matching, facial expression recognition, sex/age estimation, ethnic classification, face image generation, etc. The database can be downloaded from our http://www.cripacsir.cn/dataset/ | ['Zhenan Sun', 'Kunbo Zhang', 'Caiyong Wang', 'Yunlong Wang', 'Jawad Muhammad'] | 2021-05-08 | null | null | null | null | ['age-estimation', 'age-estimation'] | ['computer-vision', 'miscellaneous'] | [ 9.16804224e-02 -5.11421025e-01 -2.34117299e-01 -8.50807846e-01
-3.45633537e-01 -2.95664012e-01 4.67184901e-01 -4.29807991e-01
-3.11022818e-01 5.58289945e-01 -2.10455909e-01 4.99068573e-02
-5.13486750e-02 -6.73118591e-01 -1.00290671e-01 -1.11810768e+00
-6.03510849e-02 4.07635838e-01 -5.65590799e-01 -4.92329635e-02
4.76241231e-01 1.11332285e+00 -1.79986548e+00 1.26363561e-01
4.29084063e-01 1.01667142e+00 -3.69574994e-01 4.73293185e-01
-4.21920978e-02 4.55640666e-02 -5.66013098e-01 -6.18034720e-01
4.02108014e-01 -3.30925792e-01 -3.78870308e-01 2.02900916e-01
9.50242281e-01 -1.88685402e-01 2.71725729e-02 1.13919437e+00
9.89019454e-01 -5.50924353e-02 6.86501861e-01 -1.48749614e+00
-7.41107523e-01 4.47870567e-02 -1.14866984e+00 2.33024940e-01
2.46073723e-01 6.46443572e-03 1.93569720e-01 -1.28249586e+00
4.44839537e-01 1.49053729e+00 5.87246835e-01 8.64946842e-01
-8.46253276e-01 -1.47054708e+00 -3.82890850e-01 4.07884300e-01
-1.93650496e+00 -1.12941420e+00 8.23075473e-01 -4.03570294e-01
3.55447412e-01 3.08865786e-01 4.20845777e-01 7.07082331e-01
3.01827788e-02 -3.05984411e-02 1.52500069e+00 -4.55647588e-01
-2.50266165e-01 5.56217670e-01 6.85145659e-03 1.04187226e+00
1.01919129e-01 7.75822178e-02 -6.05026662e-01 -3.79684657e-01
6.32739067e-01 -8.84826109e-02 8.17417502e-02 8.17856845e-03
-5.11119008e-01 6.42990053e-01 1.42147496e-01 4.13091063e-01
-3.53225231e-01 -3.64812732e-01 4.37877327e-01 3.11945945e-01
3.48556042e-01 -4.14538294e-01 -1.70288965e-01 3.21111470e-01
-8.44680190e-01 1.06526919e-01 2.87599236e-01 7.01253891e-01
7.74195433e-01 2.04720959e-01 6.50382712e-02 1.29837692e+00
5.36235452e-01 7.85481274e-01 4.71773654e-01 -6.01562858e-01
7.54573494e-02 6.18577003e-01 -3.27997595e-01 -1.26488543e+00
-2.26045012e-01 -6.95157126e-02 -8.63198936e-01 3.33106875e-01
4.07802582e-01 -1.15298741e-01 -9.69967246e-01 1.57125592e+00
4.94796485e-01 2.49862745e-01 -1.27536850e-02 7.00850368e-01
8.84032071e-01 3.64986658e-01 3.18380654e-01 -5.36298215e-01
1.77460897e+00 -2.95493633e-01 -5.45722544e-01 1.63666487e-01
1.13155611e-01 -1.16966832e+00 6.45805240e-01 2.52286285e-01
-6.48483515e-01 -5.88512838e-01 -7.53678799e-01 3.06938171e-01
-4.81509179e-01 6.24989986e-01 6.24569595e-01 1.66229784e+00
-9.24451470e-01 -1.02218330e-01 -2.90041119e-01 -7.53889143e-01
9.88120139e-01 8.52805972e-01 -9.89483297e-01 -1.36454165e-01
-6.43783510e-01 7.31986105e-01 -1.68417823e-02 3.70539546e-01
-8.39129090e-01 -5.98051310e-01 -6.34826958e-01 -5.31779826e-01
8.38755444e-02 -1.73361808e-01 6.16539598e-01 -1.44075918e+00
-1.36607540e+00 1.70709944e+00 -6.41406298e-01 1.78945199e-01
2.02634513e-01 2.55835295e-01 -1.00379598e+00 5.39388368e-03
1.20804887e-02 6.02695405e-01 9.97832417e-01 -1.02053642e+00
-4.77640688e-01 -1.19007504e+00 -6.34542227e-01 -7.24970177e-02
-2.61977166e-01 8.66662800e-01 -1.45145774e-01 -4.19890404e-01
-4.63138781e-02 -9.45178807e-01 2.54263729e-01 4.50455584e-02
-8.02515596e-02 -3.57894838e-01 1.09312582e+00 -7.81288326e-01
9.31380987e-01 -2.23722005e+00 -3.22140396e-01 6.00804627e-01
-1.37792587e-01 2.82726169e-01 -8.34305361e-02 -1.55106867e-02
-4.43181664e-01 1.20581739e-01 -1.92639783e-01 -4.29099835e-02
-3.97545844e-01 -3.63781862e-02 1.26532882e-01 8.83682072e-01
2.55418807e-01 2.75272906e-01 -2.63712734e-01 -7.39975035e-01
2.23489478e-01 6.90444767e-01 -3.27796638e-01 1.52912438e-01
4.91723388e-01 5.63695133e-01 -3.77247751e-01 1.30717885e+00
1.20926332e+00 6.71849430e-01 1.54342681e-01 -3.36303055e-01
-7.12584183e-02 -7.33118773e-01 -1.22215199e+00 1.04986906e+00
-2.14775264e-01 4.75903392e-01 3.00960392e-01 -7.72477269e-01
1.19236553e+00 5.90083063e-01 5.50816357e-01 -3.57041329e-01
5.08842885e-01 3.57549578e-01 2.26375908e-01 -7.21226931e-01
6.17675520e-02 -3.61134470e-01 4.59340423e-01 2.85606772e-01
2.46710077e-01 2.36829355e-01 1.82506070e-01 -4.42402780e-01
1.73405960e-01 -9.48162824e-02 3.58886778e-01 -4.43822384e-01
1.16653657e+00 -2.28661552e-01 7.01600850e-01 8.67113695e-02
-6.22129798e-01 3.42356503e-01 2.60084867e-01 -5.83501816e-01
-7.72575796e-01 -7.65547752e-01 -6.68106377e-01 1.09398615e+00
-4.07628447e-01 -5.30048832e-02 -9.34778810e-01 -5.67476988e-01
2.75153853e-02 2.03317199e-02 -8.29497635e-01 8.45503435e-02
-5.44371426e-01 -1.24580026e+00 9.34722364e-01 1.16714887e-01
8.13226998e-01 -8.91064227e-01 -1.74182847e-01 -4.28799123e-01
3.72471750e-01 -7.29915738e-01 -5.22621810e-01 -6.41728640e-01
-5.19496143e-01 -1.77484155e+00 -8.10185611e-01 -9.46319044e-01
9.95389223e-01 6.63456693e-02 7.65218735e-01 2.56826967e-01
-8.56888175e-01 2.16254011e-01 1.17987655e-01 -7.87791014e-01
-2.46312544e-01 -3.96978259e-01 4.81412292e-01 6.66349947e-01
1.07166684e+00 -3.29886109e-01 -6.13373578e-01 6.32009804e-01
-6.14245236e-01 -6.70080960e-01 5.03368139e-01 6.90302610e-01
2.34434411e-01 4.79372330e-02 6.77080214e-01 -1.04712462e+00
4.12828803e-01 -4.80833173e-01 -6.28816843e-01 4.82419997e-01
-4.19073731e-01 -6.20768785e-01 1.55949607e-01 -1.67506993e-01
-1.34541261e+00 1.09115683e-01 -2.47410044e-01 -1.76666528e-01
-6.68516397e-01 -6.58038305e-03 -5.67583919e-01 -5.82510471e-01
6.36225760e-01 8.42233077e-02 4.23221618e-01 -2.77304828e-01
-8.15849453e-02 1.07589948e+00 5.01000345e-01 -7.28487611e-01
7.83271074e-01 4.28815454e-01 1.79641724e-01 -1.26377046e+00
-2.33914569e-01 -4.77480322e-01 -8.78234088e-01 -4.36769903e-01
6.76431656e-01 -7.87710965e-01 -1.02543259e+00 8.78235579e-01
-6.69554710e-01 5.16119361e-01 4.92961675e-01 4.06672895e-01
-1.07445568e-02 7.74010047e-02 -1.65232211e-01 -1.23714757e+00
-5.89369774e-01 -1.16833079e+00 6.74055994e-01 6.06754661e-01
1.05039760e-01 -6.62684083e-01 -9.14457664e-02 5.41273355e-01
5.08183300e-01 2.51747668e-01 8.86006892e-01 -4.40220088e-01
2.62545254e-02 -2.70899266e-01 -3.09445739e-01 3.41920257e-01
6.18031919e-01 5.99306643e-01 -1.46920121e+00 -1.99274227e-01
4.65680147e-03 -1.93252519e-01 4.44591671e-01 2.65199631e-01
1.14881003e+00 -9.91302580e-02 -1.48155317e-01 6.23993576e-01
1.34595442e+00 7.56959796e-01 6.71018302e-01 -2.32555136e-01
5.33867061e-01 1.08313084e+00 2.62557000e-01 4.38313335e-01
2.21982766e-02 5.97737730e-01 -9.37414169e-02 9.29901302e-02
-5.04719913e-02 1.22398570e-01 3.07822526e-01 2.70615965e-01
-6.69891775e-01 2.06931174e-01 -9.63810325e-01 2.15436861e-01
-1.07763124e+00 -1.15388763e+00 6.87499419e-02 2.16595817e+00
5.17866075e-01 -7.70102084e-01 4.86408442e-01 2.65910238e-01
1.10584235e+00 3.55339958e-04 -3.09755921e-01 -3.56193423e-01
-3.11195672e-01 5.24967492e-01 1.78547159e-01 2.15994045e-01
-1.08576417e+00 8.51861596e-01 6.16263199e+00 6.68498576e-01
-1.43962324e+00 2.21641093e-01 1.20028901e+00 -3.22691314e-02
5.03291309e-01 -3.86240989e-01 -1.01350427e+00 3.24774832e-01
9.09490287e-01 -1.57680973e-01 3.31578732e-01 7.72924125e-01
2.27269486e-01 -1.27259985e-01 -7.70816267e-01 1.36316693e+00
5.93394339e-01 -9.04515207e-01 1.28002867e-01 1.62084356e-01
6.34677112e-01 -4.88217384e-01 4.24002498e-01 -1.05604744e-02
-3.17520767e-01 -1.47028220e+00 1.71884984e-01 3.39820385e-01
1.20772052e+00 -9.99572039e-01 8.22880566e-01 -2.92220175e-01
-1.08931935e+00 -1.90849856e-01 -4.75648075e-01 2.97107786e-01
-3.23993742e-01 1.27133206e-01 -8.22349012e-01 3.93077046e-01
7.26590633e-01 3.69968116e-01 -7.57201850e-01 8.06448936e-01
3.43980819e-01 5.42589545e-01 -3.72340739e-01 2.57280141e-01
-2.14610100e-01 -5.04305601e-01 1.49125978e-01 9.00853932e-01
5.28346241e-01 2.83268839e-01 -1.90180823e-01 5.40050387e-01
-2.57057220e-01 7.71743536e-01 -7.73532927e-01 2.27932725e-02
5.17709553e-01 1.45224071e+00 -4.95847881e-01 -7.91030154e-02
-4.99606669e-01 4.12574172e-01 -2.03055128e-01 1.59398928e-01
-4.42882538e-01 -6.45764247e-02 9.33341324e-01 1.89472139e-01
-3.47306818e-01 1.86695009e-01 -2.68127006e-02 -7.37147510e-01
-1.44609615e-01 -1.20227659e+00 7.79881179e-01 -3.06860238e-01
-1.17387390e+00 6.45203650e-01 1.38358340e-01 -7.21102357e-01
6.00920469e-02 -9.52161849e-01 -6.71466708e-01 1.18879330e+00
-1.22025526e+00 -1.46691692e+00 -5.65525115e-01 9.04863417e-01
4.37194347e-01 -1.04046965e+00 1.00993359e+00 6.86282873e-01
-1.09113050e+00 9.14793134e-01 -2.37414062e-01 4.57047462e-01
9.53128994e-01 -4.80702609e-01 -3.68508786e-01 5.78131795e-01
-3.42112556e-02 8.47338378e-01 2.47218281e-01 -5.14137983e-01
-1.50744629e+00 -8.19983006e-01 6.27316952e-01 -4.78190780e-01
7.46667460e-02 -1.67134136e-01 -5.67124426e-01 5.04349411e-01
1.83765948e-01 1.56822041e-01 1.22615623e+00 1.24040961e-01
-4.57503498e-01 -6.56769216e-01 -1.77040946e+00 3.28724444e-01
7.44897664e-01 -4.99859482e-01 2.44899541e-02 2.42401600e-01
-6.27751112e-01 9.06480551e-02 -9.86931562e-01 4.36266601e-01
9.96217787e-01 -1.03104150e+00 8.70616138e-01 -5.58898866e-01
-2.27661221e-03 -2.97030777e-01 -2.58107543e-01 -7.50605404e-01
-4.72056121e-02 -4.32441175e-01 6.00019455e-01 1.65357828e+00
1.20582998e-01 -8.07459593e-01 9.60094333e-01 6.82337582e-01
4.96926486e-01 -4.40516979e-01 -1.00768983e+00 -3.23786348e-01
1.02271415e-01 -6.15408085e-03 7.67676592e-01 9.69038606e-01
-5.61861634e-01 3.02060675e-02 -3.49719644e-01 2.35766485e-01
9.44454253e-01 -7.14486092e-02 8.17415714e-01 -1.27394068e+00
4.18268621e-01 -5.46167016e-01 -6.71350121e-01 3.56586635e-01
4.25755650e-01 -7.65924573e-01 -4.96375501e-01 -5.77235401e-01
3.55480134e-01 -6.50630116e-01 -2.12912962e-01 4.42071408e-01
1.26398623e-01 8.95235538e-01 -9.92339104e-02 4.76305783e-02
4.24624056e-01 1.17948659e-01 9.15278256e-01 -2.56114323e-02
4.57255132e-02 1.05796158e-01 -6.30165637e-01 6.79129899e-01
9.61754978e-01 -1.78470522e-01 -3.46582472e-01 -6.47689700e-02
-5.33385396e-01 -2.39489794e-01 2.08909705e-01 -8.40560734e-01
9.94618535e-02 -4.90044713e-01 9.75128293e-01 -3.39788646e-01
4.61756647e-01 -8.56063485e-01 5.83576322e-01 2.78449476e-01
4.88025360e-02 2.96376228e-01 2.33475775e-01 -7.12177604e-02
-1.70136273e-01 -2.89052099e-01 1.25744462e+00 -2.35833421e-01
-7.37201035e-01 7.72918940e-01 -7.01134801e-02 -2.97490776e-01
1.37436831e+00 -4.33115214e-01 -2.59985596e-01 9.86441225e-02
-4.58457559e-01 -2.21076936e-01 4.23626065e-01 4.90825444e-01
6.15110755e-01 -1.36244810e+00 -1.11109221e+00 8.67668867e-01
2.92253047e-01 -7.15975821e-01 5.13971925e-01 8.55344355e-01
-6.30161107e-01 3.60689461e-01 -9.23790574e-01 -3.55851769e-01
-2.07466054e+00 3.01736802e-01 4.92428184e-01 6.18500650e-01
9.31422263e-02 8.64540935e-01 7.40315467e-02 -3.07037473e-01
2.37193126e-02 6.23773515e-01 -5.54401398e-01 2.75965005e-01
7.99701810e-01 4.35297877e-01 -3.35254483e-02 -1.51838362e+00
-6.65093780e-01 1.00924122e+00 -9.80438888e-02 1.62243515e-01
1.13509762e+00 1.13665365e-01 -6.76599562e-01 2.19446085e-02
1.22497690e+00 2.31902689e-01 -4.25323844e-01 -1.57747343e-02
-2.58560814e-02 -1.09165883e+00 -1.71660542e-01 -6.70605838e-01
-1.44561243e+00 8.40014815e-01 1.29950809e+00 -5.20039439e-01
1.27591658e+00 -3.31005931e-01 1.02598339e-01 1.31160855e-01
4.09067303e-01 -9.45644498e-01 -5.22750258e-01 -3.55916619e-02
9.83561218e-01 -1.50495303e+00 -5.88684417e-02 -5.47665834e-01
-4.57942009e-01 1.14842629e+00 7.86779344e-01 1.07699297e-01
9.64988768e-01 1.47124812e-01 5.68602383e-01 -1.62236214e-01
-3.28864455e-01 9.13013443e-02 2.74851143e-01 8.90354037e-01
8.28613281e-01 3.77443433e-02 -1.97952047e-01 3.61645699e-01
-3.11464757e-01 3.68154868e-02 9.24554188e-03 6.29954278e-01
-1.42279670e-01 -1.25955868e+00 -9.21665251e-01 5.72484851e-01
-9.24793780e-01 2.37162903e-01 -4.37466711e-01 6.54904127e-01
4.29452896e-01 8.97162497e-01 1.23218028e-02 -1.84386715e-01
1.30081326e-01 4.87353414e-01 7.52926648e-01 -3.88436615e-01
-6.77451789e-01 -1.13907754e-01 8.35380610e-03 -1.24723770e-01
-7.21064806e-01 -7.98794985e-01 -7.95658469e-01 -7.02678800e-01
4.56569493e-02 2.17773050e-01 9.83202755e-01 5.51828206e-01
1.67536110e-01 -3.50248843e-01 8.19037974e-01 -6.44101918e-01
-1.13028720e-01 -1.07769060e+00 -7.17211902e-01 3.76342684e-01
-1.13098333e-02 -8.08543026e-01 -1.12769663e-01 2.24357784e-01] | [13.259722709655762, 0.8940135836601257] |
afaa2aa1-d59d-490c-b9c1-d4348a233aeb | robust-a-optimal-experimental-design-for | 2305.03855 | null | https://arxiv.org/abs/2305.03855v1 | https://arxiv.org/pdf/2305.03855v1.pdf | Robust A-Optimal Experimental Design for Bayesian Inverse Problems | Optimal design of experiments for Bayesian inverse problems has recently gained wide popularity and attracted much attention, especially in the computational science and Bayesian inversion communities. An optimal design maximizes a predefined utility function that is formulated in terms of the elements of an inverse problem, an example being optimal sensor placement for parameter identification. The state-of-the-art algorithmic approaches following this simple formulation generally overlook misspecification of the elements of the inverse problem, such as the prior or the measurement uncertainties. This work presents an efficient algorithmic approach for designing optimal experimental design schemes for Bayesian inverse problems such that the optimal design is robust to misspecification of elements of the inverse problem. Specifically, we consider a worst-case scenario approach for the uncertain or misspecified parameters, formulate robust objectives, and propose an algorithmic approach for optimizing such objectives. Both relaxation and stochastic solution approaches are discussed with detailed analysis and insight into the interpretation of the problem and the proposed algorithmic approach. Extensive numerical experiments to validate and analyze the proposed approach are carried out for sensor placement in a parameter identification problem. | ['Todd Munson', 'Sven Leyffer', 'Ahmed Attia'] | 2023-05-05 | null | null | null | null | ['experimental-design'] | ['methodology'] | [ 5.02632022e-01 2.51630723e-01 1.96536735e-01 -3.55172038e-01
-8.21456075e-01 -4.42750663e-01 1.30446956e-01 -1.50441661e-01
-3.15592527e-01 8.67137730e-01 1.26974776e-01 -2.55200028e-01
-1.40295148e+00 -3.98280323e-01 -7.17953622e-01 -1.20903254e+00
1.87128082e-01 6.61340415e-01 -2.73679316e-01 1.10297039e-01
5.46973705e-01 4.74246532e-01 -1.32256508e+00 -7.34150112e-01
6.97524548e-01 1.01826572e+00 4.26516563e-01 4.07823801e-01
5.35399735e-01 -3.91243026e-02 -3.40321779e-01 2.21947894e-01
2.69489974e-01 -2.79105902e-01 -2.95268774e-01 3.56751174e-01
-4.62214828e-01 -5.85006289e-02 1.41150683e-01 1.24721229e+00
8.08849216e-01 3.42363447e-01 9.21832442e-01 -1.06918478e+00
-2.62219340e-01 5.84678710e-01 -7.58859932e-01 1.81835387e-02
1.83615219e-02 -7.13657439e-02 8.20573330e-01 -8.68602455e-01
5.31869419e-02 1.14350271e+00 5.50588548e-01 -1.26485571e-01
-1.31868351e+00 -1.28169581e-01 -1.29324302e-01 8.27298239e-02
-1.76373017e+00 -6.12061918e-01 9.59312499e-01 -5.85272789e-01
1.85804471e-01 1.93114176e-01 1.52460948e-01 5.53629816e-01
2.02388212e-01 2.43742839e-01 1.21539450e+00 -5.92581809e-01
6.78735018e-01 4.19344939e-02 4.10012960e-01 3.11598301e-01
8.60141397e-01 3.16669017e-01 -8.79039317e-02 -2.96010643e-01
4.89155650e-01 -2.48305097e-01 -3.52509409e-01 -5.62440813e-01
-1.00122499e+00 8.49149764e-01 3.71077172e-02 1.71813164e-02
-7.12073803e-01 1.27824321e-01 -2.40085963e-02 -7.35931993e-02
2.58002728e-01 4.31192070e-01 -3.77272338e-01 2.05725566e-01
-8.15109909e-01 4.82166767e-01 7.23803043e-01 9.27382410e-01
6.63265049e-01 1.46528885e-01 4.89582680e-02 7.31646478e-01
7.84210980e-01 9.50476766e-01 -1.29545599e-01 -1.16556072e+00
4.24118578e-01 -9.74275619e-02 6.21861279e-01 -1.17057860e+00
-3.37600350e-01 -7.49663472e-01 -5.81991792e-01 -3.60774174e-02
3.40354890e-01 -5.22858024e-01 -7.42125869e-01 1.64600682e+00
5.72297215e-01 2.74234921e-01 2.67105430e-01 9.49204445e-01
1.73427671e-01 6.36789262e-01 -2.69408584e-01 -6.27469659e-01
1.39412093e+00 4.35339324e-02 -1.08147919e+00 -4.02254909e-01
1.64997369e-01 -9.87658978e-01 5.02147138e-01 4.06690210e-01
-8.95020127e-01 -1.33199394e-01 -1.25924242e+00 5.56219697e-01
2.03905568e-01 2.94769615e-01 1.83671825e-02 7.62150705e-01
-5.76625824e-01 4.25887704e-01 -6.48288369e-01 -1.05035074e-01
-1.74929217e-01 5.45838714e-01 3.09275091e-01 -1.08631931e-01
-1.04441416e+00 8.98351133e-01 4.41547662e-01 8.49763095e-01
-8.36905599e-01 -9.03368354e-01 -7.46935248e-01 -1.66919716e-02
7.84219027e-01 -6.56888366e-01 1.16579354e+00 -2.59308964e-01
-1.60054529e+00 2.42617756e-01 -4.32646684e-02 -9.46984813e-02
3.50971341e-01 -9.82281864e-02 -2.66862899e-01 8.07101198e-04
1.51110247e-01 -2.57849813e-01 1.01936638e+00 -1.40335512e+00
-2.13811666e-01 -5.58362961e-01 -1.12584151e-01 1.33403884e-02
3.31414461e-01 -1.87767640e-01 -8.78887996e-02 -5.55203736e-01
7.30605423e-01 -1.05996084e+00 -5.78367233e-01 -3.33629012e-01
-5.39673805e-01 2.49465033e-01 8.01517367e-01 -7.03264594e-01
1.09208930e+00 -1.90072906e+00 3.99744958e-01 8.26379061e-01
-3.39778244e-01 -2.81974852e-01 2.73191959e-01 4.09747273e-01
-1.38239101e-01 -1.89517319e-01 -7.53124535e-01 -1.14589110e-01
1.85579523e-01 3.31177890e-01 -1.01326756e-01 1.08372402e+00
-2.09941342e-01 2.39414692e-01 -6.11221969e-01 -2.60254502e-01
5.41260958e-01 4.59274501e-01 -4.77615207e-01 1.97071895e-01
-1.69542909e-01 6.22806013e-01 -1.07263052e+00 4.96056795e-01
7.75815308e-01 7.49500096e-02 1.67190015e-01 -5.12193978e-01
-2.72335023e-01 -2.55877346e-01 -1.97799373e+00 1.12320435e+00
-6.21204734e-01 6.83584139e-02 6.30411506e-01 -1.53254294e+00
9.30382967e-01 4.26091224e-01 7.05732763e-01 -9.65841934e-02
3.96089017e-01 4.49211866e-01 8.46481845e-02 -6.49073362e-01
4.64320689e-01 -5.71847200e-01 -2.65281260e-01 4.13089544e-01
-3.09527755e-01 -4.76902843e-01 -6.15988038e-02 -1.99778959e-01
5.55671036e-01 -2.04405133e-02 6.47267222e-01 -6.81463659e-01
7.32128143e-01 -3.77289593e-01 6.23657525e-01 7.33002067e-01
1.53854430e-01 5.55891335e-01 3.01848173e-01 2.67879367e-01
-9.78555799e-01 -6.26915872e-01 -4.32825178e-01 2.20878795e-01
2.51936197e-01 3.86478335e-01 -5.68192124e-01 2.10085422e-01
1.26506567e-01 8.89328241e-01 -5.10800004e-01 1.94754116e-02
-5.60114980e-01 -1.19486332e+00 -5.69837652e-02 1.13181442e-01
3.67119372e-01 -3.46915513e-01 -7.48640776e-01 3.73159915e-01
-2.28439540e-01 -1.00757277e+00 7.36946007e-03 9.49211866e-02
-8.47628176e-01 -1.01999640e+00 -6.21209919e-01 -9.65806842e-02
8.93216252e-01 4.96130846e-02 6.42504275e-01 -3.62744153e-01
-1.57550007e-01 7.54374623e-01 -2.38593578e-01 -4.72005278e-01
-2.69149572e-01 -4.18863922e-01 2.16103896e-01 4.65776980e-01
-2.94890881e-01 -6.48760140e-01 -5.53675592e-01 6.33180201e-01
-9.17433619e-01 -5.78751564e-01 5.48222423e-01 7.24571526e-01
6.94865346e-01 3.89640659e-01 7.39848256e-01 -5.70048988e-01
6.76096618e-01 -7.67077088e-01 -1.13106799e+00 3.05398494e-01
-5.63550591e-01 5.40433288e-01 6.26213551e-02 -3.50316204e-02
-1.52796137e+00 6.80362284e-02 9.29091685e-03 -7.45714158e-02
5.95370494e-02 9.77564037e-01 -6.84275389e-01 -2.67104238e-01
3.01172823e-01 -8.74908343e-02 -2.04124246e-02 -6.76106513e-01
2.55596757e-01 7.88763702e-01 4.82443303e-01 -1.19433427e+00
7.68365443e-01 5.37720501e-01 5.61535418e-01 -1.18947542e+00
-9.17055070e-01 -5.32242835e-01 -2.84263998e-01 -4.07590419e-01
5.88513613e-01 -5.63294828e-01 -7.54865646e-01 2.08505690e-01
-9.39687073e-01 3.94890875e-01 -2.67469585e-01 8.02769184e-01
-8.92662346e-01 4.81253296e-01 3.55717629e-01 -1.42148721e+00
1.61808744e-01 -1.50848687e+00 1.06629729e+00 1.92003995e-01
-1.71868160e-01 -1.12848616e+00 1.30589381e-02 3.08039695e-01
2.04903334e-01 3.56336415e-01 6.67189896e-01 -3.57370943e-01
-6.70798123e-01 -3.34072500e-01 7.55536631e-02 1.81037709e-01
-1.15951039e-01 -2.82746613e-01 -6.93780184e-01 -2.07749948e-01
6.26872599e-01 2.02646405e-01 2.72850364e-01 1.08791816e+00
8.90391469e-01 -1.70992807e-01 -3.97136241e-01 4.46117073e-01
1.70750177e+00 2.76246935e-01 3.58986795e-01 2.05607846e-01
1.93692043e-01 8.06595147e-01 8.52070689e-01 1.02082098e+00
4.55443747e-03 8.67110431e-01 6.01455569e-01 3.55261475e-01
6.66006982e-01 1.58190846e-01 -1.33981675e-01 5.29586315e-01
1.19479606e-03 -3.64683867e-01 -6.82231784e-01 6.72411084e-01
-1.85605860e+00 -6.05808675e-01 -1.57305717e-01 2.43767834e+00
2.83388704e-01 -2.95651883e-01 -4.50111061e-01 4.42998528e-01
1.08543491e+00 7.99580812e-02 -5.15948057e-01 -1.65428162e-01
-1.30862957e-02 -5.53537905e-02 9.71327603e-01 7.32596099e-01
-7.34605789e-01 4.14113812e-02 6.28443527e+00 7.37829804e-01
-6.24766767e-01 5.78563958e-02 2.33415797e-01 3.00397277e-01
-4.88985747e-01 4.35164481e-01 -8.46406877e-01 4.69688177e-01
7.71425724e-01 -2.99704015e-01 3.38447005e-01 4.84672427e-01
9.15829241e-01 -5.78224957e-01 -8.59354496e-01 6.80153787e-01
-1.54957265e-01 -9.64862943e-01 -5.31099141e-01 2.29670644e-01
9.40514982e-01 -4.05337870e-01 1.25685379e-01 -2.32231110e-01
2.13188916e-01 -6.23882890e-01 8.82843018e-01 7.95172155e-01
2.16753706e-01 -7.11923301e-01 8.21120739e-01 3.73230994e-01
-8.03051293e-01 -3.43649179e-01 -1.40172496e-01 -1.29446879e-01
7.52156019e-01 1.08084202e+00 -5.90239108e-01 8.52740407e-01
4.05805916e-01 3.18766862e-01 8.78360197e-02 1.13864350e+00
-1.40018627e-01 6.16152823e-01 -8.25069547e-01 1.06408279e-02
2.17325166e-01 -6.63431168e-01 1.03816736e+00 6.55223846e-01
7.79421508e-01 4.45456415e-01 -2.27123275e-02 1.00631654e+00
5.58235824e-01 -9.62342173e-02 -4.63493586e-01 -7.02229589e-02
7.04417050e-01 7.81629324e-01 -6.45874321e-01 1.25067994e-01
-1.58007547e-01 -1.40176460e-01 -4.37173843e-01 7.87953138e-01
-7.59437978e-01 -1.38889223e-01 5.00650883e-01 -3.39018367e-02
5.56954384e-01 -2.70769894e-01 -5.60851336e-01 -6.26356304e-01
2.75882572e-01 -5.87894082e-01 4.12897289e-01 -5.47872543e-01
-9.01135683e-01 -1.44486487e-01 7.86296308e-01 -1.12834549e+00
-5.01633584e-01 -5.75771034e-01 -3.96365732e-01 8.96509469e-01
-1.22057414e+00 -7.48480737e-01 1.40880555e-01 2.02744827e-01
2.32088551e-01 -2.19922606e-03 2.50651956e-01 1.23404257e-01
-6.85847700e-01 2.58001722e-02 8.62819791e-01 -5.76760650e-01
1.61962926e-01 -5.96443653e-01 -6.11314952e-01 1.07197857e+00
-5.71669579e-01 7.18343675e-01 1.66603994e+00 -6.87979043e-01
-1.67152739e+00 -7.36475408e-01 5.94606400e-01 1.30579859e-01
8.94606590e-01 3.10566705e-02 -4.66077268e-01 6.23391151e-01
-1.26766473e-01 -6.89080060e-02 2.59266287e-01 -1.37531668e-01
3.35089594e-01 -1.86490729e-01 -1.33366215e+00 4.16292220e-01
4.66250271e-01 -1.19625106e-01 -5.56266785e-01 2.81457037e-01
2.17675507e-01 -4.10204649e-01 -9.83670950e-01 7.89285302e-01
5.90592921e-01 -5.44572413e-01 1.13678360e+00 -1.28643557e-01
1.67197995e-02 -6.80210769e-01 -6.23168111e-01 -1.38059974e+00
1.09298393e-01 -7.26867676e-01 1.19006827e-01 1.14146936e+00
2.72096992e-01 -6.79156005e-01 6.50728941e-01 6.98639035e-01
-1.73554584e-01 -4.45868343e-01 -1.23571503e+00 -8.06253254e-01
-2.94865936e-01 -5.86272597e-01 2.94875592e-01 3.37493896e-01
-4.05761242e-01 7.84279481e-02 -5.50760150e-01 8.13740611e-01
1.23859870e+00 1.10801749e-01 4.61632460e-01 -1.05103183e+00
-4.92837816e-01 2.85248049e-02 -1.82826191e-01 -9.85237956e-01
9.57134962e-02 -2.77956575e-01 6.93978965e-01 -1.25967276e+00
-1.14106514e-01 -5.42519987e-01 -4.95939478e-02 -4.46320474e-01
-1.34748286e-02 -2.08982006e-01 -1.24954849e-01 -1.21122770e-01
1.59835964e-02 7.41936743e-01 9.11026418e-01 -5.05437218e-02
8.51243641e-03 4.58949834e-01 -6.75278425e-01 7.98510551e-01
4.80056643e-01 -7.96941757e-01 -7.61501431e-01 -4.21897292e-01
3.86843622e-01 6.66885257e-01 3.86620969e-01 -8.11693907e-01
1.06148258e-01 -3.13799590e-01 -2.28589267e-01 -7.36449182e-01
5.05855322e-01 -1.34768832e+00 8.32814693e-01 2.79516637e-01
-1.10723950e-01 -1.10349141e-01 -1.45017743e-01 9.47239816e-01
-1.16393648e-01 -9.84816432e-01 9.79525864e-01 5.14090918e-02
-4.28113848e-01 -1.65318087e-01 -4.34180677e-01 -2.66291611e-02
9.16346669e-01 -1.75271824e-01 1.81495965e-01 -6.35479689e-01
-1.08387315e+00 3.32910180e-01 -1.17844325e-02 -8.87786224e-02
4.62141573e-01 -9.06784832e-01 -7.92604089e-01 -1.00918338e-01
-1.62195593e-01 -1.45654961e-01 5.01158357e-01 1.10292685e+00
-2.59025127e-01 5.12914121e-01 1.76113501e-01 -7.56945550e-01
-8.33162248e-01 2.57127345e-01 4.57127094e-01 -7.91018829e-02
-1.10796496e-01 5.26369750e-01 5.52421622e-02 -4.15383846e-01
3.04128072e-04 -2.33654812e-01 -5.12030236e-02 5.89199886e-02
4.35694009e-02 7.88371623e-01 9.40301940e-02 -7.91034400e-01
-2.91333109e-01 9.30835843e-01 5.92438161e-01 -5.96631825e-01
1.37761748e+00 -7.58087575e-01 -2.13483348e-01 2.88962007e-01
1.00969732e+00 -1.44273475e-01 -9.97949004e-01 -4.24193323e-01
2.42845640e-01 -3.15995753e-01 4.64142710e-01 -4.40952718e-01
-8.39681268e-01 4.06326801e-01 3.86018395e-01 -3.08998358e-02
1.01115239e+00 -2.18834043e-01 1.69693172e-01 5.28300822e-01
3.94764423e-01 -1.25100124e+00 -3.34567010e-01 4.09703523e-01
1.06996691e+00 -9.87351298e-01 5.56748927e-01 -5.15150189e-01
-1.13333665e-01 8.12006116e-01 2.69635078e-02 -1.66276082e-01
1.18093133e+00 3.86934608e-01 -1.98681831e-01 -8.48554149e-02
-1.11312792e-01 8.17871243e-02 1.28756106e-01 3.04862261e-01
1.62191689e-01 8.54276791e-02 -7.87455797e-01 7.29025960e-01
6.83673397e-02 -2.12108523e-01 5.92838287e-01 1.12513709e+00
-5.49506605e-01 -8.27443302e-01 -1.20324004e+00 1.79976627e-01
-3.45183820e-01 1.59262255e-01 2.86801428e-01 8.39335859e-01
-7.26783574e-02 1.16134799e+00 -3.16536337e-01 1.84490114e-01
5.58879495e-01 -2.03420684e-01 4.05892819e-01 -2.42314950e-01
1.28223330e-01 3.89874071e-01 2.03750074e-01 -1.88634351e-01
-6.39914930e-01 -1.06435525e+00 -1.00194848e+00 1.53628767e-01
-6.80159867e-01 7.40322590e-01 1.03994250e+00 1.31274998e+00
7.06512406e-02 5.35258412e-01 6.47792816e-01 -9.87748384e-01
-1.10071623e+00 -9.33794856e-01 -9.69281197e-01 -2.93425769e-01
3.82426560e-01 -1.16434109e+00 -7.47380316e-01 -6.21353798e-02] | [6.531422138214111, 3.516587257385254] |
45bed365-9faf-4a02-adf5-43032359d0bb | target-sound-extraction-with-variable-cross | 2303.08372 | null | https://arxiv.org/abs/2303.08372v1 | https://arxiv.org/pdf/2303.08372v1.pdf | Target Sound Extraction with Variable Cross-modality Clues | Automatic target sound extraction (TSE) is a machine learning approach to mimic the human auditory perception capability of attending to a sound source of interest from a mixture of sources. It often uses a model conditioned on a fixed form of target sound clues, such as a sound class label, which limits the ways in which users can interact with the model to specify the target sounds. To leverage variable number of clues cross modalities available in the inference phase, including a video, a sound event class, and a text caption, we propose a unified transformer-based TSE model architecture, where a multi-clue attention module integrates all the clues across the modalities. Since there is no off-the-shelf benchmark to evaluate our proposed approach, we build a dataset based on public corpora, Audioset and AudioCaps. Experimental results for seen and unseen target-sound evaluation sets show that our proposed TSE model can effectively deal with a varying number of clues which improves the TSE performance and robustness against partially compromised clues. | ['Michael Zeng', 'Yanmin Qian', 'Shujie Liu', 'Takuya Yoshioka', 'Dongmei Wang', 'Zhuo Chen', 'Yao Qian', 'Chenda Li'] | 2023-03-15 | null | null | null | null | ['target-sound-extraction'] | ['audio'] | [ 3.01564693e-01 -4.31990713e-01 2.54975736e-01 -2.05603153e-01
-1.44922614e+00 -7.90642142e-01 4.97392595e-01 1.79350656e-02
-3.67118657e-01 2.87154645e-01 3.99092108e-01 -5.42375520e-02
8.13556537e-02 -4.57120836e-01 -6.65286362e-01 -4.23720419e-01
5.30408248e-02 1.48437396e-01 9.10226166e-01 9.30583999e-02
2.20361993e-01 -4.94624861e-02 -1.86756659e+00 8.38921130e-01
2.59662151e-01 1.32571590e+00 4.48276490e-01 1.11773980e+00
5.39473407e-02 9.84361768e-01 -6.86867654e-01 -1.88160911e-01
1.44348800e-01 -4.28637117e-01 -4.41193730e-01 -4.18453932e-01
3.49425763e-01 -1.42109439e-01 -2.35891864e-01 9.03563857e-01
1.08121455e+00 1.22541301e-01 5.99040151e-01 -1.52514970e+00
-2.87916929e-01 9.46534216e-01 -1.23103879e-01 6.35693371e-01
7.59218812e-01 4.31495607e-01 1.21866107e+00 -1.15848529e+00
2.38337144e-02 1.26590896e+00 5.81482947e-01 5.00212729e-01
-7.91318059e-01 -1.06861484e+00 6.29912317e-02 5.96139610e-01
-1.25729847e+00 -7.88244128e-01 1.01691854e+00 -3.77784520e-01
6.95890784e-01 5.59495330e-01 1.44626871e-01 1.32347238e+00
-3.21642309e-01 6.76666915e-01 9.90314245e-01 -5.45120776e-01
2.86571532e-01 2.30519131e-01 1.82029903e-01 2.76024908e-01
-5.87075591e-01 1.53076693e-01 -1.25851631e+00 -4.49238151e-01
2.23341703e-01 -5.69973350e-01 -5.72504342e-01 2.73365438e-01
-1.24783921e+00 3.25140417e-01 4.11242135e-02 1.00750208e-01
-1.96330622e-01 -3.55274864e-02 5.43795764e-01 2.11459026e-01
8.59473944e-02 1.15619339e-01 -5.17687023e-01 -2.69810468e-01
-6.51986063e-01 1.56305209e-02 8.35795045e-01 6.94518089e-01
3.35749567e-01 1.02712512e-01 -5.07261276e-01 1.08788049e+00
5.41507065e-01 5.38055956e-01 4.89983082e-01 -6.44210935e-01
6.96810365e-01 1.16758011e-01 1.90678716e-01 -7.66452014e-01
-5.39075118e-03 -3.56352359e-01 -2.12055355e-01 -2.09062770e-02
2.68810600e-01 3.67069826e-03 -8.11147571e-01 2.06831574e+00
4.24122810e-01 9.46314514e-01 -9.91341695e-02 1.06932473e+00
1.10970199e+00 7.13248253e-01 1.50664806e-01 -1.10464133e-01
1.56529641e+00 -8.87606144e-01 -5.86406887e-01 -1.67638391e-01
-1.18808851e-01 -1.12349796e+00 1.34641349e+00 6.21623814e-01
-9.20542896e-01 -9.08560395e-01 -9.56508577e-01 1.53962225e-01
-1.85949773e-01 2.29421351e-02 8.05097595e-02 5.24407625e-01
-7.97658503e-01 -8.58452469e-02 -6.63073063e-01 -2.31074572e-01
6.15158901e-02 1.50401458e-01 -1.96584493e-01 -8.82860795e-02
-1.40161610e+00 5.25759339e-01 3.34975541e-01 1.02946334e-01
-1.53766310e+00 -6.16009474e-01 -5.38041174e-01 3.77927154e-01
5.35030425e-01 -5.34502625e-01 1.46173894e+00 -6.54065311e-01
-1.44111693e+00 4.99177128e-01 -1.16963692e-01 -1.88233733e-01
2.67266899e-01 -3.17598432e-01 -8.78022254e-01 3.65011871e-01
4.84386506e-03 5.41486442e-01 1.13859963e+00 -1.26372671e+00
-7.87024915e-01 5.53794689e-02 9.55184996e-02 3.41520339e-01
-1.83434352e-01 6.63285792e-01 -6.31073713e-01 -5.08679748e-01
-2.09679574e-01 -5.70917547e-01 3.39768320e-01 -2.17982724e-01
-7.74033487e-01 -2.06596062e-01 7.37890005e-01 -5.37698925e-01
1.35855758e+00 -2.43094635e+00 -3.61712337e-01 5.43726161e-02
-3.50021720e-02 2.31508613e-01 -3.17916930e-01 4.38474029e-01
-8.71843845e-02 -2.24276364e-01 1.83091387e-02 -4.21284646e-01
2.80858278e-01 -2.09013838e-02 -6.20300651e-01 -4.15648408e-02
7.12234378e-02 1.89433888e-01 -8.86583030e-01 -5.80274343e-01
-3.27062421e-02 4.09411818e-01 -4.65090305e-01 7.97674835e-01
-2.84664750e-01 1.79355651e-01 -4.54013824e-01 7.07061768e-01
4.09145772e-01 1.83453619e-01 -4.11765754e-01 -1.80827782e-01
3.35757509e-02 6.54567301e-01 -1.53463399e+00 1.52747738e+00
-3.87554586e-01 3.01691681e-01 1.57015860e-01 -4.24832821e-01
6.56228065e-01 9.05053616e-01 -5.76663613e-02 -2.74505466e-01
1.03876144e-01 1.55375391e-01 2.27954894e-01 -7.41701305e-01
-4.74807946e-03 9.60097928e-03 -9.97788683e-02 5.28437555e-01
5.35048544e-01 2.18053177e-01 9.54509750e-02 2.16273904e-01
1.37014914e+00 -7.76871741e-02 -3.07896119e-02 2.31605083e-01
5.77813983e-01 -3.66197228e-01 5.92912197e-01 1.03855002e+00
-2.48704374e-01 8.74035299e-01 3.26852977e-01 7.18853762e-03
-5.26385427e-01 -1.19067585e+00 6.75256103e-02 1.66002202e+00
-1.59744486e-01 -5.65911353e-01 -7.31990159e-01 -4.05151069e-01
-4.93743896e-01 6.91141725e-01 -4.98512268e-01 -1.50057137e-01
-2.53954142e-01 -5.00960171e-01 8.81195486e-01 4.19032782e-01
3.03780347e-01 -1.31148243e+00 -5.54146111e-01 2.57141531e-01
-5.77313185e-01 -1.15725088e+00 -9.11195576e-01 2.49083608e-01
5.02736717e-02 -1.10846186e+00 3.70487198e-02 -6.88470066e-01
2.69129630e-02 1.94126353e-01 1.14964426e+00 -1.14191204e-01
-1.40292808e-01 6.29638135e-01 -5.69340885e-01 -6.28163695e-01
-5.55432141e-01 -4.65352267e-01 1.05172880e-01 5.63949525e-01
3.49656850e-01 -8.74513924e-01 -5.53333342e-01 4.46126759e-01
-8.08927536e-01 -9.10697281e-02 4.20238227e-01 6.05868816e-01
3.30152720e-01 -1.35570290e-02 7.99180925e-01 -4.25154388e-01
6.20427549e-01 -7.98581898e-01 -1.79969877e-01 2.81119406e-01
4.56712395e-02 -3.06271225e-01 6.33409083e-01 -1.09775889e+00
-9.99339223e-01 3.42234559e-02 -2.42796734e-01 -5.78156412e-01
-4.68464255e-01 4.02282715e-01 -5.99879801e-01 3.59438121e-01
6.97425008e-01 3.43413085e-01 -6.55050933e-01 -7.11041272e-01
3.32540721e-01 1.10301089e+00 9.56065476e-01 -6.68458045e-01
7.01061785e-01 -5.80854248e-03 -3.91798049e-01 -3.87890637e-01
-7.97019124e-01 -6.84631348e-01 -2.70997703e-01 -3.94861341e-01
7.30498135e-01 -9.44142044e-01 -8.31849635e-01 3.85674179e-01
-1.16915643e+00 -8.02504458e-03 1.71341613e-01 6.66396856e-01
-3.61571074e-01 3.62689346e-02 -4.12429929e-01 -1.12055588e+00
-2.70591557e-01 -1.04960299e+00 9.73150849e-01 -3.78499515e-02
-1.55149981e-01 -4.05132145e-01 2.28641823e-01 3.54517162e-01
2.74284363e-01 -8.29455908e-03 9.06692088e-01 -1.24589622e+00
-6.02566183e-01 -2.06632704e-01 1.53653443e-01 2.84357280e-01
4.86671664e-02 -1.30095497e-01 -1.77093375e+00 5.92456758e-02
8.03682283e-02 -4.67356443e-01 7.83969164e-01 -1.25349283e-01
9.80822086e-01 -4.43649173e-01 -1.42417893e-01 2.61195987e-01
1.14174426e+00 6.11688435e-01 3.35648835e-01 1.25928018e-02
5.64191699e-01 2.75640279e-01 4.44821209e-01 3.73720109e-01
3.18060279e-01 5.84196985e-01 4.72522616e-01 1.74331039e-01
-3.54359269e-01 -5.89412630e-01 5.13972521e-01 1.00869894e+00
3.43013793e-01 -6.31754577e-01 -6.41501546e-01 5.80052733e-01
-1.51572311e+00 -1.14918566e+00 2.56671578e-01 2.33863187e+00
1.02789319e+00 3.87169838e-01 1.87591851e-01 7.14140952e-01
8.78901958e-01 -4.43318114e-02 -4.80595052e-01 7.94011168e-03
1.12727366e-01 2.07015902e-01 -4.46317852e-01 4.50792849e-01
-1.05033517e+00 5.58356583e-01 6.07488441e+00 1.04053938e+00
-1.19624686e+00 3.54340702e-01 1.54219449e-01 -3.35330904e-01
-1.45256117e-01 -2.14489207e-01 -6.49982929e-01 7.76331186e-01
1.08457649e+00 2.31009677e-01 4.68259066e-01 4.55819815e-01
1.92781955e-01 -5.39798364e-02 -1.46207905e+00 9.93769825e-01
2.01110587e-01 -7.30861008e-01 -6.35971874e-02 -5.52022040e-01
-9.31307524e-02 1.21053696e-01 1.77659065e-01 4.30547327e-01
2.77916133e-01 -7.43794262e-01 1.13727522e+00 4.28537458e-01
8.02667201e-01 -5.33621490e-01 2.41698235e-01 5.09687245e-01
-1.38431072e+00 -2.65541971e-01 9.21724960e-02 8.06403309e-02
-2.39037704e-02 -5.10237440e-02 -1.17739427e+00 4.26382899e-01
9.17764544e-01 4.25515603e-03 -5.00487924e-01 1.48803651e+00
-2.25120693e-01 1.35098588e+00 -7.43638515e-01 1.85822204e-01
-2.04639181e-01 4.89869356e-01 9.09522474e-01 1.29047894e+00
4.61599320e-01 1.15136974e-01 2.42986694e-01 6.32227898e-01
5.32614328e-02 7.91266337e-02 -3.62276912e-01 2.65979826e-01
1.11096418e+00 1.17248011e+00 -5.26786983e-01 1.16607063e-02
-5.64622045e-01 5.64821064e-01 2.64484007e-02 4.47402775e-01
-8.70949388e-01 -4.04832631e-01 1.70277119e-01 -5.49909249e-02
4.06197697e-01 3.56254876e-01 2.41929471e-01 -8.32766354e-01
1.07100412e-01 -1.07569540e+00 7.86490560e-01 -1.41685736e+00
-1.24910235e+00 1.11534059e+00 3.03041395e-02 -1.54585016e+00
-2.09627509e-01 -2.55714804e-01 -1.03467703e+00 9.40552056e-01
-1.31714845e+00 -1.32822394e+00 -2.47848034e-03 8.88134420e-01
7.43030429e-01 -2.43386537e-01 8.50235701e-01 5.03189743e-01
-4.49859560e-01 7.55818367e-01 -4.53542799e-01 1.66256055e-01
9.26685214e-01 -1.09804988e+00 -1.49747888e-02 1.10358202e+00
6.36230230e-01 3.85393530e-01 9.09675717e-01 -3.11547905e-01
-1.13581860e+00 -8.28316271e-01 7.78754652e-01 -6.46333992e-01
7.10024714e-01 -6.80078447e-01 -1.01393175e+00 6.34425104e-01
2.91001648e-01 2.57435590e-01 9.61524069e-01 -4.41550650e-02
-5.90418756e-01 -2.95670748e-01 -7.61714458e-01 2.78585285e-01
6.77865922e-01 -8.38931620e-01 -8.98489475e-01 -9.79894772e-02
9.07759190e-01 -4.05836135e-01 -3.65988016e-01 2.81281263e-01
4.54286158e-01 -9.50688183e-01 9.87501144e-01 -5.36661267e-01
6.90377317e-03 -6.78871155e-01 -5.27027905e-01 -1.21108961e+00
-8.74305367e-02 -7.95296073e-01 -2.38217264e-01 1.78567207e+00
5.09760678e-01 -1.37345985e-01 -4.10711803e-02 4.66165334e-01
-3.66479397e-01 -4.31425214e-01 -1.14571428e+00 -3.63410473e-01
-6.67186320e-01 -1.02507496e+00 7.86143720e-01 7.16563642e-01
7.58362515e-03 7.96144962e-01 -5.63578188e-01 7.53932476e-01
3.87471825e-01 -1.05536118e-01 5.09486973e-01 -1.09768796e+00
-6.34837627e-01 -1.19594924e-01 -1.85214490e-01 -8.84641647e-01
-1.15556382e-01 -6.70968533e-01 3.60514969e-01 -9.98658240e-01
1.54501304e-01 -3.42452675e-01 -1.08143330e+00 6.29507124e-01
-4.57238197e-01 2.30604574e-01 1.06730074e-01 4.32319678e-02
-8.00009727e-01 4.35503811e-01 8.51361752e-01 -1.61547568e-02
-1.72413468e-01 3.47775578e-01 -6.29864752e-01 9.03511703e-01
4.07611728e-01 -8.13155770e-01 -7.10079610e-01 -4.48319286e-01
1.74146220e-01 5.95526218e-01 5.40812314e-01 -1.10432088e+00
6.32896364e-01 -2.11049523e-03 9.57474783e-02 -6.52454019e-01
7.15906620e-01 -7.15924323e-01 4.78495657e-02 -2.10088938e-01
-6.49446368e-01 -1.87354684e-01 2.89907128e-01 7.14756489e-01
-3.11735392e-01 -1.47156194e-01 4.76589382e-01 1.95033941e-02
-4.64396209e-01 1.90074891e-01 -3.75278175e-01 1.68536097e-01
4.81191158e-01 1.09018184e-01 -2.78916180e-01 -4.03724104e-01
-8.32775593e-01 4.15247567e-02 -4.39966828e-01 5.06595254e-01
7.38946438e-01 -1.48854148e+00 -7.94432878e-01 1.74518585e-01
2.63437003e-01 -3.36123705e-01 2.60182679e-01 4.76391762e-01
3.08431953e-01 3.31572592e-02 -5.45717180e-02 -5.50169706e-01
-1.37578905e+00 7.60068476e-01 3.82846057e-01 2.74520159e-01
-3.70281041e-01 1.20512593e+00 3.46117288e-01 -1.62860647e-01
7.75932074e-01 -4.27582979e-01 -5.08910775e-01 -1.84651464e-01
7.79202938e-01 2.35580549e-01 -2.68022381e-02 -6.13596082e-01
-4.25806582e-01 2.45830312e-01 2.90523231e-01 -5.35919309e-01
9.59333241e-01 -2.31602684e-01 2.19424590e-01 9.07705426e-01
7.89506197e-01 3.04411173e-01 -1.28687382e+00 -5.70600927e-01
-2.00779259e-01 -4.63166237e-01 1.20608948e-01 -1.26643622e+00
-6.96362972e-01 9.56714690e-01 6.72337830e-01 3.40950727e-01
1.42299521e+00 7.10628107e-02 7.65895188e-01 1.78242892e-01
2.77666926e-01 -8.12144756e-01 3.45287651e-01 1.66681111e-01
9.59091306e-01 -1.13826561e+00 -7.81749249e-01 -3.11603189e-01
-6.16501749e-01 8.15467417e-01 8.28156054e-01 2.88359821e-01
7.58744240e-01 5.69852829e-01 3.91393781e-01 1.04847029e-01
-1.36039329e+00 -3.39035958e-01 4.83184308e-01 6.42084539e-01
8.85320641e-03 -2.94608116e-01 4.94566679e-01 1.31080282e+00
-2.24180862e-01 -2.24669665e-01 2.84969956e-01 4.64179635e-01
-5.85364938e-01 -9.40447748e-01 -7.34764218e-01 1.54257059e-01
-7.98760653e-01 -3.91128540e-01 -3.02917749e-01 4.05301958e-01
4.65203881e-01 1.50107312e+00 -1.07234530e-01 -6.68379486e-01
4.36668545e-01 4.78888988e-01 2.08282724e-01 -6.46253288e-01
-6.62524641e-01 5.67863524e-01 1.49152324e-01 -2.36928552e-01
-4.10914451e-01 -4.74906266e-01 -9.30565715e-01 3.89587760e-01
-3.80073398e-01 3.74976218e-01 1.94569781e-01 1.01803148e+00
2.45086923e-01 6.31929100e-01 7.94272125e-01 -5.92086077e-01
-4.85649586e-01 -1.39259946e+00 -3.92501473e-01 3.29884738e-01
6.62286758e-01 -7.52945244e-01 -5.45651436e-01 3.80074054e-01] | [15.202513694763184, 5.299941062927246] |
9a6567b2-58bf-4d2c-bd70-9955c2b2d90a | a-dynamic-evolutionary-framework-for-timeline | 1905.05550 | null | https://arxiv.org/abs/1905.05550v2 | https://arxiv.org/pdf/1905.05550v2.pdf | A Dynamic Evolutionary Framework for Timeline Generation based on Distributed Representations | Given the collection of timestamped web documents related to the evolving topic, timeline summarization (TS) highlights its most important events in the form of relevant summaries to represent the development of a topic over time. Most of the previous work focuses on fully-observable ranking models and depends on hand-designed features or complex mechanisms that may not generalize well. We present a novel dynamic framework for evolutionary timeline generation leveraging distributed representations, which dynamically finds the most likely sequence of evolutionary summaries in the timeline, called the Viterbi timeline, and reduces the impact of events that irrelevant or repeated to the topic. The assumptions of the coherence and the global view run through our model. We explore adjacent relevance to constrain timeline coherence and make sure the events evolve on the same topic with a global view. Experimental results demonstrate that our framework is feasible to extract summaries for timeline generation, outperforms various competitive baselines, and achieves the state-of-the-art performance as an unsupervised approach. | ['Guo-Hua Wang', 'Jing Nie', 'Dongyun Liang'] | 2019-05-14 | null | null | null | null | ['timeline-summarization'] | ['natural-language-processing'] | [ 2.26727828e-01 2.24160269e-01 -3.57869089e-01 -2.51036495e-01
-1.04214430e+00 -6.15378261e-01 1.28440166e+00 6.30673707e-01
1.02583796e-01 7.00185895e-01 1.00918484e+00 2.47577235e-01
-4.36628550e-01 -8.46853912e-01 -6.17753744e-01 -4.90667015e-01
-5.05687118e-01 5.62185585e-01 5.22431016e-01 -1.63232997e-01
5.44756055e-01 1.19586796e-01 -1.63432777e+00 3.24999064e-01
1.06744719e+00 3.92068356e-01 1.87361360e-01 9.34786916e-01
-2.51293808e-01 5.31581342e-01 -1.28695714e+00 -8.06664228e-02
-2.08402038e-01 -6.06851697e-01 -5.82552195e-01 2.37337947e-02
3.54352891e-01 -1.47807658e-01 -6.09468341e-01 5.95483303e-01
3.34271252e-01 4.66324538e-01 6.92475855e-01 -1.29720807e+00
-3.92173946e-01 9.43016112e-01 -5.07779956e-01 5.87105513e-01
5.38665950e-01 -3.79410237e-01 1.22994149e+00 -4.30110693e-01
1.17820227e+00 1.27336919e+00 3.75825226e-01 1.30115137e-01
-7.23299861e-01 -1.74080163e-01 7.01957583e-01 4.31497574e-01
-8.66746426e-01 -1.82242125e-01 7.89505363e-01 -1.91473171e-01
1.01676250e+00 5.23645222e-01 7.48621523e-01 1.35090184e+00
5.20342708e-01 1.11912358e+00 4.59841460e-01 -2.78136194e-01
4.06652272e-01 -2.52542615e-01 4.52815622e-01 2.30358139e-01
3.02407056e-01 -3.14641207e-01 -1.19832945e+00 -7.76778042e-01
1.01248145e-01 2.42939107e-02 -3.57297003e-01 6.18829951e-02
-1.33341122e+00 6.21315241e-01 -1.54085206e-02 3.81903529e-01
-6.29046321e-01 3.96986380e-02 5.39998591e-01 1.80850342e-01
9.94510710e-01 6.13093674e-01 -3.18936765e-01 -4.30700719e-01
-1.46213329e+00 6.14723265e-01 1.02276051e+00 9.40690458e-01
4.59567398e-01 -1.54954880e-01 -8.06678891e-01 3.70877028e-01
-1.81000367e-01 3.52288991e-01 6.28655672e-01 -6.86595559e-01
5.26413679e-01 6.93665743e-01 2.66541958e-01 -9.37874496e-01
-2.98357178e-02 -5.49649954e-01 -3.01053345e-01 -5.29513359e-01
-2.47885704e-01 -1.43674314e-01 -7.03266025e-01 1.68475378e+00
4.93383676e-01 5.54625809e-01 -3.59703414e-02 4.93432105e-01
7.30607510e-01 1.42790318e+00 -2.31395558e-01 -6.21177256e-01
1.04261649e+00 -1.06052113e+00 -9.83957767e-01 -1.54864982e-01
9.65601057e-02 -6.45275354e-01 5.72295964e-01 2.29939520e-01
-1.11652219e+00 -1.85235202e-01 -1.06580937e+00 9.16752964e-02
-3.96934956e-01 -1.90036930e-02 5.67627788e-01 5.47428653e-02
-9.63563204e-01 9.45152760e-01 -1.09471083e+00 -8.47876310e-01
-2.76213318e-01 -2.14757949e-01 2.66855180e-01 1.99360296e-01
-1.28838408e+00 7.06756294e-01 4.81647521e-01 -3.84681582e-01
-1.09855676e+00 -8.77037764e-01 -5.57515860e-01 3.17445695e-01
5.36938310e-01 -8.82764280e-01 1.42783082e+00 -3.17741513e-01
-1.28817511e+00 1.86302260e-01 -7.05743730e-01 -5.83348811e-01
6.08905911e-01 -7.30627716e-01 -4.55547512e-01 3.92585576e-01
3.27331394e-01 1.86737835e-01 7.77405858e-01 -9.58698153e-01
-9.91612196e-01 -1.85618266e-01 -3.85304153e-01 3.66560459e-01
-6.82877183e-01 2.38396302e-01 -8.98233593e-01 -8.38554442e-01
-6.67401180e-02 -6.84603512e-01 -2.58876324e-01 -7.26900041e-01
-7.73402870e-01 -7.43298709e-01 9.68654096e-01 -7.51762152e-01
1.83576334e+00 -1.59609032e+00 3.47619057e-01 -1.09084494e-01
-2.30918135e-02 -2.39723280e-01 5.35410503e-03 1.30943549e+00
2.10258260e-01 2.02245802e-01 6.78880897e-04 -4.15000260e-01
3.93218808e-02 -1.09967738e-01 -1.10989940e+00 1.82851776e-01
-1.10483751e-01 6.58488274e-01 -1.32426095e+00 -5.01826942e-01
-6.46764636e-02 1.63241148e-01 -5.62071912e-02 4.32313234e-01
-6.74251258e-01 2.13741496e-01 -9.43428397e-01 2.66036808e-01
1.56398341e-01 -4.47251439e-01 1.40220135e-01 1.58563107e-01
-4.13234472e-01 6.34256244e-01 -8.47480178e-01 1.70978880e+00
-5.18646045e-03 8.11098337e-01 -7.26666212e-01 -4.25229877e-01
9.32222903e-01 3.70258093e-01 6.31093621e-01 -3.27041090e-01
-4.13115531e-01 -1.07894666e-01 -7.14253724e-01 -3.94802481e-01
1.15125179e+00 7.56352186e-01 -2.82186091e-01 9.71881330e-01
-2.02977774e-03 -2.07101125e-02 6.17519319e-01 7.28568316e-01
1.20205617e+00 3.44485760e-01 4.99926299e-01 3.11761536e-02
-7.02914372e-02 1.25691935e-01 5.89516282e-01 1.25668287e+00
3.04016143e-01 6.28934622e-01 6.41236424e-01 -3.02352071e-01
-1.01814508e+00 -9.19767320e-01 3.17146689e-01 1.29427469e+00
1.44634381e-01 -1.12121153e+00 -5.02938747e-01 -5.78568161e-01
-3.20370942e-01 1.30492926e+00 -7.90423512e-01 -2.04579115e-01
-6.61971211e-01 -7.52322376e-01 1.95012808e-01 2.04213277e-01
1.26301646e-01 -8.43306839e-01 -8.20890844e-01 5.66496611e-01
-4.67453688e-01 -6.38010859e-01 -5.39941370e-01 -3.02112490e-01
-9.08401251e-01 -8.01532328e-01 -9.03001726e-01 -4.04662974e-02
3.76590610e-01 3.61544937e-01 1.07542789e+00 -4.34993953e-01
-1.32126644e-01 6.23531222e-01 -4.28122818e-01 -6.34243429e-01
-2.90133983e-01 3.57582122e-01 -2.24983841e-01 6.19927980e-02
-9.51812137e-03 -5.76152503e-01 -6.93932295e-01 -6.25328496e-02
-8.61793637e-01 2.46363208e-02 1.56663775e-01 5.56938052e-01
4.33823228e-01 2.63349861e-01 6.63375080e-01 -8.73779058e-01
9.18766916e-01 -8.75042796e-01 -1.46081358e-01 7.38006294e-01
-7.95906723e-01 8.83410797e-02 4.54648972e-01 -5.20936310e-01
-1.46606624e+00 -5.99017799e-01 8.01765859e-01 -2.54065484e-01
2.02029809e-01 8.82235229e-01 3.05252552e-01 1.08793950e+00
5.91002285e-01 5.79627275e-01 -7.67106771e-01 -4.25587744e-01
5.75996399e-01 2.88796216e-01 5.18858373e-01 -5.85359454e-01
6.19968295e-01 7.40134180e-01 -4.17944700e-01 -8.52570772e-01
-9.68896806e-01 -7.35580683e-01 -2.39068106e-01 -5.41947782e-01
2.17180982e-01 -8.92491102e-01 8.30710307e-02 1.99229777e-01
-1.29222322e+00 1.92028791e-01 -3.91396374e-01 4.00575936e-01
-3.45247418e-01 4.94327784e-01 -4.79478896e-01 -8.88065934e-01
-8.02896202e-01 -3.94252151e-01 1.33428741e+00 6.51458442e-01
-8.66697907e-01 -9.77050960e-01 6.55915439e-01 -3.45021427e-01
2.86768228e-01 6.95093572e-01 7.24361420e-01 -1.04226851e+00
-8.11805844e-01 -3.67043227e-01 3.44126105e-01 -6.00757062e-01
3.09990525e-01 6.74525917e-01 -8.39424610e-01 -3.80805552e-01
-1.72238559e-01 2.57837474e-01 1.20732200e+00 6.19190514e-01
8.09917688e-01 -7.46797621e-01 -8.51860285e-01 1.88936412e-01
9.49775338e-01 1.99276239e-01 2.70180285e-01 5.55525959e-01
2.46177316e-01 8.66751015e-01 8.70212972e-01 1.02061605e+00
6.24411643e-01 5.98926604e-01 8.90835896e-02 1.01954751e-01
7.89345056e-02 -6.71223521e-01 5.39036930e-01 1.12511504e+00
1.15051232e-01 -9.79941905e-01 -6.06272995e-01 1.13766527e+00
-2.35139918e+00 -1.34546328e+00 -8.89869630e-02 2.23273396e+00
7.23295093e-01 1.54345527e-01 4.63638417e-02 -1.77217945e-01
9.26926672e-01 8.40247691e-01 -7.51960814e-01 -2.34727040e-01
-1.86862528e-01 -2.27069318e-01 8.58188942e-02 2.27847233e-01
-9.33822691e-01 8.06554973e-01 6.73854113e+00 6.92967176e-01
-8.64317417e-01 -1.13886502e-03 3.18797469e-01 -3.96223843e-01
-6.64466202e-01 1.28736451e-01 -1.02506554e+00 5.74077010e-01
1.37457120e+00 -1.28877354e+00 -1.00695141e-01 6.86726153e-01
5.98463118e-01 3.47492397e-02 -1.15224516e+00 3.52793366e-01
3.26995611e-01 -1.43036962e+00 3.64629626e-01 -4.12525684e-02
1.02682054e+00 1.90498754e-02 -1.31261870e-01 1.99081004e-01
5.91741204e-01 -2.78212965e-01 8.41957510e-01 8.86493266e-01
2.59617656e-01 -7.53759205e-01 3.51662636e-01 3.80961210e-01
-1.22377467e+00 1.76566660e-01 -2.73810089e-01 3.93629491e-01
4.77201432e-01 8.50262523e-01 -1.25880671e+00 1.01728523e+00
7.20914185e-01 1.11973214e+00 -6.32540524e-01 1.21354103e+00
-3.51940334e-01 1.03832626e+00 -3.66116315e-01 -4.19899136e-01
1.60830125e-01 3.87993604e-02 1.28693891e+00 1.39402485e+00
7.44430184e-01 -6.56371638e-02 1.99378043e-01 6.23684168e-01
1.28680021e-01 6.70863912e-02 -5.97126722e-01 -3.17275137e-01
8.93906534e-01 1.11162186e+00 -7.30683506e-01 -6.45565569e-01
5.04855439e-02 9.06215668e-01 -2.37967409e-02 3.91116351e-01
-6.84165835e-01 -4.52517241e-01 3.27714384e-01 -2.70103782e-01
3.60721916e-01 2.88336035e-02 8.49152952e-02 -1.22308028e+00
9.15625170e-02 -4.91224289e-01 8.83674145e-01 -9.25420940e-01
-1.13849986e+00 7.68310845e-01 3.49780798e-01 -1.32777059e+00
-8.39474201e-01 4.65794116e-01 -1.39282334e+00 4.77020085e-01
-1.32626271e+00 -1.01245749e+00 -1.94661915e-01 1.08493574e-01
1.03666329e+00 -3.82980369e-02 5.77859402e-01 -5.01416147e-01
-4.73552793e-01 5.38946800e-02 5.74268878e-01 -6.55595303e-01
8.54906678e-01 -1.46528971e+00 9.08181846e-01 1.25614727e+00
7.23559409e-02 7.53353596e-01 1.20414639e+00 -1.10905993e+00
-1.25215197e+00 -1.12693477e+00 1.40407729e+00 -3.91460121e-01
8.16069186e-01 -1.12646550e-01 -9.30437148e-01 5.87822080e-01
6.63798928e-01 -9.39242661e-01 5.33343434e-01 2.01856211e-01
-7.99424201e-02 -1.77722815e-02 -4.64815557e-01 7.85657823e-01
8.56759667e-01 -2.24543348e-01 -1.17396247e+00 7.60143995e-01
1.15860415e+00 -2.49639779e-01 -6.20256722e-01 -6.55879779e-03
2.40466028e-01 -6.19886100e-01 7.19638109e-01 -6.90899909e-01
6.34831011e-01 -1.50820091e-01 3.19365710e-01 -1.55310428e+00
-4.04226035e-01 -1.28481245e+00 -7.23347068e-01 1.72621810e+00
1.69801980e-01 -3.85041624e-01 4.67486262e-01 2.53606379e-01
-1.92643687e-01 -4.67520416e-01 -7.91124165e-01 -5.70070386e-01
-6.47896051e-01 3.76752093e-02 8.55441570e-01 8.08093369e-01
7.65785873e-02 2.50215322e-01 -4.92276460e-01 4.23652530e-01
6.37934446e-01 7.19829798e-01 7.36783743e-01 -1.38543761e+00
-6.30334467e-02 -4.69325960e-01 1.53026804e-01 -8.64652872e-01
3.99141712e-03 -5.44297218e-01 -2.19424237e-02 -1.86592245e+00
4.49540973e-01 2.30522782e-01 -3.30406904e-01 1.34668410e-01
-5.27731955e-01 -8.84918690e-01 -1.30381525e-01 4.93124664e-01
-1.01209390e+00 8.53305876e-01 8.26204658e-01 -2.39032522e-01
-5.14581501e-01 4.51286286e-02 -5.94537616e-01 3.82763237e-01
5.66466212e-01 -6.76468134e-01 -7.00352609e-01 -2.62527794e-01
2.86722928e-01 3.81894648e-01 -1.40729025e-01 -6.96094394e-01
7.41283178e-01 -2.78184921e-01 1.14315972e-01 -1.22919464e+00
2.21860930e-01 -1.51371613e-01 4.27066684e-01 1.77600101e-01
-8.68431747e-01 2.24608541e-01 -2.67523155e-02 1.16506839e+00
-2.67269701e-01 -2.58506089e-01 -8.96383226e-02 -1.30876005e-01
-6.06517255e-01 3.36015314e-01 -4.10240501e-01 1.04448490e-01
8.89006674e-01 -7.60838836e-02 -7.28827655e-01 -8.15986037e-01
-2.48343244e-01 5.34807563e-01 3.32731485e-01 8.14026952e-01
4.57299620e-01 -9.86980021e-01 -1.13217390e+00 -5.27763605e-01
1.99144930e-01 -2.18797386e-01 2.74753660e-01 1.80367425e-01
-4.07139808e-02 5.67874312e-01 2.19547138e-01 -5.33988178e-01
-1.08225381e+00 2.99357980e-01 -1.92135692e-01 -7.03245699e-01
-1.00829554e+00 4.46962416e-01 7.45635927e-02 2.68319607e-01
2.21781939e-01 -7.47005939e-02 -6.38638616e-01 5.78048527e-01
9.50192809e-01 6.82980955e-01 1.18700340e-01 -1.97635919e-01
-1.24958426e-01 2.10852206e-01 -5.21355927e-01 -3.80129725e-01
1.72216094e+00 -3.41561913e-01 -5.92662245e-02 8.83970797e-01
7.87096918e-01 7.40789622e-02 -1.31125045e+00 -3.63101870e-01
4.13310558e-01 -2.01856092e-01 -3.24326120e-02 -6.58546388e-01
-5.03914297e-01 2.47521967e-01 4.43770736e-02 6.67871118e-01
8.22142363e-01 2.02839568e-01 8.70915830e-01 3.84652674e-01
2.35747263e-01 -1.10946333e+00 1.08518280e-01 5.04190981e-01
1.05085111e+00 -5.42352200e-01 4.93646055e-01 -1.80682033e-01
-6.24023378e-01 1.17294025e+00 3.35977167e-01 7.43976310e-02
2.27295950e-01 -1.28330559e-01 -3.59437615e-01 -3.24817419e-01
-1.59689939e+00 9.68141705e-02 5.07174611e-01 2.45744899e-01
2.01731294e-01 -5.37084974e-02 -4.92153227e-01 4.06828672e-01
-3.65215480e-01 -2.62399524e-01 7.30282009e-01 1.12919581e+00
-5.62644780e-01 -7.25332081e-01 -3.12605858e-01 5.29258132e-01
-3.72303128e-01 1.48879662e-01 -6.96576953e-01 5.36428869e-01
-6.22642219e-01 1.02061725e+00 2.15449497e-01 -1.15233869e-03
2.07985222e-01 1.69180527e-01 -5.01431338e-02 -6.48127317e-01
-7.33407140e-01 3.94559115e-01 3.19262534e-01 -4.11044866e-01
-2.01315597e-01 -1.02898371e+00 -9.74599481e-01 -4.94827069e-02
-2.99705923e-01 4.58172470e-01 6.19558692e-01 6.29094303e-01
8.28349113e-01 7.58171201e-01 8.94580007e-01 -8.29343855e-01
-3.45119268e-01 -1.10868251e+00 -1.87384531e-01 2.67863363e-01
2.99678773e-01 -3.22040021e-01 -4.17250335e-01 2.61244744e-01] | [12.616626739501953, 9.498787879943848] |
265e067f-ecb1-4c97-a27e-74bbe1b498a4 | are-the-best-multilingual-document-embeddings | 2304.14796 | null | https://arxiv.org/abs/2304.14796v1 | https://arxiv.org/pdf/2304.14796v1.pdf | Are the Best Multilingual Document Embeddings simply Based on Sentence Embeddings? | Dense vector representations for textual data are crucial in modern NLP. Word embeddings and sentence embeddings estimated from raw texts are key in achieving state-of-the-art results in various tasks requiring semantic understanding. However, obtaining embeddings at the document level is challenging due to computational requirements and lack of appropriate data. Instead, most approaches fall back on computing document embeddings based on sentence representations. Although there exist architectures and models to encode documents fully, they are in general limited to English and few other high-resourced languages. In this work, we provide a systematic comparison of methods to produce document-level representations from sentences based on LASER, LaBSE, and Sentence BERT pre-trained multilingual models. We compare input token number truncation, sentence averaging as well as some simple windowing and in some cases new augmented and learnable approaches, on 3 multi- and cross-lingual tasks in 8 languages belonging to 3 different language families. Our task-based extrinsic evaluations show that, independently of the language, a clever combination of sentence embeddings is usually better than encoding the full document as a single unit, even when this is possible. We demonstrate that while a simple sentence average results in a strong baseline for classification tasks, more complex combinations are necessary for semantic tasks. | ['Cristina Espana-Bonet', 'Josef van Genabith', 'Sonal Sannigrahi'] | 2023-04-28 | null | null | null | null | ['sentence-embeddings', 'sentence-embeddings'] | ['methodology', 'natural-language-processing'] | [ 2.23824903e-01 -1.16808854e-01 -2.97516078e-01 -6.40667021e-01
-1.10268712e+00 -6.38445973e-01 7.98946738e-01 5.26925564e-01
-8.82998228e-01 7.11475790e-01 4.86544311e-01 -3.57133120e-01
4.88629155e-02 -6.19269252e-01 -5.37179887e-01 -3.45521867e-01
1.82812363e-01 5.67790926e-01 1.35620369e-03 -4.70927566e-01
2.82436520e-01 2.99548864e-01 -1.34369278e+00 4.10250276e-01
7.61608362e-01 7.12555110e-01 3.24419528e-01 7.62206495e-01
-9.02658701e-01 3.83026928e-01 -5.92026591e-01 -5.89554191e-01
-7.82165080e-02 -1.58106074e-01 -8.80178273e-01 -1.09531268e-01
5.21240473e-01 1.11431256e-01 -8.21716487e-02 8.81218255e-01
4.76940781e-01 9.82023105e-02 8.56060684e-01 -7.88065255e-01
-1.15688848e+00 8.50907803e-01 -1.41473740e-01 2.01923907e-01
4.26417530e-01 -2.99455523e-01 1.37093019e+00 -1.13026178e+00
5.94697893e-01 1.22047043e+00 7.52754211e-01 5.15067101e-01
-1.41903841e+00 -1.77969351e-01 1.82284802e-01 1.86337650e-01
-1.29807758e+00 -5.60889781e-01 6.18319869e-01 -2.42128387e-01
1.65674376e+00 9.80554521e-02 2.34989628e-01 1.34460831e+00
7.05212429e-02 6.96761370e-01 8.76204610e-01 -9.45689380e-01
-1.35574788e-02 5.97147048e-01 3.94708663e-01 4.61090505e-01
4.30121750e-01 -3.96951556e-01 -5.35003960e-01 -7.44855851e-02
2.31430709e-01 -4.40112874e-02 -2.14316010e-01 -3.32433760e-01
-1.25007820e+00 1.17410922e+00 2.46755674e-01 8.56311023e-01
-2.76579410e-01 1.43075660e-01 8.32164884e-01 4.69008833e-01
7.41224289e-01 5.80973744e-01 -7.82744527e-01 -3.01633209e-01
-1.09012640e+00 1.85377464e-01 7.96389461e-01 7.58852005e-01
8.14473510e-01 8.63022208e-02 -1.86225355e-01 1.14062119e+00
6.33644015e-02 2.06766143e-01 8.39641392e-01 -4.70563620e-01
7.15723455e-01 4.56613481e-01 -6.45957589e-02 -8.24079990e-01
-3.31648886e-01 -3.43234092e-01 -4.37647998e-01 -1.67892486e-01
3.98619413e-01 1.60079867e-01 -8.27914238e-01 1.67135167e+00
-1.32244378e-01 -2.65722185e-01 4.71171796e-01 5.79012930e-01
6.90874755e-01 7.84277737e-01 2.55271494e-01 -3.65565270e-02
1.57705235e+00 -9.54149067e-01 -6.79892302e-01 -4.06934977e-01
1.07544398e+00 -8.76576126e-01 1.21738517e+00 2.67995030e-01
-8.69158983e-01 -6.41748130e-01 -1.06740463e+00 -6.25378788e-01
-1.00470281e+00 1.40882671e-01 5.73519647e-01 8.32407773e-01
-1.03558946e+00 6.06921136e-01 -6.06433272e-01 -7.33023822e-01
1.54017314e-01 8.58016461e-02 -6.74617231e-01 -3.29477996e-01
-1.34146571e+00 1.35373259e+00 5.94397068e-01 -8.17204863e-02
-2.16289029e-01 -6.47890806e-01 -1.41151667e+00 1.61708429e-01
-5.18363230e-02 -4.43119407e-01 1.13801467e+00 -7.38607347e-01
-1.18112731e+00 9.30968702e-01 -4.17575002e-01 -7.21045792e-01
1.48380935e-01 -3.98476928e-01 -3.20204914e-01 8.98777246e-02
-1.25200413e-02 6.59157813e-01 5.65417707e-01 -8.92095149e-01
-2.75903463e-01 -2.59419948e-01 5.68223111e-02 1.07465498e-01
-9.72777724e-01 2.75514513e-01 -1.56051576e-01 -6.49242997e-01
-8.26537684e-02 -4.91346627e-01 -1.80503070e-01 -1.06944159e-01
-5.71681000e-02 -5.89095712e-01 6.84911013e-01 -7.79879689e-01
1.17133391e+00 -2.01575875e+00 9.78565812e-02 -2.70336717e-01
-2.42772326e-01 3.17090631e-01 -3.50236297e-01 8.86753201e-01
-9.55964252e-02 3.18062305e-01 -3.57306987e-01 -7.54266918e-01
3.40473533e-01 4.12782848e-01 -5.33718884e-01 3.29784840e-01
5.39515316e-01 8.95881414e-01 -9.28366959e-01 -6.11340642e-01
4.88035262e-01 7.41228819e-01 -5.74545264e-01 -9.18576792e-02
-1.08699217e-01 -4.11755107e-02 -1.62996575e-01 4.56129611e-01
3.11497897e-01 9.72071812e-02 2.38090932e-01 -8.94751698e-02
-6.37797592e-03 7.10832894e-01 -9.05374944e-01 2.27145696e+00
-1.10405099e+00 6.99608326e-01 -1.95629373e-01 -1.41788745e+00
9.35252547e-01 5.07922947e-01 1.31919608e-01 -5.63538790e-01
1.34390956e-02 5.26996374e-01 -2.51176149e-01 -5.24910629e-01
1.01757920e+00 -3.79733413e-01 -4.79355425e-01 5.06078184e-01
5.29808342e-01 -3.49453956e-01 5.41867435e-01 1.16288461e-01
9.93373632e-01 1.86784074e-01 5.01990855e-01 -2.59756505e-01
4.96277690e-01 -7.55589805e-04 2.14496762e-01 6.10585809e-01
-1.79274846e-02 8.06449533e-01 3.91169041e-01 -3.41133356e-01
-1.23164248e+00 -1.07998610e+00 -4.34313834e-01 1.19000351e+00
-3.19861472e-01 -6.47065163e-01 -5.11783004e-01 -6.76388741e-01
3.60569954e-02 1.14787996e+00 -4.44785535e-01 -7.09606111e-02
-4.98672545e-01 -7.02663839e-01 5.67829907e-01 7.02884734e-01
-1.01144731e-01 -1.02867019e+00 -3.12421054e-01 4.13054854e-01
-8.63962516e-04 -1.38125312e+00 -3.42530400e-01 4.32252347e-01
-7.81715095e-01 -6.59773469e-01 -7.54367709e-01 -8.57231379e-01
5.91434121e-01 8.56622830e-02 1.15453398e+00 -1.19589314e-01
-3.43591422e-01 3.02765846e-01 -6.77878201e-01 -3.08399022e-01
-4.07747954e-01 3.63935649e-01 1.86121568e-01 -2.67897576e-01
8.25312555e-01 -3.34107518e-01 -1.83041364e-01 -3.83786500e-01
-9.58343387e-01 -4.57123697e-01 5.78092515e-01 1.04458487e+00
2.14418486e-01 -3.49182189e-01 5.50283551e-01 -8.96014452e-01
8.70228946e-01 -3.54544342e-01 -1.18741088e-01 3.10938299e-01
-4.49337363e-01 4.43475068e-01 8.84566724e-01 -2.97368765e-01
-8.32954943e-01 -1.70164391e-01 -4.88230377e-01 -1.32655233e-01
-3.21567237e-01 5.54171383e-01 8.27037990e-02 3.76243711e-01
7.29937434e-01 2.75518805e-01 -9.42602083e-02 -6.79887652e-01
6.30780041e-01 1.09682739e+00 2.01989248e-01 -8.18509340e-01
6.03609324e-01 1.50332257e-01 -5.48738837e-01 -1.15363324e+00
-9.53415513e-01 -5.08421719e-01 -8.65757406e-01 3.57228518e-01
9.75657046e-01 -7.54351854e-01 1.07473303e-02 -6.65430278e-02
-1.52100110e+00 1.45251349e-01 -5.46331823e-01 7.11633980e-01
-3.29191983e-01 5.53672373e-01 -5.41686893e-01 -7.16947913e-01
-4.11874443e-01 -1.07497513e+00 1.31313312e+00 -1.42429337e-01
-5.66136181e-01 -1.47004426e+00 1.89353466e-01 2.80112058e-01
6.35968983e-01 -1.00209273e-01 1.11497819e+00 -1.07176757e+00
8.34933370e-02 -4.60924357e-01 -2.39005238e-01 8.75453174e-01
1.72194600e-01 -5.21438792e-02 -1.14938104e+00 -2.66063780e-01
-1.95942402e-01 -6.04155779e-01 1.12463224e+00 1.32955670e-01
9.27856922e-01 -3.98915820e-02 -1.52593806e-01 3.23127478e-01
1.59768367e+00 -3.42139244e-01 4.05547678e-01 3.66288304e-01
5.40389180e-01 6.93196356e-01 2.41457626e-01 1.97344989e-01
3.75596553e-01 5.84965348e-01 4.36027236e-02 1.83571801e-01
-1.60204098e-01 -2.63368398e-01 6.46130562e-01 1.29233408e+00
2.49083608e-01 -2.04437912e-01 -7.74627447e-01 8.96027148e-01
-1.41055346e+00 -8.29240084e-01 -5.68736605e-02 2.21874189e+00
9.75006342e-01 2.04128444e-01 -2.78744847e-01 2.87771702e-01
4.87474948e-01 3.66798818e-01 -4.15292382e-02 -1.17893445e+00
-9.22924578e-02 6.49121523e-01 3.27437103e-01 4.93387014e-01
-8.73609066e-01 1.12381411e+00 6.65886641e+00 8.85225415e-01
-1.05272472e+00 3.55699867e-01 5.78537323e-02 -2.52417102e-02
-5.93157530e-01 -2.66440073e-03 -1.00893509e+00 2.58793652e-01
1.33247459e+00 -2.07181722e-01 7.78129846e-02 8.97388160e-01
-1.23393685e-01 1.68021023e-01 -1.22024930e+00 9.46971953e-01
4.02843088e-01 -1.22368693e+00 2.76724100e-01 -1.36582524e-01
4.69862640e-01 5.54158874e-02 -1.12388179e-01 7.05272019e-01
-2.45755054e-02 -1.09623349e+00 6.17444992e-01 1.68978408e-01
9.16357458e-01 -6.06797636e-01 8.75562847e-01 2.09243149e-01
-1.08649814e+00 9.18327719e-02 -6.96469545e-01 -1.76286355e-01
3.03788066e-01 6.89563930e-01 -6.94664896e-01 7.96485722e-01
4.63322550e-01 8.05950761e-01 -5.58834016e-01 4.96553630e-01
-3.30327123e-01 4.03705478e-01 -2.61717498e-01 -2.46134847e-01
5.12573600e-01 4.90064472e-02 1.85868576e-01 1.70401263e+00
4.89321858e-01 -5.03524184e-01 1.33429735e-03 6.39757872e-01
-1.14794694e-01 4.77578521e-01 -9.45074141e-01 -5.00034094e-01
3.54702920e-01 1.31471741e+00 -4.98167813e-01 -4.00197208e-01
-8.42909038e-01 1.06852102e+00 7.92971730e-01 2.80507594e-01
-5.70230901e-01 -8.72437835e-01 8.98383915e-01 -2.27517411e-01
4.42876697e-01 -6.43847466e-01 -2.14828625e-01 -1.37277722e+00
2.00609848e-01 -5.07376373e-01 3.30969095e-01 -4.61510956e-01
-1.56126857e+00 7.53400803e-01 -2.96529266e-03 -1.05218112e+00
-6.25913918e-01 -1.15057206e+00 -5.18786907e-01 8.55409563e-01
-1.89020765e+00 -9.89715099e-01 2.10049152e-01 2.09534615e-01
8.89438927e-01 -1.52808368e-01 1.39332783e+00 3.50769758e-01
-2.46419683e-01 6.34510279e-01 3.21379215e-01 2.17695162e-01
7.88313925e-01 -1.37435842e+00 5.24068594e-01 6.13230884e-01
7.23544717e-01 8.04744184e-01 7.36643434e-01 -5.98621815e-02
-1.40529191e+00 -8.14623296e-01 1.66697979e+00 -5.72774351e-01
8.82328391e-01 -7.98864305e-01 -1.10602796e+00 6.95696592e-01
4.73584741e-01 1.33653983e-01 8.30952227e-01 5.06218135e-01
-5.95873117e-01 2.14730203e-02 -8.69245887e-01 4.19646025e-01
8.01129282e-01 -9.11112130e-01 -1.23386002e+00 4.68985319e-01
8.68116856e-01 -6.69301823e-02 -9.08765614e-01 6.58494532e-02
4.19016123e-01 -6.72809780e-01 9.87668931e-01 -8.30236316e-01
4.79359001e-01 -5.68705872e-02 -4.44550514e-01 -1.52380443e+00
1.36606529e-01 -1.11625485e-01 1.33690313e-01 1.40524244e+00
6.24106050e-01 -7.45132804e-01 5.17610431e-01 4.05231893e-01
-2.43027002e-01 -6.40846133e-01 -1.05255342e+00 -1.07060778e+00
4.77387428e-01 -9.83288646e-01 3.71847361e-01 9.58313107e-01
1.74267679e-01 5.67740262e-01 -4.02617082e-02 -1.05379894e-01
4.02841777e-01 2.82745082e-02 4.92054075e-01 -1.08313441e+00
-6.79069012e-02 -4.87336516e-01 -6.42078340e-01 -9.49072778e-01
7.14436948e-01 -1.45600641e+00 -3.18844207e-02 -1.87152386e+00
4.84337360e-02 -2.64463723e-01 -3.72378021e-01 3.92537326e-01
1.13448761e-02 1.71169177e-01 4.87551689e-02 -2.63323933e-01
-2.94121861e-01 7.18493402e-01 6.62625849e-01 -2.62677073e-01
1.54144853e-01 -5.48716605e-01 -6.72303915e-01 6.08366251e-01
7.59713113e-01 -5.70997238e-01 -2.28248253e-01 -9.93370593e-01
2.50361055e-01 -4.97162640e-01 1.79285675e-01 -8.61661792e-01
5.93812652e-02 2.17062593e-01 2.38347307e-01 -3.77595007e-01
5.86083770e-01 -6.67820573e-01 -6.04179621e-01 3.05141598e-01
-4.69681442e-01 2.46636365e-02 1.93899825e-01 3.92332673e-01
-5.78917801e-01 -7.64942646e-01 4.26016986e-01 -2.02914357e-01
-8.18864226e-01 -3.94390617e-03 -1.89700305e-01 1.76527604e-01
6.38984561e-01 -2.63967335e-01 -2.02362955e-01 -9.75397080e-02
-5.28481662e-01 1.18067954e-02 3.36595476e-01 7.65129507e-01
5.52562773e-01 -1.28040671e+00 -8.52162182e-01 2.26605713e-01
4.52474445e-01 -3.49758118e-01 -1.86555479e-02 5.02973676e-01
-4.52643842e-01 8.40067208e-01 3.08836196e-02 -4.24846441e-01
-1.00153601e+00 5.12727022e-01 -9.40762758e-02 -3.80928695e-01
-4.97665435e-01 7.84601331e-01 -1.22366242e-01 -8.17728221e-01
1.18298538e-01 -5.68361640e-01 -2.08234295e-01 3.38742644e-01
5.88865995e-01 -6.78606331e-02 3.33719820e-01 -6.02811635e-01
-4.23964262e-01 6.11286342e-01 -2.86994308e-01 -1.60909399e-01
1.42868483e+00 -4.48431559e-02 -1.02836192e-01 7.76303232e-01
1.57054031e+00 -1.28784180e-01 -5.70539773e-01 -4.02300447e-01
2.49321952e-01 -4.14029449e-01 4.55446504e-02 -2.30568588e-01
-5.87824345e-01 1.46551204e+00 2.01148584e-01 2.73200601e-01
6.18494928e-01 5.85381612e-02 1.02494156e+00 6.59337103e-01
3.99069339e-01 -1.27378047e+00 -1.65811077e-01 7.18787372e-01
8.05397689e-01 -1.24719810e+00 3.52791101e-02 1.54805193e-02
-4.87117380e-01 1.41485834e+00 1.32728457e-01 -1.29236519e-01
4.79966581e-01 1.16768815e-01 -6.79437723e-03 8.70573372e-02
-7.59855509e-01 -3.14018011e-01 1.63767532e-01 5.35591602e-01
1.02520990e+00 -1.18348271e-01 -5.11668503e-01 5.37092924e-01
-4.04782057e-01 -2.74107158e-01 3.98002446e-01 9.89702761e-01
-5.52084446e-01 -1.46468163e+00 -2.79918283e-01 4.87399876e-01
-5.85738420e-01 -4.52730626e-01 -9.82596800e-02 8.04414570e-01
-8.78552571e-02 8.02035749e-01 2.09828287e-01 -6.47023553e-03
2.46762693e-01 5.79460800e-01 6.69150293e-01 -1.09994662e+00
-4.39704478e-01 -5.09237587e-01 3.87352586e-01 -2.08646789e-01
-3.67684841e-01 -6.81481421e-01 -1.16127741e+00 -2.09083613e-02
-2.61982948e-01 2.53449082e-01 1.12104273e+00 1.10843182e+00
2.36056805e-01 3.80176067e-01 2.33634964e-01 -9.39643979e-01
-8.20764482e-01 -1.26594162e+00 -5.89182258e-01 4.99812573e-01
1.91836834e-01 -4.80227947e-01 -4.90486860e-01 1.83323454e-02] | [10.667266845703125, 8.791193962097168] |
f50b0103-9425-45f4-8c88-199121714387 | pyramidal-fisher-motion-for-multiview-gait | 1403.6950 | null | http://arxiv.org/abs/1403.6950v1 | http://arxiv.org/pdf/1403.6950v1.pdf | Pyramidal Fisher Motion for Multiview Gait Recognition | The goal of this paper is to identify individuals by analyzing their gait.
Instead of using binary silhouettes as input data (as done in many previous
works) we propose and evaluate the use of motion descriptors based on densely
sampled short-term trajectories. We take advantage of state-of-the-art people
detectors to define custom spatial configurations of the descriptors around the
target person. Thus, obtaining a pyramidal representation of the gait motion.
The local motion features (described by the Divergence-Curl-Shear descriptor)
extracted on the different spatial areas of the person are combined into a
single high-level gait descriptor by using the Fisher Vector encoding. The
proposed approach, coined Pyramidal Fisher Motion, is experimentally validated
on the recent `AVA Multiview Gait' dataset. The results show that this new
approach achieves promising results in the problem of gait recognition. | ['R. Medina-Carnicer', 'M. J. Marin-Jimenez', 'F. M. Castro'] | 2014-03-27 | null | null | null | null | ['multiview-gait-recognition'] | ['computer-vision'] | [-2.08214372e-01 -6.07702374e-01 1.36338118e-02 -6.90869987e-02
-3.66758436e-01 -4.54638898e-01 7.67936647e-01 3.30312133e-01
-7.50854850e-01 4.53391671e-01 3.69554684e-02 5.29026568e-01
-1.15205839e-01 -8.05395246e-01 -1.70487925e-01 -8.46460760e-01
-4.15107369e-01 4.97840405e-01 5.96796095e-01 -2.45119214e-01
3.11412364e-01 7.34164178e-01 -1.74800396e+00 -3.00329756e-02
3.67988706e-01 7.18304873e-01 -1.65911019e-01 8.27307403e-01
2.47564986e-01 2.09959686e-01 -6.37371778e-01 -3.22720289e-01
6.29302338e-02 -4.84719694e-01 -5.57508767e-01 3.60638469e-01
3.97849053e-01 -5.37385345e-02 -1.71873808e-01 8.38807225e-01
3.74677092e-01 3.53082925e-01 9.00025129e-01 -1.05048656e+00
-1.66395932e-01 -1.08786866e-01 -4.42598283e-01 4.35681552e-01
8.27140510e-01 1.58001855e-01 9.70011711e-01 -9.74635720e-01
9.49636459e-01 1.28943801e+00 1.00040030e+00 1.87993392e-01
-1.36095941e+00 -1.03669710e-01 -7.87806734e-02 4.42744672e-01
-1.78956449e+00 -3.38114530e-01 6.74382269e-01 -7.60623753e-01
7.87044466e-01 8.84369761e-02 9.37670827e-01 9.00281429e-01
2.18464091e-01 6.08528793e-01 8.21889877e-01 -5.75969636e-01
2.79116184e-01 -2.80480444e-01 3.94785345e-01 1.00353253e+00
6.07517302e-01 4.32653906e-05 -3.93410206e-01 -3.31219733e-01
5.96842051e-01 1.68037657e-02 2.24260706e-02 -7.41969645e-01
-1.18347323e+00 8.14387858e-01 2.32598588e-01 7.80507684e-01
-5.49478710e-01 2.04774052e-01 4.39724118e-01 -1.49884537e-01
4.09880757e-01 -2.67078847e-01 2.73765534e-01 -2.75235325e-01
-1.40785706e+00 8.22156847e-01 7.23466396e-01 4.30938929e-01
8.70044589e-01 -1.16624527e-01 -2.57754058e-01 4.52827334e-01
4.82391924e-01 6.02727532e-01 3.46739262e-01 -7.14011848e-01
4.03655380e-01 7.26885080e-01 3.54576617e-01 -1.45791614e+00
-4.93111491e-01 -1.88889936e-01 -7.08540797e-01 1.81336597e-01
7.13220179e-01 -5.71259111e-02 -6.19221091e-01 1.34622836e+00
3.65990400e-01 1.84563518e-01 -1.80568695e-01 8.17798853e-01
2.36036465e-01 3.97766650e-01 1.20907538e-01 9.60541423e-03
1.44458389e+00 -4.82454598e-01 -5.67470729e-01 1.54968962e-01
5.92935443e-01 -5.01496255e-01 4.13986862e-01 2.09839970e-01
-8.83481741e-01 -7.56920636e-01 -1.12237108e+00 3.60099822e-01
-5.78474402e-01 4.18074489e-01 4.72060740e-02 8.75622511e-01
-9.64798033e-01 8.64389420e-01 -1.04006886e+00 -8.25520754e-01
1.68577522e-01 2.40900785e-01 -4.44561988e-01 7.76234120e-02
-1.00540853e+00 7.05652475e-01 3.15756083e-01 1.51344195e-01
-5.01540244e-01 -1.95479151e-02 -9.32782650e-01 -1.14341453e-01
4.87922169e-02 -7.22188890e-01 6.55925930e-01 -6.28212929e-01
-1.17383313e+00 1.02395749e+00 -2.08678097e-01 -4.52253222e-01
7.76487827e-01 -2.50932992e-01 -2.84969836e-01 5.68925560e-01
2.92842746e-01 3.85078549e-01 9.88696039e-01 -9.85994220e-01
-5.76128542e-01 -5.49298525e-01 -3.87727499e-01 -7.59859234e-02
-3.16602498e-01 2.06189174e-02 -4.19607610e-01 -7.86548793e-01
-1.04927428e-01 -1.03993416e+00 -1.79253116e-01 -1.35736495e-01
-2.59084910e-01 -3.01884234e-01 9.20125186e-01 -7.62910903e-01
1.42095613e+00 -2.08263493e+00 4.70408082e-01 6.00019693e-01
-6.24170876e-04 4.75970268e-01 2.28887111e-01 6.65357828e-01
2.24975124e-01 -1.59039408e-01 -3.57006550e-01 -4.99473691e-01
6.66862205e-02 1.43592134e-01 1.52420282e-01 9.34267581e-01
1.63960561e-01 6.18294418e-01 -8.97860110e-01 -7.93723106e-01
5.52111208e-01 5.67294002e-01 -3.57957900e-01 -5.33124879e-02
3.12521398e-01 2.85310477e-01 -7.24532604e-01 5.43456852e-01
5.15098333e-01 1.53124601e-01 -8.01394042e-03 -9.53581929e-02
-3.04102629e-01 -4.46350455e-01 -1.46102345e+00 1.67760968e+00
1.77024692e-01 3.65513057e-01 -2.71654427e-01 -1.04651773e+00
1.01715648e+00 3.00926834e-01 7.39096642e-01 -2.07452849e-01
7.74747282e-02 1.37969315e-01 -3.48762333e-01 -4.38774407e-01
3.79370242e-01 6.51527056e-03 -4.04367536e-01 1.03015490e-01
3.58085901e-01 5.37642300e-01 6.34310961e-01 -2.05347627e-01
1.09364748e+00 4.96536791e-01 6.97887301e-01 -4.29623872e-01
1.03487921e+00 4.93622497e-02 2.44566366e-01 6.27635419e-01
-5.19625485e-01 6.33547306e-01 2.83124387e-01 -5.62488914e-01
-1.00074005e+00 -1.26892591e+00 -1.91219673e-02 6.93533063e-01
1.21453702e-02 -7.67945945e-01 -9.33072805e-01 -4.61271286e-01
2.58365929e-01 3.24268378e-02 -6.57815576e-01 4.49698828e-02
-9.41840172e-01 -7.51731753e-01 7.23513842e-01 5.54882169e-01
6.13438308e-01 -8.34268391e-01 -1.31157315e+00 4.62418407e-01
-1.84794977e-01 -9.82752979e-01 -3.35491985e-01 -3.09904009e-01
-6.90798879e-01 -9.52950358e-01 -1.20463884e+00 -5.11781037e-01
2.73820907e-01 9.13642906e-03 8.30293179e-01 -1.39734030e-01
-7.45422721e-01 7.80572951e-01 -3.51364851e-01 3.38633269e-01
-8.89475122e-02 -1.17839925e-01 1.95391446e-01 5.62819719e-01
4.92799431e-01 -4.52560455e-01 -6.28676832e-01 1.97608799e-01
-4.77123111e-01 -6.93551183e-01 3.80430132e-01 7.59703517e-01
4.63670135e-01 -1.45118222e-01 1.31225020e-01 -2.15593696e-01
4.29012805e-01 -2.81734943e-01 -4.43814456e-01 1.06753744e-01
-1.51454449e-01 2.07370743e-01 3.71728003e-01 -3.41206849e-01
-5.99062502e-01 3.17788184e-01 -9.02175009e-02 -5.00144005e-01
-4.96856958e-01 2.13423267e-01 9.99164283e-02 -2.15221584e-01
4.56974447e-01 3.26573908e-01 -7.40314797e-02 -5.00869334e-01
3.40642512e-01 3.31890047e-01 5.54280519e-01 -4.52745914e-01
8.28417361e-01 6.46818340e-01 4.13764298e-01 -1.26245642e+00
-5.76176830e-02 -9.78577971e-01 -1.24557829e+00 -6.27321780e-01
1.28077447e+00 -7.04454541e-01 -8.99580598e-01 4.96993393e-01
-1.12751532e+00 3.77167284e-01 -2.81990856e-01 4.61337894e-01
-7.91108072e-01 7.61019468e-01 -4.27827477e-01 -1.03954625e+00
-8.56274217e-02 -9.21402037e-01 1.36956525e+00 1.56159908e-01
-5.94969571e-01 -1.12321901e+00 6.37032568e-01 -1.11379236e-01
3.66282277e-02 5.41654944e-01 4.97380435e-01 -4.76513445e-01
-3.76016259e-01 -5.55297792e-01 1.88548207e-01 1.60685942e-01
2.40556952e-02 1.84241518e-01 -6.34735107e-01 -2.88549453e-01
-2.59011209e-01 8.77989084e-02 1.10434222e+00 4.45230603e-01
3.27567518e-01 1.37284011e-01 -6.87183261e-01 3.60385209e-01
1.46687329e+00 -2.27777530e-02 4.82609928e-01 3.06020617e-01
6.56704664e-01 6.32253706e-01 7.82010674e-01 7.03596354e-01
2.45709017e-01 1.14306831e+00 -8.11034814e-03 4.16920334e-01
-2.69428808e-02 -1.09173395e-01 6.61486030e-01 3.83772999e-01
-7.16725290e-01 1.20258965e-02 -1.03344202e+00 6.68851912e-01
-2.03419518e+00 -1.38431871e+00 -1.73796445e-01 2.27551413e+00
-9.04084146e-02 1.10358909e-01 7.66426921e-01 3.73086840e-01
8.41357827e-01 4.12423998e-01 1.62039533e-01 -2.34908149e-01
1.63289169e-05 2.11621106e-01 5.20531356e-01 5.17872572e-01
-1.52128255e+00 5.87022126e-01 6.35209513e+00 5.96861005e-01
-7.87766099e-01 -3.94890793e-02 2.67276866e-03 2.91793764e-01
5.45009375e-01 -1.63637564e-01 -1.14831257e+00 4.53653932e-01
1.00038385e+00 -3.28571275e-02 -5.41608967e-02 6.88197076e-01
1.79164022e-01 -2.84221262e-01 -8.03417385e-01 1.09783423e+00
2.87067831e-01 -1.00126278e+00 -5.36128357e-02 7.94050768e-02
2.43220538e-01 -5.02913237e-01 2.97815651e-02 -6.49190620e-02
-1.10445373e-01 -6.06605947e-01 7.72306740e-01 9.11274135e-01
4.15987313e-01 -8.38494956e-01 6.73000872e-01 1.35813713e-01
-1.99258506e+00 -1.47953048e-01 -1.41049549e-01 6.99768141e-02
3.00096780e-01 3.65398198e-01 -6.85002923e-01 7.63455629e-01
5.54661453e-01 1.04378140e+00 -8.52302313e-01 1.13839972e+00
2.33720466e-01 4.46200997e-01 -3.63853484e-01 1.44022200e-02
3.24810982e-01 -4.42508578e-01 8.81869256e-01 1.61669970e+00
5.37352502e-01 -3.35425258e-01 4.70272839e-01 8.83259237e-01
5.41417539e-01 1.39958218e-01 -8.56467903e-01 -8.72488320e-02
-6.18666075e-02 1.14273727e+00 -9.78853881e-01 -3.96531910e-01
-3.14714700e-01 1.07431042e+00 -2.68615223e-02 1.93352714e-01
-4.67479378e-01 -3.41658711e-01 6.28383100e-01 2.43929312e-01
7.54467010e-01 -6.20222569e-01 1.85997665e-01 -1.16447318e+00
1.88540682e-01 -2.88337678e-01 4.44708705e-01 -1.65895849e-01
-1.20492303e+00 2.65319854e-01 4.28196281e-01 -1.46754503e+00
-7.39020467e-01 -6.64011657e-01 -6.58081949e-01 9.65968907e-01
-8.07701707e-01 -1.28251994e+00 -1.72935992e-01 7.18981504e-01
3.02215725e-01 -1.55564547e-01 9.03767645e-01 3.25415671e-01
-3.23742896e-01 1.80648118e-01 1.02321856e-01 2.91396707e-01
4.70242083e-01 -1.12842631e+00 5.30329764e-01 8.75543594e-01
8.96212384e-02 5.24276376e-01 9.60277557e-01 -8.85251105e-01
-1.24265885e+00 -8.60859573e-01 1.03250301e+00 -4.42283332e-01
6.49602532e-01 -2.74287611e-01 -7.14390159e-01 4.57959741e-01
-2.07173213e-01 2.44728789e-01 4.11915302e-01 -2.14190707e-01
-2.46961247e-02 -3.03911082e-02 -1.12940812e+00 3.51506025e-01
1.00144029e+00 -3.99164468e-01 -7.65117764e-01 -9.22597274e-02
-1.84069425e-01 2.72881746e-01 -8.61357033e-01 1.70871988e-01
7.83389211e-01 -1.08186090e+00 1.39198148e+00 -3.23413730e-01
-2.97238886e-01 -3.85960221e-01 -1.53968677e-01 -1.10024405e+00
-4.18912500e-01 -2.80053496e-01 -5.97792789e-02 1.08378363e+00
-4.52090740e-01 -3.55560273e-01 8.98506761e-01 -2.35507354e-01
3.22041363e-01 -2.38552153e-01 -1.23941875e+00 -8.87878060e-01
-2.00732693e-01 -8.72871727e-02 2.35150754e-02 4.47719783e-01
3.11779231e-02 4.67182212e-02 -3.63434464e-01 6.93074018e-02
1.10374057e+00 2.45585516e-02 6.60270214e-01 -1.54006529e+00
-3.12340230e-01 -4.42121178e-01 -1.38028264e+00 -6.35916591e-01
1.14225440e-01 -6.66036963e-01 -8.80930349e-02 -1.15578973e+00
-3.01741976e-02 6.10280111e-02 -1.32465437e-01 3.03751249e-02
-8.64374712e-02 4.61770087e-01 4.29490119e-01 1.40439838e-01
-5.60593903e-01 3.90347153e-01 7.05618024e-01 -7.64312446e-02
-1.73661374e-02 9.89653729e-03 3.30698401e-01 7.70102799e-01
4.99310791e-01 -2.91990429e-01 1.39253110e-01 3.45479190e-01
-2.95114487e-01 1.48941517e-01 9.05183136e-01 -1.66563666e+00
1.03716500e-01 1.54781029e-01 5.32488406e-01 -8.64790738e-01
6.23614728e-01 -5.97601414e-01 1.88095421e-01 9.91990805e-01
6.65928423e-02 3.77547920e-01 -1.43084511e-01 6.83548331e-01
-2.99701482e-01 -2.67075181e-01 8.27295005e-01 -2.25559294e-01
-9.90136266e-01 6.01589344e-02 -6.09427750e-01 -2.78173685e-01
1.23931813e+00 -4.20257747e-01 2.03676701e-01 -1.57357529e-01
-1.10518372e+00 -2.30847701e-01 6.41994834e-01 1.15235388e-01
7.06834853e-01 -1.41149199e+00 -6.04808152e-01 2.84801960e-01
2.07214341e-01 -8.48022819e-01 2.70874321e-01 8.72866869e-01
-6.46632314e-01 5.85554779e-01 -6.44243002e-01 -9.32379067e-01
-1.60565639e+00 6.16131008e-01 4.61229533e-01 -4.65932906e-01
-8.62018287e-01 2.96321541e-01 -2.36267954e-01 1.49639025e-01
-1.58719435e-01 -1.99042216e-01 -5.14451683e-01 4.53199655e-01
5.85579515e-01 7.90234506e-01 -4.89233173e-02 -1.35323417e+00
-7.05633521e-01 1.14845443e+00 4.22489226e-01 -4.97329414e-01
1.17842519e+00 -1.82404056e-01 8.05689842e-02 7.51875162e-01
1.21794891e+00 -6.47618622e-02 -1.01510155e+00 3.33829522e-02
4.41814899e-01 -5.05675733e-01 -5.32883823e-01 -4.48731631e-02
-5.41561782e-01 1.02021360e+00 1.12128413e+00 1.97631180e-01
7.16182649e-01 -2.51380414e-01 5.13859510e-01 1.91226497e-01
6.16183698e-01 -8.57455134e-01 5.81429116e-02 3.18923026e-01
5.01374006e-01 -7.65736282e-01 -2.03262810e-02 -3.56581450e-01
-4.85848367e-01 1.35216320e+00 -5.00913747e-02 -7.25978076e-01
5.68215370e-01 1.15214046e-02 -3.99369866e-01 -4.23772670e-02
-2.29185134e-01 -6.80148482e-01 3.41140538e-01 8.05591822e-01
1.70042753e-01 1.17479227e-01 -6.71307027e-01 2.35008433e-01
2.01931685e-01 9.82458293e-02 2.32388452e-01 1.11730504e+00
-6.83885455e-01 -1.20913053e+00 -8.19212437e-01 1.23883896e-01
-2.60093898e-01 4.76043344e-01 -2.70434767e-01 9.45198298e-01
3.04060191e-01 7.60711312e-01 -3.93745257e-03 -4.24881607e-01
4.54859912e-01 4.86850709e-01 7.29421198e-01 -3.09244543e-01
-5.78161776e-01 8.51605833e-02 5.32458425e-02 -6.58832252e-01
-7.66360223e-01 -1.20970511e+00 -1.01018131e+00 -1.16574369e-01
4.00510579e-01 9.86081362e-02 3.87790561e-01 1.00227606e+00
-1.92558058e-02 1.67049095e-01 3.18475485e-01 -1.48014629e+00
-5.74976027e-01 -6.82674170e-01 -7.62504995e-01 8.75668228e-01
4.16609049e-01 -1.07241774e+00 -1.80494979e-01 2.78501302e-01] | [14.20322322845459, 1.5042697191238403] |
d14e1feb-43ff-4d44-9198-30dce350fd8e | conspiracy-machines-the-role-of-social-bots | 2012.09536 | null | https://arxiv.org/abs/2012.09536v1 | https://arxiv.org/pdf/2012.09536v1.pdf | Conspiracy Machines -- The Role of Social Bots during the COVID-19 Infodemic | The omnipresent COVID-19 pandemic gave rise to a parallel spreading of misinformation, also referred to as an Infodemic. Consequently, social media have become targets for the application of social bots, that is, algorithms that mimic human behaviour. Their ability to exert influence on social media can be exploited by amplifying misinformation, rumours, or conspiracy theories which might be harmful to society and the mastery of the pandemic. By applying social bot detection and content analysis techniques, this study aims to determine the extent to which social bots interfere with COVID- 19 discussions on Twitter. A total of 78 presumptive bots were detected within a sample of 542,345 users. The analysis revealed that bot-like users who disseminate misinformation, at the same time, intersperse news from renowned sources. The findings of this research provide implications for improved bot detection and managing potential threats through social bots during ongoing and future crises. | ['Eric Hochstrate', 'Milad Mirbabaie', 'Felix Brünker', 'Julian Marx'] | 2020-12-17 | null | null | null | null | ['rumour-detection'] | ['natural-language-processing'] | [-5.74162789e-02 5.23198843e-01 -1.72734007e-01 6.19995594e-01
3.04933548e-01 -7.13563919e-01 1.04958391e+00 4.12625730e-01
-4.16680932e-01 6.62402570e-01 5.09331107e-01 -5.63576162e-01
4.11010593e-01 -8.83647561e-01 -6.44637719e-02 -5.04366994e-01
-1.86678357e-02 2.06797510e-01 2.35543936e-01 -5.99090576e-01
7.36218989e-01 3.10163498e-01 -7.79374659e-01 1.06759146e-01
4.91896749e-01 4.36312780e-02 1.43330976e-01 4.94663149e-01
-6.02430850e-02 1.20097840e+00 -1.31516731e+00 -3.77365351e-01
-1.69942379e-01 -4.15053576e-01 -4.89630371e-01 -8.05438608e-02
-2.89839566e-01 -4.81112003e-01 -4.10059094e-01 1.25991142e+00
4.16025639e-01 -2.69731671e-01 4.28224146e-01 -1.29528379e+00
-4.79474068e-01 8.48110974e-01 -7.90645361e-01 8.77194405e-01
4.84572530e-01 5.00590444e-01 4.65127110e-01 -4.00017232e-01
9.91531312e-01 1.75989592e+00 7.94085443e-01 2.47210830e-01
-1.06637299e+00 -9.53867078e-01 -4.44735587e-01 -2.59767681e-01
-1.04585588e+00 -2.25642532e-01 4.79022950e-01 -9.51519549e-01
6.81021035e-01 1.61501437e-01 8.25266302e-01 1.21651709e+00
6.81740701e-01 3.98403332e-02 1.04910445e+00 4.24258485e-02
1.02318162e-02 5.61842024e-01 -3.22017014e-01 4.00039643e-01
8.09325933e-01 -5.85937165e-02 -2.95899391e-01 -9.16340172e-01
3.72654855e-01 -1.53205529e-01 -1.31802009e-02 7.24398494e-01
-9.41894770e-01 1.58990693e+00 6.39939368e-01 7.67004490e-01
-8.30745935e-01 -1.85811594e-01 6.13972068e-01 4.08973426e-01
9.69469488e-01 7.72718489e-01 -2.77475221e-03 -2.77778119e-01
-1.88337445e-01 1.93404675e-01 8.38166952e-01 1.53220311e-01
7.44466960e-01 -2.95960102e-02 1.09545276e-01 5.05725443e-01
3.63067478e-01 8.86846900e-01 4.09587562e-01 -6.15144789e-01
3.34361583e-01 7.08577454e-01 1.93522945e-01 -2.21842813e+00
-6.15980864e-01 -3.59705389e-01 -4.66792107e-01 -2.59227365e-01
2.81920224e-01 -6.95044518e-01 -1.61403045e-01 1.41506469e+00
4.12562430e-01 -7.96548426e-02 -4.94088173e-01 4.61452931e-01
3.72240424e-01 7.71005511e-01 2.66499639e-01 -4.60799843e-01
1.19763756e+00 -3.27948421e-01 -8.50082874e-01 -1.92515999e-01
8.21471393e-01 -7.73965955e-01 4.74834055e-01 -1.31482154e-01
-7.15397596e-01 2.13043585e-01 -4.25247878e-01 7.45929778e-01
-7.68837631e-01 -1.10998642e+00 1.19212918e-01 9.77930725e-01
-8.55273247e-01 3.49469095e-01 -3.98094743e-01 -6.44545674e-01
3.78822148e-01 -1.06805317e-01 2.62552291e-01 2.39401504e-01
-1.20750582e+00 1.09714544e+00 2.87355274e-01 -3.08495075e-01
-5.75716972e-01 -4.09027010e-01 -2.40143090e-01 -4.88696873e-01
3.61432701e-01 -2.98876762e-01 9.18248296e-01 -9.22215700e-01
-7.93150783e-01 7.89211750e-01 2.31357321e-01 -4.66083139e-01
4.31467503e-01 2.52103478e-01 -4.90524888e-01 4.14027214e-01
8.93323600e-01 1.90501422e-01 9.44694281e-01 -1.01812708e+00
-4.21532929e-01 -5.30381441e-01 5.52912243e-02 -1.38623476e-01
-6.16536558e-01 7.97120869e-01 7.36139774e-01 -7.59016633e-01
-2.64811069e-01 -1.00810778e+00 -2.00256810e-01 -7.95316041e-01
-8.07834625e-01 -1.23965926e-01 1.38114071e+00 -8.04454505e-01
1.09924185e+00 -1.74639106e+00 -3.40854853e-01 2.23825276e-01
6.81997299e-01 6.68707073e-01 1.62057862e-01 1.23894203e+00
4.43913698e-01 7.84577489e-01 2.68359900e-01 2.20892131e-01
-4.40609157e-01 4.45038825e-02 -3.43506247e-01 8.17642510e-01
-1.02220990e-01 6.52926683e-01 -1.44707549e+00 -1.23966768e-01
-1.25531882e-01 2.60690272e-01 -3.31324041e-01 -6.97985440e-02
-1.67962775e-01 8.71532738e-01 -7.17939854e-01 1.84376776e-01
4.68324184e-01 -3.88675302e-01 1.57783613e-01 5.76821327e-01
-5.75319707e-01 5.04662514e-01 -5.95263764e-02 7.67930690e-03
9.18547809e-03 1.14839959e+00 4.98348027e-01 -4.09286588e-01
6.85109377e-01 1.71266541e-01 3.56205195e-01 -2.97900319e-01
7.32209384e-01 1.39493104e-02 2.93481648e-01 -5.33352137e-01
7.88086712e-01 -5.87152094e-02 6.01902902e-02 1.26826656e+00
-6.01743102e-01 1.06575243e-01 4.29797806e-02 5.58816135e-01
1.29130018e+00 -9.73534405e-01 4.17651623e-01 -2.81489938e-01
1.68647692e-01 3.84968281e-01 8.96611251e-03 8.64510596e-01
-4.49108750e-01 -5.62231481e-01 7.15731859e-01 -2.68625408e-01
-6.04394197e-01 -4.45703447e-01 2.43271098e-01 1.19142663e+00
2.01827779e-01 -4.19051945e-01 -8.04290175e-01 -4.60336268e-01
-1.98927615e-03 8.68576884e-01 -4.58445311e-01 -1.43866241e-01
-3.47388119e-01 -9.02806997e-01 8.19022536e-01 -5.14113128e-01
6.08483911e-01 -1.34354210e+00 -6.33415043e-01 4.47029501e-01
-6.03951514e-01 -9.19281006e-01 -2.05224320e-01 -5.21398962e-01
-4.66647029e-01 -1.40911186e+00 -4.82727140e-01 -1.11231126e-01
7.34579146e-01 9.88962650e-01 5.12144923e-01 5.33522725e-01
-1.45469010e-01 2.67706692e-01 -4.89717007e-01 -8.93713117e-01
-1.43775117e+00 1.10603020e-01 2.79780000e-01 -7.96692166e-03
3.64694327e-01 -6.01499915e-01 -2.77849823e-01 3.79535586e-01
-9.04016078e-01 -3.22880596e-01 8.60183686e-02 8.76595452e-02
-8.40157151e-01 2.05076113e-01 1.07512832e+00 -9.98930275e-01
1.27213907e+00 -1.35722697e+00 -3.66136611e-01 -4.95093048e-01
-5.17629623e-01 -8.05353105e-01 3.66609186e-01 -6.64160907e-01
-8.16413105e-01 -9.76639628e-01 4.15528923e-01 2.34711170e-01
-3.69774848e-02 7.19503820e-01 8.45896423e-01 -2.56795853e-01
1.14166892e+00 -1.56344548e-01 5.70032477e-01 -1.15747079e-01
9.79699790e-02 1.15997207e+00 -4.40951675e-01 3.36608738e-01
1.11328745e+00 6.87460244e-01 -4.48295474e-01 -1.81757188e+00
-4.04300988e-01 -6.79425895e-01 1.24883272e-01 -8.14427316e-01
8.16647589e-01 -5.89983642e-01 -8.55602205e-01 9.28242087e-01
-1.46630180e+00 -3.44805092e-01 4.92682070e-01 1.29623100e-01
2.49730483e-01 4.51180071e-01 -9.39901352e-01 -1.01819420e+00
-2.43710399e-01 -4.31796491e-01 -3.36490870e-02 9.03494470e-03
-7.35880017e-01 -1.15055358e+00 4.46767390e-01 6.44333541e-01
1.17783463e+00 4.06776696e-01 5.07892787e-01 -9.39810514e-01
-1.90361217e-01 -3.39042634e-01 -3.87762874e-01 -2.67557263e-01
6.39454424e-01 -1.12291478e-01 -6.57105267e-01 -3.03695768e-01
3.84224653e-01 -2.51332492e-01 2.76169062e-01 -4.17058878e-02
-1.05026901e-01 -1.32398129e+00 -5.56622028e-01 -5.05963445e-01
8.01951647e-01 2.04001591e-01 4.63174134e-01 4.94211257e-01
4.31344509e-01 1.06134796e+00 3.56432289e-01 8.39016259e-01
5.07164411e-02 2.83372194e-01 7.13207781e-01 3.25633228e-01
3.19819361e-01 -4.26304758e-01 6.09659076e-01 7.18588889e-01
-1.26394257e-01 -2.81904995e-01 -9.20153737e-01 3.59645545e-01
-1.14497256e+00 -1.26760161e+00 -4.03818607e-01 1.65271330e+00
5.06323993e-01 9.66627523e-02 7.60773003e-01 -6.56394735e-02
1.26213002e+00 6.56034708e-01 -1.26437142e-01 -2.93250501e-01
2.24658266e-01 -5.21619260e-01 8.35563540e-01 6.70814633e-01
-6.44129634e-01 9.02164698e-01 6.24718714e+00 4.76098776e-01
-1.11815417e+00 3.40235114e-01 4.03169572e-01 5.29396534e-01
-7.15939030e-02 -1.71123222e-01 -5.49239933e-01 5.59524775e-01
1.11144829e+00 -4.40666795e-01 7.29232252e-01 3.90626222e-01
9.44461107e-01 -3.85732472e-01 3.13053042e-01 1.92004129e-01
2.06558183e-01 -1.32028401e+00 -2.53216505e-01 5.98890960e-01
7.42963195e-01 3.14626604e-01 1.17087290e-02 -2.31614057e-02
3.73913050e-01 -6.56703889e-01 3.59440058e-01 -1.67079076e-01
3.00712615e-01 -4.23323750e-01 5.73892117e-01 9.31421697e-01
-6.12527251e-01 -4.09774691e-01 -1.94203168e-01 -7.16920018e-01
6.63090646e-01 4.96073037e-01 -1.38389170e+00 -2.24642336e-01
5.39448597e-02 7.25969911e-01 -5.94376624e-01 4.82077748e-01
-1.62542775e-01 8.17125916e-01 -9.42387525e-03 -4.23933804e-01
5.48331141e-01 -2.50325203e-01 1.37156272e+00 1.04814160e+00
-1.29434571e-01 7.76554346e-02 -2.11692050e-01 8.93835545e-01
1.57012880e-01 -5.72065637e-02 -1.25263596e+00 -1.04787135e+00
6.60971284e-01 9.55408514e-01 -1.08359754e+00 -1.61392778e-01
-1.22854328e-02 5.08489370e-01 6.81906044e-02 2.32567728e-01
-5.54862738e-01 -7.84957111e-02 2.81735212e-01 7.05436289e-01
-2.09707946e-01 -1.88099951e-01 -2.55485959e-02 -7.88160622e-01
-8.55592966e-01 -1.12796581e+00 -8.62225890e-02 -3.57680827e-01
-1.38913524e+00 4.78550345e-01 7.12719746e-04 -5.16792476e-01
-3.77497613e-01 -1.88948996e-02 -9.03572321e-01 5.35403967e-01
-7.63145566e-01 -7.07040071e-01 4.94327582e-02 2.92607516e-01
1.41623631e-01 -2.83122987e-01 5.17754555e-01 -2.18040884e-01
-2.69367337e-01 -8.40087980e-02 -3.57036144e-01 1.64201051e-01
1.36302456e-01 -3.58064651e-01 3.46703082e-01 4.26336378e-01
-5.58474839e-01 6.89791381e-01 9.65843916e-01 -1.44246769e+00
-9.17143703e-01 -1.14470005e+00 1.14951599e+00 -2.35279277e-01
1.43863046e+00 -1.06490456e-01 -5.03233552e-01 7.25935757e-01
5.33151865e-01 -9.66964424e-01 3.79808784e-01 -3.48938435e-01
-1.84197262e-01 6.92203581e-01 -1.52986968e+00 9.82698798e-01
7.77657151e-01 -5.53122044e-01 -2.88699657e-01 8.20631504e-01
7.98724890e-01 4.83671427e-01 -3.23478132e-01 -4.82182682e-01
2.80631363e-01 -1.03324068e+00 6.30003989e-01 -6.51406229e-01
3.84555727e-01 3.98208380e-01 3.15477580e-01 -1.32382131e+00
-3.61697525e-01 -1.22778392e+00 3.38060915e-01 8.98376405e-01
1.89396873e-01 -1.28930056e+00 4.19704080e-01 2.69913375e-01
3.04475516e-01 1.56247653e-02 -6.43863380e-01 -6.01215959e-01
-8.21161419e-02 1.98939383e-01 -6.77518919e-02 1.53423703e+00
4.26531196e-01 4.26360399e-01 -4.67929959e-01 4.44033146e-02
5.79340458e-01 -9.58075821e-01 1.04230976e+00 -9.94274259e-01
1.06342752e-02 -4.98493373e-01 -3.13561648e-01 -3.04624587e-01
-1.09385252e-01 -5.54474354e-01 -2.43341208e-01 -8.87171805e-01
2.93423142e-02 -2.21064672e-01 6.26800179e-01 -1.13297871e-03
1.65846333e-01 3.54255170e-01 4.88204509e-01 6.48264647e-01
-1.04379110e-01 3.45332436e-02 1.34126043e+00 -4.38945629e-02
-6.69776261e-01 1.23863332e-01 -8.83048654e-01 9.35849905e-01
1.10565281e+00 -9.25763488e-01 -2.36467421e-01 4.26790357e-01
8.34317625e-01 1.21303104e-01 3.88160437e-01 -4.93587881e-01
4.39804383e-02 -3.73486578e-01 -6.01116955e-01 -2.31173545e-01
1.08719230e-01 -5.19192815e-01 1.63546532e-01 1.18587303e+00
-8.99537206e-02 2.85538763e-01 9.89859551e-02 7.47472644e-01
1.89211458e-01 -2.46958524e-01 7.83683062e-01 -1.90417469e-01
3.77933174e-01 -1.94170222e-01 -1.74434876e+00 4.59669232e-01
1.06973720e+00 4.20457833e-02 -1.00301516e+00 -1.09459567e+00
-2.70601690e-01 -2.30223358e-01 4.23415750e-01 2.58500159e-01
1.52642220e-01 -7.40959525e-01 -8.25788677e-01 -2.02244133e-01
-3.66228551e-01 -9.32410419e-01 -9.02717933e-02 1.09689629e+00
-5.35826623e-01 2.91864574e-01 -2.98743129e-01 4.33970941e-03
-9.73741293e-01 5.78736544e-01 1.34349301e-01 -1.15598887e-01
-3.82993788e-01 2.83336997e-01 -7.29838759e-02 -1.97007746e-01
-1.94368050e-01 6.05251431e-01 -4.13749605e-01 4.75717127e-01
9.13744569e-01 1.13041878e+00 -6.42483234e-01 -1.20964718e+00
-2.41504878e-01 -4.43510115e-01 -5.29340208e-02 -1.29732803e-01
9.98352230e-01 -2.66766042e-01 -8.16773832e-01 5.11585355e-01
9.81731534e-01 4.95900750e-01 -2.01187596e-01 -2.44111568e-03
2.33582646e-01 -7.33763278e-01 -3.83126378e-01 -5.51021516e-01
-6.61992908e-01 1.88985914e-01 -3.97145122e-01 1.42547882e+00
2.12770268e-01 1.60038427e-01 9.17292833e-01 1.68500081e-01
2.43796945e-01 -7.00257838e-01 5.64116359e-01 5.70863783e-01
8.45289111e-01 -7.52583265e-01 -2.47542143e-01 -7.20881224e-01
-4.90301639e-01 7.11094618e-01 2.02492058e-01 -1.31357461e-01
7.72685587e-01 -1.19620062e-01 9.28860158e-03 -7.25312352e-01
-3.62893701e-01 3.21136236e-01 -6.08828783e-01 8.12397242e-01
1.32025212e-01 2.11231172e-01 -8.41868699e-01 -8.87487456e-02
2.19330937e-02 -3.06387931e-01 1.28762209e+00 6.93242908e-01
-9.29799676e-01 -4.97532517e-01 -7.74879277e-01 4.26404893e-01
-8.62431467e-01 2.58110818e-02 -1.16328013e+00 8.06512833e-01
-1.14310026e-01 1.57795644e+00 -1.44887269e-01 -7.01295257e-01
-5.03429115e-01 -3.35130304e-01 -4.72548455e-01 -5.99546850e-01
-9.36868906e-01 -2.11673260e-01 6.26621962e-01 -1.48168057e-01
-5.20373821e-01 -6.14220202e-01 -6.14594042e-01 -1.11365128e+00
-6.21128440e-01 1.83109984e-01 5.34329593e-01 1.04895449e+00
3.58358115e-01 -1.74772531e-01 9.42955077e-01 -7.47626007e-01
-5.65086126e-01 -1.57309496e+00 -6.39712751e-01 1.56318724e-01
2.35857576e-01 -4.49156493e-01 -9.44481492e-01 -4.95022267e-01] | [8.38931655883789, 10.027200698852539] |
959ccdae-740f-4a81-9936-64f4e38d00bd | camera-ppg-waveforms-at-the-forehead | 2306.09879 | null | https://arxiv.org/abs/2306.09879v1 | https://arxiv.org/pdf/2306.09879v1.pdf | Camera PPG waveforms at the forehead | In order to obtain insights into the feasibility of replacing ECG-guided triggering in magnetic resonance imaging (MRI) by a system based on video photoplethysmography (PPG), PPG and ECG data were collected from volunteers in an MRI scanner. PPG waveforms obtained using remote camera PPG directed at the forehead are studied in qualitative and quantitative sense over a number of volunteers. The data analysis considers variations in PPG waveforms across volunteers, modelling of the waveforms in Fourier series, dependencies of waveforms and features on the interbeat interval (IBI) and breath-holding, and models for ECG-blind estimation of R-peak position. The main findings are that the PPG waveform depends on the volunteer and that its shape changes with IBI and does not depend on breath-holding in the given scenario. Low-order harmonic models provide accurate approximations to the PPG waveform, where for higher IBI the waveform shows more temporal details. Accurate predictions (20 ms std) of the delays between markers in ECG and PPG appear feasible from a single PPG feature. | ['C. Possanzini', 'R. Springorum', 'J. Sénégas', 'M. Padalko', 'B. Balmaekers', 'J. H. Wülbern', 'H. H. Der Sarkissian', 'A. C. den Brinker'] | 2023-06-16 | null | null | null | null | ['photoplethysmography-ppg'] | ['medical'] | [ 1.76857561e-01 -1.36822954e-01 3.41781050e-01 -3.46478313e-01
-5.26578605e-01 -7.14244604e-01 3.10720921e-01 1.42466977e-01
-3.71571869e-01 5.00592113e-01 2.08184481e-01 -1.50222644e-01
-3.59386981e-01 -3.39841507e-02 -1.92225367e-01 -8.36158395e-01
-7.59748280e-01 3.30693841e-01 2.77403742e-01 1.44016147e-01
1.22482747e-01 6.76365197e-01 -8.37171376e-01 -6.29671067e-02
3.51761550e-01 9.21011984e-01 -4.23935242e-02 9.79388893e-01
4.31376725e-01 2.93139637e-01 -8.49198103e-01 -1.00577809e-01
2.74707943e-01 -7.17365801e-01 -4.66956675e-01 -1.95755810e-01
-7.45496824e-02 -2.68064320e-01 -4.11546230e-01 5.90654910e-01
1.02099180e+00 -2.00447515e-01 7.17091143e-01 -8.87982249e-01
1.62061844e-02 3.95704567e-01 -5.91499686e-01 6.91461265e-01
4.79750544e-01 2.43514046e-01 8.80873799e-02 -5.28574646e-01
8.98271322e-01 2.71119714e-01 1.29066527e+00 3.11302990e-01
-1.38910735e+00 -9.63585526e-02 -6.76257133e-01 4.09136504e-01
-1.30240762e+00 -2.35231236e-01 8.95635068e-01 -4.58994359e-01
7.86220372e-01 5.27481437e-01 8.78576040e-01 5.14767528e-01
7.48672485e-01 -2.94266224e-01 1.35767806e+00 -5.65379381e-01
6.70385286e-02 -2.98074959e-03 2.52095044e-01 1.40431777e-01
1.69527143e-01 6.67175725e-02 -4.94435698e-01 -5.34235001e-01
7.76363969e-01 -3.24644655e-01 -1.02982366e+00 -1.46858141e-01
-1.38024473e+00 3.63415390e-01 -2.58260906e-01 6.29665613e-01
-6.95541680e-01 1.40796989e-01 5.16257048e-01 3.40267181e-01
2.90944669e-02 3.31411600e-01 -2.44781122e-01 -5.48770249e-01
-1.06690776e+00 -1.16759032e-01 9.81225967e-01 4.89109218e-01
1.17235638e-01 -3.03038713e-02 -4.61500883e-01 5.44311881e-01
1.32664829e-01 5.74995220e-01 6.77544832e-01 -1.08616245e+00
-1.13774732e-01 -2.59626031e-01 2.30327144e-01 -9.25786555e-01
-9.54861939e-01 -5.55641279e-02 -3.61680895e-01 -2.65798926e-01
7.03849316e-01 -4.40599263e-01 -5.96962750e-01 1.17586267e+00
3.88441175e-01 5.18833213e-02 -2.97698349e-01 1.26353312e+00
8.71399641e-01 3.51385176e-01 -1.24884143e-01 -9.89348948e-01
1.57835257e+00 1.19276196e-02 -9.25045729e-01 8.74598473e-02
3.81185263e-01 -6.55643940e-01 6.06063366e-01 3.38137090e-01
-1.25462246e+00 -5.04571497e-01 -8.18548083e-01 5.90214729e-01
4.16023314e-01 -2.14042976e-01 -5.67723289e-02 8.52519333e-01
-1.10928369e+00 1.02595985e+00 -9.61341798e-01 -3.06269318e-01
-3.17581385e-01 2.44413480e-01 -2.90272832e-01 1.54309794e-01
-1.28645170e+00 1.09945023e+00 -2.96467133e-02 5.94768703e-01
-2.18750313e-01 -1.00176489e+00 -5.64802647e-01 -3.73270720e-01
-2.81851590e-01 -3.94499660e-01 1.05315578e+00 -3.84482294e-01
-1.52365994e+00 8.97832811e-01 -1.89052999e-01 -2.97183454e-01
6.79945171e-01 3.52636427e-01 -7.02219129e-01 8.88549507e-01
-3.63501281e-01 -9.00847614e-02 1.03266501e+00 -9.77472723e-01
4.88908291e-01 -4.94913012e-01 -7.77891695e-01 -3.32161710e-02
5.50257623e-01 3.26538295e-01 -8.59474763e-02 -2.49855459e-01
4.74959701e-01 -9.38585937e-01 -4.70618680e-02 -3.83613110e-01
-1.55810788e-02 4.94418979e-01 3.95966053e-01 -1.08633876e+00
1.39760816e+00 -2.08863425e+00 -3.61033738e-01 4.70605135e-01
2.71905124e-01 2.17561826e-01 2.81949252e-01 6.79574370e-01
-3.33509505e-01 -8.48658308e-02 -2.68280894e-01 4.03777540e-01
-3.19720060e-01 1.03859134e-01 -6.09788708e-02 1.17020702e+00
-3.75285298e-01 9.17620480e-01 -7.69828856e-01 -3.95493478e-01
2.18735516e-01 5.75124025e-01 7.58126378e-02 1.07887819e-01
7.13642061e-01 9.02235687e-01 -7.14271218e-02 3.90885890e-01
8.26445639e-01 1.77305728e-01 4.47773725e-01 -6.31626189e-01
-4.46238518e-01 1.44093692e-01 -8.70245576e-01 1.47708452e+00
-1.57949954e-01 6.95970297e-01 2.27310002e-01 -9.22912598e-01
1.03110719e+00 9.40645278e-01 6.73649549e-01 -8.54648530e-01
2.85975277e-01 2.89408386e-01 7.47205257e-01 -1.00253308e+00
-1.37754560e-01 -6.61435485e-01 3.83723348e-01 7.18710124e-01
-1.57335803e-01 -2.87933111e-01 1.98021397e-01 -4.12296280e-02
8.73905063e-01 3.59433740e-02 2.24818870e-01 -7.34782577e-01
3.98172528e-01 -3.66107821e-01 2.07017183e-01 6.90618932e-01
-3.99186641e-01 1.32133830e+00 6.47355258e-01 -6.66419506e-01
-7.15287089e-01 -6.61394358e-01 -6.66276097e-01 1.01626880e-01
2.06798792e-01 -2.38750532e-01 -7.11807370e-01 1.13102295e-01
-6.26206994e-02 5.83963931e-01 -5.42359531e-01 -8.86492282e-02
-7.68828273e-01 -1.00042272e+00 3.52551579e-01 2.84813315e-01
-1.77325040e-01 -1.03296292e+00 -1.36844826e+00 6.21098518e-01
-4.82868165e-01 -1.18260908e+00 -3.59936416e-01 -1.16024977e-02
-1.23406792e+00 -1.02880812e+00 -1.03719723e+00 3.88290845e-02
3.07178468e-01 -2.48165503e-01 1.05858958e+00 -2.55014509e-01
-8.05532694e-01 1.02949357e+00 -2.85363674e-01 -3.91843110e-01
-4.17599767e-01 -6.10221148e-01 6.89853951e-02 2.84329243e-02
1.34833813e-01 -9.28785861e-01 -9.05265152e-01 4.34120864e-01
-4.36759681e-01 -5.47206581e-01 1.17775731e-01 2.00505286e-01
4.92228031e-01 -5.48279822e-01 4.46445584e-01 -5.85182309e-01
7.45815396e-01 -3.57276589e-01 -2.47968271e-01 9.10778195e-02
-6.35296822e-01 -3.98204625e-01 2.86051482e-01 -5.79887152e-01
-6.66311979e-01 -4.61106181e-01 8.81993771e-02 -4.47272629e-01
-2.12321967e-01 3.38122100e-01 5.65763295e-01 -2.94148058e-01
8.41718435e-01 4.69131678e-01 3.32327396e-01 -1.52067587e-01
-7.42902681e-02 3.24757427e-01 9.74336445e-01 -4.89068061e-01
3.06535780e-01 4.33462709e-01 2.73108572e-01 -1.13254666e+00
9.50790122e-02 -3.24754894e-01 -9.30849254e-01 -6.66235983e-01
7.06688702e-01 -3.47928017e-01 -8.86446893e-01 3.65043670e-01
-1.18385005e+00 -1.26073673e-01 -4.08700496e-01 9.90908861e-01
-4.58130449e-01 7.82475591e-01 -6.64981365e-01 -1.11349142e+00
-7.42925286e-01 -9.27597463e-01 5.00076592e-01 2.65418850e-02
-6.90811217e-01 -1.12834108e+00 2.40518466e-01 5.36992662e-02
6.29752219e-01 8.07272792e-01 6.69872344e-01 -3.86538088e-01
-1.41811918e-03 -6.08481228e-01 3.87613535e-01 -3.25307064e-02
-2.01688819e-02 6.47654608e-02 -9.04014647e-01 -1.85548291e-01
8.57037485e-01 3.69497031e-01 6.07682504e-02 1.18055308e+00
5.67181170e-01 -1.08959757e-01 -1.82635058e-02 7.37892032e-01
1.29225647e+00 5.92890620e-01 8.94712090e-01 -2.82152127e-02
3.29624206e-01 7.26330400e-01 1.96080372e-01 5.78682661e-01
1.11604780e-01 3.75549287e-01 7.81619400e-02 5.55588156e-02
-9.62013286e-03 2.12691486e-01 1.81461558e-01 7.65011013e-01
-6.40681863e-01 3.35951358e-01 -9.61092174e-01 3.74129236e-01
-1.09466100e+00 -6.18516982e-01 -8.84102702e-01 2.62306547e+00
6.50519431e-01 -2.17683598e-01 3.27640980e-01 1.56793326e-01
9.32846010e-01 -2.52954848e-02 -3.41065943e-01 -6.70976102e-01
1.72670037e-01 3.22033465e-01 6.29119039e-01 4.58337486e-01
-2.44859010e-01 -3.26843947e-01 7.37577724e+00 -1.01629794e-01
-1.71400964e+00 2.71164924e-01 4.59448516e-01 9.83560607e-02
-2.86499768e-01 1.97094418e-02 -3.16615045e-01 7.68983603e-01
1.33934045e+00 -5.18569648e-01 2.36004651e-01 1.14137247e-01
5.93536258e-01 -5.27428329e-01 -9.05897558e-01 1.18625295e+00
1.15719087e-01 -1.00730467e+00 -7.16635287e-01 -1.23607740e-01
2.69775465e-02 -5.83066493e-02 -4.39137697e-01 -4.43871021e-01
-1.07313812e+00 -6.77065253e-01 5.65453231e-01 8.23462784e-01
1.28111243e+00 -2.70930380e-01 6.27282381e-01 2.90843248e-01
-8.30348313e-01 4.29018557e-01 -2.05159947e-01 1.06773376e-01
4.63316411e-01 7.01749742e-01 -7.77222991e-01 7.14151561e-01
3.56189370e-01 3.02130044e-01 -2.56869704e-01 1.24971402e+00
-1.85636520e-01 9.75748062e-01 -4.73170280e-01 6.20128334e-01
-3.38023335e-01 -4.49081779e-01 7.66090095e-01 1.06694722e+00
3.50870878e-01 8.73397708e-01 -4.10553247e-01 8.60911727e-01
8.31053495e-01 -1.44855380e-01 -4.65660155e-01 1.21252410e-01
3.96918833e-01 1.45463777e+00 -9.70703602e-01 -1.54396649e-02
-4.21510071e-01 3.89267653e-01 -7.44931161e-01 5.38787186e-01
-6.25528395e-01 -5.01922905e-01 -1.84677586e-01 1.01051879e+00
-9.09766033e-02 -1.25834113e-02 -2.07115486e-01 -7.89524972e-01
2.99245685e-01 -4.63724822e-01 2.76364118e-01 -9.87511039e-01
-7.45954692e-01 7.91177571e-01 3.08100879e-01 -1.25783265e+00
-6.82195842e-01 -1.47277817e-01 -1.06238782e+00 1.26005042e+00
-1.13008201e+00 -3.96713614e-01 -6.43928349e-02 6.02228165e-01
-3.83585244e-01 6.57297373e-01 7.96552896e-01 2.29790077e-01
1.24035329e-01 2.23983198e-01 -3.84243667e-01 -1.38041988e-01
6.00818455e-01 -1.06790698e+00 1.61985025e-01 5.69578767e-01
-3.14451784e-01 8.50922048e-01 1.09474516e+00 -5.17812550e-01
-1.45589650e+00 -1.73267126e-01 9.00350034e-01 -2.99003661e-01
4.41884160e-01 1.95253283e-01 -1.01457036e+00 2.17772603e-01
1.76668704e-01 5.38376033e-01 5.82925379e-01 -6.38624012e-01
2.03937918e-01 -1.96623370e-01 -1.39035082e+00 -1.77319691e-01
2.46701017e-01 -5.37114739e-01 -7.70690620e-01 7.77848512e-02
-5.03859390e-03 -6.97765052e-01 -1.37407148e+00 2.78698027e-01
9.69018400e-01 -1.01358449e+00 6.40677929e-01 1.41615411e-02
-1.83159813e-01 -2.03850985e-01 5.09733200e-01 -1.21100819e+00
-1.77627504e-01 -1.46003592e+00 8.19159597e-02 7.64469862e-01
-5.34693785e-02 -1.06587267e+00 2.76068717e-01 1.16976416e+00
-5.53018078e-02 -4.77747798e-01 -9.54295099e-01 -6.28984034e-01
-2.60222971e-01 -3.86801749e-01 -1.05357893e-01 7.35293090e-01
5.07019937e-01 -6.37322664e-02 -2.48906180e-01 1.59787491e-01
7.99089849e-01 1.45202234e-01 -1.09635271e-01 -1.15095651e+00
-2.58078277e-01 1.77468389e-01 -4.26057667e-01 -1.76635370e-01
-5.90083420e-01 -6.36112273e-01 -1.35837153e-01 -1.34451199e+00
-9.53964740e-02 -1.60808221e-01 -9.46078897e-02 -9.70376357e-02
-1.17279135e-01 3.39713246e-01 1.91609308e-01 2.29585677e-01
2.30065554e-01 -2.15894535e-01 1.24152541e+00 5.38077474e-01
-2.86905318e-01 1.65620327e-01 -4.83930670e-03 5.91113865e-01
2.91693956e-01 -7.04320252e-01 -3.75155449e-01 9.57869664e-02
1.71249822e-01 1.06351721e+00 2.68962830e-01 -8.72833610e-01
3.34675014e-01 5.53670228e-01 4.50513393e-01 -6.35518909e-01
1.47558436e-01 -7.29639173e-01 9.38978255e-01 6.53673589e-01
2.52893046e-02 4.19258475e-01 9.09868479e-02 5.00079915e-02
-3.22410941e-01 -5.50087392e-01 8.90016973e-01 -4.69362974e-01
8.35380033e-02 2.21832767e-01 -6.52764320e-01 4.14950550e-01
7.52825439e-01 -6.60633147e-01 1.45352101e-02 -5.86826205e-01
-1.24919939e+00 -2.46958941e-01 2.01222405e-01 -3.02716523e-01
4.80562925e-01 -1.10480034e+00 -7.06901550e-01 3.55312943e-01
-2.93987870e-01 -4.87182349e-01 6.82464302e-01 1.93528569e+00
-7.91094720e-01 2.80575275e-01 -1.84057355e-01 -8.19830358e-01
-1.05062044e+00 1.77715302e-01 9.31806564e-01 1.63406804e-02
-7.69812405e-01 5.82531273e-01 -1.42807722e-01 3.65966409e-01
-1.15254596e-01 -2.07076684e-01 -1.82536587e-01 3.56153905e-01
7.01149404e-01 4.32168394e-01 3.73612404e-01 -6.70815825e-01
-4.31463063e-01 9.38139021e-01 3.47027957e-01 -2.39651427e-01
1.18524981e+00 -4.36751693e-01 -2.26931959e-01 7.41611421e-01
1.00482166e+00 -1.76146328e-02 -1.05619752e+00 1.02236718e-01
-1.62003025e-01 -3.88349861e-01 -7.69208670e-02 -8.94331694e-01
-1.11469758e+00 1.03363109e+00 7.44208515e-01 5.19323945e-01
1.14288974e+00 -3.48391265e-01 5.80899596e-01 -4.07586306e-01
5.15052617e-01 -7.40928710e-01 -4.17556822e-01 -2.17850104e-01
9.34715986e-01 -2.74863869e-01 2.69346386e-01 -2.72107661e-01
-7.85832047e-01 1.58374345e+00 -3.34269524e-01 -8.64977688e-02
7.69894898e-01 3.17156732e-01 3.97394359e-01 -3.03684235e-01
-3.80460888e-01 4.16107982e-01 3.18680555e-01 8.33019733e-01
7.13526845e-01 -3.54469642e-02 -1.06606185e+00 4.49043453e-01
3.96042652e-02 3.37139189e-01 9.71916616e-01 6.80005908e-01
-3.44038457e-02 -7.90448248e-01 -4.98047143e-01 3.23062360e-01
-9.97592807e-01 -7.47982636e-02 1.04885295e-01 7.88939297e-01
-6.50152117e-02 6.22595668e-01 -2.11084262e-02 2.56214142e-01
6.14700019e-01 3.91741782e-01 1.03259611e+00 -2.09427357e-01
-7.12199092e-01 3.80244941e-01 7.07398430e-02 -6.16360843e-01
-5.75074017e-01 -9.46896434e-01 -1.36658919e+00 1.65539309e-01
-1.54470012e-01 3.64793301e-01 9.69539762e-01 7.33044803e-01
2.53139585e-01 2.46785283e-01 5.48896790e-01 -8.94211888e-01
-4.56748843e-01 -9.98918474e-01 -1.21582937e+00 3.38733852e-01
3.23462248e-01 -1.27134383e-01 -7.03080595e-01 2.86767632e-01] | [14.050289154052734, 3.0480542182922363] |
6f9cc6da-db04-4e0d-855b-173c36af2c37 | a-novel-deep-learning-model-for-hotel-demand | 2203.04383 | null | https://arxiv.org/abs/2203.04383v1 | https://arxiv.org/pdf/2203.04383v1.pdf | A Novel Deep Learning Model for Hotel Demand and Revenue Prediction amid COVID-19 | The COVID-19 pandemic has significantly impacted the tourism and hospitality sector. Public policies such as travel restrictions and stay-at-home orders had significantly affected tourist activities and service businesses' operations and profitability. To this end, it is essential to develop an interpretable forecast model that supports managerial and organizational decision-making. We developed DemandNet, a novel deep learning framework for predicting time series data under the influence of the COVID-19 pandemic. The framework starts by selecting the top static and dynamic features embedded in the time series data. Then, it includes a nonlinear model which can provide interpretable insight into the previously seen data. Lastly, a prediction model is developed to leverage the above characteristics to make robust long-term forecasts. We evaluated the framework using daily hotel demand and revenue data from eight cities in the US. Our findings reveal that DemandNet outperforms the state-of-art models and can accurately predict the impact of the COVID-19 pandemic on hotel demand and revenues. | ['Zhishan Guo', 'Arthur Huang', 'Ashkan Farhangi'] | 2022-03-08 | null | null | null | null | ['covid-19-modelling', 'time-series-prediction'] | ['time-series', 'time-series'] | [-7.25549400e-01 -2.26638079e-01 -4.11036581e-01 -4.62909669e-01
-3.52329165e-01 -5.30823350e-01 8.16679239e-01 9.46863219e-02
-1.13586538e-01 4.54270691e-01 9.11567748e-01 -9.01610732e-01
-2.09689304e-01 -8.60694051e-01 -3.81063849e-01 -5.55875003e-01
-4.83777612e-01 5.13440430e-01 -6.94002271e-01 -6.02850258e-01
3.55328053e-01 5.69280326e-01 -1.27169704e+00 1.55620977e-01
8.93989921e-01 1.08859086e+00 8.73442292e-02 2.25464642e-01
-4.86849323e-02 7.00710475e-01 -2.88565457e-01 -9.39710066e-02
5.60707748e-01 1.80686310e-01 -4.47677612e-01 -2.04508483e-01
-5.53501725e-01 -6.51241422e-01 -4.57323372e-01 2.71450132e-01
3.62584174e-01 1.30514070e-01 7.25148439e-01 -1.46484947e+00
-8.84552598e-01 5.48412919e-01 -2.21001074e-01 6.12154305e-01
6.00280762e-02 4.73630607e-01 9.00901854e-01 -8.42459977e-01
3.80396277e-01 1.05900562e+00 6.55179918e-01 -1.36540189e-01
-1.02666783e+00 -7.67299235e-01 4.92190003e-01 1.80449039e-01
-1.13547671e+00 -1.70162275e-01 5.91750383e-01 -7.18780279e-01
1.44417834e+00 2.11192816e-01 8.80816340e-01 1.11728311e+00
6.42610312e-01 6.41748726e-01 8.09694707e-01 1.11116268e-01
2.35722940e-02 -9.95729491e-02 -2.53533572e-01 1.85324684e-01
8.86460692e-02 4.29796189e-01 -3.32322791e-02 1.92773178e-01
3.24124694e-01 6.79096818e-01 2.09879696e-01 5.40422499e-01
-9.02061105e-01 1.09829581e+00 6.80860639e-01 1.10884599e-01
-9.92980659e-01 -9.42899957e-02 4.30717707e-01 3.79006237e-01
9.55013752e-01 4.76311266e-01 -8.27063560e-01 -3.31431866e-01
-6.71147346e-01 1.61171883e-01 7.77087390e-01 4.94708687e-01
4.18747962e-01 4.95750338e-01 1.74865842e-01 4.13628280e-01
1.79611012e-01 8.20973933e-01 3.64709318e-01 -4.73141879e-01
2.88653523e-01 4.56710279e-01 3.14294338e-01 -1.45963728e+00
-9.15620267e-01 -7.70887077e-01 -6.62891567e-01 -4.23182189e-01
3.85476230e-03 -4.44872200e-01 -9.71913159e-01 1.35909855e+00
-2.26170495e-01 2.72664696e-01 -4.16087266e-03 8.67449164e-01
5.09110570e-01 1.00713563e+00 -9.52919051e-02 -1.92452520e-01
7.25860655e-01 -8.50846291e-01 -7.18037724e-01 -2.36106604e-01
8.18173349e-01 -4.29998338e-01 9.11379397e-01 1.76918179e-01
-7.05868542e-01 -2.35184893e-01 -3.16242516e-01 3.12597394e-01
-8.48702669e-01 -3.08147132e-01 7.19260514e-01 3.26703861e-02
-9.66430902e-01 2.85720453e-02 -8.53946567e-01 -2.27533534e-01
1.35679483e-01 2.66369611e-01 1.53698057e-01 -5.30240498e-02
-1.25421929e+00 1.07677221e+00 1.36020139e-01 3.18112940e-01
-1.13427806e+00 -7.80510008e-01 -8.39527547e-01 3.00016463e-01
-3.99153791e-02 -4.01396871e-01 1.12698102e+00 -7.05642521e-01
-7.90824056e-01 3.03895980e-01 -9.21713039e-02 -6.44378424e-01
1.46197483e-01 1.30572379e-01 -1.13080871e+00 -4.75707769e-01
2.48735666e-01 1.41391695e-01 2.84156024e-01 -1.10073054e+00
-8.67713273e-01 -4.55210507e-01 -4.79507931e-02 4.30246666e-02
-3.14670146e-01 -6.12065159e-02 5.59563488e-02 -5.43431103e-01
-2.88038999e-01 -1.19676924e+00 -3.14551353e-01 -1.16340709e+00
-3.01416039e-01 -2.38611385e-01 7.19442368e-01 -8.76038194e-01
1.48049998e+00 -2.07602310e+00 -3.40674132e-01 3.41619551e-01
1.88178977e-03 -2.11167876e-02 -1.93738386e-01 1.04843986e+00
-2.37868447e-02 3.54574740e-01 3.75607252e-01 -3.10003757e-01
1.85913309e-01 5.64874828e-01 -8.12451124e-01 3.80507588e-01
1.89913765e-01 9.72218037e-01 -8.48945856e-01 2.64079243e-01
2.80575842e-01 4.33285236e-01 -5.94185233e-01 4.23341617e-02
-2.03918651e-01 6.00487649e-01 -3.24167013e-01 9.16187227e-01
6.48645103e-01 -3.59729290e-01 1.67565551e-02 2.50145257e-01
-7.58994997e-01 6.29240692e-01 -2.39745528e-01 9.19013143e-01
-8.64120662e-01 8.51810396e-01 -2.92582572e-01 -9.88831460e-01
1.00268614e+00 1.03199311e-01 7.36815333e-01 -1.35253346e+00
9.40303877e-02 3.60964715e-01 1.34563491e-01 -7.50721157e-01
6.27603650e-01 -4.75538895e-02 -3.99450183e-01 5.45890152e-01
-5.56064606e-01 2.38326550e-01 -5.57150692e-02 -1.25586107e-01
7.37280726e-01 -5.25145710e-01 -7.84391351e-03 -2.93116093e-01
6.56352267e-02 3.29877526e-01 7.77985692e-01 1.55217603e-01
8.41584615e-03 -8.23410749e-02 4.38068956e-01 -1.15044129e+00
-8.71190310e-01 -8.17491412e-01 -2.76453346e-02 1.20718229e+00
-2.43599713e-01 -4.70941924e-02 -6.44906834e-02 -3.20068181e-01
4.77158308e-01 1.25639784e+00 -8.54990423e-01 -1.28282785e-01
-3.82329583e-01 -7.15411484e-01 2.02436164e-01 6.08284235e-01
1.26129389e-01 -8.37008357e-01 -3.91434550e-01 2.65857041e-01
-4.30674642e-01 -7.78470695e-01 -4.23262566e-01 1.76851064e-01
-4.08648282e-01 -6.88993990e-01 -1.50039643e-01 -5.82970798e-01
5.92577338e-01 3.82040977e-01 1.15408242e+00 -1.94677293e-01
5.53840920e-02 2.95891285e-01 -1.32204995e-01 -9.14509714e-01
6.03663959e-02 3.26223046e-01 3.88522655e-01 -7.95475319e-02
7.22479582e-01 -4.88161951e-01 -8.98633003e-01 4.28574830e-01
-7.12058961e-01 -2.09630638e-01 3.77014697e-01 6.82137489e-01
3.99505854e-01 4.08304036e-01 1.19809198e+00 -1.31533608e-01
9.53688443e-01 -1.33007801e+00 -5.34994304e-01 -2.46406287e-01
-1.24850476e+00 -3.47113729e-01 8.48301649e-01 -7.56214047e-03
-8.62238824e-01 -4.73901451e-01 -1.22364670e-01 -1.81786925e-01
9.46510360e-02 1.33751774e+00 5.12599707e-01 4.65802401e-01
1.45565331e-01 3.44258100e-01 3.20385247e-02 -4.32126105e-01
1.86656564e-02 8.14476967e-01 3.81098151e-01 -1.60660539e-02
7.81546652e-01 6.81019843e-01 -2.70327568e-01 -7.26634383e-01
-6.30305350e-01 -6.72520101e-01 -4.33036536e-01 -2.73860127e-01
6.34813190e-01 -1.20673585e+00 -8.08319449e-01 3.20767313e-01
-8.45668912e-01 -4.70554769e-01 9.01507214e-02 5.15817523e-01
-3.06114376e-01 -5.99120259e-01 -5.17439306e-01 -9.96956527e-01
-1.67550951e-01 -1.15706062e+00 5.86396873e-01 6.61550462e-02
-8.41820389e-02 -1.52918482e+00 2.66429454e-01 1.23816654e-01
9.38484788e-01 6.16012454e-01 9.46624279e-01 -9.29232478e-01
-4.39646065e-01 -2.98768908e-01 -1.45207375e-01 1.49066269e-01
4.65982884e-01 7.60468096e-02 -5.75675189e-01 -3.18807572e-01
-2.56312430e-01 9.56219882e-02 7.25346565e-01 7.82747269e-01
1.01698494e+00 -9.73039687e-01 -3.69831800e-01 8.64652097e-01
1.34518993e+00 7.36564636e-01 3.31130236e-01 8.23046327e-01
3.68304342e-01 7.85956144e-01 7.63650298e-01 9.63536620e-01
1.17565966e+00 2.21336022e-01 1.09140730e+00 -3.70881736e-01
6.75816298e-01 -3.62430245e-01 3.39677185e-01 9.38214719e-01
-1.61726281e-01 -3.86948526e-01 -1.46610403e+00 9.64245856e-01
-1.55393934e+00 -1.12229371e+00 -9.94923562e-02 1.67748475e+00
9.97235626e-02 1.03606589e-01 4.11365181e-01 -3.81231695e-01
-1.12057790e-01 3.02796572e-01 -9.23062623e-01 -9.19699311e-01
-7.70156756e-02 -4.03786719e-01 8.27545524e-01 2.45326728e-01
-6.34994507e-01 8.17061603e-01 6.99799347e+00 -1.39437107e-04
-1.49215519e+00 -1.76046595e-01 9.69286501e-01 -2.76730180e-01
-8.95208836e-01 -2.99475938e-01 -6.06871784e-01 6.29993260e-01
1.47674441e+00 -6.53943300e-01 7.04123318e-01 8.61650586e-01
1.26963770e+00 3.43122363e-01 -6.06487930e-01 5.98188996e-01
-1.07158653e-01 -1.53164256e+00 -4.23632897e-02 7.61122465e-01
9.96433854e-01 6.98980033e-01 8.06603014e-01 7.11601615e-01
5.51994979e-01 -1.30700779e+00 2.87050605e-01 6.89262986e-01
5.85782170e-01 -1.22627807e+00 9.02092636e-01 3.62281382e-01
-1.02761912e+00 -8.37243855e-01 -1.00487564e-02 -5.98697662e-01
4.42612022e-01 3.62403959e-01 -1.39717388e+00 2.89333463e-01
6.19435251e-01 9.68079507e-01 -1.59988776e-01 5.78134239e-01
3.39829624e-01 7.73289323e-01 9.28033963e-02 3.20736200e-01
9.29225028e-01 -3.56813461e-01 3.11515361e-01 1.11995363e+00
6.64839029e-01 1.08702797e-02 1.75979003e-01 5.71649134e-01
6.16047457e-02 -9.60562080e-02 -1.15997887e+00 -3.53532821e-01
3.97674888e-01 9.45690393e-01 -2.10404947e-01 -3.15043330e-02
-7.23450601e-01 3.52187574e-01 -2.34402120e-01 6.77057385e-01
-9.33881998e-01 9.10119787e-02 1.34063423e+00 3.13683212e-01
3.58946651e-01 -6.87802553e-01 -2.31750563e-01 -9.65493083e-01
-3.91287953e-01 -7.73270845e-01 2.67651618e-01 -5.95993400e-01
-1.21535087e+00 4.99240786e-01 -2.37920478e-01 -1.14183402e+00
-7.82264769e-01 -4.12032038e-01 -8.89661729e-01 8.94073486e-01
-2.10575795e+00 -1.24357212e+00 -1.98175922e-01 5.03224730e-01
5.71707189e-01 -1.22138411e-01 8.22427511e-01 2.22326785e-01
-8.55046511e-01 1.21000960e-01 6.92299724e-01 4.87974882e-02
3.63806784e-01 -9.14521217e-01 6.73464358e-01 7.71063089e-01
-6.52182698e-01 7.15024114e-01 6.46413684e-01 -7.83481538e-01
-1.42393029e+00 -1.37181950e+00 1.17618346e+00 -2.46485740e-01
1.00845826e+00 -1.46653324e-01 -5.33867836e-01 1.21020710e+00
2.62653172e-01 -5.24695277e-01 1.14501870e+00 2.42892489e-01
-5.23383692e-02 -4.01327372e-01 -9.91822064e-01 4.70402420e-01
7.52333224e-01 -4.99721408e-01 -2.86099464e-01 5.01365423e-01
1.06565475e+00 2.83447579e-02 -9.13883328e-01 1.23578325e-01
6.94398344e-01 -7.52767026e-01 7.98919141e-01 -9.26194370e-01
4.78450954e-01 3.80797178e-01 -3.48826945e-01 -1.87539089e+00
-7.19752550e-01 -5.75590312e-01 9.78861898e-02 7.47599781e-01
6.63766980e-01 -9.93056059e-01 2.40459397e-01 9.71787572e-01
-6.14806950e-01 -7.30266452e-01 -7.21059918e-01 -9.59122062e-01
7.38911554e-02 -4.64126587e-01 1.42987335e+00 1.18059874e+00
-1.28209770e-01 -1.92613184e-01 -5.77948928e-01 3.07095826e-01
5.43165430e-02 3.42530578e-01 8.53800356e-01 -1.34545171e+00
3.99583310e-01 -5.16154289e-01 -1.66283287e-02 -8.33592772e-01
2.24464148e-01 -8.35614264e-01 -5.12215495e-01 -1.85802162e+00
-3.27402055e-01 -4.06119198e-01 -6.25929296e-01 5.13287961e-01
5.79675257e-01 -2.02803850e-01 -9.00945216e-02 3.12674940e-01
-9.35686529e-02 7.67918885e-01 1.12012196e+00 -9.29923505e-02
-5.38632035e-01 1.37886479e-01 -9.92990792e-01 4.91862208e-01
1.14407444e+00 -1.44599406e-02 -4.12888795e-01 -7.49618053e-01
6.13604844e-01 1.32510126e-01 1.68921977e-01 -4.12412286e-01
-1.45509811e-02 -8.54811549e-01 1.66799277e-01 -1.06163120e+00
2.19104126e-01 -1.02667832e+00 2.37432986e-01 6.41211629e-01
-8.39952901e-02 1.03058147e+00 5.01978755e-01 5.05456805e-01
-3.37153643e-01 8.35306287e-01 5.13216481e-02 1.59686357e-01
-8.04211080e-01 6.21574700e-01 -8.34763169e-01 -2.19634533e-01
1.08770597e+00 -1.47534996e-01 -5.09339035e-01 -8.18175197e-01
-5.09532034e-01 7.38735080e-01 3.21029931e-01 7.52666235e-01
6.21292770e-01 -1.32897890e+00 -8.25322509e-01 5.16225696e-01
6.53363243e-02 -2.53615588e-01 2.95547038e-01 1.09072769e+00
-3.03319603e-01 1.22459078e+00 -3.68302196e-01 -6.47878572e-02
-4.53931838e-01 7.42444932e-01 1.68508589e-01 -2.13559374e-01
-2.54158050e-01 3.13182294e-01 9.68632996e-02 -5.77510595e-01
-5.50320894e-02 -8.18794250e-01 -3.22393388e-01 4.35909182e-01
3.41133565e-01 5.74967265e-01 -6.73644096e-02 -9.11105454e-01
-4.84413445e-01 -9.73287448e-02 1.50696244e-02 9.18493867e-02
1.97345066e+00 -6.92778647e-01 3.71846855e-02 6.43122017e-01
1.39595747e+00 -2.04565570e-01 -1.20069492e+00 4.53794934e-02
-1.71908811e-02 -5.98457694e-01 3.50344300e-01 -1.07751787e+00
-1.12074494e+00 4.47044075e-01 2.08567321e-01 7.08183467e-01
1.29340720e+00 -3.66027236e-01 1.28068364e+00 5.34944311e-02
9.37357992e-02 -1.10617173e+00 -9.95334461e-02 7.01288342e-01
9.92990792e-01 -1.37715137e+00 -5.66253841e-01 4.76866394e-01
-7.91950941e-01 8.42954576e-01 3.52932960e-01 4.79284786e-02
9.16632473e-01 6.90844953e-02 8.35600719e-02 -4.59210426e-01
-1.13617373e+00 -1.66370913e-01 2.29562402e-01 5.12487352e-01
2.50273496e-01 6.74709558e-01 -1.23765737e-01 6.22264326e-01
-4.87623841e-01 -1.28145128e-01 5.56990743e-01 3.79509062e-01
-1.40540838e-01 -5.88027358e-01 -1.25019237e-01 7.35458076e-01
-3.70713741e-01 -3.63722771e-01 -1.51086882e-01 1.08920109e+00
-1.15712762e-01 1.21339321e+00 6.25227630e-01 -8.36301148e-01
2.03343078e-01 -1.24645613e-01 -6.89956367e-01 -2.55033255e-01
-4.45055306e-01 -1.54211959e-02 2.62193441e-01 -6.46205783e-01
-1.22451093e-02 -9.47078228e-01 -1.18111467e+00 -8.98794711e-01
1.35196626e-01 -8.45025703e-02 1.00612187e+00 8.74859810e-01
8.43853772e-01 5.93513727e-01 1.32968414e+00 -6.90608740e-01
-4.51589286e-01 -9.07897055e-01 -6.11579180e-01 7.45674670e-02
8.50395679e-01 -6.65593982e-01 -7.25544572e-01 -3.26189816e-01] | [6.649960517883301, 2.841557741165161] |
51832699-2bbe-49e3-89ce-047ee131b5c9 | skelevision-towards-adversarial-resiliency-of | 2204.00734 | null | https://arxiv.org/abs/2204.00734v1 | https://arxiv.org/pdf/2204.00734v1.pdf | SkeleVision: Towards Adversarial Resiliency of Person Tracking with Multi-Task Learning | Person tracking using computer vision techniques has wide ranging applications such as autonomous driving, home security and sports analytics. However, the growing threat of adversarial attacks raises serious concerns regarding the security and reliability of such techniques. In this work, we study the impact of multi-task learning (MTL) on the adversarial robustness of the widely used SiamRPN tracker, in the context of person tracking. Specifically, we investigate the effect of jointly learning with semantically analogous tasks of person tracking and human keypoint detection. We conduct extensive experiments with more powerful adversarial attacks that can be physically realizable, demonstrating the practical value of our approach. Our empirical study with simulated as well as real-world datasets reveals that training with MTL consistently makes it harder to attack the SiamRPN tracker, compared to typically training only on the single task of person tracking. | ['Duen Horng Chau', 'Sheng-Yun Peng', 'Nilaksh Das'] | 2022-04-02 | null | null | null | null | ['sports-analytics'] | ['computer-vision'] | [ 5.77684119e-02 -1.11360647e-01 1.41221136e-01 2.42116123e-01
-6.07252777e-01 -9.16009784e-01 5.73300838e-01 -1.91019908e-01
-5.25966883e-01 6.36446655e-01 1.94765832e-02 -3.25971514e-01
-3.03794146e-02 -3.03899825e-01 -9.31383729e-01 -4.92013097e-01
-6.33296132e-01 1.53930530e-01 3.93052667e-01 -1.67283416e-01
-3.67884606e-01 6.40786171e-01 -1.21398830e+00 -4.96635050e-01
3.02325010e-01 6.47421479e-01 -8.18748236e-01 1.04982424e+00
8.77524197e-01 4.69020128e-01 -9.99680638e-01 -8.64635050e-01
8.73586237e-01 1.36887506e-01 -2.54360318e-01 -1.10953152e-01
1.20030928e+00 -5.13727129e-01 -6.96120858e-01 9.91800427e-01
7.61146963e-01 1.88363880e-01 2.39851907e-01 -1.93487298e+00
-2.29772076e-01 1.70050427e-01 -6.40496552e-01 3.84631932e-01
5.53388000e-01 5.63457906e-01 6.71565056e-01 -8.77003651e-03
2.10656852e-01 1.49812746e+00 1.25042045e+00 7.30142772e-01
-1.06327748e+00 -1.17413104e+00 1.96730807e-01 -2.67826825e-01
-1.16984010e+00 -4.26719487e-01 4.86750931e-01 -4.91708934e-01
3.64664286e-01 1.93807289e-01 3.97897601e-01 1.62099755e+00
3.58729064e-01 5.92646122e-01 9.68565583e-01 -6.13980070e-02
3.49561647e-02 -8.32339451e-02 1.16681829e-01 7.21349597e-01
8.79673600e-01 9.40825343e-01 -3.34155232e-01 -7.74598420e-01
5.58868945e-01 -2.29283813e-02 6.80017769e-02 -4.03405219e-01
-9.64071572e-01 7.87664533e-01 2.77639896e-01 -2.45667785e-01
-5.23812175e-02 6.98747873e-01 5.30565739e-01 2.88526863e-01
1.24310829e-01 3.36400688e-01 -3.11823815e-01 5.97689413e-02
-6.37648165e-01 8.34792733e-01 7.80502677e-01 8.10382068e-01
8.71235803e-02 2.49885336e-01 -1.81857958e-01 5.97194135e-02
3.53417009e-01 1.04969180e+00 -1.36859626e-01 -1.11180055e+00
4.72256064e-01 3.22966464e-02 4.29260522e-01 -1.04778385e+00
-4.32583928e-01 -3.19040120e-01 -5.92431664e-01 5.39310992e-01
9.68132973e-01 -7.48049915e-01 -6.25190914e-01 1.97780645e+00
5.12304187e-01 6.39981985e-01 4.04904112e-02 7.70928502e-01
2.94700801e-01 1.28053471e-01 4.55825359e-01 2.50000477e-01
1.37209928e+00 -5.69395185e-01 -4.09128636e-01 -5.39881468e-01
3.89284134e-01 -6.37588084e-01 6.03540182e-01 2.98768222e-01
-7.84705341e-01 -4.96211737e-01 -9.78289425e-01 3.24883997e-01
-1.80550709e-01 -2.78329849e-01 6.75074220e-01 1.45207393e+00
-6.35221601e-01 2.50323653e-01 -8.80512416e-01 -3.71419370e-01
4.35910553e-01 5.22908330e-01 -3.27348799e-01 1.02904610e-01
-1.41791582e+00 9.94984925e-01 2.57494897e-02 -9.92028192e-02
-9.77884948e-01 -7.91088879e-01 -9.65616822e-01 -2.28059322e-01
5.08855700e-01 -6.97216630e-01 1.16628921e+00 -5.29578209e-01
-1.07225215e+00 5.65226436e-01 3.13231856e-01 -9.59193587e-01
8.56536329e-01 -7.35759318e-01 -5.45468330e-01 2.51869373e-02
2.49904003e-02 5.09559631e-01 1.20966482e+00 -9.27860498e-01
-5.17591238e-01 -4.68376458e-01 3.70755315e-01 -5.04298471e-02
-3.38151157e-01 1.85256064e-01 -8.10017064e-02 -8.47853541e-01
-7.94430017e-01 -1.50666988e+00 -3.35456312e-01 4.03016388e-01
-3.90603930e-01 -6.45375997e-02 1.08943474e+00 -6.37364388e-01
5.44313371e-01 -2.34234023e+00 -2.32843220e-01 2.90552229e-01
1.89104855e-01 4.04415041e-01 -3.99085619e-02 6.18384369e-02
2.24084303e-01 -2.09865585e-01 3.28075320e-01 -2.84105033e-01
2.03461111e-01 1.52497459e-02 -4.52374876e-01 8.74641776e-01
-1.67010441e-01 8.84151876e-01 -7.43189275e-01 -3.90595496e-01
1.55297473e-01 4.25878495e-01 -3.91072839e-01 -5.34285679e-02
-5.31785451e-02 4.91906852e-01 -5.68079829e-01 6.22113287e-01
4.93560076e-01 8.79400447e-02 -5.02211936e-02 7.76296407e-02
3.40250582e-01 -4.31682676e-01 -1.26226914e+00 1.16098428e+00
-1.39718419e-02 6.90172613e-01 1.80777147e-01 -3.87610942e-01
3.75631809e-01 3.10478628e-01 5.36380768e-01 -3.40939403e-01
2.02392116e-01 -2.89550394e-01 8.51730481e-02 -1.20027073e-01
3.71464401e-01 -1.06104054e-01 -6.22340798e-01 4.10608560e-01
-1.90144256e-01 4.93526369e-01 -1.30162820e-01 3.47173899e-01
1.33567047e+00 -1.40076624e-02 6.34703934e-02 -1.48201272e-01
3.31646532e-01 -4.86218445e-02 3.94819170e-01 1.14354312e+00
-8.52063656e-01 -5.04162423e-02 1.19915254e-01 -5.13170540e-01
-9.20677185e-01 -1.34302604e+00 1.86979651e-01 1.19382226e+00
3.99325430e-01 -1.83621511e-01 -7.84408212e-01 -9.29675639e-01
6.81680501e-01 3.15138072e-01 -5.98092914e-01 -5.29382885e-01
-7.94413567e-01 -5.78034580e-01 1.56009519e+00 6.68241084e-01
4.19366628e-01 -6.20248973e-01 -8.04537058e-01 -1.56450421e-01
1.59779400e-01 -1.64872694e+00 -7.31502116e-01 -5.36969960e-01
-3.35777491e-01 -1.29563236e+00 -6.43897295e-01 -2.14456275e-01
3.05875838e-01 2.42371604e-01 7.26691782e-01 1.31071866e-01
-5.15994787e-01 9.92988348e-01 1.73453957e-01 -6.63829744e-01
-4.66292351e-01 -1.44317493e-01 6.26170516e-01 -9.10162106e-02
1.45976007e-01 -1.85248017e-01 -3.96422803e-01 3.09038401e-01
-5.55538774e-01 -6.75992966e-01 3.84026915e-01 3.57979149e-01
-1.08949728e-01 2.19189972e-01 4.47475702e-01 -7.38484025e-01
5.83401978e-01 -2.66698480e-01 -8.87729704e-01 2.47969821e-01
-2.49843881e-01 -4.97728959e-02 4.39835131e-01 -8.89011800e-01
-6.70492649e-01 1.40722170e-01 7.83097669e-02 -6.41390145e-01
-2.33972311e-01 -3.93555790e-01 -1.44318715e-01 -7.84596086e-01
8.59639883e-01 -1.31035224e-01 -7.14199501e-04 -6.09370768e-02
4.13942605e-01 2.68416796e-02 7.86343575e-01 -7.12789297e-01
1.58936954e+00 6.55040503e-01 4.01359349e-01 -8.07832241e-01
-8.42306256e-01 -2.28706986e-01 -1.98361203e-01 -4.72583771e-01
8.90493155e-01 -1.00752211e+00 -1.36107302e+00 5.68790674e-01
-8.93569052e-01 -2.24384889e-01 -4.33951989e-02 3.26034546e-01
-4.23739761e-01 7.74568319e-01 -6.06452823e-01 -8.83424103e-01
-3.14424694e-01 -9.43334341e-01 1.06665266e+00 8.85367990e-02
-2.76231676e-01 -1.13465929e+00 2.43962973e-01 4.96178061e-01
2.30421916e-01 7.84510791e-01 1.61384419e-01 -7.39722073e-01
-5.72356045e-01 -5.90510190e-01 4.39052992e-02 4.23520617e-02
-1.55329421e-01 -4.40756023e-01 -9.20887709e-01 -8.99931252e-01
-2.19251052e-01 -4.74760711e-01 5.60452819e-01 2.57097125e-01
7.20142901e-01 -4.39089477e-01 -5.86786628e-01 5.47057748e-01
1.09630120e+00 -2.52420396e-01 3.40990245e-01 4.80528891e-01
8.34892094e-01 4.18998331e-01 5.71781933e-01 2.48313591e-01
1.16389908e-01 8.38826180e-01 2.56229013e-01 5.14313020e-02
-3.94382924e-02 -2.10857213e-01 6.99683666e-01 -6.00082934e-01
-1.75949745e-02 -2.18575031e-01 -7.21617997e-01 6.82327151e-02
-1.79122114e+00 -1.14281166e+00 1.92640796e-01 2.29998398e+00
3.95602733e-01 5.99050462e-01 7.19781220e-01 5.58683695e-03
7.83523381e-01 2.22621113e-01 -5.30182660e-01 2.38023922e-02
1.80642530e-01 1.17914984e-02 1.15462101e+00 4.31325078e-01
-1.50629056e+00 7.87854195e-01 7.16115570e+00 5.50863266e-01
-7.73450911e-01 3.10962647e-01 2.09284481e-02 -4.02550399e-01
5.60859144e-01 -3.20758700e-01 -1.11049640e+00 5.55817783e-01
1.01286948e+00 -1.04072683e-01 2.45027810e-01 7.23864913e-01
-6.48956373e-02 1.52252465e-01 -1.11785555e+00 7.86299586e-01
1.46651879e-01 -8.92452776e-01 -2.99254894e-01 3.30358297e-01
4.31660354e-01 -2.16412693e-01 4.39287215e-01 3.54058117e-01
7.64938712e-01 -8.29978406e-01 7.24529088e-01 2.16217980e-01
4.10547227e-01 -8.79377127e-01 5.10776460e-01 3.37646455e-01
-1.26012301e+00 -2.54429251e-01 3.46680707e-03 5.84030431e-03
3.41799587e-01 8.06427896e-02 -4.82681364e-01 3.33428591e-01
6.89646423e-01 2.60035127e-01 -6.91909492e-01 8.57461333e-01
-2.20785197e-02 5.41243672e-01 -6.11892462e-01 3.24576795e-01
9.98093784e-02 4.14388001e-01 8.92332315e-01 1.10669696e+00
-1.66560020e-02 -9.90548432e-02 6.57838583e-01 3.61310065e-01
4.74956408e-02 -5.86950958e-01 -9.48632360e-01 2.22214341e-01
3.41553897e-01 1.00690961e+00 -4.97363120e-01 -8.17387179e-02
-2.98536479e-01 7.45486081e-01 5.50128184e-02 2.35072583e-01
-1.32362175e+00 7.73070306e-02 1.08340538e+00 2.03399673e-01
2.98470885e-01 -4.51032758e-01 -1.38205200e-01 -9.18708384e-01
9.61033702e-02 -9.76742387e-01 6.85032964e-01 -1.77458838e-01
-1.36254704e+00 1.99800909e-01 1.75672248e-01 -1.04919779e+00
-2.68695295e-01 -5.50520301e-01 -6.75605536e-01 6.00269496e-01
-1.05309689e+00 -1.45219195e+00 -4.64367494e-02 9.11984801e-01
1.22861311e-01 -4.18730050e-01 4.49559957e-01 4.03174102e-01
-5.68013728e-01 1.24234855e+00 -1.99505851e-01 3.98432404e-01
7.47570813e-01 -9.09100652e-01 8.51362348e-01 1.33494258e+00
2.50146841e-04 5.28339922e-01 1.04423094e+00 -1.02724719e+00
-1.44479823e+00 -1.25101614e+00 2.13792384e-01 -9.73650217e-01
8.73109400e-01 -5.73468924e-01 -5.05847752e-01 9.06216562e-01
-2.44918838e-01 1.56442225e-01 5.15963912e-01 7.04888627e-02
-6.55396581e-01 2.44103651e-02 -1.43781817e+00 6.96281910e-01
1.03175843e+00 -5.26325822e-01 -5.11196077e-01 4.66618508e-01
7.19732225e-01 -4.58607286e-01 -7.57083952e-01 3.15034121e-01
8.48227382e-01 -5.09867311e-01 1.56351650e+00 -7.16535509e-01
-4.34921980e-01 -3.14735562e-01 8.75161439e-02 -7.49641657e-01
-2.70248473e-01 -8.66214335e-01 -5.41489482e-01 9.17593658e-01
-1.28082037e-01 -7.44858682e-01 1.03153419e+00 7.63862610e-01
4.88827109e-01 -6.93344176e-02 -1.06452942e+00 -1.32909155e+00
9.30532888e-02 -3.07263076e-01 2.76339203e-01 7.63025880e-01
-5.43403566e-01 1.94249284e-02 -8.04630220e-01 9.82734144e-01
1.41184700e+00 -3.96139503e-01 1.06933773e+00 -1.37089455e+00
-4.91015047e-01 2.57562124e-03 -8.19442213e-01 -5.32776296e-01
3.70523930e-01 -3.12062293e-01 -8.03456381e-02 -5.88957906e-01
2.66217347e-02 -3.94051045e-01 -5.46987541e-02 3.50157887e-01
-2.78704792e-01 4.59792197e-01 5.84129393e-01 1.32097304e-01
-7.56118596e-01 -1.27757052e-02 7.10409105e-01 -2.63169110e-01
1.83863819e-01 4.86345321e-01 -5.93279898e-01 8.14393520e-01
6.43943012e-01 -6.09758794e-01 -2.06373796e-01 -1.60188422e-01
-5.14737666e-02 -1.32726490e-01 1.09159923e+00 -1.40769124e+00
4.23894972e-01 7.47042522e-02 4.85929191e-01 -2.54749686e-01
5.72458088e-01 -1.06697679e+00 1.35725334e-01 9.51427817e-01
-1.57980040e-01 1.98945999e-01 5.82596600e-01 9.12255585e-01
5.50614417e-01 2.03875050e-01 8.48962486e-01 1.09392241e-01
-4.92329717e-01 4.33238626e-01 -1.82992041e-01 2.01225460e-01
1.34771335e+00 -2.21107841e-01 -4.76055235e-01 -3.97023410e-01
-7.06463635e-01 3.52971733e-01 4.84042645e-01 6.64454818e-01
3.68852377e-01 -1.21852028e+00 -6.69882059e-01 1.46076962e-01
-1.04863346e-01 -5.96832275e-01 1.36259452e-01 4.16849971e-01
-1.71879619e-01 2.66443729e-01 -2.92317629e-01 -4.05544490e-01
-1.57914734e+00 9.37134326e-01 4.56843078e-01 -2.97691643e-01
-7.76319623e-01 5.26240349e-01 1.92639157e-01 -2.73904413e-01
5.44938982e-01 1.47315294e-01 1.96155310e-01 -5.20038486e-01
5.00090480e-01 6.78501844e-01 -3.08038861e-01 -6.42311156e-01
-5.93819201e-01 4.06034291e-01 -1.78734168e-01 -2.15170577e-01
6.40264988e-01 1.65348679e-01 4.35875386e-01 -1.98958114e-01
6.64649963e-01 1.71689659e-01 -1.56454253e+00 -9.39996168e-02
4.79053892e-02 -3.72355312e-01 -3.23470086e-01 -5.20211041e-01
-9.31655467e-01 3.31880301e-01 9.92816865e-01 3.04498672e-01
6.12267494e-01 -6.86181188e-02 1.13847351e+00 6.39350951e-01
5.83928406e-01 -7.14007199e-01 1.04565717e-01 3.14005554e-01
2.78245538e-01 -1.14926064e+00 2.22520962e-01 -3.96975458e-01
-4.11452889e-01 4.18908358e-01 7.25814402e-01 -5.14781952e-01
6.47844791e-01 4.52589154e-01 4.07492183e-02 -1.41002789e-01
-2.55805373e-01 -4.78083594e-03 2.83008963e-01 9.03681755e-01
-3.09177428e-01 8.47684816e-02 3.57147008e-01 3.43128115e-01
-3.13705832e-01 -1.28736198e-01 3.17616671e-01 1.06372380e+00
-1.85776472e-01 -8.98116171e-01 -9.37995911e-01 8.88514519e-02
-6.97241664e-01 2.68462867e-01 -3.77913862e-01 9.35231924e-01
1.30611733e-01 9.05243635e-01 -2.18188241e-01 -2.74767846e-01
2.71401465e-01 -1.22729808e-01 6.39716685e-01 -1.59947217e-01
-8.61322522e-01 -4.56150681e-01 1.09322391e-01 -6.12210274e-01
-3.43708009e-01 -9.30338383e-01 -7.20130146e-01 -6.43530190e-01
-5.02743162e-02 -8.57486948e-02 2.25628629e-01 1.10742605e+00
1.69738054e-01 2.67794818e-01 3.65578234e-01 -6.80351138e-01
-9.66511190e-01 -5.33552229e-01 -3.44632864e-01 5.98907471e-01
6.35385990e-01 -9.04845476e-01 -2.85597622e-01 -3.67263183e-02] | [5.413630962371826, 7.971582412719727] |
e2e20542-2395-44f1-b55b-d95d819a9817 | banditq-no-regret-learning-with-guaranteed | 2304.05219 | null | https://arxiv.org/abs/2304.05219v2 | https://arxiv.org/pdf/2304.05219v2.pdf | $\texttt{BanditQ}:$ Fair Multi-Armed Bandits with Guaranteed Rewards per Arm | Classic no-regret online prediction algorithms, including variants of the Upper Confidence Bound ($\texttt{UCB}$) algorithm, $\texttt{Hedge}$, and $\texttt{EXP3}$, are inherently unfair by design. The unfairness stems from their very objective of playing the most rewarding arm as many times as possible while ignoring the less rewarding ones among $N$ arms. In this paper, we consider a fair prediction problem in the stochastic setting with hard lower bounds on the rate of accrual of rewards for a set of arms. We study the problem in both full and bandit feedback settings. Using queueing-theoretic techniques in conjunction with adversarial learning, we propose a new online prediction policy called $\texttt{BanditQ}$ that achieves the target reward rates while achieving a regret and target rate violation penalty of $O(T^{\frac{3}{4}}).$ In the full-information setting, the regret bound can be further improved to $O(\sqrt{T})$ when considering the average regret over the entire horizon of length $T$. The proposed policy is efficient and admits a black-box reduction from the fair prediction problem to the standard MAB problem with a carefully defined sequence of rewards. The design and analysis of the $\texttt{BanditQ}$ policy involve a novel use of the potential function method in conjunction with scale-free second-order regret bounds and a new self-bounding inequality for the reward gradients, which are of independent interest. | ['Abhishek Sinha'] | 2023-04-11 | null | null | null | null | ['multi-armed-bandits'] | ['miscellaneous'] | [ 7.53932819e-02 3.64547342e-01 -4.63205963e-01 -5.44526756e-01
-1.00080824e+00 -8.03297520e-01 -3.37054044e-01 8.32670927e-02
-5.79035282e-01 1.25140190e+00 -3.27590704e-01 -8.61560881e-01
-9.32965755e-01 -7.32167423e-01 -9.83006775e-01 -9.32662904e-01
-3.68377388e-01 5.01034677e-01 -9.13851783e-02 -2.45713890e-01
1.88889563e-01 3.67438406e-01 -9.89870429e-01 -1.45877779e-01
9.14496720e-01 1.99256599e+00 -1.88072681e-01 6.78789258e-01
-8.80461559e-02 1.11081946e+00 -5.20255148e-01 -1.13207710e+00
9.54175472e-01 -5.73953450e-01 -7.48134494e-01 -2.30956838e-01
2.45144609e-02 -8.25533509e-01 -3.45330745e-01 1.09034908e+00
4.04245645e-01 3.93431127e-01 2.93837994e-01 -1.23935187e+00
-8.56867433e-02 8.43356192e-01 -8.34940791e-01 3.21279407e-01
-2.17563853e-01 -9.77254957e-02 1.44586813e+00 1.47052005e-01
1.12808265e-01 9.72481072e-01 3.38651329e-01 6.66358709e-01
-1.19914877e+00 -8.19380343e-01 3.48965198e-01 -9.51127633e-02
-1.07710230e+00 -2.14098081e-01 3.47631603e-01 -2.17888996e-01
5.37920535e-01 8.54869425e-01 5.12636304e-01 2.29877174e-01
2.75917172e-01 8.70020509e-01 1.06025803e+00 -4.44128245e-01
5.54302573e-01 3.99912782e-02 6.35287315e-02 4.69593197e-01
3.23640883e-01 5.87703228e-01 -1.85169384e-01 -3.25528324e-01
7.50366509e-01 3.91073644e-01 -3.55931632e-02 -2.65483439e-01
-4.09345448e-01 1.02503359e+00 3.52551430e-01 -4.65040356e-01
-5.42105138e-01 6.22944474e-01 3.11763942e-01 5.41551888e-01
5.15575051e-01 7.49852061e-02 -5.98248303e-01 -1.87024102e-01
-8.23172092e-01 3.17747384e-01 7.89183438e-01 1.27215087e+00
3.93273354e-01 2.23766360e-02 -8.10469985e-01 5.71358740e-01
3.97399105e-02 7.44130552e-01 -4.39759970e-01 -1.50546420e+00
1.02128255e+00 -3.68776964e-03 1.12844181e+00 -3.94305497e-01
-1.11657731e-01 -7.60158360e-01 -4.41388279e-01 2.33663186e-01
7.43152022e-01 -6.83618784e-01 -3.80170077e-01 1.86822402e+00
2.69119143e-01 -5.16891837e-01 -3.03486109e-01 9.66700912e-01
-4.82268214e-01 5.39258420e-01 -9.63231847e-02 -8.87790501e-01
7.54550457e-01 -6.11052394e-01 -5.89595437e-01 -7.04642683e-02
4.72689182e-01 -5.18372238e-01 5.61545074e-01 4.45494860e-01
-1.39500761e+00 2.58999437e-01 -7.61201084e-01 6.34472191e-01
2.42734253e-01 -4.96873111e-01 8.48871350e-01 1.32553101e+00
-6.62529647e-01 8.66666555e-01 -4.98046488e-01 3.77899110e-01
5.96440017e-01 6.45726025e-01 3.77876610e-01 -1.37138993e-01
-8.69410694e-01 5.23045361e-01 9.18444097e-02 2.94975877e-01
-9.41277266e-01 -7.37557173e-01 -1.95789263e-01 2.35047176e-01
9.46621120e-01 -3.97436261e-01 1.52915180e+00 -1.10943043e+00
-1.66820395e+00 4.03167635e-01 1.63395420e-01 -8.84408772e-01
9.40218508e-01 -1.98963881e-01 8.24768543e-02 1.89285845e-01
-1.34183615e-01 -5.35356700e-02 7.30985880e-01 -7.89545357e-01
-9.23393011e-01 -4.98452336e-01 4.38362986e-01 5.06143086e-02
-1.09696068e-01 1.54132202e-01 2.99650013e-01 -3.92071217e-01
-2.52197921e-01 -9.22652781e-01 -6.03965342e-01 -1.23928785e-01
-2.09422871e-01 1.24061115e-01 -5.86928800e-02 -3.82806599e-01
1.19183862e+00 -1.76704073e+00 -2.61404753e-01 3.68708372e-01
-2.10148796e-01 7.76753062e-03 1.81632787e-01 3.89099032e-01
2.16567665e-01 2.58030921e-01 9.07504484e-02 1.26199305e-01
3.21950912e-01 2.36223415e-01 -5.46258688e-01 6.24842346e-01
-5.71592867e-01 5.61794996e-01 -6.77057266e-01 -1.31900191e-01
-1.10018484e-01 -5.67049980e-01 -7.22562492e-01 3.66263062e-01
-5.04893601e-01 7.14215487e-02 -7.66821742e-01 4.17005241e-01
7.63513148e-01 -1.39556497e-01 5.62304296e-02 4.60561484e-01
-1.30005077e-01 -1.32488981e-01 -1.07622731e+00 1.02323246e+00
-3.38025302e-01 -9.16035399e-02 4.83124793e-01 -1.22129297e+00
6.20662570e-01 8.18269923e-02 7.80877352e-01 -6.41037107e-01
5.18389583e-01 1.17403582e-01 -7.42885098e-02 -1.11312725e-01
3.72782052e-01 -9.27582741e-01 -3.30426306e-01 7.48805225e-01
-3.53952706e-01 2.54558891e-01 -5.11928760e-02 1.89099982e-01
1.12549126e+00 -1.97381020e-01 -2.88269725e-02 -2.54026502e-01
7.75589719e-02 -1.57301113e-01 6.81341588e-01 1.21085417e+00
-7.23040998e-01 -1.58850417e-01 8.24380040e-01 -2.69429564e-01
-1.04685652e+00 -1.06262147e+00 1.30743384e-01 1.60062885e+00
2.37049773e-01 4.02263314e-01 -4.20255095e-01 -7.08958626e-01
6.40267670e-01 1.08278859e+00 -7.70926535e-01 -3.57549898e-02
-4.92663793e-02 -7.75857866e-01 1.69223055e-01 2.12622821e-01
3.14777553e-01 -4.94397074e-01 -6.98009372e-01 3.41916144e-01
8.91717300e-02 -6.81181431e-01 -7.49289334e-01 4.85046655e-01
-7.62905180e-01 -8.08090329e-01 -8.01739991e-01 2.02591941e-01
4.72941160e-01 1.96398273e-01 8.07481110e-01 -5.39710820e-01
-8.93549770e-02 6.03790760e-01 -4.33166772e-01 -9.02955055e-01
4.05744463e-02 -5.27099788e-01 -9.97927338e-02 1.78052798e-01
1.40561331e-02 -2.20092356e-01 -9.70773876e-01 4.77300763e-01
-7.09890842e-01 -3.08382720e-01 4.19993460e-01 8.89978647e-01
7.72596538e-01 -2.41592526e-02 9.03830528e-01 -8.72033715e-01
4.95652139e-01 -2.85369486e-01 -1.20638633e+00 4.41382229e-01
-6.96278751e-01 1.74103290e-01 9.53579664e-01 -1.41865373e-01
-1.16095209e+00 -2.32005507e-01 1.79682851e-01 -5.75977385e-01
6.22488499e-01 7.37564042e-02 2.22768590e-01 -1.69147894e-01
4.46132094e-01 2.45599702e-01 2.35936955e-01 -2.53929347e-01
2.52885133e-01 4.71994102e-01 -3.63681801e-02 -9.74908233e-01
4.10442024e-01 3.42598528e-01 3.93825889e-01 1.13669578e-02
-1.43943655e+00 -2.27540746e-01 2.11575553e-01 -3.10754001e-01
1.15276538e-01 -5.44786751e-01 -1.88522255e+00 -2.88908690e-01
-6.13725364e-01 -2.83941329e-01 -7.43593574e-01 5.32149673e-01
-1.29583383e+00 1.95326552e-01 -5.13001025e-01 -1.97009170e+00
-5.64545631e-01 -5.46379507e-01 3.97919387e-01 2.92659730e-01
5.16056895e-01 -4.35212523e-01 -2.92226434e-01 6.81830466e-01
5.13352334e-01 2.75631279e-01 5.55478513e-01 -4.24981505e-01
-6.83161855e-01 -4.62272793e-01 -2.38430917e-01 6.18088484e-01
-3.31884265e-01 -6.30904377e-01 -3.39600950e-01 -6.17312312e-01
1.12219669e-01 -4.71811712e-01 4.82143521e-01 6.81551695e-01
1.54003227e+00 -9.45441127e-01 -7.52107380e-03 3.70747864e-01
1.75214314e+00 7.30903506e-01 3.51517797e-01 2.02881649e-01
-1.80939823e-01 4.03907418e-01 1.02060878e+00 1.26276970e+00
-1.71623692e-01 2.39960998e-01 9.53135073e-01 5.44721484e-01
9.40366805e-01 -9.57980230e-02 2.90029347e-01 -1.53226063e-01
-2.26755217e-01 -1.53117329e-01 -3.84648085e-01 3.49747539e-01
-1.95950592e+00 -1.27653766e+00 2.99716294e-01 3.14250493e+00
8.71781111e-01 2.39748478e-01 5.07681131e-01 -1.40722662e-01
7.59772897e-01 -3.17965031e-01 -1.06704998e+00 -1.07897580e+00
3.02685916e-01 4.36042577e-01 1.41326690e+00 3.40329766e-01
-7.59135723e-01 3.95497411e-01 5.92532730e+00 1.17795432e+00
-6.89615607e-01 1.18944451e-01 1.16777861e+00 -1.03346181e+00
-2.17714369e-01 7.70296082e-02 -7.70635366e-01 6.01973653e-01
1.25231504e+00 -6.28813207e-01 9.53543842e-01 1.07537854e+00
4.09927964e-01 -2.31973588e-01 -9.92151499e-01 7.85641730e-01
-6.76936030e-01 -1.32188320e+00 -5.44041455e-01 1.85405165e-01
6.32339716e-01 -2.82696635e-01 1.47124484e-01 4.83144045e-01
8.40225458e-01 -9.85142708e-01 8.66053402e-01 4.29346025e-01
9.10946727e-01 -1.40815997e+00 6.55986726e-01 5.46251833e-01
-7.30540037e-01 -7.07882762e-01 -4.46300447e-01 -2.81846374e-01
7.62245134e-02 6.11964405e-01 -3.63093942e-01 7.42852330e-01
7.00879037e-01 -2.00765073e-01 5.83401501e-01 1.18431211e+00
1.65521830e-01 5.74290693e-01 -5.08608282e-01 -5.56933224e-01
3.38068962e-01 -3.14585090e-01 3.57879251e-01 7.59315789e-01
2.30604917e-01 6.37103021e-01 2.14374736e-01 6.39036477e-01
-5.52686527e-02 2.71146059e-01 1.23718031e-01 1.17275594e-02
4.23363954e-01 8.57778728e-01 -2.79836804e-01 -4.54169326e-02
-3.69250737e-02 7.13629305e-01 2.76912689e-01 2.86185920e-01
-1.02911627e+00 -6.45067573e-01 7.14487255e-01 -5.87483048e-02
6.32889390e-01 1.44857153e-01 -4.33313668e-01 -5.97992957e-01
1.92634717e-01 -1.74032673e-01 7.18968034e-01 -2.72367299e-01
-1.57118046e+00 -1.53335929e-02 -1.87556654e-01 -1.17844725e+00
-9.90789458e-02 -4.27833915e-01 -2.88387418e-01 1.10147715e+00
-1.43897378e+00 -3.67645323e-01 5.84388137e-01 7.65626490e-01
-8.65992382e-02 1.80610850e-01 5.31506836e-01 1.72781363e-01
-5.46241105e-01 1.04502845e+00 1.08245838e+00 -2.14294761e-01
1.29920423e-01 -1.13020849e+00 -6.67113185e-01 5.21900356e-01
-7.71601379e-01 2.49002397e-01 9.86860931e-01 -2.14650989e-01
-1.60656273e+00 -1.07604051e+00 1.67711720e-01 1.29403710e-01
8.73327971e-01 -6.99648038e-02 -5.74257523e-02 6.57137275e-01
-2.25439280e-01 1.80703729e-01 8.28018427e-01 1.67092636e-01
-1.86237067e-01 -7.31220543e-01 -1.75281715e+00 4.07560498e-01
1.03912783e+00 -2.71975789e-02 3.14619280e-02 4.43081081e-01
9.69474196e-01 -3.60163212e-01 -1.07292557e+00 1.46706710e-02
8.68824720e-01 -1.00557292e+00 6.67952061e-01 -1.06219578e+00
1.85508996e-01 5.30526221e-01 -4.67052668e-01 -9.99888778e-01
-2.60611176e-01 -1.46643686e+00 -1.00075237e-01 7.30956674e-01
5.12041509e-01 -8.30144882e-01 1.16344428e+00 1.22007990e+00
5.80304749e-02 -1.05549145e+00 -1.56481349e+00 -1.33426881e+00
5.19643545e-01 -4.00617272e-01 3.91120106e-01 2.62290895e-01
4.07935828e-01 -2.66621679e-01 -8.57357860e-01 -1.07142374e-01
1.06685889e+00 5.97244322e-01 4.36927021e-01 -5.93469858e-01
-8.62281561e-01 -3.37378770e-01 1.49830699e-01 -1.19657326e+00
-3.11538309e-01 -5.59961319e-01 6.10161424e-02 -9.59562063e-01
3.53439480e-01 -9.06294584e-01 -7.70043790e-01 4.30560738e-01
1.66625306e-01 -6.72469795e-01 6.12499952e-01 -1.07987784e-02
-8.94078970e-01 6.41648412e-01 1.43578827e+00 -7.70881474e-02
-1.92310978e-02 7.35430896e-01 -9.74909425e-01 9.04906243e-02
7.82734096e-01 -5.58607876e-01 -3.41305882e-01 -3.08838356e-02
3.48079890e-01 1.19899666e+00 -3.94013710e-02 -3.26237708e-01
-7.04885945e-02 -9.23900843e-01 -3.66685949e-02 -5.68968117e-01
2.83182949e-01 -1.08934617e+00 5.67016900e-02 7.54372776e-01
-6.96831584e-01 -5.54877222e-01 -9.16838199e-02 1.04438484e+00
4.62291300e-01 -4.82610852e-01 8.51288855e-01 -1.91517428e-01
1.89801157e-01 6.68104172e-01 -1.27117306e-01 1.75334960e-01
1.40732682e+00 1.06721722e-01 -4.59448576e-01 -7.70277381e-01
-7.56373167e-01 6.00970328e-01 -1.90122306e-01 -2.34814569e-01
2.15209916e-01 -9.86498177e-01 -4.32740152e-01 -3.58902067e-01
-3.57846051e-01 -3.00461411e-01 6.86684966e-01 7.99424827e-01
-1.53725907e-01 7.11888313e-01 -1.38599694e-01 -6.28669932e-02
-6.91811085e-01 7.92053223e-01 6.35490239e-01 -6.25144482e-01
6.49496242e-02 9.82730389e-01 -1.83208704e-01 1.57897979e-01
4.33349550e-01 4.57556844e-02 5.89511335e-01 -3.46663564e-01
5.95569968e-01 7.69224167e-01 -2.33819380e-01 1.03077844e-01
-4.82742339e-02 -2.35271439e-01 -2.81642258e-01 -2.91951895e-01
1.23295379e+00 -4.03835386e-01 6.15974441e-02 -7.87391420e-03
1.03057384e+00 -2.48131141e-01 -1.61109066e+00 -2.65728593e-01
-9.05333161e-02 -9.82010961e-01 -7.55330846e-02 -1.16329026e+00
-1.07916725e+00 5.48805475e-01 6.54252291e-01 6.93893731e-01
1.17180729e+00 -3.01004052e-01 7.40330696e-01 2.04693437e-01
1.04999220e+00 -1.34938788e+00 -2.89325446e-01 3.31026971e-01
6.92411482e-01 -9.55888391e-01 -7.53906593e-02 -1.86428800e-01
-5.31057477e-01 8.13530803e-01 3.98724705e-01 -2.49183938e-01
6.27629817e-01 7.34637608e-04 -5.85634470e-01 3.79286826e-01
-7.94092596e-01 -1.64789483e-01 -2.39750579e-01 7.85550177e-02
2.05762863e-01 6.65191829e-01 -9.02126193e-01 1.11051488e+00
-1.21222585e-01 1.82530373e-01 5.15344203e-01 1.09538102e+00
-8.18965733e-01 -9.70354557e-01 -5.73655009e-01 9.85801578e-01
-1.04542887e+00 2.12113991e-01 1.34150818e-01 2.77064323e-01
-3.38500440e-01 1.23881650e+00 -2.59776153e-02 -6.34274781e-02
3.64042133e-01 -9.10531059e-02 7.11181164e-01 -5.62856123e-02
-5.37963748e-01 8.64899680e-02 1.43792763e-01 -6.15768194e-01
-1.78988487e-01 -4.95586067e-01 -1.08251345e+00 -8.85272563e-01
-3.44583362e-01 5.06425560e-01 4.65882659e-01 1.01005793e+00
1.71660915e-01 2.22830772e-01 1.38017547e+00 -3.38661343e-01
-1.55388653e+00 -5.61372995e-01 -1.28304482e+00 1.68067649e-01
7.22844452e-02 -4.33373451e-01 -3.91453743e-01 -6.50537312e-01] | [4.544561862945557, 3.3394603729248047] |
d1201467-c14d-438a-8838-686341d56dd2 | multi-perspective-context-aggregation-for | 1808.06289 | null | http://arxiv.org/abs/1808.06289v1 | http://arxiv.org/pdf/1808.06289v1.pdf | Multi-Perspective Context Aggregation for Semi-supervised Cloze-style Reading Comprehension | Cloze-style reading comprehension has been a popular task for measuring the
progress of natural language understanding in recent years. In this paper, we
design a novel multi-perspective framework, which can be seen as the joint
training of heterogeneous experts and aggregate context information from
different perspectives. Each perspective is modeled by a simple aggregation
module. The outputs of multiple aggregation modules are fed into a one-timestep
pointer network to get the final answer. At the same time, to tackle the
problem of insufficient labeled data, we propose an efficient sampling
mechanism to automatically generate more training examples by matching the
distribution of candidates between labeled and unlabeled data. We conduct our
experiments on a recently released cloze-test dataset CLOTH (Xie et al., 2017),
which consists of nearly 100k questions designed by professional teachers.
Results show that our method achieves new state-of-the-art performance over
previous strong baselines. | ['Liang Wang', 'Jingming Liu', 'Ruoyu Jia', 'Meng Sun', 'Kewei Shen', 'Wei Zhao', 'Sujian Li'] | 2018-08-20 | multi-perspective-context-aggregation-for-2 | https://aclanthology.org/C18-1073 | https://aclanthology.org/C18-1073.pdf | coling-2018-8 | ['cloze-test'] | ['natural-language-processing'] | [ 1.51676625e-01 1.27555177e-01 -1.85331121e-01 -6.31345332e-01
-1.45239329e+00 -6.69281662e-01 3.79620731e-01 1.05252601e-01
-3.72344732e-01 5.82516253e-01 4.55355525e-01 -3.78948510e-01
1.29878521e-01 -6.76003873e-01 -7.25864768e-01 -4.38550234e-01
7.87344515e-01 5.34353435e-01 3.89780432e-01 -1.24738686e-01
3.98669064e-01 -2.59930462e-01 -1.56616342e+00 7.96998382e-01
1.64117086e+00 5.98654389e-01 2.56538898e-01 6.84409559e-01
-1.12083949e-01 1.13243186e+00 -6.53487802e-01 -6.39769018e-01
-1.36150852e-01 -5.50630927e-01 -1.00125706e+00 3.41191851e-02
1.01217747e+00 -6.25392139e-01 -1.25086382e-01 1.14987969e+00
5.79918504e-01 2.78545260e-01 4.87039834e-01 -8.19997966e-01
-1.03436530e+00 9.97958481e-01 -4.42287296e-01 1.91618532e-01
6.29273713e-01 4.33042347e-01 1.13266563e+00 -1.23430800e+00
3.92348200e-01 1.18867004e+00 2.15814933e-01 9.38125372e-01
-9.48592246e-01 -5.46625435e-01 6.30166531e-01 4.59863961e-01
-6.73987925e-01 -2.83327132e-01 5.09363532e-01 -4.16635245e-01
4.26641107e-01 -1.06146475e-02 3.79995823e-01 1.15795434e+00
-5.19242942e-01 1.46233165e+00 1.42097509e+00 -4.90083754e-01
2.10890010e-01 -2.33710296e-02 5.66241324e-01 5.45448124e-01
6.86893985e-02 -4.18896437e-01 -6.60366058e-01 -1.36261284e-01
2.35316619e-01 -1.74758062e-02 -6.37793303e-01 -2.42590189e-01
-1.10328770e+00 7.57291675e-01 2.44031519e-01 -1.70112029e-02
-4.59233522e-02 -1.86845914e-01 2.56269462e-02 5.41867554e-01
5.41719973e-01 6.70966387e-01 -5.31048536e-01 -9.31190923e-02
-7.80017018e-01 5.22787392e-01 8.39942336e-01 1.06047332e+00
3.32094252e-01 -4.89040345e-01 -7.04660773e-01 8.44242096e-01
2.97375023e-01 4.74442899e-01 4.90906000e-01 -1.23685133e+00
1.09118366e+00 8.92871439e-01 3.13328236e-01 -3.31865698e-01
4.16266173e-02 -3.67605835e-01 -4.00672078e-01 -3.81086990e-02
8.28024387e-01 -2.63419896e-01 -7.65560448e-01 1.84490764e+00
5.38202405e-01 3.48157912e-01 -9.35323983e-02 7.06556916e-01
9.94810104e-01 6.78749621e-01 2.08187580e-01 3.03924959e-02
1.40061510e+00 -1.62047350e+00 -6.60053432e-01 -2.58334875e-01
5.70426345e-01 -5.79835236e-01 1.66806710e+00 8.94456029e-01
-1.57558441e+00 -5.57694674e-01 -8.16689551e-01 -4.78990316e-01
-1.60883859e-01 1.33887768e-01 6.91706985e-02 2.98540831e-01
-9.32288408e-01 2.53127337e-01 -4.96028334e-01 2.05412507e-01
4.39828664e-01 -2.45043233e-01 4.31521349e-02 -6.43529415e-01
-1.15313375e+00 7.53507197e-01 1.54014781e-01 -2.98566550e-01
-1.11366451e+00 -1.03921473e+00 -5.43339491e-01 2.98914790e-01
7.42865086e-01 -7.75087059e-01 1.95809865e+00 -4.93577212e-01
-1.69896412e+00 8.36798191e-01 -3.04635555e-01 1.03627540e-01
7.40824521e-01 -8.39520752e-01 -4.41979058e-02 2.03243885e-02
2.36858770e-01 4.43028301e-01 7.19540179e-01 -1.01807559e+00
-8.64491284e-01 -3.51013631e-01 1.92487851e-01 5.89994609e-01
-2.76286721e-01 6.25258610e-02 -4.01572734e-01 -4.40081954e-01
1.63914934e-01 -5.73275685e-01 -1.08762763e-01 -2.70290315e-01
-1.27365306e-01 -9.88573432e-01 2.73288250e-01 -7.27294743e-01
1.49905896e+00 -1.83643842e+00 4.80353922e-01 -3.31110984e-01
6.67434096e-01 4.02728230e-01 -2.63980508e-01 2.75127113e-01
1.23456113e-01 1.48569286e-01 -1.69718191e-01 -3.37576151e-01
1.39567032e-02 -3.77976209e-01 -4.53167200e-01 6.62592500e-02
4.78085726e-02 9.05110121e-01 -1.61552966e+00 -2.08207279e-01
-2.47238621e-01 -1.58421069e-01 -6.76435530e-01 7.31875896e-01
-7.92657793e-01 3.77636522e-01 -5.59315681e-01 5.58688998e-01
5.27500153e-01 -5.95344722e-01 -4.89681102e-02 4.87168372e-01
1.43469408e-01 8.70027781e-01 -8.79843295e-01 1.83488750e+00
-5.50836086e-01 3.44155490e-01 2.29997896e-02 -6.72975063e-01
4.35456663e-01 5.30607462e-01 -2.36526325e-01 -4.06747371e-01
1.52813524e-01 2.35286817e-01 -5.89779466e-02 -7.78894544e-01
2.94336200e-01 3.09897482e-01 -3.67344543e-02 1.04484022e+00
2.04747558e-01 -2.49740794e-01 4.01365936e-01 4.60481495e-01
1.07694077e+00 1.27367884e-01 1.45936742e-01 -5.76885417e-02
6.21860206e-01 -9.03358981e-02 2.54385412e-01 9.80176568e-01
-3.09945047e-01 5.45759678e-01 4.62397516e-01 -2.83815175e-01
-7.61628687e-01 -1.23201668e+00 2.53421187e-01 1.76849163e+00
-2.06933804e-02 -3.38052511e-01 -9.59082961e-01 -1.17584884e+00
-2.37362519e-01 8.62056434e-01 -5.38659871e-01 -9.22670960e-02
-6.84945345e-01 -2.31387198e-01 2.99273372e-01 6.37947917e-01
6.59231663e-01 -1.11445498e+00 -4.61693943e-01 2.96988308e-01
-5.96154571e-01 -1.00826502e+00 -6.63823307e-01 -1.98639750e-01
-8.59042466e-01 -1.10351050e+00 -6.73941851e-01 -8.08614075e-01
7.79901206e-01 6.40301704e-01 1.73766184e+00 3.63504559e-01
3.86716664e-01 2.54911840e-01 -6.10643864e-01 -4.03951973e-01
-3.81598473e-01 3.37553918e-01 -1.61045507e-01 -3.07116210e-01
7.79007971e-01 -4.35361594e-01 -7.50542700e-01 1.68942943e-01
-7.37458110e-01 3.60872030e-01 4.71672922e-01 8.54871750e-01
4.88804460e-01 -5.38578391e-01 6.94237649e-01 -1.25507212e+00
1.20472825e+00 -6.91376448e-01 -6.09490693e-01 7.68551528e-01
-4.25643712e-01 3.85244898e-02 8.33014846e-01 -5.77758014e-01
-1.27477026e+00 -1.67006239e-01 4.31841835e-02 -2.39474818e-01
-3.52551073e-01 4.26295549e-01 -2.98329085e-01 2.89518744e-01
6.68096900e-01 8.05572048e-02 -5.40022790e-01 -6.19434595e-01
8.37089002e-01 8.52207363e-01 6.78126216e-01 -1.08435929e+00
6.41388714e-01 -3.19865793e-02 -8.44955504e-01 -1.88723162e-01
-1.88137567e+00 -5.92586100e-01 -5.39037764e-01 -1.47947267e-01
7.63038099e-01 -1.24324894e+00 -5.68359792e-01 6.19945228e-01
-1.32386255e+00 -5.60243189e-01 5.47946133e-02 2.08158642e-01
-3.08457017e-01 7.16340244e-02 -6.99097097e-01 -5.16353071e-01
-3.36291075e-01 -1.29257047e+00 8.27308118e-01 5.86033463e-01
-4.92552035e-02 -8.25719059e-01 1.65947229e-01 1.06084824e+00
1.94968224e-01 -2.91033179e-01 1.02341461e+00 -8.73422682e-01
-8.35736692e-01 2.36868598e-02 -1.31645873e-01 4.27768409e-01
-2.29579970e-01 -1.38247982e-01 -1.12224555e+00 -2.32533187e-01
3.57513458e-01 -1.09538198e+00 9.91344750e-01 1.68629233e-02
1.44582546e+00 -3.01982433e-01 -1.94690898e-02 3.13689262e-01
1.05789328e+00 -1.62377551e-01 2.99672395e-01 1.74914673e-01
7.52913356e-01 7.52642870e-01 5.92786193e-01 2.01149005e-02
7.55429566e-01 8.52863938e-02 1.79939017e-01 4.61914986e-01
-1.89528167e-01 -6.22679591e-01 4.30318207e-01 1.23084009e+00
2.86701053e-01 -4.79268461e-01 -9.80480015e-01 7.32943833e-01
-1.70398295e+00 -9.94748712e-01 -9.50065777e-02 2.04285240e+00
1.22495961e+00 1.02772027e-01 1.45312361e-02 -2.49625579e-01
6.97947919e-01 -8.31939001e-03 -6.65390790e-01 -1.67082071e-01
3.73915881e-01 2.49255493e-01 -2.50813305e-01 7.46619880e-01
-9.03150976e-01 1.04764473e+00 6.22096872e+00 6.64291859e-01
-6.73995376e-01 3.08006257e-01 7.89420366e-01 -1.91483289e-01
-5.45671940e-01 -1.99962884e-01 -9.54588830e-01 5.22776961e-01
7.36904144e-01 -1.28820658e-01 3.02858144e-01 9.28493500e-01
-6.84068650e-02 -1.75088271e-01 -1.18970156e+00 4.91403997e-01
2.60748386e-01 -9.62689757e-01 1.45282224e-02 -3.78790438e-01
1.32069468e+00 2.31623411e-01 2.70229369e-01 5.34040630e-01
9.21516061e-01 -1.00007975e+00 5.05814016e-01 2.67080218e-01
4.92464840e-01 -4.66905683e-01 3.77482444e-01 8.45497191e-01
-7.24591553e-01 -1.43346980e-01 -2.49217376e-01 -1.84264824e-01
1.39451865e-02 5.24636686e-01 -7.42417276e-01 1.74998716e-01
6.00590050e-01 3.63451928e-01 -7.77859628e-01 1.17282176e+00
-1.07773721e+00 1.07277858e+00 -1.27851084e-01 -3.12284946e-01
3.24372202e-01 -5.10398634e-02 1.20708175e-01 7.64741063e-01
1.36383176e-01 4.63459700e-01 2.49989033e-01 1.13322556e+00
-5.76337278e-01 1.13540858e-01 -2.87910700e-01 2.48063341e-01
1.00226903e+00 1.12083173e+00 -1.40914664e-01 -7.07192600e-01
-7.70798147e-01 6.74963057e-01 1.06674433e+00 3.47289205e-01
-5.48727095e-01 -2.68605143e-01 2.60502338e-01 -5.46942763e-02
4.81748837e-04 1.62257060e-01 -4.07787919e-01 -1.63630116e+00
1.75511613e-01 -1.37173641e+00 4.96446371e-01 -8.77386689e-01
-1.61387515e+00 2.73455590e-01 -5.49945012e-02 -1.16883945e+00
-1.49464026e-01 -5.81438303e-01 -9.97478902e-01 1.05473912e+00
-1.46793759e+00 -7.17593014e-01 -5.60169101e-01 3.07385206e-01
1.01605463e+00 5.09805828e-02 5.03503859e-01 -3.03881336e-02
-6.01468623e-01 6.48288012e-01 9.62958336e-02 9.10838172e-02
1.00615668e+00 -1.77268708e+00 7.87869155e-01 1.03175163e+00
1.54098913e-01 8.03566933e-01 5.20438790e-01 -6.51605546e-01
-7.69721210e-01 -9.08509731e-01 1.15706491e+00 -1.10514069e+00
5.79380810e-01 -2.70813733e-01 -1.35628796e+00 7.30454385e-01
6.02839768e-01 -2.32630163e-01 7.40246892e-01 1.71565264e-01
-6.41204059e-01 2.97070950e-01 -9.37032580e-01 7.39292860e-01
9.55045760e-01 -5.60633779e-01 -1.18287694e+00 3.44074339e-01
1.06304920e+00 -7.73942649e-01 -6.34046435e-01 2.14369491e-01
1.91198215e-01 -1.14340091e+00 5.90026498e-01 -8.55289757e-01
1.00707924e+00 -1.30512252e-01 3.09461445e-01 -1.63485098e+00
-2.32948184e-01 -5.06400466e-01 -4.30388391e-01 1.13498235e+00
5.56320608e-01 -2.56477714e-01 7.61325002e-01 7.47208297e-01
-1.76503837e-01 -1.14004850e+00 -5.19142330e-01 -2.64284819e-01
5.59301496e-01 -1.85242534e-01 6.83127105e-01 8.33316803e-01
2.82263696e-01 7.27872074e-01 -4.48649116e-02 9.99418199e-02
5.89627504e-01 1.69959247e-01 6.97522223e-01 -1.19307971e+00
-3.21927071e-01 -4.41331297e-01 5.44345975e-01 -1.72637534e+00
2.40251407e-01 -7.48586595e-01 2.64236331e-01 -1.48587298e+00
4.30731028e-01 -2.64121383e-01 -2.53107756e-01 2.50482768e-01
-1.18722367e+00 -2.22055942e-01 1.14126943e-01 -9.59400982e-02
-1.13776708e+00 4.37066972e-01 1.74968278e+00 5.87695241e-02
-5.76913543e-02 2.24546462e-01 -1.03089309e+00 8.94672692e-01
8.42139602e-01 -3.13369662e-01 -8.22904825e-01 -1.07248700e+00
4.51539010e-01 7.10864067e-02 -5.40960021e-02 -8.55735660e-01
5.57684422e-01 -2.53381282e-01 2.06870928e-01 -7.55558431e-01
-1.12690993e-01 -3.92333269e-01 -8.57672513e-01 1.51965171e-01
-9.25255597e-01 2.28976771e-01 -3.51696253e-01 5.27321875e-01
-2.52004296e-01 -4.44497377e-01 6.30540490e-01 -4.34036583e-01
-3.65328252e-01 2.88481921e-01 -5.24541289e-02 8.08276415e-01
6.67074263e-01 3.17815512e-01 -8.77543807e-01 -5.76814234e-01
-4.60155994e-01 9.22041416e-01 2.72394598e-01 5.22904336e-01
5.53069472e-01 -1.15094566e+00 -1.00405800e+00 -4.12212424e-02
2.93992311e-01 6.60900116e-01 2.16732562e-01 4.78186131e-01
-3.29587370e-01 2.66573042e-01 3.38194430e-01 -3.68085057e-01
-1.04657745e+00 5.05500019e-01 2.88565844e-01 -5.23518503e-01
-2.85015762e-01 1.33335090e+00 1.92234561e-01 -6.78135216e-01
4.51503038e-01 -3.93430144e-01 -4.09983188e-01 8.35145190e-02
1.12894309e+00 5.21518528e-01 -2.04132404e-02 3.43690477e-02
3.78732473e-01 4.14810389e-01 -4.28055823e-01 -5.48571907e-02
1.07917523e+00 -1.82966366e-01 -6.84383586e-02 4.64927346e-01
7.73110211e-01 1.53554350e-01 -1.24785388e+00 -6.01531088e-01
2.28016362e-01 -3.25984716e-01 -4.11446571e-01 -1.32177162e+00
-7.34446347e-01 1.14327765e+00 2.42382810e-02 1.09543033e-01
8.96391988e-01 2.16620982e-01 7.92448282e-01 5.44153094e-01
1.65043145e-01 -9.81444657e-01 4.54275250e-01 7.12894857e-01
6.58259094e-01 -1.47148418e+00 -2.76342481e-01 -4.31102306e-01
-6.53275371e-01 8.96003604e-01 1.51195574e+00 -2.90769413e-02
9.43372026e-02 -2.24861771e-01 3.20840329e-01 -1.17949031e-01
-1.38176370e+00 -9.81179401e-02 2.92916626e-01 2.80452400e-01
7.60491550e-01 1.16840616e-01 -2.71344662e-01 9.04309392e-01
-1.84617236e-01 -8.24224576e-02 5.49492061e-01 7.36366987e-01
-8.45589578e-01 -1.10207689e+00 -2.07956761e-01 4.79120970e-01
-4.36916173e-01 -3.57960254e-01 -5.38725913e-01 3.34738612e-01
-8.91978443e-02 1.09697366e+00 -3.01234007e-01 -1.47440970e-01
3.31079125e-01 5.16239285e-01 5.53463101e-01 -1.14774263e+00
-8.08168173e-01 -3.63060743e-01 -1.42757997e-01 -4.43087548e-01
-1.45298257e-01 -5.19601762e-01 -8.77185106e-01 4.35463190e-02
-4.23866391e-01 3.06972742e-01 8.64475146e-02 1.15773118e+00
1.37217835e-01 5.06868601e-01 5.56333721e-01 -2.35574171e-01
-1.21728730e+00 -1.25660396e+00 -6.07629158e-02 5.29679835e-01
3.12460452e-01 -4.24637288e-01 -4.66369033e-01 -7.47248381e-02] | [11.207940101623535, 8.127957344055176] |
8f87960f-1feb-4092-8300-a4bc9be63a1a | entity-and-evidence-guided-document-level | null | null | https://aclanthology.org/2021.repl4nlp-1.30 | https://aclanthology.org/2021.repl4nlp-1.30.pdf | Entity and Evidence Guided Document-Level Relation Extraction | Document-level relation extraction is a challenging task, requiring reasoning over multiple sentences to predict a set of relations in a document. In this paper, we propose a novel framework E2GRE (Entity and Evidence Guided Relation Extraction) that jointly extracts relations and the underlying evidence sentences by using large pretrained language model (LM) as input encoder. First, we propose to guide the pretrained LM’s attention mechanism to focus on relevant context by using attention probabilities as additional features for evidence prediction. Furthermore, instead of feeding the whole document into pretrained LMs to obtain entity representation, we concatenate document text with head entities to help LMs concentrate on parts of the document that are more related to the head entity. Our E2GRE jointly learns relation extraction and evidence prediction effectively, showing large gains on both these tasks, which we find are highly correlated. | ['Jing Huang', 'Tengyu Ma', 'Guangtao Wang', 'Peng Qi', 'Kevin Huang'] | null | null | null | null | acl-repl4nlp-2021-8 | ['document-level-relation-extraction'] | ['natural-language-processing'] | [ 1.15395844e-01 9.10093606e-01 -5.78905523e-01 -4.53310519e-01
-8.47466111e-01 -3.52773070e-01 5.62286317e-01 5.77595472e-01
-5.31999230e-01 1.09491301e+00 5.33402681e-01 -3.75671476e-01
-2.53326058e-01 -8.51601601e-01 -1.10798323e+00 -6.84717000e-02
-2.43829206e-01 3.50261658e-01 3.06664437e-01 8.12430084e-02
3.68879326e-02 2.87594348e-01 -9.55610693e-01 5.90645671e-01
7.25851774e-01 1.00790858e+00 1.65578425e-01 7.62825966e-01
-3.88589054e-01 1.50911796e+00 -5.79634309e-01 -8.46751273e-01
-3.81193817e-01 -1.66024968e-01 -1.32173097e+00 2.30155177e-02
1.89547807e-01 -3.23698819e-01 -4.76406068e-01 7.97580123e-01
-1.53962690e-02 -5.76240150e-03 8.45737457e-01 -7.85584211e-01
-8.21841061e-01 1.41326177e+00 -6.25229418e-01 4.65496927e-01
3.60528767e-01 -3.51069540e-01 1.64117134e+00 -1.21852195e+00
7.88333654e-01 1.11989188e+00 1.60967603e-01 1.36316389e-01
-9.31797504e-01 -4.54420388e-01 7.16515481e-01 4.85176623e-01
-1.18758798e+00 -4.31835055e-01 7.51796067e-01 -7.69893676e-02
1.73967290e+00 2.05765888e-01 1.81804463e-01 9.58355904e-01
4.81067225e-02 1.06580663e+00 3.40532929e-01 -6.42089665e-01
-1.79428831e-01 2.95314431e-01 4.25874501e-01 7.95312464e-01
5.12449324e-01 -5.33988357e-01 -6.72038674e-01 7.90448934e-02
2.22194299e-01 -3.19345027e-01 -1.45752683e-01 3.05274665e-01
-1.02811253e+00 4.75686997e-01 4.46204305e-01 2.88976580e-01
-7.14172423e-01 2.24330083e-01 3.30563754e-01 7.77470767e-02
3.94979686e-01 4.24211591e-01 -1.00961041e+00 2.40941226e-01
-4.43165392e-01 2.01416854e-02 8.37156057e-01 1.24095345e+00
6.69341624e-01 -6.09703660e-01 -5.73006511e-01 8.28872442e-01
6.78074598e-01 2.03933269e-01 1.60635903e-01 -6.32985771e-01
1.18699729e+00 9.42793131e-01 -9.62516591e-02 -8.11394751e-01
-1.58753216e-01 -6.88064158e-01 -7.13401854e-01 -4.45032209e-01
-1.83037564e-01 -5.41622758e-01 -5.55535376e-01 1.67728817e+00
2.86783159e-01 1.73809394e-01 5.80050528e-01 3.79813701e-01
1.06870520e+00 7.89299846e-01 2.83675343e-01 -2.40222692e-01
1.40879238e+00 -1.31225729e+00 -7.93575525e-01 -6.28973901e-01
9.19835865e-01 -4.85782385e-01 7.70432830e-01 1.87834531e-01
-1.28492832e+00 -3.35799456e-01 -1.30708504e+00 -4.56242830e-01
-4.50446248e-01 4.11640674e-01 5.90148985e-01 -4.21676189e-01
-3.88179213e-01 3.97209287e-01 -5.95856667e-01 1.00783758e-01
6.36293173e-01 4.13593769e-01 -5.03187180e-01 -7.87870809e-02
-1.46314943e+00 1.06686389e+00 8.56740594e-01 1.75555602e-01
-3.35612416e-01 -2.81324744e-01 -1.28957260e+00 6.69265568e-01
8.58511388e-01 -6.59374714e-01 1.23615420e+00 -2.20991433e-01
-1.27544522e+00 6.67304873e-01 -4.77840662e-01 -5.62821925e-01
-1.80867106e-01 -6.77092731e-01 -4.57270741e-01 1.26979044e-02
-1.68539565e-02 5.42543650e-01 3.27039987e-01 -1.13028538e+00
-7.62573183e-01 1.00713328e-01 2.06132591e-01 1.32075891e-01
-3.26052934e-01 2.37182721e-01 -9.22267735e-01 -4.73742753e-01
4.96443026e-02 -3.30040187e-01 -4.46589589e-02 -6.87629282e-01
-9.77432489e-01 -7.92949021e-01 4.53981012e-01 -6.87336683e-01
1.59811449e+00 -1.78314745e+00 2.83333391e-01 3.39993298e-01
2.36632571e-01 1.39910698e-01 -3.42569798e-01 4.20886666e-01
-2.63173282e-02 3.68476868e-01 -1.10391207e-01 -4.40690964e-01
2.73891771e-03 3.82994264e-01 -5.46034157e-01 -3.13237429e-01
1.11273491e+00 1.20871902e+00 -8.78741264e-01 -1.02203274e+00
-3.37801576e-01 2.24823847e-01 -4.74319488e-01 4.20808226e-01
-5.26924372e-01 -1.06756978e-01 -6.82727933e-01 4.01379198e-01
2.62203276e-01 -7.39648163e-01 3.86287123e-01 -2.35246152e-01
3.31438422e-01 1.14846623e+00 -1.00256610e+00 1.31531835e+00
-6.47028506e-01 5.54757118e-01 -3.78807098e-01 -7.74415553e-01
8.52709830e-01 4.33068782e-01 6.92978278e-02 -2.69648314e-01
1.03848360e-01 2.39320658e-02 2.30206385e-01 -7.98492432e-01
3.47300619e-01 2.00428024e-01 1.78762879e-02 3.71775746e-01
5.28520942e-01 2.49085516e-01 5.19633949e-01 6.64232373e-01
1.37520778e+00 -3.11588943e-02 5.26773393e-01 3.42685670e-01
9.15023029e-01 -3.38780373e-01 5.59157550e-01 5.36939561e-01
5.33026278e-01 9.50939208e-02 8.96041095e-01 8.22619796e-02
-5.38954198e-01 -7.41857886e-01 7.12392032e-02 8.82411480e-01
-7.88324401e-02 -9.15751934e-01 -2.38991186e-01 -1.33619487e+00
-1.31922960e-01 8.79583359e-01 -5.87421417e-01 -2.28101715e-01
-8.70020628e-01 -5.21182656e-01 3.13634515e-01 9.77049172e-01
3.88611376e-01 -1.25697839e+00 -1.54336199e-01 4.13713962e-01
-3.58050555e-01 -1.51251554e+00 -1.82794750e-01 6.99210525e-01
-5.67063451e-01 -1.03587329e+00 -1.40646502e-01 -8.04389477e-01
7.07774937e-01 -2.51937568e-01 1.35028148e+00 3.85563761e-01
2.75330633e-01 -1.30346283e-01 -3.64177316e-01 -5.19676387e-01
-1.72189042e-01 4.43722725e-01 -4.10767168e-01 -2.95104265e-01
5.23362815e-01 -4.38379019e-01 -1.85743868e-01 -1.09845996e-01
-5.11708856e-01 1.50038332e-01 1.06674492e+00 7.74343371e-01
7.37540960e-01 2.46927649e-01 5.17915606e-01 -1.23113751e+00
7.14547157e-01 -7.76305795e-01 -2.95523256e-01 7.62820125e-01
-4.76300865e-01 5.62752366e-01 5.51477492e-01 -2.29417697e-01
-1.11622429e+00 -3.97802979e-01 -9.26650390e-02 2.47058757e-02
-6.30763993e-02 1.11205733e+00 -4.85207587e-01 7.27648199e-01
1.66598275e-01 -1.14199765e-01 -9.62559283e-01 -3.35160047e-01
3.68875325e-01 6.15221679e-01 5.96604705e-01 -7.43129134e-01
6.17895007e-01 -1.33725509e-01 1.22903168e-01 -3.06746870e-01
-1.64470243e+00 -2.78019905e-01 -8.41635883e-01 2.29495987e-01
9.29018080e-01 -8.28257620e-01 -6.38882041e-01 -3.58251870e-01
-1.57018650e+00 -3.31261158e-01 -3.17178905e-01 5.62011957e-01
2.21973145e-03 1.15434695e-02 -9.17654276e-01 -8.37276340e-01
-4.98845518e-01 -6.36909783e-01 9.93854165e-01 2.72248328e-01
-3.62851411e-01 -7.83066034e-01 8.00212175e-02 1.58245280e-01
-2.24200249e-01 -1.57402590e-01 1.13587725e+00 -9.88481998e-01
-7.74764061e-01 -4.45458204e-01 -5.84181964e-01 2.87986457e-01
2.19940513e-01 1.20533705e-01 -8.52395236e-01 3.00270855e-01
-4.27159220e-01 -5.68532050e-01 9.40374434e-01 -6.87290877e-02
1.27598453e+00 -6.63498282e-01 -6.51682556e-01 1.90816730e-01
8.93411040e-01 -7.07579181e-02 2.89475411e-01 2.94108868e-01
6.23171329e-01 8.74159098e-01 8.37093055e-01 5.42725474e-02
7.88055480e-01 4.48839694e-01 1.36930123e-01 4.36071604e-02
-2.73543745e-01 -5.61042666e-01 3.56075794e-01 1.05977392e+00
-5.91450110e-02 -6.86285079e-01 -6.96759045e-01 4.20000315e-01
-2.00294709e+00 -8.18772674e-01 -1.01344258e-01 1.63642144e+00
1.27086818e+00 6.89708054e-01 -3.94013703e-01 1.30534649e-01
2.89508611e-01 -2.59223897e-02 -3.87700051e-01 -2.79739171e-01
-2.84436196e-01 2.15824500e-01 1.30778223e-01 6.73301458e-01
-1.06687009e+00 1.11277699e+00 5.66814995e+00 6.24522448e-01
-7.60621071e-01 -1.22419946e-01 6.89530134e-01 3.88024189e-02
-4.12613481e-01 2.21045911e-01 -1.26672685e+00 1.29614606e-01
1.01053846e+00 -1.22608997e-01 -2.21711069e-01 5.70142984e-01
-2.11851716e-01 -5.28109819e-02 -1.32342136e+00 3.74947071e-01
-3.58316898e-02 -1.43126500e+00 1.39437735e-01 6.49281964e-02
5.93660235e-01 -8.18982795e-02 -3.53950113e-01 5.26370645e-01
5.91365278e-01 -9.66575742e-01 4.97646123e-01 5.83210468e-01
4.51551020e-01 -5.83783090e-01 1.00118506e+00 4.21072930e-01
-1.33189476e+00 2.19357926e-02 -4.11500096e-01 7.63966665e-02
3.75135779e-01 7.38318682e-01 -1.11747837e+00 6.66541457e-01
3.53225172e-01 7.12765098e-01 -5.70950985e-01 6.99280918e-01
-1.15622163e+00 8.11568260e-01 -2.91620553e-01 -9.33467448e-02
1.51083961e-01 2.60703117e-01 3.02971572e-01 1.52018642e+00
2.96893734e-02 3.90479028e-01 -1.39694046e-02 9.10419285e-01
-8.07957709e-01 8.37859958e-02 -4.16885018e-01 -1.91024423e-01
5.63884497e-01 1.31798935e+00 -4.25318122e-01 -8.14148009e-01
-8.02211940e-01 8.70596707e-01 1.08664942e+00 3.12595963e-01
-6.15201056e-01 -6.10303044e-01 1.62233785e-01 -3.47855717e-01
5.59092939e-01 -1.14489660e-01 -1.67931929e-01 -1.17848551e+00
2.75560409e-01 -4.12995100e-01 6.61215782e-01 -8.79545510e-01
-1.16272402e+00 7.70035505e-01 -5.26967980e-02 -7.11762130e-01
-3.70343000e-01 -6.05577052e-01 -7.70346999e-01 9.67171848e-01
-1.88460112e+00 -1.14015782e+00 1.95059523e-01 2.11087048e-01
5.83796978e-01 5.61421327e-02 6.73579693e-01 1.81330815e-01
-8.86376262e-01 6.23843193e-01 -6.63844347e-01 4.72493768e-01
5.25490999e-01 -1.54528344e+00 4.33167249e-01 9.33124006e-01
6.57343388e-01 8.87457073e-01 1.96299955e-01 -8.97348046e-01
-9.09084201e-01 -1.07927680e+00 1.79400849e+00 -4.88496929e-01
8.23946953e-01 -2.42253393e-01 -1.12391961e+00 1.09840155e+00
4.37224627e-01 1.35088339e-01 7.71360338e-01 6.02741838e-01
-4.58105117e-01 1.53004020e-01 -6.04210138e-01 6.44871891e-01
1.00286102e+00 -6.87063992e-01 -1.04709196e+00 1.97329789e-01
1.07645619e+00 -3.56948346e-01 -1.03032541e+00 6.45578682e-01
1.34653330e-01 -1.25586793e-01 9.80481625e-01 -1.08334827e+00
1.01149189e+00 -1.21062562e-01 -1.67149886e-01 -9.43302274e-01
-4.09656316e-01 -2.24648744e-01 -1.30168283e+00 1.70777380e+00
1.30215085e+00 -2.58272976e-01 6.57086074e-01 5.84261000e-01
-1.31338984e-01 -1.31733084e+00 -2.77146339e-01 -4.15811896e-01
-6.37703016e-02 -4.83326524e-01 7.23126292e-01 6.26999676e-01
5.23547232e-01 1.12170196e+00 -1.34555936e-01 5.31749845e-01
1.80194452e-01 2.95380026e-01 5.52971900e-01 -1.00080621e+00
-6.29877627e-01 -2.78334022e-01 2.49382794e-01 -1.28448784e+00
5.36937416e-01 -1.14748573e+00 1.68361068e-01 -1.84368277e+00
4.19633746e-01 -3.23630065e-01 -6.11855686e-01 6.65736496e-01
-7.24876881e-01 -2.84348279e-01 -2.45849453e-02 -1.87881931e-03
-8.69347632e-01 4.91317570e-01 1.24029684e+00 -2.92144686e-01
-1.84275191e-02 1.90984178e-02 -8.82070601e-01 8.66816938e-01
5.20166516e-01 -5.58646202e-01 -4.63907868e-01 -5.88191152e-01
6.89292133e-01 1.48393691e-01 6.21377379e-02 -5.98601460e-01
4.51677740e-01 -1.86619937e-01 6.14961982e-01 -7.85907328e-01
1.84870407e-01 -7.24454641e-01 -6.32500887e-01 9.76636559e-02
-9.38429654e-01 -1.30955428e-01 1.85792908e-01 3.99764299e-01
-4.50220376e-01 -4.93727744e-01 -4.05344479e-02 1.29513651e-01
-2.87165314e-01 3.52989674e-01 -1.48565069e-01 1.21333964e-01
5.82494140e-01 4.46475834e-01 -5.10037422e-01 -1.40357122e-01
-7.50425100e-01 5.41825175e-01 -3.60660076e-01 3.18050951e-01
7.30025589e-01 -1.21983933e+00 -7.70122230e-01 -8.16701576e-02
1.48144603e-01 5.59924901e-01 -7.35171512e-02 6.78484976e-01
1.91963300e-01 6.73198819e-01 5.49405038e-01 4.99448627e-02
-1.29477477e+00 6.98300838e-01 2.13819444e-02 -1.02405632e+00
-6.87014043e-01 1.30876756e+00 4.79009422e-03 -1.07143424e-01
3.49293947e-01 -7.55732119e-01 -6.10101879e-01 -5.61573394e-02
5.84985495e-01 -1.58442929e-01 8.71905535e-02 -2.71559477e-01
-4.21688735e-01 2.89902717e-01 -4.70669091e-01 -5.19894548e-02
1.52374279e+00 -2.59427488e-01 -2.43938163e-01 2.52451390e-01
1.03472137e+00 4.36202824e-01 -1.15237033e+00 -6.83551729e-01
7.80043423e-01 5.51101156e-02 2.94649173e-02 -9.78703558e-01
-8.09997559e-01 7.37352610e-01 -3.85106862e-01 6.21623434e-02
8.40831220e-01 7.53589094e-01 8.03703249e-01 9.08933222e-01
-1.44431084e-01 -7.67495871e-01 1.51340485e-01 7.39005268e-01
1.08988118e+00 -1.27711976e+00 1.54645652e-01 -9.00342882e-01
-6.28865898e-01 1.04244280e+00 9.41889286e-01 6.02853857e-02
5.97870052e-01 6.54491067e-01 -4.23989266e-01 -5.37542343e-01
-1.42123604e+00 -4.59579021e-01 8.18897784e-01 1.78949773e-01
8.65112126e-01 -2.03573704e-01 -2.73248315e-01 1.23710108e+00
-1.20770015e-01 -5.31384051e-02 1.66968748e-01 8.78322303e-01
-4.08612967e-01 -1.33466768e+00 1.01395622e-01 6.89477265e-01
-5.97036004e-01 -4.76268470e-01 -6.76741660e-01 4.05529350e-01
2.35253945e-01 8.50767016e-01 -4.72816899e-02 -3.51423144e-01
4.20194119e-01 3.08963358e-01 4.55723286e-01 -1.00439322e+00
-4.01571631e-01 -5.88168167e-02 7.34472454e-01 -1.98557466e-01
-1.97963089e-01 -4.42980498e-01 -1.71807289e+00 2.37765625e-01
-5.27058482e-01 3.04172099e-01 1.96731806e-01 1.55147767e+00
4.25657302e-01 1.07648790e+00 3.03801835e-01 -1.76493779e-01
-1.10935614e-01 -1.18776155e+00 -1.38316244e-01 1.55510783e-01
2.84539253e-01 -5.49233675e-01 8.78341720e-02 2.78675910e-02] | [9.303592681884766, 8.596776008605957] |
c9162ea7-6e53-4d3c-9ce2-4ad0969a5569 | joint-deep-learning-for-car-detection | 1412.7854 | null | http://arxiv.org/abs/1412.7854v2 | http://arxiv.org/pdf/1412.7854v2.pdf | Joint Deep Learning for Car Detection | Traditional object recognition approaches apply feature extraction, part
deformation handling, occlusion handling and classification sequentially while
they are independent from each other. Ouyang and Wang proposed a model for
jointly learning of all of the mentioned processes using one deep neural
network. We utilized, and manipulated their toolbox in order to apply it in car
detection scenarios where it had not been tested. Creating a single deep
architecture from these components, improves the interaction between them and
can enhance the performance of the whole system. We believe that the approach
can be used as a general purpose object detection toolbox. We tested the
algorithm on UIUC car dataset, and achieved an outstanding result. The accuracy
of our method was 97 % while the previously reported results showed an accuracy
of up to 91 %. We strongly believe that having an experiment on a larger
dataset can show the advantage of using deep models over shallow ones. | ['Seyedshams Feyzabadi'] | 2014-12-25 | null | null | null | null | ['occlusion-handling'] | ['computer-vision'] | [-4.00174893e-02 -1.50624961e-01 1.41955882e-01 -2.47887388e-01
-3.93464208e-01 -4.27817553e-01 5.77461839e-01 -2.22647965e-01
-4.73264515e-01 2.93084770e-01 -4.44872320e-01 -2.07864150e-01
1.59087643e-01 -8.07800829e-01 -7.94671297e-01 -7.78640032e-01
1.91158969e-02 5.20568252e-01 9.27984178e-01 -5.86861782e-02
2.47638151e-01 9.54677403e-01 -2.07024789e+00 3.66772234e-01
5.39498508e-01 9.61425245e-01 3.85605514e-01 5.83839476e-01
7.54223671e-03 7.12869942e-01 -5.72692275e-01 -3.47642601e-01
4.86150414e-01 2.49715105e-01 -6.31414533e-01 2.59207189e-01
5.03760338e-01 -4.82770711e-01 -2.34356165e-01 7.73127556e-01
5.65611482e-01 7.10746571e-02 5.58842480e-01 -1.08187306e+00
-3.94236922e-01 4.66917038e-01 -5.22926271e-01 -2.19760299e-01
3.54834720e-02 2.38923967e-01 5.40370941e-01 -9.30043995e-01
5.07702768e-01 1.09769809e+00 7.47349024e-01 5.70389807e-01
-9.90684807e-01 -6.61045611e-01 8.17355067e-02 2.08051071e-01
-1.29690742e+00 -3.36602062e-01 7.65502870e-01 -4.96631503e-01
8.99830878e-01 2.14837268e-01 5.87896287e-01 8.99559498e-01
1.78042892e-02 1.05682123e+00 8.21803689e-01 -4.48320150e-01
9.75710806e-04 4.62075651e-01 5.07878482e-01 7.73259044e-01
5.67679107e-01 2.95623064e-01 4.23174687e-02 8.26860666e-02
7.60714769e-01 1.29128426e-01 -4.53223400e-02 -7.15301096e-01
-1.00085735e+00 7.38044739e-01 5.06752789e-01 5.60910344e-01
-1.06379688e-01 1.11525498e-01 3.88210058e-01 1.76164135e-01
7.68527016e-02 1.17084369e-01 -4.85816985e-01 2.45090380e-01
-8.48042548e-01 1.56251416e-01 8.40660572e-01 1.07982862e+00
6.58714712e-01 1.30976021e-01 6.97536767e-02 7.86701620e-01
2.64991641e-01 2.20031366e-01 4.69236970e-01 -7.67767906e-01
-1.92817790e-03 8.27222347e-01 1.57873765e-01 -7.62965918e-01
-5.17855227e-01 -4.51083541e-01 -5.09977460e-01 7.41935134e-01
7.22817719e-01 -2.16918498e-01 -1.25717139e+00 1.33054900e+00
2.83155233e-01 1.16793416e-01 -1.06607340e-01 9.33488250e-01
9.52826381e-01 3.87620598e-01 5.79198152e-02 3.60576838e-01
1.38979840e+00 -1.04040170e+00 -5.26770592e-01 -9.01264250e-02
6.18392467e-01 -1.03099549e+00 8.39845240e-01 7.56417394e-01
-9.71483469e-01 -9.98886645e-01 -1.40180278e+00 -9.41769630e-02
-5.55205345e-01 8.76899481e-01 1.02839816e+00 7.73859739e-01
-9.93609846e-01 6.53224707e-01 -1.02602839e+00 -5.15565634e-01
4.90078956e-01 6.19269907e-01 -3.72234285e-01 1.95631042e-01
-6.28531575e-01 1.04913950e+00 4.92863327e-01 5.19399285e-01
-9.09649014e-01 -5.23232818e-02 -5.64942181e-01 9.44111347e-02
6.03128791e-01 -5.24224043e-01 1.04514694e+00 -7.81936228e-01
-1.54345810e+00 8.10436249e-01 4.19257954e-02 -4.26349968e-01
7.99188852e-01 -4.72414970e-01 -3.19557995e-01 -1.11097962e-01
-4.07521874e-01 7.95724571e-01 7.91926801e-01 -1.20790696e+00
-7.20947266e-01 -3.49061817e-01 1.08909540e-01 -3.51523846e-01
-1.70052111e-01 -6.73789904e-02 -6.05532944e-01 -3.76758039e-01
7.91393965e-02 -1.02559626e+00 -6.69813529e-02 2.16376200e-01
-1.40271604e-01 -3.40766877e-01 1.15372252e+00 -4.95304257e-01
6.83101594e-01 -2.19067383e+00 -9.93823484e-02 -6.23782650e-02
-6.98237419e-02 7.23419607e-01 -2.73172945e-01 2.23144084e-01
-9.79798287e-02 -1.00102901e-01 -7.09243640e-02 -2.37066373e-01
-1.47638813e-01 4.53744605e-02 2.74059977e-02 6.53912783e-01
4.21153516e-01 6.31739736e-01 -2.65650928e-01 -3.37024391e-01
4.96232897e-01 5.45467257e-01 -3.97445112e-01 9.19784084e-02
-6.66857809e-02 -1.29611284e-01 -3.10608715e-01 8.32232177e-01
1.01487935e+00 3.02481294e-01 -5.52564450e-02 -3.93949419e-01
-1.04173779e-01 -3.54898274e-01 -1.63063133e+00 1.28091168e+00
-2.80580848e-01 7.81727195e-01 1.27878293e-01 -1.00124943e+00
1.15523899e+00 3.10684085e-01 2.60666072e-01 -4.08314079e-01
5.29302359e-01 1.78789720e-01 4.24918383e-01 -5.78025937e-01
2.48676628e-01 6.65802956e-02 2.66531676e-01 1.33588478e-01
3.41457546e-01 -1.45155817e-01 3.87291610e-01 -2.15420611e-02
8.64023983e-01 3.41399252e-01 5.11631593e-02 -2.73153096e-01
7.22764313e-01 6.15461878e-02 1.14928626e-01 9.06755030e-01
-3.99480969e-01 7.25308537e-01 3.58564377e-01 -6.59432411e-01
-9.08629656e-01 -9.39693987e-01 -2.30225116e-01 1.00381768e+00
8.22504759e-02 7.86342844e-02 -6.23589575e-01 -8.78566384e-01
2.20685229e-02 5.73253810e-01 -5.39519846e-01 -2.05903016e-02
-7.98178077e-01 -7.82651186e-01 3.19109082e-01 8.69370937e-01
4.36789393e-01 -1.18751800e+00 -5.88061988e-01 1.84423149e-01
4.26982462e-01 -1.20750093e+00 1.34910434e-01 4.09733862e-01
-8.68899465e-01 -1.23588002e+00 -6.20524049e-01 -9.63901043e-01
4.00471061e-01 2.68348694e-01 7.67454922e-01 3.53990376e-01
-5.65369725e-01 1.25854716e-01 -2.13557094e-01 -6.61271989e-01
-3.22384387e-01 1.60036817e-01 -5.34161925e-02 -9.06286985e-02
3.33344221e-01 -3.02798003e-01 -3.26342463e-01 4.08911824e-01
-7.71595538e-01 -2.07537279e-01 7.97221005e-01 7.05132604e-01
2.02809632e-01 -7.48928487e-02 5.77543497e-01 -7.63068736e-01
2.72128493e-01 1.23405635e-01 -1.00054610e+00 3.87099050e-02
-3.17626983e-01 -7.21926391e-02 3.77949446e-01 -5.51763356e-01
-1.07478523e+00 6.38121068e-01 -4.22205389e-01 -2.15295598e-01
-5.44877529e-01 -8.41503665e-02 -3.68836671e-01 -3.73306990e-01
4.24280852e-01 1.22342192e-01 5.85047379e-02 -6.39363229e-01
2.31879592e-01 6.06235743e-01 4.50486988e-01 -2.91136891e-01
6.79889679e-01 4.26599592e-01 -1.60116524e-01 -7.71821499e-01
-4.65399623e-01 -5.22344351e-01 -8.97795320e-01 -1.87743083e-01
7.52949715e-01 -9.46132004e-01 -9.49986100e-01 7.87909687e-01
-1.25213253e+00 -8.83387700e-02 -8.79343152e-02 5.45025766e-01
-2.51737416e-01 3.85734260e-01 -5.45488119e-01 -9.37342465e-01
-1.30777732e-01 -1.27323949e+00 9.95469093e-01 2.76405543e-01
1.52318776e-01 -7.03770578e-01 -8.51052105e-02 2.68121272e-01
5.03432870e-01 5.71007691e-02 3.57612997e-01 -8.98224413e-01
-6.84358954e-01 -7.58553565e-01 -2.94482797e-01 6.01172805e-01
-1.10411793e-02 4.55640703e-01 -1.26498020e+00 -1.74155936e-01
8.20047855e-02 -2.39927128e-01 1.21974134e+00 4.95581664e-02
1.37649107e+00 2.70049185e-01 -5.05512416e-01 1.80412576e-01
1.41464424e+00 3.42446357e-01 7.79854238e-01 4.16242778e-01
6.41558886e-01 4.59967762e-01 4.91635591e-01 3.83152291e-02
-7.43835568e-02 8.60839486e-01 6.43391430e-01 -2.69154698e-01
-3.07888716e-01 3.89597058e-01 4.42661524e-01 2.75092155e-01
-2.54324645e-01 -2.11472929e-01 -8.74825716e-01 4.64668840e-01
-1.81010091e+00 -9.49812710e-01 -5.21050990e-01 1.95308232e+00
3.15311521e-01 4.84516263e-01 1.81291476e-01 1.79182917e-01
6.01895034e-01 -2.59239733e-01 -4.11293507e-01 -4.11326110e-01
-1.21614644e-02 1.52203515e-01 3.37112576e-01 4.19711351e-01
-1.50275254e+00 9.61502016e-01 6.57394838e+00 7.43626833e-01
-1.28621125e+00 -7.01560974e-02 3.89268041e-01 -2.66069286e-02
4.57402408e-01 -2.10207388e-01 -1.05647790e+00 1.15241431e-01
7.22222745e-01 4.27211434e-01 1.37316147e-02 1.16667449e+00
-7.41922408e-02 -3.10584456e-01 -1.20203722e+00 7.03781903e-01
4.50310335e-02 -1.11631155e+00 -2.02313601e-03 -1.87332049e-01
3.35197151e-01 1.59088239e-01 -3.59038740e-01 5.12013614e-01
1.57276139e-01 -9.21780825e-01 6.20649397e-01 5.16520321e-01
-1.72257107e-02 -7.68019497e-01 1.14431000e+00 5.57124496e-01
-1.11855674e+00 -9.28307921e-02 -4.49470133e-01 -4.52666879e-02
3.46643366e-02 4.29693043e-01 -8.18283021e-01 6.03686929e-01
5.81658781e-01 4.21312541e-01 -1.05727768e+00 1.40715790e+00
8.94308984e-02 5.62526822e-01 -5.68134964e-01 -1.54535145e-01
8.97377208e-02 -1.57212764e-02 3.15528750e-01 1.34426296e+00
2.31116593e-01 -4.06933308e-01 1.74613759e-01 6.76312745e-01
2.78309405e-01 6.17668629e-02 -6.43915355e-01 2.17566103e-01
-3.32608931e-02 1.60744977e+00 -1.05291939e+00 -3.79960239e-01
-5.28108120e-01 7.12704003e-01 1.91629156e-01 3.43636773e-03
-9.71702814e-01 -4.35401291e-01 3.43071640e-01 -3.88047937e-03
8.43314469e-01 -2.15173185e-01 -5.43754816e-01 -1.09422386e+00
6.98717013e-02 -7.20651329e-01 6.77680373e-02 -6.34304106e-01
-1.09109986e+00 7.90921926e-01 -5.25355600e-02 -1.25065362e+00
5.88311329e-02 -1.25975931e+00 -6.35366261e-01 6.20618343e-01
-1.14467931e+00 -1.30183399e+00 -3.94621968e-01 4.27528381e-01
4.99108255e-01 -3.34864080e-01 6.59224629e-01 4.39140528e-01
-8.80700529e-01 5.62928021e-01 -3.14934760e-01 3.49656522e-01
5.02136409e-01 -1.05847156e+00 1.13903530e-01 1.04101825e+00
3.22407275e-01 4.50549096e-01 6.18972361e-01 -4.13640201e-01
-1.28504241e+00 -8.91822696e-01 4.56899971e-01 -5.77880263e-01
5.46654880e-01 -5.91712773e-01 -1.13129985e+00 6.13859117e-01
3.66801739e-01 -3.71769704e-02 3.48721534e-01 -1.31482258e-01
3.46924290e-02 -2.90387362e-01 -1.15480864e+00 4.01591659e-01
9.26425278e-01 1.48499692e-02 -6.62802756e-01 2.48424798e-01
3.28113168e-01 -3.65333080e-01 -5.43548703e-01 6.30328536e-01
7.65398264e-01 -1.25942564e+00 9.06180739e-01 -2.23296836e-01
1.01316974e-01 -5.79606175e-01 -9.86718833e-02 -8.11461091e-01
-2.98734576e-01 9.86078754e-02 -1.36401132e-01 1.42081177e+00
3.67617011e-01 -7.16469526e-01 8.23088706e-01 3.29072058e-01
-2.18394324e-01 -6.61991954e-01 -6.65717125e-01 -7.92887449e-01
1.06615210e-02 -6.16689265e-01 5.04195929e-01 3.64676654e-01
-5.55810094e-01 2.90257722e-01 -1.40363708e-01 2.08560348e-01
3.60202044e-01 1.49167582e-01 1.00924110e+00 -1.51411247e+00
-2.68194199e-01 -4.30047065e-01 -6.32219851e-01 -7.32156754e-01
6.42947853e-02 -6.40987933e-01 7.81506021e-03 -1.41718006e+00
3.02248120e-01 -3.61529589e-01 -2.44539291e-01 7.41775274e-01
-2.62841629e-03 5.06848454e-01 4.15936142e-01 -1.10845938e-02
-1.30337164e-01 3.34206760e-01 1.23680985e+00 -9.95192900e-02
-6.56441823e-02 2.59363353e-01 -5.16400278e-01 1.03390741e+00
7.44826317e-01 -3.73091817e-01 -1.05249330e-01 -2.76282609e-01
-5.31550288e-01 -4.26740646e-01 5.37274599e-01 -1.45051873e+00
2.43084639e-01 2.81453907e-01 8.21512043e-01 -9.10597682e-01
4.42578614e-01 -1.23350906e+00 9.50191468e-02 6.65143430e-01
1.19708098e-01 -2.38106743e-01 6.31083310e-01 2.05098897e-01
-2.12376475e-01 -4.82906729e-01 9.14047241e-01 -1.19935028e-01
-1.01323938e+00 -1.02069974e-01 -4.44664925e-01 -7.37686098e-01
1.30293512e+00 -3.64788234e-01 -8.82339999e-02 -5.82564697e-02
-9.37609255e-01 -3.62243541e-02 3.37292314e-01 4.88059551e-01
4.35817689e-01 -1.01148021e+00 -5.81094742e-01 4.77717191e-01
-1.17723802e-02 -1.23413906e-01 -5.34293279e-02 7.53677547e-01
-6.50837004e-01 3.87901336e-01 -3.58733177e-01 -8.57110143e-01
-1.44808507e+00 7.74873614e-01 4.65830624e-01 -2.04986587e-01
-6.01535618e-01 5.90056121e-01 6.46903887e-02 -4.56461459e-01
4.68910694e-01 -4.65491056e-01 -4.94684994e-01 1.57422990e-01
5.07115364e-01 3.74107480e-01 4.20930862e-01 -3.45602006e-01
-5.98546922e-01 7.19572544e-01 -3.73340905e-01 1.73992306e-01
1.51461065e+00 1.83745041e-01 1.73597440e-01 3.02799910e-01
9.45231736e-01 8.29520226e-02 -1.22568107e+00 2.74527043e-01
5.39341848e-03 -1.82319835e-01 5.65597676e-02 -8.31697702e-01
-1.29276180e+00 8.82517934e-01 1.06963265e+00 1.91918582e-01
1.10430884e+00 -3.86431580e-03 3.91469628e-01 6.72920346e-01
3.62492204e-01 -7.54171729e-01 -3.49018052e-02 5.80049217e-01
9.01479304e-01 -1.41835082e+00 1.09953448e-01 -5.08415997e-01
-4.11769181e-01 1.50602520e+00 8.80780935e-01 -5.45755804e-01
6.02760077e-01 6.83561921e-01 1.32586107e-01 -8.67591724e-02
-6.61896944e-01 -5.40243030e-01 4.74627733e-01 6.30017579e-01
6.03770137e-01 -9.96016115e-02 -3.62735987e-01 5.07912755e-01
2.95389369e-02 3.21620882e-01 5.15633285e-01 9.08453822e-01
-4.84829575e-01 -1.19269812e+00 -4.08759147e-01 2.36607194e-01
-4.00563985e-01 3.64799112e-01 -3.85727406e-01 1.34288895e+00
5.20065069e-01 7.18469381e-01 -1.03348374e-01 -3.46717924e-01
5.66755712e-01 1.31855369e-01 5.41786551e-01 -5.02412677e-01
-7.67167926e-01 1.32104799e-01 9.88216475e-02 -6.00203454e-01
-3.94571602e-01 -6.81715250e-01 -1.04920614e+00 -3.16751264e-02
-6.28585637e-01 -2.70770609e-01 6.42938077e-01 9.67231631e-01
1.72756553e-01 6.05103254e-01 2.98627019e-01 -1.08873034e+00
-4.73026514e-01 -1.08182216e+00 -4.90862906e-01 1.25530437e-01
2.03768730e-01 -8.95811141e-01 -2.04434693e-01 4.08736877e-02] | [8.553661346435547, -0.5197449922561646] |
17efc550-0b09-4805-8232-635456f3eef5 | primal-dual-algorithms-for-non-negative | 1412.1788 | null | http://arxiv.org/abs/1412.1788v1 | http://arxiv.org/pdf/1412.1788v1.pdf | Primal-Dual Algorithms for Non-negative Matrix Factorization with the Kullback-Leibler Divergence | Non-negative matrix factorization (NMF) approximates a given matrix as a
product of two non-negative matrices. Multiplicative algorithms deliver
reliable results, but they show slow convergence for high-dimensional data and
may be stuck away from local minima. Gradient descent methods have better
behavior, but only apply to smooth losses such as the least-squares loss. In
this article, we propose a first-order primal-dual algorithm for non-negative
decomposition problems (where one factor is fixed) with the KL divergence,
based on the Chambolle-Pock algorithm. All required computations may be
obtained in closed form and we provide an efficient heuristic way to select
step-sizes. By using alternating optimization, our algorithm readily extends to
NMF and, on synthetic examples, face recognition or music source separation
datasets, it is either faster than existing algorithms, or leads to improved
local optima, or both. | ['Francis Bach', 'Felipe Yanez'] | 2014-12-04 | null | null | null | null | ['music-source-separation'] | ['music'] | [ 1.18252791e-01 -7.85052553e-02 -3.29790682e-01 -1.38990387e-01
-1.18985093e+00 -6.62702262e-01 3.40281814e-01 -1.18201472e-01
-3.51780802e-01 9.03620064e-01 7.68949166e-02 -4.21995133e-01
-5.05966127e-01 -5.37493169e-01 -6.11637414e-01 -9.28762376e-01
-3.24677169e-01 5.70441544e-01 -2.71097749e-01 -2.47538656e-01
3.95366624e-02 2.89324045e-01 -1.20995891e+00 2.20818788e-01
6.66320503e-01 9.35377955e-01 -8.91174078e-02 7.24875867e-01
4.00151163e-02 5.58373272e-01 -2.15108961e-01 -6.51552379e-01
6.16479218e-01 -5.04479349e-01 -7.71954596e-01 1.18853003e-01
4.20283377e-01 2.42968295e-02 3.08309346e-01 1.12529826e+00
4.62565392e-01 8.26614946e-02 6.75506651e-01 -1.41699207e+00
-4.84605819e-01 3.32113355e-01 -1.13849723e+00 -1.40924202e-02
3.15687150e-01 -3.63222420e-01 1.19854951e+00 -1.25132537e+00
3.39081585e-01 1.59682131e+00 1.19492257e+00 3.70220304e-01
-1.91180182e+00 -5.86133182e-01 1.00905634e-01 -1.37990937e-01
-1.19733024e+00 -4.07861114e-01 5.01207650e-01 -4.99795020e-01
8.38157535e-01 6.40541553e-01 5.31690419e-01 7.11201549e-01
-5.95794506e-02 8.93701971e-01 9.72425401e-01 -5.91000021e-01
8.51454139e-02 -4.00956832e-02 -5.89149334e-02 5.89882195e-01
2.30746374e-01 -1.10779494e-01 -4.80070412e-01 -8.26550663e-01
4.04534370e-01 -8.93627852e-02 -2.59651005e-01 -6.21164560e-01
-1.28568101e+00 1.09941077e+00 -6.35475107e-03 1.54319242e-01
-3.71447682e-01 2.61863530e-01 4.09162074e-01 6.15231872e-01
6.94373429e-01 2.81458467e-01 -6.03952467e-01 -2.39623398e-01
-1.11355686e+00 4.34951454e-01 1.02009213e+00 4.62701917e-01
7.87138879e-01 -9.52277407e-02 4.63665761e-02 1.00730968e+00
2.24863082e-01 4.95214462e-01 2.32757658e-01 -1.03148329e+00
4.89585757e-01 1.52521208e-01 2.86665231e-01 -1.29234171e+00
-5.76268375e-01 -3.19062114e-01 -1.15044951e+00 3.17158580e-01
5.85397303e-01 -2.80683309e-01 -5.52755594e-01 1.97166944e+00
3.65201503e-01 1.42353117e-01 -9.73357558e-02 9.55916286e-01
8.86714906e-02 6.28220379e-01 -3.73800278e-01 -8.89545083e-01
9.49265957e-01 -9.33973432e-01 -9.05104518e-01 -1.90022975e-01
5.16950130e-01 -9.93928969e-01 9.15390074e-01 8.99352849e-01
-1.27565587e+00 -1.37134790e-01 -9.27611351e-01 1.12458415e-01
-5.13048768e-02 2.55455643e-01 8.28680038e-01 8.87652040e-01
-1.19646502e+00 8.22338104e-01 -7.75309861e-01 8.08045194e-02
6.25138283e-02 6.08006358e-01 -3.30675662e-01 7.25994632e-02
-8.68052781e-01 5.05844593e-01 1.24392428e-01 2.11838931e-01
-4.13948447e-01 -9.95287001e-01 -6.16340518e-01 -1.49518162e-01
1.93641677e-01 -8.50426078e-01 8.96332145e-01 -1.43584263e+00
-1.40039814e+00 7.94612288e-01 -3.97386372e-01 -4.60135609e-01
7.13957369e-01 -2.32733935e-01 -1.77080601e-01 -8.47423300e-02
1.80000097e-01 2.83476800e-01 1.22433877e+00 -1.33517230e+00
-3.05968881e-01 -3.96610498e-01 -5.33939116e-02 4.20605801e-02
-4.49096769e-01 1.11522846e-01 -9.80971605e-02 -9.25172746e-01
3.04314494e-01 -1.13009620e+00 -5.06101727e-01 9.22091231e-02
-1.90003097e-01 9.36734723e-04 5.96941292e-01 -6.42904758e-01
1.56762362e+00 -2.18623853e+00 3.65677118e-01 4.53353137e-01
1.59703985e-01 1.34602770e-01 -2.14155748e-01 4.81440932e-01
-4.36867654e-01 9.80502553e-03 -6.19722664e-01 -5.16576350e-01
-5.68890236e-02 2.56284803e-01 -1.96207464e-01 7.05037475e-01
1.64035544e-01 4.04145777e-01 -9.78769958e-01 -1.23908944e-01
-1.31227955e-01 3.26054990e-01 -8.36634159e-01 -3.42619151e-01
1.68704420e-01 9.60176997e-03 2.45561283e-02 3.70973170e-01
1.00785875e+00 -1.29624262e-01 3.53570253e-01 -2.01752320e-01
-8.00412446e-02 8.72605518e-02 -1.83801627e+00 1.59609663e+00
-5.58322847e-01 6.29620612e-01 8.13384235e-01 -1.33341587e+00
6.72749162e-01 2.39191040e-01 7.84662008e-01 -2.37961233e-01
-1.44630432e-01 3.84169132e-01 -2.16139570e-01 -1.32369295e-01
3.44104230e-01 -5.33711076e-01 3.43281388e-01 6.01008654e-01
-4.40187305e-02 1.12313837e-01 5.49640536e-01 2.24556893e-01
9.37350035e-01 -2.69084871e-01 1.76110938e-01 -5.94641864e-01
7.13171363e-01 -3.31816733e-01 7.29469240e-01 6.01875067e-01
1.06732406e-01 8.44407082e-01 7.62207270e-01 -4.52305585e-01
-8.02554786e-01 -1.12698126e+00 -3.41183364e-01 1.14509308e+00
-1.81593999e-01 -8.90961349e-01 -6.21078551e-01 -5.40999591e-01
2.61416763e-01 3.75785738e-01 -5.00945389e-01 -5.83115146e-02
-4.14036155e-01 -1.27127552e+00 2.23119810e-01 2.73941994e-01
-3.47770341e-02 -3.34730089e-01 -1.54875393e-03 2.52302647e-01
-1.95358291e-01 -6.72384679e-01 -5.50483167e-01 3.80458444e-01
-1.18121207e+00 -9.51258183e-01 -8.85730326e-01 -6.28446102e-01
7.79103577e-01 2.73129910e-01 1.29367554e+00 -4.23883982e-02
-1.55981794e-01 3.66132021e-01 -1.81704722e-02 -4.69874702e-02
-2.78738916e-01 -1.49958327e-01 6.34902775e-01 3.14971536e-01
2.10150972e-01 -7.13788629e-01 -5.83451986e-01 3.83463860e-01
-7.54574776e-01 -2.42852390e-01 2.64563978e-01 1.30772972e+00
8.38993907e-01 6.96108192e-02 6.16104305e-01 -9.10231888e-01
1.05411339e+00 -4.61548448e-01 -5.02880752e-01 2.21718505e-01
-9.23527360e-01 2.74244279e-01 7.28665173e-01 -5.45773268e-01
-7.49449015e-01 3.15240860e-01 7.88856745e-02 -4.87172931e-01
5.76418996e-01 3.97142023e-01 5.77265956e-02 -2.27635294e-01
9.71036375e-01 -7.76655376e-02 1.44356266e-01 -5.65001369e-01
4.71732050e-01 5.25193512e-01 1.38409793e-01 -8.05076063e-01
7.86035359e-01 6.67177022e-01 1.64142668e-01 -7.52236724e-01
-8.08533072e-01 -5.73374987e-01 -4.35129404e-01 -1.55331683e-03
3.00474972e-01 -9.02184904e-01 -7.92194843e-01 1.13445960e-01
-9.04923558e-01 -1.42950506e-03 -1.93858758e-01 5.84258795e-01
-8.13337147e-01 5.69090426e-01 -5.45746326e-01 -1.12514853e+00
-4.34388727e-01 -7.48130143e-01 9.42989051e-01 -1.52276322e-01
-1.78351134e-01 -9.47385252e-01 4.04526770e-01 1.99384496e-01
7.15942234e-02 1.90953463e-02 7.56523609e-01 -1.70776963e-01
-6.22832775e-02 -1.52263135e-01 -8.07752162e-02 7.65610099e-01
1.32763132e-01 1.90240011e-01 -6.64822459e-01 -7.05405653e-01
4.07017097e-02 -3.72175366e-01 1.06287062e+00 6.25631332e-01
9.89388824e-01 -5.83403885e-01 -2.29566336e-01 8.41958582e-01
1.37664545e+00 -2.33697772e-01 3.54395568e-01 2.79865950e-01
4.33284193e-01 4.89537627e-01 5.99698663e-01 7.99176455e-01
8.33235532e-02 3.71859819e-01 2.91266650e-01 -2.48713568e-01
3.21692228e-01 6.08237134e-03 5.45882106e-01 8.07420075e-01
-1.16128005e-01 2.07864180e-01 -7.54778445e-01 5.72502434e-01
-2.10643554e+00 -1.03074014e+00 -3.95581573e-01 2.46120238e+00
8.94381464e-01 -8.71929526e-02 4.55216765e-01 3.16077977e-01
5.02526104e-01 3.93335670e-02 -4.42304790e-01 -6.62603498e-01
-2.57330060e-01 1.25808716e-01 6.98317885e-01 6.11099005e-01
-1.07613254e+00 5.66466272e-01 7.57190275e+00 9.32910502e-01
-8.61088932e-01 1.53608635e-01 4.83459890e-01 -4.28702801e-01
-5.84163368e-01 7.80209824e-02 -4.86552536e-01 2.26086691e-01
6.75682843e-01 -9.05386433e-02 8.57379377e-01 7.32962430e-01
3.57611060e-01 -2.76176408e-02 -8.68538558e-01 1.44913936e+00
-6.55198842e-02 -1.08254421e+00 -3.30756277e-01 9.69367996e-02
9.94021237e-01 -1.07277229e-01 1.52137518e-01 -6.96074441e-02
1.68308407e-01 -7.57939637e-01 5.93777537e-01 1.48524359e-01
5.03790557e-01 -1.12281489e+00 3.08533102e-01 4.17032242e-01
-9.55719173e-01 -2.92136729e-01 -6.08098328e-01 -2.58987188e-01
1.43462002e-01 1.08460724e+00 -2.91252613e-01 2.88648874e-01
6.93266630e-01 6.55044496e-01 -1.89610437e-01 8.47681880e-01
4.47216667e-02 6.36620879e-01 -7.50707150e-01 6.26863316e-02
2.43414238e-01 -8.93348992e-01 7.01517582e-01 1.25378120e+00
3.54840428e-01 -4.39063720e-02 1.40725404e-01 6.66329801e-01
1.75941791e-02 4.54848558e-01 -3.67614955e-01 -6.65452778e-02
-3.86060476e-02 1.34293652e+00 -6.53372824e-01 -6.71884418e-02
-5.59630990e-01 9.48216975e-01 3.45514148e-01 5.09682477e-01
-6.48321688e-01 -1.96690589e-01 1.04554081e+00 1.22294508e-01
2.18263492e-01 -3.17299932e-01 -3.79212976e-01 -1.40074933e+00
2.36999467e-01 -1.04434884e+00 6.17769837e-01 -3.70241374e-01
-1.42523110e+00 3.30757380e-01 -2.18801886e-01 -1.12538683e+00
-4.19521153e-01 -7.83364415e-01 -2.39888698e-01 8.37498128e-01
-1.10271430e+00 -6.44528151e-01 4.29644853e-01 7.70156384e-01
1.22251034e-01 -9.49514583e-02 8.15000415e-01 6.08955801e-01
-4.42307919e-01 7.77877986e-01 7.43511617e-01 -2.48975426e-01
7.31624007e-01 -1.28446162e+00 2.10773516e-02 8.24925959e-01
3.32156807e-01 7.57810235e-01 8.99002314e-01 -3.22850823e-01
-1.40335619e+00 -5.10381162e-01 7.09793746e-01 -1.41849682e-01
8.75147521e-01 -3.05238783e-01 -6.31429970e-01 5.40722013e-01
-7.10576400e-02 1.62440613e-02 9.50727761e-01 7.10026681e-01
-3.70724112e-01 -2.84078896e-01 -1.18359005e+00 4.41253245e-01
9.51931417e-01 -3.89304310e-01 -1.67066827e-01 6.84451282e-01
2.11855426e-01 -1.57530084e-01 -7.13401914e-01 3.61955851e-01
7.02294111e-01 -1.01727057e+00 1.16773820e+00 -8.54399562e-01
1.97832733e-01 -3.83477390e-01 -2.39831150e-01 -1.25878835e+00
-5.21656275e-01 -1.18988919e+00 -2.65928209e-01 1.08985090e+00
7.27300048e-01 -5.47182739e-01 8.00910890e-01 5.58373928e-01
2.45667264e-01 -8.03638339e-01 -1.03103280e+00 -9.53127623e-01
6.96668699e-02 -7.36835778e-01 1.19991548e-01 1.00969541e+00
6.45648688e-02 4.14056927e-01 -7.33042240e-01 -1.38467923e-01
8.64140272e-01 2.46190637e-01 6.69831693e-01 -1.06535530e+00
-5.20232856e-01 -4.77848887e-01 -3.04953009e-01 -1.10871840e+00
2.84834117e-01 -1.02747977e+00 -1.94799230e-01 -1.24386299e+00
2.27345482e-01 -5.38765490e-01 -2.45421007e-01 4.76641327e-01
-9.54907760e-02 4.29732978e-01 2.24821225e-01 8.82980134e-03
-3.89560848e-01 4.02313739e-01 1.02542162e+00 -2.20879838e-01
-4.04905260e-01 2.84367502e-01 -6.97339475e-01 8.40717435e-01
5.71962595e-01 -7.25069761e-01 -3.48612696e-01 -2.63531029e-01
8.69140208e-01 1.21950559e-01 6.63724840e-02 -6.22876644e-01
-6.92281649e-02 -1.12099461e-01 2.64660478e-01 -3.98258179e-01
2.93913573e-01 -7.46803045e-01 2.71313131e-01 3.65965873e-01
-1.77355260e-01 2.37116784e-01 1.52785286e-01 6.17595136e-01
-3.23138535e-01 -4.33028758e-01 9.15643156e-01 4.91896048e-02
-6.43781275e-02 1.66909516e-01 -4.17256474e-01 1.92791939e-01
6.90826833e-01 -1.33431539e-01 1.88089699e-01 -5.77142000e-01
-9.89153683e-01 2.80566424e-01 2.56186068e-01 2.41905659e-01
4.67491776e-01 -1.74583900e+00 -7.96354532e-01 2.40801051e-01
-2.94016004e-01 -3.82011950e-01 7.57811666e-02 1.21460176e+00
-5.20685673e-01 9.49123427e-02 2.48291031e-01 -5.30434906e-01
-1.46051157e+00 6.41797423e-01 2.87947387e-01 -2.43961588e-01
-2.10461333e-01 1.06689811e+00 1.59473091e-01 -4.80931401e-01
2.59599149e-01 -1.40854552e-01 3.30497742e-01 5.59538126e-01
8.81997108e-01 6.42166376e-01 8.57998878e-02 -6.71089947e-01
-4.67120767e-01 6.44340873e-01 1.79338738e-01 -2.85645545e-01
1.37322176e+00 -1.47086442e-01 -6.82087541e-01 4.02155668e-01
1.62165093e+00 2.40242630e-01 -1.11610687e+00 -4.63881567e-02
-4.02429327e-02 -6.80872977e-01 2.97188200e-02 -3.94166648e-01
-1.21412706e+00 6.38619840e-01 5.63495100e-01 3.82319748e-01
1.39710093e+00 -3.36302727e-01 6.92112565e-01 4.95685816e-01
2.02925682e-01 -1.13579154e+00 -1.32029071e-01 4.01684850e-01
7.77706087e-01 -1.28241038e+00 2.28655145e-01 -3.53510916e-01
-3.68394047e-01 1.31496477e+00 1.07891284e-01 -2.50088185e-01
8.78851771e-01 4.06663507e-01 5.33241183e-02 3.31556469e-01
-8.55295658e-01 -6.28285483e-02 2.79214919e-01 5.52629769e-01
5.49535096e-01 1.78834692e-01 -8.73586178e-01 5.70023715e-01
-1.97283477e-01 -1.31615639e-01 3.88231725e-01 5.01249790e-01
-1.18820295e-01 -1.24864352e+00 -7.04300582e-01 5.09749591e-01
-7.65764415e-01 -2.35609844e-01 -4.04636949e-01 4.65530217e-01
-6.29101992e-02 9.98736024e-01 -2.01986283e-01 -2.62903571e-01
2.32856959e-01 -8.60641003e-02 6.94870055e-01 -1.51126951e-01
-6.66686952e-01 3.20026517e-01 1.11112542e-01 -6.84290946e-01
-5.54249823e-01 -9.70642686e-01 -1.03122139e+00 -7.06243694e-01
-5.07745206e-01 3.50183427e-01 4.84364808e-01 6.13355696e-01
1.95929602e-01 -2.23132242e-02 7.81592667e-01 -7.34817445e-01
-7.85255194e-01 -5.65306962e-01 -7.51482725e-01 4.73005861e-01
4.82181430e-01 -5.03392577e-01 -6.54863894e-01 -3.92138176e-02] | [7.171154975891113, 4.498510360717773] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.